FIGGS: Faint Irregular Galaxies GMRT Survey - Overview, observations and first results
Ayesha Begum, Jayaram N. Chengalur, I. D. Karachentsev, M. E. Sharina, S. S. Kaisin
aa r X i v : . [ a s t r o - ph ] F e b Mon. Not. R. Astron. Soc. , 000–000 (0000) Printed 23 October 2018 (MN L A TEX style file v2.2)
FIGGS: Faint Irregular Galaxies GMRT Survey − Overview,observations and first results
Ayesha Begum ⋆ , Jayaram N. Chengalur , I. D. Karachentsev , M. E. Sharina and S. S. Kaisin Institute of Astronomy, University of Cambridge, Madingley Road, Cambridge, CB3 0HA, UK National Centre for Radio Astrophysics, Post Bag 3, Ganeshkhind, Pune 411 007, India Special Astrophysical Observatory, Nizhnii Arkhys 369167, Russia
ABSTRACT
The Faint Irregular Galaxies GMRT Survey (FIGGS) is a Giant Metrewave Radio Tele-scope (GMRT) based HI imaging survey of a systematically selected sample of extremelyfaint nearby dwarf irregular galaxies. The primary goal of FIGGS is to provide a comprehen-sive and statistically robust characterization of the neutral inter-stellar medium properties offaint, gas rich dwarf galaxies. The FIGGS galaxies represent the extremely low-mass end ofthe dwarf irregular galaxies population, with a median M B ∼ − . and median HI mass of ∼ × M ⊙ , extending the baseline in mass and luminosity space for a comparative studyof galaxy properties. The HI data is supplemented with observations at other wavelengths.In addition, distances accurate to ∼
10% are available for most of the sample galaxies. Thispaper gives an introduction to FIGGS, describe the GMRT observations and presents the firstresults from the HI observations. From the FIGGS data we confirm the trend of increasing HIto optical diameter ratio with decreasing optical luminosity; the median ratio of D HI /D Ho forthe FIGGS sample is 2.4. Further, on comparing our data with aperture synthesis surveys ofbright spirals, we find at best marginal evidence for a decrease in average surface density withdecreasing HI mass. To a good approximation the disks of gas rich galaxies, ranging over 3orders of magnitude in HI mass, can be described as being drawn from a family with constantHI surface density. Key words: galaxies: dwarf – galaxies: kinematics and dynamics – radio lines: galaxies
HI 21cm aperture synthesis observations of nearby spiral galax-ies is a mature field with over three decades of history – probablysomething of the order of a thousand galaxies have already beenimaged. However observers have tended to focus on bright ( ∼ L ∗ )galaxies with HI masses & M ⊙ . HI observations of faint dwarfgalaxies (M B & − ) generally require comparatively long inte-gration times, and such galaxies have hence not been studied insimilar numbers. While there have been some systematic HI sur-veys of dwarf galaxies (Swaters 1999; Stil & Israel 2002), thesehave generally been restricted to the brighter (M B e < − ) dwarfs.In hierarchical models of galaxy formation, nearby dwarfgalaxies would, in some ways, be analogs of the primordial build-ing blocks of large galaxies. A systematic HI survey of the faintestdwarf galaxies could provide data that would be useful for a di- ⋆ E-mail:[email protected] verse range of studies, ranging from, for example, testing the pre-dictions of cold dark matter models (e.g Simon & Geha (2007);Blanton et al. (2007)), checking if such systems could be the hostpopulation of quasar absorption line systems (e.g. Zwaan et al.(2005); Kanekar & Chengalur (2005)) etc. As the most chemi-cally unevolved systems in the present-day galaxy population, thefaintest dwarfs provide unique laboratories for understanding starformation and galaxy evolution in extreme environments, i.e. lowmetallicity, low dust content, low pressure, low shear, and low es-cape velocity (e.g. Ekta et al. (2006)).In this paper we describe and present the first results froma Giant Meterwave Radio Telescope (GMRT) based HI imagingstudy of faint dwarf galaxies − the Faint Irregular Galaxies GMRTSurvey (FIGGS). The primary goal of FIGGS is to obtain highquality observations of the atomic ISM for a large, systematicallyselected sample of faint, gas rich, dwarf irregular (dIrr) galaxies.Our GMRT HI images are supplemented by single dish HI obser-vations, HST V and I band images and ground based H α images c (cid:13) Begum et al. from the 6-m BTA telescope. Additionally, the HII region abun-dances and H α rotation curves are being obtained on the WilliamHerschel Telescope (WHT), Isaac Newton Telescope (INT) tele-scopes on La Palma and 6-m Russian BTA telescope, respectively.This paper is organised as follows. In Section 2 we describethe design and the properties of the galaxy sample. The main sci-ence drivers for FIGGS are described in Section.3. The GMRT ob-servations are described in Section 4 and the results of the surveyare presented and discussed in Section 5. The Faint Irregular Galaxies GMRT Survey − FIGGS, is a largeobserving program aimed at providing a comprehensive and statis-tically robust characterisation of the neutral ISM properties of faint,gas rich, dwarf irregular galaxies using the Giant Metrewave RadioTelescope (GMRT). The FIGGS sample forms a subsample of theKarachentsev et al.(2004) catalog of galaxies within ∼
10 Mpc.Specifically, the FIGGS sample consists of 65 faint dwarf irregular(dIrr)galaxies with:(i) absolute blue magnitude, M B e > − . ,(ii) HI flux integral > − (iii) optical B band major axis e > < − fromthe tip of the red giant branch (rgb), from membership in a groupwith known distance (grp), from the Tully-Fisher relation (tf), andfrom the Hubble flow (h). Column(9) gives the group membershipof the galaxy, Column(10) the inclination determined from opti-cal photometry (and assuming an intrinsic thickness, q o =0.2) andColumn(11) the reference for the (B-V) colour, and/or revised dis-tance. The data presented in the Table 1 (except for the colour) istaken from Karachentsev et al. (2004) catalog, except that reviseddistances have been adopted, if available. As can be seen from theTable 1, tip of the red giant branch (rgb) distances (which are gen-erally accurate to ∼ B ), distance, HI mass, and HI mass to light ratio (M HI / L B ) forthe FIGGS sample, while Figure 2 compares the distributions ofgas fraction, luminosity and dynamical mass of the FIGGS galaxieswith that of existing samples of galaxies with HI aperture synthe-sis observations. The gas fraction and the dynamical masses for theFIGGS sample have been derived from the GMRT observations. The FIGGS sample has a median M B ∼ − and a median HImass ∼ × M ⊙ , while spanning range of more than 100 in stel-lar light, gas mass and dynamical mass, and more than 4 in gas frac-tion. It can also be clearly seen that by focusing on fainter, lowermass galaxies than those observed in previous HI imaging stud-ies, FIGGS bridges the transition to rotation dominated low massspirals and provides a substantially extended baseline in mass andluminosity space for a comparative study of galaxy properties. The aim of FIGGS is to provide a large multi-wavelength databasefor a systematically selected sample of extremely faint dwarf irreg-ular galaxies. As mentioned in Section 1, such a database could beused to address a diverse range of astrophysical questions. Ratherthan attempting to enumerate all of these, in this section, we de-scribe in some detail a couple of key science drivers for the FIGGSsurvey.
