Elsevier

Geochimica et Cosmochimica Acta

Volume 257, 15 July 2019, Pages 266-283
Geochimica et Cosmochimica Acta

Diamond-forming media through time – Trace element and noble gas systematics of diamonds formed over 3 billion years of Earth’s history

https://doi.org/10.1016/j.gca.2019.05.014Get rights and content

Abstract

Ten individual gem-quality monocrystalline diamonds of known peridotite/eclogite paragenesis from Southern Africa (Koffiefontein, Letlhakane, Orapa) were studied for trace element concentrations and He and Ar abundances and isotopic compositions. In addition, two samples, consisting of pooled fragments of gem-quality peridotitic diamonds from Finsch and DeBeers Pool respectively, were analysed for noble gases. Previous studies (Richardson et al., 1984; Pearson et al., 1998; Gress et al., 2017; Timmerman et al., 2017) provided age constraints of 0.09, 1.0–1.1, 1.7, 2.3, and 3.2–3.4 Ga on mineral inclusions in the studied diamonds, allowing us to study trace elements and noble gases over 3 Gyr of geological time. Concentrations of trace elements in the diamonds are very low – a few hundred ppt to several tens of ppbs – and are likely dependent on the amount of sub-micron inclusions present. Trace element patterns and trace element/3He ratios of the studied monocrystalline diamonds are similar to those in fibrous diamonds, suggesting that trace elements and stable noble gas isotopes reside within the same locations in diamond and track the same processes that are reflected in the trace element patterns. We cannot discern any temporal differences in these geochemical tracers, suggesting that the processes generating them have been occurring over at least the past 2.3 Ga. 3He/4He ratios decrease and 4He and 40Ar* contents increase with increasing age of peridotitic and some eclogitic diamonds, showing the importance of in-situ radiogenic 4He and 40Ar ingrowth by the decay of U-Th-Sm and K respectively. For most gem-quality monocrystalline diamonds, uncertainties in the 3He/4He evolution of the continental lithospheric mantle combined with large analytical uncertainties and possible spatial variability in U-Th-Sm concentrations limit our ability to provide estimates of diamond formation ages using 4He ingrowth. However, the limited observed 4He ingrowth (low U + Th/3He) together with a R/Ra value of 5.3 for peridotitic diamond K306 is comparable to the present-day sub-continental lithospheric mantle value and supports the young diamond formation age found by Re-Os dating of sulphides in the same diamond by Pearson et al. (1998). After correction for in-situ radiogenic 4He produced since diamond formation a large variation in 3He/4He remains in ∼1 Ga old eclogitic diamonds that is suggested to result from the variable influence of subducted altered oceanic crust that has low 3He/4He. Hence, the 3He/4He isotope tracer supports an origin of the diamond-forming fluids from recycled oceanic crust for eclogitic diamonds, as indicated by other geochemical proxies.

Introduction

Rare macroscopic mineral inclusions in gem-quality (clear) monocrystalline diamonds (hereafter GMC diamonds), such as clinopyroxenes and garnets, have been used to determine the P-T conditions, growth environment and formation ages of diamonds (e.g., Pearson and Shirey, 1999, Stachel and Harris, 2008). Complimentary to this information, sub-micron fluid inclusions in fibrous diamonds directly provide information on the origin of the diamond-forming fluid (Navon et al., 1988). Abundant fluid inclusions in fibrous diamonds have made it possible to determine trace element concentrations by Instrumental Neutron Activation Analysis (INAA; Schrauder et al., 1996) and via online laser ablation- Inductively Coupled Plasma Mass Spectrometry (LA-ICP-MS) techniques (Rege et al., 2005, Rege et al., 2010) on small pits (<100 µm in diameter). Trace element data for fluid-rich growth zones of diamonds show two main different trace element patterns; ribbed (elevated Ba-U-Th-Light Rare Earth Elements, depleted Nb-Ta) and planed (flat from Cs to Pr, without large anomalies) patterns (terminology of Weiss et al., 2013), which are observed in all fluid end-members (Schrauder et al., 1996, Rege et al., 2005, Tomlinson et al., 2005, Zedgenizov et al., 2007, Weiss et al., 2008, Tomlinson and Müller, 2009, Weiss et al., 2009, Klein-BenDavid et al., 2010, Rege et al., 2010, Weiss et al., 2011, Klein-BenDavid et al., 2014).

