Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.atnf.csiro.au/pasa/16_1/kilborn/paper.pdf
Дата изменения: Tue May 25 04:53:25 1999
Дата индексирования: Sun Apr 10 04:25:09 2016
Кодировка:
Поисковые слова: m 106

Publ. Astron. Soc. Aust., 1999, 16, 811

.

HI Mass Function from HIPASS
V. Kilborn1 , R. L. Webster1 and L. Staveley-Smith2
School of Physics, University of Melbourne, Parkville, Vic. 3052, Australia vkilborn@physics.unimelb.edu.au, rwebster@physics.unimelb.edu.au Australia Telescope National Facility, CSIRO, PO Box 76, Epping, NSW 2121, Australia Lister.Staveley-Smith@atnf.csiro.au Received 1998 November 17, accepted 1999 January 28
1

2

Abstract: The HI Parkes All Sky Survey (HIPASS) is a blind search for extragalactic neutral hydrogen, covering the whole of the southern sky. We present the latest HI mass function (HIMF) constructed from a sample of 263 galaxies with declinations 1/Vmax and maximum likelihood techniques are used in the <-62 . Standard analysis. No upturn in the low-mass end of the HIMF is yet seen, though our selection procedure presently conspires against the lowest-mass galaxies. Keywords: galaxies: mass function -- radio lines: galaxies -- surveys

1 Introduction The neutral hydrogen (HI) mass function (HIMF) describes the number density of neutral hydrogen as a function of HI mass at the present epoch. Measuring the HIMF is important for understanding the evolution of galaxies and for understanding the incidence of low-redshift Ly- damped and `forest' lines against background ultraviolet sources. Until recently, the data used to determine the HIMF have come from very small samples of HI-selected ob jects (e.g. Zwaan et al. 1997; Schneider, Spitzak & Rosenberg 1998), or from HI observations of opticallyselected galaxies (e.g. Solanes, Giovanelli & Haynes 1996). This means that HIMFs typically have poor statistics in the extreme mass bins, or are inherently biased by the optical nature of the sample selection. Optical samples tend to miss low surface-brightness systems (e.g. Chengalur, Giovanelli & Haynes 1995) and tidal gas (e.g. Schneider et al. 1983). However, an extensive blind HI survey such as HIPASS gives both a large and an unbiased sample and will increase the number counts for the low and high mass ends of the HIMF. Presently, the low-mass end of the HIMF is somewhat controversial with contrary views being expressed as to the relative importance of low-mass and high-mass galaxies to the overall HI density of the Universe. There is some evidence for an upturn at the low-luminosity end of the optical luminosity function (e.g. Loveday 1997). It is important to establish whether a similar upturn exists at the low-mass end of the HIMF. Most derivations of the HIMF have found a shallow faint-end slope (Schechter parameter, 1 · 2--e.g. Zwaan et al. 1997), implying that low-mass galaxies contribute relatively little to the total HI density. Recently Schneider, Spitzak & Rosenberg (1998) claimed to see an upturn below
Astronomical Society of Australia 1999

