Kimberlite Melt Research Articles

Origin of Group 1A kimberlites: Fluid-saturated melting experiments at 45–55 kbar

The composition and amounts of fluids strongly affect phase relations near the liquidus of Group 1A kimberlites. Under CO 2-saturated conditions olivine disappears from the liquidus at 45 kbar and orthopyroxene, garnet and magnesite crystallize very close to the liquidus at 55 kbar. The addition of water stabilizes olivine to some extent, but garnet, magnesite and orthopyroxene remain near the liquidus. The Mg mole fraction of liquidus silicates is very high (> 0.94 for orthopyroxene), similar to that of strongly depleted garnet harzburgites. Clinopyroxene is absent in near-liquidus assemblages at all conditions investigated. Fluid-saturated kimberlite melts may be in equilibrium with carbonated garnet harzburgite at 60 kbar and with a CO 2-dominated fluid. Garnets obtained in the experiments are poor in Cr and rich in Ti compared with garnets from depleted harzburgite xenoliths from kimberlites. The direct melting of a depleted, magnesite-bearing garnet harzburgite is thus not a plausible mechanism of kimberlite formation. Based on our data and experimental results up to 160 kbar by Ringwood et al. [1] on a similar composition, a model is proposed for the formation of Group 1A kimberlites through the interaction of deep-seated melts from the asthenosphere or transition zone with strongly depleted harzburgites at the base of the continental lithosphere. Such interaction results in the enrichment in CO 2 in the initial fluid-undersaturated melts up to a saturation level. The separation of CO 2-rich fluid triggers kimberlite eruption before complete equilibration with the enclosing peridotite. This model explains many features of kimberlite geology and geochemistry, such as rapid ascent, small volumes and isotopic signatures similar to ocean island basalts.

Comparative geochemistry of West African kimberlites: Evidence for a micaceous kimberlite endmember of sublithospheric origin

A suite of largely unaltered, aphanitic, mica-bearing hypabyssal kimberlites from the Koidu kimberlite complex of the West African Craton have been investigated to determine their geochemical affinity relative to Group I (nonmicaceous) and Group II (micaceous) kimberlites of southern Africa. Comparison is made with altered kimberlites from Liberia, other West African and global kimberlites. Based on major element oxides, the Koidu kimberlites, though mica-bearing, show closest compositional similarity with the Group IA kimberlites of southern Africa. Based on major and trace elements, the Koidu kimberlites show an unusual geochemical signature. This signature is similar to that of the distinctive, micaceous Aries kimberlite of northwest Australia, and includes high Nb U (most samples > 46), Ce Sr(>0.4) , Ta Hf(>2) , and Nb Zr(>1) ratios and low P 2O 5/Ce ∗10 4(<27), Ba Rb(<32) , and U Th(<0.2) ratios compared with Group I kimberlites. Koidu kimberlites can be readily discriminated from Group II kimberlites by their higher Ti K(>0.4) and Mb La(>1) ratios and lower Ba Nb(<10) and Pb Ce(<0.06) ratios. The compositions of Liberian kimberlites are leached of mobile incompatible elements, but least affected samples show affinity to Group I. Guinea kimberlites appear to be of two types: one having affinity with Group IA and the other, micaceous variety, having affinity with the Aries kimberlite. Kimberlites with an Aries geochemical signature appear to exist on some other cratons, e.g., the Kundelungu kimberlites (Zaire) and two mica-bearing Group I kimberlites (S. Africa). The Koidu kimberlites exhibit compositionally-dependent isotopic heterogeneity though initial ϵ Nd and ϵ Sr values are broadly asthenospheric (i.e., near bulk earth) similar to Group I and Aries. A compositional spectrum appears to exist between nonmicaceous Group I kimberlites through mica-bearing Koidu kimberlites to extreme endmembers of the Aries type. This spectrum can be modelled as partial melts of heterogeneous peridotite sources which incorporate a potassic, high-Nb source component. The component may represent a fluoro-K-richterite-bearing peridotite residue derived by melt extraction from subduction-zone metasomatized peridotites. Such materials may have been trapped, together with former oceanic lithosphere, in the Transition Zone of the mantle. In response to lower mantle upwelling, diapiric uprise accompanying reduced volatile degassing of the kimberlite source may occur. Because of differences in oxidation potential across the 400 km discontinuity, reactions in the ascending diapir will lead to redox melting and ultimately segregation of the kimberlite melt at the base of the thermal boundary layer (P ~ 13 GPa) separating the subcratonic lithosphere from the convective mantle.

