Earth’s core is cooling, and it has been since the planet formed roughly 4.5 billion years ago. This slow loss of internal heat, called secular cooling, is not a sudden development or a crisis. It is one of the most fundamental processes shaping our planet, responsible for everything from plate tectonics to the magnetic field that shields us from solar radiation. The rate and consequences of that cooling, however, remain subjects of active research and surprisingly vigorous debate among geophysicists.
Where the Heat Comes From and How It Escapes
Earth’s deep interior holds heat from two main sources. The first is primordial heat, left over from the violent accretion and differentiation that formed the planet. When the early Earth’s iron sank to form the core, it released enormous gravitational energy as heat. The second source is radioactive decay. Elements like uranium, thorium, and potassium break down over geological time, generating heat in the process. Most of that radioactive heat production happens in the mantle and crust, but experiments have shown that potassium can dissolve into iron-sulfide melts at high pressures and temperatures, meaning potassium-40 likely serves as a meaningful heat source inside the core itself.1PubMed. Experimental evidence that potassium is a substantial radioactive heat source in planetary cores
The core cannot radiate heat directly into space. Instead, heat creeps outward through the mantle, a process governed by the thermal boundary at the core-mantle boundary, or CMB. This boundary sits about 2,900 kilometers below the surface and represents one of the most dramatic thermal contrasts anywhere in the planet. The core side is molten iron at thousands of degrees; the mantle side is solid rock that is comparatively cooler. Heat flows across this boundary at an estimated rate of roughly 6 to 17 terawatts, depending on the model and the assumptions fed into it.2Geophysical Journal International. Changes in core–mantle boundary heat flux patterns throughout the supercontinent cycle For perspective, the entire human civilization consumes energy at a rate of about 18 terawatts, so the heat leaking out of Earth’s core is in the same ballpark.
That heat does not escape uniformly. The mantle above the CMB is not a passive conductor. Subducting tectonic plates that sink to the deep mantle act as cold slabs that pull heat out of the core more efficiently in some regions. Large structures sitting on the CMB, sometimes called “blobs” by researchers, insulate parts of the boundary and reduce local heat flow. Modeling work has found that these sinking slabs are central regulators of deep mantle heat transfer, and their effect on deep mantle structure kicks in after a delay of roughly 50 to 250 million years.3Earth and Space Science Open Archive. Influence of secular cooling on core-mantle boundary heat flux and mantle plume temperature over 1.8 billion years In other words, what happens on the surface with plate tectonics today will influence how fast the core cools hundreds of millions of years from now.
The Inner Core as a Thermometer of Cooling
The most tangible evidence that the core is cooling is the existence of the inner core itself. Earth’s core is divided into two parts: a liquid outer core of molten iron alloy and a solid inner core at the center. The inner core exists because the center of the planet has cooled below the melting point of iron at those extreme pressures. As the core continues to lose heat, the inner core grows, its surface slowly crystallizing outward as liquid iron freezes onto it.
The inner core has not always been there. Paleomagnetic evidence suggests it nucleated, or first began to solidify, sometime between one billion and 1.5 billion years ago. Researchers identified this window by looking at the ancient magnetic field recorded in rocks. The most prominent change in the paleomagnetic record is an increase in both average field strength and variability during that interval, which fits neatly with the idea of a new energy source switching on inside the core as crystallization began.4PubMed. Palaeomagnetic field intensity variations suggest Mesoproterozoic inner-core nucleation More recent thermal conductivity measurements of iron-nickel-silicon alloys at core pressures have supported a similar age estimate of about 0.8 to 1.3 billion years.5Geophysical Research Letters. Moderate Thermal Conductivity of Fe‐Ni‐Si Alloy at Earth’s Core Conditions: Implications for Core Thermal Evolution and Geodynamo
That the inner core is geologically young compared to the planet is itself informative. Earth spent its first three billion years or more without a solid core at all. The cooling had to proceed long enough for temperatures at the center to drop below the freezing threshold before crystallization could begin. The exact timing depends on how well the core conducts heat, which is where one of the field’s biggest arguments comes in.
