6. Discussion: the nature of the 6.75µm sources
A total of 54 were detected in the field. Of these, 9 could not be identified with any optical counterparts to the limit of our optical images (); two sources with S/N 4 and seven more with S/N 3 remain unidentified. Of the remaining 45, there are 21 "secure identifications" (catalogues 1 & 2). Optical spectra are available for 14 of these showing that 7 are stars (mostly K and M) and 5 out of the 6 remaining sources display AGN or starburst activity. Thus, most of the strongest extragalactic sources with secure identifications are AGNs. The mid-IR flux in AGN host galaxies is believed to come directly or indirectly from the hot dust in the inner regions of a torus around the active nucleus. Two of the "secure identifications" have , the limit of the CFRS spectroscopic survey.
Analysis of the properties (photometric, spectroscopic, radio) for all sources with optical counterparts indicates that a large fraction of even the "less secure" counterparts are likely to be the correct identifications. There are 45 identified sources in total and 7 are spectroscopically confirmed stars, leaving 38 possible extragalactic sources (although 3 of these have stellar profiles). Ten of these have and are fainter than the CFRS survey limit, but spectra are available for 14 of the remaining 28 counterparts. As Table 3 indicates, 53% are classified as , showing evidence of an A star population and star formation activity, indicative of significant star formation 0.5 Gyr ago. Detection of galaxies in the mid-IR is supported by preliminary results from ISOCAM LW3 15µm data (Flores et al. 1998).
The average ratio between the energy () at µm and the visible energy (0.835/(1+z) µm) is high, 2.1 0.7 on average, for the 9 galaxies exhibiting star formation activity. It is higher than the value estimated for a local starburst galaxy after redshifting it to the median redshift (z=0.71) of our sample of star forming galaxies, for which we find =0.63. Fig. 12 shows the redshift distribution of this ratio, compared to local templates (starburst, elliptical and QSOs from Schmitt et al. 1997). Although several objects appears compatible with local galaxies, Fig. 6 indicates that Mid-IR to optical flux ratio is not sufficient to determine the nature of the source emission. On the other hand, some sources appear redder than any kind of local templates. This does not seem related to a deficiency in our flux calibration at 6.75µm (stars do not show the same excess when compared to Rayleigh-Jeans emission distribution), and the excess is apparently higher than our expectations for photometric errors. Indeed there are several bright ISOCAM sources (including the brightest one, CFRS14.1157) which present this red excess. From a purely statistical point of view, this is not unexpected, since the ISOCAM detections correspond to the small fraction (5%) of galaxies at which are the most extreme mid-IR emitters. On the other hand this might indicate that several field galaxies up to z=1 have higher Mid-IR fluxes (related to PAH or to hot dust) when compared to local galaxies.
Some of the unidentified sources and some of the sources with counterparts for which we do not yet have spectra may be at z 1. The fact that a) many of the counterparts without spectra are fainter than average and b) that some of these are very red in supports this possibility. On the other hand, they may also simply be highly reddened. A crude estimate of the number of high z () mid-IR emitters can be made by adding the following:
In other words, the maximum number of ISO sources which could be at z 1 is 14, or 30% of the 44 sources which are non-stellar (7 "spectroscopic" stars and 3 "stellar profile" objects excluded).
Perhaps the most intriguing object in our ISO sample is CFRS14.1157, which is the brightest non-stellar object in our LW2 image. It is likely to be a heavily absorbed AGN at z = 0.216. Further observations of this source are warranted. Deeper and more complete near-IR imaging and spectroscopy of a number of the optical counterparts and the unidentified sources would also be of interest.
© European Southern Observatory (ESO) 1999
Online publication: March 1, 1999