You need to remember that results of MD simulations depend crucially on the
empirical potential used. The Tersoff potential could not describe the
optical phonons very well and it is not a surprise if some fine details for
the optical phonons are missing or wrong compared to experiments. In this
case, you may need first-principles calculations. What I can say is that
the DOS calculation method implemented in GPUMD has no problem.
On Wed, Jul 8, 2020 at 11:03 AM 孙琳 <sunlin18@xxxxxxxxxx> wrote:
I have tried to use the "compute_dos" command to calculate the PDOS of a
single layer graphene with opt-Tersoff potential and EMD method. The output
is very similiar with that in ex2 of gpumd.
According to references of Balandin, there should be two peaks around 1350
cm-1 (or 40.5 THz) and 1585 cm-1 (or 47.55 THz) for the in-plane phonons,
corresponding to TO mode around BZ edge and LO/TO mode around BZ origin,
respectively. However, the 1350 cm-1 peak is missing in both my
calculations and ex2. Thus, could you help me explain that why the peak in
1350 cm-1 is missing? Is it possible to get the peat around 1350 cm-1 by
just optimizing parameters in ex2? Thank you very much!