masses of 108 M . However, their result is from a sample with only two galaxies in the lowest mass bin, so it remains to be seen whether this will persist in statistically larger samples. In this paper, we derive the HIMF for a sample of 263 HI-selected galaxies. In Section 2, we give the particulars of the HIPASS survey and the data selection process; in Section 3 we derive the HIMF. In Section 4, we briefly discuss the results. A value of H0 = 100h km s-1 Mpc-1 is used throughout. 2 Survey Parameters and Data Selection HIPASS is being conducted at the 64 m Parkes Radio Telescope in NSW, Australia. The survey uses the multibeam receiver, which has individual beam sizes on the sky of 14 . The sky is actively scanned in 3 в 8 strips, and it will be scanned five times to reach the final survey sensitivity. HIPASS will survey the southern sky between -1200 km s-1 and 12700 km s-1 , with a velocity resolution of 18 km s-1 . The RMS noise in the final data cubes is 13 mJy per beam. Bandpass subtraction is undertaken in real-time at the telescope (Barnes et al. 1998), and the data are gridded at a later stage into 8 в 8 cubes, with a pixel size of 4 в 4 and velocity spacing of 13 · 2 km s-1 . There will be 388 data cubes in the finished southern sky survey. The 3 HI mass limit for the survey is 2 1 · 4 в 106 DMpc M , and the 3 HI column density limit is 7 в 1017 cm-2 per velocity channel. A follow-up observation program is taking place at the Australia Telescope Compact Array (ATCA) to obtain better positional accuracy of interesting sources, and optical follow-up is undertaken at the Siding Spring and CTIO 40 inch telescopes. These observations are vital to the survey, as they will confirm marginal detections and provide greater
1323-3580/99/010008$05.00

HI Mass Function from HIPASS

9

Figure 1--Mass of the galaxies in the sample versus velocity. The curve represents the theoretical mass cut-off for the sample.

Figure 2--Velocity distribution for the full sample.

resolution for interesting ob jects such as intergalactic HI clouds. We have selected our sample of galaxies from the declination region -62 > > -70 and < -78 , through the full range in Right Ascension. This region has been scanned to the full multibeam sensitivity of five scans, and represents 0 · 572 sr in area (7% of the total HIPASS survey area). The galaxies were detected by eye in the data cubes; in total, 304 galaxies were found in this region. Of the sample, 15% are previously uncatalogued and 30% are new detections in HI. There are no confirmed detections of HI clouds without optical counterparts in this sample. To construct a complete sample, a cut-off was made at the integrated flux value of 4 Jy km s-1 .

This was chosen as a conservative limit, because the detection limits for the HIPASS data have not been explored fully at this stage. By choosing such a conservative cut-off, the detections with the lowest peak flux and velocity width are omitted from the sample, and we are limited to characterising the HIMF above a mass of MHI 107 M . This means that once again the low mass end of the HIMF will be determined on the data from just a couple of galaxies. A plot showing the theoretical mass cut-off versus the masses in the sample is shown in Figure 1, which shows that the sample seems to be complete to the chosen cut-off. With this integrated flux limit imposed, there were 263 galaxies left in the sample. A histogram showing the velocity distribution of the sample can be seen

10

V. Kilborn et al.

Figure 3--Mass distribution for the flux limited sample.

Figure 4--HI mass function (HIMF): the solid points represent the measured HI mass function per decade, and the solid line represents the best fit Schechter function for the 1/Vmax method from this sample: = 1 · 32, log M = 9 · 5 and = 0 · 01. The dotted line is the HIMF from Zwaan et al. (1997), with = 1 · 2 and log M = 9 · 55. The error bars show 1 errors (Marshall 1985).

in Figure 2--we have detected galaxies between 400 km s-1 and 10,000 km s-1 in this sample. The mass distribution for the sample can be seen in Figure 3. While the number statistics have been increased in the higher mass bins, the two lowest bins combined have just three galaxies. 3 Results The standard 1/Vmax method (Schmidt 1968) was used to determine the HIMF from the data. A Schechter (1976) function was then fitted to the data:

MHI MHI d M M

=

MHI M -

-

в exp

MHI MHI d M M

,

(1)

where M is the mass that defines the characteristic knee in the function, and is the normalisation. Figure 4 shows the HIMF derived using this method, and the best fit Schechter function is shown by the solid line. The slope of the low-mass end is = 1 · 32, the characteristic mass, log M = 9 · 5, and the normalisation, = 0 · 01. These are similar

HI Mass Function from HIPASS

11

to values found by Zwaan et al. (1997), who find = 1 · 2, and logM = 9 · 55 (see the dotted line in Figure 4). The normalisation found for the HIPASS sample is slightly lower than the Zwaan et al. value of 0 · 014. There is no indication from our data that there is a turn-up in the slope of the HIMF to the mass limit we have imposed. However, as we still have poor statistics in the low mass bins (a single galaxy in the lowest mass bin), the faint end slope is still somewhat uncertain.

