AGN candidates in the VVV near-IR galaxy catalogue

Galaxies beyond the Milky Way are difficult to detect and identify due to the obscuring effects of dust and high stellar densities at low Galactic latitudes (e.g., Kraan-Korteweg & Lahav 2000; Baravalle et al. 2021). However, the study of hidden galaxies is of great importance due to they might help to identify large-scale structures beyond our Galaxy (e.g.,Huchra et al. 2012; Macri et al. 2019), and obtain a more complete luminosity and density distributions (e.g., Peebles 1980; Klypin et al. 1993). In this context, the near-infrared (NIR) observations are most suitable to probe these highly obscured objects. While foreground source density is higher in the NIR, reduced foreground extinction reveals galaxies that are invisible at optical wavelengths.

In this regard, the VISTA Variables in the Vía Láctea Survey (VVV, Minniti et al. 2010), offers an unprecedented opportunity to discover and study new different kind of objects that have remained hidden behind the Galactic plane. In the last few years considerable effort has been made in the detection of extragalactic sources using this survey. The first work on searching for extragalactic sources was made by Amôres et al. (2012) from visual inspection on specific regions. Then, Baravalle et al. (2018) proposed a methodology that used a combination of morphological and photometric parameters to detect galaxies. Using this procedure, Baravalle et al. (2021) generated the VVV near-IR galaxy catalogue (VVV NIRGC) that covers the regions of the Sourthern Galactic disc. Also, Coldwell et al. (2014); Baravalle et al. (2019) and Galdeano et al. (2021, 2022, 2023) used similar procedure to study the near-infrared properties of galaxies in galaxy clusters. These works were particularly focused on the galaxy detections. Pichel et al. (2020) made the first attempt to classify four galaxies obscured by the Milky Way as blazar candidates combining near and mid-infrared data analysis.

Active Galactic Nuclei (AGN) are among the most energetic phenomena known in the Universe. Since their discovery several decades ago, studies at different frequencies have revealed the richness of AGN phenomenology observed from radio to $\gamma$-rays, giving rise to a vast and enthralling zoo of classifications. A full overview of AGN multi-wavelength properties can be found in Padovani et al. (2017), who also discussed the current understanding of the unification model and highlight some open questions. They emphasise the importance of the increasingly larger and deeper sky surveys at all wavelengths for major breakthroughs in the understanding of this AGN phenomenon.

There are different classes of AGNs such as type-1 and type-2 AGNs, blazars (BL Lac, Falomo et al. 2014) and flat spectrum radio quasars (FSRQ), among many others. According to the unified model (Antonucci, 1993), different classes of AGNs would be the same type of object, mainly depending on the orientation with respect to the observer. The central engine would in all cases be a black hole accreting matter, and particularly type-1 AGNs, i.e. quasars (QSOs) and Seyfert 1 galaxies, usually present black holes with masses $\gtrsim$10⁶M_☉ (e.g., Ferrarese & Merritt, 2000; Bentz & Katz, 2015; Schmidt et al., 2021; Prince et al., 2022). In addition, AGNs usually exhibit variability in the emission (e.g., Nandra et al., 1997; Edelson et al., 2002; Sandrinelli et al., 2014; Pichel et al., 2020; Husemann et al., 2022). This variability is different according to the type of AGNs and in general it is more pronounced and with higher amplitudes in blazars than in type-1 AGNs (e.g., Mao & Yi, 2021).

Over the past years, the population of known AGNs has grown significantly with new surveys and catalogues (e.g., Véron-Cetty & Véron, 2010; Rembold et al., 2017; Koss et al., 2017; do Nascimento et al., 2019; Husemann et al., 2022); however, the number of AGNs at lower Galactic latitudes behind the Galactic disc remains limited due to the obscuring effects typical of these dense regions (Edelson & Malkan, 2012; Pichel et al., 2020). Recently, Fu et al. (2021, 2022) studied the regions of the Galactic bulge searching for QSOs at lower latitudes. They used machine learning techniques to find candidates and confirmed 204 QSOs at |b| $<$ 20^∘ based on spectroscopic measurements.

In this work, we present a study of AGN candidates using near- and mid-IR data in the crowded regions of the southern Galactic disc at very low Galactic latitudes (|b| $<$ 2^∘). We analyse their colour-colour diagrams and variability; and apply different classification criteria in order to understand for the first time the nature of these AGN candidates. The paper is organised as follows. Section 2 presents the observational data used in this work and Section 3, the variability analysis. Section 4 presents the discussion and final remarks.

We obtained the profile fitted $K_{s}$ band light curves from which the VIVACE sources were originally classified (see Molnar et al., 2022, for a description of these data). The differential $K_{s}$ light curves of the five studied sources are shown in Figure 6. These are median subtracted curves sampled over more than 3000 days. The shape of the light curves is, in general, irregular without any clear periodicity. In some cases, peak-shaped changes in brightness are observed, which are significant with respect to the uncertainties.

In order to analyse the variability of the objects, we applied two different statistical methods. First, we studied the fractional variability amplitude $\sigma_{rms}$, defined as:

where N is the number of flux values $F_{i}$, with measurement uncertainties $\epsilon_{i}$, and $\mu$ is the average flux. This parameter is widely used to evaluate and analyse variability in light curves (e.g., Nandra et al., 1997; Edelson et al., 2002; Sandrinelli et al., 2014; Pichel et al., 2020), representing an excess of variability that cannot be only generated by flux errors.

