I want to study the deformation behavior of pure copper with free surface along the loading direction. I am using change_box command (since fix deform cant be used with shrink wrapped boundaries) to increase the strain by a factor 1.0000001 at each time step. For the system with periodic boundary conditions along all three directions, stress-strain response obtained using change box command is similar to that obtained using fix deform. But for the system with periodic boundaries along x and z and shrink wrapped boundary along y axis (loading direction) yields at large value of strain and soon becomes amorphous. I am using the same method for both ppp and psp case. Difference is because of the presence of free surface or the result is wrong? am I missing something ?. Here I am sending the LAMMPS input script and the stress-strain curves for ppp and psp boundary conditions. Please help me to solve this problem.