Controlled Experiments on Dark-Matter Halo Structure and Galaxy Morphology I: What Sets Galaxy Sizes?

Abstract

The properties of galaxies are intricately linked to the characteristics of their host dark-matter haloes. We use a suite of controlled simulations of isolated galaxies to quantify how halo spin, concentration, inner density profile, and baryon fraction regulate galaxy sizes, at fixed halo mass of $M_{\rm{vir}}=10^{11} M_\odot$. We generate initial conditions of haloes and inhabitant spherical gas distributions in equilibrium, on a parameter grid spanned by these four halo parameters, and evolve the systems with the $\texttt{GIZMO}$ code and the $\texttt{FIRE-3}$ physics. The resulting half-mass radii of stars and cold baryons depend systematically on halo structure and baryon content: galaxy size increases with halo spin, decreases with halo concentration, is weakly sensitive to the inner density slope except in highly cuspy haloes, and is strongly suppressed at high baryon fractions. We evaluate the relative importance of the halo parameters on galaxy size using different metrics including the quadratic response-surface method and random-forest regression, and consistently find halo concentration to be the most informative predictor of size. The baryon fraction shows a subtle, non-monotonic impact on size, by modulating how galaxy size depends on halo spin. Our results clarify which secondary parameters of host dark-matter haloes dominate the scatter in galaxy sizes at the massive-dwarf mass scale.

Figures (8)

• Figure 1: Projected stellar surface density and gas temperature maps for simulations in which the halo spin parameter $\lambda$ is varied from 0.02 to 0.08 (one value per row, as indicated), while all other parameters are held at their fiducial values: concentration $c_2=10$, logarithmic dark-matter-density slope $s_1=1$, and baryon fraction $f_{\rm b}=0.07$. Columns 1 (2) and 3 (4) show the face-on (edge-on) views. Circles indicate $0.02\,R_{\rm vir}$ (dashed), the stellar half-mass radius $r_{1/2,\star}$ (solid), and the half-mass radius of stars plus cold gas $r_{1/2,\star+{\rm cg}}$ (dash-dotted).
• Figure 2: Same as Fig.\ref{['fig:maps_spin']}, but varying halo concentration $c_2$ from 3 to 20 while keeping all other parameters at their fiducial values: spin $\lambda=0.04$, inner DM density slope $s_1=1$, and baryon fraction $f_{\rm b}=0.07$.
• Figure 3: Same as Fig. \ref{['fig:maps_spin']}, but varying the inner logarithmic slope of dark matter density $s_1$ from 0 to 1.5, while keeping all other parameters at their fiducial values: spin $\lambda=0.04$, concentration $c_2=10$, and baryon fraction $f_{\rm b}=0.07$.
• Figure 4: Same as Fig. \ref{['fig:maps_spin']}, but varying the baryon fraction $f_{\rm b}$ relative to the fiducial value of 0.07 Hafen19 from $0.5\times$ to $2.0\times$, while keeping all other parameters at their fiducial values: spin $\lambda=0.04$, concentration $c_2=10$, and logarithmic dark-matter-density slope $s_1=1$.
• Figure 5: Galaxy size (half-stellar mass radius) as a function of halo structural parameters and baryon fraction. Points represent individual simulation results, coloured by a secondary parameter ($f_{\rm b}$ or $c_2$, as indicated). Thin, colour-matched lines connect simulations in which only the horizontal-axis quantity is varied while all other parameters are fixed. Connected grey symbols show the mean stellar half-mass radius, $r_{1/2,\star}$, with error bars and the light-grey bands indicating the $1\sigma$ scatter. The symbol positions are slightly offset horizontally to improve visual clarity The black dashed curve shows the mean prediction of the best-fit quadratic response-surface model (RSM), obtained by varying the parameter of interest while fixing all others at their midpoint values. Pearson correlation coefficients, $\mathcal{R}$, and the corresponding $p$-values are quoted.