This Message Is From an External Sender

This message came from outside your organization.

Hi Jan,

Yes, it can be confusing to select the E0 value for spectra like this. The complication is that we basically use E0 for two purposes: for XANES and normalization, we use it as the energy where absorption starts. In this case, the first derivative at the Cu1+ position is perfectly reasonable. But E0 is also used as the k=0 value for the EXAFS. Here, having E0 on the first sharp edge peak can make the spline harder to fit.

The spline will be easier if you move E0 for the EXAFS up will probably work a little better. The place you chose is fine, though "maximum of a derivative" is not really anything more than "convenient" at this point.

In fact (and to follow up on Shelly's suggestion), I would suggest trying both places. One interpretation of that Cu1+ peak is that it really is the Fermi level, and that dip is just really large negative EXAFS. But trying a first-shell fit with a couple of E0 choices is probably a good idea.

But I would probably not bother to treat that first Cu1+ peak as a "pre-edge peak" that should be isolated from the "real" EXAFS.

But also: Feff tends to pick k=0 values that are a bit higher in energy than where the maximum of the first derivative is found. So the EXAFS fit either has a positive E0 value, or you can move the E0 for k=0 up. I don't recall off-hand, but I can believe that for Cu1+ it really does need to be as high in energy as you picked.

FWIW, this is why Larix allows separate values "E0k" for EXAFS k=0 that can be different from the one used for normalization.

From: Ifeffit <ifeffit-bounces@millenia.cars.aps.anl.gov> on behalf of Jan-Benedikt Weiss <jan-benedikt.weiss@tu-dortmund.de>
Sent: Monday, May 13, 2024 9:08 PM
To: ifeffit@millenia.cars.aps.anl.gov <ifeffit@millenia.cars.aps.anl.gov>
Subject: [Ifeffit] Choice of E0 in the presence of a huge pre-peak

This Message Is From an External Sender

This message came from outside your organization.

Dear all,

since I am new to XAFS analysis, I do not have a proper understanding of 
how to chose E0 the right way if taking the first maximum of the 
derivative of ľ(E) is not an option. In my dataset (K-edge of a Cu(I) 
coordination polymer with linear coordination environment), a huge 
pre-peak far up the edge is present. Larch automatically detects E0 to 
be in front of that peak which leads to a complicated subtraction of the 
background in the EXAFS analysis (the spline does not align with the 
shape of the edge; see attached plot) and will probably lead to other 
problems. I thought of two alternative option to determine E0 but I do 
not know if they are adequate.

Option 1: Take the second maximum of the derivative of ľ(E) which 
appears after the pre-peak (see attached plot). I guess this value is 
still flawed by the pre-peak and therefore not accurate.

Option 2: Fitting a baseline curve under the pre-peak and select E0 as 
the maximum of the derivative of that curve. Since this value will 
probably lay underneath the pre-peak, it will not be suitable as the E0 
value used in the background subtraction in the EXAFS analysis.

So do you have any suggestion on how to handle this case?

Best wishes
Jan