My PhD thesis - Condensed Matter Theory - Imperial College London
My PhD thesis - Condensed Matter Theory - Imperial College London
My PhD thesis - Condensed Matter Theory - Imperial College London
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
CHAPTER 9.<br />
ENERGY<br />
A NEW CALCULATION OF THE JELLIUM SURFACE<br />
Type of potential<br />
VMC energy (mHa)<br />
No image tail 32.54 ± 0.08<br />
Image plane at z 0 = 0.72 32.54 ± 0.08<br />
Image plane at z 0 = 1.49 32.70 ± 0.08<br />
Table 9.1: Comparison of energies calculated with different versions of the exchange-correlation<br />
potential. The DFT energies are identical for all potentials. The trial wave function for the VMC<br />
calculations contained no Jastrow factor; the simulations used a simulation cell containing 600<br />
electrons.<br />
The simulation cell is square in the xy-plane, extending from −∞ to ∞ in the<br />
z-direction. The boundary conditions therefore imply that<br />
k x = 2πm x<br />
L<br />
where m x = 0, 1, 2, . . . (9.3)<br />
with a corresponding relation for k y .<br />
However, this is not the only possible choice. The aim is to model an infinite<br />
slab, in which k ‖ takes on a continuous range of values. Recent studies [74, 55]<br />
have used ideas from band-structure theory [6, 59] in QMC simulations to improve<br />
the extrapolation to the thermodynamic limit. The idea is to introduce some fixed<br />
phase shift across the simulation cell, so that<br />
φ(r + L x ) = e iθx φ(r) (9.4)<br />
where L x is a vector of length L in the x-direction. Conventional periodic boundary<br />
conditions correspond to choosing θ x = 0. In the work of Lin et al. [55], an average<br />
over several different phase shifts is taken. Rajagopal and coworkers [74] found that<br />
using θ x = π (‘antiperiodic’ boundary conditions) gave good results. Under these<br />
boundary conditions, equation (9.3) becomes<br />
k x = π L (1 + 2m x). (9.5)<br />
This demonstrates the advantage of using antiperiodic or periodic boundary conditions:<br />
it is always possible to construct real orbitals by taking pairs of k ‖ and −k ‖ .<br />
This is not true for an arbitrary phase shift.<br />
170