Nuclear Stellar Discs: Structure & Evolution

• Nuclear Stellar Discs (NSDs) are compact, flattened, rotational stellar components at galaxy centres with sizes ranging from 10–500 pc and masses up to 10^9 M☉.
• They are identified through exponential light profile decompositions and kinematic analyses that reveal rapid rotation and distinct Fourier isophotal signatures.
• NSDs exhibit stellar age and metallicity gradients that trace bar-driven gas inflow and merger events, making them effective chronometers of galaxy assembly.

Nuclear stellar discs (NSDs) are compact, flattened, rotationally supported stellar systems located at galaxy centres and distinguished from both large-scale galactic discs and central spheroidal components. Across the literature, they span very compact structures with typical radii of $10^1\!-\!10^2$ pc in early-type galaxies and broader nuclear discs with characteristic radii of order $100\!-\!500$ pc and stellar masses of order $10^8\!-\!10^9\,M_\odot$. Their importance is twofold: they encode the dissipative and secular processes that funnel gas into galactic nuclei, and, because thin nuclear discs can be vulnerable to some merger configurations, their survival and stellar ages can constrain the assembly history of their hosts (Sarzi et al., 2015, Schultheis et al., 4 Sep 2025).

1. Definition, incidence, and distinction from other nuclear components

NSDs are flattened, dynamically cold stellar components embedded within the central few hundred parsecs of galaxies. In high-resolution imaging they appear as disc-like excesses above the surrounding bulge or bar light, and in spectroscopy they are recognized as rapidly rotating, relatively low-dispersion central components. They occur in both early- and late-type systems, and later syntheses emphasize that they are not restricted to a single host morphology and often coexist with other nuclear structures such as nuclear rings, nuclear bars, and nuclear star clusters (NSCs) (Schultheis et al., 4 Sep 2025).

A systematic HST-based census of nearby early-type galaxies provided the first large demographic baseline. In a volume-limited sample of 466 galaxies, 63 showed clear NSD signatures, yielding an observed fraction $f_{\rm obs}\approx13.5\%\pm1.6\%$. Correcting for projection effects using the detectability limit $\cos i<0.6$ raised the inferred incidence to $f_{\rm NSD}\approx22.8\%\pm2.5\%$. Within that sample, the fraction did not show a strong dependence on Hubble type or galactic environment, whereas the incidence declined in the most massive systems (Ledo et al., 2010).

The distinction between NSDs and NSCs is especially clear in the Milky Way. There, the NSC is roughly spherical, has effective radius $\sim5$ pc, and mass $\simeq2.5\times10^7\,M_\odot$, whereas the NSD is much larger, disc-like, and partially overlaps with the Central Molecular Zone. The NSD is kinematically colder and rotation-dominated relative to the NSC, and the two components have different stellar populations and star-formation histories (Nogueras-Lara et al., 2021).

2. Observational identification and dynamical characterization

The classical photometric description of an NSD is exponential. In surface brightness or surface density notation,

$I(R)=I_0\,\exp(-R/h),$

or equivalently,

$\Sigma(R)=\Sigma_0\,e^{-R/h}.$

Observed NSDs are therefore identified by decomposing the nuclear light into a rounder bulge plus an inclined, infinitesimally thin exponential disc. In practice, ellipse fits, unsharp masking, and Scorza-Bender-type decompositions are used to isolate the disc component, with a positive fourth cosine Fourier coefficient $100\!-\!500$0 or $100\!-\!500$1 marking disky isophotes (Corsini et al., 2015, Portaluri et al., 2013).

A widely used kinematic complement is KINEMETRY-based analysis of velocity fields and isophotes. For photometric maps, deviations from perfect ellipses are expanded as Fourier series, and the fourth cosine term is

$100\!-\!500$2

In merger-remnant tests designed to mimic realistic detectability, NSD-like photometric signatures were defined by a mean $100\!-\!500$3, while kinematic detection required $100\!-\!500$4 and $100\!-\!500$5. Those same simulations also showed that photometric detection rapidly becomes difficult for $100\!-\!500$6, while kinematic signatures remain more robust at lower inclinations (Sarzi et al., 2015).

