Unravelling the Post-Collision Properties of the Cartwheel Galaxy: A MUSE Exploration of its Bar and Inner Region

Stellar bars are extended linear structures frequently found in nearby disk galaxies, where they reshape bulges, regulate star formation, and drive the host galaxies’ transformation (Kormendy & Kennicutt, 2004; Sakamoto et al., 1999; Ellison et al., 2011; Kim et al., 2016; Kruk et al., 2018; Donohoe-Keyes et al., 2019). Strong bars are found in more than 60% of the disk galaxies that dominate star formation in the local universe (Eskridge et al., 2000; Menéndez-Delmestre et al., 2007; Marinova & Jogee, 2007). The stellar bar is a long-lived and robust structure; however, it is unknown when the first bar formed in disk galaxies. Many simulations show that bars are strong structures and difficult to dissolve once formed (Athanassoula et al., 2005; Kraljic et al., 2012). So, in the absence of legitimate pieces of evidence and clues, it remains a more complex problem to determine whether the same bar still exists in present-day galaxies or if it has been dissolved and rejuvenated (Bournaud & Combes, 2002; Barway & Saha, 2020).

The stellar bars are found across all environments and have a higher fraction in group/cluster environments, particularly for S0 galaxies (Barway et al., 2011). The evolution of bar in such environments is poorly studied, particularly in the case when a galaxy undergoes an interaction and/or collision. Recently, Barway et al. (2020) discovered a bar in the Cartwheel using Near-infrared (NIR) Ks-band images, making it the first collisionally formed ring galaxy to have a bar detected. The Cartwheel is the archetype of collisional ring galaxies and is formed by a compact galaxy falling head-on through a massive gas-rich disk galaxy close to its centre and almost perpendicular to it (Lynds & Toomre, 1976), and is distinguished by an outer ring that contains a chain of star-forming knots (Marston & Appleton, 1995; Appleton & Marston, 1997; Higdon et al., 2015).

The Cartwheel has a distinct inner ring that, unlike the outer ring, is gas-poor, with little presence of H I (Higdon, 1996), H_α (Amram et al., 1998), and star formation (Charmandaris & Mirabel, 1997). The inner ring’s low star formation rate has cast doubt on its interpretation as a post-collisional structure, and it could equally well be of pre-collisional origin. Several studies have attempted to simulate Cartwheel’s observed structure through a collision of two galaxies. Most of these have explored the nature of the outer ring and spokes in detail (Marcum et al., 1992; Appleton & Marston, 1997; Renaud et al., 2018). There are very few studies through numerical simulations to understand the evolution of a pre-collisional bar or the inner disc’s orientation after the collision. A theoretical study by Athanassoula et al. (1997) using N-body simulation to create the rings in galactic discs by infalling small companions attempts to address the questions mentioned earlier. These authors found that the bar survives the interaction and gradually grows fatter, eventually forming an oval structure. A small ring surrounds the bar and expands and detaches from it, forming an arm on one side.

The studies using Multi-Unit Spectroscopic Explorer (MUSE; Bacon et al. (2010)) observations at the Very Large Telescope (VLT) also focus primarily on the outer ring of the Cartwheel galaxy (Zaragoza-Cardiel et al., 2022; Mayya et al., 2023). In this paper, we present a detailed analysis of the inner region of the Cartwheel galaxy using MUSE data, which covers the entire optical extent of the galaxy at the seeing-limited angular resolution of $\sim$0$\aas@@fstack{\prime\prime}$6. Utilising the MUSE IFU data, we studied the kinematics and star formation history of the inner part to explore the existence of a pre-discovered bar in the Cartwheel. We also probed the nature of ionised gas and the gas-phase metallicity of the same part using identified emission lines in the MUSE data. Our primary aim is to provide more evidence for the presence of a bar in this collisional ring galaxy and also to understand the overall properties of its inner disk.

This article is organised as follows. Section §2 presents the photometric properties of the inner region. Section §3 describes the MUSE data set, the detailed analysis is presented in Section §4, and the results are presented in Section §5. We present our conclusions in Section §6.

In general, stellar bars are predominantly composed of older stellar populations. As a result, NIR observations, known to reveal structures that are not always visible in optical images, are frequently used to identify bars in galaxies. The NIR emission from older stars is less affected by the dust extinction and the presence of younger stars (Cardelli et al., 1989; Riad et al., 2010). Utilising this property of the NIR emission, Barway et al. (2020) reported the discovery of a stellar bar, a pseudo-bulge, and an unresolved point source in the archetype collisional ring galaxy Cartwheel through a careful morphological analysis of an excellent quality near-infrared $K_{s}$-band image from the European Southern Observatory (ESO) archive taken at the VLT-U1 telescope in the year 2000 using the Infrared Spectrometer And Array Camera (ISAAC) instrument. Due to a combination of its red colour and the presence of dusty features in the central region of the Cartwheel galaxy, the newly discovered bar is not recognisable in optical images even with a spatial resolution of the Hubble Space Telescope (HST).

