Papex

Open problem · Earth science · Open

Chemical heterogeneity at the core–mantle boundary

What is the chemical nature of the irregular structures at the boundary between Earth’s mantle and core?

Summary

Seismic imaging reveals large irregularities just above the core–mantle boundary, including two continent-sized regions of low seismic velocity; the precise nature of the chemical heterogeneity there is unresolved.

Why it is open

Except near its bottom, the lower mantle is nearly uniform at each depth, with lateral variations in seismic velocity of 1% or less; large anomalies appear in its deepest several hundred kilometres, a region called the D″ layer [3]. Two kinds of structure stand out. The large low-shear-velocity provinces (LLSVPs, also called LLVPs) beneath Africa and the Pacific extend about 1,000 kilometres up from the core–mantle boundary [7]. Ultralow-velocity zones (ULVZs) are 5 to 50 kilometres thick, with strongly reduced compressional and shear velocities and increased density [11].

Seismic waves slow down in hotter rock and also in rock of a different composition, and tomographic models based only on wave speeds have limited ability to tell the two apart [13]. Density helps to decide: Kei Hirose and colleagues wrote in 2006 that if density and velocity anomalies are anti-correlated, the cause "must be due to chemical heterogeneity" [3]. Density at that depth is hard to measure, and a phase change of the mantle silicate MgSiO3 near the boundary can explain some observations otherwise read as chemical [3]. They called separating the roles of the phase change and of chemical anomalies "an outstanding challenge" [3].

The problem is therefore several questions: whether the LLSVPs are chemically distinct, how much of their volume is, what the ULVZs are made of, and where any distinct material came from. No body sets a bar for a resolution. In 2023 Fred Richards and colleagues named two questions still unanswered: the true vertical extent of the provinces' buoyancy anomalies, and whether they are purely thermal or also compositionally distinct [10].

What is known

Two kinds of structure (1996–2023). In 1996 Quentin Williams and Edward Garnero reported an intermittent layer at the base of the mantle, up to about 40 kilometres thick, with compressional velocity about 10% lower than above, "most simply explained" by partial melt [1]. Samantha Hansen and colleagues wrote in 2023 that less than 20% of the core–mantle boundary had been surveyed for ULVZs; with an array of seismometers in Antarctica they found widespread, variable ULVZs beneath a largely unsampled part of the Southern Hemisphere, away from current or recent subduction [11]. In 2020 Doyeon Kim and colleagues applied a manifold-learning algorithm, the Sequencer, to thousands of seismograms of waves diffracted along the boundary beneath the Pacific; nearly half showed waves scattered by three-dimensional structures, among them a plume root beneath Hawaii and a previously unrecognised ULVZ beneath the Marquesas Islands [8].

A phase change (2004–2006). In 2004 Motohiko Murakami and colleagues found that MgSiO3 perovskite transforms to a post-perovskite phase at about 125 gigapascals and 2,500 kelvin, near 2,700 kilometres depth [3]. Hirose and colleagues considered this the likely cause of the seismic discontinuity at the top of D″, but noted observed velocity jumps of up to 3%, much larger than mineral physics predicts, and a sharpness that may require a chemical boundary [3].

Density (1999–2023). In 1999, from Earth's free oscillations and gravity, Miaki Ishii and Jeroen Tromp found denser-than-average material beneath the Pacific and Africa where shear waves are slow, suggesting compositional heterogeneity [2]; Barbara Romanowicz argued in 2001 that free oscillations resolve the lowermost mantle only weakly [3]. In 2017 Paula Koelemeijer and colleagues used Stoneley modes, oscillations trapped along the boundary, and found the overall density of the LLSVPs lower than the surrounding mantle, without ruling out a dense layer about 100 kilometres thick at their base [6]. The same year Harriet Lau and colleagues used the solid-Earth tide measured by GPS and found the bottom two-thirds of the provinces about 0.5% denser than average [7]. Richards and colleagues found that the geoid and surface topography limit large-scale density anomalies to about 900 kilometres above the boundary, and that dense material at the base of the provinces fits the geoid, topography, the boundary's ellipticity, tides and Stoneley modes together [10].

Shape and attenuation (2020–2025). In 2020 Anne Davaille and Romanowicz argued from a full-waveform tomographic model that the LLSVPs are not compact piles but bundles of separate upwellings, probably enriched in dense material, which look uniform only when the model is filtered to long wavelengths [9]. In 2025 Sujania Talavera-Soza and colleagues published a whole-mantle model of attenuation, the loss of seismic energy, from Earth's free oscillations: in the lower mantle the LLSVPs attenuate least and the fast "ring around the Pacific" most, which they read as warm, large-grained provinces surrounded by colder, small-grained mantle, and as long-lived, stable features [13].

