Re: [Ifeffit] Hexagonal WO3 (Carlo Segre)
Hi Carlo and Ying,
Does it means that if the sample is a random solid solution at experimental
temperature, in principle, you would be able to simulate the EXAFS spectra
whichever be the Mo/W ratio in the second shell coordination in the
Feff.inp, am I right?
Thanks, Yordy
2008/11/18
Send Ifeffit mailing list submissions to ifeffit@millenia.cars.aps.anl.gov
To subscribe or unsubscribe via the World Wide Web, visit http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit or, via email, send a message with subject or body 'help' to ifeffit-request@millenia.cars.aps.anl.gov
You can reach the person managing the list at ifeffit-owner@millenia.cars.aps.anl.gov
When replying, please edit your Subject line so it is more specific than "Re: Contents of Ifeffit digest..."
Today's Topics:
1. Re: [PGPLOT build problem] Upgrading to ifeffit-1.2.11b on Ubuntu from source (Matt Newville) 2. Ifeffit 1.2.11c (was Re: [PGPLOT build problem] Upgrading to ifeffit-1.2.11b on Ubuntu from source) (Matt Newville) 3. Re: Ifeffit 1.2.11c (was Re: [PGPLOT build problem] Upgrading to ifeffit-1.2.11b on Ubuntu from source) (Carlo Segre) 4. Re: Ifeffit 1.2.11c (was Re: [PGPLOT build problem] Upgrading to ifeffit-1.2.11b on Ubuntu from source) (Matt Newville) 5. Hexagonal WO3 (Ying Zhou, Anorganisch-Chemisches Inst.) 6. Re: Hexagonal WO3 (Carlo Segre)
----------------------------------------------------------------------
Message: 1 Date: Mon, 17 Nov 2008 12:30:20 -0600 From: "Matt Newville"
Subject: Re: [Ifeffit] [PGPLOT build problem] Upgrading to ifeffit-1.2.11b on Ubuntu from source To: "XAFS Analysis using Ifeffit" Message-ID: Content-Type: text/plain; charset=ISO-8859-1 Hi,
This is my fault... and I apologize for the sloppiness and for being late in fixing this. I've been mixing 'development/testing' and 'production' versions, and haven't been very good at keeping them cleanly separated. I think I have this mostly under control now, but I will have to work on a decent "test suite".
I think I found another reason to make a 1.2.11c version too, that the mixing of "devel" and "production" caused an experiment in how fitting was done to be the default, so that the reported chi-square doesn't actually match the documentation (instead of summing real and imaginary parts, it sums real and magnitude components of the misfit, which is part of an experiment to see how Feff8.5 worked for XAFS. Using Magnitude as well Phase for another way to report chi-square and misfits is potentially a useful approach, but it's not what is documented. Anyway, mixing devel and production versions like this (for the past couple years, really) is very sloppy on my part, and I apologize again.
I actually have an ubuntu box now, though I built pgplot from source, and have never made a .deb file. There is a new source kit at http://cars.uchicago.edu/~ifeffit/src/ifeffit-1.2.11c.tar.gz I'll post this to sourceforge this afternoon, and I'll also make a new DLL for Windows and post all of these on the download page. I'll also make a new updater for Windows.
--Matt
On Mon, Nov 17, 2008 at 10:07 AM, Mauro Rovezzi
wrote: Carlo Segre wrote:
Hi Mauro:
I have now produced a package of ifeffit 1.2.11b. It is in Debian unstable and available for Ubuntu intrepid and jaunty in my repository
deb http://debian-xray.iit.edu intrepid main contrib non-free
Carlo,
Thank you very much for the package.
Actually, I was able to build/install ifeffit 1.2.11b from source but not able to build than the Perl/Python wrappers and so no Horae&Friends from Bruce SVN.
Well, now I have "pgplot5", "ifeffit", "perl-ifeffit" and "python-ifeffit" from binaries and "libperlxray" "horae" and "demeter" built from source (Bruce SVN).
Everything seems fine but I still do not understand why ifeffit 1.2.11 has all these "IFF str = *" output as stated also by Gareth in his "Ifeffit 1.2.11b terminal output" post. Is it normal?
