A&A 449, 929-935 (2006)
G. Gavazzi1 - K. O'Neil2 - A. Boselli3 - W. van Driel4
1 - Universitá degli Studi di Milano-Bicocca, Piazza delle scienze 3, 20126 Milano, Italy
2 - NRAO, PO Box 2, Green Bank, WV 24944, USA
3 - Laboratoire d'Astrophysique de Marseille, BP 8, Traverse du Siphon, 13376 Marseille, France
4 - Observatoire de Paris, Section de Meudon, GEPI, CNRS UMR 8111 and Université Paris 7, 5 place Jules Janssen, 92195 Meudon Cedex, France
Received 18 July 2005 / Accepted 20 September 2005
High sensitivity 21-cm H I line observations with an rms noise level of 0.5 mJy were made of 35 spiral galaxies in the Coma Supercluster, using the refurbished Arecibo telescope, leading to detection of 25 objects. These data, combined with the measurements available in the literature, provide the set of H I data for 94% of all late-type galaxies in the Coma Supercluster with an apparent photographic magnitude mag. We confirm that the typical scale of H I deficiency around the Coma cluster is 2 Mpc, i.e. one virial radius. Comparing the H I mass function (HIMF) of cluster with non-cluster members of the Coma Supercluster we detected a shortage of high H I mass galaxies among cluster members that can be attributed to the pattern of H I deficiency found in rich clusters.
Key words: galaxies: distances and redshifts - galaxies: general - galaxies: ISM - galaxies: clusters: individual: Virgo - radio lines: galaxies
|Figure 1: The wedge diagram of the Coma Supercluster galaxies in the declination interval . Members are divided between late-type (large circles, filled if observed in HI) and early-type (E-S0a: empty squares). Empty triangles mark foreground and background galaxies.|
|Open with DEXTER|
H I line observations of galaxies have provided us with some of the most powerful diagnostics on the role of the environment in regulating the evolution of late-type (spiral) galaxies in the local Universe. Spiral galaxies in rich X-ray luminous clusters display a significant lack of H I gas with respect to their "undisturbed'' counterparts in the field (Haynes & Giovanelli 1984; Giovanelli & Haynes 1985). This pattern of H I deficiency can be attributed to various interaction mechanisms: ram-pressure (Gunn & Gott 1972), viscous stripping (Nulsen 1982), thermal evaporation (Cowie & Songaila 1977), or tidal interaction with the cluster potential well (Byrd & Valtonen 1990; Moore et al. 1996). Since these mechanisms have a higher efficiency in or near rich cluster cores, the H I deficiency parameter (Haynes & Giovanelli 1984 - see Sect. 5.1) is an environmental indicator that provides a clear signature of a galaxy's membership in a rich cluster.
The Coma Supercluster has received considerable attention in H I studies due to its proximity to us (100 Mpc). Since the pioneering study by Sullivan et al. (1981), various works (e.g. Chincarini et al. 1983a; Gavazzi 1987, 1989; Scodeggio & Gavazzi 1993; Haynes et al. 1997) have provided measurements of the H I content for most late-type galaxies in the Coma Supercluster. In addition to these single-dish studies of their global H I properties, the detailed mapping of galaxies in the Coma and A1367 clusters with radio synthesis telescopes was obtained by Bravo-Alfaro et al. (2000, 2001) and Dickey & Gavazzi (1991).
A high sensitivity, blind H I survey of 7000 square deg. of sky is planned for 2005-2006 with the ALFA multibeam system at Arecibo (Giovanelli et al. 2005), and even more sensitive (1 mJy rms) surveys of parts of these clusters will be obtained with the ALFA system. They will include the Coma Supercluster and the Virgo cluster. In preparation for these surveys, before the installation of the ALFA system, we used the single-beam Arecibo system to continue the pointed observation survey of late-type galaxies in the Virgo cluster (see Gavazzi et al. 2005a, Paper I) and in the Coma Supercluster area. Here we report the results of the Coma Supercluster observations that, in conjunction with the previously available H I data-set, enable us to review the properties of galaxies in this Supercluster, as obtained from optically selected H I observations.
