We demonstrate that the configuration interaction (CI) approximation recaptures essential features of the exact (Bethe-ansatz) solution to the one-dimensional (1D) Hubbard model. As such, it provides a valuable route for describing effects that go beyond mean-field theory for strongly correlated electron systems in higher dimensions. The CI method systematically describes fluctuation and quantum tunneling corrections to the Hartree-Fock approximation (HFA). HFA predicts that doping a half-filled Hubbard chain leads to the appearance of charged spin polarons or charged domain-wall solitons in the antiferromagnetic background. The CI method, on the other hand, describes the quantum dynamics of these charged magnetic solitons and quantum tunneling effects between various mean-field configurations. In this paper, we test the accuracy of the CI method against the exact solution of the one-dimensional Hubbard model. We find remarkable agreement between the energy of the mobile charged bosonic domain wall (as given by the CI method) and the exact energy of the doping hole (as given by the Bethe ansatz) for the entire $U/t$ range. The CI method also leads to a clear demonstration of the spin-charge separation in one dimension. Addition of one doping hole to the half-filled antiferromagnetic chain results in the appearance of two different carriers: a charged bosonic domain wall (which carries the charge but no spin) and a neutral spin-1/2 domain wall (which carries the spin but no charge).
We demonstrate that the configuration interaction (CI) approximation recaptures essential features of the exact (Bethe-ansatz) solution to the one-dimensional (1D) Hubbard model. As such, it provides a valuable route for describing effects that go beyond mean-field theory for strongly correlated electron systems in higher dimensions. The CI method systematically describes fluctuation and quantum tunneling corrections to the Hartree-Fock approximation (HFA). HFA predicts that doping a half-filled Hubbard chain leads to the appearance of charged spin polarons or charged domain-wall solitons in the antiferromagnetic background. The CI method, on the other hand, describes the quantum dynamics of these charged magnetic solitons and quantum tunneling effects between various mean-field configurations. In this paper, we test the accuracy of the CI method against the exact solution of the one-dimensional Hubbard model. We find remarkable agreement between the energy of the mobile charged bosonic domain wall (as given by the CI method) and the exact energy of the doping hole (as given by the Bethe ansatz) for the entire U/t range. The CI method also leads to a clear demonstration of the spin-charge separation in one dimension. Addition of one doping hole to the half-filled antiferromagnetic chain results in the appearance of two different carriers: a charged bosonic domain wall (which carries the charge but no spin) and a neutral spin-1/2 domain wall (which carries the spin but no charge).
We demonstrate from first principles that when a charge carrier is added to a $\mathrm{spin}\ensuremath{-}1/2$ antiferromagnetic Mott insulator in either a one- or two-dimensional lattice, the self-consistent, Hartree-Fock ground state consists of a magnetic soliton texture with a doubly degenerate electronic level at the center of the Mott-Hubbard charge gap. This model is appropriate to systems with weak interchain or interlayer magnetic couplings in which long-range antiferromagnetic order is observed in the absence of charge carriers (doping). These magnetic solitons mediate the destruction of the magnetic order as the charge carrier concentration is increased. In a one-dimensional lattice with nearest-neighbor hopping t, on-site Coulomb repulsion U, and self-consistent, antiferromagnetic moment amplitude S, we find that a charged, fermionic, magnetic domain wall soliton with a weakly ferromagnetic core, centered between two sites, has lower Hartree-Fock energy than a corresponding charged quasiparticle in one of the Mott-Hubbard bands. However, for US/t>2, this soliton is unstable to the formation of a lower energy charged, bosonic domain wall soliton, centered on a single site. For US/t<2, both of the above solitons are charged bosons. The self-consistent structure of these solitons exhibits no rotation of the local magnetic moments, but only a local suppression of the local moment amplitude in the vicinity of the hole. In the absence of doping, charge neutral domain wall solitons exhibit spin rotation within their core region. The equilibrium core size $\ensuremath{\rho}$ is determined by the degree of magnetic anisotropy. The ferromagnetic core soliton exhibits a pair of nondegenerate near-midgap electronic states. The antiferromagnetic core soliton exhibits a pair of nondegenerate electronic states that are symmetric about the midgap energy and that merge into the continuum as the anisotropy effects are made small and the soliton core radius $\ensuremath{\rho}$ becomes very large. The two-dimensional antiferromagnetic Mott insulator exhibits analogous behavior to the one-dimensional model. This analogy is precise for a 2D antiferromagnet exhibiting spin flux. For the undoped Mott insulator, the ferromagnetic core meron vortex (``lotus flower'' configuration of local magnetic moments) exhibits a doubly degenerate electronic midgap state in the continuum model and is the analog of the 1D neutral domain wall soliton. We demonstrate that a hole added to the 2D system can form a charged bosonic collective excitation, in which the spin background around the hole forms a planar vortex with local antiferromagnetic correlations at infinity and vanishing local-moment amplitude at the core of the soliton.
