Ambipolar Diffusion: Non-Ideal MHD in Disk Outer Layers

What ambipolar diffusion is: the breakdown of flux-freezing

In ideal MHD, magnetic field lines are 'frozen' into the gas: the plasma and field move together as one fluid. This is an excellent approximation when the gas is well ionized and ion-neutral collisions are frequent. Protoplanetary disks violate that assumption badly. They are cold (10–100 K) and dense enough to shield out the ionizing X-rays and cosmic rays that would otherwise keep them coupled to the field.

Ambipolar diffusion is what happens when the tiny charged fraction — ions and electrons tied to the field by the Lorentz force — must drag the overwhelmingly neutral gas along by collisions alone. If those collisions are too rare, the neutrals slip relative to the field. The field diffuses through the neutral gas rather than being carried by it.

• Ions/electrons follow the magnetic field.
• Neutrals feel the field only indirectly, via ion-neutral drag.
• The drift velocity between the two populations is the ambipolar diffusion.

It is called 'ambipolar' because both charge species (positive ions and electrons) drift together relative to the neutrals. It is a resistive, dissipative process — the drift generates frictional heating and erases small-scale field structure.

The mechanism and the governing relation

The coupling strength is set by how often a neutral collides with an ion within one orbital time. This defines the ambipolar Elsasser number:

Am ≡ γ ρ_i / Ω

where γ is the ion-neutral drag coefficient (≈ 3.5×10^13 cm³ g⁻¹ s⁻¹ for H₂-HCO⁺), ρ_i is the ion mass density, and Ω is the local Keplerian orbital frequency. Am is literally the number of times a neutral molecule collides with ions during a dynamical time Ω⁻¹.

• Am ≫ 1: neutrals and field are tightly coupled → effectively ideal MHD.
• Am ≲ 1: poor coupling → field diffuses, and the MRI is quenched.

The associated ambipolar magnetic diffusivity scales as η_A ≈ v_A² / (Ω · Am), where v_A is the Alfvén speed. Because ρ_i drops steeply with gas density, and v_A rises where the field is strong relative to thin gas, ambipolar diffusion grows most severe in the tenuous outer disk and in the surface layers — precisely the regions ideal MHD would predict to be most active.

Characteristic numbers and a worked example

Consider a solar-mass star with a minimum-mass-solar-nebula disk. At r ≈ 30 AU, the midplane gas density is roughly n ≈ 10^10 cm⁻³ and temperature ~30 K. Cosmic rays (ionization rate ζ ≈ 10⁻¹⁷ s⁻¹) sustain an ionization fraction of only x_e = n_e/n ≈ 10⁻¹⁰ to 10⁻⁸.

The Keplerian frequency there is Ω = √(GM/r³) ≈ 1.2×10⁻⁹ s⁻¹ (an orbital period of about 160 years). Plugging realistic ion densities into Am = γ ρ_i / Ω yields Am of order 0.1–1 at the midplane, rising to Am ≫ 1 in the ionized surface layers.

• Midplane, outer disk: Am ≈ 0.1–1 → ambipolar-quenched, weakly turbulent.
• Disk surface (a few scale heights up): Am ≫ 1 → MRI-active.
• Threshold: sustained MRI turbulence requires roughly Am ≳ 1 (Bai & Stone 2011).

The upshot: the resulting turbulent viscosity parameter α is small (α ~ 10⁻⁴ to 10⁻³) in ambipolar-dominated regions, far below the α ~ 0.01–0.1 that ideal MRI would produce. Magnetized disk winds, not turbulence, are then thought to carry most of the angular momentum.

How it is inferred and where it appears

Ambipolar diffusion cannot be photographed directly — it is diagnosed by its consequences. Several lines of evidence point to it dominating the outer disk:

• Low turbulence from ALMA: Molecular line broadening (e.g. CO isotopologues in TW Hya and HD 163296) implies non-thermal velocities below ~0.05–0.1 c_s, consistent with weak turbulence expected when Am ≲ 1 and the MRI is suppressed. The exoALMA program has pushed these limits further.
• Measured accretion rates: Ṁ ≈ 10⁻⁸ M_⊙/yr onto T Tauri stars cannot be sustained by ambipolar-quenched turbulence alone, motivating magnetically driven disk winds.
• Dust settling and rings: Thin, settled dust layers and the ubiquitous ALMA ring/gap structures require low midplane turbulence.

In the star-formation context, ambipolar diffusion is inferred from the mass-to-flux ratios of collapsing cores (via Zeeman splitting measurements of magnetic field strength) and from the resolution of the classical 'magnetic braking catastrophe' that would otherwise prevent rotationally supported disks from forming at all.

How it compares to Ohmic and Hall diffusion

All three non-ideal effects arise from finite conductivity, but they scale differently with density and field, so each rules a different zone:

• Ohmic dissipation comes from electron-neutral collisions and dominates the dense midplane (r ≈ 1–10 AU), the classic 'dead zone' of Gammie (1996). Its diffusivity η_O is field-independent.
• The Hall effect arises from the drift between electrons and ions and is peculiar: it does not dissipate energy, and its outcome depends on the sign of B·Ω, so aligned and anti-aligned fields behave differently.
• Ambipolar diffusion dominates the low-density outer disk and surface layers, and its diffusivity η_A ∝ B²/ρ_i grows with field strength and falling ion density.

Roughly: Ohmic wins at high density, ambipolar at low density, and Hall in between. Realistic global simulations (e.g. Bai 2017; Lesur et al.) include all three, which together can render both midplane and surface layers largely laminar, leaving winds as the accretion engine. Ambipolar diffusion, alone among the three, is also the classic mechanism in interstellar cloud collapse.

Significance, landmark work, and open questions

Ambipolar diffusion was first proposed by Leon Mestel and Lyman Spitzer (1956) to explain how magnetic flux leaks out of subcritical molecular clouds, allowing gravitational collapse on the ambipolar timescale τ_AD ≈ τ_ff² / τ_ni — often 10× the free-fall time. Its role in disks was developed decades later: Hawley, Gammie & Balbus formalized the MRI (1991–1995), Gammie (1996) introduced the Ohmic dead zone, and Bai & Stone (2011, 2013) and Simon et al. established through shearing-box and global simulations that ambipolar diffusion, not Ohmic dissipation, controls the outer disk.

• Wind vs. turbulence debate: If turbulence is quenched, do magnetocentrifugal winds carry angular momentum? How much?
• The Hall interplay: The sign-dependence of the Hall effect complicates every prediction.
• Planetesimal formation: Low turbulence aids dust settling and the streaming instability — but how low is 'low enough'?

These questions sit at the heart of modern planet-formation theory, and ALMA-era turbulence measurements are the observational battleground.

The three non-ideal MHD effects in protoplanetary disks and where each dominates

Effect	Physical origin	Where it dominates	Ionization / density regime
Ohmic dissipation	Electron-neutral collisions; finite resistivity	Dense midplane, r ≈ 1–10 AU	High density (n ≳ 10^14 cm^-3), very low ionization
Hall effect	Electron-ion drift; field-orientation dependent	Intermediate layers, r ≈ 1–30 AU	Moderate density; sign of B·Ω matters
Ambipolar diffusion	Ion-neutral drift; poor collisional coupling	Outer disk (r ≳ 30 AU) and surface layers	Low density (n ≲ 10^12 cm^-3), weak coupling
Ideal MHD (flux-frozen)	N/A — field locked to gas	Hot inner rim, disk corona above ~few H	Thermal/UV/X-ray ionized, Am ≫ 1