The kinematic appearance of NSDs is that of a fast-rotating, low-dispersion component embedded in a slower, more pressure-supported background. Earlier integral-field studies of external galaxies described them as high-inclination, fast-rotating, low-$100\!-\!500$7 components, and later reviews generalized this to inclination-corrected peak $100\!-\!500$8, often accompanied by central $100\!-\!500$9-drops or $10^8\!-\!10^9\,M_\odot$0-hollows in barred hosts (Corsini et al., 2015, Schultheis et al., 4 Sep 2025).

For the Milky Way, NSD dynamics have been modeled with both Jeans equations and self-consistent action-based distribution functions. The latter treat the NSD as an axisymmetric equilibrium system with a quasi-isothermal DF $10^8\!-\!10^9\,M_\odot$1, solve self-consistently for the NSD potential, and fit line-of-sight velocities and proper motions while explicitly accounting for Galactic bar contamination. These models indicate that axisymmetry provides a good representation of the currently available data and furnish a full 6D phase-space DF within the AGAMA framework (Sormani et al., 2021, Sormani et al., 2020).

3. Stellar populations, ages, and chemistry

NSDs are not defined only by geometry and kinematics; their stellar populations are often markedly distinct from their surroundings. In the SB0 galaxy NGC 1023, HST imaging plus integral-field spectroscopy showed that the NSD is younger and more metal rich than the host bulge. A constrained stellar-population decomposition yielded $10^8\!-\!10^9\,M_\odot$2 Gyr and $10^8\!-\!10^9\,M_\odot$3 dex, while the extrapolated bulge values were $10^8\!-\!10^9\,M_\odot$4 Gyr and $10^8\!-\!10^9\,M_\odot$5 dex. In that system the NSD therefore represents a chemically enriched, relatively young nuclear component superposed on an older bulge (Corsini et al., 2015).

Later syntheses of external NSDs emphasize radial stellar-population gradients. Two-dimensional age and metallicity maps from nearby systems show negative radial age gradients, with centres $10^8\!-\!10^9\,M_\odot$6 Gyr old and outskirts $10^8\!-\!10^9\,M_\odot$7 Gyr, together with negative metallicity gradients and positive $10^8\!-\!10^9\,M_\odot$8 gradients. Those patterns are interpreted as evidence for inside-out growth, in which progressively younger stars appear at larger radii (Schultheis et al., 4 Sep 2025).

The Milky Way supplies the most detailed stellar-population constraints. Near-infrared color-magnitude studies separated NSD and NSC populations along the line of sight through their different extinction layers and red-clump morphologies. The NSD shows a double red clump, whereas the NSC requires three Gaussian components. Synthetic CMD fitting indicated that the NSD contains $10^8\!-\!10^9\,M_\odot$9 of its stellar mass in an old $f_{\rm obs}\approx13.5\%\pm1.6\%$0 Gyr population and $f_{\rm obs}\approx13.5\%\pm1.6\%$1 in an intermediate-age $f_{\rm obs}\approx13.5\%\pm1.6\%$2 Gyr population, with metal-rich stars at $f_{\rm obs}\approx13.5\%\pm1.6\%$3. A later luminosity-function analysis found a line-of-sight age gradient: the sample from the closest edge of the NSD contains $f_{\rm obs}\approx13.5\%\pm1.6\%$4 of its stellar mass in intermediate-age stars ($f_{\rm obs}\approx13.5\%\pm1.6\%$5 Gyr), whereas the deeper sample is $f_{\rm obs}\approx13.5\%\pm1.6\%$6 older than 7 Gyr (Nogueras-Lara et al., 2021, Nogueras-Lara et al., 2023).

