1 Introduction
Of all the important unsolved problems in modern astrophysics, the development of a general and complete theory of star formation, applicable in all environments and at all cosmic times, would arguably have the greatest impact on our understanding of the universe. The challenge lies in the vast range of scales and processes involved, with gravity, turbulence and magnetic fields being the lead actors (e.g. McKee & Ostriker, 2007). While modern star formation theories can successfully explain the star formation efficiency of molecular clouds in environments similar to the solar neighbourhood, they fail to accurately describe star formation in more extreme environments, such as the high-pressure central regions of galaxies or the low metallicity gas that must have permeated primordial galaxies.
The reason for this predicament is that our understanding of the physics of star formation is currently vastly based on observations of the Milky Way. These observations have the highest physical resolutions possible (e.g., the Atacama Large Millimeter/submillimeter Array, ALMA, resolves ~0.1–0.2″ or ~1000 AU scales in the Galactic central molecular zone or complex substructures of Galactic clouds), but they are limited in the range of environments they can probe (in terms of density, metallicity, UV and X-ray radiation fields, cosmic rays, magnetic fields, etc), not to mention difficulties with distance estimation and confusion along the line of sight. Nearby galaxies (within ∼ 100 Mpc) are therefore of crucial importance, as they alleviate these issues while still being observable at resolution of tens of parsecs, or even better in the case of Local Group galaxies such as the Small and Large Magellanic Clouds (SMC and LMC, respectively2).
Indeed, significant progress has been made over the past decade thanks to new and upgraded (sub)millimeter facilities such as ALMA and the IRAM 30m/NOEMA starting to enable "Galactic-like" science in extragalactic environments, while also enabling observations of the ISM in large, representative galaxy samples. From such observations, a picture is emerging where both large-scale environment and local conditions impact the structure and properties of molecular clouds and their star formation outputs: we observe variations from galaxy to galaxy (e.g. Saintonge & Catinella, 2022), with environment within nearby galaxies (e.g. Leroy et al., 2021b), and even between the inner and outer regions of the Milky Way (e.g. Schruba et al., 2019). In parallel, sensitive dust polarimetry measurements with instruments such as POL2 on the JCMT and HAWC+ on SOFIA have proved beyond doubt the crucial role of magnetic fields in regulating star formation (e.g. Chuss et al., 2019; Pattle & Fissel, 2019; Pattle et al., 2023; Pillai et al., 2020). These studies show that magnetic fields channel gas flows along galactic filaments, thus affecting the structure and history of star formation history on galactic scales.
To achieve another step change in our understanding of ISM physics/chemistry and star formation, we need yet another significant increase in our observing capabilities. The Atacama Large Aperture Submillimeter Telescope (AtLAST) is a concept for a next generation 50-meter class single-dish submm telescope that would offer such opportunities (Klaassen et al., 2020; Mroczkowski et al., 2023; Mroczkowski et al., 2024). AtLAST is designed to offer very high sensitivity and exceptional mapping speeds through both its large aperture, extended 1–2 degree field of view, and the excellent atmospheric conditions of the Chilean Chajnantor plateau. With its unique combination of spatial coverage, angular resolution and sensitivity (in particular to extended, low surface brightness emission), AtLAST will enable the study of the ISM in nearby galaxies as has never been possible before.
In this case study, we explore four prominent nearby galaxies science areas where AtLAST will be transformational:
1. ISM physics and star formation in the LMC and SMC, with a focus on the important yet poorly studied diffuse gas (e.g., atomic gas, low surface density gas in the outer part of galaxies) and dust;
2. the origin and structure of magnetic fields, from tens of kpc down to sub-kpc scales, and their impact on star formation;
3. the physics and chemistry of the ISM as probed by multiple dust and gas tracers;
4. ISM properties and star formation across the full range of environments of the local Universe through large statistical surveys.
The background and key science questions are presented in Section 2, highlighting those that AtLAST will be uniquely positioned to address, with a description of possible survey strategies. The technical requirements for AtLAST to reach these science goals are summarised in Section 3. Finally, we summarize the key science goals and technical requirements for AtLAST in Section 4.
This paper is one of the series AtLAST Science Case papers that highlight the unique role of AtLAST in all the relevant astrophysics aspects, including science questions relating to Milky Way ecology, the sun, high-redshift galaxies, large-scale structure and the early Universe. We provide a comparison of AtLAST's proposed capabilities to that of other current facilities in Table 1, and highlight the irreplaceable place of AtLAST in the collecting area --- mapping speed parameter space in Figure 1.
Table 1. Comparing AtLAST with current millimeter and submillimeter telescopes.
| (1) | IRAM
(2) | LMT
(3) | NOEMA
(4) | CSO
(5) | JCMT
(6) | ASTE
(7) | APEX
(8) | GLT
(9) | CCAT-p
(10) | ALMA 12m array
(11) | AtLAST
(12) |
|---|
Antenna
diameter [m] | 30 | 50 | (12×)15 | 10.4 | 15 | 10 | 12 | 12 | 6 | (54×)12 | 50 |
Surface
Accuracy
R.M.S. [µm] | 66 | 75 | 50 | 15 | 23 | 20 | 17 | 17 | 10.7 | 16 | ≤20 |
Line
Frequency
Range [GHz] | 72–350 | 88–350 | 70–276 | 150–850 | 77–375 | 270–
500 | 159–738 | 86–345 | 460, 830 | (35)
84–950 | 84–950(?) |
Continuum
Camera | NIKA2 | TolTEC | ... | MUSIC;
SHARC
II | SCUBA2 | TES
Camera | LABOCA;
ArTéMiS | ... | PrimeCam | ... | TBD |
Continuum
Frequency
Range [GHz] | 230, 345 | 150,
210,
270 | ... | 270;
666, 850 | 350,
666 | 270,
350 | 345;
850 | ... | 850 | ... | TBD |
Field of
View (FoV)
[arcmin](a) | 6.5 | 4 | 0.28 | 14 | 6.7 | 7.5 | 11 | 0.35 | 78 | 0.35 | ≥60 |
Angular
Resolution,
νref | 11″,
260GHz | 5.0″,
300GHz | 0.2–3.9″ | 8″,
850GHz | 14″,
350GHz | 11″,
450GHz | 21″,
300GHz | 21″,
300GHz | 15″,
850GHz | 0.005–12″ | 1.7–15″ |
Continuum
noise-
equivalent
flux density
(NEFD) [mJy
s1/2] , νref | 33,
260GHz | 1.5,
280GHz | 4.11,
260GHz
(b) | 55,
270GHz | 90,
350GHz | 10,
270GHz | 110,
345GHz;
600,
850GHz | ... | 21, 280GHz;
400, 850GHz
(c) | 1.37, 280GHz
(b) | 1.05, 280GHz
(d) |
Mapping
Speed(e) [deg2
mJy-2 hr-1] | 0.0219 | 5.4 | 0.00335 | 0.0646 | 0.109 | 0.277 | 0.0078 | ... | 10.5 | 0.0465 | 2.48×103 |
Figure 1. Comparison of nominal collecting area and mapping speed of AtLAST and other current millimeter/submillimeter facilities.
Nominal collecting area is simply the antenna dish area. Mapping speed is computed as the field of view divided by the square of the noise-equivalent flux density (NEFD). See Table 1 for the detailed telescope parameters.
2 AtLAST nearby galaxies science case highlights
2.1 Dust and gas in the LMC and SMC
2.1.1 Context and open questions. The LMC and SMC are the most massive satellites of our Milky Way Galaxy. Together with Andromeda (M 31), M 33, as well as other smaller galaxies out to D ∼ 1.5Mpc, they form our Local Group. The LMC and SMC are strongly bound to the Milky Way and on a collision course, with the LMC expected to merge with our Galaxy in ∼ 2.5 Gyr (Cautun et al., 2019). The intensity of the gravitational interaction between the SMC, LMC and Milky Way is pulling significant amounts of the interstellar gas away from the galaxies, which can be observed across the sky as large Hi streams (Mathewson et al., 1974; Wannier & Wrixon, 1972; see Figure 2).
Figure 2. Full sky view of the LMC, SMC and the associated Stream system.
Left: LMC, SMC, the Stream between the LMC/SMC and the Milky Way, and the leading arm of Stream in Hi (red), on top of the optical all-sky image in Aitoff projection (West to the left and North up), adapted from Nidever et al. (2010). (Credit: D. Nidever/ NASA/ Simons Foundation URL.) Right: LMC and SMC in sky R.A. and Dec. coordinates (East to the left and North up), adapted from the ESO VISTA survey of the Magellanic Clouds system (VMC; Cioni et al., 2011; see also https://www.eso.org/sci/publications/announcements/sciann17518.html, with blue squares showing the footprints of the VISTA survey covering ~110 deg2 in the LMC, ~40 deg2 in the SMC, ~20 deg2 in the Bridge and ~3 deg2 in the Stream.
With gas-phase metallicities of only 20–50% of the solar abundance, the SMC and LMC are the nearest laboratories to study star formation in low-metallicity environments (Russell & Dopita, 1992). This is important, because metallicity is a key parameter regulating the ISM: metals influence the efficiency of gas cooling, the shielding of molecules from the harsh UV radiation field, the production of dust grains, which in turn are the formation sites of molecules.
At a distance of ∼50 kpc (de Grijs et al., 2014) and aided by the LMC’s low inclination (i ∼ 38°, Balbinot et al., 2015) we can spatially resolve the finest details of typical less metal abundant interstellar environments present. We can thus unveil physical processes at sub-parsec scales, and quantify the relative contribution of the various ISM components/tracers to the integrated, unresolved ISM observations which are available for distant galaxies. For example, determining locally (in the Milky Way and other very nearby galaxies such as the SMC and LMC) which tracers best measure the amount of high density gas is crucial in estimating the star formation efficiency of any dense gas in large extragalactic samples (see also Section 2.3).
The hot and cold ISM of the LMC, both gas and dust, have been extensively observed with radio and (sub)mm facilities such as ASKAP, MeerKAT, ALMA, APEX, and SOFIA (see details in Table 2 and Table 3). In addition, both the LMC and SMC have been observed in the mid- and far-infrared by the Spitzer SAGE and Herschel HERITAGE surveys (Gordon et al., 2011; Meixner et al., 2006; Meixner et al., 2013). The currently most powerful near-infrared space telescope, JWST, observed warm, shocked, and/or ionized gas in a tiny part of the Large Magellanic Cloud. In the future, the Roman Space Telescope will be able to map the Magellanic Clouds with its much wider field of view. These near-infrared imagers and spectrographs will provide additional information on warm, shocked, and/or ionized gas. Combined, the synergy of these observations have led or will lead to significant new insights, but important questions about the diffuse and dense ISM remain open due to lacking a large-aperture single-dish in the (sub)mm wavelength regime.
