These images were obtained using either T-ReCS at Gemini South or using Michelle at Gemini-North. They were all obtained by me except for the images of NGC 7027 and of BD+30 3639 which where obtained by Scott Fisher during the commissioning of Michelle. In all the images North is up and East is to the left, as normal for astronomical images.
Ground-based imaging is not as sensitive as satellite observations but the angular resolution is much better. Most of the images shown here have a resolution of 0.36 arc-seconds, close to the diffraction limit for an 8-meter telescope. Satellite images by Spitzer or ISO have 10 to 15 times worse angular resolution than these images, so most of the objects would not be well resolved. They are also generally too bright to observe with Spitzer.
The first plot below shows the images for 22 PNs all on the same angular and intensity scale. All the images except two are in the Si-5 filter, which has a central wavelength of 11.68 microns and a width of 1.01 microns. This filter shows only dust emission, and excludes the emission lines that PNs have in the 10 micron region of the spectrum. The two exceptions are the broad N-band filter image of NGC 6210, which is dominated by the line emission at 9.0, 10.5, and 12.8 microns, and the Nprime filter image of NGC 6881. The Nprime filter overlaps with the Si-5 filter, having a central wavelength of 11.21 microns and a width of 2.13 microns. The Nprime filter image of NGC 6881 has some contribution from an emission line at 10.5 microns, which contributes about 20% of the total detected energy.
In these images the colours are representing brightness on the sky, rather than the actual colour of the objects. All these observations are at about 20 times the wavelengths that the eye can see, so there is no real "colour" that could be assigned to these images.
Two of the objects are not resolved and appear just like stars: Hb 12 and Vo 1, down in the lower right corner of the image. On a linear brightness plot like this all the resolved objects are much fainter than these two unresolved PNs.
When I show the image with a more restricted intensity scale and with square-root colour scaling to better show the faint part of the intensity range one is able to see most of the PNs. They can be rather round or elliptical or roughly rectangular in shape, and then a few of the objects look very long and cylindrical. The object M1-78 looks very different than the other objects. It is sometimes classified as a PN and sometimes classified as a young HII region around an O-type star, and the morphology suggests that the latter is the case.
To show the full range of emission I need to display the images with logarithmic scaling as in the third figure. When shown this way various types of low-level noise show up in some of the images, depending on what the weather conditions were like when the observations were taken. Some of the PNs are very faint, such as PN G058.3-10.9 and NGC 6210. In the case of PN G058.3-10.9 the star appears to be detected along with the faint surrounding nebula which has a very strange shape (but this is the same shape as is seen in the optical images). The emission from NGC 6210 in the 11.7 micron filter is even fainter than that for PN G058.3-10.9, barely within the sensitivity limit of the Gemini telescope.
For object NGC 6210 there is only very faint dust emission, and its only seen in the upper part of the nebula. The Si-5 filter image looks completely different than the N-band image. This is unusual for PNs. In most cases the images in the Si-5 filter look very much like the optical images, which seems to indicate that the dust is mixed in the ionized gas.
If the individual images are normalized to the peak emission the mosaic looks as follows:
|Nebula||Filter||Peak Brightness (Jy/square arc-second)|
|BD+30o3639||Si-5 11.7 micron||5.69|
|Fg 3||Si-5 11.7 micron||2.25|
|H1-35||Si-5 11.7 micron||4.06|
|Hb 12||Si-5 11.7 micron||49.61|
|He2-115||Si-5 11.7 micron||0.735|
|Hu2-1||Si-5 11.7 micron||0.602|
|IC 2621||Si-5 11.7 micron||6.89|
|IC 418||Si-5 11.7 micron||0.594|
|IC 5117||Si-5 11.7 micron||6.27|
|IRAS 21282+5050||Si-5 11.7 micron||7.25|
|K3-62||Si-5 11.7 micron||0.457|
|M1-71||Si-5 11.7 micron||0.751|
|M1-78||Si-5 11.7 micron||10.91|
|NGC 6210||Si-5 11.7 micron||0.0127|
|NGC 6210||N broad||0.0530|
|NGC 6644||Si-5 11.7 micron||0.625|
|NGC 6881||N prime||0.413|
|NGC 7027||Si-5 11.7 micron||5.90|
|Pe1-7||Si-5 11.7 micron||1.78|
|PN G011.3-09.4||Si-5 11.7 micron||0.957|
|PN G058.3-10.9||Si-5 11.7 micron||1.13|
|SwSt-1||Si-5 11.7 micron||10.28|
|Vo 1||Si-5 11.7 micron||27.06|
The point of taking all these images was to find objects with dust beyond the ionized region or with very different morphology in the thermal dust emission than in the ionized gas. After looking at more than 20 objects there are only a few cases where there is any difference in the appearance of the PN in different filters: NGC 6210 is one, and NGC 6644 is another. The figure below shows how NGC 6644 looks in different filters: one on the strong line at 10.5 microns, one on the UIR feature at 11.3 microns, a filter at 18.1 microns, and the 11.7 micron filter. An HST image in the H alpha emission line is shown on the same angular scale for comparison. I have used an image sharpening technique on the three T-ReCS images, namely the Lucy deconvolution technique.
The deconvolution used on the images to sharpen them normally produces very good results (the mottled structure in the three deconvolved images is due to noise in the original images and should be ignored). The indication is that the bulk cold dust is distributed at the edges of the ionized region, while the UIR carrier is distributed quite differently. The [SIV] image looks similar to the optical H alpha image when deconvolved, which is as expected (and which implies that the deconvolution is working properly).
The conclusion seems to be that for most objects the dust is well mixed in the ionized region, and if there is dust outside the main part of the nebula it is too cold to detect at these wavelengths. Since its hard for dust to survive long in the plasma, where UV radiation and the electron plasma are expected to break up the grains on a relatively short timescale, this was not the result I was expecting when I started the project.
One of these days I will find the time to publish these results....