One of the main goals of FIGGS is to use the HI interferomet-ric images in conjunction with the optical data to study the inter-play between the neutral ISM and star formation in the faintest,lowest mass, gas rich dIrr galaxies. The gravitational binding en-ergy for very faint dwarf irregular galaxies is not much larger thanthe energy output from a few supernovae; consequently star for-mation in such galaxies could have a profound effect on the mor-phology and kinematics of the ISM of these systems. The FIGGSdata will enable us to study the ISM of most of our sample galax-ies at a linear resolution of ∼ − pc − i.e. comparable tothe scales at which energy is injected into the ISM through super-nova and stellar winds. FIGGS thus provide a unique opportunityto study the effects of feedback from star formation in low mass,gas rich galaxies, which in turn will allow us to understand the pro-cesses driving the evolution of these galaxies. For example, it hasbeen suggested that star formation in dwarf galaxies occurs onlyabove a constant threshold HI column density of N HI ∼ cm − (e.g. Skillman 1987; Taylor 1994). Such a threshold couldbe a consequence of disk dynamics (e.g. related to Toomre’s insta-bility criterion; Kennicutt (1989)) or a consequence of some otherphysical process, e.g. self shielding or thermo-gravitational insta-bility (Schaye (2004)). A preliminary study of a small subsam-ple of FIGGS (Begum et al. (2006)) suggested that unlike brighterdwarfs, the faintest dwarf galaxies do not show well defined thresh-old density. A detailed comparison of H α and UV images with HIcolumn density maps for the FIGGS sample will allow us to defini-tively answer the issue of the existence of a threshold density in thefaintest galaxies and also to check whether the recipes for star for-mation derived from larger galaxies (Kennicutt (1989)) continue tobe valid at this mass regime. These are critical issues in hierarchicalgalaxy formation models. The second major aim of this survey is to study the relation be-tween dark and baryonic matter in the smallest known star form-ing galaxies. According to several models of galaxy formation andevolution, the first burst of star formation in dwarf galaxies be-low a critical halo circular velocity ( ∼
100 kms − ) could lead tothe loss of a significant fraction of baryons (e.g. Efstathiou 2000; c (cid:13) , 000–000 IGGS: Faint Irregular Galaxies GMRT Survey Table 1.
The FIGGS sampleGalaxy α (J2000) δ (J2000) M B D Ho B-V Dist D estm Group i opt
Ref(h m s) ( ◦ ′ ′′ ) (mag) ( ′ ) (mag) (Mpc) (deg)SC 24 00 36 38.00 −
32 34 28 − − − −
22 15 01 − −
21 00 58 − − − − − − − − − − − − −
25 59 59 − −
04 12 30 − − − − − − − − − − − − − − − −
01 51 49 − ∗ − − ∗ − − − − −
33 33 29 − − −
38 13 53 − − − − − − − − − − − − −
31 31 45 − − −
30 58 20 − − −
29 14 11 − ∗ − − −
28 53 40 − − − − − − −
46 18 06 − ∗ − − − −
31 40 54 − − .c (cid:13) , 000–000 Begum et al.
Table 1. ( continued ) The FIGGS sampleGalaxy α (J2000) δ (J2000) M B D Ho B-V Dist D estm Group i opt
Ref(h m s) ( ◦ ′ ′′ ) (mag) ( ′ ) (mag) (Mpc) (deg)KK 250, UGC11583 20 30 15.30 +60 26 25 − − −
12 50 57 − −
32 23 26 − − [A] [B] [D][C] Figure 1.
