On the nature of the unusual transient AT 2018cow from HI observations of its host galaxy

TL;DR

This study investigates the host galaxy of the unusual transient AT2018cow using HI observations, finding no unusual atomic gas concentration near the transient, which suggests its progenitor may not be a massive star.

Contribution

Resolved atomic gas observations of AT2018cow's host galaxy reveal a normal gas distribution, challenging previous assumptions about its progenitor's nature.

Findings

No unusual atomic gas concentration near AT2018cow

Host galaxy has lower atomic gas mass than predicted by SFR

Environment suggests non-massive star progenitor

Abstract

Unusual stellar explosions represent an opportunity to learn about both stellar and galaxy evolution. Mapping the atomic gas in host galaxies of such transients can lead to an understanding of the conditions triggering them. We provide resolved atomic gas observations of the host galaxy, CGCG137-068, of the unusual, poorly-understood transient AT2018cow searching for clues to understand its nature. We test whether it is consistent with a recent inflow of atomic gas from the intergalactic medium, as suggested for host galaxies of gamma-ray bursts (GRBs) and some supernovae (SNe). We observed the HI hyperfine structure line of the AT2018cow host with the Giant Metrewave Radio Telescope. There is no unusual atomic gas concentration near the position of AT2018cow. The gas distribution is much more regular than those of GRB/SN hosts. The AT2018cow host has an atomic gas mass lower by 0.24 dex than predicted from its star formation rate (SFR) and is at the lower edge of the galaxy main sequence. In the continuum we detected the emission of AT2018cow and of a star-forming region in the north-eastern part of the bar (away from AT2018cow). This region hosts a third of the galaxy's SFR. The absence of atomic gas concentration close to AT2018cow, along with a normal SFR and regular HI velocity field, sets CGCG137-068 apart from GRB/SN hosts studied in HI. The environment of AT2018cow therefore suggests that its progenitor may not have been a massive star. Our findings are consistent with an origin of the transient that does not require a connection between its progenitor and gas concentration or inflow: an exploding low-mass star, a tidal disruption event, a merger of white dwarfs, or a merger between a neutron star and a giant star. We interpret the recently reported atomic gas ring in CGCG137-068 as a result of internal processes connected with gravitational resonances caused by the bar.

Tables

Table 1. Table 1: H i properties of CGCG 137-068.

Beam	$z_{\mathrm{HI}}$	$W_{50}$	$W_{10}$	$F_{\mathrm{int}}$	$\log \left(L_{\mathrm{HI}}^{\prime }\right)$	$\log \left(M_{HI}\right)$
($″$)		(km s-1)	(km s-1)	(Jy km s-1)	(${\text{K km s}}^{-1}{\text{ pc}}^2$)	($M_{\odot }$)
(1)	(2)	(3)	(4)	(5)	(6)	(7)
28	$0.013972\pm 0.000014$	$50\pm 30$	$127\pm 9$	$0.94\pm 0.09$	$10.738\pm 0.038$	$8.909\pm 0.038$
19	$0.013976\pm 0.000014$	$48\pm 22$	$127\pm 18$	$1.02\pm 0.12$	$10.772\pm 0.048$	$8.944\pm 0.048$
13	$0.013983\pm 0.000017$	$47\pm 22$	$126\pm 26$	$1.17\pm 0.15$	$10.832\pm 0.051$	$9.004\pm 0.051$
5	$0.013988\pm 0.000022$	$33\pm 14$	$126\pm 34$	$1.49\pm 0.23$	$10.939\pm 0.063$	$9.111\pm 0.063$

Table 2. Table 2: Properties of the continuum 1.397667 1.397667 1.397667  GHz sources within CGCG 137-068.

RA	Dec	$F_{1.4\mathrm{GHz}}$	SFRradio
(h m s)	(d m s)	(mJy)	($M_{\odot }$ yr-1)
16 16 00.209	+22 16 04.78	$1.239\pm 0.018$
16 16 00.729	+22 16 09.52	$0.101\pm 0.018$	$0.081\pm 0.011$

Full text

*Aims. *We provide resolved atomic gas observations of the host galaxy, CGCG137-068, of the unusual, poorly-understood transient AT 2018cow searching for clues to understand its nature. We test whether it is consistent with a recent inflow of atomic gas from the intergalactic medium, as suggested for host galaxies of gamma-ray bursts (GRBs) and some supernovae (SNe).

