<div dir="ltr">Hi All,<div><br></div><div>I am trying to learn how to calculate the IR spectrum with cp2k, and now I am testing on the most simple sample --- a single H2O molecule. The MD run was done within the NVE and the simulation time was over 30ps with a timestep of 0.5fs. The IR is obtained through the calculated wannier centers by applying the TRAVIS program. As we know, a single water molecule only has 3 vibrational modes: 2 stretching modes and 1 bending mode. However, the IR spectrum I obtained does not agree with that well. See below for details.</div>
<div><br></div><div>(i) The IR shows the bending mode split into two, and the stretching modes split also. I tried to increase the SCF convergency criterion, or tried the NVT ensembles, but none of them worked. There must be sth wrong, but I cannot figure it out. Could anybody help me out? Thanks a lot!</div>
<div><br></div><div>(ii) At the same time, in my simulation box (length of 10 A), along the MD time, the molecule is always drifting. When I increase the EPS_SCF from the default value to 1E-7, it got better but still drift dramatically. Any comment on this?</div>
<div><br></div><div>For your reference, here attached the input file and the corresponding IR plot. And the cp2k I used is the version 2.4.0.</div><div><br></div><div>Thanks again,</div><div><br></div><div>Jianjun</div></div>