Re: Techniques for Absorbance Calculation Without Reference Spectra
- From: Marvin <physchem@xxxxxxxxxxx>
- Date: Fri, 28 Mar 2008 15:58:02 GMT
Dave wrote:
Alright, I had very helpful answers last time, so hopefully someone
may be able to guide me on this. We do open-path UV spectroscopy using
custom equipment (and Ocean Optics specs). Our systems are deployed in
situations that may not guarantee the ability to take fresh references
on a good interval, and with a drift in signal strength over time
(ambient light, fog, etc.), we want to explore ways we can still
collect good data without taking a reference spectra. We have a known
set of compounds we're looking for, and we're using library files for
each that we use least squares regression to calculate concentration
and r^2 values.
Our goal is that we can set up an instrument, and have it run for days
on end without taking another reference spectra. We have heard of
other businesses that can do this, but have been unable to find what
methods they are using. We have currently implemented a synthesized
background by simply drawing a straight line over where the compound
should be. This eliminates all baseline shifting, but we lose our
minimum detection capabilities because it ignores the curvature of
what the reference spectra should be. (i.e. even a small curve that
becomes a straight line is a great difference in absorbance).
Is anyone familiar with any techniques we may use to accomplish this?
One method we have considered is taking a reference spectra in the
lab, and extrapolating real-time reference spectras from that by
comparing a particular range where nothing is (or should be)
absorbing. However, all we need is for a particular compound to wash
over that area and it ruins everything (also we are assuming the
spectral shape to be constant over a period of days or weeks, which
may be problematic).
In case you're unfamiliar with what I'm talking about, here is where
the idea came from:
http://clu-in.org/programs/21m2/openpath/uv-doas/
You will notice that they mention not having to take reference
spectra.
Again, I'm a software engineer and not an analytical chemist, so
forgive any term misuse here, and let me know if I should clarify.
Thanks in advance for any help here!
I read the article quickly, but the key to answering your question is near the top:
" * Besides the absorption of the trace gases also light extinction due to scattering on molecules and aerosols occurs.
" * In the atmosphere the absorptions of several species always add up to the total absorption. Thus in most of the cases it is not directly possible to measure only one specific species.
" * Usually the initial intensity I0 can not be measured at all or with sufficient accuracy."
They understood what caused the problem of calibration drift, and found a way to correct for it with information in the spectrum. Do you know the cause of the calibration drift in oyur application?
They also used an old principle of analytical chemistry. If you measure all the components, the sum of their percentage concentrations is 100%. That is useful information. You can ignore the concentrations of unmeasured trace components, because the error of doing that is small.
.
- References:
- Prev by Date: Techniques for Absorbance Calculation Without Reference Spectra
- Previous by thread: Techniques for Absorbance Calculation Without Reference Spectra
- Index(es):
Relevant Pages
|
