The late-time bolometric luminosity of SN 1987A

We present infrared photometry of SN 1987A in the J, H, K, L, M, 10, and 20-mu-m filters through day 1352 since the explosion of the supernova. We find that both the optical and mid-infrared magnitudes are slowly leveling out through day 1352 after their rapid decline in brightness during the epoch of dust formation. The near infrared colors (V - J, V - H, and V - K) initially reddened after dust formation around day 500, but by days 900-1000, they began to evolve to the blue, perhaps due to the thinning of the dust. Optical spectra obtained through day 1150 show no evidence for a sudden drop in the optical emission-line luminosities. This suggests that the region of optical emission has not undergone the so-called "infrared catastrophe," when the cooling of the ejecta becomes dominated by the infrared fine-structure lines, at least in the case described by the simple models of Fransson & Chevalier [ApJL, 322, L15 (1987)]. More complicated models where the higher density regions are cooled by the fine-structure lines at the same time that the nebular dust is thinning could be qualitatively consistent with the data. The infrared data have been combined with optical photoelectric and CCD UBVRI photometry obtained at CTIO to study the temporal evolution of the bolometric luminosity of SN 1987A. By day 1000, another source of energy beside the radioactive decay of Co-56 is needed to explain the slow leveling off to the bolometric luminosity decline. We can formally fit our data with the following energy sources: 5 +/- 1 times the predicted ("solar") amount of Co-57/Co-56; 10 +/- 3 times the predicted amount of Ti-44; or a constant energy input of 37.3 +/- 0.15 dex [log10 (ergs s-1)]. The energy input is unspecified; it could be a pulsar, light echo, or other unidentified energy source. Only the enhanced Co-57/Co-56 provides a good fit at all times to the data. The enhanced Co-57, however, is in conflict with published hard x-ray and infrared spectral data. A careful reanalysis of the uncertainties in the observed hard x-ray and ultraviolet-optical-infrared fluxes, and the model predictions, suggests that all the data can be fit by the energy deposition from 0.075M. of Ni-56, a Co-57/Co-56 between 2.5 and 4 times "solar", and "solar" Ti-44.