ArticlePDF Available

Ultramafic Carbonated Melt‐ and Auto‐Metasomatism in Mantle Eclogites: Compositional Effects and Geophysical Consequences


Abstract and Figures

The mineralogy, chemical composition and physical properties of cratonic mantle eclogites with oceanic crustal protoliths can be modified by secondary processes involving interaction with fluids and melts, generated in various slab lithologies upon subduction (auto‐metasomatism) or mantle metasomatism after emplacement into the cratonic lithosphere. Here, we combine new and published data to isolate these signatures and evaluate their effects on the chemical and physical properties of eclogite. Mantle metasomatism involving kimberlite‐like, ultramafic carbonated melts (UM carbonated melts) is ubiquitous though not pervasive, and affected between ~20–40% of the eclogite population at the various localities investigated here, predominantly at ~60–150 km depth, overlapping cratonic mid‐lithospheric seismic discontinuities. Its hallmarks include lower jadeite component in clinopyroxene and grossular component in garnet, an increase in bulk‐rock MgO ± SiO2 and decrease in FeO and Al2O3 contents, and LREE‐enrichment accompanied by higher Sr, Pb, Th, U and in part Zr and Nb, as well as lower Li, Cu ± Zn. This is mediated by addition of a high‐temperature pyroxene from a UM carbonated melt, followed by redistribution of this component into garnet and clinopyroxene. As clinopyroxene‐garnet trace‐element distribution coefficients increase with decreasing garnet grossular component, clinopyroxene is the main carrier of the metasomatic signatures. UM ultramafic melt‐metasomatism at >150 km has destroyed the diamond inventory at some localities. These mineralogical and chemical changes contribute to low densities, with implications for eclogite gravitational stability, but negligible changes in shear‐wave velocities, and, if accompanied by H2O‐enrichment, will enhance electrical conductivities compared to unenriched eclogites.
Effects of mantle metasomatism with LREE-enrichment, ascribed to interaction with a UM carbonated melt similar to kimberlite, on major-element systematics of mantle eclogites. Concentrations of various oxides (wt%) as a function of MgO in reconstructed eclogites and pyroxenites from various localities. Stippled rectangle highlights LREE-enriched samples at generally high MgO contents. Error bar next to panel letters is for a total of 10% modal variation (e.g. 45% clinopyroxene + 55% garnet to 55% clinopyroxene+45% garnet). Dark grey field in b. shows expected range of MgO-CaO for primitive mantle-derived melts that experienced various degrees of fractional crystallisation (Herzberg & Asimow, 2008). Fields for high-Mg and low-Mg arc cumulates (Lee et al., 2006) in panels b. and f., respectively, illustrate that mantle eclogites are unlikely to represent delaminated lower crustal arc material. Square with long stipples in d. and e. highlights samples from Lace inferred to have reacted with a sediment-derived melt in a subduction mélange, which lowered their FeO and MgO and increased their Al2O3 content. The correlation with SiO2 and anti-correlation with FeO at high MgO in c. and d., respectively, is interpreted as a pyroxene-control line, reflecting the addition of FeO-poor and MgO-rich magmatic pyroxene further characterised by low Na2O, moderately higher CaO and SiO2 and higher Cr2O3 than the pre-metasomatic bulk rock. Filled symbols are for LREE-enriched samples (NMORB-normalised Ce/Yb >1; NMORB from Gale et al., 2013) and open symbols for unenriched samples. References as in Figure 1.
Content may be subject to copyright.
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1029/2019GC008774
©2020 American Geophysical Union. All rights reserved.
Aulbach Sonja (Orcid ID: 0000-0001-7612-303X)
Ultramafic carbonated melt- and auto-metasomatism in mantle eclogites:
Compositional effects and geophysical consequences
Sonja Aulbach1,2, Malcolm Massuyeau3, Joshua Garber4, Axel Gerdes1,2, Larry
M. Heaman5, and K. S. Viljoen3
1Institut für Geowissenschaften, Goethe-Universität, Frankfurt am Main, Germany.
2Frankfurt Isotope and Element Research Center (FIERCE), Goethe-Universität
Frankfurt, Frankfurt am Main, Germany
3Department of Geology, University of Johannesburg, 2006 Auckland Park, South Africa.
4Department of Geosciences, Pennsylvania State University, University Park, PA, USA.
5Department of Earth and Atmospheric Sciences, University of Alberta, Edmonton AB,
Corresponding author: Sonja Aulbach (
Key Points:
Exploration of metasomatic effects during subduction of ancient oceanic crust and
after its emplacement into cratonic lithospheric mantle
Metasomatism by kimberlite-like ultramafic melt affected between 20-40% of mantle
eclogite suites worldwide, mostly at 2-5 GPa
Metasomatism lowers FeO, hence density in eclogite; no significant effect on
shearwave velocities.
©2020 American Geophysical Union. All rights reserved.
The mineralogy, chemical composition and physical properties of cratonic mantle
eclogites with oceanic crustal protoliths can be modified by secondary processes involving
interaction with fluids and melts, generated in various slab lithologies upon subduction (auto-
metasomatism) or mantle metasomatism after emplacement into the cratonic lithosphere.
Here, we combine new and published data to isolate these signatures and evaluate their
effects on the chemical and physical properties of eclogite. Mantle metasomatism involving
kimberlite-like, ultramafic carbonated melts (UM carbonated melts) is ubiquitous though not
pervasive, and affected between ~2040% of the eclogite population at the various localities
investigated here, predominantly at ~60150 km depth, overlapping cratonic mid-lithospheric
seismic discontinuities. Its hallmarks include lower jadeite component in clinopyroxene and
grossular component in garnet, an increase in bulk-rock MgO ± SiO2 and decrease in FeO
and Al2O3 contents, and LREE-enrichment accompanied by higher Sr, Pb, Th, U and in part
Zr and Nb, as well as lower Li, Cu ± Zn. This is mediated by addition of a high-temperature
pyroxene from a UM carbonated melt, followed by redistribution of this component into
garnet and clinopyroxene. As clinopyroxene-garnet trace-element distribution coefficients
increase with decreasing garnet grossular component, clinopyroxene is the main carrier of the
metasomatic signatures. UM ultramafic melt-metasomatism at >150 km has destroyed the
diamond inventory at some localities. These mineralogical and chemical changes contribute
to low densities, with implications for eclogite gravitational stability, but negligible changes
in shear-wave velocities, and, if accompanied by H2O-enrichment, will enhance electrical
conductivities compared to unenriched eclogites.
Plain Language Summary
Oceanic crust formed at spreading ridges is recycled in subduction zones and undergoes
metamorphism to eclogite. Some of this material is captured in the overlying lithospheric
mantle, where it is exhumed by passing magmas. Having formed in spreading ridges, these
eclogites have proven invaluable archives for the onset of plate tectonics, for the construction
of cratons during subduction/collision, as probes of the convecting mantle from which their
precursors formed and as generators of heterogeneity upon recycling into Earth’s convecting
mantle. During subduction and until exhumation, interaction with fluids and melts (called
metasomatism) can change the mineralogy, chemical composition and physical properties of
mantle eclogites, complicating their interpretation, but a comprehensive study of these effects
is lacking so far. We investigated mantle eclogites from ancient continents (cratons) around
the globe in order to define hallmarks of metasomatism by subduction-related fluids and
small-volume ultramafic carbonated mantle melts. We find that the latter is pervasive and
occurs predominantly at mid-lithospheric depths where seismic discontinuities are detected,
typically causing diamond destruction and a reduction in density. This has consequences for
their gravitational stability and for the interpretation of shearwave velocities in cratons.
©2020 American Geophysical Union. All rights reserved.
1 Introduction
Mantle eclogites are high-pressure garnet-clinopyroxene rocks that constitute a
portion of xenolith suites entrained in kimberlites world-wide, and are thought to occur as
lenses, pods or layers within cratonic lithospheres (Helmstaedt & Schulze, 1989; Jacob,
2004). The pre-metamorphic origin of most of these suites as subducted oceanic crust is
recognised based on elemental, isotopic, experimental and modelling constraints (Aulbach &
Jacob, 2016). Their protracted history in the cratonic mantle lithosphere involving
Precambrian oceanic crust subduction, metamorphism and interaction with kimberlite magma
prior to entrainment suggests the possibility that their compositions have changed
significantly. Such changes include interactions between slab-derived metasomatic fluids and
melts with portions of the oeanic crust upon subduction (“auto-metasomatism” in the sense
that the agent is internal to the slab) and/or of mantle-derived fluids and melts with the
eclogite reservoir after emplacement in the mantle lithosphere (“mantle metasomatism” in the
sense that the agent is externally-derived). Indeed, various effects of metasomatism in
eclogites have been described, such as addition of hydrous minerals or carbonates,
enrichments in MgO and incompatible elements as well as enriched isotopic compositions
(e.g. Czas et al., 2018; Heaman et al., 2002, 2006; Hills & Haggerty, 1989; Huang et al.,
2012, 2014; Ireland et al., 1994; De Stefano et al., 2009; Jacob et al., 2009; Misra et al., 2004;
Pyle & Haggerty, 1998; Shu et al., 2018; Smart et al., 2009, 2014; Spetsius & Taylor, 2002;
Taylor et al., 1996; Viljoen et al., 1996; Zedgenizov et al., 2018). External introduction or
internal production of a carbonatite- or kimberlite-like ultramafic melt has been implicated in
many of these instances, and such melts have been shown to be temporally and genetically
related to kimberlite magmatism (e.g. Giuliani et al., 2014; Jollands et al., 2018; Yaxley et
al., 2017). The ability to distinguish low-pressure (crustal) and secondary signatures is
necessary for accurate interpretation within a regional tectonomagmatic framework, which
has implications for eclogite origins, craton construction and evolution, and physical
properties of the cratonic mantle (e.g., Shirey & Richardson, 2011).
Despite sporadic investigations of metasomatism in eclogite, a comprehensive effort
to assess the compositional effects on a minor but consequential part of the lithospheric
mantle or its pervasiveness has yet to be made. Recent research, aiming to estimate the
proportion of eclogite in cratonic lithosphere, has highlighted the role of eclogite in
explaining some enigmatic geophysical observations (Garber et al., 2018). However, the
densities, seismic velocities and electrical conductivities of minerals in both metasomatised
and “pristine” eclogites and pyroxenites as a function of depth (and taking into account
regional peculiarities) have not been systematically investigated. Here, we combine published
and new data for kimberlite-hosted mantle eclogite/pyroxenite suites from Orapa (Zimbabwe
craton), Lace (Kaapvaal), and Diavik (Slave) with the aims to (1) identify elemental markers
of UM carbonated melt metasomatism in mantle eclogite/pyroxenite that distinguish this
from other types of metasomatism, (2) obtain further insights into the timing, locus and
pervasiveness of this interaction, and (3) explore physical consequences with respect to
density, seismic velocity and electrical conductivity, which may be used to refine craton-
specific estimates of eclogite/pyroxenite proportions as part of the subcontinental lithosphere.
These are critical to understanding and accurately interpreting the bulk lithospheric mantle
signature detected geophysically, which is commonly used to infer the composition, thermal
state and structure of cratonic lithosphere, with geodynamic implications.
©2020 American Geophysical Union. All rights reserved.
2 Sample classification and database
To understand compositional signatures of auto- and mantle metasomatism, we
compiled new and published data from a global suite of cratonic mantle eclogites (Fig. 1).
Various criteria, based on garnet, clinopyroxene and reconstructed bulk-rock compositions,
are used to classify garnet-clinopyroxene rocks, following Aulbach and Jacob (2016) who
focussed on non-metasomatised samples to capture the nature of their low-pressure protoliths.
To this end, samples are divided into gabbroic (whole-rock Eu/Eu* = chondrite-normalised
Eu/(Sm*Gd)^0.5 >1; Rudnick & Fountain, 1995) and non-gabbroic samples to broadly
separate cumulate protoliths (plagioclase-rich) from protoliths representing complementary
melts, and further into pyroxenites (molar Na/(Na+Ca in clinopyroxene <0.2) and eclogites.
Non-gabbroic eclogites comprise high-Ca (garnet Ca# = molar Ca/(Ca+Fe+Mg+Mn) >0.2),
high-Mg (Ca# ≤0.2 and Mg# = molar Mg/(Mg+Fetotal) >0.6) and low-Mg (Ca# ≤0.2 and Mg#
≤0.6) varieties. Although these cut-off values do not rigorously yield separate groupings, and
the classification of each individual suite must be separately evaluated, they generally reflect
different protoliths and processes in unmetasomatised samples: High-Mg and high-Ca
eclogites largely represent less- and more-differentiated crustal protoliths, respectively,
whereas low-Mg eclogites typically require Fe-rich protoliths (Aulbach & Jacob, 2016). For
the locations for which new data are presented (Diavik, Lace, Orapa), Text S1 provides an
overview of prior work, samples, methods, data filtering and a description of the results,
which are displayed in Tables S1 to S10. Literature data comprise suites from the northern
Slave craton (Jericho, Muskox, Voyageur; Smart et al., 2009, 2014, 2017), Koidu in the West
African craton (Aulbach et al., 2019a; Barth et al., 2001, 2002), and phlogopite-bearing
eclogites/pyroxenites from Kimberley in the Kaapvaal craton (Jacob et al., 2009), for which
trace-element ± isotopic compositions are available. Comparisons are also made to eclogites
from Fort a la Corne in the Sask craton (Czas et al., 2018) and to eclogites from Roberts
Victor in the Kaapvaal craton (Huang et al., 2012, 2014; Radu et al., 2019). These suites have
been selected because they are described as having some proportion of metasomatised
samples; for the sake of clarity not all suites are shown in all figures. In the following, we
will only refer to “eclogites” for brevity, with the implicit understanding that this also
includes pyroxenites unless stated otherwise.
As outlined in Supporting Text S1, metasomatised samples are classified based on
REE patterns into LREE-enriched eclogites and HREE-enriched eclogites, the latter
exemplified by phlogopite-eclogites from Kimberley, which have been enriched in both
LREE and HREE, and distinguished from unenriched eclogites (Fig. S1). Different types of
metasomatism are recognised according to the effects they produce: Patent metasomatism
entails addition of new minerals, whereas cryptic metasomatism is observable only in
incompatible element composition (Dawson, 1984), and stealth metasomatism entails
addition of minerals that are part of the pre-existing assemblage, but with a distinct
composition (O’Reilly & Griffin, 2013). Metasomatism is further distinguished according to
the locus of interaction, as outlined in the introduction: Autometasomatism, which occurs due
to metamorphic reactions in subducting slabs and involves fluids and melts from various slab
components, vs. mantle metasomatism (Roden & Murthy, 1985), which occurs after
emplacement of eclogitised oceanic crust in the subcontinental lithospheric mantle.
3 Geothermobarometry and bulk-rock reconstruction
Temperatures were derived from the garnet-clinopyroxene Mg-Fe exchange
thermometer of Krogh (1988; TKR88), which yields on average lower temperatures than that of
Krogh Ravna (2000) (Figure S4). Barometry on mantle eclogites is hampered by the small
amount of tetrahedrally coordinated Al, which is the pressure-sensitive component in
©2020 American Geophysical Union. All rights reserved.
clinopyroxene (Beyer et al., 2015; PB15) (see Text S1). To circumvent this problem, here,
pressures are estimated for all samples by iterative solution with the regional peridotite-
derived conductive geotherm. For Koidu, Lace and Roberts Victor, reported geotherms
correspond to a surface heatflow of 38 mW/m2 (Griffin et al., 2003; Mather, 2012; Smit et al.,
2016). For Diavik, samples from the shallow portion of the lithosphere (to ~4 GPa) appear to
have equilibrated to a lower geotherm (35 mW/m2) than those from the deeper lithosphere
(Griffin et al., 1999a; Grütter, 2009). At the peridotite-derived geotherm for Orapa (40
mW/m2; Stiefenhofer et al., 1997), the deepest samples were apparently derived from beyond
the geophysically determined depth to the lithosphere-asthenosphere boundary (220 km;
Miensopust et al., 2011), suggesting failure to equilibrate to the geotherm and isobaric
heating. A gradient for a surface heat flow of 36 mW/m2 was determined for samples from
the northern Slave craton (Grütter, 2009). With the exception of Smit et al. (2016), these
works describe geothermal gradients with respect to the older family of geotherms presented
by Pollack and Chapman (1977; PC77), which show too great a curvature at pressures >6
GPa in comparison to pressure-temperature arrays of cratonic mantle xenoliths (Rudnick and
Nyblade, 1999). Therefore, the family of geotherms of Hasterok and Chapman (2011; HC11)
is used for iterative calculations. Since even at pressures <6 GPa the two geotherm families
do not coincide (Fig. S4), the PC77 estimates are approximatelyconverted to HC11, and this
is done by replacing 40, 38 and 36 geotherms according to PC77 with 38, 37 and 35
geotherms according to HC11. Hasterok and Chapman (2011) determined a surface heatflow
of 40.2 mW/m2for the Kalahari craton, without distinguishing between xenoliths entrained in
Jurassic orangeites vs. Cretaceous kimberlites, and of 37 mW/m2 for the Slave craton
combining both northern and central Slave xenoliths. We prefer to keep these separate in
accord with prior work finding distinct pressure-temperature arrays (e.g. Grütter, 2009).
Advected heat, likely connected to kimberlite magmatism, is common especially near
the base of the continental lithosphere (Grütter, 2009). Such samples will have overestimated
pressures if their temperatures are solved with the steady-state geotherm. This includes five
samples from Orapa (OE1, 23,26, 33,34) and eight samples (D1056, D1530, D1503, D1541,
D1534, D1617, D1649, MX121) from the deeper central Slave lithosphere, which
equilibrated to a higher apparent conductive geotherm than those derived from the shallow
lithosphere (Griffin et al., 1999a; Grütter, 2009). These samples will not be considered for
depth estimates (section 5.1.6) or considerations of geophysical properties (section 5.2).
Iterative solutions of TKR88 with PB15, employing only samples with clinopyroxene <1.985 Si
cations per formula unit (pfu) considered to be reasonably accurate (Beyer et al., 2015), show
a good positive correlation with solutions using the regional conductive geotherm, but PB15
may be systematically slightly underestimated, judging by the pressure-temperature arrays of
individual eclogite suites relative to corresponding peridotite-derived conductive geotherms
(Fig. S4). For samples with clinopyroxene Si <1.985 pfu, the average difference between
iterative solutions of TKR88 with PB15 and with the regional conductive geotherm is -0.9 GPa.
Whole rock compositions were calculated for bimineralic eclogites assuming garnet
and clinopyroxene modes of 55 and 45 wt%, respectively (Aulbach & Jacob, 2016). This is
consistent with the average modes determined in a recent study on a large suite of eclogite
xenoliths from Siberia (55.4 ± 5.1 (1) wt% garnet, 44.3 ± 5.0 wt% clinopyroxene; Agashev
et al., 2018), and is also close to those calculated from average mantle eclogite compositions
based on thermodynamic modelling at an intermediate lithospheric pressure of 4 GPa (59 and
41 wt%, respectively; Garber et al., 2018). Higher variance may be observed in other suites in
particular comprising smaller sample sizes, but it is difficult to assess whether this reflects
compositional variability inherited from the protoliths, or is due to a combination of typically
large grain size in mantle eclogite relative to xenolith size and/or mineralogical layering. In
detail, these modes will vary depending on bulk composition (e.g. basaltic vs. picritic
©2020 American Geophysical Union. All rights reserved.
protoliths, melt-undepleted vs. depleted, pristine or metasomatised) and also on pressure.
For example, the global-average cratonic eclogite model of Garber et al. (2018) shows ~3%
modal variation between 6 and 8 GPa, whereas experimental work on subsolidus eclogites
with MORB compositions and on melt-depleted eclogites shows 8% modal variations over
the same pressure interval (Knapp et al., 2015). The average uncertainty in reconstructed bulk
rock compositions related to 10% modal variation is given in Table 1. Additional
uncertainties apply to samples containing minor mineral phases, such as kyanite or
orthopyroxene. The abundance of frequently observed accessory rutile is difficult to constrain
accurately or precisely in typically small (<10 cm) xenoliths, but its omission in bulk-rock
reconstructions leads to an underestimate of TiO2, Nb, and Ta concentrations (Aulbach et al.,
2008). As outlined in Aulbach and Jacob (2016), the TiO2 deficit is compensated by
assuming that the bulk rock has no negative Ti anomaly relative to elements of similar
compatibility during decompression melting of a dry peridotite, as applies to MORB. The
amount of rutile necessary to erase such anomaly is then determined, while Zr and Hf
concentrations in the hypothetical rutile are calculated using published distribution coeffients
for rutile-eclogite. Concentrations will be a maximum estimate if the original rock had a
negative Ti anomaly, as applies to some gabbroic eclogites with cumulate protoliths (Aulbach
& Jacob, 2016), or samples that experienced LREE enrichment unaccompanied by HFSE
addition. Finally, although accessory zircon and apatite occur in some eclogite suites (e.g.
Heaman et al., 2006; Nikitina et al., 2014; Shchukina et al., 2018), they are relatively
uncommon, and are thus not considered, as is coesite, which is generally trace element-poor
and mainly dilutes the other components (Jacob et al., 2003).