The histogram of M B (panel [A]), distance (panel [B]), logarithm of the HI mass (panel [C]) and the HI mass to light ratio, M HI / L B (panel [D])for the FIGGS sample. Dekel & Woo 2003). In fact, expulsion of gas because of energyinput from supernovae has been postulated as a possible mecha-nism to produce dwarf elliptical galaxies from gas rich progenitors(e.g. Miralda-Escude & Rees (1997)). Although a complete expul-sion of the ISM from galaxies has not been observed so far, expan-sive outward motions of the neutral gas in dwarf galaxies has beenobserved in at least two galaxies (viz. GR8, Begum & Chengalur2003; NGC 625, Cannon et al. 2004). To test these models, highspatial resolution interferometric observations are crucial.The Tully-Fisher (TF) relation demonstrates the existence ofa tight relation between dark and luminous matter in bright spi-ral galaxies. Mcgaugh et al. (2000) (see also McGaugh (2005))showed that dwarf galaxies deviate from the TF relation defined bybright spirals, but that the relationship is restored if one works withthe total baryonic mass instead of the luminosity, i.e. a “BaryonicTully Fisher” (BTF) relation. The FIGGS sample, both becauseit extends well beyond the region of rotation dominated dwarfsand because accurate distances are known for a large subsample,forms a very interesting dataset for studying TF and BTF relations.Most of the past studies have been done using the HI global ve-locity widths from the single dish observations (Geha et al. 2006;Mcgaugh et al. 2000). While for the brighter galaxies W (the ve-locity width at 20% emission, after correction for random motionsand instrumental broadening), is a good measure of the rotationalvelocity of the galaxy (Verheijen & Sancisi 2001); it is unclear ifthis would remain true in the case of faint dwarf galaxies, whererandom motions could be comparable to the peak rotational veloci-ties (e.g. Begum et al. 2003; Begum & Chengalur 2004a). For suchgalaxies, it is important to accurately correct for the pressure sup-port (“asymmetric drift” correction) for which one needs to know both the rotation curve as well as the distribution of the HI gas, bothof which can only be obtained by interferometric observations suchas in FIGGS. The FIGGS sample would thus allow us to concretelyanswer this question using actual observational data.The HI kinematics of FIGGS galaxies, in conjunction with theH α rotation curves can be used to accurately determine the den-sity distribution of the dark matter halos of faint galaxies. Sincestars generally make a minor contribution to the total mass in theFIGGS galaxies, accurate kinematical studies can provide direct in-formation on the density profiles of their dark matter halos with lessuncertainties arising from the unknown stellar mass to light ratio.Cosmological simulations of hierarchical galaxy formation predicta “universal” cusped density core for the dark matter halos of galax-ies (e.g. Navarro et al. 2004). On the other hand, observations ofdIrr galaxies indicate a constant density core for their dark matterhalos (e.g. Weldrake et al. 2003; de Blok et al. 2003); however thiscomparison remains controversial (e.g. van den Bosch & Swaters2001; de Blok 2005). FIGGS would not only provide a large sam-ple for such a comparison, but would also provide a data set thatis less subject to uncertainties due to the unknown stellar mass tolight ratio or large scale non circular motions due to bars or spiralarms. For all the GMRT HI observations, the observing bandwidth of1 MHz was divided into 128 spectral channels, yielding a spec-tral resolution of 7.81 kHz (velocity resolution of 1.65 km s − ).It is worth noting that this velocity resolution is ∼ c (cid:13) , 000–000 IGGS: Faint Irregular Galaxies GMRT Survey [A] [B] Figure 2.
The gas fraction of FIGGS galaxies (circles) plotted as a function of the absolute blue magnitude (left) and dynamical mass (right). FIGGS galaxieswith TRGB distances are shown as solid circles, whereas the remaining FIGGS galaxies are shown as empty circles. The same quantity is also plotted for thegalaxies in literature with interferometric HI maps. The gas fraction (f gas ) is defined as f gas = M gas / (M gas +M ∗ ). M gas , is computed by scaling the HI massby 1.33 to account for the primordial He fraction. No correction is made for the molecular gas. To compute the stellar mass, M ∗ , the stellar mass to light ratioin the B band ( Γ ∗ ) was derived from the observed (B-V) colour , using from the galaxy evolution models of Bell et al.(2003) and assuming a “diet” SalpeterIMF. Solid triangles are from McGaugh(2005), solid squares from Verheijen(2001), crosses from Swaters (1999) and empty triangles from Cˆot´e et al.(2000).Note how the GMRT FIGGS sample extends the coverage of all three galaxy properties. D E C L I NA T I O N ( J2000 ) RIGHT ASCENSION (J2000)07 28 30 25 20 15 10 05 0040 49484746454443 D E C L I NA T I O N ( J2000 ) RIGHT ASCENSION (J2000)07 28 25 20 15 1040 48 0047 300046 300045 300044 3000 D E C L I NA T I O N ( J2000 ) RIGHT ASCENSION (J2000)07 28 35 30 25 20 15 10 05 00 27 5540 504948474645444342 [A] [B] [C]
Figure 3.
The figure shows the integrated HI emission from one of the galaxy in FIGGS sample, DDO 43 at various resolutions viz. 46 ′′ × ′′ (panel [A] ),32 ′′ × ′′ (panel [B] ) and 15 ′′ × ′′ (panel [C] ). The first contour level and contour separation for these resolutions are (1.3,12.2), (2.5,18.1) and (4.0,26.2),respectively, in units of 10 cm − . ter than most earlier interferometric studies of such faint dwarfgalaxies (e.g.Lo et al. (1993)). This high velocity resolution is cru-cial to detect large scale velocity gradients in the faintest dwarfgalaxies (e.g. Begum et al. (2003); Begum & Chengalur (2004a)).For each observing run, absolute flux and bandpass calibration wasdone by observing one of the standard flux calibrators 3C48, 3C286 and 3C147, at the start and end of the observations. For the sam-ple galaxies with low LSR velocities, particular care was taken tochoose a bandpass calibrator which does not have any absorptionfeature in the relevant velocity range. The phase calibration wasdone once every 30 min by observing a nearby VLA phase calibra-tor source. c (cid:13) , 000–000 Begum et al.
Table 2.