Cartwheel’s stellar bar is oval-shaped, with a semi-major axis length (bar length) of 2.09 kpc and a nearly flat light distribution. The bulge is nearly circular, with an effective radius of 1.05 kpc and a Sersic index of 0.99, which are typical of late-type galaxies’ pseudo-bulges. Another important bar characteristic is the ’strength’ of a bar, which is 0.36 for the Cartwheel bar as determined by Barway et al. (2020), a value typically found in late-type spiral galaxies with a strong bar (Aguerri et al., 2009). To summarise, Cartwheel’s stellar bar is a strong and robust structure that survived the collision and was unaffected by the impact. Collisional ring galaxies offer the opportunity to study a new, less explored aspect of bar evolution, and kinematic details using MUSE will be critical to investigate.

We have used Integral Field Spectrograph (IFS) data of the Cartwheel observed with the MUSE instrument. MUSE is an optical IFS instrument installed on ESO’s Very Large Telescope (Bacon et al., 2010). The instrument operates within the wavelength range of 4750 Å to 9350 Å with 1^′ $\times$ 1^′ almost square field of view. The detector has the capability of 0$\aas@@fstack{\prime\prime}$2 spatial sampling and an average 1.25 Å  spectral sampling. The galaxy Cartwheel was observed between UT 24-08-2014 to 25-08-2014 as part of MUSE’s science verification program (ID- 60.A-9333). We have obtained the science-ready MUSE deep data cube of the galaxy from the ESO archive science portal¹¹1http://archive.eso.org/scienceportal/home (archive ID - ADP.2017-03-23T15:47:52.027). The deep cube has a sky coverage of 2.2 arcmin² with a total integration time of 8400 sec. During the observation, the effective seeing was $\sim$0$\aas@@fstack{\prime\prime}$63. In Figure 1, we have shown the 2d intensity map of the galaxy, constructed from the obtained deep cube, for the wavelength range 4800 - 5800 Å. The 2.2 square arcmin MUSE deep cube covers the entire galaxy, including the bright and distinct outer ring. In this work, we have studied only the inner $\sim$20${}^{\prime\prime}\times$20^′′ part of the galaxy as shown on the right panel of Figure 1. We produced a smaller cutout of dimension 100$\times$100 spaxels for the inner part from the larger deep cube and used that for further analysis. At the distance of Cartwheel (which is 133 Mpc as per the NASA/IPAC Extragalactic Database²²2https://ned.ipac.caltech.edu/), each MUSE spaxel corresponds to a length scale of $\sim$130 pc in the galaxy frame. The MUSE intensity map of the galaxy’s inner part clearly shows a bright central component surrounded by another ring structure. Our cutout encloses the inner ring including some parts of the spokes.

We have used the Galaxy IFU Spectroscopy Tool (GIST) pipeline (version - V3.0.2) developed by Bittner et al. (2019) for analysing the MUSE data cube. The pipeline is specifically designed to analyse reduced IFS data and has been successfully tested on several IFUs such as MUSE, PPAK (Calar Alto Legacy Integral Field spectroscopy Area - CALIFA), and Spectrograph for INtegral Field Observations in the Near Infrared (SINFONI) (Bittner et al., 2021; Yadav et al., 2021; Bittner et al., 2020; Gadotti et al., 2020; Neumann et al., 2020; Eisenhauer et al., 2003). We have used the Penalised Pixel-Fitting module (pPXF) (Cappellari & Emsellem, 2004; Cappellari, 2017) of the pipeline to study the stellar kinematics of the galaxy. To apply the pPXF routine, we spatially binned the data using the Voronoi binning algorithm (Cappellari & Copin, 2003). We fixed the targeted S/N at 30 for the binning, allowing us to retain a good spatial resolution of the binned data. The S/N of each spaxel is calculated within the wavelength range 4800 - 5800 Å while performing the Voronoi binning. The spaxels with S/N greater than 30 remain unbinned, while those with S/N less than 3 are excluded and not considered in the binning. The integrated spectrum of each binned unit is then fitted with the MILES single stellar population (SSP) model spectra (Vazdekis et al., 2010) combining with a line-of-sight velocity distribution (LOSVD) of the stellar populations. The LOSVD function is composed of four velocity parameters, i.e., line-of-sight velocity (V), velocity dispersion ($\sigma$), and h₃ and h₄ Gauss-Hermite moments. The fitting is performed for the continuum of the spectra between the wavelength range 4800 - 5800 Å. We adopted this wavelength range to avoid contamination from the redder wavelength part as suggested by Bittner et al. (2021, 2019). We have masked several prominent spectral lines falling within the considered wavelength range while performing the fit. We also considered an 8th-order multiplicative Legendre polynomial during the fitting to account for small deviations between the model and the observed continuum shape. The best fit is obtained with a least-square minimisation technique. Following the fitting using pPXF, we obtained the values of the four parameters (V, $\sigma$, h₃, h₄) for each bin. With the best-fit values, we constructed the spatial map of V, $\sigma$, V/$\sigma$, and h₃ for the inner part of the galaxy as shown in Figure 2. As we have studied only the inner part of the galaxy (Figure 1) in this work, hence by noting ’galaxy’ we mean ’the inner region’ in the rest of our text.