4 Trace element homogeneity and distribution
To assess the compositional homogeneity of the minerals, which may have been
affected by entrainment in the host kimberlite and pre-entrainment heating and
metasomatism, trace-element abundances were measured from the outermost rim to ~500 m
into the interior of touching garnet and clinopyroxene grains where possible, depending on
the presence of alteration (data and supporting figures in Table S10). Two unenriched
eclogites (high-Ca and low-Mg class) and two LREE-enriched eclogites (both high-Mg) from
Orapa were chosen. Several spots in clinopyroxene have implausibly high Ba contents (>1
ppm) compared to Ba-poor neighbouring spots and are interpreted to have been affected by
kimberlite contamination in the sampled volume (Aulbach & Viljoen, 2015). The trace-
element profiles show that elements that are more compatible in clinopyroxene, such as Sr
and Ce, are relatively homogeneous in clinopyroxene, whereas they increase from garnet
cores to rims in the two high-Mg eclogites (Table S10). Importantly, the difference in
concentration between garnet and clinopyroxene in the LREE-enriched, metasomatised
eclogites is systematically much larger than in the unenriched samples. This suggests that the
trace-element enrichment recorded in clinopyroxene preceeded final kimberlite entrainment
and related heating, which caused some diffusion of Sr, Ce and other similarly incompatible
elements into garnet, either from the clinopyroxene or from the host kimberlite or precursory
melt percolation. In contrast, Y, which is more compatible in garnet than in clinopyroxene,
shows no consistent zoning in either phase. As a fast-diffusing element (e.g., Dohmen et al.,
2010), Li concentrations vary little in all of the grains, but are conspicuously higher in
clinopyroxene from the unenriched eclogites than in those from the LREE-enriched eclogites.
This broad homogeneity is suggested to reflect chemical equilibrium, and the consistent
difference between LREE-enriched vs. unenriched eclogites is again interpreted to reflect a
pre-entrainment event. In summary, given the small degree of variation away from the
outermost rim (see typically small relative standard deviations for trace element abundances
obtained from multiple spots and grains per sample in Table S5), the samples are considered
©2020 American Geophysical Union. All rights reserved.
to be in trace-element equilibrium, allowing distribution coefficients (clinopyroxene/garnetDelement)
to be calculated and interpreted with respect to their dependence on temperature (and
pressure along the conductive geotherm) and composition.
5 Discussion
5.1 Compositional and mineralogical effects
5.1.1 Effects of mantle and auto- metasomatism on trace-element distribution
The compositional dependence of clinopyroxene-garnet trace-element distribution in
eclogite has been previously highlighted (O’Reilly & Griffin, 1995; Harte & Kirkley, 1997;
Aulbach et al., 2016). In the eclogite suites under consideration here, the jadeite and grossular
components are lowest (jadeite <0.2; Ca#<0.15) in LREE-enriched samples (Fig. 2a).
Focusing on the dependence of trace-element distribution on the grossular component to
illustrate the effect of LREE-enrichment, it is evident that with decreasing garnet Ca contents,
elements that are compatible (Y), mildly incompatible (Sr) and strongly incompatible (Ce) in
garnet are increasingly partitioned into clinopyroxene (Fig. 2b-d). The distribution takes on a
natural logarithmic form (Harte & Kirkley, 1997), such that very high values are attained for
low garnet Ca#. Since UM carbonated melt metasomatism decreases garnet Ca# and the
partitioning of incompatible trace elements into garnet is limited at low Ca#, this leads to the
apparently paradoxical observation that garnet in LREE-enriched samples shows few if any
trace-element differences compared to unenriched samples (Fig. 2e-f). This is also evident
from Fig. S2 where REE patterns of garnet in LREE-enriched and unenriched samples from
Orapa are similar but those of clinopyroxene are distinct. This illustrates that clinopyroxene is
the main carrier of the metasomatic signature in LREE-enriched eclogites.
5.1.2 Patent and stealth effects of auto- and mantle metasomatism
As outlined above, metasomatism can lead to patent, cryptic and/or stealth effects
(O’Reilly & Griffin, 2013). The most commonly described examples of patent mantle
metasomatism in mantle eclogite or pyroxenite are the addition of amphibole, phlogopite,
apatite, zircon and/or carbonate and diamond as well as the appearance of melt pools,
unequilibrated microstructures and an increase in fluid inclusions (Heaman et al., 2002, 2006;
Hills & Haggerty, 1989; Huang et al., 2012, 2014; Jacob et al., 2009; Misra et al., 2004; Pyle
& Haggerty, 1998; Smart et al., 2009; Spetsius & Taylor, 2002). Rutile and sulphide may be
metasomatic (e.g., Hills & Haggerty, 1989), but can be primary minerals, as rutile saturates
during metamorphism of oceanic crust (e.g. Gaetani et al., 2008), and sulphide may be
inherited from the sulphide-saturated oceanic crust (e.g., Patten et al., 2013). In contrast,
zircon formation in mantle eclogite xenoliths from the Grib kimberlite in the East European
platform was an auto-metasomatic process, linked to fluids emanating from a
Palaeoproterozoic subduction zone (Shchukina et al., 2018). This may also apply to zircon
and Nb-rich rutile in a group of mantle eclogites from Jericho recording crystallisation ages
close to the time of ca. 1.8 Ga metamorphism (Heaman et al., 2006). Stealth metasomatism
has been recognised (though not referred to as such) in eclogites from Diavik and Ekati in the
central Slave craton, expressed as the addition or overgrowth of garnet from
deserpentinisation fluids (Aulbach et al., 2011). At Lace in the Kaapvaal craton it is
expressed as the addition of aluminous clinopyroxene precipitated from a sediment-derived
melt in a subduction mélange. This clinopyroxene later exsolved garnet and kyanite (Aulbach
et al., 2016). At Koidu in the West African craton, it is expressed as addition of diopside-rich
©2020 American Geophysical Union. All rights reserved.
clinopyroxene from a UM carbonated melt, which converted low-Mg and gabbroic eclogites
to high-Mg eclogites and pyroxenites (Aulbach et al., 2019a).
A metasomatic agent containing a sufficient amount of solutes to effect patent or
stealth mineral growth should also cause significant changes in whole-rock major-element
and mineralogical compositions. The major-element relationships of the various eclogite
suites examined here indicate that many of the LREE-enriched samples are MgO- and Cr2O3-
rich (Fig. 3a), with high CaO compared to near-primary MgO-rich mantle melt products (Fig.
3b). Those with the highest MgO contents also tend to show the highest SiO2 contents (Fig.
3c) and low FeO, Al2O3 and Na2O (Fig. 3d-f). These changes are explainable by addition of a
high-temperature pyroxene from an external UM carbonated melt, such as those reported in
Mallik and Dasgupta (2013) and shown in Figure 15 in Aulbach et al. (2019a). In particular,
we interpret the anti-correlation of MgO with FeO (Fig. 3d) and correlation with SiO2 at high
MgO contents as a pyroxene control line (Fig. 3c). If so, the modes assumed for LREE-
enriched samples may be better represented by higher clinopyroxene modes of, for example,
50% than the 45% used to reconstruct bulk rocks. This is because not all added components
can be equally accommodated by garnet (in particular SiO2 or Na2O) and need to be added as
more cpx. We account for this by considering, hypothetically, the effect of addition of 5%
eclogitic clinopyroxene. The uncertainty resulting from a 10% modal variation (i.e. 45%
clinopyroxene-55% garnet and 55% clinopyroxene-45% garnet) is displayed in Figure 3 and,
with the exception of SiO2, is small relative to the variation displayed. An increase in
clinopyroxene mode by 5% would shift the bulk rock up by half the length of the error bar
with respect to SiO2, Na2O and CaO and down with respect to FeO and Al2O3 at nearly
constant MgO and Cr2O3. More drastic changes in garnet-clinopyroxene modes appear
unlikely because high-MgO eclogites and low-MgO eclogites have indistinguishable modal
proportions, e.g. in the Koidu suite (Hills and Haggerty, 1989). Even if a high-temperature
pyroxene was added, re-equilibration with the pre-existing assemblage is indicated by
concomitant changes in both garnet and clinopyroxene, as, for example, both have higher
MgO and lower FeO contents. Auto-metasomatised eclogites from Lace show CaO-Al2O3
enrichment and FeO-MgO dilution (Fig. 3d-e).
5.1.3 Auto- and mantle metasomatism and diamond
Diamond in eclogite is regarded as a metasomatic mineral because the direct
conversion of intrinsic carbon in oceanic crust is kinetically inhibited, and influx of fluids or
melts is required for diamond growth (Stachel & Luth, 2015). As diamond formation
involves the migration of melts and fluids containing various carbon species and abundances
as a function of pressure, temperature and oxygen fugacity (Stachel & Luth, 2015), there are
varied relationships between metasomatism and the formation and stability of diamond. The
close temporal relationship between eclogitic diamond formation or growth and accretionary
or collisional processes during craton amalgamation or at craton margins (Shirey &
Richardson, 2011; Timmermann et al., 2017) suggests an auto-metasomatic origin of
diamond accompanying recycling of oceanic crust, as recognised in the central Slave craton
and at Lace (Aulbach et al., 2011, 2016). This constitutes an efficient process to achieve high
diamond modes, aided by the high volume of carbonaceous fluids and melts present in this
tectonic setting (e.g., Poli, 2015; Tumiati et al., 2017). Moreover, the low oxygen fugacity of
eclogitised oceanic crust (Smart et al., 2017; Aulbach et al., 2017b, 2019c) may cause CO2 in
the autometasomatic fluid to be reduced to diamond.
An autometasomatic origin for some eclogitic diamond suites does not preclude that
significant new diamond growth or overgrowth on pre-existing diamond can occur during
later metasomatic events (e.g. Taylor et al., 1996, 1998). Although eclogitic diamond can
form from reduced or mixed carbon species (Smit et al., 2019), oxidised carbon is inferred to
©2020 American Geophysical Union. All rights reserved.
be involved in diamond formation during interaction with oxidising UM carbonated melts,
such as kimberlite (Fedortchouk & Canil, 2004). It has also been shown that melts with dilute
CO2 content, such as kimberlite, can coexist with diamond to low fO2 where fluids would be
water-maximum (Stagno et al., 2013; Stamm and Schmidt, 2017; Luth and Stachel, 2014).
Thus, diamond in eclogite from Siberia has been linked to mantle metasomatism involving
UM carbonated melts or high-density fluids. This model is supported by the occurrence of
diamonds in secondary alteration zones in eclogite (Schulze et al., 1996; Shatsky et al., 2008;
Taylor et al., 2000; Zedgenizov et al., 2018). Carbonatitic to silicic fluids have been
identified in cloudy or fibrous diamond with eclogite inclusions (Izraeli et al., 2001; Weiss et
al., 2009, 2015). In the Slave craton, interaction with a UM carbonated melt produced
extremely diamond-rich high-Mg eclogites or pyroxenites (Smart et al., 2009), and it has
been suggested based on experimental observations that both monocrystalline and fibrous
lithospheric diamonds form by redox reactions from carbonated fluids or melts (Bureau et al.,
2018). Conversely, in mantle eclogites from Orapa and Koidu, the diamond-destructive
nature of UM carbonated melt metasomatism is indicated by the association of diamond with
high-Ca and low-Mg eclogites, but absence in LREE-enriched high-Mg eclogites and
pyroxenites (Aulbach et al., 2017a; 2019a). Indeed, it has been shown experimentally that
interaction with oxidising small-volume carbonated silicate melts, such as kimberlite,
eventually causes diamond resorption and associated carbon remobilisation at mantle depth
(Fedortchouk et al., 2019). These seemingly contrasting observations may relate to the local
buffering capacity and fO2 in the system. Given the aforementioned low fO2 of mantle
eclogites, the mechanism causing the observed diamond resorption by kimberlite-like melts
remains unclear.
The association of diamond with high Na2O in eclogitic garnet, whereas eclogites
with low-Na2O garnet are barren (McCandless & Gurney, 1989), has been explained by the
pressure-dependent incorporation of Na2O in garnet, as exemplified by the eclogite suite from
Kaalvallei, Kaapvaal craton (Viljoen et al., 2005). Since UM carbonated melt metasomatism
lowers the jadeite component in clinopyroxene and therefore bulk-rock Na2O (Table 1), the
diamond-destructive effect of metasomatism, especially near the top of the diamondiferous
lithosphere, may also play a role. On the simplistic assumption that all LREE-enriched
eclogites at Orapa and Koidu are barren, whereas all unenriched eclogites are
diamondiferous, ~20 to 40% of the eclogitic diamond inventory was destroyed due to UM
carbonated melt metasomatism. This also has physical ramifications as the absence or
presence of just a few vol% diamond can significantly affect the bulk shear velocity of
eclogite (Garber et al., 2018), as will be explored further below.
5.1.4 Cryptic effects of auto- and mantle metasomatism
Two types of auto-metasomatism in subduction settings have been described. Stepped
REE patterns in mantle eclogites from the central Slave craton are very similar to those
reported for some Phanerozoic orogenic eclogites and are inconsistent with melt depletion;
they may result from garnet overgrowths due to interaction with deserpentinisation fluids,
where LREE are strongly excluded and removed with the fluids (Aulbach et al., 2011). At
Lace, kyanite- and/or diamond-bearing eclogites formed through interaction of a pelite-
derived melt with subducted oceanic crust in a mélange setting show enrichments in LREE
and Th, and depletions in Y, Zr, Hf and HREE (Aulbach et al., 2016). Strong Pb addition
during this event, resulting in low U/Pb for some eclogites, has allowed the retention of
highly unradiogenic clinopyroxene Pb isotopic compositions (206Pb/204Pb <14), yielding ca.
3.0 Ga single-stage model Pb-Pb ages (Aulbach et al., 2019b). Because of the close temporal
relationship of zircon crystallisation in eclogite xenoliths from the northern Slave craton and
Eastern European Platform with accretionary processes (Heaman et al., 2006; Shchukina et
©2020 American Geophysical Union. All rights reserved.
al., 2018), the HFSE enrichment accompanying zircon formation is also suggested to reflect
an auto-metasomatic process unrelated to mantle metasomatism sensu stricto.
Cryptic metasomatism involving UM carbonated melts has been widely recognised,
based, inter alia, on LREE- ±LILE and HFSE-enrichment (e.g. Czas et al., 2018; Heaman et
al., 2006; Huang et al., 2012; Jacob et al., 2009; Smart et al., 2009). Volumes sampled during
laser ablation can comprise extraneous kimberlite material in optically invisible cracks or
otherwise altered material; however, as outlined in Text S1, rigorous reduction of the data
with respect to such kimberlite contamination ensures that the enrichments discussed here
have a pre-entrainment origin. The coupled enrichment of LREE and MgO in the
reconstructed whole rocks (Fig. 4a) clearly supports the addition of both from a small-volume
(hence strongly incompatible element- and volatile-enriched) UM carbonated melt, probably
from the kimberlite-carbonatite spectrum, which will be referred to here as “kimberlite-like”
for simplicity. Focusing on clinopyroxene as the main carrier of the metasomatic signature, as
discussed above, Pb, Th and U were enriched along with LREE in all suites (Fig. 4b-c), with
Lace showing the lowest degree of incompatible element enrichment. Again with the
exception of Lace, Nb concentrations are on average higher in LREE-enriched eclogites (Fig.
4d), whereas Zr is only enriched in some samples from Diavik and most samples from Orapa
(not shown). Similarly, TiO2 and Y are not consistently co-enriched with LREE, but
concentrations of both are elevated in HREE-enriched eclogites from Kimberley, whereas
clinopyroxene in LREE-enriched Orapa eclogites has higher Y but not TiO2 contents (Fig.
4e). The absence of TiO2 and Y enrichments results from their low concentrations in
kimberlite-like melts, even if elevated experimental mineral-melt partition coefficients for
clinopyroxene are assumed (Table 1). In detail, such coefficients can vary significantly, and
enrichment or dilution depends on trace element contents, e.g. inherited during the low-
pressure history of the sample (prior to metasomatism) and degree of enrichment of the
kimberlite-like melt, which may grade into basanite or carbonatite. In addition, selective TiO2
enrichment, for example via ilmenite addition, would not necessarily be detected in small
samples and would not show up in bulk rocks reconstructed with only garnet and
clinopyroxene. Compared to only LREE-LILE enrichment ascribed to small-volume UM
carbonated melt metasomatism, coupled TiO2-Y-Zr-LREE-HREE enrichment is suggested to
indicate the involvement of a higher melt fraction with more dissolved SiO2 a silicate melt
analogous to that causing “melt metasomatism” recognised in peridotites (Griffin et al.,
1999b). Although the enrichment patterns vary in detail, it appears that metasomatic intensity
increases from Lace to Diavik, Jericho and Orapa, with the strongest enrichment in high-
temperature HREE-enriched phlogopite-bearing eclogites from Kimberley. This is also
mirrored in REE patterns of clinopyroxene in LREE-enriched eclogites, which converge with
those in unenriched eclogites at Nd or Sm for Diavik, but at Ho or Er for Orapa. Since
kimberlites have considerably higher contents of Sr, Nb, LREE, Ta, Pb, Th and U, and to a
lesser degree Zr and Hf than unenriched eclogites (Table 1), they are considered likely agents
for the cryptic enrichment observed in LREE-enriched eclogites.
A puzzling characteristic of LREE-enriched eclogites is their low Li, Cu and in part
Zn abundances, as reflected in clinopyroxene (Fig. 4f; Table 1) and reconstructed whole
rocks, lower than expected in fresh or seawater-altered MORB (Jenner & O’Neill, 2012;
Staudigel, 2005). These elements are mildly incompatible during differentiation of the
inferred oceanic crustal protoliths and would therefore be expected to have low abundances
in less-differentiated high-MgO and gabbroic (cumulate) protoliths, in particular at higher
melt fractions facilitated by higher Archaean TP. Nevertheless, the association of low Li, Cu
and Zn not only with high MgO but also with LREE-enrichment suggests a link to
metasomatism. This implies that they were either leached out of eclogites due to low mineral-
melt distribution coefficients, and/or diluted by stealth magmatism via addition of Li-, Cu-
©2020 American Geophysical Union. All rights reserved.
and Zn-poor clinopyroxene. Lithium concentrations in kimberlite/orangeite are seldom
reported; taking that from Lace with ~30 ppm (Howarth et al., 2011) as representative, ~1 to
8 ppm would be expected in clinopyroxene for D = 0.04 and 0.27 respectively, as reported for
equilibrium with carbonatite (Blundy & Dalton, 2000) or basanite melts (Adam & Green,
2006) (Fig. 4h). Zinc and Cu may be mildly incompatible or compatible (clinopyroxene/basaniteDZn
of 0.25 to 0.69 and DCu 0.47 to 1.5, respectively; Adam & Green, 2006). Median Zn and Cu
concentrations in kimberlites/orangeites from Lace, Jericho, Lac de Gras and southern Africa
vary from 38 to 75 and 50 to 77 ppm, respectively (Howarth et al., 2011; Le Roex et al.,
2003; Tappe et al., 2013), compared to ~40 to 100 ppm and 2.2 to 8.5 ppm (median) in
unenriched eclogite from various suites (Table 1). Thus, precipitation and equilibration of
stealth clinopyroxene (and garnet) with kimberlite may explain the low observed Li, and
perhaps Zn, concentrations in clinopyroxene from eclogites inferred to have interacted with a
small-volume UM carbonated melt. The low Cu concentrations may imply sulphide
precipitation from melt or fluid in the system, consistent with the near-ubiquity of this
mineral in eclogite xenoliths (personal observation) and evidence for partial metasomatic
origin (Aulbach et al., 2009, 2019a).
Type I eclogites from Roberts Victor have microstructures interpreted as reflecting
disequilibrium (rounded garnet grains in a “matrix” of clinopyroxene) whereas Type II
eclogites have equilibrium microstructures with interlocking garnet and clinopyroxene grains,
lower Na2O in garnet and K2O in clinopyroxene (MacGregor & Carter, 1970; McCandless &
Gurney, 1989). Type I eclogites have been suggested to be the product of carbonated melt
metasomatism of Group II eclogites (Huang et al., 2012, 2014), but they do not bear the
hallmarks of such metasomatism identified above: Few are LREE-enriched relative to the
HREE, their HREE contents are lower than those of HREE-enriched eclogites from
Kimberley and eclogitic clinopyroxenes are not consistently jadeite-poor nor do they have
high Mg#. Both groups display strong relationships between Eu/Eu* and total HREE contents
indicative of low-pressure accumulation and fractional crystallisation as exemplified by
MORB (e.g. Gale et al., 2013; Jenner and O’Neill, 2012) and gabbros (e.g. Hart et al., 1999),
albeit at a different slope perhaps reflecting lower oxygen fugacities (Aulbach & Viljoen,
2015) (Fig. S3). Recent analysis of a large suite of Group II eclogites (Radu et al., 2019)
shows that their whole-rock REE patterns, clinopyroxene K2O contents and garnet Na2O
contents overlap those of Group I eclogites investigated by Huang et al. (2012, 2014). We
suggest that the compositional signatures (LREE-depletion, HREE-enrichment, low Sr, Na2O
and K2O) of Type II eclogites are mostly explained by partial melt loss and that many Type I
eclogites are, in fact, relatively pristine, although we concur with Radu et al. (2019) that the
relationship between the different eclogite types at Roberts Victor is not straightforward.