Parameters of the GMRT observationsGalaxy Date of Observations Velocity Coverage Time on Source synthesised Beam Noise Phase Cal Cont Noise(km s − ) (hours) (arcsec ) (mJy) (mJy)DDO 226 8 July 2004 257 −
469 3.5 52 × , 26 × , 19 × −
401 5.0 50 × , 26 × , 16 × −
263 3.5 42 × , 36 × , 27 × − −
166 4.0 41 × , 30 × , 16 × −
529 5.0 41 × , 28 × , 16 × −
415 3.0 46 × , 21 × , 13 × − −
176 3.7 47 × , 32 × , 16 × − − × , 26 × , 16 × − −
60 3.0 53 × , 31 × , 20 × − −
216 4.5 53 × , 34 × , 22 × −
616 3.5 49 × , 35 × , 24 × −
421 5.0 42 × , 28 × , 18 × −
460 3.5 46 × , 32 × , 15 × −
385 5.0 41 × , 27 × , 19 × −
447 4.0 42 × , 34 × , 18 × −
569 4.0 56 × , 42 × , 28 × −
657 5.0 41 × , 27 × , 15 × −
644 4.5 41 × , 27 × , 18 × −
650 3.0 41 × , 34 × , 22 × − −
69 3.0 46 × −
846 5.0 42 × , 34 × , 20 × − − × , 23 × , 16 × −
356 5.0 42 × , 28 × , 21 × −
318 4.0 49 × , 27 × , 18 × −
348 3.5 45 × , 28 × , 19 × −
746 4.5 54 × , 33 × , 19 × −
258 3.5 49 × −
719 3.5 51 × , 28 × , 15 × − −
174 6.5 45 × , 27 × , 18 × −
422 3.0 45 × , 28 × , 24 × −
589 4.5 41 × , 28 × , 19 × −
301 4.0 45 × , 31 × , 20 × −
416 7.0 43 × , 29 × , 16 × −
724 6.0 40 × , 27 × , 19 × −
324 6.0 46 × , 28 × , 17 × −
269 3.5 51 × , 29 × , 19 × −
672 4.5 62 × , 30 × , 18 × −
593 5.0 48 × , 32 × , 21 × − −
167 7.0 42 × , 32 × , 18 × −
686 2.5 48 × , 26 × , 18 × −
380 2.5 44 × , 25 × , 16 × −
308 5.5 50 × , 26 × , 17 × −
686 2.7 48 × , 26 × , 15 × −
298 6.5 42 × , 31 × , 17 × −
333 3.5 41 × , 30 × , 21 × −
259 2.5 46 × , 30 × , 19 × −
497 3.0 48 × , 26 × , 15 × −
469 2.5 61 × , 36 × , 15 × − − −
32 6.0 42 × , 32 × , 15 × The GMRT data were reduced in the usual way using the stan-dard tasks in classic AIPS. For each run, bad visibility points wereedited out, after which the data were calibrated. The GMRT doesnot do online doppler tracking – any required doppler shifts haveto be applied during the offline analysis. However, for all of thesample galaxies, the differential doppler shift over our observinginterval was much less than the channel width, hence, there was noneed to apply any offline correction. The GMRT has a hybrid con-figuration (Swarup et al. 1991) with 14 of its 30 antennas located in a central compact array with size ≈ ≈ λ at 21cm) andthe remaining antennas distributed in a roughly “Y” shaped con-figuration, giving a maximum baseline length of ≈
25 km ( ≈ λ at 21 cm). The baselines obtained from antennas in the centralcompact array are similar in length to those of the “D” array of theVLA, while the baselines between the arm antennas are compara-ble in length to the “B” array of the VLA. A single observation withthe GMRT hence yields information on both large and small angu-lar scales. Data cubes at a range of angular resolutions were made c (cid:13) , 000–000 IGGS: Faint Irregular Galaxies GMRT Survey K il o W av l ng t h Kilo Wavlngth-80 -60 -40 -20 0 20 40 60 80806040200-20-40-60-80
Figure 4.
The figure shows the (u,v) coverage for UA 92, the sample galaxywith the shortest on source integration time (viz. 2.1 hours) using appropriate (u,v) ranges and tapers. In this paper we presentonly the low resolution HI images, i.e. made using (u,v) ranges of ∼ − λ , 0 −
10 k λ and 0 − k λ . Higher resolution observationsof the FIGGS sample will be presented in the companion paper.To obtain the low resolution HI images for the sample galaxies,the uv-taper at each (u,v) range was adjusted to achieve as close aspossible to a circular synthesized beam. A low resolution data cubewas generated for each galaxy, using the AIPS task IMAGR, andthe individual channels were inspected using the task TVMOVIEto identify the channels with HI emission. Emission was detectedfrom all of the galaxies in our sample, except for SC 24, HS 117,KK 127 and KKR 25. Apart from HS 117, all of these galaxies werepreviously claimed to be detected by single dish observations. TheGMRT data suggest that the previous flux measurements were spu-rious, probably as a result of confusion with galactic emission. Thegalaxies KK 127 and SC 24 are likely to be distant dwarf irregu-lar galaxies whereas KKR 25 is a normal dwarf spheroidal galaxy(Begum & Chengalur (2005); Karachentsev et al. (2006)). In thecase of HS 117, single dish observations did not detect this galaxy(Huchtmeier & Skillman (1998)). The HI data given in Karachent-sev et al. (2002) is a result of misidentifying galactic HI emission asemission from HS 117. For the rest of the galaxies in the sample,frequency channels with emission were identified and the contin-uum maps were made at both low (26 ′′ × ′′ ) and high (5 ′′ × ′′ )resolutions using the average of the remaining line free channels.No extended or compact emission was detected from the disk of anyof our sample galaxies. All other continuum sources lying with thefield of view were subtracted using the task UVSUB. After contin-uum subtraction, deconvolved data cubes of the line emission weremade at a range of resolutions using the AIPS task IMAGR.HI images at both high and low spatial resolutions are crucialfor a complete understanding of the properties of the atomic ISM offaint dwarf galaxies. As an example, Figure 3 shows the integratedHI column distribution at various resolutions for one of the FIGGSgalaxies DDO 43. This galaxy shows a faint, extended HI envelopewhich is only seen clearly in the lowest resolution HI maps. On theother hand, DDO 43 also has a large hole in the center (see also the VLA observations in Simpson, Hunter, & Nordgren (2005)), whichis seen in the high resolution HI map. However this hole in the HIdistribution is not at all obvious in the low resolution HI maps dueto the beam smearing.The setup and observational results for 49 galaxies from theFIGGS sample are given in Table 2. For the remaining 15 samplegalaxies, the details of the observations and data analysis can befound in Begum et al. 2003; Begum & Chengalur 2003, 2004a,b;Begum et al. 2005; Begum & Chengalur 2005; Begum et al. 2005,2006 and Chengalur et al. 2008 (in preparation). In the case ofUGCA 438, most of the short baselines were missing because ofthe non availability of some of the GMRT antennas during theobserving run, thus missing the diffuse, extended emission fromthe galaxy. Future observations of this galaxy are planned. Wehave not considered this galaxy for the analysis in this paper. Thecolumns in Table 2 are as follows: Column(1) the galaxy name,Column(2) the date of observations, Column(3) the velocity cov-erage of the observation, Column(4) the total integration time onsource, Column(5) the synthesized beam sizes of the data cubes,Column(6) the rms noise per channel for the different resolutiondata