The Thermal Conductivity Problem
How quickly the core loses heat depends heavily on how well its iron alloy conducts that heat. If the thermal conductivity of the core is high, heat escapes faster, and the core cools more rapidly. If it is low, the core retains heat longer. This number might sound like a technical detail, but it ripples through nearly every question about Earth’s deep history, including when the inner core formed, how old the magnetic field is, and how much longer both can last.
For years, researchers assumed the core had a relatively modest thermal conductivity. Then, starting around 2012, a series of experiments and theoretical calculations suggested the conductivity could be much higher than previously thought, potentially two to three times the old estimates. If those high values were correct, the core would be losing heat so fast that maintaining a magnetic field before the inner core formed would have been extremely difficult. The core simply would not have had enough thermal energy to drive the convection needed to generate a dynamo for billions of years before crystallization began.
This created a genuine paradox, because we have paleomagnetic evidence that Earth has had a magnetic field for at least 3.5 billion years, long before the inner core existed. One group of researchers measured the thermal conductivity of solid iron at inner core conditions and found a total value of about 200 watts per meter per kelvin when accounting for both electron and lattice vibration contributions, noting that the lattice component becomes significant under extreme pressure and should not be ignored.6Chinese Physics B. Thermal conductivity of iron under the Earth’s inner core pressure But the inner core is solid; the outer core, where the dynamo actually operates, is liquid and has a different composition. Recent measurements of iron-nickel-silicon alloy compressed to outer core pressures found a more moderate thermal conductivity of about 73 watts per meter per kelvin for the liquid outer core.5Geophysical Research Letters. Moderate Thermal Conductivity of Fe‐Ni‐Si Alloy at Earth’s Core Conditions: Implications for Core Thermal Evolution and Geodynamo That moderate value helps resolve the paradox: it means the heat loss through conduction alone at the CMB would be around 11 terawatts, which is less than the estimated total CMB heat flow of about 15 terawatts. The difference between the two represents heat that can drive convection and, by extension, the magnetic field.
The debate is far from settled. Different research groups, using different compositions and experimental setups, have reported values spanning a wide range. High-pressure experiments using a diamond-anvil cell remain the primary tool for probing these conditions, but replicating the core’s exact chemistry and temperature simultaneously is extraordinarily difficult.7PubMed Central. The thermal conductivity of the Earth’s core and implications for its thermal and compositional evolution Until the community converges on a narrower range, the pace of core cooling remains uncertain by a factor of roughly two to three.
How Cooling Powers the Magnetic Field
Earth’s magnetic field is generated by the geodynamo, a self-sustaining process in which convective motion of the electrically conducting liquid iron in the outer core produces electric currents that generate magnetic fields. This convection has two drivers, both consequences of the core cooling down. The first is straightforward thermal convection: the outer core loses heat to the mantle, the cooled fluid near the CMB sinks, and hotter fluid rises to replace it. The second driver is compositional convection, and it is tied directly to the growth of the inner core.
When iron crystallizes onto the inner core, it preferentially incorporates the heavier iron while rejecting lighter elements like sulfur, silicon, and oxygen into the surrounding liquid. This makes the liquid near the inner core boundary buoyant relative to the fluid above it, driving vigorous upward flow. Both latent heat released during crystallization and this chemical separation contribute to powering the dynamo.8Comptes Rendus Geoscience. The inner core and the geodynamo In fact, many geophysicists consider the compositional contribution to be more efficient at driving convection than the thermal one.9Physics of the Earth and Planetary Interiors. Geomagnetic field and the growth of the Earth’s inner core: Past, present and future
This creates a somewhat counterintuitive situation: the core cooling down is what keeps the magnetic field alive. If the core stopped cooling, the inner core would stop growing, compositional buoyancy would cease, and the thermal driving force would eventually weaken. The magnetic field would decay and disappear. So when people ask whether core cooling is bad news, the honest answer is that core cooling is what makes Earth habitable in the first place.
