Shortly after its outburst, we suggested that supernova 1987a might condense a dust shell of substantial visual optical thickness as many classical novae do and predicted that dust might form within a year after the explosion. A critical examination of recent optical and infrared observations reported by others confirms that dust grains had begun to grow at a temperature of 1000 K after 300 days and that the dust shell had become optically thick by day 600. After day 600, the infrared luminosity closely followed the intrinsic luminosity expected for thermalized 56Co y rays, demonstrating that the luminosity is powered by radioactivity and that the dust is outside the radioactivity zone. The infrared luminosity sets an upper limit to the soft intrinsic bolometric luminosity of a pulsar central engine. This upper limit for the pulsar in supernova 1987a is the same luminosity as the Crab pulsar has today 936 years after its formation. It is unlikely that the rotation rate for a pulsar in supernova 1987a can be much higher than %30 revolutions per sec. The relatively long time required for the shell to grow to maximum optical depth as compared with the dust in nova shells may be related to the relatively low outflow velocity of the condensible ejecta.that dust may have formed (11), but the shell was probably optically thin because the short wavelength brightness of the supernova remained fairly high. A similar infrared excess combined with low luminosity at short wavelengths suggested the formation of an optically thick shell in SN 1979c (see paper 1). In neither case did the optical/infrared observations appear to cover the initial stages of dust formation; nor did they give any information on the extinction by the dust. Estimates of the dust shell optical depth in these supernovae are dependent entirely on assumptions about the temporal evolution ofthe luminosity of an unspecified central engine. Some investigators suggested that the infrared excesses in these cases might come from light echoes (12, 13).SN 1987a presents a much more favorable opportunity to observe dust formation.
Review of Our Prediction of Dust Formation in SN 1987aIn paper 1, we argued that an optically thick dust shell could form in a time tc:In Gehrz and Ney (ref. [2]where M is the mass ofthe ejected shell. Novae that condense optically thick dust shells were found to have Pc 2 3 X 10-16