Value of Rw Given no d(G(r))

[Fahd]

Hi everyone!

I am working on developing protocol for getting PDF scans at my university. I was refining the standard at my lab using PDFgui when I noticed that I'm unable to get decent values for Rw. Could this be because I don't have an uncertainty column in my PDF .gr file? The nickel tutorial files refined much more cleanly and I was able to get those down to about 6-7% Rw with the same prompt appearing. I'm sorry if my question is rudimentary, I am new to this.

I am using a D8 Venture diffractometer with a silver anode and PHOTONIII-C14 flat panel detector, and I've measured at distances of 100mm and 60mm so far for 4 and 3 hours respectively, although here I'm only looking at the 60mm scan data. I usually refine my parameters in the following order:

1. data-scale factor

2. lattice parameters a,b,c (no angles)

3. Qdamp

4. Uiso (U11-U33, no anistropic U12-23s adjusted)

5. Qbroad

6. delta2 (sample collected at room temp.)

Thanks for any help and let me know if I can clarify anything!

[Mikkel Juelsholt]

Hi 

To your initial question: No Rwp is not influenced by the lack of uncertainties.

However Rwp is background and noised biased. And in the PDF you sent here your noise is just high.

That is why you do not get the same low Rwp as the Ni standard. You are using a Lab instrument so it is expected that you get get worse data.

Furthermore looking at your PDF it looks like you have a fairly nanostructured/disordered sample and the Kieserite structure is just more complicated than the FCC Ni structure. Usually that makes it harder to get all the details of the PDF perfect, thus higher Rwp. 

So in total if you want to lower your Rwp you need to measure better data. 3-4 hours on a lab instrument is not very long. 

Here is an example of a PDF from a lab instrument: https://chemistry-europe.onlinelibrary.wiley.com/doi/full/10.1002/cctc.202101220 

In this paper we measured the PDFs for something like 24 hours. Normally you would also use a measurement scheme that collects more data at high Q. So for 24 hours you would have a program that collects data for 4 hours at 0-8 Å-1, 8 hours at 8-16 Å-1 and 12 hours at 16-20 Å-1. Or something like that.

But remember if your sample is less crystalline you might need longer than that. For lab sources good data is usally measured in days and not hours. 

If you measured 2-3 days you would get much lower Rwp, but likely never as low as Ni since Ni is just easier to fit. 

Even for synchrotron sources and Rwp of 10-25 is common for more complicated samples.

Cheers Mikkel

[Fahd]

Hey Mikkel,

Just sending this to double check if you got my response (you have helped a ton).

Thanks!

Fahd

--
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/4059cad2-5147-423f-b835-fa15c5862508n%40googlegroups.com.