Cheers, Mauro _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
------------------------------
Message: 2 Date: Mon, 17 Nov 2008 17:06:35 -0600 From: "Matt Newville"
Subject: [Ifeffit] Ifeffit 1.2.11c (was Re: [PGPLOT build problem] Upgrading to ifeffit-1.2.11b on Ubuntu from source) To: "XAFS Analysis using Ifeffit" Message-ID: Content-Type: text/plain; charset=ISO-8859-1 Hi Folks,
I've uploaded source kits for Ifeffit 1.2.11c to the cars.uchicago.edu and sourceforge.net download sites.
I have a Windows dll made for this bug fix version. If anyone is interested, this is currently at http://cars.uchicago.edu/~ifeffit/updates/011/bin/ifeffit_12.dll
This should be a drop-in replacement for the IFEFFIT_DIR\bin\ifeffit_12.dll on your system. Artemis.exe will not be updated from this dll, as it has this dlls "compiled in". Bruce will have to make a new Artemis.exe, and then we'll make a automatic update.
Sorry for all the trouble,
--Matt
On Mon, Nov 17, 2008 at 12:30 PM, Matt Newville
wrote: Hi,
This is my fault... and I apologize for the sloppiness and for being late in fixing this. I've been mixing 'development/testing' and 'production' versions, and haven't been very good at keeping them cleanly separated. I think I have this mostly under control now, but I will have to work on a decent "test suite".
I think I found another reason to make a 1.2.11c version too, that the mixing of "devel" and "production" caused an experiment in how fitting was done to be the default, so that the reported chi-square doesn't actually match the documentation (instead of summing real and imaginary parts, it sums real and magnitude components of the misfit, which is part of an experiment to see how Feff8.5 worked for XAFS. Using Magnitude as well Phase for another way to report chi-square and misfits is potentially a useful approach, but it's not what is documented. Anyway, mixing devel and production versions like this (for the past couple years, really) is very sloppy on my part, and I apologize again.
I actually have an ubuntu box now, though I built pgplot from source, and have never made a .deb file. There is a new source kit at http://cars.uchicago.edu/~ifeffit/src/ifeffit-1.2.11c.tar.gz I'll post this to sourceforge this afternoon, and I'll also make a new DLL for Windows and post all of these on the download page. I'll also make a new updater for Windows.
--Matt
On Mon, Nov 17, 2008 at 10:07 AM, Mauro Rovezzi
wrote: Carlo Segre wrote:
Hi Mauro:
I have now produced a package of ifeffit 1.2.11b. It is in Debian unstable and available for Ubuntu intrepid and jaunty in my repository
deb http://debian-xray.iit.edu intrepid main contrib non-free
Carlo,
Thank you very much for the package.
Actually, I was able to build/install ifeffit 1.2.11b from source but not able to build than the Perl/Python wrappers and so no Horae&Friends from Bruce SVN.
Well, now I have "pgplot5", "ifeffit", "perl-ifeffit" and "python-ifeffit" from binaries and "libperlxray" "horae" and "demeter" built from source (Bruce SVN).
Everything seems fine but I still do not understand why ifeffit 1.2.11 has all these "IFF str = *" output as stated also by Gareth in his "Ifeffit 1.2.11b terminal output" post. Is it normal?
Cheers, Mauro _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
------------------------------
Message: 3 Date: Mon, 17 Nov 2008 20:03:06 -0600 (CST) From: Carlo Segre
Subject: Re: [Ifeffit] Ifeffit 1.2.11c (was Re: [PGPLOT build problem] Upgrading to ifeffit-1.2.11b on Ubuntu from source) To: Matt Newville Cc: XAFS Analysis using Ifeffit Message-ID: Content-Type: text/plain; charset=us-ascii; format=flowed Debian packages have been uploaded to unstable and I am in the process of building Ubuntu packages for my repository. i expect that they will be up tonight.
carlo
On Mon, 17 Nov 2008, Matt Newville wrote:
Hi Folks,
I've uploaded source kits for Ifeffit 1.2.11c to the cars.uchicago.edu and sourceforge.net download sites.
I have a Windows dll made for this bug fix version. If anyone is interested, this is currently at http://cars.uchicago.edu/~ifeffit/updates/011/bin/ifeffit_12.dll
This should be a drop-in replacement for the IFEFFIT_DIR\bin\ifeffit_12.dll on your system. Artemis.exe will not be updated from this dll, as it has this dlls "compiled in". Bruce will have to make a new Artemis.exe, and then we'll make a automatic update.
