ebook img

Stability thresholds and calculation techniques for fast entangling gates on trapped ions PDF

0.91 MB·
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Stability thresholds and calculation techniques for fast entangling gates on trapped ions

Stability thresholds and calculation techniques for fast entangling gates on trapped ions C D B Bentley,1,∗ R L Taylor,1 A R R Carvalho,1,2 and J J Hope1 1Department of Quantum Science, Research School of Physics and Engineering, The Australian National University, Canberra, Australia 2ARC Centre for Quantum Computation and Communication Technology, The Australian National University, Canberra, Australia (Dated: January 14, 2016) Fast entangling gates have been proposed for trapped ions that are orders of magnitude faster than current implementations. We present here a detailed analysis of the challenges involved in 6 performing a successful fast gate. We show that the RWA is a stable approximation with respect 1 to pulse numbers: the timescale on which we can neglect terms rotating at the atomic frequency is 0 negligiblyaffectedbythenumberofpulsesinthefastgate. Incontrast,weshowthatthelaserpulse 2 instabilitydoesgiverisetoapulse-numberdependenteffect; thefastgateinfidelityiscompounded withthenumberofappliedimperfectpulses. Usingadimensionalreductionmethodpresentedhere, n we find bounds on the pulse stability required to achieve two-qubit gate fidelity thresholds. a J 3 I. INTRODUCTION GZC and FRAG gates are independent of the ini- 1 tial motional states, errors in the gate are enhanced ] Atwo-qubitentanglinggateisanessentialcompo- accordingtothemeanvibrationalmodeoccupation, h nent of any quantum information processing (QIP) as shown in [4, 6]. Certain gate error sources have p system [1]. Fast gates for trapped ions using con- been considered: the effects of trap anharmonicity - t trolled large momentum kicks offer a significantly on both schemes [2, 4, 6], dissipation effects on the n faster operation timescale than traditional gates re- GZCscheme[4]andlasercontrolerrorsintheFRAG a u quiring spectral resolution of sidebands [2–7]. This schemesuchasinsufficientlaserrepetitionratesand q in turn leads to simpler gate adaption for long ion pulse timing or direction errors [5]. Only a prelimi- [ crystals or more complex geometries [3, 8–10], with nary analysis of pulsed fast gate errors due to pulse 1 relatively invariant schemes required for sufficiently area imperfections has been performed, despite the v fast gates [6, 11]. There has been recent progress conclusionthatsucherrorsaresignificant[2,5]. Fur- 0 towards the implementation of pulsed fast gates in thermore, a phase stability analysis is still required. 1 the production of the required high repetition-rate Laser phase dependence in each momentum kick 1 pulsed lasers [12] and their application to perform comprising the fast gates arises when the rotating 3 a single-qubit gate [13], as well as spin-motion en- wave approximation (RWA) is no longer valid due 0 tanglement [14]. In this paper we outline challenges to short pulse durations relative to the atomic tran- . 1 for performing a complete fast gate protocol, and sition frequency. This leads to imperfect popula- 0 present both the techniques for quantifying gate fi- tion transfer between internal states. Short pulse 6 delity subject to imperfections, as well as the re- durationsarenecessarysuchthatthetotalmotional 1 quired thresholds in laser stability and pulse times evolution during the pulses is negligible. Thus the : v to perform high-fidelity gates. combined duration from all of the applied pulses is Xi For implementation of a fast gate, and certainly required to be much shorter than the trap period, for considering their application to large-scale al- which is on the order of 1 µs. Few pulses or very r a gorithms, detailed analysis of the stability require- short pulse durations seem preferable, however in- ments for the trap and control are critical. Er- creasing the number of pulse pairs in a fast gate im- ror correction can be applied between gate oper- proves the gate speed, fidelity and scaling with the ations, however individual gate operations are re- numberofions. TheRWAprovidesthelowerbound quired to meet high-fidelity thresholds [15–17]. The forpulsedurationsforfastgates,andexaminingthe scheme proposed by Garc´ıa-Ripoll, Zoller and Cirac dependence of this lower bound on the number of (GZCscheme)[2]andtheFastRobustAntisymmet- pulses in a gate is essential for applications of fast ricGate(FRAGscheme)[5]wereshowntohavevery gates to QIP. high fidelity and robustness in [6], and we focus on Significantinfidelityalsoarisesfromimperfectap- these schemes for our error analysis. While perfect plied pulses. Ideal pulses keep the internal qubit states invariant throughout the phase gate and re- storetheinitialmotionalstateattheendofthegate, however imperfect pulses cause internal state trans- ∗ [email protected] fers as well as occupation of a range of motional 2 levels after the gate. Random errors in the pulse trap frequency, Ω (cid:29) ν. In this regime, multiple duration were considered in [2] for a four-kick se- numberstatesofeachsharedmotionalmodeareex- quence, and in [5] a worst-case error bound was cal- cited by pairs of counter-propagating laser pulses, culatedusingperturbationtheoryforsmallerrorsin as shown in Figure 1. These π-pulse pairs provide the pulse area and low numbers of pulse pairs. The momentum kicks such that a