Structure and energetics of Xe−n : Many-body polarization effects

Abstract

In a previous paper, Martyna and Berne, J. Chem. Phys. 8 8, 4516 (1988), diffusion Monte Carlo simulations were performed to determine the absolute binding energies of an excess electron to small clusters of xenon atoms (n≤19) using a pair additive pseudopotential. In this approximation, the electron–xenon polarization energy is treated as pair additive and therefore ignores the induced dipole–induced dipole interactions. Here we treat the many‐body polarization problem in the dipole approximation. It is found that while the smallest stable cluster anion is Xe⁻₆ for the pair polarization model this increases to Xe⁻₇ for the many‐body polarization model. In fact, the electron binding energy corresponding to the pair‐polarization model was found to be a factor of 2.7 larger than for the many‐body polarization model for all the clusters studied. In accord with this very large destabilization of electron binding energy (induced by many‐body polarization), the spatial extent of the electronic ground state in the many‐body polarization model increases compared to that of the pair polarization model. We also compare our results for both the many‐body polarization and the two‐body polarization models to corresponding dielectric continuum models developed by Stampfli and Bennemann, Phys. Rev. A 7 1, 1674 (1988). In the many‐body polarization case, the continuum model agrees well with our results. However, the agreement in the pair polarization case is rather poor for all cluster sizes. If parameters of the continuum model are adjusted to obtain agreement for small clusters sizes, the model is found to break down for large clusters sizes where the spatial extent of the electron is small enough that the microscopic details of the cluster become extremely important. A new variant of the fast Fourier transform projector method suitable for use in problems involving electron attachment to clusters is also developed. The results obtained with this new method are shown to agree with those of diffusion Monte Carlo.