Before the Inner Core Existed
If the inner core is only about a billion years old, but the magnetic field has persisted for over three billion years, something else must have powered the dynamo for a very long time before crystallization began. This is one of the trickier puzzles in deep Earth science. Without compositional convection from inner core growth, the early dynamo had to run entirely on thermal convection, meaning the core had to lose heat fast enough to keep the liquid churning.
One proposed solution involves the style of tectonics on early Earth. Modeling suggests that maintaining a thermally driven dynamo before inner core nucleation, with realistic mantle and core temperatures, requires that Earth operated in a “sluggish-lid” mode early in its history rather than the vigorous plate tectonics we see today. This mode depends on the existence of an asthenosphere, the weak layer in the upper mantle that allows tectonic plates to move.10PubMed Central. Coupled fates of Earth’s mantle and core: Early sluggish-lid tectonics and a long-lived geodynamo Under sluggish-lid conditions, the mantle would still extract enough heat from the core to sustain convection without requiring extremely high core temperatures that would be difficult to justify physically.
Other researchers have explored whether radioactive heating within the core itself could have supplemented the thermal budget. If enough potassium-40 dissolved into the core during Earth’s formation, its decay heat could have partially offset the cooling requirement, effectively subsidizing the dynamo during the long era before inner core crystallization switched on the compositional engine.
The Role of the Mantle as a Bottleneck
A key insight from decades of research is that the mantle, not the core, controls how fast the core cools. The core is a good thermal conductor and can deliver heat to the CMB efficiently. But the mantle above is a poor conductor and convects sluggishly by comparison. The rate at which the mantle carries that heat upward to the surface is what ultimately sets the pace of core cooling. This is sometimes described as the mantle being a “thermal bottleneck.”
Estimates for the present-day mantle cooling rate span a wide range, from about 7 to 210 kelvins per billion years.2Geophysical Journal International. Changes in core–mantle boundary heat flux patterns throughout the supercontinent cycle That enormous spread reflects genuine uncertainty about mantle viscosity, composition, and the efficiency of convection at different depths. But even at the high end, this is an incredibly slow process. The core is not going to freeze solid anytime relevant to human civilization. The inner core grows at a rate measured in fractions of a millimeter per year. At that pace, the core will remain partially liquid for billions of years to come.
Mineral physics also matters here. At the base of the mantle, a phase transition in the dominant mineral, magnesium silicate perovskite, converts it into a denser form called post-perovskite. This transition releases energy and destabilizes the thermal boundary layer, which increases heat flow out of the core and raises temperatures in the interior mantle.11Geophysical Research Letters. Effects of a perovskite‐post perovskite phase change near core‐mantle boundary in compressible mantle convection Experiments pinning down this phase transition boundary suggest a CMB temperature of roughly 3,700 kelvins, somewhat lower than earlier estimates, with a minimum global heat flow from the core of about 6.6 terawatts.12Earth and Planetary Science Letters. Determination of post-perovskite phase transition boundary up to 4400 K and implications for thermal structure in D″ layer That lower bound on heat flow is important because it confirms that a significant amount of heat continues to escape the core even under the most conservative estimates.
What a Cooled Core Looks Like on Mars
If you want to see what happens when a rocky planet’s core cools past a critical threshold, Mars offers a cautionary example. Crustal rocks on Mars carry ancient magnetic signatures, evidence that the planet once had a global magnetic field generated by a core dynamo. But that dynamo died roughly four billion years ago.13Journal of Geophysical Research: Planets. Hydrogenation of the Martian Core by Hydrated Mantle Minerals With Implications for the Early Dynamo Without a magnetic field, Mars lost its protection from the solar wind, which gradually stripped away much of its atmosphere. The thin atmosphere that remains cannot support liquid water on the surface, contributing to the barren landscape we see today.
