The High Molecular Gas Content, and the Efficient Conversion of Neutral into Molecular Gas, in Jellyfish Galaxies

What ultimately regulates galaxy evolution is the availability of gas prone to star formation, and its efficiency in forming new stars (Schmidt 1959; Kennicutt 1998). Therefore, in the last years, a lot of effort has been dedicated to building statistically significant samples of galaxies with observational coverage able to map all the gas phases, with the aim of discovering the underlying scaling relations that correlate the neutral and molecular gas content with the stellar mass, if any (Cortese et al. 2011; Saintonge et al. 2011a, 2011b; Corbelli et al. 2012; Catinella et al. 2018). More recent results also include the dust contribution (Casasola et al. 2020).

Scaling relations have been derived, in fact, for the H i content of nearby galaxies (Bigiel et al. 2008, 2010) while only small samples exist at higher redshifts (Fernández et al. 2016; Cortese et al. 2017; Catinella et al. 2018), suggesting that the gas in the molecular phase was the predominant fraction at early epochs.

As for the cold gas content, the largest statistical sample for which a homogeneous set of data has been collected so far is the COLDGASS (Saintonge et al. 2011a) sample together with its low mass extension xCOLDGASS (Saintonge et al. 2017). Galaxies in the xCOLDGASS sample are mass-selected at z = 0.01–0.05, without any a priori bias on the IR/UV fluxes, and cover a range in stellar mass between 10⁹ and ${10}^{11.5}{M}_{\odot }$. While this sample gives a wonderful insight into the gas properties of galaxies in general, it has not been designed to cover galaxies in different environmental conditions, i.e., it does not distinguish between galaxy properties in clusters and those in the field, which are expected to behave differently. While early works on the Virgo and Coma clusters (Kenney & Young 1989; Boselli et al. 1997) have revealed no significant molecular gas deficiency in cluster galaxies (including the most H i deficient), more recent studies that make use of larger samples and better resolution find, instead, that H i gas stripped galaxies also have a lower H₂ content, albeit in lower proportion (Boselli et al. 2014). Similar results are also given by Corbelli et al. (2012), who study 35 metal-rich spiral galaxies that are part of the Herschel Virgo Cluster Survey. These galaxies have a well determined H i deficiency parameter, which has been found to anticorrelate with the molecular gas fraction, i.e., the ratio between molecular gas mass and stellar mass. Indeed, the molecular gas fraction has been found to decrease as the H i deficiency increases, while the ratio between the molecular and the total gas increases. This has been interpreted as due to the fact that both neutral and molecular gas are stripped, but the former is stripped more easily than the latter.

Single dish and, more recently, interferometric ALMA data have confirmed a normal molecular gas content in the disk of ESO137-001, a jellyfish galaxy in the nearby Norma cluster (Jáchym et al. 2014, 2019), albeit confined in a very small central region (with ∼1.5 kpc radius), with a similar amount of molecular gas detected along the stripped tail. The D100 galaxy close to the Coma cluster center, instead, shows a higher than expected molecular gas fraction within the central ∼2 kpc radius (Jachym et al. 2017), and a very H₂ rich gas tail, as traced by the CO emission.

A clear view on the molecular gas content of cluster galaxies subject to ram-pressure stripping is still missing, and we can now start to cast light on the subject by using the multiwavelength data set collected by the GASP sample. The GASP survey⁹ (Poggianti et al. 2017b) has started exploring the effects of environmental interactions on nearby ($z\sim 0.05$) cluster/group galaxies thanks to a dedicated VLT MUSE Large Program (GAs Stripping Phenomena in galaxies with MUSE, P. I. B. Poggianti) mainly tracing the ionized gas, but complementary data sets at different wavelengths are being collected and have started offering a clear view on all the connected gas phases (see also George et al. 2018; Poggianti et al. 2019; Ramatsoku et al. 2019, 2020; Deb et al. 2020). In particular, in Moretti et al. (2018) we observed with the APEX telescope 4 GASP galaxies detecting molecular gas both in the galaxy disks and in the ionized gas tails.

