1. FIGGS: The Faint Irregular Galaxy GMRT Survey
Jayaram N Chengalur
NCRA/TIFR
Ayesha Begum
I. D. Karachentsev
Sambit Roychowdhury
S. Kaisin, M. Sharina

2. Near field cosmology from Dwarf Galaxies
Extremely faint dwarfs are particularly interesting in the context of hierarchical galaxy formation models
The smallest objects collapse first, larger galaxies from by merger of smaller ones
The process of galaxy merger is highly inefficient
Every large galaxy should be surrounded by dozens of left over “dwarf galaxies”
Detailed study of these galaxies could lead to insights into the formation and evolution of galaxies in the early universe.
Numerical simulation: Each large Dark Matter halo is surrounded by several, as yet unmerged, smaller halos.

3. The Faint Irregular Galaxy GMRT Survey: FIGGS
GMRT observations of the neutral hydrogen (HI) in a well defined sample of nearby dwarf galaxies.
Selected from KK catalog
65 galaxies identified using the selection criteria
MB > -14.5, HI Flux > 1 Jykm/s,
Dopt > 1', δ > -40o
Accurate distances are known for a large fraction of the sample
complimentary multi-wavelength data (HST, Ground based optical broad band, Ha, GALEX UV…)
By far the largest interferometric HI study of dwarf galaxies

4. The Giant Meterwave Radio Telescope (GMRT)
Hybrid configuration
12 dishes in central compact array
Remaining along 3 “Y” arms
Allows one to simultaneously make low and high resolution images
14 km
1 km x 1 km
Aperture Synthesis Radio Telescope (interferometer)
30 Antennas each 45m in diameter
About 70 km N of Pune, 160 km E of Mumbai.

5. Properties of the FIGGS sample
<D>~ 4Mpc
<MHI> ~ 3 x 107 Msun
<MB> ~ -13

6. FIGGS vs. other large interferometric HI surveys
Average gas fraction of FIGGS galaxies < fgas >~ 0.7
 FIGGS probes the regime of faint, very low mass, gas rich galaxies extending the baseline for a comparative study of galaxy properties

7. FIGGS Data Products HI Spectrum Velocity Field HI (11”) + GALEX UV
CCD R+B
HI (5”) + Ha
HI (28”) +
CCD V + B

8. FIGGS: Expected Outcomes
Study the relationship between dark and baryonic matter in the smallest gas rich galaxies
Dark Matter Halos, Tully-Fisher relation, Baryonic Tully Fisher relation, baryon fraction….
Study the modes of conversion of gas into stars in the smallest units of star formation
Star formation thresholds, recipes, effective yield…

9. Scaling relations for gas disks
Begum et al.  MNRAS,386,1667 (2008)
HI mass correlates tightly with DHI (1 Msun/pc2)
Relationship matches within error bars with that found for large spirals (Broeils & Rhee 1997)
Over 3 orders of magnitude of HI mass, disks of galaxies can be described as being drawn from a family with fixed average SHI
Correlation of HI mass with optical diameter weaker than for spirals
Coupling between gas and star formation in dwarfs not as tight as in spirals?

10. Dark Matter

11. DDO 210 (MB mag)
Is there an ordered component to the gas motions in the
faintest dwarf galaxies?
V ~ 1.6 km/s (GMRT)
Begum & Chengalur 2004 A&A, 413, 525
DV ~ 6.5 km/s (VLA Darray)
Lo et al AJ, 106, 507

12. Large Scale Kinematics of the HI
All FIGGS galaxies show large scale velocity gradients
- Not always consistent with that expected from a rotating disk

13. HI in Cam B (MB ~ -12.0)
Begum, Chengalur & Hopp New Ast, 2003, 8, 267

14. Cam B : Rotation Curve
Begum et al. New Ast, 2003, 8, 267
Peak rotation velocity comparable to velocity dispersion
Important to correct for pressure support before determining the dynamical mass (Model Dependent!)

15. Mass model for CamB
Begum et al. New Ast, 2003, 8, 267
Constant density core provides a good fit, but “NFW” type halo does not
» NWF halos in general do not provide a good fit to our sample galaxies.
Rotation curve derived at resolutions ranging from 300pc to 750 pc
» beam smearing does not seem to be important

16. TF relations
Begum et al., MNRAS,386,138, (2008)
Vrot
Baryonic Mass
I band Magnitude
Compare FIGGS sample with HST large spiral Sample (both samples have accurate distances)
FIGGS galaxies are slightly under luminous for their velocity width, but have about the right amount of baryons (Model dependent conclusion!)
Dwarf galaxies have been less efficient at converting gas into stars

17. Scatter about the TF
Begum et al., MNRAS,386,138, (2008)
HST
FIGGS
The scatter about the BTF/TF (normalized by the measurement error) is much larger for the FIGGS sample as compared to the HST sample
Star formation efficiency is lower and shows more scatter in dwarf galaxies as compared to large spirals

