Causal optimization method for imaginary-time Green's functions in interacting electron systems

The imaginary-time Green's function is a building block of various numerical methods for correlated electron systems. Recently, it was shown that a model-independent compact orthogonal representation of the Green's function can be constructed by decomposing its spectral representation. We investigate the performance of this so-called intermediate representation (IR) from several points of view. First, we develop an efficient algorithm for computing the IR basis functions of arbitrary high degree. Second, for two simple models, we study the number of coefficients required to represent the Green's function within a given tolerance. We show that the number of coefficients grows only as $O(ln\ensuremath{\beta})$ for fermions and converges to a constant for bosons as temperature $T=1/\ensuremath{\beta}$ decreases. Third, we show that this remarkable feature is ascribed to the properties of the physically constructed basis functions. The fermionic basis functions on the real-frequency axis have features whose width is scaled as $O(T)$, which is consistent with the low-$T$ properties of quasiparticles in a Fermi liquid state. On the other hand, the properties of the bosonic basis functions are consistent with those of spin/orbital susceptibilities at low $T$. These results demonstrate the potential wide applications of the IR to calculations of correlated systems.

We show that the imaginary-time Green's functions of a system of quantum charged particles at equilibrium satisfy a set of sum rules which express the screening of a classical external charge. We also provide a new derivation of the perfect screening property ${\mathrm{lim}}_{\mathrm{k}\ensuremath{\rightarrow}0}$${\mathrm{\ensuremath{\epsilon}}}^{\mathrm{\ensuremath{-}}1}$(k)=0 where \ensuremath{\epsilon}(k) is the static dielectric function. The proof relies on exact equations for the Green's functions and certain assumptions of spatial clustering. It is nonperturbative, independent of the statistics, and applies to jellium as well as to multicomponent systems.

Exchange energy and specific heat of the electron gas at low temperature

Ab initio green's function calculations on highly conducting polymers: Effects of electron correlation and aperiodicity

The green's function method in quantum statistics

<sec><p indent="0mm">Exotic nuclei far from the β stability line have become important scientific goals in the studies of experimental physics at large scientific facilities and theoretical research due to the rich new physics. In these exotic nuclei, the neutron or proton Fermi level is close to the continuum threshold. The pairing correlation could scatter the valence neutrons or protons into the continuum. This leads to the extended neutron or proton density distributions in these exotic nuclei. Therefore, properly describing the pairing correlation and continuum is crucial for studying the structures and properties of these exotic nuclei. The Hartree-Fock-Bogoliubov (HFB) theory is one of the promising tools for describing exotic nuclei. In this theory, people usually solve the HFB equation directly in the coordinate space or the Woods-Saxon basis. In the coordinate space, when employing box boundary conditions to discretize the continuum, the behavior of the wave functions at the boundaries will be affected by the size of the box. Meanwhile, one could not obtain the energy and width of the resonant states directly from the discretized continuum states. The Green’s function (GF) is a simple and effective tool for handling the continuum, which has been widely used in nuclear structure research. The GF can be established by using the wave functions that satisfy the equation of motion and the proper bound boundary conditions for the bound and continuum states. Then one can use the loop integral of the GF on the complex energy plane to construct the particle density and level density. In this way, the continuum states can be included in the density with proper boundary conditions, and thus the extended density distribution can be properly described. Furthermore, the GF on the complex energy plane and the level density can be used to identify the resonant states directly. </sec><sec> This paper briefly reviews the development of the GF method in the continuum density functional theory, focusing on its application in describing halo phenomena and single-particle resonance states in exotic nuclei. For the description of the halo phenomena, this review takes the neutron-rich Zr as examples to show the results given by Skyrme HFB calculations with the GF method. The extended neutron particle and pair density distributions can be more properly described by the GF method, compared to the box-discretized method. The self-consistency to deal with the pairing correlation and the continuum is important to determine the asymptotic neutron particle and pair density distributions. Furthermore, with the level density obtained by the loop integral of the GF on the complex energy plane, one can describe both the bound and resonant states on the same footing. Recently, in the relativistic mean field theory, a new method to obtain precise information for the resonant energy and width is proposed by using directly the poles of the GF on the complex energy plane. This helps explain the conservation and breaking of pseudospin symmetry in the nucleon single-particle levels. Additionally, Green’s function method is easily compatible with various theoretical models, thus in the future it can be further applied to describe the exotic nuclear collective resonances and nuclear reaction processes. </sec>

Efficient calculational techniques for reflection high energy electron diffraction (RHEED) are reported for surfaces with large periodic supercells. A fast Fourier transform approach enables the computer time scaling of a conventional RHEED calculation to be reduced to n2 log (n), where n is the number of diffracted beams used in the calculation. The special technique needed to implement this for arbitrary incident azimuths with symmetry optimization is detailed. A Green's function method is also introduced which is particularly suitable for calculations for imperfect surfaces. This combines the conventional approach to RHEED for dealing with substrate diffraction with a Green's function treatment for an imperfect surface of supercells and has n log (n) time scaling. Techniques for matching the results of the conventional and Green's function treatments at the substrate–surface interface are given. In addition, numerical procedures for solving the resulting equations are described and a selection of illustrative results is presented.

