key: cord-0278825-525al5e3
authors: Michiyama, Tomonari; Ueda, Junko; Tadaki, Ken-ichi; Bolatto, Alberto; Molina, Juan; Saito, Toshiki; Yamashita, Takuji; Zhuang, Ming-Yang; Nakanishi, Kouichiro; Iono, Daisuke; Wang, Ran; Ho, Luis C.
title: Discovery of a [CI]-faint, CO-bright Galaxy: ALMA Observations of the Merging Galaxy NGC 6052
date: 2020-06-17
journal: nan
DOI: 10.3847/2041-8213/ab9d28
sha: 7a9cb09de19eedee6d19194c2ad28fbcf2024ad8
doc_id: 278825
cord_uid: 525al5e3

We report sensitive [ion{C}{1}]~$^3P_1$--$^3P_0$ and $^{12}$CO~$J$=4--3 observations of the nearby merging galaxy NGC 6052 using the Morita (Atacama Compact) Array of ALMA. We detect $^{12}$CO~$J$=4--3 toward the northern part of NGC 6052, but [ion{C}{1}]~$^3P_1$--$^3P_0$ is not detected with a [ion{C}{1}]~$^3P_1$--$^3P_0$ to $^{12}$CO~$J$=4--3 line luminosity ratio of$~lesssim0.07$. According to models of photodissociation regions, the unusual weakness of [ion{C}{1}]~$^3P_1$--$^3P_0$ relative to $^{12}$CO~$J$=4--3 can be explained if the interstellar medium has a hydrogen density larger than $10^5,{rm cm}^{-3}$, conditions that might arise naturally in the ongoing merging process in NGC 6052. Its [ion{C}{1}]~$^3P_1$--$^3P_0$ emission is also weaker than expected given the molecular gas mass inferred from previous measurements of $^{12}$CO~$J$=1--0 and $^{12}$CO~$J$=2--1. This suggests that [ion{C}{1}]~$^3P_1$--$^3P_0$ may not be a reliable tracer of molecular gas mass in this galaxy. NGC 6052 is a unique laboratory to investigate how the merger process impacts the molecular gas distribution.

The atomic carbon emission line with the lower forbidden 3 P fine structure ([C I] 3 P 1 -3 P 0 , hereafter [C I] (1-0)) has a frequency of 492 GHz (λ 609 µm). In the classical framework of photodissociation region (PDR) theory (e.g., Tielens & Hollenbach 1985) , the ultraviolet (UV) photons from young massive stars control the properties of the interstellar medium (ISM) and [C I] emission mostly arises from a narrow transition layer between [C II] and CO. On the other hand, [C I] observations for Galactic molecular clouds (0.1-1 pc scales) frequently show spatial coincidence between [C I] and CO emission lines and brightness temperatures very similar to that of 13 CO J=1-0 (e.g., Ojha et al. 2001; Oka et al. 2001; Ikeda et al. 2002; Kramer et al. 2008; Shimajiri et al. 2013; Burton et al. 2015) . This suggests that [C I] emission can be used to trace the bulk molecular gas mass in clouds. proposed the ubiquitous distribution of atomic carbon throughout a typical Giant Molecular Cloud is due to the combination of dynamics and the non-equilibrium chemistry processes, and Papadopoulos & Greve (2004) showed agreement between molecular gas masses estimated from [C I] and 12 CO emission lines in Arp 220 and NGC 6240.

An Transform Spectrometer (SPIRE/FTS) mounted on the Hershel Space Observatory measured [C I] luminosities for Ultra/ Luminous Infrared Galaxies (U/LIRGs) (Kamenetzky et al. 2016; Lu et al. 2017) . The operation of high-frequency receivers (e.g., Band 8 and 10 receivers) on the Atacama Large Millimeter/submillimeter Array (ALMA) has also enabled investigations of the ∼ 10 pc resolution distribution of atomic carbon in the nucleus (< 1kpc) of nearby galaxies (e.g., Krips et al. 2016; Izumi et al. 2018; Miyamoto et al. 2018; Salak et al. 2019) and ∼ 100 pc scale maps of U/LIRGs such as NGC 6240 (Cicone et al. 2018) and IRAS F18293-3413 (Saito et al., in prep) . In addition, there is increasing availability of [C I] measurements of bright high-z galaxies, such as gravitational lensing galaxies and submillimter galaxies (SMGs), that have been conducted using millimeter/submillimeter interferometers (e.g., Walter et al. 2011; Bothwell et al. 2017; Tadaki et al. 2018; Valentino et al. 2018 Valentino et al. , 2020 .

The relation between [C I] and CO has been investigated using these data. Jiao et al. (2017 Jiao et al. ( , 2019 show a typical integrated brightness temperature ratio (i.e., units in [K km s −1 pc 2 ]) of [C I] (1-0)/CO (1-0)∼ 0.2 and possible enhancement of [C I] abundance in U/LIRGs with respect to normal star-forming galaxies. Salak et al. (2019) find variations of the [C I] (1-0)/CO (1-0) ratio in 3050 pc resolution data in the starburst galaxy NGC 1808, suggesting it is due to different excitation conditions and/or carbon abundance in different regions. Valentino et al. (2018 Valentino et al. ( , 2020 shows a line luminosity ratio of [C I] (1-0)/CO (4-3)∼ 0.1 − 3 (in units of [L ] ) in z ∼ 1 − 2 main sequence galaxies, suggesting this is driven by changes of the density in the range n ∼ 10 3 -10 5 [cm −3 ] based on a simple PDR model (see also Bothwell et al. 2017) .

The determination of the conversion factor from [C I] luminosity to molecular gas mass is a current research topic (Israel et al. 2015; Crocker et al. 2019; Heintz & Watson 2020) . Recent theoretical models posit that atomic carbon is well-mixed with H 2 in several environments, such as metal-poor and/or sources with highcosmic ray fluxes (e.g., Papadopoulos et al. 2018 ). In such environments [C I]-rich/CO-poor gas may be very abundant, suggesting that [C I] could be better than CO as a molecular gas mass tracer.

Summarizing the previous work, [C I] emission is expected to be an indicator of the cold molecular gas but its strength may depend on various parameters. To investigate it further it is necessary to collect a wider range of observations. The Morita Array (the official name of the Atacama Compact Array of 7 m diameter dishes) of ALMA enables us to investigate nearby galaxies with [C I] fluxes that were too low to be detectable with SPIRE/FTS. In this letter, we report the non-detection of [C I] (1-0) and the detection of CO (4-3) in NGC 6052 using the Morita Array.

NGC 6052 (Arp 209, UGC 10182, VV 86, Mrk 297) is a system of two colliding galaxies (east and west components) at a redshift of z = 0.01581 1 (the corresponding luminosity distance is D L = 70.8 Mpc, and the scale is 1. 0 ∼ 333 pc) 2 . The systemic velocity of V sys = 4666 km s −1 is calculated based on the radio velocity convention. The background of Figure 1 is a three-color composite image of NGC 6052 obtained by the Wide Field Camera 3 (WFC3) onboard the NASA/ESA Hubble Space Telescope 3 . The positions of the West and East nuclei are shown by X symbols. Modeling by Nbody simulations suggests that the evolutionary phase of this colliding system is about ∼ 150 Myr after the first impact (Taniguchi & Noguchi 1991) . The total infrared luminosity (8-1000 µm, L TIR ) of NGC 6052 is ∼ 10 11.0 L based on Infrared Astronomical Satellite (IRAS) photometry (Sanders et al. 2003) , which corresponds to a star formation rate (SFR) of SFR∼ 18 M yr −1 using the calibration by Mo et al. (2010) . The metallicity of NGC 6052 has been calculated in literature, but the values vary depending on the method: e.g., 12 + log([O]/[H]) = 8.85 (Sage et al. 1993) 8.65 (James et al. 2002 ), 8.34 (Shi et al. 2005 , and 8.22-8.80 (Rupke et al. 2008) . For the sake of simplicity, we assume the Milky Way metallicity for PDR modeling in SECTION 4.1. filters. The red contours show the CO (4-3) spatial distribution obtained from the Morita Array observation at 3, 6, 9, 12, and 15 sigma levels with an achieved r.m.s of 3.6 Jy km s −1 beam −1 . The grey circle shows the FoV (full width at half maximum of the primary beam ∼ 23. 5 at the CO (4-3) sky frequency). The West and East nuclei are shown as X symbols. The pointing center is indicated by the western black X symbol (16 h 05 m 12. s 88, +20 • 32 32. 61 in ICRS frame). The position of the Western and Eastern nucleus is based on NED. The synthesized beam (3. 8 × 3. 1) and the spatial scale are shown at the left-and right-bottom corners, respectively.

imaging was performed using tclean in CASA with a velocity resolution of 15 km s −1 , cell size of 0. 2, image size of 200 pixel, and Briggs weighting with a robust parameter of 0.5. The clean masks for each channel were determined manually. The synthesized beam size and sensitivity at the velocity resolution of 15 km s −1 are shown in columns (3) and (4) of Table 1 , respectively.

The intensity map integrated within the velocity range of [-72, 63] km s −1 is presented in Figure 2 . The CO (4-3) emission line is detected in the northern part of the galaxy, and it is offset from the optical peak. The CO (4-3) distribution is similar to CO (2-1) obtained by Submillimeter Array (SMA) (Ueda et al. 2014 

is not detected within the FoV, particularly it is undetected in the region detected on the CO (4-3) and CO (2-1) peaks. We measure a velocity-integrated CO (4-3) flux of S CO(4−3) ∆v = 414 ± 124 Jy km s −1 Note-(1) Line name. (2) Table 1 ). In this letter, we adopt 30 % as a conservative systematic uncertainly error, but this is not critical for our conclusions. We note that the error can be estimated from the r.m.s level of the channel map (7 Jy km s −1 ) with a systematic uncertainty associated with the expected absolute flux accuracy of 10 % for ALMA Band 8 4 . However, the actual performance of the flux calibration has not been explored for the ACA stand-alone mode in Band 8 by the ALMA observatory. We obtain a 3σ upper limit for the [C I] (1-0) velocity integrated intensity using

where σ ch = 61.5 [mJy beam −1 ] is the r.m.s revel of the channel map (Table 1) 

where S line ∆v is the velocity-integrated flux in Jy km s −1 , ν rest is the rest frequency of the observing line emission in GHz, z is redshift, D L is the luminosity distance in Mpc, and ν obs = ν rest /(1 + z) is the sky frequency for the observing line emission in GHz (Solomon & Vanden Bout 2005) . We use L line to investigate the molecular gas mass (SECTION 4.2) and L line to investigate the luminosity ratios for a PDR modeling (SEC-TION 4.1).

In this section, we explain low L [CI] (1−0) /L CO(4−3) of < 0.07, which is not well documented in previous observations such as Israel & Baas (2002) ; Valentino et al. (2020) . We use the Photodissociation Region Toolbox (PDRT 5 ) (Kaufman et al. 1999 (Kaufman et al. , 2006 Pound & Wolfire 2008) to investigate density (n H ) and incident FUV radiation field (U uv corresponding to photons with 6 eV≤ hν < 13.6 eV) in units of the average interstellar radiation field in the vicinity of the Sun (G 0 = 1.6 × 10 −3 ergs cm −2 s −1 ) (Habing 1968 ). The modeling predicts line intensities for combinations of n H and U uv by self-consistently solving for chemical processes, radiation transfer, and thermal balance. 2016), and the upper limits we obtain for NGC 6052. The red solid and blue dashed lines indicate the tracks for constant n H and U uv , respectively. According to the plane-parallel PDR model, the UV radiation cannot penetrate deeply into the molecular medium if the region is dense, yielding the narrow [C I] layer and hence low [C I] flux. The upper limit for NGC 6052 (L [CI] (1−0) /L CO(4−3) < 0.07) suggests a dense PDR (n H > 10 5 [cm −3 ]), which is denser than those seen in the comparison galaxies (e.g., Bothwell et al. 2017; Valentino et al. 2018 Valentino et al. , 2020 .

A possible explanation of high densities is compression of the ISM by the ongoing merger because the region associated with the CO (4-3) emission corresponds to the collision front based on the merger model calculation (Taniguchi & Noguchi 1991) . Furthermore, planeparallel model might be suitable for young off-nuclear starburst regions that have not had time to erode the medium. The age of the stellar population at CO detected region is estimated to be ∼ 5.5 Myr by modeling of the mid-infrared atomic lines (Whelan et al. 2007) , suggesting the presence of young off-nuclear starbursts triggered by the ongoing merger. The dense PDR regions associated with young off-nuclear starburst activities might be a hint to understand the [C I]-poor, COrich region in NGC 6052.

The PDRT calculations suggest that the small L [CI] (1−0) /L TIR is indicative of a high U uv . However, it is not possible to obtain U uv from the L [CI] (1−0) /L TIR obtained from our observations. This is because the ratio of the TIR luminosity in the observable region of the Morita Array to the galaxy integrated total TIR luminosity (determined by IRAS measurement) is unknown. Instead, we can access the galaxy integrated value of ([O I] 63 + [C II] 158 )/TIR 6 that can also characterize the PDR. The ratio suggests U uv = 4 − 5. ).

In this section, we investigate whether [C I] (1-0) can trace molecular gas mass inferred from previous CO (1-0) measurements. Albrecht et al. (2007) reported the CO (1-0) integrated intensity of 82 K km s −1 within the 24 single-dish beam. The expected molecular mass is M CO H2 = 8 × 10 8 M assuming a CO (1-0) to H 2 conversion factor α CO =0. Figure 3(b) shows the the relation between CO (1-0) and [C I] (1-0) luminosities for other nearby galaxies taken from Kamenetzky et al. (2016) . The upper limit we measure for the [C I] (1-0)/CO (1-0) ratio in NGC 6052 is located an order of magnitude below the global relation obtained in previous surveys (e.g., Jiao et al. 2017 Jiao et al. , 2019 . This suggests that [C I] (1-0) may not be a reliable tracer of molecular gas mass in NGC 6052. Although both α [CI] and α CO depend on metallicity, such a [C I]-poor region cannot be explained under the assumption of any metallicity (e.g., Bolatto et al. 2013; Glover & Clark 2016; Heintz & Watson 2020) . For example, at lower metallicities CO is would be more easily dissociated due to the lack of dust shielding, and consequently more carbon would be observed as [C I].

An alternative explanation for the big difference between NGC 6052 and other systems in Figure 3 is resolving out spatially extended diffuse gas emission. The emission from [C I](1-0) and CO(1-0) are usually coextensive, and both are likely to originate predominantly from extended, low excitation gas. On the other hand, higher-J CO emission is likely dominated by compact, high excitation gas and the contribution from more diffuse gas. Thus, the [C I](1-0)/CO (4-3) and [C I] (1-0)/CO (1-0) ratio may represent properties of very different phases of the gas, and it is possible that the Morita Array resolved out most extended emission from [C I] (1-0). In such a case, the large difference seen in Figure 3 (b) can be explained if >88 % of the flux of [C I] (1-0) was missing due to its extended structure, possibly larger than the Maximum Recoverable Scale of ∼ 4.3 kpc. While such large missing flux is not generally seen in other galaxies, follow-up mosaic mapping observations including 12m, Morita Array, and total power would be important to check the possibility of an extremely extended [C I] (1-0) distribution in NGC 6052. In either case, whether [C I]-poor or with an extremely extended [C I] distribution, NGC 6052 is a unique laboratory to investigate how the merger process impacts the use of [C I] as a mass tracer.

We report [C I] (1-0) and CO (4-3) observations of the nearby merging galaxy NGC 6052 using the ALMA Morita Array. We detect CO (4-3) with high significance (signal-to-noise ratio of > 40), but [C I] (1-0) is undetected to a stringent upper limit of 0.07 times the strength of the CO (4-3) emission. Models of PDRs can explain the weakness of [C I] as the result of gas densities that are unusually high (n H > 10 5 cm −3 ), which might arise naturally in the collision front of the ongoing merger. In addition, [C I](1-0) is far weaker than expected for the amount of molecular gas inferred from the existing measurements of CO (1-0) and CO (2-1) in NGC 6052. This may suggest [C I]-poor, CO-rich sys-tem and/or extremely extended diffuse molecular gas distribution that is not well documented in literature.

A. NOTES ABOUT SPIRE/FTS MEASUREMENTS NGC 6052 was observed by SPIRE/FTS (Kamenetzky et al. 2016; Lu et al. 2017) . We downloaded the science data products automatically generated by the data processing pipelines from "herschel science archive 7 ". Same as Lu et al. (2017) , we adopt point-source calibration and fit simultaneously the continuum and emission lines using a 5th order polynomial and sinc profiles following the "Spectrometer Line Fitting" script in Herschel Interactive Data Processing Environment (HIPE). Figure A. 1 shows the continuum subtracted SPIRE/FTS spectrum (black solid lines) and the fitting results (green dashed lines). Only CO (4-3) is marginally detected with the S/N of 3.9 and both [C I] (1-0) and [C I] (2-1) are not detected (S/N 3). In our analysis, S/N is calculated using the ratio between peak flux density obtained by spectrum fitting and the rms value of the flux density in the line free frequency range (i.e., 460-480 GHz for CO(4-3) and [C I] (1-0) and 800-820 GHz for [C I] (2-1) emissions), which is not the same method with Kamenetzky et al. (2016) and Lu et al. (2017) . Since the S/N is ∼3 for [C I] (1-0) and [C I] (2-1), these lines may be considered as "detection" in different methods. For example, [C I] (1-0) emission line is considered as "detection" in the analysis by Kamenetzky et al. (2016) , but "non-detection" by Lu et al. (2017) and our analysis. In Table A .1, the velocity integrated flux (and 3σ upper limits) measured by our analysis, Kamenetzky et al. (2016) , and Lu et al. (2017) are shown. Comparison of this SPIRE measurements with our Morita Array result implies that the CO (4-3) flux recovered by the Morita Array may be ∼ 30% of the total. But, given the fact that the Herschel beam is four times the solid angle of our FoV and the considerable uncertainties in the FTS measurement (i.e., The intensity is an analytical solution assuming sinc function for the spectroscopically unresolved line. 8 ), it is impossible to directly compare SPIRE flux and our Morita Array's results. Further discussion of SPIRE/FTS measurements is beyond the aim of this letter because the line detection is marginal even for the CO (4-3) emission. The best fit is shown as green dashed line. The peak flux density and the frequency offsets from the systematic velocity measured by CO (4-3) are shown at the right upper corners for each panel. The S/N is calculated by our own definition (see SECTION A). Supplementary, the CO (4-3) (blue) and [C I] (1-0) (red) spectrum obtained by this project are shown in (a) and (b), respectively. The grey arrows indicate the frequencies of the CO (4-3) and [C I] (1-0), and [C I] (2-1) corresponding to the systemic velocity. The second peak fitted in (c) is for CO (7-6) emission line. 

Note-(1) Line name.

(2)-(4) The velocity integrated line flux measured by this project, Kamenetzky et al. (2016) , and Lu et al. (2017) .

Astronomical Data Analysis Software and Systems XVI

Galaxy Formation and Evolution

Astronomical Data Analysis Software and Systems XVII

This work was supported by the National Science Foundation of China (11721303, 11991052) and the National Key R&D Program of China (2016YFA0400702). We are grateful to the anonymous referee for useful comments which helped the authors to improve the paper. The authors appreciate Prof. Masami Ouchi (The University of Tokyo) and Prof. Takuya Hashimoto (Tsukuba University) who provided a comfortable and fruitful research environment during the COVID-19 pandemic time. This paper makes use of the following ALMA data: ADS/JAO.ALMA #2018.1 00994. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada), MOST 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.