d-symmetry superconductivity due to valence bond correlations

It is shown that d-symmetry superconductivity due to valence bond correlations is possible. Valence bond correlations are compatible with antiferromagnetic spin order. In order to explictly construct a homogeneous state with the valence bond structure in the two-dimensional Hubbard model for an arbitrary doping, we have used the variational method based on unitary local transformation. Attraction between holes in the d-channel is due to modulation of hopping by the site population in course of the valence bond formation, and corresponding parameters have been calculated variationally. An important factor for the gap width is the increase in the density of states on the Fermi level due to antiferromagnetic splitting of the band. The gap width and its ratio to the T _c are 2Δ≃0.1t and 2Δ/kT _c≃4.5−4 for U/t≃8. The correspondence between the theoretical phase diagram and experimental data is discussed. The dependence of T _c on the doping δ=|n−1| and the Fermi surface shape are highly sensitive to the weak interaction t′ leading to diagonal hoppings. In the case of t′>0 and p-doping, the peak on the curve of T _c(δ) occurs at the doping δ _opt, when the energy of the flattest part of the lower Hubbard subband crosses the Fermi level at k∼(π,0). In underdoped samples with δ<δ _opt, the anisotropic pseudogap in the normal state corresponds to the energy difference |E(π,0)−μ| between this part of the spectrum and the Fermi level. Zh. éksp. Teor. Fiz. 114, 985–1005 (September 1998)     

ARPES spectral functions and Fermi surface for La<Subscript>1.86</Subscript>Sr<Subscript>0.14</Subscript>CuO<Subscript>4</Subscript> compared with LDA + DMFT + Σ<Subscript>k</Subscript> calculations

The slightly underdoped high-temperature system La_1.86Sr_0.14CuO₄ (LSCO) is studied by means of high-energy high-resolution angular resolved photoemission spectroscopy (ARPES) and the combined LDA + DMFT + Σ _k computational scheme. The corresponding one-band Hubbard model is solved via dynamical mean field theory (DMFT), and the model parameters needed are obtained from first principles in the local density approximation (LDA). An “external” k-dependent self-energy Σ _k describes the interaction of correlated electrons with antiferromagnetic (AFM) pseudogap fluctuations. Experimental and theoretical data clearly show a “destruction” of the LSCO Fermi surface in the vicinity of the (π, 0) point and formation of “Fermi arcs” in the nodal directions. ARPES energy distribution curves as well as momentum distribution curves demonstrate a deviation of the quasiparticle band from the Fermi level around the (π, 0) point. The same behavior of spectral functions follows from theoretical calculations suggesting the AFM origin of the pseudogap state.     

Pseudogap behavior in Bi<Subscript>2</Subscript>Ca<Subscript>2</Subscript>SrCu<Subscript>2</Subscript>O<Subscript>8</Subscript>: Results of the generalized dynamical mean-field approach

Pseudogap phenomena are observed for the normal underdoped phase of different high-T _c cuprates. Among others, the Bi₂Sr₂CaCu₂O_{8 − δ} (Bi2212) compound is one of the most studied experimentally. To describe the pseudogap regime in Bi2212, we use a novel generalized ab initio LDA + DMFT + Σ_k hybrid scheme. This scheme is based on the strategy of one of the most powerful computational tools for real correlated materials: the local density approximation (LDA) + dynamical mean-field theory (DMFT). Conventional LDA + DMFT equations are here supplied with an additional (momentum-dependent) self-energy Σ_k in the spirit of our recently proposed DMFT + Σ_k approach taking into account pseudogap fluctuations. In the present model, Σ_k describes nonlocal correlations induced by short-range collective Heisenberg-like antiferromagnetic spin fluctuations. The effective single-impurity problem of the DMFT is solved by the numerical renormalization group (NRG) method. Material-specific model parameters for the effective x ² − y ² orbital of Cu-3d shell of the Bi2212 compound, e.g., the values of intra-and interlayer hopping integrals between different Cu sites, the local Coulomb interaction U, and the pseudogap potential Δ were obtained within the LDA and LDA + DMFT schemes. Here, we report on the theoretical LDA + DMFT + Σ_k quasiparticle band dispersion and damping, Fermi surface renormalization, momentum anisotropy of (quasi)static scattering, densities of states, spectral densities, and angular-resolved photoemission (ARPES) spectra, taking into account pseudogap and bilayer splitting effects for normal (slightly) underdoped Bi2212 (δ = 0.15). We show that LDA + DMFT + Σ_k successfully describes strong (pseudogap) scattering close to Brillouin zone boundaries. Our calculated LDA + DMFT + Σ_k Fermi surfaces and ARPES spectra in the presence of pseudogap fluctuations are almost insensitive to the bilayer splitting strength. However, our LDA-calculated value of bilayer splitting is rather small to describe the experimentally observed peak-dip-hump structure. The results obtained are in good semiquantitative agreement with various recent ARPES experiments. The article was submitted by the authors in English.     

