Collective excitations are one of the most common and interesting features of many-body systems. Of particular interest are those collective modes which can be interpreted in terms of vibrations. Nuclei show a large variety of such modes, both low-lying and high-lying. In particular the giant dipole resonance is due to the coherent motion of protons against neutrons. Its analogue in metal clusters is the dipole plasmon excitation, due to the oscillation of electrons against the positive ions. The Random Phase Approximation (RPA) is extensively used as a microscopic theory to study the basic properties of these collective excitations. In the derivation of RPA use is made of the Quasi Boson Approximation (QBA). It consists in replacing the expectation values in the correlated ground state with those in the uncorrelated (Hartree-Fock) one. Strictly related to QBA is the feature of RPA of predicting a harmonic spectrum. On the other hand, the existence of anharmonicities in the multiphonon spectra of nuclei and their relevance in various physical processes are well established. Overcoming QBA has represented the starting point of many attempts aiming at improving RPA. A recently developed approach in this line will be presented and discussed. A very nice property of RPA is to satisfy Energy Weighted Sum Rules (EWSR). However, some violations are present in several extensions of RPA. We will show how to cure this shortcoming.
Collective excitations are one of the most common and interesting features of many-body systems. Of particular interest are those collective modes which can be interpreted in terms of vibrations. Nuclei show a large variety of such modes, both low-lying and high-lying. In particular the giant dipole resonance is due to the coherent motion of protons against neutrons. Its analogue in metal clusters is the dipole plasmon excitation, due to the oscillation of electrons against the positive ions. The Random Phase Approximation (RPA) is extensively used as a microscopic theory to study the basic properties of these collective excitations. In the derivation of RPA use is made of the Quasi Boson Approximation (QBA). It consists in replacing the expectation values in the correlated ground state with those in the uncorrelated (Hartree-Fock) one. Strictly related to QBA is the feature of RPA of predicting a harmonic spectrum. On the other hand, the existence of anharmonicities in the multiphonon spectra of nuclei and their relevance in various physical processes are well established. Overcoming QBA has represented the starting point of many attempts aiming at improving RPA. A recently developed approach in this line will be presented and discussed. A very nice property of RPA is to satisfy Energy Weighted Sum Rules (EWSR). However, some violations are present in several extensions of RPA. We will show how to cure this shortcoming.
The second random-phase approximation (SRPA) is the simplest and most natural extension of the RPA. It enlarges the space of the elementary modes introduced to describe the collective states by adding 2 particle -2 hole excitations to the 1 particle -1 hole ones of the RPA. In deriving the SRPA equations, use is made, as in the RPA, of the so-called quasi-boson approximation (QBA) where expectation values in the ground state of the system are approximated by their values in the uncorrelated reference state. This, however, has been shown to imply a degree of approximation worse than that in the RPA. It is, therefore, necessary to improve the QBA by considering a reference state which contains some correlations. Having in mind to perform such calculations for realistic systems, we consider a simple extension of the SRPA in which the reference state contains 2 particle - 2 hole correlations. The quality of such an extension is tested by applying it to a solvable three-level model and found to be good.
A broad variety of many-body methods exists for the investigation of ground-state properties, ranging from sophisticated ab initio approaches to traditional, phenomenological models. The description of low-lying excited states of medium-mass nuclei with ab initio methods has also become possible through recent progress in many-body theory. For collective modes at higher energies, however, these methods usually cannot be applied. Therefore, when describing collective excitations either completely phenomenological, macroscopic models are employed or microscopic models using phenomenological interactions. One of the microscopic models well suited for the calculation of collective properties is the random-phase approximation (RPA). In the past, the use of phenomenological interactions for RPA has shown promising results. However, the application of chiral NN interactions yielded transitions at significantly too high energies, far from agreement with experimental data. This thesis focuses on the description of collective modes using both RPA and its second-order extension, SRPA. In contrast to previous research endeavors, we employ chiral NN+3N interactions. The use of chiral interactions is an important first step for describing ground-state, excitation and collective properties on an equal foundation. We find that the inclusion of 3N terms is crucial for RPA calculations and the prediction for collective modes is drastically improved through the 3N terms. For SRPA we show first-ever results with chiral interactions, again leading to an improvement in the predictions. For a successful ab initio description of ground-state properties the inclusion of correlations is of paramount importance. Past RPA calculations have been performed using the quasi-boson approximation, effectively neglecting ground-state correlations. Using RPA, the next step along the path towards an ab initio description of collective properties will, therefore, be the inclusion of correlations. To that end, we extend the RPA formalism to include ground-state correlations from two different many-body methods, the in-medium similarity renormalization group (IM-SRG) and coupled-cluster theory with singles and doubles excitations (CCSD). Both methods have been applied with great success for the calculation of ground-state energies. We develop a formalism based on density matrices for CC-RPA that enables RPA based on an CCSD ground state. The use of IM-SRG transformed matrix elements gives us the possibility to include ground-state correlations even at the level of SRPA. For both methods we observe a strong upward shift in the strength distributions, and, unexpectedly, we find a good agreement between IM-RPA and CC-RPA results. The structure of the transitions remains largely unchanged. We conclude that correlations have significant impact on the energetic positions, but not on the structure of the strength distributions. Employing IM-SRPA we find a strong downward shift in energy similar to the case of SRPA. The agreement of both methods with experiment is comparable.
