Abell 66: A Comparative Analysis of Images Taken Now Versus Taken 'Back Then'
(A.K.A: …. PN G 019.8-23.7, ESO 595-4)

OTA: CDK24"
Camera: Morovian 61000 Pro
Observatory: Heaven's Mirror, Lat -30.4708, Long -70.7647
Date of Capture: Oct 2023
Date of Processing: Oct 2023

Exposures:
R: 10 x 180 sec
G:   9 x 180
B: 10 x 180
H: 22 x 1200 sec
O: 28 x 1200
Total Exposure time: 14 hours
Image Width: 30'

Processing Tools:
1.    Commercial: PixInsight and Topaz Studio2
2.    Pixinsight Addons: NoiseXTerminator, BlurXTerminator, and StarXTerminator
3.    My Scripts: NB_Assistant,  Subframe Weighting Tool (Excel w/ J. Hunt)

Background:
  The faint planetary nebula Abell 66 lies at a declination of -21.5 degrees. While observable from much of the northern hemisphere, imaging it from there can be challenging. Being considered an "old" and faint planetary nebula undoubtedly contributes to the sparsity of studies it has received. However, for the Heaven's Mirror 24" scope, in N. Chile, this target transits almost over-head.

   This discussion compares the analyses of Abell 66 made by Hua et al. (1) in 1998 using 2m telescopes and early model CCD cameras with new results from a 24" amateur telescope (CDK24) hosting a modern CMOS camera (Moravian 61000 Pro downsampled 2x at time of capture) with modern narrowband filters (3 micron) and benefiting from modern image-processing tools and techniques (largely PixInsight and Topaz Studio2, in this study).

   Study (1) utilized several New Zealand and French telescopes in the 2m-aperture range to capture images through Ha, OIII, and NII narrowband filters with a 1.5-nanometer bandpass. Although not reported, their Techtronics cameras' specifications are unlikely to be competitive with modern astro-camera resolution, bit depth, or QE. Similarly, image-processing technologies have vastly advanced on many fronts since (1) published their results. For example, the new mages underwent AI denoising and a small amount of AI sharpening. Further enhancement comes from "precision contrast" and "precision detail" in Topaz Studio2. These algorithms operate on distinct segments of the dynamic ranges of their corresponding parameters. Hua et al. report seeing as 1.5" to 2". FWHM of the image used here was measured as ~1.5" on a single, randomly selected, red subframe. Also important, (1)'s images of Abell 66 apparently targeted the faintest out-ring of Abell 66 and badly over-exposed much of the bright inner ring, making recognition of detail there untenable.

   Hua et al. discussed the models for planetary nebula (PNe) formation forwarded by Dgani and Soker 1977 (2), Dwarkadas and Balick 1998 (3), and Kwok, et al. 1978 (4). Those studies strove to explain the shapes of planetary nebulae and ascribe parts of those shapes to interaction with those winds as well as being products of two different phases of the nebula's formation. The first wind that ejects the outer ring of ionized gas travels relatively slowly and arises while the star resides on the "asymptotic giant branch" of the Russel-Hertzsprung diagram. A faster wind arises during the star's PN stage, accounting for the inner, more massive ring. Among several PNe, (1) imaged Abell 66, attempting to evaluate the presence/absence of the slow wind/outer loop and to characterize the morphologies produced by the two types of winds.

   This study revisits (1) 's conclusions using Ha and OIII images with their stars removed by StarXTerminator. Removing the stars before post-processing facilitates better-quality results from the subsequent processing steps and visually clearer comparison of the images. Furthermore, I applied an image-tuning method that approximately separates the omnipresent continuum radiation from the Ha and OIII emission lines by solving simultaneous equations involving broadband images in R, G, and B and the Ha and OIII narrowband images. Post-processing the images (after stretching the starless images) focused on maximizing fine-scale contrasts and resolution but did not intentionally prefer linear structures to other structure geometries.

   As in (1), my images have North up and East to the right. In the Astrobin presentation, the Ha image appears when the mouse is outside the picture area, and the OIII image appears at the same scale and alignment when the mouse hovers over the image. Each image can be viewed in more detail by selecting it from the thumbnail images on the right.

Observations:
The Ha and OIII images show a bipolar geometry of the outer shell. The outer shell's emission appears more substantial in the Northern and Southern parts of the nebula. The inner (high-speed) cloud has less bipolarity than the outer (slow-speed) cloud. However, the OIII component of both the inner and outer rings appears more extensive than the Ha component. This may reflect a different Ha/OIII ratio with OIII dominating.

   The early, slower ring displays almost no internal structure in the Ha component, with more, but still not pervasive, structure appearing in the OIII component. The apparent lack of well-defined structures in the outer ring could result from insufficient image exposure, the observation geometry, or simply due to the way it formed. The current observations do not suffice to resolve these alternatives.

   The inner ring OIII and Ha images show considerable structure but not identical structures at identical locations. The OIII inner ring has radial structures dominating. Many individual sharp radial features host crossing structures with a more puffy, cloud like appearance. Similar abundant radial structures were reported for the Helix Nebula by O'Dell, et al. (5); see their Figure 6. The radial flow in the Helix Nebula's Ha cloud segments into finer radial structures or streams ("cometary knots"?) that largely lack the crossing patterns seen in the Abell 66. The presence of interstellar winds impacting Abell 66 and not impacting the Helix Nebula might account for this difference. The OIII cloud's structures in Abell 66 also become less distinct in the E-W direction location, perhaps another effect of the interstellar medium/winds. Still, the bipolar origin of the rings could also cause these distortions.