This thesis examines the properties of a new microscopic model for non-Fermi-liquid behavior and d-wave pairing in a strongly correlated quasi-two-dimensional (2D) electron gas with purely repulsive interactions. Themicroscopic Hamiltonian for this system involves a nearest neighbor electron hopping matrix element t, an on-site Coulom b repulsion U , and a nearest neighbor Coulomb repulsion V. We suggest that nearest neighbor Coulomb repulsion on the energy scale of t stabilizes a state in which electrons undergo a �somersault� in their internal spin-space (spin-flux). Spin-flux is a new form of spontaneous symmetry breaking in a strongly correlated electron system in which the Hamiltonian acquires a term with the symmetry of spin-orbit coupling at the mean-field level. Spin-flux modifies the single quasi-particle dispersion relations from that of a conventional antiferro-magnet (AFM ). When this spin-1/2 A FM insulator is doped, the charge carriers are mobile, charged, bosonic meron-vortex solitons accompanied by unoccupied states deep inside the Mott-Hubbard charge-transfer gap. This model provides a unified microscopic basis for (i) non-Fermi-liquid transport properties for low and intermediate doping, (ii) mid-infrared optical absorption, (iii) destruction of A FM long range order with doping, (iv) angled resolved spectroscopy (ARPES), (v) d-wave preformed charge carrier pairs, an d (vi) a transition from the non-Fermi liquid state to a conventional Fermi-liquid metal for large doping. The approximations used to study the 2D spin-flux Hubbard model are the unrestricted Hartree-Fock Approximation and the Configuration Interaction (C l) Method. In ID, the C l approximation leads to excellent agreement with the exact Bethe Ansatz solution of the Hubbard model, as well as a clear demonstration of the spin-charge separation: the charge of the doping hole is carried by charged bosonic domain-walls, while the spin of the doping hole is carried by neutral fermionic domain-walls. In 2D, the C l method suggests a precursor to the spin-charge separation: addition of a single hole leads to the appearance of a charge-carrying vortex (the meron-vortex) bound to a spin-carrying antivortex.
We investigate the effect of hole doping on the strong-coupling Hubbard model at half filling in spatial dimensions D\ensuremath{\ge}1. We start with an antiferromagnetic mean-field description of the insulating state, and show that doping creates solitons in the antiferromagnetic background. In one dimension, the soliton is topological, spinless, and decoupled from the background antiferromagnetic fluctuations at low energies. In two dimensions and above, the soliton is nontopological, has spin quantum number 1/2, and is strongly coupled to the antiferromagnetic fluctuations. We derive the effective action governing the quasiparticle motion, study the properties of a single carrier, and comment on a possible description at finite concentration. \textcopyright{} 1996 The American Physical Society.
In this paper we numerically study the single-band Hubbard model on a two-dimensional square lattice using a mean-field iteration approach. We calculate the energies and configurations of finite-size systems for various doping densities up to 20%. The results show that there exists an attractive interaction among the holes, which forces the holes to form domain walls. The most stable configuration of domain-wall structure is determined, and its stability against transverse spin fluctuations is studied. When the nearest-neighbor repulsion V is considered, the system becomes richer in phases. The binding energy is reduced and vanishes completely at Vc=0.17t (for U/t=5, with t and U being the hopping integral and the on-site Hubbard repulsion, respectively). For small value (but larger than Vc) of V, we find that the holes form well-defined spin bags, which are separated from each other for a very small doping density. For a doping density larger than a critical value, they form domain walls. The physical implications of these results and their relation to experiment are discussed.
The internal structure of stripes in the two-dimensional Hubbard model is studied by going beyond the Hartree-Fock approximation. Partially filled stripes with inhomogeneous magnetization and charge along the stripe core, consistent with experimental observations, are stabilized by quantum fluctuations, included through the configuration interaction method. Similar configurations result to be stable for filled stripes, and have a lower energy than the standard paramagnetic core solution. Away from half-filling, solitonlike solutions show up within the Hartree-Fock approximation. The hopping of short regions of the stripes in the transverse direction is comparable to the bare hopping element. The hole-hole correlation shows maxima along the stripe direction, while the integrated value of ${n}_{\stackrel{\ensuremath{\rightarrow}}{k}}$ compares satisfactorily with experimental results.
We investigate a microscopic model for strongly correlated electrons with both on-site and nearest-neighbor Coulomb repulsion on a two-dimensional (2D) square lattice. This exhibits a state in which electrons undergo a ``somersault'' in their internal spin space (spin flux) as they traverse a closed loop in external coordinate space. When this spin-21 antiferromagnetic (AFM) insulator is doped, the ground state is a liquid of charged, bosonic meron vortices, which for topological reasons are created in vortex-antivortex pairs. The magnetic exchange energy of the distorted AFM background leads to a logarithmic vortex-antivortex attraction which overcomes the direct Coulomb repulsion between holes localized on the vortex cores. This leads to the appearance of preformed charged pairs. We use the configuration interaction (CI) method to study the quantum translational and rotational motion of various charged magnetic solitons and soliton pairs. The CI method systematically describes fluctuation and quantum tunneling corrections to the Hartree-Fock approximation. We find that the lowest-energy charged meron-antimeron pairs exhibit d-wave rotational symmetry, consistent with the symmetry of the cuprate superconducting order parameter. For a single hole in the 2D AFM plane, we find a precursor to spin-charge separation in which a conventional charged spin polaron dissociates into a singly charged meron-antimeron bound pair. This model provides a unified microscopic basis for (i) non-Fermi-liquid transport properties, (ii) d-wave preformed charged carrier pairs, (iii) midinfrared optical absorption, (iv) destruction of AFM long-range order with doping and other magnetic properties, and (v) certain aspects of angle-resolved photoemission spectroscopy.