Entanglement Spectroscopy and its application to the fractional - - PowerPoint PPT Presentation

entanglement spectroscopy and its application to the
SMART_READER_LITE
LIVE PREVIEW

Entanglement Spectroscopy and its application to the fractional - - PowerPoint PPT Presentation

Entanglement Spectroscopy and its application to the fractional quantum Hall phases N. Regnault Ecole Normale Sup erieure Paris and CNRS, Department of Physics, Princeton Global Scholar C O L E N O R M A L E S U P R I E U R E


slide-1
SLIDE 1

Entanglement Spectroscopy and its application to the fractional quantum Hall phases

  • N. Regnault

Ecole Normale Sup´ erieure Paris and CNRS, Department of Physics, Princeton Global Scholar

É C O L E N O R M A L E S U P É R I E U R E

slide-2
SLIDE 2

Acknowledgment

  • A. Sterdyniak (MPQ, Germany)
  • B. Estienne (Universit´

e Pierre et Marie Curie, France)

  • Z. Papic (University of Leeds)

F.D.M. Haldane (Princeton University)

  • R. Thomale (W¨

urzburg, Germany)

  • M. Haque (Maynooth, Ireland)

A.B. Bernevig (Princeton University)

slide-3
SLIDE 3

Topological phases

What is topological order ? phases that can’t be described by a broken symmetry. No local order parameter. At least one physical (i.e. measurable) quantity related to a topological invariant (like the surface genus). A system with a gapped bulk and gapped or gapless surface (or edge) modes. Simplest example : the integer quantum Hall effect (quantized Hall conductance). Since 2005, the revolution of topological insulators 2D TI : 3D TI :

slide-4
SLIDE 4

Making things harder : strong interactions

Most celebrate example : the fractional quantum Hall effect. Alliance of a non-trivial band structure (Landau levels) and strong interactions. An exotic place : emergent fractional charges with fractional statistics or non-abelian.

Rxx Rxy B

No classification of fractional phases (as opposed to the non-interacting case). Non-perturbative problem → variational methods and numerical simulations. No local order parameter → which phase is emerging ?

slide-5
SLIDE 5

Outline

Entanglement Spectrum Fractional Quantum Hall Effect FQHE and Entanglement Spectrum ES and Fractional Chern Insulators

slide-6
SLIDE 6

Entanglement Spectrum

slide-7
SLIDE 7

Entanglement spectrum (Li and Haldane 2008)

Start from a quantum state |Ψ Create a bipartition of the system into A and B Reduced density matrix ρA = TrB |Ψ Ψ| Entanglement Hamiltonian : ρA = e−Hent A

B

1D:

A B

2D:

L

The eigenvalues of Hent are the entanglement energies {ξi}. Lower entanglement energies ≃ higher weights in ρA. If O = OA + OB and , the ξi can be labeled by the OA quantum numbers. Entanglement entropy SA = −TrA [ρA ln ρA], area law for gapped systems (i.e. SA ∝ Ld−1).

slide-8
SLIDE 8

Entanglement spectrum

Example : system made of two spins 1/2

B A

Entanglement spectrum : ξ as a function of Sz,A (z projection of the spin A) |↑↑

0.5 1 1.5 2

  • 1
  • 0.5

0.5 1

ξ Sz,A

1 √ 2 (|↑↓ + |↓↑)

0.5 1 1.5 2

  • 1
  • 0.5

0.5 1

ξ Sz,A

1 √ 4 |↑↓ +

  • 3

4 |↓↑

0.5 1 1.5 2

  • 1
  • 0.5

0.5 1

ξ Sz,A

The counting (i.e the number of non zero eigenvalue) also provides information about the entanglement

slide-9
SLIDE 9

The AKLT spin chain

A prototype of a gapped spin-1 chain. HAKLT =

  • j
  • Sj ·

Sj+1 + 1 3

  • j
  • Sj ·

Sj+1 2 The ground state of the AKLT Hamiltonian is the valence bond state.

