home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!agate!physics3!aephraim
- From: aephraim@physics3 (Aephraim M. Steinberg)
- Newsgroups: sci.physics
- Subject: Re: photon 'detectors' - how reliable?
- Date: 21 Jan 1993 19:03:00 GMT
- Organization: /etc/organization
- Lines: 62
- Message-ID: <1jms14$ibn@agate.berkeley.edu>
- References: <wwadge.727584610@csr>
- NNTP-Posting-Host: physics3.berkeley.edu
-
- In article <wwadge.727584610@csr> wwadge@csr.UVic.CA (Bill Wadge) writes:
- >I was browsing through one of the many 'pop' physics books,
- >where they are describing one of the stock quantum mechanics
- >experiments, and noticed something along the lines of
- >
- > .. towards a detector which records every photon ..
- >
- >Is this possible? Can one really build a device so sensitive that it will
- >detect 100% reliably every photon that enters it, and never
- >go off by accident?
- >
- >Seems unlikely, but then I'm not a physicist ...
-
- Does seem unlikely, doesn't it? But it will happen one day (just about).
-
- Some photodiodes currently quote quantum efficiencies very close to
- 100%, but not when operating in single-photon detection mode. That is, they
- yield a continuous current proportional to the incoming intensity and
- supposedly nearly every incoming photon yields an electron-hole pair. (I
- don't know how-- or whether-- this last fact is confirmed.)
-
- The SINGLE-photon detectors of preference until recently were photomultiplier
- tubes, with efficiencies ranging from less than 1% to perhaps 5%. This is
- one of the reasons (along with the geometric nature of the decay process
- originally used) that the "detection loophole" was not closed by experiments
- like those of Clauser and of Aspect. In recent years, pmt's have been
- superceded by avalanche photodiodes, frequently with quantum efficiencies
- of 10 or 20%. We have one photodiode from EG&G which is about 40% efficient.
-
- VERY recently, however, there has been some progress towards nearly-100%
- efficient detectors. EG&G offers rather expensive "single-photon counting
- modules" which we have measured to have efficiencies of approximately 75%
- when they are sufficiently overbiased-- the active element in these modules
- is still an APD. On a different tack, Rockwell has been working on
- "solid state photomultipliers" which theoretically should have quantum
- efficiencies as high as 90 or 95%, although unlike the APD's, they need to
- be cooled to near-liquid Helium temperatures. We have measured efficiencies
- close to 70% for these devices as well, and it is quite likely that there
- were other losses present during that measurement, so the actual efficiency
- could well be higher. Needless to say, both companies (and others) are
- working towards further improvements.
-
- As far as I know, most people only care in a "linear" fashion about
- such efficiencies. For tests of Bell inequalities, on the other hand,
- there is a sharp cutoff of 83% (ignoring all other experimental imperfections)
- for eliminating the detection loophole. (Although some recent papers
- propose ways of reducing this cutoff to around 70%, it is nearly certain
- that there will always be some sharp cutoff, and that these schemes will
- be less feasible than those which require 83%.)
-
- Anyway, don't trust everything you read. The same physics book that
- mentioned 100% efficient photodetectors could well have mentioned frictionless
- pulleys and massless ropes: With many of the former and one of the latter,
- one could construct a 100% efficient photodetector simply by amplifying
- the "light pressure" acting on a ball attached to one end of the rope!
-
-
- --
- Aephraim M. Steinberg | "WHY must I treat the measuring
- UCB Physics | device classically?? What will
- aephraim@physics.berkeley.edu | happen to me if I don't??"
- | -- Eugene Wigner
-