Fluid inclusions within GMC diamonds are scarce (Jablon and Navon, 2016). Rege et al. (2010) showed that although cloudy monocrystalline diamonds sometimes have trace element concentrations above the detection limit of “online” laser ablation ICP-MS, clear diamonds, without abundant fluid inclusions, have concentrations too low to permit quantitative measurement via this method. McNeill et al. (2009) developed an “offline” laser ablation and solution mode ICP-MS technique to sample larger quantities of GMC diamond, concentrating the analytes by ablations lasting between 20 minutes and several hours, such that analyte levels were above the limit of quantification by solution-mode ICP-MS. This approach also avoids the problems inherent to the lack of a matrix-matched reference material for direct ablation diamond analysis. The data available using the off-line ablation approach (Rege et al., 2010, Melton et al., 2012, Krebs et al., 2019) suggest that the trace element patterns of GMC diamonds are similar to those of fibrous diamonds, showing LREE enrichment over HREE, negative Sr anomalies, enriched Th, U, and both ribbed and planed patterns. In contrast, most GMC diamonds analysed by Melton et al. (2012) showed trace element patterns different to those of fibrous diamonds and the authors suggested that trace elements in these GMC diamonds may be present in ‘complex and variable mixtures of mineral inclusions’.

Like trace elements, noble gases are probably primarily located in fluid inclusions (e.g., Johnson et al., 2000, Broadley et al., 2018) and noble gas isotopic compositions and elemental abundances may thus possibly be related to trace element ratios and contents. As with trace elements, noble gas data for GMC diamonds are scarce. For a better understanding of the structure, volatile cycles, and evolution of the mantle, samples of different ages need to be investigated for both trace elements and noble gases (as 4He and 40Ar are decay products of the trace elements U-Th-Sm and K). The present-day noble gas compositions of the convecting upper mantle and a deeper mantle source contaminated by subducted material have been characterised by studying fresh basaltic glasses from mid-ocean-ridge (MORBs) and ocean islands (OIBs) (e.g., McDougall and Honda, 1998). Diamonds, on the other hand, could shed light on the noble gas composition of the sub-continental lithospheric mantle in the past. Diamond is an ideal mineral to study noble gases over time, as diamonds have a wide range in formation ages (0.07–3.5 Ga; Pearson and Shirey, 1999, Shirey et al., 2013), are physically and chemically inert, and have extremely low He diffusion rates (4He ∼ 4 × 10−21 cm2/s; Shelkov et al., 1998). Thus, diamonds can preserve trapped noble gas compositions over long timescales. Noble gases analysed from polycrystalline diamonds with minimum age constraints that are greater than or equal to the kimberlite eruption age, suggest heterogeneity in the mantle or evolution of the composition over time (Honda et al., 2011). However, no GMC diamonds with better age constraints have been studied for noble gases so far.

In this study, we present trace element and noble gas data for ten Southern-African GMC diamonds. Trace element abundances were determined using offline laser ablation followed by ICP-MS, He-Ne-Ar data was obtained utilising a high temperature furnace for graphitisation. The combined trace element and He-Ar data can provide new insights in the location of trace elements and noble gases in gem diamond, U-Th-Sm/He dating, diffusion and fractionation between trace elements and noble gases. The ten samples are of known parageneses (E-type n = 9, P-type n = 1) and have Re-Os or Sm-Nd mineral inclusion ages that range from 2.3 Ga to 0.09 Ga (Pearson et al., 1998, Gress et al., 2017, Timmerman et al., 2017). In addition, fragments of diamonds from Finsch and DeBeers Pool were measured for noble gases. These samples are remnants of many peridotitic monocrystalline diamonds that were broken open to extract inclusions (∼500 garnets for Finsch and ∼600 garnets for DeBeers Pool) for Sm-Nd and Rb-Sr dating, and yielded inclusion isochron ages of 3.2–3.3 Ga (Richardson et al., 1984). The wide range in mineral inclusion ages of the GMC diamonds in this study provides the unique opportunity to investigate the temporal changes of trace elements and He and Ar isotope compositions in the sub-continental lithospheric mantle.

Section snippets

Methods

Ten GMC diamonds from Koffiefontein (n = 3), Orapa (n = 4) and Letlhakane (n = 3) were analysed for trace elements and noble gases. The diamonds were selected based on homogeneous carbon isotope values and nitrogen concentrations and aggregation states. Descriptions and available N data are provided in Pearson et al., 1998, Timmerman et al., 2017. Carbon isotope values for the Koffiefontein diamonds were measured in this study, using the standard set-up for in-situ C analyses by SHRIMP as

Trace elements

Concentrations of trace elements in the studied GMC diamonds are extremely low, ranging from a few ppt up to 10 s of ppb, except for a few hundred ppb for Zr and for La and Nd in the Orapa samples (Table 3). Due to the low concentrations, the uncertainties on the element concentrations are large. The element Ce was not present above detection limit (blank plus 3σ) and is therefore not plotted or shown. In the Orapa samples the La/Nd and Nd/Pr ratios are not chondritic and appear to be