5.1.5 Muting of crustal signatures and isotopic effects of UM carbonated melt
LREE-enriched eclogites are dominated by high-Mg eclogites and pyroxenites, and
the difference between metasomatised and unmetasomatised pyroxenite varieties tends to be
small (Table 1). This suggests that part of the high-Mg eclogites and most pyroxenites are the
product of metasomatism. This may be accompanied by a muting or complete erasure of
typical crustal signatures. The effects of mantle metasomatism on stable or radiogenic isotope
compositions have been previously discussed in the literature, with varying conclusions, and
are reviewed here together with the effects on Eu/Eu* as a typical crustal signature in
particular in eclogites with gabbroic protoliths. If the inference of stealth metasomatism by
addition of clinopyroxene from an ultramafic kimberlite-like melt is correct, the added
clinopyroxene had no or small negative Eu anomalies. Taking Ce/Yb as a measure for the
intensity of UM carbonated melt metasomatism and clinopyroxene as the main carrier of the
©2020 American Geophysical Union. All rights reserved.
metasomatic signature, it is evident that the range of Eu/Eu* is much larger at low Ce/Yb
than at high Ce/Yb (Fig. 5a). When plotted against MgO content in reconstructed bulk rocks,
NMORB-normalised Eu/Eu >1.05 is retained in some gabbroic eclogites with moderate MgO
contents, but in most samples with MgO >15 wt%, reflecting the greatest stealth and cryptic
metasomatic overprint, Eu/Eu* is <1.05 (Fig. 5b). These systematics indicate erasure of the
gabbroic signature in eclogites strongly affected by mantle metasomatism, possibly aided by
the experimentally demonstrated higher Eu diffusivity relative to other REE under reducing
conditions (Szumila et al., 2019), such as those recorded by mantle eclogite (Aulbach et al.,
2017b, 2019c; Smart et al., 2017; Stagno et al., 2015).
Oxygen isotope compositions lower and higher than the canonical mantle range (5.1-
5.9; Mattey et al., 1994) have been interpreted as reflecting high- and low-temperature
seawater alteration, respectively (Muehlenbachs and Clayton, 1972a,b), although mantle-like
18O is not evidence against a low-pressure origin (e.g. Gregory and Taylor, 1981; Smart et
al., 2012). Though Huang et al. (2014) see evidence for an increase in 18O in response to
mantle metasomatism in Type I eclogites at Roberts Victor, Riches et al. (2016) concluded
that there is no detectible effect of metasomatism on 18O and Korolev et al. (2018) consider
the effect to be small (generally <1.2 ‰). As discussed by Radu et al. (2019), mass balance
requires high fluid/melt-rock ratios to change the isotopic composition of the major element
O. Since 18O in garnet from unenriched eclogites from the northern Slave, Lace and Koidu
kimberlites are largely within the canonical mantle range, they cannot be used to assess the
effect of mantle metasomatism. There is, however, some evidence that this has occurred in
the sample suite from Orapa, where mantle-like 18O is associated with radiogenic Sr
ascribed to overprint by a small-volume melt derived from an aged lithospheric mantle
metasome (Aulbach et al., 2017a). Figure 3a shows that UM carbonated melt metasomatism
is accompanied by an increase in Cr2O3, suggested to reflect stealth addition of Cr2O3-rich
pyroxene. This component has been taken up by both eclogitic clinopyroxene and garnet.
Thus, taking Cr2O3 in garnet as a proxy for the degree of metasomatic alteration, it is evident
that many unenriched eclogites have mantle-like 18O and some LREE-enriched mantle
eclogites retain 18O outside canonical mantle, but that at high metasomatic intensity (high
Cr2O3) values are exclusively mantle-like (Fig. 5c). The association of low 18O and high
Eu/Eu* in some metasomatised eclogites from the Sask craton (reported in Czas et al., 2018)
appears consistent with a deep crustal gabbroic protolith that experienced high-temperature
seawater alteration (not shown), whereas metasomatised samples with high MgO content
have heavier, mantle-like 18O, consistent with muting of the crustal signature during strong
metasomatism by a UM carbonated melt (Fig. 5d). Czas et al. (2018) identify chemical
gradients, including in 18O, from primary to metasomatised zones in eclogitic garnet in this
sample suite, showing conclusively that O isotopes can be modified at high melt/rock ratios,
but only towards and not away from canonical mantle values.
Effects of mantle metasomatism on radiogenic isotopes have been recognised and
must be considered on a suite-by-suite basis because of local variations in the isotopic
composition of the metasomatic agent, the timing of metasomatism with consequent isotopic
ingrowth and the low-pressure history of the samples. Thus, positive trends of Y and negative
trends of Eu/Eu* with 87Sr/86Sr may result from accumulation (low time-integrated Rb/Sr,
low Y, high Eu/Eu*) and differentiation (high time-integrated Rb/Sr, high Y, low Eu/Eu*), as
discussed for Orapa (Aulbach et al., 2017a) (Fig. 6a). Interaction with a young kimberlite-like
(sensu stricto) agent with OIB-like isotopic characteristics (Becker & Le Roex, 2006; Tappe
et al., 2017) would increase low 87Sr/86Sr of eclogites with ancient cumulate protoliths, but
decrease that of those with ancient evolved protoliths. Proto-kimberlite and orangeite
metasomatism mobilises aged and isotopically evolved lithospheric metasomes with enriched
©2020 American Geophysical Union. All rights reserved.
isotope characteristics (unradiogenic Nd, radiogenic Sr; Becker & Le Roex, 2006; Giuliani et
al., 2015; Tappe et al., 2008) and would therefore likely decrease 143Nd/144Nd and increase
87Sr/86Sr in affected eclogites. At Lace, although LREE-enriched samples nearly span the
local range of initial 143Nd/144Nd, a broad trend towards less radiogenic Nd with increasing Cr
abundances is recognised, the lowest initial value, in pyroxenites, being below that of the host
orangeite (Fig. 6b). This may indicate that pyroxenitisation preceded Cretaceous orangeite
magmatism and involved an agent sampling an aged lithospheric mantle metasome, as
suggested also for Orapa. At Koidu, metasomatism by a kimberlite-like melt is associated
with an increase of 87Sr/86Sr in clinopyroxene and convergence to a common value of
~0.7035, clearly linked to low Li abundances, while initial 143Nd/144Nd in garnet decreases
from radiogenic values (>0.5145) to a relatively constant value of ~0.5130 (Fig. 6c),
associated with high Cr contents. At Orapa, strongly radiogenic Sr, indicative of a
metasomatic agent derived from an isotopically aged lithospheric mantle reservoir with
87Sr/86Sr >0.7060 and similar to orangeites (Becker & Le Roex, 2006), is mostly associated
with LREE-enriched eclogites (Fig. 6a). In contrast, in eclogites from the northern Slave
craton, the metasomatic agent had less radiogenic Sr, whereas in the central Slave craton a
link between LREE-enrichment and 87Sr/86Sr is not evident. Conversely, the preservation of
unradiogenic Sr in the eclogite suite from the Victor kimberlite, Superior craton, was
suggested to indicate negligible effects of such metasomatism (Smit et al., 2014).
5.1.6 Depth, extent and timing of UM carbonated melt metasomatism
The nature and intensity of metasomatism varies with depth and may be localised due
to differences in the density of fluids vs. melts allowing the former to rise to shallower depths
(Griffin et al., 2003). Further effects are related to the solidus of the metasomatised lithology:
harzburgites with high solidi are more likely to be percolated by fluids whereas lherzolites
and eclogites with lower solidi may be percolated more readily by melts (Stachel & Luth,
2015). Keeping the uncertainties outlined in chapter 3 in mind, the LREE-enrichment in
eclogites from Orapa, Diavik, northern Slave and Kimberley is focused at mid-lithospheric
depths (Fig. 7a-b). This is similar to the depths where continental mid-lithospheric
discontinuities are recorded, which have been ascribed to the accumulation of metasomatic
minerals, such as phlogopite and carbonate, which are directly (xenoliths) or indirectly
(mantle melts) observed in cratons (Aulbach et al., 2017c; Rader et al., 2015). The inferred
silicate melt metasomatism causing HREE-enrichment in Kimberley eclogites is largely
restricted to greater depths of ~120 to 150 km (Fig. 7c), closer to where intense melt
percolation and refertilisation of the peridotitic lithosphere are observed (e.g. O’Reilly &
Griffin, 2013), and is rare to absent in the other eclogite suites considered here. In the Sask
craton, LREE-enrichment in eclogites, unaccompanied by HREE addition, occurs at similar
depths (Fig. 7d). Nevertheless, in common with HREE-enriched Kimberley eclogites, Sask
samples are conspicuously Nb-enriched, whereas in the northern Slave craton, this
enrichment is restricted to a very narrow pressure interval (Fig. 7d).
Bearing in mind that the number of samples is probably too low for statistical
significance, the proportion of pristine vs. metasomatised eclogite in each of the suites (as
gauged by bulk-rock Ce/YbN >1) may be used to assess the extent of UM carbonated melt
metasomatism. This indicates that between ~20% (Lace, Diavik) and ~40 % (northern Slave,
Orapa) of the eclogite reservoir was affected by UM carbonated melt metasomatism (Table
1). The observation that ascending kimberlites do not sample the mantle lithosphere in a
representative manner (Moss et al., 2018) adds further uncertainty to this estimate.
As is true for their peridotitic counterparts (e.g., Griffin et al., 1999b; Jollands et al.,
2017), LREE enrichment in some eclogites appears to be temporally related to kimberlite-like
magmatism (though not directly caused by the host kimberlite). This is evidenced by
©2020 American Geophysical Union. All rights reserved.
diamond brecciation and elemental heterogeneity in deeply-derived eclogites from the Sask
craton, which occurred immediately prior to entrainment, as modelled from diffusion profiles
(Czas et al., 2018). In the absence of such heterogeneity and given that zircon formation is
related to auto-metasomatism during emplacement of oceanic crust (as discussed above), the
timing of mantle metasomatic effects in LREE-enriched eclogites is difficult to pin down
precisely due to their multi-stage evolution. Some metasomatism is plausibly related to
reactivation of older metasomes related to lamproite magmatism (Mitchell, 2006) and failed
protokimberlite magmatism (e.g., Giuliani et al., 2014). This is the case for Koidu, where Sr-
Nd isotopes in eclogites indicate a Neoproterozoic age for metasomatism, possibly due to
extension during Rodinia break-up (Aulbach et al., 2019a). Southern Africa presents a special
case because older lamproite/orangeite magmatism (>110 Ma) was superseded by kimberlite
magmatism sensu stricto (mostly ≤95 Ma), accompanied by strong metasomatism, warming
and loss of the deepest part of the lithospheric root (Griffin et al., 2003; Kobussen et al.,
2009). As outlined above, eclogites from the older Lace kimberlite are only mildly
overprinted (20% affected), whereas higher proportions have been affected at Orapa and
Kimberley, plausibly between 110 and 95 Ma. Circumstantial evidence for a young event at
Lace (relative to host orangeite emplacement) comes from the inverse relationship of initial
143Nd/144Nd with Sm/Nd in metasomatic pyroxenites (not shown), which cannot persist for
extended periods of time, as radiogenic ingrowth would cause a counter-clockwise rotation of
the relationship. At Orapa, several samples show significant Sr isotopic heterogeneity (Table
S7), e.g. from 0.7010 to 0.7043 in diamondiferous eclogite 801; low measured Rb/Sr
excludes ablation of kimberlite-contaminated areas as the origin of this variability. This
variability also qualitatively suggests relatively recent metasomatic overprint, with
insufficient time and/or heat to allow rehomogenisation. Of note, a phlogopite-rich eclogite
(30 vol%) occurs in the ca. 1150 Ma Premier (Cullinan) kimberlite (Dludla et al., 2006). The
age of this kimberlite is similar to the age determined by Ar-Ar dating on metasomatically
introduced phlogopite in peridotite xenoliths from various Kaapvaal kimberlites (Hopp et al.,
2008). This provides bona fide evidence that significant hydration in the Kaapvaal craton is
not a phenomenon restricted to relatively recent deep protokimberlite activity, and
circumstantial evidence that phlogopite addition precursory to kimberlite magmatism is a
recurrent process since at least the Mesoproterozoic.
5.2 Effects of mantle metasomatism on eclogite physical properties
In order to assess the effects of mantle metasomatism on density and shear-wave
velocity of mantle eclogites, we used the same methods as in Garber et al. (2018) [cf.
Connolly, 2009; Stixrude & Lithgow-Bertelloni, 2005; 2011], with the difference that we (1)
used measured rather than calculated mineral compositions, (2) did not consider accessory
phases, such as coesite or kyanite (which due to their low abundance are not necessarily
exposed at thin section scale), (3) used pressure-temperature estimates obtained by iterative
solution of temperatures derived from the garnet-clinopyroxene Mg-Fe exchange
thermometer of Krogh (1988) with regional peridotite-derived geotherms, as parameterised
from Hasterok and Chapman (2011) and (4) applied mineral modes of 45% clinopyroxene
and 55% garnet, compared to ~4045% and ~5560%, respectively, employed in Garber et
al. (2018), the effect of which is explored in Fig. S5. Briefly, the molar volume, density, and
shear moduli of each relevant clinopyroxene and garnet compositional endmember were
calculated over a large pressure-temperature grid in Perple_X (using the elastic parameters
and equations of state in Stixrude and Lithgow-Bertelloni (2005; 2011)), followed by
extraction of the relevant endmember data from the grid at the pressure and temperature of
each xenolith using MATLAB. Clinopyroxene and garnet endmember proportions were
calculated directly from EPMA major-element oxide data, with clinopyroxene site
©2020 American Geophysical Union. All rights reserved.
assignation and endmember inversion performed using the methods of Morimoto et al. (1988)
and Dietrich and Petrakakis (1986), respectively. Fe3+ and Cr contents are relatively minor
across the sample suite, thus these endmembers were not considered, nor would they
significantly affect the results if they were included (cf. discussion in Garber et al., 2018).
Molar volumes and densities were calculated for clinopyroxene and garnet using Voigt
averages of their compositional endmember properties, whereas shear moduli (Gs) were
calculated for the same phases with molar-volume weighted Reuss averages of endmember
shear moduli (cf. Stixrude and Lithgow-Bertelloni, 2005, eqs. 911). Finally, bulk-rock
densities were calculated using Voigt averages of the solid-solution densities; these densities
were then used in combination with Voigt-Reuss-Hill averaged bulk-rock shear moduli to
calculate shear-wave velocities. Seismic attenuation was not considered, and thus calculated
shear-wave velocities are maxima. Electrical conductivity was calculated using Hashin
Shtrikman lower (HS−) and upper (HS+) bounds based on formulations in Jones et al. (2009)
and using the pressure-temperature-conductivity relationship for dry garnet in Dai and Karato
(2009) and the temperature-H2O-conductivity relationship for wet clinopyroxene in Zhao and
Yoshino (2016).
5.2.1 Density
One of the most conspicuous effects of UM carbonated melt metasomatism is FeO
dilution (Fig. 3d), with a corresponding decrease in the almandine and hedenbergite
endmembers. Hence, the density of metasomatized eclogites is systematically lower than
unmetasomatised eclogites from the same locality (Fig. 8a-b; Table 1). The calculated
densities for LREE-enriched eclogites are maxima if bulk rocks have clinopyroxene-garnet
modes closer to 50:50 than to the 45:55 employed, as discussed above. Despite this reduction,
densities of metasomatised eclogites are still considerably higher than those of peridotites at
similar depths, which are restricted to <3.5 g/cm3 even for the maximum bound determined
for cold cratonic peridotite by Garber et al. (2018) (i.e., using the same methods as in this
study). Kopylova et al. (2004) report lower densities for mantle eclogite xenoliths from the
northern Slave craton based on laboratory-determined derivatives for eclogite bulk rocks
(~3.4-3.5 g/cm3; see comparison in Fig. S5), but these rocks may have been affected by late-
stage changes in the xenoliths (partial melting of clinopyroxene due to decompression and
chemical interaction with host kimberlite; Gao et al., 2000; Pan et al., 2018). Apart from
metasomatic effects, several regional peculiarities emerge. For example, even unenriched
eclogites from Lace and Diavik are relatively FeO-poor and therefore have densities much
lower than the average eclogite at 38 or 40 mW/m2 from Garber et al. (2018) (Fig. 8a-b).
Conversely, unenriched eclogites from the northern Slave craton, and most eclogites from
Koidu regardless of their enrichment, are more FeO-rich and hence more dense. Even
recognising these metasomatic and regional differences, the range of eclogite densities
calculated in this study scatters about the range calculated by Garber et al. (2018) using
thermodynamic minimisation techniques along a similar geotherm.
It is worth highlighting that, since mantle metasomatism occurs after emplacement of
oceanic crust into the lithospheric mantle, the density changes described above do not pertain
to eclogite deeply recycled into the convective mantle. In contrast, “auto-metasomatised”
eclogite, which interacted with pelite-derived melts in subduction mélanges, also is markedly
FeO-poor (Fig. 3d), hence less dense. Such interactions may decrease slab pull and also be a
factor causing slabs to stagnate at mid-mantle depths (Xu et al. 2019), as the density
difference with the dominant mantle consisting of pyrolite is much smaller for low-Fe than
for high-Fe eclogite, although the extent of this type of metasomatism is difficult to assess.
©2020 American Geophysical Union. All rights reserved.
5.2.2 Shearwave velocity (VS)
At constant composition and pressures ~3 to 3.5 GPa kyanite disappears and eclogite
assemblages become dominated by garnet and clinopyroxene, with phase compositions and
modal proportions depending on bulk-rock eclogite composition. At higher pressures,
eclogite VS decreases along the geotherm as temperature increases, and is lower for warm
than for cold conductive geotherms (e.g., Garber et al., 2018). Shear-wave velocity is also
sensitive to composition as it increases with decreasing hedenbergite and almandine
component (cf. Garber et al., 2018, their Fig. 6), which are lowered by UM carbonated melt
metasomatism, and increases weakly with bulk-rock SiO2 content, which is higher in many
LREE-enriched eclogites inferred to have experienced stealth clinopyroxene addition (Fig.
3c). However, because metasomatism serves to deplete both the seismically slowest
endmembers (hedenbergite and almandine) and the seismically fastest endmembers (jadeite
and grossular), metasomatised eclogite VS is typically similar to velocities calculated for
unmetasomatised eclogites from the same region (Fig. 8c-d). As for density, there are distinct
regional differences in eclogite VS: eclogites from Lace are seismically fast (≥4.8 km/s)
regardless of enrichment, as are unenriched eclogites from Diavik at pressures > 5 GPa,
which may be explained by the generally SiO2-rich and FeO-poor compositions of these
suites (Fig. 3c,d). Eclogites from the northern Slave craton, presumably equilibrated to a cold
conductive geotherm, give some of the highest VS at mid-cratonic lithosphere depth (Fig. 8d).
Notably, with the exception of the Lace and N. Slave eclogite suites, and even allowing for a
small model-dependent uncertainty in VS (≤0.02 km/s; Garber et al., 2018), most eclogite
samples in this study are seismically slower than the range of eclogite compositions
considered in Garber et al. (2018). Such differences arise from a slight discrepancy between
clinopyroxene and garnet compositional endmembers calculated thermodynamically (Garber
et al., 2018) relative to those observed in mantle eclogites (this study), but a more significant
effect arises from differences in the calculated modes of each phase between studies (Fig. S5)
particularly because eclogite bulk compositions do not deviate significantly between the
two different approaches.
5.2.3 Electrical conductivity
Electrical conductivity, which can be extracted from regional magnetotelluric studies,
is strongly temperature-dependent, but also increases with increasing mineral H2O content
(e.g., Jones et al., 2009; Yang et al., 2011; Zhao & Yoshino, 2016). With the exception of the
Roberts Victor and Udachnaya (Siberia) eclogites studied by Huang et al. (2014) and
Kolesnichenko et al. (2018), respectively, there are no comprehensive published studies on
H2O content in mantle eclogite suites. Garnet in both studies has very low H2O contents, but
whereas clinopyroxene in eclogite from Udachnaya has a low median content of ~40 ppm,
that from Roberts Victor contains ~760 ppm H2O. Although the relationship between
metasomatism and H2O content in peridotitic minerals is tenuous at best (Marshall et al.,
2018), we calculate H2O content in clinopyroxene from LREE-enriched and HREE-enriched
eclogites assuming equilibrium with a melt containing 3 wt% H2O (similar to estimates for
primitive southern African kimberlites and orangeites), whereas 100 ppm is assigned to
clinopyroxene from unenriched ecogites. A comparison of expected H2O concentrations
using the approach outlined above vs. measured concentrations in eclogites from Roberts
Victor (Huang et al. 2014) shows that, unsurprisingly, there is only a very crude
correspondence, though median values do not differ strongly (760 ppm measured vs. 620
ppm calculated). Combined with dry garnet, we obtain bulk eclogite conductivities as
outlined in the caption to Figure 8. Our aim is to obtain a sense of the importance of this
parameter to bulk eclogitic electrical conductivity. For these hypothetical H2O contents,
©2020 American Geophysical Union. All rights reserved.
