ɛ(k) = h2 k 2 2m, k F = (3π 2 n) 1/3

Transcription

1 4D-XY Quantum Criticality in Underdoped High-T c cuprates M. Franz University of British Columbia franz@physics.ubc.ca February 22, 2005 In collaboration with: A.P. Iyengar (theory) D.P. Broun, D.A. Bonn (experiment)

2 4D-XY QUANTUM CRITICALITY 1 Landau s Fermi liquid paradigm Lev Davidovich Landau: Electron states in solids are adiabatically connected to the states of noninteracting electron gas SLIDES CREATED USING FoilTEX & PP 4

3 4D-XY QUANTUM CRITICALITY 2 Despite enormous Coulomb forces (U C e2 a eV) at low energies most metals behave like a free electron gas [Landau, 1957] k z ky kx ɛ(k) = h2 k 2 2m, k F = (3π 2 n) 1/3 N electrons in a box Fermi sphere Ground state (T = 0): all levels below Fermi momentum k F are filled; levels above k F empty.

4 4D-XY QUANTUM CRITICALITY 2 Despite enormous Coulomb forces (U C e2 a eV) at low energies most metals behave like a free electron gas [Landau, 1957] k z ky kx ɛ(k) = h2 k 2 2m, k F = (3π 2 n) 1/3 N electrons in a box Fermi sphere Ground state (T = 0): all levels below Fermi momentum k F are filled; levels above k F empty. Root cause: Pauli exclusion principle phase space for scattering near FS is severely limited. SLIDES CREATED USING FoilTEX & PP 4

5 4D-XY QUANTUM CRITICALITY 3 Structure of electron propagator in FL G(k, ω) = 1 ω ɛ k Σ(k, ω) = z k ω E k + iγ k + G incoh (k, ω) with z 1 k = [1 ReΣ ω ] ω=e k, quasiparticle weight τ 1 Γ k (E k E F ) 2, quasiparticle lifetime

6 4D-XY QUANTUM CRITICALITY 3 Structure of electron propagator in FL G(k, ω) = 1 ω ɛ k Σ(k, ω) = z k ω E k + iγ k + G incoh (k, ω) with z 1 k = [1 ReΣ ω ] ω=e k, quasiparticle weight τ 1 Γ k (E k E F ) 2, quasiparticle lifetime Spectral function: A(k, ω) coherent quasiparticle A(k, ω) = 2ImG(k, ω) z k δ(ω E k ) + A incoh (k, ω) A incoh ω SLIDES CREATED USING FoilTEX & PP 4

7 4D-XY QUANTUM CRITICALITY 4 Exceptions to FL paradigm 1D interacting systems (a.k.a. Luttinger Liquids )

8 4D-XY QUANTUM CRITICALITY 4 Exceptions to FL paradigm 1D interacting systems (a.k.a. Luttinger Liquids ) Quantum Hall Fluids

9 4D-XY QUANTUM CRITICALITY 4 Exceptions to FL paradigm 1D interacting systems (a.k.a. Luttinger Liquids ) Quantum Hall Fluids Systems near Quantum Criticality

10 4D-XY QUANTUM CRITICALITY 4 Exceptions to FL paradigm 1D interacting systems (a.k.a. Luttinger Liquids ) Quantum Hall Fluids Systems near Quantum Criticality High-T c Cuprate Superconductors (?) SLIDES CREATED USING FoilTEX & PP 4

11 4D-XY QUANTUM CRITICALITY 5 Cuprate superconductors Cuprates are layered quasi-2d materials with electronic properties dominated by the CuO 2 layers. T T* Pseudogap AF dsc x Crystal structure of La 2 x Sr x CuO 4 Phase diagram of cuprates. SLIDES CREATED USING FoilTEX & PP 4

12 4D-XY QUANTUM CRITICALITY 6 d-wave superconductivity in cuprates Superconducting order parameter in cuprates exhibits d-wave symmetry i.e. changes sign upon 90 o rotation. k = 1 2 0(cos k x cos k y ),

13 4D-XY QUANTUM CRITICALITY 6 d-wave superconductivity in cuprates Superconducting order parameter in cuprates exhibits d-wave symmetry i.e. changes sign upon 90 o rotation. k = 1 2 0(cos k x cos k y ), Low-energy excitations, E k = ɛ 2 k + 2 k, occur near 4 nodal points: d-wave order parameter Dirac cone Dirac Fermions

