Researchers discovered a distinct fourth contributor to mantle heterogeneities\u2014one that arises from the quantum behavior of electrons in iron ions.<\/p>\n
Recently, researchers discovered a distinct fourth contributor to mantle heterogeneities\u2014one that arises from the quantum behavior of electrons in iron ions.<\/p>\n
This quantum-scale process occurs within individual atoms, but recent findings show that it can shape planetary-scale structures by influencing the buoyancy, viscosity, and flow of rock in the lower mantle. Research on quantum effects in this region is still in the early stages, but this phenomenon is clearly essential to a fuller understanding of Earth\u2019s deep interior.<\/p>\n
Spins Under Pressure<\/p>\n
Deep in the lower mantle, iron ions in the two dominant minerals, bridgmanite ((Mg,Fe)(Si,Fe)O3) and ferropericlase ((Mg, Fe)O), undergo a pressure-driven rearrangement of their electronic structure known as a spin transition. In this process, some electrons in the iron ions change from a \u201chigh-spin\u201d configuration to a \u201clow-spin\u201d one (more details are given below). No chemical bonds are broken, and mineral crystal symmetries remain unchanged. But the affected iron ions shrink, their bonds stretch, and their volume collapses.<\/p>\n
High-pressure experiments revealed this electronic transformation in the laboratory 2 decades ago [e.g., Badro et al.<\/a>, 2003; Lin et al.<\/a>, 2005; Komabayashi et al.<\/a>, 2010]. At the time, the potential implications for mantle structure and behavior were considered potentially important but speculative. If this phenomenon is subtle, can it matter at the scale of continents and convection flows? The answer has emerged only gradually.<\/p>\n Spin is an intrinsic property of single ions. The populations of high- and low-spin iron ions in the mantle change continuously with pressure and temperature [Tsuchiya et al.<\/a>, 2006], a phenomenon known as the iron spin crossover (ISC). As the proportions of these spin states change, the compressibility of the ions\u2019 host mineral phases is altered [Wentzcovitch et al.<\/a>, 2009].<\/p>\n Because the iron spin crossover (ISC) affects a mineral\u2019s compressibility, it leaves a distinctive imprint on P wave speeds, whereas S wave speeds change much less.<\/p>\n Ferropericlase makes up about 20% of the lower mantle volume, but it is the mantle mineral with the highest per-mole iron content. Therefore, the ISC is most detectable in this phase. Meanwhile, the ISC in bridgmanite is less detectable and may have a less significant effect on compressibility, even though bridgmanite constitutes up to 80% of the lower mantle.<\/p>\n The explanation may be that in the bridgmanite crystal structure, iron usually occupies a less compressed position within a silica framework. To be compressed sufficiently to undergo the ISC, iron in bridgmanite has to replace some of the silicon ions in the mineral\u2019s smaller octahedral site. This replacement does not happen easily because the mantle contains enough aluminum and silicon to complete the octahedral framework without stretching to accommodate the larger iron ions.<\/p>\n Compressional seismic waves (P waves) respond to changes in the compressibility of the material through which they are traveling more than shear waves (S waves) do. Because the ISC affects a mineral\u2019s compressibility, it leaves a distinctive imprint on P wave speeds, whereas S wave speeds change much less [Wu et al.<\/a>, 2013]. This difference is hidden in plain sight. Only when mineral physics and 3D seismic imaging were brought together did the consequences become clear.<\/p>\n The Physics of a Spin Crossover<\/p>\n The ISC is strictly a quantum phenomenon. In a high-spin state, electrons occupy separate atomic orbitals with parallel spins, resulting in a larger ion with a larger volume. In a low-spin state, electrons pair within lower-energy orbitals, reducing both spin and ionic volume. The temperature at any given depth determines the relative abundances of these different spin states (Figure 1).<\/p>\n At lower pressures, iron in both ferropericlase and bridgmanite remains in the high-spin state, but increasing pressure stabilizes the low-spin configuration. Laboratory measurements have shown that the spin states of iron ions begin to change at depths near 1,000 kilometers and continue well past 2,000 kilometers, covering most of the lower mantle [Lin et al.<\/a>, 2007].<\/p>\n Experiments and ab initio (first-principles) calculations have similarly indicated that the ISC does not create a sharp seismic discontinuity in the mantle [Lin et al.