Crystallization Dynamo Theory in White Dwarfs

• Crystallization dynamo theory is a framework where magnetic fields in carbon–oxygen white dwarfs are generated by compositional convection triggered during phase separation at core crystallization.
• It distinguishes between a brief, efficient overturning convection phase that robustly amplifies fields and a long-term, slow thermohaline regime with minimal magnetic impact.
• The theory connects observed magnetic field distributions in cool, metal-polluted and binary white dwarfs to dynamo action while addressing challenges in dynamo scaling, rotation effects, and magnetic diffusion.

Crystallization dynamo theory is a white-dwarf magnetogenesis framework in which magnetic fields are generated by convection associated with core crystallization in carbon–oxygen white dwarfs. Its central premise is that crystallization is not merely a cooling milestone: phase separation enriches the solid in oxygen, leaves the adjacent liquid relatively carbon-rich, and thereby creates a buoyant, compositionally unstable liquid mantle above the growing solid core. In the original phenomenological form of the theory, rotation is the second essential ingredient, because it organizes the convective flow into a large-scale dynamo; in more recent formulations for isolated white dwarfs, the decisive magnetic amplification occurs during a short efficient-convection episode at the onset of crystallization, while later convection is too weak to sustain megagauss fields. The theory has been invoked to explain the temperature distribution of magnetic polluted white dwarfs, the stronger magnetism of accreting binaries, and the delayed appearance of surface magnetism relative to crystallization onset (Schreiber et al., 2021).

1. Physical basis of the dynamo

In a cooling carbon–oxygen white dwarf, crystallization begins at the center and proceeds outward. Because the carbon–oxygen phase diagram is spindle-shaped, the solid phase is oxygen-enriched relative to the liquid. The newly formed solid is therefore denser, while the surrounding liquid becomes relatively carbon-rich and hence lighter. That lighter liquid is redistributed by Rayleigh–Taylor instability and compositionally driven convection, producing a convective or compositionally unstable shell outside the growing solid core (Schreiber et al., 2021).

This structure is the white-dwarf analogue of the configuration that powers dynamos in planets and fully convective low-mass stars: a solid inner region plus a conducting convective fluid shell. The 2022 close-double-white-dwarf study summarized the geometry as “an outer carbon rich convective zone and a solid inner oxygen rich nucleus,” and emphasized that the mechanism requires a carbon–oxygen white dwarf in which crystallization has already begun (Schreiber et al., 2022).

The geodynamo analogy is structurally important but not exact. In white dwarfs the buoyancy source is chemical separation during C/O freezing, the convective mantle is buried deep inside the star, and the observable surface field depends on later transport through stable outer layers. A plausible implication is that crystallization dynamo theory is simultaneously a theory of interior field generation and a theory of magnetic visibility.

2. Rotation, spin-up, and magnetic scaling

In the original rotation-plus-crystallization picture, convection alone is insufficient: rotation sets the dynamo regime. The relevant rotational measure is the Rossby number, defined as the rotation period divided by the convective turnover time. Saturation is expected when the Rossby number is below about $0.1$, which for white dwarfs corresponds to spin periods of order seconds to minutes. In that framework, accreting white dwarfs in close binaries can be driven into a saturated regime and produce very strong fields, up to several $100$ MG, whereas polluted isolated white dwarfs should generally spin up less and therefore produce weaker fields (Schreiber et al., 2021).

The proposed angular-momentum source differs by population. For metal-polluted isolated white dwarfs, the theory uses disc-mediated accretion of planetary debris. Solving the adopted angular-momentum balance equation for a $0.6\,M_\odot$ white dwarf with initial spin periods of 1–3 days and total accreted masses in the range $10^{-6}$–$10^{-3}\,M_\odot$ over several Gyr yields final spin periods from several minutes to a few hours once more than about $10^{-5}\,M_\odot$ is accreted. That is short enough to activate a dynamo, but generally not short enough to reach the seconds-to-minutes saturation threshold associated with the strongest fields in close binaries (Schreiber et al., 2021).