Crystallization of diamond from a silicate melt of kimberlite composition in high-pressure and high-temperature experiments

In high-pressure and high-temperature experiments (1800- 2200 °C and 7.0-7.7 GPa), diamond crystallized and grew in a volatile-rich silicate melt of kimberlite composition. This diamond has well- developed {111} faces, and its morphologic characteristics resemble those of natural diamond but differ from those of synthetic diamond grown from metallic solvent-catalysts. The kimberlite melt has a strong solvent-catalytic effect on diamond formation, supporting the view that some natural diamonds crystallized from volatile-rich melts in the upper mantle.

The petrogenesis of group 2 ultrapotassic kimberlites from Finsch Mine, South Africa

Major, trace element and radiogenic isotope results are presented for a suite of hypabyssal kimberlites from a single pipe, at the Finsch Mine, South Africa. These are Group 2 kimberlites characterised by abundant phlogopite ± serpentine ± diopside; they are ultrabasic (SiO 2 < 42 wt.%%) and ultrapotassic (K 2O/Na 2O > 6.9) igneous rocks, they exhibit a wide range in major element chemistry with SiO 2 = 27.6−41.9 wt. % and MgO = 10.4−33.4 wt. %. ( 87 Sr/ 86 Sr) i =0.7089 to 0.7106, ϵ Nd is −6.2 to −9.7 and they have unradiogenic ( 207Pb/ 204Pb) i contents which ensure that they plot below the Pb-ore growth curve. They have high incompatible and compatible element contents, a striking positive array between Y and Nb which indicates that garnet was not involved in the within suite differentiation processes, and a negative trend between K/Nb and Nb contents which suggests that phlogopite was involved. In addition, some elements exhibit an unexpected order of relative incompatibility for different trace elements which suggests that the intra-kimberlite variations are not primarily due to variations in the degree of partial melting. The effects of fractional crystallization are difficult to establish because for the most part they have been masked by the entrainment of 50–60% mantle peridotite. Thus, the Finsch kimberlites are interpreted as mixtures of a melt component and entrained garnet peridotite, with no evidence for significant contamination with crustal material. The melt component was characterised by high incompatible element contents, which require both very small degrees of partial melting, and source regions with higher incompatible element contents than depleted or primitive mantle. Since the melt component was the principal source of incompatible elements in the kimberlite magma, the enriched Nd, Sr and Pb isotope ratios of the kimberlite are characteristic of the melt source region. The melt fractions were therefore derived from ancient, trace elements enriched portions of the upper mantle, most probably situated within the sub-continental mantle lithosphere, and different from the low 87Sr/ 86Sr garnet peridotite xenoliths found at Finsch. Within the sub-continental mantle lithosphere old, incompatible element enriched source regions for the kimberlite melt fraction are inferred to have been overlain by depleted mantle material which became entrained in the kimberlite magma.

Argon isotope and halogen chemistry of phlogopite from South African kimberlites: a combined step-heating, laser probe, electron microprobe and TEM study

