Surface-Accretion Disk Models

• Surface-accretion disk models are frameworks in which the disk’s vertical stratification creates a dynamically critical surface or upper layer that regulates mass transport and observable properties.
• They are applied across various systems—from black-hole disk–outflow interfaces and layered protoplanetary accretion to white-dwarf debris disks and circumplanetary shocks—each using distinct physical criteria for surface definition.
• Methodologies range from vertically averaged approximations and MHD simulations to radiative and hydrodynamic analyses, with key diagnostics including spin-dependent outflow power, redshifted molecular absorption, and emission spectra.

“Surface-accretion disk model” is not a single universally fixed formalism in the arXiv literature. The expression is used for several related constructions in which the disk surface, the disk–outflow interface, or vertically localized upper layers control mass transport, angular-momentum extraction, radiation, or compositional evolution. In some works the “surface” is a hydrodynamically defined boundary $z=h(r)$ with $P|_{z=h}=0$ and $v_r|_{z=h}=0$ (Mukhopadhyay, 2013); in others it is an MRI-active ionized skin above a dead midplane (Perez-Becker et al., 2010, Perez-Becker et al., 2011, Zhu et al., 2010), a wind-driven accretion layer at $z \approx z_s \sim 2\!-\!4H_g$ (Okuzumi, 2024, Ikeda et al., 26 May 2026), a coronal inflow channel in global MHD disks (Zhu et al., 2017), or the emitting photosphere of a geometrically thick inner flow (Gu et al., 2016). This suggests that the common element is vertical stratification: the surface is treated as a dynamically privileged zone rather than as a passive boundary.

1. Terminological scope and unifying idea

In the cited literature, the same phrase is attached to several non-identical but structurally related models.

Context	Surface role	Representative papers
Black-hole disk–outflow coupling	Upper boundary $z=h(r)$ where inflow terminates and outflow is launched	(Mukhopadhyay, 2013)
Layered protoplanetary accretion	Ionized active skin above a magnetically decoupled midplane	(Perez-Becker et al., 2010, Perez-Becker et al., 2011, Zhu et al., 2010)
Wind-driven young-star disks	Gas accretes in a surface layer while settled dust remains below it	(Okuzumi, 2024, Ikeda et al., 26 May 2026, Zhu et al., 2017)
Compact-object radiative surface models	Disk surface emits, obscures, or redirects radiation and magnetic flux	(Koutsantoniou et al., 2014, Gu et al., 2016, Karakonstantakis et al., 8 Dec 2025)

The mathematical implementation is therefore context dependent. Some models are vertically averaged and transport based, some are explicitly two-dimensional in $(r,z)$, and some are radiative or geometric surface constructions. The surface may be defined by ionization physics, hydrostatic balance, optical depth, magnetic dominance, or a photospheric boundary.

2. Disk surfaces as dynamical interfaces in compact-object accretion

In the hydrodynamic black-hole disk–outflow model of Ghosh and Mukhopadhyay, the surface is the upper boundary of a coupled advective disk–outflow region. The flow is sub-Keplerian, advective, geometrically thick, axisymmetric, and inviscid in the inner region, with a polytropic equation of state $P=K\rho^\gamma$ and $c_s=\sqrt{\gamma P/\rho}$. Vertical structure is closed through $\partial/\partial z \equiv 1/z$, and the vertical velocity is prescribed as

$v_z = l \left(\frac{z}{r}\right)^\mu c_s .$

The disk–outflow surface is determined by

$P|_{z=h}=0$0

together with the vertical momentum balance at $P|_{z=h}=0$1. In this framework the outflow power is computed from the surface energy flux,

$P|_{z=h}=0$2

and the solutions show that the coupled region becomes thinner and extends closer to the black hole as spin increases, while the outflow power increases strongly with spin (Mukhopadhyay, 2013).

