Studies of IRAS sources at high galactic latitudes - II. Results from a redshift survey at b > 60 : distribution in depth, luminosity function, and physical nature of IRAS galaxies

Abstract

We have identified 530 IRAS sources from a complete sample, flux limited at |$S(60 \mu\text m)=0.5$| Jy, in an area of sky covering 844 square degrees at the North Galactic Pole. After exclusion of 34 bright stars, 99 per cent of the remainder coincide with galaxies, mostly in the magnitude range 14–17, but with a tail of faint objects. Spectroscopic observations have been made of candidates for 183 IRAS sources. Together with published data, we have redshifts for two overlapping complete sub-samples, totalling 303 galaxies. We derive several important results: (i) A large fraction of the IRAS catalogue consists of unremarkable, bright, spiral galaxies. Fainter galaxies however mostly show strong low-excitation lines (confirming the preliminary work of Allen, Roche & Norris), probably indicating violent star formation. From the size of the images, this is typically occurring in relatively normal spiral galaxies, rather than, for example, isolated extragalactic H II regions. One IRAS galaxy in ∼ 30 has a 60μm luminosity in excess of |$10^{12}L_\odot$|⁠, i.e. is as luminous as a quasar. About 7 per cent show high excitation lines. (ii) The median redshift is 0.030 (180 Mpc, |$H_0=50$|⁠). Assuming that the density of IRAS sources is proportional to the overall density of galaxies, the IRAS catalogue is then an unbiased survey of cosmological structure on a scale 2–300 Mpc, more than 10 times as large as the Local Supercluster. (iii) The 60μm luminosity function shows a change of slope at about |$10^{11}L_\odot$|⁠. We fit it with a simple two power-law function. Below and above the knee the best fit power-law slopes are −0.8 and −2.5 respectively. (iv) Examination of the nearest neighbour distribution for luminous IRAS galaxies shows them to exist in a fairly normal range of environments. Although there certainly exist disturbed systems whose interaction may be the cause of a violent burst of star formation, galaxy interaction is unlikely to be the major cause of the IRAS galaxy phenomenon. (v) Comparison of our IRAS sample with non-Seyfert Markarian galaxies, shows the two populations to be overlapping, but not identical. If the dominant underlying phenomenon in both populations is violent star formation, this implies that the strength of the burst is not the only physical parameter involved. (vi) From a complete hard X-ray flux-limited sample of Seyfert galaxies, we calculate that IRAS galaxies are ∼50 times as common as Type 1 Seyferts, at the same IR output. The space density of Type 2 Seyferts is not well known, but IRAS galaxies are probably 10−25 times as common as all types of Seyfert galaxy.