Ab initio electron propagator calculations in molecular transport junctions: predictions of negative differential resistance.

In this work we study current-voltage characteristics in transport molecular junctions with a 1,4-benzene dithiol molecule as a bridge by using different ab initio electron propagator methods such as OVGF and P3 which are both programs in a Gaussian software package. The current-voltage characteristics are calculated for different values of Fermi energy in various basis sets such as 6-311++G(p,d) and cc-pVDZ and are compared with the experimental data. A good agreement is found in almost the entire voltage range. In addition, the results of our calculations indicate that the accuracy of ab initio electron propagator methods is in the range of 0.2-0.3 eV. Since the computational methods are truly ab initio, implying no adjustable parameters, functions, or functionals, the theoretical predictions can be improved only by changing the model of a transport device. The current-voltage characteristics predict peaks, i.e., negative differential resistances, for the various values of Fermi energy. As shown, the origin of the negative differential resistances is related to the voltage dependences of overlap integrals for the active terminal orbitals, expansion coefficients of partial atomic wavefunctions in Dyson orbitals, and the voltage dependences of Dyson poles (ionization potentials). We find that two peak behavior in the current-voltage characteristics can be explained by the anharmonicity of potential energy surfaces. As a result of our studies, we predict that negative differential resistances can be experimentally found by changing a position of Fermi level, i.e., by using different metal electrodes coated by a gold atomic monolayer.