|Annu. Rev. Astron. Astrophys. 1997. 35:
Copyright © 1997 by . All rights reserved
Over the past 15 years, a number of surveys have succeeded in locating low surface brightness galaxies. The distribution in surface brightness is continuous, but operationally we choose to define galaxies with µ0 23 mag arcsec-2 as low surface brightness. In terms of the narrow surface brightness distribution of Freeman (1970), a disk galaxy this diffuse should be extremely rare. In practice, low surface brightness galaxies are a mixed bag, including objects as diverse as giant gas-rich disks and dwarf spheroidals. This of course is the reason for their importance: such galaxies offer a new window onto the diversity of galaxy morphology and evolution.
3.1. Photographic Surveys
Unsurprisingly for someone who left his idiosyncratic mark on much of extragalactic astronomy, Zwicky (1957) was one of the first to speculate on the existence of low surface brightness galaxies. His claim of a steeply inceasing tail of faint galaxies was at odds with Hubble's (1936) earlier Gaussian form for the luminosity function. In retrospect, both were correct; Hubble had identified mostly galaxies of high surface brightness, Zwicky had discovered an exponential tail of mostly low surface brightness dwarfs. The David Dunlap Observatory (DDO) catalog (van den Bergh 1959) was the first catalog to contain significant numbers of diffuse galaxies, although most of them had low mass and so were not representative of the full range of LSB types. Meanwhile, Reaves (1956) and Arp (1965) had identified the selection effect which might lead low surface brightness galaxies to be missed. This selection bias was first clearly formulated by Disney (1976).
The discovery of low surface brightness galaxies advanced considerably in the 1980s. In a prescient piece of work, Longmore et al (1982) obtained optical and 21 cm data on a sample of 151 LSB galaxies selected by visual inspection of UK Schmidt plates. Many early studies were based on the diameter limited Uppsala General Catalog of Galaxies (UGC; Nilson 1973). With no explcit surface brightness selection, the UGC contains for example an order of magnitude more LSB galaxies than the catalog of Fisher & Tully (1981). LSB galaxies from the UGC catalog were subsequently studied by Romanishin et al (1982), who noted that they had relatively large amounts of gas for their luminosity.
Large numbers of LSB dwarfs were detected in the monumental photographic survey of the Virgo cluster by Binggeli, Sandage, & Tammann (1985) using plates taken with the Du Pont 100-inch telescope. Another large survey was carried out in the nearby Fornax cluster using both du Pont plates (Ferguson & Sandage 1988, Ferguson 1989) and sky survey plates from the UK Schmidt Telescope (Phillipps et al 1987). The surface brightness limits of both these surveys were µlim 25 B mag arcsec-2, however the spatial resolution and morphological classification is superior on the du Pont plates, which gives an advantage in defining cluster membership in the absence of redshifts.
The next improvement was offered by visual searches of the POSS-II plates, which reach to a deeper limiting isophote, µlim 26 B mag arcsec-2 (Schombert & Bothun 1988, Schombert et al 1992). Binggeli, Tarenghi & Sandage (1990) used deep Palomar plates to identify several hundred predominantly LSB dIm and dE galaxies. Impey, Bothun & Malin (1988) used photographically amplified images of Virgo to push the limiting isophote down to µlim 27.5 B mag arcsec-2 , and a similar surface brightness can be reached by automated scans of UK Schmidt plates (Irwin et al 1990a). These studies yielded new samples of extremely LSB galaxies in Virgo and Fornax (Davies et al 1988; Bothun, Impey & Malin 1991). A further gain in sensitivity can be achieved by digitally stacking scans of existing sky survey plates, or by using Tech Pan emulsions; large-scale surveys are currently underway with limits of µlim 27 R mag arcsec-2 (Schwartzenberg, Phillipps & Parker 1995b).
3.2. CCD Surveys
Digital detectors can survey for LSB galaxies in the field down to a much lower limiting isophote, but over much smaller areas (Schwartzenberg et al 1995a). CCD surveys of nearby (Turner et al 1993, Bernstein et al 1995) and more distant clusters have been undertaken (Driver et al 1994). The Texas Survey for field LSB galaxies adds a new dimension with red selection down to µlim 27.5 R mag arcsec-2 (O'Neil et al 1997). Dalcanton et al (1997b) have used strip scans made with the Palomar 200-inch telescope operating in transit mode to find galaxies with 23 < µ0 < 27 V mag arcsec-2 . This approaches the limit below which individual galaxies cannot be distinguished from distant clusters of galaxies (Schectman 1973). At the limit of deep surveys, LSB galaxies are being mined from the WFPC2 images of the Hubble Deep Field by several groups. The next major step forward in large area surveys will come with the Sloan Digital Sky Survey (Gunn & Knapp 1993).
The diversity of galaxies uncovered by these surveys is striking. An early surprise was the accidental discovery of the giant LSB disk galaxy Malin 1 in a survey of the Virgo cluster (Bothun et al 1987). This remarkable galaxy is the prototype of systems which have low surface mass density stellar disks, large physical sizes, and enormous amounts of neutral hydrogen (Impey & Bothun 1989, Knezek 1993). These galaxies are extreme cousins of the gas-rich LSB galaxies discussed by Longmore et al (1982), characterized by large exponential scale lengths (s > 10 kpc) and low central surface brightnesses (µ0 > 25 B mag arcsec-2 ). Further examples have recently been found (Bothun et al. 1990, Sprayberry et al. 1993, 1995a). Dwarf spirals have also been detected (Schombert et al 1995); both the smallest and the largest spiral galaxies known have low surface brightness. At the other extreme are LSB dwarfs, with similar surface brightness but much smaller scale lengths (s ~ 1-2 kpc). Surface brightness selection accounts for the fact that dwarf members of the Local Group continue to be discovered (Irwin et al 1990b). Even at the modest distance of the Coma cluster, many Local Group dwarf spheroidals would be too low in surface brightness to be detected in a shallow survey, and too compact to be distinguished from stars in a deep survey.