18. Star formation in extremely faint dwarfs

19. Star formation recipes for spirals and starburst galaxies
(Kennicutt ApJ 498, 541, 1998)
Star formation occurs only above a fixed threshold gas column density
Star formation rates determined from Ha observations
Related to instabilities in thin disks?
(Safronov AnAp 23, 979, 1960;Toomre ApJ 139, 1217, 1964)
Above this threshold column density, the star formation rate has a power law dependence on the gas column density
SSFR = (2.5 ± 0.7)10−4(Sgas/1Msunpc−2)1.4±0.15Msunyr−1kpc−2
(Kennicutt ApJ 498, 541, 1998)
Unclear if recipes derived from large spirals apply to dwarf (“primordial”) galaxies

20. Star formation in Dwarfs (As traced by Ha)
(Begum et al MNRAS )
Star formation recipes have traditionally used Ha as the tracer of star formation rate Ssfr
Since observed gas density SHI varies with resolution, all HI images have to be constructed at the same linear resolution
Linear resolution used here is 300pc
There is no one to one correspondence between HI column density and star formation (as traced by Ha)
The observed “threshold” column density is much lower than that predicted from disk instability model

21. Ha vs. HI on fine (~ 20pc scales)
(Begum et al MNRAS )
On very small scales Ha emission is sometimes co-incident with HI peaks, and sometimes with HI “holes”

22. Measuring SFR in faint Galaxies
At low SFR, Ha measurements of SFR are susceptible to stochastic errors
Ha is produced by massive OB stars, i.e. traces the instantaneous SFR
The number of OB stars in small clusters is susceptible to large Poisson fluctuations even if the IMF is universal
UV may be a better indicator of SFR

23. Pixel by Pixel Correlation with FUV
Roychowdhury et al. MNRAS,397,1435, (2009)
Pixel by Pixel Correlation with FUV
Several galaxies show “Schmidt Law” type behaviour but the slope is
Generally much steeper than 1.4
Highly variable from galaxy to galaxy
Power law continues until one reaches the sensitivity limit of the GALEX data
At the highest Sgas, the SSFR approaches the value predicted by the “Kennicutt-Schmidt” law

24. Fine scale structure of the ISM in dwarf galaxies

25. Fine Structure in the HI gas
The HI emission is clumpy and shows considerable structure even on scales of pc.
Expected in turbulent ISM models – BUT Can this be quantified?

26. Power spectrum of HI fluctuations in the faintest dwarf galaxies
HI signal is very weak – estimators used earlier won’t work
Suggest a new estimator, based on correlation between visibilities on adjacent baselines which is well suited to dragging the signal out of the noise
Similar estimator is used in interferometric CMBR work
e.g. Hobson et al. MNRAS 275, 863 (1995)

27. HI fluctuation PowerSpectrum in DDO 210 (MB ~ -10.9)
The power spectrum has a slope a ~ ± 0.45
over scales from 80 – 500 pc
the intensity fluctuations appear to come from modulation of the density
similar slope to that seen in the Milkyway
In thin disks expect a ~ -1.6
Since fluctuations happen in 2D
DDO210 must have a relatively fat disk
Begum et al. MNRAS,372,L33 (2006)
See also Dutta et al. MNRAS,398,887 (2009)

28. Intrinsic axial ratio of dwarf HI disksb
Face on
Edge on
Intermediate
q=b/a
Observed Distribution
Intrinsic Distribution
Invert directly, or using Lucy deconvolution
p=c/a

29. Intrinsic axial ratio of HI disks
Observed axial ratios
Intrinsic axial ratios
Dwarfs have quite fat HI disks <q> ~ 0.57
--Consistent with their higher s/Vrot ratios

30. Thank you

31. Caveat’s
Roychowdhury et al. MNRAS,397,1435, (2009)
Correcting the SFR calibration for low metalicity will increase the deviation from the “Kennicutt-Schmidt” SFR law
Correcting Sgas for possible molecular gas content will have the same effect
To bring agreement with the “Kennicutt-Schmidt” law one will need truncation of the IMF at the upper end.

32. Star formation recipes and galaxy formation simulations
Numerical simulations accurately follow the gravity driven merger of the dark matter halos
Physics of the baryonic material is complex and poorly understood
heating/cooling, collapse to form stars, feed back from star formation
Most simulations of galaxy formation use “recipes” for following the star formation process
Recipes are generally derived from observations of nearby galaxies.
Numerical simulations of galaxy formation in LCDM cosmology are able to produce realistic disk galaxies. While the gravity driven infall and merger are accurately computed, star formation and feedback are incorporated largely via “recipes” (e.g. Fabio et al. MNRAS 2007, 374,1479)