Detection of novel electronic order in Fe based superconducting (and related) materials with point contact spectroscopy

Transcription

1 Detection of novel electronic order in Fe based superconducting (and related) materials with point contact spectroscopy H.Z. Arham, W.K. Park, C.R. Hunt, LHG UIUC Z. J. Xu, J. S. Wen, Z. W. Lin, Q. Li, G. Gu BNL P. H. Tobash, F. Ronning, E.D. Bauer, J.L. Sarrao, J.D. Thompson LANL J. Gillett, S. Sebastian Cambridge, UK A. Thaler, S. L. Bu dko, P. C. Canfield Ames ISU Center for Emergent Superconductivity

2 After a couple of decades HTS is NOT UNIQUE to Cuprates!!! And we still do not understand the mechanism of SC in any of the HTS and related novel superconductors!

3 Generic HTS phase diagrams Pressure and doping tuning Cuprates Fe Based

4 Similar phase diagrams of related materials organic and heavy-fermion SCs 4 3 (K) 2 * T AFM 1 SC SC 0 SC? Co 0.5 Rh 0.5 Ir X 0.5 Co 4

5 Outline Quasiparticle Scattering Spectroscopy [Point Contact Spectroscopy] Heavy Fermion CeCoIn 5 Detecting phase orderings and broken symmetries Heavy Fermion URu 2 Si 2 Detecting the Hybridization Gap Ba(Fe 1 x Co x ) 2 As 2 and Fe 1+y Te Detecting orbital ordering?

6 Quasiparticle Scattering Spectroscopy (Point Contact Spectroscopy PCS ) ev/2 +ev/2 I di Au d CeMIn5 V dv (+) bias (+) CeMIn 5 Point Contact Spectroscopy, Naidyuk & Yanson (2005) If two bulk metals are in contact with each other and the contact size is smaller than electronic mean free paths, quasiparticle energy gain/loss mostly occurs at the constriction. Nolinearities in current voltage characteristics reflect energydependent quasiparticle scatterings in the contact region.

7 Quasiparticle Scattering Spectroscopy (Point Contact Spectroscopy PCS ) Transparent Contact between two materials (not tunneling). Contact area less than elastic mean free path. Measure di/dv vs. V bias. Soft PCS Needle-anvil PCS Microshorts through insulating layer Thermally stable X Little control over junction resistance Electrochemically etched nanoscale gold Tunable junction resistance (piezoelectric bimorphs / screw) X Thermally less stable

8 Quasiparticle scattering (PCS): What we detect 1. Superconductivity (intro): Gap, order parameter symmetry, density of state, this is well understood in the framework of BTK and Andreev reflection. 2. Kondo Impurity / Kondo Lattice effects (intro): Fano lineshape consistent with the growth of the Kondo liquid (starts at T*, and grows until T c ): microscopic explanation? 3. Antiferromagnetic ordering (intro): Conductance enhancement at T N in some AF systems (not all) 4. The hybridization gap & Fano resonance in a heavy fermions 5. Orbital ordering? This talk: Enhancement at high T in Co:BaFe 2 As 2 and FeTe.

9 Superconductivity: PCS to Tunneling BTK model Three fitting parameters Δ = superconducting gap Γ = Dynes broadening factor (qp scattering rate) = barrier strength at the N/S interface Z eff Assuming Γ = 0 and Δ = 1 Z = 0 Andreev Reflection Z = 0.5 Z = 5.0 Tunneling

10 The Heavy Fermion Superconductor CeCoIn 5 : Phase diagram of series Ce M In 5 (M = Co, Rh, In) & transport T c = 2.3 K (high for many HFS) Superconductivity in clean limit (mfp = 810Å >> ξ 0 ) 4 3 T * (K) 2 AFM 1 SC SC 0 SC? Co 0.5 Rh 0.5 Ir 0.5 Co CeCoIn X 5 CeRhIn 5 CeIrIn 5 CeCoIn 5 Pagliuso et al., Phys. Rev. B 64 (2001) (R)

11 Superconductivity: Gap and OP Symmetry Data: Consistency Along Three Orientations Normalized Conductance K 1.52 K (001) K 1.57 K (110) K 1.47 K (100) Voltage (mv) Voltage (mv) Voltage (mv) (+) CeCoIn 5 ; ( ) Au => Adding electrons to CeCoIn 5 above the Fermi energy is more difficult than removing them Note the shapes of the conductance curves

12 Spectroscopic Evidence for d x 2 -y 2 Symmetry Normalized di/dv Exp. Data Compare shapes for data with calculation. Flat vs. cusp-like Z is always (100) finite. -2 Nodes -1 are 0 along 1 2 (110) direction. V (mv) (110) SC surface Normalized di/dv antinodal Calc ev/δ Z nodal Andreev Bound States (ABS) k y k x WK Park et al., PRL 100,

13 Kondo: Background Conductance Asymmetry of CeCoIn 5 Normalized Conductance Voltage (mv) T(K) T * T c Background develops an asymmetry* at the heavyfermion liquid coherence temperature, T* ~45 K. This asymmetry gradually increases with decreasing temperature until the onset of superconducting coherence, T c =2.3 K. di/dv( 2 mev) / di/dv (+2 mev) log T T (K) * Temperature (K)