Sorry for all the trouble,
--Matt
On Mon, Nov 17, 2008 at 12:30 PM, Matt Newville
wrote: Hi,
This is my fault... and I apologize for the sloppiness and for being late in fixing this. I've been mixing 'development/testing' and 'production' versions, and haven't been very good at keeping them cleanly separated. I think I have this mostly under control now, but I will have to work on a decent "test suite".
I think I found another reason to make a 1.2.11c version too, that the mixing of "devel" and "production" caused an experiment in how fitting was done to be the default, so that the reported chi-square doesn't actually match the documentation (instead of summing real and imaginary parts, it sums real and magnitude components of the misfit, which is part of an experiment to see how Feff8.5 worked for XAFS. Using Magnitude as well Phase for another way to report chi-square and misfits is potentially a useful approach, but it's not what is documented. Anyway, mixing devel and production versions like this (for the past couple years, really) is very sloppy on my part, and I apologize again.
I actually have an ubuntu box now, though I built pgplot from source, and have never made a .deb file. There is a new source kit at http://cars.uchicago.edu/~ifeffit/src/ifeffit-1.2.11c.tar.gz I'll post this to sourceforge this afternoon, and I'll also make a new DLL for Windows and post all of these on the download page. I'll also make a new updater for Windows.
--Matt
On Mon, Nov 17, 2008 at 10:07 AM, Mauro Rovezzi
wrote: Carlo Segre wrote:
Hi Mauro:
I have now produced a package of ifeffit 1.2.11b. It is in Debian unstable and available for Ubuntu intrepid and jaunty in my repository
deb http://debian-xray.iit.edu intrepid main contrib non-free
Carlo,
Thank you very much for the package.
Actually, I was able to build/install ifeffit 1.2.11b from source but not able to build than the Perl/Python wrappers and so no Horae&Friends from Bruce SVN.
Well, now I have "pgplot5", "ifeffit", "perl-ifeffit" and "python-ifeffit" from binaries and "libperlxray" "horae" and "demeter" built from source (Bruce SVN).
Everything seems fine but I still do not understand why ifeffit 1.2.11 has all these "IFF str = *" output as stated also by Gareth in his "Ifeffit 1.2.11b terminal output" post. Is it normal?
Cheers, Mauro _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
-- Carlo U. Segre -- Professor of Physics Associate Dean for Special Projects, Graduate College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 segre@iit.edu http://www.iit.edu/~segre segre@debian.org
------------------------------
Message: 4 Date: Mon, 17 Nov 2008 21:10:51 -0600 From: "Matt Newville"
Subject: Re: [Ifeffit] Ifeffit 1.2.11c (was Re: [PGPLOT build problem] Upgrading to ifeffit-1.2.11b on Ubuntu from source) To: segre@agni.phys.iit.edu Cc: XAFS Analysis using Ifeffit Message-ID: Content-Type: text/plain; charset=ISO-8859-1 Thanks, Carlo!
--Matt
On Mon, Nov 17, 2008 at 8:03 PM, Carlo Segre
wrote: Debian packages have been uploaded to unstable and I am in the process of building Ubuntu packages for my repository. i expect that they will be
up
tonight.
carlo
On Mon, 17 Nov 2008, Matt Newville wrote:
Hi Folks,
I've uploaded source kits for Ifeffit 1.2.11c to the cars.uchicago.edu and sourceforge.net download sites.
I have a Windows dll made for this bug fix version. If anyone is interested, this is currently at http://cars.uchicago.edu/~ifeffit/updates/011/bin/ifeffit_12.dll
This should be a drop-in replacement for the IFEFFIT_DIR\bin\ifeffit_12.dll on your system. Artemis.exe will not be updated from this dll, as it has this dlls "compiled in". Bruce will have to make a new Artemis.exe, and then we'll make a automatic update.
Sorry for all the trouble,
--Matt
On Mon, Nov 17, 2008 at 12:30 PM, Matt Newville
wrote: Hi,
This is my fault... and I apologize for the sloppiness and for being late in fixing this. I've been mixing 'development/testing' and 'production' versions, and haven't been very good at keeping them cleanly separated. I think I have this mostly under control now, but I will have to work on a decent "test suite".