The selection of the cluster targets for H I observations is described in Sect. 2, the observations and the data reduction are presented in Sect. 3 and the results are given in Sect. 4 and discussed in Sect. 5. A Hubble constant of is assumed throughout this paper.
Galaxies in the present study were selected from the CGCG Catalogue (Zwicky et al. 1961-68) in the region . There are 1127 CGCG galaxies listed in this region with an apparent photographic magnitude . Their wedge-diagram is given in Fig. 1, where the structure of the Coma-A1367 Supercluster stands out clearly as the pronounced density enhancement near 7000 km s-1 and as part of one of the largest known coherent structures in the local universe, named the "Great Wall'' (Ramella et al. 1992). Other conspicuous features are: (i) the "Fingers of God'' of Coma and A1367, that span the interval 4000<V<10 000 km s-1, mostly traced by early-type objects, and (ii) the large ( ) "void'' in front of the Supercluster, with a density that is 150 times lower than the mean galaxy density in the universe. Other remarkable features, also known as the "legs of the homunculus'' (de Lapparent et al. 1986) are two filaments pointing toward the Coma cluster in the interval 4500<V<7000 km s-1. A third "homunculus leg'' surrounds the void on the western side, projected near A1367. Objects in the interval 6000<V<8000 km s-1 form a bridge between the two clusters with a narrow velocity distribution. Above V>8000 km s-1, the Coma Supercluster fades into the background, so setting a boundary between its members and the projected background objects is rather arbitrary. In assigning the membership of individual galaxies to the various sub-structures we follow the criteria of Gavazzi et al. (1999).
|Figure 2: The sky projection of Coma Supercluster members (including the "homunculus legs'' HL). Spirals are represented by circles (filled if observed in H I) and E-S0a objects by empty squares.|
|Open with DEXTER|
After the present high sensitivity (rms mJy) observations of 33 additional CGCG galaxies (and of two fainter objects: FOCA 610, 636), 295/315 (94%) late-type members (including HL) were observed (see Table 1), of which 259 were detected and 36 have upper limits. Of the remaining 20 unobserved late-type targets, 13 are members of double or multiple systems that could not be resolved by the Arecibo beam (97-111S, 97-129E, 127-051N, 128-029E, 127-025N, 159-049S, 128-031N, 128-002, 97-036, 98-072, 98-073, 98-081, 98-087). The Coma cluster (160-243) and 6 galaxies (127-121, 129-016, 157-077, 158-046, 160-180, 161-029) belonging to the HL were not observed due to scheduling constraints. To fill in a scheduling hole, 13 galaxies in the Virgo cluster were also observed, and all but one were detected. The results of these observations are given in Table 3, and the HI profiles of the detected galaxies are shown in Fig. A.2. These objects will not be considered further in this paper.
All data on the Coma Supercluster and Virgo galaxies are collected and made available worldwide via the "Goldmine'' website (http://Goldmine.mib.infn.it; see Gavazzi et al. 2003).
Table 1: Sample completeness.
Using the refurbished 305-m Arecibo Gregorian radio telescope, we observed 35 galaxies in the Coma Supercluster (plus 13 in the Virgo cluster) (see Sect. 2) in February 2004 and January-March 2005. Data were taken with the L-band Wide receiver, using a nine-level sampling with two of the 2048 lag subcorrelators set to each polarization channel. All observations were taken using the position-switching technique, with each blank sky (or OFF) position observed for the same duration, and over the same portion of the telescope dish as the on-source (ON) observation. Each 5 min + 5 min ON+OFF pair was followed by a 10 s ON+OFF observation of a well-calibrated noise diode. The overlaps between both sub-correlators with the same polarization allowed a contiguous radial velocity search range of sufficient width from -1000 to 8500 km s-1. The velocity resolution was 2.6 km s-1, the instrument's HPBW at 21 cm is and the pointing accuracy was about 15''. The pointing positions used were the optical center positions of the target galaxies listed in Table 1. Flux density calibration corrections were good to within 10% (and often much better); see the discussion of the errors involved in O'Neil (2004).
