• If you are citizen of an European Union member nation, you may not use this service unless you are at least 16 years old.

  • Files spread between Dropbox, Google Drive, Gmail, Slack, and more? Dokkio, a new product from the PBworks team, integrates and organizes them for you. Try it for free today.


Galactic Sub-group Discussions

Page history last edited by pamela Klaassen 1 year, 5 months ago

Galactic Science Case Sub-group Discussions


Below are the discussion points that have come about through discussion of the Galactic science working groups (primarily via telecons)




Polarisation Studies

(contents taken from a summary by Sarah Sadavoy)


1) Potential Capabilities for Polarization Observations

- large-scale mapping of polarization at high resolution.

- based on the transmission curves on the plateau and high site (e.g., from this paper: https://arxiv.org/ftp/arxiv/papers/1807/1807.04354.pdf), AtLAST could cover 350 GHz (850 um), 460 GHz (650 um), 650 GHz (460 um), and 800 GHz (375 um).  Thus, AtLAST can bridge a regime in FIR/SMM that is not well covered by other instruments.

2) Synergy with other facilities (in terms of polarization)

- dust polarization at 1-2 mm: LMT/TolTEC, IRAM/NIKA-2 (resolutions of 6-10 arcsec) - not yet available
- dust polarization at 850 um (450 um?): JCMT/POL-2 (resolutions of 10-14 arcsec)
- dust polarization at 50-214 um: SOFIA/HAWC+ (resolutions of 4-19 arcsec) - unable to do mapping
- dust polarization at 850 um: Planck (all sky survey, but resolution is 10 arcmin)
- line polarization at 1-3 mm: IRAM/XPOL (resolutions of 10-30 arcsec) - unclear if upgrades are complete
- dust/line polarization at 850 um - 2 mm: ALMA (resolutions vary, but largest angular scale typically < 30 arcsec)

AtLAST can bridge the scales between Planck and ALMA, which can be very important for magnetic field structures.  ALMA does not have large-scale polarization capabilities (unclear if polarization will be made available for the ACA) and the Planck polarization data have low resolution.  Depending on the size of the detector, AtLAST should detect emission on scales of < 4 arcsec to several arcmin.  These scales are important to trace magnetic field structure within filaments and down to the scales of cores in nearby clouds.  Planck showed that magnetic fields tend to be perpendicular to dense filaments, but could not resolve how the field changes within the dense filaments, particularly toward the regions of star formation.

3) New capabilities

- facilities with line polarization are limited.  Observations can be done with XPOL at the IRAM 30m telescope, but the telescope site is not great for polarization which will limit survey capabilities.  So a good survey instrument for line polarization at a really good location would be unique.

- full Stokes parameters to get linear polarization (Stokes Q/U) for dust polarization / line polarization (Goldreich Kylafis effect) and circular polarization (Stokes V) for Zeeman observations.  There is also some indication that circular polarization can be detected in non-Zeeman molecules due to the conversion of background linear polarization to circular polarization (Houde+ 2013 ApJ, 764, 24; Hezareh+ 2013 A&A, 558, 45).  This conversion may complicate comparisons between linear polarization measurements from dust and line observations so having the Stokes V could help clear this up.

- Zeeman observations also limited.  This will be a new capability at ALMA.  Depending on the line tracer, larger-scale emission could be used with AtLAST

4) Potential Science

- High resolution single-dish polarization observations;  < 4 arcsec will allow us to resolve a scale of 0.1 pc (important for cores and filaments) out to ~5 kpc.  Bridge gap between Planck and ALMA, study how flux-frozen fields are brought in by gravity, the influence of outflows, the effect of fields in building up mass into filaments and cores, energy balance, etc.

- Dust polarization from 350 GHz to 800 GHz can be used to study polarization efficiency.  Previous observations show a trend to lower polarization fractions at ~ 350 um (see Figure 5 in Vaillancourt & Matthews 2012, ApJS, 201, 13) which isn't fully understood.  Only a few regions have been studied at these frequencies, AtLAST is well positioned to conduct large, multi-wavelength polarization surveys to determine whether this trend is ubiquitous, why it is happening, and what it means for dust properties/magnetic field measurements.  Is there a way to observe several bands simultaneously?

- Line polarization could be superior to dust polarization toward the Galactic plane, where multiple clouds along the line of sight can be in issue for optically thin dust emission. Problem with the 90 degree ambiguity, however.  Are there reliable ways to break this?  (e.g., Cortes+ 2005, ApJ, 628, 780).

