|
Subject: New GSAS kits From: "Bob Von Dreele" [vondreele@lanl.gov] To: [Rietveld_l@ill.fr] Date: Mon, 16 Dec 2002 10:32:19 -0800 Hi everyone, New Linux & Windows versions of gsaskit have been placed on the LANL ftp server (ftp://ftp.lanl.gov/public/gsas). Both are identical. See notes below for latest updates (there have been a few important ones). All the best for the Holidays & have a Happy New Year. Bob Von Dreele GSAS News....... Dec. 15, 2002 Still more fixes & an addition... 1. A new constant wavelength X-ray and neutron profile function (#5) has been defined. This has the same form as #2 except that the "Lij" are now "Dij". These are coefficients for a macroscopic strain on the lattice parameters for the phase in a given histogram with respect to the lattice parameters given for that phase. The strain is given by del-2th = (D11*h^2+D22*k^2+D33*l^2+D12*h*k+D13*h*l+D23*k*l)*d^2*tan(th) This function has it's uses; e.g. for combining data sets taken at slightly different temperatures on the same sample. |
|
Subject: Re: New GSAS kits Date: Tue, 17 Dec 2002 12:15:09 +0100 To: rietveld_l@ill.fr From: Jonathan WRIGHT [wright@esrf.fr] Reply-To: rietveld_l@ill.fr Bob, This sounds great - just what we are after for using anisotropic thermal expansion data! I'm assuming the distance-angle restraints are calculated based on the unit cell parameters, so we should try to use a "averaged" values for cell par's in multipattern fits? One other question - are these parameters forced to retain the symmetry of the lattice? For example, a strained sample with a "cubic" lattice might have different values for D11, D22 and D33, so that the (100), (010) and (001) peaks no longer exactly overlap. Are you calculating the contributions of all of the various peaks using a peaklist generated P-1, regardless of the symmetry, or does the user need to be consistent with their choice of space group so that the program can use a multiplicity factor for symmetry equivalent reflections? This could be very useful for refinements of small distortions. Often one wants to use a space group which has a higher symmetry than the lattice, which can present a challenge for getting the list of atoms and constraints correct. Cheers, Jon |
|
Subject: Re: New GSAS kits From: "Bob Von Dreele" [vondreele@lanl.gov] To: rietveld_l@ill.fr Date: Tue, 17 Dec 2002 08:36:46 -0800 Hi Jon, Couple of things. This function is only effective if the lattice parameter changes are small (at most 0.1% or so) so the effect on things like bond length restraints are negligible. The assumption is that the effect is like a strain so that the atom positions are unaffected. There are no symmetry constraints; all the Dij are independent. My experience with it for lysozyme (tetragonal) is that D11 & D22 follow each other quite well with D33 different (& opposite sign with pH changes). I've varied lattice parameters & Dij for one out of two histograms - works fine. I haven't tried it with data showing small distortions (e.g. cubic -> rhombohedral in a perovskite) from symmetry but should work very well. Bob |
|
Subject: Re: New GSAS kits From: "Bob Von Dreele" [vondreele@lanl.gov] To: rietveld_l@ill.fr Date: Tue, 17 Dec 2002 09:08:17 -0800 Hi Jon (again), Reflecting on your comment & thinking about it a it more - the symmetry effect is tricky. A change in symmetry makes reflections not equivalent & makes them appear in different places in the pattern. Consequently, reducing the reflection set symmetry to P-1 and processing over all of them is the right way to account for the effect of changing symmetry with Dij refinement in function #5. Down side is the "expense" of doing all those profile function calculations. I'll have to think about this some more. However, function #5 will be OK for those cases where the symmetry doesn't change. Bob |