All writing
Physics breakthroughs

How neutral atoms in optical tweezers build a quantum computer

Arrays of tightly focused laser beams trap individual atoms, while laser-driven Rydberg states provide the strong interactions needed to entangle and shuttle qubits without physical wires.

Illustration of neutral atoms trapped in an array of optical potentials created by focused laser beams
Neutral-atom qubits trapped in optical potentials formed by laser light. Credit: NIST.

The neutral-atom approach to quantum hardware

Building quantum computing hardware requires balancing two opposing physical demands. A qubit must decouple from stray environmental noise during idle storage, yet it must interact strongly with selected partners to execute multi-qubit logic gates.

Different physical platforms resolve this trade-off in distinct ways. Superconducting processors rely on lithographic planar circuits coupled through microwave resonators. This design permits fast gate operations, but micro-fabrication inhomogeneities produce site-to-site variations, and fixed on-chip wiring constrains connectivity. Trapped ions provide identical atomic transitions and long coherence times, but their shared positive charge causes long-range Coulomb repulsion, which complicates dense two-dimensional scaling.

Neutral atoms in their electronic ground states offer an alternative starting point. Carrying zero net electric charge, they do not experience Coulomb repulsion. At separations of several micrometers, their permanent electric dipole moments vanish, yielding negligible mutual interactions. Every atom of a chosen isotope is identical by fundamental symmetry, removing the device variations found in lithographic circuits.

The platform addresses the conflict between isolation and interaction by operating in two physical regimes. During storage and transport, atoms remain in electronic ground states, where they interact weakly and decouple from stray background fields. When entangling gates are required, resonant laser pulses promote valence electrons to high-lying Rydberg states. These extended orbitals introduce strong dipole-dipole interactions, which persist for the laser pulse duration and vanish once atoms return to the ground state.

Optical tweezers and the dipole force

Neutral atoms are confined without physical walls using optical tweezers: far-off-resonance optical dipole traps formed by focusing laser beams to sub-micrometer waists through high numerical aperture objectives.

The trapping mechanism relies on the AC Stark effect. An oscillating optical electric field E(r,t)\mathbf{E}(\mathbf{r}, t) induces an electric dipole moment p=α(ω)E\mathbf{p} = \alpha(\omega) \mathbf{E} in the atom, where α(ω)\alpha(\omega) is the dynamic polarizability at laser frequency ω\omega. The time-averaged interaction energy yields the conservative optical dipole potential:

Udip(r)=−12ϵ0cRe[α(ω)]I(r)U_{\text{dip}}(\mathbf{r}) = -\frac{1}{2\epsilon_0 c} \text{Re}[\alpha(\omega)] I(\mathbf{r})

where I(r)=12ϵ0c∣E(r)∣2I(\mathbf{r}) = \frac{1}{2} \epsilon_0 c |\mathbf{E}(\mathbf{r})|^2 is the cycle-averaged laser intensity. In a two-level atom model with transition frequency ω0\omega_0 and spontaneous decay rate Γ\Gamma, the real dynamic polarizability is:

Re[α(ω)]≈6πϵ0c3Γ/ω02ω02−ω2\text{Re}[\alpha(\omega)] \approx 6\pi\epsilon_0 c^3 \frac{\Gamma/\omega_0^2}{\omega_0^2 - \omega^2}

Tuning the laser below atomic resonance (ω<ω0\omega < \omega_0) produces red detuning (Δ=ω−ω0<0\Delta = \omega - \omega_0 < 0) and a positive polarizability, Re[α(ω)]>0\text{Re}[\alpha(\omega)] > 0. The resulting negative potential draws polarizable atoms toward the intensity maximum.