- Survey of compact sources in Zeeman observations.  Lots of potential lines (most common are OH and CN, can also do other lines like SO, SiO).  Large survey is not beneficial here (CN is mostly compact) but a series of pointings would be good.  If AtLAST can efficiently take snapshots (single-pointings) of high density cores, can measure Zeeman splitting on small scales and look for variations in the line-of-sight field direction across clouds and filaments.

Questions and Comments (from e-mails and during the last telecon)

1) What areal coverage do we need from AtLAST and to what depth (e.g., few clouds at 10s of sq degree, all of the GB at 100s sq degree, etc.).  Do we need to observe just the dense molecular gas or also the diffuse outer parts of clouds (Av > 10, Av > 3 etc.)
-  for dust polarization, we will be primarily limited to the dense parts of clouds where there is sufficient emission to measure polarization.  For diffuse parts of clouds, one can use NIR polarization instead.  So going to an Av of 3 may not be doable with dust.  We may have better luck with line polarization, but line polarization in diffuse cloud structure has not been very well tested.

2) For continuum polarimetry, is areal coverage more important that angular resolution?  - smaller dish easier to build with larger FOV, larger dish gives better resolution but is that important?
- lots of discussion about smaller dish versus larger dish. Unclear that we would benefit greatly from very high resolution given that there will be some overlap with ALMA in compact configurations.  May be better to have a larger FOV to achieve maximum survey capabilities.

3) How does AtLAST compete/collaborate with other facilities? Ground-based polarimeters have difficulty recovering large scales, HAWC+ can’t map.  Is it better to go to space (e.g., SPICT/OST - poor resolution but great sensitivity)
- agreement that we will need to fold in what TolTEC can do, as there will be a lot of overlap in science.  If we go with a smaller dish, larger maps and fainter dust emission (because we’ll be observing at higher frequencies) could be the benefit.  TolTEC is also targeting specific filaments, with a bias toward nearby clouds.
- Line polarization in the Galactic plane (if we can get it to work) would be an unique prospect as line emission will enable us to measure polarization for multiple clouds along the line of sight.  Issues with dust polarization too (we don’t know grain alignment well).  As Thomas suggested, we need to concentrate on big science goals and new science, not just the same thing but better.
- Zeeman observations (not just with CN, but with other more abundant molecules) would also be an unique capability.

4) How does AtLAST continuum polarimetry compare with ALMA?  Will it uncover new regions for consideration (e.g., via surveys and large scale mapping), larger scales around sources of interest (reasonable beam match is important)?
- agreement that AtLAST should do both.  AtLAST should do ALMA follow-up (large-scale observations for interesting targets) and also new science for ALMA (surveys that identify regions that ALMA can follow-up on at higher resolution).

5) Is it practical to go for line polarization?   Line polarization levels are typically low (few percent) and it is difficult to calibrate/measure instrument effects.
- Wayne suggested a better instrument design could improve instrument calibration.  The ROVER instrument at the JCMT was not ideal.  It used a stepping wave plate and from Mauna Kea the weather can change rapidly in the time it takes to go through a complete 360 deg rotation.
- Is there a chance to do this better?

6) Is it practical to try and observe Zeeman splitting?
- Jens points out that Zeeman signatures are very faint and will need a lot of integration time for an upper limit.  May be worth mentioning as a science capability, but not making it a main motivation.
- Can be improve Zeeman observations in the telescope design?  We need to discuss with an expert.



Transient Science

(contents taken from a summary by Doug Johnstone)


Monitoring Known Sources:


1) Efficient Observing Strategies require a very versatile telescope:

- if sources are clustered, like protostars, then it is important to efficiently observe a field of view similar to the clustering scale

- for nearby molecular cloud protostars the required field of view is ~ 0.5 x 0.5 square degrees

- relative calibration accuracy can be achieved through the many sources in the field

- if sources are sparse across the sky then we need a telescope that slews quickly

- for high mass star formation regions the required field of view is the size of the region < 0.5x0.5 square degrees

- relative calibration accuracy may need to be done separate from the observation of the field (too few bright point sources)


2) Wavelength Range, Sensitivity, and Spatial Resolution

- monitoring a wide-range of wavelengths near the peak of the SED is preferable (at least three independent wavelengths)

- for protostars this argues for 350 microns (200 microns - goal) through 850 microns

- high sensitivity is required in order to uncover small variations - varies inversely as SNR; to uncover 10% variations requires a SNR ~ 30

- reach only brightness protostars with ~10 mJy sensitivity, to get the many faint sources desire sensitivity of ~0.1 mJy !!