LREE-enriched eclogites have predictably higher conductivities than unenriched eclogites at
a given depth (hence temperature). As hypothetical H2O abundances were calculated for both
LREE- and HREE-enriched eclogites from Kimberley, they plot along a single trend (Fig. 8e-
f). Conductivities are higher than those expected for average eclogite used in Garber et al.
(2018) along applicable geotherms (36 mW/m2 for the northern Slave craton, 38 mW/m2 for
Lace, Roberts Victor, Diavik and Koidu, 40 mW/m2 for Kimberley), which directly reflects
their choice of purely dry assemblages.
5.3 Implications and conclusions
The role of mantle eclogite in the density, seismic velocity and electrical conductivity
signatures of cratonic lithosphere has recently been highlighted, using average eclogite and
peridotite compositions as well as high-Mg and low-Mg extremes with depth (Garber et al.,
2018). The present study aims to shed more light on the physical properties of mantle
eclogite, and also the effect of UM carbonated melt- and auto-metasomatism on eclogite
composition, using a wide range of observed mantle eclogite compositions and their
distribution with depth for various cratons.
The regional extent of mantle metasomatism remains poorly constrained. On the one
hand, it has frequently been cautioned that the mantle sampled by kimberlites is anomalous
because magmatism occurs along, in cases multiply, reactivated lithospheric pathways (e.g.,
Afonso et al., 2008). Consistent with this, the positive buoyancy observed in cratonic areas
away from kimberlite provinces, as opposed to near-neutral buoyancy within these areas, has
been ascribed to anomalous lithosphere enrichment associated with kimberlite magmatism,
indicating the presence of more refractory mantle away from those regions (Artemieva et al.,
2019). On the other hand, cratonic mid-lithospheric discontinuities, ascribed to deposition of
metasomes (e.g. Rader et al., 2015) are not restricted to kimberlite provinces and suggest that
metasomatism widely affects deep and warm lithospheric keels underlying cold and viscous
cratonic cores (e.g., Aulbach et al., 2017c). Regardless of the extent of metasomatism,
eclogite emplacement into the cratonic mantle starting in Mesoarchaean to Palaeoproterozoic
time (Shirey & Richardson, 2011) relates to large-scale geodynamic processes and is
unrelated to areas of kimberlite magmatism where subduction relics are accidentally
exhumed as xenoliths. In addition to uncertainties regarding regional extent, there are also
questions regarding the relevance to present-day geophysical observations. Thus, although
the effects of older metasomatism may persist through time (e.g. isotopically aged, ancient
metasomes tapped during Phanerozoic lamproite magmatism; Mitchell, 2006), all xenoliths
provide a snapshot of lithospheric conditions at the time of entrainment in the host kimberlite,
in cases several 100 Ma ago, and subsequent events may have modified the lithosphere up
until the present. This is exemplified in the southern African case, where significant warming,
refertilisation and thinning occurred between older (~Jurassic) orangeite and younger
(~Cretaceous) kimberlite emplacement (e.g. Kobussen et al., 2009).
If the metasomatism identified in the various eclogite suites under consideration here
is geographically restricted, it implies that (1) in areas of kimberlite activity, metasomatic
densification of mantle peridotite (e.g. Lee et al., 2011) may be partially offset by the lowered
density of metasomatised eclogite, and (2) in areas away from kimberlite activity,
positive/excess buoyancy of low-density refractory peridotite (Poudjom Djomani et al., 2001)
may be balanced by the presence of dense (i.e. unmetasomatised) eclogite, as suggested by
Kelly et al. (2003) and Garber et al. (2018). Except in HREE-enriched Kimberley eclogites,
where high FeO contents at higher pressures reflect infiltration of asthenosphere-derived
silicate melt, there is no correlation of FeO with temperature, suggesting that no density
sorting has taken place within the lithosphere. Very dense eclogites at Koidu and Diavik
©2020 American Geophysical Union. All rights reserved.
persist between ~2.5 and 5.5 GPa (Fig. 8b), possibly attesting to the strength, buoyancy, and
high viscosity of the dominant peridotitic lithosphere.
Although Kopylova et al. (2016) conclude that eclogites from the Slave craton do not
derive from depths corresponding to geophysically-detected discontinuities, the concentration
of seismically fast, LREE-enriched, SiO2-rich. and FeO-poor eclogites at mid-lithospheric
depths at several localities may help explain the high maximum velocities observed at similar
depths in the shearwave tomographic model of French and Romanowicz (2014) (Fig. 8c-d).
However, the results presented here also complicate interpretations that significant eclogite
fractions may help explain the fast shear-wave velocity signature of cratonic mantle
lithosphere globally (e.g., Garber et al., 2018). On the one hand, the decreased density but
unmodified VS signature of metasomatised eclogite permits significant fractions of
chemically modified, stiff, and fast-VS eclogite to be hosted in cratonic lithosphere perhaps
far more than postulated using gravity and density limitations (≤20 vol%: Garber et al.,
2018). On the other hand, even this theoretical limit exceeds those from empirical studies, for
example those using surveys of garnet xenocrysts (4%; McLean et al., 2007) or those based
on thermal state and heat production constraints (0.17 to 1.5%; Russell et al., 2001). Further,
though there are particular suites with Vs significantly higher than the cratonic average (Fig.
8; cf. French and Romanowicz, 2014), most calculated eclogite shear-wave velocities in this
study are slower than velocities calculated for thermodynamically modelled eclogites,
reflecting a key difference in the modal proportions of clinopyroxene and garnet. Therefore,
though small fractions of either metasomatised or unmetasomatised eclogite increase the bulk
shear-wave velocity of cratonic lithosphere above that of peridotite alone, our results
emphasise that such contributions cannot reconcile the observed VS; assuming that xenolith
P-T conditions accurately represent craton thermal structures, other constituents with fast
shear-wave velocities (e.g., diamond) must be present. Notably, we have not considered
diamond in concert with our eclogite results, but recognise the potential for VS modifications
driven by either diamond growth or destruction during metasomatism. Any diamond in
unenriched eclogites, and diamond in the extremely diamond-rich LREE-enriched eclogites
from the northern Slave craton (Smart et al., 2009) will increase bulk eclogite VS and
decrease the fraction of eclogite in explaining bulk mantle VS.
Eclogite conductivity is much more temperature-dependent, hence also dependent on
the geothermal gradient in addition to the depth of xenolith derivation, than on degree of
hydration. Nevertheless, hypothetical enrichment in H2O may have occurred along with
LREE around the depths of cratonic mid-lithospheric discontinuities, focused at ~60 to 160
km and potentially linked to the formation of amphibole-/phlogopite-and carbonate-bearing
metasomes (Rader et al., 2015; Aulbach et al., 2017c). Concentrations of H2O-rich eclogite
have been suggested to cause conductivity anomalies in the upper mantle (Liu et al., 2019)
and may help explain anomalous conductivities observed in the Slave cratonic mid-
lithosphere (Jones et al., 2001) and elsewhere (Selway, 2014). If H2O enrichment occurred
along with LREE around the depths of cratonic mid-lithospheric discontinuities as observed
in several localities (Diavik, Orapa, Kimberley), it may establish a link to the formation of
amphibole-/phlogopite-and carbonate-bearing metasomes, which have been suspected of
acting as weak zones where future decratonisation may occur (Aulbach et al. 2017c; Rader et
al. 2015). Perhaps not coincidentally, olivine in peridotite xenoliths from the Kaapvaal craton
reaches a maximum H2O content at ~160 km depth (Baptiste et al. 2012), and upward
increasing intensity of hydrous metasomatism has been implicated in generating cratonic
seismic velocity profiles as well as MLDs (Eeken et al. 2018). However, the conductivities of
dry or hypothetical wet eclogites in this study are much higher than bulk conductivities
reported for the Kaapvaal cratonic lithosphere, suggesting that either eclogites are generally
©2020 American Geophysical Union. All rights reserved.
much drier than inferred here and/or placing limits on the proportion of eclogite that may be
present in the regional lithospheric mantle.
Acknowledgments, Samples and Data
Discussions with Yana Fedortchouk are greatly appreciated. Jens Fiebig is thanked for
access to the stable isotope laboratory, and Dominik Gudelius for carrying out the analyses.
We gratefully acknowledge reviews by Thomas Stachel and Maya Kopylova, as well as
editorial comments by Ulrich Faul, all of which significantly improved the manuscript.
Funding to SA by the Deutsche Forschungsgemeinschaft under DFG-grant AU356/10 is
greatly appreciated. FIERCE is financially supported by the Wilhelm and Else Heraeus
Foundation, which is gratefully acknowledged. This is FIERCE contribution No. XX. KSV
and MM acknowledge financial support from the SA Department of Science and Technology
(DST) through their Research Chairs initiative, as administered by the National Research
Foundation (NRF), as well as financial support from the Centre of Excellence for Integrated
Mineral and Energy Resource Analysis (CIMERA) at the University of Johannesburg. JMG
acknowledges financial support from National Science Foundation Grant OISE-1545903 and
Penn State University.
The authors declare no conflicts of interest. Geochemical data are available at IEDA
online database (
Abers, G. A., & Hacker, B. R. (2016). A MATLAB toolbox and Excel workbook for
calculating the densities, seismic wave speeds, and major element composition of
minerals and rocks at pressure and temperature. Geochemistry Geophysics
Geosystems, 17(2), 616-624. doi:10.1002/2015gc006171
Adam, J., & Green, T. (2006). Trace element partitioning between mica- and amphibole-
bearing garnet lherzolite and hydrous basanitic melt: 1. Experimental results and the
investigation of controls on partitioning behaviour. Contributions to Mineralogy and
Petrology, 152(1), 1-17. doi:10.1007/s00410-006-0085-4
Afonso, J. C., Fernandez, M., Ranalli, G., Griffin, W. L., & Connolly, J. A. D. (2008).
Integrated geophysical-petrological modeling of the lithosphere and sublithospheric
upper mantle: Methodology and applications. Geochemistry Geophysics Geosystems,
9. doi:Q05008 10.1029/2007gc001834
Agashev, A. M., Pokhilenko, L. N., Pokhilenko, N. P., & Shchukina, E. V. (2018).
Geochemistry of eclogite xenoliths from the Udachnaya Kimberlite Pipe: Section of
ancient oceanic crust sampled. Lithos, 314, 187-200. doi:10.1016/j.lithos.2018.05.027
Allsopp, H. L., Bristow, J. W., Smith, C. B., Brown, R., Gleadow, A. J. W., Kramers, J. D.,
& Garvie, O. G. (1989). A summary of radiometric dating methods applicable to
kimberlites and related rocks. In J. Ross, A. L. Jacques, J. Ferguson, D. H. Green, S.
Y. O’Reilly, R. V. Danchin, & A. J. A. Janse (Eds.), Kimberlites and Related Rocks:
Their Composition, Occurrence, Origin and Emplacement, (pp. 343-357): Geological
Society of Australia Special Publication. Blackwell Scientific.
Artemieva, I. M., Thybo, H., & Cherepanova, Y. (2019). Isopycnicity of cratonic mantle
restricted to kimberlite provinces. Earth and Planetary Science Letters, 505, 13-19.
Aubaud, C., Hirschmann, M. M., Withers, A. C., & Hervig, R. L. (2008). Hydrogen
partitioning between melt, clinopyroxene, and garnet at 3 GPa in a hydrous MORB
©2020 American Geophysical Union. All rights reserved.
with 6 wt.% H2O. Contributions to Mineralogy and Petrology, 156(5), 607-625.
Aulbach, S., & Jacob, D. E. (2016). Major- and trace-elements in cratonic mantle eclogites
and pyroxenites reveal heterogeneous sources and metamorphic processing of low-
pressure protoliths. Lithos, 262, 586-605. doi:10.1016/j.lithos.2016.07.026
Aulbach, S., & Viljoen, K. S. (2015). Eclogite xenoliths from the Lace kimberlite, Kaapvaal
craton: From convecting mantle source to palaeo-ocean floor and back. Earth and
Planetary Science Letters, 431, 274-286. doi:10.1016/j.epsl.2015.08.039
Aulbach, S., O'Reilly, S. Y., Griffin, W. L., & Pearson, N. J. (2008). Subcontinental
lithospheric mantle origin of high niobium/tantalum ratios in eclogites. Nature
Geoscience, 1(7), 468-472. doi:10.1038/ngeo226
Aulbach, S., Creaser, R. A., Pearson, N. J., Simonetti, S. S., Heaman, L. M., Griffin, W. L.,
& Stachel, T. (2009). Sulfide and whole rock Re-Os systematics of eclogite and
pyroxenite xenoliths from the Slave Craton, Canada. Earth and Planetary Science
Letters, 283(1-4), 48-58. doi:10.1016/j.epsl.2009.03.023
Aulbach, S., Stachel, T., Heaman, L. M., & Carlson, J. A. (2011). Microxenoliths from the
Slave craton: Archives of diamond formation along fluid conduits. Lithos, 126(3-4),
419-434. doi:10.1016/j.lithos.2011.07.012
Aulbach, S., Gerdes, A., & Viljoen, K. S. (2016). Formation of diamondiferous kyanite-
eclogite in a subduction melange. Geochimica Et Cosmochimica Acta, 179, 156-176.
Aulbach, S., Jacob, D. E., Cartigny, P., Stern, R. A., Simonetti, S. S., Worner, G., & Viljoen,
K. S. (2017a). Eclogite xenoliths from Orapa: Ocean crust recycling, mantle
metasomatism and carbon cycling at the western Zimbabwe craton margin.
Geochimica Et Cosmochimica Acta, 213, 574-592. doi:10.1016/j.gca.2017.06.038
Aulbach, S., Woodland, A. B., Vasilyev, P., Galvez, M. E., & Viljoen, K. S. (2017b). Effects
of low-pressure igneous processes and subduction on Fe3+/ Fe and redox state of
mantle eclogites from Lace (Kaapvaal craton). Earth and Planetary Science Letters,
474, 283-295. doi:10.1016/j.epsl.2017.06.030
Aulbach, S., Massuyeau, M., & Gaillard, F. (2017c). Origins of cratonic mantle
discontinuities: A view from petrology, geochemistry and thermodynamic models.
Lithos, 268, 364-382. doi:10.1016/j.lithos.2016.11.004
Aulbach, S., Höfer, H. E., & Gerdes, A. (2019a). High-Mg mantle eclogites from Koidu
(West African craton): Neoproterozoic ultramafic melt metasomatism of subducted
Archaean plateau-like oceanic crust. Journal of Petrology, 60, 723-754. doi:
Aulbach, S., Heaman, L. M., Jacob, D. E., & Viljoen, K. S. (2019b). Ages and sources of
mantle eclogites: ID-TIMS and in situ MC-ICPMS Pb-Sr isotope systematics of
clinopyroxene. Chemical Geology, 503, 15-28. doi:10.1016/j.chemgeo.2018.10.007
Aulbach, S., A. B. Woodland, R. A. Stern, P. Vasilyev, L. M. Heaman, & Viljoen, K. S.
(2019c). Evidence for a dominantly reducing Archaean ambient mantle from two
redox proxies, and low oxygen fugacity of deeply subducted oceanic crust Scientific
Reports, 9, 20190.
Barth, M. G., Rudnick, R. L., Horn, I., McDonough, W. F., Spicuzza, M. J., Valley, J. W., &
Haggerty, S. E. (2001). Geochemistry of xenolithic eclogites from West Africa, Part I:
A link between low MgO eclogites and Archean crust formation. Geochimica Et
Cosmochimica Acta, 65(9), 1499-1527.
Barth, M. G., Rudnick, R. L., Horn, I., McDonough, W. F., Spicuzza, M. J., Valley, J. W., &
Haggerty, S. E. (2002). Geochemistry of xenolithic eclogites from West Africa, part
2: Origins of the high MgO eclogites. Geochimica Et Cosmochimica Acta, 66(24),
©2020 American Geophysical Union. All rights reserved.
Becker, M., & Le Roex, A. P. (2006). Geochemistry of South African on- and off-craton,
Group I and Group II kimberlites: Petrogenesis and source region evolution. Journal
of Petrology, 47(4), 673-703. doi:10.1093/petrology/egi089
Bell, D. R., Rossman, G. R., & Moore, R. O. (2004). Abundance and partitioning of OH in a
high-pressure magmatic system: Megacrysts from the Monastery kimberlite, South
Africa. Journal of Petrology, 45(8), 1539-1564. doi:10.1093/petrology/
Beyer, C., D. J. Frost, and N. Miyajima (2015), Experimental calibration of a garnet-
clinopyroxene geobarometer for mantle eclogites, Contributions to Mineralogy and
Petrology, 169(2), 18.
BlichertToft, J., & Albarède, F. (1997). The Lu-Hf isotope geochemistry of chondrites and
the evolution of the mantle-crust system. Earth and Planetary Science Letters, 148(1-
2), 243-258. doi:10.1016/s0012-821x(97)00040-x
Blichert-Toft, J., & Puchtel, I. S. (2010). Depleted mantle sources through time: Evidence
from Lu-Hf and Sm-Nd isotope systematics of Archean komatiites. Earth and
Planetary Science Letters, 297(3-4), 598-606. doi:10.1016/j.epsl.2010.07.012
Bureau, H., Remusat, L., Esteve, M., Pinti, D. L., & Cartigny, P. (2018). The growth of
lithospheric diamonds. Science Advances, 4(6). doi:10.1126/sciadv.aat1602
Connolly, J. A. D. (2009), The geodynamic equation of state: What and how, Geochemistry
Geophysics Geosystems, 10, Q10014
Creaser, R. A., Grutter, H., Carlson, J., & Crawford, B. (2004). Macrocrystal phlogopite Rb-
Sr dates for the Ekati property kimberlites, Slave Province, Canada: evidence for
multiple intrusive episodes in the Paleocene and Eocene. Lithos, 76(1-4), 399-414.
Czas, J., Stachel, T., Pearson, D. G., Stern, R. A., & Read, G. H. (2018). Diamond
brecciation and annealing accompanying major metasomatism in eclogite xenoliths
from the Sask Craton, Canada. Mineralogy and Petrology, 112, 311-323.
Dai, L. D., & Karato, S. (2009). Electrical conductivity of pyrope-rich garnet at high
temperature and high pressure. Physics of the Earth and Planetary Interiors, 176(1-2),
83-88. doi:10.1016/j.pepi.2009.04.002
Davis, G. L. (1977). The ages and uranium content of zircons from kimberlites and
associated rocks. Paper presented at the 2nd International Kimberlite Conference,
Santa Fe, USA.
Dawson, J. B. (1984). Contrasting types of upper mantle metasomatism. In J. Kornprobst
(Ed.), Kimberlites. II. The mantle and crust-mantle relationships (pp. 289-294).
Amsterdam: Elsevier.
Day, H. W. (2012), A revised diamond-graphite transition curve, American Mineralogist,
97(1), 52-62.
De Stefano, A., M. G. Kopylova, P. Cartigny, & Afanasiev, V. (2009). Diamonds and
eclogites of the Jericho kimberlite (Northern Canada), Contributions to Mineralogy
and Petrology, 158(3), 295-315.
Dietrich, H., & Petrakakis, K. (1986). A Linear Algebraic Method for the Calculation of
Pyroxene Endmember Components. Tchermaks Mineralogische und Petrographische
Mitteilungen, 35: 275-285.
Djomani, Y. H. P., O'Reilly, S. Y., Griffin, W. L., & Morgan, P. (2001). The density structure
of subcontinental lithosphere through time. Earth and Planetary Science Letters,
184(3-4), 605-621.
Dludla, S., le Roex, A. P., & Gurney, J. J. (2006). Eclogite xenoliths from the Premier
kimberlite, South Africa: geochemical evidence for a subduction origin. South African
©2020 American Geophysical Union. All rights reserved.
Journal of Geology, 109(3), 353-368. doi:10.2113/gssajg.109.3.353Fedortchouk, Y.,
and D. Canil (2004), Intensive variables in kimberlite magmas, Lac de Gras, Canada
and implications for diamond survival, Journal of Petrology, 45(9), 1725-1745.
Dohmen, R., Kasemann, S. A., Coogan, L., & Chakraborty, S. (2010). Diffusion of Li in
olivine. Part I: Experimental observations and a multi species diffusion model,
Geochimica Et Cosmochimica Acta, 74, 274-292.
Fedortchouk, Y., Liebske, C., & McCammon, C. (2019). Diamond destruction and growth
during mantle metasomatism: An experimental study of diamond resorption features.
Earth and Planetary Science Letters, 506, 493-506. doi:10.1016/j.epsl.2018.11.025
French, S. W., & Romanowicz, B. A. (2014). Whole-mantle radially anisotropic shear
velocity structure from spectral-element waveform tomography. Geophysical Journal
International, 199(3), 1303-1327. doi:10.1093/gji/ggu334
Gaetani, G. A., Asimow, P. D., & Stolper, E. M. (2008). A model for rutile saturation in
silicate melts with applications to eclogite partial melting in subduction zones and
mantle plumes. Earth and Planetary Science Letters, 272(3-4), 720-729.