Momentum dependence of pseudogap and superconducting gap in variation theory

To consider the origin of a pseudogap and a superconducting (SC) gap found in the high-T_c cuprates, we evaluated the momentum dependence of the singlet gap corresponding to the pseudogap and the SC gap in the t–J model, using an optimization variational Monte Carlo (VMC) method. In the underdoped regime, the singlet gap is significantly modified from the simple d_x²-y²(d)-wave gap (∝ cos k_x − cos k_y) by the contribution of long-range pairings. Its angular dependence at the quasi Fermi surface is qualitatively consistent with those experimentally observed in both hole and electron-doped cuprates. On the other hand, a SC gap is almost unchanged, preserving the original simple d-wave form. Thus, it seems that the incoherent part of the singlet gap mainly influences the forms of observed gaps.     

Superconductivity in the presence of fermion condensation

Fermion condensation (FC) is studied within the density functional theory. FC can fulfill the role of a boundary, separating the region of strongly interacting electron liquid from the region of strongly correlated electron liquid. Consideration of the superconductivity in the presence of FC shows that, under certain circumstances, at temperatures above T _c the superconductivity vanishes and the superconducting gap smoothly transforms into a pseudogap. The pseudogap occupies only a part of the Fermi surface, and one that shrinks with increasing temperature and vanishes at T=T*, and the single-particle excitations of the gapped area of the Fermi surface have a width γ ∼(T-T _c ). Pis’ma Zh. éksp. Teor. Fiz. 68, No. 6, 491–496 (25 September 1998) Published in English in the original Russian journal. Edited by Steve Torstveit.     

Doping-dependent electronic structure of cuprates studied using angle-scanned photoemission

Full k -maps of the electronic structure near the Fermi level of differently doped cuprates measured with angle-scanned photoelectron spectroscopy are presented. The valence band maximum of the antiferromagnetic insulator Sr₂CuO₂Cl₂, which is taken as a representative of an undoped cuprate, and the Fermi surfaces of overdoped, optimally doped and underdoped Bi₂Sr₂CaCu₂O_8+δ high-temperature superconductors are mapped in the normal state. The results confirm the existence of large Luttinger Fermi surfaces at high doping with a Fermi surface volume proportional to (1+x), where x is the hole concentration. At very low doping, however, we find that this assumption based on Luttinger's theorem is not fulfilled. This implies a change in the topology of the Fermi surface. Furthermore the intensity of the shadow bands observed on the Fermi surface of Bi₂Sr₂CaCu₂O_8+δ as a function of the doping is discussed. Received 12 October 1999 and Received in final form 12 April 2000     

Origin of “Hot Spots” in the pseudogap regime of Nd1.85Ce0.15CuO4: An LDA + DMFT + Σk study

The material-specific electronic band structure of the electron-doped high- T _c cuprate Nd_1.85Ce_0.15CuO₄ (NCCO) is calculated in the pseudogap regime using the recently developed generalized LDA + DMFT + Σ _k scheme. The LDA/DFT (density-functional theory within local density approximation) provides model parameters (hopping integral values and local Coulomb interaction strength) for the one-band Hubbard model, which is solved by the DMFT (dynamical mean-field theory). To take pseudogap fluctuations into account, the LDA + DMFT is supplied with an “external” k-dependent self-energy Σ _k that describes interaction of correlated conducting electrons with nonlocal Heisenberg-like antiferromagnetic (AFM) spin fluctuations responsible for the pseudogap formation. Within this LDA + DMFT + Σ _k approach, we demonstrate the formation of pronounced hot spots on the Fermi surface (FS) map in NCCO, opposite to our recent calculations for Bi₂Sr₂CaCu₂O_{8 − δ} (Bi2212), which have produced a rather extended region of the FS “destruction.” There are several physical reasons for this fact: (i) the hot spots in NCCO are located closer to the Brillouin zone center; (ii) the correlation length ξ of AFM fluctuations is longer for NCCO; (iii) the pseudogap potential Δ is stronger than in Bi2212. Comparison of our theoretical data with recent bulk-sensitive high-energy angle-resolved photoemission (ARPES) data for NCCO provides good semiquantitative agreement. Based on that comparison, an alternative explanation of the van Hove singularity at −0.3 eV is proposed. Optical conductivity for both Bi2212 and NCCO is also calculated within the LDA + DMFT + Δ _k scheme and is compared with experimental results, demonstrating satisfactory agreement. The text was submitted by the authors in English.     