*Methods. *We observed the Hi hyperfine structure line of the AT 2018cow host with the Giant Metrewave Radio Telescope.

*Results. *There is no unusual atomic gas concentration near the position of AT 2018cow. The gas distribution is much more regular than those of GRB/SN hosts. The AT 2018cow host has an atomic gas mass lower by 0.24 dex than predicted from its star formation rate (SFR) and is at the lower edge of the galaxy main sequence. In the continuum we detected the emission of AT 2018cow and of a star-forming region in the north-eastern part of the bar (away from AT 2018cow). This region hosts a third of the galaxy’s SFR.

*Conclusions. *The absence of atomic gas concentration close to AT 2018cow, along with a normal SFR and regular Hi velocity field, sets CGCG137-068 apart from GRB/SN hosts studied in Hi. The environment of AT 2018cow therefore suggests that its progenitor may not have been a massive star. Our findings are consistent with an origin of the transient that does not require a connection between its progenitor and gas concentration or inflow: an exploding low-mass star, a tidal disruption event, a merger of white dwarfs, or a merger between a neutron star and a giant star. We interpret the recently reported atomic gas ring in CGCG 137-068 as a result of internal processes connected with gravitational resonances caused by the bar.

1 Introduction

Unusual, luminous, and rare stellar explosions provide an opportunity to learn about stellar evolution and also about galaxy evolution in a broader context. An example of the latter approach is the possibility to select galaxies that experience a recent inflow of gas from the intergalactic medium (IGM) using host galaxies of long gamma-ray bursts (GRBs) and some types of supernovae (SN). Atomic gas concentrations away from the galaxy centres towards GRB/SN positions suggest an external origin of the gas (Michałowski et al. 2015, 2016, 2018a), and a potential deficiency in molecular gas (Hatsukade et al. 2014; Stanway et al. 2015; Michałowski et al. 2016, 2018b). Studying gas inflows in such a direct way is important because they are required to fuel star formation in all galaxies, as implied from observations (Sancisi et al. 2008; Sánchez Almeida et al. 2014; Spring & Michałowski 2017; Elmegreen et al. 2018; Combes 2018) and simulations (Schaye et al. 2010; van de Voort et al. 2012; Narayanan et al. 2015). Recently Thöne et al. (2019) also suggested that in GRB hosts gas outflows are very common.

Observations of atomic gas in host galaxies of unusual and/or unclassified transients can therefore bring us closer to understanding the nature of these events. Similar atomic gas properties around the position of a transient to those of GRBs would suggest that the explosion mechanism is similar, i.e., an explosion of a massive star.

With this in mind, we report an analysis of gas properties in the host galaxy of the unusual and poorly-understood transient AT 2018cow111It was initially designated ATLAS 18qqn by the ATLAS discovery team.. The transient was discovered on 16 June 2018 by the Asteroid Terrestrial-impact Last Alert System (ATLAS; Tonry et al. 2018) surveying the entire visible sky every two nights (Smartt et al. 2018; Prentice et al. 2018). It was classified as a broad-lined type Ic (Izzo et al. 2018; Xu et al. 2018), type Ib (Benetti et al. 2018), or interacting type Ibn (Fox & Smith 2019) supernova and given the designation SN 2018cow. However, it is unclear whether this really was a supernova (see below). It has been detected at (sub)millimetre (de Ugarte Postigo et al. 2018; Smith et al. 2018; Ho et al. 2019) and radio (Dobie et al. 2018; Bright et al. 2018; Nayana & Chandra 2018; Margutti et al. 2019) wavelengths, including very long-baseline interferometry (VLBI) from which the most precise position has been derived: RA (J2000) = 16:16:00.2243, Dec. (J2000) = +22:16:04.893 with a $\sim 1\,$mas uncertainty (An 2018; Bietenholz et al. 2018; Horesh et al. 2018). AT 2018cow had several unusual characteristics: high peak luminosity, blue colour/high temperature even a month after the explosion, very fast initial flux rise (Prentice et al. 2018; Perley et al. 2019), high decline rate, no spectral features up to four days after the explosion, very broad short-lived absorption and emission spectral features (Perley et al. 2019), variability of the X-ray light curve (Rivera Sandoval et al. 2018), a month-long plateau at millimetre wavelengths (see Michałowski et al. 2018c for another example), as well as high radio flux (Ho et al. 2019).