Gale, A., Dalton, C. A., Langmuir, C. H., Su, Y. J., & Schilling, J. G. (2013). The mean
composition of ocean ridge basalts. Geochemistry Geophysics Geosystems, 14(3),
489-518. doi:10.1029/2012gc004334Gao, S., H. Kern, Y. S. Liu, S. Y. Jin, T. Popp,
Z. M. Jin, J. L. Feng, M. Sun, & Zhao, Z. B. (2000). Measured and calculated seismic
velocities and densities for granulites from xenolith occurrences and adjacent exposed
lower crustal sections: A comparative study from the North China craton, Journal of
Geophysical Research-Solid Earth, 105(B8).
Garber, J. M., Maurya, S., Hernandez, J. A., Duncan, M. S., Zeng, L., Zhang, H. L. L., . . .
Stixrude, L. (2018). Multidisciplinary Constraints on the Abundance of Diamond and
Eclogite in the Cratonic Lithosphere. Geochemistry Geophysics Geosystems, 19(7),
2062-2086. doi:10.1029/2018gc007534
Giuliani, A., Phillips, D., Kamenetsky, V. S., Kendrick, M. A., Wyatt, B. A., Goemann, K.,
& Hutchinson, G. (2014). Petrogenesis of Mantle Polymict Breccias: Insights into
Mantle Processes Coeval with Kimberlite Magmatism. Journal of Petrology, 55(4),
831-858. doi:10.1093/petrology/egu008
Giuliani, A., Phillips, D., Woodhead, J. D., Kamenetsky, V. S., Fiorentini, M. L., Maas, R., . .
. Armstrong, R. A. (2015). Did diamond-bearing orangeites originate from MARID-
veined peridotites in the lithospheric mantle? Nature Communications, 6, 6837.
doi:10.1038/ncomms7837Gregory, R. T., & Taylor, H. P. (1981). An oxygen isotope
profile in a section of cretaceous oceanic crust, Samail ophiolite, Oman Evidence
for δ18O buffering oft he oceans by deep (>5 km) seawater-hydrothermal circulation
at mid-ocean ridges, Journal of Geophysical Research, 86(NB4), 2737-2755.
Griffin, W. L., O'Reilly, S. Y., Natapov, L. M., & Ryan, C. G. (2003). The evolution of
lithospheric mantle beneath the Kalahari Craton and its margins. Lithos, 71(2-4), 215-
241. doi:10.1016/j.lithos.2003.07.006
Griffin, W. L., B. J. Doyle, C. G. Ryan, N. J. Pearson, S. Y. O'Reilly, R. Davies, K. Kivi, E.
Van Achterbergh, & L. M. Natapov (1999a), Layered mantle lithosphere in the Lac de
Gras area, Slave Craton: Composition, structure and origin, Journal of Petrology,
40(5), 705-727.
Griffin, W. L., Shee, S. R., Ryan, C. G., Win, T. T., & Wyatt, B. A. (1999b). Harzburgite to
lherzolite and back again: metasomatic processes in ultramafic xenoliths from the
Wesselton kimberlite, Kimberley, South Africa. Contributions to Mineralogy and
Petrology, 134(2-3), 232-250. doi:10.1007/s004100050481Grütter, H. S. (2009).
Pyroxene xenocryst geotherms: Techniques and application. Lithos, 112, 1167-1178.
©2020 American Geophysical Union. All rights reserved.
Hamilton, P. J., Onions, R. K., Bridgwater, D., & Nutman, A. (1983). Sm-Nd studies of
Archean metasediments and metavolcanics from West Greenland and their
implications for the Earths early history. Earth and Planetary Science Letters, 62(2),
263-272. doi:10.1016/0012-821x(83)90089-4
Hart, S. R., J. Blusztajn, H. J. B. Dick, P. S. Meyer, and K. Muehlenbachs (1999), The
fingerprint of seawater circulation in a 500-meter section of ocean crust gabbros,
Geochimica Et Cosmochimica Acta, 63(23-24), 4059-4080.
Harte, B., & Kirkley, M. B. (1997). Partitioning of trace elements between clinopyroxene and
garnet: Data from mantle eclogites. Chemical Geology, 136(1-2), 1-24.
Hasterok, D., & Chapman, D. S. (2011). Heat production and geotherms for the continental
lithosphere. Earth and Planetary Science Letters, 307(1-2), 59-70.
Hayman, P. C., Cas, R. A. F., & Johnson, M. (2009). Characteristics and alteration origins of
matrix minerals in volcaniclastic kimberlite of the Muskox pipe (Nunavut, Canada).
Lithos, 112, 473-487. doi:10.1016/j.lithos.2009.06.025
Heaman, L. M., Creaser, R. A., & Cookenboo, H. O. (2002). Extreme enrichment of high
field strength elements in Jericho eclogite xenoliths: A cryptic record of
Paleoproterozoic subduction, partial melting, and metasomatism beneath the Slave
craton, Canada. Geology, 30(6), 507-510.
Heaman, L. M., Kjarsgaard, B. A., & Creaser, R. A. (2004). The temporal evolution of North
American kimberlites. Lithos, 76(1-4), 377-397. doi:10.1016/j.lithos.2004.03.047
Heaman, L. M., Creaser, R. A., Cookenboo, H. O., & Chacko, T. (2006). Multi-stage
modification of the northern Slave mantle lithosphere: Evidence from zircon- and
diamond-bearing eclogite xenoliths entrained in Jericho kimberlite, Canada. Journal
of Petrology, 47(4), 821-858. doi:10.1093/petrology.egi097
Helmstaedt, H., & Schulze, D. J. (1989). Southern African kimberlites and their mantle
sample; implications for Archean tectonics and lithosphere evolution. In R. Rushmer
(Ed.), Evolution and Differentiation of the Continental Crust (pp. 67-91). Cambridge:
Cambridge University Press.
Herzberg, C., & Asimow, P. D. (2008). Petrology of some oceanic island basalts:
PRIMELT2.XLS software for primary magma calculation. Geochemistry Geophysics
Geosystems, 9. doi:10.1029/2008gc002057
Hills, D. V., & Haggerty, S. E. (1989). Petrochemistry of eclogites from the Koidu kimberlite
complex, Sierra-Leone. Contributions to Mineralogy and Petrology, 103(4), 397-422.
Hopp, J., Trieloff, M., Brey, G. P., Woodland, A. B., Simon, N. S. C., Wijbrans, J. R., . . .
Reitter, E. (2008). 40Ar/39Ar-ages of phlogopite in mantle xenoliths from South
African kimberlites: Evidence for metasomatic mantle impregnation during the
Kibaran orogenic cycle. Lithos, 106(3-4), 351-364. doi:10.1016/j.lithos.2008.09.001
Howarth, G. H., Michael, E., Skinner, W., & Prevec, S. A. (2011). Petrology of the
hypabyssal kimberlite of the Kroonstad group II kimberlite (orangeite) cluster, South
Africa: Evolution of the magma within the cluster. Lithos, 125(1-2), 795-808.
Huang, J. X., Greau, Y., Griffin, W. L., O'Reilly, S. Y., & Pearson, N. J. (2012). Multi-stage
origin of Roberts Victor eclogites: Progressive metasomatism and its isotopic effects.
Lithos, 142, 161-181. doi:10.1016/j.lithos.2012.03.002
Huang, J. X., Li, P., Griffin, W. L., Xia, Q. K., Greau, Y., Pearson, N. J., & O'Reilly, S. Y.
(2014). Water contents of Roberts Victor xenolithic eclogites: primary and
metasomatic controls. Contributions to Mineralogy and Petrology, 168(6).
©2020 American Geophysical Union. All rights reserved.
Ireland, T. R., Rudnick, R. L., & Spetsius, Z. (1994). Trace elements in diamond inclusions
from eclogites reveal link to Archean granites. Earth and Planetary Science Letters,
128(3-4), 199-213.
Izraeli, E. S., Harris, J. W., & Navon, O. (2001). Brine inclusions in diamonds: a new upper
mantle fluid. Earth and Planetary Science Letters, 187(3-4), 323-332.
Jacob, D. E. (2004). Nature and origin of eclogite xenoliths from kimberlites. Lithos, 77(1-4),
295-316. doi:10.1016/j.lithos.2004.03.038Jacob, D. E., B. Schmickler, & Schulze, D.
J. (2003), Trace element geochemistry of coesite-bearing eclogites from the Roberts
Victor kimberlite, Kaapvaal craton, Lithos, 71(2-4), 337-351.
Jacob, D. E., Viljoen, K. S., & Grassineau, N. V. (2009). Eclogite xenoliths from Kimberley,
South Africa - A case study of mantle metasomatism in eclogites. Lithos, 112, 1002-
1013. doi:10.1016/j.lithos.2009.03.034
Jenner, F. E., and H. S. O'Neill (2012), Analysis of 60 elements in 616 ocean floor basaltic
glasses, Geochemistry Geophysics Geosystems, 13.
Jollands, M. C., Hanger, B. J., Yaxley, G. M., Hermann, J., & Kilburn, M. R. (2018).
Timescales between mantle metasomatism and kimberlite ascent indicated by
diffusion profiles in garnet crystals from peridotite xenoliths. Earth and Planetary
Science Letters, 481, 143-153. doi:10.1016/j.epsl.2017.10.021
Jones, A. G., Ferguson, I. J., Chave, A. D., Evans, R. L., & McNeice, G. W. (2001). Electric
lithosphere of the Slave craton. Geology, 29(5), 423-426.
Jones, A. G., Evans, R. L., & Eaton, D. W. (2009). Velocity-conductivity relationships for
mantle mineral assemblages in Archean cratonic lithosphere based on a review of
laboratory data and Hashin-Shtrikman extremal bounds. Lithos, 109(1-2), 131-143.
Kelly, R. K., Kelemen, P. B., & Jull, M. (2003). Buoyancy of the continental upper mantle.
Geochemistry Geophysics Geosystems, 4. doi:10.1029/2002gc000399
Klemme, S., Vanderlaan, S. R., Foley, S. F., & Gunther, D. (1995). Experimentally-
determined trace and minor element partitioning between clinpyroxene and
carbonatite melt under upper-mantle conditions. Earth and Planetary Science Letters,
133(3-4), 439-448. doi:10.1016/0012-821x(95)00098-w
Knapp, N., Woodland, A. B., & Klimm, K. (2015). Experimental constraints on coesite
abundances in eclogite and implications for the X seismic discontinuity. Journal of
Geophysical Research-Solid Earth, 120(7), 4917-4930. doi:10.1002/2015jb011933
Kobussen, A. F., Griffin, W. L., & O'Reilly, S. Y. (2009). Cretaceous thermo-chemical
modification of the Kaapvaal cratonic lithosphere, South Africa. Lithos, 112, 886-895.
Kolesnichenko, M. V., Zedgenizov, D. A., Ragozin, A. L., Litasov, K. D., & Shatsky, V. S.
(2018). The role of eclogites in the redistribution of water in the subcontinental
mantle of the Siberian craton: results of determination of the water content in minerals
from the Udachnaya pipe eclogites. Russian Geology and Geophysics, 59(7), 763-779.
Kopylova, M. G., Lo, J., & Christensen, N. I. (2004). Petrological constraints on seismic
properties of the Slave upper mantle (Northern Canada). Lithos, 77, 493-510.
Kopylova, M. G., Beausoleil, Y., Goncharov, A., Burgess, J., & Strand, P. (2016). Spatial
distribution of eclogite in the Slave cratonic mantle: The role of subduction.
Tectonophysics, 672, 87-103. doi:10.1016/j.tecto.2016.01.034
Korolev, N. M., Melnik, A. E., Li, X. H., & Skublov, S. G. (2018). The oxygen isotope
composition of mantle eclogites as a proxy of their origin and evolution: A review.
©2020 American Geophysical Union. All rights reserved.
Earth-Science Reviews, 185, 288-300. doi:10.1016/j.earscirev.2018.06.007Krogh, E.
J. (1988), The garnet-clinopyroxene Fe-Mg geothermometer - a reinterpretation of
existing experimental data, Contributions to Mineralogy and Petrology, 99(1), 44-48.
Krogh Ravna, E. J. (2000). The garnet-clinopyroxene Fe2+-Mg geothermometer: an updated
calibration. Journal of Metamorphic Geology, 18, 211-219.
Le Roex, A. P., Bell, D. R., & Davis, P. (2003). Petrogenesis of group I kimberlites from
Kimberley, South Africa: Evidence from bulk-rock geochemistry. Journal of
Petrology, 44(12), 2261-2286. doi:10.1093/petrology/egg077
Lee, C. T. A., Cheng, X., & Horodyskyj, U. (2006). The development and refinement of
continental arcs by primary basaltic magmatism, garnet pyroxenite accumulation,
basaltic recharge and delamination: insights from the Sierra Nevada, California.
Contributions to Mineralogy and Petrology, 151(2), 222-242. doi:10.1007/s00410-
Lee, C. T. A., Luffi, P., & Chin, E. J. (2011). Building and Destroying Continental Mantle. In
R. Jeanloz & K. H. Freeman (Eds.), Annual Review of Earth and Planetary Sciences,
Vol 39 (Vol. 39, pp. 59-90).
Liu, H., Q. Zhu, and X. Yang (2019), Electrical conductivity of OH-bearing omphacite and
garnet in eclogite: the quantitative dependence on water content, Contrib Mineral
Petrol, 174, 57.
Lugmair, G. W., & Marti, K. (1978). Lunar initial Nd-143-Nd-144 - Differential evolution of
lunar crust and mantle. Earth and Planetary Science Letters, 39(3), 349-357.
doi:10.1016/0012-821x(78)90021-3Luth, R. W., & Stachel, T. (2014). The buffering
capacity of lithospheric mantle: implications for diamond formation, Contributions to
Mineralogy and Petrology, 168(5).
MacGregor, I. D., & Carter, J. L. (1970). The chemistry of clinopyroxenes and garnets of
eclogite and peridotite xenoliths from the Roberts Victor mine, South Africa. Physics
of the Earth and Planetary Interiors, 3, 391397. Mallik, A., & Dasgupta, R. (2013).
Reactive Infiltration of MORB-Eclogite-Derived Carbonated Silicate Melt into Fertile
Peridotite at 3 GPa and Genesis of Alkalic Magmas. Journal of Petrology, 54, 2267-
Marshall, E. W., Lassiter, J. C., & Barnes, J. D. (2018). On the (mis)behavior of water in the
mantle: Controls on nominally anhydrous mineral water content in mantle peridotites.
Earth and Planetary Science Letters, 499, 219-229. doi:10.1016/j.epsl.2018.07.033
Mather, K. A. (2012). A xenolith-based lithospheric transect of the Slave craton, N.W.T.,
Canada. (Doctor of Philosophy), Durham University, Durham.
Mattey, D., Lowry, D., & Macpherson, C. (1994). Oxygen-isotope composition of mantle
peridotite. Earth and Planetary Science Letters, 128(3-4), 231-241. doi:10.1016/0012-
McCandless, T. E., & Gurney, J. J. (1989). Sodium in garnet and potassium in clinopyroxene:
Criteria for classifying mantle eclogites. In J. Ross, A. L. Jacques, J. Ferguson, D. H.
Green, S. Y. O’Reilly, R. V. Danchin, & A. J. A. Janse (Eds.), Kimberlites and
Related Rocks, Vol 2. Their Mantle/Crust Setting, Diamonds and Diamond
Exploration (Vol. 14, pp. 827-832). Carlton: Geological Society of Australia Special
Publication No. 14. Blackwell Scientific.
McLean, H., Banas, A., Creighton, S., Whiteford, S., Luth, R. W., & Stachel, T. (2007).
Garnet xenocrysts from the Diavik mine, NWT, Canada: composition, color, and
paragenesis. Canadian Mineralogist, 45, 1131-1145.
Miensopust, M. P., Jones, A. G., Muller, M. R., Garcia, X., & Evans, R. L. (2011).
Lithospheric structures and Precambrian terrane boundaries in northeastern Botswana
revealed through magnetotelluric profiling as part of the Southern African
©2020 American Geophysical Union. All rights reserved.
Magnetotelluric Experiment. Journal of Geophysical Research-Solid Earth, 116.
Misra, K. C., Anand, M., Taylor, L. A., & Sobolev, N. V. (2004). Multi-stage metasomatism
of diamondiferous eclogite xenoliths from the Udachnaya kimberlite pipe, Yakutia,
Siberia. Contributions to Mineralogy and Petrology, 146(6), 696-714.
Mitchell, R. H. (2006). Potassic magmas derived from metasomatized lithospheric mantle:
Nomenclature and relevance to exploration for diamond-bearing rocks. Journal of the
Geological Society of India, 67(3), 317-327.
Morimoto, N., Gabries, J., Ferguson, A.K., Ginzburg, I.V., Ross, M., Seifert, F.A., Zussman,
J., Aoki, K., & Gottardi, G. (1988). Nomenclature of Pyroxenes. American
Mineralogist, 73, 1123-1133.
Moss, S. W., Kobussen, A., Powell, W., & Pollock, K. (2018). Kimberlite emplacement and
mantle sampling through time at A154N kimberlite volcano, Diavik Diamond Mine:
lessons from the deep. Mineralogy and Petrology, 112, 397-410. doi:10.1007/s00710-
018-0630-7Muehlenbachs, K., & Clayton, R. N. (1972a). Oxygen isotope studies of
fresh and weathered submarine basalts, Canadian Journal of Earth Sciences, 9(2),
Muehlenbachs, K., & Clayton, R. N. (1972b). Oxygen isotope geochemistry of submarine
greenstones, Canadian Journal of Earth Sciences, 9(5), 471-478.
Nikitina, L. P., N. M. Korolev, V. N. Zinchenko, & J. T. Felix (2014), Eclogites from the
upper mantle beneath the Kasai Craton (Western Africa): Petrography, whole-rock
geochemistry and U-Pb zircon age, Precambrian Research, 249, 13-32.
O'Reilly, S. Y., & Griffin, W. L. (1995). Trace-element partitioning between garnet and
clinopyroxene in mantle-derived pyroxenites and eclogites - P-T-X controls.
Chemical Geology, 121, 105-130.
O'Reilly, S. Y., & Griffin, W. L. (2013). Mantle Metasomatism. In D. E. Harlov & H.
Austrheim (Eds.), Metasomatism and the Chemical Transformation of Rock (pp. 471-
533). Berlin Heidelberg: Springer Verlag.Pan, S., J. Zheng, Z. Yin, W. L. Griffin, M.
Xia, A. Lin, & Zhang, H. (2018). Spongy texture in mantle clinopyroxene records
decompression-induced melting, Lithos, 320-321, 144-154.
Patten, C., Barnes, S. J., Mathez, E. A., & Jenner, F. E. (2013). Partition coefficients of
chalcophile elements between sulfide and silicate melts and the early crystallization
history of sulfide liquid: LA-ICP-MS analysis of MORB sulfide droplets. Chemical
Geology, 358, 170-188. doi:10.1016/j.chemgeo.2013.08.040
Poli, S. (2015). Carbon mobilized at shallow depths in subduction zones by carbonatitic
liquids. Nature Geoscience, 8(8), 633-+. doi:10.1038/ngeo2464Pollack, H. N.,
Chapman, D. S. (1977). On the regional variation of heat-flow, geotherms, and
lithospheric thickness, Tectonophysics, 38(3-4), 279-296.
Pyle, J. M., & Haggerty, S. E. (1998). Eclogites and the metasomatism of eclogites from the
Jagersfontein Kimberlite: Punctuated transport and implications for alkali magmatism.
Geochimica Et Cosmochimica Acta, 62(7), 1207-1231. doi:10.1016/s0016-
Rader, E., Emry, E., Schmerr, N., Frost, D., Cheng, C., Menard, J., . . . Geist, D. (2015).
Characterization and Petrological Constraints of the Midlithospheric Discontinuity.
Geochemistry Geophysics Geosystems, 16(10), 3484-3504.
Radu, I. B., Harris, C., Moine, B. N., Costin, G., & Cottin, J. Y. (2019). Subduction relics in
the subcontinental lithospheric mantle evidence from variation in the O-18 value of
eclogite xenoliths from the Kaapvaal craton. Contributions to Mineralogy and
©2020 American Geophysical Union. All rights reserved.
Petrology, 174(3). doi:10.1007/s00410-019-1552-z
Riches, A. J. V., Ickert, R. B., Pearson, D. G., Stern, R. A., Jackson, S. E., Ishikawa, A., . . .
Gurney, J. J. (2016). In situ oxygen-isotope, major-, and trace-element constraints on
the metasomatic modification and crustal origin of a diamondiferous eclogite from
Roberts Victor, Kaapvaal Craton. Geochimica Et Cosmochimica Acta, 174, 345-359.
doi:10.1016/j.gca.2015.11.028Roden, M. F., and R. V. Murthy (1985), Mantle
Metasomatism, Annual Review of Earth and Planetary Sciences, 13, 269-296.