The elastic response of a vacancy in a semi-infinite fcc copper lattice containing a free surface is calculated by using a new multiscale Green's function method. The method treats the lattice distortion near the vacancy at the atomistic level and the free surface at the macroscopic continuum level in the same formalism. The lattice is modeled using the lattice statics Green's function that fully accounts for the discrete atomistic structure of the lattice and can model a large crystallite containing a million atoms without excessive CPU effort. The method is especially useful for modeling the elastic response of nanocrystals containing point defects in which surfaces and interfaces play important roles. The method bridges the length scales seamlessly by relating the microscopic lattice distortion near a point defect to measurable macroscopic parameters of the solid such as the strain and the displacement field at a free surface. Using the interatomic potential derived by Cleri and Rosato, the lattice distortion, relaxation energy, and relaxation volume due to a vacancy are calculated in an otherwise perfect copper lattice for a million-atom model containing a free (100) surface. The calculated value of the relaxation volume is in excellent agreement with the observed value. Numerical results are also presented for the strain and the displacement fields at the free surface due to a vacancy and the interaction energy between a vacancy and the free surface in anisotropic semi-infinite copper.

Simulations that are meant to determine the steady-state distribution of a diffusible solute such as oxygen in tissues have typically used finite difference methods to solve the diffusion equation. Finite difference methods require a tissue mesh with enough points to resolve oxygen gradients near and between discrete blood vessels. The large number of points that are typically required can make these calculations very slow. In this paper, we investigate a numerical method known as the Green's function method which is not bound by the same constraint. The Green's function method is expected to yield an accurate oxygen distribution more quickly by requiring fewer mesh points. Both methods were applied to calculate the steady state oxygen distribution in a model simulation region. When the Green's function calculation used meshes with 1/2, 1/4 and, 1/8 of the resolution required for the finite-difference mesh, there was good agreement with the finite difference calculation in all cases. When the volume of the domain was increased 8-fold the Green's function method was able to calculate the O2 field in 22 minutes, whereas the finite difference calculation is expected to take approximately 1 week. The number of steps required for the Green's function calculation increases quadratically with the number of points in the tissue mesh. As a result, small meshes are calculated very quickly using Green's functions, while for larger mesh sizes this method experiences a significant decrease in efficiency.

The Green’s function (GF) method, which makes use of GFs, is an important and elegant tool for solving a given boundary-value problem for the differential equation from a real engineering or physical field. Under a concentrated source, the solution of a differential equation is called a GF, which is singular at the source location, yet is very fundamental and powerful. When looking at the GFs from different physical and/or engineering fields, i.e. assigning the involved functions to real physical/engineering quantities, the GFs can be scaled and applied to large-scale problems such as those involved in Earth sciences as well as to nano-scale problems associated with quantum nanostructures. GFs are ubiquitous and everywhere: they can describe heat, water pressure, fluid flow potential, electromagnetic (EM) and gravitational potentials, and the surface tension of soap film. In the undergraduate courses Mechanics of Solids and Structural Analysis, a GF is the simple influence line or singular function. Dropping a pebble in the pond, it is the circular ripple traveling on and on. It is the wave generated by a moving ship in the opening ocean or the atom vibrating on a nanoscale sheet induced by the atomic force microscopy. In Earth science, while various GFs have been derived, a comprehensive review is missing. Thus, this article provides a relatively complete review on GFs for geophysics. In section , the George Green’s potential functions, GF definition, as well as related theorems and basic relations are briefly presented. In section , the boundary-value problems for elastic and viscoelastic materials are provided. Section is on the GFs in full- and half-spaces (planes). The GFs of concentrated forces and dislocations in horizontally layered half-spaces (planes) are derived in section in terms of both Cartesian and cylindrical systems of vector functions. The corresponding GFs in a self-gravitating and layered spherical Earth are presented in section in terms of the spherical system of vector functions. The singularity and infinity associated with GFs in layered systems are analyzed in section along with a brief review of various layer matrix methods. Various associated mathematical preliminaries are listed in appendix, along with the three sets of vector function systems. It should be further emphasized that, while this review is targeted at geophysics, most of the GFs and solution methods can be equally applied to other engineering and science fields. Actually, many GFs and solutions methods reviewed in this article are derived by engineers and scientists from allied fields besides geophysics. As such, the updated approaches of constructing and deriving the GFs reviewed here should be very beneficial to any reader.

