Dear all,
I have alpha-quartz powders and I did 3 hours XRD scan for these powders both at 30 degrees Celsius and 450 degrees Celsius (there exists no phase transformation at this temperature). After refining by TOPAS-5, I wonder why my Rwp is 14% for 30 degrees and 19% for 450 degrees. Would you please check to see if the steps mentioned below are correct and Is the attached Rietveld refinement for silica done by TOPAS publishable?
I want to use TOPAS to find out the effects of temperatures on unit cell parameters (trying to calculate the thermal expansion coefficient). I calculated the thermal expansion coefficient of alpha-quartz powders according to the change of the variation in unit cell parameters, as shown in the following table and this result is in accordance with other researcher's value.
Steps:
First for 30 degrees:
1- I loaded the scan file (30 degrees). 2- load the emission file.
3- set the background order to 4
4- loaded instrument details
5- chose the correction (zero error, sample displacement, LP factor)
6- start X from 22.5 because there exists a large disagreement between the actual scan pattern and fitted scan pattern before 22.5 degree
6- load the quartz.str file (PDF Card - 00-046-1045)
7- I changed all the occ and beg from fix to refine.
8- Run and check the Rwp value
9-Save structure in str format for 450 degrees (named as fitted-30degree.str)
Then for 450 degrees:
1- I loaded the scan file (450 degrees). 2- load the emission file. 3- set the background order to 4 4- loaded instrument details 5- chose the correction (zero error, sample displacement, LP factor) 6- start X from 22.5 6- load the fitted-30degree.str file 7- I changed all the occ and beg from fix to refine. 8- Run and check the Rwp value
Regards, Zhaoyang
Your Rwp is going to be dominated by the main peak. Use the cumulative chi2 button to see what is contributing to your Rwp.
What is misfitting before 22.5°? peak position, intensity, shape? What was the instrument geometry? if Bragg_Brentano, was the beam wholely within the sample for the entire angular range?
I wouldn't refine both zero error and specimen displacement.
How are you calculating the thermal expansion? just between your two data points, or is there other data?
why are you refining occupancies?
Generally speaking, your Rwp values are too high for such a trivial structure. Unfortunately, it is impossible to understand from your plots what went wrong in the refinement.
However, there are some obvious things to improve:
1. The experimental data between 20 and 23 degrees cannot be simply excluded. This range should be readily accessible with a normal laboratory equipment, so you need a very convincing explanation of the reasons why you could not get good data.
2. The accuracy of the numbers in the table looks overestimated for me. The standard deviations should be provided.
3. Thermal expansion is never linear. So the value obtained from the difference in the lattice parameters at 450C and room temperature is not very informative.
4. Thermal expansion is different in different directions. So it should be clearly stated which coefficient you show. Is it volumetric or linear? If linear, along which axis?
The short answer is no. There are many parameters you shouldn't have refined. One of them is the zero error (you have lab data and we assume that your diffractometer was already aligned). In any case, one should never refine the zero error and the sample displacement simultaneously. What is the reason you excluded the 2 theta range < 22 deg.? It is hard to say what is wrong with your refinement just by looking at the plots. We need to see a .pro file. What peak shape model do you use? Also, you should calculate the errors.
- Badges
- Science topic
- Similar topics
- Engineering
- Materials Engineering
- Powders