Argon isotopic analyses were undertaken on phlogopite from the Swartruggens kimberlite dyke and a Premier kimberlite (1200 Ma) peridotite xenolith. Groundmass phlogopite from Swartruggens yields a plateau age of 145.0± 0.4 Ma, consistent with previous age determinations. Phlogopite phenocrysts from Swartruggens and macrocrysts from the Premier xenolith yield complex age spectra, with anomalously old ages, attributed to incorporation of excess radiogenic argon. Laser probe analyses on single phlogopite grains reveal systematic zonations in excess Ar and CI concentrations across (001) cleavage surfaces. One Premier macrocryst exhibits ages in excess of 2.3 Ga and CI levels of 1600 ppm at its centre. These values decrease systematically to 1.2 Ga and 1300 ppm Cl along grain margins. Similar results were obtained from a single Swartruggens phenocryst, which exhibits a range in values of 340—800 Ma and 300—1300 ppm Cl. A second Swartruggens phenocryst is characterised by smaller variations in age (140–230 Ma) and CI content (390–470 ppm ). Fluorine concentrations, determined by electron microprobe, are relatively constant or increase slightly towards grain edges. The laser probe profiles cannot be reconciled with the step-heating results, probably due to phlogopite degradation during invacuo furnace heating. Transport of Ar and CI in kimberlitic phlogopite appears to be dominated by radial diffusion (cylindrical geometry). The variety of laser probe profiles obtained suggests that Ar and Cl diffusion is governed by factors such as lattice diffusion, diffusion anisotropy, and structural defects, which reduce the effective radii of diffusion and may impart a component of pipe diffusion. It is suggested that the xenolith phlogopite entrapped excess 40Ar and halogens in the mantle lithosphere, in response to elevated Ar and halogen fluid pressures. Swartruggens phenocrysts appear to have crystallised from a volatile-rich kimberlite melt. Subsequent magma devolatilisation prior to emplacement reduced Ar partial pressure and CI content. Possible reasons for enhanced F levels after devolatilisation include increased F solubility in the kimberlite melt, extraction of F from infiltrating hydrothermal fluids and local heterogeneities in fluid composition. The final distributions of Ar, Cl and F in kimberlitic phlogopite are variably dependent on several parameters, including local fluid composition, timing of melt devolatilisation, diffusion/ exchange mechanisms, and mineral composition.

Geochemistry of mafic volcanic rocks from the Lake Kivu (Zaire and Rwanda) section of the western branch of the African Rift

The volcanic rocks from the Lake Kivu volcanic province (southeastern Zaire and Rwanda) are part of the western branch of the east African Rift system and were emplaced from early Miocene to sub-Recent times. The lavas are mainly of basaltic composition and range from quartz-normative tholeiites through alkali basalts and basanites to melilite nephelinites. They were derived from a heterogeneous upper mantle source by variable degrees of melting. The inferred source compositions are highly variable in K, Rb, Ti, V, P, Th and Hf but relatively uniform in light REE, Sr and Zr. Compared with primordial mantle, the source beneath the western branch of the East African Rift is strongly enriched in highly incompatible elements. The enrichment and heterogeneity probably were caused by interaction of the mantle with a kimberlite melt.

Lamproites and other potassium-rich igneous rocks: a review of their occurrence, mineralogy and geochemistry