14 4D-XY QUANTUM CRITICALITY 6 d-wave superconductivity in cuprates Superconducting order parameter in cuprates exhibits d-wave symmetry i.e. changes sign upon 90 o rotation. k = 1 2 0(cos k x cos k y ), Low-energy excitations, E k = ɛ 2 k + 2 k, occur near 4 nodal points: d-wave order parameter Dirac cone Dirac Fermions SLIDES CREATED USING FoilTEX & PP 4

15 4D-XY QUANTUM CRITICALITY 7 Superfluid density Superfluid density ρ s is a fundamental characteristic of a superconductor reflecting its ability to carry supercurrent in response to applied magnetic field H = A: j s = ρ s 4e 2 h 2 c A.

16 4D-XY QUANTUM CRITICALITY 7 Superfluid density Superfluid density ρ s is a fundamental characteristic of a superconductor reflecting its ability to carry supercurrent in response to applied magnetic field H = A: j s = ρ s 4e 2 h 2 c A. ρ s is related to London penetration depth λ, a fundamental lengthscale in a superconductor describing penetration of magnetic field H into the bulk via vacuum superconductor λ ρ s = h2 c 2 16πe 2 λ 2 j H SLIDES CREATED USING FoilTEX & PP 4

17 4D-XY QUANTUM CRITICALITY 8 Superfluid density in cuprates, ab-plane: the old story In the optimally doped and moderately underdoped region experiments show ρ ab s (x, T ) λ 2 ab (x, T ) ax bk BT, with a 244meV and b 3.0 [Lee and Wen, PRL 78, 4111 (1997)].

18 4D-XY QUANTUM CRITICALITY 8 Superfluid density in cuprates, ab-plane: the old story In the optimally doped and moderately underdoped region experiments show ρ ab s (x, T ) λ 2 ab (x, T ) ax bk BT, with a 244meV and b 3.0 [Lee and Wen, PRL 78, 4111 (1997)]. In a BCS d-wave superconductor one would have ρ ab s (x, T ) a(1 x) 2 ln 2 v F v k B T. SLIDES CREATED USING FoilTEX & PP 4

19 4D-XY QUANTUM CRITICALITY 9 The linear T -dependence is known to arise from thermally excited nodal quasiparticles which exhibit linear density of states at low energies. DOS T ω

20 4D-XY QUANTUM CRITICALITY 9 The linear T -dependence is known to arise from thermally excited nodal quasiparticles which exhibit linear density of states at low energies. DOS T ω The linear x-dependence (cf. Uemura plot) reflects proximity to the Mott-Hubbard insulator at half filling.

21 4D-XY QUANTUM CRITICALITY 9 The linear T -dependence is known to arise from thermally excited nodal quasiparticles which exhibit linear density of states at low energies. DOS T ω The linear x-dependence (cf. Uemura plot) reflects proximity to the Mott-Hubbard insulator at half filling. Problem: models that give correct x-dependence (e.g. RVB-type theories) generally yield strong ( x 2 ) dependence of the coefficient b. SLIDES CREATED USING FoilTEX & PP 4

22 4D-XY QUANTUM CRITICALITY 10 3D-XY critical scaling Optimally doped YBCO shows 3D-XY critical behavior near T c : ρ s (T ) (T c T ) 2/3 with relatively wide critical region T 10K. Such critical behavior is characteristic of phase disordering transition to the normal state. In 3d this is known to be caused by vortex loop unbinding. Kamal et al., PRL (1994)

23 4D-XY QUANTUM CRITICALITY 10 3D-XY critical scaling Optimally doped YBCO shows 3D-XY critical behavior near T c : ρ s (T ) (T c T ) 2/3 with relatively wide critical region T 10K. Such critical behavior is characteristic of phase disordering transition to the normal state. In 3d this is known to be caused by vortex loop unbinding. Kamal et al., PRL (1994) 3D-XY critical behavior has also been observed in other high-t c compounds using transport and thermodynamic probes. It thus appears to be a fundamental universal feature of cuprates. SLIDES CREATED USING FoilTEX & PP 4

24 4D-XY QUANTUM CRITICALITY 11 Superfluid density in cuprates: the new story Recent UBC Group data on ultra-pure YBCO single crystals for doping levels as low as T c = 5K show ab-plane show tantalizing qualitative deviations from the old phenomenology.

