MUSE-DARK

I. Dark matter halo properties of intermediate-z star-forming galaxies

¹ Univ. Lyon1, Ens de Lyon, CNRS, Centre de Recherche Astrophysique de Lyon (CRAL), 69230 Saint-Genis-Laval, France
² Aix Marseille Univ, CNRS, CNES, LAM, Marseille, France
³ Leibniz-Institut für Astrophysik Potsdam (AIP), An der Sternwarte 16, 14482 Potsdam, Germany
⁴ Institut de Recherche en Astrophysique et Planétologie (IRAP), Université de Toulouse, CNRS, UPS, CNES, 31400 Toulouse, France

^★ Corresponding author: This email address is being protected from spambots. You need JavaScript enabled to view it.

Received: 24 September 2025
Accepted: 12 February 2026

Abstract

Context. The core–cusp problem is a key challenge to the Lambda cold dark matter (ΛCDM) model. Stellar feedback and other baryonic processes have been proposed as solutions to reconcile the tension between the predicted cuspy profiles and the observed cores of low- and intermediate-mass galaxies. However, at z > 1, disk–halo decompositions usually focus on massive star-forming galaxies (SFGs) with log(M_★/M_⊙) > 10, as observations typically lack the depth and sensitivity to probe lower-mass systems, where cores are predicted to occur.

Aims. For this study we analysed the dark matter (DM) halo properties of 127 intermediate-redshift (0.3 < z < 1.5) SFGs down to low stellar masses (8 < log(M_★/M_⊙) < 11), using the highest signal-to-noise data from the MUSE Hubble Ultra Deep Field Survey, as well as photometry from the Hubble Space Telescope and James Webb Space Telescope.

Methods. We employed a traditional 2D line fitting algorithm and a 3D forward modelling approach to analyse the kinematics of our sample, enabling us to measure individual rotation curves extending up to two to three times the effective radii. We performed a disk–halo decomposition with a 3D parametric model, which includes stellar, DM, and gas components, as well as corrections for pressure support. We tested our methodology using mock data cubes generated from idealised disk simulations, and we performed several cross-checks, such as comparing our 3D disk–halo decomposition with results obtained from external data, as well as 1D decompositions, finding good agreement. We tested six DM density profiles, including the Navarro–Frenk–White, Burkert, Einasto and the generalised αβγ profile of (Di Cintio, A., Brook, C. B., Dutton, A. A., et al. 2014, MNRAS, 441, 2986, DC14), as well as a baryon-only model using a Bayesian analysis.

Results. Marginalising against the unknown neutral gas content, we find that the mass-dependent DC14 DM profile, which accounts for the response of DM to baryonic processes such as stellar feedback, performs as well as or better than the other halo models for most of the sample (≳80%), and that baryon-only models seem to be disfavoured with respect to DM models for 84% (107/127) of the galaxies. We find that 89% of the SFGs have DM fractions (f_DM(< R_e)) larger than 50%, and we extend the f_DM(< R_e)−Σ_M★ relation at lower masses. Using DC14 DM profiles, we infer DM inner slopes γ < 0.5 for 66% of the sample, indicative of cored DM density profiles. The stellar–halo mass and concentration–halo mass relations inferred from our 3D modelling agree with the theoretical expectations, albeit with larger scatter. Our results confirm the anti-correlation between the halo scale radius and DM density with a slope of ∼ − 1, which seems to evolve with redshift. While the halo scale radii are z-invariant, we find tentative evidence that DM halos of z ∼ 1 SFGs were denser (by ∼0.3 dex) than those in the local Universe.

Conclusions. We measured DM halo properties of intermediate-z SFGs down to 10⁸ M_⊙, and find that a substantial fraction of the sample can be described by cored DM density profiles. This may point towards core formation driven by baryonic processes in the context of ΛCDM.