Towards a Scalable and More Efficient Ytterbium Atom Array Quantum Device

Arrays of neutral atoms trapped in optical tweezers have emerged as a promising platform for quantum simulation, computing, and metrology. Recent advances in these systems have enabled shorter duty cycles and increasingly complex yet error-correctable quantum operations with high fidelities, thanks to improvements in system-specific controllability, such as real-time qubit transport and erasure techniques. In this work, we present progress toward a scalable and more efficient Yb atom array quantum processor, realized through novel AMO-physics techniques and integrated with state-of-the-art photonic devices.

To enable rapid atom loading in tweezers and improve experimental repetition rates, we implement a dual-wavelength magneto-optical trap in a core-shell configuration. Additionally, we utilize fast light modulators to engineer the geometry and transport of optical tweezers, introducing new modalities for running quantum operations with improved scalability and efficiency. We encode quantum information into nuclear spin states within the metastable clock manifold at a magic trapping wavelength, which facilitates high-fidelity imaging and cooling, rapid (MHz-scale) single-qubit rotations via Raman transitions, motional decoherence-free rearrangement, and erasure detection for high-fidelity quantum operations.

Using over 1,000 qubits based on this long-lived nuclear spin state, we aim to explore and benchmark quantum many-body dynamics in a regime where classical computers struggle. This will be achieved through both digital circuit and analog Hamiltonian dynamics approaches, with optimally co-designed quantum circuits and Floquet engineering involving trappable Rydberg states. These efforts open up new opportunities to address key questions in modern condensed matter physics, particularly those concerning magnetism and topological phases of matter.