Table 2. Observations of ISM tracers in the LMC (an incomplete collection).
| ISM phase | LMC |
|---|
| Atomic gas (HI) | ATCA+Parkes HI survey, beam~ 1′, area~10°×10°, RMS~2.4 K (Kim et al., 2003; Staveley-Smith et al., 2003) |
| GASKAP: an HI survey with the ASKAP telescope, beam~30″, area~250 deg2, RMS~2 K (Dempsey et al., 2022; Dickey et al., 2013; Pingel et al., 2022) |
| MagiKAT: a MeerKAT large survey of HI and OH, beam~20″, RMS~0.49 K per 1 km s−1 (van Loon et al., 2010) |
| Molecular gas (CO) | CO(1–0) at beam~ 8.8′ over 6° ×6° with Columbia 1.2-m Millimeter-Wave Telescope, RMS~0.06 K per 1.3
km s-1 (Cohen et al., 1988) |
| Various CO(1–0) observations in selected regions with the SEST 15-m, with beam~45″ at 115 GHz and
typical RMS~0.1–0.5 K per 0.1 km s-1 (Caldwell & Kutner, 1996; Israel et al., 1986; Israel et al., 2003;
Johansson et al., 1994; Johansson et al., 1998; Kutner et al., 1997) |
| NANTEN complete survey of CO(1–0) at beam~ 2.6′ with NANTEN 4-m, covering ~ 30 deg2, RMS~0.07 K per
0.65 km s−1 (Fukui et al., 1999; Fukui et al., 2008; Mizuno et al., 2001) |
| MAGMA survey of CO(1–0) at beam~ 33″ with Mopra 22-m, covering 3.6 deg2 RMS~0.24 K per 0.53 km s−1
(Hughes et al., 2010; Pineda et al., 2009; Wong et al., 2011; see Data Release URL) |
| APEX CO(2–1) at beam~27″, CO(3-2) at ~18″, CO(6-5) at ~9″, and isotopologue lines in 3–13 arcmin2 areas
(Okada et al., 2019) |
| APEX CO(3–2) at beam~ 20″ (~ 5 pc) over 13.8 deg2 RMS~150 mK per 0.5 km s-1 (Grishunin et al., 2024) |
| Herschel SPIRE high-J CO at beam~17–42″ over ~5′×5′ (Lee et al., 2016) |
| Radio continuum | ASKAP 888 MHz survey, beam~ 13′, area~ 120 deg2 , RMS~ 58 µ Jy beam−1 (ASKAP-EMU; Pennock et al., 2021) |
| PDR tracers ([CI]) | AST/RO in N159/N160 H ii regions at beam~ 3.8′, line RMS ~0.1 K km s-1 (Bolatto et al., 2000a) |
| APEX [CI](1–0) at beam~13″ and [CI](2–1) at ~8″ in 3–13 arcmin2 areas (Okada et al., 2019) |
| Herschel/HIFI at ~ 44″ for [CI](1–0) and ~ 26.6″ for [CI](2–1) along 36 lines of sight (Pineda et al., 2017) |
Ionized gas & PDR
([CII]) | Herschel/HIFI at ~ 12″ along 36 lines of sight (Pineda et al., 2017) |
| SOFIA/GREAT at ~ 16″ in 3–13 arcmin2 areas (Okada et al., 2019) |
| SOFIA observations over ~40' × 20' (60 pc × 30 pc) area (Madden & LMC+ Consortium, 2023) |
Ionized gas & H ii
regions (Hα) | Hα maps with ≲ 5′′ resolution covering 8° × 8° in the LMC (MCELS; Smith & MCELS Team, 1998; Smith & MCELS Team, 1999; Winkler et al., 2015) |
| Hα maps with 1° resolution covering ~25° × 25° area centered at the LMC (Wisconsin Hα Mapper, WHAM;
Smart et al., 2023) |
| SDSS-V Local Volume Mapper (LVM) survey of optical emission lines, area~78.5 deg2, spatial resolution~10
pc (fiber diameter 0.5°) (Drory et al., 2024) |
| Dust (IR/submm) | Spitzer SAGE (Meixner et al., 2006): all 7 IRAC and MIPS bands |
| Herschel HERITAGE (Meixner et al., 2013): beam~ 7″–36″, area~ 7° × 7°, RMS~0.2–9 MJy sr−1 (see Data Release URL) |
| QUEST maps from Clark et al. (2021) that includes the diffuse dust from a combination with Planck |
| high resolution dust reddening at 25″ resolution from Chen et al., 2022 |
| South Pole Telescope (and Planck) at 1.4, 2.1, and 3.0 mm, beam~ 1.0–1.7′, (5–10′ for Planck data), area~
7.5° × 7.5° (Crawford et al., 2016) |
| SOFIA/HAWC+ dust continuum and polarization at 53–214 µm with beam =5–18″ (Tram et al., 2021;
Tram et al., 2023) |
Table 3. Observations of ISM tracers in the SMC (an incomplete collection).
| ISM phase | SMC |
|---|
| Atomic gas (HI) | ATCA+Parkes HI survey, beam~ 1.6′ (30 pc), RMS~2 K (Stanimirovic et al., 1999) |
| GASKAP: an HI survey with the ASKAP telescope, beam~30″, area~250 deg2, RMS~2 K (Dickey et al., 2013; Dempsey et al., 2022;
Pingel et al., 2022) |
| MagiKAT: a MeerKAT large survey of HI and OH, beam~20″, RMS~0.49 K per 1 km s-1 (van Loon et al., 2010) |
| HI outflows with ASKAP at ~ 30″ (Dempsey et al., 2020; McClure-Griffiths et al., 2018) |
| Molecular gas (CO) | CO(1–0) at ~ 8.8′ (~ 160 pc) with Columbia 1.2-m Millimeter-Wave Telescope covering 2°×2°, RMS~0.012 K
per 1.3 km s-1 (Rubio et al., 1991) |
| Various CO(1–0) and (2–1) observations in selected regions with the SEST 15-m, with beam~45″/22″ at
115/230 GHz and typical RMS~ 0.1–0.2K per ~ 0.1kms−1 (Israel et al., 1986; Israel et al., 1993; Rubio et al., 1993; Rubio et al., 1996) |
| CO(1–0) at ~ 42″ with the Mopra 22-m, RMS∼210 mK per 0.9 km s−1 (Muller et al., 2010) |
| CO(1–0), 13CO(1–0) and C18O(1–0) at ~ 33″ with Mopra 22-m along 18 lines of sight, plus CO(3–2) and
13CO(3–2) at ~ 17.5″ with APEX 12-m (Pineda et al., 2017) |
| CO(2–1) at ~ 27″ (~ 9 pc) in two regions (10′×10′and 7′×10′) with APEX, RMS∼ 32–37 mK per 0.25 km s−1 (Di Teodoro et al., 2019) |
| CO(2–1) at ~ 27″ (~ 9 pc) in discrete regions totaling ~ 0.4 deg2 (0.47 kpc2) region with APEX, RMS∼ 0.1–0.7
K per 0.25 km s−1 (Saldaño et al., 2023a) |
| CO(3–2) at ~ 20″ (~ 6 pc) in ~ 2.5 deg2 region with APEX, RMS~ 0.1–0.7 K per 0.25 km s−1 (Saldaño et al., 2023b) |
| Radio | ASKAP 960/1320 MHz survey, beam~30″/16″, area~30 deg2, RMS~186/165 µJy beam−1 (ASKAP-EMU; Joseph et al., 2019) |
| MeerKAT 1.3 GHz survey, beam~ 13″, area~ 7°×7°, RMS~ 11 µJy beam−1 (Cotton et al., 2024) |
| PDR tracers ([CI]) | ISO satellite at 3.8′ (Bolatto et al., 2000b) |
| Herschel/HIFI at 44″ for [CI](1–0) and 26.6″ for [CI](2–1) along 18 lines of sight (Pineda et al., 2017) |
Ionized gas & PDR
([CII]) | Herschel/HIFI at 12″ along 18 lines of sight (Pineda et al., 2017) |
Ionized gas & H ii
regions (Hα) | H α with ≲ 5′′ resolution covering 4.5° × 3.5° in the SMC maps (MCELS; Smith & MCELS Team, 1998; Smith & MCELS Team, 1999;
Winkler et al., 2015) |
| Hα maps with 1° resolution covering ∼20° × 20° area centered at the SMC (Wisconsin Hα Mapper, WHAM;
Smart et al., 2019) |
| SDSS-V Local Volume Mapper (LVM) survey of optical emission lines, area~16.4 deg2, spatial resolution~10
pc (fiber diameter 0.5°) (Drory et al., 2024) |
| MUSE observation of optical emission lines, area~ 1′ × 1′ , resolution~ 1.3″ (Bodensteiner et al., 2020) |
| Dust (IR/submm) | Spitzer SAGE (Meixner et al., 2006): all 7 IRAC and MIPS bands |
| Herschel HERITAGE (Meixner et al., 2013): beam~ 7″–36″, area~ 7° × 7°, RMS~0.2–9 MJy sr−1 (see Data Release URL) |
| South Pole Telescope (and Planck) at 1.4, 2.1, and 3.0 mm, beam∼ 1.0–1.7′, (5–10′ for Planck data), area~ 5°
× 5° (Crawford et al., 2016) |
| QUEST maps from Clark et al. (2021) |
| AzTEC 1.1 mm continuum over 4.5 deg2 with ASTE 10-m (Takekoshi et al., 2017) |
How much molecular gas is there? These existing observing campaigns yield important constraints on the properties of the ISM in the LMC and SMC (see Figure 3). For example, Fukui et al. (2008) show that most of the molecular gas in the LMC is contained in over a hundred individual giant molecular clouds (GMCs; Hughes et al., 2010), with a mean molecular mass surface density ∼ 50M⊙ pc−2, making those clouds only half as dense as those in the inner Milky Way. At the distance of LMC, resolving a few to tens of parsec-scale GMCs would need an angular resolution ≲4″. On the other hand, observations of the [CII] line emission in the LMC with SOFIA/GREAT suggest that ≳ 75% of the cold gas may be “CO-dark”, that is, missed from observing the tranditional tracer CO lines (Chevance et al., 2020a). Accurately quantifying the molecular gas contents is of foremost importance, as a significant open question in the field is whether the ability of an individual GMC to form stars (i.e. its star formation efficiency) is determined by the intrinsic properties of the cloud, or dependent on interstellar environments (e.g. Chevance et al., 2020b; Girichidis et al., 2020; Krumholz & McKee, 2005). The first step to answer this fundamental question is therefore to characterise the molecular gas content (both CO-dark and CO-bright) of GMCs of various properties in a broad range of environments characterised by different levels of large-scale turbulence, pressure, external shear, and magnetic field strengths (e.g. André, 2017; Pattle et al., 2023).
Figure 3.
Bottom four panels: Stellar light, Hi, and CO emission in the LMC. Panels from left to right show the optical RGB image from the ESA Sky; Hi emission in gray with labels of regions from Ochsendorf et al. (2017); CO(1–0) emission (green) from the NANTEN telescope (at 2.6′ ∼ 40 pc resolution; Fukui et al., 2008) on top of Hi column density gray image (at ∼ 14′ ∼ 200 pc resolution; Kim et al., 2003) adapted from Harada et al. (2019); and the ALMA small mosaic observations at a few locations highlighted in green and red, adapted from Sawada et al. (2018). Top-left panel shows the JWST NIRCam RGB image of the Tarantula Nebula star-forming region at ultra-high resolution (0.06′′ ∼ 3000 AU), adapted from NASA website. Top-right five panels show the ALMA observations of the 12CO(1–0) emission in five GMCs in the LMC (field of view 2.5′ × 2.5′ each) at sub-parsec scales (at 3′′ ∼ 0.7pc resolution) from Sawada et al. (2018).
How much dust is there? Using common methods and calibrations, it can be demonstrated that the interstellar dust masses calculated for the SMC are too large to be explained by the dust production from evolved stars The infrared to submm emission of Magellanic Clouds have been observed either globally or point by point and modelled with simple phenomenological (Gordon et al., 2014; Leroy et al., 2007) or full state of the art dust models (Bot et al., 2010; Chastenet et al., 2017). These studies have inferred quite large dust masses of ∼ 2×104 −105 M⊙ for the SMC and ∼ 4×105 –3×106 M⊙ for the LMC, compared to the known dust production from evolved stars (1.3×10−6 M⊙yr−1; Schneider et al., 2014; Srinivasan et al., 2016), assuming that dust destruction and star formation are non-negligible sinks of interstellar dust in the ISM. Similarly, excessively large dust masses are also observed in high redshift (and therefore relatively low metallicity) galaxies (e.g., Dwek & Cherchneff, 2011; Donevski et al., 2020; Rowlands et al., 2014), with these observations particularly hard to explain at z ≳ 6 (e.g., Witstok et al., 2023), when the Universe is too young for Asymptotic Giant Branch stars to have evolved to the dust forming stage.