It has been shown that it could not have been powered by radioactive decay (Prentice et al. 2018; Margutti et al. 2019; Perley et al. 2019). Several models have been proposed to explain the observed properties: a stellar collapse leading to the formation of a magnetar (Prentice et al. 2018; Margutti et al. 2019), a luminous blue variable exploding in a non-uniform circum-stellar medium (Rivera Sandoval et al. 2018), a SN from a low-mass hydrogen-rich star, a failed SN from a blue supergiant (Margutti et al. 2019), a tidal disruption event (TDE; Liu et al. 2018; Kuin et al. 2018; Perley et al. 2019), a jet driven by an accreting neutron star colliding with a giant star (Soker et al. 2019), or a merger of white dwarfs (Lyutikov & Toonen 2018). However, the constraints on the nature of this explosion set by the host galaxy properties have not been explored thoroughly.

AT 2018cow exploded within a spiral galaxy of type Sc (Willett et al. 2013), CGCG 137-068, at a redshift of $z=0.014$ (Perley et al. 2019). It has an inclination from the line of sight of $24.4^{\circ}$ (Makarov et al. 2014).222http://leda.univ-lyon1.fr/ledacat.cgi?CGCG%20137-068. It has a bar and weak spiral arms (Perley et al. 2019). Its stellar mass and SFR are 1.42^{+0.17}_{-0.29}\times 10^{9}\,\mbox{M_{\odot}} and 0.22^{+0.03}_{-0.04}\,\mbox{\mbox{M_{\odot}}\ yr{}^{-1}}, respectively (Perley et al. 2019). The galaxy was claimed to be asymmetric with more near-IR emission in the south-west, i.e., in the part of the galaxy where AT 2018cow exploded (Kuin et al. 2018), $\sim 1.7$ kpc from the galaxy centre (Kuin et al. 2018; Perley et al. 2019).

The objectives of this paper are: i) to provide a resolved measurement of the atomic gas properties of the host galaxy of AT 2018cow in order to learn about its nature, and ii) to test whether these properties are consistent with a recent inflow of atomic gas from the intergalactic medium.

We use a cosmological model with $H_{0}=70$ km s*-1* Mpc*-1*, $\Omega_{\Lambda}=0.7$, and $\Omega_{m}=0.3$, implying that AT 2018cow, at $z=0.014$, is at a luminosity distance of 60.6 Mpc and $1\arcsec$ corresponds to 286 pc at its redshift. We also assume the Chabrier (2003) initial mass function (IMF).

2 Data

On 8 and 9 February 2019 the field of AT 2018cow was observed for 14 hrs with the Giant Metrewave Radio Telescope (GMRT)333Project no. 35_021, PI: M. Michałowski. For calibration of the flux and the bandpass 3C286 was observed for 15 min at the start and the end of the run. For the phase calibration 1609+266 was observed every 40 min. The correlator was setup with 33 Mhz bandwidth and 512 channels centred around 1400 MHz.

After the submission of this paper, additional GMRT data were reported by Roychowdhury et al. (2019). Hence, for the Hi analysis we also included these archival data444Project no. DDTC022, PI: M. Arabsalmani. As part of that program 7 hrs of data were obtained on 27 August 2018. The channel width was half as wide as for our observations. The same calibrators were observed.

The data were reduced with a range of data reduction packages. We downloaded the FITS files with the raw data from the GMRT archive. These FITS files were then loaded into the Common Astronomy Software Applications (CASA) package (McMullin et al. 2007) with the importgmrt task without applying the online flags. Further data reduction was done with the meerkathi555https://github.com/ska-sa/meerkathi, a private repository for the time of development pipeline which is being developed for Hi data reduction of MeerKAT data. The pipeline is setup in a modular fashion using the platform-independent radio interferometry scripting framework stimela666https://github.com/SpheMakh/Stimela/wiki. In practice this means that the calibrator data are initially flagged with AOflagger (Offringa 2010) and calibrated and transferred to the target with CASA. For the data in this paper the phase calibrator was used as a bandpass calibrator, as the GMRT bandpass clearly fluctuated over the time of the observations. As the phase calibrator was bright (4.8 Jy), this leads to an improved bandpass calibration over the course of the observations.

