Key new pieces of the HIMU puzzle from olivines and diamond inclusions


Mantle melting, which leads to the formation of oceanic and continental crust, together with crust recycling through plate tectonics, are the primary processes that drive the chemical differentiation of the silicate Earth. The present-day mantle, as sampled by oceanic basalts, shows large chemical and isotopic variability bounded by a few end-member compositions1. Among these, the HIMU end-member (having a high U/Pb ratio, μ) has been generally considered to represent subducted/recycled basaltic oceanic crust2,3,4,5. However, this concept has been challenged by recent studies of the mantle source of HIMU magmas. For example, analyses of olivine phenocrysts in HIMU lavas indicate derivation from the partial melting of peridotite, rather than from the pyroxenitic remnants of recycled oceanic basalt6. Here we report data that elucidate the source of these lavas: high-precision trace-element analyses of olivine phenocrysts point to peridotite that has been metasomatized by carbonatite fluids. Moreover, similarities in the trace-element patterns of carbonatitic melt inclusions in diamonds7 and HIMU lavas indicate that the metasomatism occurred in the subcontinental lithospheric mantle, fused to the base of the continental crust and isolated from mantle convection. Taking into account evidence from sulfur isotope data8 for Archean to early Proterozoic surface material in the deep HIMU mantle source, a multi-stage evolution is revealed for the HIMU end-member, spanning more than half of Earth’s history. Before entrainment in the convecting mantle, storage in a boundary layer, upwelling as a mantle plume and partial melting to become ocean island basalt, the HIMU source formed as Archean–early Proterozoic subduction-related carbonatite-metasomatized subcontinental lithospheric mantle.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Figure 1: Compositions of olivine phenocrysts from HIMU (Mangaia, Tubuai) and intermediate HIMU–EM1 (Karthala) lavas.
Figure 2: Normalized trace-element patterns comparing HIMU–EM1 lavas (Mangaia, Tubuai and Karthala), Group I kimberlites, high-Mg carbonatitic fluids from diamond inclusions and the calculated melt of SCLM source compositions.
Figure 3: Mg# versus major element abundances, trace-element ratios, and isotope ratios in cratonic xenoliths.
Figure 4: Conceptual model for the evolution of the HIMU mantle source.


  1. 1

    Zindler, A. & Hart, S. Chemical geodynamics. Annu. Rev. Earth Planet. Sci. 14, 493–571 (1986)

  2. 2

    Zindler, A., Jagoutz, E. & Goldstein, S. Nd, Sr and Pb isotopic systematics in a three-component mantle: a new perspective. Nature 298, 519–523 (1982)

  3. 3

    Chauvel, C., Hofmann, A. W. & Vidal, P. HIMU-EM – the French Polynesian connection. Earth Planet. Sci. Lett. 110, 99–119 (1992)

  4. 4

    Hofmann, A. W. Mantle geochemistry: the message from oceanic volcanism. Nature 385, 219–229 (1997)

  5. 5

    Hanyu, T. et al. Geochemical characteristics and origin of the HIMU reservoir: a possible mantle plume source in the lower mantle. Geochem. Geophys. Geosyst. 12, Q0AC09 (2011)

  6. 6

    Herzberg, C. et al. Phantom Archean crust in Mangaia hotspot lavas and the meaning of heterogeneous mantle. Earth Planet. Sci. Lett. 396, 97–106 (2014)

  7. 7

    Weiss, Y., Griffin, W. L., Bell, D. R. & Navon, O. High-Mg carbonatitic melts in diamonds, kimberlites and the sub-continental lithosphere. Earth Planet. Sci. Lett. 309, 337–347 (2011)

  8. 8

    Cabral, R. A. et al. Anomalous sulphur isotopes in plume lavas reveal deep mantle storage of Archaean crust. Nature 496, 490–493 (2013)

  9. 9

    Hauri, E. H. & Hart, S. R. ReOs isotope systematics of HIMU and EMII oceanic island basalts from the south Pacific Ocean. Earth Planet. Sci. Lett. 114, 353–371 (1993)

