Label for the Intermediate Scattering Function

[Anders-Markvardsen]

The intermediate scattering function is currently, within the code, labelled FQt, which is shorthand for F(Q,t). This is consistent with the e.g. the nMOLDYN paper (Gerald R. Kneller et al. “nMOLDYN: A program package for a neutronscattering oriented analysis of Molecular Dynamics simulations” https://www.sciencedirect.com/science/article/abs/pii/001046559500048K).

However, if you have worked in the field of crystallography (and with ref to the International Tables of Crystallography) using F is confusing because structure factor and this label refers to a different quantity (the Fourier transform of the density), and F(Q,t=0) in the crystallography community equals the intensity of a Bragg reflection and is denoted I_Q. You can't really win unfortunately because of the widespread use of the two different meanings of structure factor in two very closely related fields (see also https://en.wikipedia.org/wiki/Structure_factor).

Before we fix #269 and allow users to refine against intermediate scattering function data we should discuss / agree on what label to use for this quantity long-term, and if different from F(Q,t).

In the spin-echo community it seems I(Q,t) is used e.g. see https://tsapps.nist.gov/publication/get_pdf.cfm?pub_id=922151.

What do people think?

[hbordallo]

Suggestion.
If we are going to use nMoldyn, I think we should use the nomenclature that is defined in the program paper. I have attached it. Heloisa
Rog et al. • Vol. 24, No. 5 • Journal of Computational Chemistry
JCC_24_p657.pdf

[Anders-Markvardsen]

In that paper

would you be able to ask Gerald why they are using that symbol rather than a "normal" typed capital F? In particular since for dynamical scattering function they use a "normal" typed S, so why different for the intermediate scattering function? Thanks

[hbordallo]

Here what I got from Gerald:

Dear Heloisa

This is my answer to the question answered by Anders : I use the curly F abs S if the
scattering lengths are involved and normal F and S if not. The reason is that the
physical units are different - they differ by the units of a cross section (length^2).
Is this clear ?

All the best
Gerald

I think it is clear no, no?
Heloisa

[Anders-Markvardsen]

Yes and no. No, in the sense that in https://pubs.acs.org/doi/abs/10.1021/acs.jpcc.1c04322 where Gerald is also an author, the notation F is used for curly F in paper JCC paper above, hence inconsistent. Yes, in that both are F's in some ways, and maybe that is good enough!

[hbordallo]

I saw it also. I will ask him again:-) now I am curious! I was hoping that he would extend the explanation.

[hbordallo]

E-mail sent:-)