Signatures of the precursor superconductivity above T c

Transcription

1 Dresden, 18 April 2007 Signatures of the precursor superconductivity above T c T. DOMANSKI M. Curie-Skłodowska University, Lublin, Poland

2 Outline

3 Outline Introduction / pairing in the many-body systems /

4 Outline Introduction / pairing in the many-body systems / Conventional superconductors / effects of small pair fluctuations /

5 Outline Introduction / pairing in the many-body systems / Conventional superconductors / effects of small pair fluctuations / HTSC cuprates

6 Outline Introduction / pairing in the many-body systems / Conventional superconductors / effects of small pair fluctuations / HTSC cuprates basic properties

7 Outline Introduction / pairing in the many-body systems / Conventional superconductors / effects of small pair fluctuations / HTSC cuprates basic properties description of pair fluctuations above T c

8 Outline Introduction / pairing in the many-body systems / Conventional superconductors / effects of small pair fluctuations / HTSC cuprates basic properties description of pair fluctuations above T c results for ARPES, STM and collective phenomena

9 Outline Introduction / pairing in the many-body systems / Conventional superconductors / effects of small pair fluctuations / HTSC cuprates basic properties description of pair fluctuations above T c results (ARPES, collective features, etc) Summary

10 I. Introduction

11 Pairing is a common phenomenon which occurs between various kinds of fermions such as: quarks, electrons, nucleons or atoms.

12 Pairing is a common phenomenon which occurs between various kinds of fermions such as: quarks, electrons, nucleons or atoms. The underlying pairing mechanism can be driven by:

13 Pairing is a common phenomenon which occurs between various kinds of fermions such as: quarks, electrons, nucleons or atoms. The underlying pairing mechanism can be driven by: 1. exchange of phonons / classical superconductors, MgB 2, diamond,... /

14 Pairing is a common phenomenon which occurs between various kinds of fermions such as: quarks, electrons, nucleons or atoms. The underlying pairing mechanism can be driven by: 1. exchange of phonons / classical superconductors, MgB 2, diamond,... / 2. exchange of magnons / superconductivity of the heavy fermion compounds /

15 Pairing is a common phenomenon which occurs between various kinds of fermions such as: quarks, electrons, nucleons or atoms. The underlying pairing mechanism can be driven by: 1. exchange of phonons / classical superconductors, MgB 2, diamond,... / 2. exchange of magnons / superconductivity of the heavy fermion compounds / 3. strong correlations / high T c superconductors /

16 Pairing is a common phenomenon which occurs between various kinds of fermions such as: quarks, electrons, nucleons or atoms. The underlying pairing mechanism can be driven by: 1. exchange of phonons / classical superconductors, MgB 2, diamond,... / 2. exchange of magnons / superconductivity of the heavy fermion compounds / 3. strong correlations / high T c superconductors / 4. Feshbach resonance / ultracold superfluid atoms /

17 Pairing is a common phenomenon which occurs between various kinds of fermions such as: quarks, electrons, nucleons or atoms. The underlying pairing mechanism can be driven by: 1. exchange of phonons / classical superconductors, MgB 2, diamond,... / 2. exchange of magnons / superconductivity of the heavy fermion compounds / 3. strong correlations / high T c superconductors / 4. Feshbach resonance / ultracold superfluid atoms / 5. other / pairing in nuclei, gluon-quark plasma /

18 Pairing is a common phenomenon which occurs between various kinds of fermions such as: quarks, electrons, nucleons or atoms. The underlying pairing mechanism can be driven by: 1. exchange of phonons / classical superconductors, MgB 2, diamond,... / 2. exchange of magnons / superconductivity of the heavy fermion compounds / 3. strong correlations / high T c superconductors / 4. Feshbach resonance / ultracold superfluid atoms / 5. other / pairing in nuclei, gluon-quark plasma / Very often formation of the fermion pairs goes hand in hand with superconductivity/superfluidity but it needs not be the rule.