- note that this may be below the confusion limit but that is okay as we are looking for time variability at this brightness

- for almost all expected variability cases, the brightening source should be compact (think light crossing times)

- thus an advantage to having higher spatial resolution - crowding, substructure, etc. 

- however, follow up of interesting sources with ALMA, etc. is also reasonable (AtLAST finding sources for ALMA)



Monitoring Unknown Sources:


I am not, yet, an expert on this topic and would like to have others chime in on the requirements for these type of searches. Most of the arguments that I have seen so far are for extragalactic sources but maybe there are galactic counterparts we should consider here ...



Competition/Collaboration with other Telescopes:


JCMT Transient Survey, using SCUBA-2 at 850 microns, requires ~50 hrs a year and monitors about 50 protostellar sources in 8 nearby star-forming regions (0.5x0.5 square degrees) with monthly cadence to a depth of ~5% variability per year.  The survey has uncovered about 10% of the sources to be variable! The survey also observes at 450 microns but the weather and telescope are not as capable and the data, while still useful, is not as sensitive.


CCAT-p will be more powerful by observing with a very sensitive and much larger instantaneous field of view camera, at 350 microns where the variability signature from protostars is also significantly stronger. It is not yet clear just how deep the observations will reach.


Space missions (SPICA/OST) will have incredible sensitivity and a very wide range of wavelengths over the peak of the SED (likely covering ~50 microns - 500 microns). The resolution, at the longer wavelengths, will be poor and efficient observing strategies are not obvious for the fields of view required (space telescopes tend to be inefficient to move about). As importantly, space missions in the far IR are limited in lifetime.


From the above, the niche for AtLAST will be in monitoring for variability (transients) over long periods of time (tens of year), at the shortest wavelengths that can be commonly observed from the site (need to reach 350 microns!), with an excellent sensitivity (at least 1 mJy rms but aim for 0.1 mJy), and with an efficient observing strategy across 0.5x0.5 square degree fields and with a fast slew between fields.  Smaller beam sizes are always preferred but mapping speed and short wavelength observations are more important.



Thoughts on Calibration:


Stability of the instrument, allowing a relative calibration between epochs to be performed to high accuracy (1% or better preferred), is necessary. We can already reach about 2% relative calibration with SCUBA-2 at the JCMT so by being careful this should be straightforward for AtLAST (assuming that there are non-varying sources in the field to calibrate against).


Dynamic range is also important as the brightness of the sources in a field may range over (even more than) three orders of magnitude and we desire a signal to noise of ~100 for the faintest of these sources. Thus we need a signal to noise of at least 10^5!






Galactic continuum surveys

(contents taken from a summary by Thomas Stanke)


General remarks:
We need a science case, that is really new and not feasible with present/upcoming
telescopes, not just some thing that we can do somewhat better.

The key question we have to answer is:
Which science question can we think of, that needs to be answered, and needs a
significant gain in submm single dish observing capability?

We should have a telescope in mind that is very significantly more powerful
than present day (size, FOV, instrumentation capabilities, frequency coverage),
but leave details about telescope/instrumentation/site to the experts.

Continuum surveys:
- to be benchmarked against existing surveys:
   Spitzer/Herschel Galactic plane surveys
   JCMT Gould Belt survey, other surveys of individual regions

- improved sensitivity -> extend mass range to lower masses
- improved spatial resolution -> resolve smaller structures
- caveat: cloud structures are not point sources, confusion might be the limit

- what would be the great new science? (just better core mass functions?
 filament-core relation? do we really need a _much_ better telescope for that?)

Line surveys (full (accessible) galactic plane?):
- to be benchmarked against existing surveys:
 low-J CO surveys (e.g., SEDIGISM; Dame et al (?))
 emerging wide-field (several square degrees) multi-line surveys (e.g., Bron ea 2018)

- extend frequency coverage to higher frequencies
 -> higher excitation lines
 -> atomic lines (e.h., CI (which might be CCATprime stuff, actually)

- what would be the great new science?
 Chemistry on galactic scales?
 "Warm" molecular gas, e.g. in shocks - build up of clouds/filaments?



