- Ifeffit - millenia.cars.aps.anl.gov

background subtractions
by Bruce Ravel 19 May '03

19 May '03

Hey Carlo, Here is my answer to your other two questions, also CCed to the mailing list. > 2. What is the justification for fitting the background during the fitting > process. I understand that it is important to remove background large k > oscillations but it seems to me that this is better done without the bias > of simultaneously fitting FEFF shells. I know that Steve Wasserman and > Jeff Terry have a Mathematica background subtraction algorithm which > allows you to back FT the region below Rbkg and subtract it from chi. > This seems to be roughly the same idea as the background fit but is less > biased by the fitting. I would appreciate your comments. Also, is this > now a part of FEFFIT or is it solely in Artemis? It is roughly the same idea, but I would say that you *want* the background to be biased by the fitting. The issue here needs to be restated in terms of what is meant by "a good background removal". Implicit in the back transform algorithm that you describe is the assumption that a good background removal makes the low-R region "look pretty". That is, one assumes that the low-R regions should tail away in some monotonic fashion. Basically, this is an esthetic argument rather than a statistical or scientific argument. On some level, subtracting away a portion of the spectrum via a back transform algorithm is not horribly different from modifying the plot of the chi(R) spectrum with Photoshop or The Gimp. First off, I should explain how the background co-refinement in Ifeffit works. You define a lower bound of the fitting range in R-space. This Rmin value takes the role of Rbkg in the autobk algorithm. That is, Rmin is the cutoff in R-space above which you will use Fourier components from feffNNNN.dat files to fit the data and below which you will use a spline to fit the data. The spline is determined in exactly the same way that the spline in autobk is determined. Is this cheating in terms of number of `independent' point? No. The splin has exactly the number of knots corresponding to the band whose width in R is from 0 to Rmin. That is, different information is used by the spline than by the rest of the fit. Why is another pass of the autobkl algorithm needed at this stage? Why couldn't autobk just do the job correctly up front? Well, this is sort of a numerical issue. Autobk has a really hard job to perform. If you Fourier transform the entire mu(E) spectrum, you will find that the Fourier contribution from the mu0(E) function is much much huger than the contribution from the chi(K) function. Thus autobk is attempting to isolate something small out of something large. By the time Artemis sees the data, mu0(E) is all gone. Further refining the background is thus a matter of separating something small from something small. The REAL REASON to do the background co-refinement is because Ifeffit (and thus Artemis) will report on the **correlations between** the background parameters and the things you are really interested in, such as delta_R and sigma^2. This is improtant. Even more it is a way of defining "goodness" of a background removal that removes esthetic concerns. A "well-removed" background is one that is not too highly correlated with the fitting parameters. Of course, "not to highly" is a rather subjective, but it atleast is a number to work with. So what is the right value of Rmin? Well, you want to choose an Rmin that is high enough to make a pretty picture (esthetics is not entirely without merit, after all) but low enough so as not to introduce excessive correlation with the fitting parameters. Rmin, therefore, is one of the parameters that you need to fiddle with in your data. For oxides (and carbides, nitrides, etc) Rmin might be quite low. Indeed, there may not be a value which gives a spline that is loosely correlated with the fitting parameters and which also satisfies the esthetic criteria of a pretty low-R region taht tails off monotonically. To avoid harming the data by excessive correlation with the spline, you may be forced to accept low-R peaks even in your published data. Finally, how does this discussion relate to the background removal done in Athena? Well, it should be clear by now that you cannot know the right value of Rmin a priori. You have to do some fits and measure the correlation between the background and the fitting parameters to judge what Rmin should be. Good practice with Athena is to choose a rather low value of Rbkg and not worry that some icky little low-R peaks might remain in the data. If you choose a low Rbkg in Athena, you can always choose a higher Rmin value in Artemis. But you cannot go the other way. You should not choose a value of Rmin that is *lower* than the Rbkg used in Athena because then you will not correctly evaluate the correlations. > 3. When it comes time to report results, is it better to be able to > subtract the background? Is it legitimate to report fitting parameters > without including the "hidden" background parameters? I say yes, it is better to subtract the background found by Artemis. (That's what the "use previous background" button does.) Esthetics, as I confessed above, is still important. And yes, reporting fitting parameters from a fit with a background co-refinement is most certainly legitamate. Indeed, it is more so because you can also report the level of correlation with the background spline. This gives you reader MORE confidence in your results and a better ability to interpret their validity. Sorry I was so long winded, but these are subtle issues. When I give a series of lectures on data analysis, I usually spend about 30 minutes discussing background issues. HTH B PS: Background co-refinement is, like everything else mathematical, something done by Ifeffit. Ifeffit does all the heavy lifting, Artemis just keeps it all organized. -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu Code 6134, Building 3, Room 222 Naval Research Laboratory phone: (1) 202 767 5947 Washington DC 20375, USA fax: (1) 202 767 1697 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b, X24c, U4b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

