J.W. Harris

Brine inclusions in diamonds: a new upper mantle fluid

Micro-inclusions in cloudy diamonds from Koffiefontein consist of three main types: silicates, carbonates and brine inclusions. The silicates belong to either the eclogitic or the peridotitic paragenesis and both are associated with carbonates and brine. The brine composition is roughly (K,Na)8(Ca,Fe,Mg)4SiO(CO3)4Cl10(H2O)28–44. Average mass proportions are about 30–42% water, 19–22% chlorine, 14–17% sodium and potassium, 22–25% Fe–Ca–Mg–carbonates, and 3–4% silica. The brine composition is distinct from that of fluids trapped in fibrous diamonds mainly by its high chlorine and low silica content. The close association of carbon-bearing brine, silicate minerals and diamonds suggests that such brines are important for diamond growth in both eclogitic and peridotitic environments. The similarity of brine composition in both environments may indicate that diamonds of both suites grew in a single event.

Diamond genesis, mantle fractionations and mantle nitrogen content: a study of δ13C–N concentrations in diamonds

A compilation of more than 1200 δ13C–N data from well-characterised diamonds show a correlation of the maximum diamond nitrogen content (i.e. a limit sector) with δ13C over the full diamond δ13C range (i.e. more than 30‰). Diamonds with low δ13C values are characterised by rather low N contents (∼0 ppm at δ13C<−30‰), whereas diamonds with high δ13C have more variable nitrogen contents, with a much higher upper limit (∼3500 ppm at δ13C=−4.5‰). This correlation defines a concave trend that is therefore incompatible with a mixing relationship, such as would be produced by the admixture of subducted and primordial components. The limit sector more likely reflects the evolution of mantle melts (or fluids) during differentiation. Nitrogen uptake is seen as a kinetic process, depending mostly on the diamond rate of growth; at a given δ13C value, as a result of slow growth conditions, diamonds with nitrogen contents lower than the maximum value are interpreted as having fractionated the N/C ratio relative to their growth medium. The limit sector is applicable to every diamond paragenesis (peridotitic, eclogitic and fibrous) suggesting that every diamond type may derive from a similar isotopic source. Assuming a mantle δ13C value of −4.5‰, we deduce that the initial C/N ratio of mantle melts (i.e. the diamond growth medium) from which diamonds crystallise ranges between 200 and 500, which is surprisingly similar to that of mid-ocean ridge basalts. Therefore, in spite of their different context and age, it appears that subcontinental and oceanic mantles give samples with similar δ13C, δ15N and C/N, suggesting an overall homogeneity of volatiles within these parts of the Earth since the Archaean. Diamonds also demonstrate that carbon and nitrogen do not behave similarly during the evolution of the diamond growth medium. Accordingly, mantle nitrogen concentration cannot be deduced in a simple way. If N behaved as an incompatible element during partial melting, a mantle nitrogen concentration of about 2 ppm could be expected, provided that the mantle carbon content is about 400 ppm. However, from several lines of evidence presented in this study, nitrogen is not regarded as a totally incompatible element, and a higher mantle nitrogen concentration (perhaps up to 40 ppm) is preferred.

Archean subduction recorded by Re-Os isotopes in eclogitic sulfide inclusions in Kimberley diamonds

Eclogitic sulfide minerals encapsulated in diamonds originating from the deepest part of the continental mantle keel beneath the Kaapvaal craton, southern Africa, and brought to the surface by the Kimberley kimberlites, show low Ni and Os contents and high Re/Os ratios characteristic of a basaltic protolith. The sulfide inclusions with the lowest Os contents give late Archean single grain absolute ages while those with higher Os contents yield a well-constrained 2.9 Ga isochron age and radiogenic initial Os isotope composition (γOs=+45). This indicates a significant time gap between basaltic precursor generation and eclogitic diamond crystallization, consistent with extended residence in a near-surface environment prior to subduction associated with accretion of the Kimberley block to the rest of the craton and subsequent diamond formation. These results suggest that subduction-related crustal recycling was already a viable process during continent formation in the middle Archean and may have been implicated in eclogitic diamond formation ever since.

Sulphide inclusions in diamonds from the Koffiefontein kimberlite, S Africa: constraints on diamond ages and mantle Re–Os systematics