cubes, Column(7) the phase calibrator used, Column(8) the σ limits on continuum emission from the galaxy at resolutions of(26 ′′ × ′′ ) and (5 ′′ × ′′ ) respectively. We note that although forsome of the sample galaxies the on-source integration time is short( ∼ − . hours), the hybrid configuration of the GMRT leads toa reasonable sampling of the (u,v) plane. As an example, Figure 4shows the (u,v) coverage for UA 92, the sample galaxy with theshortest on source integration time (viz. 2.1 hours).We examined the line profiles at various locations in thegalaxy and found that they were (to zeroth order) symmetric andsingle peaked. For some galaxies, in the very high column densityregions, a double gaussian and/or a gauss-hermite fit does providea somewhat better description of the data, but even in these regions,the mean velocity produced by the moment method agrees withinthe errors with the peak velocity of the profile. Since we are inter-ested here mainly in the systematic velocities, moment maps pro-vide an adequate description of the data. Moment maps (i.e. mapsof the total integrated flux (moment 0), the flux weighted velocity(moment 1) and the flux weighted velocity dispersion (moment 2))were made from the data cubes using the AIPS task MOMNT. Toobtain the moment maps, lines of sight with a low signal to noise ra-tio were excluded by applying a cutoff at the σ level, ( σ being therms noise level in a line free channel), after smoothing in velocity(using boxcar smoothing three channels wide) and position (usinga gaussian with full width at half maximum (FWHM) ∼ timesthat of the synthesized beam). Maps of the velocity field were alsomade in GIPSY using single gaussian fits to the individual profiles.The velocities produced by MOMNT in AIPS are in reasonableagreement with those obtained using a single gaussian fit. Note thatthe AIPS moment 2 map systematically underestimates the velocitydispersion (as obtained from gaussian fitting) particularly near theedges where the signal to noise ratio is low. This can be understoodas the effect of the thresholding algorithm used by the MOMNTtask to identify the regions with signal. A detailed analysis of FIGGS data will be presented in companionpapers. Here we restrict ourselves to a preliminary analysis of theglobal HI and optical properties of the FIGGS sample.The global HI profiles for our sample galaxies, obtained from c (cid:13) , 000–000 Begum et al.
Table 3.
Results from the GMRT observationsGalaxy FI
GMRT V sys ∆ V D HI M HI M HI L B D HI D Ho FI GMRT FI SD i HI Ref(Jy kms − ) (kms − ) (kms − ) ( ′ ) (10 M ⊙ ) (deg)And IV 19.5 ± ± ± ± ± ± ± ± − ± ± ± ± ± ± ± ± ± ± ± ± ± ± − ± ± − ± ± ± ± ± ± ± ± ± − ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± − ± ± − ± ± ± ± ± ± ± − ± ± ± − ± − ± ± ± − ± ± − ± ± − ± ± − ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± − ± ± ± ± ± ± ± ± ± ± ± ± ± ± − ± ± − ± ± − ± ± ± ± ± − ± ± − ± ± ± ± ± − ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± − ± ± ± − ± ± (cid:13) , 000–000 IGGS: Faint Irregular Galaxies GMRT Survey Figure 5.
The global HI profiles of the sample galaxies obtained from the lowest resolution data cubes (see Table 2). the coarsest resolution data cubes (see Table 2) are shown in Fig-ure 5. The parameters derived from the global HI profiles for thewhole FIGGS sample are listed in Table 3. The columns are as fol-lows: (1) the galaxy name, (2) the integrated HI flux along withthe errorbars, (3) the central heliocentric velocity (V sys ), (4) thevelocity width at 50% of the peak ( ∆ V ), (5) the HI diameter (inarcmin) at a column density of ∼ atoms cm − (D HI ), (6) the derived HI mass (M HI ), (7) the HI mass-to-light ratio (M HI / L B ),(8) the ratio of the HI diameter to the Holmberg diameter. (9) the ra-tio of the GMRT flux to the single dish flux (FI/FI SD ), (10) the incli-nation as measured from the HI moment 0 maps (i HI ), and (11) thereference for the single dish fluxes.As seen in Column(9) in Table 3, the HI flux measured fromthe GMRT HI profiles for most FIGGS galaxies, in general, agree c (cid:13) , 000–000 Begum et al.
Figure 5. ( continued )The global HI profiles of the sample galaxies obtained from the lowest resolution data cubes (see Table 2). (within the errorbars) with the values obtained from the single dishobservations. The average ratio of GMRT flux to single dish fluxis . . This indicates that in general no flux was missed becauseof the missing short spacings in our interferometric observations.However, for some galaxies the integrated flux derived from theGMRT observations is significantly smaller than the single dishvalues. The GMRT fluxes could be lower than those obtained from single dish measurements either because of (i) a calibration error or(ii) a large fraction of HI being in an extended distribution that isresolved out, or (iii) the single dish flux is erroneous, possibly be-cause of confusion with galactic emission. However, the flux of thepoint sources seen in the GMRT images are in good agreement withthose listed in NVSS, indicating that our calibration is not at fault.We note that in the case of KKH 12, KKH 6, UGC 6456, UGCA 92 c (cid:13) , 000–000 IGGS: Faint Irregular Galaxies GMRT Survey Figure 5. ( continued )The global HI profiles of the sample galaxies obtained from the lowest resolution data cubes (see Table 2). and KK 41 there is a strong local HI emission at velocities veryclose to the systemic velocities, making it likely that the single dishintegrated flux measurements for these galaxies were contaminatedby blending of their HI emission with that of the galactic emission.In the case of KK 246, its HI spectrum was near the edge of theGMRT observing band, hence the flux could not be reliably esti-mated.The GMRT integrated HI emission of the sample galaxies, ob-tained from the coarsest resolution data cubes (see Table 2), over-layed on the optical Digitized Sky Survey (DSS) images are shownin Figure 6.The HI morphological inclinations (i HI ) for our sample galax-ies were estimated from the integrated HI maps by fitting ellipticalannuli to the HI images at various resolutions. For sample galaxieswhich have HI disks less extended than 2 synthesised beams (acrossthe diameter of the galaxy) at the lowest HI resolution, could inprinciple be derived from the higher resolution HI maps. However,for most sample galaxies ellipse fitting to the high resolution HImaps is not reliable because of clumpiness in the central high col-umn density regions. The derived inclination (without applying anycorrection for the intrinsic thickness of the HI disk) is given in Col-umn(10) in