Transport quantum logic gates for trapped ions

• The approach targets difficult trapped-ion optics, including near-UV wavelengths, limited fiber technology, and demanding power requirements.Source: §I · p. 2
• Controlled transport of ions can replace precise temporal control of laser intensity while implementing universal quantum computation.Source: §I · p. 2
• Global laser switching transfers individual operation control to ion motion, relaxing timing and on/off-ratio requirements.Source: §I · p. 2
• One laser-beam set can support parallel operations, potentially reducing the total laser power required without higher power for parallelism.Source: §I · p. 2
• A proposed extension uses microfabricated permanent magnets to implement a transported two-qubit gate without laser beams.Source: §I · p. 3

• Repeating spatial preconfiguration and transport operations realizes one-qubit rotations, two-qubit gates, and measurements between arbitrary qubits, sufficient for universal computation [17].Source: §II · p. 3
• Laser beams are switched collectively, while scheduled ions or ion pairs are transported through them for one- and two-qubit operations.Source: §II · p. 4
• Measurements can run in parallel with position-resolving detectors or serially through scheduled transports when position resolution is unavailable.Source: §II · p. 4
• Transport gates use ion motion for individual control, relaxing laser timing and enabling one beam set to serve parallel operations across the array.Source: §II · pp. 4–5
• Transport through the beam separately controls each qubit and makes gate interactions insensitive to slow in-plane pointing instabilities, requiring stabilization mainly perpendicular to ion motion.Source: §II · p. 5

• Transported ions can use the stationary-ion interaction Hamiltonians when their coordinates are transformed as r → r_lab + v t during motion.Source: §III · p. 6
• The qubit is encoded in two hyperfine ground states, and stimulated Raman transitions couple these states through two laser beams after eliminating off-resonant levels.Source: §III · p. 6
• The effective interaction contains sideband transitions, motional-state couplings, and AC-Stark level shifts with respective Rabi frequencies or shifts.Source: §III · p. 6
• Choosing δ0 ≃ ω0 with co-propagating beams favors motion-independent spin flips, whereas δ0 ≃ ωm and |Δk| ≃ |k| favors state-dependent motional driving.Source: §III · p. 7
• AC-Stark shifts can be balanced through laser intensity or polarization choices without appreciably reducing the relevant Rabi frequencies.Source: §III · p. 7

• The analysis assumes all laser beams have Gaussian TEM00 transverse profiles.Source: §IV · p. 8
• Near the beam waist, where |z| ≪ |z_r|, the model approximates w(z) ≈ w0 and neglects Gouy phase and wavefront-curvature terms.Source: §IV · p. 8
• Under these approximations, the Gaussian-beam expression simplifies to capture the spatial dependence relevant to the subsequent analysis.Source: §IV · p. 8

V. ONE-QUBIT ROTATIONS

• The rotation angle θ is set by integrating the time-dependent coupling during a beam traverse, so transport speed controls the resulting rotation.A π-pulse example uses w0 = 20 µm, Ωm = 250 kHz, and velocity v ≃25 m/s.Source: §V · pp. 9–10
• For perpendicular beams, the rotation infidelity 1−F ≃(Δθ/θ)^2 is smaller than 10^-4 when the initial ion-beam-center distance satisfies D > 2.6w0.Source: §V · p. 9
• The rotation phase φ is controlled by the relative laser phase, while propagation produces a phase shift Δφp ≃2π(s0/Λ0) between zones separated by s0.Because the shift cannot be neglected across an extended array, phase bookkeeping is required.Source: §V · pp. 9–10
• Using the same beam pair at multiple locations removes independent phase control, requiring ion repositioning, multiple phase settings, or sequential switching.The added sequential overhead may be limited because one-qubit rotations are typically faster than two-qubit gates.Source: §V · p. 10
• Arbitrary one-qubit rotations remain possible with a fixed relative phase by sandwiching a rotation between two Z operations.Transport through a single Raman beam produces the required Z(φ) operations via a time-dependent AC Stark shift.Source: §V · p. 10
• Universal one-qubit operations can be implemented by transporting ions through one laser beam and then two beams with suitable detuning and velocity.Source: §V · pp. 8, 11

Motional disentanglement

• Small final motional displacements are obtained when the pulse envelope has small Fourier components at the detuning, allowing non-Gaussian beam shapes as well.This suppresses residual entanglement between motional and internal states without requiring a strictly Gaussian beam profile.Source: §VI · p. 14

Phase-space trade-off

• The displacement decays exponentially in p^2, while the logical phase decays more slowly, so appreciable phase remains even when residual displacement is very small.Figure 3 presents the normalized displacement and phase versus p, exposing the trade-off between motional closure and phase accumulation.Source: §VI · p. 15

Alternative implementation

• Using two transits can cancel motional displacements while adding gate phases, but it requires timing controlled to a small fraction of 2π/δ.A sufficiently small residual displacement is instead readily achieved for reasonably large p.Source: §VI · p. 15