closed trajectory in perturbation technique was used for just four pulse phase-space is described for the centre of a coherent pairs, and fails well before 100 pulse pairs with er- state, as in Figure 1 (h). The area enclosed in each ror on the order of 1% in the pulse area [5]. It was mode’s phase-space determines a conditional phase concludedthatimperfectpulseareawilllimitthefi- applied to different two-qubit computational states. delity of fast gates; a more complete analysis of the The evolution of an ideal fast gate can be de- required pulse stability is necessary for gate imple- scribedasalternatingdisplacementandrotationop- mentation. erators in phase space for each motional mode. A It is possible to model the full dynamics of a displacement operator for mode p is described by gate without the RWA or with pulse area imper- fections, and thereby directly calculate the gate fi- Dˆp(α)=exp[αa†p−α∗ap], (1) delity. However, the Hamiltonian operator required for this calculation has a dimension given by the for a displacement of α, where ap is the mode anni- square of the full state vector dimension, which in- hilationoperator. UndertheRWA,pairsofcounter- cludes both the internal qubit states and the vibra- propagatingπ-pulsesgiverisetomodedisplacement tional mode states for each shared mode. In the operators as follows [6]: ideal-pulsecase, thecomplexityisvastlyreducedby simplifying the requirements for performing a high- Ukick =e−2izk(x1σ1z+x2σ2z) (2) fidelity gate to three control conditions [2]. In this =ΠL Dˆ (−2iz(b(p)σz+b(p)σz)η ), (3) paper,wepresentasimplificationmethodforimper- p=1 p 1 1 2 2 p fect gates that permits fidelity calculation for large when there is negligible motional evolution between momentumkicksintrapswithmanyionsandacor- the two pulses. Here z is the direction of the first responding number of shared motional modes. Our pulseinthepair,kisthelaserwavenumber,x isthe i method is presented in Section II following a review position operator for ion i and σz is the usual Pauli i offastgates. InSectionIII,weapplythisfidelitycal- Z operator acting on ion i. There are L motional culationtechniquetoexploretheeffectofpulsenum- modes corresponding to L ions in the crystal, and ber on the phase dependence of the gate with short b(p) is the ion-mode coupling coefficient between ion pulsedurations. Thisprovidesaminimumpulsedu- i i and mode p. The Lamb-Dicke parameter η for p rationboundforhigh-fidelityfastgatescomposedof mode p is given by varying numbers of pulses. In Section IV, we apply our method to consider imperfect pulse areas com- (cid:115) (cid:126) prehensively. We introduce the errors in the atom- η =k , (4) p 2Mν lightevolutionunitaries,andconstructtheimperfect p gate evolution operators to directly compute the fi- for ion mass M and mode frequency ν . delity. This gives us an accurate measure of fidelity p Two main causes of imperfect momentum kicks for large numbers of pulses. Finally, we present our to the ions come from counter-propagating pulses conclusions in Section V. applied with area not equal to π, or from breaking the RWA through short pulse durations. These im- perfections are shown in Figure 1, which illustrates II. QUANTIFYING FAST GATE ERRORS their effect at each stage in the fast gate evolution process. We present the fast gate mechanism and sum- The free motion of the ions, and their motional marise the GZC and FRAG gate schemes, followed modes, corresponds to rotation operators in each by a general fidelity calculation method for two mode p’s phase space: trapped ions as well as a two ion gate in a longer ion crystal. Up,mot =e−iνpδtka†pap, (5) where δt is the time between the kth and (k+1)th k A. Gate dynamics and limitations momentum kicks. The displacements and free rotations are deter- Fast gates operate in the strong-coupling regime, mined according to particular pulse schemes. These where the laser coupling is much greater than the schemes satisfy the required gate conditions: (1) 3 (cid:3)(cid:5)(cid:1) (cid:3)(cid:1)(cid:2) (a) b (h) (cid:3)(cid:5)(cid:1) (cid:1) (g) (cid:1)(cid:5)(cid:8) (cid:1)(cid:2)(cid:3) (cid:1)(cid:5)(cid:8) (cid:3) (cid:21)(cid:22)(cid:23)(cid:13)(cid:24)(cid:20)(cid:19)(cid:25)(cid:22)(cid:12)(cid:1)(cid:1)(cid:1)(cid:5)(cid:5)(cid:5)(cid:4)(cid:6)(cid:7) (cid:12)(cid:9)(cid:18)(cid:16) (cid:1)(cid:1)(cid:2)(cid:2)(cid:5)(cid:4) (cid:21)(cid:22)(cid:23)(cid:13)(cid:24)(cid:20)(cid:19)(cid:25)(cid:22)(cid:12)(cid:1)(cid:1)(cid:1)(cid:5)(cid:5)(cid:5)(cid:4)(cid:6)(cid:7) (cid:2) (cid:1)(cid:5)(cid:1)(cid:1)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2) (cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:3)(cid:1)(cid:2)(cid:1)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:3)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:4)(cid:3)(cid:1)(cid:2)(cid:1)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:4)(cid:3)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2) (cid:16)(cid:8)(cid:16)(cid:17) (cid:1)(cid:2)(cid:6) d (cid:1)(cid:5)(cid:1)(cid:1)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2) (cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:3)(cid:1)(cid:2)(cid:1)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:3)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:4)(cid:3)(cid:1)(cid:2)(cid:1)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:4)(cid:3)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2) (cid:18)(cid:19)(cid:20)(cid:10)(cid:21)(cid:17)(cid:16)(cid:22)(cid:19)(cid:9)(cid:1)(cid:1)(cid:1)(cid:1)(cid:1)(cid:1)(cid:5)(cid:5)(cid:5)(cid:5)(cid:5)(cid:5)(cid:1)(cid:1)(cid:3)(cid:3)(cid:4)(cid:4)(cid:1)(cid:2)(cid:1)(cid:2)(cid:1)(cid:2)(cid:1)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:9)(cid:3)(cid:1)(cid:2)(cid:10)(cid:3)(cid:1)(cid:2)(cid:3)(cid:3)(cid:1)(cid:11)(cid:2)(cid:1)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:12)(cid:13)(cid:3)(cid:1)(cid:2)(cid:14)(cid:3)(cid:1)(cid:2)(cid:3)(cid:3)(cid:1)(cid:2)(cid:15)(cid:2)(cid:3)(cid:1)(cid:2)(cid:16)(cid:17)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:18)(cid:19)(cid:3)(cid:1)(cid:2)(cid:20)(cid:4)(cid:3)(cid:1)(cid:2)(cid:19)(cid:1)(cid:16)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:4)(cid:3)(cid:1)(cid:2)(cid:2)((cid:3)(cid:1)(cid:2)b(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2))(cid:3)(cid:1)(cid:2) (cid:7)(cid:8)(cid:9)(cid:10)(cid:9)(cid:11)(cid:12)(cid:13) --(cid:1)(cid:1)(cid:1)(cid:2)(cid:2)(cid:2)(cid:5)(cid:6)(cid:1) f -(cid:1)(cid:2)(cid:3) -(cid:1)(cid:2)(cid:4) -(cid:1)(cid:2)(cid:5) -(cid:1)(cid:2)(cid:6) g(cid:1)(cid:2)a(cid:1) c(cid:1)(cid:2)(cid:6) (cid:1)(cid:2)(cid:5)e (cid:18)(cid:19)(cid:20)(cid:10)(cid:21)(cid:17)(cid:16)(cid:22)(cid:19)(cid:9)(cid:1)(cid:1)(cid:1)(cid:1)(cid:1)(cid:1)(cid:5)(cid:5)(cid:5)(cid:5)(cid:5)(cid:5)(cid:1)(cid:1)(cid:3)(cid:3)(cid:4)(cid:4)(cid:1)(cid:2)(cid:1)(cid:2)(cid:1)(cid:2)(cid:1)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:9)(cid:3)(cid:1)(cid:2)(cid:10)(cid:3)(cid:1)(cid:2)(cid:3)(cid:11)(cid:3)(cid:1)(cid:2)(cid:1)(cid:3)(cid:1)(cid:2)(cid:12)(cid:3)(cid:1)(cid:2)(cid:13)(cid:3)(cid:1)(cid:2)(cid:14)(cid:3)(cid:1)(cid:2)(cid:3)(cid:3)(cid:1)(cid:2)(cid:15)(cid:2)(cid:3)(cid:1)(cid:2)(cid:16)(cid:17)(cid:3)(cid:1)(cid:2)(cid:18)(cid:3)(cid:1)(cid:2)(cid:19)(cid:3)(cid:1)(cid:2)(cid:20)(cid:4)(cid:19)(cid:3)(cid:1)(cid:2)(cid:1)(cid:16)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:4)(cid:3)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)((cid:3)(cid:1)(cid:2)f(cid:3)(cid:1)(cid:2))(cid:3)(cid:1)(cid:2) (cid:6)(cid:7)(cid:8)(cid:9)(cid:10)(cid:11)(cid:12)(cid:13)(cid:14)(cid:15)(cid:16)(cid:17)(cid:16)(cid:13) (cid:7)(cid:8)(cid:9)(cid:10)(cid:9)(cid:11)(cid:12)(cid:13)(cid:14)(cid:8)(cid:15)(cid:11)(cid:9)(cid:11)(cid:8)(cid:12) (cid:6)(cid:7)(cid:8)(cid:9)(cid:10)(cid:11)(cid:12)(cid:13)(cid:14)(cid:15)(cid:16)(cid:17)(cid:16)(cid:13) (cid:1)(cid:5)(cid:3)(cid:2)(cid:3)(cid:1)(cid:3)(cid:1)(cid:2) (c) (cid:1)(cid:5)(cid:3)(cid:1) (cid:3)(cid:1)(cid:3)(cid:1) (d) (cid:1)(cid:5)(cid:3)(cid:2)(cid:3)(cid:1)(cid:3)(cid:1) (e) (cid:18)(cid:19)(cid:20)(cid:10)(cid:21)(cid:17)(cid:16)(cid:22)(cid:19)(cid:9)(cid:1)(cid:1)(cid:1)(cid:5)(cid:5)(cid:5)(cid:1)(cid:1)(cid:3)(cid:1)(cid:2)(cid:1)(cid:2) (cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2) (cid:21)(cid:22)(cid:23)(cid:13)(cid:24)(cid:20)(cid:19)(cid:25)(cid:22)(cid:12)(cid:1)(cid:1)(cid:1)(cid:1)(cid:1)(cid:5)(cid:5)(cid:5)(cid:5)(cid:5)(cid:1)(cid:1)(cid:1)(cid:1)(cid:1)(cid:1)(cid:4)(cid:6)(cid:7)(cid:8)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2) (cid:18)(cid:19)(cid:20)(cid:10)(cid:21)(cid:17)(cid:16)(cid:22)(cid:19)(cid:9)(cid:1)(cid:1)(cid:1)(cid:5)(cid:5)(cid:5)(cid:1)(cid:1)(cid:3)(cid:1)(cid:2)(cid:1)(cid:2)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2)(cid:3)(cid:1)(cid:2) (cid:1) (cid:2) (cid:3)(cid:1) (cid:3)(cid:2) (cid:4)(cid:1) (cid:4)(cid:2) (cid:1) (cid:2) (cid:3)(cid:1) (cid:3)(cid:2) (cid:4)(cid:1) (cid:4)(cid:2) (cid:1) (cid:2) (cid:3)(cid:1) (cid:3)(cid:2) (cid:4)(cid:1) (cid:4)(cid:2) (cid:6)(cid:7)(cid:8)(cid:9)(cid:10)(cid:11)(cid:12)(cid:13)(cid:14)(cid:15)(cid:16)(cid:17)(cid:16)(cid:13) (cid:9)(cid:10)(cid:11)(cid:12)(cid:13)(cid:14)(cid:15)(cid:16)(cid:17)(cid:18)(cid:19)(cid:20)(cid:19)(cid:16) (cid:6)(cid:7)(cid:8)(cid:9)(cid:10)(cid:11)(cid:12)(cid:13)(cid:14)(cid:15)(cid:16)(cid:17)(cid:16)(cid:13) (cid:1) (cid:1)=(cid:1)(cid:2)(cid:3)(cid:4)(cid:5) (cid:2) (cid:1)=(cid:6)(cid:7)(cid:8)(cid:9)(cid:2)(cid:3)(cid:4)(cid:5) (cid:3) (cid:1)=(cid:1)(cid:2)(cid:2)=(cid:9)(cid:10)(cid:11)(cid:2)(cid:3)=(cid:12)(cid:4)/(cid:9) FIG. 1. (h): Centre-of-mass (COM) phase-space trajectory for the centre of a coherent state during the gate operation. The sides of the trajectory correspond to momentum kicks. The angle of each vertex corresponds to free evolution between kicks, and marks the gate evolution point for the other subfigures, from a→g. (a)→(g): Population occupying COM mode number states, for both ions in the excited state, at different points during the GZC gate operation, with n = 1 (14 total pulse pairs). Blue circles represent an ideal gate, satisfying the rotating wave approximation (RWA) and with no pulse area imperfections given by (1−ξ). A gate with systematic pulse imperfections (ξ = 0.95) is also considered (yellow squares), as well as a gate where the RWA is invalid (green diamonds) such that the pulse duration and laser phase must be defined (τ = 5 fs, φ = 3π/5). After the non-ideal gates, population is lost to other internal states, and some of the population is imperfectly restored to the initial COM state, |2(cid:105) , the second excited number state. Values were chosen to illustrate the effects of these errors. The c point in the gate operation described by subfigures (a)→(g) corresponds to a→g marked in subfigure (h). conditional phase evolution according to the two- TheintegerndeterminesthegatetimeT ,which G qubit gate described by scales optimally with