For a Gaussian beam propagating along the zz-axis with power PP, waist w0w_0, and Rayleigh range zR=πw02/λz_R = \pi w_0^2 / \lambda, the intensity profile is:

I(r,z)=2Pπw2(z)exp⁡(−2r2w2(z))I(r, z) = \frac{2P}{\pi w^2(z)} \exp\left( -\frac{2r^2}{w^2(z)} \right)

with w(z)=w01+(z/zR)2w(z) = w_0 \sqrt{1 + (z / z_R)^2}. Expanding near the focal maximum gives harmonic trap oscillation frequencies:

ωr=4U0mw02,ωz=2U0mzR2\omega_r = \sqrt{\frac{4 U_0}{m w_0^2}}, \quad \omega_z = \sqrt{\frac{2 U_0}{m z_R^2}}

where U0=∣Udip(0,0)∣U_0 = |U_{\text{dip}}(0, 0)| is the trap depth and mm is the atomic mass. In rubidium experiments (P≈5 to 10 mWP \approx 5\text{ to }10\,\text{mW}, w0≈0.8 to 1.0 μmw_0 \approx 0.8\text{ to }1.0\,\mu\text{m}, λ≈850 nm\lambda \approx 850\,\text{nm}), trap depths reach U0/kB≈1 mKU_0 / k_B \approx 1\,\text{mK} (≈20 MHz\approx 20\,\text{MHz}). Atoms pre-cooled via optical molasses to 10 to 20 μK10\text{ to }20\,\mu\text{K} reside deep within the conservative potential.

In addition to the conservative potential, atoms scatter photons from the trapping laser through Rayleigh scattering. The off-resonance scattering rate is:

Γscatt(r)=1ℏϵ0cIm[α(ω)]I(r)≈ΓℏΔUdip(r)\Gamma_{\text{scatt}}(\mathbf{r}) = \frac{1}{\hbar\epsilon_0 c} \text{Im}[\alpha(\omega)] I(\mathbf{r}) \approx \frac{\Gamma}{\hbar\Delta} U_{\text{dip}}(\mathbf{r})

Because the dipole potential scales inversely with detuning as 1/Δ1/\Delta while the photon scattering rate scales as 1/Δ21/\Delta^2, optical tweezers operate in the far-detuned regime. Large red detunings (often tens of nanometers away from the rubidium D1 and D2 resonance lines) ensure that optical trap depths remain sufficient for deep spatial confinement while holding spontaneous scattering rates below several photons per second. This low scattering rate prevents photon recoil heating from ejecting atoms during computational cycles.

Loading single atoms and assembling defect-free arrays

Focusing an optical tweezer waist below one micrometer induces light-assisted collisional blockade when loading from a laser-cooled background vapor. This collisional loss restricts tweezer occupation to either zero or one atom.

Under near-resonant cooling light, two atoms in the same microtrap absorb a photon and collide along an attractive molecular potential:

Rb+Rb+ℏωcooling⟶Rb2∗⟶Rb+Rb+ΔEkin\text{Rb} + \text{Rb} + \hbar\omega_{\text{cooling}} \longrightarrow \text{Rb}_2^* \longrightarrow \text{Rb} + \text{Rb} + \Delta E_{\text{kin}}

When the quasimolecule decays back to the repulsive ground curve, the acquired kinetic energy ΔEkin\Delta E_{\text{kin}} splits between both atoms. Because this energy gain exceeds the millikelvin trap depth U0U_0, both atoms escape the tweezer. Consequently, whenever a second atom enters an occupied site, both are lost. This pairwise mechanism bounds single-atom loading efficiency to approximately 50%50\%.

Assembling deterministic, defect-free arrays requires active rearrangement:

  1. Static arrays of hundreds of microtraps are formed using a spatial light modulator that imprints computed phase holograms on a trapping laser, generating two-dimensional grids of diffraction-limited focal spots.
  2. The initial stochastic occupancy is mapped within tens of milliseconds using fluorescence imaging. Cooling beams scatter photons from trapped atoms, which are collected by an electron-multiplying charge-coupled device or scientific CMOS camera.
  3. A control processor calculates collision-free relocation paths. Mobile tweezers driven by two-dimensional acousto-optic deflectors extract atoms from filled sites and transport them to vacant target sites.

This rearrangement produces defect-free target arrays containing hundreds of resolved atoms with filling fractions exceeding 99%99\% in under fifty milliseconds.

Encoding qubits in ground-state hyperfine levels

Long-lived computational states are encoded in the ground electronic manifold 52S1/25^2S_{1/2} of alkali atoms such as rubidium-87 (87Rb^{87}\text{Rb}, nuclear spin I=3/2I = 3/2).

