Annu. Rev. Astron. Astrophys. 1998. 36: 189-231 Copyright © 1998 by Annual Reviews. All rights reserved

2.4. Forbidden Lines

The H emission line is redshifted out of the visible window beyond z ~ 0.5, so there is considerable interest in calibrating bluer emission lines as quantitative SFR tracers. Unfortunately, the integrated strengths of H and the higher order Balmer emission lines are poor SFR diagnostics because the lines are weak and stellar absorption more strongly influences the emission-line fluxes. These lines, in fact, are rarely seen in emission at all in the integrated spectra of galaxies earlier than Sc (Kennicutt 1992a; also see Figure 1).

The strongest emission feature in the blue is the [OII]3727 forbidden-line doublet. The luminosities of forbidden lines are not directly coupled to the ionizing luminosity, and their excitation is sensitive to abundance and the ionization state of the gas. However, the excitation of [OII] is sufficiently well behaved that it can be calibrated empirically (through H) as a quantitative SFR tracer. Even this indirect calibration is extremely useful for lookback studies of distant galaxies because [OII] can be observed in the visible out to redshifts z ~ 1.6, and it has been measured in several large samples of faint galaxies (Cowie et al 1996, 1997;, Ellis 1997, and references therein).

Calibrations of SFRs in terms of [OII] luminosity have been published by Gallagher et al (1989), based on large-aperture spectrophotometry of 75 blue irregular galaxies, and by Kennicutt (1992a), using integrated spectrophotometry of 90 normal and peculiar galaxies. When converted to the same IMF and H calibration, the resulting SFR scales differ by a factor of 1.57, reflecting excitation differences in the two samples. Adopting the average of these calibrations yields

(3)

where the uncertainty indicates the range between blue emission-line galaxies (lower limit) and samples of more luminous spiral and irregular galaxies (upper limit). As with Equations 1 and 2, the observed luminosities must be corrected for extinction, in this case the extinction at H, because of the manner in which the [OII] fluxes were calibrated.

The SFRs derived from [OII] are less precise than those from H because the mean [OII] / H ratios in individual galaxies vary considerably, over 0.5-1.0 dex in Gallagher et al's (1989), Kennicutt's (1992a) samples, respectively. The [OII]-derived SFRs may also be prone to systematic errors from extinction and variations in the diffuse gas fraction. The excitation of [OII] is especially high in the diffuse ionized gas in starburst galaxies (Hunter & Gallagher 1990, Hunter 1994, Martin 1997), enough to more than double the L[OII]/SFR ratio in the integrated spectrum (Kennicutt 1992a). On the other hand, metal abundance has a relatively small effect on the [OII] calibration, over most of the abundance range of interest (0.05 Z Z 1 Z). Overall the [OII] lines provide a very useful estimate of the systematics of SFRs in samples of distant galaxies, and they are especially useful as a consistency check on SFRs derived in other ways.