the total number of pulses in the scheme N as T ∝N−2/3 [4, 6]. Ugate =eiπ4σ1zσ2z (6) The FRAGpand GGZC scphemes consist of 10n and 14n pulses respectively. The FRAG scheme has a for a gate applied to ions 1 and 2, and (2) no mo- state-averaged fidelity, as defined in [6], of 0.96 for tional dependence, such that the initial motional n = 1, and 0.995 for n = 2, while for higher n the state is restored following the gate operation. The infidelity is below 10−8. We neglect the low-fidelity GZCandFRAGschemesarecharacterizedbypulse n = 1 case of the FRAG scheme, which obscures pairszappliedattimest,interspersedwithfreeevo- thestabilityanalysis. TheGZCscheme,withhigher lution. For the FRAG scheme [5]: totalnumbersofpulsesforeachn,achievesinfidelity on the order of 10−5 for n=1, and infidelity below z = (−n, 2n, −2n, 2n, −2n, n) 10−8 for higher n. t = (−τ , −τ , −τ , τ , τ , τ ). 1 2 3 3 2 1 The scaling of errors with the number of pulses is examined for both the FRAG and GZC schemes. At time −τ , n counter-propagating pulse pairs 1 Weexploretheeffectsoferrorsonschemeswithlow are applied along the trap axis (aligned with the z pulse-numbers using the n = 1 GZC scheme due to axis) to provide a 2n(cid:126)k momentum kick in the −z its high fidelity. While more robust for lower num- direction. bers of pulses, the GZC scheme is slower than the TheGZCscheme[2,4]ischaracterisedasfollows: FRAG scheme for n ≥ 2, as shown in [5]. The ef- fects of finite laser repetition rates on these schemes z = (−2n, 3n, −2n, 2n, −3n, 2n) were explored in [5, 6], where it is shown that for t = (−τ1, −τ2, −τ3, τ3, τ2, τ1). repetition rates of around 300 MHz, even a gate 4 with perfect π-pulses has non-negligible infidelity. stretch modes respectively. The motional inner Faster repetition rates have robust fidelities, partic- product is stricter than the computational fidelity ularly for the two-qubit case. Errors due to imper- of [6], with a stronger motional restoration require- fectpulseareasorfrombreakingtheRWAaffectany ment that population must be restored to the ini- schemeregardlessofrepetitionrate;inthispaperwe tial number state for each mode at the end of the consider these errors independently by assuming an gate operation. This is a convenient choice for our infinite repetition rate. Furthermore, this approx- number basis, and directly considers effective heat- imation allows a clear analysis of the relationship ing caused by the gate to be infidelity. between these errors and number of applied pulses. The ideal gate operation of equation (6), with Themethodsinthispapercanbeappliedusingpar- duration T , applies a state-dependent phase while G ticular, finite repetition rates to model the errors in preserving the internal and motional states: an experiment more precisely. (cid:16) To model the effect of imperfect pulses or an in- U |φ (cid:105)= eiπ/4a |gg(cid:105)+e−iπ/4a |ge(cid:105) id i 00 01 valid RWA, we expand the appropriate unitary op- (cid:17) erator for the applied gate, U , in the number ba- + e−iπ/4a |eg(cid:105)+eiπ/4a |ee(cid:105) re 10 11 sis. We can then observe the phase-space evolution during the gate process, and calculate the fidelity ⊗e−iTG(νcnc+νrnr)|ncnr(cid:105), (12) of the gate. While coherent states are preserved by where the motional component is global phase cor- the momentum kicks and rotations, the momentum responding to free evolution for each mode. kicksdeformaninitialnumberstatetospreadacross Therealgateisamorecomplexoperationonboth manymodes. Attheendofahigh-fidelitygate,how- thecomputationalandmotionalstates, andwecon- ever, this spread resolves back into the initial num- sider the error to first order in the small gate im- ber state, as shown in Figure 1(a)-(g). perfection. Since the ideal gate does not transform the basis states but just applies a phase, a real gate approximating the ideal operation has only small B. Fidelity calculation: dimensional reduction population transfer between internal states. The gate schemes are designed to restore the motional To assess the impact of particular errors, we use states for preserved internal states; only a fraction the state-averaged fidelity as a measure of the gate of the motional state population (to second order performance. The fidelity of a pure state |ψ(cid:105) with in the error) will be restored for altered internal respect to a density matrix σ is given by the state states with changed state-dependent displacement overlap [18]: operators. Thesetermswithchangedinternalstates thusprovideasecond-ordercorrectiontothefidelity, F =(cid:104)ψ|σ|ψ(cid:105). (7) which we neglect here. For an initial state |φ (cid:105), the final state following the Similarly, an ideal counter-propagating pair of i ideal gate operation U is given by pulses acting on two ions with the same internal id stateaffectsonlytheCOMmode. Animperfectpair |ψ(cid:105)=Uid|φi(cid:105), (8) of pulses may alter the stretch mode to some small degree; a first-order error term. This perturbation and the final density matrix following the real, im- to the stretch mode also has only a small effect on perfect, operation U is given by re the fidelity; only a fraction of the perturbation is σ =U |φ (cid:105)(cid:104)φ |U† . (9) expected to be restored to its initial motional state re i i re andthissecond-ordercontributionisneglected. Ions The state-averaged fidelity is thus withoppositeinternalstatesareassumedtohavean (cid:90) invariant COM mode, with the gate acting on the F = |(cid:104)φ |U†U |φ (cid:105)|2, (10) stretch mode. i id re i φi For the basis state integrating uniformly over the unit hypersphere de- a |gg(cid:105)⊗|n n (cid:105)≡a |ggn n (cid:105), (13) 00 c r 00 c r scribedbytheinitialstatewitharbitrarycoefficients a : the fidelity inner product element is thus jk |φi(cid:105)=(a00|gg(cid:105)+a01|ge(cid:105)+a10|eg(cid:105)+a11|ee(cid:105)) |a00|2(cid:104)ggnc|Ui†dUre|ggnc(cid:105) ⊗|ncnr(cid:105). (11) =|a00|2e−iπ/4(cid:104)ggnc|Ure|ggnc(cid:105), (14) The initial motional state is the number product wherethestretchmodeisallowedtoevolvefreelyby state |n (cid:105)⊗|n (cid:105) for the centre-of-mass (COM) and the ideal and real unitaries, and thus cancels from c r 5 the inner product. Only the population retained in during the gate operation. This truncation occurs the computational ground state of both ions is re- for higher phonon numbers with larger numbers of tained in the fidelity term. The unitaries act sym- applied pulses in a gate corresponding to larger mo- metrically on |ee(cid:105), and the same symmetry between mentum kicks. We truncate our number basis at 50 |eg(cid:105) and |ge(cid:105) allows us to simplify our full fidelity statesforn=1foreachgate,70statesforn=2and expression: n = 5, and 130 states for n = 10. The single-mode analysis thus reduces the dimensions for calculating (cid:90) (cid:12) thestatevectorevolutionbyafactorofaround100, F = (cid:12)(|a |2+|a |2)eiνcTGnc−iπ/4 (cid:12) 00 11 depending on the number of applied pulses. The φi state vector’s dimension is four times the dimension ×(cid:104)ggn |U |ggn (cid:105)+(|a |2+|a |2)eiνrTGnr+iπ/4 c re c 01 10 of the number basis, due to the two basis internal ×(cid:104)gen |U |gen (cid:105)|2. (15) states for each ion. r re r For each internal state of the two qubits, the effect of the real gate on just a single motional mode con- C. Extending fidelity calculation to larger tributes to the fidelity expression. We thus expand traps U for each mode independently, and accordingly re cancel the motional phase term in the ideal unitary Performingfastgateswithinalargequantumpro- from free evolution of the other mode. cessorrequiresanalysisofgateimperfectionsintraps Ageneralpositionoperatordecompositionfortwo with larger numbers of ions. The number of mo- trapped ions is described by tionalmodesisequaltothenumberoftrappedions. Thepositiondecompositioncanbegeneralisedfrom kx =b(c)η (a +a†)+b(r)η (a +a†), (16) equation (16) for arbitrary numbers of ions, and we i i c c c i r r r againapplytheapproximationofseparablemotional wherethesubscriptcdescribestheCOMmode,and modes in a harmonic trap. subscriptrdescribesthestretchmode. Thecoupling The ideal and real unitaries can be written as a operators for two trapped ions are separable product of the operation on each mode: b(c) =(√1 ,√1 ) (17) (cid:89)L 2 2 Uid = Uid,p (21) 1 1 p=1 b(r) =(−√ ,√ ), (18) 2 2 (cid:89)L U = U , (22) re re,p with the jth vector element representing the √cou- p=1 pling for ion j, and mode frequencies of ν and 3ν for the COM and stretch modes respectively. where L is the number of modes and The real gate unitary U is applied to both modes, howevertoconsiderfirrest-ordererrorswecan Uid,p =eiφpσ1zσ2z, (23) treat it as separable for each mode. This allows us up to global phase. This simplifies the fidelity F to apply a single-mode expansion of kx for each in- i expression to require only the product of the mode- ternal state of the qubits. For the states |gg(cid:105) and dependent fidelities F up to first order in the error p |ee(cid:105), the terms in U contributing to the fidelity re term: equation (15) are given by (cid:90) F (cid:39) |Π F |2, (24) kx =b(c)η (a +a†), (19) p p i i c c c φi F =(cid:104)φ |U† U |φ (cid:105). (25) while for |ge(cid:105) and |eg(cid:105), p i id,p re,p i We can calculate the contributions U of each kx =b(r)η (a +a†). (20) re,p i i r r r modetotherealgateunitary,usinganappropriately truncated single-mode number basis for each calcu- Our separable representation of U reduces the re lation. For a high-fidelity gate scheme, the ideal dimension of the state vector by a factor given by phase contribution for each mode is given by [6] the number of required motional basis states. The Hamiltonian is reduced in dimension by the square N m−1 ofthisfactor. Weusethenumberbasistomodelthe φ =8η2σzσzb(p)b(p) (cid:88) (cid:88) z z sin(ν (t −t )), p p 1 2 1 2 m k p m k stateevolution,andtruncatethebasissuchthatneg- m=2 k=1 ligible population occupies the maximal basis states (26) 6 wherez isthenumberofpulsepairsappliedattime is: k t determinedbythegatescheme. Forahigh-fidelity gkate, (cid:80)pφp ≈ π4. Upulse,1 =exp(cid:20)−2iτπ (cid:18)(cid:90) tf σ1+e−i(kx1−2ωatt+φ)dt Our method to estimate fidelity in this fashion is ti to first calculate the real unitaries for each mode (cid:90) tf + σ−ei(kx1−2ωatt+φ)dt+τσ+ei(kx1+φ) U and the ideal phases φ . The individual 1 1 re,p p ti mode expansion of the gate unitary ensures a low- (cid:17)(cid:105) dimensional state vector, as required for computa- +τσ1−e−i(kx1+φ) , (28) tion. for a pulse of duration τ =t −t , where f i (cid:90) tf σ+e−i(kx1−2ωatt+φ)dt 1 ti −iσ+ (cid:16) (cid:17) III. BREAKING THE RWA = 1 e−i(kx1−2ωattf+φ)−e−i(kx1−2ωatti+φ) . 