The computational basis states are mapped to two hyperfine sublevels:

∣0⟩≡∣F=1,mF=0⟩,∣1⟩≡∣F=2,mF=0⟩|0\rangle \equiv |F = 1, m_F = 0\rangle, \quad |1\rangle \equiv |F = 2, m_F = 0\rangle

These states are separated by the ground-state hyperfine splitting:

ΔνHF≈6.8346826109 GHz\Delta\nu_{\text{HF}} \approx 6.8346826109\,\text{GHz}

This transition provides an atomic clock reference. Because both states have magnetic quantum numbers mF=0m_F = 0, their differential Zeeman shift vanishes to first order in the magnetic field:

∂ΔE∂B∣B→0=0\left. \frac{\partial \Delta E}{\partial B} \right|_{B \to 0} = 0

Residual magnetic sensitivity arises at second order through the Breit-Rabi formula:

ΔE(2)=βB2\Delta E^{(2)} = \beta B^2

where β≈575 Hz/G2\beta \approx 575\,\text{Hz/G}^2 for 87Rb^{87}\text{Rb}. Surrounding the chamber with mu-metal shielding and applying stable Helmholtz bias fields suppresses magnetic dephasing.

Qubit register initialization begins with optical pumping. By illuminating the laser-cooled atoms with circularly polarized light tuned to the 52S1/2,F=2→52P3/2,F′=25^2S_{1/2}, F=2 \to 5^2P_{3/2}, F'=2 transition alongside a repumping beam, atoms undergo spontaneous optical cycling until they settle into the dark state ∣F=1,mF=0⟩|F=1, m_F=0\rangle, achieving state preparation fidelities above 99%99\%.

Two-photon Raman transitions provide coherent local control over single-qubit rotations. Two laser beams with frequency difference ω1−ω2≈ΔνHF\omega_1 - \omega_2 \approx \Delta\nu_{\text{HF}} drive transitions via a virtual intermediate level detuned by several gigahertz from the 52P1/25^2P_{1/2} or 52P3/25^2P_{3/2} manifolds. Aligning the two Raman beams in a co-propagating geometry sets the effective wavevector to Δk=k1−k2≈0\Delta\mathbf{k} = \mathbf{k}_1 - \mathbf{k}_2 \approx 0, which suppresses Doppler sensitivity and motional dephasing during single-qubit rotations. The bare dephasing time T2∗T_2^* for these clock qubits spans one to four seconds, and dynamical decoupling sequences such as Carr-Purcell-Meiboom-Gill or XY-8 pulse trains extend coherence times past ten seconds.

The Rydberg state and van der Waals interactions

Ground-state neutral atoms separated by several micrometers exhibit negligible mutual interaction. To generate entanglement for multi-qubit logic, atoms are excited to high-lying Rydberg states using coherent laser pulses.

A Rydberg state is an electronic configuration where a valence electron is promoted to a high principal quantum number nn, typically between n=60n = 60 and n=75n = 75 (such as ∣70S1/2⟩|70S_{1/2}\rangle). Single-photon ultraviolet light near 297 nm297\,\text{nm}, or two-photon transitions near 420 nm420\,\text{nm} and 1013 nm1013\,\text{nm}, drives these excitations.

Rydberg states exhibit extreme physical properties that scale steeply with nn:

  • The mean electron orbital radius scales as ⟨r⟩≈12[3n2−l(l+1)]a0∝n2a0\langle r \rangle \approx \frac{1}{2} [3n^2 - l(l+1)] a_0 \propto n^2 a_0, with Bohr radius a0≈0.0529 nma_0 \approx 0.0529\,\text{nm}. For n=70n = 70, the orbital radius exceeds 250 nm250\,\text{nm}.
  • The electric dipole matrix element to adjacent states scales as μ=⟨ns∣er∣n′p⟩∝n2ea0\mu = \langle n s | e r | n' p \rangle \propto n^2 e a_0.
  • The atomic polarizability scales as αR∝n7\alpha_R \propto n^7, making Rydberg levels responsive to weak electric fields.
  • The radiative lifetime scales as τ∝n3\tau \propto n^3. At room temperature, blackbody-induced transitions reduce this lifetime to 150 to 300 μs150\text{ to }300\,\mu\text{s}, which remains much longer than microsecond gate operations.