3 2

Re: [Ifeffit] Windows Installer for Ifeffit 1.2.0
by Mark Jensen 19 May '03

19 May '03

Bruce, Matt, and the curious: Thanks for clarifying your distribution philosophy. >I think we haven't figured this out yet. Is it OK to download >ifeffit-1.2.0.exe as the basic installer, and then install >whatever updates to athena/artemis may exist on top of that? >>I thought that people were, by and large, happy with the solution of >>dropping new athena & artemis exes on top of the old ones rather than >>downloading the 30MB installer every time I make an update. Since I I very much like the current system of being able to 'plug-in' new versions of Athena and Artemis exe's as they are made available. I just wasn't sure if that approach was being abandoned. >>In any case, either you got confused updating or we got confused >>building the installer. It *was* our intention that the new installer >>have Athena 0.8.019 and Artemis 0.6.004. As far as what happened in the installation goes I'm not sure who was confused. Maybe I was pushing the installer a little too hard: I definitely chose the correct program to install Ifeffit 1.2.0 because: 1) Its filename is 'ifeffit-1.2.0.exe' which makes the choice pretty obvious 2) Ifeffit 1.2.0 was installed 3) When I repeated the procedure on Monday morning, making absolutely certain that I used ifeffit-1.2.0.exe, I was still left with Athena 0.8.015 and Artemis 0.5.008. (Those must have been the most recent versions of Athena and Artemis that I was using on that computer... I thought I was more up to date.) As I have done for Ifeffit installations in the past, I did no uninstallation or file deletion before trying the new install and I installed into the same directory as the existing version. When this Ravelware trouble caused me to look more closely at the directory contents, the only files that had May 2003 dates were the *.bat files and the new Ifeffit files. The dates on Athena and Artemis were still March 2003! However, during the install, I saw the installer say that it was unpacking Artemis and Athena. Apparently it did not install over the old versions (but it did install Ifeffit 1.2.0?!). The Solution: So, I uninstalled Ifeffit and deleted the remaining contents of the Ifeffit directory (Keeping my fingers crossed that the new install would work). The ensuing activation of ifeffit-1.2.0.exe now gives me the correct versions of Ifeffit, Artemis, and Athena! Now let's do some fitting. Thanks for your thoughts, Mark ------------------------------ Mark P. Jensen Chemistry Division Argonne National Laboratory 9700 South Cass Avenue Argonne, Illinois 60439-4831 USA 630-252-3670 630-252-7501 FAX mjensen(a)anl.gov

1 0

Windows Installer for Ifeffit 1.2.0
by Mark Jensen 17 May '03

17 May '03

Bruce and Matt and everybody else, I am probably missing something, but the Artemis and Athena splash screens from the new Ifeffit 1.2.0 windows installer that I just used say that it installed Athena 0.8.015 and Artemis 0.5.008 and Ifeffit 1.2.0. Initially, I thought that the windows installer would (as in the past) include the latest (and greatest) versions of Athena and Artemis. But now that I think about it more, Matt seemed to indicate that he wants a stable distribution, which might mean older versions of Athena and Artemis. Bruce's website, on the other hand, seems to say that the windows installer for Ifeffit 1.2.0 would have the most recent windows versions of Artemis and Athena. I politely ask, which of these ideas is winning and when can windows users hope for some new versions of the Ravel-ware. Mark Jensen ------------------------------ Mark P. Jensen Chemistry Division Argonne National Laboratory 9700 South Cass Avenue Argonne, Illinois 60439-4831 USA 630-252-3670 630-252-7501 FAX mjensen(a)anl.gov