A different compact-object usage treats the disk surface as the radiation source itself. In Kerr spacetime, disk-surface photons traced backward from the ISCO generate a strongly anisotropic radiation field once Doppler beaming, frame dragging, and light bending are included. For razor-thin disks the azimuthal radiation force is PR-like and negative at modest spin, but for finite-thickness disks and tori it becomes positive and increases steeply with spin, providing the azimuthal electromotive source for the Cosmic Battery (Koutsantoniou et al., 2014). In the thick-disk model for ultraluminous supersoft sources, the outer photosphere of a geometrically thick inner disk and the outer thin disk dominate the observed emission when the inner hard region is shaded for $P|_{z=h}=0$3. In that model,

$P|_{z=h}=0$4

with $P|_{z=h}=0$5, and the saturated luminosity is $P|_{z=h}=0$6, implying $P|_{z=h}=0$7 (Gu et al., 2016).

A geometrically thin but metric-sensitive surface construction appears in the Kerr-mimicker wormhole problem. There, the circular-orbit quantities at fixed circumferential radius are identical to Kerr if the metric differs only in $P|_{z=h}=0$8, but the proper annulus area is different, so the surface temperature obeys

$P|_{z=h}=0$9

with $v_r|_{z=h}=0$0 at the throat. The observable difference is therefore not orbital kinematics on circular equatorial orbits but the surface-emission geometry itself (Karakonstantakis et al., 8 Dec 2025).

3. Ionized active layers and layered accretion in protoplanetary disks

A major branch of the subject concerns layered protoplanetary disks in which only the surface is sufficiently ionized to couple to magnetic stresses. In the X-ray/cosmic-ray models of transitional and conventional disks, the central result is that MRI activity in the surface layer is limited primarily by ambipolar diffusion rather than Ohmic diffusion. The controlling parameter is

$v_r|_{z=h}=0$1

and X-ray irradiated layers typically give $v_r|_{z=h}=0$2, far below the $v_r|_{z=h}=0$3 threshold quoted from Hawley and Stone for robust MRI turbulence in predominantly neutral gas. Even allowing weak MRI at $v_r|_{z=h}=0$4, the active surface columns are only $v_r|_{z=h}=0$5 and the implied accretion rates are generally too small to explain most observed T Tauri disks (Perez-Becker et al., 2010).

The FUV-ionized surface-layer model changes that conclusion by ionizing abundant atomic carbon and sulfur rather than relying on the low-ionization X-ray chemistry. In that case the active layer has

$v_r|_{z=h}=0$6

with values up to $v_r|_{z=h}=0$7 in dust-depleted cases, while the electron and ion fractions reach $v_r|_{z=h}=0$8. The resulting $v_r|_{z=h}=0$9 is $z \approx z_s \sim 2\!-\!4H_g$0 above the carbon front, $z \approx z_s \sim 2\!-\!4H_g$1 peaks near $z \approx z_s \sim 2\!-\!4H_g$2, Hall diffusion is negligible, and MRI can operate in an ideal-like regime. The predicted accretion rates are comparable to observed T Tauri values at $z \approx z_s \sim 2\!-\!4H_g$3 AU, although the inner $z \approx z_s \sim 2\!-\!4H_g$4 AU still require additional transport unless mixing thickens the active layer (Perez-Becker et al., 2011).

The two-zone layered-accretion model of protostellar disks formalizes this vertical separation. The active layer has prescribed surface density $z \approx z_s \sim 2\!-\!4H_g$5, the total surface density is $z \approx z_s \sim 2\!-\!4H_g$6, MRI acts in the active layer and in the midplane only when $z \approx z_s \sim 2\!-\!4H_g$7 K, and GI enters through

$z \approx z_s \sim 2\!-\!4H_g$8

During infall, low-angular-momentum material initially lands in hot inner radii and accretes quasi-steadily; later infall reaches beyond $z \approx z_s \sim 2\!-\!4H_g$9 AU, mass piles up in the dead zone, GI heating raises $z=h(r)$0, and FU Ori-like gravo–magneto outbursts are triggered. After infall ends, the outer active disk evolves from $z=h(r)$1 to $z=h(r)$2, while a massive inner dead-zone belt persists (Zhu et al., 2010).