I think I found another reason to make a 1.2.11c version too, that the mixing of "devel" and "production" caused an experiment in how fitting was done to be the default, so that the reported chi-square doesn't actually match the documentation (instead of summing real and imaginary parts, it sums real and magnitude components of the misfit, which is part of an experiment to see how Feff8.5 worked for XAFS. Using Magnitude as well Phase for another way to report chi-square and misfits is potentially a useful approach, but it's not what is documented. Anyway, mixing devel and production versions like this (for the past couple years, really) is very sloppy on my part, and I apologize again.
I actually have an ubuntu box now, though I built pgplot from source, and have never made a .deb file. There is a new source kit at http://cars.uchicago.edu/~ifeffit/src/ifeffit-1.2.11c.tar.gz I'll post this to sourceforge this afternoon, and I'll also make a new DLL for Windows and post all of these on the download page. I'll also make a new updater for Windows.
--Matt
On Mon, Nov 17, 2008 at 10:07 AM, Mauro Rovezzi
wrote:
Carlo Segre wrote:
Hi Mauro:
I have now produced a package of ifeffit 1.2.11b. It is in Debian unstable and available for Ubuntu intrepid and jaunty in my
repository
deb http://debian-xray.iit.edu intrepid main contrib non-free
Carlo,
Thank you very much for the package.
Actually, I was able to build/install ifeffit 1.2.11b from source but not able to build than the Perl/Python wrappers and so no Horae&Friends from Bruce SVN.
Well, now I have "pgplot5", "ifeffit", "perl-ifeffit" and "python-ifeffit" from binaries and "libperlxray" "horae" and "demeter" built from source (Bruce SVN).
Everything seems fine but I still do not understand why ifeffit 1.2.11 has all these "IFF str = *" output as stated also by Gareth in his "Ifeffit 1.2.11b terminal output" post. Is it normal?
Cheers, Mauro _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
-- Carlo U. Segre -- Professor of Physics Associate Dean for Special Projects, Graduate College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 segre@iit.edu http://www.iit.edu/~segre segre@debian.org
------------------------------
Message: 5 Date: Tue, 18 Nov 2008 17:20:24 +0100 From: "Ying Zhou, Anorganisch-Chemisches Inst."
Subject: [Ifeffit] Hexagonal WO3 To: ifeffit@millenia.cars.aps.anl.gov Message-ID: Content-Type: text/plain;charset=iso-8859-1;format="flowed" Dear All
Recently, I got one compound, W/Mo oxide, which has the similar structure with hexagonal WO3. We tested the EXAFS of W L3 edge and Mo K edge of this compound. So, we use the hexagonal WO3 model to refine the structure of our compounds. My question is that how can I get the FEFF file of this compound. Can I directly change some of W to Mo in FEFF file? or Do I need to add the Mo in the atom file? I do appreciate your help.
Best wishes Ying
------------------------------
Message: 6 Date: Tue, 18 Nov 2008 10:33:35 -0600 (CST) From: Carlo Segre
Subject: Re: [Ifeffit] Hexagonal WO3 To: XAFS Analysis using Ifeffit Message-ID: Content-Type: TEXT/PLAIN; charset=US-ASCII; format=flowed Hi Ying:
On Tue, 18 Nov 2008, Ying Zhou, Anorganisch-Chemisches Inst. wrote:
Dear All
Recently, I got one compound, W/Mo oxide, which has the similar structure with hexagonal WO3. We tested the EXAFS of W L3 edge and Mo K edge of this compound. So, we use the hexagonal WO3 model to refine the structure of our compounds. My question is that how can I get the FEFF file of this compound. Can I directly change some of W to Mo in FEFF file? or Do I need to add the Mo in the atom file? I do appreciate your help.
There are several ways to do this, depending on what you are trying to find out. If you change W to Mo in the atoms input file than you are really making MoO3. What it seems like you need to do is to generate the FEFF input file from the WO3 structure, then replace the central atom with Mo. This will give you the first path which is likely oxygen. The second longest set of paths are likely to be W/Mo and you can do two things, one is to generate a FEFF calculation with W in this position and then make a second calculation with Mo and use both paths to fit the Mo environment. You will have to be careful about multiplicities though since by starting with the WO3 structure, you are assuming some degeneracies in the paths. The advantage is that you potentially have some multiple scatteing paths available to use.