However, recent work on dust in the LMC, SMC and other nearby resolved galaxies suggests that this “dust budget crisis” might be attributed to our lack of understanding of cosmic dust in low density, low metallicity environments and wrongly applying assumptions based on observations of dust in high metallicity environments. Indeed, submm and mm emission excesses have been observed in nearby galaxies, especially mm excesses are found at low metallicities and/or in diffuse environments such as the ISM of the LMC (Galliano et al., 2011), and even more prominently in the SMC (Bot et al., 2010; Gordon et al., 2014; Planck Collaboration et al., 2011). The excess refers to the emission observed in the (sub)mm regime that is higher than what is extrapolated by single modified black-body emission or full dust models fitted on the far-infrared peak (Bot et al., 2010; Galliano et al., 2011; Gordon et al., 2011). It can be either due to an excess above standard emissivity index (β = [1.5–2]) in the shorter wavelengths, or a flattening (decreasing β index) or break in the spectral energy distribution toward longer wavelengths. The excess is also often associated to a lower dust mass surface density, and low metallicities or dwarf galaxies (Galametz et al., 2014; Galliano et al., 2018; Kirkpatrick et al., 2013). In the Magellanic Clouds, an excess has been observed at even longer wavelengths in the mm to centimeter range, above the expected dust emission fitted to the far-infrared to submm energy range (Bot et al., 2010; Israel et al., 2010; Planck Collaboration et al., 2011).
This (sub)mm dust emission excess can be interpreted as an indication of a different dust composition, grain properties or emissivity at longer wavelengths in low metallicity environments (Galliano et al., 2018), or the emission from magnetic nanoparticles (Draine & Hensley, 2013). Furthermore, Clark et al. (2021) have shown that in such large galaxies, Herschel observations suffered from spatial filtering on large scales, compounding the problem. After correcting for this missing large scale, diffuse emission by adding data from the Planck satellite, part of the tensions between observations and models however appear to be resolved (Clark et al., 2023). This highlights the vital importance of getting submm and mm observations at an angular resolution (of a few arcseconds) that resolves individual environments and that also account for diffuse, large-scale dust emission, if we are to characterise the properties of interstellar dust in low metallicity environments.
2.1.2 The role of AtLAST in the study of the LMC and SMC. Interferometers like ALMA and current single dish submillimeter telescopes like APEX are very inefficient at mapping large-scale, low-density interstellar medium at multiple bands, either due to filtering of flux on large scales, poor sensitivity, poor spatial resolution, or low mapping efficiency3. The characterization of the diffuse molecular and atomic ISM, where GMCs are born and remain embedded, is therefore impossible without a next-generation (sub)millimeter single dish telescope such as AtLAST. We give an overview here of how AtLAST will tackle the core questions raised above regarding the ISM in the LMC/SMC, together with a description of the corresponding technical requirements in Section 2.1.3.
What are the properties of interstellar dust in the diffuse and dense ISM? So far, studies of the distribution and properties of dust in the LMC/SMC have relied heavily on the full maps observed with Herschel at wavelengths of ∼ 70–500 µm (spatial resolution ∼ 7–36″; see summary in Table 2 and Table 3). To expand on these results, higher spatial resolution (e.g., tens of parsecs cloud-scale) measurements are required to trace spatial variations in dust properties, being sensitive to large scale emission is required to properly trace the very diffuse dust component, and longer wavelength data are required to increase the precision to which we can measure the dust emissivity index (and as a consequence dust temperature) and the dust surface density. The advantages of such an approach are demonstrated for instance for M 31, which was targeted by the HASHTAG programme with SCUBA-2 on the JCMT at 450 and 850 µm (Smith et al., 2021). Unfortunately, it is currently not possible to make large scale continuum maps (7° × 7° for the LMC and 3° × 5° for the SMC) at submm or mm wavelengths. The exception is the survey performed with the South Pole Telescope, which was expanded to include the LMC and SMC at 1.4, 2.1 and 3 mm (Crawford et al., 2016). However, this survey has a spatial resolution of only 1.0′ to 1.7′ (∼ 400–800 pc), and contains contributions from free-free and synchrotron radiation complicating the measurement of thermal dust emission at these wavelengths.
With AtLAST, we will characterise the shape of the dust spectral energy distribution from the far-infrared to the millimeter in several bands (e.g., with enerby-balancing code like MAGPHYS (da Cunha et al., 2008), CIGALE (Boquien et al., 2019), and many others), from dense regions to diffuse environments. This will make it possible to separate the cold dust component from free-free emission, but also from background emission from distant galaxies or CMB fluctuations. Studying this emission in the two different metallicity regimes offered by the LMC and SMC will enable us to understand how dust varies with environment and hence to model it precisely and ultimately get better dust properties (including the most precise dust temperature, emissivity index β, and dust masses in LMC/SMC-like enviornments). The significantly (up to ∼ 15×) higher spatial resolution compared to Herschel will make it possible to trace substructures, and characterise the multi-temperature dusty medium, in principle in any nearby galaxies.
How can we trace dark molecular gas, and what is its importance? As metallicity decreases, CO is restricted to the highest density parts of the molecular clouds, leaving a substantial molecular layer of H2 without CO that has been dubbed "CO-dark" (e.g. Wolfire et al., 2010). This CO dark component is often conceptualised as the outer layer of giant molecular complexes, but in reality the geometry of the ISM is much more complicated. Given the filamentary structure of the ISM, only small CO-emitting clumps survive the photoionisation, making the whole CO-dark/CO-emitting components very difficult to detect without sensitive, high resolution and large field-of-view observations (as observed for example with ALMA in the low metallicity dwarf galaxy WLM, Rubio et al., 2015). Observations of the [CII] line emission in the LMC with SOFIA/GREAT suggest that ≳ 75% of the cold gas may be CO-dark (Chevance et al., 2020a). To reach our goal of understanding how star formation proceeds in low density, low metallicity environments, it is vital that we accurately resolve the molecular gas reservoirs of galaxies such as the SMC and LMC, and therefore we must look beyond CO for molecular ISM tracers.
The neutral carbon’s [CI](1–0) and (2–1) fine structure lines are popular tracers of CO-dark gas, proposed to be very promising for tracing the CO dissociation layer of PDRs and the low-metallicity gas without much CO (Glover & Clark, 2016; Ramambason et al., 2024). By observing both [CI] lines, it is furthermore possible to measure the temperature, density and optical depth of the gas. Significant mapping of the [CI] in the LMC and SMC is currently impossible: ALMA’s field of view is far too small4, and a 12m single dish like APEX does not offer enough sensitivity and resolution5. A survey of the LMC and SMCs with AtLAST in Bands 8 and 96 would therefore allow us to map not just the distribution of the diffuse molecular gas, but its physical conditions as well, at unprecedented sensitivity and resolution. This would be a game-changer in our understanding of the ISM and of star formation in diffuse, low metallicity extragalactic environments. For example, accurate mapping of the full molecular gas reservoir (CO and CO-dark) will allow us to answer the very important question of the extent to which star formation can occur in diffuse ISM environments.
How does metallicity drive the gas-to-dust ratio? Given the low metallicity ISM of especially the SMC, key physical and chemical processes that regulate cooling, heating, dust and molecule formation are different than in the Milky Way. Studies of nearby galaxies have shown that such processes do not scale linearly with the amount of metals available. For example, the gas-to-dust ratio is observed to evolve linearly as expected only at high metallicities. Below a critical metallicity threshold, the amount of dust per gas unit breaks down from the expected relationship (e.g. Galliano et al., 2018; Rémy-Ruyer et al., 2014; Roman-Duval et al., 2017). This critical metallicity of ∼ 0.2Z⊙ is usually interpreted as the threshold above which grain growth becomes efficient (Zhukovska et al., 2016). An alternative interpretation is that the break corresponds to a regime change where diffuse gas becomes more important (Clark et al., 2023). Straddling this metallicity transition region, the LMC/SMC and lower-metallicity, fainter dwarf galaxies will offer us a unique opportunity to accurately measure both dust and gas abundances to significantly improve our understanding of how these two components of the ISM interrelate. AtLAST will be essential in achieving these goals, as they can only be achieved through a complete census of the metals (e.g., with the optical integral-field Local Volume Mapper survey; Kollmeier et al., 2019), dust and gas (with AtLAST).
2.1.3 AtLAST surveys of the LMC and SMC. At declinations of ∼ −70 degrees, the LMC and SMC are only observable from the Southern hemisphere (see Figure 2). They have already been mapped by a wide range of radio and (sub)mm telescopes, as summarised in Table 2 and Table 3, but suffer from lack of resolution, confusion, and/or lack of sensitivity to large-scale flux. AtLAST will be transformational, by allowing full mapping of the LMC and SMC over nearly a hundred square degrees in continuum emission and moderately bright lines, scanning through the electromagnetic spectrum from Band 3 (∼ 100 GHz) to Band 10 (∼ 850 GHz) to determine accurate dust and gas properties. The best spatial resolution can be achieved in the highest-frequency Band 10, about 1.5–2 arcsec (0.4–0.6 pc).
Continuum observations With AtLAST we will be able to obtain multi-band submillimeter continuum maps with sufficient sensitivity and area coverage to trace emission from bright star forming regions to the faint, diffuse ISM. Coverage in multiple bands will be crucial: it is needed in the first place to derive the spectral shape of the dust emission to constrain the dust composition and grain sizes, and to discriminate between dust and free-free emission in regions where ionized gas is important. Due to the sensitivity and large field of view of AtLAST, "contamination" by Cosmic Microwave Background (CMB) fluctuations will be an issue, but with multi-band observations we will be able to disentangle them, given the well characterised spatial and spectral characteristics of the CMB. Similarly, we can expect to be affected by confusion with more distant sub-millimeter/millimeter galaxies, resolved or unresolved (Cosmic Infrared Background), which will create correlated noise that will have to be estimated, but can be separated from dust emission using its spatial scale properties and wavelet phase harmonics (Auclair et al., 2024).
Figure 4 and Figure 5 show the footprint of a proposed mapping survey of the LMC and SMC with AtLAST. Assuming a field of view of 1 deg2 for the first-generation AtLAST continuum camera, we will need ∼ 40 pointings to cover the LMC (∼ 7° × 7°) and ∼ 8 pointings to cover the SMC (∼ 4° × 2°). For each pointing, the integration time depends on the bandwidth and depth. Here we adopt a high-spatial-resolution multi-wavelength SMC dust simulation by C. Bot et al. (in prep.) to estimate the expected observing time. The simulation uses the dust parameters from Clark et al. (2023) to extrapolate the observed dust spectral energy from the Herschel HERITAGE survey (Meixner et al., 2013) combined with Planck on large scales (Clark et al., 2021) to AtLAST wavelengths and sampled at the AtLAST angular resolutions (Table 4). CMB fluctuations contributing on large scales were also added using the SMICA map from Planck data (Planck Collaboration et al., 2020). From these maps, we find that the detection limit can be set by the faintest dust structures in the SMC bar region, which has flux densities of the order of 0.01 MJy/sr at 100 GHz (Band 3), to 0.25 MJy/sr at 400 GHz (Band 8) and 0.5 MJy/sr at 680 GHz (Band 9). Given the angular resolution of AtLAST at these bands, we can derive the RMS to achieve at least 3σ per angular resolution unit: which is 1.5 µJy/beam, 22 µJy/beam and 19 µJy/beam at 100 GHz, 400 GHz and 680 GHz, respectively.
Figure 4. Dust emission at 100 µm in the LMC from the Herschel space observatory (Meixner et al., 2013) The left panel shows the full ∼ 7° × 7° area of LMC.
Grids are 1° × 1° fields of view of AtLAST continuum camera. The right panel shows the zoom-in 30′ view around the 30 Doradus Hii region. The fields of view of AtLAST, ALMA (12-m array and 7-m array), JWST NIRCam and Roman Space Telescope WFI are shown for comparison. Only the future-generation AtLAST and Roman are able to map the full near to far-infrared and sub-millimeter dust emission in the LMC.
Figure 5. Similar to Figure 4 but for the SMC.
