Non-Fermi Liquids and Bad Metals in NdNiO3 Thin Films

Transcription

1 Non-Fermi Liquids and Bad Metals in NdNiO3 Thin Films Susanne Stemmer Materials Department University of California, Santa Barbara Workshop on Bad Metal Behavior in Mott Systems Schloß Waldthausen, Germany June 9, 5

2 Acknowledgements Graduate students and postdocs: Evgeny Mikheev Adam Hauser Nelson Moreno Jinwoo Hwang (now at OSU) Jack Zhang Junwoo Son (now at Postech) Collaborators: Burak Himmetoglu, Chris Van de Walle (UCSB Materials) Funding

3 Quantum Criticality Temperature Non-Fermi Liquid QCP Fermi Liquid Unconventional superconductors Quantum Hall systems Heavy-fermions Quantum magnets Cold atom gases Tuning parameter Non-thermal Continuous phase transition at zero-temperature Quantum fluctuations influence the material over a wide range of temperatures and across phase diagram A non-thermal control parameter causing large changes in properties: metal-insulator transitions, magnetic transitions,... 3

4 Quantum Criticality Resistivity ~ AT n Temperature Non-Fermi Liquid QCP Fermi Liquid Resistivity ~ AT Non-Fermi liquid behavior: Tuning parameter Poorly understood power laws in transport coefficients Non-saturating resistances that escalate past the Mott-Ioffe- Regel limit: Strange Metals Strong temperature dependence of the Hall coefficient (not reflecting a real change in Fermi surface)

5 Bad Metals Mott-Ioffe-Regel limit: mean free path length ~ lattice spacing Neither saturating nor nonsaturating metals are understood P. B. Allen, Physica B 38, -7 (); O. Gunnarsson, et al., Nature 5, 7 (); B. Chakraborty, and P.B. Allen, Phys. Rev. Lett., (979); M. Calandra, and O. Gunnarsson, Phys. Rev. B, 55 (); P. B. Allen, Nature 5, 7 (); Millis, A.J., Hu, J., & Sarma, S.D., Phys. Rev. Lett. 8, (999); O. Gunnarsson, et al., RMP 75, 85 (3); N. E. Hussey, K. Takenaka, & H. Takagi, Phil. Mag. 8, 87 (). Resistivity Non-saturating Bad metal (cuprate) Temperature Saturating ρ MIR = 3 π ħ q k F a Mott-Ioffe-Regel limit Mott-Ioffe-Regel limit Saturating metals are described by: Electron-electron scattering: n ~ ρ = + ρ + AT n ρ sat Non-Fermi liquids: n < ρ: residual resistance (disorder) Saturation resistance connected in parallel, serves to reduce resistance

6 This talk: NdNiO3 Substrate Non-Fermi liquid behavior in strained NdNiO3 thin films: NFL power law coefficients in resistivity Quantum critical point Bad metal behavior Vary strain and confinement (thickness): relationship to electronic structure

7 Metal-Insulator Transitions in RNiO3 5 Paramagnetic metal NdNiO3 LaNiO3 Paramagnetic insulator Antiferro magnetic insulator La Nd Pr Nd Pr Sm Eu J. B. Torrance, et al., Phys. Rev. B 5, 89 (99). J. M. Rondinelli, S. J. May, and J. W. Freeland, MRS Bulletin (March ). Resistivity (mω cm) NdNiO3 film A. J. Hauser, et al., Appl. Phys. Lett., 9 (5). Quantum phase transition Size of the rare earth ion is the tuning parameter: band-width driven metal-insulator transition Expect the transition to be sensitive to strain Is the transition quantum critical? Non-Fermi liquid behavior? 7

8 Orbital Engineering in RNiO3 R. Eguchi, et al., Phys. Rev. B 79, 5 (9).

9 Orbital Engineering in RNiO3 A. Frano, Spin Spirals and Charge Textures in Transition-Metal-Oxide Heterostructures (Ph.D. thesis, ), Springer. R. Eguchi, et al., Phys. Rev. B 79, 5 (9). O. E. Peil et al., Phys. Rev. B 9, 58 (); E. Benckiser, et al., Nat. Mater., 89 (); J.W. Freeland, et al., EPL 9, 57 (); A.V. Boris, et al., Science 33, 937 (); M. Wu et al., Phys. Rev. B 88, 5 (3); H. K. Yoo, et al., Sci. Rep. 5, 87 (5); S. B. Lee et al., Phys. Rev. Lett., 5 (). Tensile strains and confinement favor the large hole surface Promotes a spin density wave instability and insulating state Confinement and tensile strains have qualitatively similar effects 9

