Towards new states of matter with atoms and photons

Transcription

1 Towards new states of matter with atoms and photons Jonas Larson Stockholm University and Universität zu Köln Aarhus Cold atoms and beyond 26/6-2014

2 Motivation Optical lattices + control quantum simulators. Hubbard models, spin models, 2

3 Motivation High control monitoring hybrid systems manybody systems beyond condensed matter paradigm models. 3

4 Outline 1. Cavity QED in five or so minutes. 2. Many-body cqed: (quantum) optical bistability 3. Collective phenomena, Dicke physics. 4. SU(3) Dicke model. 5. And then 4

5 Cavity Quantum Electrodynamics 5

6 Cavity QED Jaynes-Cummings physics Cavity QED = coupling between few material (atomic) and few electromagnetic degrees of freedom. Cavity atom-field coupling g~ 1 V (V effective mode volume). Strong coupling regime g n γ, κ (γ/κ atom/photon decay rates). Haroche & Raimond, Exploting the Quantum (OUP). 6

7 Cavity QED Jaynes-Cummings physics Jaynes-Cummings Hamiltonian H jc = ωa + a + Ω 2 σ z + g a + σ + σ + a Field energy Atomic energy Interaction energy Dressed states (polaritons) ψ n+ = cos θ e, n + sin θ g, n + 1, ψ n = cos θ e, n sin θ g, n + 1, E n± = ωn ± g2 n + 1, tan 2θ = 2g n + 1/, = Ω ω. Haroche & Raimond, Exploting the Quantum (OUP). 7

8 Cavity QED Jaynes-Cummings physics Vacuum Rabi splitting, E 0± = ±g ( = 0) Schoelkopf & Girvin, Nature 451 (2008). 8

9 Cavity QED Jaynes-Cummings physics Photon blockade Kimble, Nature 436 (2005). 9

10 Cavity QED Jaynes-Cummings physics Atom-atom, atom-field, field-field entanglement. Logic gates. Cats. Tomography. Quantum-classical correspondence. Zeno and measurement phenomena. Field quantization. 10 Haroche & Raimond, Exploting the Quantum (OUP).

11 Cavity QED Dicke physics Dicke 1954: How does N atoms radiate? H d = ωa + a + N i=1 Ω 2 σ i z + g a + + a σ i x ωa + a + Ω 2 S z + g a + + a S x Dicke states S z S, m = m S, m (maximal spin sector S = N/2). ψ H d S, S, 0 2 ~N 2 ψ Superradiance! Enhanced radiation by a factor N! Dicke, Phys. Rev. 93 (1954). 11

12 Cavity QED Dicke physics Thermodynamic limit (g g/ N). Z 2 -parity symmetry breaking. g c = ωω/2 Normal phase: a = 0, S z = N/2 Superradiant phase: a 0, S z > N/2 Classical PT (Ising type), survives also at T = 0 ( quantum PT). No-go theorem. Minimal coupling Hamiltonian, assume ground state atoms, no dipole-dipole interaction Dicke PT forbidden (Rzaźewski, Phys. Rev. Lett. 35, (1975)). Large atom-field coupling by driving openness PT? Wang & Hioe, Phys. Rev. A 7 (1973). 12

13 Many-body cqed 13

14 Many-body cqed Thermal atomic gases had been loaded into resonators, next step a BEC. 2007; Stamper-Kurn, Reichel, Esslinger. (no strict proof for BEC at this stage). Vacuum Rabi splitting ~ N. Optical bistability Brennecke, Nature 450 (2007). 14

15 Many-body cqed Low temperature coupling to atomic motion. Hamiltonian in dispersive regime ( g n, U 0 = g 2 / ) H = dx ψ + x d2 dx 2 + U 0cos 2 x a + a ψ x c a + a iη a a +. Rotating frame ( c = ω ω p ), standing wave of resonator, longitudinal cavity pumping (amplitude η, frequency ω p ), neglecting atom-atom interaction. Steady state photon number n s = Optical bistability η 2 κ 2 + c U 0 dx ψ(x) 2 cos 2 (kx) 2. Ritsch, Rev. Mod. Phys. 85 (2013). 15

16 Many-body cqed Optical bistability Lowest vibrational modes of the BEC: H = ω rec c + c + ( c + g c + c + )a + a iη a a +. Paradigm optomechanical Hamiltonian. Multi-stability and strong coupling regime Brennecke, Science 322 (2008). Venkatesh, Phys. Rev. A 83 (2011). Larson, Phys. Rev. Lett. 100 (2008). 16

17 Many-body cqed Transverse pumping: atoms mediate scattering of photons from pump to cavity. Expand in lowest vibrational modes Effective model = Dicke Hamiltonian in the Schwinger representation. Effective potential Dicke physics V eff x, z = V 0 cos 2 z + U 0 α 2 cos 2 x + W 0 α + α cos x cos (z) Field phase angle α + α = 0, π Z 2 symmetry breaking of Dicke. Condensates in a checkerboard supersolid phase: populates every second site. Nagy, Phys. Rev. Lett. 104 (2010). 17

