local minima
Posted: Fri Feb 29, 2008 4:29 pm
Hi,
I am trying to simulate a structure under high pressures. When I go from 7 and 8 GPa to like 11 GPa, I calculate DOS my system looks metallic, However when I jump from 9 and 10 Gpa to 11 GPa, The DOS changes and shows a small band gap at fermi level. I checked bond lengths and they decrease in all jumps and if I compare the total energies they become smaller as I jump from 7->11Gpa to 10->11Gpa. Also the structure is metallic in the range 7to10Gpa and abruptly shows a bandgap at 11GPa. I am using the EDIFF= 1E-6 and EDIFFG = -.001. I changed the algorithm from conjugate gradient to IBRION = 1, still same thing. The structure doesn't change too much in these jumps . Is it possible that there are many many local minima and I get stuck up in one of them? How can I verify that?
I am trying to simulate a structure under high pressures. When I go from 7 and 8 GPa to like 11 GPa, I calculate DOS my system looks metallic, However when I jump from 9 and 10 Gpa to 11 GPa, The DOS changes and shows a small band gap at fermi level. I checked bond lengths and they decrease in all jumps and if I compare the total energies they become smaller as I jump from 7->11Gpa to 10->11Gpa. Also the structure is metallic in the range 7to10Gpa and abruptly shows a bandgap at 11GPa. I am using the EDIFF= 1E-6 and EDIFFG = -.001. I changed the algorithm from conjugate gradient to IBRION = 1, still same thing. The structure doesn't change too much in these jumps . Is it possible that there are many many local minima and I get stuck up in one of them? How can I verify that?