A B

}

S=1

}

S=0

S=1/2 S=1/2

For an open chain, the two extreme unpaired spin- 1

2 correspond to

the edge excitations (4-fold degenerate ground state)

slide-10
SLIDE 10

The AKLT spin chain, the Li-Haldane conjecture

5 10 15 20 25 30 35

  • 3
  • 2
  • 1

1 2 3 4

ξ Sz,A

1 2 1

ξ Sz,A

AKLT Heis.

5 10 15 20 25 30 35

  • 3
  • 2
  • 1

1 2 3 4

ξ Sz,A

ξ

D

ES for an open chain with 8 sites and lA = 4. Sz,A : z-projection of A total spin. Reduced density matrix is 81 × 81 but only two non-zero eigenvalues for the AKLT model. The trace has introduced an artificial edge → a spin- 1

2 edge

  • excitation. The ES mimics the edge spectrum of the model.

Away from the model state : An entanglement gap ∆ξ between a low (entanglement) energy structure related the model state and a high energy structure. ∆ξ should stay finite at the thermodynamical limit if the two phases are in the same universality class.

slide-11
SLIDE 11

Fractional Quantum Hall Effect

slide-12
SLIDE 12

Fractional Quantum Hall effect

Landau levels (spinless case)

hwc hwc

N=0 N=1 N=2

Rxx Rxy B

Cyclotron frequency : ωc = eB

m ,

Filling factor : ν = hn

eB = N NΦ

Partial filling + interaction → FQHE Lowest Landau level (ν < 1) : zm exp

  • −|z|2/(4l2

B)

  • N-body wave function :

Ψ = P(z1, ..., zN) exp(− |zi|2/(4l2

B))

What are the low energy properties ? Gapped bulk, Massless edge Strongly correlated systems, emergence of exotic phases :fractional charges, non-abelian braiding.

slide-13
SLIDE 13

The Fractional QHE

FQHE is a hard N-body problem : a single Landau level (the lowest one for ν < 1, no spin) the effective Hamiltonian is just the (projected) interaction ! H = PLLL

  • i<j

V ( ri − rj) PLLL (insert in V your favorite interaction plus screening effect, finite width, Landau level,...) Two major methods : variational method : find a wave functions describing low energy physics (symmetries, CFT, model...) numerical calculation : exact diagonalizations on different geometries (sphere, plane, torus, ...), DMRG

Z

NF L =+3

z

L =+2

z

L =-3

z

L =-2

z

L =0

z

L =+1

z

L =-1

z

Nbr orb. ≃ NΦ

slide-14
SLIDE 14

The Laughlin wave function

A (very) good approximation of the ground state at ν = 1

3

ΨL(z1, ...zN) =

  • i<j

(zi − zj)3e−

i

|zi|2

4l2

x ρ

The Laughlin state is the unique (on genus zero surface) densest state that screens the short range (p-wave) repulsive interaction. Topological state : the degeneracy of the densest state depends on the surface genus (sphere, torus, ...)

slide-15
SLIDE 15

The Laughlin wave function : quasihole excitations

Add one flux quantum at z0 = one quasi-hole Ψqh(z1, ...zN) =

  • i

(z0 − zi) ΨL(z1, ...zN)

ρ x

Locally, create one quasi-hole with fractional charge +e

3

Quasi-holes obey fractional statistics (fractional charge + flux) Adding quasiholes/flux quanta increases the size of the droplet For given number of particles and flux quanta, there is a specific number of qh states that one can write These numbers/degeneracies can be classified with respect some quantum number (angular momentum Lz) and are a fingerprint of the phase (related to the statistics of the excitations).

slide-16
SLIDE 16

The torus geometry : topological degeneracy

N = 4 N=4 bosons F

k = 0 y

k = 1 y

k = 2 y

k = 3 y

k = y

2 1 1

1 2 3 0.02 0.04 0.06 0.08 0.1 2 4 6 8 10 12 14 16

Energy KT

y

Model interaction for the Laughlin state, N = 6, NΦ = 18 In the LLL, the one-body wf are :

  • k∈Z e

2π Ly (ky+kNφ)(x+iy)e− x2 2 e

− 1

2

Ly

2 (ky+kNφ)2

The Laughlin ν = 1/m is m-fold degenerate

  • n the torus.

