ARPES investigation of the bad metallic behavior in Fe 1.06 Te

Transcription

1 ARPES investigation of the bad metallic behavior in Fe 1.06 Te Véronique Brouet, Ping-Hui Lin, Maria Fuglsang Jensen Yoan Texier, David Le Bœuf, Joseph Mansart Laboratoire de Physique des Solides d Orsay SOLEIL synchrotron, CASSIOPEE beamline Amina Taleb-Ibrahimi, Patrick Le Fèvre, François Bertran Sample synthesis FeTe: Enrico Giannini, University of Geneva, Swizerland Pnictides: Dorothée Colson, Anne Forget, Florence Rullier- Albenque SPEC, CEA-Saclay, France

2 A new familyof correlatedsystems LiFeAs BaFe 2 As 2 FeTe Effective masses calculated by DMFT Yin, Haule, Kotliar, Nature Materials 2011

3 A new familyof correlatedsystems Effective masses calculated by DMFT FeTe BaFe 2 As 2 Rullier-Albenque et al., PRL 2009 Yin, Haule, Kotliar, Nature Mat FeTe is the most correlated case and behavesas a «badmetal» quite differentfrombafe 2 As 2. Wang et al., New J. Phys Chen et al., PRB 2009 Moon et al., PRL 2012

4 Outline WhatcanwelearnwithARPES about the strengthand nature of electronic correlations? How do they evolve from family to family? => Imaging the differentbands of the electronic structure and defining renormalization values for each orbital Orbital differentiation in LiFeAs => Characterizingthe «badmetallicstate» in FeTe. There isa large «pseudogap» (60meV) in the paramagnetic state of FeTe. Spin-freezing regime? Are pseudogaps ubiquitous in pnictides?

5 Electronicstructure studiedby ARPES hv Z Electron analyser θ e - E X φ = hν W Crystal kin E B hk = 2mEkin sinθ Y CASSIOPEE beamline, SOLEIL synchrotron => ev => high energy and angular resolution

6 Genericelectronicstructure expected for iron-basedsuperconductors Fermi Surface Band dispersion along the diagonal Electrons X X Γ Holes X X Γ X

7 Electronicstructure viewedby ARPES Ba(Fe 0.92 Co 0.08 ) 2 As 2 Hole pockets Electron pockets LiFeAs Hole pockets Electron pockets Modulation of intensities=> V. Brouet et al. PRB 2012, L. Moreschiniet al. PRL 2014

8 ARPES dispersions vsband calculations Ba(Fe 0.92 Co 0.08 ) 2 As 2 (kz=0.7) LiFeAs(kz=0) NB : Band calculationsare shownrenormalizedby a factor 2. => «modest» correlations «Shrinking» of FS => Smaller hole and electron pockets => Driven by interband transitions => Brouet et al., PRL 13 Orbital polarization => Holes are transferred from dxz/dyz to dxy

9 Renormalizationvalues Ba(Fe 0.92 Co 0.08 ) 2 As 2 (kz=0.7) Global renormalization of meV shift LiFeAs(kz=0) Global renormalization of 2 With individual shift and normalization V. Brouet et al., PRL 2013 DMFT ~ 2,5 for both 2,5 for dxz/dyzand 3 for dxy => In LiFeAs, dxy and dxz/dyz orbitals start to differentiate

10 The case of FeTe Fe 1.06 Te Resistivity Much larger magnetic moment than Fe pnictides => M=2,1µ B Bao, PRL 2009 «Double stripe» magnetic structure NOT corresponding to FS nesting q AF

11 Previousstudiesof FeTe(T>T N ) Zhang, DL Feng et al. PRB 2012 Liu, ZX Shenet al PRL 2013

12 70eV FeTeFermi Surface (80K) A dxz/dyz hole pockets dxz/dyz electron pocket LiFeAs No trace of d xy band Strange shape for electron pocket! Hole pockets Electron pockets

13 FeTeFermi Surface (80K) dxz/dyz hole pockets dxz/dyz electron pocket No trace of d xy band Strange shape for electron pocket! Dotted white lines: band calculation shifted and renormalized by 2

14 FeTe: evolutionof the bands with temperature T N P.H. Lin, V. Brouet et al., PRL 2013

15 Transition to the magneticstate 80 K 20 K Splittingof LEED spot vs T LEED evidences a transition to the magnetic monoclinic phase thattakesplace similarlyas in the bulk.