(contents taken from a summary by Jens Kaufmann)


Concerning science… I will not go into details, because our group will anyway come up with more good science drivers than I can generate here. I would, however, think that AtLAST has a science niche in the following fields, compared to ALMA:

  • the study of dense cores of <0.1 pc size, because ALMA takes a lot of time to sample many of these;
  • the study of dense filaments of ~0.1 pc size (feel free to put your favorite filament width here…), again because ALMA is not good in covering many of these;
  • the connection of cores and filaments to overall cloud structure on spatial scales >1 pc, because ALMA does not cover many clouds and is bad in recovering extended emission; and
  • an improved multi–wavelength characterization of dust continuum emission, because ALMA has no multi–band observing capability.

This results in technical requirements… Consider you want to study targets out to a distance d.

  • In that case, AtLAST needs to resolve angular scales <21 arcsec * (s / 0.1 pc) / (d / 2 kpc). I would thus say that one should target an angular resolution of 10 arcsec or better.
  • AtLAST also has to have a large instantaneous field of view (FoV). Well–calibrated maps of continuum emission with a resolution of 5 arcmin are available from Planck. Thus I would argue that AtLAST needs an instantaneous FoV >5 arcmin.
  • In fact, the FoV also must be large enough to beat ALMA in mapping speed. I have no easy way to evaluate that criterion.
  • The wavelength coverage should be instantaneous, and it should span a wavelength range that permits to reliably derive dust temperatures. I assume this requires measurements from about 1 mm down to about 500mu wavelength, but this requires more detailed consideration.

These requirements dictate telescope and camera specifications…

  • The beam size at wavelength w and dish size D is b = 5 arcsec * (w / 1mm) / (D / 50m). So, to fulfill b < 10 arcsec at w < 1 mm, we can work with D > 25 m.
  • Provided the field of view is F, we need about N_pix ~ (F / b)^2 detector pixels. This results in N_pix ~ 3550 * (F / 5 arcmin)^2 * (D / 50 m)^2 * (w / 1 mm)^–2. Note that this is an order–of–magnitude estimate, because the chosen detector technology will control the pixel spacing.
  • Assuming we need to go to 450mu wavelength (i.e., 670 GHz) and require a transmission of at least 50%, we need a PWV < 0.6 mm for a reasonable amount of time. On the plateau you get this for 30% of the time, or so.
  • Correspondingly, the surface accuracy should be of order 450 mu / 20 = 22 micron.

I would think that AtLAST would have a particular potential when pushing for short wavelengths. I do not have up–to–date numbers, but I understand that the LMT and the IRAM 30m–telescope have surface accuracies in the range 50–100 mu. AtLAST could beat this well. But APEX achieves 15mu or so over a dish size of 12m. So, AtLAST at 25m dish size is only a factor 2 “better” than APEX. That sounds like a hard sell, unless we can come up with an order–of–magnitude improvement in camera technology and FoV.





Connection to the CGM (Circum-Galactic Medium)

(contents taken from a summary by Claudia Cicone)




Imaging the CGM near and far with AtLAST


Study of extended (radii>10 kpc up to ~100 kpc) molecular and atomic gas reservoirs in 

the so-called "circum-galactic medium" (CGM), so far studied mostly in absorption and in 

other gas phases/at other wavelengths (most current CGM works based on HST-COS 

and more recently MUSE observations).


The CGM is the interface between the ISM and the IGM, and is broadly defined as the 

metal-enriched medium (metallicity, Z>0) within 3 Virial radii. It is the scene where large-scale 

galactic inflows and outflows take place, and it was metal-enriched (and perhaps even generated)

 through multiple explosive feedback episodes during the evolution of the host galaxy and its 

satellite progenitors. A small fraction of the CGM is due to ISM contribution from tiny 



The CGM is clumpy and multiphase: not only warm/ionised gas but also dense clouds. 

There is indication that CGM especially at high-z embeds large fractions of cold and dense 

molecular gas. -> this is where AtLAST comes into play. Sensitive mapping of extended 

regions (a few arcmins)  of the sky with a high sensitivity to diffuse and large-scale 

structures (>20”) is prohibitive for any mm/submm interferometer, including ALMA.


Some CGM-related science questions: 


(i) Feedback (from starbursts and AGNs) and outflows:  galactic-scale outflows can be

generated as a result of powerful explosive feedback episodes from intense star formation

and/or AGN activity. Extended galactic-scale outflows with a significant (and in some

cases even dominant in mass) component of cold and dense molecular gas

have been detected in galaxies near and far up to radii a few kpc: how far do these

outflows extend? Do they escape the halo or rather stall above the disk (thus contributing

to forming the CGM)? Are these outflows responsible for the formation of the CGM?