Re–Os isotope compositions of syngenetic sulphide inclusions in both eclogite suite (E-type) and peridotite suite (P-type) parageneses in diamonds from the Koffiefontein mine, South Africa have been analysed using a technique capable of analysing single inclusion grains, or, in some cases multiple inclusions from the same diamonds. Sulphide inclusion Ni contents broadly correlate with Os abundances in that low-Ni (6.8–8.7% Ni), E-type sulphides have 4.7 to 189 ppb Os whereas the two high-Ni (25%), P-type sulphides have 5986 and 6097 ppb Os. Two P-type sulphides from the same diamond define the first mineral isochron obtained for a single diamond which has an age of 69±30 Ma with chondritic initial 187Os/188Os. This indicates that the sulphides, and hence the host diamond, crystallised close to the time of kimberlite emplacement (90 Ma), in the Mesozoic. This is supported by Pb isotopic measurements of a fragment from one of the sulphides, together with the absence of significant Type IaB nitrogen aggregation in the host diamond lattice. E-type sulphide inclusions have radiogenic Os isotopic compositions, 187Os/188Os 0.346 to 2.28, and Re–Os model ages from 1.1 to 2.9 Ga. They define an array on a Re–Os isochron diagram that may be interpreted as defining a single period of E-type sulphide growth at 1.05±0.12 Ga, with an elevated initial 187Os/188Os. Alternatively, two episodes of sulphide crystallisation, from a chondritic reservoir, may be invoked in the Archaean and in the Proterozoic. The results for both P- and E-type diamonds point to a spectrum of diamond crystallisation ages. High contents of both Re and Os, and the similarity of Re/Os ratios of sulphide inclusions in diamonds to whole rock eclogite and peridotite xenoliths indicate that small amounts of sulphides can dominate the mantle budget of both these elements during melting. Recent addition to the lithospheric mantle of high-Os material similar to that from which the P-type sulphides crystallised may explain the variable, sometimes young Os model ages seen in whole rock xenolith Re–Os data.

Subduction-related diamonds? — The evidence for a mantle-derived origin from coupled δ13C–δ15N determinations

Abstract For the first time, δ 13 C – δ 15 N of more than 150 diamonds of known paragenesis, originating from different localities are reported. These coupled δ 13 C – δ 15 N determinations argue against a direct formation from subducted biogenic carbon for most of these diamonds. Our data confirm the apparent isotopic disequilibrium of nitrogen between the internal (mantle) and external (atmosphere+crust) reservoirs of the Earth, and thus support the heterogeneous accretion model of the Earth inferred from earlier fibrous diamond studies. We suggest that non-fibrous diamonds have recorded the evolution of the nitrogen isotopic composition starting from a δ 15 N -value lower than −25‰ to the present δ 15 N -value around −5‰. This isotopic composition change of mantle nitrogen was built up by large recycling of nitrogen during the early history of the Earth.

Archean emplacement of eclogitic components into the lithospheric mantle during formation of the Kaapvaal Craton

Re-Os data for selected whole-rock eclogites and eclogitic sulfide inclusions in diamonds from seven kimberlites on the Kaapvaal-Zimbabwe cratons yield ages around 3 Ga. The presence of such old eclogitic components in the subcontinental lithospheric mantle is more common than previously thought. Overlap of these ages with those of many cratonic peridotites demonstrates that incorporation of eclogitic materials accompanied or closely followed cratonic lithosphere stabilization. Many of these eclogitic components have fractionated C, N, O, and S isotopic compositions which could have been created at low-temperatures, implying that some style of Archean subduction involving emplacement of oceanic lithosphere was part of craton keel development.

Raman barometry of diamond formation

Abstract Pressures and temperatures of the diamond source region are commonly estimated using chemical equilibria between coexisting mineral inclusions. Here we present another type of geobarometer, based on determination of the internal pressure in olivine inclusions and the stresses in the surrounding diamond. Using Raman spectroscopy, pressures of 0.13 to 0.65 GPa were measured inside olivine inclusions in three diamonds from the Udachnaya mine in Siberia. Stresses in the diamond surrounding the inclusions indicated similar pressures (0.11–0.41 GPa). Nitrogen concentration and aggregation state in two of the diamonds yielded mantle residence temperatures of ∼1200°C. Using this temperature and the bulk moduli and thermal expansion of olivine and diamond, we calculated source pressures of 4.4–5.2 GPa. We also derived a linear approximation for the general dependence of the source pressure ( P 0 , GPa) on source temperature ( T 0 , °C) and the measured internal pressure in the inclusion ( P i ): P 0 =(3.259×10 −4 P i +3.285×10 −3 ) T 0 +0.9246 P i +0.319. Raman barometry may be applied to other inclusions in diamonds or other inclusion–host systems. If combined with IR determination of the mantle residence temperature of the diamond, it allows estimation of the pressure at the source based on a non-destructive examination of a single diamond containing a single inclusion.

Noble gas and halogen geochemistry of mantle fluids: comparison of African and Canadian diamonds

