6.3 Other Calibrations
6.3.1 Historical Galactic Supernovae
Early attempts to use the historical galactic supernovae to calibrate SNe Ia were based on Tycho's supernova of 1572 (van den Bergh 1970, de Vaucouleurs 1985). Recent improved estimates of distances to the remnants of historical supernovae (Strom 1988) imply, however, that SN 1572 (MB = -18.0) was significantly less luminous than SN 1006 (MB = -20.0) and Kepler's supernova of 1604 (MB = -19.7), which has led Green (1986) and Strom (1988) to suggest that SN 1572 may have been a Type Ib. In any case SN 1572 should not be used to calibrate SNe Ia, which are the brightest supernovae. Distances from Strom, apparent magnitudes from Clark and Stephenson (1977, 1982), and extinction estimates from various sources lead to a mean absolute magnitude for SNe 1604, 1006, and the (extremely uncertain) SN 185 (MB = -20.2) of MB = -20.0 ± 0.5. It seems unlikely that the remnant distance scale will decrease, because currently favored distances such as Strom's lead to a surprisingly high local supernova frequency that could be reduced by increasing the remnant distances (van den Bergh and Tammann 1991).
6.3.2 Nearby Extragalactic SNe Ia
Among the SNe Ia that have appeared in galaxies within a few megaparsecs, SN 1937C in IC 4182 and SNe 1895B and 1972E in NGC 5253 were observationally normal and not highly reddened, while SN 1986G in NGC 5128 was observationally peculiar and highly reddened and therefore cannot be trusted as a calibrator.
Sandage and Tammann (1982) derived a distance modulus µ0 = 28.21 ± 0.3 (4.4 ± 0.6 Mpc) for IC 4182 by assuming an average <MV(3)> = -7.72 for its three brightest red stars. SN 1937C had B = 8.7 ± 0.2 which, if extinction is negligible (Sandage and Tammann 1982), gives MB = -19.51 ± 0.36. Very recently Pierce et al. (1992) have derived a shorter distance to IC 4182 using I- and K-band photometry of its brightest red stars. Their calibration implies MB = -18.8. Observation of Cepheids in IC 4182 are needed to resolve the discrepancy.
Van den Bergh (1989) reviews the distance to the NGC 5128/5236 group, which he takes to include NGC 5253. Based on the planetary-nebula distance for NGC 5128 relative to that of M31 (Jacoby et al. 1988), van den Bergh recommends µ0 = 27.95 ± 0.13 (3.9 ± 0.2 Mpc) for NGC 5128 and concludes that most of the group members have distances in the range 3.2-4.6 Mpc. With a mean B = 8.35 ± 0.2 for SN 1972E and 1895B, a foreground extinction AB = 0.12 (Burstein and Heiles 1978), and negligible parent-galaxy extinction (Caldwell and Phillips 1989), the distance range of 3.2-4.6 Mpc leads to MB = -19.77 ± 0.45.
With standard assumptions SN 1986G comes out to be less luminous. Adopting B = 12.5 ± 0.1 and E (B - V) = 0.9 ± 0.1 (Phillips et al. 1987, Rich 1987), R = AV / E (B - V) = 3.0 ± 0.5, and µ0 = 27.95 ± 0.13 gives MB = -19.05 ± 0.55. This is fainter than SNe 1972E and 1895B unless NGC 5253 is in front of the NGC 5128/5236 group (de Vaucouleurs 1979, but see also van den Bergh 1980) or the extinction has been underestimated.
Current estimates based on thermal emission assume that the optical flux at the supernova photosphere can be approximated by the flux emitted by a blackbody having the same optical (or optical-infrared) color temperature. The radius of the photosphere can be derived by a method analogous to Baade's (1926) method for variable stars (Branch and Patchett 1973) or simply taken to be the product of the expansion velocity at the photosphere and an estimated rise time from explosion to maximum light (Arnett 1982b, Branch et al. 1983). During most of their evolution the energy distribution of SNe Ia is non-Planckian, but for a brief time near 25 days after maximum light the energy distribution from the U to the K band is close to that of a blackbody having T = 6000 ± 1000 K (Leibundgut 1988). With a velocity at the photosphere of 9500 ± 500 km s-1 (Branch et al. 1988), a rise time to maximum light of 17 ± 3 days, and the assumption of blackbody emissivity, the absolute magnitude 25 days after maximum light becomes -18.2 ± 1.0, which corresponds to MB = -20.4 ± 1.0 at maximum light. Very recently Jeffery et al. (1992) have applied this method to near-maximum-light spectra to derive MB = -19.8 for SN 1990N and -20.0 for the peculiar SN 1991T. These simple thermal emission estimates have the virtue of being independent of all intermediate astronomical distance calibrations but a critical uncertainty is how well the blackbody flux approximates the true optical flux (cf. Harkness 1987). The external error in the thermal emission estimates will become known when detailed spectrum calculations for SNe Ia (Branch et al. 1991) have been performed.