Secondly, we analysed the slope of the light curves taking into account that Cioni et al. (2013) showed that 75% of the QSOs have a slope variation in the $K_{s}$ passband $>10^{-4}$ mag/day. These authors derived the slope of the overall $K_{s}$ variation in light curves sampled over 300$-$600 days, 40$-$80 days or shorter range. In order to carry out this analysis, we performed a linear fit of the $K_{s}$ light curves considering a range of days defined by the highest and lowest variations. In all cases, the time range considered is longer than 1200 days.

Table 4 shows the $K_{s}$ variability of the five studied sources. As mentioned above, the shape of the light curves is in certain cases irregular and they do not show any clear periodicity. The VIVACE catalogue contains some mis-classifications of aperiodic variable sources as they were not included in the training set and the classification system. Indeed, Molnar et al. (2022) noted that aperiodic YSOs and/or AGNs, for example, might contaminate the long period variable category. However, for completeness, we added in the table the periodicity from VIVACE, which is a statistical parameter (see Molnar et al., 2022, for more details). We visually inspected the light curves phase-folded with the VIVACE period and did not find them to be particularly convincing fits to these data.

In this work we studied the five sources from the VVV NIRGC that also have WISE counterparts and present variability in the VIVACE catalogue. In the near-IR colour-colour diagrams, our objects present values ($H$ - $K_{s}$) $>$ 0.5 mag and ($J$ - $H$) $>$ 0.55 mag, being in general redder than most of the remaining sources (Fig. 3). The two galaxies with $Y$ photometric data (VVV-J131828.35-635442.3 and VVV-J134416.68-632638.7) are both located in the QSO region following Cioni et al. (2013).

Analysing the mid-IR colour-colour diagram (Fig. 4), we found that the five sources satisfy both the Assef et al. (2018) and Stern et al. (2012) criteria, suggesting that our sources could be AGNs. In addition, the objects VVV-J121313.50-613155.1 and VVV-J134416.68-632638.7 are located in the AGN region defined empirically by Jarrett et al. (2011). The object VVV-J125631.40-611626.0 is in the YSO region of the mid-IR colour-colour diagram of Koenig et al. (2012) and also it is not located in the region populated by galaxies in Wright et al. (2010). Using the WGS method, from Figure 5 the WISE sources coincident with VVV-J114556.04-635628.0, VVV-J121313.50-613155.1 and VVV-J131828.35-635442.3 satisfy only one of the criteria to be a blazar candidate. The sources VVV-J125631.40-611626.0 and VVV-J134416.68-632638.7 do not satisfy any of the criteria. Based on this, none of the objects present sufficient evidence to be blazar candidates.

On the other hand, we analysed the variability using two different statistical methods: the fractional variability amplitude $\sigma_{rms}$ (Eq. 1) and the slope of the light curve (see Sect. 3). The values of $\sigma_{rms}$ are in the range 12.6 $-$ 33.8, which are in accordance to those found in type-1 AGNs (e.g., Nandra et al., 1997; Edelson et al., 2002). These values are, however, lower than those found in blazars (e.g., Sandrinelli et al., 2014; Pichel et al., 2020). Considering that in general type-1 AGNs present smaller variability amplitudes than blazars (e.g., Mao & Yi, 2021), these results suggest that these objects are type-1 AGN candidates; i.e. quasars and/or Seyfert 1 galaxies. Additionally, the slopes of the light curves, which are in the range (2.6$-$4.7) $\times 10^{-4}$ mag/day, are well within the limit value ($>$ 1 $\times 10^{-4}$ mag/day) reported by Cioni et al. (2013) for quasars.

The object VVV-J125631.40-611626.0 presents the highest $\sigma_{rms}$ value and following Lucas et al. (2017) the YSOs show the highest variations in the $K_{s}$ light curves. In this sense, typical YSOs exhibit variability that is stochastic, bursting or semi-periodic in short or long time scales. Indeed, these objects have been previously characterised in detail in the VVV survey (e.g., Lucas et al. 2017; Contreras Peña et al. 2017; Medina et al. 2021; Guo et al. 2020, 2022).

In summary, Table 5 shows the various observational evidences for each studied source in favour or against the criteria used by different authors to be considered AGN candidates. All these results suggest that four sources could be type-1 AGNs and the source VVV-J125631.40-611626.0 might be a YSO.

This kind of work allow us to find AGNs at lower Galactic latitudes where there is a lack of these studies. Future data coming from Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST Science Collaboration et al., 2009) and the eROSITA X-ray telescope (Brunner et al., 2022) would complement the VVV and VVVX databases. This synergy among data coming from different wavelengths allow us to deeply investigate these interesting galaxies. Further work involving spectroscopy is also crucial to shed light on their nature and active properties.

We would like to thank the anonymous referee for the useful comments and suggestions which has helped to improve this paper. This work was partially supported by Consejo de Investigaciones Científicas y Técnicas (CONICET) and Secretaría de Ciencia y Técnica de la Universidad Nacional de Córdoba (SecyT). The authors gratefully acknowledge data from the ESO Public Survey program IDs 179.B-2002 and 198.B-2004 taken with the VISTA telescope, and products from the Cambridge Astronomical Survey Unit (CASU). D.M. gratefully acknowledges support by the ANID BASAL projects ACE210002 and FB210003 and by Fondecyt Project No. 1220724.

The VVV NIRGC used in this article is available at https://catalogs.oac.uncor.edu/vvv_nirgc/. The VIVACE catalogue is available at https://people.ast.cam.ac.uk/~jls/data/vproject/vivace_catalogue.fits.