  10. 10

    Stracke, A. Earth’s heterogeneous mantle: a product of convection-driven interaction between crust and mantle. Chem. Geol. 330–331, 274–299 (2012)

  11. 11

    Class, C., Goldstein, S. L., Altherr, R. & Bachelery, P. The process of plume–lithosphere interactions in the ocean basins: the case of Grande Comore. J. Petrol. 39, 881–903 (1998)

  12. 12

    Sobolev, A. V. et al. The amount of recycled crust in sources of mantle-derived melts. Science 316, 412–417 (2007)

  13. 13

    Matzen, A. K., Baker, M. B., Beckett, J. R. & Stolper, E. M. The temperature and pressure dependence of nickel partitioning between olivine and silicate melt. J. Petrol. 54, 2521–2545 (2013)

  14. 14

    Balta, J. B., Asimow, P. D. & Mosenfelder, J. L. Manganese partitioning during hydrous melting of peridotite. Geochim. Cosmochim. Acta 75, 5819–5833 (2011)

  15. 15

    De Hoog, J. C. M., Gall, L. & Cornell, D. H. Trace-element geochemistry of mantle olivine and application to mantle petrogenesis and geothermobarometry. Chem. Geol. 270, 196–215 (2010)

  16. 16

    Coogan, L. A., Saunders, A. D. & Wilson, R. N. Aluminum-in-olivine thermometry of primitive basalts: evidence of an anomalously hot mantle source for large igneous provinces. Chem. Geol. 368, 1–10 (2014)

  17. 17

    Agee, C. B. & Walker, D. Aluminum partitioning between olivine and ultrabasic silicate liquid to 6 GPa. Contrib. Mineral. Petrol. 105, 243–254 (1990)

  18. 18

    Saal, A. E., Hart, S. R., Shimizu, N., Hauri, E. H. & Layne, G. D. Pb isotopic variability in melt inclusions from oceanic island basalts, Polynesia. Science 282, 1481–1484 (1998)

  19. 19

    Jackson, M. G. & Dasgupta, R. Compositions of HIMU, EM1, and EM2 from global trends between radiogenic isotopes and major elements in ocean island basalts. Earth Planet. Sci. Lett. 276, 175–186 (2008)

  20. 20

    Castillo, P. R. The recycling of marine carbonates and sources of HIMU and FOZO ocean island basalts. Lithos 216–217, 254–263 (2015)

  21. 21

    Willbold, M. & Stracke, A. Trace element composition of mantle end-members: Implications for recycling of oceanic and upper and lower continental crust. Geochem. Geophys. Geosyst. 7, Q04004 (2006)

  22. 22

    Becker, M. & Le Roex, A. P. Geochemistry of South African on- and off-craton, Group I and Group II kimberlites: petrogenesis and source region evolution. J. Petrol. 47, 673–703 (2006)

  23. 23

    Gregoire, M., Bell, D. R. & Le Roex, A. P. Garnet lherzolites from the Kaapvaal craton (South Africa): trace element evidence for a metasomatic history. J. Petrol. 44, 629–657 (2003)

  24. 24

    McDonough, W. F. Contraints on the composition of the continental lithospheric mantle. Earth Planet. Sci. Lett. 101, 1–18 (1990)

  25. 25

    Hart, S. R., Gerlach, D. C. & White, W. M. A possible new Sr–Nd–Pb mantle array and consequences for mantle mixing. Geochim. Cosmochim. Acta 50, 1551–1557 (1986)

  26. 26

    Shirey, S. B. & Walker, R. J. The Re-Os isotope system in cosmochemistry and high-temperature geochemistry. Annu. Rev. Earth Planet. Sci. 26, 423–500 (1998)

  27. 27

    Rudnick, R. L., McDonough, W. F. & Chappell, B. W. Carbonatite metasomatism in the northern Tanzanian mantle: petrographic and geochemical characteristics. Earth Planet. Sci. Lett. 114, 463–475 (1993)

