One of the papers I read by Toby, where (title of the paper was "R factors in Rietveld analysis: How good is good enough?"), he tells us that to get good chi square value, you must have low Rwp, where this Rwp factor is related to background. So for noisy background, one will get good chi square and for very less background noise, chi square will be high, which is happening in my case. So in the papers when one publish them, what should one do, as most of the references in papers shows chi square should be less than 2. In these cases how should one do refinement and how can one in journals raise this point as background noise are very less so our chi square is high!!
Hello, Anshul Sharma.
Background is not the only factor that affects chi2 and Rwp. If you have a very noisy background in a sample, the Rwp and chi2 would be higher than a sample with low background, but there is more to explore. Factors that affects too much these parameters are: unindexed peaks, wrong intensities of the model, shape differences, etc. Unindexed peaks seems to make the most impact on Rwp and chi2. So, to make a complete analysis of your sample, you could share your result, so we could analyze what is happening in your situation.
Feel free to send a reply if you need help.
Best regards,
Ricardo Tadeu
Ricardo Tadeu Maia I would disagree. If you have very noisy background or let´s say really bad signal to noise ratio, you can easily get GoF/chi2 down to "reasonable" levels. The better quality of your data the more will each imperfection in the refinement add to your statistics.
Michal Hegedus Thanks for your valuable answer. But what is the solution then to this problem?
Anshul Sharma I do not think there´s ultimate answer to your question. Your refinement should be acceptable while you have reasonable quality of your data/low Rwp factors and the structural model makes sense in terms of bond angles/lengths. If for sufficiently good data you are getting high Rwp or GoF - there is a problem. If you have good statistics but your structure is distorted and bong angles/lengths do not make sense - there is a problem again with your refinement.
- Badges
- Science topic
- Similar topics
- Filing