Constraining the location of trace elements and noble gases in GMC diamonds

The Orapa diamonds have on average lower trace element concentrations in their garnet and clinopyroxene mineral inclusions than the mineral inclusions in Letlhakane diamonds (Timmerman et al., 2017). In contrast, the bulk diamonds (lattice + micro-inclusions) of the same Orapa samples have higher concentrations than bulk diamonds of the Letlhakane samples. Therefore, the concentrations in the diamond are possibly highly dependent on the amount of micro-inclusions and/or number of twinning

Conclusions

Changes in the trace element and noble gas compositions in GMC diamonds over >3 billion years of Earth’s history were investigated in previously dated diamonds (3.4–0.09 Ga; 3P-type and 9 E-type diamonds), to track changes in gem-diamond forming fluid compositions and source regions through geological time. This study shows that trace element patterns of GMC diamonds are similar to fibrous diamonds and that the processes that produce variations in fluid trace element systematics, from planed

Acknowledgements

We thank Yannick Bussweiler for help with the laser, Sarah Woodland for assistance with the ICP-MS and Xiaodong Zhang for help with the Helix-MC Plus. Diamond samples from Botswana were provided by Prof. Dr. GR Davies and Dr. IL Chinn through the Debswana Diamond Company and the Finsch, DeBeers Pool, and Koffiefontein samples were provided by Prof. Dr. JW Harris through the De Beers Consolidated Mines. Both companies are members of the De Beers Group of Companies. Without this support this

References (85)

  • L. Johnson et al.

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

    Geochim. Cosmochim. Acta

    (2000)
  • M.A. Kendrick et al.

    Subduction zone fluxes of halogens and noble gases in seafloor and forearc serpentinites

    Earth Planet. Sci. Lett.

    (2013)
  • O. Klein-BenDavid et al.

    Mixed fluid sources involved in diamond growth constrained by Sr–Nd–Pb–C–N isotopes and trace elements

    Earth Planet. Sci. Lett.

    (2010)
  • O. Klein-BenDavid et al.

    The sources and time-integrated evolution of diamond-forming fluids – trace elements and isotopic evidence

    Geochim. Cosmochim. Acta

    (2014)
  • M. Krebs et al.

    A common parentage-low abundance trace element data of gem diamonds reveals similar fluids to fibrous diamonds

    Lithos

    (2019)
  • M.D. Kurz et al.

    Helium isotopic variability within single diamonds from the Orapa kimberlite pipe

    Earth Planet. Sci. Lett.

    (1987)
  • J.-Y. Lee et al.

    A redetermination of the isotopic abundances of atmospheric Ar

    Geochim. Cosmochim. Acta

    (2006)
  • K. Li et al.

    Diamonds isotope compositions indicate altered igneous oceanic crust dominates deep carbon recycling

    Earth Planet. Sci. Lett.

    (2019)
  • H.J. Lippolt et al.

    (Uranium+ thorium)/helium dating of apatite: experience with samples from different geochemical environments

    Chem. Geol.

    (1994)
  • P. McConville et al.

    Implanted 3 He, 4 He, and Xe in further studies of diamonds from Western Australia

    Geochim. Cosmochim. Acta

    (1991)
  • W.F. McDonough et al.

    The composition of the Earth

    Chem. Geol.

    (1995)
  • G. Melton et al.

    Trace elements in gem diamond from Akwatia, Ghana and DeBeers Pool, South Africa

    Chem. Geol.

    (2012)
  • M. Ozima et al.

    Solar-type Ne in Zaire cubic diamonds

    Geochim. Cosmochim. Acta

    (1988)
  • M. Ozima et al.

    3He4He ratio, noble gas abundance and K-Ar dating of diamonds—an attempt to search for the records of early terrestrial history

    Geochim. Cosmochim. Acta

    (1983)
  • D.G. Pearson et al.

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

    Earth Planet. Sci. Lett.

    (1998)
  • D. Porcelli et al.

    The evolution of He isotopes in the convecting mantle and the preservation of high 3 He/4 He ratios

    Earth Planet. Sci. Lett.

    (2008)
  • S. Rege et al.

    Trace-element patterns of fibrous and monocrystalline diamonds: insights into mantle fluids

    Lithos

    (2010)
  • M. Schrauder et al.

    Trace element analyses of fluid-bearing diamonds from Jwaneng, Botswana

    Geochim. Cosmochim. Acta

    (1996)
  • A. Seta et al.

    Concurrent evolution of 3He/4He ratio in the Earth’s mantle reservoirs for the first 2 Ga

    Earth Planet. Sci. Lett.

    (2001)
  • D.A. Shelkov et al.

    The radial distribution of implanted and trapped 4He in single diamond crystals and implications for the origin of carbonado

    Chem. Geol.

    (1998)
  • S. Shilobreeva et al.

