Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates

We study the doping evolution of the electronic structure in the normal phase of high-Tc cuprates. The electronic structure and the Fermi surface of cuprates with a single CuO2 layer in the unit cell like La2-xSrxCuO4 have been calculated by the LDA+GTB method in the regime of strong electron correl...

Full description

Saved in:
Bibliographic Details
Date:2010
Main Authors: Ovchinnikov, S.G., Korshunov, M.M., Shneyder, E.I.
Format: Article
Language:English
Published: Відділення фізики і астрономії НАН України 2010
Subjects:
Online Access:https://nasplib.isofts.kiev.ua/handle/123456789/13286
Tags: Add Tag
No Tags, Be the first to tag this record!
Journal Title:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Cite this:Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates / S.G. Ovchinnikov, M.M. Korshunov, E.I. Shneyder // Укр. фіз. журн. — 2010. — Т. 55, № 1. — С. 55-64. — Бібліогр.: 74 назв. — англ.

Institution

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id nasplib_isofts_kiev_ua-123456789-13286
record_format dspace
spelling Ovchinnikov, S.G.
Korshunov, M.M.
Shneyder, E.I.
2010-11-04T10:00:03Z
2010-11-04T10:00:03Z
2010
Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates / S.G. Ovchinnikov, M.M. Korshunov, E.I. Shneyder // Укр. фіз. журн. — 2010. — Т. 55, № 1. — С. 55-64. — Бібліогр.: 74 назв. — англ.
2071-0194
PACS 71.27.+a,71.18.+y, 74.25.Jb
https://nasplib.isofts.kiev.ua/handle/123456789/13286
We study the doping evolution of the electronic structure in the normal phase of high-Tc cuprates. The electronic structure and the Fermi surface of cuprates with a single CuO2 layer in the unit cell like La2-xSrxCuO4 have been calculated by the LDA+GTB method in the regime of strong electron correlations (SEC) and compared to ARPES and quantum oscillations data. We have found two critical concentrations, xc1 and xc2, where the Fermi surface topology changes. Following I.M. Lifshitz’s ideas of the quantum phase transitions (QPT) of the 2:5-order, we discuss the concentration dependence of the low-temperature thermodynamics. The behavior of the electronic specific heat δ(C/T ) ~ (x - xc)^½ is similar to the Loram and Cooper experimental data in the vicinity of xc1 ≈ 0.15. In the superconducting state of cuprates, we consider both magnetic and phonon contributions to the d-wave pairing and found that there is no dominant mechanism of superconductivity. Magnetic and phonon contributions to the critical temperature are of the same order.
У цiй работi обговорено змiни електронної структури в нормальнiй фазi високотемпературних надпровiдникiв – шаруватих купратiв. Результати розрахункiв електронної структури та поверхнi Фермi одношарових купратiв методом LDA+GTB iз врахуванням сильних кореляцiй порiвнюються з даними ARPES та квантових осциляцiй. Виявлено двi критичнi точки xc1 та xc2, в яких вiдбувається перебудова поверхнi Фермi. В околi критичних точок у межах iдеологiї I.М. Лiфшица про квантовi фазовi переходи 2,5 роду знайдено змiни термодинамiчних властивостей за низьких температур. Особливiсть електронної теплоємностi δ(C/T ) ~ (x - xc)^½ достатньо добре узгоджується з вiдомими експериментальними даними в околi xc1 ≈ 0.15. Якiсно обговорюються змiни знака константи Холла з допуванням. Також розглянуто надпровiдний стан з урахуванням магнiтного i фононного механiзмiв спарювання.
We thank A. Kordyuk for the discussion of the results and T.M. Ovchinnikova for the technical assistance. This work was supported by project 5.7 of the programm “Quantum physics of the condensed matter” of the Presidium.
en
Відділення фізики і астрономії НАН України
Тверде тіло
Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates
Вплив квантових фазових переходів Ліфшица на нормальний та надпровідний стани купратів
Article
published earlier
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
title Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates
spellingShingle Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates
Ovchinnikov, S.G.
Korshunov, M.M.
Shneyder, E.I.
Тверде тіло
title_short Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates
title_full Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates
title_fullStr Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates
title_full_unstemmed Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates
title_sort effect of lifshitz quantum phase transitions on the normal and superconducting states in cuprates
author Ovchinnikov, S.G.
Korshunov, M.M.
Shneyder, E.I.
author_facet Ovchinnikov, S.G.
Korshunov, M.M.
Shneyder, E.I.
topic Тверде тіло
topic_facet Тверде тіло
publishDate 2010
language English
publisher Відділення фізики і астрономії НАН України
format Article
title_alt Вплив квантових фазових переходів Ліфшица на нормальний та надпровідний стани купратів
description We study the doping evolution of the electronic structure in the normal phase of high-Tc cuprates. The electronic structure and the Fermi surface of cuprates with a single CuO2 layer in the unit cell like La2-xSrxCuO4 have been calculated by the LDA+GTB method in the regime of strong electron correlations (SEC) and compared to ARPES and quantum oscillations data. We have found two critical concentrations, xc1 and xc2, where the Fermi surface topology changes. Following I.M. Lifshitz’s ideas of the quantum phase transitions (QPT) of the 2:5-order, we discuss the concentration dependence of the low-temperature thermodynamics. The behavior of the electronic specific heat δ(C/T ) ~ (x - xc)^½ is similar to the Loram and Cooper experimental data in the vicinity of xc1 ≈ 0.15. In the superconducting state of cuprates, we consider both magnetic and phonon contributions to the d-wave pairing and found that there is no dominant mechanism of superconductivity. Magnetic and phonon contributions to the critical temperature are of the same order. У цiй работi обговорено змiни електронної структури в нормальнiй фазi високотемпературних надпровiдникiв – шаруватих купратiв. Результати розрахункiв електронної структури та поверхнi Фермi одношарових купратiв методом LDA+GTB iз врахуванням сильних кореляцiй порiвнюються з даними ARPES та квантових осциляцiй. Виявлено двi критичнi точки xc1 та xc2, в яких вiдбувається перебудова поверхнi Фермi. В околi критичних точок у межах iдеологiї I.М. Лiфшица про квантовi фазовi переходи 2,5 роду знайдено змiни термодинамiчних властивостей за низьких температур. Особливiсть електронної теплоємностi δ(C/T ) ~ (x - xc)^½ достатньо добре узгоджується з вiдомими експериментальними даними в околi xc1 ≈ 0.15. Якiсно обговорюються змiни знака константи Холла з допуванням. Також розглянуто надпровiдний стан з урахуванням магнiтного i фононного механiзмiв спарювання.
