Stress sign convention

woates · Post by **woates** » Sun Jun 30, 2013 3:18 am

Hello,

I've been calculating the energy and stress in a unit cell of barium titanate. I run a clamped ion simulation in the cubic state and several additional simulations using linear increments in ion positions towards tetragonal. I was expecting the stress to be negative near the cubic state indicating compression. All three normal stress components are positive. As a neophyte to DFT, but knowledgable in mechanics, does the stress calculated from DFT follow standard mechanics sign convention?

I tried checking this against 1/Volume*dE/d(strain_33)=sigma_33 where dE is the change in total energy between increments in clamped ion positions. The signs are clearly flipped between stress output in Abinit versus this change in energy with respect to strain calculation. Also note that not only the signs but the magnitudes of the stress are also different which I have yet to resolve.

Thanks,
Billy