Silicate Dynamo: Planetary Magnetic Fields

• The Silicate Dynamo Hypothesis is the proposal that convecting silicate melts, particularly in basal magma oceans, can produce self-sustaining magnetic fields.
• It identifies key conditions such as high electrical conductivity, vigorous convection, and a magnetic Reynolds number above roughly 40 as essential for dynamo action.
• The hypothesis is applied to early Earth and massive super-Earths, highlighting diverse geodynamic regimes and contrasting transport properties in silicate melts.

Searching arXiv for the cited papers to ground the article in current records. arXiv lookup: (Zaghoo, 2019, Lherm et al., 2024, Mazevet et al., 2014, Bolis et al., 2016, Pahlevan et al., 2010, Dragulet et al., 1 Aug 2025). The Silicate Dynamo Hypothesis is the proposition that a planetary magnetic field can be generated in convecting silicate melt, rather than exclusively in a liquid metallic core. In current research usage, the hypothesis usually refers to a basal magma ocean—a layer of silicate melt sitting directly atop the metallic core and beneath the mostly solid mantle—or, more broadly, to a deep magma ocean whose electrical conductivity, convective vigor, and thickness permit self-sustaining magnetohydrodynamic induction. The hypothesis has been developed in two partially distinct settings: massive super-Earths, where deep silicate melt may be the only possible dynamo layer once iron cores become solid or subadiabatic, and the early Earth, where a long-lived, iron-rich basal magma ocean has been proposed as a precursor to the core dynamo (Zaghoo, 2019, Lherm et al., 2024, Dragulet et al., 1 Aug 2025).

1. Physical basis of a silicate dynamo

A silicate dynamo requires the same elements as any fluid dynamo: an electrically conducting fluid, sustained convection in a rotating shell, and a sufficiently large magnetic Reynolds number. In the formulation used for basal magma oceans, the controlling induction parameter is

$R_m = \mu_0\, v\, l\, \sigma,$

where $\mu_0$ is the magnetic permeability of free space, $v$ is a characteristic flow speed, $l$ is a characteristic layer thickness, and $\sigma$ is the electrical conductivity. Numerical dynamo studies adopted in the Earth basal magma ocean literature use a canonical threshold of about $R_m \gtrsim 40$ for self-sustaining field generation (Dragulet et al., 1 Aug 2025).

Thermal viability is a separate requirement. For a basal magma ocean, convection is sustained only if the total heat flux out of the layer exceeds conductive transport along the adiabat,

$$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$

with $k$ the thermal conductivity and $\nabla T_{\rm ad}$ the adiabatic temperature gradient. In coupled core–basal-magma-ocean models, dynamo activity is then assessed using both magnetic Reynolds numbers and entropy budgets; a positive entropy available for dissipation, $E_\Phi>0$, is taken as a necessary condition for ohmic dissipation and therefore for dynamo operation. Core-dynamo suppression is commonly formulated by comparing heat flow across the core–mantle boundary with adiabatic conduction, $\mu_0$0 (Lherm et al., 2024).

This framework immediately implies that the silicate dynamo problem is materially contingent. A deep silicate layer may convect vigorously and still fail as a dynamo if $\mu_0$1 is too low, while a highly conducting silicate layer may still fail if $\mu_0$2 is large enough to suppress buoyant overturn. Much of the literature therefore turns on how pressure, temperature, phase state, Fe enrichment, and spin state alter $\mu_0$3 and $\mu_0$4.

2. Thermodynamic preconditions and planetary architecture

For terrestrial-like super-Earths in the range $\mu_0$5–$\mu_0$6, interior models with an Earth-like core–mantle mass ratio of $\mu_0$7 show that the key thermodynamic comparison is between the mantle adiabat and silicate melting curves. Using a Simon-law fit for MgSiO$\mu_0$8 perovskite melting and an adiabatic gradient

$\mu_0$9

the decisive result is a pressure crossover: “the adiabat becomes steeper than the liquidus above 4–5 Mbar for both SiO$v$0 and MgSiO$v$1.” In practice this yields deep basal magma oceans whose thickness increases strongly with mass: for about $v$2, a basal magma ocean about $v$3 km thick at the base of the mantle, and for about $v$4, a basal magma ocean exceeding $v$5 km, or about $v$6 of planetary radius. The same calculations predict that iron cores behave oppositely with increasing mass: planets larger than about $v$7 develop an entirely solid Fe core, while planets more massive than about $v$8 have internal cores that are subadiabatic and non-convecting because iron thermal conductivity rises strongly with pressure (Zaghoo, 2019).

