THE INTERNAL STRUCTURE OF MOLECULAR CLOUDS .3. EVIDENCE FOR MOLECULARDEPLETION IN THE NGC-2024 CONDENSATIONS

Maps of the star forming region NGC 2024 in the mm-wave transitions of (CS)-S-34, (CO)-O-17 and (CO)-O-18, obtained at the IRAM 30-m telesco pe, with angular resolutions of 21" to 12" are presented. Using the Ve ry Large Array, NGC2024 has been imaged in the (J,K) = (1,1) and (2,2) inversion lines of NH3 and the lambda = 1.3 cm continuum with a resol ution of 3". The mm-wave dust emission peaks NGC2024:FIR 1 to 7 report ed by Mezger et al. (1988, 1992) to be candidates for protostellar obj ects are not seen in (CO)-O-18. The contrast in the (CS)-S-34 maps is similar to that found in dust emission, but positions and angular size s of the maxima do not agree very well. The emission of the CS and CO isotopes arises from a ridge which has a minimum coinciding with two c rescent shaped regions of free-free emission, the northern continuum p eak (NCP) and the southern continuum peak (SCP). This ridge has a coun terpart in the dust emission. For each of the compact dust emission ma xima FIR 1 to 7, we find a corresponding compact NH3 source with good positional agreement. The sizes and orientations of the NH3 peaks are similar to the values derived for the dust emission peaks with the exc eption of FIR/NH3 4 and 5 where the NH3 emission arises from a smaller area. From a comparison with single dish data, the bulk of the NH3 em ission must come from the more extended molecular ridge rather than fr om the compact condensations. Only for NH3/FIR 4 and 5 are kinetic tem peratures derived from ammonia greater-than-or-equal-to 40K. For the o ther NH3 cores T(kin) is < 18 K to 23 K. This is significantly colder than the molecular ridge, where the the kinetic temperature is greater -than-or-similar-to 30 K and implies that those dust continuum peaks a re column density maxima rather than temperature maxima. Toward FIR 1, 2, 3, 6, and 7, the linewidths of the ammonia profiles are much small er than those of single dish CS or CO isotopes, and lead to virial mas ses which are approximately 1/10 the masses determined from the dust c ontinuum emission. The average relative abundance of NH3, calculated f rom the virial masses is greater-than-or-similar-to 10(-9). It is even lower for the molecular regions NH3/FIR 4 and 5, nearest the NCP and SCP. If the models of grain emission properties are not seriously in e rror, our data imply that for the medium density ridge, molecules are abundant in gas phase. This leads to similar H-2 column density estima tes from dust and from molecules. NH3 emission might arise merely from part of the surface of the compact condensations; the cold, dense gas phase environment of the condensations is, however, devoid of NH3 and other polar molecules. One explanation is to postulate large dust/gas temperature gradients in the FIR sources and a chemical destruction o f NH3 and CS, which would imply that stars are already present in the condensations. (CO)-O-18 is, however, not affected by such chemistry, which makes it more straightforward that the molecules are depleted du e to freezing out onto grain surfaces.