Re: [Ifeffit] Fitting EXAFS with out any atomic coordinates?
Hi Mario, There are a couple of different ways to approach the kind of problem you describe, although you must be willing to guess at what the local environment looks like. For example, maybe the chemical composition tells you that you have an iron oxide. At a minimum, you could use Atoms to make a crystal structure for, say, hematite, and then only fit the nearest-neighbor paths, perhaps with an unknown coordination number. This can give you some information. Or, following the same example, maybe you think your material might be like a disordered hematite. Then you could try to fit the hematite structure out for several shells, but with free parameters to account for disorder (e.g. you could include vacancies). The key questions that you should not lose track of: --What would you like to know about your material? --What do you think your material might be like? I think sometimes people forget in particular about the first question. EXAFS doesn't answer generic questions very well (tell me all about my sample), but it often answers particular questions quite effectively (are my iron atoms bonded to oxygen, or sulfur?). So if you know what your particular questions are you've got a good chance of getting help using Bruce's and Matt's software. --Scott Calvin Sarah Lawrence College At 02:15 PM 3/28/2007, you wrote: Content-class: urn:content-classes:message Content-Type: multipart/alternative; boundary=_=_NextPart_001_01C77165.0DE07268 Hi Bruce my name is Mario and I was just wondering if your program can fir the EXAFS spectra without any atomic positions. Obviously the answer is no if you use atoms to create the FEFF file but what does one do when we have a semi-crystalline material for which no one has attempted the fractional coordinates. I mean the XRD pattern shows some crystalline but due to the high background I would think it would be very hard to do the Rietveld analysis. So what can one do, can we fit the EXAFS spectra knowing only the chemical composition and perhaps the lattice geometry(orthorhombic or hexagonal)? thanks ___ 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
Re: [Ifeffit] Fitting EXAFS with out any atomic coordinates?
On Wednesday 28 March 2007 13:15, Mario Gomez wrote: Hi Bruce my name is Mario and I was just wondering if your program can fir the EXAFS spectra without any atomic positions. Obviously the answer is no if you use atoms to create the FEFF file but what does one do when we have a semi-crystalline material for which no one has attempted the fractional coordinates. I mean the XRD pattern shows some crystalline but due to the high background I would think it would be very hard to do the Rietveld analysis. So what can one do, can we fit the EXAFS spectra knowing only the chemical composition and perhaps the lattice geometry(orthorhombic or hexagonal)? Mario, Scott's answer was excellent, but I like the sound of my own voice so I thought I'd answer as well ;-) The Feff+Ifeffit+Artemis zeitgeist requires a list of atomic coordinates to begin analysis. Except in the rare situations, we don't know the actual arrangement of atoms. Indeed, to find out is one of the reasons we might do EXAFS. As Scott suggested, thinking about what you sample resembles is usually a good place to start. Consider amorphous germanium as an example. aGe is typically described as a continuous random network of bonds which are about the same length as the bonds in crystalline Ge. Because of that, crystalline Ge is a good place to start. In fact, I suspect that you could make considerable progress in analyzing aGe data with that one Feff calculation on the well-ordered material. EXAFS analysis is something of a creative endeavor -- the trick is to figure out how to use things that you know to model things that you don't know. Reading papers by the folks who answer questions on this list (Scott Calvin, Shelly Kelly, Anatoly Frenkel, Paul Fons) is a really good source of ideas for your own work with IfeffitArtemis. B -- Bruce Ravel -- [EMAIL PROTECTED] Molecular Environmental Science Group, Building 203, Room E-165 MRCAT, Sector 10, Advanced Photon Source, Building 433, Room B007 Argonne National Laboratory phone and voice mail: (1) 630 252 5033 Argonne IL 60439, USAfax: (1) 630 252 9793 My homepage:http://cars9.uchicago.edu/~ravel EXAFS software: http://cars9.uchicago.edu/~ravel/software/ ___ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit
Re: [Ifeffit] Fitting EXAFS with out any atomic coordinates?
Since I like the sound of my voice too, I will translate the zeitgeist from Bruce's email. According to Wikipedia, It is originally a German expression that means the spirit (Geist) of the time (Zeit). Can it be loosely translate here as poltergeist? Anatoly -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Bruce Ravel Sent: Wednesday, March 28, 2007 3:41 PM To: XAFS Analysis using Ifeffit Subject: Re: [Ifeffit] Fitting EXAFS with out any atomic coordinates? On Wednesday 28 March 2007 13:15, Mario Gomez wrote: Hi Bruce my name is Mario and I was just wondering if your program can fir the EXAFS spectra without any atomic positions. Obviously the answer is no if you use atoms to create the FEFF file but what does one do when we have a semi-crystalline material for which no one has attempted the fractional coordinates. I mean the XRD pattern shows some crystalline but due to the high background I would think it would be very hard to do the Rietveld analysis. So what can one do, can we fit the EXAFS spectra knowing only the chemical composition and perhaps the lattice geometry(orthorhombic or hexagonal)? Mario, Scott's answer was excellent, but I like the sound of my own voice so I thought I'd answer as well ;-) The Feff+Ifeffit+Artemis zeitgeist requires a list of atomic coordinates to begin analysis. Except in the rare situations, we don't know the actual arrangement of atoms. Indeed, to find out is one of the reasons we might do EXAFS. As Scott suggested, thinking about what you sample resembles is usually a good place to start. Consider amorphous germanium as an example. aGe is typically described as a continuous random network of bonds which are about the same length as the bonds in crystalline Ge. Because of that, crystalline Ge is a good place to start. In fact, I suspect that you could make considerable progress in analyzing aGe data with that one Feff calculation on the well-ordered material. EXAFS analysis is something of a creative endeavor -- the trick is to figure out how to use things that you know to model things that you don't know. Reading papers by the folks who answer questions on this list (Scott Calvin, Shelly Kelly, Anatoly Frenkel, Paul Fons) is a really good source of ideas for your own work with IfeffitArtemis. B -- Bruce Ravel -- [EMAIL PROTECTED] Molecular Environmental Science Group, Building 203, Room E-165 MRCAT, Sector 10, Advanced Photon Source, Building 433, Room B007 Argonne National Laboratory phone and voice mail: (1) 630 252 5033 Argonne IL 60439, USAfax: (1) 630 252 9793 My homepage:http://cars9.uchicago.edu/~ravel EXAFS software: http://cars9.uchicago.edu/~ravel/software/ ___ 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
