Origin of sub-lithospheric diamonds from the Juina-5 kimberlite (Brazil): constraints from carbon isotopes and inclusion compositions

Abstract

interpreted as evidence of diamond growth from abiotic organic carbon added to the oceanic crust during hydrothermal alteration. The bulk isotopic composition of the oceanic crust, carbonates plus organics, is equal to the composition of mantle carbon (−5 ‰), and we suggest that recycling/mixing of subducted material will replenish this reservoir over geological time. Several exposed, syngenetic inclusions have bulk compositions consistent with former eclogitic magnesium silicate perovskite, calcium silicate perovskite and NAL or CF phases that have reequilibrated during their exhumation to the surface. There are multiple occurrences of majoritic garnet with pyroxene exsolution, coesite with and without kyanite exsolution, clinopyroxene, Fe or Fe-carbide and sulphide minerals alongside single occurrences of olivine and ferropericlase. As a group, the inclusions have eclogitic affinity and provide evidence for diamond formation at pressures extending to Earth's deep transition zone and possibly the lower mantle. It is observed that the major element composition of inclusions and isotopic compositions of host Juina-5 diamonds are not correlated. The diamond and inclusion compositions are intimately related to subducted material and record a polybaric growth history across a depth interval stretching from the lower mantle to the base of the lithosphere. It is suggested that the interaction of slabderived melts and mantle material combined with subsequent upward transport in channelised networks or a buoyant diapir explains the formation of Juina-5 diamonds. We conclude that these samples, despite originating at great mantle depths, do not provide direct information about the ambient mantle, instead, providing a snapshot of the Earth's deep carbon cycle. Keywords Diamonds · Sub-lithospheric mantle · Carbon cycle · SubductionAbstract Forty-one diamonds sourced from the Juina-5 kimberlite pipe in Southern Brazil, which contain optically identifiable inclusions, have been studied using an integrated approach. The diamonds contain <20 ppm nitrogen (N) that is fully aggregated as B centres. Internal structures in several diamonds revealed using cathodoluminescence (CL) are unlike those normally observed in lithospheric samples. The majority of the diamonds are composed of isotopically light carbon, and the collection has a unimodal distribution heavily skewed towards δ 13 C ~ −25 ‰. Individual diamonds can display large carbon isotope heterogeneity of up to ~15 ‰ and predominantly have isotopically lighter cores displaying blue CL, and heavier rims with green CL. The light carbon isotopic compositions are Communicated by O. Müntener. Electronic supplementary materialThe online version of this article

Stable isotope evidence for crustal recycling as recorded by superdeep diamonds

“…All lines of evidence point to a related, sub-lithospheric, origin for the diamonds studied. The diamonds have δ 13 C values from -26.4 ± 0.1 to -2.8 ± 0.1 ‰ and all measurements are consistent with previously published data for these and other samples from the same localities , Thomson et al, 2014. Mineral inclusion δ 18 O values range from 7.5 ± 0.3 to 12.9 ± 0.3 ‰ and are anticorrelated with δ 13 C of the host diamonds (Fig.…”

“…Our model is also compatible with all other observations of melting of carbonated subducted crust and the mineralogy and compositions of inclusions in superdeep diamonds, including the scarcity of carbide and metal (Bulanova et al 2010, Thomson et al 2014). Moreover, a study of the coupled behaviour of C and N isotopes in Juina diamonds similarly found evidence for a mixing relationship between crustal and mantle components (Palot et al, 2012).…”

“…Mixing C whilst preserving the crustal δ 18 O could be achieved by ingress of a mantle-derived H 2 O-and CO 2 /carbonate-free fluid into the slab: these chemical constraints rule out all C-bearing fluids except CH 4 . Although methane-rich fluids have been associated with diamond formation (Thomassot et al, 2007), they seem improbable for our suite of samples because of the existence of carbonate inclusions (Bulanova et al, 2010;Thomson et al, 2014) that would not be stable with the significantly more reducing conditions required for methane to predominate in the fluid.…”

“…Although the zonation varies (Bulanova et al 2010;Thomson et al 2014), nearly all the diamonds that contain very light carbon (-22 to -27 ‰) are either unzoned or have a light core and heavier rim. This suggests initial diamond growth from a fluid dominated by the crustal-derived carbon component, and progressive dilution of crustal carbon in the fluid with mantle carbon as the mixing process and diamond growth progress together.…”

“…and Collier-4 kimberlites; Bulanova et al, 2010;Walter et al, 2011;Thomson et al, 2014 and the Machado River alluvial deposit; Bulanova et al, 2008). We use these data to test the hypothesis that many superdeep diamonds are formed from recycled material.…”