19 Hamiltonian of the pairing interactions

20 Hamiltonian of the pairing interactions The momentum representation: Ĥ = k,σ ξ k ĉ kσĉkσ + k,k V k,k ĉ k ĉ k ĉ k ĉ k where V k,k < 0 (at least for some k, k states)

21 Hamiltonian of the pairing interactions The momentum representation: Ĥ = k,σ ξ k ĉ kσĉkσ + k,k V k,k ĉ k ĉ k ĉ k ĉ k where V k,k < 0 (at least for some k, k states) The real space representation: Ĥ = i,j t i,j ĉ iσĉjσ + i,j V i,j ĉ i ĉi ĉ j ĉj σ with attractive potential V i,j < 0

22 II. Conventional superconductors

23 Major property

24 Major property Pair formation and onset of their. coherence coincide at T c (0) (T) 0 temperatura T c Pairing is responsible for the gap in the single particle spectrum

25 Major property Pair formation and onset of their. coherence coincide at T c (0) (T) 0 temperatura T c Pairing is responsible for the gap in the single particle spectrum The order parameter 2-nd order phase transition c V (T) 0 0 ~ e /T T c temperatura / as classified by Landau /

26 Pair fluctuations

27 Pair fluctuations Possible sources:

28 Pair fluctuations Possible sources: 1. quantum fluctuations  ˆB =  ˆB +  ˆB  ˆB + δâ δ ˆB where δâ =   fluctuation neglected in the BCS theory.

29 Pair fluctuations Possible sources: 1. quantum fluctuations  ˆB =  ˆB +  ˆB  ˆB + δâ δ ˆB where δâ =   fluctuation neglected in the BCS theory. 2. topological In the low dimensional (dim 2) systems ODLRO does not establish. There can appear only the power-law behaviour ˆψ (r 1 ) ˆψ (r 2 ) r 1 r 2 θ(t ) whith the phase stiffness θ 0 for T T KT. J.M. Kosterlitz and P.J. Thouless, J. Phys. C 6, 1181 (1973).

30 Historical remarks: 1

31 Historical remarks: 1 First estimation of the fluctuation effects has been done by V.L. Ginzburg (1963).

32 Historical remarks: 1 First estimation of the fluctuation effects has been done by V.L. Ginzburg (1963). He predicted smearing of the specific heat jump near T c but in an extremely narrow temperature region δt T c ( ) a ξ a interatomic distance, ξ correlation length. V.L. Ginzburg, Sov. Solid State 2, 61 (1968).

33 Historical remarks: 2

34 Historical remarks: 2 First experimental observation of the sc fluctuations above T c has been obtained for granular aluminium.

35 Historical remarks: 2 First experimental observation of the sc fluctuations above T c has been obtained for granular aluminium. Tunneling conductance revealed a small pseudogap above T c.

36 Historical remarks: 2 First experimental observation of the sc fluctuations above T c has been obtained for granular aluminium. Tunneling conductance revealed a small pseudogap above T c. R.W. Cohen and B. Abels, Phys. Rev. 168, 444 (1968).

37 Historical remarks: 3

38 Historical remarks: 3 One can study the fluctuations in a systematic way using the perturbative theory for conductivity. The leading diagrams are:

39 Historical remarks: 3 One can study the fluctuations in a systematic way using the perturbative theory for conductivity. The leading diagrams are: Aslamazov-Larkin Maki - Thompson diagrams giving corrections to the density of states

40 Historical remarks: 3 One can study the fluctuations in a systematic way using the perturbative theory for conductivity. The leading diagrams are: Aslamazov-Larkin Maki - Thompson diagrams giving corrections to the density of states V.V. Dorin et al, Phys. Rev. B 48, (1993).

41 III. HTSC cuprates

42 1. The parent compounds are quasi-2d Mott insulators

43 1. The parent compounds are quasi-2d Mott insulators Important remark: Spatial extent of the pairs is very short ξ ab 5 Å

44 2. Superconductivity appears upon doping by

45 2. Superconductivity appears upon doping by electrons or holes O. Fisher et al, Rev. Mod. Phys. 79, 353 (2007).