Table 4. Angular resolution (AR) and maximum recoverable scale (MRS).
| frequency | Band | AtLAST 50m | ALMA 43×12m AR/MRS |
|---|
| | AR | C1 | C3 | C5 |
|---|
| (GHz) | | (arcsec) | (arcsec) | (arcsec) | (arcsec) |
|---|
| 100 | 3 (84–116) | 15.088 | 3.38/28.5 | 1.42/16.2 | 0.55/6.70 |
| 200 | 5 (163–211) | 7.544 | 1.69/14.3 | 0.71/8.10 | 0.28/3.35 |
| 300 | 7 (275–373) | 5.029 | 1.13/9.50 | 0.47/5.40 | 0.18/2.23 |
| 400 | 8 (385–500) | 3.772 | 0.84/7.12 | 0.35/4.05 | 0.14/1.68 |
| 500 | 8 (385–500) | 3.018 | 0.68/5.70 | 0.28/3.24 | 0.11/1.34 |
| 600 | 9 (602–720) | 2.515 | 0.56/4.75 | 0.24/2.70 | 0.092/1.12 |
| 700 | 9 (602–720) | 2.155 | 0.48/4.07 | 0.20/2.31 | 0.079/0.96 |
| 800 | 10 (787–950) | 1.886 | 0.42/3.56 | 0.18/2.02 | 0.069/0.84 |
| 900 | 10 (787–950) | 1.676 | 0.38/3.17 | 0.16/1.80 | 0.061/0.74 |
Using the AtLAST sensitivity calculator, with a bandwidth of 32 GHz, elevation 45 deg, and H2O percentiles of 70, 40 and 10 for the aforementioned bands, we obtain an on-source integration time of 27.0 h, 5.2 h, 28.5 h, respectively, for one pointing. The Band 8 observation (calculated at 400 GHz) has the best observing efficiency in terms of the significantly shorter integration time to achieve the detection limit. This is set by both the dust SED shape and the angular resolution at that frequency. Other bands are more time-consuming, but are still crucial to constrain the shape of dust SED at each position. To map the whole LMC and SMC, we need to multiply the above times by the number of pointings. Thus the total amount of on-source time will be about ∼ 2900 h. Spreading the observing time into multiple years, and in each year to obtain a full mapping of LMC and SMC, will be a wise option to ease the scheduling pressure and also to allow for transient studies (see companion case study by Orlowski-Scherer et al., 2024).
This AtLAST LMC/SMC survey will eventually provide multi-band (e.g., 3 bands, from Band 3 ∼ 100 GHz to Band 9 ∼ 700 GHz) deep dust continuum mapping, with angular resolutions ≳ 10× better than Herschel and ≳ 4× better than APEX, with unprecedented sensitivity to diffuse emission that ALMA will completely miss, and with important frequency and spatial coverages not covered by Herschel, ALMA, APEX, LMT, etc.
Line observations The molecular gas in the LMC has (in an area-covering way) so far only been observed in CO main isotopologue lines, e.g., in the CO(1-0) line by Fukui et al. (1999, 2008) and Kawamura et al. (2009) with the NANTEN 4-m telescope at 40 pc (2.6′), and further at 11 pc (30′′) by Wong et al. (2011) with the Mopra 22-m telescope for selected small regions (3.6 deg2 in total). An on-going large survey with the APEX 12-m telescope targets the 12CO(3–2) and 13CO(3–2) lines at 5 pc resolution, aiming to cover 17.4 deg2 or 13.3 kpc2 (Grishunin et al., 2024; Weiß et al., 2023). Deeper and/or higher spatial resolution observations of CO and other tracers remain limited to individual pointings or small specific regions (e.g., Sawada et al., 2018; Wong et al., 2017; Wong et al., 2022). The lack of contiguous, sensitive, high-resolution multi-line coverage of at least significant portions of the LMC and SMC, in order to sample the full range of environments they contain, is still a very important obstacle preventing us from understanding the physical and chemical structure of our neighbour low-metallicity galaxies.
We here outline prospects (and limitations) for possible line mapping surveys of the LMC and SMC with AtLAST. First, we describe a CO(3–2) survey that would correspond to the APEX 12CO(3–2)/13CO(3–2) survey of Weiß et al. (2023) and Grishunin et al. (2024), but 4× higher spatial resolution and an order of magnitude higher sensitivity. Such a survey would reveal CO-bright molecular gas in a much more complete way (including intrinsically faint structures as well as small features suffering from beam dilution if observed with poorer spatial resolution). It also serves as an illustration as to which extent extensive mapping of correspondingly fainter lines from other tracers will be possible. Second, we describe a survey in the neutral carbon [CI] lines at 492 and 809 GHz, which will be key in understanding the contribution of "CO-dark" molecular gas and atomic gas at the transition between the diffuse ionized ISM and molecular clouds. For the [CI] surveys we set the target sensitivity to be able to achieve 4-5 σ detections of a 0.1 K line brightness, as the [CI] lines detected by Pineda et al. (2017) and Bolatto et al. (2000a) are at levels of a few times 0.1 K (with line-widths of typically a few km/s).
Table 5 summarizes our example LMC/SMC line surveys in terms of the time that would be required to cover one square degree. For the "APEX-like" CO(3–2) survey, AtLAST (if equipped with a large multi-pixel heterodyne array) will provide mapping speeds comparable with contemporary facilities due to the combined effect of spatial resolution and field of view, but the resulting data will have an order of magnitude improvement in sensitivity and a factor of several in spatial resolution. Full cloud scale maps in a number of lines significantly fainter than the main CO lines will be possible, on scales ranging from the full galaxies down to about one parsec, i.e., the scales of clumps within clouds that are thought to form clusters or groups of stars.
Table 5. Telescope time needed to cover one square degree for the APEX CO(3-2) survey of Weiß et al., (2023) (Weiß2023) using the 7-pixel LASMA receiver array, a corresponding deeper AtLAST survey at similar frequencies (AtLAST 13CO), and higher frequency AtLAST [CI] surveys (assuming 1000 pixel heterodyne arrays on a 50 m diameter AtLAST).
AtLAST telescope times were estimated using the AtLAST sensitivity calculator, assuming an elevation of 40° (accounting for the low declination of the LMC/SMC).
| Survey | frequency
(GHz) | resolution
spatial/velocity | sensitivity
1 σ | pwv H2O
percentile | time/
square degree(a) |
|---|
| Weiß2023 | 345.8, 330.6 | 19″/0.5 km/s | 200 mK | | 36.7 h |
| AtLAST 13CO | 330.6 | 4.6″/0.5 km/s | 20 mK/38 mJy | 40 | ~ 40 h |
| AtLAST [CI] 3P1–3P0 | 492.16 | 3″/1.0 km/s | 22 mK/40 mJy | 20 | ~260 h |
| | | | 10 | ~150 h |
| | | | 40 | ~1000 h |
| AtLAST [CI] 3P2–3P1 | 809.34 | 1.8″/1.0 km/s | 22 mK/40 mJy | 10 | ~1500 h |
At higher frequencies, aiming for lines such as the [CI] lines or higher-J CO lines, the situation is more challenging. It will still be possible to cover significant area at sufficient sensitivity in the brightest lines (particularly if observations can be spread over several observing periods) at moderately high frequencies (e.g., [CI] 3P1–3P0 in ALMA Band 8). At even higher frequencies (ALMA Bands 9 and 10) observations will be restricted to partial coverage. In terms of [CI] 3P2–3P1 observations (required to constrain the excitation conditions of the gas), we will likely need to adopt the observing strategy of targeting only regions where the 3P1–3P0 line has been detected.
Table 5 also shows the time estimate for the [CI] 3P1–3P0 492 GHz line for three different values of the H2O pwv percentile. While at the 10 and 20 percentile (corresponding to about 0.5 and 0.8 mm pwv, respectively, according to the documentation of the ETC) significant (multisquare degree) coverage is still feasible, this clearly becomes prohibitive at the 40 percentile level (corresponding to about 1.5 mm pwv). Going to a high and dry site, as AtLAST will, is absolutely mandatory if any significant sensitive coverage at these frequencies is to be achieved.
2.2 Extragalactic magnetic fields
2.2.1 Context and open questions. Cosmic magnetic fields are one of the most elusive forces in our Universe, even though their importance has become increasingly clear in all areas of astrophysics. Magnetic fields are one of the main regulators of local star formation efficiency (Krumholz & Federrath, 2019) and play a crucial role in the ISM dynamics in galaxies (Beck, 2015). They affect the transport of ionized gas, dust, and metals outward to the circumgalactic medium (CGM) via galactic outflows (Lopez-Rodriguez, 2023; Lopez-Rodriguez et al., 2021a; Lopez-Rodriguez et al., 2023; Thompson et al., 2006), and they are central to the dynamics of cosmic rays (Rodríguez Montero et al., 2023). B-fields are a crucial component in theoretical galaxy formation (e.g., Hennebelle & Iffrig, 2014; Hopkins et al., 2023; Martin-Alvarez et al., 2020; Moss et al., 2000), with the new generation of numerical magnetohydrodynamical (MHD) simulations (e.g. Wissing et al. (2022)) posing magnetic fields and cosmic rays as alleviating forces for challenges such as matching the halo mass-stellar mass or mass-metallicity relations.
Magnetic fields in galaxies are thought to have originated as a weak seed during the early stages of the Universe (Subramanian, 2019; Widrow, 2002). These magnetic fields have to be amplified during galaxy formation to explain the observed ∼ µG strengths in present-day galaxies (Beck et al., 2020). Turbulence-driven dynamos are expected to be responsible for the generation of most of the magnetic energy required to reach these magnetic field strengths (Subramanian, 2019). Dynamos convert kinetic energy into magnetic energy. Mergers within the cosmic web trigger considerable star formation activity through starbursts. During these starburst phases, the magnetic fields of galaxies can be further enhanced via turbulent dynamos driven by stellar feedback such as supernova (SN) explosions. This generates tangled B-fields at and below the turbulent scale of approximately < 50 − 100 pc for SN feedback (Brandenburg & Subramanian, 2005; Haverkorn et al., 2008). All of this leads to a theoretical scaling of Btot ∝ SFR1/3 in models of a galaxy dominated by SN-driven turbulence (Schober et al., 2013). Thus, an amplification of the B-fields from the early universe to present-day galaxies is expected to be driven by star formation activity.
Most of our knowledge about magnetic fields in galaxies has been established using radio polarimetric observations in the 3–20 cm wavelength range (Beck, 2013). These observations trace the magnetic fields via synchrotron emission and Faraday rotation measurement arising from high-energy particles passing through a magneto-ionic ISM in the warm and diffuse phase of the ISM (Martin-Alvarez et al., 2023). This method is widely used to measure the magnetic field strengths and directions in our Milky Way as well as nearby galaxies. Magnetic field strengths are estimated assuming equipartition between magnetic energy and cosmic ray electrons. Measuring the Rotation Measure (RM) requires at least three frequency bands with large frequency separations to accurately estimate the magnetic field direction along the line-of-sight (LOS). These remarkable efforts have shown that all spiral galaxies have kpc-scale ordered magnetic fields with an average strength of 5±2 µG (Beck et al., 2020). The total magnetic field strength, measured by synchrotron total intensity and assuming equipartition between the total B-field and total cosmic-ray electron density, is estimated to be ∼17±14 µG (Beck et al., 2020; Fletcher, 2010). However, these measurements have been performed at low resolution (>15′′), which corresponds to a spatial scale of ∼ 300 pc at 20 Mpc, and tracing the warm ISM at vertical heights > 200 pc from the galaxy’s midplane, missing the information about the small-scale magnetic field closely related to the star formation activity. The SKAO will provide sub-arcsecond resolutions, thus enabling complementary high-resolution L-/C-band radio polarization studies, although the small-scale Faraday-rotation and Faraday-depolarization due to foreground clouds of diffuse thermal gas and magnetic fields can be very challenging to account for, for instance, the so-called “canals” in the maps of polarized intensity (e.g., Haverkorn & Heitsch, 2004; Haverkorn et al., 2004).