Another approach is to generate the FEFF input file from WO3, look at the near neighbors and then generate some "quick first shell" paths for the nearest neighbor (oxygen) and then two fro the next near neighbor (one with W and one with Mo). The difference here is that you now have uncorrelated paths with no implied degeneracies. You cannot use this to fit multiple scattering paths.
I guess that my approach would be to start with the latter, more simple approach first to try to understand the problem. If I see that the environment looks like it is possible to fit with multiple scattering, I would move to the first approach, being careful to realize that in a solid sloution compound, the occupancy of the second shell of metal atoms is probably random and not perfectly ordered like a FEFF input file with substitutions would be.
Hope this helps.
Carlo
-- Carlo U. Segre -- Professor of Physics Associate Dean for Special Projects, Graduate College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 segre@iit.edu http://www.iit.edu/~segre segre@debian.org
------------------------------
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
End of Ifeffit Digest, Vol 69, Issue 13 ***************************************
-- Yordy E. Licea Fonseca Departamento de Físico-Química Instituto de Química-IQ/UFRJ Laboratório de Catálise Heterogênea Rio de Janeiro, RJ, CEP 21.949-909 Brasil Email: yliceafonseca@gmail.com
Hi Yordy: I have removed all the previous emails to keep the message short... On Sat, 22 Nov 2008, Yordy Licea Fonseca wrote:
Hi Carlo and Ying,
Does it means that if the sample is a random solid solution at experimental temperature, in principle, you would be able to simulate the EXAFS spectra whichever be the Mo/W ratio in the second shell coordination in the Feff.inp, am I right?
I am not quite sure what you are getting at but I think that the answer is yes. What you needto do is to have a single scattering path for each of the two possibilities in the second shell (Mo or W) along with the first shell path to O and then start fitting with minimal parameters. For example, if you know that the total number of atoms in the second shell is 12 for this structure (and you are reasonably sure that this is the correct structure), you can constrain the sum of the two occupation factors to be equal to 12 provided that you have a reasonable understanding of what the amplitude reduction factor might be. You can also start by constraining the delta R and sigma-squareds for the two paths until you are sure that the fit makes sense and is stable. If you have data from both W and Mo edges, you have more leeway in your fitting and you can begin to make a model which tries to figure out if there is clustering of one kind of atom in the structure. There are a number of papers in the literature that do this (off the top of my head I can think of one in particular by Shibata and Bunker for Au/Ag nanoparticles but there are surely many more). Ultimately you can develop a nice model but the one thing that you can't really do too much is include multiple scattering paths which include the atoms which are substituted for one another. Hope this answers your question, Carlo -- Carlo U. Segre -- Professor of Physics Associate Dean for Special Projects, Graduate College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 segre@iit.edu http://www.iit.edu/~segre segre@debian.org
Dear Carlo
Thanks very much for your nice suggestions. This is really what I want to
know.
Best wishes
Ying
On Mon, 24 Nov 2008 11:33:02 -0600 (CST)
Carlo Segre
Hi Yordy:
I have removed all the previous emails to keep the message short...
On Sat, 22 Nov 2008, Yordy Licea Fonseca wrote:
Hi Carlo and Ying,
Does it means that if the sample is a random solid solution at experimental temperature, in principle, you would be able to simulate the EXAFS spectra whichever be the Mo/W ratio in the second shell coordination in the Feff.inp, am I right?
I am not quite sure what you are getting at but I think that the answer is yes. What you needto do is to have a single scattering path for each of the two possibilities in the second shell (Mo or W) along with the first shell path to O and then start fitting with minimal parameters. For example, if you know that the total number of atoms in the second shell is 12 for this structure (and you are reasonably sure that this is the correct structure), you can constrain the sum of the two occupation factors to be equal to 12 provided that you have a reasonable understanding of what the amplitude reduction factor might be. You can also start by constraining the delta R and sigma-squareds for the two paths until you are sure that the fit makes sense and is stable.