Nearly antiferromagnetic Fermi liquids: a progress report

I describe recent theoretical and experimental progress in understanding the physical properties of the two dimensional nearly antiferromagnetic Fermi liquids (NAFL's) found in the normal state of the cuprate superconductors. In such NAFL's, the magnetic interaction between planar quasiparticles is strong and peaked at or near the commensurate wave vector, Q ≡ (fy fy). For the optimally doped and underdoped systems, the resulting strong antiferromagnetic correlations produce three distinct magnetic phases in the normal state: mean field above T_cr, pseudoscaling between T_cr and T_*, and pseudogap below T_*. I present arguments which suggest that the physical origin of the pseudogap found in the quasiparticle spectrum below Tcr is the formation of a precursor to a spin-densitywave- state, describe the calculations based on this scenario of the dynamical spin susceptibility, Fermi surface evolution, transport, and Hall effect, and summarize the experimental evidence in its support.     

Theory of dual fermion superconductivity in hole-doped cuprates

Since the discovery of the cuprate high-temperature superconductivity in 1986, a universal phase diagram has been constructed experimentally and numerous theoretical models have been proposed. However, there remains no consensus on the underlying physics thus far. Here, we theoretically investigate the phase diagram of hole-doped cuprates based on an itinerant-localized dual fermion model, with the charge carriers doped on the oxygen sites and localized holes on the copper d _{x2 ? y2} orbitals. We analytically demonstrate that the puzzling anomalous normal state or the strange metal could simply stem from a free Fermi gas of carriers bathing in copper antiferromagnetic spin fluctuations. The short-range high-energy spin excitations also act as the “magnetic glue” of carrier Cooper pairs and induce d-wave superconductivity from the underdoped to overdoped regime, distinctly different from the conventional low-frequency magnetic fluctuation mechanism. We further sketch out the characteristic dome-shaped critical temperature T _c versus doping level. The emergence of the pseudogap is ascribed to the localization of partial carriers coupled to the local copper moments or a crossover from the strange metal to a nodal Kondo-like insulator. Our work provides a consistent theoretical framework to understand the typical phase diagram of hole-doped cuprates and paves a distinct way to the studies of both non-Fermi liquid and unconventional superconductivity in strongly correlated systems.     

Effect of interlayer single-particle hoppings on the superconducting transition temperature

It is well known that the superconducting transition temperature of high-T _c cuprates depends on the number of CuO₂ planes in the unit cell. The multilayer structure implies the possibility of interlayer hopping. Under the assumption that the interlayer hopping can be specified by the parameter t _⊥(k) = t _⊥(cos(k _x ) − cos(k _y ))², the quasiparticle excitation spectrum for the bilayer cuprate in the superconducting state has been determined in the framework of the t − t′ − t″ − t _⊥ − J* model using the generalized mean-field approximation. It turns out that the interlayer hoppings does not create any additional mechanism of the Cooper paring and does not lead to an increase in T _c . The splitting of the upper Hubbard quasiparticle band attributed to the interlayer hoppings is manifested as two peaks in the doping dependence of the superconducting transition temperature at temperatures below the maximum T _c value for a single-layer cuprate. It has been found that antiferromagnetic interlayer correlations suppress the interlayer splitting. This probably leads to the common doping dependence of T _c for both single-layer and bilayer cuprates.     

Suppression of hole doping in fine-grained HTSC YBa<Subscript>2</Subscript>Cu<Subscript>3</Subscript>O<Subscript><Emphasis Type="Italic">y</Emphasis></Subscript> due to structural disordering

It was proven experimentally that the structural disordering inherent to fine-grained high-temper- ature YBa₂Cu₃O _y superconductors (with an average grain size of 〈D〉 < 2 μm) leads to a reduction of the level of hole doping and the creation of features inherent to the pseudogap state (antiferromagnetic correlations and the lowered density of states at the Fermi level) even in samples with optimum oxygen content y ≈ 6.92.     