Isotopes in plume lavas (2017–2025). Andrea Mundl and colleagues found ocean island basalts from Hawaii, Samoa and Iceland with tungsten-182 ratios up to 18 parts per million below those of the ambient upper mantle, lower where the ratio of helium-3 to helium-4 is higher; they inferred source domains that formed within the first 60 million years of the Solar System and contain metal, and suggested very large ULVZs beneath those hotspots as the likely source [5]. In 2025 Nils Messling and colleagues reported that Hawaiian basalts have larger anomalies in ruthenium-100 than the ambient mantle, which together with their low tungsten-182 they call "diagnostic of a core contribution" [14]. What the core itself contains is a related problem, the light elements in Earth's outer core.

Main approaches

Mostly thermal provinces. The LLSVPs are hot upwellings, with at most a thin dense base [6, 10]. Strongest evidence: the low average density from Stoneley modes [6]; Richards and colleagues conclude that the provinces are "dominantly thermal structures" [10]; the attenuation model derives their long life from grain size and temperature [13]. Main difficulty: dense basal material is still needed to fit the geodynamic and geodetic data [10], and the tidal result implies dense chemical components [7].

Subducted oceanic crust. Former basaltic crust is about 3% denser than average mantle at the base of the mantle, and simulations show it can segregate there [3]. Strongest evidence: in 2025 Ewa Krymarys and colleagues measured shear velocities of the silica phases of basalt beyond core–mantle pressures, found them 7–14% lower than predicted, and concluded that 23–33% oceanic crust by volume could explain the typical −1.5% shear-velocity anomaly of the LLSVPs without extreme temperatures [15]; subducted material spread along the boundary could explain the range of ULVZs [11]. Main difficulty: computational models had predicted high shear velocities for basaltic crust and so questioned it as the sole source [15]; the new values are from one group's experiments [15].

Primordial reservoirs. Dense material formed early in Earth's history survives at the base of the mantle [5, 10]. Strongest evidence: the tungsten and helium signatures of domains formed in the first 60 million years [5]; Richards and colleagues find early-formed, chondrite-enriched basalt in the deepest 100–200 kilometres of the provinces most compatible with the geophysical data [10]. Main difficulty: low tungsten-182 can also come from the core, as Messling and colleagues argue for Hawaii [14].

Remnants of the Moon-forming impactor. In 2023 Qian Yuan and colleagues proposed that the provinces are buried relics of the mantle of Theia, the body in the Moon-forming giant impact [12]. Strongest evidence: in their impact simulations part of Theia's mantle reaches Earth's solid lower mantle, and in their convection models dense blobs tens of kilometres across sink into piles that survive to the present [12]. Main difficulty: the 2.0–3.5% density excess rests on models of Theia's mantle and on the Moon's higher iron-oxide content [12]; whether such an impact formed the Moon is itself an open problem, the giant-impact origin of the Moon.

Melt or iron-rich solids in ULVZs. Strongest evidence: partial melt fits the approximately 3:1 ratio of shear to compressional velocity reductions found in some studies [11]; iron-rich (Mg,Fe)O, measured to 121 gigapascals at room temperature, has sound velocities so low that a small amount can greatly lower the average [4]. Main difficulty: some ULVZs lie far from the presumably hottest regions, and their thickness does not always follow the pattern expected from melting alone [11].

Open sub-questions

  1. Where the density sits. Can a joint inversion of body tides and Stoneley modes decide whether the excess density of the LLSVPs is confined to their deepest 100–200 kilometres [6, 7, 10]?
  2. Attenuation at finer scale. The 2025 attenuation model constrains even spherical harmonics only up to degree four; do regional body-wave measurements inside the provinces also show the low attenuation expected for large grains [13]?
  3. Basalt fraction and density. Does 23–33% oceanic crust by volume, which fits the shear-velocity anomaly, also give a density excess compatible with the tidal and Stoneley-mode results [6, 7, 15]?
  4. Ruthenium beyond Hawaii. Do basalts from Samoa and Iceland, which also have low tungsten-182, carry the ruthenium-100 excess found in Hawaii [5, 14]?
  5. ULVZs away from the provinces. Applied to diffracted waves beneath poorly sampled regions, does the Sequencer method find ULVZs far from the LLSVPs as often as a subducted-material origin predicts [8, 11]?
  6. A Theia signature. What ruthenium or tungsten isotope signature would distinguish Theia-derived material from subducted crust in plume lavas [12, 14]? See also the giant-impact origin of the Moon.

Work on this in Papex

If you have a testable answer to one of the sub-questions above, you can draft it as a private hypothesis in Papex, or prepare the draft from Claude or ChatGPT with the Papex connector. You review the exact version, sign it and decide whether to publish it; a public record that names this problem is listed on this page. A Papex record shows who signed which version and when. It is not peer review and does not mean the hypothesis is correct.

