Geometrically Thick Disks

• Geometrically thick disks are disk-like astronomical structures with a substantial vertical extent relative to their radial size, serving as key indicators of formation history and dynamics.
• In galaxies, these disks diagnose stellar assembly, age gradients, and chemical diversity, as exemplified by studies of the Milky Way and M31.
• In accretion theory, thick disks are modeled as pressure-supported tori whose structure influences black-hole spin evolution, energy feedback, and observable spectral features.

Geometrically thick disks are disk-like astronomical structures whose vertical extent is not negligible relative to their radial extent. In galactic stellar systems, the term denotes populations with large scale heights or stars located high above the mid-plane; in accretion theory, it denotes flows for which the vertical scale height is not small compared with cylindrical radius, commonly written as $H/R \sim 1$, $H/R \sim 0.1-1$, or, in several GRMHD settings, $H/R \approx 0.3$ (Dalcanton et al., 2023, Sadowski et al., 2015, Lowell et al., 2023). Taken together, the literature treats thickness not as a purely morphological label but as a diagnostic of formation history, turbulence, pressure support, radiation trapping, magnetic structure, and observational appearance.

1. Definitions and geometric formalism

The phrase “geometrically thick disk” is used in two major astrophysical settings. For stellar disks in galaxies, thickness is quantified by a vertical density law, a scale height, or a geometric cut in distance from the mid-plane. For accretion disks, thickness is quantified by the semi-height $H$, by the ratio $H/R$, or by the shape of relativistic equipotential surfaces that define a pressure-supported torus (Lian et al., 15 Apr 2025, Martig et al., 2016, Riaz et al., 2019).

Setting	Thickness measure	Representative formulation
Stellar disks	$h_z$, $h_Z$, or large $|z|$	$\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$; $\rho(Z)=\rho_0\,\mathrm{sech}^2(Z/2h_Z)$
Accretion disks	$H/R \sim 0.1-1$0, $H/R \sim 0.1-1$1, equipotential shape	geometrically thick when $H/R \sim 0.1-1$2 or $H/R \sim 0.1-1$3; often $H/R \sim 0.1-1$4 for constant specific angular momentum

In stellar-disk work, several parameterizations coexist. M31 was modeled with an axisymmetric exponential disk,

$H/R \sim 0.1-1$5

while the Milky Way young thick-disk analysis measured mono-abundance populations with a $H/R \sim 0.1-1$6 vertical law,

$H/R \sim 0.1-1$7

and the AURIGA simulations used a double-$H/R \sim 0.1-1$8 decomposition to separate thin and thick components in edge-on projections (Dalcanton et al., 2023, Lian et al., 15 Apr 2025, Pinna et al., 2023).

In the Milky Way literature, “thick disk” may be defined chemically, kinematically, by age, or geometrically. One of the central conceptual points in recent work is that geometric thickness and chemical thick-disk classification are not always equivalent: a chemically low-$H/R \sim 0.1-1$9 population can still be geometrically thick, and a geometrically defined thick disk can mix multiple physical components (Martig et al., 2016, Lian et al., 15 Apr 2025, Pinna et al., 2023).

2. Geometrically thick stellar disks in galaxies

A direct resolved-stellar measurement was obtained for M31 by exploiting dust geometry. In a moderately inclined disk with a thin dust midplane, the fraction of reddened stars would be $H/R \approx 0.3$0 everywhere, but in a thickened stellar disk projection effects cause systematic near-side and far-side asymmetries. Mapping the fraction of reddened red giant branch stars across M31 yielded values from $H/R \approx 0.3$1 to $H/R \approx 0.3$2, implying

$H/R \approx 0.3$3

a thickness far larger than the Milky Way thin disk and comparable to the Milky Way thick disk (Dalcanton et al., 2023).

The Milky Way itself now shows more than one geometrically thick component. One study identified a relatively young thick disk with median age $H/R \approx 0.3$4 and scale height

$H/R \approx 0.3$5

while the canonical old high-$H/R \approx 0.3$6 thick disk has

$H/R \approx 0.3$7

and median age $H/R \approx 0.3$8. Thin-disk populations in the same analysis have typical thickness

$H/R \approx 0.3$9

The young thick component is more radially extended than the old thick disk and becomes the predominant thick component beyond $H$0 (Lian et al., 15 Apr 2025).

Independent Milky Way work based on APOGEE giant-star ages showed that the geometrically defined thick disk has a strong radial age gradient: at roughly $H$1–$H$2 above the plane, median ages decline from about $H$3–$H$4 in the inner disk to about $H$5 in the outer disk, and the fraction of stars younger than $H$6 rises from $H$7 to $H$8 (Martig et al., 2016). This directly establishes that a geometrically thick disk need not be uniformly old.

