Hi all,

Recently I have a problem getting proper elastic modulus from energy-strain curve using COMPASS forcefield in discover module.

My model is a periodic model along two directions(y and z) and a vaccum slab along the direction(x) in which modulus is calculated.

After an initial run in NPT to determine the equilibrium structure, atoms at two ends of the x direction are fixed and small displacement is applied to one end. Then I just let the structure get equilibrium in NVT with a subsequent minimization to get the potential energy. But somehow the modulus from this method is much higher than experiment result. Is there anybody know why? Is this related to the boundary conditions applied? How can I get a more reasonable result?

Cheers,

Bruce