Dear Users,
I am working with a core–shell BTO ferroelectric system in LAMMPS and trying to compute piezoelectric and dielectric properties using a static (0 K) approach.
My general workflow is:
- First, I relax/minimize the structure
- Then I apply a small electric field
- For dij: I allow box relaxation (e.g., “fix box/relax aniso”) and measure the induced strain
- For the dielectric response: I apply a small electric field and perform energy minimization, then compute the polarization
However, I observe a consistent issue:
If I relax or minimize the structure too strongly before applying the electric field, I do not observe any meaningful response afterward.
On the other hand, if the structure is not properly relaxed, the response to the electric field also disappears.
In both cases, the result is unphysical: even when I vary the applied field, the strain (for dij) and the polarization (for dielectric response) remain almost unchanged.
It seems that either over-minimization or insufficient relaxation destroys the correct ferroelectric response.
Could you please clarify how the initial relaxation should be properly performed in static (0 K) calculations for ferroelectric core–shell systems in order to obtain reliable piezoelectric and dielectric properties?
Thank you very much for your help.