Edge-on external galaxies show comparable complexity. In FCC 153 and FCC 177 in the Fornax cluster, geometrically defined thick disks are old, relatively metal poor, and $H$9-enhanced, with dominant ages of $H/R$0–$H/R$1 and a secondary younger component around $H/R$2; outer thick-disk regions are slightly less metal poor and less $H/R$3-enhanced, consistent with flaring and heating of thin-disk stars (Pinna et al., 2019). In the 24-galaxy AURIGA sample, thick disks are on average $H/R$4 older, $H/R$5 more metal poor, and $H/R$6 more $H/R$7-enhanced than thin disks, with weighted-mean thick-disk scale height $H/R$8 (Pinna et al., 2023).

3. Formation channels, age structure, and interpretive issues

The formation of stellar thick disks remains contested. A recurrent question is whether disk thickness is inherited from the birth environment or established later by secular dynamical heating. The recent Milky Way detection of a young geometric thick disk explicitly frames the issue in those terms and argues that the discontinuities in age–scale height and age–velocity dispersion strongly discount secular dynamical heating as the primary mechanism, instead favoring a turbulent, bursty birth environment and a new phase of upside-down disk formation (Lian et al., 15 Apr 2025).

At the same time, thick-disk assembly is not described by a single channel across all systems. The Fornax 3D study proposed three populations and three mechanisms—early in-situ formation, later accretion, and disk heating—for FCC 153 and FCC 177, and emphasized that environment can have a strong effect on the galaxy evolutionary path (Pinna et al., 2019). The AURIGA simulations likewise concluded that thick disks result from early initial in-situ formation followed by later growth driven by the combination of direct accretion of stars, some in-situ star formation fueled by mergers, and dynamical heating of stars, with mergers contributing an average accreted mass fraction of $H/R$9 in thick-disk dominated regions; in two galaxies, about half of the geometric thick-disk mass was directly accreted (Pinna et al., 2023).

The radial age gradient of the Milky Way’s geometrically thick disk provides an important reconciliation of apparently conflicting observations. In simulations discussed alongside the APOGEE data, old stars are centrally concentrated while younger mono-age populations flare in the outer disk, bringing younger stars to high $h_z$0. The resulting geometric thick disk is therefore old in the center and younger in the outskirts (Martig et al., 2016). A plausible implication is that a single geometric thick disk can arise from different mechanisms in different radial zones.

Several common simplifications are therefore misleading. One is the assumption that a geometric thick disk is synonymous with a chemically defined thick disk; another is that red broad-band colors imply uniform old age. The Milky Way work on the young thick disk shows that a chemically low-$h_z$1 population can be geometrically thick (Lian et al., 15 Apr 2025), while the radial-age-gradient study shows that an age change from $h_z$2 to $h_z$3 at metallicity $h_z$4 alters $h_z$5 or $h_z$6 by only about $h_z$7, making substantial age gradients difficult to see photometrically (Martig et al., 2016).

4. Relativistic thick accretion disks and equilibrium tori

In accretion theory, geometrically thick disks are often modeled as stationary, axisymmetric, pressure-supported tori. A standard framework is the Polish doughnut approximation, in which the fluid is a perfect fluid with purely azimuthal motion,

$h_z$8

and, under stationarity, axisymmetry, and barotropy, the relativistic Euler equation can be integrated in terms of an effective potential. For constant specific angular momentum, that potential reduces to

$h_z$9

and its level surfaces coincide with pressure and density surfaces; minima define disk centers and maxima define cusps. In this regime, the disk is geometrically thick with

$h_Z$0

and the inner edge can lie inside the ISCO, with the cusp between the marginally stable and marginally bound orbits when $h_Z$1 (Riaz et al., 2019).

More general spacetimes preserve the same equilibrium logic while altering topology. Around a Schwarzschild black hole in a swirling universe, the metric acquires an odd $h_Z$2 symmetry, the torus loses equatorial symmetry, and closed equilibrium tori exist only for sufficiently small swirl. In that model, a finite torus requires

$h_Z$3

and when the swirling parameter exceeds

$h_Z$4

the function becomes monotonic and no equilibrium torus exists. The same repulsive deformation of the effective potential can create an outer cusp and an excretion channel (Chen et al., 2024).

For Kerr black holes in a swirling background, spin–spin coupling stabilizes prograde circular orbits and destabilizes retrograde ones, produces an outer marginally stable orbit, and bends the equilibrium structures into bowl-like forms. The thick-disk solutions are classified by cusp structure: Type 1a, 1b, 1c, and Type 2. Because the background breaks equatorial reflection symmetry, the extrema of $h_Z$5 are generally off the equatorial plane and the disk cross section becomes concave for prograde orbits and convex for retrograde ones (Gjorgjieski et al., 2024).

Viscous extensions preserve the importance of the effective-potential backbone while altering the detailed shape of the torus. In a causal relativistic Navier–Stokes treatment around Schwarzschild, with shear viscosity, relaxation time, and curvature corrections, the corrected pressure surfaces are obtained from

$h_Z$6

and the paper’s central result is that viscosity and curvature can qualitatively alter the shape and stability of geometrically thick accretion disks, producing cusps and changing equipressure contours relative to the classic inviscid torus picture (Lahiri et al., 2019).