Given a degenerating family of Riemann surfaces of genus g > 2, it is a natural and important question to study their spectral behaviors, in particular, Green's functions, during degeneration (see [9, 10, 17, 11, 13 and 5]). By the uniformization theorem, any Riemann surface of genus g > 2 has a canonical metric of constant curvature 1 . From now on, we call a Riemann surface with this canonical metric a hyperbolic surface, and a degenerating family of hyperbolic surfaces is denoted by St (l> 0). The degeneration of St is caused by the pinching of several simple, closed geodesics 71(l), ..., 7,,(1) on St, that is, the lengths of these geodesics go to zero as l--+0. Let At be the Laplacian of St, then the kernel of At consists of functions which are constant on each connected component of St, and At is invertible on the orthogonal complement of the kernel. The Schwartz kernel of this inverse is called the Green function of St, and denoted by G(z, w; t) for z, w~Sl. We would like to study the asymptotic behavior of G(z, w; I) as l ~O. In order to compare functions on S O and St, we use the harmonic map of infinite energy ~ : So --* S~\{71 (l) . . . . ,7,,(l)} constructed by Wolf El4]. Intuitively speaking, this map ~z opens up each node (a pair of cusps) of So into a simple closed pinching geodesic of SF. At first, we assume that there is only one pinching geodesic ~'(l)=7~(/) of length l on S~. If 7'([) separates St, then So has two components Se, ~ and So,z, of genus g~, gz respectively.

This paper presents a Green's function approach to the calculation of the pi-electronic states of molecules. The Green's functions are written in terms of the molecular orbital. Physical quantities are more easily treated in this approach than the previous method using the AO scheme. Formulae for the excitation and the ground state energies are obtained by RPA. Using the semi-empirical parameters, several examples are calculated. However, the well-known instability in RPA appears when the excitation and the ground state energies are calculated with the non-empirical parameters. From the viewpoint of diagramatic method, we discuss how this instability can be removed by introducing doubly excited configurations are renormalization of orbital energies.

The method of broadband Green's functions with low wavenumber extractions (BBGFL) is used to calculate Green's function for inhomogeneous waveguides filled with different dielectrics and with irregular boundaries. To construct the BBGFL modal solutions, we derive governing equations of the linear eigen-matrix problem and orthonormalization condition. In BBGFL, the Green's function is represented in modal expansions with convergence accelerated by higher order low wavenumber extractions. To obtain a linear eigenvalue problem for the modes, we use two BBGFLs of rectangular waveguides with two dielectric wavenumbers. The orthonormalized mode functions are used to construct the Green's function. Current wavenumber derivatives and Green's function wavenumber derivatives are computed by a single low wavenumber MoM impedance matrix. The wavenumber derivatives are used to accelerate the convergence of modal summations to 6th order. Numerical results are illustrated and compared with the direct MoM method of using free space Green's function. Results show accuracies and computation efficiencies for broadband simulations of Green's functions.

We discuss approximations of the Riemannian geometry near the horizon. If a (D + 1)-dimensional manifold has a bifurcate Killing horizon then we approximate by a product of the two-dimensional Rindler space and a (D − 1)-dimensional Riemannian manifold . We obtain approximate formulae for scalar Green's functions. We study the behavior of the Green's functions near the horizon and their dimensional reduction. We show that if is compact then the Green's function near the horizon can be approximated by the Green's function of the two-dimensional quantum field theory. The correction term is exponentially small away from the horizon. We extend the results to black brane solutions of supergravity in 10 and 11 dimensions. The near-horizon geometry can be approximated by . We discuss the Euclidean Green's functions on and their behavior near the horizon.

I compute the spectral functions of a 1D Holstein polaron moving in a modulated potential, to examine the effects of electron-phonon interactions on band gaps. The imaginary time Green function is computed using diagrammatic quantum Monte Carlo, which exactly sums the diagrammatic series. From the imaginary time Green function, spectral functions are computed. The electron-phonon interaction flattens the electronic dispersion and leads to an increase in the gap at momentum k = π/2. At strong coupling, polaron sidebands form in the gap. These results demonstrate the strong effect that electron-phonon interactions can have on band gaps.