A superluminous supernova in high surface density molecular gas within the bar of a metal-rich galaxy

TL;DR

This study uses ALMA observations to analyze the molecular gas distribution in a metal-rich galaxy hosting a superluminous supernova, revealing high-density gas regions conducive to massive star formation.

Contribution

First detailed ALMA study of molecular gas in a superluminous supernova host galaxy, linking gas dynamics to supernova progenitor formation.

Findings

Molecular gas confined within the galaxy's bar structure.

Gas flows along lanes likely lead to massive cloud and star formation.

High surface density of molecular gas near the supernova location.

Abstract

We report the Atacama Large Millimeter/submillimeter Array (ALMA) observations of the metal rich host galaxy of superluminous supernova (SLSN) PTF10tpz, a barred spiral galaxy at z=0.03994. We find the CO(1-0) emission to be confined within the bar of the galaxy. The distribution and kinematics of molecular gas in the host galaxy resemble gas flows along two lanes running from the tips of the bar towards the galaxy center. These gas lanes end in a gaseous structure in the inner region of the galaxy, likely associated with an inner Lindblad resonance. The interaction between the large-scale gas flows in the bar and the gas in the inner region plausibly leads to the formation of massive molecular clouds and consequently massive clusters. This in turn can result in formation of massive stars, and thus the likely progenitor of the SLSN in a young, massive cluster. This picture is consistent with SLSN PTF10tpz being located near the inner structure. We find the molecular gas in the vicinity of the SLSN to have high surface densities, comparable with those in interacting galaxies or starburst regions in nearby galaxies. This lends support to high densities being favorable conditions for formation of SLSNe progenitors.

Tables

Table 1. Table 1: Global properties of the host galaxy of SLSN PTF10tpz.

RA	Dec	redshift	12+${\log }_{10}\left[O/H\right]$	$M_{\ast }$	${\left(S\Delta v\right)}_{\mathrm{CO}\left(1-0\right)}$	$W_{20,\mathrm{CO}\left(1-0\right)}$	$M_{\mathrm{mol}}$
(J2000)	(J2000)			($M_{\odot }$)	($\mathrm{Jy}{\mathrm{kms}}^{-1}$)	(km s-1)	($M_{\odot }$)
21:58:31.77	-15:33:02.61	0.03976$\pm$0.00007	9.22	${10}^{10.85}$	($18.39\pm 0.32$)	470	${10}^{9.8}$

Full text

ISM: molecules, ISM: kinematics and dynamics, (stars:) supernovae: individual (SLSN PTF10tpz), galaxies: ISM, submillimeter: galaxies

,

1 Introduction

Superluminous Supernovae (SLSNe) are rare and extremely bright transients (Smith et al., 2007; Quimby et al., 2013). Their luminous emission remains brighter than the peak magnitude of most supernovae (SNe) for hundreds of days. Their peak visual absolute magnitudes are 10-100 times brighter than those of typical core-collapse supernovae (e.g., Quimby et al., 2011), making it possible to detect them at $z>2$ (Cooke et al., 2012; Curtin et al., 2019). Up to date, the underlying explosion mechanism behind these recently identified energetic events remains speculative. However, their very large energy releases suggest that they arise from the stars at the very high mass end of the stellar initial mass function (IMF) with $\rm M\gtrsim 100\,\,M_{\odot}$ (Gal-Yam & Leonard, 2009). Studying the close environments of SLSNe can provide valuable insights into understanding the physical conditions in which their progenitors form, and thus put constraints on progenitor models for their formation.

Similarly to the conventional classification of SNe according to their spectroscopic properties, SLSNe are grouped into two main subclasses: hydrogen-poor events (SLSN-I), and events with spectroscopic signatures of hydrogen (SLSN-II; see Gal-Yam, 2012). Both groups appear to trace the ultraviolet light of their host galaxies which supports the association between their progenitors and massive stars (Neill et al., 2011; Lunnan et al., 2015). Their hosts are found to be mainly irregular, compact, dwarf galaxies, with high star formation rate (SFR) surface densities, and high specific star formation rates (sSFR, Lunnan et al., 2013, 2014a, 2015; Perley et al., 2016; Angus et al., 2016). Although, the host galaxies of type II SLSNe seem to cover a larger range in galaxy masses and metallicities compared to type I SLSN hosts (Neill et al., 2011; Perley et al., 2016; Angus et al., 2016).

