APOD: Globular Cluster M15 from Hubble (2013 Nov 19)

[Anthony Barreiro]

I have often imagined that some cleverly-written software could selectively reject areas of obvious cosmic ray hits if the data were somehow constantly streaming instead of delivered all at once. This is probably a manifestation of my lack of real understanding about how CCD's work, though.

I can certainly imagine something like that, although not with a CCD. But you could have some hypothetical new type of detector that output a signal each time a photon hit it, containing the time and coordinates of that event. With such a rich dataset, all sorts of interesting processing would be possible. With a CCD (and other current spatial detectors), however, most of the time information is lost. At best, you can determine that a certain number of photons hit a specific pixel between two known times, where the difference in those times is typically large compared to the time between photon strikes.

Such a detector is no longer purely hypothetical.

http://www.skyandtelescope.com/communit ... 70371.html

[Chris Peterson]

Intuitively (because I've got nothing else) it sounds like a job for lots of short exposures, with the new zero readout-noise sensors you mentioned a few weeks back.

Yes, but those detectors are still scanned in order to read them out. Since you'd need a readout time on the order of a few billion per second, you'd then have a pixel scan rate of trillions per second. I don't see that happening. So instead, I imaging a detector that simply provides data for every photon. The output rate would still be very high, of course.

[Chris Peterson]

Such a detector is no longer purely hypothetical.

http://www.skyandtelescope.com/communit ... 70371.html

Indeed, detectors that work this way have been experimented with for a while now. They would also allow for interferometry using a pair of telescopes (which can currently only be done by tying those scopes together optically).

Still, I think it will be a while before we see optical detectors with a high spatial resolution and a large dynamic range that truly give information about each photon (time, position, energy). But I'm sure it will happen. It's "just" an engineering problem, not one of any fundamental physics. It's also something that would be immensely valuable for ordinary consumer cameras, and that's always good for driving development.

[Nitpicker]

Such a detector is no longer purely hypothetical.

http://www.skyandtelescope.com/communit ... 70371.html

Indeed, detectors that work this way have been experimented with for a while now. They would also allow for interferometry using a pair of telescopes (which can currently only be done by tying those scopes together optically).

Still, I think it will be a while before we see optical detectors with a high spatial resolution and a large dynamic range that truly give information about each photon (time, position, energy). But I'm sure it will happen. It's "just" an engineering problem, not one of any fundamental physics. It's also something that would be immensely valuable for ordinary consumer cameras, and that's always good for driving development.

I'm glad to read this. Until now, I was under the impression that the money currently going into CMOS technology was going to end up winning the day across the board. Now I'm not so sure. Thanks Anthony and Chris.

[MarkBour]

Thanks for continuing to answer. There is just so much one can learn from you folks !
Here's another blown-up bit from that image. I took from pixel (1500,2100) for about 140w x 130h pixels.
Now that's what I call a constellation! I'd call it the serpent if I lived in the vicinity.

[geckzilla]

I have often imagined that some cleverly-written software could selectively reject areas of obvious cosmic ray hits if the data were somehow constantly streaming instead of delivered all at once. This is probably a manifestation of my lack of real understanding about how CCD's work, though.

I can certainly imagine something like that, although not with a CCD. But you could have some hypothetical new type of detector that output a signal each time a photon hit it, containing the time and coordinates of that event. With such a rich dataset, all sorts of interesting processing would be possible. With a CCD (and other current spatial detectors), however, most of the time information is lost. At best, you can determine that a certain number of photons hit a specific pixel between two known times, where the difference in those times is typically large compared to the time between photon strikes.

I just read something really cool about JWST's detectors. They are not CCDs and they can partially correct for cosmic rays. So far I don't know what the detectors are called. They are just calling them detectors or the focal plane assembly, which isn't very specific.
http://www.stsci.edu/jwst/doc-archive/handbooks/

[Chris Peterson]

I just read something really cool about JWST's detectors. They are not CCDs and they can partially correct for cosmic rays. So far I don't know what the detectors are called. They are just calling them detectors or the focal plane assembly, which isn't very specific.

The primary imager uses a mosaic of fairly conventional HgCdTe arrays. They operate in two different wavelength bands. By constructing the entire array as a mosaic, they can do some interesting things- fast readout of a small area, for instance, can allow the camera to act as a wavefront sensor, correcting the position of the mirror segments.

The sensors function similarly to CCDs, in that no timing information is present, only the knowledge that a photon arrived between the start and end of the exposure (typically a long interval). But it is possible to place the same target on different parts of the array, to allow simultaneous imaging at the long and short wavelength bands. There is no processing on the sensor, so I doubt there is any ability to compensate for cosmic rays at the detector level. However, when you you have simultaneous collection of data on two sensor arrays, cosmic rays are readily identified by simple post processing, either as the data is read or afterward. But I don't think there's any way to restore the data lost to a cosmic ray hit.

[geckzilla]

https://en.wikipedia.org/wiki/Mercury_cadmium_telluride

HgCdTe or mercury cadmium telluride (also cadmium mercury telluride, MCT, MerCad Telluride, MerCadTel, MerCaT or CMT) is narrow direct bandgap zincblende II-VI ternary alloy of CdTe and HgTe with a tunable bandgap spanning the shortwave infrared to the very long wave infrared regions. The amount of cadmium (Cd) in the alloy (the alloy composition) can be chosen so as to tune the optical absorption of the material to the desired infrared wavelength. CdTe is a semiconductor with a bandgap of approximately 1.5 eV at room temperature. HgTe is a semimetal, hence its bandgap energy is zero. Mixing these two substances allows one to obtain any bandgap between 0 and 1.5 eV.

I think I would be hard pressed to find another article in Wikipedia with more jargon. "This here takes infrared pictures."