sratio on Diffpy-CMI ?

55 views
Skip to first unread message

Clara Doisneau

unread,
May 13, 2025, 9:10:43 AMMay 13
to diffpy-users
Dear all,

I have a question concerning fitting a PDF with Diffpy-CMI.
The material that we are analyzing is extremely covalent (boron phosphide B1P1), and fitting with delta2 is not working well. However I did not see sratio and rcut like on PDFgui so I wanted to know if it was implemented ? 
I checked the documentation and I did not see it. 

Thank you in advance,
Best regards,
Clara DOISNEAU 

Simon Billinge

unread,
May 13, 2025, 9:53:00 AMMay 13
to diffpy...@googlegroups.com
Thanks for the question Clara,

I am pretty sure that Long Yang looked at that.  I will forward this to him in case he is not monitoring diffpy-users....

On Tue, May 13, 2025 at 9:10 AM Clara Doisneau <doisnea...@gmail.com> wrote:
Dear all, I have a question concerning fitting a PDF with Diffpy-CMI. The material that we are analyzing is extremely covalent (boron phosphide B1P1), and fitting with delta2 is not working well. However I did not see sratio and rcut like on
ZjQcmQRYFpfptBannerStart
This Message Is From an External Sender
This message came from outside your organization.
 
ZjQcmQRYFpfptBannerEnd
--
You received this message because you are subscribed to the Google Groups "diffpy-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diffpy-users...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/diffpy-users/e1824314-5193-4e3d-96e5-e9ec5979b546n%40googlegroups.com.


--
Simon Billinge
Professor, 
Department of Applied Physics and Applied Mathematics
Columbia University

Long Yang

unread,
May 13, 2025, 10:52:09 AMMay 13
to diffpy-users
Dear Clara,

Thanks for your question. Unfortunately, `sratio` and `rcut` are not implemented in the CMI. You may refer to a previous post for more details. https://groups.google.com/u/2/g/diffpy-users/c/JcFu2R6ZlTw/m/OCHLT528BAAJ

Or you may try wrapping it up in python to obtain a similar effect in the PDFgui, as follows:

If r_ij < rcut:
sigma_ij -> sigma_ij * sratio   (where sigma_ij is the width of the peak jointing atom i and atom j)

In addition, as you may have known, `sratio` is introduced for describing the correlated motion effects, so it is tricky if you use `delta` parameter in the meantime.


Bests,
Long Yang
Assistant Professor
School of Materials Science and Engineering
Tongji University, Shanghai 201804, China
https://www.yanglonggroup.com/

Simon Billinge

unread,
May 13, 2025, 10:58:20 AMMay 13
to diffpy...@googlegroups.com
As a general matter, the rcut/sratio is a kind of brual fix that can make subsequent regression a bit unstable.  It would be nice to come up with a more elegant solution to this and implement it in cmi if anyone wants to have a go, we would be delighted to have a PR from the community!

We have a project for this summer to bring diffpy-cmi up to latest Python and working on windows natively, so now is a great time to get new features merged!

Best wishes,

Simon

On Tue, May 13, 2025 at 10:52 AM Long Yang <dragon...@gmail.com> wrote:
Dear Clara, Thanks for your question. Unfortunately, `sratio` and `rcut` are not implemented in the CMI. You may refer to a previous post for more details. https: //groups. google. com/u/2/g/diffpy-users/c/JcFu2R6ZlTw/m/OCHLT528BAAJ Or you may

Clara Doisneau

unread,
May 15, 2025, 9:48:33 AMMay 15
to diffpy-users
Thank you both for your answer ! Working with covalent material is a bit complicated and delta is definetely not the way to tackle it (for our material at least), so I will try the part of code from Long to see if it works in the meantime, 
Thank you a lot, this forum is extremely helpfull ! 

Best regards,
Clara DOISNEAU  

Simon Billinge

unread,
May 15, 2025, 9:54:34 AMMay 15
to diffpy...@googlegroups.com
Thanks Clara,

Your question has motivated us to seek a more robust solution too....  we are resource limited (not enough people) but we have ideas how to do it and will try.  If anyone wants to help please reach out.

S

On Thu, May 15, 2025 at 9:48 AM Clara Doisneau <doisnea...@gmail.com> wrote:
Thank you both for your answer ! Working with covalent material is a bit complicated and delta is definetely not the way to tackle it (for our material at least), so I will try the part of code from Long to see if it works in the meantime, Thank
Reply all
Reply to author
Forward
0 new messages