The typical low metallicity of SLSN hosts could imply a strong metallicity bias for SLSN progenitors (Lunnan et al., 2014b; Perley et al., 2016; Chen et al., 2017), favouring single-star progenitor models with low mass loss due to stellar winds (Langer et al., 2007). However the detection of a fraction of SLSNe (of both types) in quite large, massive and metal-rich host galaxies raises questions about whether low metallicity is indeed necessary for the formation of a SLSN progenitors (see Lunnan et al., 2015; Perley et al., 2016; Nicholl et al., 2017). In some of the metal-rich cases, the location of the SLSN is coincident with the central regions – the most metal-rich regions of the host galaxy (Gal-Yam, 2012). A more likely, but less addressed, factor is formation of SLSN progenitor stars in dense environments. Massive stars are more likely to be found in massive star clusters which form in dense giant molecular clouds (GMCs). This picture is consistent with the occurrence of SLSNe in the central regions of metal-rich galaxies as gas flows, driven by galaxy interactions or bars, can create dense concentrations of molecular gas in the central regions of galaxies (Sheth et al., 2005; Gallagher et al., 2018). It is also compatible with progenitor models in which massive stars in dense and young star cluster merge into a star with sufficient mass for forming a SLSN (Portegies Zwart & van den Heuvel, 2007; Pan et al., 2012; van den Heuvel & Portegies Zwart, 2013).

SLSN PTF10tpz (type II SLSN) at $z=0.03994$ is one of the SLSNe which occurred close to the nucleus of a massive and metal-rich galaxy. In this letter we study the properties of molecular gas, the fuel for star formation, in the close environment of this SLSN, and investigate whether its progenitor is formed in a dense environment. We present deep carbon monoxide (CO) emission line observations of the host galaxy of SLSN PTF10tpz obtained by ALMA. The high spatial and spectral resolution of our ALMA observations and the fairly close distance of the host galaxy (183 Mpc using a flat $\Lambda$CDM with cosmological parameters from Planck Collaboration et al., 2018) allow us to model the dynamics of gas in the galaxy and to investigate the conditions in which the SLSN progenitor formed. This is the first study of molecular gas in a SLSN host galaxy.

2 Observations and data analysis

We used the Band-3 receivers of the ALMA 12-m Array in C43-6 configuration to map the CO(1-0) emission of the host galaxy on January 04, 2018 (project code: 2017.1.01568.S; PI: Arabsalmani). The observations used a 2 GHz band, centred on 110.883 GHz, covering the redshifted CO(1-0) emission line. This band was sub-divided into 1920 channels, yielding a velocity resolution of $\sim 2.6$ km s*-1*, and a total velocity coverage of about 5400 km s*-1*. The initial calibration of the data is done by the ALMA support staff, using the ALMA data pipeline in the Common Astronomy Software Applications (casa) package (McMullin et al., 2007). We use the casa package tclean to produce a spectral cube with a velocity resolution of $\sim$ 16 km s*-1*. With Natural weighting, we yield a synthesised beam size of 0.36*′′$\times$0.52′′*, and a root-mean-square noise of 0.3 mJy per channel in the cube.

We create a secondary data cube from our spectral data cube by applying Hanning smoothing across blocks of three consecutive velocity channels (using casa-specsmooth task), followed by spatial smoothing by convolving with a Gaussian kernel of full width at half maximum equal to 6 pixels (using casa-imsmooth task). The smoothing ensures that any localized noise peaks are ignored and only emission correlated spatially and in velocity is chosen. We then mask out pixels in the original spectral data cube which lie below a threshold flux in the secondary data cube. The threshold flux used to select pixels is approximately twice the noise in a line-free channel of the original spectral data cube. We obtain the total intensity and the intensity-weighted velocity field maps of the CO(1-0) emission line by applying the immoment tasks in casa package on the original spectral data cube with masked pixels.