25 4D-XY QUANTUM CRITICALITY 11 Superfluid density in cuprates: the new story Recent UBC Group data on ultra-pure YBCO single crystals for doping levels as low as T c = 5K show ab-plane show tantalizing qualitative deviations from the old phenomenology.

26 4D-XY QUANTUM CRITICALITY 11 Superfluid density in cuprates: the new story Recent UBC Group data on ultra-pure YBCO single crystals for doping levels as low as T c = 5K show ab-plane show tantalizing qualitative deviations from the old phenomenology. Data from UBC/SFU group [Broun et al. unpublished] SLIDES CREATED USING FoilTEX & PP 4

27 4D-XY QUANTUM CRITICALITY 12 New ab-plane phenomenology

28 4D-XY QUANTUM CRITICALITY 12 New ab-plane phenomenology New features: No visible 3D-XY critical regime.

29 4D-XY QUANTUM CRITICALITY 12 New ab-plane phenomenology New features: No visible 3D-XY critical regime. ρ ab s (x, T ) Ax 2 BxT with x T c, a parameter roughly proportional to doping

30 4D-XY QUANTUM CRITICALITY 12 New ab-plane phenomenology amplitude: ρ ab s (x, 0) T 2 c New features: No visible 3D-XY critical regime. ρ ab s (x, T ) Ax 2 BxT with x T c, a parameter roughly proportional to doping

31 4D-XY QUANTUM CRITICALITY 12 New ab-plane phenomenology amplitude: ρ ab s (x, 0) T 2 c New features: No visible 3D-XY critical regime. ρ ab s (x, T ) Ax 2 BxT with x T c, a parameter roughly proportional to doping slope SLIDES CREATED USING FoilTEX & PP 4

32 4D-XY QUANTUM CRITICALITY 13 Understanding the puzzle...

33 4D-XY QUANTUM CRITICALITY 13 Understanding the puzzle...

34 4D-XY QUANTUM CRITICALITY 13 Understanding the puzzle... With underdoping T c falls and the 3D-XY classical critical region shrinks to zero as the system approaches a quantum critical point at x = x c.

35 4D-XY QUANTUM CRITICALITY 13 Understanding the puzzle... With underdoping T c falls and the 3D-XY classical critical region shrinks to zero as the system approaches a quantum critical point at x = x c. This QCP lives in (3+1) dimensions, 1 standing for the imaginary time τ.

36 4D-XY QUANTUM CRITICALITY 13 Understanding the puzzle... With underdoping T c falls and the 3D-XY classical critical region shrinks to zero as the system approaches a quantum critical point at x = x c. This QCP lives in (3+1) dimensions, 1 standing for the imaginary time τ. For xy-type models 4 is the upper critical dimension; one thus expects mean field critical behavior. SLIDES CREATED USING FoilTEX & PP 4

37 4D-XY QUANTUM CRITICALITY 14 Consistency check If the underdoped region is controlled by (3+1)D-XY quantum critical point then we should be able to understand the behavior of ρ ab s (x, T ) based on very general scaling arguments.

38 4D-XY QUANTUM CRITICALITY 14 Consistency check If the underdoped region is controlled by (3+1)D-XY quantum critical point then we should be able to understand the behavior of ρ ab s (x, T ) based on very general scaling arguments. Indeed, elementary scaling analysis gives ρ ab s (x, 0) T c (d 2+z)/z with z the dynamical critical exponent (z 1).

39 4D-XY QUANTUM CRITICALITY 14 Consistency check If the underdoped region is controlled by (3+1)D-XY quantum critical point then we should be able to understand the behavior of ρ ab s (x, T ) based on very general scaling arguments. Indeed, elementary scaling analysis gives ρ ab s (x, 0) T c (d 2+z)/z with z the dynamical critical exponent (z 1). In d = 2 we have ρ ab s (x, 0) T c independent of z (the Uemura scaling ).