Number of orbitals is Nφ. Each orbital is labeled by its quantum number ky. Invariant under the magnetic translations. K T

y = ( i ky,i) modNΦ.

There is another quantum number (purely many-body) related the center of mass degeneracy.

slide-17
SLIDE 17

The Haldane’s exclusion principle

The number of quasihole states per momentum sector can be predicted by a generalization of the Pauli’s principle. For the Laughlin ν = 1/m, no more than 1 particle in m consecutive orbitals (including periodic boundary conditions

  • n the torus).

Example Laughlin ν = 1/3 state with 9 flux quanta

k = y

1 0 0

1 2

1 0 0

3 4 5

1 0 0

6 7 8 k = y

1 0 0

1 2

0 1 0

3 4 5

1 0 0

6 7 8 k = y

0 1 0

1 2

0 1 0

3 4 5

0 1 0

6 7 8 k = y

1 0 0

1 2

0 1 0

3 4 5

0 1 0

6 7 8

Can be generalized to the Moore-Read or Read-Rezayi states (non-abelian excitations) : no more than k particles in k + 2 consecutive orbitals.

slide-18
SLIDE 18

The Haldane’s exclusion principle

Example : Finding back the 3-fold degeneracy of the Laughlin ν = 1/3 with Nφ = 3 × N = 9

k = y

1 0 0

1 2

1 0 0

3 4 5

1 0 0

6 7 8

K =0+3+6 mod 9=0 y

T k = y

0 1 0

1 2

0 1 0

3 4 5

0 1 0

6 7 8

K =1+4+7 mod 9=3 y

T k = y

0 0 1

1 2

0 0 1

3 4 5

0 0 1

6 7 8

K =2+5+8 mod 9=6 y

T 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 1 2 3 4 5 6 7 8 9

Energy KT

y

Laughlin GS

slide-19
SLIDE 19

The Laughlin wave function : edge excitations

B B

One dimensional chiral mode with a linear dispersion relation E ≃ 2πv

L n

The degeneracy of each many-body energy level Et is given by the sequence 1, 1, 2, 3, 5, 7, .... Non-interacting case (i.e. IQHE)

(a) E = 0 (b) E = 1 (c) E = 2 (d) E = 2

energy momentum t t t t

slide-20
SLIDE 20

The Laughlin wave function : edge excitations

B B

One dimensional chiral mode with a linear dispersion relation E ≃ 2πv

L n

The degeneracy of each many-body energy level Et is given by the sequence 1, 1, 2, 3, 5, 7, .... Interacting case (Laughlin ν = 1

3)

(a) E = 0 (b) E = 1 (c) E = 1 (d) E = 2 (e) E = 2

t t t t t energy momentum

slide-21
SLIDE 21

FQHE and Entanglement Spectrum

slide-22
SLIDE 22

Orbital entanglement spectrum

FQHE on a cylinder (Landau gauge) : orbitals are labeled by ky, rings at position 2πky

L l2 B

Divide your orbitals into two groups A and B, keeping Norb,A

  • rbitals : orbital cut ≃ real space cut (fuzzy cut)

K =01234567....

y

A B

Ky

A

1 0 1

K = y 1 2

}

1

4 5

}

B

1

6 7 3

1

Laughlin state N = 12, half cut

10 20 30 40 50 60 2 4 6 8 10 12 14 16 18

ξ Ky,A

OES Laughlin N= 12, NA= 6 on a cylinder L= 15 Ky 1 1 2 3 5 7 ...

Fingerprint of the edge mode (edge mode counting) can be read from the ES. ES mimics the chiral edge mode spectrum. For FQH model states, nbr. levels is exp. lower than expected.

slide-23
SLIDE 23

ES on various geometries

10 20 30 40 50 5 10 15 20 25 30 35

NF

(a)

ξ

Lz,A

10 20 30 40 50 5 10 15 20 25 30 35

ξ

Lz,A (d)

NF

R= 10 20 30 40 50 2 4 6 8 10 12 14 16 18

ξ

(b) Ky,A

10 20 30 40 50 60 2 4 6 8 10 12 14 16 18

ξ Ky,A

OES Laughlin N= 12, NA= 6 on a cylinder L= 15 Ky 1 1 2 3 5 7 ...