issn 2071-0194
url https://nasplib.isofts.kiev.ua/handle/123456789/13286
citation_txt Effect of Lifshitz Quantum Phase Transitions on the Normal and Superconducting States in Cuprates / S.G. Ovchinnikov, M.M. Korshunov, E.I. Shneyder // Укр. фіз. журн. — 2010. — Т. 55, № 1. — С. 55-64. — Бібліогр.: 74 назв. — англ.
work_keys_str_mv AT ovchinnikovsg effectoflifshitzquantumphasetransitionsonthenormalandsuperconductingstatesincuprates
AT korshunovmm effectoflifshitzquantumphasetransitionsonthenormalandsuperconductingstatesincuprates
AT shneyderei effectoflifshitzquantumphasetransitionsonthenormalandsuperconductingstatesincuprates
AT ovchinnikovsg vplivkvantovihfazovihperehodívlífšicananormalʹniitanadprovídniistanikupratív
AT korshunovmm vplivkvantovihfazovihperehodívlífšicananormalʹniitanadprovídniistanikupratív
AT shneyderei vplivkvantovihfazovihperehodívlífšicananormalʹniitanadprovídniistanikupratív
first_indexed 2025-11-25T10:11:02Z
last_indexed 2025-11-25T10:11:02Z
_version_ 1850509844235157504
fulltext EFFECT OF LIFSHITZ QUANTUM PHASE TRANSITIONS EFFECT OF LIFSHITZ QUANTUM PHASE TRANSITIONS ON THE NORMAL AND SUPERCONDUCTING STATES IN CUPRATES S.G. OVCHINNIKOV,1, 2 M.M. KORSHUNOV,1, 3, 4 E.I. SHNEYDER1, 5 1L.V. Kirensky Institute of Physics, Siberian Branch of Russian Academy of Sciences (Krasnoyarsk 660036, Russia; e-mail: sgo@ iph. krasn. ru ) 2Siberian Federal University (Krasnoyarsk 660041, Russia) 3Max-Planck-Institut für Physik komplexer Systeme (D-01187 Dresden, Germany) 4Department of Physics, University of Florida (Gainesville, Florida 32611, USA) 5Reshetnev Siberian State Aerospace University (Akademgorodok 50, Krasnoyarsk 660036, Russia) PACS 71.27.+a,71.18.+y, 74.25.Jb c©2010 We study the doping evolution of the electronic structure in the normal phase of high-Tc cuprates. The electronic structure and the Fermi surface of cuprates with a single CuO2 layer in the unit cell like La2−xSrxCuO4 have been calculated by the LDA+GTB method in the regime of strong electron correlations (SEC) and compared to ARPES and quantum oscillations data. We have found two critical concentrations, xc1 and xc2, where the Fermi surface topology changes. Following I.M. Lifshitz’s ideas of the quantum phase transitions (QPT) of the 2.5-order, we discuss the concentration dependence of the low-temperature thermody- namics. The behavior of the electronic specific heat δ(C/T ) ∼ (x− xc) 1/2 is similar to the Loram and Cooper experimental data in the vicinity of xc1 ≈ 0.15. In the superconducting state of cuprates, we consider both magnetic and phonon contributions to the d-wave pairing and found that there is no dominant mecha- nism of superconductivity. Magnetic and phonon contributions to the critical temperature are of the same order. 1. Introduction Nowadays, high-Tc cuprates is the second most studied class of condensed matter after semiconductors. Both the nature of superconductivity and the abnormal pseu- dogap feature of the normal phase are not clear yet [1–9]. A lot of experimental data on the electronic structure have been obtained by ARPES that reveals the dop- ing evolution of the Fermi surface (FS) from small arcs near (π/2, π/2) at a small doping to the large FS around (π, π) at a large doping [10]. Quantum oscillations mea- surements in strong magnetic fields on the single crystals YBa2Cu3O6.5 [11] and YBa2Cu4O8 [12] have proved the existence of small hole pockets in the underdoped (UD) cuprates that looks as a contradiction to the ARPES arcs. This contradiction has been explained by the inter- action between holes and spin fluctuations in the pseu- dogap state with the existing short-range antiferromag- netic (AFM) order [13–16]. It occurs that a part of the hole pocket related to the shadow band has a smaller quasiparticle (QP) lifetime due to the QP scattering on spin fluctuations. Recently, the VUV laser ARPES [17] has found a closed FS pocket in the UD La-Bi2201 with a small intensity at the shadow band part. The strong interaction of electrons with spin fluctuations is a gen- eral property of SEC systems and takes place not only in cuprates but also, e.g., in manganates [18]. The conventional LDA (local density approximation) approach to the electronic structure in the regime of SEC fails. Various realistic multiband models of a CuO2 layer in cuprates in the low-energy region result in the effec- tive Hubbard and t − J models [19–23]. In the hybrid LDA+GTB scheme [24] that combines the LDA calcula- tions of the multiband p − d model parameters and the generalized tight-binding (GTB) treatment of SEC, the low-energy effective t− t′ − t′′ − J∗ model has been ob- tained from a microscopic approach with all parameters being calculated ab initio. Small hole pockets in the UD case with area ∼ x ap- pear in a theory considering the hole dynamics in the AFM spin background and have been obtained by the ex- act diagonalization [25] and the quantum Monte Carlo studies of finite clusters [26, 27], as well as by various variational and perturbation calculations for the infinite- ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1 55 S.G. OVCHINNIKOV, M.M. KORSHUNOV, E.I. SHNEYDER dimensional lattice [28–32]. Once the long-range AFM order disappears with doping, the electronic structure calculations in the paramagnetic phase results in the dispersion of the valence band with the top at (π, π) and the large FS [33]. Still there