Chemical and chemo-dynamical work further decomposes the Milky Way NSD into a dominant metal-rich component and a minority metal-poor one. A two-Gaussian metallicity model yielded $f_{\rm obs}\approx13.5\%\pm1.6\%$7 dex with $f_{\rm obs}\approx13.5\%\pm1.6\%$8 dex and $f_{\rm obs}\approx13.5\%\pm1.6\%$9 dex with $\cos i<0.6$0 dex; the metal-rich population constitutes $\cos i<0.6$1 of the stars and is kinematically cooler, with proper-motion dispersions $\cos i<0.6$2 mas yr$\cos i<0.6$3 and $\cos i<0.6$4 mas yr$\cos i<0.6$5, whereas the metal-poor population is hotter and more frequently occupies chaotic or box orbits. Metal-rich stars are predominantly on $\cos i<0.6$6-tube orbits, while chaotic/box orbits are more common among metal-poor stars. Detailed high-resolution abundance analysis of nine M giants extended this picture to 18 elements: NSD abundance trends broadly match the inner bulge and NSC, 17 elements follow local thick-disc behaviour at subsolar metallicities, and sodium is the only element with a clearly distinct trend, being enhanced in the NSD and NSC at supersolar metallicity. No typical globular-cluster abundance signatures were detected in the metal-poor NSD stars (Nogueras-Lara et al., 2024, Ryde et al., 21 May 2025).

The transition between the NSC and NSD also carries a chemical discontinuity. Integral-field spectroscopy at $\cos i<0.6$7 pc found that metallicity decreases from the central NSC through the transition region toward the NSD, with a steep gradient of $\cos i<0.6$8 dex per 10 pc inside $\cos i<0.6$9 pc, much steeper than the NSD gradient of $f_{\rm NSD}\approx22.8\%\pm2.5\%$0 dex per 10 pc at $f_{\rm NSD}\approx22.8\%\pm2.5\%$1 pc. That sharp change argues that the NSC and NSD are chemically distinct components with different formation histories, even if they may exchange gas or stars over time (Feldmeier-Krause, 2022).

4. Formation pathways and secular evolution

The dominant formation scenario for NSDs is in-situ star formation from gas driven into the nucleus by non-axisymmetric torques, especially bars. In high-resolution barred-galaxy simulations, once the stellar bar strengthens it removes angular momentum from the gas and drives a central inflow. The inflowing gas settles into a rotating star-forming nuclear disc that is thinner, younger, kinematically cooler, and more metal rich than the surrounding bar. In that simulation the nuclear disc is elliptical and orthogonal to the bar, and its signature is subtle in kinematics but strong in age and metallicity maps (Cole et al., 2014).

The Milky Way analogue simulated with SWIFT extends this secular picture. There the NSD is a rotationally supported stellar structure confined to $f_{\rm NSD}\approx22.8\%\pm2.5\%$2 pc and $f_{\rm NSD}\approx22.8\%\pm2.5\%$3 pc, with an inner NSC at $f_{\rm NSD}\approx22.8\%\pm2.5\%$4 pc. Its star-formation history displays a main burst at bar formation time, with $f_{\rm NSD}\approx22.8\%\pm2.5\%$5 Gyr, peak $f_{\rm NSD}\approx22.8\%\pm2.5\%$6, and $f_{\rm NSD}\approx22.8\%\pm2.5\%$7 over $f_{\rm NSD}\approx22.8\%\pm2.5\%$8 Gyr. After bar formation, secondary bursts recur when spiral arms and the bar reconnect while rotating at different pattern speeds, temporarily enhancing gas inflow into the inner kiloparsec. This suggests that NSD star-formation histories can retain a record not only of bar formation but also of later bar-spiral coupling (Boin et al., 22 Jun 2026).

Observationally, the Milky Way NSD supports the same secular-inflow picture. VLT/KMOS spectroscopy showed that the metal-rich population rotates much faster than the metal-poor one, and the metal-rich stellar rotation agrees with the rotation of molecular gas tracers in the Central Molecular Zone to within $f_{\rm NSD}\approx22.8\%\pm2.5\%$9 over $\sim5$0. The metal-poor stars rotate much more slowly and show hints of counter-rotation. The metal-rich stars are therefore interpreted as a dynamically cool, chemically distinct population plausibly formed from CMZ gas funnelled inward by the bar (Schultheis et al., 2021).