Table 3. Figure 7 shows a comparison between the mor-phological inclination derived from the optical and HI isophotes ofthe galaxy. No correction has been applied for the intrinsic thick-ness of the disk in both cases. The solid line shows a case whenboth inclinations are the same. We find that for 6 galaxies the HIinclination is significantly greater than the optical inclination (viz.KKH 11, UGC 6456, NGC 3741, UGC 8055, KK 230, KK 250).On the other hand, for many galaxies the optical inclination is foundto be systematically higher than the inclination derived from the HImorphology. This result, if interpreted literally, suggest that the HIdisks of these galaxies are thicker than the disks of their opticalcounterparts. However, we caution that a proper analysis using de-convolved angular sizes of the the HI disks needs to be done beforea firm conclusion can be drawn.The diameter of the HI disk at a column density of N HI ∼ atoms cm − (except for UGCA 92 where the HI di-ameter is measured at N HI ∼ atoms cm − ) estimated fromthe lowest resolution integrated HI emission maps is given in Col-umn(7) of Table 3. The ratio of the HI diameter to the optical(Holmberg) diameter for the sample is also given in Column(9) ofthe same table. Figure 8 shows the histogram of the derived HI ex-tent of FIGGS at N HI ∼ cm − , normalised to the Holmbergdiameter of the galaxy. The median HI extent of the FIGGS sam-ple (normalised to Holmberg diameter of the galaxy) is 2.4. For acomparison, Hunter (1997) using the data compiled from the lit-erature for comparatively bright Im galaxies found that the ratioof D HI / D Ho is somewhat smaller, viz. 1.5 −
2. The extreme out-liers in Figure 8 is NGC 3741, our FIGGS data show it to havean HI extent of ∼ Ho (Holmberg diameter). Follow-up WSRT+DRAO+GMRT observations resulted in HI being de-tected out to ∼ Ho − NGC 3741 has the most extended HIdisk known. For NGC 3741 the rotation curve could be derived outa record of ∼
44 times the disk scale length and from the last mea-sured point of the rotation curve we estimate the dynamical mass tolight ratio, M D / L B ∼ − which makes it one of the “darkest”irregular galaxies known (Begum et al. 2005, 2008).Figure 9 shows a tight correlation between HI mass and the HIdiameter, measured at N HI of 1 × cm − for FIGGS sample.The galaxies in FIGGS sample with accurate distances are shownas solid points, whereas the remaining galaxies are shown as opencircle. The best fit to the whole FIGGS sample shown as a solid linegives log(M HI ) = (1 . ± . HI ) + (6 . ± . (1)The best fit relation was also derived using only the galaxies withTRGB distances, however no significant difference was foundbetween the best fit parameters derived in this case and thatderived using the whole sample. Eqn.(1) implies that the HI disksof the FIGGS sample are well described as having an constantaverage surface mass density ∼ . M ⊙ pc − . A tight correlationbetween HI mass and the size of the HI disk has been noted c (cid:13) , 000–000 Begum et al.
Figure 6.
The GMRT integrated HI column density distribution (contours) overlayed on the optical DSS images (grey scales) of FIGGS galaxies from thelowest resolution data cubes (see Table 2). The contours are uniformly spaced. The first contour level and the contour separation are printed below the galaxyname in units of 10 cm − . c (cid:13) , 000–000 IGGS: Faint Irregular Galaxies GMRT Survey Figure 6. ( continued ) The GMRT integrated HI column density distribution (contours) overlayed on the optical DSS images (grey scales) of FIGGS galaxiesfrom the lowest resolution data cubes (see Table 2).The first contour level and the contour separation are printed below the galaxy name in units of 10 cm − .c (cid:13) , 000–000 Begum et al.
Figure 6. ( continued ) The GMRT integrated HI column density distribution (contours) overlayed on the optical DSS images (grey scales) of FIGGS galaxiesfrom the lowest resolution data cubes (see Table 2). The first contour level and the contour separation are printed below the galaxy name in units of 10 cm − . earlier for spiral galaxies (e.g. Broeils & Rhees (1997)) and forbrighter dwarf galaxies (Swaters 1999). For these samples the HIdiameter was measured at a slightly higher column density, viz.1 M ⊙ pc − . For the FIGGS galaxies, the relationship betweenthe HI mass and the HI diameter measured at 1 M ⊙ pc − is log(M HI ) = (1 . ± . HI ) + (6 . ± . ,for comparison, Broeils & Rhees (1997) measure log(M HI ) = (1 . ± . HI ) + (6 . ± . . Thefit coefficients overlap within the error bars. Hence from theFIGGS data we find that there is at best marginal evidence for adecrease in average HI surface density with decreasing HI mass;to a good approximation, the disks of gas rich galaxies, rangingover 3 orders of magnitudes in HI mass, can be described as beingdrawn from a family with constant HI surface density. The HImass also correlates with the optical (Holmberg) diameter (shownin Fig. 10), although with a larger scatter. A linear fit with a slopeand intercept of 1.74 ± ± HI and the optical diameter, also seen in sample of brighter dwarfs(e.g. Swaters (1999)), is probably indicative of a looser couplingbetween the gas and star formation in dwarfs, compared to that inspiral galaxies.Figure 11 shows the HI mass to light ratio for the FIGGS sam-ple plotted as a function of the HI extent, D HI / D Ho . A trend of anincrease in the M HI / L B with an increase in the HI extent of thegalaxies is clearly seen. The best fit to the FIGGS sample shown asa solid line gives log( M HI L B ) = (1 . ± . HI D Ho ) + ( − . ± . (2)van Zee et al.(1995) from a HI mapping of a sample of low lumi-nosity galaxies also found an evidence of an extended HI extent forhigh M HI / L B galaxies.Figure 12 shows M HI / L B for the FIGGS sample as a function Figure 7.
A comparison of the morphological optical and HI inclination ofthe FIGGS sample. The solid lines shows the case when the two inclinationsare the same. of M B . The same quantity for several other spiral and dwarf galax-ies, spanning a range in absolute B magnitude from M B ∼ − to M B ∼ − is also plotted. The sample from which these galax-ies have been drawn are listed in the figure caption. The galaxiesin FIGGS sample with TRGB distances are shown as solid circles,whereas the remaining FIGGS galaxies are shown as open circles.The solid line shows an empirically determined upper envelope forM HI / L B as a function of a M B from Warren et al. (2007). This up-per envelope can be interpreted as a minimum fraction of the total c (cid:13) , 000–000 IGGS: Faint Irregular Galaxies GMRT Survey Figure 8.