14 Model fits magnitude of AR, asymmetry and T dep! Normalized Conductance Data (circles) and fit (red line) is excellent 400 mk data best fit (001) CeCoIn 5 Normalized Conductance Best fit over wide T range with a Fano lineshape 400 mk Lorentzian 2.6 K Fano line f (ε) F(ε) ε q F F ε Voltage (mv) Voltage (mv) Fano may be explained by interference between f electrons and conduction electrons via spin flip (Kondo) scattering. W. K. Park et al., PRL 08 and Y.-F. Yang et al., PRL 08

15 Fano Resonance in Single Impurity and Quantum Dot V. Madhavan et al., Science 280, 567 (1998) Co atoms on Au(111) K. Kobayashi et al., PRL 88, (2002) The Fano effect is essentially a single-impurity problem describing how a localized state embedded in the continuum acquires itinerancy over the system.

16 Antiferromagnetism: Cd:CeCoIn 5 : Anomalous Conductance Zero-bias Conductance (Ω -1 ) T c di/dv (Ω -1 ) Voltage (mv) T N Temperature (K) L. Pham et al. PRL 97, (2006) WK Park et al., Physica B 403, 731 (2008) non-monotonic; enhancement below T N, competition below T c0

17 Quasiparticle scattering (PCS): What we detect 1. Superconductivity (intro): Gap, order parameter symmetry, density of state, this is well understood in the framework of BTK and Andreev reflection. 2. Kondo Impurity / Kondo Lattice effects (intro): Fano lineshape consistent with the growth of the Kondo liquid (starts at T*, and grows until Tc): microscopic explanation? 3. Antiferromagnetic ordering (intro): Conductance enhancement at T N in some AF systems (not all NOW: 1. The hybridization gap & Fano resonance in a heavy fermion This talk: Helps explain #2. 2. Orbital ordering? This talk: Enhancement at high-t in Co:BaFe 2 As2and FeTeSe.

18 The Heavy Fermion / Kondo Lattice URu 2 Si 2 Δ = 115 K (9.9 mev) Palstra et al., PRL (86); PRB (87) A Phase Diagram Hassinger et al., PRB 77, (2008)

19 Hybridization Picture of a Kondo Lattice Periodic Anderson model ( ) σ σ ( ) σ σ V ( f σc σ ckσ fkσ ) H = ε μ c c + ε μ f f U n n PAM k k k f k k k k k f, i f, i kσ kσ kσ i Mean field solution { ( ) } ε λ ε λ = 1 ± 2 + ± + = E 4 V, V z 2 V k k k 0 e.g., Newns & Read, Adv. Phys. (1987) μ: chemical potential λ : renormalized f-level V: renormalized hybridization amplitude z = 1- n f (n f : occupancy)

20 Electron co-tunneling in a Kondo Lattice Provides a model to account for transport / tunneling taking itinerant and re-normalized f-electron band Single-impurity: Fano ineshape Kondo lattice hybridization gap: Double peak Maltseva, Dzero, Coleman, PRL (09)

21 Fano Resonance: Single Impurity vs. Kondo Lattice Single Impurity Kondo Lattice A distinct double-peak structure in a Kondo lattice: signature of a hybridization gap, distinguishable from a single impurity Fano resonance. Asymmetry due to interference between renormalized heavy bands and conduction band.

22 Typical Conductance Data 0.07 T = 2.07 K di/dv (Ω -1 ) Sample bias (mv)

23 Temperature Dependence Is this hybridization gap the longthought hidden order parameter? The answer lies in the temperature dependence. T HO Conductance spectra (filled circles) along with fitted curves (solid lines). Top three curves on a magnified vertical scale. The double-peak structure persists well above T HO.

24 Hybridization Gap Hybridization gap opening temperature T hyb = 27 K >> T HO. Δ hyb is unlikely to be the hidden order parameter, as opposed to recent theoretical claims. Renormalized f-level shows characteristic temperature dependence. Crossing the chemical potential at T HO?

25 FINALLY: Measurement of ordering in the normal state of underdoped Ba(Fe 1 x Co x ) 2 As 2 and Fe 1+y Te AFM orthorhombic SC Chu et al. PRB 79, (2009)

26 Agreement with published data for T<T c K Exp Data 1 gap fit, 5.45mV 2 gap fit, 5.0mV, 9.0mV di/dv (1/Ω) G (1/Ω) T = 7.3 K V bias (mv) Exp Data 2 band BTK Fit Δ 1 =4.99mV Γ 1 =1.8mV Z 1 =0.41 Δ 2 =9.95mV Γ 2 =2.25mV Z 2 =0.435 w=0.5 Tortello et al. PRL 105, (2010) V bias (mv) Βα(Φε 0.92 Χο 0.08 ) 2 Ασ 2