16 Evolution of the electronicstructure in the magneticstate CalculationdoneusingWien2K package, converging to a moment M=2.2µ B. AF side ferro side diagonal Q AF doesnot correspond to FS nesting No gap opening(no FS nesting) Large shifts of the bands Rehybridizations into new conduction channels.

17 FeTe: evolutionof the bands with temperature Energy (ev)

18 Hole/electronpocketnearGamma Liu et al PRL 2013

19 FeTe: evolutionof the bands with temperature Energy (ev)

20 «Pseudogap» onthe electronpocket 80K 60K 40K 20K Spectra at k F

21 Good agreement withtransport Good metallic features in the magnetic state contrary to PM state Dominatedby electrons in the magneticstate contraryto PM state (ok withr H ). Change of the signof R H Chen et al., PRB 2009

22 Can itbeunderstoodtheoretically? «Hund s metals» Specific heat coefficient γ vs J J not U isresponsiblefor the mass enhancement Hauleand Kotliar, New J. Of physics2009 Electronic correlations are usually governed by Coulomb repulsion U In iron-based superconductors Hund s couplings play a major role. controlled by U Fe 1 Fe 2 New type of correlatedstate?

23 Hund scouplingsenlargeregionsof «badmetallic» behaviors Quasiparticle weightas a functionof U and J L. demedici, J. Mravlje, A. Georges, PRL 2011 J decreases the quasiparticle weight but destabilizes the Mott transition. Non Fermi liquid behaviors, Spin freezing regime Werner et al., PRL 2008

24 Is a pseudogapexpectedin the «spin freezing» regime? FeTeisexpectedto bea metalswithlowcoherencetemperature. Fractional power laws for the self-energy. DMFT simulations for FeTe at various temperatures Yin, Haule, Kotliar, PRB 2012

25 Is a pseudogapexpectedin the «spin freezing» regime? FeTeisexpectedto bea metalswithlowcoherencetemperature. Fractional power laws for the self-energy. DMFT simulations for FeTe at various temperatures For Liebsch, J opens a pseudogap associated to a collective mode in the selfenergy due to spin-fluctuations. Bare DOS Yin, Haule, Kotliar, PRB 2012 Liebsch, PRB 2011

26 (π, π) and (π, 0)fluctuations One characteristic of the FeTe pseudogap is that it only develops on the electron pocket. (π, π ) fluctuations (FePn) (π, 0) fluctuations (FeTe) 80K X (π,π) Γ X X X 20K

27 Are «pseudogaps» a generalfeatureof iron-basedsuperconductors? For Xu et al. (ARPES study), there isa 18meV PG in Ba 0.75 K 0.25 Fe 2 As 2 related to (π, π) fluctuations. For Moon et al. (infraredstudy), thereisa PG in underdoped Ba(Fe,Co) 2 As 2, alsoconnectedto (π, π) fluctuations. Xu et al. Nat Com 2011 Moon et al. PRL 2012

28 Conclusions Correlationsdeveloppreferrentiallyin dxy, whichismore renormalized in LiFeAs than dxz/dyz and not detected in FeTe A large pseudogap of 60meV on the dxz/dyz elctron pocket characterizesthe «badmetallic» phase of FeTe It disappears in the magnetically ordered phase. It mayberelatedto (π,π) fluctuations and alreadypresentin other iron pnictides It maybeemphasizedin FeTedue to the largerfrozenmagnetic moments and the incoherence of dxy. Thanksto : L. de Medici, S. Bierman, M. Casula, A. Georges and M. Grioni