The histogram of the extent of the HI disk (measured at N HI =1 × cm − ), normalised to the Holmberg diameter for the FIGGSsample. Figure 9.
The HI mass for the FIGGS sample versus the HI diameter (mea-sured at N HI ∼ cm − ). The solid line represents the fit to the datapoints. Galaxies in FIGGS sample with TRGB distances are shown as solidpoints, while the remaining galaxies in the sample are shown as open cir-cles. baryonic mass which needs to be converted into stars in order fora galaxy of a given baryonic mass to remain gravothermally sta-ble (Warren et al. (2007)). It is interesting to note that except forAnd IV, all FIGGS galaxies lie much below this upper envelope.This implies that these galaxies have converted much more baryonsinto stars than the minimum required for remaining stable. In thiscontext, it is interesting to note that the average gas fraction for theFIGGS sample is 0.7. Thus, for the majority of the dwarf galaxiesin our sample, the baryonic mass is dominated by gas, rather thanstars. Figure 10.
The HI mass for the FIGGS sample versus the Holmberg diam-eter. The solid line represents the fit to the data points. Galaxies in FIGGSsample with TRGB distances are shown as solid points, while the remaininggalaxies in the sample are shown as open circles.
UA292 DDO154E215-G009
Figure 11.
The log of HI mass to B band light ratio for the FIGGSsample plotted as a function of the extent of the HI disk (measured atN HI ∼ × cm − ) normalised to the Holmberg diameter. Crossesshow three additional galaxies from the literature with high M HI / L B andextended HI disks, UA292 (Young et al. 2003), ESO215-G?009 (Warren etal. 2004) and DDO 154 ( Carignan & Purton 1998). In order to investigate the environmental dependence of the HIcontent for FIGGS galaxies, we plot M HI / L B for FIGGS sampleas a function of tidal index (TI) (Figure 13). Some additional gasrich galaxies with known HI extent are also plotted in the figure. TIis taken from Karachentsev et al. (2004) and it represents the localmass density around a given galaxy, estimated using a large sampleof galaxies within ∼
10 Mpc of the Milky Way. A negative value ofTI for a galaxy indicates that the galaxy is isolated, whereas a pos- c (cid:13) , 000–000 Begum et al.
Figure 12.
The log of HI mass to light ratio vs. B band absolute mag-nitude. Galaxies from FIGGS sample with TRGB distances are shown assolid points whereas the remaining FIGGS galaxies are shown as open cir-cles. Crosses are galaxies from Warren et al.(2007) and solid triangles fromVerheijen (2001). The solid line marks the locus of an upper envelope forthe H I mass-to-light ratio at a given luminosity from Warren et al.(2007).
UA292DDO154E215-G009
Figure 13.
The log of HI mass to light ratio as a function of the tidalindex for the FIGGS sample. Additional galaxies from literature with highM HI / L B are also marked in the plot. itive number indicates that the galaxy is in a dense environment.Figure 13. shows that most of the FIGGS galaxies are in less denseenvironments, and that all the galaxies with high M HI / L B (i.e e > D HI / D Ho > . ) are iso-lated.To summarize, we have presented the first results from the DDO154E215-G009
Figure 14.
The HI extent of the FIGGS sample (normalised to the Holm-berg radius) plotted as a function of the tidal index for the FIGGS sample.Additional galaxies from literature with very extended HI disks viz. DDO154 and ESO215-G?009 and are also marked in the plot.
Faint Irregular Galaxies GMRT Survey (FIGGS). FIGGS is a largeimaging program aimed at providing a comprehensive and statis-tically robust characterisation of the neutral ISM properties of ex-tremely faint, nearby, gas rich, dIrr galaxies using the GMRT. TheGMRT HI data is supplemented with observations at other wave-lengths. The HI images in conjunction with the optical data will beused to investigate a variety of scientific questions including the starformation feedback on the neutral ISM, threshold for star forma-tion, baryonic TF relation and dark matter distribution in low massgalaxies. The optical properties of the FIGGS sample, GMRT ob-servations and the main science drivers for the survey are described.The GMRT integrated HI column density maps and the HI spectrafor the sample galaxies are presented. The global HI properties ofthe FIGGS sample, derived from the GMRT observations, and theircomparison with the optical properties of the sample galaxies arealso presented. A detailed comparison of the gas distribution, kine-matics and star formation in the sample galaxies will be presentedin the companion papers.
ACKNOWLEDGMENTS
The observations presented in this paper were made with the GiantMetrewave Radio Telescope (GMRT). The GMRT is operated bythe National Center for Radio Astrophysics of the Tata Institute ofFundamental Research. Partial support for this work was providedby ILTP grant B-3.13.
