Hi, Thanks for nice and clear videos about refinement. But one thing, I couldn't find any Bruker made XRD instruments in the instruments configuration list. If its already there can you please suggest best suitable one for "D8 Advance X-Ray diffractometer" !!
Question: I use .xrdml files, from Highscore, with Device file: PAN-Excel Auto slits. After identification, the calculated profile fits perfectly, meaning different peaks display the phase profiles. Doing Rietveld refinement the phase profiles fall flat - get no fitting at all. X2 = 59.06 and Rwp = 28.05. I refine the same pattern with Highscore use the same phases - all peaks are resolved, fitting is perfect. Rwp = 5.9 and GOF = 1.67. Am I missing something in Profex, I must do something wrong?
Something must be going wrong, but I can't tell what based on your description. If you would like me to have a look, please go to the Profex website, find my email address on the contact page, and send me your scan file and instrument files.
Would it be possible to calculate the amorphous content using the external standard method (O'Connor and Raven, 1988) in Profex? I am able to calculate the K-Factor from the scan of the Corundum, but not sure how to implement it in the scan with amorphous content. Would Phase1 - (GEWICHT1 * MACsample)/K-Factor give the actual quantity of phase 1? How would I get the amorphous content from this? Is there any normalisation of the GEWICHT parameters required? Thank you for this amazing software.
You are on the right track. Assuming that corundum has quantity 1.0: 1.0 = (GEWICHTcor * MACcor) / K you get K from: K = GEWICHTcor * MACcor The weight fraction of phase 1 in the other sample is: Wphase1 = GEWICHTphase1 * MACsample / K And your amorphous content is: Wamorph = 1 - Wphase1
hi, thanks for the amazing software and easy tutorials, if i do refinement of multiple phases like cement and some phases show 0% phases quantity (chemical composition) then it means absence of that phase or present in low amount ?
Excellent!! Thank you very much for your help!! I wonder what happens when you have a new material that doesn't have known phases... I mean, all of the peaks in the XRD spectrum don't fit with any of the phases in the database?? What can I do in that case? Thank a lot for your answer.
It is unlikely that the structure is really unknown, considering that the big databases (eg the COD) contain more than 400k entries. The Profex database, however, only has around 1000 entries. Check out the video on "phase identification with QualX" or so. It explains how to search the big COD and import the matching structures in Profex.
Hi Greetings. Is there any way around if we are not sure of the instrument configuration in which the test was performed ? To what extent will this detail affect the refinement of the results ?
It depends on how close the configuration is to your true hardware setup. If it is totally wrong, you may not get anywhere close to a good fit. But if it is mostly correct, maybe differs in the positions of the slits or some other minor discrepancies, phase quantities are usually still ok. The first parameters affected are those related to peak width, i.e. crystallite size and micro-strain. I recommend to check your configuration by refining a highly crystalline reference sample, such as Al2O3 or LaB6. If you get a good fit and realistic crystallite sizes, you are using a correct configuration file.
It means that you are over-refining your dataset. There must be correlating parameters. As long as chi^2 is just a little bit below 1, e.g. 0.97, I wouldn't worry, it seems to be an excellent refinement. But if it goes further below 1, you should reduce the number of refined parameters.
Thanks a lot for this amazing tutorial. It is extremely clear and didactic. I have one question: when I look line by line at my control file I don't see the line related to the background. As I am measuring samples containing clays from 3 (2 theta degrees), the background refinement at low angles is only feasible to manage from that line in the code (As you stated in other tutorial). Do you know if this is a bug? Is it related to the instrument configuration I chose? Thanks a lot for your feedback.
If you don't specify the number of background coefficients (RU=), it will be set automatically. You can see the number in the "statistics" section of the "results" window. You can always just override it by adding the keyword "RU=n" if it is missing in your control file. Hope that answers your question.
@@profextutorials2325 Thank you very much for your prompt answer. What I noticed in the linked files is that indeed with "RU=n" you can modify the number of background coefficients, but the second line related to the file that links the device is missing. In the files, for the device I use (Bruker D2) there is no an .xy file for that specific device, so I could adapt the background for my refinement, but not include an additional background file. Could I also just override the code to include this file?
