6 Minutes
New constraint on core chemistry: carbon as a freezing agent
Cartoon of the Earth with cutaway showing the mantle and inner and outer core. Magnetic field lines produced by the geodynamo extend into space and interact with the solar wind.
A multinational team from the University of Oxford, University of Leeds and University College London reports that carbon could be the decisive ingredient that allowed Earths molten outer core to crystallize into the solid inner core we observe today. Published in Nature Communications on 4 September 2025, the study derives a quantitative constraint on core composition: about 3.8 percent of the core's mass as carbon would reduce the degree of supercooling required to initiate freezing to values consistent with palaeomagnetic and geophysical evidence.
Understanding why and how the inner core formed matters because inner core growth drives the convection in the outer core that powers the geodynamo — the magnetic field that shields Earth from solar and cosmic particle radiation and has helped sustain life on the surface.
Scientific background: supercooling, nucleation and core composition
The formation of a solid inner core is not simply a question of temperature crossing a melting point. Instead, nucleation physics controls the onset of crystallization. Molten iron alloys can be supercooled well below their equilibrium melting temperature before crystals appear; classic analogies include cloud droplets that can cool far below 0 degrees C without freezing.
Earlier theoretical work indicated that a pure iron core would require extreme supercooling — on the order of 800–1000 °C — to nucleate crystals. That scenario conflicts with geophysical constraints: if the Earths core had been that deeply supercooled, models predict a very rapid inner core growth and collapse of the planetary magnetic field, neither of which match the paleomagnetic or seismic record. Instead, independent studies suggest past supercooling was limited to roughly 250 °C below the melting temperature.
This discrepancy points to the importance of alloying elements in the core. Seismology already implies the core is less dense than pure iron, so elements such as silicon, sulphur, oxygen and carbon have long been proposed as light components. The new study explicitly models how these light elements affect nucleation and the freezing pathway at pressures and temperatures equivalent to the inner core boundary.
Methods: atomic-scale simulations of nucleation under core conditions
The team used large-scale atomic simulations, tracking roughly 100,000 atoms under the extreme pressure and temperature conditions of Earths deep interior. These molecular-scale calculations allow direct estimation of nucleation rates: how frequently small crystalline clusters form and grow out of a supercooled liquid.
By varying the alloy chemistry, the researchers quantified how each candidate light element changes the required supercooling for nucleation. They tested mixtures with representative concentrations of silicon, sulphur, oxygen and carbon to establish which compositions can both nucleate at modest supercooling and match the observed size of the inner core today.
The simulations show that silicon and sulphur, contrary to some expectations, tend to retard nucleation and therefore require larger degrees of supercooling. Carbon, however, has the opposite effect: it catalyses nucleation in these iron-rich alloys, lowering the barrier to freezing and permitting inner core formation with far less supercooling.
Key results and implications for Earth evolution
When the team modeled a core containing about 2.4 percent carbon by mass, the predicted required supercooling dropped to roughly 420 °C — an improvement but still higher than palaeomagnetic limits. Extrapolating the simulations to a core composition with approximately 3.8 percent carbon produced a required supercooling near 266 °C, consistent with independent constraints and with the present-day size of the inner core.
This composition is the only one identified in the study that simultaneously explains both nucleation behavior and seismically inferred inner core dimensions. The result therefore implies carbon may be a more significant component of Earths core than many previous models assumed, and that carbon played a crucial role in allowing the inner core to nucleate in Earths early history.
Additional implications include:
Magnetic field stability
Inner core nucleation and subsequent growth provide energy and compositional buoyancy that drive outer-core convection and sustain the geodynamo. A delayed or absent inner core would alter the magnetic field history, with consequences for atmospheric retention and surface habitability.
Constraints on planetary formation and volatile delivery
If Earths core contains several percent carbon, this constrains models of accretion and volatile partitioning during planetary formation, and may inform comparisons with other terrestrial bodies.
The study also finds that inner core freezing could occur without external nucleation seeds, since candidate particles tested in prior models would melt or dissolve at core conditions, reinforcing the importance of alloy chemistry itself as the primary control.
Expert Insight
Dr Amelia Reyes, a planetary physicist not involved in the study, notes that the work highlights how atomic-scale physics scales up to shape planetary evolution. She comments that the carbon-rich scenario is plausible given constraints from meteoritic chemistry and offers a coherent link between seismic observations, paleomagnetism and core thermodynamics. In her view, the result motivates lab and observational tests that can further narrow the carbon budget of the deep Earth.
Lead researchers emphasize the broader significance of linking microscopic nucleation kinetics to macroscopic planetary behavior. They also underline the value of high-fidelity simulations in exploring regions of Earth that are inaccessible to direct sampling. The research was supported by the Natural Environment Research Council (NERC).
Conclusion
New atomic-scale simulations indicate that a modest but significant carbon fraction in Earths core — on the order of 3.8 percent by mass — would have lowered the supercooling threshold enough to permit inner core nucleation consistent with geophysical constraints. This finding reframes the role of carbon in core chemistry, strengthens a mechanistic connection between core composition and magnetic field evolution, and provides a testable target for future experiments and planetary-formation models. By clarifying how the solid inner core could form, the work advances our understanding of the deep Earth and the long-term stability of the geodynamo.
Comments