The intragroup light in KiDS+GAMA groups: A stacking analysis

TL;DR

This work tackles the challenge of measuring intragroup light (IGL) in galaxy groups by performing a large-sample stacking analysis using KiDS imaging of GAMA groups. A custom reprocessing pipeline is developed to preserve low-surface-brightness features, enabling robust IGL measurements that are compared to Hydrangea simulations after accounting for PSF effects. The study provides upper and lower bounds on the IGL fraction $f_{ m IGL}$ across $0.09\leq z\leq 0.27$ and $12.5\leq \log_{10}[M_{200}/\mathrm{M}_\odot]\leq 14.0$, finding values in the few to ~20% range and no strong redshift evolution, with results broadly consistent with simulations and prior observational work. The analysis demonstrates the feasibility and value of stacked LSB studies in current and upcoming wide-field surveys (e.g., Euclid and LSST) for tracing the growth of diffuse stellar components in group-scale haloes. Key methodological insights include the importance of background modeling, satellite masking, and PSF characterization in recovering faint diffuse light from ensemble measurements.

Abstract

Galaxy groups and clusters assembled through dynamical interactions of smaller systems, resulting in the formation of a diffuse stellar halo known as the intragroup or intracluster light (IGL/ICL). By preserving the records of these interactions, the IGL/ICL provides valuable insight into the growth history of galaxy groups and clusters. Groups are especially interesting because they represent the link between galactic haloes and massive clusters. However, the low surface brightness of this diffuse light makes it extremely challenging to detect individually. Recent deep wide-field imaging surveys allow us to push such measurements to lower brightness limits by stacking data for large ensembles of groups. In this work, we present a special-purpose pipeline to reprocess individual $r-$band Kilo-Degree Survey (KiDS) exposures to optimise the IGL detection. Using an initial sample of 2385 groups with at least five spectroscopically-confirmed member galaxies from the Galaxy and Mass Assembly (GAMA) survey and reprocessed deep images from the KiDS, we present the first robust measurement of IGL from a large group sample (~ 750) down to 31-32 mag/arcsec$^2$ (varying in different stacked bins). We also compare our stacked IGL measurements to predictions from matched mock observations from the Hydrangea cosmological hydrodynamic simulations. Systematics in the imaging data can affect IGL measurements, even with our special-purpose pipeline. However, with a large sample and optimised analysis, we can place well-constrained upper and lower limits on the IGL fraction (3 - 21 per cent) for our group ensemble across $0.09\leq z\leq 0.27$ and $12.5\leq \log_{10}[M_{200}/\mathrm{M}_\odot] \leq 14.0$. This work explores the potential performance of stacked statistical analysis of diffuse light in large samples of systems from next-generation observational programs, such as $Euclid$ and LSST.

Figures (12)

• Figure 1: Distributions of different properties of the GAMA groups with $N_{\textrm{FoF}}\geq5$ in our sample. The absolute $r-$band magnitudes ($M_{\mathrm{r}}$) and redshifts ($z$) of the central galaxies (CG) were directly obtained from the GAMA-II Galaxy Group Catalogue Robotham2011. The halo masses were computed from the total $r-$band group luminosity using eqn. 37 of viola2015. The vertical lines in the left panels indicate the magnitude range of the group CGs used in this work.
• Figure 2: Flatfield (delta flat) in the $r$-band obtained by averaging the science observations that were already flatfielded using the standard AstroWISE pipeline. The values shown in the colorbars represent the (dimensionless) relative change with respect to the original flatfield. The top panel shows the full mosaic. The bottom figures show the bottom-left chip with a zoom-in of the top-left corner of that chip (red square). Some structure is visible, likely due to variations in the illumination, as well as some low-level fringing.
• Figure 3: Left: histogram of the background values in randomly placed annuli on the field images with polynomial sky subtractions of order 0 (blue), 1 (purple), and 2 (green). The vertical lines denote the corresponding 1 $\sigma$ values, as shown in the upper right corner. The scatter to the background values is smaller for higher-order polynomial estimation of the background. It is clear that higher-order polynomials remove spatial variation in the background more efficiently. Right: mean and scatter of the mean background values for the different polynomial background estimations (indicated in the upper right corner). This again demonstrates how the scatter is gradually reduced for higher-order polynomial estimation to the background values.
• Figure 4: Surface brightness profiles of the group central galaxies beyond 20 arcseconds radial distance from the centre for 0$^{\mathrm{th}}$ (blue), 1$^{\mathrm{st}}$ (purple), and 2$^{\mathrm{nd}}$ (red) order polynomial background estimations, respectively. In all the profiles, error bars indicate $1\sigma$ uncertainties on the mean. All three profiles are the same within 20 arcseconds (not shown). Beyond that, however, higher-order background estimations over-subtract the background compared to the lower-order ones. This is the most prominent for the 2$^{\mathrm{nd}}$-order polynomial estimation of the background.
• Figure 5: Stitched point spread function (PSF) from combining stars within different magnitude ranges (see text) from all fields with updated background-subtracted images. The colours and line styles indicate the radial range where different groups of stars (core, intermediate, outer 1, and outer 2) contributed to constructing the total PSF. The brown solid line shows the PSF constructed in the same way as above from standard KiDS data-release 4 images. The excess of faint light beyond 200 pixels in the PSF from the updated pipeline indicates the missing light in the standard KiDS pipeline.