    Insights into C and H storage in the altered oceanic crust: results from ODP/IODP Hole 1256D

    Geochim. Cosmochim. Acta

    (2011)
  • T. Stachel et al.

    The origin of cratonic diamonds—constraints from mineral inclusions

    Ore Geol. Rev.

    (2008)
  • F. Stuart et al.

    Resolving mantle and crustal contributions to ancient hydrothermal fluids: He-Ar isotopes in fluid inclusions from Dae Hwa W-Mo mineralisation, South Korea

    Geochim. Cosmochim. Acta

    (1995)
  • C. Thomazo et al.

    Biological activity and the Earth's surface evolution: insights from carbon, sulfur, nitrogen and iron stable isotopes in the rock record

    Comptes Rendus Palevol

    (2009)
  • S. Timmerman et al.

    N-δ13C-inclusion profiles of cloudy diamonds from Koffiefontein: evidence for formation by continuous Rayleigh fractionation and multiple fluids

    Chem. Geol.

    (2018)
  • S. Timmerman et al.

    Dated eclogitic diamond growth zones reveal variable recycling of crustal carbon through time

    Earth Planet. Sci. Lett.

    (2017)
  • S. Timmerman et al.

    U-Th/He systematics of fluid-rich ‘fibrous’ diamonds–Evidence for pre-and syn-kimberlite eruption ages

    Chem. Geol.

    (2019)
  • I. Tolstikhin et al.

    The evolution of terrestrial volatiles: a view from helium, neon, argon and nitrogen isotope modelling

    Chem. Geol.

    (1998)
  • E. Tomlinson et al.

    Trace element compositions of submicroscopic inclusions in coated diamond: a tool for understanding diamond petrogenesis

    Geochim. Cosmochim. Acta

    (2005)
  • E. Tomlinson et al.

    A snapshot of mantle metasomatism: trace element analysis of coexisting fluid (LA-ICP-MS) and silicate (SIMS) inclusions in fibrous diamonds

    Earth Planet. Sci. Lett.

    (2009)
  • N. Wada et al.

    A noble gas study of cubic diamonds from Zaire: constraints on their mantle source

    Geochim. Cosmochim. Acta

    (1998)
  • Y. Weiss et al.

    High-Mg carbonatitic melts in diamonds, kimberlites and the sub-continental lithosphere

    Earth Planet. Sci. Lett.

    (2011)
  • Cited by (10)

    • Coupled deep-mantle carbon-water cycle: Evidence from lower-mantle diamonds

      2021, Innovation
      Citation Excerpt :

      The nitrogen contents from A- and B-type defects range from 52 to 96 ppm and from 49 to 134 ppm, respectively. The corresponding percentage of A-type defects and the approximate age of the diamonds (Table 1) were used to estimate average mantle residence temperatures.23,24 The fitted defect profiles and observed spectra are shown in Figure S1.

    • Deep carbon through time: Earth's diamond record and its implications for carbon cycling and fluid speciation in the mantle

      2020, Geochimica et Cosmochimica Acta
      Citation Excerpt :

      Conversely, Premier/Cullinan has multiple growth ages for peridotitic paragenesis diamonds (see Fig. 1) and since the carbon isotope data cannot be tied to specific growth events, data from Cullinan diamonds are not considered here. A few studies with age constraints have utilized SIMS analysis, with multiple points per sample, namely the Zimmi (Sierra Leone) data from Smit et al. (2019a), the Koffiefontein data (Timmerman et al., 2019a) and some of the Finsch data (Palot et al., 2013), the latter two both from South Africa. In these cases, the number of samples is only 5, 3 and 10, but the number of individual analyses equate to 179, 12 and 113, respectively.

    • Contrasting noble gas compositions of peridotitic and eclogitic monocrystalline diamonds from the Argyle lamproite, Western Australia

      2019, Lithos
      Citation Excerpt :

      The formula of Graham et al. (1987); 4He* = 2.8 x 10-8 x 7.35 x U (ppm) x t (Myr)) was used for modelling the ingrowth, with a Th/U ratio of 3, 1600 Myr (E-type) and 2000 Myr (P-type) and varying U contents. While the modelled curves (Fig. 3b) can reproduce most of the data, the amount of U (12 ppb for P-type, 15 ppb for E-type) needed to produce the lowest observed R/Ra values is an order of magnitude higher than the bulk of the literature range for gem-quality diamonds (<1.2 ppb U n = 28, only one sample had 15 ppb; Timmerman et al., 2019; Krebs et al., 2019). In addition, the plotted modelled curves are R/Ra values versus the ingrowth of 4He, which is shifted relative to curves of R/R values versus total 4He (initial+ingrowth) and would result in a poor fit for the peridotitic diamonds.

    View all citing articles on Scopus
    View full text