If you have data from both W and Mo edges, you have more leeway in your fitting and you can begin to make a model which tries to figure out if there is clustering of one kind of atom in the structure. There are a number of papers in the literature that do this (off the top of my head I can think of one in particular by Shibata and Bunker for Au/Ag nanoparticles but there are surely many more). Ultimately you can develop a nice model but the one thing that you can't really do too much is include multiple scattering paths which include the atoms which are substituted for one another.
Hope this answers your question,
Carlo
-- Carlo U. Segre -- Professor of Physics Associate Dean for Special Projects, Graduate College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 segre@iit.edu http://www.iit.edu/~segre segre@debian.org _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Dear Carlo,
Thank you very much. The point is that we have obtained a Co Mo/W O4 solid
solution. It seems to me a very similar problem to that posted by Ying.
Good luck in yor work,
Yordy
2008/11/25
Send Ifeffit mailing list submissions to ifeffit@millenia.cars.aps.anl.gov
To subscribe or unsubscribe via the World Wide Web, visit http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit or, via email, send a message with subject or body 'help' to ifeffit-request@millenia.cars.aps.anl.gov
You can reach the person managing the list at ifeffit-owner@millenia.cars.aps.anl.gov
When replying, please edit your Subject line so it is more specific than "Re: Contents of Ifeffit digest..."
Today's Topics:
1. Re: Hexagonal WO3 (Carlo Segre) (Ying Zhou, Anorganisch-Chemisches Inst.)
----------------------------------------------------------------------
Message: 1 Date: Mon, 24 Nov 2008 19:47:48 +0100 From: "Ying Zhou, Anorganisch-Chemisches Inst."
Subject: Re: [Ifeffit] Hexagonal WO3 (Carlo Segre) To: segre@agni.phys.iit.edu, XAFS Analysis using Ifeffit Message-ID: Content-Type: text/plain;charset=iso-8859-1;format="flowed" Dear Carlo
Thanks very much for your nice suggestions. This is really what I want to know.
Best wishes Ying
On Mon, 24 Nov 2008 11:33:02 -0600 (CST) Carlo Segre
wrote: Hi Yordy:
I have removed all the previous emails to keep the message short...
On Sat, 22 Nov 2008, Yordy Licea Fonseca wrote:
Hi Carlo and Ying,
Does it means that if the sample is a random solid solution at experimental temperature, in principle, you would be able to simulate the EXAFS
spectra
whichever be the Mo/W ratio in the second shell coordination in the Feff.inp, am I right?
I am not quite sure what you are getting at but I think that the answer is yes. What you needto do is to have a single scattering path for each of the two possibilities in the second shell (Mo or W) along with the first shell path to O and then start fitting with minimal parameters. For example, if you know that the total number of atoms in the second shell is 12 for this structure (and you are reasonably sure that this is the correct structure), you can constrain the sum of the two occupation factors to be equal to 12 provided that you have a reasonable understanding of what the amplitude reduction factor might be. You can also start by constraining the delta R and sigma-squareds for the two paths until you are sure that the fit makes sense and is stable.
If you have data from both W and Mo edges, you have more leeway in your fitting and you can begin to make a model which tries to figure out if there is clustering of one kind of atom in the structure. There are a number of papers in the literature that do this (off the top of my head I can think of one in particular by Shibata and Bunker for Au/Ag nanoparticles but there are surely many more). Ultimately you can develop a nice model but the one thing that you can't really do too much is include multiple scattering paths which include the atoms which are substituted for one another.
Hope this answers your question,
Carlo
-- Carlo U. Segre -- Professor of Physics Associate Dean for Special Projects, Graduate College Illinois Institute of Technology Voice: 312.567.3498 Fax: 312.567.3494 segre@iit.edu http://www.iit.edu/~segre segre@debian.org _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
------------------------------
_______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
End of Ifeffit Digest, Vol 69, Issue 17 ***************************************
-- Yordy E. Licea Fonseca Departamento de Físico-Química Instituto de Química-IQ/UFRJ Laboratório de Catálise Heterogênea Rio de Janeiro, RJ, CEP 21.949-909 Brasil Email: yliceafonseca@gmail.com
participants (3)
-
Carlo Segre
-
Ying Zhou, Anorganisch-Chemisches Inst.
-
Yordy Licea Fonseca