40 4D-XY QUANTUM CRITICALITY 14 Consistency check If the underdoped region is controlled by (3+1)D-XY quantum critical point then we should be able to understand the behavior of ρ ab s (x, T ) based on very general scaling arguments. Indeed, elementary scaling analysis gives ρ ab s (x, 0) T c (d 2+z)/z with z the dynamical critical exponent (z 1). In d = 2 we have ρ ab s (x, 0) T c independent of z (the Uemura scaling ). In d = 3 we have ρ ab s (x, 0) T (1+z)/z c.

41 4D-XY QUANTUM CRITICALITY 14 Consistency check If the underdoped region is controlled by (3+1)D-XY quantum critical point then we should be able to understand the behavior of ρ ab s (x, T ) based on very general scaling arguments. Indeed, elementary scaling analysis gives ρ ab s (x, 0) T c (d 2+z)/z with z the dynamical critical exponent (z 1). In d = 2 we have ρ ab s (x, 0) T c independent of z (the Uemura scaling ). In d = 3 we have ρ ab s (x, 0) T (1+z)/z c. For 1 z 2 this is consistent with experiment! SLIDES CREATED USING FoilTEX & PP 4

42 4D-XY QUANTUM CRITICALITY 15 Problems The 4D-XY QCP idea seems to naturally explain: Lack of visible classical critical region in very underdoped YBCO.

43 4D-XY QUANTUM CRITICALITY 15 Problems The 4D-XY QCP idea seems to naturally explain: Lack of visible classical critical region in very underdoped YBCO. Violations of the Uemura scaling ρ s T c.

44 4D-XY QUANTUM CRITICALITY 15 Problems The 4D-XY QCP idea seems to naturally explain: Lack of visible classical critical region in very underdoped YBCO. Violations of the Uemura scaling ρ s T c. There are, however, two serious issues: Cuprates are strongly anisotropic; it is unclear how broad the (3+1)D critical region is.

45 4D-XY QUANTUM CRITICALITY 15 Problems The 4D-XY QCP idea seems to naturally explain: Lack of visible classical critical region in very underdoped YBCO. Violations of the Uemura scaling ρ s T c. There are, however, two serious issues: Cuprates are strongly anisotropic; it is unclear how broad the (3+1)D critical region is. There is strong (linear) T -dependence of the bare superfluid density coming from quasiparticles which may invalidate the scaling laws. SLIDES CREATED USING FoilTEX & PP 4

46 4D-XY QUANTUM CRITICALITY 16 2D - 3D crossover At T = 0, away from QCP, fluctuations in CuO 2 layers are decoupled and system behaves 2 dimensionally.

47 4D-XY QUANTUM CRITICALITY 16 2D - 3D crossover At T = 0, away from QCP, fluctuations in CuO 2 layers are decoupled and system behaves 2 dimensionally. Close to QCP the c-axis coherence length grows, ξ c (x x c ) ν, and ultimately exceeds the interlayer spacing d 0. At that point system starts behaving 3 dimensionally.

48 4D-XY QUANTUM CRITICALITY 16 2D - 3D crossover At T = 0, away from QCP, fluctuations in CuO 2 layers are decoupled and system behaves 2 dimensionally. Close to QCP the c-axis coherence length grows, ξ c (x x c ) ν, and ultimately exceeds the interlayer spacing d 0. At that point system starts behaving 3 dimensionally. For YBCO we estimate, using ξ c λ 2 ab /λ cκ, T 3D 5 10K. SLIDES CREATED USING FoilTEX & PP 4

49 4D-XY QUANTUM CRITICALITY 17 Quantum XY model To address the details of ρ s temperature and doping dependence in a fluctuating d-wave superconductor we need a model.

50 4D-XY QUANTUM CRITICALITY 17 Quantum XY model To address the details of ρ s temperature and doping dependence in a fluctuating d-wave superconductor we need a model. The simplest model showing XY-type critical behavior is given by the Hamiltonian H XY = 1 ˆn i V ijˆn j 1 J ij cos( ˆϕ i ˆϕ j ). 2 2 ij ij

51 4D-XY QUANTUM CRITICALITY 17 Quantum XY model To address the details of ρ s temperature and doping dependence in a fluctuating d-wave superconductor we need a model. The simplest model showing XY-type critical behavior is given by the Hamiltonian H XY = 1 ˆn i V ijˆn j 1 J ij cos( ˆϕ i ˆϕ j ). 2 2 ij Here ˆn i and ˆϕ i are the number and phase operators representing Cooper pairs on site r i of a cubic lattice and are quantum mechanically conjugate variables: [ˆn i, ˆϕ j ] = iδ ij. ij