Reconstruction of the Fermi surface in the pseudogap state of cuprates

Reconstruction of the Fermi surface of high-temperature superconducting cuprates in the pseudogap state is analyzed within a nearly exactly solvable model of the pseudogap state, induced by short-range order fluctuations of the antiferromagnetic (AFM), spin-density wave (SDW), or a similar charge-density wave (CDW) order parameter, competing with the superconductivity. We explicitly demonstrate the evolution from “Fermi arcs” (on the “large” Fermi surface) observed in the ARPES experiments at relatively high temperatures (when both the amplitude and phase of the density waves fluctuate randomly) towards the formation of typical “small” electron and hole “pockets,” which are apparently observed in the de Haas-van Alphen and Hall resistance oscillation experiments at low temperatures (when only the phase of the density waves fluctuate and the correlation length of the short-range order is large enough). A qualitative criterion for the quantum oscillations in high magnetic fields to be observable in the pseudogap state is formulated in terms of the cyclotron frequency, the correlation length of fluctuations, and the Fermi velocity. The text was submitted by the authors in English.     

The non-Fermi-liquid nature of the metallic states of the Hubbard Hamiltonian

It has long since been argued that the metallic states of the single-band Hubbard Hamiltonian ? in two spatial dimensions (i.e. for d = 2) should be non-Fermi liquid, a possibility that would lead to the understanding of the observed anomalous behaviour of the doped copper-oxide-based superconducting compounds in their normal metallic states. Here we present a formalism which enables us to express, for arbitrary d, the behaviour of the momentum-distribution function nσ(k) pertaining to uniform metallic ground states of ? close to S _F; σ (the Fermi surface of the fermions with spin index σ, σ = ↑, ↓) in terms of a small number of constant parameters which are bound to satisfy certain inequalities implied by the requirement of the stability of the ground state of the system. These inequalities restrict the range of variation in nσ(k) for k infinitesimally inside and outside the Fermi sea pertaining to fermions with spin index σ and consequently the range of variation in the zero-temperature limit of nσ(k) for k on S _F; σ On the basis of some available accurate numerical results for nσ(k) pertaining to the Hubbard and the t-J Hamiltonian, we conclude that, at least in the strong-coupling regime, the metallic ground states of ? for d = 2 cannot be Fermi-liquid nor can they in general be purely Luttinger or marginal Fermi liquids. We further rigorously identify the pseudogap phenomenon, or 'truncation' of the Fermi surface, clearly observed in the normal states of underdoped copper-oxide-based superconductors, as corresponding to a line of resonance energies (i.e. these energies strictly do not relate to quasiparticles) located below the Fermi energy, with a concomitant suppression to zero of the jump in nσ(k) over the 'truncated' parts of the Fermi surface. Our analyses make explicit the singular significance of the non-interacting energy dispersion ε _k underlying ? in determining the low-energy spectral and transport properties of the metallic ground states of ?.     

Magnetization of optically polarized gas atoms by an optical pulse train

The properties of the density matrix and the multipole moments arising in oriented and aligned atoms with zero nuclear spin through the interaction with strong resonant ultrashort pulses with wave vector k ₀ and circular or linear polarization have been found. Calculations have been made for the time-dependent light-induced magnetization μ(t′) of a gas of pre-oriented and prealigned atoms following the passage of a weak resonant elliptically polarized pulse with frequency ω and wave vector k collinear with k ₀. It is shown that for oriented atoms, μ(t′) is an even function of the detuning from resonance, ω-ω _ba, and can be split into two terms whose directions are a consequence of symmetry and are determined by the vectors k ₀ and k as well as by the direction of rotation of the electric fields corresponding to the pulses. For aligned atoms the vector μ(t′) is collinear with k, and the first term is an even function of ω-ω _ba. However, the second term is an odd function of ω-ω _ba and reverses direction when the sign of ω-ω _ba changes, as well as when the orientation of the axes of the polarization ellipse is changed. It is shown that if a series of weak linearly polarized pulses pass through the gas, the light-induced magnetization of the oriented and aligned gas atoms can be decomposed into three factors: the first determines the direction and is a consequence of the symmetry; the second (with the dimensions of magnetic moment) depends on the characteristics of the resonant transitions; and the third is a universal function of t′ and ω-ω _ba that does not depend on the underlying characteristics of the resonant transition. These vector factors and the universal functions are in principle different for oriented and aligned atoms. Zh. éksp. Teor. Fiz. 111, 63–92 (January 1997)     

Anomalous thermodynamics of the doped Mott-Hubbard insulators

The concentration dependence of the entropy of doped Mott-Hubbard insulators has been considered within the t-J model. It has been shown that a change in the type and statistics of charge carriers as compared to the Fermi gas leads to a radical change in the entropy s, in particular, to the giant growth of the entropy upon doping. The quantity ∂s/∂x ≈ k _B is approximately consistent with the experimental data for HTSC cuprates in the pseudogap phase.     

