2000ApJ...531..624D

We describe an object in the Hubble Deep Field North with very unusual near-infrared properties. It is readily visible in Hubble Space Telescope NICMOS images at 1.6 µm and from the ground at 2.2 µm, but it is undetected (with S/N≲2) in very deep WFPC2 and NICMOS data from 0.3 to 1.1 µm. The f_ν flux density drops by a factor ≳8.3 (97.7% confidence) from 1.6 to 1.1 µm. The object is compact but may be slightly resolved in the NICMOS 1.6 µm image. In a low-resolution, near-infrared spectrogram, we find a possible emission line at 1.643 µm, but a reobservation at higher spectral resolution failed to confirm the line, leaving its reality in doubt. We consider various hypotheses for the nature of this object. Its colors are unlike those of known Galactic stars, except perhaps the most extreme carbon stars or Mira variables with thick circumstellar dust shells. It does not appear to be possible to explain its spectral energy distribution as that of a normal galaxy at any redshift without additional opacity from either dust or intergalactic neutral hydrogen. The colors can be matched by those of a dusty galaxy at z≳2, by a maximally old elliptical galaxy at z≳3 (perhaps with some additional reddening), or by an object at z≳10 whose optical and 1.1 µm light have been suppressed by the intergalactic medium. Under the latter hypothesis, if the luminosity results from stars and not an AGN, the object would resemble a classical, unobscured protogalaxy, with a star formation rate ≳100 M_☉.yr^–1. Such UV-bright objects are evidently rare at 2<z<12.5, however, with a space density several hundred times lower than that of present-day L^* galaxies.