are apologists of the “universal metal dispersion” calculating the LDA band structure and the FS and claiming the rigid band be- havior with a Fermi level shift of the fixed band disper- sion [34]. After the small hole pockets were discovered in the Landau oscillations experiments [11, 12], the rigid band scenario becomes evidently unconvincing. In place of the conventional Fermi liquid state of a normal metal, the pseudogap state appears in the phase diagram of cuprates beside the long-range AFM phase. Though the origin of the pseudogap state is still debated, the con- tribution of the fluctuating short-range AFM order is clear [5]. The short-range AFM order is essential not only in the UD region. Even at the optimal doping, the AFM correlation length ξAFM ≈ 10 Å [35]. At low tem- peratures T ≤ 10 K, spin fluctuations are slow with a typical time scale of 10−9 s and on the spatial scale of ξAFM (size of the AFM microdomain) [36]. This time is large in comparison with the fast electronic lifetime in ARPES (∼ 10−13 s) [37] and the cyclotron period T ∼ 2πω−1 c ∼ 10−12, ωc being a cyclotron frequency in quantum oscillations experiments [11, 12]. Thus, we safely consider that the spin fluctuations are frozen at low T and take only the spatial dependence of the short- range AFM order into account. This means that the electronic self-energy Σ (k, ω) will depends only on the momentum, Σ (k, ω)→ Σ(k). We use this approach to study the concentration de- pendence of the electronic structure and the FS. In Sec- tion 2, we present the electronic structure and the change of the FS topology within the t− t′− t′′−J∗ model. The FS area and the Luttinger theorem are also discussed. In Section 3, we give the qualitative picture based on the interaction between hole and spin fluctuations. In Sec- tion 4, we use the Lifshitz ideas [38, 39] on the QPT to study the low-temperature thermodynamics. The elec- tronic specific heat singularity near a QPT is compared with the experimental data [40]. In Section 5, we extend the same approach to the superconducting d-wave state. We found that both magnetic and phonon contributions to Tc are similar in magnitude. 2. The Fermi Surface of La2−xSrxCuO4 and Its Doping Evolution Within the LDA+GTB approach, we start from the ab initio LDA calculations and construct the Wannier func- tions in the basis of oxygen p-orbitals and copper eg- orbitals. The multiband p−d model [41] parameters are calculated ab initio. Then we apply the cluster pertur- bation approach [19, 42] and introduce the Hubbard X- operators constructed within the full set of eigenstates of the unit cell (a CuO6 cluster) that is obtained by the exact diagonalization of the multiband p − d model Hamiltonian of the cluster. By the GTB method, we construct the low-energy effective Hubbard model with U = ECT, where ECT is the charge transfer gap [43]. In the Hubbard model, the X0σ f operator describes the hole annihilation at the site f in the lower Hubbard band (LHB) of holes that corresponds to an electron at the bottom of the conductivity band. (Here, f enumerates the CuO6 unit cells.) The hole annihilation in the upper Hubbard band (UHB) is given by the X σ̄2 f operator and corresponds to an electron at the top of the valence band. In the limit of SEC, Ueff � t (t is the effective intersite hopping), we may exclude either the two-hole state |2〉, obtain the effective Hamiltonian for LHB describing the electron-doped cuprates or two-electron state |0〉 (hole vacuum d10p6), and get the effective Hamiltonian for UHB. The latter case is interesting for the hole-doped cuprates. We emphasize that the effective t− t′− t′′−J∗ model is derived from the microscopic approach and its parameters are calculated ab initio. Here, J∗ means that we take into account not only the superexchange inter- action of localized spins J but also the 3-site correlated hopping that is of the same order as J and has to be included in the theory [45]. The model Hamiltonian is given by Ht−J∗ = Ht−J +H(3), Ht−J = ∑ f,σ [ (ε− µ)Xσσ f + (ε2 − 2µ)X22 f ] + + ∑ f 6=g,σ t11fgX 2σ̄ f X σ̄2 g + ∑ f 6=g Jfg ( Sf · Sg − 1 4 nfng ) H(3) = ∑ f 6=m6=g,σ t01fmt 01 mg Ueff ( Xσ2 f X σ̄σ m X2σ̄ g −X σ̄2 f Xσσ m X2σ̄ g ) . Here, Jfg = 2 ( t01fg )2 /Ueff , t01fg is the interband (LHB↔ UHB) hopping parameter between two sites f and g, Sf is the spin operator, ε and ε2 are one- and two-hole local energies, and µ is the chemical potential. The intraband 56 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1 EFFECT OF LIFSHITZ QUANTUM PHASE TRANSITIONS hopping parameters t11fg have been calculated up to the 6- th nearest neighbors. It appears that only 3 coordination spheres are important. The dispersion with hoppings only to the nearest neighbors and to the next-nearest neighbors is qualitatively different from the dispersion calculated for 3 coordination speres. The contribution of the more distant neighbors to the hole dispersion is negligible. The parameters calculated ab initio within the model for La2−xSrxCuO4 are (in eV): t = 0.932, t′ = −012, t′′ = 0.152, J = 0.298, J ′ = 0.003, J ′′ = 0.007. We introduce the hole Green function in the UHB (here, σ̄ ≡ −σ) Gσ(k, E) = 〈〈 X σ̄2 k ∣∣X2σ̄ k 〉〉 E . (1) The analysis of the whole set of diagrams in the X- operators diagram technique results in the exact gen- eralized Dyson equation [44] Gσ(k, E) = Pσ(k, E) E − ε0 + µ− Pσ(k, E)tk − Σσ(k, E) . Here, tk