Abstract Volatile-bearing fluids in diamond have been characterised using extension of the 40Ar-39Ar technique to simultaneously measure noble gas isotopes, halogens (Cl, Br and I), K and U. Samples investigated include opaque cubic and fibrous diamonds from the North West Territories, Canada, the Democratic Republic of Congo, (DRC, formerly Zaire), and Jwaneng, Botswana. These are compared with results obtained from metasomatised mantle xenoliths from Bultfontein, South Africa. Diamonds and xenoliths show a narrow range of 40Ar∗/Cl values between 506–1347 × 10−6 molar (M) with mean values that overlap for African diamonds and xenoliths (831 ± 218 × 10−6 M) and Canadian diamonds (965 ± 273 × 10−6 M). These values are consistent with the estimated MORB value and support the presence of a widespread 40Ar and Cl-rich fluid in the mantle. Canadian diamonds have high and variable halogen ratios, with Br/Cl = 1.3–63.0 × 10−3 M and I/Cl = 9.8–1703.5 × 10−6 M. In contrast, African diamonds, have less variable Br/Cl = 1.0–2.0 × 10−4 M and I/Cl = 13.6−176.4 × 10−6 M (with most I/Cl between 20–70 × 10−6 M). Fluids in diamonds from DRC and Botswana, have Ar and halogen compositions close to those estimated for the present-day MORB source. The concentrations of noble gases (Ar, Kr and Xe), halogens, K and U in Canadian coated stones are 10–30 times higher than in African coated stones probably due to a higher inclusion population density. The large variation in halogen ratios measured in Canadian diamonds is the first evidence for significant halogen fractionation in the mantle. The Br/Cl ratios are notably above the range reported for crustal fluids. The high halogen ratios in Canadian diamonds are consistent with crystallisation of a Cl-bearing mineral, possibly involving apatite, from a fluid with starting composition similar to that in African diamonds.

SULFIDE INCLUSION CHEMISTRY AND CARBON ISOTOPES OF AFRICAN DIAMONDS

Significant differences in the composition of sulfide mineral inclusions among diamond suites from Koffiefontein, Orapa, Premier, Roberts Victor, Jagersfontein, Sierra Leone, Star, and Mwadui have been found. The mode of the Ni content of the monosulfide (mss) inclusions lies between 8 and 10 wt%, i.e., between the means for mss from Siberian diamonds with inclusion of the eclogitic (3 wt% Ni) and peridotitic (23 wt% Ni) paragenesis. Considering the NiFe ratios of the diamond mss inclusions and mantle olivines, together with experimental and naturally observed NiFe distribution coefficients, we conclude that less than 20% of the mss inclusions of the African diamonds (mostly from Koffiefontein) could have been in chemical equilibrium with mantle olivine. This observation is in sharp contrast with the reported relative abundance of silicate inclusions in Koffiefontein diamonds (93% peridotitic, 7% eclogitic) and lends support to the proposal that a separate sulfide diamond paragenesis should be recognized. The δ 13C distributions of sulfide containing diamonds differs among kimberlites, however, for each kimberlite sulfide and silicate inclusion containing diamonds cover the same δ 13C range. Sulfides with high Ni concentrations can occur in diamonds of low as well as high 13C content. The current observations, in conjunction with other chemical properties of diamonds suggest that fluid reactions rather than silica melt equilibria may be important in diamond formation. A dominance of fluid processes would have significant implications for the interpretation of the chemical and geochronological record of diamond inclusions.

Depth-related carbon isotope and nitrogen concentration variability in the mantle below the Orapa kimberlite, Botswana, Africa

Abstract Cubic diamonds from the Orapa kimberlite have a very restricted δ 13 C range with a mean of −5.98 ± 0.57‰ , a relatively high and constant nitrogen content 897 ± 171 ppm and low nitrogen aggregation state. They are thought to have been derived from shallow mantle depth. Diamonds with peridotitic inclusions (P-Type) range in δ 13 C between−4.2 and − 18.9‰ with a major mode between −5 and−7‰. Their nitrogen content is low and state of aggregation high. Pressure-temperature ( P - T ) equilibration conditions estimated for their inclusions are close to the wet peridotite solidus. Eclogitic inclusions in Orapa diamonds (E-Type) indicate higher P - T equilibration conditions than P-Type minerals. The δ 13 C values of their hosts vary from −2.6 to−18.0%. with a major mode between −13 and −15‰. The nitrogen content of E-type diamonds is significantly higher than that of P-Type diamonds. Comparison of the mineral and carbon isotopic composition of Orapa diamond eclogites with those of E-Type diamonds indicates that only part of the E-Type diamonds could have been derived from a physical breakup of diamond eclogites. 13 C-depleted E-Type diamonds and eclogite xenoliths with low δ 13 C diamonds were formed in a similar and restricted P - T range. An inclusion mineral paragenesis with compositions transitional between those of peridotitic and eclogitic minerals (websteritic, W-Type) has been recognized. It is characterized by a unique combination of a relatively high chrome and Fe content. The δ 13 C values of W-Type diamonds lie, with one exception (−6.9%.), in the range of −15.2 to −22.4‰. The nitrogen content of these diamonds is significantly lower than that of E-Type and is indistinguishable from that of P-Type diamonds. The P - T equilibration conditions estimated for them are similar to those of the peridotitic paragenesis. The collective data indicate a higher frequency of occurrence of 13 C-depleted diamonds, a decrease in the N content of diamonds, and an increase in the N aggregation state with increasing mantle depth. The decline in N content is not monotonic however, because P-Type diamonds tend to have lower N contents than E-Type diamonds. The combined high chrome and Fe content of the websteritic diamond inclusions are very difficult to reconcile with a subduction origin of the low δ 13 C host and indicate that 13 C-depleted C may be a primary mantle feature.