<\/a>, 2007; Wu and Wentzcovitch<\/a>, 2014; Zhuang and Wentzcovitch<\/a>, 2024]. Instead, it produces a broad, mixed-spin region in which high- and low-spin states coexist (Figure 1b), each with its characteristic ionic volume and local strains.<\/p>\n Those structural subtleties and the changes in the population of these iron spin states under pressure alter the minerals\u2019 responses to compression.<\/p>\n From Quantum Physics to Seismic Velocities<\/p>\n The gradual, pressure-induced ISC reduces ferropericlase\u2019s bulk modulus, or its resistance to compression (Figure 2). Meanwhile, the mineral\u2019s shear modulus, its resistance to shape deformation, gently increases [Wu et al.<\/a>, 2013]. Because P wave velocities depend on both moduli and S wave velocities depend on the shear modulus but not the bulk modulus, the ISC reduces P wave velocities with increasing depth, whereas S wave velocities increase slightly with depth.<\/p>\n A breakthrough in understanding these trends came when advanced ab initio simulations revealed that the elastic signature of the ISC (i.e., its response to applied force) is very diffuse [Zhuang and Wentzcovitch<\/a>, 2024]. This signature spans most of the lower mantle, broadly reducing P wave speeds rather than causing an abrupt drop. Because the ISC is temperature dependent, these broad variations are even less obvious when 3D seismic wave speeds are spherically averaged, making them less apparent in 1D models (Figure 3).<\/p>\n These updated simulations also showed that the ISC-induced reduction in P wave speeds is unavoidable. Models that infer mantle temperatures from seismic velocities typically do not account for this reduction, leading to unrealistically low temperatures and extreme compositions [Cobden et al.<\/a>, 2024]. The ISC acts as a background correction in calculations of temperature and composition: essential but invisible unless explicitly accounted for (Figure 3).<\/p>\n This insight set the stage for a new way of looking at seismic data, not through 1D radial averages, but through the 3D relationships between P and S wave speeds.<\/p>\n A Seismological Signature Noticeable in 3D<\/p>\n Where the ISC becomes strikingly evident is in the difference between compressional wave and shear wave speed anomalies induced by temperature variations (Figure 4). Because temperature affects P wave speeds differently than S wave speeds, regions containing ferropericlase should show muted P wave anomalies relative to their S wave anomalies.<\/p>\n Slabs of subducted oceanic lithosphere, for example, have faster wave speeds than the ambient mantle because they are colder and harder to compress. Thermodynamic models suggest that the ISC should dampen P wave speed anomalies but not S wave speed anomalies in these slabs, producing a diagnostic pattern of decorrelation between P and S wave speeds.<\/p>\n This effect was recognized by Wu and Wentzcovitch<\/a> [2014], but the first global-scale evidence of it appeared when Shephard et al.<\/a> [2021] used \u201cvote maps\u201d to highlight consistent features across many independent tomography models (i.e., how many models agree on, or \u201cvote\u201d for, the presence of a feature). The vote maps showed that in the lower mantle, P and S wave structures diverge precisely where ferropericlase is expected to be in a mixed-spin state.<\/p>\n These findings form a consistent narrative: The ISC is not expressed as a sharp or diffuse \u201clayer\u201d in the mantle, but through 3D contrasts between P and S wave structures.<\/p>\n More recently, Cobden et al.<\/a> [2024] used full-waveform tomography<\/a> to obtain absolute values of P and S wave speeds. Their results showed that reproducing observed velocities with realistic temperatures and compositions requires inclusion of the ISC; without it, the middle mantle would have to be unrealistically cold and strongly depleted of silicon.<\/p>\n Although the influence of the ISC on the ratio of temperature-induced S wave to P wave variations was previously recognized, a full treatment of the ISC in ferropericlase has been shown to quantitatively reconcile mineral physics predictions with seismological observations [Wu and Wentzcovitch<\/a>, 2014; Zhuang and Wentzcovitch<\/a>, 2024].<\/p>\n Together, these findings form a consistent narrative: The ISC is not expressed as a sharp or diffuse \u201clayer\u201d in the mantle, but through 3D contrasts between P and S wave structures.