A later reanalysis changed the dynamical emphasis. It argued that crystallization-driven convection is much slower than previously estimated, with $t_{\rm conv}\sim 10^6$–$10^7\ {\rm s}$, so most white dwarfs satisfy $P\ll t_{\rm conv}$ and therefore lie in the fast-rotation regime. In that treatment the field is at least in equipartition with convective motion and may be further enhanced by $B\propto (t_{\rm conv}/P)^{1/2}$, depending on the assumed scaling law (Ginzburg et al., 2022).

These two formulations are not identical. The earlier one tied strong-field production directly to very short spin periods and low Rossby number, whereas the 2022 reanalysis argued that low Rossby number is almost generic once the convective timescale is revised upward. This suggests that the unresolved part of the theory lies less in the existence of a buoyancy source than in the correct dynamo saturation law for white-dwarf interiors.

3. Population evidence and benchmark systems

The strongest observational support for crystallization dynamo theory originally came from cool, metal-polluted white dwarfs. Magnetic polluted white dwarfs are overwhelmingly cool: all magnetic DZ white dwarfs and all but one magnetic DAZ white dwarfs have effective temperatures below about $100$0 K, and three isolated white dwarfs with Zeeman-split Balmer emission lines also cluster around $100$1 K. For typical white-dwarf masses, crystallization begins roughly in the $100$2–$100$3 K range, and for a $100$4 white dwarf starts near $100$5 K. The observed magnetic DAZ and DZ stars almost all fall below these onset temperatures, and at low temperatures about $100$6 of known DZ white dwarfs are magnetic whereas no magnetic DZ has been found in the $100$7–$100$8 K interval (Schreiber et al., 2021).

The polluted samples are weakly to moderately magnetic rather than extreme objects. The compiled DZ sample spans effective temperatures from about $100$9 K to $0.6\,M_\odot$0 K and surface field strengths from about $0.6\,M_\odot$1 MG up to $0.6\,M_\odot$2 MG, while the DAZ sample includes objects from about $0.6\,M_\odot$3 K to $0.6\,M_\odot$4 K with fields from $0.6\,M_\odot$5 to $0.6\,M_\odot$6 MG, plus WD 2105–820 at $0.6\,M_\odot$7 K and $0.6\,M_\odot$8 MG. This matches the expectation that planetary accretion generally supplies enough angular momentum for a dynamo, but not the extreme spin-up needed for the strongest binary-like fields (Schreiber et al., 2021).

Close double white dwarfs provide a different test because they isolate the joint requirement of crystallization and prior spin-up. Among 57 close double white dwarfs with periods below 35 days, only one strongly magnetic white dwarf is known: the magnetic component of NLTT 12758. That star has a field of about 3.1 MG, a mass of $0.6\,M_\odot$9, a spin period of 23 min, and an effective temperature of $10^{-6}$0 K; it is paired with a non-magnetic $10^{-6}$1 DA companion in a 1.154 d orbit. The magnetic component lies near or within the crystallization regime, while the vast majority of known close double white dwarfs are too hot to be crystallizing (Schreiber et al., 2022).

The evolutionary interpretation of NLTT 12758 is specifically dynamo-motivated. A revised formation scenario invokes stable, highly non-conservative mass transfer, then a symbiotic phase, then one common-envelope event. In the best-fitting MESA model, modest wind accretion during the symbiotic phase is enough to spin the first-formed white dwarf to about 23 min before it later crystallizes. The magnetic field is then interpreted not as a direct product of mass transfer or common-envelope evolution, but as a later consequence of crystallization acting on a previously spun-up white dwarf (Schreiber et al., 2022).

4. Convective regimes and the onset-of-crystallization revision

A major revision of crystallization dynamo theory concerns the actual convective regime above the crystallizing core. A generalized mixing-length theory that self-consistently includes thermal diffusion and composition gradients showed that crystallization-driven convection is not generically fast overturning convection through the entire crystallization epoch. Instead, the same local theory has two limits: a slow thermohaline regime at low Peclet number and a fast overturning regime at large Peclet number. In flux form, the controlling quantity is the dimensionless composition-flux parameter $10^{-6}$2; $10^{-6}$3 yields fast overturning, whereas $10^{-6}$4 yields slow thermohaline convection. In the white-dwarf models analyzed there, fast overturning occurs only for a short time, $10^{-6}$5 Myr, near crystallization onset, while most of the later evolution is thermohaline (Castro-Tapia et al., 2024).