Far-infrared/submillimeter polarimetry, on the other hand, observes the polarized thermal emission from magnetically aligned dust grains associated with the cold and dense phase of the ISM in galaxies. Elongated dust grains irradiated by starlight spin along the axis of their greatest moment of of inertia (i.e., the minor axis of the dust grain) producing a magnetic moment that aligns with the local magnetic field in the ISM. Thus, polarized thermal emission produce a measurable polarization with their position angle perpendicular to the local B-field. This technique has successfully traced the magnetic fields in more than a dozen nearby (<20 Mpc) galaxies (i.e., spirals, starbursts, dwarfs, mergers, active galactic nuclei) in the 53–250 µm using SOFIA/HAWC+ at resolutions of 4.8-18′′ (90 pc - 1 kpc) (see most of the latest results from the SOFIA SALSA large survey; Borlaff et al., 2023; Lopez-Rodriguez et al., 2022b), as well as in starburst galaxies, M82 at 850 µm with the JCMT at a resolution of 15′′ (277 pc) (Pattle et al., 2021) and NGC 253 at 860 µm with ALMA at a resolution of ∼ 0.3′′ (5 pc) (Lopez-Rodriguez, 2023). FIR polarimetric observations of galaxies trace the magnetic fields as a density-weighted average within the beam size sensitive to the densest gas in the cold neutral phase of the ISM (Martin-Alvarez et al., 2023). Specifically, SOFIA/HAWC+ observations trace FIR magnetic fields associated with dense, log10(NH[cm−2]) = 19 − 23, and cold, Td = 19 − 48 K, dusty, turbulent star-forming regions (those traced at FIR) and these are less ordered than those in the warmer, less dense interstellar medium (those traced at radio) (Borlaff et al., 2023; Lopez-Rodriguez et al., 2022b).
FIR polarimetric observations show that all spiral galaxies have kpc-scale magnetic fields spatially correlated with the molecular gas and with large angular dispersion (i.e., turbulent magnetic field) arising from the star-forming regions in the spiral arms (Borlaff et al., 2021; Lopez-Rodriguez, 2021; Lopez-Rodriguez et al., 2023; see examples in Figure 6). SOFIA/HAWC+ 154 µm observations of the Antennae, a merger system of two spiral galaxies, showed a circumgalactic medium permeated with a 9 kpc scale ordered magnetic field connecting both pair members (Lopez-Rodriguez et al., 2023). This large-scale ordered magnetic field is not detected at radio wavelengths due to Faraday depolarization and the star formation activity in this region. In addition, FIR polarimetric observations are sensitive to the magnetic fields in the cold phase of the galactic outflows in starburst galaxies (Lopez-Rodriguez et al., 2021a; Lopez-Rodriguez, 2023). Note that radio polarimetric observations suffer from the short lifetime of CR and Faraday rotation in the galactic outflows challenging the characterization of magnetic fields in starburst galaxies (Adebahr et al., 2017). For M82, the magnetic field strength in the cold phase of the galactic outflow was estimated to be 305 ± 15 µG at a resolution of 90 pc (4.8”) at 53 µm. This magnetic field was estimated to be ‘open’ into the CGM enriching it with dust metals and astrophysical magnetic fields amplified by supernovae explosions. The magnetic field strength was estimated using a modified version of the Davis-Chandrasekhar-Fermi (DCF) method (Chandrasekhar & Fermi, 1953; Davis & Greenstein, 1951). The original DCF was developed to measure the magnetic field strength for a steady state and uncompressible medium using the measurements of the gas velocity dispersion and the dispersion of the polarization angles. The modified DCF method includes large-scale flows and shearing to account of the several components in the galactic disk and galactic outflow (Guerra et al., 2023; Lopez-Rodriguez et al., 2021a). However, the low resolution observations achieved by SOFIA and JCMT impede the quantification of the magnetic energy in galaxies because the turbulent coherence length of the magnetic field is not resolved. Therefore, the energy budget in the ISM of host galaxies responsible for the gas dynamics and star formation history is still unknown, and the theoretical scaling of Btot ∝ SFR1/3 is still unconfirmed.
The FIR/submm polarized spectrum provides unique information about the properties of dust and physical conditions of the star-forming regions. In the Milky Way, the polarization spectrum of star forming regions has been observed to be wavelength dependent in the FIR/submm regime (e.g., Hildebrand et al., 1999; Michail et al., 2021; Vaillancourt et al., 2008). This dependency is thought to arise from temperature gradients where only the warm silicate dust grains with high emissivity index embedded in the star-forming regions are aligned, while a colder component mainly following a black body emission is not aligned. Although dust models (e.g., Guillet et al., 2018; Hensley & Draine, 2023) have been able to explain the polarized spectrum in the diffuse ISM of the Milky Way, these models do not reproduce the wavelength dependence of the polarized spectrum in regions of strong radiation fields in the Milky Way or nearby starburst galaxies. The diffuse ISM is characterized by a roughly flat polarized spectrum with a homogeneous mix of dust components. However, the strong radiation fields in star-forming regions can change the dust composition, dust grain sizes, and dust temperatures. In nearby starbursts, the unresolved gradient of temperatures and tangled magnetic field along the LOS makes the interpretation of the polarized spectrum very complex (Lopez-Rodriguez, 2023). However, the shape of the polarized spectrum has been shown useful to distinguish between disk- and starburst-dominated galaxies for unresolved polarimetric observations, and to characterize the dust composition on the cold phase of the galactic outflows. multi-wavelength polarimetric observations are required to characterize the dust composition in the ISM of galaxies.
Thus, despite all these efforts so-far, we still do not fully understand how the measured B-fields in galaxies originated, are amplified, or are affected by star formation activity in galaxies. This is mainly due to the fact that radio wavelengths trace the warm ISM in a volume-filling medium and suffer from Faraday rotation and short lifetimes of cosmic rays in dense regions. The B-fields in the cold and dense ISM, where most of the mass and star formation reside, have been overlooked.
Even when SKA becomes available to map a large number of nearby galaxies’ B-fields at (sub)arcsecond scales, we will still lack the crucial high-resolution far-infrared/submillimeter view of the magnetic fields. The ALMA interferometer can help us to achieve subarcsecond-resolution dust polarization imaging over square-arcminute scales, however, it is very inefficient in surveying a large number of galaxies, and more importantly, missing substantial diffuse ISM where magnetic fields play a crucial role. A complete view of magnetic fields in nearby galaxies, as a benchmark to understand distant Universe as well as ISM and MHD physics, is something AtLAST can uniquely deliver.
2.2.2 The role of AtLAST in the study of extragalactic magnetic fields. The concept for the AtLAST 50-m single dish submillimeter telescope (Mroczkowski et al., 2024) has transformational capabilities for characterizing and understanding magnetic fields in our local Universe. First is its unprece dented survey speed for a large number (∼ hundreds) of galaxies thanks to the large field of view. Nearby (< 20 Mpc) galaxies are typically tens of arcmin in diameter (except for the LMC/SMC and M31), which can be rapidly covered by AtLAST using a single pointing (∼ 1 deg2). Note that ALMA requires multiple pointings to cover a single galaxy making it very inefficient and even for the brightest galaxies, unfeasible (see Table 6). Second is its superb antenna collecting area and sensitivity to map the faint, diffuse dust emission in nearby galaxies and their circumgalactic media. This extended and diffuse emission is filtered out by interferometers. Third is its capability of observing the high-frequency multi-bands achieving 1.5–3′′ resolution to resolve nearby (<20 Mpc) galaxies at giant molecular cloud scales (20–300 pc). These resolutions perfectly match existing and future large surveys of nearby galaxies at cloud scales (e.g., the PHANGS large program; Leroy et al., 2021a; the EDGE-CALIFA survey; Bolatto et al., 2017, etc.), which allows a comprehensive study of the multi-phase ISM and correlates the magnetic field with the gas flows in galaxies7.
Table 6. Common extragalactic molecular/atomic lines.
| Band | species & transition | rest-frame frequency | resolution |
|---|
| Band 3 | HCN J = 1–0 | 88.632 | 17.0″ |
| HCO+ J = 1–0 | 89.189 | 16.9″ |
| HNC J = 1–0 | 90.664 | 16.6″ |
| N2H+ J = 1–0 | 93.174 | 16.2″ |
| CS J = 2–1 | 97.981 | 15.4″ |
| C18O J = 1–0 | 109.782 | 13.7″ |
| 13CO J = 1–0 | 110.201 | 13.7″ |
| C17O J = 1–0 | 112.359 | 13.4″ |
| CN J = 3/2–1/2 | 113.491 | 13.3″ |
| CO J = 1–0 | 115.271 | 13.1″ |
| Band 4 | H2CO 21,2–11,1 | 140.840 | 10.7″ |
| CS J = 3–2 | 146.969 | 10.3″ |
| Band 5 | HCN J = 2–1 | 177.261 | 8.5″ |
| HCO+ J = 2–1 | 178.375 | 8.5″ |
| HNC J = 2–1 | 181.325 | 8.3″ |
| N2H+ J = 2–1 | 186.345 | 8.1″ |
| CS J = 4–3 | 195.954 | 7.7″ |
| Band 6 | C18O J = 2–1 | 219.560 | 6.9″ |
| 13CO J = 2–1 | 220.399 | 6.8″ |
| C17O J = 2–1 | 224.714 | 6.7″ |
| CO J = 2–1 | 230.538 | 6.5″ |
| CS J = 5–4 | 244.935 | 6.2″ |
| HCN J = 3–2 | 265.886 | 5.7″ |
| HCO+ J = 3–2 | 267.558 | 5.6″ |
| HNC J = 3–2 | 271.981 | 5.5″ |
| N2H+ J = 3–2 | 279.512 | 5.4″ |
| Band 7 | CS J = 6–5 | 293.912 | 5.1″ |
| C18O J = 3–2 | 329.331 | 4.6″ |
| 13CO J = 3–2 | 330.588 | 4.6″ |
| C17O J = 3–2 | 337.061 | 4.5″ |
| CS J = 7–6 | 342.883 | 4.4″ |
| CO J = 3–2 | 345.796 | 4.4″ |
| HCN J = 4–3 | 354.505 | 4.3″ |
| HCO+ J = 4–3 | 356.734 | 4.2″ |
| HNC J = 4–3 | 362.630 | 4.2″ |
| N2H+ J = 4–3 | 372.673 | 4.0″ |
| Band 8 | CO J = 4–3 | 461.041 | 3.3″ |
| [CI] 3P1–3P0 | 492.161 | 3.1″ |
| (fainter lines omitted hereafter) |
| Band 9 | CO J = 6–5 | 691.473 | 2.2″ |
| Band 10 | CO J = 7–6 | 806.652 | 1.9″ |
| [CI] 3P2–3P1 | 809.342 | 1.9″ |
Taking full advantage of AtLAST in observing nearby galaxies at a few arcsecond resolution, the following key science questions will be eventually addressed (to name a few):
How do magnetic fields affect star formation at cloud scales?
Does the star formation efficiency correlate with the magnetic field strength?
Is the magnetic field the support of supersonic turbulence in the ISM?
What determines the diversity and evolution of magnetic field strengths among galaxies?
2.2.3 AtLAST surveys to map magnetic fields in nearby galaxies. In addition to their immediate scientific objectives, polarization studies will simultaneously provide very deep and multi-wavelength dust continuum observations. For example, to achieve an uncertainty of 0.3% in the degree of polarization a signal-to-noise ratio of ∼ 500 in the total intensity is required. Therefore through polarimetric observations, deep dust continuum maps will be provided to the community to perform more general studies of dust emission and dust grain properties.
To enable significant statistics, a survey of dust polarization in over a hundred nearby galaxies with AtLAST 50m is needed. Note that each galaxy will be resolved at scales of 20 − 300 pc providing thousands of statistically independent polarization measurements that will be used to study trends between the magnetic field structure and strength with the star formation activity and gas kinematics in the disk of galaxies.
We assume a conservative polarization of 1% by means of magnetically aligned dust grains in resolved polarimetric observations of nearby galaxies, based on the thermal dust polarization in spiral galaxies at 53–220 µm (Lopez-Rodriguez et al., 2022a) and that the polarization is roughly constant in the disk of galaxies in the 50–800 µm wavelength range (Lopez-Rodriguez, 2023). Star-forming regions show a wavelength dependence in the polarized spectrum in the FIR/submm regime with a minimum polarization of 1% in the 100 − 200µm range and increasing to ∼ 4% at shorter and longer wavelengths. The shape of the polarized spectrum is still under active study as it provides unique information about dust properties and composition. Thus, in the following, we estimate the required observing time to map ∼ 100 nearby galaxies at the fiducial frequencies of 230 GHz (Band 6), 345 GHz (Band 7), 460 GHz (Band 8), 690 GHz (Band 9) and 806 GHz (Band 10) with AtLAST 50-m.