3 2

Re: Ifeffit Digest, Vol 3, Issue 8
by k-kupiecki＠northwestern.edu 16 May '03

16 May '03

John and Bruce, Thank you for your kind help. I'm still a little perplexed... Basically I'm taking SEXAFS on a Ni(110) single crystal substrate (with a fraction of a monolayer of Au which substitutes into the first layer... SEXAFS spectra were obtained at the AuL3 edge). I've taken data in four different posistions: "grazing incidence" with the E vector of the x-rays parallel to the surface normal 1) with x-rays along the (001) direction 2) with x-rays along the (-110) direction and "normal incidence" with the E vector of the x-rays perpendicular to the surface normal (E parallel to the surface plane) 3) with x-rays along the (001) direction 4) with x-rays along the (-110) direction After receiving both Bruce and John's emails I tried three different feff.inp files, using my original hemisphere cluster, each using one of the following: POLARIZATION 1 1 0 or POLARIZATION 0 0 1 or POLARIZATION -1 1 0 but upon ploting out chi(k) vs. k from the chi.dat's for each of these they looked identical... did I miss something? Thanks, Kristine Witkowski Northwestern University > > Hi Kristine, > I basically agree with what Bruce Ravel wrote. We've occasionally done > FEFF calculations on surfaces (e.g. with K. Baberschke and students), > and the key to the calculation is to use appropriate Polarization > cards (and Ellipticity cards) in the calculation as described in the > doc. Please let me know if you have problems doing this. > Cheers, > John >

1 0

1.2.0 windows installer released
by Matt Newville 16 May '03

16 May '03

Hi Everyone, The Windows installer for ifeffit 1.2.0 is now available at the Ifeffit homepage. This is now officially the "stable" version, but I want to stress that there may still be bugs (in fact, there is a serious bug still in Feff6L!!!), and that "stable" means that changes to the core library will attempt to follow the principle of "first, do no harm". Also, the sourceforge.net site for Ifeffit now has the latest versions of the stable source and windows executables at http://sourceforge.net/projects/ifeffit/ --Matt

1 0

new versions of Athena and Artemis using Ifeffit 1.2.0
by Bruce Ravel 16 May '03

16 May '03

Hi folks, Today I posted a new source code tarball for Athena and Artemis to my web site. I did *not* post new Windows executables, however. Linux, Unix, and OSX users should first upgrade Ifeffit by grabbing the Ifeffit 1.2.0 tarball from Matt's web site and then grabbing the tarball from my web site or run the horae_update script. Windows users should wait until Matt announces the new Windows installer (which is almost ready to be released) and download that. Future Windows executables of Athena and Artemis will assume that you have already updated the base Ifeffit package. New stuff this time: 1. The horae_update script is a bit smarter in that it compares the version on the server to the version on your computer and does nothing if you seem to be up to date. 2. New features of Ifeffit 1.2.0 are used, such as rebinning in Athena and fits in Artemis with multiple-data-sets, multiple-k-weights, and/or background co-refinements. 3. Better reporting of ill-posed variables in Artemis (based on Ifeffit's reporting of the same). 4. Better reporting of best-fit path lengths after a fit finishes. 5. Check for mismatched parentheses in math expressions in Artemis (a HUGE problem in prior versions). 6. Examples of multiple-data-set fits in the form of artemis project files. Look for the files examples/artemis/EuTiO3/eto.apj and examples/artemis/Cu/temparatures.apj in the source code tarball or C:\Program Files\Ifeffit\examples\EuTiO3\eto.apj and C:\Program Files\Ifeffit\examples\Cu\temparatures.apj on Windows. Be sure to read the comments in the project journals of those two files. + several bug fixes Regards, B -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu Code 6134, Building 3, Room 222 Naval Research Laboratory phone: (1) 202 767 5947 Washington DC 20375, USA fax: (1) 202 767 1697 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b, X24c, U4b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