Rudnick, R. L., & Fountain, D. M. (1995). Nature and composition of the continental crust -
a lower crustal perspective. Reviews of Geophysics, 33(3), 267-309.
doi:10.1029/95rg01302Rudnick, R. L., & Nyblade, A. (1999). The thickness and heat
production of Archean lithosphere: constraints from xenolith thermobarometry and
sur- face heat flow, in Mantle Petrology: Field Observations and High Pressure
Experimentation: A Tribute to Francis R. (Joe) Boyd, edited by Y. Fei, C. Bertka and
B. Mysen, Geochemical Society.
Russell, J. K., Dipple, G. M., & Kopylova, M. G. (2001). Heat production and heat flow in
the mantle lithosphere, Slave craton, Canada. Physics of the Earth and Planetary
Interiors, 123(1), 27-44. doi:10.1016/s0031-9201(00)00201-6
Schmidberger, S. S., Simonetti, A., Heaman, L. M., Creaser, R. A., & Whiteford, S. (2007).
Lu-Hf, in-situ Sr and Pb isotope and trace element systematics for mantle eclogites
from the Diavik diamond mine: Evidencefor Paleoproterozoic subduction beneath the
Slave craton, Canada. Earth and Planetary Science Letters, 254(1-2), 55-68.
Schulze, D. J. (1989). Constraints on the abundance of eclogite in the upper mantle. Journal
of Geophysical Research-Solid Earth and Planets, 94(B4), 4205-4212.
Schulze, D. J., Wiese, D., & Steude, J. (1996). Abundance and distribution of diamonds in
eclogite revealed by volume visualization of CT X-ray scans. Journal of Geology,
104(1), 109-114.
Selway, K. (2014). On the Causes of Electrical Conductivity Anomalies in Tectonically
Stable Lithosphere. Surveys in Geophysics, 35(1), 219-257. doi:10.1007/s10712-013-
Shatsky, V., Ragozin, A., Zedgenizov, D., & Mityukhin, S. (2008). Evidence for multistage
evolution in a xenolith of diamond-bearing eclogite from the Udachnaya kimberlite
pipe. Lithos, 105(3-4), 289-300. doi:10.1016/j.lithos.2008.04.008
Shchukina, E. V., Agashev, A. M., & Zedgenizov, D. A. (2018). Origin of zircon-bearing
mantle eclogites entrained in the V. Grib kimberlite (Arkhangelsk region, NW
Russia): Evidence from mineral geochemistry and the U-Pb and Lu-Hf isotope
compositions of zircon. Mineralogy and Petrology, 112, 85-100. doi:10.1007/s00710-
Shirey, S. B., & Richardson, S. H. (2011). Start of the Wilson Cycle at 3 Ga Shown by
Diamonds from Subcontinental Mantle. Science, 333(6041), 434-436.
Shu, Q., Brey, G. P., & Pearson, D. G. (2018). Eclogites and garnet pyroxenites from
Kimberley, Kaapvaal craton, South Africa: their diverse origins and complex
metasomatic signatures. Mineralogy and Petrology, 112, 43-56. doi:10.1007/s00710-
Skinner, E. M. W., Apter, D. B., Morelli, C., & Smithson, N. K. (2004). Kimberlites of the
Man craton, West Africa. Lithos, 76(1-4), 233-259. doi:10.1016/j.lithos.2004.04.034
Smart, K. A., Heaman, L. M., Chacko, T., Simonetti, A., Kopylova, M., Mah, D., & Daniels,
D. (2009). The origin of high-MgO diamond eclogites from the Jericho Kimberlite,
Canada. Earth and Planetary Science Letters, 284(3-4), 527-537.
©2020 American Geophysical Union. All rights reserved.
Smart, K. A., Chacko, T., Stachel, T., Tappe, S., Stern, R. A., Ickert, R. B., & Eimf. (2012).
Eclogite formation beneath the northern Slave craton constrained by diamond
inclusions: Oceanic lithosphere origin without a crustal signature. Earth and
Planetary Science Letters, 319, 165-177. doi:10.1016/j.epsl.2011.12.032
Smart, K. A., Chacko, T., Simonetti, A., Sharp, Z. D., & Heaman, L. M. (2014). A Record of
Paleoproterozoic Subduction Preserved in the Northern Slave Cratonic Mantle: Sr-Pb-
O Isotope and Trace-element Investigations of Eclogite Xenoliths from the Jericho
and Muskox Kimberlites. Journal of Petrology, 55(3), 549-583.
Smart, K. A., Tappe, S., Simonetti, A., Simonetti, S. S., Woodland, A. B., & Harris, C.
(2017). Tectonic significance and redox state of Paleoproterozoic eclogite and
pyroxenite components in the Slave cratonic mantle lithosphere, Voyageur kimberlite,
Arctic Canada. Chemical Geology, 455, 98-119. doi:10.1016/j.chemgeo.2016.10.014
Smit, K. V., Stachel, T., Creaser, R. A., Ickert, R. B., DuFrane, S. A., Stern, R. A., & Seller,
M. (2014). Origin of eclogite and pyroxenite xenoliths from the Victor kimberlite,
Canada, and implications for Superior craton formation. Geochimica Et
Cosmochimica Acta, 125, 308-337. doi:10.1016/j.gca.2013.10.019
Smit, K. V., Shirey, S. B., & Wang, W. Y. (2016). Type Ib diamond formation and
preservation in the West African lithospheric mantle: Re-Os age constraints from
sulphide inclusions in Zimmi diamonds. Precambrian Research, 286, 152-166.
Smit, K. V., T. Stachel, R. W. Luth, and R. A. Stern (2019), Evaluating mechanisms for
eclogitic diamond growth: An example from Zimmi Neoproterozoic diamonds (West
African craton), Chemical Geology, 520, 21-32.
Sokol, A. G., Kupriyanov, I. N., Palyanov, Y. N., Kruk, A. N., & Sobolev, N. V. (2013).
Melting experiments on the Udachnaya kimberlite at 6.3-7.5 GPa: Implications for the
role of H2O in magma generation and formation of hydrous olivine. Geochimica Et
Cosmochimica Acta, 101, 133-155. doi:10.1016/j.gca.2012.10.018
Spetsius, Z. V., & Taylor, L. A. (2002). Partial melting in mantle eclogite xenoliths:
Connections with diamond paragenesis. International Geology Review, 44(11), 973-
987. doi:10.2747/0020-6814.44.11.973
Stachel, T., & Luth, R. W. (2015). Diamond formation - Where, when and how? Lithos, 220,
200-220. doi:10.1016/j.lithos.2015.01.028Stagno, V., D. O. Ojwang, C. A.
McCammon, & Frost, D. J. (2013). The oxidation state of the mantle and the
extraction of carbon from Earth's interior, Nature, 493(7430), 84-88.
Stagno, V., Frost, D. J., McCammon, C. A., Mohseni, H., & Fei, Y. (2015). The oxygen
fugacity at which graphite or diamond forms from carbonate-bearing melts in
eclogitic rocks. Contributions to Mineralogy and Petrology, 169(2), 16.
doi:10.1007/s00410-015-1111-1Stamm, N., & Schmidt, M. W. (2017).
Asthenospheric kimberlites: Volatile contents and bulk compositions at 7 GPa, Earth
and Planetary Science Letters, 474, 309-321.
Staudigel, H. (2005), Hydrothermal alteration processes in the oceanic crust, in Treatise on
Geochemistry. The Crust., edited by H. D. Holland and K. K. Turekian, pp. 511-535,
Elsevier/Pergamon, Amsterdam.
Stiefenhofer, J., Viljoen, K. S., & Marsh, J. S. (1997). Petrology and geochemistry of
peridotite xenoliths from the Letlhakane kimberlites, Botswana. Contributions to
Mineralogy and Petrology, 127(1-2), 147-158. Stixrude, L., and C. Lithgow-
Bertelloni (2005), Thermodynamics of mantle mineralsI. Physical properties,
Geophysical Journal International, 162, 610-631.
©2020 American Geophysical Union. All rights reserved.
Stixrude, L., and C. Lithgow-Bertelloni (2011), Thermodynamics of mantle mineralsII.
Phase equilibria, Geophysical Journal International, 184, 1180-1213.
Sun, S.-s., & McDonough, W. F. (1989). Chemical and isotopc systematics of oceanic
basalts: implications for mantle composition and processes. In A. D. Saunders, M. J.
Norry, & D. S. a. M. J. Norr (Eds.), Magmatism in the Ocean Basins (pp. 313-345).
London: Geological Society.
Szumila, I., Trail, D., & Danielson, L. (2019). The effect of fO(2) on the diffusion of redox-
sensitive elements in haplobasaltic melt at 1 GPa and 1300 degrees C. Chemical
Geology, 512, 107-120. doi:10.1016/j.chemgeo.2019.02.018
Tappe, S., Brand, N. B., Stracke, A., van Acken, D., Liu, C. Z., Strauss, H., . . . Mitchell, R.
H. (2017). Plates or plumes in the origin of kimberlites: U/Pb perovskite and Sr-Nd-
Hf-Os-C-O isotope constraints from the Superior craton (Canada). Chemical Geology,
455, 57-83. doi:10.1016/j.chemgeo.2016.08.019
Tappe, S., Foley, S. F., Kjarsgaard, B. A., Romer, R. L., Heaman, L. M., Stracke, A., &
Jenner, G. A. (2008). Between carbonatite and lamproite - Diamondiferous Torngat
ultramafic lamprophyres formed by carbonate-fluxed melting of cratonic MARID-
type metasomes. Geochimica Et Cosmochimica Acta, 72(13), 3258-3286.
Tappe, S., Pearson, D. G., Kjarsgaard, B. A., Nowell, G., & Dowall, D. (2013). Mantle
transition zone input to kimberlite magmatism near a subduction zone: Origin of
anomalous Nd-Hf isotope systematics at Lac de Gras, Canada. Earth and Planetary
Science Letters, 371, 235-251. doi:10.1016/j.epsl.2013.03.039Taylor, L. A., H. J.
Milledge, G. P. Bulanova, G. A. Snyder, & Keller, R. A. (1998). Metasomatic
eclogitic diamond growth: Evidence from multiple diamond inclusions, International
Geology Review, 40(8), 663-676.
Taylor, L. A., G. A. Snyder, G. Crozaz, V. N. Sobolev, E. S. Yefimova, & Sobolev, N. V.
(1996). Eclogitic inclusions in diamonds: Evidence of complex mantle processes over
time, Earth and Planetary Science Letters, 142(3-4), 535-551.
Taylor, L. A., Keller, R. A., Snyder, G. A., Wang, W. Y., Carlson, W. D., Hauri, E. H., . . .
Bezborodov, S. M. (2000). Diamonds and their mineral inclusions, and what they tell
us: A detailed "pull-apart" of a diamondiferous eclogite. International Geology
Review, 42(11), 959-983.
Timmerman, S., Koornneef, J. M., Chinn, I. L., & Davies, G. R. (2017). Dated eclogitic
diamond growth zones reveal variable recycling of crustal carbon through time. Earth
and Planetary Science Letters, 463, 178-188. doi:10.1016/j.epsl.2017.02.001
Tumiati, S., Tiraboschi, C., Sverjensky, D. A., Pettke, T., Recchia, S., Ulmer, P., . . . Poli, S.
(2017). Silicate dissolution boosts the CO2 concentrations in subduction fluids.
Nature Communications, 8. doi:10.1038/s41467-017-00562-z
Viljoen, K. S., Smith, C. B., & Sharp, Z. D. (1996). Stable and radiogenic isotope study of
eclogite xenoliths from the Orapa kimberlite, Botswana. Chemical Geology, 131(1-4),
Viljoen, K. S., Schulze, D. J., & Quadling, A. G. (2005). Contrasting group I and group II
eclogite xenolith petrogenesis: Petrological, trace element and isotopic evidence from
eclogite, garnet-websterite and alkremite xenoliths in the Kaalvallei kimberlite, South
Africa. Journal of Petrology, 46(10), 2059-2090. doi:10.1093/petrology/egi047
Weiss, Y., Kessel, R., Griffin, W. L., Kiflawi, I., Klein-BenDavid, O., Bell, D. R., . . . Navon,
O. (2009). A new model for the evolution of diamond-forming fluids: Evidence from
microinclusion-bearing diamonds from Kankan, Guinea. Lithos, 112, 660-674.
Weiss, Y., McNeill, J., Pearson, D. G., Nowell, G. M., & Ottley, C. J. (2015). Highly saline
©2020 American Geophysical Union. All rights reserved.
fluids from a subducting slab as the source for fluid-rich diamonds. Nature,
524(7565), 339-342. doi:10.1038/nature14857
Xu, J. G., Zhang, D. Z., Fan, D. W., Dera, P. K., Shi, F., & Zhou, W. G. (2019).
Thermoelastic Properties of Eclogitic Garnets and Omphacites: Implications for Deep
Subduction of Oceanic Crust and Density Anomalies in the Upper Mantle.
Geophysical Research Letters, 46(1), 179-188. doi:10.1029/2018gl081170Yang, X.
Z., H. Keppler, C. McCammon, H. W. Ni, Q. K. Xia, and Q. C. Fan (2011), Effect of
water on the electrical conductivity of lower crustal clinopyroxene, Journal of
Geophysical Research-Solid Earth, 116.
Yaxley, G. M., Berry, A. J., Rosenthal, A., Woodland, A. B., & Paterson, D. (2017). Redox
preconditioning deep cratonic lithosphere for kimberlite genesis - evidence from the
central Slave Craton. Scientific Reports, 7. doi:10.1038/s41598-017-00049-3
Zedgenizov, D. A., Ragozin, A. L., Shatsky, V. S., & Griffin, W. L. (2018). Diamond
formation during metasomatism of mantle eclogite by chloride-carbonate melt.
Contributions to Mineralogy and Petrology, 173(10). doi:10.1007/s00410-018-1513-y
Zhao, C. C., & Yoshino, T. (2016). Electrical conductivity of mantle clinopyroxene as a
function of water content and its implication on electrical structure of uppermost
mantle. Earth and Planetary Science Letters, 447, 1-9. doi:10.1016/j.epsl.2016.04.028
©2020 American Geophysical Union. All rights reserved.
Figure 1. Criteria used to distinguish eclogite and pyroxenite classes. a. Mg# (molar
Mg/(Mg+Fetotal) as a function of Na/(Na+Fetotal) in clinopyroxene, separating pyroxenites
from eclogites. b. Total HREE content (HREE, summing abundances from Tb to Lu in
ppm) as a function of Eu/Eu* (chondrite-normalised Eu/(Sm*Gd)^0.5; chondrite of Sun &
McDonough, 1989) in garnet, separating gabbroic from non-gabbroic eclogites and c. Ca#
(molar Ca/(Ca+Mg+Fe+Mn)) as a function of Mg#, classifying non-gabbroic eclogites into
high-Mg, low-Mg and high-Ca classes. Samples are from the northern Slave craton (Jericho,
Muskox, Voyageur; Smart et al., 2009, 2014, 2017), central Slave craton (Diavik;
Schmidberger et al., 2007; this study), West African craton (Koidu; Barth et al., 2001, 2002;
Aulbach et al., 2019a), Zimbabwe craton (Orapa; Viljoen et al., 1996; Aulbach et al., 2016;
this study), and the Kaapvaal craton (Kimberley, Jacob et al., 2009, and Lace, Aulbach &
Viljoen, 2015, Aulbach et al., 2016).
©2020 American Geophysical Union. All rights reserved.
Figure 2. Effects of mantle metasomatism with LREE-enrichment, ascribed to
interaction with a UM carbonated melt similar to kimberlite, on mineral compositions and
trace-element distribution in mantle eclogites. a. Jadeite-content in clinopyroxene as a
function of grossular content (Ca#: 100Ca/(Ca+Fe+Mg+Mn) in garnet. Note the generally
low values observed for LREE-enriched bulk rocks (filled symbols). b.-d. Distribution of
various trace elements between clinopyroxene and garnet, expressed as Di =
clinopyroxeneCi/garnetCi, as a function of grossular content in garnet. With decreasing garnet Ca#,
all trace elements partition more strongly into clinopyroxene than into garnet; Di become
extreme for highly incompatible elements (Ce) in LREE-enriched samples with grossular-
poor garnet. e.-f. Sr and Ce concentration in garnet as a function of garnet Ca#, illustrating
that absolute concentrations of incompatible elements are as high or higher in unenriched
eclogites and pyroxenites than in LREE-enriched eclogites, due to decreased partitioning into
garnet at metasomatically-induced low Ca#. Filled symbols are for LREE-enriched samples
(NMORB-normalised Ce/Yb >1; NMORB from Gale et al., 2013) and open symbols for
unenriched samples. References as in Figure 1.
©2020 American Geophysical Union. All rights reserved.
Figure 3. Effects of mantle metasomatism with LREE-enrichment, ascribed to
interaction with a UM carbonated melt similar to kimberlite, on major-element systematics of
mantle eclogites. Concentrations of various oxides (wt%) as a function of MgO in
reconstructed eclogites and pyroxenites from various localities. Stippled rectangle highlights
LREE-enriched samples at generally high MgO contents. Error bar next to panel letters is for
a total of 10% modal variation (e.g. 45% clinopyroxene + 55% garnet to 55%
clinopyroxene+45% garnet). Dark grey field in b. shows expected range of MgO-CaO for
primitive mantle-derived melts that experienced various degrees of fractional crystallisation
(Herzberg & Asimow, 2008). Fields for high-Mg and low-Mg arc cumulates (Lee et al.,
2006) in panels b. and f., respectively, illustrate that mantle eclogites are unlikely to represent
delaminated lower crustal arc material. Square with long stipples in d. and e. highlights
samples from Lace inferred to have reacted with a sediment-derived melt in a subduction
mélange, which lowered their FeO and MgO and increased their Al2O3 content. The
correlation with SiO2 and anti-correlation with FeO at high MgO in c. and d., respectively, is
interpreted as a pyroxene-control line, reflecting the addition of FeO-poor and MgO-rich
magmatic pyroxene further characterised by low Na2O, moderately higher CaO and SiO2 and
higher Cr2O3 than the pre-metasomatic bulk rock. Filled symbols are for LREE-enriched
samples (NMORB-normalised Ce/Yb >1; NMORB from Gale et al., 2013) and open symbols
for unenriched samples. References as in Figure 1.
©2020 American Geophysical Union. All rights reserved.
Figure 4. Cryptic effects of mantle metasomatism with LREE-enrichment, ascribed to
interaction with a UM carbonated melt similar to kimberlite, on mantle eclogites. a.
NMORB-normalised Ce/Yb as a function of MgO (wt%) in reconstructed whole rocks,
illustrating that LREE-enrichment is linked to an increase in MgO. b.-f. Minor- (wt%) and
trace-element (ppm) relationships in clinopyroxene, as the main carrier of the metasomatic
signature (see Fig. 2). LREE-enrichment is accompanied by enrichment in Sr, Nb, Ce, Pb, Th
and U in most eclogite suites, whereas Y and Zr are only enriched in some. Enrichment in Y,
HREE and HFSE is ascribed to interaction with a low-volume mafic silicate melt. In contrast,
TiO2 is not higher in LREE-enriched than in unenriched samples. Copper and Li were diluted
by addition of low-Li pyroxene or were leached out of the rock, in particular Kimberley (Cu
is not reported) for which median Li is shown as solid line. Stippled lines show expected Li
concentrations in clinopyroxene in equilibrium with melt using high (D = 0.27, in equilibrium
with basanite; Adam & Green, 2006) and low (D = 0.04, in equilibrium with carbonatite
“carbon.”; Blundy & Dalton, 2000) experimental distribution coefficients (“D”). Filled
symbols are for LREE-enriched samples (NMORB-normalised Ce/Yb >1; NMORB from
Gale et al., 2013) and open symbols for unenriched samples. References as in Figure 1.
©2020 American Geophysical Union. All rights reserved.
Figure 5. Muting of crustal signatures in LREE-enriched eclogites, ascribed to
interaction with a UM carbonated melt similar to kimberlite (“kimberlite metasomatism”).
Crustal signatures are gauged as Eu/Eu* (chondrite-normalised Eu/(Sm*Gd)^0.5; chondrite
of Sun & McDonough, 1989) and 18O (‰) in garnet reflecting crustal differentiation and
seawater alteration in low-pressure protoliths, respectively, whereas metasomatic intensity is
gauged with Ce/Yb, Cr2O3 (wt%) and MgO (wt%). The range of Eu/Eu* and 18O is larger in
unenriched than in LREE-enriched eclogites. Canonical mantle 18O range from Mattey et al.
(1994). Filled symbols are for LREE-enriched samples (NMORB-normalised Ce/Yb >1;
NMORB from Gale et al., 2013) and open symbols for unenriched samples. References as in
Figure 1, plus Czas et al. (2018) for samples from Fort a la Corne (Sask craton), Huang et al.
(2012, 2014) for Type I and Type II eclogites and Radu et al. (2019) for Type II eclogites
from Roberts Victor (Kaapvaal).
©2020 American Geophysical Union. All rights reserved.