To estimate the sensitivity requirements, we use as a reference the JCMT SCUBA-2 HASHTAG/DOWSING large programs which reach a rms noise level of ∼ 2.0 mJy/beam (0.41 MJy/sr) with a 13.5′′ beam at 850 µm (Smith et al., 2021). Given the large SCUBA-2 beam,it is expected that the dust emission filling factor is ≪ 1, resulting into a flux density that would be much higher when measured with the smaller AtLAST beam. We use the PHANGS-ALMA CO(2–1) 1′′ imaging data (Leroy et al., 2021b) of a typical spiral galaxy NGC 4321 (M 100) to estimate the beam filling factor. We smooth the CO moment-0 image to the 3.3′′ AtLAST beam and to the 13.5′′ SCUBA-2 beam, respectively, then examine the emission at the outer disk and find a beam filling factor of 0.2. Furthermore, considering a greybody dust spectral energy distribution of S/S850µm ∝ , we can estimate an SCUBA-2-equivalent RMS 5.7 MJy/sr (∼ 4.5 mJy/beam) with the AtLAST beam at 460 GHz. In order to detect the 1% polarized dust emission, we need to go 100× deeper than the SCUBA-2 surveys with AtLAST. Thus an RMS of ∼ 16 µJy/beam is needed.
We use the AtLAST sensitivity calculator to estimate the required on-source integration time. Assuming a source elevation of 45 deg, observing frequency 460 GHz, bandwidth 32 GHz, single polarization sensitivity, and H2O profile percentile of 20 (30), we need 28.6 h (47.3 h) onsource integration time per galaxy. Similarly, at 345 GHz and 660 GHz (RMS ∼ 10 µJy/beam and ∼ 31 µJy/beam), with 50-th and 20-th H2O percentile, we need 18 h and 36 h on-source integration time, respectively. This sets the condition to detect the polarized dust emission across extragalactic disks like the HASHTAG/DOWSING sample, but is 100× deeper than JCMT. Luminous and ultra-luminous infrared galaxies ((U)LIRGs) are brighter than the HASHTAG/DOWSING spiral galaxies, and thus they require shorter observing times (integration time ∝ flux−2) of order of a few hours. For a survey of roughly a hundred nearby galaxies including ~10–20 spirals ~50–100 and (U)LIRGs, the total on-source integration time will be about one to two thousand hours. A wider bandwidth and/or a multiband continuum camera will reduce the required observing time correspondingly.
2.3 Physics and chemistry of the interstellar medium
2.3.1 Context and open questions. The relationship between the cold ISM of galaxies and their star formation output is commonly studied and parametrised as the Kennicutt-Schmidt relation (KS relation; Kennicutt, 1989; Schmidt, 1959). This empirical relation links the star formation rate (surface) density to the total HI and H2 gas (surface) density, through a power law with a measured index of NKS,HI+H2 ∼ 1.4 across large samples of nearby galaxies (de los Reyes & Kennicutt, 2019; Kennicutt & De Los Reyes, 2021). This power index, however, is found to be about unity in tens of nearby spiral galaxies on resolved kpc-scales (Bigiel et al., 2008; Leroy et al., 2008; Leroy et al., 2013; Schruba et al., 2011; Wong & Blitz, 2002), with significant variations in both slope and normalisation of the relation at GMC scales and from galaxy to galaxy (Pessa et al., 2021; Pessa et al., 2022; Sánchez et al., 2021). At high enough sub-kpc scales, the KS relation eventually breaks down, a feature that can be exploited to place constraints on important timescales in the star formation process (e.g. Chevance et al., 2020b; Kruijssen & Longmore, 2014).
In addition to this range of physical scales, the KS relation can be studied in multiple gas tracers, each sensitive to a different phase of the ISM. For example, it is found that it is the H2 (molecular) gas rather than the HI (atomic) gas that determines the linear KS relation, even in atomic gas-dominated regions (e.g. Schruba et al., 2011). However, when considering the flaring disk scale heights, the volumetric gas density and star formation rate density are more strongly correlated in both atomic and molecular phases (e.g. Bacchini et al., 2019). Furthermore, Gao & Solomon (2004) first found that the dense gas, as traced by high-dipole-moment molecules like HCN, linearly traces the star formation rate (NKS,dense ∼ 1.0), with fewer galaxy-to-galaxy variations. This may indicate that the denser phase traced by HCN is the fundamental unit where H2 gas is turning into stars at a constant efficiency (Gao & Solomon, 2004; Gao et al., 2007; Greve et al., 2014; Lada et al., 2010; Lada et al., 2012; Liu et al., 2015; Wu et al., 2005; Zhang et al., 2014).
Despite these breakthroughs, we are far from having a full picture yet. In the past ten years, Galactic observations of an even higher-critical-density molecule, N2H+, and observations of HCN and HCO+ in our galaxy as well as in central regions of nearby galaxies have brought into question whether some of the most popular extragalactic dense gas tracers, HCN and HCO+, are indeed tracing the highest-density gas as commonly assumed (e.g. Kauffmann et al., 2017; Pety et al., 2017; Tafalla et al., 2023). Differences between the spatial distributions of N2H+ and HCN in individual galactic giant molecular clouds (GMCs), and the abnormally enhanced HCN in nearby galaxy centers (Bigiel et al., 2016; Gallagher et al., 2018a; Gallagher et al., 2018b; Jiang et al., 2020; Jiménez-Donaire et al., 2017; Jiménez-Donaire et al., 2019; Usero et al., 2015) all suggest that the chemical abundances and excitation conditions are playing crucial roles.
Such results make it obvious that further progress requires moving away from single-line observations, in favor of deep, high resolution, multi-tracer and multi-transition observations, alongside sophisticated non-local-thermaldynamic-equilibrium (non-LTE) excitation and radiative transfer modeling (Leroy et al., 2017; Leroy et al., 2022; Liu et al., 2021; Neumann et al., 2023; Priestley et al., 2023; Teng et al., 2022; Teng et al., 2023).
2.3.2 The role of AtLAST in probing the physics and chemistry of the ISM. Sensitive, large-area and high-resolution multi-species, multi-transition line observations of nearby galaxies are a critical resource for obtaining an accurate understanding of the ISM and star formation physics in nearby galaxies. What is also needed are large surveys that provide significant statistics over a broad range of galactic environments (see example in Figure 7), from low-density regions at the transition between the atomic and molecular phases, to giant star-forming complexes, and up to extreme environments such as circumnuclear regions, including gas exposed to radiative and mechanical feedback from starbursts and active galactic nuclei, and material affected by galaxy mergers.
Figure 7. Artistic rendering (informed by real observations) of the key physical environments in the central kiloparsec of the nearest starburst galaxy NGC 0253 from Martín et al. (2021).
Such a goal is impossible to reach with current facilities. First, spatially resolving multiple lines at physical scales of hundreds of pc or better is required to clearly distinguish physical processes that operate on a large range of spatial scales, from a hundredparsec cloud scales to kilo-parcsec sub-galactic scales (Schinnerer & Leroy, 2024). In the submillimeter wavelengths (ν ≳ 300 GHz), where critical high-energy-level transitions of important molecules are found (see Table 6), the resolution from current submm single-dish telescopes (all ≤ 15 m in size) is insufficient8. Even the largest current millimeter telescope, the Large Millimeter Telescope (LMT), with an aperture of 50-m cannot come close to achieving the desired observations due to a lack of frequency coverage in the submm.
Secondly, the largest submillimeter interferometer, ALMA, has the sensitivity to achieve incredible multi-species, multi-transition observations, however, it has two severe limitations that will not be resolved even by planned upgrades: (1) extended emission is filtered out, and (2) ALMA has a very small field of view compared to the angular sizes of nearby galaxies (see Table 6). For example, the ALMA Cycle 5 large program ALCHEMI has conducted a comprehensive survey of spectral lines in Bands 3, 4, 6, and 7 at 1′′ resolution in the nearest starburst galaxy, NGC 0253 (Martín et al., 2021). Figure 8 shows how these observations reveal an astonishing number of spectral lines, from simple species all the way to complex organic molecules, but only within its central molecular zone (850 × 340 pc, see Figure 7), leaving a vast range of other galactic environments unexplored at such depth. Another important issue is the missing flux. Despite the large collecting area and sensitivity of ALMA’s main array, the total power (TP) antennas with a diameter of 15m is less than one-tenth as efficient as a 50m antenna for scanning faint lines in a large sample of galaxies. Without adequate single-dish data, it has been demonstrated that missing flux can be severe for faint lines (e.g., Leroy et al., 2021b; Liu et al., 2023a), especially missing line flux at GMC edges and in inter-arm regions by more than 50% thus propagating to a huge error in line ratios. Figure 9 shows an example of missing extended flux when mapping a nearby galaxy with ALMA. In conclusion, because of the limitations listed above, ALMA is not the right instrument to expand a survey like ALCHEMI to many more galaxies (to reveal galaxy-galaxy variations), as it cannot recover the multi-line physics accurately at large scales.
Figure 8. Spectrum of the central kiloparsec of the nearest starburst galaxy NGC 0253 from the ALCHEMI survey using the ALMA ACA 7-m array (Martín et al., 2021).
Figure 9. Figure from Leroy et al. (2021b) showing the effect of missing flux.
The left panel shows a short-spacing corrected image of NGC 0628 in the CO(2-1) line, and the right panel shows a 7-m only cleaned image (i.e., without short-spacing correction using any TP data).
AtLAST will uniquely enable such surveys of the multi-species, multi-transition molecular/atomic lines in hundreds of nearby galaxies, representing all kinds of galactic environments and thus physical mechanisms, at GMC/star-forming complex scales. ALMA will be highly complementary to achieving sub-cloud resolution in a limited number of targeted regions of interest. We summarize below some of the key science questions that can only be answered with AtLAST.
Is the gas-star formation relation universal? This question is fundamental in our understanding of star formation. The HERACLES (Leroy et al., 2009), PHANGS (Leroy et al., 2021a) and EDGE-CALIFA (Bolatto et al., 2017; Sánchez et al., 2021) surveys have achieved sub-kpc scale studies of the KS relation in a single 12C16O line in about a hundred nearby star-forming galaxies representing the local star-forming main sequence. The EMPIRE (Jiménez-Donaire et al., 2019) and MALATANG (Jiang et al., 2020; Tan et al., 2018) surveys have achieved ≥ kpc-scale dense-gas KS relation studies in about ten IR-bright galaxies. However, the debate about the emissivity of dense gas tracers and the scatter in star formation efficiency at low densities are strongly limiting our theoretical understanding of the star formation processes. Is there a density threshold? Is local gas pressure playing a role? What drives the star formation efficiency / timescale variations? How well can we trust the surface mass densities from CO or HCN, using conversion factors? These are all related questions that need AtLAST’s sensitivity, GMC-scale resolution, mapping speed and large field of view.
What are the gas densities and temperatures? Multi-transition non-LTE modeling of molecular/atomic lines is the most powerful tool to measure gas densities, temperatures and line excitation conditions for the multiphase gas in galaxies. Fifteen years ago, only very few galaxies had ground-based (sub)millimeter observations of the CO spectral line energy distribution (SLED) up to J = 6–5. The Herschel space observatory opened a new window to studying the CO SLEDs in over 160+ nearby galaxies (Greve et al. (2014); Kamenetzky et al. (2014); Kamenetzky et al. (2018); Kamenetzky et al. (2016); Kamenetzky et al. (2017); Liu et al. (2015); Lu et al. (2017); Rosenberg et al. (2015)), however, with ∼ 30–40′′ resolution these observations do not resolve external GMCs. Moreover, the limited spectral resolution of the instruments on board the Herschel satellite prevents us from using these data to study spectral line profile variations in different H2 tracers and CO transitions. The latter is a promising avenue for understanding the relation between extreme H2 gas excitation and feedback mechanisms at play in galaxies, as shown by Montoya Arroyave et al. (2023) and Montoya Arroyave et al. (2024) using sensitive multi-J CO and [CI] APEX observations of tens of local (U)LIRGs.