Using standard IDL data reduction software available at Arecibo, corrections were applied for the variations in the gain and system temperature with zenith angle and azimuth. A baseline of order one to three was fitted to the data, excluding those velocity ranges with H I line emission or radio frequency interference (RFI). The velocities were corrected to the heliocentric system, using the optical convention, and the polarizations were averaged. All data were boxcar-smoothed to a velocity resolution of 12.9 km s-1 for further analysis. For all spectra, the rms noise level was determined and for the detected objects the central line velocity, the line widths at, respectively, the 50% and 20% level of the peak, and the integrated line flux were determined. No flux correction depending on the source size was applied because the optical extent of all detected targets does not significantly exceed the Arecibo HPBW.
In order to identify those sources whose H I detections could have been confused by nearby galaxies, we queried the NED, HyperLeda, and Goldmine databases and inspected DSS images over a region of 10' radius surrounding the central position of each source, given the telescope's sidelobe pattern. Quoted values are weighted averages from the HyperLeda database, unless otherwise indicated.
The H I spectra of both the clearly and the marginally detected galaxies are shown in Figs. A.1 and A.2, and the global H I line parameters are listed in Table 3. These are directly measured values; no corrections have been applied to them for, e.g., instrumental resolution. Table 3 is organized as follows:
The newly obtained H I data were combined with those available from the literature for the late-type galaxies in the Coma Supercluster, as listed in Table 4. The sample comprises 315 galaxies, of which 295 were observed, 259 were detected, and 36 remain undetected. Table 4 is organized as follows:
For the late-type galaxies in the present study, we estimated the H I deficiency parameter following Haynes & Giovanelli (1984) as the logarithmic difference between of a reference sample of isolated galaxies and as actually observed in individual objects: Def . Then, is computed from the galaxies optical linear diameter d as: , where d is estimated consistently with Haynes & Giovanelli (1984); and a and b are weak functions of the Hubble type, as listed in Table 3 of Paper I (notice that across the Hubble sequence, i.e. increases approximately as the galaxy linear diameter squared).
|Figure 3: The late-type Coma Supercluster members (including HL) coded according to their H I deficiency parameter: Def (empty circles); Def (filled circles).|
|Open with DEXTER|
Figure 3 shows that H I deficient galaxies segregate around the center of the Coma cluster and, to a much lesser degree, around A1367. One important issue that can be addressed with the present data-set, given its completeness, is on what scale the phenomenon of H I ablation holds. The H I deficiency parameter of individual galaxies is given in Fig. 4 as a function of the projected linear separation from the X-ray center of the Coma cluster, out to 15 Mpc, along with average values taken in bins of 0.5 (from 0 to 2) and in bins of 1 further out (see also Table 2). It is apparent that significant H I deficiency occurs out to approximately 3 Mpc radius. At one virial radius (i.e. at 2.2 Mpc, Girardi et al. 1998; or 2.9 Mpc Lokas & Mamon 2003; Neumann et al. 2001), the average H I content of the supercluster galaxies becomes indistinguishable from that of the field, in agreement with Solanes et al. (2001).
The star-formation rate, as derived from H observations, shows a significant cut-off near 1 to 2 projected virial radii from the Coma cluster (Gavazzi et al. 2005b) and from over-density peaks in the SDSS (Nichol 2004; Miller 2004). What the H I observations show is the driver of the SFR decline, i.e. the lack of gas to sustain it.
|Figure 4: The projected distribution about the X-ray center of the Coma cluster of the H I deficiency of Late-type Supercluster detected (dots) and undetected (triangles) members (excluding HL). Large circles represent averages in bins of 0.5 (1). One errorbars are given.|
|Open with DEXTER|
Table 2: The radial distribution of the mean binned H I deficiency of Late-type Supercluster members (excluding HL) about the Coma cluster.
Table 3: Parameters of the newly observed galaxies.