  28. 28

    Graham, S., Lambert, D. & Shee, S. The petrogenesis of carbonatite, melnoite and kimberlite from the Eastern Goldfields Province, Yilgarn Craton. Lithos 76, 519–533 (2004)

  29. 29

    Veizer, J. & Compston, W. 87Sr/86Sr in Precambrian carbonates as an index of crustal evolution. Geochim. Cosmochim. Acta 40, 905–914 (1976)

  30. 30

    Farquhar, J. et al. Mass-independent sulfur of inclusions in diamond and sulfur recycling on early earth. Science 298, 2369–2372 (2002)

  31. 31

    McKenzie, D. & O’Nions, R. K. Mantle reservoirs and ocean island basalts. Nature 301, 229–231 (1983)

  32. 32

    Stein, M., Navon, O. & Kessel, R. Chromatographic metasomatism of the Arabian–Nubian lithosphere. Earth Planet. Sci. Lett. 152, 75–91 (1997)

  33. 33

    Class, C., Goldstein, S. L. & Shirey, S. B. Osmium isotopes in Grande Comore lavas: a new extreme among a spectrum of EM-type mantle endmembers. Earth Planet. Sci. Lett. 284, 219–227 (2009)

  34. 34

    Class, C., Goldstein, S. L., Stute, M., Kurz, M. D. & Schlosser, P. Grand Comore Island: a well-constrained “low He-3/He-4” mantle plume. Earth Planet. Sci. Lett. 233, 391–409 (2005)

  35. 35

    Hanyu, T., Tatsumi, Y. & Kimura, J.-I. Constraints on the origin of the HIMU reservoir from He–Ne–Ar isotope systematics. Earth Planet. Sci. Lett. 307, 377–386 (2011)

  36. 36

    Kramers, J. D. Lead and strontium isotopes in Cretaceous kimberlites and mantle-derived xenoliths from Southern Africa. Earth Planet. Sci. Lett. 34, 419–431 (1977)

  37. 37

    Collerson, K. D., Williams, Q., Ewart, A. E. & Murphy, D. T. Origin of HIMU and EM-1 domains sampled by ocean island basalts, kimberlites and carbonatites: the role of CO2-fluxed lower mantle melting in thermochemical upwellings. Phys. Earth Planet. Inter. 181, 112–131 (2010)

  38. 38

    Tilton, G. R. & Bell, K. Sr–Nd–Pb isotope relationships in Late Archean carbonatites and alkaline complexes: applications to the geochemical evolution of Archean mantle. Geochim. Cosmochim. Acta 58, 3145–3154 (1994)

  39. 39

    Bell, K. & Tilton, G. R. Nd, Pb and Sr isotopic compositions of East African carbonatites: evidence for mantle mixing and plume inhomogeneity. J. Petrol. 42, 1927–1945 (2001)

  40. 40

    Bizimis, M., Salters, V. J. M. & Dawson, J. B. The brevity of carbonatite sources in the mantle: evidence from Hf isotopes. Contrib. Mineral. Petrol. 145, 281–300 (2003)

  41. 41

    McNutt, M. K. Marine geodynamics: depth-age revisited. Rev. Geophys. 33, 413–418 (1995)

  42. 42

    Braun, M. G. & Kelemen, P. B. Dunite distribution in the Oman ophiolite: implications for melt flux through porous dunite conduits. Geochem. Geophys. Geosyst. 3, 8603 (2002)

  43. 43

    Herzberg, C. & Asimow, P. D. Petrology of some oceanic island basalts: PRIMELT2.XLS software for primary magma calculation. Geochem. Geophys. Geosyst. 9, Q09001 (2008)

  44. 44

    Korenaga, J. & Kelemen, P. B. Major element heterogeneity in the mantle source of the North Atlantic igneous province. Earth Planet. Sci. Lett. 184, 251–268 (2000)