Separated by distance RR without static electric fields, two Rydberg atoms interact through the dipole-dipole Hamiltonian:

V^dd=14πϵ0R3[d^1⋅d^2−3(d^1⋅n^)(d^2⋅n^)]\hat{V}_{\text{dd}} = \frac{1}{4\pi\epsilon_0 R^3} \left[ \hat{\mathbf{d}}_1 \cdot \hat{\mathbf{d}}_2 - 3(\hat{\mathbf{d}}_1 \cdot \hat{\mathbf{n}})(\hat{\mathbf{d}}_2 \cdot \hat{\mathbf{n}}) \right]

where d^i\hat{\mathbf{d}}_i is the dipole operator and n^\hat{\mathbf{n}} is the interatomic unit vector. For two atoms in the pair state ∣ns,ns⟩|ns, ns\rangle, the dipole-dipole Hamiltonian couples them to intermediate pair states ∣n′p,n′′p⟩|n'p, n''p\rangle with an energy defect ΔF=2E(ns)−E(n′p)−E(n′′p)\Delta_F = 2E(ns) - E(n'p) - E(n''p). When this energy defect is non-zero, second-order perturbation theory produces an isotropic van der Waals potential:

V(R)=−∑n′,n′′∣⟨ns,ns∣V^dd∣n′p,n′′p⟩∣2ΔF=C6R6V(R) = -\sum_{n', n''} \frac{|\langle ns, ns | \hat{V}_{\text{dd}} | n'p, n''p \rangle|^2}{\Delta_F} = \frac{C_6}{R^6}

The dispersion coefficient C6C_6 scales with transition dipole moments squared divided by the energy spacing to adjacent manifolds:

C6∝∣μ∣4ΔF∝(n2)4n−3=n11C_6 \propto \frac{|\mu|^4}{\Delta_F} \propto \frac{(n^2)^4}{n^{-3}} = n^{11}

This n11n^{11} scaling amplifies interactions by twelve orders of magnitude over ground states. For 87Rb^{87}\text{Rb} in ∣70S1/2⟩|70S_{1/2}\rangle, C6/h≈860 GHz⋅μm6C_6 / h \approx 860\,\text{GHz}\cdot\mu\text{m}^6. At R=3 μmR = 3\,\mu\text{m}, the interaction energy exceeds 1 MHz1\,\text{MHz}, producing strong coupling on microsecond timescales. Small external electric fields can tune the energy defect ΔF\Delta_F toward zero through DC Stark shifts. At these electric-field-induced Förster resonances, the interaction switches from the 1/R61/R^6 van der Waals regime to the longer-range 1/R31/R^3 resonant dipole-dipole regime.

The Rydberg blockade and two-qubit entangling gates

The van der Waals interaction between Rydberg atoms gives rise to the Rydberg blockade, which provides the mechanism for two-qubit entangling gates.

An ensemble of neutral atoms driven by a laser field coupling qubit state ∣1⟩|1\rangle to Rydberg state ∣r⟩|r\rangle with Rabi frequency Ω(t)\Omega(t) and detuning Δ(t)=ωL−ω0r\Delta(t) = \omega_L - \omega_{0r} is described by the many-body Hamiltonian:

H^=∑i=1NℏΩi(t)2σ^x,i−∑i=1NℏΔi(t)n^i+∑i<jC6∣ri−rj∣6n^in^j\hat{H} = \sum_{i=1}^N \frac{\hbar \Omega_i(t)}{2} \hat{\sigma}_{x,i} - \sum_{i=1}^N \hbar \Delta_i(t) \hat{n}_i + \sum_{i < j} \frac{C_6}{|\mathbf{r}_i - \mathbf{r}_j|^6} \hat{n}_i \hat{n}_j

where σ^x,i=∣1⟩i⟨r∣i+∣r⟩i⟨1∣i\hat{\sigma}_{x,i} = |1\rangle_i \langle r|_i + |r\rangle_i \langle 1|_i and n^i=∣r⟩i⟨r∣i\hat{n}_i = |r\rangle_i \langle r|_i.