<\/p>\n The Iron Spin Crossover Quietly Reshapes the Mantle<\/p>\n New understanding from recent research affirms several consequential points for geophysical studies. First, the ISC reshapes interpretations of deep-mantle seismic structure. Seismic heterogeneity reflects not only temperature, phase, and composition changes but also variations in spin state abundances. Temperature-dependent changes in high- and low-spin iron populations directly influence seismic wave speeds.<\/p>\n Second, accounting for ISC effects yields more realistic mantle temperatures in interpretations based on full-waveform tomography. Without these effects, models indicate unrealistically cold or extreme compositions.<\/p>\n Third, mixed-spin ferropericlase densifies more rapidly with increasing pressure. This densification invigorates mantle convection [Bower et al.<\/a>, 2009] and may reduce its viscosity, potentially influencing tectonic processes such as slab sinking and stagnation, as well as plume rising and morphology.<\/p>\n The ISC illustrates how quantum-scale processes shape planetary-scale behavior.<\/p>\n On a broader scale, the ISC illustrates how quantum-scale processes shape planetary-scale behavior. Rather than producing sharp seismic boundaries, it introduces subtle, pervasive effects on wave speeds and their relationships in 3D.<\/p>\n The ISC provides seismologists with a framework for reconciling mantle temperatures, compositions, and wave speed anomalies. It also highlights for mineral physicists the importance of electronic structure and suggests to geodynamicists that electron spin may influence buoyancy, viscosity, and flow. Further, it shows that key processes in Earth\u2019s interior can remain hidden until multiple disciplines integrate their expertise and efforts.<\/p>\n As seismic imaging improves and mineral physics models become more accurate, the ISC\u2019s role in shaping lower mantle structure and dynamics will become clearer. Indeed, revealing its influence on prominent but not fully understood phenomena<\/a> such as large low shear velocity provinces<\/a>, the D\u02ba discontinuity and layer, and slab stagnation zones is an active area of research. What is already clear is that the ISC is a global, depth-spanning process essential to understanding Earth\u2019s deep interior.<\/p>\n References<\/p>\n Badro, J., et al. (2003), Iron partitioning in Earth\u2019s mantle: Toward a deep lower mantle discontinuity, Science, 300(5620), 789\u2013791, https:\/\/doi.org\/10.1126\/science.1081311<\/a>.<\/p>\n Bower, D. J., et al. (2009), Enhanced convection and fast plumes in the lower mantle induced by the spin transition in ferropericlase, Geophys. Res. Lett., 36, L10306, https:\/\/doi.org\/10.1029\/2009GL037706<\/a>.<\/p>\n Cobden, L., et al. (2024), Full-waveform tomography reveals iron spin crossover in Earth\u2019s lower mantle, Nat. Commun., 15, 1961, https:\/\/doi.org\/10.1038\/s41467-024-46040-1<\/a>.<\/p>\n Komabayashi, T., et al. (2010), High-temperature compression of ferropericlase and the effect of temperature on iron spin transition, Earth Planet. Sci. Lett., 297, 691\u2013699, https:\/\/doi.org\/10.1016\/j.epsl.2010.07.025<\/a>.<\/p>\n Lin, J.-F., et al. (2005), Spin transition of iron in magnesiow\u00fcstite in the Earth\u2019s lower mantle, Nature, 436, 377\u2013380, https:\/\/doi.org\/10.1038\/nature03825<\/a>.<\/p>\n Lin, J.-F., et al. (2007), Spin transition zone in Earth\u2019s lower mantle, Science, 317, 1,740\u20131,743, https:\/\/doi.org\/10.1126\/science.1144997<\/a>.<\/p>\n Shephard, G. E., et al. (2021), Seismological expression of the iron spin crossover in ferropericlase in the Earth\u2019s lower mantle, Nat. Commun., 12, 5905, https:\/\/doi.org\/10.1038\/s41467-021-26115-z<\/a>.<\/p>\n Tsuchiya, T., et al. (2006), Spin transition in magnesiow\u00fcstite in Earth\u2019s lower mantle, Phys. Rev. Lett., 96, 198501, https:\/\/doi.org\/10.1103\/PhysRevLett.96.198501<\/a>.<\/p>\n Wentzcovitch, R. M., et al. (2009), Anomalous compressibility of ferropericlase throughout the iron spin cross-over, Proc. Natl. Acad. Sci. U. S. A., 106, 8,447\u20138,452, https:\/\/doi.org\/10.1073\/pnas.0812150106<\/a>.<\/p>\n Wu, Z., and R. M. Wentzcovitch (2014), Spin crossover in ferropericlase and velocity heterogeneities in the lower mantle, Proc. Natl. Acad. Sci. U. S. A., 111, 10,468\u201310,472, https:\/\/doi.org\/10.1073\/pnas.1322427111<\/a>.<\/p>\n Wu, Z., et al. (2013), Elastic anomalies in a spin-crossover system: Ferropericlase at lower mantle conditions, Phys. Rev. Lett., 110, 228501, https:\/\/doi.org\/10.1103\/PhysRevLett.110.228501<\/a>.