Figure 6. a. Eu/Eu* (chondrite-normalised Eu/(Sm*Gd)^0.5; CI chondrite of Sun &
McDonough, 1989) as a function of 87Sr/86Sr in clinopyroxene from various localities; for
samples with variable 87Sr/86Sr from Orapa, the lowest value is shown, with the assumption
that the least radiogenic value represents the least compromised one with respect to later
metasomatism. Eclogites with gabbroic (cumulate) protoliths are expected to have higher
Eu/Eu* than those with protoliths originating as residual melts, and to evolve to lower and
higher 87Sr/86Sr, respectively. Depending on this pre-history, metasomatism by young
asthenosphere-derived kimberlites, and by orangeites having an aged lithospheric metasome
component, may increase or decrease the original 87Sr/86Sr. Cr content (ppm) in garnet, as a
measure of metasomatic intensity, as a function of initial Nd isotopic composition
(143Nd/144Ndi) in eclogites from b. Lace (this work) and c. Koidu (Aulbach et al., 2019a).
Shown for comparison is range of 143Nd/144Ndi of the Lace orangeite (Howarth et al., 2011)
and of Kaapvaal kimberlites (sensu stricto; Le Roex et al., 2003). The 143Nd/144Nd of the
compositionally transitional Koidu kimberlite is unknown, but possibly approximated by the
unradiogenic Nd recorded in garnet from metasomatised high-Mg eclogite.
©2020 American Geophysical Union. All rights reserved.
Figure 7. Depth focussing of metasomatism with LREE-enrichment (“LREE-enr.”)
and with both LREE- and HREE-enrichement (“HREE-enr.”) gauged by NMORB-
normalised Ce/Yb (NMORB of Gale et al., 2013), Y (ppm) and Nb (ppm) in reconstructed
whole rocks as a function of pressure (GPa) obtained by iterative solution of temperature
(Krogh, 1988) with the regional geotherm (see section 3 and text S1), as parameterised from
Hasterok and Chapman (2011). Pressure range of cratonic mid-lithospheric discontinuities
(MLDs) in a. from Rader et al. (2015). As applies to all incompatible elements (with varying
sensitivity), Y contents decrease with increasing mantle potential temperature (TP) and melt
fraction during melt generation in the protolith stage and are excluded during accumulation,
whereas they increase with increasing degree of igneous differentiation of the crustal
protoliths. Partial melting upon subduction, modelled as melt extraction from rutile eclogite
with a melt fraction F of 0.3, increases Y (see also Fig. S3); estimates for various NMORB
from Sun and McDonough (1989) and Gale et al. (2013). Nb contents are minimum because
rutile, a frequent accessory mineral in eclogites known to control Nb abundances, has not
been considered in bulk rock reconstruction (see text for details). Uncertainty on the pressure
along a 38 mW/m2 geothermal gradient is 0.38 GPa propagated for a 60°C uncertainty at a
temperature of 1000°C (“prop. from temp.”). Filled symbols are for LREE-enriched samples
(NMORB-normalised Ce/Yb >1 and open symbols for unenriched samples. References as in
Figure 1.
©2020 American Geophysical Union. All rights reserved.
Figure 8. Variation of a.-b. density (rho, g/cm3), c.-d. shearwave velocity (VS, km/s)
and e.-f. hypothetical (average of upper and lower bound) bulk electrical conductivity (S/m)
with pressure (GPa) or depth (km) in eclogites; suites have been split into two panels for
clarity. Density and shear-wave velocity were calculated for eclogite mineral endmember
proportions weighted by modes using the same methods and data as in Garber et al. (2018)
[cf. Connolly, 2009; Stixrude & Lithgow-Bertelloni, 2005; 2011], but at 5% lower garnet
modes, 5% higher clinopyroxene modes, and at pressure-temperature estimates obtained as
outlined in the text. Electrical conductivity was calculated based on formulations in Jones et
al. (2009), Dai and Karato (2009) and Zhao and Yoshino (2016) (see text for further details).
To illustrate the effect of water and in accord with rare observational data from eclogites
(Huang et al., 2014), it is assumed that all garnet is dry and that clinopyroxene in LREE-
enriched eclogites, and in HREE-enriched eclogites from Kimberley, has equilibrated with a
kimberlite-like melt containing 3 wt% H2O, whereas clinopyroxene in unenriched eclogites
was assumed to contain 100 ppm H2O. The distribution coefficient clinopyroxene/meltDH depends
on Al2O3 in clinopyroxene and was calculated using the relationship reported in Aubaud et al.
©2020 American Geophysical Union. All rights reserved.
(2008). Conductivities are hypothetical values except for eclogites from Roberts Victor where
it was calculated for measured H2O contents in clinopyroxene. Shown for comparison in a.-d.
are results for average eclogite as well as low-Mg and high-Mg extremes from Garber et al.
(2018) for 38 mW/m2 as interpolated from their results for 35 and 40 mW/m2. In panels c.
and d. the average and bounds of average cratonic VS in the model of French and
Romanowicz (2014) is shown with black and grey long stipples, respectively. Uncertainties
related to a 10% modal variation (0.45 gt plus 0.55 clinopyroxene, as opposed to the 0.55 gt
plus 0.45 clinopyroxene used in bulk rock reconstruction) are shown as error bars.
Uncertainty on the pressure along a 38 mW/m2 geothermal gradient is 0.38 GPa propagated
for a 60°C uncertainty at a temperature of 1000°C (“prop. from temp.”). Model-dependent
differences between VS are ≤0.02 km/s (Garber et al., 2018). Solid and stippled black lines in
e. and f. are conductivities calculated by Garber et al. (2018) for average eclogite and
peridotite, respectively, for conductive geotherms corresponding to a surface heat flow of 35
(lower conductivity) and 40 mW/m2 (higher conductivity); black bars are conductivities
extracted from magnetotelluric studies in the Slave, Zimbabwe (Zim) and Kaapvaal cratons
as reported in Garber et al. (2018). The uncertainty arising from a 10% modal variation is not
discernible at the scale displayed. Filled symbols are for LREE-enriched samples (NMORB-
normalised Ce/Yb >1; NMORB from Gale et al., 2013) and open symbols for unenriched
samples. References as in Figure 1, plus Huang et al. (2014) for samples from Roberts Victor
and Czas et al. (2018) for samples from Fort a la Corne (Sask craton).
©2020 American Geophysical Union. All rights reserved.
Table 1. Salient compositional and physical characteristics of unenriched and metasomatised
cratonic eclogites and pyroxenites
... In contrast to most DI, the cratonic eclogite reservoir continued to evolve subsequent to diamond formation, mostly through interaction with fluids and kimberlite-like small-volume melts 9 . This type of metasomatism affected 20-40% of the eclogite reservoir and caused diamond destruction at most xenolith localities 21 , although it can occasionally be associated with abundant diamond growth (e.g., northern Slave craton 22 ). ...
... Data sources in Supplementary Data 1. 69 ). a Based on two-tailed t-test for null hypothesis that the means of two populations are equal, and imposing alpha = 0.05 (below which the null hypothesis is rejected), using either equal or unequal variances depending on F-test outcomes b Bulk rocks are reconstructed with 0.55 garnet, 0.45 clinopyroxene minus half the weight of rutile each (e.g., for 1 wt.% rutile 0.545 garnet, 0.445 clinopyroxene); for V reconstruction, estimated rutile modes and concentrations in rutile were considered. c Oxygen fugacities as reported or recalculated relative to the Fayalite-Magnetite-Quartz (FMQ) buffer using the eclogite oxybarometer of ref. 70 ; temperatures are derived using the thermometer of ref. 51 in iterative solution with regional conductive model geotherms (see ref. 21 ). ...
... This alone cannot account for the difference between DI and xenolithic garnet, and therefore crystal-chemical effects in addition to rutile-solubility effects are required. Higher CaO contents in garnet, here gauged by its molar Ca# (Ca/(Mg + Fe total + Ca + Mn)), facilitate incorporation of incompatible trace elements 21,25,27 . Consideration of xenoliths only, to minimise superposition of effects from higher average temperatures recorded for DI, reveals a significant negative correlation of garnet V abundance with Ca# for two of six sample suites (Table 2), which translates into a negative correlation of D(V) cpx-gt with garnet Ca# (Supplementary Fig. 1e-f). ...
Full-text available
Cratonic eclogite is the product of oceanic crust subduction into the subcontinental lithospheric mantle, and it also is a fertile diamond source rock. In contrast to matrix minerals in magma-borne xenoliths, inclusions in diamond are shielded from external fluids, retaining more pristine information on the state of the eclogite source at the time of encapsulation. Vanadium is a multi-valent element and a widely used elemental redox proxy. Here, we show that that xenolithic garnet has lower average V abundances than garnet inclusions. This partly reflects crystal-chemical controls, whereby higher average temperatures recorded by inclusions, accompanied by enhanced Na2O and TiO2 partitioning into garnet, facilitate V incorporation at the expense of clinopyroxene. Unexpectedly, although diamond formation is strongly linked to metasomatism and xenoliths remained open systems, V concentrations are similar for bulk eclogites reconstructed from inclusions and from xenoliths. This suggests an oxygen-conserving mechanism for eclogitic diamond formation, and implies that eclogite is an efficient system to buffer fO2 over aeons of lithospheric mantle modification by subduction-derived and other fluids. Billions of years ago, pieces of subducted oceanic crust were trapped as eclogites in the mantle keels beneath continents. Comparing eclogitic minerals protected inside and evolving outside of ancient diamonds reveals no change in redox state.
Full-text available
This work provides an introduction to: - A history of inclusion discovery in lithospheric diamonds; - The inclusion–diamond relationship; - Inclusion suites and parageneses; - Origin and evolution of subcratonic lithospheric mantle. It subsequently reviews, discusses and provides key observations on: - Major and trace element composition of peridotitic suite inclusions and comparison with cratonic peridotite xenolith minerals; - Diamond formation in eclogitic-pyroxenitic substrates.
Full-text available
Here we report the nitrogen characteristics and composition of high-density fluid (HDF) trapped in microinclusions in a suite of fibrous diamonds from the ~142 Ma Chidliak CH-7 kimberlite pipe, the Hall Peninsula, southern Baffin Island, Nunavut. Within these diamonds, we observe three populations based on the chemistry of the encapsulated HDFs, the diamond's nitrogen aggregation states, and the diamond color. ‘Chidliak C' diamonds contain highly silicic HDFs, have nitrogen in A- and C-centers (with 5–20% in C-centers), and a characteristic intense yellow color. ‘Chidliak A' diamonds contain silicic to low-Mg carbonatitic HDFs, carry nitrogen solely in A-centers, and are mostly colorless. A third population, ‘Chidliak B', has grey color and distinctive low-K2O silicic to low-Mg carbonatitic HDF compositions and overall smoother and less fractionated trace element pattern relative to ‘Chidliak C' and ‘Chidliak A' diamonds; they carry nitrogen in A- and B-centers (with ~15% in B-centers) and are characterized by a grey hue. An eclogitic paragenesis of all diamonds is evident by the HDF compositional variation as well as the presence of omphacitic clinopyroxene inclusions. The appearance of a diamond with A- and B-centers in its octahedral core and A- and C-centers in its coat suggests formation at two distinct events at a similar depth. Combined with pressure and mantle residence estimates based on nitrogen aggregation considerations, we argue that the three diamond populations formed at the relatively shallow region of the lithosphere (likely <180 km) during distinct metasomatic events in the North Atlantic Craton (NAC) since the Proterozoic. The youngest event by silicic HDFs took place close in time to kimberlite activity at 142–157 Ma, as evident by the preservation of nitrogen C-centers in ‘Chidliak C' diamonds. A link between this event and the mid-lithosphere discontinuity (MLD) in eclogitic portions of the cratonic lithosphere in Chidliak is plausible. The timing of ‘Chidliak A' diamonds formation by more carbonatitic HDFs is less well constrained, but can be related to Ca-rich metasomatism observed in local peridotite xenoliths and/or alkaline magmatism between 610 and 550 Ma. A possible link between the formation of ‘Chidliak B' diamonds and the timing of Mesoproterozoic olivine lamproite magmatism ca. 1400 Ma is suggested based on the HDF trace element composition and the aggregated nature of nitrogen in these diamonds. The nitrogen systematics and eclogitic source of the fibrous diamonds are comparable with those observed for previously studied gem-quality diamonds from Chidliak. We suggest that these similarities show a temporal connection and mutual crystallization of the two diamond types. This strengthens the involvement of HDFs in the formation of gem-quality diamonds.
Full-text available
The origin of the eclogites that reside in cratonic mantle roots has long been debated. In the classic Roberts Victor kimberlite locality in South Africa, the strongly contrasting textural and geochemical features of two types of eclogites have led to different genetic models. We studied a new suite of 63 eclogite xenoliths from the former Roberts Victor Mine. In addition to major- and trace-element compositions for all new samples, we determined 18O/16O for garnet from 34 eclogites. Based on geochemical and textural characteristics we identify a large suite of Type I eclogites (n = 53) consistent with previous interpretations that these rocks originate from metamorphosed basaltic-picritic lavas or gabbroic cumulates from oceanic crust, crystallised from melts of depleted MORB mantle. We identify a smaller set of Type II eclogites (n = 10) based on geochemical and textural similarity to eclogites in published literature. We infer their range to very low δ18O values combined with their varied, often very low Zr/Hf ratios and LREE-depleted nature to indicate a protolith origin via low-pressure clinopyroxene-bearing oceanic cumulates formed from melts that were more depleted in incompatible elements than N-MORB. These compositions are indicative of derivation from a residual mantle source that experienced preferential extraction of incompatible elements and fractionation of Zr-Hf during previous melting.
Reconstructed whole-rock and mineral major- and trace-element compositions, as well as new oxygen isotope data, for 22 mantle eclogite xenoliths from the Catoca pipe (Kasai Craton) were used to constrain their genesis and evolution. On the basis of mineralogical and major-element compositions, the Catoca eclogites can be divided into three groups: high-alumina (high-Al) (kyanite-bearing), low-magnesian (low-Mg#), and high-magnesian (high-Mg#) eclogites. The high-Al Catoca eclogites contain kyanite and corundum; high Al2O3 contents in rock-forming minerals; rare earth element (REE) patterns in garnets showing depleted LREEs, positive Eu anomalies (1.03–1.66), and near-flat HREEs; and high Sr contents in garnets and whole-rock REE compositions. All of these features point to a plagioclase-rich protolith (probably gabbro). Reconstructed whole-rock compositions (major elements, MREEs, HREEs, Li, V, Hf, Y, Zr, and Pb) and δ18O of 5.5–7.4‰ of the low-Mg# Catoca eclogites are in good agreement with the compositions of picrite basalts and average mid-ocean ridge basalt (MORB). The depleted LREEs and NMORB-normalised Nd/Yb values of 0.07–0.41 indicate that the degree of partial melting for the majority of the low-Mg# eclogites protolith was ≥30%. The narrow δ18O range of 5.5–7.4‰ near the ‘gabbro–basalt’ boundary (6‰) obtained for the high-Al and low-Mg# Catoca eclogites reflects the influence of subduction-related processes. This case shows that mantle eclogites represented by two different lithologies and originating from different protoliths — plagioclase-rich precursor, presumably gabbro (for high-Al eclogites), and basalt (low-Mg# eclogites) — can provide similar and overlapping δ18O signatures on account of the influence of subduction-related processes. Chemical compositions of the high-Mg# eclogites indicate a complicated petrogenesis, and textural signatures reveal recrystallisation. The presence of Nb-rich rutile (8–12 wt% of Nb2O5) enriched with HFSE (Zr/Hf of 72.6–75.6) and multiple trace-element signatures (including reconstructed whole-rock NMORB-normalised Ce/Yb of 3.9–10.6 and Sr/Y of 5.8–9.6, MgO contents of 15.7–17.9 wt%, and high Ba and Sr) provide strong evidence for deep metasomatic alteration. High Cr contents in clinopyroxene (800–3740 ppm), garnet (430–1400 ppm), and accessory rutile (700–2530 ppm), together with extremely low Li contents of 1.0–2.4 ppm in clinopyroxene, may indicate hybridisation of the eclogites with peridotite. Comparison of the chemical compositions (major and trace elements) of (1) unaltered fresh cores of coarse-grained garnets from the low-Mg# eclogites, (2) secondary garnet rims (ubiquitous in the low-Mg# eclogites), (3) proto-cores in the coarse-grained garnet (high-Mg# eclogites), and (4) homogeneous recrystallised fine-grained garnets (high-Mg# eclogites) suggests that the high-Mg# eclogites formed through recrystallisation of low-Mg# eclogite in the presence of an external fluid in the mantle. Four of the five high-Mg# samples show that mantle metasomatism inside the Kasai craton mantle beneath the Catoca pipe occurred at a depth range of 145–160 km (4.5–4.8 GPa).
The central region of the Kaapvaal craton is relatively understudied in terms of its lithospheric mantle architecture, but is commonly believed to be significantly impacted by post-Archean magmatism such as the ca. 2056 Ma Bushveld large igneous event. We investigate a collection of 17 eclogite xenoliths from the Cretaceous Palmietfontein kimberlite at the Western Limb of the Bushveld Complex for their mineralogical compositions (major and trace elements, plus Fe³⁺ contents), as well as stable oxygen and radiogenic Pb isotopic compositions to gain further insights into the nature and evolution of the central Kaapvaal cratonic mantle lithosphere. New U/Pb age determinations on mantle-derived zircon yield a magma emplacement age of ca. 75 Ma for the Palmietfontein Group-1 kimberlite, which means that the entrained eclogite xenoliths may record a protracted metasomatic history from the Proterozoic through to most of the Phanerozoic eon. Garnet δ¹⁸O values of up to 6.9‰ and positive Eu anomalies for the bulk rocks suggest seawater-altered oceanic crustal protoliths for the Palmietfontein eclogite xenolith suite, which is typical for the eclogitic components of the Kaapvaal root and other cratonic mantle sections worldwide. However, several features of the Palmietfontein eclogites are commonly not observed in other mantle-derived eclogite xenolith suites. Firstly, the samples studied yield relatively low equilibration pressures and temperatures between 2.7 and 4.5 GPa and 740–1064 °C, indicating a relatively shallow residence between 90 and 150 km depths. Secondly, euhedral coarse amphibole is present in several eclogite nodules where it is in equilibrium with ‘touching’ garnet, supporting eclogite residence within the amphibole stability field at uppermost lithospheric mantle conditions. Thirdly, primary omphacitic clinopyroxene is often overgrown by diopside, and is significantly enriched in incompatible trace elements. The clinopyroxene is also characterized by elevated ²⁰⁶Pb/²⁰⁴Pb of 17.28–19.20 and ²⁰⁷Pb/²⁰⁴ Pb of 15.51–16.27, and these Pb isotopic compositions overlap with those of Mesozoic Group-2 kimberlites from the Kaapvaal craton. Our results show that eclogites reside at ~85 km depth beneath the central Kaapvaal craton as part of a layer that corresponds to an approximately 50 km thick seismically-detected mid-lithospheric discontinuity. Mid-lithospheric discontinuities have been interpreted as metasomatic fronts formed by focussed crystallization of hydrous mineral phases from enriched volatile-bearing melts, and as such the strongly overprinted amphibole-bearing eclogite xenoliths from Palmietfontein may represent a physical expression of such seismically anomalous metasomatic layer at mid-lithospheric depth. Our Pb isotope data suggest that the focussed metasomatism can be attributed to volatile-rich melts reminiscent of potassic Group-2 kimberlites, which have been invoked in MARID-style metasomatic overprinting of the lower lithospheric mantle beneath the western Kaapvaal craton. However, the relatively low fO2 recorded by the Palmietfontein eclogites (minimum FMQ-4.5) suggests that the metasomatism at mid-lithospheric depth was less protracted compared to the more intensive and oxidizing metasomatism typically observed near the base of cratonic mantle roots. While it is possible that Proterozoic magmatic events were responsible for the focussed mid-lithospheric metasomatism of the Kaapvaal mantle, on the basis of the Pb isotope constraints the Palmietfontein eclogites were most likely overprinted during ca. 120 Ma Group-2 kimberlite magmatism.