That distinction strongly affects dynamo viability. In the MLT-based convection study, the thermohaline regime gives $10^{-6}$6–$10^{-6}$7 for most of the cooling history, whereas the early overturning phase gives $10^{-6}$8–$10^{-6}$9. The paper explicitly connected this to the heuristic saturated-dynamo estimate $10^{-3}\,M_\odot$0, concluding that the slow regime is far too weak for megagauss white dwarfs and that only the early fast regime is plausibly dynamo-relevant (Castro-Tapia et al., 2024).

A companion 2024 study pushed this further by combining white-dwarf cooling models with MAC balance and Davidson’s low-$10^{-3}\,M_\odot$1 scaling. It argued that crystallization-driven convection spans two physically distinct epochs: a brief high-$10^{-3}\,M_\odot$2, efficient overturning phase at onset and a long low-$10^{-3}\,M_\odot$3 thermohaline phase afterward. For the $10^{-3}\,M_\odot$4 model, $10^{-3}\,M_\odot$5–$10^{-3}\,M_\odot$6 at crystallization onset and remains above unity for $10^{-3}\,M_\odot$7 Myr, then drops to $10^{-3}\,M_\odot$8 for $10^{-3}\,M_\odot$9 Gyr. In that framework, the thermohaline regime yields velocities $10^{-5}\,M_\odot$0–$10^{-5}\,M_\odot$1 and fields $10^{-5}\,M_\odot$2, while the efficient regime yields $10^{-5}\,M_\odot$3–$10^{-5}\,M_\odot$4 and $10^{-5}\,M_\odot$5–$10^{-5}\,M_\odot$6 (Fuentes et al., 2024).

The same paper made a sharper claim than the earlier rotation-controlled literature: in the efficient regime the field estimate becomes independent of the stellar rotation rate. That result follows from the adopted MAC scaling, in which the explicit $10^{-5}\,M_\odot$7 dependence cancels after substituting the efficient-convection expression for $10^{-5}\,M_\odot$8. The authors presented this as consistent with the lack of a strong observed correlation between white-dwarf field strength and rotation (Fuentes et al., 2024).

5. Breakout, diffusion, and post-dynamo field evolution

Crystallization dynamo theory now includes a distinct transport problem: interior field generation does not imply immediate surface magnetism. In the breakout formulation, the dynamo operates in a compositionally convective mantle above the crystallization front, deep within the carbon–oxygen core. The field remains buried until magnetic diffusion from the outer edge of that mantle to the surface becomes fast enough that the field can emerge. The onset of crystallization scales as $10^{-5}\,M_\odot$9, but the breakout time is longer by a few Gyr and scales as $t_{\rm conv}\sim 10^6$0, where $t_{\rm conv}\sim 10^6$1 depends on the pre-crystallization C/O profile (Blatman et al., 2023).

The physical reason for the delay is geometric. Early in crystallization, the outer edge of the mixed convective region is trapped deep inside the core by a stabilizing C/O-induced $t_{\rm conv}\sim 10^6$2-step, so the diffusion time is of order $t_{\rm conv}\sim 10^6$3 Gyr. Only after enough of the star has crystallized that the convective mantle approaches the helium layer does the diffusion time fall to $t_{\rm conv}\sim 10^6$4 Gyr. Numerically, roughly half the stellar mass must crystallize before breakout, and at lower masses an even larger fraction is required. In volume-limited samples, the first appearance of strong magnetic fields $t_{\rm conv}\sim 10^6$5 approximately coincides with the predicted $t_{\rm conv}\sim 10^6$6, although some magnetic white dwarfs are slightly younger than the model allows (Blatman et al., 2023).