If you have a scan that only contains your background, you can convert it to ASCII *.xy format and reference it with UNT=.xy. You shouldn't use a background scan from a different instrument, because they will be highly instrument dependent. In my case, some waves in the background originated from the polymeric sample holders that were hit by the beam. So I flipped a sample holder around and measured the pure polymer on the backside. That is the background I am using with the RMS-D8... instrument. Maybe the same works for you. Or if you want to include an amorphous bump in the background, you need to measure the pure amorphous phase and use this scan as a background scan. You can either reference the background scan in the instrument editor. Then the line "UNT=.xy" will be added to the project template and it will be applied to new projects automatically. Or you can add the UNT line manually to an existing refinement, but then you will have to copy the .xy scan to the project directory. Else it will not be found during the refinement and will silently be ignored (IIRC there will be no error message).
Sir, firstly thankyou so much for such a wonderful explanation. One ques: for as my material is cuco2s4, during phase match, what should i select, as i m selecting mineral and then select elements from periodic table then it come error that no phase matches. Plz reply at earliest
Your phase is probably not yet in the default repository. In that case, search-match makes no sense. Since you already know what phase it is, you can instead search for it in the COD database (file > import structure files > query COD database). You need to make sure that the COD is installed and connected. There is a video explaining it on this channel.
@Nilesh You can create a refinement report from the menu "Results -> Generate Report". That will create a HTML report and open it in a web browser. From there you can print it to PDF. To customize the content and layout of the report, check out the user manual Part 2 Section 5: www.profex-xrd.org/wp-content/uploads/2022/04/Profex-BGMN-Part-2-Application-EN.pdf
Firstly: Fantastic Software!... I am blown away. And you Tuts are just on point. I just have a quick question (a silly one perhaps): When I was doing the third refinement (releasing the B1 parameter) I noticed that my X2 was slightly lower than the one you calculated (mine = 1.17) Are these differences to be expected even though I am following your steps to the letter? What could be the reason?(Just curious).
Thanks! I noticed that using different computers to do exactly the same refinement can lead to slightly different results. I don't know why, my only explanation is that it is an accumulation of rounding errors. There are a lot of floating point operations going on, and these are not infinitely accurate. Maybe different computers or operating systems reach refinement convergence at a slightly different number of steps. But that's just a guess.
I used the scherr formula to calculate the grain size and it gave me a reasonable result but when I used profex it returned and error. Is there an explanation?????
Check in the *. lst file if the B1 parameter was refined as 0.0. When that happens, there is no crystallite size related peak broadening, the crystals appear "infinitely large" to Profex, and as a result of this the GrainSize reports ERROR.
An awesome video on a great software. Thanks a lot! One small question, what's a good resolution for capturing diffractograms to use Profex successfully? I am using data with 0.025 degrees resolution. Is that good enough? Thanks.
The sharpest peak should be sampled with at least 5 data points. I am using 0.0122 degrees, with the previous instrument 0.02, both worked equally well. 0.025 is in the ballpark, but whether or not it is ideal depends on the minimum peak width your instrument produces.
This can have different reasons: - if the goal reported an error (e.g. in case of infinitely large crystals / B1=0), it will not be written to the report. This will change with the upcoming version 4.2.6, which will allow you to either skip errors or write them to the report anyway. - or maybe you don't have the instruction to write crystallite sizes to the report in your Profex preferences. To check that, go to "Edit -> Preferences -> BGMN -> Summary Tables" and check your "Local Parameters and GOALs". There should be a line "^GrainSize\(\d+,\d+,\d+\)$" (without the quotes) in your listed GOALs. If not, add the line and try creating the report again.
Excellent!! Thank you very much for posting this tutorial.
Hi, Thanks for nice and clear videos about refinement. But one thing, I couldn't find any Bruker made XRD instruments in the instruments configuration list. If its already there can you please suggest best suitable one for "D8 Advance X-Ray diffractometer" !!
Thanks for detailed video.
Really will be helpful for my work. Full of information.
Question: I use .xrdml files, from Highscore, with Device file: PAN-Excel Auto slits. After identification, the calculated profile fits perfectly, meaning different peaks display the phase profiles. Doing Rietveld refinement the phase profiles fall flat - get no fitting at all. X2 = 59.06 and Rwp = 28.05. I refine the same pattern with Highscore use the same phases - all peaks are resolved, fitting is perfect. Rwp = 5.9 and GOF = 1.67.
Am I missing something in Profex, I must do something wrong?
Something must be going wrong, but I can't tell what based on your description. If you would like me to have a look, please go to the Profex website, find my email address on the contact page, and send me your scan file and instrument files.