Taking advantage of the large sample of optically selected galaxies with H I measurements and optical (B-band) photometry (accurate to within 0.1 mag), we study the H I mass vs. relation for unperturbed galaxies. However, since the Coma Supercluster sample ( ) contains only giant galaxies brighter than , we have included the Virgo galaxies (Paper I) for the purpose of extending this correlation over a broader magnitude range. In doing so we must obviously exclude galaxies with perturbed H I contents. Conservatively, we excluded galaxies with Def (the Def parameter is determined from diameters (Sect. 5.1), independent of the B luminosity, as in Haynes & Giovanelli 1984). Figure 5 shows the relation for the 465 late-type galaxies that were selected accordingly, and plotted in three bins of Hubble type to stress the consistency among them. The three given linear regressions, obtained combining all Hubble types, are the direct one ( ), the inverse one ( ), and the one adopted in Paper I ( ). The residual of the (direct) correlation is ; i.e., the H I mass of disk galaxies can be predicted within a factor of 1.8 uncertainty from their B luminosity.
|Figure 5: The correlation between log and for Sa-Sb (filled circles), Sbc-Sc (empty circles), Scd-Im-BCD (filled squares). The direct fit (solid line), the inverse fit (dashed) (obtained considering all Hubble types), and the relation used in Paper I (dotted) are given.|
|Open with DEXTER|
Figure 6 shows the frequency distribution of for the cluster (Coma + A1367) late-type galaxies (solid histogram) and for the isolated galaxies (including HL), the latter normalized to the former by the ratio of the number of galaxies. The H I Mass Function (HIMF) so obtained cannot be meaningfully compared with the one of Virgo, obtained in Paper I, or even with the one of isolated galaxies by Zwaan et al. (2003), because of the shallowness of the optical and 21 cm observations available for the Coma Supercluster. Given the relation , discussed in the previous section, the limiting magnitude ( ) of the CGCG implies that, on average, only galaxies with log are targeted. This imposes a completeness cut-off that is even shallower than the limiting detectable H I mass of log that derives from the typical noise figure of the 21 cm observations ( mJy). In Paper I we showed that significant differences between the HIMF of the Virgo cluster and of the field occur for log , i.e. below both the present cut-off lines. For , however, there is in Fig. 6 a barely significant excess in the frequency of non-cluster members of the Coma Supercluster with respect to cluster members. This can be understood as a signature of the deficiency pattern of cluster galaxies, as found in Paper I for Virgo.
|Figure 6: The MHI function for the members of the Coma+A1367 clusters (solid) and for the non-cluster Supercluster members (dotted). The optical (log ) and radio (log ) completeness limits are drawn.|
|Open with DEXTER|
We observed 35 galaxies in the Coma Supercluster and 13 in the Virgo cluster in the 21-cm H I line with the refurbished Arecibo telescope. The high sensitivity of our observations (rms noise 0.5 mJy) resulted in the detection of 37 objects, and significant upper limits were obtained for the remaining ones. Including the present observations, the H I survey of the Coma Supercluster has reached virtual completion (94% among the late-type members). And, by combining all data, we have determined with high significance that the typical scale of H I deficiency around the Coma cluster is 2-3 Mpc, i.e. one virial radius.
With the present data, a meaningful determination of the HIMF can be only obtained for log , insufficient for comparing it with the deeper HIMF of Virgo (Paper I) and of isolated galaxies (Zwaan et al. 2003). Comparing cluster with non-cluster Supercluster members, we do, however, detect a shortage of high H I mass galaxies among cluster members which can be attributed to the pattern of H I deficiency found in rich clusters.
The Arecibo Observatory is part of the National Astronomy and Ionosphere Center, which is operated by Cornell University under a cooperative agreement with the National Science Foundation. This research also made use of the Lyon-Meudon Extragalactic Database (LEDA), recently incorporated in HyperLeda and the NASA/IPAC Extragalactic Database (NED), operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration and the Goldmine database.
Table 4: Basic H I properties of late-type members to the Coma Supercluster (including HL).
CGCG 127-018: Our H I detection ( = 6922 2 km s-1, W50 =147 km s-1 and = 2.1 0.06 Jy km s-1) is consistent with our previous (van Driel et al. 2000) Nançay detection ( = 6936 16 km s-1, W50 =131 km s-1 and = 2.7 0.3 Jy km s-1) and with the Arecibo detection of Gavazzi (1987) at = 6935 km s-1.