We also use a R-band image of SLSN PTF10tpz host galaxy presented in Perley et al. (2016). This image was obtained on 26-November-2011 by the Low-Resolution Imaging Spectrometer (LRIS, Oke et al., 1995) on the Keck I 10 m telescope (Program PI: S. Kulkarni, proposal ID: C219LA). We fit elliptical isophotes to the optical image using the stsdas package in iraf, and find the observed axial ratios of outermost isophotes converge to be 0.42. This yields an inclination angle of 68o for the optical disk of the galaxy assuming a typical intrinsic axial ratio of 0.2 for the stellar disk.

We use archival data of the observations of PTF10tpz with the Near-Infrared Camera on Keck II (NIRC-2) in adaptive optics mode on UT 2010-10-29 between 06:01 and 06:56 UT (Program PI: A. Boden, proposal ID: C246N2L) in order to establish the location of PTF10tpz. Five dithered exposures were acquired in each of the $J$, $H$, and $K$’ filters. Each exposure consisted of 5$\times$30s coadds in $H$ and $K$’ and 8$\times$30s coadds in $J$. A basic reduction with NIR sky subtraction and co-addition of dithered frames is performed. We measure the astrometric offset of PTF10tpz relative to the galaxy nucleus to be 0.25*′′* $\pm$ 0.01*′′*  W and 0.16*′′* $\pm$ 0.01*′′*  N, relying on the telescope orientation keywords in the header.

3 Results and Discussion

The Keck/LRIS R-band image of the host galaxy of SLSN PTF10tpz is shown in the left and middle panels of Fig. 1. A bright stellar bar in the centre of the galaxy, along with a faint spiral-like structure suggest that the host is a SB type galaxy (see Table LABEL:the-only-table for the global properties of the host). The right panel of Fig. 1 shows the inner region of the galaxy at the time when the SLSN PTF10tpz was bright. We find the CO(1-0) emission to be confined within the stellar bar of the host galaxy, and extending over a projected physical size of $\sim$ 10 kpc. The emission has a $\rm W_{20}$ of 470 km s*-1*  and is centred at a redshift of $z=0.03976\pm 0.00007$. The maps of the total intensity and the intensity-weighted velocity field of the CO(1-0) emission line are shown in the left and middle panels of Fig. 2, respectively. The right panel of Fig. 2 shows the R-band image of the galaxy bar within the same frame as in the left and the middle panels (also, same frame as in the middle panel of Fig. 1).

From the total intensity map it is clear that the CO emission has a significant contribution from the inner regions. Specifically, the central 1 kpc of the galaxy appears to contain about $25\%$ of the total flux of the CO(1-0) emission. The intensity-weighted velocity field of the CO(1-0) emission line is reminiscent of a rotating structure with a distinct fast rotation in the central 1 kpc of the galaxy. Notably, the molecular gas in this inner region extends over $>500$ km s*-1* in velocity space.

We measure a velocity-integrated flux density of $18.39\pm 0.32$ Jy km s*-1*  over a total velocity range of 570 km s*-1*, for the CO(1-0) emission from the galaxy. This implies a brightness temperature luminosity (as defined in Obreschkow et al., 2009) of $L^{\rm T}_{\rm CO(1-0)}=(1.45\pm 0.03)\times 10^{9}\,\rm K\,km\,s^{-1}\,pc^{2}$. We assume a Galactic CO-to-molecular-gas conversion factor of $\alpha_{\rm CO}=4.36\,\rm M_{\odot}\,(\rm K\,km\,s^{-1}\,pc^{2})^{-1}$, based on the super-solar metallicity of the host (Perley et al., 2016) and obtain a molecular gas mass of $10^{9.8}\,\rm M_{\odot}$ for the host galaxy. With the stellar mass of $\rm 10^{10.85}\,M_{\odot}$ (Perley et al., 2016), this yields a molecular gas mass–to–stellar mass ratio of $\sim 0.1$ for the SLSN host, typical of nearby star forming galaxies with similar stellar masses (Bothwell et al., 2014). We note that the conversion factor in the central kpc of nearby galaxies is found to be a factor of 2 lower than the mean value (Sandstrom et al., 2013). Since a significant fraction of the CO emission is from the central few kpc of the host galaxy, its molecular gas mass could be lower by a factor of 2. However, for consistency and to compare our results with the literature we use the molecular gas mass estimated by assuming a Galactic conversion factor. This does not affect our conclusions.

