Intraband plasma frequency of metal thin film
Posted: Sun Jun 15, 2014 5:42 pm
Dear all,
I have calculated the dielectric function for both bulk Al and Al(111) thin film. For bulk calculation, the plasma frequency is about 12.4eV, in good agreement with existing papers.
However, for the thin film plasma frequency, I found (1) the inplane plasma frequency is about 10.4eV, which is converged with the slab thickness. Also I did the right volume normalization excluding the vacuum. The smaller inplane plasma frequency was mentioned in the dissertation of J. Harl and also in Dr. D. Kresse's Phys. Rev. B 76, 035436 (2007). It seems reasonable with explanation that the additional interband transitions with small energy difference leads to the reduction of the plasma frequency compared to the bulk value.
(2) in slab setup of metal thin film. the out-of-plane intraband plasma frequency calculated from VASP is always Zero (or very small, can be rounded to Zero). Because the expression for plasma frequency is summation of square (velocity of the states) in the partially filled bands over band index and K point. A zero value of the summation means each individual term in the summation is ZERO. So this means the velocity is Zero.
Physically, this doesn't look right. Because Zero out-of-plane intraband plasma frequency basically means that there is only interband contribution to the dielectric function. However, in bulk case, the out-of-plane always has Drude-like intraband dielectric + interband dielectric funtion. For the Drude term, it has a peak around w=0eV.
How the slab dielectric function with only interband transition as mentioned converges to the bulk dielectric function. It looks it will never get there. But physically it has to converge to the bulk value.
Any idea is appreciated,
Wenmei
I have calculated the dielectric function for both bulk Al and Al(111) thin film. For bulk calculation, the plasma frequency is about 12.4eV, in good agreement with existing papers.
However, for the thin film plasma frequency, I found (1) the inplane plasma frequency is about 10.4eV, which is converged with the slab thickness. Also I did the right volume normalization excluding the vacuum. The smaller inplane plasma frequency was mentioned in the dissertation of J. Harl and also in Dr. D. Kresse's Phys. Rev. B 76, 035436 (2007). It seems reasonable with explanation that the additional interband transitions with small energy difference leads to the reduction of the plasma frequency compared to the bulk value.
(2) in slab setup of metal thin film. the out-of-plane intraband plasma frequency calculated from VASP is always Zero (or very small, can be rounded to Zero). Because the expression for plasma frequency is summation of square (velocity of the states) in the partially filled bands over band index and K point. A zero value of the summation means each individual term in the summation is ZERO. So this means the velocity is Zero.
Physically, this doesn't look right. Because Zero out-of-plane intraband plasma frequency basically means that there is only interband contribution to the dielectric function. However, in bulk case, the out-of-plane always has Drude-like intraband dielectric + interband dielectric funtion. For the Drude term, it has a peak around w=0eV.
How the slab dielectric function with only interband transition as mentioned converges to the bulk dielectric function. It looks it will never get there. But physically it has to converge to the bulk value.
Any idea is appreciated,
Wenmei