ABINIT Discussion Forums

Hello!

I'm trying to do some grouptheoretical calculations with the wavefunctions which I can get from cut3d.

My calculations with materials, which can be calculated without spin-orbit-coupling (like Cu or GaP) show perfect results compared with theory. But when I want to do grouptheoretical projections of spinor wavefunctions (for e.g. SnTe), I notice, that I have to "rotate" the spinor components by a certain angle to get the results, which I expect.

To rotate the spinor, I extract both components with cut3d and place the field in a vector. The I apply a standard rotationmatrix to get my modified spinor. I noticed, that an angle of pi/4 suits my needs for every k-point in a bandstructure along W-L-G-X-K-G.

Are the spinor-components not aligned to the z-axis, or is there something else, I didn't notice?

Thanks in advance!

Sunny

BTW: The names of the output files for the two spinor components are the same and thus are overwritten by eachother...

Hi, sounds interesting. Would you consider showing us what you use to project the wf?

At any rate, the default L in L.s is oriented according to the normal axes, with z along z... The rotation you describe does sound suspicious in that it is the same for all k along a path with very different directions. Which unit cell are you using? How do you know you get what you expect? I would have imagined the expressions for the irreps would be summed over spinor index and independent of the rotation, but perhaps not.

PS: I have added a suffix for the spinor component in an upcoming version of abinit.

best

Matthieu