Hi Joseph, I have a book chapter that explains how the different k-weighting affect the chi(k) data produced in the background removal. I'll send it to you directly, so as not to fill up everyone's mail box. Here is the reference for others that might be interested. Kelly, S.D., Hesterberg, D., and Ravel, B., in "Analysis of Soils and Minerals Using X-ray Absorption Spectroscopy", (A.L. Ulery and L.R. Drees, Eds.) p. 367. Soil Science Society of America, Madison, WI, USA, 2008. Shelly -----Original Message----- From: ifeffit-bounces@millenia.cars.aps.anl.gov [mailto:ifeffit-bounces@millenia.cars.aps.anl.gov] On Behalf Of Joseph Washington Sent: Tuesday, July 08, 2008 6:03 AM To: ifeffit@millenia.cars.aps.anl.gov Subject: Re: [Ifeffit] Athena: k-weight=3 in background removal problem Hi Carlo, Since much of our data is located at high k values, it would be an improvement to be using a high k weight in the background removal. Whether or not this will be a vast improvement might be questionable. My main concern is why this occurs for all our Ge edge data across the board, and if this is indicative of an underlying issue in data collection at the beamline or preprocessing. Please see an attached project file containing some scans and a merge. I appreciate the assistance, Joseph Carlo Segre wrote:
Hi Joseph:
On Mon, 7 Jul 2008, Joseph Washington wrote:
Greetings all, I am an APS user and a graduate of XAFS Summer School - class of 2007
We have been doing EXAFS analysis on Ge[2]Sb[2]Te[x] thin films. I noticed that at the Ge edge in particular, the k-weight value used in the background removal cannot equal 3, or else the background function no longer follows the data and oscillates wildly and/or it shoots up at the edge energy dominating the plot window. Values of 1 and 2 work fine. Even more puzzling is that this problem is inherent for the Ge edge of data taken over numerous runs, and for various samples (GST-224, GST-225, 226, 227).
COuld you tell us why you would like to use a k-weight of 3 in the background removal process? I find that 1 (or 2 sometimes) is sufficient to get a good result. I just checked on a recent data set and I agree that k-weight 1 nd 2 give nearly identical results but
:) that
k-weight 3 for the same parameters is slightly different.
It would help a lot if you oculd send a project file for us to look at.
Carlo
-- Joseph Washington Research Assistant Department of Physics North Carolina State University 431 Riddick Hall Raleigh, NC 27695-8202 email: jswashin@ncsu.edu