2ω at (29) Pulses with duration on the order of the atomic transition period 2π/ω render the RWA invalid, Fortwoions,theinteractionHamiltonianisH(cid:48)+H(cid:48) at 1 2 and cause infidelity in fast gate schemes which rely for ions 1 and 2, using H(cid:48) from equation (27). The i ontheRWA.Inthissectionweapplyourfidelitycal- pulseunitaryoperatorUpulse,1 issimilarlyextended, culation method to explore the tradeoff in pulse du- and combined with the motional free evolution uni- ration between performing large numbers of pulses tary,equation(5),toconstructtherealgateunitary in a short time for fast, high-fidelity gates, while Ure. This allows us to explore the validity of the staying in the regime where the RWA holds. Gates RWA for different pulse lengths by solving for the significantly faster than the trap evolution period phase dependence and fidelity. (∼1µs)requirelargenumbersofpulses,whichmust Figure 2 shows the effect of short pulse duration thus have very short durations. We demonstrate ongatefidelity. Forshortpulseduration,thefidelity that the valid RWA regime is altered little by the decreases as more pulses are applied for the FRAG number of applied pulses in a gate. and GZC gates. The mean infidelity is plotted for varying phase φ, and error bars mark a standard We perform fast gates using short pulses of vary- deviationininfidelityduetophasedependence. The ing duration without performing the RWA to inves- FRAG gate with n = 2 has fidelity of 0.988, and tigate the regime where the approximation holds. approachesthisvaluewithastandarddeviationless The gate should also be independent of the optical than 10−3 for pulse lengths τ > 40 fs. The GZC phase φ [19], and we quantify the pulse lengths re- schemeandtheFRAGschemeforn=5havefidelity quired for phase-independence. In the interaction above 0.999 and standard deviation less than 10−4 frame with respect to the internal states of a sin- for τ ≥60 fs. gle ion, ion 1 marked by subscripts, the atom-light Pulsesmuchlongerthantheatomictransitionpe- interaction Hamiltonian is riod are accurately described under the RWA, and the number of pulses in the gate does not signifi- (cid:126)Ω cantlyalterthisthreshold. Forgateswithincreasing H(cid:48) = (σ+e−i(kx1−(ωL+ωat)t+φ)+σ+ei(kx1−δt+φ) speed or scalability with the number of ions, large 1 2 1 1 numbers of pulses must be performed much faster +σ1−e−i(kx1−δt+φ)+σ1−ei(kx1−(ωL+ωat)t+φ)), (27) than the trap motional frequency, ν/(2π)(cid:39)1 MHz, or even much faster than 10 ns for momentum ap- plication schemes exciting short-lived atomic lev- where δ = (ω − ω ). Typical atomic frequency L at els [20]. This provides five orders of magnitude be- transitions are on the order of ω ∼ 2π×1015 Hz, at tweenasafepulseduration∼100fsandthelifetime and the fast rotating terms can be neglected fol- of typical short-lived levels, such as P in 40Ca+. lowing the RWA. Pulse durations are typically as- 3/2 sumed to be much longer than the rotation period, τ(π×1015)(cid:29)1. IV. IMPERFECT PULSES We focus on resonant transitions where δ = 0 for simplicity, such that ω = ω . Assuming constant Significant errors also arise from imperfect π- L at Ω and a perfect π-pulse, such that Ω = π/τ, the pulses, which construct the momentum kicks fun- unitary operator from equation (27) for a single ion damental to fast gates. π-pulses with arbitrarily 7 (a) � ◆◆●■●■ (b) � ●■●■ ◆ ■ ����� ●■◆◆●■ ����� ●●■ ■ ■ ������ ����� ● ◆■ ��●������� ������ ����� ● ●■ ● ● ● ● ● �������� -��� ����� ● ◆●■ ◆●■ ◆●■ ◆■ �� -��� ����� ■ ■ ■ ●■ �� ◆ ●■ ◆● ■ ■ ■ ��-� ��-� � �� �� �� �� ��� � �� �� �� �� ��� �������������(��) �������������(��) FIG. 2. Infidelity following the (a) GZC and (b) FRAG gate operations with different numbers of pulses, governed by n. The effect of changing the duration of the pulses composing the gate is shown. The initial motional state is |2(cid:105) |2(cid:105) , the second excited number state for each mode. We determine the mean and the standard deviation (error c r bars) by varying the phase φ for a given pulse duration τ. high fidelity can be constructed using composite ing to the pulse applied to a single ion follows: pulses [21, 22]; laser repetition rates must be suffi- ciently high to accommodate the pulse components Upulse =e−i2ξπ(σ+ei(kx+φ)+σ−e−i(kx+φ)). (31) inthisapproach. Inthissection,weconsidertheim- pact of infidelity in the π rotations on the full gate Reversing the pulse direction changes the sign of k fidelity. Imperfect π-pulses cause imperfect state intheevolutionoperator. Thepulserotationfidelity transfer,errantmomentumkicksandacquiredphase canbefoundforidealandrealpulseunitariesU(cid:48) pulse infidelity. and U respectively: pulse While different methods for performing π-pulses hpauvlseevraortyaitniognrofibduelsittnyesfsortoanlyasmeretflhuocdtuhaatisonas,fixtehde R. Fid.=Minψi(cid:12)(cid:12)(cid:104)ψi|(Up(cid:48)ulse)†Upulse|ψi(cid:105)(cid:12)(cid:12)2 (32) (1−ξ)2π2 relation to the full gate fidelity. We consider here (cid:39)1− , (33) the simplest case of square pulses to calculate the 4 relation between rotation fidelity and gate fidelity. up to third order in (1−ξ)π/2. To model the imperfect gate process, we assume a Assumingthatthesamelaserproduceseachpulse, suitable pulse length for the RWA, with δ =0: and that phase drift is minimal during the gate du- (cid:126)Ω ration (< 1 µs), φ is fixed. We fix ξ to be constant HR(cid:48)WA = 2 (σ+ei(kx+φ)+σ−e−i(kx+φ)). (30) during a gate operation to find the systematic error effects. ForΩconstantintime,aπ-pulsesatisfiesΩτ =π, The unitary for a counter-propagating pulse pair, for a pulse duration τ. An approximate π-pulse sat- with first pulse direction z, can be expressed in the isfies Ωτ =ξπ, with ξ (cid:39)1. The unitary correspond- computational basis {e,g}: (cid:18)e−izkx(cos(kx)cos(πξ)+izsin(kx)) cos(kx)sin(πξ)(−icos(φ)+sin(φ))(cid:19) U (z,ξ)= , (34) pair cos(kx)sin(πξ)(−icos(φ)−sin(φ)) eizkx(cos(kx)cos(πξ)−izsin(kx)) such that ξ =1 gives states, and represents the angle of rotation on the (cid:18)e−2izkx 0 (cid:19) Bloch sphere. It does not affect the magnitude of U (z,1)=− , (35) rotation which provides the error, and we set φ=0 pair 0 e2ikx for simplicity. withtheexpectedstate-dependentmomentumkicks and no φ-dependence. The φ-dependence for im- perfect pulses is in the terms of equation (34) cor- Themotionalandinternaloperatorscommutefor responding to population transfer between internal separate ions, and the unitary for a two-ion imper- 8 increaseswithn,theerrorsineachpulsecausecom- (a)��� pounding gate infidelities, shown in Figure 4. For ����������� ������◆◆◆◆ ◆ ◆ ◆ ◆ ◆ ◆ ◆ ��������●������������� bmofoattthhicethimnecoFrteiRoanAseaGslsiantnatdtheGefoZmlCleoawgnaintagensad,Fgsatigatuenrdaeas4rtdhshedoenwvuismadtbiroeanr- ������� ������■■■ ■ ■ ■ ■ ■ ■ ■ ■ ◆■ �� oroftpautilosnesiinnficdreealisteys.leFssortheaanch1s0c−h5emise,rewqiuthirend(cid:46)for10a, ���●●● ● ● ● ● ● ● ● ● gate fidelity above 0.99, or rotation infidelity less ������ ������ ������ ������ ������ ������ than 10−4 is required for a gate fidelity above 0.9. �-���������������� Both schemes are similarly affected by pulse error (b)��������◆●■◆●■◆●■ ◆●■ ◆●■ ◆■ ◆■ comUspionugnsdqiunagrweiptuhlspeusl,sewhnuermebtehre. pulse area is pro- ������������ ������������ ● ● ◆●■ ◆■ ◆■ ◆■ ��������●■�������������� preoqrutiiorenmalentots.ξ,Swysetecmanatfiicnpdutlhseeapruealseerarorera(1st−abξ)iliotny ���� ● ◆ � the order of 0.4% is permissible for fidelity better ���� ● than 0.9 and n (cid:46) 10 for each scheme. Pulse area ���� ● error (1−ξ) ≤ 0.2% is required for a fidelity above ������ ������ ������ ������ ������ ������ �-���������������� 0.98. Figure 5 demonstrates the impact of system- aticpulse-areaerrorsontheinternalstateandmode occupation following a GZC gate with n = 1; pop- FIG. 3. A GZC gate with n=1 is applied with varying rotationfidelityforindividualpulsesanddifferentinitial ulation is lost to other internal states with variable motional occupation. (a) The mean and standard devi- motional mode occupation. ation (error bars) in the occupation of motional states following the gate are shown. (b) Gate fidelity is shown as a function of pulse rotation fidelity. V. CONCLUSIONS The duration of fast gates directly impacts their fect π-pulse is given by fidelity and scalability with the number of trapped ions. We have presented a technique for calculat- U (z) 2pulse ing gate fidelities to first order in the error for large =e−i2ξπ(σ1+eizkx1+σ1−e−izkx1+σ2+eizkx2+σ2−e−izkx2). numbers of applied pulses. Applying this technique to two ions, we have demonstrated that pulse er- (36) rors cause compounding infidelity with the number Using this unitary we construct the evolution from of pulses composing the gate. Gate duration scales pulse pairs, which we intersperse with the motional with the number of pulses, so this pulse fidelity re- free evolution unitaries to build up our gate oper- quirementisofgreatimportanceforusingfastgates ations. The necessary pulse times for gates with for scalable QIP. Pulse infidelity less than 10−5 is varying numbers of pulses are found according to required for gate fidelity above 0.99 with up to 140 the applied scheme [2, 6]. pulse pairs in the FRAG and GZC gate schemes. We have also shown that different numbers of ap- Figure 3 shows the effect of the initial motional pliedpulsesdonotsignificantlyalterthevalidRWA state on final mode occupation and gate fidelity for regime: pulsedurationsmuchlongerthantheatomic aGZCgatewithn=1. Increasinginfidelityinindi- transition period are required. Experimental imple- vidual π-pulses, or rotation infidelity, damages the mentation of a fast gate, which requires fast and full gate fidelity and increases both the mean and robust π-pulses, will be a significant step towards standarddeviationofthemodeoccupationafterthe large-scale QIP with ions. gate. The initial motional state before the gate is applied affects the magnitude of the gate infidelity. There is not a clear relationship between initial mo- tional state and infidelity; however each initial state VI. ACKNOWLEDGMENTS isharmedbypulseerrors. Rotationinfidelityaround 3×10−4 is required for gate fidelity better than 0.9, This work was supported by the Australian Re- or rotation infidelity around 10−5 for a gate fidelity search Council Centre of Excellence for Quan- above 0.99. tum Computation and Communication Technol- Higher numbers of perfect π-pulses provide faster ogy (Project number CE110001027) (ARRC), gate times, more stability, and improved scalabil- Australian Research Council Future Fellowship ity. However, as the number of pulses in the gate (FT120100291)(JJH)aswellasDP130101613(JJH, 9 (a) �� (c)�� ������������� �� ● ● ● ● �●�■��������� ������������� �� �●�■��������� ����� �◆▲●■◆▲●■◆▲●■ ◆▲●■ ◆▲●■ ◆▲●■ ◆▲●■ ◆▲■ ◆▲■ ◆▲■ ◆▲■ ◆▲ �� ����� �◆●■◆●■◆●■ ◆●■ ◆●■ ◆●■ ◆●■ ◆●■ ◆●■ ◆●■ ◆●■ ◆ � � � ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ �-���������������� �-���������������� (b������������) ������������◆▲●■◆▲●■◆▲●■ ◆▲●■ ◆▲●■ ◆▲●■ ◆▲●■ ◆▲●■ ◆▲●■ ◆▲■ ◆▲ �◆●�■���������� (d������������)������������◆●■◆●■◆●■ ◆●■ ◆●■ ◆●■ ◆●■ ◆●■ ◆●■ ◆●■ ◆●■ �●�■��������� ● ■ ◆ � ��� ● ▲ � ��� ��� ��� ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ ������ �-���������������� �-���������������� FIG. 4. A GZC (a,b) and FRAG (c,d) fast gate are applied to |ee(cid:105)|1(cid:105) |1(cid:105) with varying n and pulse rotation error. c r (a,c): The mean and standard deviation (error bars) in the occupation of the COM mode are shown following the gate applied to the |ee(cid:105) internal state. (b,d): Gate fidelity is shown as a function of pulse rotation infidelity. ARRC). [1] D.P.DiVincenzo,“ThePhysicalImplementationof Modes,”PhysicalReviewLetters,vol.97,p.050505, Quantum Computation,” Fortschritte Der Physik, Aug. 2006. vol. 48, no. 9-11, pp. 771–83, 2000. [9] G.