2026A&A...705A.242B

Context. The Circinus galaxy is the nearest type-2 Seyfert galaxy, located at a distance of 4.2 Mpc. Its environment is challenging to explore because of the low Galactic latitudes of its location, behind the disc of the Milky Way.
Aims. The long-term goal of this work is to characterise the Circinus galaxy halo. We are also interested in determining the possible presence of dwarf satellites using near-infrared (NIR) data.
Methods. We selected 1542 galaxies from the VISTA Variables in the Vía Láctea (VVV) NIR galaxy catalogue within a 2-degree radius around the Circinus galaxy, representing roughly 2/3 of the virial radius. We used structural parameters, such as the half-light radii and colours, and correlations were examined. A neural network was trained with 486 galaxies with known spectroscopic redshifts to estimate photometric redshifts for all of these galaxies. The potential satellites of the Circinus galaxy were defined on the basis of half-light radii that were compatible with the typical sizes of dwarf satellites in galaxies and combined with photometric redshifts.
Results. The galaxy properties have been reliably characterised down to K_s ∼ 15.5 mag, representing about 90% completeness of the detections. At the distance of the Circinus galaxy, this limiting magnitude corresponds to a K_s absolute magnitude of -12.6 mag, allowing us to locate dwarf galaxies. In total, there are 20 galaxies with half-light radii larger than 2.45 arcsec, but only 8 of them have photometric redshifts below 0.04. None of these galaxies are close to the Circinus galaxy, which has a redshift of 0.0015 and shows no evidence of associated clustering. The ANNz model exhibited a high degree of accuracy in the range of 0.001 < z_phot < 0.023, thereby validating this method in these extinct regions.
Conclusions. The presence of dwarf satellites associated with the Circinus galaxy could not be confirmed with the available data in the studied region. The apparent lack of satellites could be genuine, possibly related to AGN feedback effects. Nevertheless, this work demonstrates the effectiveness of combining near-infrared data and machine learning techniques to estimate photometric redshifts at low Galactic latitudes, thereby providing useful information for future spectroscopic follow-up campaigns.