52 4D-XY QUANTUM CRITICALITY 17 Quantum XY model To address the details of ρ s temperature and doping dependence in a fluctuating d-wave superconductor we need a model. The simplest model showing XY-type critical behavior is given by the Hamiltonian H XY = 1 ˆn i V ijˆn j 1 J ij cos( ˆϕ i ˆϕ j ). 2 2 ij Here ˆn i and ˆϕ i are the number and phase operators representing Cooper pairs on site r i of a cubic lattice and are quantum mechanically conjugate variables: [ˆn i, ˆϕ j ] = iδ ij. ij The sites r i do not necessarily represent individual Cu atoms; rather one should think in terms of coarse grained lattice model valid at long lengthscales where microscopic details no longer matter. SLIDES CREATED USING FoilTEX & PP 4

53 4D-XY QUANTUM CRITICALITY 18 The first term in H XY describes interactions between Cooper pairs; we take e 2 V ij = Uδ ij + (1 δ ij ) r i r j.

54 4D-XY QUANTUM CRITICALITY 18 The first term in H XY describes interactions between Cooper pairs; we take e 2 V ij = Uδ ij + (1 δ ij ) r i r j. The second term represents Josephson tunneling of pairs between the sites: { J, for n.n. along a, b J ij = J, for n.n. along c 0, otherwise

55 4D-XY QUANTUM CRITICALITY 18 The first term in H XY describes interactions between Cooper pairs; we take e 2 V ij = Uδ ij + (1 δ ij ) r i r j. The second term represents Josephson tunneling of pairs between the sites: { J, for n.n. along a, b J ij = J, for n.n. along c 0, otherwise In the absence of interactions J clearly must be identified as the superfluid density. We thus take J = J 0 αt with α = (2 ln 2)v F /v, as in the BCS d-wave superconductor. SLIDES CREATED USING FoilTEX & PP 4

56 4D-XY QUANTUM CRITICALITY 19 The claim: As formulated above the quantum XY model captures the experimentally observed phenomenology of underdoped YBCO.

57 4D-XY QUANTUM CRITICALITY 19 The claim: As formulated above the quantum XY model captures the experimentally observed phenomenology of underdoped YBCO. Namely, it predicts ab-plane superfluid density of the form ρ ab s (x, T ) Ax 2 BxT.

58 4D-XY QUANTUM CRITICALITY 19 The claim: As formulated above the quantum XY model captures the experimentally observed phenomenology of underdoped YBCO. Namely, it predicts ab-plane superfluid density of the form ρ ab s (x, T ) Ax 2 BxT. It also naturally yields the shrinking classical fluctuation region with decreasing T c. SLIDES CREATED USING FoilTEX & PP 4

59 4D-XY QUANTUM CRITICALITY 20 Self-consistent harmonic approximation The idea is to replace the XY Hamiltonian H XY = 1 2 ˆn i V ijˆn j J ij cos( ˆϕ i ˆϕ j ). ij by the trial harmonic Hamiltonian H har = 1 2 ij ˆn i V ijˆn j K ij ( ˆϕ i ˆϕ j ) 2.

60 4D-XY QUANTUM CRITICALITY 20 Self-consistent harmonic approximation The idea is to replace the XY Hamiltonian H XY = 1 2 ˆn i V ijˆn j J ij cos( ˆϕ i ˆϕ j ). ij by the trial harmonic Hamiltonian H har = 1 2 ˆn i V ijˆn j K ij ( ˆϕ i ˆϕ j ) 2. ij Constant K, which is identified as the renormalized superfluid density, is determined from the requirement that E har H XY har is minimal.

61 4D-XY QUANTUM CRITICALITY 20 Self-consistent harmonic approximation The idea is to replace the XY Hamiltonian H XY = 1 2 ˆn i V ijˆn j J ij cos( ˆϕ i ˆϕ j ). ij by the trial harmonic Hamiltonian H har = 1 2 ˆn i V ijˆn j K ij ( ˆϕ i ˆϕ j ) 2. ij Constant K, which is identified as the renormalized superfluid density, is determined from the requirement that E har H XY har is minimal. This is just a variational principle which can be extended to T > 0 case using the Gibbs-Bogolyubov inequality F F har + H H har har. SLIDES CREATED USING FoilTEX & PP 4

62 4D-XY QUANTUM CRITICALITY 21 H har is quadratic in ˆn i and ˆϕ j and can thus be easily diagonalized: H har = q hω q (a qa q ) with the frequencies hω q = 2 KZ q V q, Z q = sin 2 (q x /2) + sin 2 (q y /2) + sin 2 (q z /2).

