If we aim to compute the surface phonon lifetime of a Cu(111) surface, the second and third-order force constants are essential (to put in Boltzmann transport equation). However, to my knowledge, it seems that no one has performed such calculations before. Theoretically speaking, is it possible to compute the third-order force constant matrix of a slab using DFT?
Edit: It seems that calculations for 2D materials have already been performed by many people. From certain perspectives, 2D materials and surface systems (slabs) are quite similar in terms of DFT modelling. However, their phonon modes can be very different, for example, the Rayleigh mode exists only in surface systems.
I am wondering if it might be possible to apply the methods used for 2D materials to surface systems by simply adding a few more layers.
Even if you get the 2nd and 3rd force constants, it is very time consuming to solve the Boltzmann transport equation. So the feasible approach is to use the phonon spectral energy density (SED) to obtain phonon lifetimes, based on molecular dynamics (e.g., LAMMPS). If you are worried about the accuracy of molecular dynamics (which relies on the accuracy of the potential function), you can train a machine learning potential for your system from DFT calculations, which has high efficiency and accuracy.
Many codes, for example phono3py but I believe the ph.x by quantum espresso are able to calculate the 3rd order force constants. Notice that for these codes a slab is nothing more than a periodic system with a large vacuum on top. This is somewhat inefficient and you might find better solutions.
Hi Che Junwei Thank you very much for your comment. May I ask if the time-consuming nature of solving the Boltzmann transport equation for slab is primarily due to the size of the vacuum layer or the size of the slab?
Hi Roberto D'Agosta , Thank you very much for your answer. So I guess we could simply apply the methods used for 2D materials to surface systems