4. Wind-driven, coronal, and magnetically braked surface accretion

In wind-driven protoplanetary models, the surface is not primarily an ionized active column but a vertically localized gas-accretion channel. The 1D transport framework of Okuzumi and collaborators assumes that wind-driven inflow is concentrated near $z=h(r)$3 while settled dust resides below it. Gas evolves through

$z=h(r)$4

with

$z=h(r)$5

whereas the dust does not co-accrete with the wind layer when $z=h(r)$6. The resulting retention criterion is

$z=h(r)$7

which in the surface-accretion regime reduces to

$z=h(r)$8

Proof-of-concept simulations show that $z=h(r)$9 can be reached for Myr timescales, enabling strong particle clumping and potentially planetesimal formation even when grains are small and drift slowly (Okuzumi, 2024).

The same vertical geometry has been applied to volatile transport. When fragile icy grains remain settled below a wind-driven surface inflow, surface accretion selectively removes ice-free gas and raises the ice-to-gas ratio outside the snow line. In the fiducial model with $(r,z)$0, the inner-disk water vapor concentration rises only to $(r,z)$1 wt% in the vertically uniform model but reaches $(r,z)$2 wt% by $(r,z)$3 Myr in the surface-accretion model, producing an anti-correlation between inner-disk vapor concentration and residual disk gas mass (Ikeda et al., 26 May 2026).

Global ideal-MHD simulations with net vertical magnetic flux supply a complementary view. They show that most inward accretion occurs in magnetically dominated upper layers, while the midplane can be slow or even outward. For $(r,z)$4 and $(r,z)$5, the vertically integrated $(r,z)$6 is $(r,z)$7, coronal inflow is supersonic, the magnetic geometry is strongly pinched at the surface, and only about $(r,z)$8 of the angular momentum transport comes from the wind torque, with wind mass flux only $(r,z)$9 of the accretion rate (Zhu et al., 2017). Observational support for this kind of regime is provided by the mid-infrared absorption spectrum of GV Tau N, where C$P=K\rho^\gamma$0H$P=K\rho^\gamma$1, HCN, NH$P=K\rho^\gamma$2, and H$P=K\rho^\gamma$3O show redshifted absorption interpreted as warm inner-disk surface inflow, with inferred $P=K\rho^\gamma$4 and Mach numbers $P=K\rho^\gamma$5 (Najita et al., 2020).

A more idealized magnetic-surface variant treats braking directly through a static vertical field. There the surface torque is

$P=K\rho^\gamma$6

the Keplerian inflow speed is

$P=K\rho^\gamma$7

and the accretion rate becomes

$P=K\rho^\gamma$8

This formulation makes the “surface-accretion” character explicit: angular momentum is extracted vertically by surface Maxwell stresses rather than by an internal viscous $P=K\rho^\gamma$9 stress (Liffman, 18 Feb 2026).

5. Boundary layers, debris disks, and circumplanetary applications

The surface-accretion concept also appears in other source classes. In white-dwarf debris disks, solids drift inward by PR drag, sublimate at $c_s=\sqrt{\gamma P/\rho}$0, and feed a gaseous disk that spreads outward and overlaps the solids. If aerodynamic coupling is sufficiently strong or viscosity sufficiently weak, the coupled evolution enters runaway once the feedback parameter satisfies $c_s=\sqrt{\gamma P/\rho}$1. In that regime the accretion rate rises from the PR-limited $c_s=\sqrt{\gamma P/\rho}$2 to $c_s=\sqrt{\gamma P/\rho}$3, and the surface-accretion interpretation is the transport of material from the debris disk to the white-dwarf surface through the gas–solid overlap region (Metzger et al., 2012).