is the Fourier transform of the hopping, Pσ(k, E) and Σσ(k, E) are the strength and the self- energy operators. In the simplest Hubbard I approxi- mation Σσ = 0, Pσ = Fσ̄2 = 〈X σ̄σ̄ f 〉 + 〈X22 f 〉. The QP spectral weight is determined by the filling factor Fσ̄2. In the diagram technique, Fσ̄2 corresponds to the so-called “terminal factors” [46]. To incorporate the effect of the short-range AFM or- der on the QP dynamics, we go beyond the Hubbard I approximation. The calculation scheme is given in [47]. We use the Mori-type method to project the higher or- der Green functions to the single particle function (1). A similar approach that took the spin dynamics into ac- count was used in [23, 48]. The hole concentration in La2−xSrxCuO4 (LSCO) per unit cell is nh = 1 + x. The completeness condition for the local Hilbert space in the t− J model is∑ σ Xσσ f +X22 f = 1. Thus, we easily obtain 〈Xσσ f 〉 = (1 − x)/2, 〈X22 f 〉 = x, and Fσ̄2 = (1 + x)/2. The Green function (1) becomes Gσ(k, E)= (1 + x)/2 E−ε0+µ− 1+x 2 tk− 1−x2 4 (t01k )2 Ueff −Σ(k) , (2) and the self-energy is given by Σ (k) = 2 1 + x 1 N ∑ q { K (q)× × ( tq− 1−x 2 Jk−q−x ( t01q )2 Ueff − (1+x) t01k t 01 q Ueff ) + 3 2 C (q)× × ( tk−q − 1− x 2 ( Jq − (t01k−q)2 Ueff ) − (1 + x)t01k t 01 k−q Ueff )} . Here, K(q) and C(q) stand for the Fourier transforms of the static kinetic and spin correlation functions, K(q) = ∑ f−g e−i(f−g)q 〈 X2σ̄ f X σ̄2 g 〉 , C(q) = ∑ f−g e−i(f−g)q 〈 Xσσ̄ f X σ̄σ g 〉 = = 2 ∑ f−g e−i(f−g)q 〈 Szf S z g 〉 . (3) For the LHB which corresponds to the electron-doped cuprates, the similar Green function has been obtained previously [49]. We assume that the spin system is an isotropic spin liquid with any averaged component of the spin being zero and the equal correlation functions for any component of the spin, 〈Sαf Sαg 〉, α = x, y, z. We calculate this correlation function following [49] by the method developed previously for the Heisenberg model [50, 51]. The resulting static magnetic suscep- tibility agrees with the other calculations for the t − J model [52, 53]. As for the kinetic correlation function, it is expressed via the same Green function (1). The self-consistent treatment of the electronic and spin systems results in the evolution of the correlation functions (3), the chemical potential, and the FS as a function of the doping (Fig. 1). At a small doping, we get 4 hole pockets close to the (π/2, π/2) point, as was expected for the AFM state. At the critical con- centration xc1 ≈ 0.15, the connection of these pockets appears along the (π, 0)− (π, π) line, and the FS topol- ogy changes. At xc1 < x < xc2 ≈ 0.24, we obtain two FS centered around the (π, π) point. The smaller one is the electronic FS; it shrinks with doping and collapses, when x → xc2. The larger one is the hole FS; with in- creasing x, it becomes more rounded. At x > xc2, only a ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1 57 S.G. OVCHINNIKOV, M.M. KORSHUNOV, E.I. SHNEYDER Fig. 1. Calculated Fermi surface for a single-layer cuprate for different doping levels x. Fermi surface topology changes at xc1 = 0.15 and xc2 = 0.24. ARPES data from [57] and [17] are shown in the lower left and lower right corners of the Brillouin zone, respectively large hole FS remains. Finally, there is one more change of the topology at x = xc3, when the FS touches the (π, 0) point and becomes of the electronic type centered around the (0, 0) point. Note that the values of critical concentrations are ob- tained with a finite accuracy. First of all, the model parameters are deduced by a complicated procedure in- volving the projection of the LDA wave functions into the Wannier function basis and may vary with a change of this basis. Second, the equation for the Green func- tion (2) is approximate and, with regard for higher order corrections, can change quantitatively values of the critical concentrations. On the other hand, the qualitative picture should remain unchanged, since it is due to the general properties of the electron scat- tering by AFM fluctuations. A qualitatively similar transformation of the FS with doping has been ob- tained within the Hubbard model in the regime of SEC (Fig. 15 in [48]), in the spin-density wave state of the Hubbard model [54], within the spin-fermion model [55], and in the ab initio multielectron quan- tum chemical approach [56]. The qualitative agree- ment of our results and the results of calculations in different approximations [48, 54–56] is basically due to the common underlying idea: the change of the electron dispersion caused by the interaction with the short-range AFM order. However, both magnetic and electronic properties are treated in our approach self- consistently. In Fig. 1, we also show the ARPES data on Bi2Sr2−xLaxCuO6+y (Bi2201) from [57] and the recent data [17] on LSCO for doping concentrations of 0.10, 0.12, and 0.16. The single crystals of Bi2201 have been studied experimentally with different hole concentra- tions, 0.05 < p < 0.18. This crystal has one CuO2 layer in the unit cell. That is why our calculations appropri- ate for LSCO can be used for Bi2201 with the condition x = p. The question arises whether the model parame- ters are the same or different for the two crystals? In the conventional single electron tight-binding model used in [57] to fit the ARPES data, the hopping parameters de- pend on doping significantly. That