   The intervening space between the source star and the inner ring also differs between the OIII and Ha images. However, both images show the region bisected by an NE-SW direction nebula. Although this bisection could have arisen due to a binary star system, it appears more likely to be a foreground segment of the inner cloud along our line of sight.

   Further evidence for this interpretation comes from the faint structures observed on this inner "floor" in both the Ha and OIII images. These structures have the same general pattern observed in the surrounding inner ring. They may indicate continuing or lagging radial outward flow in the plane of the inner ring—assuming the inner ring is planar. A planar geometry fits the dynamic measurements made by O'Dell, et al. (5) for Helix Nebula. The same geometry may apply for Abell 66. (PNe we see more edge-on could have the geometries more like that expressed by SaWe 3 and NGC 6302.)

   Hua, et al. also comment on the overall attributes of the nebula observed in their narrowband images. They conclude that the OIII has no significant indications of an outer slow-wind shell, and the inner ring lacks detailed structures. Furthermore, they noted the "tail" (the brighter part of the northern loop of the outer shell) they observed in the Ha image was absent in the OIII. (See discussions of their Figure 3.) However, the images in this study suggest otherwise on both counts. The OIII image plainly shows an extended outer ring, which probably surrounds the entire inner ring. As noted above, Hua, et al.'s data generally have an over-exposed inner ring. Apparently, they infer the presence/absence of detail from the complexity observed along the outer margin of that ring. Given the improved images presented here, neither of these conclusions appears to be valid.

   Why do the Ha and OIII ionized clouds have markedly different fine-scale structures, often with different orientations? These features may arise from different coupling strengths between the different ion specs and the ambient magnetic field (if present) or with the interstellar medium and wind.

   A final thought for imagers: combining the Ha and OIII images shown here blends the details in both images. Given the significant differences in the detail in these two images, blending them together will confuse the detail geometries and may result in an image having neither Ha or OIII properties but instead display a hodgepodge of both.

Are the Image Details Reliable?
   Preprocessed in PixInsight, given the long history of reliability, that processing stage will likely produce reliable representations of the image content. A variety of Topaz programs facilitated refinement/post-processing, including Topaz Studio2, Denoise AI, and Sharpen AI. Topaz is a very aggressive image-processing tool. However, after similar image processing, I could discern most of the exact details using an online program.

   I could not obtain the same level of detail in less aggressive image-processing programs—but that could have been expected because these features appear close to the images' resolution limits, although clearly above that limit. However, the broader structures showing the radial spikes appear in other processing, just not the fine details. I varied the order in which Topaz tools were applied and generated the same results.

   Several ancillary facts suggest that the new images accurately record the fine structures. First, the Rayleigh Criterion gives a theoretical resolution of 0.27" for this telescope at the Ha wavelength. Even adding a generous allowance for the Nyquist limit, the expected resolution exceeds 0.8". The 1.5" measured FWHM of the stars yields a Gaussian distribution standard deviation of 0.63" that does not reach the Nyquist limit, making the Nyquist limit the controlling factor. The finest-scale radial stringers have about a 2" measured width, which should be resolvable in these images. Furthermore, these images have been sharpened using AI methods (which do not impart the ringing artifacts associated with deconvolution). The 1.5" FWHM measured on the raw images has been reduced to a measured 0.9" median value in the final, fully processed images. (Based on parallel processing of the starless and star-retaining images.) Overall, we should expect the final image to reveal finer-scale details than visible in the raw images.

   Logical arguments also suggest the validity of the fine-scale radial stringers. Namely, their orientation coincides with the approximate radial and tangential directions associated with the star's eruptions, as might be the expected dominant flow directions of physical events that caused them. The structures do not align along pixel rows or columns of the images; they are not rigidly parallel to one another, nor do they have simple, absolutely straight morphology for their entire lengths, as might be expected from a capture or processing artifact. Furthermore, they are of varying lengths, separations, and intensities; again, variabilities that might not arise from artifacts. Finally, the boundaries between stringers are not abrupt. The darkest areas grade through several pixels into the brighter areas, then grade back again into the dark areas. This situation suggests the physically natural behavior of a closely spaced dynamic material.

   All these observations and calculations make the physical reality of these fine-scale features the most likely explanation, but they do not absolutely prove them. We might hope that images from large earth and space telescopes might have already probed for features. I have not seen any such image. But that does not rule out their existence. These publically-funded research telescopes commonly use unsubscribed time to create public outreach images and meticulous image processing probably holds a low priority for those images. Perhaps this perspective is incorrect, but professional scientists rely more on state-of-the-art instruments and burning research issues. If not on some scientist's to-do list, the simple things discussed here may well be overlooked.

REFERENCES:
1.    Hua, C.T., M.A. Dopita, J. Martinis, Astron. Astrophys. Suppl. Ser. 133, 1998. (https://aas.aanda.org/articles/aas/pdf/1998/21/ds7743.pdf)
2.    Dgani, R, N. Soker, 1977 (only appeared as a preprint?)
3.    Dwarkadasm /V.V., B. Balick, AstroJ, 1988.
4.    Kwok S., CR Purton, M.P. FitzGerald, ApJ, 1978.
5.    O'Dell, C.R., P.R. McCullough, M. Meixner, AstroJ, 2004. (https://iopscience.iop.org/article/10.1086/424621/pdf)

Statistics:
Distance: 1800 ly
Apparent Magnitude: 14.9
Average Surface Magnitude: 27
Pixel Span at Target: 3 Billion km

Alex Woronow, PhD