From: Phil Hobbs
Subject: Re: measure the intensity of light
Date: Mon, 18 Nov 2002 11:40:33 -0500
Organization: IBM T. J. Watson Research Center
NNTP-Posting-Date: 18 Nov 2002 16:40:49 GMT
X-Mailer: Mozilla 4.61 [en] (OS/2; U)
> I have to design a system to measure the intensity of fluorescence
> light (emission is around 520-530 nm).
> The fluorescence is bright enough and I can see it visually.
> In short I have to measure the intensity of a yellow-green light.
> The measure must be reliable.
> Which kind of sensor is more appropriate?
> Can I found calibrated photodiode or photodetector for that task.
> What about the linearity of the response?
The fact that it's visible doesn't mean that it's bright enough to be
trivial to measure. Room lights, residual pump light, thermal drifts,
and so on, make low level photodetection more an optical problem than a
How are you pumping the sample?
Do you have an estimate of the quantum yield (i.e. fluorescent
efficiency) of the fluor? It can be 10%, 10**-7, or even less.
What is the ambient environment? A system in a black box is easier to
make than one that has to work in sunlight.
What accuracy do you need?
Typical detectors for fluorescence measurements are photomultiplier
tubes, avalanche photodiodes used in Geiger (photon-counting) mode, or,
if the signal is bright enough or the environment dark enough, a PIN
photodiode or even a CCD.
Photon counting is very linear up to the point where pulses start to
overlap. PIN photodiodes are the most linear things going, and CCDs are
a close second. Analog-mode PMTs and APDs are rather less linear, as
well as more vulnerable to accuracy drifts due to non-optical ambient
effects such as temperature changes and magnetic fields.
Chopping helps to reject the dc background from ambient light, but
doesn't help pump-dependent effects such as stray pump light or filter
(If you're using coloured glass filters, many of them fluoresce.
Putting the filters in the wrong order can degrade your signal to noise
ratio by 30 dB.)