is why the authors of [57] claim that the ratio t′/t is different for Bi2201 and LSCO. However, as evident from Figs. 5,a and 5,b of [57], the hopping values are close to each other for the lowest doping for both substances. The reason is that the hopping parameters depend on the interatomic dis- tance that is almost the same in Bi2201 and LSCO. That is why we use the same parameters for all doping concen- trations. The doping dependence of the band structure and its non-rigid behavior comes up as the effect of SEC. One of the main players in this game is the filling factor Fσ̄2. Comparing our calculated FS with the experimental data in Fig. 1, we notice that, for x =0.05, 0.07, 0.10, and 0.12, the experimental Fermi arc position is close to the calculated inner part of the hole pocket (the part which is near the (0, 0) point). The outer part appears 58 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1 EFFECT OF LIFSHITZ QUANTUM PHASE TRANSITIONS as a low-intensity signal at x = 0.10 and x = 0.12 in ARPES. After the Lifshitz QPT at xc1 = 0.15, we see the two parts of the FS in agreement with ARPES data [57]. Usually, the outer FS (the nearest to (π, π) point) in Bi- cuprates is ascribed as the superlattice reflection. It may be that the superlattice signal simply masks a part of the FS that we obtain in the calculation. Another most probable explanation is that the scattering by AFM fluc- tuations suppresses the intensity of the spectral peaks corresponding to the outer FS. We will discuss this sce- nario in the next section. At a higher doping, the ARPES results in a large hole pocket centered around the (π, π) point [58], e.g., in Tl2Ba2CuO6+y with p = 0.26. Our calculations result in such a topology for x > xc2. According to [59], there is an electron pocket for LSCO at x = 0.30. We now discuss the FS area and the Luttinger the- orem. In Fig. 2, we give the FS area as a function of the doping. Note that the standard formulation of the Luttinger theorem does not work for Hubbard fermions. For free electrons, each quantum state in the k-space contains 2 electrons with opposite spins. The spectral weight of the Hubbard fermion is determined by the strength operator, Pσ = Fσ̄2, and each quantum state contains 2Fσ̄2 = 1 + x electrons. A generalized Lut- tinger theorem for the SEC system [60] takes the spec- tral weight of each |k〉 state into account. For LSCO, the hole concentration nh = 1 + x, so the electron con- centration ne = 1 − x. Using the dispersion law (see Fig. 3,b below), we calculate the number of occupied electronic states Ne k below the Fermi level. The elec- tronic concentration ne = 2Fσ̄2N e k = 1 − x. It gives us Ne k = (1 − x)/(1 + x). Then the number of free (occu- pied by holes) k-states is Nh k = 1 − Ne k = 2x/(1 + x), and the FS area in Fig. 2 is determined by this num- ber. The FS area obtained by the direct calculation of the occupied k-state under the Fermi level is shown by crosses. Two available FS areas from the quantum os- cillations data [11, 12] are also marked in Fig. 2. It is evident and very important that the Luttinger theorem is not applicable in the standard formulation. On the other hand, its generalization for the case of correlated Hubbard fermions describes the experimental data very well. 3. Qualitative Analysis of the Electron Dispersion and ARPES in a System with the Short-Range AFM Background It was shown earlier [13–16] that AFM fluctuations transform the closed hole pocket into an arc. We will 0 10 20 30 40 50 60 70 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 A re a un de r F S ( % o f B Z ) x 2x/(1+x) Calculated area under FS YBa2Cu3O6.5 (p=0.1) YBa2Cu4O8 (p=0.125) Fig. 2. Doping dependence of the FS area (in % of the Bril- louin zone area) calculated directly (+) and from the generalized Luttinger theorem (solid line). The experimental values from the quantum oscillations data [11, 12] are shown as well extent the same arguments to the doping region where AFM fluctuations are strong. The electron Green func- tion on the square lattice with the electron scattering by Gaussian fluctuations that imitate the short-range AFM order with Q = (π, π) is equal to [16] GD(k, E) = E − ε(k + Q) + ivk (E − ε(k)) (E − ε(k + Q) + ivk)− |D|2 . (4) Here, |D| stands for the amplitude of the fluctuating AFM order, ε(k) is the electron energy in the paramag- netic phase, and v = |vx(k + Q)|+ |vy(k + Q)| , vx,y(k) = ∂ε(k) ∂kx,y . In the absence of the damping, the Green function (4) describes an electron in the spin-density wave state with the long-range order, where the Umklapp shadow band is given by ε(k + Q). On the other hand, there is a dynamical transition ε(k) → ε(k + Q) for the AFM spin-liquid with a finite lifetime 1/τ ∼ vk. The paramagnetic dispersion is shown in Fig. 3,a by a thin green curve and a shadow band by a dotted curve to indicate that it has the finite lifetime as fol- lows from Eq. (4). The resulting QP dispersion in the short-range AFM state is given by a thick blue curve. With increasing the doping, the Fermi level moves down from its initial value “0” in Fig. 3,a. The first intersec- tion occurs along the (0, 0)− (π, π) direction and results in 4 small hole pockets. The inner part of the FS is ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1 59 S.G. OVCHINNIKOV, M.M. KORSHUNOV, E.I. SHNEYDER k x k y −3 −0.5 −4 0 0 00 0.5 0.5 0.50.5 −4 −4 −4 −0.5 −0.5 −0.5 −1.5 −1.5 −1.5 −1.5 −1.5−1.5 0 π/2 π 3π/2 2π 0 π/2 π 3π/2 2π Fig. 3. Qualitative scheme of the band structure of an electron against the fluctuating AFM background (a), our calculations (b), and the constant-energy cuts for x = 0.10 (c). The zero energy in (b) and (c) corresponds to the Fermi level. The constant en- ergy contours in (c) are labeled by the values of the corresponding energies (in units of t) formed mainly by the non-damped electrons from the ε(k) band, while the outer part is formed mainly by the damped shadow band. That is why the outer part has a very small spectral weight and was not seen in ARPES data until the recent discovery by the laser ARPES with the ultra-high energy resolution [17] (see Fig. 1). This qualitative analysis reproduce the calculations in [13– 16, 48]. We now proceed to higher doping concentrations. For x = 0.16 where AFM correlation length ξAFM ≈ 10 Å, we have two large FS centered around (π, π). Those can be deduced from Fig. 3,a by a further decrease of the Fermi energy, µ. The critical point xc1 appears, when µ touches the second peak along the (π, π)− (π, 0) direction. It is clear from Fig. 3,a that the inner FS will be of the electronic type and is formed by the damped shadow band. Thus, the corresponding spec- tral peaks are very small. The outer FS is of the hole type and is formed by the non-damped states. That is why its intensity is much larger than that of the in- ner part (see [57] and Fig. 1 for x = 0.16). With a further decrease of µ, it will cross the bottom of the band at x = xc2, which corresponds to the collapse of the electronic FS. Finally, at x > xc2, the crossing of µ with a saddle point at (π, 0) results in the transfor- mation of the FS from the hole to the electron type at x = xc3. The latter takes place in a strongly OD regime; this effect can be obtained in any conventional single electron approach and has been discussed before [61]. For comparison, we present our calculated band struc- ture for various doping concentrations in Fig. 3b and the constant energy cut in Fig. 3,c. It is clear that the rigid band approach of Fig. 3,a may give the correct se- quence of the FS reconstruction, but it is quantitatively wrong. 4. Low Temperature Thermodynamics near the Lifshitz Transition According to Lifshitz results [38, 39], both FS transfor- mations at xc1 and xc2 are 2.5-order electronic phase transitions (nowadays, the term QPT is used). The ap- pearance of a new FS sheet at ε = εc gives the additional density of state δg(ε) = α(ε − εc)1/2, with α ∼ 1 in a 3D system. In spite of a strong anisotropy in cuprates, they are 3D crystals. The weak interlayer hopping re- sults in a FS modulation along the kz axis that has been measured by ARPES [34]. That is why we can use re- sults of [38, 39] with a minimal modification due to the QP spectral weight in the strongly correlated system Fσ̄2 = (1 + x)/2. Near the critical point, the thermodynamical potential gains additional contribution Ω(µ, T ) = Ω0(µ, T ) + δΩ. 60 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1 EFFECT OF LIFSHITZ QUANTUM PHASE TRANSITIONS This singular contribution is induced by a new FS sheet at ε > εc and is equal to δΩ = − ∞∫ 0 δN(ε)fF (ε)dε, where fF (ε) is the Fermi function. The number of states is given by δN(ε) = { 0, ε < εc 2 3α 1+x 2 (ε− εc)3/2, ε > εc. At low temperature, T � z, and z = µ − εc. Near the QPT at z = 0, we get δΩ = { − √ π 4 (1 + x)αT 5/2e−|z|/T , z < 0, − 2 15 (1 + x)α|z|5/2 − π2 12 (1 + x)T 2|z|1/2, z > 0. It is the z5/2 singularity that tells about the 2.5-phase transition. In our case, z depends on the doping, so z(x) = 0 at x = xc1 and x = xc2. The singular contribution to the Sommerfeld param- eter γ = Ce/T , where Ce is the electronic specific heat, has the form δγ = −∂ 2δF ∂T 2 = = { √ π 4 (1 + x)α |z| 2 T 2 ( 1 + 3 T |z| + 15 4 T 2 |z|2 ) e−|z|/T , z < 0, π2 6 (1 + x)αz1/2, z > 0. We have deduced the z(x) dependence near each crit- ical point from our band structure calculations. The obtained δγ at T = 10K near xc1 is shown in Fig. 4. We also plot the experimental data [40] for LSCO, where Ce has been obtained by extrapolation of the high tem- perature data for T > Tc to the low-T region. The experimental points in Fig. 4 correspond to the total γ, γ(x) = γ0(x) + δγ, where γ0 is a smooth function at x ≈ xc1. Since the electron FS pocket disappears for x > xc2, our theory produces a singular behavior of γ(x) for x < xc2 corresponding to the case of z > 0. Measurements of the electronic specific heat [62] in NdBa2Cu3O6+y revealed two weak maxima of γ(x) at p = 0.16 and p = 0.23 that are close to our xc1 and xc2. To stay away from the superconductivity, the measurements in [62] were carried out at T = 200K, which explains why singularities appear as weak maxima. 