apw247
Dear Topas fellows:
When I tried to do the Rietveld refinement for Si standard sample with both GSAS-I and Topas using the same peak profile function (TOF_PV convoluted with GSAS back-to-back), I couldn't reproduce the results in between them. Especially in the high-Q region, the result from Topas is broader than that from GSAS-I.
Here following is the macro I used in Topas, trying to reproduce the GSAS type-3 function:
"""
TOF_PV
macro GSAS_btb 'GSAS back to back function type 3, GSAS manual page 150
{
push_peak
exp_conv_const = alpha0 + alpha1 / D_spacing;
bring_2nd_peak_to_top
exp_conv_const = beta0 + beta1 / D_spacing^4;
add_pop_1st_2nd_peak
}
"""
After all, what I want to achieve is: if I grab peak profile parameters in GSAS-I (say, using the GSAS type-3 function) and put them in Topas, when I run 0 iteration calculation, I will get identical calculation result from both. Can anyone provide some guidance on how to write macros in Topas to achieve this? Many thanks!
Cheers,
Yuanpeng
alancoelho
Hi Yuanpeng
I will have a look; can you send me by e-mail your INP and data file please.
alan
apw247
Thanks a lot Alan! I will send you the files!
Cheers,
Yuanpeng
alancoelho
I think the peak shape as described in test_examples\tof_bank2_1.inp should be used; ie.
Replace the following:
GSAS_btb
with
TOF_Exponential(a0, 100, ,0, 1, difc_b5, -)
TOF_Exponential(b0, 100, b1, 100, 4, difc_b5, +)
apw247
Hi Alan,
Thank you so much for pointing that out!
Cheers,
Yuanpeng
johnsoevans
Alan: it's a long time ago, but my recollection was the profile functions gsas vs topas were essentially identical but I didn't think there was a direct correspondence between the actual numerical values of the coefficients between topas and gsas. I think Yuanpeng wanted exact equivalence (I'm guessing to tie in with rmcprofile?).