Add "sample" component that changes neutron energy

[jokasimr]

This was suggested as a simple extension that would make it possible to simulate the distribution of wavelength / toa in indirect geometry spectroscopy instruments.

Note that it's not necessary for time of flight calculations, but it can be useful to better understand the distribution of wavelengths in the detector, without going all the way using a McStas.

[nvaytet]

This sounds like a pretty complicated extension... couldn't there be many different sample compositions, that would act differently, also if you apply different rotations to the sample?

Or did you have something else in mind?

[jokasimr]

In this case I was thinking something relatively simple, the sample component changes the wavelength of the neutrons according to some setting, for example $\Delta E \in Uniform(-E, E)$. Then the neutrons are propagated with the new wavelength, but we still keep the old value in the data array, for making figures etc.

Yeah it's a bit messy, but an IDS told me they would find it useful , and we could make it simpler by demanding that the beamline only has a single such component.

[nvaytet]

Ok, so just applying a DeltaE to neutrons. We don't care what it's made of, just a DeltaE distribution.
Sounds feasible, but it would require changes in many places, as we probably rely a lot on the fact that wavelengths don't change (also in the plotting...).

[jl-wynen]

Little correction to the OP: This is for direct geometries where we cannot propagate time back to the sample using known analyzer d-spacings.

From what I understand, it is correct that we only need rays in a uniform delta E because we just need to know the potential toa coverage for each incident wavelength. In an earlier discussion, Bing and I thought that this would be the simplest solution. But of course there might be a simpler one. (I'm anyway sceptical about tof getting good statistics for a monochromatic incident beam.)

[jokasimr]

we just need to know the potential toa coverage for each incident wavelength.

Yes that is the purpose. But I think we can estimate the toa coverage well by only simulating the neutrons up the sample position, and then calculate toa coverage directly based on some minimum/maximum energy transfer.

The suggested component in tof would only really improve on this simpler toa coverage estimate if the wavelength/time_of_arrival distribution at the sample is not "pointlike", which it should be.

[jokasimr]

@bingli621