Re: PMTs.

From: Howard Shapiro (hms@shapirolab.com)
Date: Sat Jan 31 1998 - 20:22:59 EST


>Which photo-multiplier tubes (PMT) are the best to use when building a flow
>cytometer?
>Which ones should be avoided?
>Any and all recommendations of manufacturers' models and part numbers are
>of interest.
>
See the 3rd edition of Practical Flow Cytometry (by the way, Mark, the parts
of the 2d Ed I promised you should arrive in a day or two).
Most people now use 1 1/8" diameter side-on tubes.
For forward or 90 degree scatter below 670 nm, a 931B will do; these can be
bought from Burle (formerly RCA) and Hamamatsu and are usually under $100.
They also work for fluorescence at 530 nm and below, but, even at 530 and
definitely at all longer wavelengths, fluorescence is best measured using a
tube with enhanced red sensitivity; Hamamatsu is essentially the only
supplier.  Hamamatsu's R928 is only slightly less expensive than their
R1477, which is a selected R928 with substantially better specs; both are
around The Hamamatsu R3896 offers about 20% improvement over the R1477 and
costs over $1,000; you be the judge (although the R3896 is better than that
above 800 nm if you're interested). Many people use tubes with GaAs
photocathodes at really long (e.g. around 800 nm, where Cy7 emits)
wavelengths, but the R666 and R636 GaAs tubes from Hamamatsu have maximum
gains of only about 10,000, while an R928, R1477, or R3896 has a gain of
over a million, and, at this wavelength, quantum efficiencies of the
multialkali cathodes in the latter tubes are almost as high as that of a
GaAs cathode.

Hamamatsu now makes tubes sockets incorporating Cockcroft-Walton HV supplies
for the PMT's; I recommend these, as they run on much lower current than
would be needed for a dynode chain, and decrease nonlinearity problems.  You
get the PMT voltage from a 12-15 V supply, drawing tens of milliamps at
most, and can control PMT voltage either with 0-1.25 VDC or with a
potentiometer.

Hamamatsu also makes modules (HC120-08) incorporating an R4457 1/2" side-on
tube and a programmable HV supply; these are comparable in sensitivity to a
928 (slightly better, actually) up to about 700 nm, have low noise, and are
very compact, but they cost about $900.  I've had one die on me, but
Hamamatsu should be replacing it shortly. 


>Also, are there any opinions about the use of avalanche photo-diodes in
>place of PMTs? When and where might this be appropriate?
>

Avalanche photodiodes (APD's) have higher cathode quantum efficiency than
PMT's, but lower gain (100-1,000).  They require HV bias and should be run
at constant temperature (and preferably chilled) to stabilize gain and
reduce noise.  I worked for a while with Advanced Photonix, which makes some
of the best APD-amplifier modules, and couldn't get anything near the
sensitivity of a PMT.  E G & G inherited RCA's APD business via RCA; I
haven;t had much luck with their linear APD's either.  However, E G & G
(Vaudreuil, Quebec, Canada) does make a photon counting APD module which
operates in Geiger or breakdown (nonlinear) mode; the Los Alamos group uses
this in their slow flow DNA molecule sizing system.  The price range is
$8,000 down to $2,000; Hamamatsu is said to be working on similar devices at
similar prices.  Unless you're sizing DNA fragments, APD's aren't ready for
prime time, although Oddbjorn Gjelsnes used one (a cheap one, relatively) as
the fluorescence detector in his design of the MicroCyte for Aber
Instruments (Aberystwyth, Wales, UK).

-Howard



This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:51:44 EST