CCDs vs. PMTs

How does the sensitivity of an average/high end photomultiplier tube
compare with a CCD? I was wondering if I could hook up a PMT to my
scope with some type of device that blocks out all light surrounding a
single point, then drift scan around that point. Is this possible?

Sincerely,
Michael S.

In article ,
Michael S. wrote:
How does the sensitivity of an average/high end photomultiplier tube
compare with a CCD? I was wondering if I could hook up a PMT to my
scope with some type of device that blocks out all light surrounding a
single point, then drift scan around that point. Is this possible?

A CCD is something like 70% quantum efficient at its peak wavelength
(IIRC photomultiplier tubes work better in the blue and ultra-violet
whilst CCDs are better in the red and IR), so the photomultiplier
can't win by more than a factor 1.5 or so.

Drift scan gets you a line, which sounds as if you want to rig up some
dispersive element at the right angle and scan down spectra; I'd have
thought the accuracy in tracking required to do 2D scanning would be
much, much more costly to set up than a CCD.

Infra-red astronomers used to do this sort of thing with bolometers,
and people working in the submillimetre region still do, but if you
read the papers you can almost hear the sigh of relief as focal plane
arrays became available.

Tom

How does the sensitivity of an average/high end photomultiplier tube
compare with a CCD? I was wondering if I could hook up a PMT to my
scope with some type of device that blocks out all light surrounding a
single point, then drift scan around that point. Is this possible?

Sincerely,
Michael S.

I think you would have some difficultly achieving the resolution needed. And
then just image a simple CCD (such as mine) with its 640x496 resolution. You
would have to scan each of those 250,000 points individually so while you were
scanning one, about 249,999 would be idle.

Given that astro CCDs are something like 30-70% efficiently, the PMT would be
loosing are great deal of light in comparison.

Thats my story and I'm stickin with it...

jon

On 8 Oct 2004 06:14:40 -0700, (Michael S.) wrote:

How does the sensitivity of an average/high end photomultiplier tube
compare with a CCD? I was wondering if I could hook up a PMT to my
scope with some type of device that blocks out all light surrounding a
single point, then drift scan around that point. Is this possible?

Unless you are interested in short wavelengths (UV), silicon detectors almost
always outperform PMTs. A typical PMT has about half the QE of a typical CCD.
The PMT is much more difficult to work with, given its strong sensitivity to
magnetic fields, high voltage, and the need to cool the cathode if you want to
maximize the S/N (it is much simpler to cool a CCD, which has a blind side,
unlike a PMT photocathode).

The most common reason for using a PMT over silicon is for measuring very short
wavelengths. This is not usually something done with ground-based astronomy,
however.

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com

How does the sensitivity of an average/high end photomultiplier tube
compare with a CCD

I think the first few responders missed an important point, i.e.
photomultiplier that operates about 70% efficiency. A photomultiplier has a
built in amplifier. In other words, a photon collides with a surface and an
electron is emitted. This electron then collides with another surface and
several electrons are now emitted. This process cascades along the tubes so
that 1 photon may yield 100 to 1000 electrons. The cascade is enhanced and
controlled with the application of a high voltage, i.e. of the order of 100 to
1000 volts.This process is extremely fast so that the net gain is many fold
over a single CCD cell. In a scanning device that you suggest, the exposure
times may well be shortened so that you can scan quite rapidly. I suppose you
plan to do the whole thing with mirrors (and no smoke). Hope this helps.

On 09 Oct 2004 01:41:55 GMT, (HAVRILIAK) wrote:

How does the sensitivity of an average/high end photomultiplier tube
compare with a CCD

I think the first few responders missed an important point, i.e.
photomultiplier that operates about 70% efficiency. A photomultiplier has a
built in amplifier. In other words, a photon collides with a surface and an
electron is emitted. This electron then collides with another surface and
several electrons are now emitted. This process cascades along the tubes so
that 1 photon may yield 100 to 1000 electrons. The cascade is enhanced and
controlled with the application of a high voltage, i.e. of the order of 100 to
1000 volts.This process is extremely fast so that the net gain is many fold
over a single CCD cell. In a scanning device that you suggest, the exposure
times may well be shortened so that you can scan quite rapidly. I suppose you
plan to do the whole thing with mirrors (and no smoke). Hope this helps.

I disagree. Both CCDs and PMTs are inherently photon counting devices. CCDs
generally have a higher QE (PMTs with a 70% QE are rare). Normally, both are
operated in a similar fashion, with the charge created by the detected photons
stored and then amplified. A PMT does make it possible to count individual
photons that arrive at a low rate, although more than half will normally be
missed, and the noise is high.

In the case of astronomical imaging, the fundamental limitation is the number of
photons available. That's why you need long exposure times. CCDs are very
efficient collectors- 70%-90% of electrons will be captured, and with a long
exposure the noise will be a very small part of the signal. If you try to do the
same thing with a PMT, you will detect fewer photons in the first place, so the
statistical noise will be higher. Furthermore, unless you cool the photocathode,
many spurious events will be generated. PMTs are fairly noisy at room
temperature. Building a cryostat for a PMT is not a simple (or inexpensive)
task.

I can't see any way that a scanning PMT could possibly approach a CCD in
sensitivity (as an imaging device). At the least, it would require a very
expensive PMT, complex cooling, and excellent electronics. And, of course, a
scanning system. For the cost of all that, you could be using a professional
grade thinned CCD with 90%+ QE.

_________________________________________________

Chris L Peterson
Cloudbait Observatory
http://www.cloudbait.com