Superconductivity in a simple model of the pseudogap state

An analysis is made of characteristics of the superconducting state (s-and d-pairing) using a simple, exactly solvable model of the pseudogap state produced by fluctuations of the short-range order (such as antiferromagnetic) based on a Fermi surface model with “hot” sections. It is shown that the superconducting gap averaged over these fluctuations is nonzero at temperatures higher than the mean-field superconducting transition temperature T _c over the entire sample. At temperatures T > T _c superconductivity evidently exists in isolated sections (“ drops”). Studies are made of the spectral density and the density of states in which superconducting characteristics exist in the range T > T _c however, in this sense the temperature T = T _c itself is no different in any way. These anomalies show qualitative agreement with various experiments using underdoped high-temperature superconducting cuprates.     

Effect of next-nearest-neighbour interaction on dx2?y2-wave superconducting phase in 2D t-J model

An exact diagonalization calculation of the t-J model on 2D square cluster has been studied for the ground state properties of HTSC. Effect of next-nearest-neighbour hopping and magnetic (both antiferromagnetic and ferromagnetic) interaction on d _x ²−y ²-wave pairing has been shown. Relative strength of the next-nearest-neighbour interaction with respect to that of near-neighbour interaction for the strongest d _x ²−y ²-wave pairing has been estimated. A schematic phase diagram is shown. It is shown that a two-sublattice model with antiferromagnetic interaction between them and a small intra-ferromagnetictype interaction in one sublattice favours d _x ²−y ²-wave superconductivity and moderate negative type NNN hopping adds flavours to this phase.     

Spin-orbit coupling in the superconducting phase and DDW states of high-T<Subscript>c</Subscript> cuprates

The effects of the spin-orbit coupling are considered for the high T _c cuprates with asymmetric superconducting gap (SC) and the d-density wave (DDW) phase due to its vital role in the experimental determination of the DDW state. Experiments predict an anisotropy in the DSC gap where |Δ(0,π)|>|Δ(π,0)| and the gap node deviates from the diagonal direction towards the k _x axis. Measurements also demonstrate DDW to be a possible candidate for the pseudogap in the underdoped phase. Due to the spin-orbit (SO) coupling in the low temperature orthorhombic (LTO) phase, the phase diagram of the cuprates suffers a change due to the modification of the T* value, the temperature characteristic of pseudogap, although T _c remains unaltered. Moreover, for a more generalized SO coupling, the DDW gap decreases with the angle but has no effect on the SC gap. We calculate the density of states in the various regimes of doping for the mixed SC+DDW states in the underdoped (UD) phase, SC state in the overdoped phase and also the DDW state in the UD phase and compare them with various theoretical and experimental works. The temperature dependence of the specific heat does not exhibit any qualitative change due to the SO coupling.     

Unified model of pseudogap features of conductivity in HTSCs

The pseudogap phenomenon in underdoped and optimally oxygen-doped high-temperature superconductors (HTSCs) of the Y₁Ba₂Cu₃O_x system is explained from a unified point of view within the model of negative U centers. It is shown that the pseudogap features of conductivity are not related directly to the superconductivity but arise due to the existence of statistical interaction of negative U centers with valence-band holes. Specifically due to this interaction, the hole density in the valence band does not remain constant. It differently changes with temperature for different mutual positions of the Fermi level and the valence band top. These differences lead to different temperature dependences of conductivity for underdoped and optimally doped samples.     

Extended -wave superconductivity. Flat nodes in the gap and the low-temperature asymptotic properties of high- superconductors

Remarkable anisotropic structures have been recently observed in the order parameter of the underdoped superconductor Bi₂Sr₂CaCu₂O . Such findings are strongly suggestive of deviations from a simple d _{x2 - y2} -wave picture of high- superconductivity, i.e. . In particular, flatter nodes in are observed along the directions in -space, than within this simple model for a d-wave gap. We argue that nonlinear corrections in the -dependence of near the nodes introduce new energy scales, which would lead to deviations in the predicted power-law asymptotic behaviour of several measurable quantities, at low or intermediate temperatures. We evaluate such deviations, either analytically or numerically, within the interlayer pair-tunneling model, and within yet another phenomenological model for a d-wave order parameter. We find that such deviations are expected to be of different sign in the two cases. Moreover, the doping dependence of the flatness of the gap near the nodes is also attributable to Fermi surface effects, in addition to possible screening effects modifying the in-plane pairing kernel, as recently proposed. Received 19 November 1999