Paul and Shelly, I do not think Einstein temperature for Zn-S bond should be recycled as the Einstein temperature for a Zn-O bond. To my opinion, the best approximation for Zn-O Einstein temperature would be to look up for EXAFS data on Zn2+ complexes in aqueous solution, assume that what they publish for sigma2 is dominated by thermal disorder, and calculate Einstein temperature for Zn-O from the published value of sigma2. Then, extrapolate the sigma2 to low temperature by an Einstein model. That would probably give a lower bound of the total disorder in Zn-O that you expect for this bond in your complex, since some static disorder may be present as well. Anatoly ________________________________ From: ifeffit-bounces@millenia.cars.aps.anl.gov on behalf of Kelly, Shelly D. Sent: Fri 7/27/2007 8:55 PM To: XAFS Analysis using Ifeffit Subject: Re: [Ifeffit] Debye Waller factors for Water Molecules Hi Paul, I was hoping someone else would answer your question. I have an idea. It might not be a good idea. Look up the MSRD term that you get from Grant's DFT parameterizations for the first Zn-S bond. Play around with the Einstein model for a Zn-S bond of the correct distance for the first Zn-S bond, to determine the correct Einstein temperature that gives the MSRD value that you looked up from Grant's DFT. Then use that Einstein temperature and the Einstein model for the Zn-O bond of the water molecule. --It might work.... Cheers, Shelly ________________________________ From: ifeffit-bounces@millenia.cars.aps.anl.gov [mailto:ifeffit-bounces@millenia.cars.aps.anl.gov] On Behalf Of Paul Fons Sent: Thursday, July 26, 2007 9:29 PM To: XAFS Analysis using Ifeffit Subject: [Ifeffit] Debye Waller factors for Water Molecules I am analyzing some Zn organo-metallic protein complexes and have take EXAFS data at low temperature at the Zn edge. I am now in the midst of trying to fit the data. For the mean squared relative displacement terms, I am using Grant's DFT parameterizations for the Zn-cysteine system. This is working fine. In one of my complexes, I also have a water molecule within the "first shell". Does anyone have a suggestion for a reasonable MSRD term for this (at 20 K)? Paul

Thanks everyone (Shelly, Matt, and Anatoly) for the info. I thought it would be fun to learn while trying to work out a (relatively) simple system. I have EXAFS data about a Zn site. The Zn is bonded to three cystedines so I have the DW pretty much under control, but for the water molecules. I will investigate the approaches mentioned here and see which one the data is most consistent with (I did have the foresight to choose a system for which I have high resolution x-ray crystallography data). Thanks Paul On Jul 29, 2007, at 12:57 AM, Matt Newville wrote:

Hi Paul,

A. Vogel et al (Biochemistry 43, p10379, 2004) report a Zn water (or OH) at 2.0Ang with sigma2=0.007Ang^2 for data collected at 30K.

This is for a binuclear Zn compound, and the OH is a bridging ligand (with the other ligands being dominated by histidine-like rings), so I don't know how close it is to your structure. But it might be a good place to start.

Also, the results came from EXCURVE. That's not to suggest that the results are not reliable, but you might want to check carefully the meaning of sigma2. I'm not 100% sure of this (or what the Vogel et al paper is reporting), but I think the EXCURVE sigma2 might be a factor of 2 bigger than the Ifeffit sigma2.

I'd be a little wary of using an Einstein model for "water" ligands down to 30K. It would be interesting to hear how well an Einstein model compares to the DFT modeling for any ligands in a bio-molecule.

--Matt _______________________________________________ Ifeffit mailing list Ifeffit@millenia.cars.aps.anl.gov http://millenia.cars.aps.anl.gov/mailman/listinfo/ifeffit

Paul, In addition to my previous email, before you decide whether to set and fix the Zn-O DWF or not, I suggest that you look for a spread in reported values of DWF in Zn2+ -O in aqueous solutions - most of them were done at room temperature, but to get an idea about the spread it only matters that the temperature is the same between different publications, so RT would be fine. Irit Sagi and James Penner Hahn published many papers on Zn2+ complexes, so that could give you a good starting point. If you find that the spread is significant, it is either because the amount of static disorder differs between these cases, or the Einstein model doesn't work, or both... - but it is important to know before you decide whether to keep it fixed or not in your fit. As far as keeping things fixed. Some things, like E0 or coord. numbers may be fixed if you analyze a series of compounds where you do not expect them to change between the samples. Sometimes, DWF of a certain bond is not expected to vary between different samples especially if they are all mixtures of different complexes of the same element, with the mixing fraction varied between the samples, common in time-resolved experiments with reaction intermediates. In many cases, DWF of the same bond varies from sample to sample. It is really not possible to say what's best to do in your case without knowing much detail. However, if you don't have a room for varying this parameter independently, as I assume you don't since you have a mixed first shell, it may be a good idea to fix it to something reasonable (e.g., as discussed in previous emails) and see if you get a reasonable fit. Anatoly ________________________________ From: ifeffit-bounces@millenia.cars.aps.anl.gov on behalf of Paul Fons Sent: Mon 7/30/2007 6:31 AM To: XAFS Analysis using Ifeffit Subject: Re: [Ifeffit] Debye Waller factors for Water Molecules Hi Matt, I took a look at the paper you mentioned and as in my case, the water is appears to be pretty much only bonded to the Zn at 1.96 Angstroms for the isolated enzyme. I guess this implies that Anatoly's idea of using a Zn+2 solution as a starting point might make sense. I have no idea how much the DW varies for water in these types of systems. Is the usual idea to just set it at some "reasonable" value (e.g. from the idea of Anatoly) and then keep it fixed? Paul On Jul 30, 2007, at 3:51 PM, Paul Fons wrote: Thanks everyone (Shelly, Matt, and Anatoly) for the info. I thought it would be fun to learn while trying to work out a (relatively) simple system. I have EXAFS data about a Zn site. The Zn is bonded to three cystedines so I have the DW pretty much under control, but for the water molecules. I will investigate the approaches mentioned here and see which one the data is most consistent with (I did have the foresight to choose a system for which I have high resolution x-ray crystallography data). Thanks Paul On Jul 29, 2007, at 12:57 AM, Matt Newville wrote: Hi Paul, A. Vogel et al (Biochemistry 43, p10379, 2004) report a Zn water (or OH) at 2.0Ang with sigma2=0.007Ang^2 for data collected at 30K. This is for a binuclear Zn compound, and the OH is a bridging ligand (with the other ligands being dominated by histidine-like rings), so I don't know how close it is to your structure. But it might be a good place to start. Also, the results came from EXCURVE. That's not to suggest that the results are not reliable, but you might want to check carefully the meaning of sigma2. I'm not 100% sure of this (or what the Vogel et al paper is reporting), but I think the EXCURVE sigma2 might be a factor of 2 bigger than the Ifeffit sigma2. I'd be a little wary of using an Einstein model for "water" ligands down to 30K. It would be interesting to hear how well an Einstein model compares to the DFT modeling for any ligands in a bio-molecule. --Matt _______________________________________________ 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