Issues Adding Second Phase During Refinement

[ksc165]

Hi,

I've been having trouble adding a second phase while trying to do refinements. I can fit one phase perfectly fine with no issues, but once I add a second phase, no matter what phase it is, it won't refine properly. For example, the following appeared in once scenario:

The resulting graph has peaks well beyond the viewing window. I also had another situation where these warnings didn't appear and instead I get this:

Other times, when I add another phase the Rw value doesn't change no matter how I change certain parameters for that specific phase.

Not sure if this is related in any way, but at some point I accidentally deleted some files related to GSAS-II and decided to delete the entire application. I then reinstalled it, and since then I've been having issues. I haven't been able to find any duplicate files or anything related to what could be causing this problem.

Side note as well, I'm a Mac user and when I copied the command into the terminal from 1.2. GSAS2MAIN installer for MacOS, it didn't work and returns "zsh: parse error near `\n'." However, when I copy the command from 1.2.1. MacOS: Detailed GSAS2MAIN Installation Instructions it worked. I wasn't able to find any discrepancies between the two commands so not if this is an issue on only my end.

Thank you!

[briantoby]

Hi @ksc165,

Thanks for pointing out the difference between 1.2 and 1.21. In 1.2 you are expected to change the string <install-loc> to where you want GSAS-II to be installed or the command will not work. In 1.21, I put in a default location. I will make 1.2 the same in 1.2.1. This will show up the next time I update the web pages.

There is likely something wrong with what you are doing with the software. This is not likely to be anything related to installation. The message about the Marquardt factor is saying that you are trying to fit variables that are highly correlated and that GSAS-II could not fix this. In fact after everything that was tried, the refinement process made things worse. So it gave up. The other message says that refining a bunch of parameters for the first phase had no effect at all on the quality of the fit, so those parameters were not used. The most likely way that would happen is if the unit cell parameters do not match the positions of the peaks. It is certain that your input is not describing a structure that matches your data. Please take a good look at all the values that the program is using. If you are really stuck, please get me your .gpx file (you can zip it an upload it here or e-mail it to me) and I'll see what exactly is wrong.

Note that Rietveld fitting is often poorly behaved. You need to add variables to the fit slowly. GSAS-II is much better in this than its predecessor, GSAS/EXPGUI, but still needs to be used with care, not as a black box.

Please get back to me if this is not addressing your problem.

[ksc165]

Hi,

Admittedly, I'm definitely not the best at using GSAS and was doing everything the same as I usually do until it suddenly stopped working. I had someone else who's better at using the application help me out again because we couldn't find an issue before and was able to work it out. Sorry for any inconvenience and thank you for the response!

[briantoby]

@ksc165 glad you were able to get some help. I had a lot of good mentors when I was getting started.