THERMAL DUST IMAGING OF THE ULTRACOMPACT H-II REGION G29.96-0.02

We have obtained mid-infrared images of the ultracompact H II region G 29.96-0.02 at 1 '' resolution. These include three broadband images at 8.5, 9.7, and 12.0 mu m and three narrowband images covering the spec trum around the 12.8 mu m [Ne II] emission line. The appearance of the object in all bands is basically similar to the arcuate or ''cometary '' shape seen in previously published high-resolution radio maps. At o ur shortest wavelength of 8.5 mu m, the faint, diffuse emission from t he cometary ''tail'' is enhanced compared to its brightness at other w avelengths. Taking advantage of well-matched passbands, we have used t he ratio between the images at 12.0 and at 8.5 mu m to construct maps of the dust color temperature T-d and emission optical depth tau(em) i n the object. The optical depth is sharply peaked along a long, thin r idge near the leading edge of the cometary are. The shape of this ridg e is nearly identical to that seen in the high-resolution 2 cm maps, s trongly suggesting that the dust which is warm enough to emit at 8 and 12 mu m is inside or extremely close to the zone of ionized gas in G2 9.96-0.02. This conclusion is further bolstered by the essentially ide ntical distribution of the [Ne II] emission and the thermal dust conti nuum at adjacent wavelengths. The key feature of the dust temperature map is its lack of marked variation. The great majority of the nebula shows temperatures between 190 and 240 K. The small variations that do occur present a clear pattern. At the center of the ridge of high tau (rmem), T-d is around 210 K. Moving outward to the wings of the are, t he temperature falls slowly to about 190 K at the extreme ends. On the other hand, moving back along the symmetry axis into the nebula's tai l we find that mean temperatures increase to around 230 K, with a good deal of scatter. We have tried modeling these temperature variations using some simple concepts derived from earlier radiative transfer wor k on related objects, together with a published model of G29.96-0.02 a s a bow shock. Unlike our previous success in applying such a framewor k to our observations of the ultracompact source G5.89-0.39, the model s fail to match the temperature pattern in G29.96-0.02. Of several pos sible processes which may account for the discrepancy, an increased ab undance of small grains caused by shock processing of the dust seems t o present the fewest obvious difficulties. (C) 1996 American Astronomi cal Society.