Slim Disks and Supercritical Accretion

A slim disk is the transonic, optically thick continuation of disk accretion into the regime where heat cannot all escape locally. Radial advection, pressure support, photon trapping, and often mass loss then enter the leading-order energy budget. The central lesson is not that luminosity may exceed the Eddington value—it is that the supplied mass rate, the rate reaching the compact object, and the radiation seen along one line of sight cease to be interchangeable.

Required background. Fluid and Kinetic Theory: continuum kinematics and balance laws supplies conservation laws and advective transport.

Helpful background. Black-Hole Geometry supplies the relativistic inner boundary; Relativistic Matter and Compact Objects supplies radiative closure and opacity.

When local cooling fails

Write the vertically integrated energy equation as

$$Q^+=Q^-+Q_{\rm adv},$$

where all three terms are energy rates per unit disk area. Here $Q^+$ is mechanical heating, $Q^-$ is radiative cooling from the two disk surfaces in the chosen convention, and $Q_{\rm adv}$ is the energy-advection term involving the radial entropy gradient. Positive $Q_{\rm adv}$ means that some released heat is carried inward instead of radiated locally; its sign can reverse where advection heats the gas. A radiatively efficient thin disk has $|Q_{\rm adv}|\ll Q^+$. A slim disk permits $Q_{\rm adv}/Q^+$ to be order unity, includes radial pressure gradients, and crosses a sonic point. It therefore cannot be constructed by assigning an independent circular annulus at every radius.

Photon trapping gives the useful physical estimate. For vertical optical depth $\tau$, thickness $H$, radius $R$, and inflow speed $|v_R|$,

$$t_{\rm diff}\sim\frac{\tau H}{c}, \qquad t_{\rm acc}\sim\frac{R}{|v_R|}.$$

Radiation is trapped where $t_{\rm diff}\gtrsim t_{\rm acc}$. Combining mass conservation with electron-scattering opacity gives the scaling

$$\frac{R_{\rm tr}}{r_g}\sim \dot m\frac{H}{R}, \qquad r_g=\frac{GM}{c^2},$$

up to a coefficient that depends on vertical structure and on the definition of $\dot m$. Here $\dot m\equiv \dot M c^2/L_{\rm Edd}$; a convention normalized by $L_{\rm Edd}/(\eta_0c^2)$ is smaller by $\eta_0$. The definition must travel with every quoted trapping radius.

For example, take $\tau=\kappa\Sigma/2$ from midplane to one surface and $\dot M=2\pi R\Sigma|v_R|$. Then

$$\frac{t_{\rm diff}}{t_{\rm acc}} \sim\frac{\kappa\dot M H}{4\pi cR^2} =\dot m\,\frac{H}{R}\frac{r_g}{R}.$$

This derives the scaling and shows which mass rate to use: the rate through that annulus. In a wind it cannot be replaced by the outer supply rate at every radius.

Abramowicz and Fragile 2013, § 6 develops the transonic slim-disk equations and their departure from local cooling. The limiting checks are equally important: as $Q_{\rm adv}\rightarrow0$ and $H/R\rightarrow0$, the solution should approach the thin-disk flux; if optical depth falls below unity, an optically thick slim model is no longer the appropriate closure.

Winds and the global budget

At high supply rates the inward mass flux will generally depend on radius. A closed steady budget is

$$\dot M_{\rm supply} =\dot M_{\rm H}+\dot M_{\rm wind}+\dot M_{\rm jet},$$

where the positive outward wind and jet rates count disjoint rest-mass channels. A jet with negligible matter inertia can carry substantial energy even when its mass term is negligible. The energy counterpart may be organized as

$$\dot E_{\rm supply} =\dot E_{\rm H} +L_{\rm rad}+\dot E_{\rm wind}+\dot E_{\rm jet}.$$

All terms here are total energy fluxes measured with the same asymptotic time normalization. The supplied energy is approximately $\dot M_{\rm supply}c^2$ only for cold, weakly bound supply. The outward wind and jet terms include their rest mass as well as mechanical and electromagnetic energy. The signed total horizon flux $\dot E_{\rm H}$ includes binding, advected heat, and electromagnetic extraction; adding a separate $P_{\rm spin}$ would count that extraction twice. If one instead quotes wind kinetic power above rest mass, the rest-mass flux must first be subtracted consistently from both sides.

This budget prevents a common error: inferring the horizon rate by dividing the observed luminosity by a fixed thin-disk efficiency. Advection can carry energy through the horizon; a wind can remove mass, momentum, and radiation; a funnel can make the isotropic-equivalent luminosity depend strongly on inclination.

For an idealized mass-losing supercritical disk, integrating roughly Eddington-limited radiation over a logarithmic radial interval motivates

$$L\sim L_{\rm Edd}\left(1+\ln\dot m\right).$$

This is a scaling, not a universal slim-disk luminosity law. It changes with the adopted $\dot m$, radial mass loss, photon escape, beaming, and mechanical power. Poutanen et al. 2007, § 3 and Summary gives one explicit supercritical disk-and-wind construction; it should not be silently substituted for a wind-free transonic solution.

A broadband observational test

Use the coordinated 2012 epoch XN1 of NGC 1313 X-1 as a reproducible case study: XMM–Newton ObsIDs 0693850501 and 0693851201 and NuSTAR ObsIDs 30002035002 and 30002035004. The source distance is taken as $4.2\,\mathrm{Mpc}$. This data release spans approximately $0.3$–$35\,\mathrm{keV}$ before signal-to-noise cuts and is public through the two mission archives. Walton et al. 2020, § 2 and Table 1 records the observation identifiers, exposures, reduction, response products, energy selections, and distance.