What is the role of SBs and AGNs in driving these outflows and enriching the CGM?


(ii) Feeding of galaxies and cold streams: cosmological simulations predict that 

massive galaxies in the early Universe are constantly replenished by filamentary 

streams that, by delivering cold gas directly inside the virial radius, can efficiently 

feed star formation. Direct imaging of such streams would constitute an enormous 

advantage compared to metal absorption line observatios, which have so far provided the main 

test-bed of cosmological simulations despite suffering from the limited spatial information and

from the (expected) low covering fraction of the inflows.


(iii) Galaxy assembly through mergers and interactions: galaxy mergers are at the 

basis of our hierarchical model of galaxy formation and evolution. They can be 

responsible for severe galaxy  transformations, either directly - through gravity 

and tidal forces - or indirectly - by activating starbursts and AGNs, hence 

prompting their respective feedback. Galaxy mergers re-distribute material within 

galaxies, on the one hand by inducing inflows of gas that can trigger nuclear 

starbursts and AGNs, and on the other hand by producing off-nuclear gas reservoirs. 

Previous mm/sub-mm campaigns focused mostly on the nuclear regions of local mergers, 

but very little is known about the molecular gas beyond the central few kpc: 

how far does the H2 gas reservoir extend in these sources, and how is it affected by the 

tidal forces during the merging process? Tidal forces can spread molecular gas over very 

extended (> 10 kpc) scales, leading to off-nuclear reservoirs and even generating 

small companions denominated tidal dwarfs.


Observations of the CGM are therefore critical to test theories of galaxy evolution.



Technical feasibility and requirements for AtLAST:


By reading your emails I think the strongest constraints come from galactic science. However,

below I list some requirements that according to me are needed in order to study the CGM

near and far. We are talking mostly line observations, especially CO, [CI] and [CII] (because

need to trace low excitation gas) but continuum observations are also useful to detect any 

star formation activity and small dwarfs embedded in the CGM: 


(i) Very high sensitivity (= deep integrations on targeted sources -> cannot be done in survey mode). 

Ideally line sensitivity should be much better (can we aim x10 times better?) than heterodyne 

receivers currently available on APEX in order to have reasonable integration times. Ok with only a few 

selected atmospheric windows: Band 6, 7, 8 (+9) are crucial. E.g.: 

- CO(2-1) at z~0 in Band 6; 

- [CII] at high z in Bands 6-8(+9): at z~2 in Band9 (630 GHz), at z~3 in Band8 (475 GHz), at z~4-5 in Band 8 and at z~6 in Band 6;

- The two [CI] lines both local and high z are also covered with Bands 6-9. 


(ii) large FoV. The FoV is set by the nearby sources. The most straigthforward targets (from my

perspective) would be local ULIRGs and quasars at z~0.02 and above -> at z~0.02 the scale 

is 0.4 kpc/arcsec, hence to ~100-200 kpc CGM one would need a single-pointing FoV of at least 4-5 arcmin

(these are quite long exposures so better have large FoV than having to do multiple pointings). 

More local sources e.g. Andromeda, magellanic streams, or NGC253, M82 are much brighter

and so can be more easily observed in mapping mode.


(iii) Spatial resolution: this is set by high-z observations: at z~2 (e.g. Spiderweb galaxy) the scale

is 8 kpc/arcsec, hence one needs a beam of ~5 arcsec in order to resolve spatially a structure

of ~80-90 kpc. (but do not need to do evderything with AtLAST: for higher resolution one should 

resort to ALMA).


(iii) Stable baselines since we expect broad lines (~1000 km/s FWZI): what about wobbler/chopping 



(iv) Synergies: ALMA of course: both for follow up but also for directly combining AtLAST

+ ALMA data (need zero spacing baselines and ACA and TP antennas are too low in sensitivity for this

kind of science). Synergy with SKA (HI 21cm at high-z) and VLA (CO1-0 at high-z). 

Synergy with MUSE in the optical domain (Lyalpha nebulae observed with MUSE at z~3-4). 





LST White Papers 


Mapping Spectral Line Survey (Watanabe et al.)

Astrochemistry (Yamamato et al.)


Comments (0)

You don't have permission to comment on this page.