Perley et al. (2016) performed spectroscopic observations for the host galaxy using a 1 ′′ slit passing through the galaxy centre along its major axis and measured the fluxes of the bright emission lines in the central regions of the galaxy (within the 1 ′′  slit). They performed a BPT diagram analysis using the log[Oiii/H$\beta$] vs. log([Nii]/H$\alpha$] diagram and found the emission line fluxes to be consistent with the presence of an active galactic nucleus (AGN) in the galaxy centre. They reported an average ${\rm H}\alpha$-based SFR of $\rm 6.0\,M_{\odot}\,yr^{-1}$ for the central regions of the host and stated this to be an upper limit due to a (subdominant) contribution from the AGN to the ${\rm H}\alpha$  flux. However, we note that the host galaxy on the log[Oiii/H$\beta$] vs. log([Nii]/H$\alpha$] plane lies in a region which is at the border between the AGN and star forming galaxy regimes, and hence it is hard to confirm or rule out the presence of an AGN based on this diagram. On the other hand, the fluxes of the S[ii] lines reported by Perley et al. (2016) place the host galaxy amongst the star-forming galaxies on the log[Oiii/H$\beta$] vs. log([Sii]/H$\alpha$] plane (see Kewley et al., 2006; Singh et al., 2013). We therefore use the $\rm 6.0\,M_{\odot}\,yr^{-1}$ as the reference value for the SFR in the central regions of the galaxy. By measuring the molecular gas mass in the same region we estimate an average molecular-gas-depletion time ($M_{\rm mol}/\rm SFR$) of 0.7 Gyr, similar to depletion times in nearby starburst galaxies (see e.g., Bigiel et al., 2008).

We obtain a mean molecular gas surface-density of $\Sigma_{\rm mol}\sim 160\,\rm M_{\odot}pc^{-2}$ for the galaxy, averaged over the area within which CO(1-0) emission is detected (as defined based on the total intensity map) and corrected for the inclination of the host. This is comparable with the mean molecular gas surface-densities of starburst galaxies in the local Universe (also determined by averaging within the radius of the central molecular disk as determined from their CO maps; see Kennicutt, 1998). In order to investigate the nature of the high mean surface density of the host, we measure the molecular gas surface-densities over individual beams, or equivalently over $\sim$ 350 pc scales. We find these to be as high as $\sim 1400\,\rm M_{\odot}pc^{-2}$ in the inner region of the galaxy. In particular, we measure a $\Sigma_{\rm mol}$ of $\sim 700\,\rm M_{\odot}pc^{-2}$ over a beam centred on the SLSN location. These values are larger or comparable to the molecular gas surface densities obtained for the central regions of strongly barred nearby galaxies over smaller scales (e.g., 120 pc scales, Sun et al., 2018). We use the molecular gas clumping factors from Leroy et al. (2013) to convert the surface densities averaged over the scales probed by a single beam to the surface densities on the scale of individual clouds ($\sim$ 50 pc). We predict the GMCs in the inner region of the host galaxy to have surface densities as high as $\sim 5000-10,000\,\rm M_{\odot}pc^{-2}$. These are comparable with the surface densities of GMCs in the Antennae galaxy or the nuclear starburst of NGC 253 (Leroy et al., 2015; Sun et al., 2018). The high molecular gas surface densities in the vicinity of the SLSN supports the hypotheses in which high densities are favorable conditions for formation of SLSNe progenitors (see also Arabsalmani et al., 2015; Roychowdhury et al., 2019; Arabsalmani et al., 2019).