The Rydberg blockade radius RbR_b marks the distance where the interaction energy matches the laser excitation linewidth:

V(Rb)=C6Rb6=ℏΩ  ⟹  Rb=(C6ℏΩ)1/6V(R_b) = \frac{C_6}{R_b^6} = \hbar \Omega \implies R_b = \left( \frac{C_6}{\hbar \Omega} \right)^{1/6}

For Ω/2π=4 MHz\Omega / 2\pi = 4\,\text{MHz} and n=70n = 70, the blockade radius is approximately Rb≈8.5 μmR_b \approx 8.5\,\mu\text{m}.

When two atoms are held at R<RbR < R_b, excitation of one atom shifts the doubly excited state ∣rr⟩|rr\rangle by ΔEint=C6/R6≫ℏΩ\Delta E_{\text{int}} = C_6 / R^6 \gg \hbar \Omega. This energy shift takes the laser out of resonance for the second excitation, preventing simultaneous promotion to ∣r⟩|r\rangle. Dynamics are confined to the single-excitation subspace, oscillating between ∣gg⟩|gg\rangle and (∣gr⟩+∣rg⟩)/2(|gr\rangle + |rg\rangle)/\sqrt{2} with collective Rabi frequency 2Ω\sqrt{2}\Omega.

The Levine-Pichler controlled-phase (CZCZ) gate employs this blockade under a modulated laser pulse:

  • State ∣00⟩|00\rangle does not couple to the light and accumulates no phase.
  • States ∣01⟩|01\rangle and ∣10⟩|10\rangle have one atom in ∣1⟩|1\rangle, which executes a 2π2\pi rotation through ∣r⟩|r\rangle, acquiring geometric phase ϕ1\phi_1.
  • State ∣11⟩|11\rangle couples to the field, but blockade forbids transition into ∣rr⟩|rr\rangle. The single excitation traverses a restricted path, returning to ∣11⟩|11\rangle with phase ϕ11\phi_{11}.

Selecting detuning Δ\Delta and pulse duration τ\tau to satisfy 2ϕ1−ϕ11=π2\phi_1 - \phi_{11} = \pi implements the diagonal unitary:

U^CZ=∣00⟩⟨00∣+∣01⟩⟨01∣+∣10⟩⟨10∣−∣11⟩⟨11∣\hat{U}_{CZ} = |00\rangle\langle 00| + |01\rangle\langle 01| + |10\rangle\langle 10| - |11\rangle\langle 11|

The fidelity of the resulting entangling gate is quantified through quantum state tomography and parity oscillations. Starting from the separable input state (∣0⟩+∣1⟩)⊗(∣0⟩+∣1⟩)/2(|0\rangle + |1\rangle) \otimes (|0\rangle + |1\rangle) / 2, applying the controlled-phase gate followed by a global π/2\pi/2 analysis pulse creates the maximally entangled Bell state:

∣Φ+⟩=∣00⟩+∣11⟩2|\Phi^+\rangle = \frac{|00\rangle + |11\rangle}{\sqrt{2}}

Scanning the optical phase of the analysis pulse produces oscillations in the parity operator Π^=σ^z,1σ^z,2\hat{\Pi} = \hat{\sigma}_{z,1}\hat{\sigma}_{z,2}. The amplitude of these parity oscillations, combined with populations measured in the computational basis, determines the entangled state fidelity. In 2023, Evered and colleagues reported two-qubit gate fidelities of 99.5%99.5\% in neutral-atom arrays, reaching fault-tolerance thresholds.

Dynamic shuttling and non-local circuit connectivity

In fixed planar architectures such as superconducting circuits, qubits interact only with nearest neighbors on a chip. Executing gates between distant nodes requires chains of intermediate SWAP operations that accumulate depth and error.

Neutral-atom processors avoid this bottleneck through dynamic atom transport. Because tweezers are generated optically, atoms can be translated across the register during circuit execution:

  1. Target atoms sit in static traps produced by a spatial light modulator. Mobile tweezers generated by a two-dimensional acousto-optic deflector overlap with target sites.
  2. Mobile trap power is ramped up as static traps extinguish, transferring atoms into mobile potentials without heating.
  3. Frequency-swept radiofrequency drives translate the optical focal spots across tens to hundreds of micrometers, positioning atoms adjacent to target partners.
  4. Laser pulses apply entangling gates within the blockade radius RbR_b, after which atoms can be relocated or returned to storage.