18 Experimental realizations: Many-body cqed Dicke physics Experimental measurement of field phase. Schematic picture of supersolid states (selforganization transition). Baumann, Nature 464 (2010), Phys. Rev.L ett. 107 (2011). 18

19 SU(3) Dicke model 19

20 Esslinger group, new setup 1: Many-body cqed Beyond regular Dicke physics Schematic picture of the double-cavity setup. Venkatesh & Larson, underway 20

21 Many-body cqed Beyond regular Dicke physics Only cavity A: Only cavity B: Classical potential Classical potential Venkatesh & Larson, underway 21

22 Many-body cqed Beyond regular Dicke physics Cavity A + B: Classical potential Venkatesh & Larson, underway 22

23 Many-body cqed Beyond regular Dicke physics Cavity A + B: No classical frustration! Z 2 Z 2 symmetry. Classical potential Venkatesh & Larson, underway 23

24 Physics beyond mean-field. Many-body cqed Identify the low energy vibrational states Ψ x, z = c0ψ 0 x, z + c1ψ 1 x, z +c2ψ 2 x, z, ψ 0 x, z = 1, ψ N 1,2 x, z = 2 cos 3 x ± 1 z cos(z), N 2 2 Generalized Schwinger representation H su3 = ω A a + a + ω B b + b + ΩΛ 8 + g A N Λ i, i = 1,, 8 SU(3) algebra. Beyond regular Dicke physics Λ 4, Λ 6 and Λ 8 not an SU(2) subgroup. a + a+ Λ 4 + g B N b + b+ Λ 6. Venkatesh & Larson, underway 24

25 Many-body cqed Beyond regular Dicke physics Symmetries: 1. Z 2 Z 2 generalized Dicke parity symmety Π A = exp iπ n A + Λ Λ 8 2, Π B = exp iπ n B Λ Λ If parameters X A = X B an emergent continuous U(1) symmetry U θ = exp iθ a + b b + a iθλ 2. Venkatesh & Larson, underway 25

26 Many-body cqed Beyond regular Dicke physics Mean-field of SU(3) Dicke Λ α =, α = 1,2,, 8, a =, b =. Steady state solutions (U(1) symmetry X A = X B ): Λ 3 Λ 4 = cos θ/2 1 g N c 4 /g ( Dipole 01 ), Λ 6 = sin θ/2 1 g N c 4 /g ( Dipole 02 ), Λ 1 N cos θ = N a+a N b+b N = sin θ g c 2 /g ( Dipole 12 ), 1 g c 2 /g ( Inversion 12 ), = 2ωg κ 2 +ω 2 1 g c 4 /g 4 (Quadrature A), = 2ωg κ 2 +ω 2 1 g c 4 /g 4 (Quadrature B). Venkatesh & Larson, underway 26

27 Many-body cqed Beyond regular Dicke physics Critical coupling g c = 1 2 Ω κ 2 +ω 2 ω, cavity losses κ. x A, x B ~ 1 g c g 1/2 β = 1/2 as in open SU(2) Dicke model. For the Z 2 Z 2 identical critical coupling g c and exponent β. Breaking of Z 2 Z 2 one cavity empty in superradiant phase. Breaking of U(1) both civities populated, amplitude set by θ. Venkatesh & Larson, underway 27

28 What s next? 28

29 Long range interaction Glassy states? Esslinger group, new setup 2: Atoms in (classical) 2D optical lattice 1D tubes. Setup confined in a resonator. Tubes independently interact with the same cavity mode. ~1500 tubes, atoms/tube. 29

30 Long range interaction Glassy states? N-channel Dicke realization H Nd = ωa + N a + Ω i=1 S (i) z + g i a + a + S (i) x + US (i) zs (i) z. Trace out boson field H rlmg = N 2 i=1 S (i) z + U S (i) z + g ij S (i) xs (j) x i,j. N-channel (quasi random) Lipkin-Meshkov-Glick model. g ij Gaussian distributed? glassy states (Strack & Sachdev, Phys. Rev. Lett. 107, (2011); Buchhold, Phys. Rev. A. 87, (2013)). 30

31 Long range interaction Localization Optical lattice + cavity field (standing wave). Incommensurate lattices (cavity field weak), mean-field (atomic part) t φ i = J φ i 1 + φ i+1 + μ i a + aφ i, t a = i a i i μ i φ i φ i a + η. t φ i = J φ i 1 + φ i+1 + μ i φ i, onsite disorder localization (Kramer & MacKinnon, Rep. Prog. Phys. 56, (1993)). t φ i = J φ i 1 + φ i+1 + μ i φ i + φ i 2 φ i, weak nonlinearity localization lost at large times (Pikovsky & Shepelyansky, Phys. Rev. Lett. 100, (2008)). t φ i = J φ i 1 + φ i+1 + μ i f( μ i φ i 2 ) i φ i, localization? 31

32 Thanks! Summary Classical vs. Quantum choas. Possibilities to study closed quantum dynamics. Equilibration and thermalization. Not well understood: i. Criteria for equilibration/thermalization. ii. Mechanism behind thermalization. iii. Definition for Quantum integrability. iv. Open systems 32