1 0

Artemis questions
by Bruce Ravel 16 May '03

16 May '03

Hi Carlo, Hope you don't mind my responding to the list. These are questions that I think will be of general interest. I am actually going to split my answers into two emails, both of which will still be rather long -) > I have been working with Artemis over the past few days and I have a > couple of questions that I cannot seem to find answers to in the manuals > for FEFFIT and Artemis. OK! You have me up against the wall ;-) Documentation is currently the Artemis' great shortcoming. I apologize that the doc is coming along slowly. FWIW, I have been thinking alot about the doc, but haven't found a lot of time to write. > 1. What is the difference between a SET and a DEF quantity. I suspect > that it is an issue of being updated or not but I am not sure. A SET parameter is evaluated at the start of a fit and remains invarient throughout the fit. A DEF parameter is re-evaluated at every iteration of the fit. Typically a def parameter depends on one or more guess parameters. Let's consider a simple example. Suppose that we have a crystal with a nominal lattice constant "a0" and we want to consider the possibility in the fit that the actuall lattice constant deviates from "a0". To do so, we may consider a parameter "delta_a": set a0 3.61 guess delta_a 0.0 Now suppose we want to use "a" (the measured lattice constant) in math expressions for some of our path parameters. Do we say set a a0 + delta_a or def a a0 + delta_a ? The answer is the second one. If make "a" a SET parameter, then a will be fixed to 3.61 at the beginning of the fit and not change. If, however, we make "a" a DEF parameter, then it will be updated every time that "delta_a" is updated during the fit. At the risk of confusing the issue, I feel I should make a connection with Matt's old feffit program. In the old feffit, set parameters behaved like Ifeffit's def parameters. That is, in the old feffit, set parameters were updated throughout the fit. As a rule of thumb for Artemis, if your parameter is a number, make it a set parameter. If it is a math expression, make it a def parameter. HTH, B -- Bruce Ravel ----------------------------------- ravel(a)phys.washington.edu Code 6134, Building 3, Room 222 Naval Research Laboratory phone: (1) 202 767 5947 Washington DC 20375, USA fax: (1) 202 767 1697 NRL Synchrotron Radiation Consortium (NRL-SRC) Beamlines X11a, X11b, X23b, X24c, U4b National Synchrotron Light Source Brookhaven National Laboratory, Upton, NY 11973 My homepage: http://feff.phys.washington.edu/~ravel EXAFS software: http://feff.phys.washington.edu/~ravel/software/exafs/

1 0

Anyone out there do SEXAFS?
by k-kupiecki＠northwestern.edu 15 May '03

15 May '03

HI, I'm trying to analyze my SEXAFS data... if anyone out there has done this before could you please drop me a line at k-kupiecki(a)northwestern.edu.... I could really use someone that I can directly question and bounce procedures off of.... Inparticular I'm currently looking for suggestions for how to simulate with FEFF SEXAFS in the "grazing incidence" orientation (with the E vector of the synchrotron beam close (~0.2 deg) to the surface normal of the sample (incident beam ~.2 deg incident on surface). Thus far I've been trying to analyze my samples with SEXAFS taken from a "normal geometry" (~0.2 angle incident to the beam and the E vector of the synchrotron beam in the plane of the surface)... I managed to simulate this with FEFF by giving it a hemispherical cluster representing my surface... But I'm a little confused what I need to change in FEFF to replicate the information coming from just the inter planar bonding (in "grazing incidence") instead of both inter and intra planar bonding as found in the "normal incidence" situation.... Any suggestions would be greatly appreciated! Thanks, Kristine (Kupiecki) Witkowski Northwestern University

3 2

FEFF6 errors?
by Carlo U. Segre 15 May '03

15 May '03

Matt: I have been having problems with the FEFF6 that is distributed with Ifeffit. If I try to run the executable with the feff.inp the TEST directory, it gives the following output ---- copper Calculating potentials and phases... free atom potential and density for atom type 0 free atom potential and density for atom type 1 overlapped potential and density for unique potential 0 overlapped potential and density for unique potential 1 muffin tin radii and interstitial parameters phase shifts for unique potential 0 phase shifts for unique potential 1 Preparing plane wave scattering amplitudes... nncrit in prcrit 9 Searching for paths... Rmax 0.0000 keep and heap limits 0.0000000 0.0000000 Preparing neighbor table nfound nheap nheapx nsc r Internal path finder limit exceeded -- path list may be incomplete. Paths found 0 (nheapx, nbx 1 0) Eliminating path degeneracies... Plane wave chi amplitude filter 2.34% STOP no input statement executed ---- It only generates the 4 files: geom.dat paths.bin paths.dat phase.bin If I look at paths.dat, I see: ---- copper Rmax 0.0000, keep limit 0.000, heap limit 0.000 ^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@^@ Plane wave chi amplitude filter 2.34% ------------------------------------------------------------------------------- ---- I am a bit confused about this because when building the Debian package, everything compiles properly and all the libraries seem to be there properly. Any ideas? Carlo -- Carlo U. Segre -- Professor of Physics Associate Dean for Research, Armour College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 Carlo.Segre(a)iit.edu http://www.iit.edu/~segre

2 1

Athena - plotting limits
by Carlo U. Segre 15 May '03

15 May '03

Bruce: Why does Athena only plot from 0-15 in k-space even though there is data beyond that. Now that I look back at the mu(E) plots, they are restricted in energy too. I understand that we would normally not use data to far beyond k=15 but it still might be nice to see it. Perhaps, I am just not seeing some kind of parameter that could be set? Carlo -- Carlo U. Segre -- Professor of Physics Associate Dean for Research, Armour College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 Carlo.Segre(a)iit.edu http://www.iit.edu/~segre

2 1