10 Strain as a Tuning Parameter ε xx (%) mtorr mtorr YAlO LaAlO a sub (Å) NdGaO 3 LSAT SrTiO DyScO 3. T MIT (K) A. J. Hauser, et al., Appl. Phys. Lett., 9 (5). Epitaxial strain systematically shifts the MIT 5 Metals at all temperatures - - ε xx (%) Insulator at all temperatures MIT is relatively independent of deposition conditions, which affect the stoichiometry Low-pressure films can be strained to larger strains without relaxing

11 Orbital Engineering in RNiO3 A. J. Hauser, et al., Appl. Phys. Lett., 9 (5). Tensile R H ( - cm 3 /C) T - T Crossover (K) Compressive NdGaO 3 9 mtorr mtorr LaAlO 3 mtorr A. Frano, Spin Spirals and Charge Textures in Transition-Metal-Oxide Heterostructures (Ph.D. thesis, ), Springer. O. E. Peil et al., Phys. Rev. B 9, 58 (); E. Benckiser, et al., Nat. Mater., 89 (); J.W. Freeland, et al., EPL 9, 57 (); A.V. Boris, et al., Science 33, 937 (); M. Wu et al., Phys. Rev. B 88, 5 (3); H. K. Yoo, et al., Sci. Rep. 5, 87 (5); S. B. Lee et al., Phys. Rev. Lett., 5 (). Tensile strains favor the large hole surface Promotes a spin density wave instability and the insulating state

12 Strain as a Tuning Parameter 3 NNO/DSO (+.9%) 35 5 Insulator at all temperatures T MIT (K) 5 on DyScO Metals at all temperatures - - ε xx (%) A. J. Hauser, et al., Appl. Phys. Lett., 9 (5). Likely more complex reasons for films with insulating state at all temperature than simple band width tuning Role of disorder? Criterion for Anderson transition in this system?

13 Metal-Insulator Transitions in NdNiO3 on YAlO3 NNO/LAO (-.%) 3 NNO/NGO (+.85%) on NdGaO NNO/YAO (-3. %) u.c. u.c. 8 u.c. u.c. 5 u.c on LaAlO NNO/LSAT (+.9%) NNO/STO (+.7%) NNO/DSO (+.9%) 3 on LSAT on SrTiO3 5 5 on DyScO E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el]

14 Metal-Insulator Transitions in NdNiO on YAlO3 NNO/LAO (-.%) NNO/YAO (-3. %) u.c. u.c.. 8 u.c. u.c. on LaAlO3-5 u.c of thickness and strain Complex behavior as a function NNO/LSAT (+.9%) on LSAT NNO/STO (+.7%) NNO/NGO (+.85%) on NdGaO3 5 NNO/DSO (+.9%) on SrTiO3 5 5 on DyScO E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el]

15 MITs and Non-Fermi Liquids ρ = + ρ + AT n ρ sat.. dρ/dt (mohmcm/k) n=5/3, no ρ SAT n=, no ρ SAT n=5/3, with ρ SAT ρ NFL (mohmcm) ρ + AT n T (K) T 5/3 (K 5/3 ) E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el] Metallic phase exhibits saturating resistance Need to take into account to get correct NFL exponent 5

16 MITs and Non-Fermi Liquids R = R + AT n 3 n ~. (NFL) 3 n ~. (NFL) MIT << 5 K R/R( K) R/R( K) T (K) T (K) 3 n = (FL) MIT ~ 5 K 3 ρ > ρ sat Anderson Insulator R/R( K) R/R( K) T (K) T (K) E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el]

17 MITs and Non-Fermi Liquids ρ = + ρ + AT n ρ sat. Non-Fermi liquid exponent Substrate: YAO LAO NGO LSAT STO 5 5 n = n = 5/3 E. Mikheev, et al., arxiv: 57.9 [cond-mat.str-el] t(nno) (u.c.) Non-Fermi liquid behavior only if MIT is suppressed Same exponent (5/3) across the entire sample series n is independent of disorder Need to take resistance saturation into account to get correct n 7

18 Phase Diagram ρ = + ρ + AT n ρ sat 5 u.c. PM (NFL) u.c. PM (NFL) 8 u.c. PM (NFL) u.c. PM (FL) AFI PM (FL) AFI PM (FL) PM (FL) AFI Quantum critical point shifts to more compressive strains with decreasing film thickness Confinement promotes x -y orbital polarization, which favors spin density wave and the AFM insulator Suppression of MIT leads to NFL behavior E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el] AFI thickness u.c. - - AFI A. Frano, Spin Spirals and Charge Textures in Transition-Metal-Oxide Heterostructures (Ph.D. thesis, ), Springer. compressive tensile Strain (%)

