Chemical analysis of the Milky Way's Nuclear Star Cluster: Evidence for a metallicity gradient

TL;DR

This work reanalyzes the Milky Way's Nuclear Star Cluster using KMOS spectra with an updated line list tailored for cool M giants, achieving precise stellar parameters and [$\alpha$/Fe] by a two-step Bayesian spectral fitting workflow that includes NLTE corrections. Validation against high-resolution data shows typical uncertainties of $\Delta T_{\rm eff}\simeq 150$ K, $\Delta \log g\simeq 0.4$ dex, $\Delta$[M/H] $\simeq 0.2$ dex, and $\Delta [\alpha/Fe]\simeq 0.10$ dex, enabling a robust two-component MDF: metal-poor at $[M/H]\approx -0.77$ dex and metal-rich at $[M/H]\approx +0.26$ dex. A clear radial metallicity gradient of $\sim -0.109^{+0.018}_{-0.016}$ dex/pc is detected, consistent with inside-out formation of the MWNSC and potentially connected to the MWNSD. The results imply a substantial metal-poor population in the southern outer MWNSC and underscore the importance of improved line lists and NLTE treatment for accurate chemical mapping in the Galactic center.

Abstract

The Milky Way nuclear star cluster (MWNSC) is located together with its surrounding nuclear stellar disc (MWNSD) in the Galactic centre and they dominate the gravitational potential within the inner 300\,pc. However, the formation and evolution of both systems and their possible connections are still under debate. We reanalyse the low-resolution KMOS spectra in the MWNSC with the aim to improve the stellar parameters ($\rm T_{eff}$, $\rm \log\,g$, and $\rm [M/H])$ for the MWNSC. We use an improved line-list, especially dedicated for cool M giants allowing to improve the stellar parameters and to obtain in addition global $\rm α$-elements. A comparison with high-resolution IR spectra (IGRINS) gives very satisfactory results pinning down the uncertainties to $\rm T_{eff} \simeq 150\,K$, $\rm log\,g \simeq 0.4\,dex$, and $\rm [M/H] \simeq 0.2\,dex$. Our $\rm α$-elements agree within 0.1\,dex compared to the IGRINS spectra. We obtain a high-quality sample of 1140 M giant stars where we see an important contribution of a metal-poor population ($\rm \sim 20\,\%$) centered at $\rm [M/H] \simeq -0.7\,dex$ while the most dominant part comes from the metal-rich population with $\rm [M/H] \simeq 0.26\,dex$. We construct a metallicity map and find a metallicity gradient of $\rm \sim -0.1 \pm 0.02 \,dex/pc$ favouring the inside-out formation scenario for the MWNSC.

Figures (9)

• Figure 1: GALACTICNUCLEUS $JHK_s$ false-colour image of the region covered by FK1720. Blue circles and white diamonds mark stars classified as belonging to the MWNSD and MWNSC, respectively, according to the criterion described in Sect. \ref{['membership']}. The white dashed line indicates the effective radius of the MWNSC, and the compass shows Galactic coordinates.
• Figure 2: H--K vs. K colour magnitude diagram of our data sample. Stars member of the MWNSC are indicated in red, in blue stars of being member of the MWNSD are indicated and in white we show the foreground objects. For our work here we only use stars in the MWNSC.
• Figure 3: Completeness of the KMOS FK1720 sample. Upper panel: $\rm K_s$ luminosity functions from the KMOS sample and the reference sample of GALACTICNUCLEUS (GNS) stars in the region. The associated uncertainty was estimated as the square root of the number of stars per magnitude bin. Lower panel: Completeness function obtained by comparing the two $\rm K_s$ luminosity functions.
• Figure 4: Example of a spectral fit using STARKIT for starId 30001001. The normalized, observed spectrum is in black, the best synthetic fit is indicated in red. The vertical bands show the masks applied: Blue is the $\rm Br_{\gamma}$ region, green the region around the NaI line, magenta the region around the CaI line and orange the CO band heads
• Figure 5: Comparison between stellar parameters of high-resolution spectra to degraded spectra at the resolution of KMOS. The y-axis shows the difference between low-resolution and high-resolution spectra, respectively while the x-axis displays the value from the low-resolution work. The shaded gray are shows $\rm \pm 1\,\sigma$ levels for each parameter comparison. Mean difference and standard deviations are indicated on the upper right. Black dots show the SN sample while red dots the NSC stars from Nandakumar2025