Alternative channels remain important in some systems. In NGC 1023, the NSD’s young age and super-solar metallicity support in-situ star formation from metal-enriched gas driven to the centre, either from the main disc or from external gas associated with the nearby satellite NGC 1023A; the dissipationless assembly of the NSD through star-cluster migration is rejected there. By contrast, dissipationless $\sim5$1-body experiments show that when multiple star clusters are accreted onto a pre-existing rapidly rotating nuclear disc, the remnant can display the photometric and kinematic properties of an observed NSD, and as much as $\sim5$2 of the NSD mass can be assembled from accreted clusters. This suggests that the formation channel is not unique, but depends on the dynamical state of the pre-existing nuclear component and on whether gas dissipation is available (Corsini et al., 2015, Portaluri et al., 2013).

5. Merger resilience and use as assembly chronometers

A central reason NSDs are astrophysically useful is that they can function as merger clocks. If a thin NSD is destroyed by a sufficiently violent merger and is not later rebuilt, then the age of its stars sets a lower bound on the time since the last disruptive event. Early arguments in this direction were strengthened by dedicated $\sim5$3-body experiments in which a nuclear disc embedded in a bulge and central black hole is perturbed by the supermassive black hole of an incoming secondary galaxy (Sarzi et al., 2015).

Mass ratio	$\sim5$4 (raw)	$\sim5$5 (raw)
1:1	19.0%	38.1%
1:5	61.9%	71.4%
1:10	85.7%	100.0%

These experiments show a clear mass-ratio dependence. Major $\sim5$6 mergers leave little trace of thin NSDs, especially photometrically, whereas $\sim5$7 and $\sim5$8 encounters preserve disc signatures in the majority of remnants, with kinematic signatures surviving more robustly than photometric ones. Prograde encounters with $\sim5$9 are more destructive because resonant torques are stronger and orbital decay is faster; retrograde encounters with $\simeq2.5\times10^7\,M_\odot$0 are less destructive. Photometric detections also drop rapidly as the system approaches face-on orientation, while some major-merger remnants retain disc-like kinematics only when viewed close to edge-on. On that basis, the presence of a thin NSD with both photometric and kinematic signatures implies a low $\simeq2.5\times10^7\,M_\odot$1 probability of a subsequent major merger (Sarzi et al., 2015).

The merger-clock interpretation is not, however, universal across all encounter classes. A later study of the Fornax galaxy FCC 170 examined dry $\simeq2.5\times10^7\,M_\odot$2 encounters and found that the NSD is comparatively resilient: its $\simeq2.5\times10^7\,M_\odot$3 expands by only $\simeq2.5\times10^7\,M_\odot$4 pc in most runs, its thickness parameter $\simeq2.5\times10^7\,M_\odot$5 rises only from $\simeq2.5\times10^7\,M_\odot$6 to $\simeq2.5\times10^7\,M_\odot$7 in four cases, and the central velocity-dispersion dip is preserved. In that system, the larger kpc-scale thin disc is much more strongly affected than the NSD. The interpretation offered there is that NSDs are better tracers of major encounters, while larger thin galactic discs are more informative about intermediate-mass-ratio dry mergers (Anta et al., 2023).

Taken together, these results establish that NSD fragility is regime-dependent rather than absolute. Thin NSDs are clearly vulnerable to major mergers, but not every intermediate-ratio dry interaction erases them. The utility of NSDs as assembly chronometers therefore depends on the merger mass ratio, orbital geometry, gas content, and on whether post-merger gas inflow can rebuild a new nuclear disc (Sarzi et al., 2015, Anta et al., 2023).