We investigate the kinematics and structure of molecular gas using channel maps of the CO(1-0) emission at 16 km s*-1*  velocity resolution presented in Fig. 3. These maps cover the inner region of the bar. The two white ellipses, following the orientation and flattening of the outermost isophotes of the disk, are there to help follow the path of the emission through the velocity channels. The CO emission is mainly concentrated along the path traced by these two ellipses. The emission coincident with the outer ellipse is confined mainly to the top-left and bottom-right quadrants of the ellipse. This emission extends beyond the frame covered in the channel maps in several channels, forming the faint extended wings which can be seen in the left panel of Fig. 2. The emission coincident with the inner ellipse on the other hand is mainly restricted to the top-right and bottom-left quadrants of the ellipse, and is present in all channels with velocities between -250 km s*-1*  and 257 km s*-1*. This emission forms a steep profile in the Position-Velocity (P-V) diagram of the CO(1-0) emission (Fig. 4), making an X-shape pattern when combined with the emission from the outer structure (see Bureau & Freeman, 1999; Merrifield & Kuijken, 1999).

The spatial distribution and kinematics of the molecular gas shown in Figures 2 and 3 are consistent with the presence of two gas lanes, running from the tips of the bar towards the galaxy centre. The specific orbital structures imposed by the bar, and the dissipative nature of the gas lead to two symmetric well-defined gas lanes, often associated with shocks (see e.g., Maciejewski et al., 2002; Emsellem et al., 2015). Fig. 5 shows a simple sketch of such structures in the bar of the host galaxy of SLSN PTF10tpz. In this picture, the innermost parts of the two gas lanes encounter an inner Lindblad resonance associated with the pattern-speed of the bar, close to the location of the SLSN. The transition of gas from the gas lanes to the inner resonance results in the formation of high density gas with large velocity dispersions. This in turn leads to the formation of large GMCs at the intersection of the gas lanes and the inner resonance (similar to what is seen in NGC 1365, Elmegreen et al., 2009). The Keck/NIRC2 Kp-band image of the inner region of the host (the right panel of Fig. 1) suggests the presence of a stellar ring with similar dimensions as those of the inner ellipse in Fig. 3. With a semimajor axis of about 0.1 times the length of the bar, the size of this ring is consistent with the observed median value of 0.11 for the ratio of the nuclear ring size to the bar size in the catalog of Comerón et al. (2010). This lends support to the association of the molecular gas in the inner region of the galaxy with a resonance ring. Nuclear rings are common in spiral bar galaxies, and are shown to be the sites of intense starbursts where massive and compact star clusters can form (Elmegreen, 1994; Benedict et al., 2002; Elmegreen et al., 2009; Combes et al., 2013). Note that deeper observations with higher spatial resolutions are required in order to precisely determine the location of the resonance. The existing observations do not provide enough information on the pattern speed of the bar, the epicycle frequency, and the mass distribution, to allow a detailed dynamical modeling of the galaxy.

Although the spatial resolution of data presented here does not allow us to reveal the exact nature of the inner structure of molecular gas in the host galaxy, it is likely that the interaction between the large-scale gas flows in the bar and the gas in the inner region has led to cloud-flow or cloud-cloud collisions. Observations and simulations show that cloud-cloud collisions lead to the assembly of giant molecular complexes, and consequently to the formation of massive star cluster(s) (e.g., Inoue & Fukui, 2013). This is accompanied by an increase in the SFR (e.g., Tan, 2000), and even a decrease in the depletion time (Renaud et al., 2015), similar to starbursts, like those typically hosted by interacting galaxies. In addition, it has been proposed that cloud-cloud collisions trigger the formation of massive stars (Motte et al., 2014; Takahira et al., 2018).

All these, together with the high molecular gas surface densities and short depletion time we report, suggest that massive clusters have formed near the intersection regions of the gas lanes and the inner structure that we have identified in the host galaxy of SLSN PTF10tpz. Such clusters are the birth-place of massive stars, potential progenitors of SLSNe, either directly and/or through a runaway process in their core (Portegies Zwart & van den Heuvel, 2007; Pan et al., 2012; van den Heuvel & Portegies Zwart, 2013). These clusters would then continue their motion along the ring (clockwise in Fig. 5), which is compatible with the location of the SLSN (see Fig. 3 and Fig. 5). The orbital velocity along the ring together with the separation between the formation and explosion sites of the SLSN progenitor can provide constraints on the lifetime and thus the mass of the progenitor star.