Key references

  1. Williams, Q. & Garnero, E. J. (1996). Seismic evidence for partial melt at the base of Earth's mantle. Science 273(5281), 1528–1530. 10.1126/science.273.5281.1528 · Checked 15 Oct 2026
  2. Ishii, M. & Tromp, J. (1999). Normal-mode and free-air gravity constraints on lateral variations in velocity and density of Earth's mantle. Science 285(5431), 1231–1236. 10.1126/science.285.5431.1231 · Checked 15 Oct 2026
  3. Hirose, K., Karato, S., Cormier, V. F., Brodholt, J. P. & Yuen, D. A. (2006). Unsolved problems in the lowermost mantle. Geophysical Research Letters 33(12), L12S01. 10.1029/2006GL025691 · Checked 15 Oct 2026
  4. Wicks, J. K., Jackson, J. M. & Sturhahn, W. (2010). Very low sound velocities in iron-rich (Mg,Fe)O: implications for the core-mantle boundary region. Geophysical Research Letters 37(15). 10.1029/2010GL043689 · Checked 15 Oct 2026
  5. Mundl, A., Touboul, M., Jackson, M. G., Day, J. M. D., Kurz, M. D., Lekic, V. et al. (2017). Tungsten-182 heterogeneity in modern ocean island basalts. Science 356(6333), 66–69. 10.1126/science.aal4179 · Checked 15 Oct 2026
  6. Koelemeijer, P., Deuss, A. & Ritsema, J. (2017). Density structure of Earth's lowermost mantle from Stoneley mode splitting observations. Nature Communications 8, 15241. 10.1038/ncomms15241 · Checked 15 Oct 2026
  7. Lau, H. C. P., Mitrovica, J. X., Davis, J. L., Tromp, J., Yang, H.-Y. & Al-Attar, D. (2017). Tidal tomography constrains Earth's deep-mantle buoyancy. Nature 551(7680), 321–326. 10.1038/nature24452 · Checked 15 Oct 2026
  8. Kim, D., Lekić, V., Ménard, B., Baron, D. & Taghizadeh-Popp, M. (2020). Sequencing seismograms: a panoptic view of scattering in the core-mantle boundary region. Science 368(6496), 1223–1228. 10.1126/science.aba8972 · Checked 15 Oct 2026
  9. Davaille, A. & Romanowicz, B. (2020). Deflating the LLSVPs: bundles of mantle thermochemical plumes rather than thick stagnant "piles". Tectonics 39(10), e2020TC006265. 10.1029/2020TC006265 · Checked 15 Oct 2026
  10. Richards, F. D., Hoggard, M. J., Ghelichkhan, S., Koelemeijer, P. & Lau, H. C. P. (2023). Geodynamic, geodetic, and seismic constraints favour deflated and dense-cored LLVPs. Earth and Planetary Science Letters 602, 117964. 10.1016/j.epsl.2022.117964 · Checked 15 Oct 2026
  11. Hansen, S. E., Garnero, E. J., Li, M., Shim, S.-H. & Rost, S. (2023). Globally distributed subducted materials along the Earth's core-mantle boundary: implications for ultralow velocity zones. Science Advances 9(14), eadd4838. 10.1126/sciadv.add4838 · Checked 15 Oct 2026
  12. Yuan, Q., Li, M., Desch, S. J., Ko, B., Deng, H., Garnero, E. J. et al. (2023). Moon-forming impactor as a source of Earth's basal mantle anomalies. Nature 623(7985), 95–99. 10.1038/s41586-023-06589-1 · Checked 15 Oct 2026
  13. Talavera-Soza, S., Cobden, L., Faul, U. H. & Deuss, A. (2025). Global 3D model of mantle attenuation using seismic normal modes. Nature 637(8048), 1131–1135. 10.1038/s41586-024-08322-y · Checked 15 Oct 2026
  14. Messling, N., Willbold, M., Kallas, L., Elliott, T., Fitton, J. G., Müller, T. et al. (2025). Ru and W isotope systematics in ocean island basalts reveals core leakage. Nature 642(8067), 376–380. 10.1038/s41586-025-09003-0 · Checked 15 Oct 2026
  15. Krymarys, E., Murakami, M., Saha, P. & Liebske, C. (2025). The presence of ancient subducted oceanic crust contributes to seismic anomalies in Large Low Shear Velocity Provinces. Communications Earth & Environment 6, 713. 10.1038/s43247-025-02700-8 · Checked 15 Oct 2026

Page reviewed 15 Oct 2026

Source

Open, as listed by Wikipedia · Unsolved problems in geoscience. Checked 08 Oct 2026.

The status is the source’s. Papex does not decide whether a problem is solved.

Records on this problem

No public record addresses this problem yet.

More in Earth science