We have investigated the functional derivative of the nonadditive kinetic-energy bifunctional, which appears in the embedding potential that is used in the frozen-density embedding formalism, in the limit that the separation of the subsystems is large. We have derived an exact expression for this kinetic-energy component of the embedding potential and have applied this expression to deduce its exact form in this limit. Comparing to the approximations currently in use, we find that while these approximations are correct at the nonfrozen subsystem, they fail completely at the frozen subsystem. Using test calculations on two model systems, a H 2 O¯Li + complex and a cluster of aminocoumarin C151 surrounded by 30 water molecules, we show that this failure leads to a wrong description of unoccupied orbitals, which can lead to convergence problems caused by too low-lying unoccupied orbitals and which can further have serious consequences for the calculation of response properties. Based on our results, a simple correction is proposed, and we show that this correction is able to fix the observed problems for the model systems studied.

The electric field gradient (EFG) at the gold nucleus is calculated using a finite field approach, to make the extraction of the nuclear quadrupole moment Q from experimental nuclear quadrupole coupling constants possible. The four-component Dirac-Coulomb Hamiltonian serves as the framework, 51 of the 79 electrons are correlated by the relativistic Fock-space coupled cluster method with single and double excitations, and the contribution of the Gaunt term, the main part of the Breit interaction, is evaluated. Large basis sets (up to 26s22p18d12f8g5h uncontracted Gaussians) are employed. Energy splittings of the 2D5/2 and 2D3/2 levels, rather than level shifts, are used to extract the EFG constants, as the former remain linear with Q up to 10(-5) a.u., whereas the latter display significant nonlinearity even at Q=10(-8) a.u. Larger Q values lead to larger energy changes and better precision. Excellent agreement (0.1%) is obtained between Q values derived from 2D5/2 and 2D3/2 data. Systematic errors connected with neglecting triple and higher excitations, truncating the basis and orbital active space, and approximating the Gaunt contribution are evaluated. The final value of Q(197Au) is 521(7) mb. It is lower than the muonic 547(16) mb and agrees within error bounds with the recent value of 510(15) mb obtained from molecular calculations.

Quantum dots with three-dimensional isotropic harmonic confining potentials and up to 60 electrons are studied. The Dirac-Coulomb Hamiltonian serves as a framework, so that relativistic effects are included, and electron correlation is treated at a high level by the Fock-space coupled cluster method, with single and double excitations summed to all orders. Large basis sets composed of spherical Gaussian functions are used. Energies of ground and excited states are calculated. The orbital order is 1s, 2p, 3d, 3s, 4f, 4p, 5g, ... , and closed-shell structures appear for 2, 8, 18, 20, 34, 40, and 58 electrons. Relativistic effects are negligible for low strengths of the harmonic potential and increase rapidly for stronger potentials. Breit contributions, coming from the lowest order relativistic correction to the interelectronic repulsion terms, are also studied. Correlation effects are significant for these systems, in particular for weak confining potentials and for small systems, where they constitute up to 6% of the total energies. Their relative weight goes down (although they increase in absolute value) for larger systems or confining potentials. Planned applications to quantum dots with impurities are discussed briefly.

No abstract

Electric field gradients at the nuclei of gallim and indium are determined by finite field calculations of the atomic energies as functions of the nuclear quadrupole moments. The four-component Dirac–Coulomb–Gaunt Hamiltonian serves as framework, and all electrons are correlated by Fock-space coupled cluster with single and double excitations or by single reference coupled cluster with approximate triples. Large, converged basis sets (e.g., 28s24p20d13f5g4h for In) and virtual spaces are used. Together with experimental nuclear quadrupole coupling constants, known with high precision, the calculated electric field gradients yield the nuclear quadrupole moments. For 69Ga, we get Q = 174(3) mb, in agreement with the earlier 171(2) mb obtained from molecular calculations. The 115In moment is Q = 772(5) mb, considerably lower than the previously accepted 810 mb, and in good agreement with the recent molecular value of 770(8) mb.