Different eigenvalues of ρA (shape of the ES) but the same number of non-zero eigenvalues (counting) The counting IS the important feature. For model states in the FQHE , exponentially lower than expected

slide-24
SLIDE 24

Away from model states

Groundstate of the Coulomb interaction at ν = 1/3 for N = 12 on a sphere/thin annulus ν = 1/3 Laughlin state

10 20 30 40 50 5 10 15 20 25 30 35

ξ

Lz,A

3 6 9 0 1 2 3 4 5 6

ξ

Lz,A

1 1 2 3 5 7 11

VS GS of the Coulomb interaction

5 10 15 20 25 30 35 40 10 20 30 40

ξ Lz,A

ξ

Lz,A 3 6 9 22 24 26

Dx

A low ent. energy structure identical to the Laughlin state.. An entanglement gap that does not spread over the full spectrum but protects the region mimicking the edge mode. An additional structure in the high energy part related to the neutral excitations (quasihole-quasielecton pairs).

slide-25
SLIDE 25

How to cut the system ?

The system can be cut in different ways : real space

  • rbital (or momentum) space

particle space Each way may provide different information about the system (ex : trivial in momentum space but not in real space)

NF geometrical partition particle partition NF/2 edge physics quasihole physics NF

Real space partitioning : extracting the edge physics Particle partitioning : extracting the bulk physics

slide-26
SLIDE 26

Particle entanglement spectrum

Ground state Ψ for N particles, remove N − NA, keep NA

ρA(x1, ..., xNA; x′

1, ..., x′ NA)

=

  • ...
  • dxNA+1...dxN

Ψ∗(x1, ..., xNA, xNA+1, ..., xN) × Ψ(x′

1, ..., x′ NA, xNA+1, ..., xN)

5 10 15 20 5 10 15 20

Ky,A

ξ

ν = 1/3 Laughlin N = 8, NA = 4

5 10 15 20 5 10 15 20

ξ

Ky,A

ξ

D

Coulomb GS at ν = 1/3 on a torus

Counting is the number of quasihole states for NA particles on the same geometry → the fingerprint of the phase.

slide-27
SLIDE 27

ES and Fractional Chern Insulators

slide-28
SLIDE 28

Chern insulators

A Chern insulator is a zero magnetic field version of the QHE (Haldane, 88). Topological properties emerge from the band structure. At least one band is a non-zero Chern number C, Hall conductance σxy = e2

h |C|

What about the strong interacting regime ? → Fractional Chern insulators. Neupert et al. PRL 106, 236804 (2011), Sheng et al. Nat. Comm. 2, 389 (2011), NR and BAB, PRX (2011)

e1 e2

1 2 3 4 5 6 0 1 2 3 4 5 6

  • 10
  • 5

5 10 15

kx ky

E(kx,ky)

C=-1 C=0 C=1

slide-29
SLIDE 29

PES and FCI

Filling the lowest band ν = 1/3 plus strong interaction, do we get a Laughlin-like state or a charge density wave ? FCI phase

0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 5 10 15 20

E (arb. unit) Kx + Nx * Ky (a) 5 10 15 20 5 10 15 20 ξ Kx,A + Nx * Ky,A

CDW phase

1e-05 0.0001 0.001 0.01 0.1 1 2 4 6 8 10 12 14 16

E (arb. unit) Ky (a) 5 10 15 20 2 4 6 8 10 12 14 16

ξ Ky,A (b)

Energy spectrum with similar features (3-fold degenerate groundstate) but a different entanglement spectrum.

slide-30
SLIDE 30

Conclusion

For many quantum phases, the ground state contains a surprisingly large amount of information about the excitations. The entanglement spectroscopy is a way to probe (or extract) this information. Seeing the bulk-edge correspondence. Different partitions give access to different types of excitations. Entanglement spectroscopy is a concrete tool, requiring only the ground state (example of the fractional Chern insulators). What is the meaning of the counting exponential reduction ? Efficient description using a matrix product state representation (but this is another story...).

slide-31
SLIDE 31

Conclusion

Counting is great !

slide-32
SLIDE 32

Conclusion

Counting is great ! (you just have to be careful...)