46 2. Superconductivity appears upon doping by electrons or holes A puzzle: O. Fisher et al, Rev. Mod. Phys. 79, 353 (2007). Why this diagram is asymmetric?

47 3. Inhomogeneities

48 3. Inhomogeneities The new STM spectroscopy finds that the energy gap is inhomogenous. / K. McElroy et al, Phys. Rev. Lett. 94, (2005) /

49 3. Inhomogeneities The new STM spectroscopy finds that the energy gap is inhomogenous. / K. McElroy et al, Phys. Rev. Lett. 94, (2005) / Notice: Inhomogenous structure promotes the fluctuations

50 4. Pseudogap

51 4. Pseudogap T. Nakano et al, J. Phys. Soc. Jpn. 67, 2622 (2002).

52 4. Pseudogap T. Nakano et al, J. Phys. Soc. Jpn. 67, 2622 (2002). What kind of mechanism is responsible for the pseudogap?

53 Theoretical concepts

54 Theoretical concepts

55 Theoretical concepts (a) Pseudogap is a precursor of the superconducting gap which is due to strong fluctuations (because of the reduced dimensionality, local pairing, etc). (b) Pseudogap is not related to sc gap. It represents some other type of ordering which is competing with the sc phase.

56 Pair fluctuations

57 Pair fluctuations Usual mean field solution (i.e. the saddle point) neglects an influence of any fluctuations.

58 Pair fluctuations One can describe the small fluctuations via Gaussian corrections.

59 The effective physics

60 The effective physics Without a specific pairing mechanism being established we propose to describe the coherent (for T < T c ) or incoherent (for T > T c ) fermion pairs using the phenomenological boson-fermion model.

61 The effective physics Without a specific pairing mechanism being established we propose to describe the coherent (for T < T c ) or incoherent (for T > T c ) fermion pairs using the phenomenological boson-fermion model. The dark areas denote such parts of the I-st Brillouin zone where the local (electron or hole) pairs exist.

62 The effective physics Without a specific pairing mechanism being established we propose to describe the coherent (for T < T c ) or incoherent (for T > T c ) fermion pairs using the phenomenological boson-fermion model. The dark areas denote such parts of the I-st Brillouin zone where the local (electron or hole) pairs exist. V.B. Geshkenbein, L.B. Ioffe and A.I. Larkin, Phys. Rev. B 55, 3173 (1997).

63 Anisotropic energy gap d-wave type symmetry k = (cos k x cos k y ) of the gap J.E. Hoffman et al, Science 297, 1148 (2002).

64 The BF model (J. Ranninger, 1985): H = kσ (ε k µ) c kσ c kσ + q (E q 2µ) b q b q + 1 N k,q ] v k,q [b q c k, c q k, + h.c.

65 The BF model (J. Ranninger, 1985): H = kσ (ε k µ) c kσ c kσ + q (E q 2µ) b q b q + 1 N k,q ] v k,q [b q c k, c q k, + h.c. has been explored in the context of HTSC:

66 The BF model (J. Ranninger, 1985): H = kσ (ε k µ) c kσ c kσ + q (E q 2µ) b q b q + 1 N k,q ] v k,q [b q c k, c q k, + h.c. has been explored in the context of HTSC: R. Micnas, J. Ranninger, S. Robaszkiewicz, Rev. Mod. Phys. 62, 113 (1990).

67 The BF model (J. Ranninger, 1985): H = kσ (ε k µ) c kσ c kσ + q (E q 2µ) b q b q + 1 N k,q ] v k,q [b q c k, c q k, + h.c. has been explored in the context of HTSC: R. Micnas, J. Ranninger, S. Robaszkiewicz, Rev. Mod. Phys. 62, 113 (1990). R. Friedberg and T.D. Lee, Phys. Rev. B 40, 423 (1989).

