Page 238 - Photodetection and Measurement - Maximizing Performance in Optical Systems

Measurand Modulation

Measurand Modulation 231

reflectance equipment can be very large. Hence we should certainly try to mod-
ulate our measurand fast enough.
It would be a great help to understand the noise-and-drift spectrum down to
1/year frequencies to optimize recalibration regimes, and up to kilohertz fre-
quencies to optimize measurand modulation. FIA systems, in-situ photolysis,
and the spinning disk of Fig. 10.14 can all be considered as absorption modu-
lators for liquid samples. We can change the modulation frequency in FIA
through the segment volume and the flow-rate, and for the disk via spot-size
and rotational velocity w. To make best use of the modulation, we need to choose
the modulation frequency appropriately. The considerations involved in the
choice are the same as those considered in temporal modulation in a lock-in
amplifier measurement: what is the frequency distribution of noise sources, and
can we choose the modulation frequency to avoid the worst regions? For
example, independent of the 1/f noise character, in the case of photolytic absorp-
tion modulation of a drinking water sample we might find that the background
matrix changes significantly over a period of an hour due to quality changes.
Hence we should certainly modulate with a frequency faster than 1/hour. If we
can make one flash/measure measurement cycle per minute that will probably
be fast enough. With the development of low cost MEMS the use of microflu-
idic devices, which can rapidly switch nanoliters (nL) volumes of sample, may
open up new applications of high-speed FIA-type measurand modulation.
In the case of moving disk or tape reflectance measurements, we have a lot
of control over modulation frequency through the rotational velocity. Just as
with source modulation and synchronous detection it is probably a good idea
to modulate at a frequency well above the level of industrial noise sources,
say above 500Hz. With a 50mm diameter disk and ten sample spots per cir-
cumference this is possible with a rotational velocity above 3000rpm (linear
velocity >8m/s). However, this may not be optimum in terms of spatial fre-
quencies. As the disk rotates, even without any sample present, the detected
intensity will not be perfectly constant. Due to bearing wobble and imperfect
centration there is likely to be a detectable intensity variation at the funda-
mental frequency w. If the filter paper has been calendered or otherwise formed
between rollers it is likely to have uniaxial surface structure and different
angular scattering characteristics along and perpendicular to the calender direc-
tion, giving a frequency component at 2w. At very high spatial frequencies we
will begin to see the individual paper fibers. Hence we have to address both
temporal and spatial frequencies. A single 50mL sample absorbed to cover one
half of the filter disk is likely to suffer from the low spatial frequency reflectance
variations. If it is divided into thousands of 100mm spots fired on around
the circumference from an ink-jet actuator fiber noise may provide the limit to
LOD. We may find that optimum performance will be obtained by having 20
1-mm sized spots distributed around the circumferential track, 1/mm being
well above the dominant manufacturing nonuniformity frequencies, but below
paper-fiber noise frequencies. This choice is of course independent of the rota-
tional velocity.

Downloaded from Digital Engineering Library @ McGraw-Hill (www.digitalengineeringlibrary.com)
Copyright © 2004 The McGraw-Hill Companies. All rights reserved.
Any use is subject to the Terms of Use as given at the website.