Requirements for investigating IMF of Massive Stars

ESO 3 May 2006 AndreasWicenec MatteoGuainazzi AnitaRichards

InitialMassFunction

Stars form in clusters and we can either look at I clusters where most high-mass stars are already on the main sequence, or at II starforming regions.

1 Young massive main sequence stars.

Selection of Clusters

We start to develop the methods by selecting one or more clusters in the Milky Way with known angular extent, distance modulus and negligible reddening, and examining the spectral energy distribution (series of photometric points) for each star to establish its mass. There are probably tens - hundreds of candidate massive stars per cluster (out of perhaps ten or more times as many detectable stars).

Select interesting but not too well 'solved' clusters with manageable properties e.g. (experts please revise!)

1. Low reddening e.g. E(B-V) < 0.2
2. Distance > 1kpc
3. Angular size < 1 degree
4. Reasonably well-populated

e.g. VOTable of 28 clusters from VII/229A meeting above criteria All 1697 positions from VII/229A

Colour selection

The simplest way to identify massive stars is by colour. F A B and O stars have temperatures from 8 000 to 30 000 K or hotter. This gives (approximately):

• B-V < 0.4
• U-B < 0.0

See Colour Indices for typical colours of stars of each main spectral type.

Construct observed Spectral Energy Distributions

Summary of SED Builder constructed by Pierfederici et al., work ongoing at ESO.

• Python scripting
• Uses ACR
• Current custom GUI interface
  • MySpace unsatisfactory when data and tools are not collocated
  • Could be provided as a series of applications
  • Investigate how to allow parameterised workflow or workflow builder to be used
• Currently searches for data in fixed area; could be user-specified polygon
  • VOTable list of data returned (images, spectra, catalogues)
    • AstroScope search would need to be restricted by waveband, maybe resolution and sensitivity to make list manageable.
    • Useful to be able to retain/edit list of resources matching search
  • User selects desireable data sets (see Data, above and below)
• Run SExtractor on images
  • We need archives to provide SExtractor service and suitable configuration files collocated with their images
• Cross-match SExtractor and other catalogues including lists of spectra (by sky position including errors).
  • Is best way to use TopCat method/STILTs?
• Uses VOSpec method to convert to same flux density/wavelength scale (see Photometry)
  • Generic conversion good enough for getting OBAF but not e.g. distinguish OIV from OV..
  • If high-resolution spectra (e.g. FORS) are to be included in SED, can they be pre-averaged to a single point to improve SNR?
• Produces VOTable of SEDs, UCDs provided

Obtain models

Start with library of precomputed model SEDs for various masses/temperatures etc.

• New Grids of ATLAS9 Model Atmospheres (F. Castelli, R.L. Kurucz, 2004, IAU Symp. No 210). This model can be accessed through the Theroretical Spectra Access Protocol. An interface exists in VOSpec, as well as a web server supported by the Spanish VO
• De-reddeining correction for spectra not available in any spectral display tools, but being currently implemented in VOSpec. It is likely to be available by the IAU (MG plans to present the implementation on the framework of one of the ESA/VO science cases there)
• Also see GAIA models?
• Need to scale models to data. Simple additive constant on flux axis? (choose units carefully?)
  • Ideally incorporate known DistanceModulus for each cluster
  • Crude method - find brightest point in observed SED; find model point at closest wavelength; subtract to calculate offset
  • Ultimate method - find standard star in field of cluster at same distance.

Compare model SEDs with data

• Discard fore/background objects
  • Use published distances, parallax etc. if available
  • SEXtractor star-galaxy discrimination (not complete)
  • Use photometry (see below)
• Manually using VOSpec
• Combine VOSpec and STILTS methods to scale and shift models to each observed SED
• For each observed SED, estimate which model fits best
  • By eye?
  • UCL huntfit3
  • STILTS? R?
• For candidate massive stars of similar type, calculate average magnitude
  • Discard anomalous magnitude objects
• VOSpec recipes

High-resolution spectra

• VOSpec incorporates a first implementation of the Simple Line Access Protocol (SLAP). It allows identification of lines in a IVOA Line Data Model- (LDM-) compliant database on top of a Spectral Energy Distribution displayed vith VOSPec. Currently, there is no elegant way to access the SLAP in a non-interactive way (i.e., through a script - the general lack of web-services in the VO world. One can include in a script an HTTP call, and parse the stream though, a not elegant but in principle possible solution. Of course, any line catalogue to be searched through the SLAP has to have a LDM-compliant interface layer first!
• Models
  • CMFGen, which generates large (5E+08 points) synthetic spectra and associated properties - T, G etc. e.g. output SEDs and Martins et al. A&A 2005
• Spectral fitting - equivalent widths etc. - SpecView ?

Data

Catalogues of clusters etc.

1. accessible to AstroScope (from HEASARC)
2. an alternative solution, allowing access to source catalogues through IVOA (draft) Source Data Model prescriptions, will be presented at the IVOA Victoria INTEROp: VOQuest. Fully integrated in PLASTIC, it allows to retrieve catalogues, and feed them to any PLASTIC-compatible VO applications. The current technological demostratos access Vizier catalogues such as 2MASS or USNO2
3. Vizier
   • VII/229A Optically visible open clusters and Candidates (Dias+ 2002-2005)
   • II/62 Catalogue of Extinction Data (Neckel+, 1980) no distances but all sky
   • J/A+A/359/347 Reddening in open and globular clusters (Dutra+, 2000)
4. A site Devoted to Stellar Clusters in the Galaxy and the Magellanic Clouds much data including distance modulus - is this data available in any form potentially VO-accessible?