To prevent motional heating, transport trajectories must maintain adiabaticity in the trap. Trajectories follow quintic polynomials:

x(t)=d[10(tT)3−15(tT)4+6(tT)5](0≤t≤T)x(t) = d \left[ 10\left(\frac{t}{T}\right)^3 - 15\left(\frac{t}{T}\right)^4 + 6\left(\frac{t}{T}\right)^5 \right] \quad (0 \le t \le T)

with transport distance dd and duration TT. This parameterization ensures velocity and acceleration vanish at the endpoints.

At velocities of v≈0.5 μm/μsv \approx 0.5\,\mu\text{m}/\mu\text{s}, a 100 μm100\,\mu\text{m} translation completes in roughly 200 μs200\,\mu\text{s}, with vibrational heating measured below 0.10.1 quanta per round trip. Differential AC Stark shifts induced by moving traps are calibrated and canceled in real time through phase updates on single-qubit drive lasers. This transport enables arbitrary connectivity across large arrays without physical wires or SWAP overhead.

Logical qubits and the path to fault tolerance

Scaling quantum computation to deep circuits requires quantum error correction, where quantum information is encoded across collective entangled states of multiple physical qubits.

Neutral-atom arrays are well suited to topological and surface error-correcting codes. Dynamic shuttling allows physical qubits to be partitioned across dedicated functional zones:

  • Storage zones hold idle data qubits in dark optical potentials away from scattered light.
  • Entangling zones receive selected atoms for parallel multi-qubit Rydberg gates.
  • Readout zones isolate ancilla qubits for mid-circuit state detection.

This spatial separation prevents resonant imaging photons during ancilla measurement from decohering neighboring data qubits. Ancillae are shuttled to the readout zone, illuminated to detect error syndromes, and recooled without disturbing data registers.

Transport also permits the execution of transversal gates, which act bitwise between corresponding physical qubits of separate code blocks. Because transversal gates do not spread errors within a code block, they are fault-tolerant by design. In neutral-atom systems, entire two-dimensional arrays of physical qubits belonging to one logical block can be shuttled into position and entangled with a second block using a single global Rydberg pulse.

In 2024, Bluvstein and collaborators operated a logical quantum processor based on reconfigurable arrays of rubidium atoms. The processor supported up to 48 logical qubits using distance-two detection codes ([[8,3,2]][[8, 3, 2]] codes) and color codes ([[7,1,3]][[7, 1, 3]] Steane codes). The demonstration showed fault-tolerant state preparation, logical feedforward, and the creation of logical Greenberger-Horne-Zeilinger states. Logical error rates systematically dropped below physical error rates during error detection, showing the viability of the architecture for fault-tolerant quantum computing.

Systematic decoherence and physical limits

While neutral-atom processors have achieved high gate fidelities and reconfigurable connectivity, several physical mechanisms constrain coherence:

  1. Finite Rydberg lifetime: Spontaneous emission and blackbody radiation limit Rydberg state lifetimes. For rubidium at n=70n = 70, the lifetime at 300 K300\,\text{K} is 150 to 300 μs150\text{ to }300\,\mu\text{s}. During a 200 ns200\,\text{ns} gate pulse, decay induces an intrinsic error floor near 10−310^{-3}, capping two-qubit gate fidelities near 99.9%99.9\%. Cryogenic environments at 4 K4\,\text{K} reduce blackbody radiation, extending lifetimes past one millisecond.
  2. Laser phase noise: Coherent Rydberg driving requires optical fields in the ultraviolet or blue spectrum. High-frequency laser phase noise transfers to qubit state dephasing during excitation. Mitigating this error requires locking gate lasers to ultra-low-expansion reference cavities with sub-kilohertz linewidths.
  3. Thermal motion and Doppler dephasing: At 10 μK10\,\mu\text{K}, atoms retain thermal velocities vth=kBT/m≈3 cm/sv_{\text{th}} = \sqrt{k_B T / m} \approx 3\,\text{cm/s}. Motion along the Rydberg beam propagation axis produces Doppler shifts Δω=k⋅v≈2π×50 kHz\Delta\omega = \mathbf{k} \cdot \mathbf{v} \approx 2\pi \times 50\,\text{kHz}, introducing trajectory-dependent phase errors. Raman sideband cooling to the three-dimensional motional ground state suppresses this dephasing.
  4. Background gas collisions: In ultra-high vacuum (P<10−11 mbarP < 10^{-11}\,\text{mbar}), collisions with residual gas molecules eject atoms on timescales of ten to one hundred seconds. Cryogenic pumping further reduces residual gas pressures, extending trapping lifetimes to hours.