6. The Milky Way NSD as a resolved benchmark

Because individual stars can be resolved in the Galactic centre, the Milky Way provides the most complete physical characterization of an NSD. Axisymmetric dynamical models based on line-of-sight velocities and VIRAC2 proper motions give $\simeq2.5\times10^7\,M_\odot$8, $\simeq2.5\times10^7\,M_\odot$9 pc, $I(R)=I_0\,\exp(-R/h),$0 pc, and a velocity dispersion $I(R)=I_0\,\exp(-R/h),$1 that declines with radius. Earlier Jeans models found $I(R)=I_0\,\exp(-R/h),$2 and $I(R)=I_0\,\exp(-R/h),$3, and suggested the possibility of a vertically biased disc with $I(R)=I_0\,\exp(-R/h),$4, although that inference depends on the inner-dispersion profile, axisymmetry, and tracer selection. In both approaches the NSD dominates the gravitational field over $I(R)=I_0\,\exp(-R/h),$5 pc and rotates at roughly $I(R)=I_0\,\exp(-R/h),$6 (Sormani et al., 2021, Sormani et al., 2020).

Resolved-star studies constrain the NSD’s three-dimensional structure. Proper-motion separation of eastward- and westward-moving red-clump stars gives a near-to-far edge separation of $I(R)=I_0\,\exp(-R/h),$7 pc, consistent with a projected radius $I(R)=I_0\,\exp(-R/h),$8 pc and therefore with an axisymmetric structure along and across the line of sight. Independent CMD and extinction-layer analyses found de-reddened red-clump distances $I(R)=I_0\,\exp(-R/h),$9 kpc for the NSD and $\Sigma(R)=\Sigma_0\,e^{-R/h}.$0 kpc for the NSC, implying that the NSD stars are $\Sigma(R)=\Sigma_0\,e^{-R/h}.$1 pc closer than the NSC population along that sightline. The same work measured $\Sigma(R)=\Sigma_0\,e^{-R/h}.$2 mag and $\Sigma(R)=\Sigma_0\,e^{-R/h}.$3 mag, demonstrating a differential extinction of $\Sigma(R)=\Sigma_0\,e^{-R/h}.$4 mag between the two components (Nogueras-Lara, 2022, Nogueras-Lara et al., 2021).

A continuing observational question is whether the Milky Way hosts a true nuclear bar or instead an approximately axisymmetric NSD. Current stellar kinematics do not require a nuclear bar, and the symmetry of the red-clump proper-motion structure is consistent with a roundish, rotating stellar disc, although a weak bar cannot be ruled out. This leaves the issue open to future higher-precision astrometry and spectroscopy (Nogueras-Lara, 2022, Schultheis et al., 4 Sep 2025).

The Milky Way NSD is also characterized across the electromagnetic spectrum. In hard X-rays, SRG/ART-XC observations in the 4–12 keV band show that the Galactic-centre emission on scales of a few hundred parsecs is dominated by a regularly shaped NSD aligned with the Galactic plane, with projected longitudinal and latitudinal scale heights of approximately 100 pc and 20 pc. The measured absorbed flux is $\Sigma(R)=\Sigma_0\,e^{-R/h}.$5, corresponding to an unabsorbed luminosity of $\Sigma(R)=\Sigma_0\,e^{-R/h}.$6 at 8.178 kpc. Using $\Sigma(R)=\Sigma_0\,e^{-R/h}.$7, the mean mass-normalized emissivity is $\Sigma(R)=\Sigma_0\,e^{-R/h}.$8, exceeding the canonical Galactic-ridge value by a factor $\Sigma(R)=\Sigma_0\,e^{-R/h}.$9. The close match between X-ray luminosity density and stellar mass density leaves only a small room for any diffuse $100\!-\!500$00 keV plasma component in the NSD region (Nezabudkin et al., 6 Jul 2025).

The Milky Way therefore serves as the reference laboratory in which NSD morphology, dynamics, extinction structure, chemo-dynamics, and even integrated high-energy emission can all be studied simultaneously. That uniquely resolved view has made the Galactic NSD central to current attempts to connect nuclear discs with bar-driven gas inflow, inside-out growth, and the long-term evolution of galaxy nuclei (Sormani et al., 2021, Schultheis et al., 4 Sep 2025).