5. Imaging, spectroscopy, energetics, and spin evolution of thick accretion flows

Finite thickness changes observables even when the spacetime is fixed. In relativistic reflection spectroscopy, geometrically thick disks differ from razor-thin Novikov–Thorne disks because their inner edge can lie inside the ISCO and the near-side upper layers can obscure the far inner region. Ray-traced iron lines and full reflection spectra are therefore “definitively different” from thin-disk profiles. When thick-disk synthetic spectra are fit with thin-disk models, the fits can still be statistically good, but spin is systematically overestimated, the emissivity index $h_Z$7 is biased, and the non-Kerr deformation parameter $h_Z$8 can be driven away from zero (Riaz et al., 2019).

The global energetics of thick accretion flows are likewise distinctive. Three-dimensional GRRMHD simulations found that for non-rotating black holes, both optically thin ADAF-like flows and optically thick supercritical flows return

$h_Z$9

of the accreted rest-mass energy to their surroundings, roughly half the standard thin-disk efficiency. For the spinning case $|z|$0, the total feedback efficiency is about

$|z|$1

The papers emphasize that the jet is collimated and likely to interact only weakly with the environment, whereas the outflow and radiation components cover a wide solid angle (Sadowski et al., 2015).

In the magnetically arrested disk regime, thick disks also control black-hole spin evolution. Non-radiative 3D GRMHD simulations of radiatively inefficient and geometrically thick MADs with

$|z|$2

found that the equilibrium spin is very low,

$|z|$3

because of both disk angular-momentum starvation and jet-mediated extraction of rotational energy. A maximally spinning black hole spins down from $|z|$4 to $|z|$5 after accreting only about $|z|$6 of its own mass (Lowell et al., 2023).

Misalignment introduces another geometric layer. Idealized 3D GRMHD simulations of tilted thick disks around Kerr black holes found

$|z|$7

warped and twisted steady states, and standing shocks near

$|z|$8

These shocks partially align the inner disk with the black-hole spin, with stronger alignment at higher spin, but the alignment timescale remains larger than the inflow timescale by a factor of roughly $|z|$9, so full alignment is not achieved before accretion (Gupta et al., 2024).

Millimeter imaging studies show a related consequence: thick geometry can fill in the image interior and obscure the classical inner shadow, while anisotropic synchrotron radiation can enhance photon-ring detectability by suppressing direct primary emission more than lensed emission (Zhang et al., 2024). This suggests that thickness and emissivity anisotropy are coupled observational degrees of freedom rather than separable corrections.

6. Nonstandard spacetimes and other astrophysical realizations

Geometrically thick disks are frequently used as probes of non-Kerr spacetimes. In Kerr-MOG imaging, thick magnetized equilibrium tori around a black hole with MOG parameter $\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$0 become smaller as $\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$1 increases, yet the total flux density and peak brightness in their images rise. Comparison with M87* was used to constrain $\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$2 and led to the conclusion that the presence of the MOG parameter broadens the allowable range of black-hole spin (Zhang et al., 2024). In Kerr-Sen imaging with optically thin thick flows, increasing charge $\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$3 shrinks both the photon rings and the central dark regions, while frame dragging makes brightness asymmetry more pronounced with increasing $\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$4 and observer inclination $\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$5; the BAAF model, because of its conical approximation, produces narrower higher-order images and a clearer separation between direct and higher-order structure than the RIAF model (Wang et al., 9 Apr 2026).

Thick-disk theory also changes the interpretation of compact-object spin bounds. In a Polish-doughnut treatment of accretion onto non-Kerr or exotic compact objects, the inner edge need not have a single equatorial cusp; two cusps can appear above and below the equatorial plane. Combining thick-disk spin-equilibrium calculations with the empirical requirement $\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$6 yields an observational bound

$\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$7

only slightly weaker than the earlier thin-disk bound (Li et al., 2012).

The concept is not confined to black-hole accretion. White-dwarf debris disks can also be geometrically thick. In one recent model, the vertical structure is Gaussian,

$\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$8

with conservative values

$\rho_\star(R,z)=\rho_0 \exp(-R/h_r)\exp(-|z|/h_z)$9

and inclinations $\rho(Z)=\rho_0\,\mathrm{sech}^2(Z/2h_Z)$0 in an “edge-off” configuration. In that regime, transit depth and color become strong functions of inclination; the heated inner rim can contribute substantially to the infrared excess even at high inclination; and the same thick-disk framework can reproduce the reddening in the transit of WD J1013$\rho(Z)=\rho_0\,\mathrm{sech}^2(Z/2h_Z)$10427, the infrared excesses of WD 1145$\rho(Z)=\rho_0\,\mathrm{sech}^2(Z/2h_Z)$2017 and WD 1232$\rho(Z)=\rho_0\,\mathrm{sech}^2(Z/2h_Z)$3563, and the silicate feature in G29-38 (Bhattacharjee, 28 Jul 2025).

Across these very different applications, a common theme emerges: once the vertical structure is comparable to the relevant radial scale, geometry ceases to be a perturbation. In galaxies, that geometry records assembly, accretion, flaring, and heating; in accretion flows, it controls equilibrium topology, self-occultation, energy transport, jet production, and the interpretation of high-precision imaging and spectroscopy.