Download references


We thank S. Lambart, D. G. Pearson, S. Aulbach and D. Walker for discussions, A. W. Hofmann for discussions and an informal review, J. Gross and B. A. Goldoff for help with the EPMA analyses, and K. Lehnert and A. Johansson for help with accessing the xenolith data in EarthChem’s PetDB database. This work was supported by an LDEO Postdoctoral Fellowship for Y.W., NSF grant EAR13-48045 to Y.W., C.C. and S.L.G., and the Storke Endowment of the Columbia University Department of Earth and Environmental Sciences. This is Lamont–Doherty Earth Observatory contribution number 8046.

Author information




Y.W., C.C. and S.L.G. conceived the project, developed the model and wrote the paper. Y.W. preformed the EPMA analyses. T.H. provided the olivine samples for the study and contributed to the paper.

Corresponding authors

Correspondence to Yaakov Weiss or Cornelia Class or Steven L. Goldstein or Takeshi Hanyu.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Additional information

The new data (Supplementary Data 1, as well as the compiled xenolith data set used here (Supplementary Data 2, have been submitted to EarthChem (

Reviewer Information Nature thanks E. Hauri and W. White for their contribution to the peer review of this work.

Extended data figures and tables

Extended Data Figure 1 Sr–Nd–Pb isotopes of Mangaia, Tubuai and Karthala samples in this study, African Group I kimberlites and continental carbonatites.

a, 143Nd/144Nd versus 87Sr/86Sr. b, 208Pb/204Pb versus 206Pb/204Pb. The data illustrate the compositional range between the global mantle endmembers (DMM, depleted MORB mantle; EM1; EM2 and HIMU); mixing on this diagram is linear. Data sources: Mangaia and Tubuai (ref. 5), Karthala (ref. 11), kimberlites (refs 22, 36, 37) and continental carbonatites (refs 38, 39, 40). The mantle end-member components are from ref. 1; the grey and black data points are from Hawaii, Iceland, St Helena, Cook–Austral Islands, Samoa, Society, Marquesas, Pitcairn and Tristan (from the compilation of ref. 10).

Extended Data Figure 2 Pressure–temperature corrections to the Ni contents of Mangaia, Tubuai and Karthala olivine phenocrysts.

a, 300 p.p.m. Ni. b, 500 p.p.m. Ni. Following ref. 13, we calculated Ni in near-surface olivine in equilibrium with Mangaia and Karthala primary magmas at different segregation pressures (ΔT in the model). On the basis of the age of the oceanic crust of Mangaia and Tubuai (80–120 Ma) and Karthala (approximately 140 Ma), which are equivalent to estimated lithosphere thicknesses of around 75–100 km (ref. 41), we assume that the difference between the pressures (P) of segregation of these lavas compared to MORBs (around 1 GPa; ref. 42) is 1.25–2.00 GPa, equivalent to ΔT = 70–110 °C. This difference in the segregation pressures corresponds to 300–500 p.p.m. Ni (the arrow pointing to lower Ni indicates the effect of increasing pressure and/or temperature (‘P/T effect’) on Ni-in-olivine). When corrected, Mangaia, Tubuai and Karthala data agree with the modelled olivine liquid line of descent for Mangaia primary magmas6 (that is, the green dashed line; L, liquid; Ol, olivine; Cpx, clinopyroxene; Plag, plagioclase), and lie on an extension of the crystallization trend of MORB phenocryst compositions. These relationships support the connection between the olivine phenocryst compositions and a peridotite source lithology. For the calculations, we used PRIMELT2 (ref. 43) to determine the parental magma compositions, the Fo content of the olivine in equilibrium with this liquid and the liquidus temperature at 1 bar; and Ni = 0.37 wt% for the residual peridotite olivine, which is the mean global value for olivines from spinel and garnet peridotites44. The yellow ellipses encompass >95% of the data points for Karthala olivine (see Supplementary Data 1).