g cm-3 . The cosmological mass density of HI at z = 0, HI , is the ratio of the HI density over the critical density at the present epoch (c = 1 · 88 в 10-29 h2 g cm-3 , Padmanabhan 1993). This gives a value of HI = 1 · 5 в 10-4 h-1 . Assuming that the mass percentage of HeI is 25%, the total cosmological mass density of neutral gas at the present epoch is g = 1 · 88 в 10-4 h-1 : this is slighty lower than the Zwaan et al. (1997) value of g (z = 0) = (2 · 7 ± 0 · 5) в 10-4 h-1 . 4 Discussion and Conclusions So far, we have taken a conservative cut-off to derive the HIMF. This has excluded the low mass galaxies which we are particularly interested in, and in the future we will work on determining the sensitivity limits of the survey more fully so that all the available data can be included in the analysis. The preliminary results from HIPASS show an HIMF that does not indicate a steep rise at the faint end, which is a similar result to previous surveys. We derive a cosmological mass density of g = 1 · 88 в 10-4 h-1 , which is also similar to past results. With better understanding of the survey sensitivity we will be able to push the limits of the low mass end of the HIMF.

Figure 5--Contours from the maximum likelihood method. The maximum value is seen at = 1 · 15, log M = 9 · 74. The contours are 68 · 3% and 99 · 9% confidence limits.

Acknowledgments We acknowledge the multibeam working group for their help with observations, and thank the Parkes Observatory staff for their constant support. References
Barnes, D. G., et al. 1998, in ADASS VII (San Fransisco: ASP), p. 89 Binggeli, B., Sandage, A., & Tammann, G. A. 1988, A&A, 26, 509 Chengalur, J. N., Giovanelli, R., & Haynes, M. P. 1995, AJ, 109, 2415 Loveday, J. 1997, ApJ, 489, 29 Marshall, H. 1985, ApJ, 299, 109 Padmanabhan, T. 1993, in Structure Formation in the Universe (Cambridge University Press), p. 29 Sandage, A., Tamman, G. A., & Yahil, A. 1979, ApJ, 232, 352 Schechter, P. 1976, ApJ, 203, 297 Schmidt, M. 1968, ApJ, 151, 393 Schneider, S. E., Helou, G., Salpeter, E. E., & Terzianm, Y. 1983, ApJ, 273L, 1 Schneider, S. E., Spitzak, J. G., & Rosenberg, J. L. 1998, ApJ, 507, L9 Solanes, J., Giovanelli, R., & Haynes, M. 1996, ApJ, 461, 609 Zwaan, M. A., Briggs, F. H., Sprayberry, D., & Sorar, E. 1997, ApJ, 490, 173

A maximum likelihood method (Sandage, Tammann & Yahil 1979) was also used to determine Schechter parameters for the data. The Schechter parameters determined using this method are slightly different from the best fit parameters found using the 1/Vmax method, with = 1 · 15 and log M = 9 · 74. Figure 5 shows a plot of contours of likelihood for the sample. It is possible that the difference in value of the 1/Vmax method and the maximum likelihood method is due to clustering of the galaxies in this region. The maximum likelihood method is not dependent on the distribution of galaxies in the sample, whereas the 1/Vmax varies according to the clustering of the galaxies in the sample (Binggeli, Sandage & Tammann 1988). A full investigation of selection effects and different computation methods will be presented in a later paper. Using the 1/Vmax Schechter function parameters derived for this sample we can determine the HI density at the present epoch, which is HI = (2 - )M , where is the Euler gamma function. We obtain HI = 4 · 2 в 107 h M Mpc-3 , or 2 · 8 в 10-33 h