Amazing tutorial! Very helpful and didatic.
Would it be possible to calculate the amorphous content using the external standard method (O'Connor and Raven, 1988) in Profex? I am able to calculate the K-Factor from the scan of the Corundum, but not sure how to implement it in the scan with amorphous content. Would Phase1 - (GEWICHT1 * MACsample)/K-Factor give the actual quantity of phase 1? How would I get the amorphous content from this? Is there any normalisation of the GEWICHT parameters required?
Thank you for this amazing software.
You are on the right track. Assuming that corundum has quantity 1.0:
1.0 = (GEWICHTcor * MACcor) / K
you get K from:
K = GEWICHTcor * MACcor
The weight fraction of phase 1 in the other sample is:
Wphase1 = GEWICHTphase1 * MACsample / K
And your amorphous content is:
Wamorph = 1 - Wphase1
hi, thanks for the amazing software and easy tutorials, if i do refinement of multiple phases like cement and some phases show 0% phases quantity (chemical composition) then it means absence of that phase or present in low amount ?
It means the phase is below your detection limit. It could be completely absent, or present in an amount too low to generate peaks above the noise.
did anyone use this software, I am getting some errors. My Chi2 value is coming close to 40 and is not getting reduced. Can anyone help ?
Excellent!! Thank you very much for your help!! I wonder what happens when you have a new material that doesn't have known phases... I mean, all of the peaks in the XRD spectrum don't fit with any of the phases in the database?? What can I do in that case? Thank a lot for your answer.
It is unlikely that the structure is really unknown, considering that the big databases (eg the COD) contain more than 400k entries. The Profex database, however, only has around 1000 entries. Check out the video on "phase identification with QualX" or so. It explains how to search the big COD and import the matching structures in Profex.
Hi
Greetings. Is there any way around if we are not sure of the instrument configuration in which the test was performed ? To what extent will this detail affect the refinement of the results ?
It depends on how close the configuration is to your true hardware setup. If it is totally wrong, you may not get anywhere close to a good fit. But if it is mostly correct, maybe differs in the positions of the slits or some other minor discrepancies, phase quantities are usually still ok. The first parameters affected are those related to peak width, i.e. crystallite size and micro-strain.
I recommend to check your configuration by refining a highly crystalline reference sample, such as Al2O3 or LaB6. If you get a good fit and realistic crystallite sizes, you are using a correct configuration file.
@@profextutorials2325 Many Thanks for such a clear reply. I express my sincere gratitude
Hi, Thanks for the nice software. Can one do only the hkl fitting like TOPAS to get the cell parameter only?
Yes, it's called LeBail refinement:
ruclips.net/video/YXjszm2ZZ0Y/видео.html
Many thanks for the great tutorial! Why can Rwp be less than Rexp and GOF less than 1? for multiphase sample. Is it reasonable?
It means that you are over-refining your dataset. There must be correlating parameters. As long as chi^2 is just a little bit below 1, e.g. 0.97, I wouldn't worry, it seems to be an excellent refinement. But if it goes further below 1, you should reduce the number of refined parameters.
Thanks a lot for this amazing tutorial. It is extremely clear and didactic.
I have one question: when I look line by line at my control file I don't see the line related to the background. As I am measuring samples containing clays from 3 (2 theta degrees), the background refinement at low angles is only feasible to manage from that line in the code (As you stated in other tutorial).
Do you know if this is a bug? Is it related to the instrument configuration I chose?
Thanks a lot for your feedback.
If you don't specify the number of background coefficients (RU=), it will be set automatically. You can see the number in the "statistics" section of the "results" window. You can always just override it by adding the keyword "RU=n" if it is missing in your control file. Hope that answers your question.
@@profextutorials2325 Thank you very much for your prompt answer.
What I noticed in the linked files is that indeed with "RU=n" you can modify the number of background coefficients, but the second line related to the file that links the device is missing.
In the files, for the device I use (Bruker D2) there is no an .xy file for that specific device, so I could adapt the background for my refinement, but not include an additional background file.
Could I also just override the code to include this file?
If you have a scan that only contains your background, you can convert it to ASCII *.xy format and reference it with UNT=.xy. You shouldn't use a background scan from a different instrument, because they will be highly instrument dependent. In my case, some waves in the background originated from the polymeric sample holders that were hit by the beam. So I flipped a sample holder around and measured the pure polymer on the backside. That is the background I am using with the RMS-D8... instrument. Maybe the same works for you. Or if you want to include an amorphous bump in the background, you need to measure the pure amorphous phase and use this scan as a background scan.
