This image shows the famous circumstellar disc of debris and dust orbiting the star Beta Pictoris 63.4 light years away. This is a very young system thought to be only around 12 million years old and is essentially similar to how our own Solar System must have formed some 4.5 billion years ago. The disc is seen edge-on from our perspective and appears in professional images as thin wedges or lines protruding radially from the central star in opposite directions.
For the last couple of years I have been wondering if it was possible for amateurs to capture this special target but have never come across any such images. The main difficulty is the overwhelming glare from Beta Pictoris itself which completely drowns out the dust disc that is circling very close to the star. Images of the disc taken by the Hubble Space Telescope, and from big observatories, are usually made by physically blocking out the glare of Beta Pictoris itself within the optical path.
But I then found this excellent 1993 paper 'Observation of the central part of the beta Pictoris disk with an anti-blooming CCD' (Lecavelier des Etangs, A., Perrin, G., Ferlet, R., Vidal-Madjar, A., Colas, F., et al., 1993, A&A, 274, 877)
Full article available here: http://adsabs.harvard.edu/abs/1993A%26A...274..877L

I realised that with this technique it might not be entirely impossible to also record the debris disc with relatively modest equipment. I followed the technique described in the paper above, which basically consists of imaging Beta and then taking another image of a similar reference star under the same conditions. The two images are subtracted from each other to eliminate the stellar glare, and the dust disc should then hopefully reveal itself. However, since the two stars have different magnitudes I needed to calculate how long to expose Alpha for in order to get a similar image which I could subtract from the Beta image:

The magnitude difference between the stars is 3.86(Beta) - 3.30(Alpha) = 0.56
Due to the logarithmic nature of the magnitude scale we know that a difference of 1 magnitude equals a brightness ratio of 2.512. Therefore 2.512 to the power of the numerical magnitude difference then equals the variation in brightness.
2.512^0.56 = 1.67, so it appears Alpha is 1.67 times brighter than Beta. This means that exposure for Alpha should be 1/1.67 = 0.597x that of Beta.

With this approach I managed to capture the first amateur image of the debris disc on 16th November 2011 and the result can be seen here: https://pbase.com/rolfolsen/image/139722640/original
This image generated a lot of attention from both professional and amateur astronomers and was reported in the media all over the world.

After the initial success of capturing the disc I was contacted by several people in the astronomy community, among which were Dr Grant Christie of Auckland's Stardome Observatory. Following his suggestions I had a go at imaging it again using shorter exposures. This was to minimise the area saturated by Beta Pictoris itself and potentially reveal more of the debris disc closer to the star. The ICX098BQ chip in the ToUCam saturates fairly quickly for bright stars, even at very short exposures. Since long exposures with my ToUCam can only be controlled in 0.5s increments I decided to use 7.0s and 4.0s for Beta and Alpha respectively, which translates to a factor of 0.571. This was very close to the calculated brightness factor of 0.597 and still significantly shorter than the 30.0s I used for the first image on 16/11/2011.

So for this second image I collected 344 images of Beta Pictoris at 7.0 seconds each and 299 images of Alpha Pictoris at 4.0s each. Both sets of images were dark subtracted and stacked separately in Registax. I then subtracted the Alpha image from the Beta image in PixInsight LE, and also created the absolute difference between the two.
The absolute difference image is simply easier to work with, but the subtraction image was important as a reference to examine which of the stars had contributed the various parts of the difference. Below are both the subtraction result (left) and the absolute difference (right):







In the subtraction image on the left contributions from Alpha Pictoris are dark, while contributions from Beta Pictoris are light. It is clear that there is a fairly strong signal corresponding to the exact location of the debris disc and that it is coming from the Beta image.

I created a more natural looking final image by taking the original stacked Beta image and then blending in the central parts from a stretched version of the absolute difference image that showed the dust disc. I decided to also keep the black spot of the central glare from the difference image since the contrast with the protruding disc just seems better this way. So there is no occulting disc involved, it is simply for the sake of presentation. I created a couple of versions which can be seen in the gallery below. This is a vastly better image than the first one taken on 16/11/2011. I believe the higher number of subframes (344 versus 55) coupled with the shorter exposure times are responsible for the improvement.

I used MaximDL to produce some more in-depth illustrations of what is going on in the difference image. First a area plot of the intensity immediately around Beta Pictoris. The circular plateau in the centre corresponds to the saturated area caused by Beta Pictoris itself (The narrow trough immediately surrounding it is an artifact of the image processing). The debris disc is visible as the elevated red areas on each side of the star:



And profile plots taken both through the debris disc plane and perpendicular to it. The horizontal scale is Astronomical Units and the area saturated by Beta itself is highlighted on the plots:



The visible part of the debris disc seems to extend to roughly 250-300AU before it falls below the background noise levels.
I have found that the limiting magnitude with the ToUCam from my location is around 20.0. So how far out should the debris disc theoretically be visible in my image? This is a plot of the magnitude pr. square arcsecond for the debris disc (Smith, B. A. & Terrile, R. J. 1984, Science, 226, 1421 A Circumstellar Disk around Beta Pictoris):



It seems from the figure above that if a limiting magnitude of approximately 20.0 is assumed, then the debris disc should be visible out to somewhere around 250-300AU. This corresponds with what is seen in the profile plots above.

This annotated crop of the final image shows the extent of the debris disc on both sides of Beta itself:



Some notes on ToUCam and IR sensitivity
The disc is most prominent in IR and fortunately the ICX098BQ CCD chip is very good at picking up signals in IR, probably on par with or even better than some modern CCD's in this aspect.
Another property of the ICX098BQ that helps here is actually the Bayer colour matrix. This is because each pixel either receives only red, green or blue light due to the overlaid Bayer matrix. But since the matrix dyes are practically transparent to IR light, every pixel therefore receives signal from the IR band. So the IR S/N ratio in the final image can be assumed to be correspondingly higher than the RGB S/N. This could easily explain why the ToUCam seems to pick up IR light so well; the Bayer matrix effectively acts like a IR pass filter and allows a proportionally higher number of pixels to be IR illuminated than RGB illuminated.
While I have not tested this in practice it could mean that the limiting magnitude of the ToUCam is higher in IR than in RGB.


Rolf Wahl Olsen, 04/12/2011