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 γ x Theory Experiment Fig. 4. Sommerfeld parameter near the Lifshitz QPT. Experimen- tal data for γ = Ce/T at T = 10 K were taken from [40] 5. Interplay of Magnetic and Phonon Contributions to d-Wave Pairing A magnetic mechanism of pairing within the tJ model has been studied within various approaches. We use here the “no double occupation” constraint mean-field BCS- like version [23] and add the phonon contribution to the pairing. The resulting total coupling parameter is given by a sum of magnetic and phonon contributions [63] λtot (q) = 1− x 2 J + λphϑ (|ξq − µ| − ωD) , where λph = f (x)G, f (x) = (1 + x) (3 + x)/8− 3C01/4 depends on the nearest neighbor spin correlation func- tion C01 < 0, and G is the effective electron-phonon matrix element. The obtained doping dependence of Tc has maximum at the optimal doping xc1 = 0.15 due to the van Hove singularity in the density of states induced by the Lifshitz transition. This maximum of Tc results in a minimum of the isotope effect exponent. Fitting the calculated isotope effect exponent at the optimal dop- ing to the experimental one, we found G/J = 1.1 and Tmax c (J 6= 0, G 6= 0) ≈ Tmax c (J = 0, G 6= 0) (see details in [64]). Thus, the phonon and magnetic contributions to Tc are of the same order of magnitude. This means that there is no dominant mechanism of pairing, both are equally important. 6. Conclusion Previously, transformations of the FS has been discussed within a variational approach to the t−J model [29]. The small hole pocket near the (π/2, π/2) point has been ob- tained in the UD AFM. At a large doping, the electronic ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1 61 S.G. OVCHINNIKOV, M.M. KORSHUNOV, E.I. SHNEYDER FS around (0, 0) point also has been obtained. The same scenario of the FS evolution with doping was analyzed in the recent paper [65], where the authors considered the free dispersion of a doped antiferromagnet. Nevertheless, the FS for intermediate x in [29] does not correspond to our FS and to the experimental data in [17]. Recently, there were a lot of discussions on the change of the carrier sign upon doping. At large x, the FS be- comes of the electronic type: in LSCO, this happens at x > 0.30 [66]. As was mentioned above, it is rather a trivial fact. More unusual are the experimental data on the change of the Hall coefficient (RH) sign in the UD systems. This effect was observed (under a strong mag- netic field of 50÷ 60T that suppresses the superconduc- tivity) in YBa2Cu3Oy with p =0.10, 0.12, and 0.14 [67], and in LSCO with p = 0.11 [68]. All these crystals be- long to the region x < xc1 and, according to our theory, should have the small hole FS pockets. We believe that the arguments of [69] can explain the negative total Hall coefficient due to opposite partial contributions to RH of the FS with opposite curvatures in a two-dimensional metal. The low-temperature transport measurements on La1.6−xNd0.4SrxCuO4 in a strong magnetic field up to 35T reveal a change of the FS topology at p∗ ≈ 0.23 [70]. This critical point is very close to our xc2 = 0.24. Also, our theory agrees with the data of [70] in the sense that, at p = 0.24, the RH indicates the large cylindrical FS with 1+p holes. At p = 0.20 that corresponds to x < xc2, the RH(T ) increase at a low temperature leads to the conclusion that the FS reconstruction and the pseudo- gap formation happen at p < p∗ [70]. The critical con- centration xc2 agrees with the concentration pc = 0.23, where the van Hove singularity in Bi2201 has been found in ARPES [71, 72]. There is a wide discussion in the literature on the quantum critical point Pcrit, where the pseudogap char- acteristic temperature T ∗(P ) → 0. According to [73], Pcrit = 0.19, but Pcrit = 0.27 according to [74]. All these values are obtained by the extrapolation from a finite- T regime. On the contrary, the two critical points xc1 and xc2 obtained here are the properties of the ground state and result from the Lifshitz QPT. It is possible that our xc2 is somehow related to the Pcrit; at least, p∗ = 0.24 is related to the pseudogap formation at p < p∗ [70]. We thank A. Kordyuk for the discussion of the re- sults and T.M. Ovchinnikova for the technical assistance. This work was supported by project 5.7 of the programm “Quantum physics of the condensed matter” of the Pre- sidium of the RAS, RFFI grant 09-02-00127, and inte- gration project N 40 of SB RAS. 1. E. Dagotto, Rev. Mod. Phys. 66, 763 (1994). 2. E.G. Maksimov, Phys. Usp. 43, 965 (2000). 3. M. Imada, A. Fujimori, and Y. Tokura, Rev. Mod. Phys. 70, 1039 (1998). 4. S.G. Ovchinnikov, Phys. Usp. 40, 993 (1997). 5. M.V. Sadovskii, Phys. Usp. 44, 515 (2001). 6. V.M. Loktev, R.M. Quick, S.G. Sharapov, Phys. Rep. 349, 2 (2001). 7. V.F. Elesin, V.V. Kapaev, and Yu.V. Kopaev, Phys. Usp. 47, 949 (2004). 8. Yu.A. Izyumov and E.Z. Kurmaev, Phys. Usp. 51, 23 (2008). 9. P.A. Lee, Rep. Prog. Phys. 71, 012501 (2008). 10. A. Damascelli, Z. Hussein, and Z.X. Shen, Rev. Mod. Phys. 75, 473 (2003). 11. N. Doiron-Leyrand et al., Nature 447, 565 (2007). 12. E.A. Yelland, J. Singleton, C.H. Mielke, N. Narrison, F.F. Balakirev, B. Dabrowski, and J.R. Cooper, Phys. Rev. Lett. 100, 047003 (2008). 13. E.Z. Kuchinskii, I.A. Nekrasov, M.V. Sadovskii, JETP Lett. 82, 198 (2005). 14. E.Z. Kuchinskii and M.V. Sadovskii, JETP 103, 415 (2006). 15. N. Harrison, R.D. McDonald, and J. Singleton, Phys. Rev. Lett. 99, 206406 (2007). 16. E.Z. Kuchinskii and M.V. Sadovskii, JETP Lett. 88, 192 (2008). 17. J. Meng, G. Liu, W. Zhang et al., arXiv:0906.2682 (2009). 18. M.Yu. Kagan and K.I. Kugel, Phys. Usp. 44 553 (2001). 19. S.G. Ovchinnikov and I.S. Sandalov, Physica C 161, 607 (1989). 20. S.V. Lovtsov and V.Yu. Yushankhai, Physica C 179, 159 (1991). 21. J.H. Jefferson, H. Eskes, and L.F. Feiner, Phys. Rev. B 45, 7959 (1992). 22. V.I. Belinicher, A.L. Chernyshev, and V.A. Shubin, Phys. Rev. B 53, 335 (1996). 23. N.M. Plakida and V.S. Oudovenko, Phys. Rev. B 59, 11949 (1999). 