Summary In this paper the geological occurrence, geochemistry and mineralogy of ultrapotassic (K 2 O/Na 2 O &gt; 3 (molar ratio)) and perpotassic (K 2 O/Al 2 O 3 &gt; 1 (molar ratio)) igneous rocks, especially lamproites, are reviewed and discussed in the context of compositionally-similar mantle-derived melts. Lamproites are K- and Mg-rich igneous rocks (typically K 2 O &gt; 5 wt.%, MgO &gt; 5 wt.%) which possess an exotic and diagnostic mineralogy and geochemistry. Known lamproites occur in 21 major suites or localities in continental regions with a variety of geological and tectonic environments; they range in age from the early Proterozoic dykes at Holsteinsborg, W Greenland, and Chelima, India, and Precambrian pipe at Argyle, W Australia, to the Middle Pleistocene flows and the Recent volcanics of the Leucite Hills, Wyoming, and Gaussberg, Antarctica, respectively. Intrusive and extrusive forms of lamproites include flows, a variety of pyroclastics (welded tuffs, piperno, air-fall tuffs, volcanic breccias etc.), cinder cones, dykes, sills and diatremes. Whereas kimberlite diatremes tend to be carrot shaped, the shape of olivine lamproite diatremes approximates a sherbet-glass. The recent discovery of diamondiferous lamproties of large volumetric proportion in the E and W Kimberleys, NW Australia, and the reclassification of the diamondiferous micaceous peridotite at Prairie Creek, Arkansas, as a lamproite substantiate their economic importance. The 21 lamproite suites considered here tend to be localized marginal to continental craton cores in areas that overlie fossil Benioff zones, in contrast with the general occurrence of kimberlites more interior to continental cratons. The petrographic diversity of lamproites has historically hindered the development of a concise and universal classification and nomenclature. Lamproites are distinguished from kimberlites and alkali basalts (and lamprophyres) in terms of mineralogy, mineral chemistry, geochemistry and volcanic extrusive character. Relative to kimberlites, lamproites are enriched in K, Si, Ti, Al, Rb, Sr, Zr and Ba and depleted in CO 2 , Ca, Mg, Fe, Ni, Co and Cr. Lamproites are characterized by the general presence of phlogopite, diopside, leucite and K-richterite, occasional glass, olivine, sanidine, priderite, perovskite, wadeite, apatite and chrome spinel, and very rare ilmenite. Lamproite amphiboles, diopsides and phlogopites are distinctly depleted in Al 2 O 3 relative to those of nearly all other igneous rocks. Lamproite magmas are produced by the partial melting of old refractory mantle peridotite (approaching a dunite or harzburgite in mineralogy) that was enriched in K-bearing and other incompatible-element-enriched phases, such as phlogopite and apatite, most probably as a result of some metasomatic event which occurred prior to melting. In contrast with alkali basalt and kimberlite melts which are apparently produced from the partial melting of a CO 2 -enriched mantle periodtite (i.e. a source with a relatively high CO 2 /H 2 O ratio), water is the key volatile species involved with lamproite petrogenesis (source with a low CO 2 /H 2 O ratio).

Groundmass oxide minerals in the Koidu kimberlite dikes, Sierra Leone, West Africa

Petrology and oxide mineral chemistry are presented on 5 kimberlite dikes that are classified into three groups: (1) one dike is highly carbonated and highly oxidized (> MH) and is characterised by chlorite+Mn-titanomagnetite+rutile+hematite (after chlorite)+maghemite (after titanomagnetite), with ilmenite and perovskite being absent; (2) three dikes are typified by atoll-textured spinels +phlogopite+euhedral Mn-picroilmenite, of intermediate oxidation state (WM-FMQ) with coexisting deuterically serpentinized olivine+Ni-Fe alloys and magnetite; (3) the remaining dike records an early crystallization event under very low (≤ WM) redox conditions that precipitated anionic-deficient spinels and a Mg-Ti-Cr-wustite-type phase, followed by late stage more oxidizing (=FMQ) Mnpicroilmenite. Spinels are complexly zoned and crystallization trends among the dikes are diverse, underscoring the fact that no single compositional trend, or evolutionary sequence is typical of kimberlites. Ilmenites are euhedral, and criteria for groundmass crystallization are established. Extraordinarily high MnO (max 17 wt%) contents and high geikielite (62 mole%) concentrations expand the ilmenite field typically assigned to that of kimberlites. Zirconium, Nb and Cr are present in concentrations of 0.5–3 wt% (as oxides) in ilmenite. These highly incompatible elements, along with Mn, are concentrated in late stage melt fractions. The high pyrophanite contents, which are more typical of silicic alkali suites, are accompanied by phlogopite in the Koidu dikes. Objective evaluations of kimberlite-carbonatite relations, as outlined in the literature, cannot be made based on the oxide mineral group. Much of the compositional data for oxides in kimberlites are on mantle-derived xenolith suites and are not from oxides derived from the crystallization of kimberlitic melts. Assessments of the fO2's of kimberlites have considerable potential in evaluating diamond survival through redox reactions. Manganese-rich (+Nb, Zr, Cr) ilmenites are typical of many kimberlites and should be considered in the suite of index minerals employed in prospecting.