63 4D-XY QUANTUM CRITICALITY 21 H har is quadratic in ˆn i and ˆϕ j and can thus be easily diagonalized: H har = q hω q (a qa q ) with the frequencies hω q = 2 KZ q V q, Z q = sin 2 (q x /2) + sin 2 (q y /2) + sin 2 (q z /2). For short range interactions V q const as q 0; we have ω q q, i.e. acoustic phase mode. For Coulomb interactions V q 1/q 2 as q 0; we have ω q ω pl, i.e. gapped plasma mode. SLIDES CREATED USING FoilTEX & PP 4

64 4D-XY QUANTUM CRITICALITY 22 Simple power counting shows that at low T the contribution from the phase mode to the superfluid density is δρ ph s { T 3, short range interaction e ωpl/t, Coulomb interaction

65 4D-XY QUANTUM CRITICALITY 22 Simple power counting shows that at low T the contribution from the phase mode to the superfluid density is δρ ph s { T 3, short range interaction e ωpl/t, Coulomb interaction In either case the low-t behavior of ρ s will be dominated by the quasiparticle contribution included via J = J 0 αt.

66 4D-XY QUANTUM CRITICALITY 22 Simple power counting shows that at low T the contribution from the phase mode to the superfluid density is δρ ph s { T 3, short range interaction e ωpl/t, Coulomb interaction In either case the low-t behavior of ρ s will be dominated by the quasiparticle contribution included via J = J 0 αt. However, as we shall see, quantum fluctuations will strongly renormalize both the amplitude J 0 and the slope α. SLIDES CREATED USING FoilTEX & PP 4

67 4D-XY QUANTUM CRITICALITY 23 SCHA: the results Using the identity cos( ˆϕ i ˆϕ j ) har = exp [ 1 2 ( ˆϕ i ˆϕ j ) 2 har ], valid for harmonic Hamiltonians, we obtain E har = H XY har = KS Je S/K, with S = (4N) 1 q Vq Z q.

68 4D-XY QUANTUM CRITICALITY 23 SCHA: the results Using the identity cos( ˆϕ i ˆϕ j ) har = exp [ 1 2 ( ˆϕ i ˆϕ j ) 2 har ], valid for harmonic Hamiltonians, we obtain E har = H XY har = KS Je S/K, with S = (4N) 1 q Vq Z q. Parameter S describes the aggregate strength of interactions that are responsible for quantum fluctuations and reduction of ρ s.

69 4D-XY QUANTUM CRITICALITY 23 SCHA: the results Using the identity cos( ˆϕ i ˆϕ j ) har = exp [ 1 2 ( ˆϕ i ˆϕ j ) 2 har ], valid for harmonic Hamiltonians, we obtain E har = H XY har = KS Je S/K, with S = (4N) 1 q Vq Z q. Parameter S describes the aggregate strength of interactions that are responsible for quantum fluctuations and reduction of ρ s. Minimizing E har with respect to K we obtain K = Je S/K J(1 S/J). SLIDES CREATED USING FoilTEX & PP 4

70 4D-XY QUANTUM CRITICALITY 24 To obtain the leading temperature dependence substitute J = J 0 αt and expand to leading order in T : ρ s (x, T ) = K J 0 (1 S J 0 ) αt ( 1 1 S 2 J 0 ).

71 4D-XY QUANTUM CRITICALITY 24 To obtain the leading temperature dependence substitute J = J 0 αt and expand to leading order in T : ρ s (x, T ) = K J 0 (1 ) ( S J 0 αt 1 1 S 2 J 0 ). As expected, both the amplitude and the slope are reduced by quantum fluctuations.