Investigating individual Main Sequence hot stars

1. Wavelength range approx. 100 - 2000+ nm (based on models)
   • Optical - UBV for colour-based pre-selection
   • Any available bands (need at least 3 but ideally many more)
   • UV if possible
   • Best available large sky surveys are IR - can we use these?
2. Positional accuracy and resolution better than c. 5 arcsec
   • Aperture corrections probably not needed; available surveys do not resolve hot stars (resolution coarser than 0.1 arcsec, mostly at least 0.5-1 arcsec); worry about close binaries later.
3. Good photometry
4. Minimum sensitivity ** - Galactic hot stars are bright!

• Optical
  • The only optical surveys with good coverage are based on photographic plates. GSC has better photometry than USNO-B, ESO has recent version. Need ADQL access if possible for ColourCutter
    • Bands include optical J, B, F which can be roughly converted to Johnson U B V but coverage is not all-sky in all bands
  • Can we find enough interesting clusters in SDSS, INT-WFS etc?
  • DSS?
• UV
  • The IUE slit is 10"x20" wide. It is too large for studies on interesting cluster of the sample.
  • Can we find clusters with HST or ?? coverage?

Flux/magnitude conversions

• For initial demo use optical data in whatever photometric systems/conversions we can get, e.g. VOSpec can convert magnitudes to flux densities by assuming that they are all on Johnson system.
• For colour cut using GSC or other photgraphic bands, could convert colour cut criteria rather than data?
• Convert between physical units using built in formulae (STILTS?)

Devlop tool for more precise interconversion/conversions:

• VOSpec has an implementation, which deald with standard Johnson-Cousin photometry, constant (i.e., color-independent) zero-point. This can be extended to other photometric system if needed, but it is not deemed to be the final solution.
• Real solution through proper Photometry Data Model is being proposed, but unlikely to be implemented before nect InetrOP meeting (October 2006)
  • Andreas Wicenec to liaise with Eduardo Gonzalez-Solares and Mark Allen over science requirements and available information
    • Review
    • UK SCHMIDT TELESCOPE Passbands for photographic plate photometry * Final requirements to be feedback to the Chair of the Photometry Data Model Working Group (if created at the Victoria Meeting IVOA Interop) - most likely Pedro Osuna
• Optical (also UV and IR when in magnitudes)
  • Magnitudes to physical units
  • Between common magnitude systems
• Produce tool which has sensible defaults at all stages but potentially could handle all of (in order of priority)
  • Apply Zero Point and conversion factor (source-independent instrumental properties)
  • Correct for filter leakage which depends on object brightness
    • and reddening
• X-ray counts to physical units (not needed for basic Galactic case)
  • Three parallel efforts ongoing: XMM-Newton SOC (M.Guainazzi), XMM-Newton SSC (D.Fyfe), Chandra CXC (J.Mc Dowell). None of these project is currently funded
  • Conversion on the basis of fixed spectral index and photoelectic column density already available in X-ray catalogue (surely in the 1XMM and the XMM-Newton Slew Survey Catalogue)
  • Preliminary algorithm to convert X-ray spectra from instrumental count to physical units taking into account the proper color correction (spectral shape) presented at ADASSXVI by M.Guainazzi. It has not bee widely tested or benchmarked, but the software is available upon request.
  • PIMMS/HEASARC does it but the other way round - it cannot be really used.

More useful links

• Definitions
• Evolutionary tracks and isochrones for alpha-enhanced stars

Post-main-sequence stars

A complete census must include stars which are so massive that they have turned off the main sequence very rapidly - WR, RSG, YHG, LBV...

Some (WR, LBV) cn be recognised from optical/UV spectra (SEDs?). RSG are easy to find by IR colours (see AVO case 2005) although additional criteria are needed to distiguish from low-mass ones - but the latter would be fainter and not part of young cluster.

II Ongoing SFR - Identify signs of massive star formation

To follow - see Massive star formation

Action summary

Need to attatch developers!

1. SED builder Andreas + AstroGrid - who?
   • User-specified position and search radius
     • AstroScope but restricted by wavelength regime etc.?
     • Need VOTable of selected data resources
   • Locate SExtractor with suitable configs at image data archives
   • Other MySpace problems?
   • STILTS crossmatch?
2. Magnitude/flux conversions Andreas producing spec, liase rest of VOTECH SWG, Pedro Osuna, and Mark Taylor - or with the IVOA Photometry Working Group, if formally constituted after the Victoria IVOA InterOp meeting
3. Models
   • Kurucz library of stellar atmospheres available as TSAP, and through VOSpec interface
   • GAIA Andreas
   • Fitting to data Anita, Mark Taylor, Jeremy Yates
4. Data
   • What is already available via harvestable registry? Anita
     • It would be much easier if more data was accessible via
       • ADQL or at least waveband-restricted search
       • By sending a list of positions (etc.) as for Vizier
   • Vizier catalogues ?
   • Progress on ESO data Andreas
   • Neither XMM-Newton nor IUE data are scientifically useful for the case
   • Strategy to make the best of current availability
     • Take entire Dias VII/229A list (~1700 clusters) and investigate what data are available within 30 arcmin radius of each cluster in major surveys in 4 or more optical bands for any given cluster:
       • INT-WFS - send list as Vizier query, obs. log only complete to 2002 - SA111 looks suitable
       • SDSS
       • ESO - Andreas
     • Select clusters which have suitable data and are the least reddened, not too close/extended etc.
     • Cone search GCS for each cluster and make colour cut to select candidate massive stars
     • Use multi-band data to construct SEDs for candidates and apply magnitude criteria to eliminate non-members.
5. Develop case Anita and all
6. Present at IAU ?
   • MG, AMSR, maybe AW attending IAU during VO session
   • Likely to focus on use of tools - not definative science result.
7. Write up for deliverable by ?? Anita

-- AnitaRichards - 04 May 2006