Hello Artur,
We always appreciate feedback from our users and take this kind of request seriously when deciding on future release plans, so thank you for taking the time to post. The native capability to model these kinds of systems would be beneficial to OptiFDTD and is something worth considering, though is limited to very specific applications.
In the mean time I believe we actually can design the system you are describing through the use of our scripting capabilities, provided that the tensor you are looking to model is diagonal. To do this you would create a series of “materials” that had different permittivity tensors defined within the profile designer. You could then run a script that would assign these predefined materials either to small spheres assigned to each grid point(a computationally costly proposition depending on the size and resolution of your system) or create structures that would mimic the size and shape of individual structures within your crystal and assign the material to them.
I hope this gives you a chance to try to model your systems and encourage you to share how this worked out for you.
Regards,
Scott