These mass thresholds produce a characteristic super-Earth architecture. Between roughly $v$9 and $l$0, the expected state is Earth-like: liquid outer Fe core, inner solid core, and an iron-core dynamo. Between about $l$1 and $l$2, some liquid Fe may remain, but thermal convection becomes increasingly difficult. Above about $l$3, the predicted structure is an “inverted” one: solid iron core plus deep basal liquid silicate mantle. In that regime, if any dynamo exists, it must be silicate rather than metallic (Zaghoo, 2019).

A distinct thermodynamic setting is the post-impact early Earth. Giant-impact energetics imply a global magma ocean and a silicate vapor atmosphere; for a Mars-mass impactor the mean temperature rise is estimated as about $l$4 K, and for a $l$5 impactor about $l$6 K. The convecting silicate envelope is represented by a specific entropy of about $l$7, and radiative cooling from a photosphere near $l$8 K drives vigorous convection throughout the system. Although this work does not build a magnetic model, it provides a thermodynamic backdrop for whole-mantle magma-ocean dynamos and for later basal-magma-ocean scenarios (Pahlevan et al., 2010).

3. Electrical conductivity, metallization, and transport constraints

The central empirical dispute in the silicate dynamo literature concerns the transport properties of silicate liquids at extreme pressure and temperature. First-principles work on SiO$l$9 found that silica remains “a poor electrical conductor up to 10 Mbar due to an increase in the Si–O coordination with pressure.” Along isotherms, conductivity does not increase monotonically with compression; instead, increased Si–O coordination opens a more pronounced pseudo-gap around the Fermi energy. At conditions relevant to the deepest mantles of super-Earths, the reported DC electrical conductivity is of the order of $\sigma$0–$\sigma$1, and “silica will not contribute significantly to either magnetic field generation in Earth like exoplanets up to several times the Earth mass.” Only above about $\sigma$2–$\sigma$3 K does SiO$\sigma$4 reach conductivity comparable to that of a poor metal (Mazevet et al., 2014).

Laser-driven decaying-shock experiments on MgO, MgSiO$\sigma$5, and Mg$\sigma$6SiO$\sigma$7 sharpened that constraint by showing that melting and metallization do not coincide. In MgO, a melting signature was observed at $\sigma$8 TPa and $\sigma$9 K, while reflectivity begins to increase only near $R_m \gtrsim 40$0 TPa and $R_m \gtrsim 40$1 K. For MgSiO$R_m \gtrsim 40$2 and Mg$R_m \gtrsim 40$3SiO$R_m \gtrsim 40$4, increasing reflectivity of the liquids begins at $R_m \gtrsim 40$5 TPa and $R_m \gtrsim 40$6 K, and at $R_m \gtrsim 40$7 TPa and $R_m \gtrsim 40$8 K, respectively. The paper’s explicit inference is that “poorly electrically conducting liquid” exists in the phase diagram in the vicinity of the melting line, a result described as having “important implications for the generation of dynamos in Super-earths mantles” (Bolis et al., 2016).

A different picture emerges for iron-rich silicate liquids at Earth basal-magma-ocean conditions. Ab-initio molecular-dynamics calculations at $R_m \gtrsim 40$9–$$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$0 GPa and $$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$1–$$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$2 K find that a pyrolitic liquid with $$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$3 already has electronic conductivity greater than $$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$4 S/m, while an Fe-dominated liquid with $$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$5 exceeds $$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$6 S/m. The density of states at the Fermi level increases roughly linearly with $$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$7, and the electrical conductivity scales approximately as $$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$8. In the pyrolitic case, electrons contribute $$Q_{\rm total} > Q_{\rm cond} = 4\pi r^2\,k\,\nabla T_{\rm ad},$$9–$k$0 of total conductivity; at high Fe content, they contribute more than $k$1. At the same time, total thermal conductivity remains modest, with $k$2 and $k$3 for Fe-rich silicate liquid (Dragulet et al., 1 Aug 2025).