REFERENCES
Begum, A., Chengalur, J.N. & Hopp, U., 2003, New Astronomy,8, 267Begum, A & Chengalur, J.N., 2003, A&A, 409, 879Begum, A & Chengalur, J.N., 2004, A&A, 413, 525Begum, A & Chengalur, J.N., 2004, A&A, 424, 509Begum, A & Chengalur, J.N., 2005, MNRAS, 362, 609 c (cid:13) , 000–000 IGGS: Faint Irregular Galaxies GMRT Survey Begum, A, Chengalur, J.N. & Karachentsev, I. D., 2005, A&A,433, 1LBegum, A, Chengalur, J.N., Karachentsev, I. D. & Sharina, M. E.,2005, MNRAS, 359, 53LBegum, A, Chengalur, J.N., Karachentsev, I. D., Kaisin, S. S. &Sharina, M. E., 2006, MNRAS, 365, 1220Begum, A, Chengalur, J.N., Kennicutt, R. C., Karachentsev, I. D.& Janice, L. C., 2008, MNRAS, 383, 809Bell, E. F., McIntosh, D. H., Katz, N. & Weinberg, M. D., 2003,ApJSS, 149, 289Blanton, M. R., Geha, M. & West, A. A. 2007 (astro-ph/0707.3813)Bouchard, A., Jerjen, H., Da Costa, G. S. & Ott, J., 2007, AJ, 133,261Bremnes, T., Binggeli, B. & Prugniel, P., 1999, A&AS, 137, 337Bremnes, T., Binggeli, B. & Prugniel, P., 2000, A&AS, 141, 211Broeils, A. H. & Rhees, M. H., 1997, A&A, 324, 877Carignan, C., & Purton, C. 1998, ApJ, 506, 125Cannon, J. M., McClure-Griffiths, N. M., Skillman, Evan D. &Cˆot´e S., 2004, ApJ, 607, 274Corbin, M. R., Kim, H., Jansen, R. A., Windhorst, R. A. & Fer-nandes, R. C., 2008 ApJ (in press) (astro-ph-0710.2557v1)Cote, S., Freeman, K. C., Carignan, C. & Quinn, P. J., 1997, AJ,114, 1313Cˆot´e S., Carignan, C. & Freeman, K. C., 2000, AJ, 120, 3027de Blok, W. J. G., Bosma, A. & McGaugh, S., 2003, MNRAS,340, 657de Blok, W. J. G., 2005, ApJ, 634, 227Dekel, A. & Woo, J., 2003, MNRAS, 344, 1131Efstathiou, G. 2000, MNRAS, 317, 697Ekta, Chengalur, J. N., Pustilnik, S. A., 2006, MNRAS, 372, 853Ferguson, A. M. N., Gallagher, J. S. & Wyse, R. F. G., 2000, AJ,120, 821Geha, M., Blanton, M. R., Masjedi, M. & West, A. A., 2006, ApJ,653, 240Giovanelli, R. et al. 2005, AJ, 130, 2613Hoffman, G. L., Salpeter, E. E., Farhat, B., Roos, T., Williams, H.& Helou, G., 1996, ApJS, 105, 269Hopp, U. & Schulte-Ladbeck, R. E., 1995, A&AS, 111, 527Hunter, D., 1997, PASP, 109, 937Hunter, D. A. & Elmegreen, B. G., 2006, ApJS, 162, 49Huchtmeier, W. K. & Richter, O. G., 1986, A&AS, 63, 323Huchtmeier, W. K. & Skillman, E., 1998, A&AS, 127, 269Huchtmeier, W. K., Karachentsev, I. D. & Karachentseva, V. E.,2003 A&A, 401, 483Kanekar, N. & Chengalur, J. N., 2005, A&A, 429, L51Karachentseva I. D., Prugniel, P., Vennik, J., Richter, G. M.,Thuan, T. X., & Martin, J. M., 1996, A&AS, 117, 343Karachentsev I. D., Makarova, L. N. & Andersen, M. I., 1999,MNRAS, 307, 37LKarachentsev I. D., Karachentseva V. E., Huchtmeier W. K. &Makarov D. I., 2004, AJ, 127, 2031Karachentsev I. D. et al., 2006, AJ, 131, 1361Kennicutt, R. C., 1989, ApJ, 344, 685Lo, K. Y., Sargest, W. L. W. & Young, K., 1993, AJ, 106, 507Makarova, L., 1999, A&AS, 139, 491Makarova, L., Karachentsev, I. D., Grebel, E. K. & Barsunova, O.Y., 2002, A&A, 384, 71Makarova, L., Karachentsev, I. D., Grebel, E. K., Harbeck, D.,Korotkova, G. G. & Geisler, D., 2005, A&A, 433, 751McGaugh, S. S., Schombert, J. M., Bothun, G. D., de Blok, W. J.G., 2000, ApJL, 533, 99 McGaugh, S. S., 2005, ApJ, 632, 859Matthews, L. D., Gallagher, J. S., Littleton, J. E., 1995, AJ, 110,581Miralda-Escude, J. & Rees, M. J., 1997, ApJ, 478,57Navarro et al., 2004, MNRAS, 349, 1039Parodi, B. R., Barazza, F. D. & Binggeli, B., 2002, A&A, 388, 29Pustilnik, S. A. & Martin, J. M., 2007, A&A, 464, 859Schaye, J., 2004, ApJ, 609, 667simon, J. D. & Geha, M., 2007, ApJ, 670, 313Simpson C. E., Hunter D. A., Nordgren T. E., 2005, AJ, 130, 1049Skillman, E. D., 1987, in Star Formation in Galaxies, edited by C.J. Lonsdale Persson, NASA, 263Springob, C. M., Haynes, M. P., Giovanelli, R. & Kent, B. R.,2005, ApJS, 160, 149Stil, J. M. & Israel. F. P., 2002, A&A, 389, 29Swarup, G., Ananthakrishnan, S., Kapahi, V.K., Rao, A.P., Sub-rahmanya, C.R. & Kulkarni, V.K. 1991, Current Science, 60, 95Swaters, R., 1999, PhD thesis, Rijksuniversiteit GroningenTaylor, C. L., Brinks, E., Pogge, R. W. & Skillman, E. D., 1994AJ, 107, 971Taylor, V. A., Jansen, R. A., Windhorst, R. A., Odewahn, S. C. &Hibbard, J. E., 2005, ApJ, 630, 784Tully, B. et al. 2006, AJ, 132, 729van Zee, L., Haynes, M. P. & Giovanelli, R., 1995, AJ, 109, 990van Zee, L., 2000, ApJ, 119, 2757van den Bosch, F. C., Swaters, R. A., 2001, MNRAS, 325, 1017Verheijen, M.A.W & Sancisi, R., 2001, A&A, 370, 765Verheijen, M.A.W, 2001, ApJ, 563, 694Warren, B. E., Jerjen, H. & Koribalski, B. S., 2004, AJ, 128, 1152Warren, B. E., Jerjen, H. & Koribalski, B. S., 2007, AJ, 134, 1849Weldrake, D. T. F., de Blok, W. J. G. & Walter, F., 2003, MNRAS,340, 12Young, L. M., van Zee, L., Lo, K. Y., Dohm-Palmer, R. C., &Beierle, M. E. 2003, ApJ, 592, 111Zwaan, M. A., Meyer, M. J., Steveley-Smith, L. & Webster, R. L.,2005, MNRAS, 359, L30 c (cid:13)000