[Ifeffit] Is it legitimate to merge transmission data withfluorescence data?

[Matthew]

A trick I've used when the fluorescence data are good but affected by overabsorption, and the transmission data are noisy is to 
reduce both data sets to post-edge normalized form (0 in pre-edge, oscillating about 1 post-edge), then apply a simple 
overabsorption model to 'correct' the fluorescence so it matches the transmission.  This is a bit risky in that some of the same 
samples which exhibit overabsorption will also show hole (thickness) effect, which looks enough like overabsorption so that you 
under-correct the fluorescence.

Otherwise, if the goal is to reduce noise, perhaps the best way is to reduce both to post-edge normalized form, then do a weighted 
sum.  What weights?  Look at the noise level on each and weight it accordingly.  One way to do that is to fit a polynomial and look 
at the magnitude of the residuals and ask "how many counts/point would an ideal Poissonian counter need to get the residual this 
low?".  The "2-column Editor" program available from the ALS Beamline 10.3.2 website 
http://xraysweb.lbl.gov/uxas/Beamline/Software/Software.htm -> Utilities has that as a capability.  If you want to use it, download 
the program and reply to me off-line for an explanation of how to use the noise analysis; the on-line doc is outdated and doesn't 
include that feature.
    mam
----- Original Message ----- 
From: "Bruce Ravel" <bravel at bnl.gov>
To: "XAFS Analysis using Ifeffit" <ifeffit at millenia.cars.aps.anl.gov>
Sent: Monday, August 10, 2009 3:20 PM
Subject: Re: [Ifeffit] Is it legitimate to merge transmission data withfluorescence data?


> On Monday 10 August 2009 06:13:32 pm jrkizews at ncsu.edu wrote:
>> Dear XAFS community members,
>> I have a question concerning data merge. We normally collect XAFS data in
>> both transmission mode and fluorescence mode. We normally have to merge a
>> few scans to enhance signal-to-noise ratio. For one particular sample, I
>> want to know if it is legitimate to merge its transmission scans with its
>> fluorescence scans to improve data quality?
>
> Hi Fiona,
>
> There is no a priori reason not to do so and certainly not a numerical
> reason.  From a numerical perspective, you can merge anything with
> anything!
>
> My one concern is that the fluorescence and transmission data are
> really equivalent.  That is, if the fluo data are affected by
> significant self-absorption attenuation or if the transmission data
> suffer from significant pin-hole effects, then you run the risk of
> degrading the entire data set by doing the merge.  But if the data are
> merely noisy, then I think you are safe doing so.
>
> Think about it this way: when you use a multi-element detector, you
> are making several measurements which are presumed to be identical.
> We routinely merge the channels of an MED.  You situation is, from a
> measurement theory perspective, analogous.  Assuming your data do not
> suffer from the problems mentioned above, it seems all right to me.
>
> B
>
>
>
> -- 
>
> Bruce Ravel  ------------------------------------ bravel at bnl.gov
>
> National Institute of Standards and Technology
> Synchrotron Methods Group at NSLS --- Beamlines U7A, X24A, X23A2
> Building 535A
> Upton NY, 11973
>
> My homepage:    http://xafs.org/BruceRavel
> EXAFS software: http://cars9.uchicago.edu/~ravel/software/exafs/
>
> _______________________________________________
> Ifeffit mailing list
> Ifeffit at millenia.cars.aps.anl.gov
> http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
>