The accuracy of total electronic energies obtained using the fixed-node diffusion quantum Monte Carlo (FN-DMC) method is determined by the choice of the many-body nodal surface. Here, we perform a systematic comparison of the quality of FN-DMC energies for a selection of atoms and diatomic molecules using nodal surfaces defined by single determinants of Hartree-Fock, B3LYP, and LDA orbitals. Through comparison with experimental results, we show that the use of Kohn-Sham orbitals results in significantly improved FN-DMC atomization energies over those obtained using Hartree-Fock orbitals. We also discuss the effect of spin contamination in the orbitals.

We study theoretically the nature of correlations in space and time of the current in a one-dimensional bilinear array of tunnel junctions in the normal conduction limit, using the kinetic Monte Carlo (KMC) method. The bilinear array consists of two parallel rows of tunnel junctions, capacitively coupled in a ladder configuration. The electrostatic potential landscape and the charge-charge interaction length both depend on the circuit capacitances, which in turn influence transport and charge correlations in the array. We observe the formation of stationary charge states when only one rail is voltage biased. When a symmetric bias is applied to both rails, the site at which the positive and negative charge carriers recombine can drift throughout the array. We also calculate charge densities and auto- and cross-correlation functions.Comment: 11 pages, 16 figure

We simulate one-dimensional arrays of tunnel junctions using the kinetic Monte Carlo method to study charge filling behaviour in the large charging energy limit. By applying a small fixed voltage bias and varying the offset voltage, we investigate this behaviour in clean and disordered arrays (both weak and strong disorder effects). The offset voltage dependent modulation of the current is highly sensitive to background charge disorder and exhibits substantial variation depending on the strength of the disorder. We show that while small fractional charge filling factors are likely to be washed out in experimental devices due to strong background charge disorder, larger factors may be observable.