24. M.M. Korshunov, V.A. Gavrichkov, S.G. Ovchinnikov, I.A. Nekrasov, Z.V. Pchelkina, and V.I. Anisimov, Phys. Rev. B 72, 165104 (2005). 25. W. Stephan and P. Horsch, Phys. Rev. Lett. 66, 2258 (1991). 26. R. Preuss, W. Hanke, and W. von der Linden, Phys. Rev. Lett. 75, 1344 (1995). 62 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1 EFFECT OF LIFSHITZ QUANTUM PHASE TRANSITIONS 27. V.F. Elesin and V.A. Koshurnikov, J. Exp. Theor. Phys. 79, 961. (1994). 28. B.I. Shraiman and E.D. Siggia, Phys. Rev. Lett. 61, 467 (1988). 29. S.A. Trugman, Phys. Rev. Lett. 65, 500 (1990). 30. A.F. Barabanov, Superconductivity: Physics, Chemistry and Technology 3, 8 (1990) (in Russian). 31. A.F. Barabanov, R.O. Kuzian, and L.A. Maksimov, J. Phys.: Condens. Matter. 39, 129 (1991). 32. A.P. Kampf, Phys. Rev. 249, 219 (1994). 33. G. Dorf, A. Muramatsu, and W. Hanke, Phys. Rev. B 41, 9264 (1990). 34. S. Sahrakorpi, R.S. Markiewicz, H. Lin et al., Phys. Rev. B 78, 104513 (2008). 35. T.R. Thurston, R.J. Birgeneau, M.A. Kastner et al., Phys. Rev. B 40, 4585 (1989). 36. S.M. Haden et al., Phys. Rev. Lett. 66, 821 (1991). 37. D. Mihailovic and V.V. Kabanov, Superconductivity in Complex Systems. Series: Structure and Bonding Vol. 114, edited by K. A. Müller and A. Bussmann-Holder (Springer, Berlin, 2005), p. 331. 38. I.M. Lifshitz, Sov. Phys. JETP 11, 1130 (1960). 39. I.M. Lifshitz, M.Ya. Asbel, and M.I. Kaganov, Electron Theory of Metals (Consultants Bureau, New York, 1973). 40. J.W. Loram, J. Luo, J.R. Cooper, W.Y. Liang, and J.L. Tallon, Phys. Chem. Solids 62, 59 (2001). 41. Ya.B. Gaididei, V.M. Loktev. Phys. Stat. Sol. B 147, 307 (1988). 42. V.A. Gavrichkov, S.G. Ovchinnikov, А.А. Borisov, and E.G. Goryachev, JETP 91, 369 (2000). 43. J. Zaanen, G.A. Sawatzky, and J.W. Allen, Phys. Rev. Lett. 55, 418 (1985). 44. S.G. Ovchinnikov and V.V. Val’kov, Hubbard Operators in the Theory of Strongly Correlated Electrons (Imperial College Press, London, 2004). 45. V.V. Val’kov, T.A. Val’kova, D.M. Dzebisashvili, and S.G. Ovchinnikov, JETP Lett. 75, 378 2002. 46. R.O. Zaitsev, Sov. Phys. JETP 41, 100 (1975). 47. M.M. Korshunov and S.G. Ovchinnikov, Eur. Phys. J. B 57, 271 (2007). 48. N.M. Plakida and V.S. Oudovenko, JETP 104, 230 (2007). 49. V.V. Val’kov and D.M. Dzebisashvili, JETP 100, 608 (2005). 50. H. Shimahara and S. Takada, J. Phys. Soc. Jpn 60, 2394 (1991); 61, 989 (1992). 51. A.F. Barabanov and V.M. Berezovskii, JETP 79 627 (1994). 52. A. Sherman and M. Schreiber, Phys. Rev. B 65, 134520 (2002). 53. A.A. Vladimirov, D. Ihle, and N.M. Plakida, Theor. Math. Phys. 145, 1576 (2005). 54. S. Sachdev, A.V. Chubukov, and A. Sokol, Phys. Rev. B 51, 14874 (1995). 55. A.F. Barabanov, A.A. Kovalev, O.V. Urusaev, A.M. Belemuk, and R. Hain, JETP 92, 677 (2001). 56. L. Hozoi, M.S. Laad, and P. Fulde, Phys. Rev. B 78, 165107 (2008). 57. M. Hashimoto, T. Yoshida, H. Yagi et al., Phys. Rev. B 77, 094516 (2008). 58. M. Plate, J.D.F. Mottershead, I.S. Elfimov et al., Phys. Rev. Lett. 95, 077001 (2005). 59. A. Ino, C. Kim, M. Nakamura et al., Phys.Rev. B 65, 094504 (2002). 60. M.M. Korshunov and S.G. Ovchinnikov, Phys. Sol. St. 45, 1415 (2003). 61. F. Onufrieva and P. Pfeuty, Phys. Rev. B 61, 799 (2000). 62. U. Tutsch, P. Schweiss, H. Wühl, B. Obst, and Th. Wolf, Eur. Phys. J. B 41, 471 (2004). 63. E.I. Shneyder and S.G. Ovchinnikov, JETP Lett. 83, 394 (2006). 64. E.I. Shneyder and S.G. Ovchinnikov, to be published in JETP (2009). 65. V.M. Loktev and V. Turkowski, J. Low Temp. Phys. 154, 117 (2009). 66. I. Tsukada and S. Ono, Phys. Rev. B 74, 134508 (2006). 67. D.Le Boeuf, N. Doiron-Leyraud, J. Levallois et al., Na- ture 450, 533 (2003). 68. T. Adachi, T. Noji, and Y. Koike, Phys. Rev. B 64, 144524 (2001). 69. N.P. Ong, Phys. Rev. B 43, 193 (1991). 70. R. Daou, N. Doiron-Leyrand, D. Le Boeuf et al., Nature Phys. 5, 31 (2009). 71. A. Kaminski, S. Rosenkranz, N.M. Fretweel et al., Phys. Rev. B 73, 174511 (2006). 72. A.A. Kordyuk, S.V. Borisenko, M. Khupfer, and J. Fink, Phys. Rev. B 67, 064504 (2003). 73. J.G. Storey, J.L. Tallon, and G.V.M. Williams, Phys. Rev. B 78, 140506(R) (2008). 74. S. Hufner, M.A. Hossain, A. Damascelly, and G.A. Sawatzky, Rep. Prog. Phys. 71, 062501 (2008). Received 20.09.09 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1 63 S.G. OVCHINNIKOV, M.M. KORSHUNOV, E.I. SHNEYDER ВПЛИВ КВАНТОВИХ ФАЗОВИХ ПЕРЕХОДIВ ЛIФШИЦА НА НОРМАЛЬНИЙ ТА НАДПРОВIДНИЙ СТАНИ КУПРАТIВ С.Г. Овчинников, М.М. Коршунов, Е.I. Шнейдер Р е з ю м е У цiй работi обговорено змiни електронної структури в нор- мальнiй фазi високотемпературних надпровiдникiв – шарува- тих купратiв. Результати розрахункiв електронної структури та поверхнi Фермi одношарових купратiв методом LDA+GTB iз врахуванням сильних кореляцiй порiвнюються з даними ARPES та квантових осциляцiй. Виявлено двi критичнi точки xc1 та xc2, в яких вiдбувається перебудова поверхнi Фермi. В околi критичних точок у межах iдеологiї I.М. Лiфшица про квантовi фазовi переходи 2,5 роду знайдено змiни термодина- мiчних властивостей за низьких температур. Особливiсть еле- ктронної теплоємностi δ(C/T ) ∼ (x − xc)1/2 достатньо добре узгоджується з вiдомими експериментальними даними в око- лi xc1 ≈ 0, 15. Якiсно обговорюються змiни знака константи Холла з допуванням. Також розглянуто надпровiдний стан з урахуванням магнiтного i фононного механiзмiв спарювання. 64 ISSN 2071-0194. Ukr. J. Phys. 2010. Vol. 55, No. 1