>Weve been following the discussion about Bio-Rads enhanced pmts with some
>interest, as we have been investigating this subject in connection with
>low-level light detection from photoproteins, and we feel that a positive
>email from us might be helpful to anyone else who is interested in this
>topic (with the possible exception of Bio-Rad [:-) but were not out to
>knock them either).  The general idea has been around for a long time, and
>anyone who is interested in it could well do worse than to consult the
>paper Optical Devices to Increase Photocathode Quantum Efficiency  by W.D.
>Gunter, G.R. Grant and S.A. Shaw, in Applied Optics Vol 9, pp251-257,
>published way back in the mists of 1970.
>
>This paper describes the technique of increasing sensitivity by multiple
>reflections within the photocathode, and suggests a variety of external
>optical attachments, based on prisms or prism-mirror combinations, that
>allows it to be applied to any standard end-window photomultiplier.  The
>simplest implementations have the photocathode at a 45 degree angle to the
>incoming light, but other possible arrangements allowing the photocathode
>to be at its conventional orientation of 90 degrees are also described
>there.  Those are more complicated, but are likely to be more convenient in
>practice.
>
>We dont know what Bio-Rad are doing, but it looks as if anyone else could
>achieve equivalent results to theirs by implementing one of those
>arrangements or an optically equivalent one, and putting it onto a
>conventional photomultiplier.  The data in the paper shows that the
>improvement is greater at longer wavelengths, and also appears greater for
>an S20 than for a bialkali photocathode.  With a bialkali tube, and using
>what they called a two-bounce device, they got a 25% improvement at 500nm,
>and 50% at 600nm, increasing to 100% at 800nm (but youre talking about an
>actual quantum efficiency of twice almost nothing at this wavelength with
>bialkali!).  The improvement with an S20 tube was greater, being about 100%
>(i.e. twofold) at 500nm, 140% at 600nm and 200% at 700nm.  The quantum
>efficiency of bialkali exceeds that of S20 at wavelengths shorter than
>about 550nm (data from Electron Tubes Ltd, formerly Thorn EMI), so the
>greater enhancement with S20 would tend to suggest under these conditions
>that bialkali would not overtake S20 until a somewhat shorter wavelength
>(about 500nm?) was reached.
>
>In case this sounds like a free lunch (although it can be sometimes), the
>downside is that both the permissible area and the angle of the incoming
>light beam are considerably reduced by these arrangements - unlikely to be
>a problem for confocal, but it may be for other applications.  We expect to
>be playing around with the method ourselves during the next few weeks, and
>well be happy to share all results and technical details on request.
>
>                Martin
>
>Dr. Martin Thomas
>Cairn Research Ltd
>Unit 3G, Brents Shipyard Industrial Estate,
>Faverahsm, Kent, ME13 7DZ
>Tel: 01795 590140  Fax: 01795 590150
>E-mail: [log in to unmask]
>http://ourworld.compuserve.com/homepages/cairnresearch

I encourage you read this reference and do the experiment. Just a couple of
caveats.  The PMT housings in the Biorad don't have a lot of extra room for
the prism/mirrors you need to bring an axial ray around to the side and
across the window (the new tubes that they recommend are shorter) and it is
important that the PMT have a flat window.

Most MRC-600s were shipped with PMTs having prismatic inner photocathode
surfaces.  This design produced 2 internal reflections and 4 interactions
with the photocathode, for any axial ray.  The problem is that the PE
emitted on the inside have a much (2x) better chance of being collected and
multiplied if they emerge from a "peak" than if they emerge from a "valley"
in this prismatic surface. So you can have large local (2mm) variations in
QE.  (i.e. the effective QE all depends how well your pinhole lines up with
a "peak" in the PC).

Obviously a prismatic surface will not make a good light-pipe in the Gunter
sense. However, even photocathodes that appear smooth may in fact be
spherically curved on the inner surface and again these make poor light
pipes because the angle of incidence can not be easily kept below the
critical angle.



                   ****************************************
Prof. James B. Pawley,                                       Ph.  608-263-3147
Room 1235, Engineering Research Building,                    FAX  608-265-5315
1500 Engineering Dr., Madison, WI, 53706
[log in to unmask]