In the LOS velocity map (Figure 2, top-left), we noticed that the north-western region of the galaxy has a blue-shifted velocity, whereas the south-eastern part is red-shifted. The velocity shown in the figure is measured in the rest frame of the galaxy. In the top-right panel of Figure 2, we have displayed the velocity dispersion map of the galaxy. The dispersion values mostly range between 40 - 80 km/s. In the southeastern part, in between the ring and the central disk, we noticed a region with high dispersion velocity (red bins). The average velocity dispersion in the eastern half is noticed to be relatively higher compared to the western half of the galaxy. Using these V and $\sigma$ maps, we have produced the ratio of LOS velocity and velocity dispersion (i.e., V/$\sigma$) map of the galaxy (bottom-left). The values of V/$\sigma$ mostly lie in the range $\sim$ $-2$ to $+2$, whereas only for the bar region it ranges between $\sim$ $-1$ and $+1$. In the bottom-right panel of Figure 2, we have shown the h₃ velocity moment of the galaxy. We found the values of h₃ mostly have a range between $-$0.15 and 0.10 with a slightly lopsided distribution in the central region with respect to the zero velocity line seen in the LOS velocity map. For each kinematic parameter, here we note the typical formal error that arises from the pPXF fitting. The error in the estimated LOS velocity ranges between $\sim$ 5 - 12 km/s, whereas the mean error in the velocity dispersion is around 12 km/s. The error in h₃ ranges between $\sim$ 0.05 - 0.18 with a median value $\sim$0.12. For spaxels belonging to the bar, the error in h₃ is relatively smaller with a range of $\sim$ 0.05 - 0.11 and median value $\sim$ 0.08. We note here that the formal errors in the derived V, $\sigma$, and $h_{3}$ values in Cartwheel are slightly higher compared to the nearby systems which have a better sampling on the physical scales with the MUSE.

In this work, we have used the MUSE IFU data to study stellar kinematics, SFH, and the nature of ionised gas in the inner region of the Cartwheel, which is an ideal laboratory to test the pre- and post-collisional structure. Using the GIST pipeline, we have produced stellar LOS velocity, velocity dispersion, h₃ velocity moment, mass-weighted age, and emission line maps for the inner region of the Cartwheel. The main results of the study are summarised below -

• •

  The intensity, LOS velocity, and velocity dispersion profiles of the Cartwheel individually show characteristic signatures of a bar as reported in the simulation by Bureau & Athanassoula (2005).

• •

  We found a trend of weak correlation between V/$\sigma$ and h₃ for the bar region, which further supports the presence of a pre-discovered bar in the galaxy. The kinematic signature of the bar signifies that the past encounter of the Cartwheel, which resulted in the formation of a ring structure in the system, has not affected the stellar bar to the extent of destroying it.

• •

  For the inner part of the Cartwheel outside the bar region, we noticed no specific trend of anti-correlation between V/$\sigma$ and h₃. It indicates that the collision has affected the stable ordered motion of stars in the inner disk of the galaxy.

• •

  We produced a BPT map and showed that the ionisation in the galaxy’s central region is possibly due to star formation, and the presence of an AGN is unlikely. However, given the dust in the inner region, the higher resolution JWST NIR observations carry a better potential to probe the nature of the central source which we plan to do in the future.

• •

  The stellar populations within the bar are found to be older (with mass-weighted age constrained between $\sim$ 9 and 13 Gyr). Such older populations in the bar support its pre-collisional existence, which signifies that the bar was already in form when the Cartwheel underwent the collision.

In this work, we have addressed whether the bar in Cartwheel was pre-existing or formed after the collision by studying the stellar populations in the bar region, using integral field spectroscopy data from the MUSE. Our finding suggests that the central structures existed before the collision, putting impetus to include inner regions such as bars in future numerical and theoretical studies of collisional ring galaxies. Given the amount of dusty structure, the JWST in the near-IR wavelengths would be ideal for studying the inner region. Thus, collisional ring galaxies provide opportunities to learn a new, less explored aspect of the evolution of the inner region, both observationally and theoretically, which usually survived the collision and remained unaffected by the impact.