Title: Cluster Dynamical Mean Field Approach to Strongly Correlated Materials
1Cluster Dynamical Mean Field Approach to Strongly
Correlated Materials
- K Haule
- Rutgers University
2References and Collaborators
- Strongly Correlated Superconductivity a
plaquette Dynamical mean field theory study, K.
H. and G. Kotliar, Phys. Rev. B 76, 104509
(2007). - Nodal/Antinodal Dichotomy and the Energy-Gaps of
a doped Mott Insulator, M. Civelli, M. Capone, A.
Georges, K. H., O. Parcollet, T. D. Stanescu, G.
Kotliar, Phys. Rev. Lett. 100, 046402 (2008). - Modelling the Localized to Itinerant Electronic
Transition in the Heavy Fermion System CeIrIn5,
J.H. Shim, K. Haule and G. Kotliar, Science 318,
1615 (2007), - Quantum Monte Carlo Impurity Solver for Cluster
DMFT and Electronic Structure Calculations in
Adjustable Base, K. H., Phys. Rev. B 75, 155113
(2007). - Optical conductivity and kinetic energy of the
superconducting state a cluster dynamical mean
field study, K. H., and G. Kotliar, Europhys
Lett. 77, 27007 (2007). - Doping dependence of the redistribution of
optical spectral weight in Bi2Sr2CaCu2O8delta,
F. Carbone, A. B. Kuzmenko, H. J. A. Molegraaf,
E. van Heumen, V. Lukovac, F. Marsiglio, D. van
der Marel, K. H., G. Kotliar, H. Berger, S.
Courjault, P. H. Kes, and M. Li, Phys. Rev. B 74,
064510 (2006). - Avoided Quantum Criticality near Optimally Doped
High Temperature Superconductors, K.H. and G.
Kotliar, Phys. Rev. B 76, 092503 (2007).
Thanks to Ali Yazdani for unpublished data!
3Standard theory of solids
Band Theory electrons as waves Rigid band
picture En(k) versus k Landau Fermi Liquid
Theory applicable Very powerful quantitative
tools LDA,LSDA,GW
- Predictions
- total energies,
- stability of crystal phases
- optical transitions
4Strong correlation Standard theory fails
- Fermi Liquid Theory does NOT work . Need new
concepts to replace rigid bands picture! - Breakdown of the wave picture. Need to
incorporate a real space perspective (Mott). - Non perturbative problem.
5Non perturbative methods
On site correlations usually the strongest
-gt Mott phenomena at integer fillings
Successful theory which can describe Mott
transition Dynamical Mean Field Theory
6Mott phenomena at half filling
Single site DMFT
Georges, Kotliar, Krauth, Rozenber, Rev. Mod.
Phys. 1996
7Dynamical Mean Field Theory
For a given lattice site, DMFT envisions the
neighboring sites on the lattice as a Weiss
field of conduction electrons, exchanging
electrons with that site.
Maps lattice model to an effective quantum
impurity model
More rigorously DMFT sumps up all local diagrams
(to all orders in perturbation theory)
8DMFT in single site approximation
Successfully describes spectra and response
functions of numerous correlated materials
Mott transition in V2O3 LaTiO3 actinides
(Pu,) Lanthanides (Ce,) and far to many to
mention all
, KH et al. 2007
, KH et al. 2003
Recent review
(G. Kotliar S. Savrasov K.H., V. Oudovenko O.
Parcollet and C. Marianetti, RMP 2006).
9Band structure and optics of heavy fermion CeIrIn5
Non-f spectra at 10K
300K
J.H. Shim, KH, and G. Kotliar, Science 318, 1618
(2007).
10Later verified by Yang Pines
Remarkable agreement with Y. Yang D. Pines
cond-mat/0711.0789!
11High Tc Need non-local self-energy
d-wave pairing 2x2 cluster-DMFT necessary to
capture the order parameter
Fermi surface evolution with doping can not be
understood within single site DMFT.
12Cluster DMFT approaches
- Momentum space approach-Dynamical cluster
approximation - (Hettler, Maier, Jarrel)
- Real space approach Cellular DMFT
(Kotliar,Savrasov,Palson)
In the Baym Kadanoff functional, the interacting
part F is restrictied to the degrees of freedom
(G) that live on the cluster.
Maps the many body problem onto a self consistent
impurity model
FGplaquette
- Impurity solvers
- Continuous time QMC
- Hirsh-Fye QMC
- NCA
- ED
periodization
13An exact impurity solver, continuous time QMC -
expansion in terms of hybridization
K.H. Phys. Rev. B 75, 155113 (2007) P Werner,
PRL (2007) N. Rubtsov PRB 72, 35122 (2005).
General impurity problem
Diagrammatic expansion in terms of hybridization
D Metropolis sampling over the diagrams
- Exact method samples all diagrams!
- No severe sign problem
14Approach
- Understand the physics resulting from the
proximity to a Mott insulator in the context of
the simplest models.
- Construct mean-field type of theory and follow
different states as a function of parameters
superconducting normal state. - Second step compare free energies which will
depend more on the detailed modeling and long
range terms in Hamiltonian..