72 4D-XY QUANTUM CRITICALITY 24 To obtain the leading temperature dependence substitute J = J 0 αt and expand to leading order in T : ρ s (x, T ) = K J 0 (1 ) ( S J 0 αt 1 1 S 2 J 0 ). As expected, both the amplitude and the slope are reduced by quantum fluctuations. Crucially, observe that the T = 0 amplitude decays faster than the slope.

73 4D-XY QUANTUM CRITICALITY 24 To obtain the leading temperature dependence substitute J = J 0 αt and expand to leading order in T : ρ s (x, T ) = K J 0 (1 ) ( S J 0 αt 1 1 S 2 J 0 ). As expected, both the amplitude and the slope are reduced by quantum fluctuations. Crucially, observe that the T = 0 amplitude decays faster than the slope. In particular, for small S/J 0 the above expression is consistent with experimentally observed behavior if we identify x (1 1 2 S ρ ab s (x, T ) J 0 x 2 αxt, J 0 ). SLIDES CREATED USING FoilTEX & PP 4

74 4D-XY QUANTUM CRITICALITY 25 How about the region S J 0? In this regime one can construct a critical theory of strong phase fluctuations [Doniach, PRB 24, 5063 (1981)] as an expansion in small order parameter ψ(x, τ). This leads to a quantum Ginzburg-Landau action S = β 0 dτ d 3 x { r ψ u ψ ψ } 2c 2 τψ 2, with parameters r, u and c given as functions of J and S.

75 4D-XY QUANTUM CRITICALITY 25 How about the region S J 0? In this regime one can construct a critical theory of strong phase fluctuations [Doniach, PRB 24, 5063 (1981)] as an expansion in small order parameter ψ(x, τ). This leads to a quantum Ginzburg-Landau action S = β 0 dτ d 3 x { r ψ u ψ ψ } 2c 2 τψ 2, with parameters r, u and c given as functions of J and S. Combining SCHA with this critical theory provides a consistent picture for the suppression of ρ s by quantum fluctuations. SLIDES CREATED USING FoilTEX & PP 4

76 4D-XY QUANTUM CRITICALITY 26 Summary The combined SCHA+critical analysis of the Quantum XY model predicts doping and temperature dependence of ρ ab s (x, T ) that is consistent with recent experiments on strongly underdoped YBCO.

77 4D-XY QUANTUM CRITICALITY 26 Summary The combined SCHA+critical analysis of the Quantum XY model predicts doping and temperature dependence of ρ ab s (x, T ) that is consistent with recent experiments on strongly underdoped YBCO. We are currently investigating implications of this model for other physical observables, namely specific heat, c-axis superfluid density, fluctuation diamagnetism, thermal and electrical conductivity. SLIDES CREATED USING FoilTEX & PP 4

78 4D-XY QUANTUM CRITICALITY 27 Conclusions Phenomenology of the underdoped cuprates suggests that phase fluctuations play important role as the doping is reduced.

79 4D-XY QUANTUM CRITICALITY 27 Conclusions Phenomenology of the underdoped cuprates suggests that phase fluctuations play important role as the doping is reduced. In this region single particle gap grows but superfluid density becomes vanishingly small, approximately as ρ ab s (x, T ) J 0 x 2 αxt, reflecting the approach to Mott insulating phase.

80 4D-XY QUANTUM CRITICALITY 27 Conclusions Phenomenology of the underdoped cuprates suggests that phase fluctuations play important role as the doping is reduced. In this region single particle gap grows but superfluid density becomes vanishingly small, approximately as ρ ab s (x, T ) J 0 x 2 αxt, reflecting the approach to Mott insulating phase. We have shown that quantum XY model in 3 spatial dimensions captures the observed behavior, including the unusual T dependence, deviations from the Uemura scaling, and the apparent lack of classical critical fluctuations.

81 4D-XY QUANTUM CRITICALITY 27 Conclusions Phenomenology of the underdoped cuprates suggests that phase fluctuations play important role as the doping is reduced. In this region single particle gap grows but superfluid density becomes vanishingly small, approximately as ρ ab s (x, T ) J 0 x 2 αxt, reflecting the approach to Mott insulating phase. We have shown that quantum XY model in 3 spatial dimensions captures the observed behavior, including the unusual T dependence, deviations from the Uemura scaling, and the apparent lack of classical critical fluctuations. This phenomenology puts severe constraints on microscopic models of underdoped cuprates. SLIDES CREATED USING FoilTEX & PP 4