In global 3D resistive MHD models of low-mass protostars accreting in the boundary-layer regime, the radial structure divides into an MRI-active disk, a transition layer, and a boundary layer. Angular momentum transport changes character across these zones: MRI and coronal accretion dominate farther out, spiral shocks and vertical fluxes become important in the transition layer, and Maxwell stresses dominate in the boundary layer itself. The simulated system yields $c_s=\sqrt{\gamma P/\rho}$4 and a boundary-layer luminosity $c_s=\sqrt{\gamma P/\rho}$5, while also showing decretion flows in the midplane and frequent reconnection-driven explosive events in the stellar atmosphere and disk corona (Takasao et al., 19 Mar 2025).

Circumplanetary disks provide yet another variant. Two-dimensional global hydrodynamics coupled to one-dimensional radiation–hydrodynamic shock calculations show that vertical inflow from the circumstellar disk first shocks above the circumplanetary disk, then converges toward a narrow annulus on the planetary surface. The circumplanetary-disk shocks are weaker H$c_s=\sqrt{\gamma P/\rho}$6 emitters, by approximately $c_s=\sqrt{\gamma P/\rho}$7 orders of magnitude in luminosity, than the planetary-surface shocks. The dominant emission comes from localized planetary surface regions at latitudes $c_s=\sqrt{\gamma P/\rho}$8, producing $c_s=\sqrt{\gamma P/\rho}$9 and line widths of order $\partial/\partial z \equiv 1/z$0 for PDS 70-like parameters (Takasao et al., 2021).

6. Diagnostics, limitations, and broader significance

Across these models, the principal diagnostics are likewise surface-centered. They include outflow power extracted from a disk–outflow interface and its strong spin dependence (Mukhopadhyay, 2013); compact-object spectra shaped by disk-surface emission, obscuration, or photospheric area (Koutsantoniou et al., 2014, Gu et al., 2016, Karakonstantakis et al., 8 Dec 2025); redshifted molecular absorption from warm surface inflow (Najita et al., 2020); enhanced inner-disk water vapor and anti-correlation with residual gas mass in wind-driven protoplanetary disks (Ikeda et al., 26 May 2026); and high midplane dust-to-gas ratios created indirectly by gas removal from dust-poor surface layers (Okuzumi, 2024). In resolved young-star disks, global MHD calculations imply that “accretion rate” cannot be identified with a single midplane radial velocity, because inward coronal flow and midplane decretion can coexist (Zhu et al., 2017, Takasao et al., 19 Mar 2025).

The main limitations are equally recurrent. Many formulations are vertically averaged or 1D in radius, with the surface entering through closures such as $\partial/\partial z \equiv 1/z$1, prescribed $\partial/\partial z \equiv 1/z$2, or constant $\partial/\partial z \equiv 1/z$3 (Mukhopadhyay, 2013, Okuzumi, 2024, Zhu et al., 2010). Several neglect full non-ideal MHD, radiative transfer, or chemistry beyond a reduced network (Perez-Becker et al., 2010, Perez-Becker et al., 2011). Compact-object surface models often adopt pseudo-Newtonian gravity, idealized photospheres, or blackbody emission (Mukhopadhyay, 2013, Gu et al., 2016, Karakonstantakis et al., 8 Dec 2025). Even the global MHD calculations that explicitly recover coronal accretion remain ideal-MHD or resistive-MHD approximations and are sensitive to net field strength, thickness, and thermodynamics (Zhu et al., 2017, Takasao et al., 19 Mar 2025).

A common misconception is that “surface accretion” always means the same transport mechanism. The cited literature does not support that reading. Sometimes the surface is an ionized MRI layer, sometimes a wind-torqued coronal channel, sometimes a hydrodynamic outflow boundary, sometimes an emitting photosphere, and sometimes a geometry-induced proper-area effect. The more defensible synthesis is narrower: surface-accretion disk models are models in which the disk surface or surface-adjacent layers control the dominant observable or transport channel. Under that broader reading, the concept spans black-hole outflows, protoplanetary layered accretion, wind-driven volatile evolution, white-dwarf debris disks, circumplanetary shocks, and protostellar boundary layers.