68 The BF model (J. Ranninger, 1985): H = kσ (ε k µ) c kσ c kσ + q (E q 2µ) b q b q + 1 N k,q ] v k,q [b q c k, c q k, + h.c. has been explored in the context of HTSC: R. Micnas, J. Ranninger, S. Robaszkiewicz, Rev. Mod. Phys. 62, 113 (1990). R. Friedberg and T.D. Lee, Phys. Rev. B 40, 423 (1989). C. Lannert, M.P.A. Fisher and T. Senthil, Phys. Rev. B 64, (2001).

69 The BF model (J. Ranninger, 1985): H = kσ (ε k µ) c kσ c kσ + q (E q 2µ) b q b q + 1 N k,q ] v k,q [b q c k, c q k, + h.c. has been explored in the context of HTSC: R. Micnas, J. Ranninger, S. Robaszkiewicz, Rev. Mod. Phys. 62, 113 (1990). R. Friedberg and T.D. Lee, Phys. Rev. B 40, 423 (1989). C. Lannert, M.P.A. Fisher and T. Senthil, Phys. Rev. B 64, (2001). E. Altman and A. Auerbach, Phys. Rev. B 65, (2002).

70 The BF model (J. Ranninger, 1985): H = kσ (ε k µ) c kσ c kσ + q (E q 2µ) b q b q + 1 N k,q ] v k,q [b q c k, c q k, + h.c. has been explored in the context of HTSC: R. Micnas, J. Ranninger, S. Robaszkiewicz, Rev. Mod. Phys. 62, 113 (1990). R. Friedberg and T.D. Lee, Phys. Rev. B 40, 423 (1989). C. Lannert, M.P.A. Fisher and T. Senthil, Phys. Rev. B 64, (2001). E. Altman and A. Auerbach, Phys. Rev. B 65, (2002). and for description of the superfluid ultracold atoms:

71 The BF model (J. Ranninger, 1985): H = kσ (ε k µ) c kσ c kσ + q (E q 2µ) b q b q + 1 N k,q ] v k,q [b q c k, c q k, + h.c. has been explored in the context of HTSC: R. Micnas, J. Ranninger, S. Robaszkiewicz, Rev. Mod. Phys. 62, 113 (1990). R. Friedberg and T.D. Lee, Phys. Rev. B 40, 423 (1989). C. Lannert, M.P.A. Fisher and T. Senthil, Phys. Rev. B 64, (2001). E. Altman and A. Auerbach, Phys. Rev. B 65, (2002). and for description of the superfluid ultracold atoms: R.A. Duine and H.T.C. Stoof, Phys. Rep. 396, 115 (2004).

72 The BF model (J. Ranninger, 1985): H = kσ (ε k µ) c kσ c kσ + q (E q 2µ) b q b q + 1 N k,q ] v k,q [b q c k, c q k, + h.c. has been explored in the context of HTSC: R. Micnas, J. Ranninger, S. Robaszkiewicz, Rev. Mod. Phys. 62, 113 (1990). R. Friedberg and T.D. Lee, Phys. Rev. B 40, 423 (1989). C. Lannert, M.P.A. Fisher and T. Senthil, Phys. Rev. B 64, (2001). E. Altman and A. Auerbach, Phys. Rev. B 65, (2002). and for description of the superfluid ultracold atoms: R.A. Duine and H.T.C. Stoof, Phys. Rep. 396, 115 (2004). Q. Chen, J. Stajic, S. Tan and K. Levin, Phys. Rep. 412, 1 (2005).

73 Outline of the procedure

74 Outline of the procedure In order to study the many-body effects we construct

75 Outline of the procedure In order to study the many-body effects we construct the continuous canonical transformation e ˆ S(l) Ĥe Ŝ(l)

76 Outline of the procedure In order to study the many-body effects we construct the continuous canonical transformation e ˆ S(l) Ĥe Ŝ(l) to decouple the boson from fermion degrees of freedom.

77 Outline of the procedure In order to study the many-body effects we construct the continuous canonical transformation e ˆ S(l) Ĥe Ŝ(l) to decouple the boson from fermion degrees of freedom. Hamiltonian at l = 0 Ĥ F + Ĥ B + ˆV BF

78 Outline of the procedure In order to study the many-body effects we construct the continuous canonical transformation e ˆ S(l) Ĥe Ŝ(l) to decouple the boson from fermion degrees of freedom. Hamiltonian at 0 < l < Ĥ F (l) + Ĥ B (l) + ˆV BF (l)

79 Outline of the procedure In order to study the many-body effects we construct the continuous canonical transformation e ˆ S(l) Ĥe Ŝ(l) to decouple the boson from fermion degrees of freedom. Hamiltonian at l = Ĥ F ( ) + Ĥ B ( ) + 0

80 Outline of the procedure In order to study the many-body effects we construct the continuous canonical transformation e ˆ S(l) Ĥe Ŝ(l) to decouple the boson from fermion degrees of freedom. Hamiltonian at l = Ĥ F ( ) + Ĥ B ( ) + 0 T. Domański and J. Ranninger, Phys. Rev. B 63, (2001).

81 (a) Energy T c < T Intensity E F k F Wave vector ε k Intensity (b) v k 2 u k 2 E F Energy T <T c k F Wave vector -E k 2 E k Single particle spectrum of conventional superconductors consists of two Bogoliubov branches gapped around E F (no fluctuation effects are here taken into account).

82 Bogoliubov-like spectrum T = 0 T=0 A F (k,ω) ω Below the critical temperature T c there exist two branches of the single-particle excitation energies which (according to the BCS theory) occur at ω k = ± (ε k µ) k. T. Domański and J. Ranninger, Phys. Rev. Lett. 91, (2003).

83 Bogoliubov-like spectrum T c < T < T T=0.004 A F (k,ω) Above T c the Bogoliubov-type spectrum still survives but one of the branches ( the shaddow ) becomes overdamped. This means that fermion pairs have no longer an infinite life-time. T. Domański and J. Ranninger, Phys. Rev. Lett. 91, (2003).

84 Bogoliubov-like spectrum T c < T < T T=0.007 A F (k,ω) Above T c the Bogoliubov-type spectrum still survives but one of the branches ( the shaddow ) becomes overdamped. This means that fermion pairs have no longer an infinite life-time. T. Domański and J. Ranninger, Phys. Rev. Lett. 91, (2003).

85 Bogoliubov-like spectrum T > T T=0.02 A F (k,ω) For temperatures far above T c the Bogoliubov modes are gone and there remains only a single quasiparticle peak surrounded by an incoherent background. T. Domański and J. Ranninger, Phys. Rev. Lett. 91, (2003).

86 Experimental data (Japanese group) Energy (mev) Coherence factor 50 E F (a) A B C T < T c -ε k ε k (c) E k 140 K -E k Momentum u k 2 v k 2 u k 2 + v k ε k (mev) Intensity (arb. units) A B C E F -50 E F (b) A B C Binding Energy (mev) H. Matsui, T. Sato, and T. Takahashi, Phys. Rev. Lett. 90, (2003).

87 Experimental data (Dresden group) T < T c T c < T T. Eckl et al, Phys. Rev. B 70, (2004).

88 The peak-dip-hump structure A.G. Loeser, Z.-X. Shen et al, Phys. Rev. B 56, (1997).

89 Spectral function for T < T c Schematic view of the spectral function in the antinodal direction for temperatures T < T c obtained using the boson-fermion model. T. Domański and J. Ranninger, Phys. Rev. B 70, (2004).

90 The phenomenon of waterfalls J. Graf et al, Phys. Rev. Lett. 98, (2006).

91 Physical implications of the waterfalls D.S. Ionosov et al, cond-mat/ ; A.A. Kordyuk et al, cond-mat/

92 Waterfalls can result from coupling to bosonic mode A d F(k,ω) T > T 0.7π k=0.6π k=0.5π k=0.4π k=0.3π=k F k=0.2π k=0.1π ω k=0 T. Domański and J. Ranninger, Phys. Rev. B 70, (2004).