Mars is smaller than Earth, which matters. A smaller planet has a higher surface-area-to-volume ratio, meaning it loses heat faster relative to its total thermal budget. Mars also lacks plate tectonics, which on Earth helps regulate the pace of mantle cooling. The standard model for Mars’s dynamo death involves the heat flow at the Martian core-mantle boundary dropping below the threshold needed to sustain thermal convection. Some researchers have proposed an alternative: that hydrogen from hydrated mantle minerals could have dissolved into the Martian core, changing its composition in ways that impeded convective motions even while heat flow was still high enough to theoretically sustain a dynamo.13Journal of Geophysical Research: Planets. Hydrogenation of the Martian Core by Hydrated Mantle Minerals With Implications for the Early Dynamo
Earth is not Mars. Our planet is larger, retains more heat, has an actively growing inner core pumping compositional energy into the outer core, and has plate tectonics helping to regulate the whole thermal system. But Mars demonstrates that core cooling is not purely academic. For a rocky planet, the long-term trajectory of its core temperature determines whether it gets to keep a magnetic field, an atmosphere, and potentially life on its surface.
How Fast Is the Cooling Really Happening
Pinning a single number on the rate of core cooling is surprisingly hard, because different approaches yield different answers. Mantle convection models produce present-day CMB heat flux values ranging from about 7 terawatts to 15 terawatts. Three-dimensional spherical models of mantle convection that account for the post-perovskite phase transition find that the best agreement with seismic observations depends on the steepness of that transition: at a moderate slope, the best fit comes with core heat flow around 15 terawatts, while at a steeper slope the best match drops to about 7.5 terawatts.14PubMed Central. Topology of the postperovskite phase transition and mantle dynamics That factor-of-two uncertainty is large, but both ends of the range tell the same story: heat is leaving the core at a geologically significant rate.
The heat flux is also not spatially uniform. Modeling of how mantle upwellings and downwellings create localized heat flux anomalies at the CMB shows that the ratio of the peak-to-peak variation to the average heat flux can easily exceed a factor of two, and under some assumptions about material properties can reach values above 30.15arXiv Central. Changes in core-mantle boundary heat flux patterns throughout the supercontinent cycle This means some regions of the core are cooling much faster than others at any given time, influenced by the geometry of subduction zones and deep mantle structures far above.
For a given set of material properties, these spatial variations themselves change by 30 to 50 percent over geological time, driven by the supercontinent cycle. When continents cluster together, subduction patterns shift, the distribution of cold slabs at the base of the mantle rearranges, and the pattern of heat extraction from the core changes accordingly. Core cooling is real and ongoing, but it is not a simple, steady draining of a thermal battery. It is a dynamic process modulated by everything from mineral phase transitions at the CMB to the arrangement of continents at the surface.
The Far Future of Earth’s Core
At current rates, the inner core will continue to grow for billions of years. If nothing else changes, eventually the entire core would solidify and compositional convection would stop. The magnetic field would weaken and eventually cease, as it did on Mars. But “eventually” here means something like two to four billion years in the future, depending on which thermal conductivity and heat flux values turn out to be correct. By that time, the Sun will be well on its way to becoming a red giant, so the loss of the magnetic field will be the least of Earth’s problems.
There is also the question of whether the cooling will remain gradual or could produce surprises. Some modeling work has suggested that as the inner core grows, changes in its crystallization pattern could alter the style of convection in the outer core, potentially leading to periods of unusual magnetic field behavior. The paleomagnetic record already shows that the field’s reversal frequency has varied dramatically over geological time, with long periods of stability punctuated by intervals of frequent reversals. Whether some of that variability connects to changes in the cooling rate or inner core growth dynamics is an open question.
What is not in question is the basic narrative. Earth formed hot, has been losing that heat ever since, and the consequences of that cooling include the solidification of the inner core, the convective motions that generate the magnetic field, and indirectly the plate tectonics that recycle the surface. The core is cooling. That cooling is one of the engines that makes Earth the planet it is.