<\/p>\n Zhuang, J., and R. Wentzcovitch (2024), Fe2+ partitioning in Al-free pyrolite: Consequences for seismic velocities and heterogeneities, Geophys. Res. Lett., 51, e2024GL108967, https:\/\/doi.org\/10.1029\/2024GL108967<\/a>.<\/p>\n Author Information<\/p>\n Renata Wentzcovitch ([email\u00a0protected]<\/a>), Department of Earth and Environmental Science and Department of Applied Physics and Applied Mathematics, Lamont-Doherty Earth Observatory, Columbia University, New York; Laura Cobden, Utrecht University, Utrecht, Netherlands; Christine Houser, Earth-Life Science Institute, Tokyo; Grace Shephard, Research School of Earth Sciences, Australian National University, Canberra; and Jingyi Zhuang, Department of Earth and Environmental Science, Lamont-Doherty Earth Observatory, Columbia University, New York<\/p>\n Citation:\u00a0Wentzcovitch, R., L. Cobden, C. Houser, G. Shephard, and J. Zhuang (2026), A quiet quantum revolution in Earth\u2019s deep interior,\u00a0Eos, 107, https:\/\/doi.org\/10.1029\/2026EO260199<\/a>. Published on 22 June 2026.<\/p>\n![\"Two-panel](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/06\/iron-ions-spin-mantle-transition-1024x439.png\")
Fig. 1. (a) Iron ions in lower mantle minerals can change from a high-spin state to a low-spin state with increasing pressure and depth. The competition between crystal field splitting (Ec, favoring low spin) and Hund\u2019s rule coupling (EX, favoring high spin) determines which state is stable. Black arrows and horizontal lines represent electrons and d electronic subshells, respectively, in an iron ion. (b) Changes in the populations of high-spin and low-spin iron occur continuously over a broad, transitional pressure range, the width of which increases with temperature [Zhuang and Wentzcovitch<\/a>, 2024]. n refers to the proportion of low-spin to high-spin iron ions present, with 0 representing no low-spin ions and 1 representing all low-spin ions. Dashed lines refer to analyses by Lin et al.<\/a> [2007] and Komabayashi et al.<\/a> [2010]. \u201cGeotherm\u201d refers to the geothermal gradient.<\/p>\n![\"Data](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/06\/volume-change-ferropericlase-pressure-iron-spin.png\")
Fig. 2. Curves showing calculated volume changes (\u0394V) of ferropericlase (with an iron mole fraction, XFecalc, of 18.75%) with increasing pressure for different proportions (n) of low-spin (LS) versus high-spin (HS) iron atoms. Data points (black plus signs) represent experimental data from Lin et al.<\/a> [2005] for ferropericlase with an XFeexp of 17%. These data show that the mineral\u2019s volume decreases faster in the mixed-spin, or iron spin crossover (ISC), region between about 40 and 60 gigapascals (GPa) than outside this region. This effect decreases P wave speeds across a broad depth interval in the mantle while leaving S wave speeds nearly unchanged, producing characteristic P-S wave patterns in 3D tomography.<\/p>\n![\"Data](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/06\/velocities-seismic-waves-mantle-pressure.png\")
Fig. 3. Predicted velocities for compressional (VP), shear (VS), and bulk (V\u03c6) seismic waves in pyrolite mantle rock (75% bridgmanite, 18% ferropericlase, and 7% davemaoite by volume, with XFe = 10%) compared with velocities from the preliminary reference Earth model (dots). \u03c1 is the density of the rock. Light and dark shades of each color represent, respectively, velocities without and with accounting for the ISC in ferropericlase. Solid and dashed lines indicate the presence and absence, respectively, of iron partitioning effects. The mixed-spin region of ferropericlase spans most of the lower mantle, producing subtle but widespread effects on seismic velocities [Zhuang and Wentzcovitch<\/a>, 2024]. This diffuse depth range explains why 1D global models do not exhibit obvious anomalies. Credit: Zhuang and Wentzcovitch<\/a> [2024], CC BY 4.0<\/a><\/p>\n![\"Multipart](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2026\/06\/global-shear-compressional-wave-tomography-1024x761.png\")
Fig. 4. Comparisons of global S and P wave tomography using the \u201cvote map\u201d method show the predicted signature of the ISC in different parts of the mantle, including seismically fast, cold, ferropericlase-rich slabs (top and middle) and seismically slow, hot plume areas (bottom). These regions exhibit weaker P wave anomalies but stronger S wave anomalies. The P-S decorrelation is consistent with predictions from mineral physics and is observed in both the vote map [Shephard et al.<\/a>, 2021] and full-waveform models [Cobden et al.<\/a>, 2024]. Credit: Adapted from Shephard et al.<\/a> [2021], CC BY 4.0<\/a><\/p>\n