Taken together, these results define two materially different regimes. Mg-rich endmembers near their melting curves tend to be weakly conducting, whereas Fe-enriched multicomponent basal magma ocean liquids can enter a conductivity range explicitly treated as dynamo-capable in Earth models. This suggests that the hypothesis is not a statement about “silicate” as a single class of fluid, but about specific compositional and thermodynamic subspaces.

4. Earth basal magma ocean models

The Earth-specific form of the hypothesis is motivated by the timing problem of the geodynamo. Paleomagnetic evidence indicates a strong field at at least $k$4 Ga, whereas inner-core growth is likely younger than $k$5 Ga or at most $k$6 Ga. If core thermal conductivity is large, a purely thermal early core dynamo becomes difficult, and an alternative source of magnetic induction is required (Lherm et al., 2024, Dragulet et al., 1 Aug 2025).

In one recent formulation, the basal magma ocean is modeled as an ideal MgSiO$k$7–FeO mixture coupled thermally to the core and solid mantle. The model integrates internal structure, uses energy budgets for mantle, basal magma ocean, and core, and evaluates convective stability through mass-anomaly fluxes and dynamo activity through entropy budgets and magnetic Reynolds numbers. Its conservative nominal case begins with an initial core–mantle-boundary temperature of $k$8 K, an initial basal magma ocean thickness of $k$9 km, and an initial FeO mass fraction of $\nabla T_{\rm ad}$0 wt%. The basal magma ocean remains fully convective while it exists, but the core is initially stably stratified; in the nominal model, the basal magma ocean hosts a dynamo from $\nabla T_{\rm ad}$1 to about $\nabla T_{\rm ad}$2 Gyr, the core dynamo begins at about $\nabla T_{\rm ad}$3–$\nabla T_{\rm ad}$4 Gyr, the basal magma ocean fully crystallizes by $\nabla T_{\rm ad}$5 Gyr, and inner-core nucleation occurs at about $\nabla T_{\rm ad}$6 Gyr. The same model states that a basal magma ocean dynamo requires a conductivity of at least $\nabla T_{\rm ad}$7; below that, convection and positive entropy production may persist, but $\nabla T_{\rm ad}$8 does not exceed the threshold for self-sustaining induction (Lherm et al., 2024).

A complementary Earth model uses ab-initio transport properties in a time-dependent thermal–chemical evolution scheme. With a fixed bulk-silicate-Earth Fe content, the modeled basal magma ocean dynamo satisfies $\nabla T_{\rm ad}$9 for only the first $E_\Phi>0$0 Gyr of Earth’s history. When Fe enrichment during crystallization is included, conductivity rises, the minimum thickness required for dynamo action decreases from about $E_\Phi>0$1 km to less than $E_\Phi>0$2 km as $E_\Phi>0$3 increases from $E_\Phi>0$4 to $E_\Phi>0$5, and the lifetime of the silicate dynamo extends to about $E_\Phi>0$6 Gyr under mixing-length scaling. Under the more restrictive Coriolis–inertial–Archimedean scaling, the lifetime shortens by about $E_\Phi>0$7 Gyr. The same study concludes that Fe enrichment raises conductivity into a regime where a thermal silicate dynamo is robust while leaving thermal conductivity low enough that convection is not choked off (Dragulet et al., 1 Aug 2025).

The Earth literature therefore treats the basal magma ocean not merely as a molten residue, but as a distinct dynamical layer that can bridge the interval between post-accretion magma-ocean states and the later thermo-compositional core dynamo. A plausible implication is that the early geomagnetic field may have involved a transition from silicate to metallic induction rather than an uninterrupted core-only history.

5. Super-Earth formulations and mineral-physics classification

In super-Earth studies, the hypothesis arises from a different constraint: the progressive failure of metallic core dynamos with increasing planetary mass. For planets larger than about $E_\Phi>0$8, high iron thermal conductivity keeps the core subadiabatic and non-convecting; for planets larger than about $E_\Phi>0$9, the Fe liquidus no longer intersects the planetary temperature profile, so the core is entirely solid. At the same time, the shallow adiabaticity of perovskite or stishovite silicates relative to their liquidus permits deep, long-lived basal magma oceans. The paper states explicitly that these deep magma oceans raise “an intriguing prospect of a possibly convecting and conducting silicate shell that, in principle, is capable of sustaining a dynamo” (Zaghoo, 2019).