We present a method which allows us to treat correlations in finite Fermi systems in a more consistent way than the random-phase approximation (RPA). The quasiboson approximation (QBA), where expectation values in the ground state of the system are approximated by their values in the uncorrelated reference state, is avoided in our approach where the correlated ground state is always used. We derive a closed, nonlinear set of equations which determines the energies and wave functions of the excited states as well as the single-particle occupation numbers in the ground state. As an example we apply to metallic clusters a simplified version of the approach which represents, however, a significant improvement over previous attempts to go towards a self-consistent RPA. We show that our method allows one to correct for the inadequacy of standard RPA in cases where the use of QBA becomes questionable.
Elucidating the impact of strong electronic interactions on the collective excitations of metallic systems has been of longstanding interest, mainly due to the inadequacy of the random phase approximation (RPA) in the strongly correlated regime. Here, we adopt our newly developed radial Kotliar and Ruckenstein slave boson representation to analyze the charge excitation spectrum of a Hubbard model, extended with long range interactions. Working on the face centered cubic lattice, at half filling, and in different coupling regimes ranging from uncorrelated to the metal-to-insulator transition, we compare our results to conventional RPA as a benchmark. We focus on the influence of the local and long range couplings on the particle-hole excitation continuum and the plasmon and upper Hubbard band collective modes. Beyond the weak coupling regime, we find numerous quantitative and even qualitative discrepancies between our method and standard RPA. Our work thus deepens the understanding of charge collective modes in correlated systems and lays the foundations for future studies of a broad series of materials.
A nonlinear generalization of the quasiparticle random phase approximation is constructed with the aim of describing anharmonic effects in the low-lying spectra of vibrational nuclei. Angular momentum zero pairing correlations of the ground state are described by a generalized BCS calculation based on a variational method. The dynamics of the vibrational states are described by the Heisenberg equations of motion for fermion pair operators; matrix elements of products of pair operators are treated by sum rule methods. Kinematical constraints are expressed by the algebra of the pair operators and by boson-like polynomial expansions for the quasiparticle conserving multipole operators in terms of the quasiparticle pair changing operators, these expansions enforcing the content of the Pauli principle. The physical basis for the choice of approximations is described. The final algorithm is organized so as to involve little more than repeated BCS-like and RPA-like calculations, converging to a self-consistent solution of the full nonlinear scheme. Detailed discussion is carried out for a modified form of the pairing plus quadrupole interaction Hamiltonian, but the method is not confined to this model.NUCLEAR STRUCTURE Approximate many-body methods applied to modified pairing plus quadrupole model. Description of vibrational spectrum of single closed shell nuclei. Equations of motion and algebraic methods.
A limitation common to all extensions of random-phase approximation including only particle-hole configurations is that they violate to some extent the energy weighted sum rules. Considering one such extension, the improved RPA (IRPA), already used to study the electronic properties of metallic clusters, we show how it can be generalized in order to eliminate this drawback. This is achieved by enlarging the configuration space, including also elementary excitations corresponding to the annihilation of a particle (hole) and the creation of another particle (hole) on the correlated ground state. The approach is tested within a solvable 3-level model.