CGCG 127-039: The integrated line intensity of our H I detection ( = 6911 1 km s-1, W50 = 37 km s-1 and = 0.82 0.04 Jy km s-1) is 3.4 times lower than that of our previous Nançay detection (van Driel et al. 2000), with = 6919 9 km s-1, W50 = 40 km s-1 and = 2.8 0.2 Jy km s-1. The latter was made with an elongated 21' ( )HPBW, which is considerably larger than the round Arecibo beam. This difference can be due to the 13.6 B mag SBbc spiral NGC 3832, due south of the target galaxy, whose mean H I line parameters ( = 6909 6 km s-1, W50 = 171 km s-1 and = 10.4 Jy km s-1) are based on 5 spectra, all obtained at Arecibo (Chincarini et al. 1983a; Giovanardi & Salpeter 1985; Lewis 1985; Lewis et al. 1985; Sullivan et al. 1981).
CGCG 127-055: Our H I detection ( = 6626 3 km s-1, W50 = 215 km s-1 and = 1.9 0.08 Jy km s-1) is consistent with our previous measurement (van Driel et al. 2000) made at Nançay, with = 6656 21 km s-1, W50 = 183 km s-1 and = 2.1 0.3 Jy km s-1. Our new data have a 3.6 times better rms noise level.
CGCG 128-072: Given the uncertainties, the global parameters of our H I detection ( = 6795 6 km s-1, W50 = 119 km s-1 and = 0.82 0.07 Jy km s-1) are consistent with those of our previous Nançay detection (van Driel et al. 2000), with = 6848 81 km s-1, W50 = 213 km s-1 and = 1.4 0.6 Jy km s-1. Our new data have a 5.4 times better rms noise level.
CGCG 129-004: discrepant values for the optical redshift have been published, 4847 and 6729 km s-1 (Gavazzi et al. 1999a, and the compilation by Falco et al. 1999); we find an H I value of 6736 7 km s-1, consistent with the latter value.
CGCG 130-003: there are two discrepant literature values for its optical redshift, 7094 and 22,425 km s-1 (Straus et al. 1992; Gregory et al. 1988); we find an H I value of 7140 4 km s-1, consistent with the former value.
CGCG 157-044: our H I detection ( = 6607 4 km s-1, W50 = 240 km s-1 and = 0.84 0.08 Jy km s-1) has a 1.7 times lower line intensity than our previous Nançay detection (van Driel et al. 2000), with = 6628 36 km s-1, W50 = 309 km s-1 and = 1.5 0.4 Jy km s-1. The difference is less than two times the uncertainty in the latter value, however, and therefore not significant.
CGCG 159-071: the global parameters of our H I detection ( = 6971 1 km s-1 and W50 = 189 km s-1 and = 2.6 0.1 Jy km s-1) are consistent with those of our previous Nançay detection (van Driel et al. 2000), with = 6985 6 km s-1, W50 = 164 km s-1 and = 2.4 0.4 Jy km s-1.
CGCG 159-097: Its optical redshift, 6573 190 km s-1, is not well determined. We measured an H I value of 6424 30 km s-1, consistent with 3 of the published optical values - only the optical velocity of 6883 75 km s-1 measured by van Haarlem et al. (1993) is in disagreement with all other values.
CGCG 160-128: Our detection ( = 7920 1 km s-1, W50 = 115 km s-1 and = 2.5 0.07 Jy km s-1) is consistent with our previous Nançay detection ( = 7940 5 km s-1, W50 = 100 km s-1 and = 2.2 0.3 Jy km s-1), which was based on data with a 4 times higher rms of 2.8 mJy (van Driel et al. 2000).
CGCG 161-051: We did not confirm our previous, quite tentative Nançay detection (van Driel et al. 2000), with = 6993: km s-1, W50 = 235: km s-1 and = 1.5: Jy km s-1. Our new spectrum has a 2.4 times better rms, of 1.5 mJy.
CGCG 161-054: Our detection ( = 6756 3 km s-1, W50 = 284 km s-1 and = 1.3 0.1 Jy km s-1) is consistent with our previous, tentative Nançay detection, with = 6760: km s-1, W50 = 335: km s-1 and = 1.8: Jy km s-1, which was based on data with a 4 times higher rms of 3.2 mJy (van Driel et al. 2000).