In order to have a rough estimate for the mass of the progenitor star we assume the orbital velocity along the ring to be same as the radial velocity at the apocenters of the inner ellipse in Fig. 3, i.e. about 250 km s*-1*. If the progenitor star was indeed formed at the intersection of the gas lanes and the inner structure, it would have an age of about 3 Myr when it reached the explosion site. This would predict a stellar mass of $\sim$ 100 $\rm M_{\odot}$  for the SLSN progenitor star (see Leitherer et al., 1999; Crowther et al., 2010). It is however likely that a delay, of the order of a few times the free-fall time, exists between the assembly of the dense cloud at this intersection and the formation of the star itself. For a surface density of a few thousands $\rm M_{\odot}pc^{-2}$ on the scale of individual molecular clouds, we estimate a volume density of $\sim 50\,\rm M_{\odot}pc^{-3}$ and thus a free-fall time of $\sim 1$ Myr. Considering this delay, the estimated 100 $\rm M_{\odot}$  is in fact a lower limit for the mass of the SLSN progenitor star. The uncertainty on the progenitor mass dominantly comes from the mass–lifetime relation of very massive stars. The theoretical models for this relation suffer from substantial uncertainties, in particular due to dependency on parameters such as the spin and the metallicity of the star and also the internal mixing (e.g., Decressin et al., 2007). However, according to our current knowledge and state of the art models, timescales of the order of 1-2 Myr correspond to the most massive stars known (see e.g., Köhler et al., 2015). The very short timescale that we estimate between the onset of formation and the explosion of the progenitor star ($\lesssim$ 1 Myr) therefore indicates that the progenitor star of SLSN PTTF10tpz is amongst the most massive stars.

4 Summary and Conclusions

We study the spatial distribution and kinematics of molecular gas in the metal-rich and barred host galaxy of SLSN PTF10tpz though CO(1-0) emission line observations with ALMA. We detect the CO(1-0) emission line within the bar of the galaxy. The molecular gas in the inner region of the bar, where the SLSN is located, have surface densities as high as $\sim 1400\,\rm M_{\odot}pc^{-2}$ at $\sim 350$ pc scales. This predicts surface densities of $\sim 5000-10,000\,\rm M_{\odot}pc^{-2}$ for individual GMCs, comparable with those of GMCs in the Antennae galaxy or the starburst regions in the nearby Universe. We find the distribution and kinematics of gas to be consistent with two gas lanes running from the tips of the bar towards the galaxy center. These lanes encounter a gaseous structure in the central $\sim$ 1 kpc of the galaxy. The position of SLSN PTF10tpz is very close to the the intersection regions of the gas lanes and the inner structure. This structure is plausibly associated with an inner Lindblad resonance ring, also suggested by the Keck/NIRC2 image of the inner region of the host. It is likely that the interaction between the large-scale gas flows in the bar and the gas at the inner resonance leads to the assembly of giant molecular complexes, and consequently to the formation of massive star clusters. This is supported by the short depletion time and high surface densities of molecular gas that we measure in the central regions of the galaxy. Our findings therefore suggest in-situ formation of massive star clusters (in one of which the SLSN progenitor was formed) due to the internal dynamics of the host galaxy and without any external contribution such as interaction or merger. This demonstrates the important role that the kpc-scale (hydro-)dynamics of galaxies could play in the formation of massive stars and hence massive star explosions.

We would like to thank Nissim Kanekar for valuable contributions to this project. We thank Robert Quimby, Andy Boden, and Gaspard Duchesne for triggering and acquiring the Keck/NIRC2 observations. Many thanks to Ariane Lancon and Fabrice Martins for valuable insights on the lifetime of massive stars. We also thank Peter Erwin for helpful discussions. M.A. acknowledges support from the Australian Research Council Centre of Excellence for All Sky Astrophysics in 3 Dimensions (ASTRO 3D) through project number CE170100013. F.R. acknowledges support from the Knut and Alice Wallenberg Foundation. D.C. is supported by the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 702622. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada), NSC and ASIAA (Taiwan), and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ.