Optically controlling quantum dots for quantum computing Dan Gammon Naval Research Laboratory Washington DC σ- Collaborations Naval Research Lab Allan Bracker Alexander Efros Tom Kennedy Morgan Ware (NRC postdoc) Eric Stinaff (NRC postdoc) Andre Shabaev (NRC postdoc) Alex Ukhanov (NRC postdoc) Duncan Steel (U. Michigan) Vladimir Korenev (Ioffe inst.) Igor Merkulov David Gershoni (Technion U.) Lu Sham (UCSD) Contents • Exciton as qubit • Spin as qubit (from the optical point of view) • Hyperfine interactions and nuclear spin Single QD Spectroscopy Single Qdot Hole Luminescence Intensity 0.2 µm • • • 10 Qdots Apertures in Al mask directly on sample Diameters: 200 nm - 25 um Backscattering or Transmission 1.5 µm 11 105 Qdots 10 mono. 25 µm 1.67 1.68 1.69 Energy (eV) 1.70 AlGaAs GaAs AlGaAs ~3 nm z x 10-100 nm Fine Structure Splittings • Doublets: ∆ = 10 - 100 µeV Spin degeneracy • Linear Polarization along [110] axes STM of GaAs [110] [110] 100 nm [Gammon, et al., PRL 76, 3005 (1996)] Exciton Spin Structure |+2> |+1> |−1> |−2> |−1/2> |+1/2> σ− |−3/2> 1 exciton states σ+ Change basis |+3/2> Spin interactions lift the degeneracies: 1. Exchange 2. Zeeman 3. Hyperfine vacuum δb δ0 x y bright dark 1 exciton states δd vacuum Spectral and spatial addressing of QDs [Qiang Wu, et al., PRB 62, 13022 (2000)] Quantum Computing • • Quantum bit Coherence time • Probe and control individual qubit – initialize – “rotate” – measure • Entangle and operate on two qubits • Scalability |1> |0> E1 E0 Crystal GS Exciton qubit 1 Ground state |0> 0 C0 Ψ = c0 0 + c1 1 • • Gammon and Steel, Physics Today 55(10), 36 (2002) Xiaoqin Li, et al., Optic & Photonic News (2004) in press Exciton |1> + C1 Homogeneous: – No inhomogeneous broadening – Fully resolved (fine structure?) – No line wandering (spectral diffusion) => Linewidth is determined by coherence time (T2) Γhom = 1 T2 = 1 + γ dep 2T1 Energy relaxation “inelastic” E1 -100 0 100 ∆E (µeV) E1 E0 Crystal GS Pure dephasing “elastic” Γhom~23 µeV <=> T2~60 ps PL Intensity Linewidths, Dephasing, and Coherence • • • Gammon, et al., Science 273, 87 (1996) Bonadeo, et al., PRL 81, 2759 (1998) Guest et al PRB 65, 24131 (2002) Coherent Control Y τ ε (t+ τ ) E1 X ε (t ) • Excite QD with pair of coherent pulses (5 ps pulsewidth) as a function of delay (τ) 1 2 E0 ε(t) PL GS E 0 τc=40 ps 1 E 1 E PL Int. Luminescence intensity (a.u.) Laser 2 0.5 0 0 1622 1624 1626 1628 1630 2 4 τf (fs) 6 Energy (meV) Bonadeo, et al., Science 282, 1473 (1998) Decoherence E 1Y PL oscillations decay because of dephasing Γ 34 µeV -100 Tc=0 ps PL Intensity • 0 ∆E (µeV) 100 1 T2=40 ps Tc=20 ps 0.5 Tc=40 ps τ Y Tc=60 ps 0 0.8 Tc= 80 ps 1.6 2.4 3.2 4 τf (fs) 4.8 5.6 X ε2(t+τ) ε1(t) 0 6.4 0 20 40 60 80 Delay time (ps) 100 120 Quantum Superposition of Spin States • Rotate polarization to create superposition of x and y states 1 E1X E1Y Y E0 εX(t) εY(t) X Tosc= 69 ps 0.8 hω 0 GS 0.6 ∆ (60 µeV) E -100 1Y E 0 1X ∆E (µeV) 0.4 0.2 100 0 0 20 40 60 80 Time delay (ps) 100 120 3-level coherence Ground state |00> C00 + Exciton |01> Exciton |10> C10 Ψ = c0 0 1 0 2 + + c1 1 1 0 2 C01 + c2 0 1 1 2 0 11 2 110 2 010 2 Biexciton • 2 excitons 1 exciton 0 excitons Biexciton state is shifted down by ~4 meV [Brunner et al., PRL 73, 1138 (1994)] [Gang Chen et al., PRL 88,117901 (2002)] Biexciton Ground state |00> C00 + Exciton |01> Exciton |10> C10 + C01 Biexciton |11> + C11 111 2 0 11 2 110 2 010 2 If no biexciton shift, then no entanglement: the excitons factor! Ψ = c00 0 1 0 2 + c10 1 1 0 2 + c01 0 1 1 2 + c11 1 1 1 2 = (c0 0 1 + c1 1 1 )(c0 0 2 + c1 1 2 ) For a short discussion along these lines, see; Gammon, et al. Physica E 9, 99 (2001). •For a review; Gang Chen et al., Coherence in Semiconductor Q.Wires and QDs, ed. T. Takagahara (Gordon and Breach 2003). Chapter 9 Qubit Rotations Ground State Ù Exciton Rabi Oscillations Approach τ • Laser pulse (4 ps) to pump • Probe population through differential transmission • Rabi Oscillations of single exciton • 4 ps Π pulse => fast operations ε (t+ τ ) Differential transmission Results Y 2 2π 4π Pulse area Stievater et al. PRL 87, 133603 (2001). ε (t ) 1 See also: Kamada et al. PRL (2001) Htoon et al., PRL (2001) Zrenner et al., Nature (2002) X Rabi Oscillations Ground State Ù Exciton Rabi Oscillations 2π 4π Pulse area Stievater et al., PRL 87, 133603 (2001). Exciton Ù Biexciton Rabi Oscillations 0π π (a.u.) 22π E Field Pulse area Elaine Wu et al., Science 301, 809 (2003). A QNOT gate (exciton) QNOT Gate • Coherent control of two interacting excitons provides an example of a QNOT gate Control bit Target bit 00 → 00 01 → 01 10 → 11 π pulse |11> |10> |01> |00> 11 → 10 Qubit: exciton Exciton decoherence / π pulse |x> T2 ~ 60 ps 1 ns τpulse ~ 4 ps 100 fs T2/τpulse ~ 10 pulses |0> 104 Use Spin!? DT (arb units) Exciton 1.615 Laser Biexciton 1.616 1.617 1.618 1.619 Energy (eV) 1.620 1.621 Contents • Exciton as qubit • Spin as qubit (from the optical point of view) • Hyperfine interactions and nuclear spin A spin qubit qubit: spin electron trion |i> trion trion electron |0> |1> electron Luminescence from single dot in a diode AlGaAs GaAs QDs n+-GaAs N+- I diode with QDs PL Energy (eV) X+ X° 1.656 1.654 X- X+ - X 1.652 XX° 1.650 XX- XX+ -1 0 X° 1 Diode Bias (V) 2 The quantum toolbox Initialization (Optical Pumping) Rotations (coherent Raman) A. Shabaev, et al. PRB 68, 201305 (2003). P. Chen et al., PRB 69, 075320 (2004). Quantum Toolbox Entanglement Measurement (recycling transitions) J1σ1.σc Immamoglu et al PRL 83, 4204 (1999) Piermarcchi et al., PRL 89, 167402 (2002) Troiani et al., PRL 90, 206802 (2003) Pazy et al., Europhysics Lett.62, 175 (2003) Shabaev et al., PRB (2003) Calarco et al., PRA 68, 12310 (2003) Selection Rules -1 +1 +2 -2 σ- Exchange Splitting ~100 µeV σ+ − 3 2 3 2 σ- σ+ forbidden 0 allowed − 1 2 1 2 Exciton (X0) vs. Trion (X-) in B-field E E X° ge - gh |0〉 Bz Voigt Faraday BZ Y X σ+ Y s σ ge s + Bz Bx Voigt Bx Faraday BZ Y X X gh e- ge Bx Bx X- σ+ Y X s σ+ [Tischler, et al., PRB 66, 08131 (2002)] [M. Bayer, et al., PRB 65, 195315 (2002)] s Optical Orientation For GaAs quantum dots: Nonresonant excitation into QW ~30 meV above QD energy Fast hole spin flip, Slow electron spin flip X σ+ QD - σ- PL polarization of X°, X+, XPL Energy (meV) 1 X° 0 PL polarization -1 + -2 I −I P= + − I +I - X -3 X+ -4 Polarization (%) 60 X+ 40 X- X+ Large positive (long e-spin lifetime) X° Zero or small positive at Bz=o (anisotropic exchange) X- Positive or negative, depends on conditions 20 0 X° -20 3 4 Diode Bias (V) − 5 Optical pumping of electron spin X- polarization: 1. h-spin orientation 2. Ground state e-spin pumping bright σ+ X- Polarization (%) 100 W/cm2 10 0 1500 W/cm2 -10 dark σ- -20 4.0 Need fast hole flip rate compared to electron spin flip rate 4.2 4.4 4.6 4.8 5.0 Diode Bias (V) Previous work: Dzhioev et al., Phys. Solid State, 40 1587 (1998) Cortez et al., PRL 89, 207401 (2002) The Hanle effect and electron spin lifetimes geµBBx E Z Polarization ρz (%) 20 BX 10 X+ Y σ+ S e e Ts−1 X° Bx 0 Polarization: -10 ρ z ( B) = ρ z (0) X- -20 -0.5 0.0 BX 0.5 (Tesla) Halfwidth: 1 1 + ( B B1 2 ) 2 ( ) B1 2 = h µ B g e Ts−1 Hanle Halfwidth (G) ge~0.2 X° : No polarization X+ : B1/2 = 3.5kG ⇒ TS ≈ 160 psec X- : B1/2 = 20G ⇒ TS ≈ 30 nsec X- 100 50 0 0 1 2 3 4 Power (mW) 5 Bracker et al., cond-mat/0408466 Quantum Beats Xσ+ Charged QD σ+ Bx Decay ~ 1.5 nsec DT (a.u.) Single Electron! 0 200 Delay(ps) 400 600 Decay ~ 260 psec X0 Decay ~ 66 psec Bz πx Neutral QD πx Ensemble Spin C oherence @ 0.9 T Ensemble Spin C oherence @ 6.6 T Single QD X C oherence @ 0.8 T 0 Gurudev Dutt et al. subm. 500 1000 1500 D elay (psec) 2000 2500 Measurement trion trion |i> trion electron electron electron |0> |1> Contents • Exciton as qubit • Spin as qubit (from the optical point of view) • Hyperfine interactions and nuclear spin s++ σ Nuclear spin Nuclear spin => effective magnetic field • Average => Overhauser shifts • Fluctuations => spin dephasing h e •Brown et al. PRB 54, R17339 (1996) •Gammon et al., Science 277, 85 (1997) •Gammon et al., PRL 86, 5176 (2001) •For a review see; Gammon et al., Coherence in Semiconductor Q.Wires and QDs, ed. T. Takagahara (Gordon and Breach 2003) Chap. 8. •Bracker et al., cond-mat/0408466 Nuclear spin fluctuations Electrons dephase in the effective magnetic field produced by a net fluctuation of the nuclear spin Merkulov et al., PRB 66 153409 (2002) Khaetskii et al., PRL 88, 186802 (2002) Dzhioev et al., PRB 65 205309 (2002) δBeff Inhomogeneous dot-to-dot variation OR Long-time variation in a single dot (spectral diffusion) Ts ≈ h N / A N ≈ 105 atoms A=90 µeV (hyperfine constant in GaAs) expect: Ts ≈ 6 ns measure: 4-30 ns Pumping of nuclei; Overhauser effect Oriented e-h pair + Hyperfine Interaction (electron-nuclear spins) mutiple cycles Overhauser effect (hyperfine) BN Bext Effective B-field σ+ Electron spin splitting depends on laser pol. Key Point: BN affects only electron spin • Not hole spin (p-like)! • Not orbital (i.e. no diamagnetic shift)! Observed: Max. Overhauser shift ∆EN= (∆Eσ +-∆Eσ -)/2 = 81 µeV Nuclear Polarization PN = ∆EN / (3/2)*Ahf = 60% Effective B-Field BN = ∆EN / 2geµB = 14 Tesla X°, X+, X- in longitudinal magnetic field σσ+ PL polarization: Spin splittings vary with laser polarization: PL Intensity (a.u.) X° → signature of optical pumping of nuclear spins Bext = 5T σ- σ Laser: + σ 1.6560 1.6565 X PL Intensity (a.u.) 40 BZ - Y Praw (%) 1.6555 60 1.6530 PL Intensity (a.u.) X+ σ+ s Spin splitting (meV) 1.6525 πx σ - B X° 3.9 V X- 4.1 V X+ 2.5 V X+ 20 X° 0 XC 0.6 X 0.5 Amplitude: Polarization “memory” Offset: Thermal polarization -20 -40 X 1.6520 π y σ+ - X° Amplitude: Overhauser shift X+ Offset: Intrinsic splitting 0.4 0.3 0.25 0.50 0.75 1.00 1.25 1.6530 1.6535 1.6540 PL Energy (eV) Retardance (λ) Bias dependence of polarization, Bz=5T -1 -2 -3 -4 XX+ By increasing the bias and injecting unpol. e’s: • X- polarization increases • Nuclear polarization decreases Electron polarization is the link ! X° 40 X+ 20 X- 0 -20 0 C 0 X° 20 40 X- X+ 3 -40 4 Diode Bias (V) Allan Bracker et al., cond-mat/0408466 -20 5 Nuclear pol. (%) X° 0 Polarization (%) 60 Overhauser shift (µeV) PL Energy (meV) 1 Future Directions • Beyond the perturbative regime – π pulses, etc. • Ground state (spin) coherences – Λ vs. V transitions • Decoherence physics • QD molecules and more complex structures – entanglement • Quantum computing?