High-resolution, multi-transition, multi-species non-LTE studies have only been made very recently in a few IR-bright galaxies using ALMA (e.g., CO isotopologues J = 1–0 to 3–2 Teng et al. (2022, 2023); CO J = 1–0 to 4–3 + [CI] 3P1–3P0 Liu et al. (2023a, 2023b)). Also with ALMA, the PHANGS survey has mapped of order one hundred nearby galaxies in CO J = 2–1, but obtaining similar coverage of additional lines, e.g., CO J = 3–2, is still very time consuming given the small field of view and reduced atmospheric transmission at high frequencies. The ALMA studies are limited to a few galaxy centers, or a few off-center regions (e.g., like the GMC study in M83’s XUV disk; Koda et al., 2022), which do not probe the variety of galactic environments. Simulations and theory have shown that the emissivity of different species can vary significant with galactic environments, thus the conversion factors do vary accordingly. The gap from single-line conversion factor-based studies to multi-line non-LTE studies has to be filled in order to understand how gas density and temperature play a role across the star-forming disks/environments, and to anchor simulations and theories (e.g., photodissociation region, cosmic ray dominated region, X-ray dominated region and shock physics (Bisbas et al., 2015; Bisbas et al., 2017; Kaufman et al., 1999; Kaufman et al., 2006; Meijerink & Spaans, 2005; Meijerink et al., 2007; Papadopoulos, 2010; Papadopoulos et al., 2011; Wolfire et al., 2022).
2.3.3 AtLAST surveys of multiple gas/dust tracers in the nearby Universe. A complete survey of multiple (sub)millimetre lines in a large number of galaxies at hundreds of parsec scales will enable unique studies of gas and chemistry in galaxies. The brightest (sub)millimetre lines are listed in Table 6, but many fainter lines from species like HC3N, HC3HO, HOCN, HOCO+ (Martín et al., 2021) are also expected. The sample of nearby galaxies should include a variety of southern nearby galaxies. For main-sequence types, there are ∼ 100 galaxies from the PHANGS-ALMA large program which have full-disk CO(2–1) line mapping at ∼ 1′′ (about 30–300 pc). This survey (the large program and a few pilot programs) used about a hundred hours of the ALMA main array and 600+ hours of ALMA ACA including the total power observations. The 13CO(2–1) isotopologue line is observed simultaneously, but due to the limited sensitivity to the extended emission, only brightest clouds are imaged in high quality. AtLAST multi-line mapping of these galaxies will be an unprecedented legacy survey at moderately high resolution (∼ 2–4′′) starting from Band 7, or the CO(3–2) transition. The 3–2 transitions of CO isotopologues have been found crucially useful in determining the CO-to-H2 conversion factors (Teng et al., 2022; Teng et al., 2023). The Band 8 CO(4–3) and [C I] lines are of great importance in constraining gas density and even possibly stellar and AGN feedback at resolved scales (Liu et al., 2023a; Liu et al., 2023b; Saito et al., 2020). An ALMA ACA survey of the CO(4–3) and [C I](1–0) lines was conducted in 36 (ultra-)luminous galaxies at ∼ 3′′ (Michiyama et al., 2021), leading to the finding of unusually low-[C I] galaxies which are still puzzling. However, the deep, multi-tracer single-dish APEX observations of local (U)LIRGs presented by Montoya Arroyave et al. (2023) include [CI](1-0) coverage for 17 sources (some of which in common with the ACA sample of Michiyama et al. (2021)) and show a significantly higher [CI] flux recovered by APEX, resulting in an average [CI](1-0)/CO(1-0) luminosity ratio as high as 0.210 ± 0.009 in local (U)LIRGs. The AtLAST survey of lines will be highly complementary to ALMA main array and ACA surveys which either have much higher resolution (and usually missing fluxes or would require hundreds of hours total power observing time) or only have the sensitivity to observe the most infrared-luminous starbursts.
We can consider a sample of 100 nearby galaxies with a variety of types:
main-sequence spirals
starbursts
AGNs/composite galaxies
dwarfs
and we observe the Band 6 (CO 2–1 isotopologue), Band 7 (CO 3–2), Band 8 (CO 4–3 and [C I] 1–0), Band 9 (CO 6–5), and Band 10 (CO 7–6 and [C I] 2–1) lines. In each Band, we assume that the receiver has sufficient bandwidth to cover all the key lines simultaneously, that is, ∼ 16–20 GHz per sideband for a double-sideband receiver. We also require a multi-beam receiver or integral field unit (IFU) to efficiently map the galaxies. While both multi-beam receiver and IFU achieve a high mapping efficiency thanks to focal-plane multiplexing, multi-beam receiver uses superconducting heterodyne technique thus usually has a much higher spectral resolution than an IFU using superconducting kinetic induced detectors (KIDs). This spectral resolution capability is essential to spectral line science. In comparison, the large-format camera proposed for previous science cases has no spectral resolution thus is only for dust continuum science (but is sensitive to faint polarized signal). In the case of a multi-beam heterodyne receiver, assuming a 919-beam (18-layer) array, as shown in Figure 10, a galaxy like NGC 3627 at a distance of ∼ 11 Mpc will be covered by roughly ten pointings. For slightly more distant local galaxies like those in the EDGE-CALIFA or MaNGA samples, fewer pointings are needed. In the case of an IFU, the field of view is much larger, allowing for these nearby galaxies to be covered in one pointing, but the spectral resolution is much lower, of only ∼ 300–3000 km/s (R ∼ 100–1000). Our survey of line ISM physics do not strongly require resolving the lines, but require accurate integrated line fluxes and ratios, therefore an IFU with a spectral resolution R ≳ 1000 will be more efficient than a multi-beam receiver (although a multi-beam receiver will enable kinematic studies that would add additional value).
Figure 10. Footprints of various assumed instruments (from left to right): a 919-beam (18-layer) heterodyne receiver, a 217-beam (9-layer) heterodyne receiver, an IFU with 50 × 50 pixels, and an IFU with 900 × 900 pixels, on top of the nearby galaxy NGC 3627’s CO(2–1) emission (Leroy et al., 2021a).
The heterodyne receiver’s hexagon units and the IFU pixels are assumed to have a spacing/size of 4′′, which is about the AtLAST 50-m angular resolutions at Band 7–10, and is about half of the angular resolutions at Band 4–5. A large-format IFU as shown in the last panel can most efficiently observe multiple galaxies, with each galaxy fully sampled at resolutions as shown in the first three panels.
Then, we estimate the line fluxes and sensitivity requirements. Taking NGC 3627 as an example (Figure 10), its CO(2–1) surface brightness ranges from ≳ 1000 K km s−1 to ∼ 30 K km s−1 at ∼ 8 kpc away from the center along the spiral arm. The peak temperature ranges from ∼ 8 K to ∼ 1 K from the center to the ∼ 8 kpc arm. The 13CO lines are in general ∼ 10× fainter than the 12CO lines (e.g., Cao et al., 2023; Cormier et al., 2018; Henkel et al., 2014; Jiménez-Donaire et al., 2017; Taniguchi & Ohyama, 1998; Wilson & Rood, 1994), and the [C I](1–0) lines are generally ∼ 5–10× fainter than 12CO(2–1) (Liu et al., 2023a). The ratios of CO(1–0) line fluxes to those of HCN and HCO+ 1–0 are typically ∼ 30–100 at resolved kpc scales (Bigiel et al., 2016; Gallagher et al., 2018a; Gallagher et al., 2018b; Jiménez-Donaire et al., 2019; Neumann et al., 2023). Thus, in general we will need to reach an RMS of ∼ 0.1 K km s−1 to achieve S/N ≳ 3 detections for all the aforementioned lines. Given a line width of ∼ 30 km s−1 and the AtLAST beam of 2′′–7′′ from Band 10 to Band 6 (Jy/K ∼ 1.5–2), this RMS corresponds to ∼ 5–7 mJy at these bands.
Using the AtLAST sensitivity calculator, we compute that at an elevation of 45 deg, an observing frequency 230/345/461/492/691/809 GHz, with a bandwidth of 0.023/0.035/0.05/0.05/0.07/0.08 GHz (∼ 30 km s−1), and an H2O profile percentile 80/80/50/50/20/20, the on-source integration time will be ∼ 1/10/24/84/25/60 minutes. Thus the expected typical on-source observing time is only ∼ 3.5 hours for all the major Band 6–10 lines in one pointing in a galaxy like NGC 3627. The full survey of ∼ 100 nearby galaxies with one IFU pointing (or multiple pointings with multi-beam heterodyne receiver) for each galaxy will be at scales of hundreds to a thousand hour.
2.4 An “SDSS-like” submillimeter survey for galaxy evolution studies in the nearby Universe
2.4.1 Context and open questions. The evolution of galaxies is a fundamental question and a complex process in astrophysics. The amount of cold gas, the rate of star formation, and the environment of galaxies all play a critical role in shaping galaxy evolution. How the cold gas is accreted into galaxy systems, how the star formation is set by the turbulent ISM properties, and how the gas and dust are influenced by the external gravitational forces in dense environments are all key processes demanding detailed understanding with large-sample surveys. While millions of galaxies have been observed in optical spectroscopic surveys such as SDSS and more recently DESI, only hundreds of them have constraints on the molecular gas reservoir — thus leading to a giant gap in our knowledge of galaxy evolution (Liu et al., 2019; Saintonge et al., 2011; Saintonge et al., 2017; Saintonge & Catinella, 2022; Tacconi et al., 2018; Tacconi et al., 2020). An extra-large submillimeter survey in the local Universe will be a substantial breakthrough. It will help bridge the gap with optical, near-infrared (e.g., Euclid, LSST, Roman), and radio (e.g., with the SKA) surveys, which would otherwise become even greater. Extending galaxy samples with molecular gas measurements by orders of magnitude will lead to the most important statistical understanding of cold gas and star formation, and environments and galaxy quenching. Possibly tens of thousands of molecular line detections could be achieved with an “SDSS-like” survey with AtLAST, highly complementary to the atomic gas surveys with the future SKA.
The statistical properties of the molecular ISM gas in low-mass (M⋆ ≲ 109 M⊙) galaxies is one of the major unknowns in galaxy surveys. Low-mass galaxies have sub-solar metallicities and correspondingly low CO abundance compared to the H2. With less dust shielding, radiation fields created by the young massive stars are in general stronger and harder. Many low-mass galaxies are also satellites of massive galaxies, e.g., the LMC and SMC as described in Section 2.1. The xCOLDGASS survey (Saintonge et al., 2011; Saintonge et al., 2017) is the largest millimetre CO survey of local galaxies with the IRAM 30 m telescope. This survey has been able to constrain the local-Universe CO luminosity distribution down to 107.5 K km s−1 pc2 (or a stellar mass limit of 109.0M⊙). The CO luminosity function (and the consequent H2 mass function) in the local universe is a important tool in constraining models (e.g. Vallini et al., 2016) and anchoring studies of the redshift evolution of gas reservoirs (e.g., Dessauges-Zavadsky et al., 2020; Liu et al., 2019; Walter et al., 2020). The scaling relations between molecular gas mass and other global galaxy properties (stellar mass, SFR, morphology,...) are also proving to be very constraining for hydrodynamical simulations (e.g. Davé et al., 2020). Despite all these successes, with only ∼ 1000 galaxies, combining xCOLDGASS and other similar surveys (Boselli et al., 2014a; Cicone et al., 2017; Colombo et al., 2020; Wylezalek et al., 2022), we can only scratch the surface, as significantly more statistics are required to disentangle the competing effects of different evolutionary mechanisms. It is particularly crucial to improve statistical constraints of the H2 content of low-M* galaxies, not only because these are the most common galaxies in the Universe, but also because including these sources allows us to expand the dynamic range of current scaling relations and so as to provide more reliable constraints for theoretical models, as shown by pilot studies using ∼ 330 hours of APEX observations (ALLSMOG survey: Cicone et al., 2017; Hagedorn et al., 2024).