- Approach the problem from high temperatures where
- physics is more local. Address issues of finite
frequency and finite temperature crossovers.
- Leave out disorder, electronic structure, phonons
- CDMFTLDA second step, under way
15S(iw) with CTQMC
next nearest neighbor important in underdoped
regime
on-site largest
nearest neighbor smaller
Hubbard model, T0.005t
16Momentum space differentiation
gets replaced by coherent SC state
with large anomalous self-energy
t-J model, T0.005t
17SC Tunneling DOS
K. H. and G. Kotliar, Phys. Rev. B 76, 104509
(2007).
NM d0.20
SC d0.08
SC d0.20
NM d0.08
Large asymmetry at low doping
Gap decreases with doping DOS becomes more
symmetric
Asymmetry is due to normal state DOS -gt Mottness
Computed by the NCA for the t-J model
18Ratio AS/AN
Ratio more universal, more symmetric
With decreasing doping gap increases, coherence
peaks less sharp-gtNon BCS
19Ratio AS/AN
CDMFT calculation
ratio almost symmetric
Pronounced dip-hump feature can not be fitted
with BCS
20Non-BCS
DOS in normal state decreases
Gap increases
21Normal state DOS and SC gap
CDMFT
Not using realistic band structure (t)
22Gap changes, mode does not
J.C. Davis, Nature 442, 546 (2006)
23Where does the dip-hump structure come from?
24Eliashberg theory
Real part constant
phonon frequency
W
D0
Gap up to D0W
No scattering up to D0W
25Dip-hump structure
26Phenomenology
Similar frequency dependence of gap recently
introduced by W.Sacks and B. Doucot PRB
74,174517 (2006) to fit experiments.
27Fermi surface
d0.09
Cumulant is short in ranged
Arcs FS in underdoped regime pocketslines of
zeros of G arcs
Single site DMFT PD
28Nodal quasiparticles
29Nodal quasiparticles
Vnod almost constant up to 20
the slopevnod almost constant
vD dome like shape
Superconducting gap tracks Tc!
M. Civelli, cond-mat 0704.1486
30Two energy scales in Raman Spectrum in the SC
State of Underdoped Cuprates
doping
Energy scale of peak in antinodal (nodal) region
increases (decreases) with decreasing doping in
underdoped cuprates.
Le Tacon et al, Nat. Phys. 2, 537 (2006)
31Evolution of Nodal and Antinodal energy scales
with doping
Le Tacon et al, Nat. Phys. 2, 537 (2006)
32Antinodal gap two gaps
M. Civelli, using ED, PRL. 100, 046402 (2008).
33Optical conductivity
- Low doping two components
- Drude peak MIR peak at 2J
- For xgt0.12 the two components merge
- In SC state, the partial gap opens causes
- redistribution of spectral weight up to 1eV
34Optical spectral weight - Hubbard versus t-J model
f-sumrule
Excitations into upper Hubbard band
Drude
t-J model
J
no-U
35Optical spectral weight Optical mass
mass does not diverge approaches 1/J
Bi2212
F. Carbone,et.al, PRB 74,64510 (2006)
Weight increases because the arcs increase and Zn
increases (more nodal quasiparticles)
Basov et.al., PRB 72,60511R (2005)
36Temperature/doping dependence of the optical
spectral weight
37Optical weight, plasma frequency
Weight bigger in SC, K decreases (non-BCS)
Weight smaller in SC, K increases (BCS-like)
A.F. Santander-Syro et.al, Phys. Rev. B 70,
134504 (2004)
F. Carbone,et.al, PRB 74,64510 (2006)
38Kinetic energy change
Kinetic energy increases
cluster-DMFT, Eu. Lett. 77, 27007 (2007).
Kinetic energy decreases
Phys Rev. B 72, 092504 (2005)
Kinetic energy increases
Exchange energy decreases and gives largest
contribution to condensation energy
same as RVB (see P.W. Anderson Physica C, 341, 9
(2000)
39Origin of the condensation energy
- Resonance at 0.16t5Tc (most pronounced at
optimal doping) - Second peak 0.38t120meV (at opt.d)
substantially contributes to condensation energy
40Conclusions
- Plaquette DMFT provides a simple mean field
picture of the - underdoped, optimally doped and overdoped regime
- One can consider mean field phases and track them
even in the - region where they are not stable (normal state
below Tc) - Many similarities with high-Tcs can be found in
the plaquette DMFT - Strong momentum space differentiation with
appearance of arcs in UR - Superconducting gap tracks Tc while the PG
increases with underdoping - Nodal fermi velocity is almost constant
- Tunneling DOS As/An has a dip hump dip structure
-gt comes from the structure - in the anomalous self-energy
- Optical conductivity shows a two component
behavior at low doping - Optical mass 1/J at low doping and optical weigh
increases linearly with d - In the underdoped system -gt kinetic energy saving
mechanism - overdoped system -gt kinetic energy loss
mechanism - exchange energy is always optimized in SC
state