The stellar and dark matter distributions in elliptical galaxies measured by stacked weak gravitational lensing

We investigate stellar mass and central dark matter density profiles of photometric luminous red galaxies with stellar masses of $\sim10^{10}-10^{12}M_\odot$ using weak gravitational lensing measurements from the Hyper Suprime-Cam Subaru Strategic Program data obtained with the Subaru Telescope. By stacking weak lensing signals from a large number of galaxies, we obtain average tangential shear profiles down to $\sim 10,\mathrm{kpc}/h$, which are fitted assuming a two-component model consisting of stellar and dark matter components to constrain their central dark matter distribution. We find a preference for non-zero core radii of dark matter distributions in galaxies with stellar masses of $\sim 10^{11}M_\odot$. Our results imply a stronger feedback effect than that typically predicted by current hydrodynamical simulations. In addition, we provide a new constraint on the stellar-to-halo mass relation, where both stellar and halo masses are, for the first time, directly constrained by weak gravitational lensing. Our results prefer the stellar initial mass function (IMF) that is more bottom-heavy than the Salpeter IMF.

💡 Research Summary

This paper presents the first systematic study of the stellar and dark‑matter density profiles in the centers of elliptical (luminous red) galaxies using stacked weak gravitational lensing from the Hyper Suprime‑Cam Subaru Strategic Program (HSC‑SSP). The authors select a photometric LRG sample covering ~1200 deg², with redshifts 0.05 ≤ z ≤ 1.25 and stellar masses 10^{10.3}–10^{12} M⊙ (initially derived assuming a Salpeter IMF). Using the HSC‑SSP three‑year shear catalog (≈430 deg², source density ≈23 arcmin⁻²), they measure the differential surface density ΔΣ(r) around the lenses by stacking the tangential shear of background galaxies. Background sources are assigned photometric redshifts from a DNN‑based estimator, and a stringent P(z) cut (Δz = 0.1, P_cut = 0.95) is applied to suppress dilution by physically associated galaxies.

The inner mass profile (0.013–0.1 Mpc h⁻¹) is modeled as the sum of a stellar component (Hernquist profile, with scale radius tied to the observed half‑light radius) and a dark‑matter component described by a cored power‑law density: ρ_DM ∝ (r² + r_c²)^{-(γ+2)/2}. This functional form allows analytic expressions for the lensing quantities Σ(r) and ΔΣ(r), facilitating a χ² fit to the stacked data. The free parameters are the fitted stellar mass M_*_fit, the dark‑matter normalization A, the outer slope γ, and the core radius r_c. The outer region (0.04–3 Mpc h⁻¹) is fitted with a truncated NFW halo for central galaxies, a satellite term derived from the Fourier transform of the same profile, and a 2‑halo term, allowing simultaneous constraints on the average halo mass M, the satellite fraction f_sat, and the typical host halo mass of satellites M_h.

Key results: (1) For galaxies with M_* ≈ 10^{11} M⊙, the data strongly prefer a non‑zero core radius, r_c ≈ 0.02–0.05 Mpc h⁻¹, over a pure NFW cusp (Δχ² ≈ 5). The outer slope γ is close to 1, indicating an NFW‑like behavior at larger radii. (2) The fitted stellar masses are ∼30 % higher than the photometric estimates, implying a stellar IMF heavier than Salpeter (i.e., a bottom‑heavy IMF). (3) The derived stellar‑to‑halo mass relation lies above the canonical SHMR from abundance‑matching studies, suggesting that feedback in these systems is less efficient than predicted by current hydrodynamical simulations. (4) Comparison with state‑of‑the‑art simulations (EAGLE, IllustrisTNG) shows that the observed core sizes are significantly larger than the ≲0.01 Mpc h⁻¹ cores predicted for halos of this mass, indicating that stronger central feedback (e.g., more energetic AGN outflows) may be required.

Systematic tests—including varying the P(z) cut, propagating uncertainties in half‑light radii and stellar masses, and excluding the innermost 0.013 Mpc h⁻¹ where PSF and blending dominate—demonstrate that the core detection and IMF inference are robust. The study proves that stacked weak lensing can probe galaxy interiors down to ∼10 kpc, providing an independent avenue to disentangle stellar and dark‑matter contributions without relying on strong lensing or stellar dynamics. The authors conclude that their measurements place new constraints on the central dark‑matter structure of massive ellipticals and favor a bottom‑heavy IMF, offering valuable benchmarks for future surveys (LSST, Euclid, Roman) and for improving galaxy formation models.