27 PCS on Parent Compound: BaFe 2 As 2 Increasing T Raw Data

28 Raw Data

29 Electronic Orthorhombicity Fisher Group: Science 329, 824 (2010).

30 ARPES STM Shen Group: arxiv: v1 Davis Group: Science 327, 181 (2010) PCS sensitive to orbital ordering?

31 New line in the Phase Diagram T S, T M determined from peaks in dr/dt

32 Quasiparticle scattering / Point Contact Spectroscopy 1. Superconductivity (intro): Gap, order parameter symmetry, density of state, this is well understood in the framework of BTK and Andreev reflection. 2. Kondo Impurity / Kondo Lattice effects (intro): Fano lineshape consistent with the growth of the Kondo liquid (starts at T*, and grows until Tc): microscopic explanation? 3. Antiferromagnetic ordering (intro): Conductance enhancement at T N in some AF systems (not all) 4. The hybridization gap & Fano resonance in a heavy fermion This talk: Helps explain #2. 5. Orbital ordering? This talk: Enhancement at high-t in Co:BaFe 2 As2and FeTeSe. Thank You!

33 T onset reflected in zero bias conductance

34 Similar spectra observed when Co doping is low enough for magnetic order to exist (underdoped regime)

35 Fe 1+y Te shows similar curves

36 Other experimental techniques that have observed signals above T S : ARPES: Splitting between the d xz and d yz bands that develops above the magnetic transition temperature. Yi et al. arxiv: v1 (to be published PNAS) Inelastic Neutron Scattering: Anisotropic high energy spin excitations in the paramagnetic phase. Harriger et al. arxiv: v1 Resistivity: In plane resistive anisotropy above T S. Chu et al. Science 329, 824 (2010); Optical Conductivity: Anisotropic charge dynamics above T S. Dusza et al. arxiv: v1 These signals are attributed to unequal occupation of the d xz and d yz bands, leading to orbital ordering above T M. The unequal occupation is speculated to occur due to spin excitations or some sort of instability preferring one band over the other.

37 Modified Phase Diagram & Conclusion AFM orthorhombic SC % Co T N T S T onset T c_mid K 132 K 170±10 K K 119 K 165±10 K K 107 K 160±10 K 5 70 K 78 K 155±15 K 8.9 K K 75 K 145± K PCS is detecting K a signal 8 only for samples with 24 K magnetism. However, the signal survives well above T N. We speculate that orbital ordering might be responsible for our conductance spectra K

38 Antiferromagnetism (& Kondo): CeMIn K CeCoIn K CeIrIn 5 di/dv (Ω -1 ) K 57.9 K Voltage (mv) Asymmetry: robust & reproducible in pure and Hg- & Cd-doped Co: asymmetry follows HF spectral weight qualitatively (two-fluid model, Nakatsuji-Pines-Fisk). Rh: enhancement due to AFM di/dv (Ω -1 ) di/dv (Ω -1 ) K 48.4 K K Voltage (mv) CeRhIn K 15.1 K 35.9 K Voltage (mv)

39 Quasiparticle Tunneling into a Kondo Lattice Yang, PRB (2009) Figgins & Morr, PRL (2010) Fogelström, Park, Greene, Goll, Graf, PRB (2010) Wölfle, Dubi, Balatsky, PRL (2010) q F : Fano asymmetry parameter ( A/B; A = probability for tunneling into heavy bands, B = into cond. band) W = πn(0)v 2 : width of the Kondo resonance λ: renormalized f-level -D 1, D 2 : cond. band edges; 2D = D 1 + D 2 : band width If D 1 = D 2, Δ hyb = 2V 2 / D (hybridization gap) t c : amplitude of direct tunneling into cond. band Fano resonance!

40 Conjecture on a Fano Resonance Data taken from different junctions, showing a systematic variation. Asymmetric double-peak structure is reproducibly observed. Positive-bias conductance peak is always higher ( q F > 0). V min = -3 ~ -0.5 mv < T << T HO These observations lead us to conjecture on a Fano resonance in a Kondo lattice, as predicted by Maltseva-Dzero-Coleman (PRL 2009). Interference between channels into the hybridized heavy bands (A) and the conduction band (B). q F A/B.

41 Analysis Using a Fano Resonance Model a c di di di = + ω dv dv dv FR bg b d Fano resonance Background (Maltseva et al., 2009) Assume a parabolic background Energy-dep. quasiparticle broadening due to correlation effects, γ(e) (Wölfle et al., PRL, 2010) Fig. # a b c d T (K) R J (W) q F Δ hyb (mev) V (mev) λ (mev) Average Δ hyb = 13 mev, consistent with recent optical spectroscopy results by Levallois et al. (arxiv: )

42 Relation to Gaps from Other Measurements R = R + at + b T Δ + T Δ Δ T 2 0 ( / )[1 2 / ]exp( / ) Fitting gives rise to Δ = 6.7 mev Δ hyb /2. 1.5% Rh-doped (T HO = 12.8 K): Δ = 4.7 mev, Δ hyb = 10 mev. Δ Δ hyb /2. Finite conductance within Δ hyb indicates finite DOS (e.g., due to correlation effects) with the chemical potential close to the gap center. Δis approximately half of Δ hyb.