93 Collective phenomena

94 Collective phenomena We also investigated the pair correlation function k c k (τ )c q k (τ ) k c q k c k

95 Collective phenomena We also investigated the pair correlation function k c k (τ )c q k (τ ) k c q k c k and found the corresponding spectral function N q δ ( ω Ẽ q ) + A inc k (ω) consisting of:

96 Collective phenomena We also investigated the pair correlation function k c k (τ )c q k (τ ) k c q k c k and found the corresponding spectral function N q δ ( ω Ẽ q ) + A inc k (ω) consisting of: the quasiparticle peak at ω = Ẽ q

97 Collective phenomena We also investigated the pair correlation function k c k (τ )c q k (τ ) k c q k c k and found the corresponding spectral function N q δ ( ω Ẽ q ) + A inc k (ω) consisting of: the quasiparticle peak at ω = Ẽ q and the incoherent background A inc k (ω).

98 Collective phenomena We also investigated the pair correlation function k c k (τ )c q k (τ ) k c q k c k and found the corresponding spectral function N q δ ( ω Ẽ q ) + A inc k (ω) consisting of: the quasiparticle peak at ω = Ẽ q and the incoherent background A inc k (ω). T. Domański and J. Ranninger, Phys. Rev. B 70, (2004).

99 The pair spectrum for T < T c inc. background incoherent background ~ E q q The quasiparticle peak is well separated from the incoherent background and, in the limit q 0, has a characteristic dispersion Ẽ q = c q. This Goldstone mode is a hallmark of the symmetry broken state. Such a unique situation could be observed in the case of ultracold fermion atoms, otherwise the Coulomb repulsions lift this mode to the high plasmon frequency.

100 The pair spectrum for T > T > T c incoherent background ~ E q incoherent background q crit q Above the transition temperature (for T > T > T c ): the qusiparticle peak overlaps at small momenta with the incoherent background, for q 0 the Goldstone mode disappears, remnant of the Goldstone mode is seen above q crit.

101 The pair spectrum for T > T > T c ~ de q /dq x 0.2 normal state pseudogap state q x Remnant of the Goldstone branch in the dispersion of fermion pairs above T c. T. Domański and A. Donabidowicz, (2007) in print.

102 Experimental evidence: 1 Residual Meissner effect observed experimentally with use of the ultrafast magnetic fields. J. Corson et al, Nature 398, 221 (1999).

103 Experimental evidence: La 2-x Sr x CuO 4 T (K) T onset 20 T c Sr content x The large Nernst effect measured above T c. Y. Wang et al, Science 299, 86 (2003).

104 SUMMARY

105 SUMMARY Formation of the fermion needs not be accompanied by the onset of superfluidity/superconductivity.

106 SUMMARY Formation of the fermion needs not be accompanied by the onset of superfluidity/superconductivity. Appearance of fermion pairs leads to a (partial) depletion of the single particle states near the Fermi energy.

107 SUMMARY Formation of the fermion needs not be accompanied by the onset of superfluidity/superconductivity. Appearance of fermion pairs leads to a (partial) depletion of the single particle states near the Fermi energy. Strong quantum fluctuations suppress the long-range coherence (ordering) while fermion pairs are preserved.

108 SUMMARY Formation of the fermion needs not be accompanied by the onset of superfluidity/superconductivity. Appearance of fermion pairs leads to a (partial) depletion of the single particle states near the Fermi energy. Strong quantum fluctuations suppress the long-range coherence (ordering) while fermion pairs are preserved. Besides the single particle features (Bogoliubov modes) there should be visible also the collective features which are characteristic for the superfluid state.

109 SUMMARY Formation of the fermion needs not be accompanied by the onset of superfluidity/superconductivity. Appearance of fermion pairs leads to a (partial) depletion of the single particle states near the Fermi energy. Strong quantum fluctuations suppress the long-range coherence (ordering) while fermion pairs are preserved. Besides the single particle features (Bogoliubov modes) there should be visible also the collective features which are characteristic for the superfluid state. Precursor effects are robust for all superconductors but their temperature extent varies from case to case.

110 Resonating Valence Bonds Why not? P.W. Anderson, Phys. Rev. Lett. 96, (2006).