-D. Lin, S.-L. Zhu, R. Islam, K. Kim, M.-S. [2] J.Garc´ıa-Ripoll,P.Zoller,andJ.Cirac,“SpeedOp- Chang, S. Korenblit, C. Monroe, and L.-M. Duan, timizedTwo-QubitGateswithLaserCoherentCon- “Large-scale quantum computation in an anhar- trolTechniquesforIonTrapQuantumComputing,” monic linear ion trap,” EPL (Europhysics Letters), PhysicalReviewLetters,vol.91,pp.2–5,Oct.2003. vol. 86, p. 60004, June 2009. [3] L.-M.Duan,“ScalingIonTrapQuantumComputa- [10] P.Zou,J.Xu,W.Song,andS.-L.Zhu,“Implemen- tionthroughFastQuantumGates,”PhysicalReview tationoflocalandhigh-fidelityquantumconditional Letters, vol. 93, p. 100502, Sept. 2004. phasegatesinascalabletwo-dimensionaliontrap,” [4] J. Garc´ıa-Ripoll, P. Zoller, and J. Cirac, “Coherent Physics Letters A, vol. 374, pp. 1425–1430, Mar. control of trapped ions using off-resonant lasers,” 2010. Physical Review A, vol. 71, p. 062309, June 2005. [11] S.-L.Zhu,C.Monroe,andL.-M.Duan,“Arbitrary- [5] C. D. B. Bentley, A. R. R. Carvalho, D. Kielpinski, speed quantum gates within large ion crystals and J. J. Hope, “Fast gates for ion traps by split- through minimum control of laser beams,” Euro- ting laser pulses,” New Journal of Physics, vol. 15, physics Letters, vol. 73, pp. 485–91, Feb. 2006. p. 043006, Apr. 2013. [12] M.Petrasiunas,M.Hussain,J.Canning,M.Steven- [6] C.D.B.Bentley,A.R.R.Carvalho,andJ.J.Hope, son,andD.Kielpinski,“Picosecond554nmyellow- “Trapped ion scaling with pulsed fast gates,” New green fiber laser source with average power over Journal of Physics, vol. 17, p. 103025, Oct. 2015. 1 W,” Optics Express, vol. 22, no. 15, pp. 17716– [7] J. Mizrahi, B. Neyenhuis, K. G. Johnson, W. C. 17722, 2014. Campbell, C. Senko, D. Hayes, and C. Monroe, [13] W. Campbell, J. Mizrahi, Q. Quraishi, C. Senko, “Quantum control of qubits and atomic motion D. Hayes, D. Hucul, D. Matsukevich, P. Maunz, using ultrafast laser pulses,” Applied Physics B, and C. Monroe, “Ultrafast Gates for Single Atomic vol. 114, pp. 45–61, Nov. 2014. Qubits,” Physical Review Letters, vol. 105, pp. 1–4, [8] S.-L. Zhu, C. Monroe, and L.-M. Duan, “Trapped Aug. 2010. IonQuantumComputationwithTransversePhonon 10 [14] J.Mizrahi,C.Senko,B.Neyenhuis,K.G.Johnson, (a)��� ◆○▼▲ W. C. Campbell, C. W. S. Conover, and C. Mon- ■ ��� ● roe, “Ultrafast Spin-Motion Entanglement and In- |〉�������������� ������ ◆●■▲ ����ξ��������������� [15] MtLtieeor.tnftAeerar.osnNm,diveeQotlsrlue.yna1wn1at0inut,dhmpIa..I2LnS0.fio3nCr0ghm0leu1aa,AtniMtogon,am.yQ,Cu”2a0aP1nm3htub.ymrsiidcCagleo:mRCepvuaitmeaw-- ��� ▼ ����� bridge University Press, 2000. ○ � [16] E. Knill, “Quantum computing with realistically ��� ◆●○■▲▼◆●○■▼▲ ◆●○■▲▼◆●○■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼ noisy devices,” Nature, vol. 434, pp. 39–44, Mar. � �� �� �� �� �� 2005. �������������� (b)�����● [17] J.Benhelm,G.Kirchmair,C.F.Roos,andR.Blatt, “Towards fault-tolerant quantum computing with ■ trapped ions,” Nature Physics, vol. 4, pp. 463–6, |〉�������������� ��������������������◆○▲▼◆●○■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▲▼◆○●■▼▲◆○●■▼▲◆○●■▲▼◆○●■▲▼◆○●■▼▲◆○●■▲▼◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▲▼◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▲▼◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▲▼◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲ ◆●○■▲▼ ������ξ������������������� [[1189]] AAAmc2e0..pes0Gasr5e.sM.isu2l,cr.”0he0rsP8Sish.tttoye,asNicnco.eam,LlapRnGaegrv.efioerrwIdema,Arlae,nahvdn,odMl.Ji.7d.N1e,aiPepl.l.sqe0uHn6a,2on“3mtD1ue0mi,s,tJapaunrnncodee- � �� �� �� �� �� D.Leibfried,“Pulsedforcesequencesforfastphase- �������������� insensitive quantum gates in trapped ions,” New (c)�����● Journal of Physics, vol. 16, p. 053049, May 2014. [20] R. Gerritsma, G. Kirchmair, F. Z¨ahringer, J. Ben- ����� |〉�������������� ��������������������◆■●■●■●■●■●■●● ◆●■▲▼ �����ξ������������������� [21] Bhsdvuoee.rlcl.meTam5y,.0eoT,Rnfpot.CrpoBo.afsl1otaI3hvIt–,te”,1ab9nTar,dnahDndeNceEhCc..iu.nV2rgoF.0p0f.Vre8aaRi.tcnoationPosonh,vsy,“soPi“fcSratemlhceJioso4oiuoptnrhn2Pamcol3emD/a2--, �����○▲▼◆○▲▼◆○▼▲◆○▼▲◆○▼▲◆○▼▲◆○■▲▼◆○■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▲▼◆○●■▼▲◆○●■▲▼◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲◆○●■▼▲ ○ � pproosciteesspinugls,e”sPfohryshicigahl-RfiedveileitwyAqu,avnotlu.m83,inpfo.r0m53a4ti2o0n, � �� �� �� �� �� �������������� May 2011. [22] S.S.IvanovandN.V.Vitanov,“High-fidelitylocal addressingoftrappedionsandatomsbycomposite FIG. 5. Population in (a) |ee(cid:105), (b) |eg(cid:105) and (c) |gg(cid:105) sequences of laser pulses,” Optics Letters, vol. 36, states after a GZC gate applied to |ee(cid:105)⊗|2(cid:105)c with n = no. 7, pp. 1275–1277, 2011. 1. The fraction ξ of a perfect square π-pulse performed determinestherestorationoftheinternalstateandCOM motional mode to the initial state.

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.