A defensible reproduction freezes the XMM Science Analysis System and NuSTAR HEASoft/CALDB versions, extracts source and background from declared regions, generates response matrices and effective-area files, and retains the good-time and pile-up checks. The selection must record the good-time thresholds, the energy interval in which source counts dominate the background, and the signal-to-noise threshold used for grouping. Fit native counts rather than an unfolded plot. Include detector cross-normalizations, the Galactic and host absorption columns, and background uncertainty. The likelihood can be Poisson in ungrouped source-plus-background channels or an explicitly justified grouped statistic; the published analysis instead groups to signal-to-noise at least five and uses $\chi^2$.

The physical comparison should use at least three forward models on the same selected channels:

1. an absorbed thin multicolor disk plus a hard component;
2. a radially advective or $p$-free disk approximation;
3. a disk-plus-wind or two-temperature photospheric model.

Record luminosity and color temperature jointly, rather than declaring advection from one best-fit exponent. In the multi-epoch campaign, the hotter component occupies two luminosity–temperature tracks; each is compatible within the limited dynamic range with both $L\propto T^4$ and $L\propto T^2$. The high-energy flux is also much less variable than the lower-energy spectrum Walton et al. 2020, § 3.2 and Figures 3–5. These are constraints on a model, not a unique detection of a slim disk.

Statistical uncertainty comes from finite counts and parameter covariance. Model and systematic uncertainty includes cross-calibration, absorption, the distance, color correction, disk thickness, inner stress, inclination, wind scattering, geometric beaming, and the unknown nature of the accretor. Refit under all three model families, repeat with each instrument removed, and compare posterior-predictive residuals and the luminosity–temperature relation. A good count-space fit that changes physical interpretation under a plausible wind is evidence of non-identifiability, not confirmation.

The observational comparison includes results through 2026-08-18. Later calibration releases, reprocessed event files, or results require the pipeline and conclusion to be revisited.

What the model does and does not establish

Slim-disk physics establishes how advection alters a thickening, optically thick flow and why efficiency need not be constant. A luminous ULX spectrum may be consistent with that regime, but luminosity alone does not determine compact-object mass or even distinguish a black hole from a neutron star. Pulsation nondetection is not proof of a black-hole accretor; wind orientation and magnetic truncation supply competing explanations.

Detailed spectra and instrument likelihoods are beyond the physical model developed here; the observational case is included only to show how the model is confronted with data. Radiation, Spectra, and Variability develops radiative components and count-space systematics, while Electromagnetic Mass and Spin Inference develops conditional parameter inference.

Common pitfalls

Calling every source above $L_{\rm Edd}$ a slim disk. Beaming, an incorrect mass, opacity, a magnetized neutron star, or a wind photosphere can alter the apparent luminosity.

Holding $\dot M$ constant through a wind. The supplied rate and horizon rate must be reported separately, along with the mass and energy carried away.

Reading a color temperature as the effective temperature. Spectral hardening, scattering, inclination, and a moving photosphere change the mapping.

Locate the trapping region

A flow at $R=30r_g$ has $H/R=0.5$, $|v_R|=0.02c$, and midplane-to-surface optical depth $\tau=300$. Compute $t_{\rm diff}/t_{\rm acc}$. Use the same mass-conservation and optical-depth conventions as above to infer $\dot m$ and estimate $R_{\rm tr}$ if $\dot M$ and $H/R$ stay constant over the trapping region. Is local radiation of all dissipated heat self-consistent at $30r_g$?

Solution

The ratio is independent of the black-hole mass:

$$\frac{t_{\rm diff}}{t_{\rm acc}} =\frac{\tau H/c}{R/|v_R|} =\tau\frac{H}{R}\frac{|v_R|}{c} =300(0.5)(0.02)=3.$$

Vertical photon escape takes three times the radial inflow time, so advection cannot be neglected. Consistent mass conservation gives $\dot m=\tau(R/r_g)(|v_R|/c)=180$. The scaling estimate then gives

$$R_{\rm tr}\sim \dot m\frac{H}{R}r_g =180(0.5)r_g=90r_g.$$

The chosen annulus lies inside the estimated trapping boundary. A model imposing $Q^-=Q^+$ locally there is not self-consistent; it must include an advected energy flux. The estimate assumes constant $\dot M$ and $H/R$ between $30r_g$ and $90r_g$; a quantitative answer would solve the coupled vertical and transonic structure and check any wind loss.

References

• Abramowicz, Marek A., and P. Chris Fragile. “Foundations of Black Hole Accretion Disk Theory.” Living Reviews in Relativity 16 (2013): 1. DOI.
• Poutanen, Juri, Galina Lipunova, Sergey Fabrika, Alexander G. Butkevich, and Pavel Abolmasov. “Supercritically Accreting Stellar Mass Black Holes as Ultraluminous X-Ray Sources.” Monthly Notices of the Royal Astronomical Society 377 (2007): 1187–1194. DOI.
• Walton, Dominic J., et al. “The Unusual Broad-band X-Ray Spectral Variability of NGC 1313 X-1 Seen with XMM–Newton, Chandra, and NuSTAR.” Monthly Notices of the Royal Astronomical Society 494 (2020): 6012–6029. DOI.