19 Bad Metals Mott-Ioffe-Regel limit: mean free path length ~ lattice spacing Neither saturating nor nonsaturating metals are understood P. B. Allen, Physica B 38, -7 (); O. Gunnarsson, et al., Nature 5, 7 (); B. Chakraborty, and P.B. Allen, Phys. Rev. Lett., (979); M. Calandra, and O. Gunnarsson, Phys. Rev. B, 55 (); P. B. Allen, Nature 5, 7 (); Millis, A.J., Hu, J., & Sarma, S.D., Phys. Rev. Lett. 8, (999); O. Gunnarsson, et al., RMP 75, 85 (3); N. E. Hussey, K. Takenaka, & H. Takagi, Phil. Mag. 8, 87 (). Resistivity Non-saturating Bad metal (cuprate) Temperature Saturating ρ MIR = 3 π ħ q k F a Mott-Ioffe-Regel limit Mott-Ioffe-Regel limit Saturating metals are described by: Electron-electron scattering: n ~ ρ = + ρ + AT n ρ sat Non-Fermi liquids: n < ρ: residual resistance (disorder) Saturation resistance connected in parallel, serves to reduce resistance

20 Saturating Metallic Phase In the -K limit: = + ρ (K) ρ ρ sat. 5 u.c. ρ( K) ρsat ρ ρ MIR = 3 π ħ q k F a. - - ρmir ρsat ρ(k) ρ SAT YAO LAO NGO LSAT STO ρ ( K) YAO LAO NGO LSAT STO ε xx (%) t(nno) (u.c.) ρsat and ρ depend on the magnitude of the mismatch strain ρsat is independent of thickness (confinement, disorder) ρ strongly depends on thickness E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el]

21 Saturation and Orbital Polarization Δ (ev) u.c ρ sat (mωcm) ρsat depends on the magnitude of orbital polarization A parallel channel due to interband scattering Conversely, predict that materials with a single band at the Fermi level show no saturation (i.e., cuprates) ε xx (%) Consistent with theoretical predictions (modified Boltzmann transport models that account for interband scattering and QMC simulations) P. B. Allen, Physica B 38, -7 (); O. Gunnarsson, et al., Nature 5, 7 (); B. Chakraborty, and P.B. Allen, Phys. Rev. Lett., (979); M. Calandra, et al., Phys. Rev. B, 55 (). ρsat is tunable by orbital polarization E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el]

22 Criterion for the Anderson Insulator : ρsat ~ ρ ρsat ρ(k) ρ SAT YAO LAO NGO LSAT STO ρ ( K) YAO LAO NGO LSAT STO ρ = + ρ + AT n ρ sat 5 5 t(nno) (u.c.) ρsat is independent of thickness but depends on strain ρ depends on thickness Predict that material is insulating at all temperatures when ρsat ~ ρ Correct to within u.c. This Anderson insulator is strain and thickness tunable E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el]

23 MITs and Non-Fermi Liquids ρ = + ρ + AT n ρ sat 3 n ~. (NFL) 3 n ~. (NFL) MIT << 5 K R/R( K) R/R( K) T (K) T (K) 3 n = (FL) MIT ~ 5 K 3 ρ > ρ sat Anderson Insulator R/R( K) R/R( K) T (K) T (K) 5

24 Phase Diagram Thickness (u.c.) 5 5 PM (NFL) PM (NFL) AFI PM (FL) AFI Anderson Insulator - - ρsat ~ ρ Strain (%) E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el]

25 Summary Single non-fermi liquid exponent (5/3) across the entire series, if MIT is suppressed Non-Fermi liquid phase or quantum critical point? Need to take resistance saturation into account to get correct NFL exponent NdNiO3 is a saturating metal, but Ioffe Regel Limit is exceeded Degree of bad metal behavior depends on orbital polarization (strain) Large orbital polarization increases ρsat and makes the material increasingly non-saturating resistivity escalates past the Mott-Ioffe-Regel limit Can be tuned by strain and confinement Metals with large degeneracy have large ρsat and are thus non-saturating Quantitative understanding of the role of electronic structure in strongly correlated phenomena is desirable New routes to controlling MITs: Confinement stabilizes spin-density wave and insulating state modulate confinement Strain can modulate transition between Anderson insulator and metal E. Mikheev, et al., arxiv:57.9 [cond-mat.str-el] 5

26 Thank you