Extended Data Figure 3 Ca and Al in olivine from garnet and spinel peridotite xenoliths as a function of temperature.

Top, Al concentration. Middle, Ca concentration. Bottom, Ca/Al ratio. Xenolith data and temperatures are from ref. 15. The data show that while partitioning of both Ca and Al into olivine increase with temperature, the Ca/Al ratios in olivine initially decrease and become nearly constant with increasing temperature.

Extended Data Figure 4 Ca and Al partitioning with temperature in olivines in basalts.

a, Ratio of Ca and Al partition coefficients (D). b, Ca/Al ratios in olivine phenocrysts. The experimentally determined D values for Ca- and Al-in-olivine, in equilibrium with a peridotite-derived melt, are from ref. 17; they were determined for increasing temperature and pressure. Data for Ca and Al in olivine phenocrysts from MORB (NW Pacific, Hess Deep, Gulf of California, Siqueiros Transform) and large igneous provinces (LIPs; SE Greenland, Baffin Island, Gorgona, Madagascar) are from ref. 16; temperatures were determined on the basis of the Al-in-olivine thermometer16. Both the experimental results and the measured Ca and Al contents in crystallizing olivines show decreasing Ca/Al ratios with increasing temperature. On the basis of these data and the observations in Extended Data Fig. 3, Ca/Al ratios in olivine phenocrysts from magmatic systems can be expected to decrease with increasing temperature. In contrast, our data from HIMU hot plumes show very high Ca/Al ratios in olivine phenocrysts compared with low Ca/Al ratios in phenocrysts from cooler MORB lavas (or other hot OIBs) (see Fig. 1). Temperature differences therefore cannot explain the Ca/Al variations in olivine from OIB and MORB lavas. We thus conclude that the high Ca/Al ratios of HIMU olivine phenocrysts and lavas instead reflect the compositions of the HIMU magma source.

Extended Data Figure 5 187Os/188Os evolution of mixtures of lherzolite and metasomatized harzburgite.

The metasomatic event in the SCLM took place in the Archean or early Proterozoic and involved materials that had been at the Earth’s surface (on the basis of the observation that HIMU lavas contain sulfur that shows mass-independent fractionation)8. We use 2.7 Ga for this example. Our study determined that the metasomatizing agent is subduction-derived carbonatitic fluids/melts. The primitive upper mantle (PUM) evolution line is based on present-day values of 187Os/188Os = 0.129 and 187Re/188Os = 0.43 (ref. 26); at 2.7 Ga the 187Os/188Os of PUM was 0.109. The figure shows evolution lines from the PUM at 2.7 Ga for mixtures of lherzolite and metasomatized harzburgite ranging from 50% to 100% lherzolite. The Re/Os ratios of the normal (lherzolite) and metasomatized (harzburgitic) SCLM are from the averages of 143 cratonic lherzolite xenoliths (187Re/188Os = 0.4) and 31 harzburgite xenoliths (187Re/188Os = 2.44), respectively, all with Mg# between 0.910 and 0.935 from the mantle xenolith database in PetDB (see Supplementary Data 2). For a metasomatic event at 2.7 Ga, only 18%–30% metasomatized harzburgite is required to reproduce the 187Os/188Os ratios of HIMU lavas.

Supplementary information

Supplementary Information

This file contains the Supplementary Text and Data for: (1) Trace and element Melting Model and (2) Os Evolution Model of Old Metasomatized SCLM. Additional references are also included. (PDF 309 kb)

Supplementary Data 1

This file contains the EPMA analyses of olivine, which is the Source Data for the RED, PINK and YELLOW data points in Figure 1. (XLSX 161 kb)

Supplementary Data 2

This file shows the Cratonic Xenoliths, which is the Source Data for all the data points in Figure 3. (XLSX 632 kb)

PowerPoint slides

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Weiss, Y., Class, C., Goldstein, S. et al. Key new pieces of the HIMU puzzle from olivines and diamond inclusions. Nature 537, 666–670 (2016).

Download citation

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.