Full-text available
Kimberlite-borne mantle eclogites represent an important diamond source rock. Although the origin and stability of diamond, as opposed to its low-pressure polymorph graphite, have been studied for decades, their relationship in rare natural samples where both polymorphs coexist remains poorly constrained. In order to shed new light on this issue, seven graphite-diamond-bearing eclogites from the kimberlite pipe Udachnaya, Siberian craton were comprehensively investigated with respect to their petrography, mineral chemical composition and omphacite 87Sr/86Sr, acquired in situ by laser-ablation multicollector ICPMS. The calculated P-T conditions for basaltic-group eclogites (Eu/Eu* < 1) correspond to a pressure range of 4.8 – 6.5 GPa and temperatures of 1060 – 1130 °C, while gabbroic eclogites with positive Eu- and Sr-anomalies have a smaller pressure variation (4.8 – 5.8 GPa), but a larger range in temperature (990 – 1260 °C). Reconstructed bulk compositions for gabbroic eclogites indicate an oceanic crustal origin for their protoliths, with accumulation of plagioclase and olivine ± clinopyroxene (gabbronorite or olivine gabbro). The protoliths of basaltic eclogites probably formed from the complementary residual melt. The presence of coesite and low Mg# in basaltic eclogites suggest that their LREE-depletion was the result of < 10 % partial melting during subsequent subduction and emplacement into the cratonic lithosphere. Extremely unradiogenic 87Sr/86Sr (0.70091 – 0.70186 for six of seven samples) not only provide new evidence for the Archean age (2.5 – 2.9 Ga) of Yakutian graphite-diamond-bearing eclogites and for formation of their protoliths in a depleted mantle source, but also suggest that they were not significantly metasomatically overprinted after their formation, despite their extended residence in the cratonic mantle lithosphere. The mineralogical and petrographic features indicate that the primary mineral association includes garnet, omphacite, ± coesite, ± kyanite, ± rutile, graphite, and diamond. Graphite occurs in the samples in the form of idiomorphic crystals (the longest dimensions being up 0.4 to 1 mm) in garnet and kyanite and extends beyond their grain boundaries. Diamonds occur as octahedral cubic transparent, slightly colored or bright-yellow crystals as big as 0.1 to 2 mm. Furthermore, idiomorphic and highly ordered graphite occurs as inclusions in diamond in four samples. The carbon isotope composition for diamond and graphite has a narrow range (-4 to -6.6 ‰) for both groups (gabbroic and basaltic), indicating a mantle source and limiting the role of subducted isotopically light biogenic carbon or reduction of isotopically heavy carbonate in diamond crystallization. Importantly, the presence of graphite and diamond inclusions in garnet, omphacite, and kyanite in three samples indicates a co-formation close in time to eclogitization. Combined, the petrographic and geochemical evidence suggests that both polymorphic carbon modifications can form in the diamond stability field, as also suggested by experiments and some natural examples, although the exact mechanism remains unresolved. Furthermore, this study provides natural evidence that graphite can be preserved (metastably) deep within the diamond stability field, without recrystallizing into diamond, for a long time – ≥ 2.5 Ga.
Very few zircon-bearing, kimberlite-hosted mantle eclogite xenoliths have been identified to date; however, the zircon they contain is crucial for our understanding of subcratonic lithospheric mantle evolution and eclogite genesis. In this study, we constrain the characteristics of zircon from mantle eclogite xenoliths based on existing mineralogical and geochemical data from zircons from different geological settings, and on the inferred origin of mantle eclogites. Given the likely origin and subsequent evolution of mantle eclogites, we infer that the xenoliths can contain zircons with magmatic, metamorphic and xenogenic (i.e. kimberlitic zircon) origins. Magmatic zircon can be inherited from low-pressure mafic oceanic crust precursors, or might form during direct crystallization of eclogites from primary mantle-derived melts at mantle pressures. Metamorphic zircon within mantle eclogites has a number of possible origins, ranging from low-pressure hydrothermal alteration of oceanic crustal protoliths to metasomatism related to kimberlite magmatism. This study outlines a possible approach for the identification of inherited magmatic zircon within subduction-related mantle eclogites as well as xenogenic kimberlitic zircon within all types of mantle eclogites. We demonstrate this approach using zircon grains from kimberlite-hosted eclogite xenoliths from the Kasai Craton, which reveals that most, if not all, of these zircons were most likely incorporated as a result of laboratory-based contamination.
Full-text available
Oxygen fugacity (ƒO2) is an intensive variable implicated in a range of processes that have shaped the Earth system, but there is controversy on the timing and rate of oxidation of the uppermost convecting mantle to its present ƒO2 around the fayalite-magnetite-quartz oxygen buffer. Here, we report Fe3+/ΣFe and ƒO2 for ancient eclogite xenoliths with oceanic crustal protoliths that sampled the coeval ambient convecting mantle. Using new and published data, we demonstrate that in these eclogites, two redox proxies, V/Sc and Fe3+/ΣFe, behave sympathetically, despite different responses of their protoliths to differentiation and post-formation degassing, seawater alteration, devolatilisation and partial melting, testifying to an unexpected robustness of Fe3+/ΣFe. Therefore, these processes, while causing significant scatter, did not completely obliterate the underlying convecting mantle signal. Considering only unmetasomatised samples with non-cumulate and little-differentiated protoliths, V/Sc and Fe3+/ΣFe in two Archaean eclogite suites are significantly lower than those of modern mid-ocean ridge basalts (MORB), while a third suite has ratios similar to modern MORB, indicating redox heterogeneity. Another major finding is the predominantly low though variable estimated ƒO2 of eclogite at mantle depths, which does not permit stabilisation of CO2-dominated fluids or pure carbonatite melts. Conversely, low-ƒO2 eclogite may have caused efficient reduction of CO2 in fluids and melts generated in other portions of ancient subducting slabs, consistent with eclogitic diamond formation ages, the disproportionate frequency of eclogitic diamonds relative to the subordinate abundance of eclogite in the mantle lithosphere and the general absence of carbonate in mantle eclogite. This indicates carbon recycling at least to depths of diamond stability and may have represented a significant pathway for carbon ingassing through time.
Full-text available
Here we present high‐resolution 2‐D coupled tectonic‐surface processes modeling of extensional basin formation. We focus on understanding feedbacks between erosion and deposition and tectonics during rift and passive margin formation. We test the combined effects of crustal rheology and varying surface process efficiency on structural style of rift and passive margin formation. The forward models presented here allow to identify the following four feedback relations between surface processes and tectonic deformation during rifted margin formation. (1) Erosion and deposition promote strain localization and enhance large offset asymmetric normal fault growth. (2) Progressive infill from proximal to more distal half grabens promotes the formation of synthetic sets of basinward dipping normal faults for intermediate crustal strength cases. (3) Sediment loading on top of undeformed crustal rafts in weak crust cases enhances middle and lower crustal flow resulting in sag basin subsidence. (4) Interaction of high sediment supply to the distal margin in very weak crust cases results in detachment‐based rollover sedimentary basins. Our models further show that erosion efficiency and drainage area provide a first‐order control on sediment supply during rifting where rift‐related topography is relatively quickly eroded. Long‐term sustained sediment supply to the rift basins requires elevated onshore drainage basins. We discuss similar variations in structural style observed in natural systems and compare them with the feedbacks identified here.
Full-text available
Eclogite is potentially an important constituent in local regions in the deep crust and upper mantle. The electrical conductivity of omphacite and garnet in eclogite has been measured at 1 GPa and 350–800 °C with pre-annealed OH-bearing samples. The conductivities were determined using a piston–cylinder apparatus and a Solartron-1260 Impedance/Gain Phase Analyser in the frequency range of 10⁶–1 Hz. The sample water contents show almost no change before and after the experimental runs. The conductivity of both omphacite and garnet increases with temperature, and the activation enthalpy is ~ 82 kJ/mol for omphacite and 90 kJ/mol for garnet, which is nearly independent of water content in each mineral. The conduction is probably dominated by protons, and for both minerals, the conductivity increases linearly with water content, with a water content exponent of ~ 1. These data are used to model the bulk conductivity of an eclogite with different water contents and modal compositions. In combination with reported data, the conductivity of the eclogite is similar to that of typical granulites above 600 °C, but is much larger than that of olivine, assuming small to moderate water contents. This would mean that the contribution of eclogites, if present, to the electrical structure of the deep continental crust cannot be easily separated from that of granulites, and that the regional enrichments of eclogites in the upper mantle may cause high electrical anomalies. The results also provide information for the electrical property of orogen-related thickened deep crust where eclogites may be locally abundant, e.g., in the Dabieshan region and the Tibet plateau. At mantle depths, eclogitized portions of subducted slabs are usually of very low conductivities as suggested by geophysical observations, implying small water contents in the constitutive omphacite and garnet and the limited ability of these minerals in recycling water into the deep mantle.
Full-text available
Here we present SIMS data for a suite of Zimmi sulphide-bearing diamonds that allow us to evaluate the origin and redox-controlled speciation of diamond-forming fluids for these Neoproterozoic eclogitic diamonds. Low δ ¹³ C values below −15‰ in three diamonds result from fluids that originated as carbon in the oceanic crust, and was recycled into the diamond-stable subcratonic lithospheric mantle beneath Zimmi during subduction. δ ¹³ C values between −6.7 and −8.3‰ in two diamonds are within the range for mantle-derived carbon and could reflect input from mantle fluids, serpentinised peridotite, or homogenised abiogenic and/or biogenic carbon (low δ ¹³ C values)and carbonates (high δ ¹³ C values)in the oceanic crust. Diamond formation processes in eclogitic assemblages are not well constrained and could occur through redox exchange reactions with the host rock, cooling/depressurisation of CHO fluids or during H 2 O-loss from CHO fluids. In one Zimmi diamond studied here, a core to rim trend of decreasing δ ¹³ C (−23.4 to −24.5‰)and decreasing [N]is indicative of formation from reduced CH 4 -bearing fluids. Unlike mixed CH 4 -CO 2 fluids near the water maximum, isochemical diamond precipitation from such reduced CHO fluids will only occur during depressurisation (ascent)and should not produce coherent fractionation trends in single diamonds that reside at constant depth (pressure). Furthermore, due to a low relative proportion of the total carbon in the fluid being precipitated, measurable carbon isotopic variations in diamond are not predicted in this model and therefore cannot be reconciled with the 1‰ internal core-to- rim variation. Consequently, this Zimmi eclogitic diamond showing a coherent trend in δ ¹³ C and [N]likely formed through oxidation of methane by the host eclogite, although the mineralogical evidence for this process is currently lacking.
Full-text available
Mantle eclogites are commonly accepted as evidence for ancient altered subducted oceanic crust preserved in the subcon-tinental lithospheric mantle (SCLM), yet the mechanism and extent of crustal recycling in the Archaean remains poorly constrained. In this study, we focus on the petrological and geochemical characteristics of 58 eclogite xenoliths from the Roberts Victor and Jagersfontein kimberlites, South Africa. Non-metasomatized samples preserved in the cratonic root have variable textures and comprise bimineralic (garnet (gt)-omphacite (cpx)), as well as kyanite (ky)-and corundum (cor)-bearing eclogites. The bimineralic samples were derived from a high-Mg variety, corresponding to depths of ~ 100-180 km, and a low-Mg variety corresponding to depths of ~ 180-250 km. The high-Al (ky-, cor-bearing) eclogites originated from the lowermost part of the cratonic root, and have the lowest REE abundances, and the most pronounced positive Eu and Sr anomalies. On the basis of the strong positive correlation between gt and cpx δ 18 O values (r 2 = 0.98), we argue that δ 18 O values are unaffected by mantle processes or exhumation. The cpx and gt are in oxygen isotope equilibrium over a wide range in δ 18 O values (e.g., 1.1-7.6‰ in garnet) with a bi-modal distribution (peaks at ~ 3.6 and ~ 6.4‰) with respect to mantle garnet values (5.1 ± 0.3‰). Reconstructed whole-rock major and trace element compositions (e.g., MgO variation with respect to Mg#, Al 2 O 3 , LREE/HREE) of bimineralic eclogites are consistent with their protolith being oceanic crust that crystallized from a picritic liquid, marked by variable degrees of partial melt extraction. Kyanite and corundum-bearing eclogites, however, have compositions consistent with a gabbroic and pyroxene-dominated protolith, respectively. The wide range in reconstructed whole-rock δ 18 O values is consistent with a broadly picritic to pyroxene-rich cumulative sequence of depleted oceanic crust, which underwent hydrothermal alteration at variable temperatures. The range in δ 18 O values extends significantly lower than that of present day oceanic crust and Cretaceous ophiolites, and this might be due to a combination of lower δ 18 O values of seawater in the Archaean or a higher temperature of seawater-oceanic crust interaction.
Full-text available
Synchrotron-based high-pressure/high-temperature single-crystal X-ray diffraction experiments to ~24 GPa and 700 K were conducted on eclogitic garnets (low-Fe: Prp28Alm38Grs33Sps1 and high-Fe: Prp14Alm62Grs19Adr3Sps2) and omphacites (low-Fe: Quad57Jd42Ae1 and high-Fe: Quad53Jd27Ae20), using an externally heated diamond anvil cell. Fitting the pressure-volume-temperature data to a third-order Birch-Murnaghan equation of state yields the thermoelastic parameters including bulk modulus (KT0), its pressure derivative (K′T0), temperature derivative ((∂KT/∂T)P), and thermal expansion coefficient (αT). The densities of the high-Fe and low-Fe eclogites were then modeled along typical geotherms of the normal mantle and the subducted oceanic crust to the transition zone depth (550 km). The metastable low-Fe eclogite could be a reason for the stagnant slabs within the upper range of the transition zone. Eclogite would be responsible for density anomalies within 100–200 km in the upper mantle of Asia.
Bimineralic eclogites and pyroxenites (n ¼ 75 6 accessory rutile, ilmenite, sulphide, apatite) from the Koidu kimberlite (West African Craton) were investigated for mineral major and trace elements and mineral Sr-Nd isotope compositions to constrain (1) the nature and age of their basaltic to pi-critic protoliths, and (2) the effect, timing and source of mantle metasomatism. Consistent with published work, samples are grouped into low-Mg eclogites with Mg# from 0Á49 to 0Á73 (median 0Á59; n ¼ 40) and high-Mg eclogites with Mg# from 0Á60 to 0Á88 (median 0Á75; n ¼ 14), plus pyroxen-ites [clinopyroxene Na/(Na þ Ca) <0Á2; n ¼ 8] and gabbroic eclogites and pyroxenite (Eu/Eu* of reconstructed bulk-rocks >1Á05; n ¼ 8), with five unclassifiable samples. Reconstructed low-Mg and gabbroic eclogites have major and trace element systematics (Eu/Eu*-heavy rare earth elements-Y) indicating crustal protolith crystallisation, confirming an origin as subducted oceanic crust. Their high FeO contents at MgO >$10 wt % require an Fe-rich source, the high melt productivity of which led to the formation of thicker crust, perhaps in a plateau-like setting. This is consistent with SiO 2-MgO relationships indicating differentiation at $0Á5 GPa. Unradiogenic Sr in some clinopyroxene (87 Sr/ 86 Sr of 0Á7010-0Á7015), combined with light rare earth element (LREE) depletion relative to normal mid-ocean ridge basalt (N-MORB) for the majority of samples (average N-MORB-normal-ised Nd/Yb of unmetasomatised samples ¼ 0Á51), suggests eclogitisation and partial melt loss in the Neoarchaean, possibly coeval with and parental to 2Á7 Ga overlying continental crust. Most reconstructed high-Mg eclogites and some pyroxenites formed by metasomatic overprinting of low-Mg eclogites and gabbroic eclogites, as indicated by the preservation of positive Eu anomalies in some samples, and by the Mg-poorer composition of included versus matrix minerals. Coupled enrichment in MgO, SiO 2 and Cr 2 O 3 and in incompatible elements (Sr, LREE, Pb, Th and U) is ascribed to metasomatism by a kimberlite-like, small-volume, carbonated ultramafic melt, mediated by addition of clinopyroxene from the melt (i.e. stealth metasomatism). Strontium-Nd isotope systematics suggest a Neoproterozoic age for this metasomatic event, possibly linked to Rodinia break-up, which facilitated intrusion of asthenospheric carbonated melts with an ocean island basalt-like 87 Sr/ 86 Sr i of $0Á7035. Cretaceous kimberlite magmatism (including Koidu), with more radiogenic 87 Sr/ 86 Sr ($0Á7065, intermediate between Kaapvaal kimberlites and orangeites), may have been partially sourced from associated Neoproterozoic metasomes. The presence of diamonds in low-Mg eclogites, but absence in high-Mg eclogites, indicates the diamond-destructive nature of this event. Nevertheless, the moderate proportion of affected eclogites ($35%) suggests preservation of a sizeable diamond-friendly mantle eclogite reservoir beneath Koidu.
Diffusivities for elements (P, Sc, V, Mn, Co, Zn, Cr, Nb, Mo, La, Ce, Pr, Sm, Eu, Gd, Ta, W) at 1300 °C and 1 GPa in basaltic melts were characterized across a range of oxygen fugacity (fO 2 ) conditions. Experiments were carried out using either a reducing (FMQ-3.1), intermediate (~FMQ-1.2) or oxidizing (FMQ + 6) fO 2 . For each fO 2 , three experiments were conducted for durations of 20, 40, and 80 min. For a given time series, changes in diffusivity are typically within 3 standard error at a single fO 2 . The magnitudes of the elemental diffusivities can be grouped into those of the High-field Strength Elements (HFSE), the Rare Earth Elements (REE), the transition elements, and P. Vanadium and Sc have diffusivities more similar to the REEs and HFSEs respectively, than the other transition elements. The best fits of diffusivities for P also suggest that the diffusivity of this element is more in line with those of the HFSEs. At oxidizing conditions, a fractionation of Nb from Ta with greater magnitude than that at the other oxygen fugacities is seen. Across oxygen fugacities explored here, Eu exhibits unique changes in diffusion. At more reducing conditions, the diffusivity of Eu increases relative to the neighboring REE elements Sm and Gd, with this effect most pronounced at FMQ-3.1 and present in experiments conducted at intermediate fO 2 conditions. This demonstrates that an Eu anomaly can be generated by diffusion alone. In oxidizing conditions, because Eu likely is present as mostly Eu ³⁺ , the signal vanishes as Eu diffusivity becomes similar to that of other trivalent REEs. There are small systematic changes in element diffusivities for both redox-sensitive and non-redox sensitive elements as fO 2 is varied. Averages of the 20, 40, and 80 min diffusivities for all elements done in the intermediate fO 2 experiments have the slowest diffusivities of the three oxygen fugacities explored. On average, the diffusivities of the entire contingent of elements studied from the more reducing (FMQ-3.1) conditions are faster than those from the intermediate fO 2 by about a factor of 1.5. The elemental diffusivities recovered from the oxidizing experiments are, on average, about ~2 times as fast of those recovered from the intermediate experiments. For elements fit in these experiments, an order of magnitude change in element diffusivities, even for redox-sensitive elements, is never seen over the range of oxygen fugacities explored at 1300 °C. These experiments demonstrate that oxygen fugacity can have an important effect on the diffusivity of certain redox-sensitive elements (e.g. Eu) and that fO 2 might play a role in element transport generally.
Most diamonds found in kimberlites show complex patterns of growth and dissolution (resorption) surface features. Populations of diamonds from within single kimberlite bodies commonly contain a large diversity of diamond surface forms, some of which are a result of dissolution in kimberlite magma and others are inherited from the mantle. Morphological studies of natural diamonds differentiated features produced during dissolution in kimberlite magma and during mantle metasomatism. The former features were experimentally reproduced at 1–3 GPa and used to infer the presence and composition of magmatic fluid in different kimberlites. However, the mantle-derived resorption features have not been reproduced experimentally and the composition and origins of their formative solvents are unknown. Here we report the results of diamond dissolution experiments conducted in a multi-anvil apparatus at 6 GPa and 1200 to 1500 °C in synthetic CaO–MgO–SiO2–CO2–H2O system. The experiments produced very different diamond resorption morphologies in COH fluid, in silicate-saturated fluid, and in silicate and carbonate melts. Dissolution in SiO2-free COH fluid developed rounded crystal forms with shallow negative trigons, striations and hillocks, which are commonly observed on natural diamonds and are similar in 6 GPa and in 1–3 GPa experiments. However, silicate-saturated fluid produced very different resorption features that are rarely observed on natural diamonds. This result confirms that natural, SiO2-poor fluid-induced resorption develops under the comparatively low-pressures of kimberlite ascent, because at mantle pressures the high content of SiO2 in fluids would produce features like those from the silicate-saturated experiments. Comparison of the experimental products from this study to natural diamond resorption features from the literature suggests that natural diamonds show no record of dissolution by fluids during mantle metasomatism. Diamond resorption morphologies developed in experiments with silicate–carbonate melts closely resemble many of the mantle-derived resorption features of natural diamonds, whose diversity can result from variable SiO2 concentration in carbonatitic melts and temperature variation. The experimental results imply that metasomatism by fluids does not dissolve diamond, whereas metasomatism by melts is diamond-destructive. The repetitive growth-dissolution patterns of natural diamonds could be due to diamond growth from fluids in harzburgitic lithologies followed by its dissolution in partial melts.
The isopycnicity hypothesis states that the lithospheric mantle of ancient platforms has a unique composition such that high density due to low lithosphere temperature is nearly compensated by low-density composition of old cratonic mantle. This hypothesis is supported by petrological studies of mantle xenoliths hosted in kimberlite magmas. However, the representativeness of the kimberlite sampling may be questioned, given that any type of magmatism is atypical for stable regions. We use EGM2008 gravity data to examine the density structure of the Siberian lithospheric mantle, which we compare with independent constraints based on free-board analysis. We find that in the Siberian craton, geochemically studied kimberlite-hosted xenoliths sample exclusively those parts of the mantle where the isopycnic condition is satisfied, while the pristine lithospheric mantle, which has not been affected by magmatism, has a significantly lower density than required by isopycnicity. This discovery allows us to conclude that our knowledge on the composition of cratonic mantle is incomplete and that it is biased by kimberlite sampling which provides a deceptive basis for the isopycnicity hypothesis.