Corrections in Topas

dalby39

Hi all

I am doing Rietveld refinements on operando neutron diffraction data on battery cells using a curved PSD where the cell was mounted in different geometries. Specifically, the cell was mounted so the neutron beam passed through the electrode pellet from the side, i.e. transmission through the diameter of the pellet. Afterwards, the cell was tilted so that the beam hit the pellet at a 20 degree angle, i.e. a reflection geometry. I have attached an image of the two geometries. VTG is vertical transmission geometry, while HRG is horizontal reflection geometry. The pellet is in the bottom of the dark grey cylinder.

I am interested in the corrections that I need to do to compare refinements on data in the two geometries. I find it difficult to see which corrections are used per default in TOPAS. When ‘neutron_data’ is written in the INP-file, TOPAS knows to use neutron scattering lengths, but what else is TOPAS doing? According to the technical reference, a Lorentz correction is required when using ‘neutron_data’:

scale_pks = 1 / (Sin(Th)^2 Cos(Th));

But is TOPAS automatically using the Lorentz_Factor macro when ‘neutron_data’ is specified, or does the user need to add it? When I tried to add the Lorentz_Factor macro, the refinement got much worse.

For the reflection geometry, I have been using the Fixed_Incident_Beam macros by Matthew Rowles, and they seem to be working well. However, I am unsure about what TOPAS assumes about the geometry when nothing is specified. Matthew’s macros calculates an intensity correction compared with a Bragg-Brentano geometry, but is TOPAS assuming a Bragg-Brentano geometry by default? In Fullprof, the geometry has to be specified when constructing a PCR-file, but in TOPAS I don’t see what the geometry is assumed to be if no macros are used.

rowlesmr

Hi Dalby

How absorbing was your electrode? If it's anything like (the small amount) I've done, you could probably treat your setup as capillary geometry, which simplifies things quite a bit; neutrons are quite penetrating.

You do need to explicitly add Lorentz_Factor. neutron_data just uses scattering lengths instead of form factors.

The comparison with the fixed incident beam intensities with Bragg-Brentano is a little misleading, as what it is doing is comparing what happens when you change the volume of material (as a function of 2Th) that diffracts to when there is no change in volume (as a function of 2Th); Bragg-Brentano fulfills this last requirement of no volume change with angle.

The geometry assumed with no macros is that the volume of diffraction doesn't alter with angle, there is no Lorentz, there is no polarisation, there is no instrument contribution at all, and so on. Rather bare-bones.

Hope this helps.

Matthew

johnsoevans

Dalby39 and I had some prior off line chat about this. A couple of points in case of use to others:

There are some notes on the common Lorentz/Polarisation factors at:

http://topas.dur.ac.uk/topaswiki/doku.php?id=lorentz_and_polarisation_factors

There’s also more information in the book linked in the tab at the top of this page.

In your “unusual” geometry I’d always suggest measuring a standard in exactly the same conditions as the best way of making sure you have corrections correct.