To address these limits, architectures are expanding to include alkaline-earth atoms such as ytterbium-171 (171Yb^{171}\text{Yb}) and strontium-88 (88Sr^{88}\text{Sr}), as well as dual-element mixtures (87Rb^{87}\text{Rb} and 133Cs^{133}\text{Cs}). Alkaline-earth atoms have two valence electrons, providing optical clock transitions and nuclear-spin qubits with I=1/2I = 1/2. This structure allows erasure error conversion: Rydberg decay predominantly populates metastable states that can be detected via optical cycling without destroying the computational state. Converting physical decay into localized erasure errors raises fault-tolerance thresholds, supporting the development of scalable quantum processors.

References

  1. D. Jaksch, J. I. Cirac, P. Zoller, S. L. Rolston, R. Côté, and M. D. Lukin, "Fast Quantum Gates for Neutral Atoms," Physical Review Letters 85, 2208 (2000). DOI: 10.1103/PhysRevLett.85.2208
  2. M. D. Lukin, M. Fleischhauer, R. Côté, L. M. Duan, D. Jaksch, J. I. Cirac, and P. Zoller, "Dipole Blockade and Quantum Information Processing in Mesoscopic Atomic Ensembles," Physical Review Letters 87, 037901 (2001). DOI: 10.1103/PhysRevLett.87.037901
  3. H. Levine, A. Keesling, G. Semeghini, A. Omran, H. Bernien, S. Schwartz, V. Vuletić, and M. D. Lukin, "Parallel Implementation of High-Fidelity Multiqubit Gates with Neutral Atoms," Physical Review Letters 123, 170503 (2019). DOI: 10.1103/PhysRevLett.123.170503
  4. D. Bluvstein, A. Omran, H. Levine, A. Keesling, G. Semeghini, S. Ebadi, T. T. Wang, A. A. Michailidis, N. Maskara, W. W. Ho, S. Choi, M. Serbyn, M. Greiner, V. Vuletić, and M. D. Lukin, "A quantum processor based on coherent transport of entangled atom arrays," Nature 604, 451-456 (2022). DOI: 10.1038/s41586-022-04592-6
  5. S. J. Evered, D. Bluvstein, M. Kalinowski, S. Ebadi, T. Manovitz, H. Zhou, S. H. Li, A. A. Geim, T. T. Wang, N. Maskara, H. Levine, G. Semeghini, M. Greiner, V. Vuletić, and M. D. Lukin, "High-fidelity parallel entangling gates on a neutral-atom quantum computer," Nature 622, 268-272 (2023). DOI: 10.1038/s41586-023-06481-y
  6. D. Bluvstein, S. J. Evered, A. A. Geim, S. H. Li, H. Zhou, T. Manovitz, S. Ebadi, M. Cain, M. Kalinowski, D. Hangleiter, J. P. Bonilla Ataides, N. Maskara, I. Cong, X. Gao, P. Sales Rodriguez, T. Karolyshyn, G. Semeghini, M. J. Gullans, M. Greiner, V. Vuletić, and M. D. Lukin, "A logical quantum processor based on reconfigurable atom arrays," Nature 626, 58-65 (2024). DOI: 10.1038/s41586-023-06927-3
  7. S. Ma, G. X. Meng, Y. Zhou, P. Scholl, E. F. K. W. Wong, N. Bhatt, J. Choi, and J. D. Thompson, "High-fidelity gates and mid-circuit erasure conversion in an atomic qubit," Nature 622, 279-284 (2023). DOI: 10.1038/s41586-023-06516-4
  8. Cover illustration credit: National Institute of Standards and Technology (NIST), Optical Lattice Image Gallery (2017).