After the initial calibration the target was split out of the measurement set, further flagged with AOflagger, imaged with WSclean (Offringa et al. 2014) in Stokes I, then the sources in the field were extracted and modeled with pyBDSF777https://github.com/lofar-astron/PyBDSF, after which this model was used in Cubical888https://github.com/ratt-ru/CubiCal/ (Kenyon et al. 2018) for the self-calibration. This step was repeated until a phase-only self-calibration no longer improved the extracted models.

After the calibration, the three separate days were mapped on to the same channel grid with the CASA task mstransform and the modelled continuum was subtracted from the data. Any residual continuum was subtracted with uvlin. At this stage the data were also doppler-corrected and projected onto a barycentric velocity frame.

The visibilities were weighted according to a Briggs weighting scheme with Robust = 0.0 and uvtapers of 4, 6, 8 and 20 k$\lambda$ were applied to attain cubes with varying spatial resolution. The cubes were inverted and cleaned with the CASA task TCLEAN. The cleaning was performed in an iterative process where we first clean the full cube to a 10$\sigma$ threshold then create a mask with SoFiA (Serra et al. 2015), and then clean within this mask to 0.5$\sigma$. This last step was done outside the pipeline as currently it can not deal with the frequency increments of opposite sign in the different datasets.

The final cubes have a resolution of $\textnormal{FWHM}=28\aas@@fstack{\prime\prime}9\times 26\aas@@fstack{\prime\prime}2$, $19\aas@@fstack{\prime\prime}2\times 18\aas@@fstack{\prime\prime}1$, $13\aas@@fstack{\prime\prime}12\times 12\aas@@fstack{\prime\prime}9$, and $5\aas@@fstack{\prime\prime}5\times 4\aas@@fstack{\prime\prime}6$ and a channel width of 65.1 kHz. The frequency axis was converted into a velocity axis using the relativistic definition which results in a channel width of 13.9 km s*-1* with an error of $\sim$ 0.01 km s*-1* on the outermost channels of the cube.

For our data from February 2019 (excluding those from August 2018, due to the variability of AT 2018cow) we also imaged together all channels of the entire 33 MHz bandwidth (before continuum subtraction) to produce a continuum image at an observed frequency of $1.397667$ GHz. The beam size is $2.0\arcsec\times 1.8\arcsec$ and the noise is $17.5\,\mu$Jy beam*-1*. The Hi line spans $\sim 0.6$ MHz, so it should not affect this continuum image based on the 33 MHz bandwidth. Indeed, when we exclude the channels with the line emission we obtain an almost identical map.

In order to correct the astrometry we identified 16 sources in the Faint Images of the Radio Sky at Twenty-Centimeters (FIRST) survey (Becker et al. 1995; White et al. 1997) that are point-like in our continuum map. On average these sources were found to be shifted on our map with respect to the FIRST position by $(+1.30\pm 0.14)\arcsec$ in right ascension and $(+0.51\pm 0.14)\arcsec$ in declination. We shifted our continuum and Hi maps by this offset. This has very little effect on Hi maps, as their beam sizes are much larger. This offset also implies that the positional uncertainty in our continuum map is $0.14\arcsec$ in both directions.

3 Results

The Hi fluxes at each frequency element were determined by aperture photometry with an aperture radius of $45^{\prime\prime}$. The spectra are shown in Fig. 1. The Hi emission maps derived from the collapsed cubes within the dotted lines in Fig. 1 are shown in Fig. 2. This range was selected to encompass the full velocity width of the line. It was also used to obtain integrated Hi emission ($F_{\rm int}$ in Jy km s*-1*) directly from the spectra. The line luminosity ($L^{\prime}_{\rm HI}$ in K km s*-1*pc2) was calculated using Eq. 3 in Solomon et al. (1997) and transformed to $M_{\rm H{\sc I}}$ using Eq. 2 in Devereux & Young (1990). The Hi zeroth and first moment maps (integrated emission and velocity field) are also shown on Fig. 2.