Logical gate

• The resulting phase gate differs from the standard two-qubit phase gate only by one-qubit z-rotations.Its logical phase is ΦL = Φ↑↑ + Φ↓↓ − (Φ↑↓ + Φ↓↑).Source: §VI · p. 15

Gate optimization

• Maximum logical-phase modulus occurs when ∆k_zd = nπ, requiring the ion spacing to equal an integer multiple of the standing wave’s half-period.The transported harmonic well is adjusted to produce this spacing and maximize the logical phase.Source: §VI · p. 15

• For p →∞, no motional energy is deposited, making p a gate-adiabaticity parameter that controls proximity to the phase-space origin.Increasing p also increases the number of windings per time interval τ.Source: §VI · p. 16
• For n < 10, the necessary detuning δ ranges from 29.63 MHz to 768 kHz, too large to satisfy |δ| ≪ωCOM reliably.Meaningful results in this regime require including the faster rotating terms in Eq.(12).Source: §VI · p. 17
• For n ≥10, transport speed v_n is reasonable and gate durations τ_n on the order of 1 µs are achievable if sufficient laser power produces the required Rabi frequencies Ω↓.The assumed parameters are ωCOM = 2π × 4 MHz, w0 = 20 µm, Ω↑= −1/2Ω↓, wavelength 313 nm, and p = 3.48 for ε ≤10−4.Source: §VI · p. 17
• Gate times are around 5 µs, comparable to present state-of-the-art phase gates; γ must be between 60 and 40 degrees, with Rabi frequencies close to values realized in previous experiments.Choosing γ close to 90 degrees reduces the number of polarization-gradient periods encountered, but requires compensation, for example by a higher Rabi rate requiring more laser power.Source: §VI · pp. 16–17

VII. EFFECTS OF SYMPATHETIC COOLING

• Adding refrigerator ions complicates the normal-mode description, especially for substantially different ion masses, but does not change the method’s essential features.Source: §VII · p. 17
• For reflection-symmetric ion configurations, normal modes satisfy v1 = ±v2, and suitable modes require large qubit amplitudes while keeping η2 ∝1/ωv sufficiently large.Stretch-type modes with v1 = −v2 are identified as optimal candidates under the stated mode criteria.Source: §VII · p. 18
• In the strongly adiabatic regime, small motional excursions and rapid phase-space revolutions make gates more robust to photon recoil and heating.Reasonable fidelity requires the heating rate to remain small compared with the phase-space revolution rate.Source: §VII · p. 18
• Gates can be executed during sympathetic recooling when the combined phonon-scattering rates from heating and cooling remain small compared with the detuning δ.Source: §VII · p. 18
• Sympathetic cooling can preserve the Lamb-Dicke regime during a gate even when unchecked heating would drive ions out of it.Source: §VII · p. 18

Alternative laser implementation

• The scheme works for qubits with field-insensitive qubit transition frequencies, but creating the required beam frequencies entails added technical difficulty; for 9Be+ hyperfine-state ion-qubits, the example requires acousto-optical modulation at ω0/2, where the smallest typically considered ω0 is 2π × 1.25 GHz.Source: §VIII · p. 19

Magnetic transport gates

• Periodic magnetic field patterns combined with ion transport can exert state-dependent forces for a two-qubit gate without laser mediation.The proposal uses alternating current-carrying wires or permanently magnetized strips and tunes transport near a motional-mode resonance.Source: §VIII · p. 20

Constraints and scaling

• Magnetic gates require adiabatic spin following, which limits gate speed through the trap frequency and transport-induced field variation.Closer ion-electrode proximity strengthens field modulation but might increase anomalous heating, while enabling smaller trap electrodes.Source: §VIII · p. 21
• Unlike laser-mediated gates, magnetic gates avoid spontaneous emission and may permit stronger confinement, faster gates, and more flexible trap-array scaling.The relevant Lamb-Dicke parameter can be adjusted through the microfabricated pattern period rather than a laser wavelength.Source: §VIII · pp. 21–22

IX. SUMMARY AND CONCLUSIONS

• Controlled ion transport enables parallel quantum logic operations with relaxed temporal and spatial control requirements for laser beams.The approach is designed to simplify the optics needed for scaling multizone trap arrays.Source: §IX · p. 22
• Universal one-qubit rotations and two-qubit gates can be implemented with existing technology in a multizone transport architecture.The paper analyzes these operations after outlining the basic trap-array architecture.Source: §IX · p. 22
• Periodic magnetic field patterns combined with ion transport provide a possible laser-free route to state-dependent forces.The authors present this as an extension of the two-qubit gate mechanism.Source: §IX · p. 22
• The magnetic approach might enable more flexible trap-array scaling and reduce limitations on fidelity and gate speed posed by laser-mediated two-qubit gates.This is presented as a possible benefit rather than an established demonstrated result.Source: §IX · p. 22