We then started an observational campaign devoted to measuring the molecular gas content of these galaxies using ALMA interferometric data at ∼1 kpc resolution. We have described the observations and the data analysis in Moretti et al. (2020), where we have shown the results for the JW100 galaxy. Surprisingly, in this galaxy we have found an anomalously large content of molecular gas, even excluding the new molecular gas possibly born in the tail from the stripped neutral gas. Assuming the standard conversion factor, we tentatively concluded that the star formation efficiency (SFE), i.e., the star formation rate (SFR) surface density divided by the H₂ mass density, on scales of 1 kpc shows a gradient moving from the central part of the disk toward the stripped tail, and the corresponding depletion time is always longer than the typical ${\tau }_{\mathrm{dep}}$ of nearby disk galaxies (i.e., ∼1–2 Gyr, Bigiel et al. 2011; Leroy et al. 2013), increasing from the disk to the tail.

In this paper we analyze the ALMA Band 3 data of the four GASP galaxies observed in cycle 5 to understand whether their molecular gas content is following the scaling relations of normal galaxies, or if instead they are depleted/enriched, in molecular gas.

All the GASP galaxies here analyzed are operatively defined as jellyfishes (as they possess ionized gas tails whose length is comparable to the galaxy disk diameter, Poggianti et al. 2017a), and they also have a central AGN. Stellar masses and redshifts are given in Moretti et al. (2018), while results from MUSE data are described in Poggianti et al. (2017b), Bellhouse et al. (2017), Gullieuszik et al. (2017), Poggianti et al. (2019), and Moretti et al. (2020).

Throughout this paper we will make use of the standard cosmology ${H}_{0}=70\,\mathrm{km}\,{{\rm{s}}}^{-1}\,{\mathrm{Mpc}}^{-1}$, ${{\rm{\Omega }}}_{M}=0.3$, and ${{\rm{\Omega }}}_{{\rm{\Lambda }}}=0.7$. As in the other GASP papers, our stellar masses are calculated adopting a Chabrier (2003) initial mass function (IMF).

Observations of the CO(1–0) emission (rest frequency 115.271 GHz) of our four galaxies, namely JO201, JO204, JO206, and JW100, have been obtained with ALMA during Cycle 5 (project 2017.1.00496.S). The spectral configuration used yields a velocity resolution of 3.1 km s⁻¹, which has been smoothed to 20 km s⁻¹ in the final data cubes used for the following analysis. In order to cover with homogeneous sensitivity the entire area observed with MUSE, including disk and tails, mosaics have been necessary. The actual configurations used for the observations allowed us to sample scales up to the maximum recoverable scale (MRS), reported in Table 1 for each galaxy.

Table 1.  Properties of CO(1–0) Line Images: Observed Frequency, Synthesized Beam (${\theta }_{\mathrm{maj}}$, ${\theta }_{\min }$, and PA), rms, and Maximum Recoverable Scale (MRS; Equation (7) in the ALMA Technical Handbook)

Galaxy	${\nu }_{\mathrm{obs}}$	${\theta }_{\mathrm{maj}}$	${\theta }_{\min }$	PA	rms	MRS
	GHz	''	''	deg	mJy b−1	''
JO201	110.307	1.99	1.57	−84.6	0.5	20
JO204	110.625	1.62	1.36	81.5	0.5	20
JO206	109.677	1.60	1.30	−87.4	0.7	23
JW100	108.644	2	1.7	8.3	0.9	24

Download table as:  ASCIITypeset image

The data have been calibrated and imaged using the CASA software (version 5.4.0-7; McMullin 2007), as described in Moretti et al. (2020). The details of the obtained line images are reported in Table 1

From the cleaned data cubes we obtained moment-zero maps, using the SoFiA software (Serra et al. 2015) to construct detection masks for each ALMA datacube, as described in Moretti et al. (2020) for JW100. Figure 1 shows the CO(1–0) moment-zero for the disks of the four galaxies analyzed here.

Zoom In Zoom Out Reset image size

For each galaxy we measure both the amount of molecular gas within the stellar disk and outside it, where the CO is located in correspondence with the ionized gas tails found in Poggianti et al. (2017b), Bellhouse et al. (2017), Gullieuszik et al. (2017), and Poggianti et al. (2019). We use here as a disk definition the one given in Gullieuszik et al. (2020) (shown as a red contour in Figure 1), which is based on the stellar isophote at 1σ above the sky background level measured on the undisturbed side of the galaxy and then symmetrized to exclude the contribution from the stripped tail.

Table 1 shows that we will base our results on the detection of molecular gas on scales between ∼1 and ∼20 kpc, thus neglecting the contribution from any gas diffuse on larger scales. We do not have single dish data at the CO(1–0) frequency to assess the possible flux loss, and therefore our data are, strictly speaking, lower limits. The uncertainty on the measured fluxes is of the order of 10%.

As shown in Figure 1, the molecular gas distribution in the galaxy disk is quite consistent with the Hα emission (colored contours) derived from MUSE data (Bellhouse et al. 2017; Gullieuszik et al. 2017; Poggianti et al. 2017b, 2019).

For each galaxy we have used the CO(1–0) emission to derive the CO flux and the H₂ mass using the following equation from Watson & Koda (2017):

$$\begin{eqnarray}&&\left(\displaystyle \frac{{M}_{{{\rm{H}}}_{2}}}{{M}_{\odot }}\right)=1.1\times {10}^{4}\left(\displaystyle \frac{{\alpha }_{\mathrm{co}}}{4.3}\right)\left(\int {S}_{10}{dv}\right){\left({D}_{L}\right)}^{2},\end{eqnarray} \tag{ 1 }$$

where ${\alpha }_{\mathrm{co}}$ is the CO-to-H₂ conversion factor expressed in ${M}_{\odot }{\left({\rm{K}}\mathrm{km}{{\rm{s}}}^{-1}{\mathrm{pc}}^{2}\right)}^{-1}$, S₁₀ is the CO flux density in janskys and D_L is the luminosity distance in megaparsecs.

As is well known, the H₂ mass is strongly dependent on the ${\alpha }_{\mathrm{CO}}$ factor. Table 2 gives the molecular gas masses for different assumptions on ${\alpha }_{\mathrm{CO}}$, as well as the SFRs measured within the disk using the star-forming spaxels from Vulcani et al. (2018) and the corresponding H₂ depletion times, defined as ${M}_{{{\rm{H}}}_{2}}/{SFR}$.

Table 2.  Molecular and Atomic Gas Properties of the Observed Galaxies

M⋆	SFR	${\alpha }_{\mathrm{CO}}$	${M}_{{{\rm{H}}}_{2},\mathrm{in}}$	${M}_{{{\rm{H}}}_{2},\mathrm{out}}$	${\tau }_{\mathrm{dep},{{\rm{H}}}_{2}}$	MH i	Rmol	${M}_{\mathrm{gas}}/{M}_{\star }$	${\tau }_{\mathrm{dep},\mathrm{tot}}$
1e9 ${M}_{\odot }$	${M}_{\odot }$ yr−1		1e9 ${M}_{\odot }$		Gyr	1e9 ${M}_{\odot }$			Gyr
				JO201
		4.3	16.5[46%]	2%	3.3		14	0.53	3.5
35.5	5 ± 1	3.0a	11.5[32%]		2.3	1.15	10	0.39	2.6
		1.3b	5.0[14%]		1		4	0.21	1.3
				JO204
		4.3	8.1[20%]	1%	5.4		⋯	⋯	⋯
40	1.5 ± 0.3	3.0a	5.7[14%]		3.9	⋯	⋯	⋯	⋯
		1.4b	2.7[7%]		1.8		⋯	⋯	⋯
				JO206
		4.3	8.7[10%]	0.4%	1.8		12	0.14	1.9
90	4.8 ± 0.9	2.8a	5.6[6%]		1.2	0.7	8	0.10	1.3
		2.0b	4.0[4%]		0.8		6	0.08	1.0
				JW100
		4.3	23.7[8%]	0.6%	9.1		132	0.09	9.2
300	2.6 ± 0.5	3.0a	16.5[5%]		6.3	0.2	92	0.07	6.4
		0.9b	5.0[2%]		1.9		28	0.03	2.0

Note. For each galaxy and ${\alpha }_{\mathrm{CO}}$ assumption (^a from Accurso et al. 2017 and ^b from Amorin et al. 2016) we list stellar masses; SFRs from Hα MUSE emission within the disks from Vulcani et al. (2018); molecular gas masses within the galaxy disk assuming different ${\alpha }_{\mathrm{CO}}$ and, in parenthesis, the molecular gas to stellar mass fractions in the disk; molecular gas to stellar mass fractions in the stripped tail; H₂ depletion times; H I masses within the disk from Ramatsoku et al. (2019, 2020) and T. Deb et al. (2020, in preparation); ${R}_{\mathrm{mol}}={M}_{{{\rm{H}}}_{2}}/{M}_{\mathrm{HI}}$ in the disks; and total gas fractions (molecular+neutral, disks+tails) and corresponding total depletion times for the different ${\alpha }_{\mathrm{CO}}$.

Download table as:  ASCIITypeset image

First, we used the Milky Way ${\alpha }_{\mathrm{CO}}$ that is equal to 4.3 ${M}_{\odot }{\left({\rm{K}}\mathrm{km}{{\rm{s}}}^{-1}{\mathrm{pc}}^{2}\right)}^{-1}$ (Bolatto et al. 2013), including the helium correction, which is the standard value used in the literature. We used this value to calculate the molecular gas masses and the corresponding gas fractions with respect to the galaxy stellar mass, ${f}_{{{\rm{H}}}_{2}}={M}_{{{\rm{H}}}_{2}}/{M}_{\star }$, shown as filled symbols in Figure 2.

Zoom In Zoom Out Reset image size

We used the same assumption to calculate the molecular gas mass present in the tail clumps. As described in detail in A. Moretti et al. (2020, in preparation), which deals with the extraplanar emission, the molecular gas masses in correspondence to the tail star-forming clumps ($0.4\mbox{--}1.7\times {10}^{9}{M}_{\odot }$) amounts to only a small fraction of the total stellar mass (see column 5 in Table 2).

As a second step, in order for our data to be comparable with the xCOLDGASS sample, we calculated the H₂ masses assuming an ${\alpha }_{\mathrm{CO}}$ variable with the metallicity, following the relation found for the same sample by Accurso et al. (2017), that depends both on the galaxy metallicity and on its distance from the star formation main sequence. This relation, though, only holds up to metallicities of 12+log(O/H) = 8.8.

We used the [O iii]/[S ii+] versus [N ii]/[S ii+] line ratio to derive the spatially resolved gas phase metallicities (Franchetto et al. 2020) for our galaxies and derived the mean ${\alpha }_{\mathrm{CO}}$ within the disk. As our method to estimate the gas metallicity is different from the one used by Accurso et al. (2017) our galaxies might be skewed toward higher values. This does not bias the results, though, as we mostly use the asymptotic value from the Accurso et al. (2017) relation.

As for the distance from the main sequence of the SFR–mass relation, we assumed the average value found for jellyfish galaxies in Vulcani et al. (2018), i.e., 0.15 dex.

Both using the Milky Way ${\alpha }_{\mathrm{CO}}$ (filled symbols in Figure 2) and the metallicity-dependent one (empty symbols), we find that the total molecular gas fractions in our jellyfish galaxies are significantly larger than the mean values found for star-forming main-sequence galaxies in the xCOLDGASS sample by Saintonge et al. (2017), shown as a reference in Figure 2. Gray dots in Figure 2 are the single data points from Saintonge et al. (2017) for galaxies within 0.4 dex from the main sequence, demonstrating indeed that our jellyfishes lie at the upper edge of the observed distribution. We note that the results obtained by Bolatto et al. (2017) on a subsample of the CALIFA galaxies observed with the CARMA interferometer are perfectly in agreement with the Saintonge relation. The Virgo cluster data by Corbelli et al. (2012), represented as black filled dots in Figure 2, having rescaled the stellar masses to correct for the different assumption in the IMF, lie below the mean field relations, which led the authors to conclude that H₂ has been stripped from the galaxies.

We notice that the ram pressure acting on the infalling galaxies could in principle unbind the already existing giant molecular clouds (GMCs) in the galaxy disk, increasing the CO/H₂ ratio. If this were the case, for a correct estimate of the molecular gas mass we should be using a lower ${\alpha }_{\mathrm{CO}}$, more similar to that found, for example, in ULIRGs (see Bolatto et al. 2013; Sandstrom et al. 2013; Israel et al. 2015) where the underlying cold gas distribution is more diffuse. To mimic this effect, we also used a second metallicity-dependent formulation of ${\alpha }_{\mathrm{CO}}$ (Amorin et al. 2016; which gives results compatible with the studies by Schruba et al. 2012 and Genzel et al. 2012 for higher redshift galaxies) where this factor reaches values that go from 0.9 (in JW100) to 2.0 (in JO206). The corresponding molecular gas masses are also given in Table 2.

With these assumptions we find the molecular gas fractions, shown in Figure 2 as arrows, which are closer to the literature relations. We caution, however, that these extremely low conversion factors have been extrapolated using relations meant to include primarily low metallicity galaxies. The behavior of the ${\alpha }_{\mathrm{CO}}$ in the high metallicity range is still debated, as some authors do not find significant deviations from the MW value (Wolfire et al. 2010; Sandstrom et al. 2013).

We conclude that the molecular gas fractions we derive for our jellyfish galaxies are much higher than in both field galaxies and in the Corbelli et al. Virgo cluster galaxies. Using a constant MW-like ${\alpha }_{\mathrm{CO}}$, the molecular gas fractions at given stellar mass are higher than the mean values in xCOLDGASS galaxies by an average factor of ∼5, while using the Accurso et al. (2017) relation by an average factor of ∼4, with molecular gas fractions ranging between ∼8% and ∼50%. The derived H₂ masses are extremely high, ranging between 8 and 24 $\times {10}^{9}\,{M}_{\odot }$. Using a second metallicity-dependent ${\alpha }_{\mathrm{CO}}$ (plausible if the molecular gas were much more diffuse than usual due to the ram pressure actually disrupting GMCs, but highly uncertain at the high metallicities of our galaxies) yields molecular gas fractions closer to the literature values for nonstripped galaxies, though still well above the mean for field galaxies and still higher than those observed in Virgo spirals. Thus, unless the ${\alpha }_{\mathrm{CO}}$ is close to the value mostly found in ULIRGs (in the sense that most of the molecular gas is diffuse), jellyfish galaxies have huge reservoirs of H₂.

The standard molecular depletion time in galaxy disks, resolved on an ∼1 kpc scale, is ∼2 Gyr (Bigiel et al. 2008; Leroy et al. 2008). The H₂ depletion timescales of our jellyfishes are significantly longer if we consider MW-like ${\alpha }_{\mathrm{CO}}$, except for JO206 (2). When using the Accurso et al. (2017) scaling relation, the timescales for JW100 and JO204 are still longer than normal galaxies, while the other two are more consistent with the literature values (JO201 and JO206). Assuming a very low ${\alpha }_{\mathrm{CO}}$ over the entire extent of the disk leads to timescales generally shorter than in normal galaxies.

3.1.  ${{\rm{H}}}_{2}/{\rm{H}}\,{\rm{I}}$ Mass Ratio, Total Gas Mass Fractions, and Depletion Times

Having established that molecular gas is abundantly present in jellyfish galaxies, we now proceed to evaluate the proportion of the different gas phases within the galaxy disk to better characterize the star-forming cycle. Three of our galaxies (Ramatsoku et al. 2019, 2020; T. Deb et al. 2020, in preparation) also possess an estimate of the H i gas content (preliminary from MeerKat data for JW100), while for the fourth one (JO204) this has been hampered by H i absorption due to the presence of a continuum source at the center of the galaxy (Deb et al. 2020). We note that within the disk the neutral and the molecular gases have similar distributions, at the observed spatial scales. For the JO206 galaxy we had to recalculate the mass within the disk using the disk definition adopted here.

In Figure 3 we show the disk molecular gas fraction ${R}_{\mathrm{mol}}={M}_{{{\rm{H}}}_{2}}/{M}_{{\rm{H}}{\rm{I}}}$ of these three galaxies, compared with the values in spiral disks derived from the xGASS sample assuming a radial fit to the H i mass distribution (Wang et al. 2020). R_mol in the disk of our galaxies turns out to be very high (ranging from 4 to ∼100 ), while the average expected value in galaxy disks is ${R}_{\mathrm{mol}}=1$ (black horizontal line in Figure 3). Our data are clearly significantly offset compared to normal galaxies with the same stellar mass. This result persists for any assumption on ${\alpha }_{\mathrm{CO}}$.

Zoom In Zoom Out Reset image size

Since R_mol can be considered proportional to the ratio between the typical GMC lifetime over the conversion time between the neutral and the molecular phase (Leroy et al. 2008), our results strongly suggest that the conversion of H i into molecular gas is very efficient in jellyfish galaxies. We note that similar trends have been found also in interacting galaxies (Casasola et al. 2004).

Strikingly, summing up the molecular and neutral gas masses and considering the total (disk and tail) ratio of gas and stellar masses, we obtain gas mass fractions (shown in the right panel of Figure 3) that are within the observed range of normal galaxies. In other words, the total gas (molecular+neutral) associated to our jellyfish galaxies (considering both what is left in the disk and what is in the tail) is normal for their stellar mass. These results do not change if we restrict these comparisons to normal galaxies within the range of stellar mass surface density similar to ours (that goes from 8 to 9 ${M}_{\odot }\,\,{\mathrm{kpc}}^{-2}$).

Finally, considering depletion times, we note that in Ramatsoku et al. (2020), for JO201 and JO206, we found that H i depletion times are much shorter than the average value found in galaxy disks, apparently implying a very efficient star formation (as far as H i is concerned) and/or a low H i content for their SFR. Evaluating the global (H i + H₂) depletion time for the three galaxies where we have both measurements, it turns out that within the galaxy disk the estimated depletion time is in agreement with literature values (Leroy et al. 2008) only for the JO206 galaxy, while JO201 and JW100 have unusually long depletion times, unless we assume extremely low values for the ${\alpha }_{\mathrm{CO}}$.

Our analysis of the molecular gas content of four GASP ram-pressure stripped galaxies measured on 1 kpc scale with ALMA reveals a very high molecular gas fraction with respect to both isolated/field galaxies and Virgo cluster galaxies with a similar stellar mass.

Our results are dependent on the assumptions on the ${\alpha }_{\mathrm{CO}}$, but even assuming a very low ${\alpha }_{\mathrm{CO}}$ (∼1, so far observed mostly in ULIRGs) the H₂ content is higher than typical values for normal galaxies of the same mass.

When considering the neutral gas still present in the disk, we find an enhanced molecular-to-neutral gas ratio R_mol with respect to undisturbed galaxies (Wang et al. 2020; even when using the lowest ${\alpha }_{\mathrm{CO}}$), and at the same time a total gas fraction, which is in good agreement with the scaling relations for normal spirals found by Catinella et al. (2018).

These results strongly suggest that the gas compression caused by the ram pressure in the peak stripping phase causes the conversion of large amounts of H i into the molecular phase in the disk, possibly implying that only part of the H i gets efficiently stripped.

We acknowledge funding from the agreement ASI-INAF n.2017-14-H.0, as well as from the INAF main-stream funding program. B.V., M.G., and R.P. also acknowledge the Italian PRIN-Miur 2017 (PI A. Cimatti). This project has received funding from the European Research Council (ERC) under the European Union's Horizon 2020 research and innovation program (grant agreement No. 833824, GASP project and grant agreement No. 679627, FORNAX project). We acknowledge S. Tonnesen and Y. Jaffé for the stimulating discussion and J. Wang for sharing in electronic format her data. This paper makes use of the following ALMA data: ADS/JAO.ALMA#2017.1.00496.S. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada) and NSC and ASIAA (Taiwan), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ. Based on observations collected by the European Organisation for Astronomical Research in the Southern Hemisphere under ESO program 196.B-0578 (VLT/MUSE). This research made use of APLpy, an open-source plotting package for Python (Robitaille and Bressert, 2012; Robitaille, 2019).

Facilities: ALMA - Atacama Large Millimeter Array, VLT(MUSE). -

Software: CASA (McMullin & Waters 2007).