We detected and resolved the Hi emission of the host of AT 2018cow. The atomic gas disk is larger than the stellar disk with a centre (moment 0 ‘centre of mass’) offset from the optical centre by $\sim 1$–$2\arcsec$ ($\sim 0.3$–$0.6\,$kpc in projection), and $\sim 5$–$6\arcsec$ or $\sim 1.4$–$1.7$ kpc from the position of AT 2018cow. Using the formula of Ivison et al. (2007)111111$r=0.6\times\mbox{FWHM}_{\rm beam}/(\mbox{S}/\mbox{N})$, where $\mbox{FWHM}_{\rm beam}$ is the FWHM of the beam and $\mbox{S}/\mbox{N}$ is the signal-to-noise ratio. the positional uncertainty is $\sim 0.5$–$1.5$″, so the offset to the galaxy centre is at most $2\sigma$, but to the AT 2018cow position it is significant at $\sim 5\sigma$.

The Hi maps (Fig. 2) do not show strong evidence of recent gas inflows. The gas distribution is much more regular than those of the hosts of GRB 980425 (Arabsalmani et al. 2015), GRB 060505 (Michałowski et al. 2015), and SN 2009bb (Michałowski et al. 2018a), which exhibit strong gas concentrations close to the GRB/SN positions, away from the galaxy centres.

On the two zeroth moment maps with the highest resolutions we see the ring-like structure reported by Roychowdhury et al. (2019). In Sect. 4 we provide evidence that this structure is of internal origin.

On the outskirts of CGCG 137-068 there are gas plumes in both the zeroth moment and the collapsed maps (Fig. 2), but they are of low signal-to-noise, so they cannot be interpreted as real structures with confidence. Moreover, they can be spiral structures.

Moreover, from one resolution to another the ‘centre of mass’ of the zeroth moment maps moves only by $\sim 1\arcsec$ ($\sim 3\arcsec$ for the $28\arcsec$ map with the worst positional uncertainty). This suggests that the distribution is symmetric. The velocity fields (bottom row of Fig. 2) and the double-horn profiles of the Hi spectra (Fig. 1) are consistent with a rotating disk.

The SFR-$M_{\rm H{\sc I}}$ relation (Eq. 1 in Michałowski et al. 2015), predicts \log(\mbox{M_{\rm H{\sc I}}}/\mbox{M_{\odot}})=9.14^{+0.04}_{-0.07} for \mbox{SFR}=0.22\,\mbox{\mbox{M_{\odot}}\ yr{}^{-1}} of CGCG 137-068 (the errors include both the uncertainty in the SFR and in the parameters of the relation). This is $0.24$ dex, i.e. $\sim 3\sigma$, higher than the measured value (Table 1), which is within the scatter of this relation (0.38 dex at $1\sigma$). Hence CGCG 137-068 has a normal atomic gas content for its SFR and is located close to most gas-poor galaxies within this relation. The relation has been established using over 1500 galaxies, also covering the SFR range relevant here.

We present the continuum map at an observed frequency of $1.397667$ GHz in Fig. 3. We detected two point sources within CGCG 137-068: AT 2018cow in the south-west and a second source at the north-eastern part of the bar, just outside the bulge. The positions, fluxes, and SFRs using the conversion of Bell (2003) are listed in Table 2. For the first source we show the SFR as an upper limit, as it is dominated by AT 2018cow. This is motivated by a small offset ($0.24\arcsec$) of this source to the VLBI position of AT 2018cow and the fact that a variable $1.4\,$GHz flux at this level has been reported by Margutti et al. (2019).

The SFR of the second source (0.081\,\mbox{\mbox{M_{\odot}}\ yr{}^{-1}}) is $37^{+7}_{-8}$% of the total SFR of the galaxy as measured from the spectral energy distribution modelling (0.22\,\mbox{\mbox{M_{\odot}}\ yr{}^{-1}}; Perley et al. 2019). This source is coincident with one of the peaks of the Hi maps. The analogous continuum source is not present in the other half of the bar on the other side of the bulge. Star formation along the bar and differences between the two halves of the bar are common among local spirals, but regions inside bars do not dominate the total SFR (Regan et al. 1996; Sheth et al. 2000, 2002; Koda & Sofue 2006; Momose et al. 2010; Hirota et al. 2014; Yajima et al. 2019). Moreover, barred spirals always exhibit significant star formation in the galaxy centre, which is not evident for CGCG 137-068.

The SFR and stellar mass of CGCG 137-068 (Perley et al. 2019) imply a specific SFR (\mbox{sSFR}\equiv\mbox{SFR}/\mbox{M_{*}}) of $\sim 0.15\,\mbox{Gyr}^{-1}$. At this stellar mass, the sSFR of a main-sequence galaxy is $\sim 0.2\,\mbox{Gyr}^{-1}$ (Speagle et al. 2014). Hence, CGCG 137-068 is a main-sequence galaxy at the bottom of the scatter of this relation with no enhancement or strong suppression of star formation.

The atomic gas and star formation properties of CGCG 137-068 are summarised in Fig. 4 and compared with GRB/SN hosts with Hi measurements (Michałowski et al. 2015, 2018a). For each galaxy we also show the predicted gas depletion time from the $M_{\rm H{\sc I}}$-SFR relation (Michałowski et al. 2015). GRB/SN hosts occupy two regions of this diagram: either on/below the main-sequence and abundant with atomic gas (high gas depletion timescale well above the prediction), or above the main-sequence with low gas depletion timescale, due to elevated SFR. In contrast, CGCG 137-068 is below the main sequence, but it has lower gas content than predicted from the $M_{\rm H{\sc I}}$-SFR relation. In particular it is different than the hosts of GRB 060505 and 111005A, which have 0.3–0.5 dex more atomic gas than predicted from their SFR. In terms of the $M_{\rm H{\sc I}}$/SFR ratio, the AT 2018cow host is most similar to the GRB 980425 host, which is, however, at the upper boundary of the main sequence and exhibits a strong gas concentration close to the GRB position, unlike the AT 2018cow host.

4 Discussion

The atomic gas distribution of CGCG 137-068 does not show strong unusual features (especially not at the AT 2018cow position), in contrast to the off-centre gas concentrations and irregular velocity fields of the host galaxies of GRBs or relativistic SNe (Arabsalmani et al. 2015; Michałowski et al. 2014, 2015, 2016, 2018a). Moreover, there is no enhancement of the SFR, which could be a signature of a gas inflow. The environment of AT 2018cow therefore suggests that its progenitor may not have been a massive star (Prentice et al. 2018; Margutti et al. 2019; Rivera Sandoval et al. 2018; Fox & Smith 2019). However, the GRB/SN host sample with atomic gas measurements is small, so we cannot rule out this hypothesis.

The asymmetry in the distribution of atomic gas in the case of the host of the relativistic SN 2009bb may be a result of interaction (Michałowski et al. 2018a), as has also been observed for other galaxies (Sancisi et al. 2008; Rasmussen et al. 2006). Hence to investigate this further we analysed the large-scale environment of CGCG 137-068 using the NASA/IPAC Extragalactic Database (NED) of CGCG 137-068. It seems fairly isolated, with no other galaxies within 500 kpc projected distance and 1000 km s*-1* velocity. The nearest galaxy is UGC 10322, more than 500 kpc away in projected distance. Hence, in the current catalogues there is no galaxy which is close enough to significantly influence the properties of CGCG 137-068. We found that CGCG 137-068 is $\sim 700\,$kpc to the west of a possible galaxy group extending several hundred kpc across and containing six galaxies. Similarly to the host of SN 2009bb (Michałowski et al. 2018a), this could mean that there is a supply of intergalactic gas available for inflow onto CGCG 137-068, but we did not find any evidence of such process.

On the other hand, all other proposed explosion mechanisms of AT 2018cow, apart from massive-star core-collapse, should not result in a connection between its progenitor and gas concentration or inflow: an exploding low-mass hydrogen-rich star (Margutti et al. 2019), a TDE (Liu et al. 2018; Kuin et al. 2018; Perley et al. 2019), and a merger of white dwarfs or a neutron star and a giant star (Lyutikov & Toonen 2018; Soker et al. 2019). Hence, the normal atomic gas distribution of CGCG 137-068 is consistent with these mechanisms.

After the submission of this paper, the results of Roychowdhury et al. (2019) on the atomic gas distribution in the host galaxy of AT 2018cow were published. They found a ring of gas, also visible in our combined dataset (Fig. 2).

As claimed by Roychowdhury et al. (2019), such a gas ring could be the result of a minor merger. However, most rings in galaxies has been shown to be the result of resonances caused by the presence of a bar (gravitational torques; see the review by Buta & Combes 1996) and other internal mechanisms, like viscous torques (Icke 1979; Buta 1986; Lesch et al. 1990; Combes & Gerin 1985; Armillotta et al. 2019). Similarly, Díaz-García et al. (2019) found an increasing fraction of ringed galaxies with increasing bar Fourier density amplitude (also for galaxies with stellar masses similar to that of CGCG 137-068).

Indeed CGCG 137-068 exhibits a strong bar, which can be the cause of the appearance of the gas ring. Moreover, the Hi velocity fields presented here and by Roychowdhury et al. (2019) do follow a rotation pattern, and do not show any sign of disturbances, given the errors in the measurements. Finally, almost all spiral galaxies (including those with similar masses to CGCG 137-068) exhibit central depressions of atomic gas (likely due to conversion to the molecular phase) or enhancement at the location of the spiral arms (Leroy et al. 2008; Bigiel & Blitz 2012; Martinsson et al. 2016). This feature, combined with low sensitivity (as in the highest-resolution map of Roychowdhury et al. 2019) would give rise to a ring-like structure in the data, which would have a purely internal origin. Hence, the presence of the gas ring in CGCG 137-068 without any sign of disturbance is not strong evidence of a recent merger.

5 Conclusions

We observed the Hi atomic hydrogen line emission of the AT 2018cow host galaxy with the Giant Metrewave Radio Telescope. There is no unusual atomic gas concentration near the position of AT 2018cow. The gas distribution is much more regular than those of the hosts of GRBs and SNe. The AT 2018cow host has an atomic gas mass lower by 0.24 dex than the prediction from its SFR and is at the lower edge of the galaxy main sequence. In the continuum we detected the emission of AT 2018cow and of a star-forming region in the north-eastern part of the bar (away from AT 2018cow). This region hosts a third of the galaxy star formation rate (SFR).

The absence of atomic gas concentration close to AT 2018cow, along with a normal SFR and regular Hi velocity field sets CGCG137-068 apart from GRB/SN hosts studied in Hi. The environment of AT 2018cow therefore suggests that its progenitor may not have been a massive star. Our findings are consistent with an origin of the transient that does not require a connection between its progenitor and gas concentration or inflow: an exploding low-mass star, a tidal disruption event, or a merger of white dwarfs or of a neutron star and a giant star. We interpret the recently reported atomic gas ring in CGCG 137-068 as a result of internal processes connected with gravitational resonances caused by the bar.

Acknowledgements.

We thank Joanna Baradziej for help in improving this paper. M.J.M. acknowledges the support of the National Science Centre, Poland, through the POLONEZ grant 2015/19/P/ST9/04010 and SONATA BIS grant 2018/30/E/ST9/00208; this project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No. 665778. P.K. is supported by the BMBF project 05A17PC2 for D-MeerKAT J.H. was supported by a VILLUM FONDEN Investigator grant (project number 16599). D.A.K. acknowledges support from the Juan de la Cierva Incorporación fellowship IJCI-2015-26153. A.d.U.P. and C.C.T. acknowledge support from Ramón y Cajal fellowships (RyC-2012-09975 and RyC-2012-09984). D.A.K., A.d.U.P., and C.C.T. acknowledge support from the Spanish research project AYA2017-89384-P. L.K.H. acknowledges funding from the INAF PRIN-SKA program 1.05.01.88.04. The Cosmic Dawn Center is funded by the DNRF. R.L. acknowledges support from the grant EMR/2016/007127 from the Dept. of Science and Technology, India. This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (grant agreement no. 679627; project name FORNAX; PI Paolo Serra). We thank the staff of the GMRT who have made these observations possible. GMRT is run by the National Centre for Radio Astrophysics of the Tata Institute of Fundamental Research. Based on observations made with the Gran Telescopio Canarias (GTC), in the Roque de los Muchachos Observatory. We acknowledge the usage of the HyperLeda database (http://leda.univ-lyon1.fr). This research has made use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration; SAOImage DS9, developed by Smithsonian Astrophysical Observatory (Joye & Mandel 2003); and NASA’s Astrophysics Data System Bibliographic Services.