Due to these sample size limitations, current surveys are unable to accurately quantify the impact of large-scale environment and the position of a galaxy with respect to the cosmic web on the molecular ISM. In addition, compared to low-mass galaxies in the field, those in dense regions are experiencing both secular evolution and environmental processes which are not fully understood. We know from HI observations that galaxies in dense regions (such as clusters) generally have a reduced atomic gas content and star formation activity, compared to their counterparts in low density environments (e.g. Haynes & Giovanelli, 1986). For example, Figure 11 shows an example of stripping of HI gas in a galaxy infalling into the Virgo cluster. A range of mechanisms, from gravitational interactions to hydrodynamic processes, can explain the impact of the dense cluster environment on the HI contents of galaxies. What is still not totally understood is their effect on the molecular gas component, as molecular gas is much more embedded into the gravitational potential of the disc and mainly located within giant molecular clouds, which means it is much less sensitive to external perturbations than the atomic phase (e.g., Boselli et al., 2014b; Boselli et al., 2022; see also recent APEX and ALMA studies from the GASP survey; Bacchini et al., 2023; Moretti et al., 2020).
Figure 11. The right panel shows the Virgo cluster’s optical and CO images from the VERTICO ALMA large program (Brown et al., 2021, see the website) and the Astronomy Picture of the Day.
The CO maps of each individual member galaxies are scaled up for visualization. The top left panel is adapted from Boselli et al. (2016) and Boselli et al. (2022), showing the very extended stripped materials from the Virgo cluster member galaxy NGC 4569. The image size is 50′ ×25′ or 250kpc×125kpc. The red image shows the stripped ionised gas extending up to 150 kpc. The bottom left panel is adapted from Brown et al. (2021), showing the cluster members NGC 4567 and NGC 4568. The Virgo cluster extends at least 10° × 10° in the sky.
An extra-large survey of > 10,000 local galaxies to measure accurate molecular gas masses, especially if extending the coverage of current surveys into the dwarf galaxy regime (M* < 109.0 M⊙) and covering the full range of extragalactic environments, from field to clusters, is the only approach to achieve the statistics required to address several important questions in galaxy evolution, such as those described below.
2.4.2 The role of AtLAST in the study of the local galaxy population. AtLAST is ideally suited to shed light on these important topics. Thanks to its high sensitivity and large field of view, it can be used to map in a blind survey a sufficiently large fraction of the nearby sky to measure with unprecedented sensitivity and statistics the molecular gas mass function of galaxies in the local Universe for objects located in different density regions, from voids and filaments to groups and massive clusters.
What is the abundance of molecular gas in the local Universe and its distribution? The molecular gas mass function is a very useful constraint to semi-analytic models and cosmological simulations (e.g. Davé et al., 2020; Diemer et al., 2019; Lagos et al., 2015). Like for atomic gas, where large blind HI surveys have allowed for accurate measurement of the HI mass function in the local Universe (Jones et al., 2018), we can eventually have a deep and complete census of molecular gas in the nearby Universe with AtLAST. Therefore, a key next step is largely improving the constraints on the z = 0 molecular gas mass functions which have been exclusively based on optically- or IR-selected samples (e.g. Andreani et al., 2020; Fletcher et al., 2021; Keres et al., 2003). AtLAST is the only facility which would enable a blind survey for molecular gas in the nearby Universe, which is essential to anchor the all important studies of the redshift evolution of the baryonic mass of galaxies and their star formation output (e.g. Walter et al., 2020).
Molecular gas scaling relations in the dwarf galaxy regime The first large systematic molecular gas surveys of the past 15 years, both at low and high redshifts, have been transformational in our understanding of the central roles of the baryon cycle and star formation efficiency in the galaxy evolution process (e.g., Carilli & Walter, 2013; Gao & Solomon, 2004; Kennicutt, 1998; Leroy et al., 2008; Leroy et al., 2021a; Saintonge et al., 2011; Saintonge et al., 2017; Tacconi et al., 2018). With these surveys, we now have a good overview of the basic scaling relations between total molecular gas mass and other galaxy observables (e.g., de los Reyes & Kennicutt, 2019; Kennicutt & De Los Reyes, 2021; Tacconi et al., 2020), but only in relatively massive galaxies (M* > 109.5M⊙). A key next step is systematically surveying CO and other species in lower mass galaxies (e.g. ALLSMOG, Cicone et al., 2017). With the increased sensitivity of AtLAST and improved survey strategies enabled by its very large field of view (see Section 2.4.3), it will be possible to expand molecular gas scaling relations well into the regime of dwarf galaxies (M* < 109M⊙), either through direct line detection or stacking of the many undetected objects that would be covered by a blind CO line survey.
Why do galaxies stop forming stars? This question, often referred to as "star formation quenching", is central to our understanding of galaxy growth, and as of yet remains unanswered. A galaxy can stop forming stars if it runs out of cold gas, or if any cold gas it has is some-how prevented from forming stars. All the physical mechanisms that are invoked to explain quenching work by either affecting the gas contents of the galaxies, or the star formation efficiency out of any available gas. Therefore, only by having direct information about the cold atomic and molecular gas mass of galaxies can we accurately pinpoint the mechanisms that are causing star formation activity to shut down. Current evidence suggests that both a reduction of gas reservoirs and a decrease in star formation efficiency are at play (e.g. Colombo et al., 2020; Saintonge & Catinella, 2022), but the sample size limitations are currently preventing us from understanding why these changes are happening. AtLAST will enable the assembly of a sample of ∼ 105 galaxies with CO detections (two orders of magnitude more than current samples, see Section 2.4.3) which will finally allow us to disentangle the complex physical processes behind star formation quenching.
2.4.3 AtLAST surveys to probe star formation and galaxy evolution in the nearby Universe. The primary goal of this survey is to detect CO lines in a significantly large sample that is a factor of > 10–100 × larger than any existing survey in the local Universe. This consists of two types of targets: a mass-complete sample of field galaxies in the local Universe (e.g., redshift z ∼ 0.01–0.38, or distances ≲ 2, 000 Mpc, and a stellar mass limit ~ 108–9 M⊙); and a sample of (super)clusters for the dense environments.
The field targets will naturally be drawn from on-going and future large optical/near-infrared and radio H I surveys, for example, the Dark Energy Survey (DES) and the LSST in the southern hemisphere. The sample selection will follow the COLDGASS and xCOLDGASS strategy but have great improvements in the volume completeness. As mentioned in Section 2.4.1, the COLDGASS and xCOLDGASS IRAM 30-m large programs randomly trim a large, mass-complete, H I-detected sample from the GALEX Arecibo SDSS Survey (GASS; Catinella et al., 2010; Catinella et al., 2013) down to a few hundred galaxies for the CO observations. The final combined sample is 532 galaxies with a total of ∼ 950 hr of IRAM 30-m observing time. The receiver has a single beam thus the field of view of each IRAM 30-m observation is limited to ∼ 22′′ at 115 GHz and ∼ 11′′ at 230 GHz. With a large-format submillimeter IFU taking advantage of AtLAST’s very large field of view, surveys on unprecedented scales will become possible. We boldly assume that an expensive, large-format IFU can be build to fill about 1 deg2 field of view on AtLAST, with a bandwidth of 32 GHz and a spectral resolution of R~1000 (~40 km/s at 230 GHz). This requires a spaxel unit separation of 4″ to achieve a Nyquist sampling of the ~7″ beam at 230 GHz, that is, a KID pixel array of 900 × 900 and channel number ~1000 along the third dimension. The cost scales up quickly with the number of KIDs, that is, 3D spaxel units. Under this assumption, each targeted observation at a selected galaxy will have an 1 deg2 field of view to allow for detections of CO line emitters in the vicinity. According to our best constraint of the CO luminosity function so far (Fletcher et al., 2021; Saintonge et al., 2017), we predict that with 30 min on-source integration of the IFU, the number of CO line emitters will be as large as ∼ 150 in a single 1 deg2 field of view as shown in Figure 12 (this estimate assumed metallicity-dependent conversion factors, see Figure 12 caption). Observing the CO(2–1) line will achieve more galaxies per redshift bin and a much lower MH2 mass limit. Therefore, it is better than observing the CO(1–0) line for studying the faintest galaxies. If we observe a southern xCOLDGASS-like mass-complete sample of ∼ 500 primary targets, each with 30 min on-source integration with the IFU, then we will eventually obtain a combined sample of 500 × 150 ∼ 75, 000 galaxies at 0 < z < 0.16 (i.e., out to a distance of 200 Mpc) with > 5σ CO(2–1) detections. The exact number depends on the actual CO luminosity distribution to be measured and may vary by a factor of a few, but this is truly transformational compared to any existing CO surveys.
Figure 12. Predicted numbers of CO emitters in an 1 deg2 field of view for 30 min on-source integration with the Band 3 CO(1–0) line (lower panels) and with the Band 6 CO(2–1) line (upper panels).
With a bandwidth of 32 GHz, the CO(1–0) line frequencies are ∼ 115–83 GHz, corresponding to z ∼ 0–0.38, Similarly, for the CO(2–1) line the frequencies are∼ 230–198 GHz, corresponding to z ∼ 0–0.16. The numbers of CO emitters are calculated by integrating the CO luminosity distribution in (Saintonge et al., 2017, see their Figure 6) from the 5σ detection limit to 1011 K km s−1 pc2. Left panels are the predicted numbers of CO emitters per redshift bin of Δz = 0.02 for CO(2– 1) and Δz = 0.01 for CO(1–0). The total predicted CO detections above 5σ are ∼ 150–160 for the CO(2–1) and CO(1–0) lines. Right panels are the molecular hydrogen mass MH2 detection limit computed from the 5σ detection limit using a metallicity-dependent αCO conversion factor following Bolatto et al. (2013). A half-solar metallicity is assumed for the MH2 limits, and a CO(2– 1)/(1–0) brightness temperature ratio of 0.8 is assumed for the excitation.
In the dense environments, our (super)cluster targets should contain a statistically significant number of galaxies in the densest parts of our local Universe such as those located in the Coma/A1367 supercluster (e.g. Gavazzi et al., 2010). This particular region of the sky includes ∼ 4000 objects down to the SDSS r-band mag 17.7 at a typical distance of 70–100 Mpc over ∼ 400 square degrees, with galaxies located in two rich clusters (Coma and A1367), in several groups and filaments, and in voids. It is in the northern hemisphere though (Dec ∼ +19). Another widely-studied cluster, the Virgo cluster, is also in the north (Dec ∼ +12), spreading over 100–200 deg2 in the sky with over 1300 member galaxies. To facilitate the observation at the AtLAST site, southern (super)clusters can be selected from the Abell et al. (1989) catalog, where over 4000 clusters are listed and each has at least 30 member galaxies. The AtLAST’s unprecedented sensitivity to the extended emission will make the survey of nearby galaxy clusters not only for counting galaxies number densities and luminosity function, but also for mapping any possible stripped gas at tens of kpc scales (Figure 11). If a gas-rich satellite galaxy has 50% of its gas stripped, then an extended gas structure of a mass Mgas ∼ 108M⊙ will be spread over tens of beams. Our detection limit of 30 min on-source integration is already ∼ 2 × 107 M⊙ (for CO-traced H2 gas at 0.5 Z⊙) out to ∼ 100 Mpc, thus it is very promising for AtLAST to open a new era for studying the extended stripped gas structures in clusters. These cluster observations can be much deeper than the field galaxy observations, as they require fewer pointings.
Combining the field and cluster CO surveys, this will provide us with an unrivaled sample of 105 CO emitters in the local Universe, enabling the transformative science depicted above. Such a survey will require 500 hours for the field galaxies and another 500 hours for the clusters, including calibration and telescope overheads, covering ∼ 500–1000 deg2 areas. Since the observations are done in Band 6, the observing time is not very sensitive to weather conditions and therefore such a survey could make good use of time in poor weather conditions. In the calculation for this survey, we assumed a water vapour percentile of 85 and 95 for the Band 6 and Band 3 sensitivity estimates using the AtLAST sensitivity calculator. We also assumed line widths of 200 km/s, which is 0.16 GHz and 0.073 GHz at Band 6 and Band 3, respectively.
