Dear Darin & others, The trick here is to realize the effect of the different site multiplicities. Moreover, remember that the constraint applies to the shifts & not the values. So start the site fractions with the right stereochemistry - you can put all Co in one site & all Fe in the other with the duplicate sites set to zero FRAC. Then put in constraint that forces amount of Fe lost from A-site to appear in the B-site and vise versa for the Co; a single constraint with 4 terms should suffice. The constraint factors will be 1.0 or 0.5 to account for the site multiplicity changes. All this assumes no vacancies - you can't really determine that from a single experiment, too many variables & not enough information. Bob Von Dreele
________________________________ From: Darin Hoffman [mailto:[EMAIL PROTECTED] Sent: Mon 8/9/2004 5:23 PM To: [EMAIL PROTECTED] Hello all: I am currently working on a rietveld refinement of CoFe2O4 using GSAS. The structural model I am using is Site A 1/8, 1/8, 1/8 Site B 1/2, 1/2, 1/2 Oxygen site 1/4+d,1/4+d,1/4+d On site A and B both Iron and Cobalt are present. I am trying to create a constraint that forces the ratio of the Co:Fe to be 1:2 as well as force the fractional occupancies of each site to inverse each other . In other words percent Iron goes up percent Cobalt goes down. This part I have been able to do quite easily. I thought I was able to set up the 1:2 ratio constraint correctly but I don't believe the answer because I am getting the Cobalt on both sites to have a Fraction value of ~0.28 and the Iron fraction to = ~ 0.57. I know that I may have vacancies but I don't believe both site should be the same. Can anyone please tell me how to set up a chemical ratio constraint either going through the EXPGUI or DOS interface? Sincerely, Darin Hoffman ################################################# Darin Hoffman, TSPA Research Intern, NPDF Lujan Neutron Scattering Center Los Alamos National Laboratory e-mail: [EMAIL PROTECTED] PH: 505-667-8704 #####################################################