The galaxy bias profile of cosmic voids

Abstract

Cosmic voids are underdense regions within the large-scale structure of the Universe, spanning a wide range of physical scales - from a few megaparsecs (Mpc) to the largest observable structures. Their distinctive properties make them valuable cosmological probes and unique laboratories for galaxy formation studies. A key aspect to investigate in this context is the galaxy bias, $b$, within voids - that is, how galaxies in these underdense regions trace the underlying dark-matter density field. We want to measure the dependence of the large-scale galaxy bias on the distance to the void center, and to evaluate whether this bias profile varies with the void properties and identification procedure. We apply a void identification scheme based on spherical overdensities to galaxy data from the IllustrisTNG magnetohydrodynamical simulation. For the clustering measurement, we use an object-by-object estimate of large-scale galaxy bias, which offers significant advantages over the standard method based on ratios of correlation functions or power spectra. We find that the average large-scale bias of galaxies inside voids tends to increase with void-centric distance when normalized by the void radius. For the entire galaxy population within voids, the average bias rises with the density of the surrounding environment and, consequently, decreases with increasing void size. Due to this environmental dependence, the average galaxy bias inside S-type voids - embedded in large-scale overdense regions - is significantly higher ($\langle b\rangle_{\rm in} > 0$) at all distances compared to R-type voids, which are surrounded by underdense regions ($\langle b\rangle_{\rm in} < 0$). The bias profile for S-type voids is also slightly steeper. Since both types of voids host halo populations of similar mass, the measured difference in bias can be interpreted as a secondary bias effect.

Figures (11)

• Figure 1: The spatial distribution of galaxies and cosmic voids in a 10-$h^{-1}$Mpc slice ($100 < z[h^{-1}$Mpc$] < 110$) extracted from the TNG300 simulation box. Gray dots represent galaxies in the parent catalog within this region (see Sect. \ref{['sec:data']}), while purple circles indicate the cross-sectional radii of the corresponding voids from the fiducial sample that intersect this slice (see Sect. \ref{['sec:voids']}).
• Figure 2: The fiducial void sample extracted from TNG300, based on the selection described in Section \ref{['sec:voids']}. Left: Distribution of void radii in the sample. Right: Relation between the overdensity parameters $\Delta(R_{\rm void})$ and $\Delta_{\rm max}$, where symbols are color-coded by void radius, and their sizes are also proportional to void radius.
• Figure 3: The galaxy bias profile inside voids as a function of void size. Left: The average galaxy bias as a function of distance to the center of each void in our fiducial void sample. Three different void radius ranges are indicated by color. The average galaxy bias profile for each of these subsets are also shown, where uncertainties correspond to the errors on the means. Right: The average galaxy bias profile for the same subsets but using the void-centric distance normalized by void radius. Only central galaxies are employed in this analysis. A non-uniform binning has been adopted to avoid discreteness issues in the innermost regions of voids (see text). This binning is maintained in all figures.
• Figure 4: The average galaxy bias profile for R-type and S-type voids, where the distance to the center of voids have been normalized by the void radius, $R_{\rm void}$. The uncertainties correspond to the errors on the means.
• Figure 5: The average galaxy bias as a function of void radius, $R_{\rm void}$ (left), and maximum overdensity in the surrounding region, $\Delta_{\rm max}$ (right). The uncertainties correspond to the errors on the means. Note that, as shown in Fig. \ref{['fig:sample']}, voids with high values of $\Delta_{\rm max}$, which also tend to have high bias, are very scarce. At fixed $R_{\rm void}$, the population is dominated by lower-$\Delta_{\rm max}$ (lower-$\langle b \rangle$ voids), which explains why the average bias only reaches $\sim$0.5 in the left panel of this figure.