Question on local distortion of crystals

[bjkoz]

Hello,
I want to simulated a spherically bent crystal imaging optic in near-normal Bragg reflection geometry. I have made the ideal optic and would like to extend it to include a distorted crystal. For context, the crystal is approximately 50 micro-meters thick and initially flat. It is then bonded to a spherical substrate, with the crystal planes bent to the shape of the substrate. Imperfections in the substrate and binding will likely distort the entire crystal, both in the bulk as well as the surface.

The documentation for local_n indicates that it's possible to specify the normal for the crystal planes independently of the surface normal. However, local_n_distorted and local_z_distorted do not appear to include both surface and crystal planes as separate options. If I specify local_n_distorted, will XRT apply the distortion to the crystal planes and the surface, or just the surface?

Worst case, I can redefine the toroidal mirror class and define local_n and local_z with my distortion built in.

Thanks as always.
Bernie

[kklmn]

Hi Bernie,

Firstly, have you tried our xrtBentXtal.py calculator? If not, you may get surprised about the effect of elastic crystal deformation. Here, for example, is the effect of a 1 m spherical bending compared with the flat case:

The broadening is of the order of 100 µrad, that is huge compared with "standard" figure errors. May be, you don't really need them on the scale of this effect?

As to your question, the result of local_n_distorted is a 2- or 3-item tuple applied to the surface normal, not a 4- or 6-item tuple that would be expected for two normals. I think that one cannot fiddle with the crystal normal on a microscale at will, whereas one has that freedom with the physical surface. And you're right, one can still redefine local_n and local_z in any imaginary way.