You can either reference the background scan in the instrument editor. Then the line "UNT=.xy" will be added to the project template and it will be applied to new projects automatically. Or you can add the UNT line manually to an existing refinement, but then you will have to copy the .xy scan to the project directory. Else it will not be found during the refinement and will silently be ignored (IIRC there will be no error message).
@@profextutorials2325 Thank you very much. I tried this way and it has certainly improved the refinement. Looking forward to more tutorials.
Sir, firstly thankyou so much for such a wonderful explanation.
One ques: for as my material is cuco2s4, during phase match, what should i select, as i m selecting mineral and then select elements from periodic table then it come error that no phase matches.
Plz reply at earliest
Your phase is probably not yet in the default repository. In that case, search-match makes no sense. Since you already know what phase it is, you can instead search for it in the COD database (file > import structure files > query COD database). You need to make sure that the COD is installed and connected. There is a video explaining it on this channel.
How to get file in pdf or other with all data likes peaks list, hkl values etc.
@Nilesh You can create a refinement report from the menu "Results -> Generate Report". That will create a HTML report and open it in a web browser. From there you can print it to PDF.
To customize the content and layout of the report, check out the user manual Part 2 Section 5:
www.profex-xrd.org/wp-content/uploads/2022/04/Profex-BGMN-Part-2-Application-EN.pdf
Firstly: Fantastic Software!... I am blown away. And you Tuts are just on point. I just have a quick question (a silly one perhaps): When I was doing the third refinement (releasing the B1 parameter) I noticed that my X2 was slightly lower than the one you calculated (mine = 1.17) Are these differences to be expected even though I am following your steps to the letter? What could be the reason?(Just curious).
Thanks! I noticed that using different computers to do exactly the same refinement can lead to slightly different results. I don't know why, my only explanation is that it is an accumulation of rounding errors. There are a lot of floating point operations going on, and these are not infinitely accurate. Maybe different computers or operating systems reach refinement convergence at a slightly different number of steps. But that's just a guess.
Hello, thanks a lot for super nice videos about refinement. I'm wondering Global goals Qbeta TCP 0.8 means 80 wt% or mol%?
@Xinfnag Li The quantities are given in wt-%.
@@profextutorials2325 Thanks for your nice reply
I used the scherr formula to calculate the grain size and it gave me a reasonable result but when I used profex it returned and error. Is there an explanation?????
Check in the *. lst file if the B1 parameter was refined as 0.0. When that happens, there is no crystallite size related peak broadening, the crystals appear "infinitely large" to Profex, and as a result of this the GrainSize reports ERROR.
Is it possible to do Rietveld Refinement just inserting two-theta and intensity file? If yes, then why it is not processing?
Yes it should work. Please follow the instructions here to send a support request:
www.profex-xrd.org/?page_id=1171
Thanks Sir, I will check!
An awesome video on a great software. Thanks a lot! One small question, what's a good resolution for capturing diffractograms to use Profex successfully? I am using data with 0.025 degrees resolution. Is that good enough? Thanks.
The sharpest peak should be sampled with at least 5 data points. I am using 0.0122 degrees, with the previous instrument 0.02, both worked equally well. 0.025 is in the ballpark, but whether or not it is ideal depends on the minimum peak width your instrument produces.
Thank You for all you do...
why is it that I'm getting the same grain size value for different (hkl) positions?
most likely a cubic structure
The grain size value wasn't exported to the report file and I tried several times and I did not work can you help me?
This can have different reasons:
- if the goal reported an error (e.g. in case of infinitely large crystals / B1=0), it will not be written to the report. This will change with the upcoming version 4.2.6, which will allow you to either skip errors or write them to the report anyway.
- or maybe you don't have the instruction to write crystallite sizes to the report in your Profex preferences. To check that, go to "Edit -> Preferences -> BGMN -> Summary Tables" and check your "Local Parameters and GOALs". There should be a line "^GrainSize\(\d+,\d+,\d+\)$" (without the quotes) in your listed GOALs. If not, add the line and try creating the report again.
@@profextutorials2325 thank you
amazing tutorial, thank you!
41:10 Export of refinement results
56:00 Export the graph with SVG format