The same super-Earth analysis is, however, structurally skeptical. Shock experiments show that silicates metallize along the Hugoniot above about $\mu_0$00 TPa, but at $\mu_0$01–$\mu_0$02 Mbar and $\mu_0$03–$\mu_0$04 K the reported conductivity of mineral silicates is only “few tens S/cm,” and even in this “metallic regime” the liquid is described as “at best, a poor semiconductor.” Because liquid iron is about $\mu_0$05 times more conductive, a silicate dynamo would have to compensate with much larger flow speeds or shell thicknesses. The paper further notes that an underlying more conducting iron shell could severely attenuate any field through an electromagnetic skin-depth effect (Zaghoo, 2019).

Experimental work on MgO, MgSiO$\mu_0$06, and Mg$\mu_0$07SiO$\mu_0$08 reinforces that caution by locating a regime of poorly conducting liquids near the melting line, rather than immediate metallization upon melting. First-principles results on SiO$\mu_0$09 also argue against pressure-driven metallization across much of the super-Earth range. In this combined reading, the super-Earth form of the hypothesis is best understood as a structural possibility rather than an established transport outcome: planets above about $\mu_0$10 may possess the correct geometry for a silicate dynamo, but the conductivity bottleneck remains severe (Bolis et al., 2016, Mazevet et al., 2014).

This yields a mineral-physics-centered classification of terrestrial-like super-Earths. Planets of about $\mu_0$11–$\mu_0$12 are expected to be Earth-like dynamo hosts; those between about $\mu_0$13 and $\mu_0$14 may retain some liquid iron but have increasingly weak or short-lived core convection; those above about $\mu_0$15 may have thick, vigorously convecting basal magma oceans but no iron-core dynamo. In that final regime, any long-lived intrinsic magnetic field would likely require silicate induction.

6. Controversies, limitations, and current research directions

The principal controversy is not whether silicate melt can ever be electrically conducting, but whether it can be conducting enough, for long enough, and under the correct buoyancy regime to sustain a planetary-scale field. Older super-Earth-oriented studies emphasized low conductivity, delayed metallization, and electromagnetic screening by underlying iron. More recent Earth basal-magma-ocean studies emphasize Fe enrichment, spin-state effects, and conductivities exceeding $\mu_0$16–$\mu_0$17 S/m under $\mu_0$18–$\mu_0$19 GPa and $\mu_0$20–$\mu_0$21 K conditions. These are not simply contradictory results; they probe different compositional regimes, different pressure–temperature windows, and different planetary architectures (Mazevet et al., 2014, Dragulet et al., 1 Aug 2025).

The hypothesis also depends strongly on model closure. Earth evolution calculations are sensitive to the initial core–mantle-boundary temperature, the parameterization of mantle convection, the Fe partition coefficient in the basal magma ocean, the radiogenic content of the planet, and the adopted convective-velocity and magnetic-scaling laws. In one Earth model, a transient silicate dynamo followed by a core dynamo after $\mu_0$22 billion years is the nominal outcome; in another, Fe enrichment can sustain $\mu_0$23 for about $\mu_0$24 Gyr under mixing-length scaling but for a shorter interval under Coriolis-constrained scaling (Lherm et al., 2024, Dragulet et al., 1 Aug 2025).

A broader open question concerns the relation between global magma-ocean dynamos and basal magma-ocean dynamos. Thermodynamic work on the post-giant-impact Earth describes a fully molten mantle, a silicate vapor atmosphere, and vigorous convection driven by radiative loss from a $\mu_0$25 K photosphere over at least the $\mu_0$26 year disk-cooling interval. This suggests that the “silicate dynamo” label may cover more than one temporal regime: an early whole-mantle magma-ocean regime and a later basal-magma-ocean regime (Pahlevan et al., 2010).

Current research directions follow directly from these unresolved dependencies. The literature repeatedly calls for improved high-pressure measurements and calculations of electrical and thermal conductivity in Fe-rich multicomponent silicate liquids, better constraints on Fe–Mg–Si partitioning and liquidus relations, and fully three-dimensional numerical dynamos for thin, rotating, conducting silicate shells coupled to metallic cores. Until those issues are resolved, the Silicate Dynamo Hypothesis remains a physically grounded but materially selective explanation: robust in some Earth basal-magma-ocean models, structurally possible but transport-limited in massive super-Earths, and fundamentally controlled by the transport coefficients of deep silicate liquids.