Post-dynamo diffusion imposes an additional constraint. A 2024 induction-equation study followed the magnetic field after the efficient dynamo phase ends and showed that the eventual surface field differs from the initial buried field $t_{\rm conv}\sim 10^6$7 by at least a factor of $t_{\rm conv}\sim 10^6$8. The reduction depends strongly on the initial outer boundary of the convective region, $t_{\rm conv}\sim 10^6$9: when $10^7\ {\rm s}$0, the resulting surface field satisfies $10^7\ {\rm s}$1. As the solid core grows, the conductivity increases, the magnetic field becomes progressively frozen in, and outward transport slows substantially (Castro-Tapia et al., 2024).

That transport calculation narrows the phenomenological scope of the theory. Across $10^7\ {\rm s}$2–$10^7\ {\rm s}$3, crystallization-driven dynamos can explain only magnetic C/O white dwarfs with surface fields less than a few MG. The same diffusion bottleneck applies to buried fossil fields: if crystallization-driven convection is responsible for transporting a fossil field from deep in the core, that buried field must be at least 100 times stronger than the eventual observed surface field (Castro-Tapia et al., 2024).

A central tension in the recent literature follows directly. The onset-convection calculations yield interior fields of $10^7\ {\rm s}$4–$10^7\ {\rm s}$5, whereas the breakout-and-diffusion calculations imply that only a reduced remnant generally reaches the surface. This suggests that crystallization dynamo theory is now best understood as a coupled theory of generation, burial, and delayed emergence rather than as a one-step prediction for the observed field amplitude.

6. Scope, alternatives, and unresolved problems

The theory is broad but not universal. It is explicitly formulated for carbon–oxygen white dwarfs; helium-core white dwarfs and many hotter or more massive objects fall outside its nominal domain. It has been used to explain the high incidence of weaker magnetic fields in old metal-polluted white dwarfs, the stronger fields of accreting binaries whose white dwarfs were spun up more efficiently, and the particular status of NLTT 12758 among close double degenerates. It has also been proposed as a channel for strong fields in some isolated white dwarfs, provided that the field is generated during the short efficient-convection phase at crystallization onset rather than during the long thermohaline phase (Schreiber et al., 2021).

At the same time, the papers are explicit that crystallization is probably not the only route to magnetic white dwarfs. Hot magnetic white dwarfs likely require some other origin, with double white-dwarf mergers suggested for at least part of that population. Slowly rotating single magnetic white dwarfs with fields above 100 MG, and systems such as V471 Tau, CC Ceti, and HY Eri, are identified as cases likely requiring fossil fields, mergers, or some other mechanism rather than a crystallizing C/O-core dynamo (Schreiber et al., 2022).

Several limitations recur across the literature. There is no white-dwarf-specific three-dimensional dynamo simulation. The early convective shell geometry and its outer boundary are uncertain. Quantitative field-strength laws remain unsettled because different studies adopt different balances—equipartition, Lorentz–Coriolis, or MAC—and the white-dwarf magnetic Prandtl number is very different from that of planets and low-mass stars. Observational samples of magnetic DAZ and DZ stars are still small and incomplete, while individual masses for magnetic polluted white dwarfs are difficult to measure because Zeeman splitting, Stark broadening, atmospheric metal blanketing, and magnetic suppression of convection complicate standard analyses (Schreiber et al., 2021).

The breakout calculations introduce additional sensitivities. The predicted timing depends on the pre-crystallization C/O profile inherited from progenitor evolution, on the CO phase diagram, on the conductivity prescription, and on uncertain stellar-evolution inputs such as $10^7\ {\rm s}$6. The breakout study therefore argued that magnetic emergence may probe both the CO phase diagram and uncertainties during core helium burning, while also noting that sample incompleteness, atmosphere mismatch, and younger magnetic outliers prevent a definitive claim (Blatman et al., 2023).

In its current form, crystallization dynamo theory is therefore a family of related models rather than a single closed formalism. Its most robust element is the thermodynamic source of compositional buoyancy during C/O crystallization. Its most contentious elements are the saturated dynamo scaling, the duration of efficient convection, and the mapping from a buried interior field to an observable surface field. The literature consistently describes the mechanism as physically coherent and empirically suggestive, but not yet definitive.