Oxidation states of the upper mantle recorded by megacryst ilmenite in kimberlite and type A and B spinel lherzolites

The intrinsic oxygen fugacities of homogeneous, inclusion-free, megacryst ilmenites from the Frank Smith, Excelsior, Sekameng and Mukorob kimberlite pipes in southern Africa, and the alnoitic breccia in the Solomon Islands have been determined. Similar measurements have been made of the type A and B spinel peridotites from San Carlos in Arizona. The type A peridotites are characterised by oxygen fugacities close to the iron-wustite buffer, similar to those of equivalent peridotite specimens from other continental and island arc environments. In strong contrast, the type B peridotites and all of the ilmenite megacrysts range between the oxygen fugacities defined by the nickelnickel oxide and fayalite-magnetite-quartz buffers. A close relationship between type B peridotites, oxidized metasomatizing fluids in the upper mantle and oxidized, silicaundersaturated magma types is suggested. It is unlikely that a solid elemental carbon phase can be an equilibrium crystallization product of kimberlite magmas if the ilmenite megacrysts represent the redox state of kimberlite melts. The ultimate source of the oxidizing fluids and the development of such a wide dispersion (>4 orders of magnitude) in oxygen fugacities of the upper mantle is not clear, but may involve recycled lithosphere, fluids from the lower mantle or result from the relatively rapid diffusion of H2, compared with other potential volatile species, in the mantle.

Instability of garnet from the mantle: Glass as evidence of metasomatic melting

Instability of garnets is manifest by incongruent melting in the kimberlite of Fayette County, Pennsylvania. Quenching of this reaction has resulted in the formation of alkali-rich Ne- and Qz-normative glasses, spinel, aluminous Ca-rich and Ca-poor pyroxenes, and olivine. The crystalline products have a range of metastable textures and compositions; these represent quenching of the reaction at various stages of its completion during a progressive increase in cooling-rate consequent upon the final rise of the kimberlite through the upper mantle and lower crust. The melting was not isochemical; addition of alkalis and probably volatiles are required to balance the overall reaction. These metasomatic reactants may have their origin in the host kimberlitic melt or may have infiltrated from the adjacent mantle. Similarities of some of the features described herein with other reported kelyphite occurrences suggest that incongruent melting of garnet may be more commonplace than heretofore appreciated in “kelyphite” formation.

The significance of groundmass ilmenite and megacryst ilmenite in kimberlites

Criteria are suggested for distinguishing xenocrystic ilmenites from those indigenous to the host kimberlite. For instance, in contrast to groundmass grains, ilmenite xenocrysts usually are larger, have reaction rims of leucoxene and perovskite, exhibit strong magnesium enrichment outward, and sometimes have exsolution lamellae and deformation features. Most of the abundant ilmenite macrocrysts found in kimberlite appear to have been phenocrysts in a crystal mush unrelated to kimberlite. On the other hand, kimberlitic groundmass ilmenite is rare, but consistently more magnesian than the cores of macrocrysts. Strong Mg-enrichment patterns evident in the ilmenite macrocrysts probably developed during their attempt to equilibrate with the more magnesian, fractionating kimberlitic liquid. The hypothesis of extensive reaction of ilmenite with kimberlite melt/ fluid has implications with regard to the following: (1) the degree of differentiation of kimberlite melts; (2) the genesis of mantle megacrysts; (3) the reactivity of kimberlite; and (4) the usefulness of groundmass ilmenite as a petrogenetic indicator.

Mineralogical-geochemical features of zircons from kimberlites and problems of their origin

The present rounded aspect of zircons from kimberlites has resulted from the effect of brittle deformation, plastic deformations, and intense solution. The geochemical features of the zircons (distribution of uranium, thorium, and rare earth elements) exclude the possibility of their crystallization in the kimberlitic melt. It was concluded that the zircons in the kimberlites are of xenogenic origin. The most likely parent rocks for the zircons could have been certain high-temperature peridotites, corresponding to a restite after melting out of the kimberlitic magma from mantle rocks of lherzolitic composition.