Session S10: Focus Session: Quantum Computing with Fermionic Arrays

Quantum register of fermion pairs

Quantum control of motional states is crucial for quantum science, ranging from quantum metrology to information processing. However, the motional coherence of individual particles can be fragile to maintain, as external degrees of freedom couple strongly to the environment. Systems in nature can host robust motional coherence by harnessing the power of the symmetry of fermion pairs, such as electrons in helium and Cooper pairs. In this work, we demonstrate a novel architecture to encode information in the motional state of a pair of fermions. Coherent control is realized via modulation of interactions between the atoms and the nonlinearity of the trapping potential. The energy difference between the two motional states is set by the atomic recoil energy, is dependent on only the mass and the lattice wavelength, and is insensitive to the noise of the confining potential. In this experiment, we observe quantum coherence beyond ten seconds. The methods presented here will enable coherently programmable quantum simulators of many-fermion systems, precision metrology based on atom pairs and molecules and, by implementing further advances, digital quantum computation using fermion pairs.   

Programmable fermionic quantum simulation using optical tweezer arrays

Optical tweezer arrays have recently found wide-ranging applications in quantum simulation, computation, and metrology due to their flexibility and programmability. We discuss advances using tweezer arrays to study itinerant fermionic systems such as the Hubbard model, realizing a software-programmable, “bottom-up” approach toward quantum simulation. By implementing a 1D Fermi-Hubbard chain at half filling with Li-6 atoms, we create a Mott insulator with strong antiferromagnetic correlations. Two-dimensional arrays of arbitrary geometries are realized with a novel stroboscopic technique that allows for independent tuning of each site. Furthermore, we take advantage of our spin- and density- resolved quantum gas microscope for readout, allowing state-of-the-art entropies to be achieved upon post-selection. Progress toward realizing low-temperature phases of matter in geometries such as triangular ladders will be discussed, opening the door to understanding exotic quantum spin liquid states. These quantum simulations will enhance our fundamental understanding of strongly correlated quantum systems.    

Quantum computation and simulation based on fermion pair registers

Recent experiments demonstrated that an ensemble of qubits can be realized by using vibrational modes of fermion pairs localized on optical lattices [1]. This approach has the advantages of long qubit coherence time, robustness against experimental imperfections such as laser intensity noise, and the scalability to large system sizes. In this work, we develop methods to engineer interactions between neighboring qubits, enabling quantum computation  and simulation in this new architecture. We utilize the combination of Feshbach resonance and the particle tunneling to construct entangling gates, as well as quantum Ising-type Hamiltonians with tunable coupling strength. Furthermore, we propose explicit experimental protocols to characterize and optimize the engineered quantum gates and Hamiltonian in realistic settings, demonstrated by our extensive numerical simulations. This work enables quantum computation and simulation in existing experimental platforms involving fermionic quantum gas microscopes.

Reference:

[1] Hartke, T. et al. Quantum register of fermion pairs. Nature 601, 537–541 (2022).   

Stroboscopic fermion tweezer arrays: heating and Hubbard parameters

The realization of Fermi-Hubbard tweezer arrays with lithium-6 atoms (?arXiv:2110.15398) opens a new stage for studying fermionic matter and fermionic quantum computing, where programmable lattice geometries and Hubbard parameters are paired with single-site imaging. Creating useful 2D tweezer arrays requires accurate tuning of individual lattice sites to compensate for disorder, which may be accomplished by using time-averaged potentials of rapidly strobed tweezers. Here we will present calculations for 2D stroboscopic tweezer arrays with a numerically exact discrete variable representation (DVR) methods, and compare results with experimental measurements. In particular, we will describe how heating from the stroboscopic potential depends on strobe frequency, and quantify how stroboscopic potentials modify Hubbard parameters such as the interaction U and tunneling t in multi-tweezer configurations. Our calculations enable evaluation and optimization of 2D tweezer array experiments.   

Low entropy, programmable optical tweezer arrays for quantum simulation of strongly correlated fermions

Studying Fermi-Hubbard physics with optical tweezer arrays offers the advantage of being able to work with arbitrary lattice geometries and initialize low entropy states. Building on previous work in one dimensional arrays [1], we use crossed acousto-optical modulators to generate Li-6 tweezer arrays in two dimensions. Using a stroboscopic technique, we generate initial configurations of square lattices with up to 50 fermions and non-square lattices such as ring, triangular and Lieb lattices. To post-select on starting with a zero-entropy band insulator initial state, we implement a full spin-charge readout bilayer imaging scheme. By adiabatically ramping on additional tweezers in between the loading sites, we can create tunnel-coupled two-dimensional Fermi-Hubbard simulators outside the realm of modern computational efforts. This opens the door for microscopic studies of low-temperatures fermionic phases in novel lattice geometries that can give rise to spin or kinetic frustration, flat-bands, Dirac points and topological band-structures.

[1] B. Spar et. al. arXiv:2110.15398 (2021)   

Local adressing on the ultranarrow 1S0-3P2 transition in Sr

Alkaline-earth atoms have ultranarrow optical transitions between their ¹S₀ ground state and metastable triplet states. The frequency of the transition to the ³P₀ state is the basis of optical lattice clocks, due to its insensitivity to magnetic fields. In contrast, the ³P₂ state possesses a large magnetic moment, which is advantageous for proposed quantum simulation and computation schemes with neutral atoms.

Here, we report on the first absolute frequency measurement of the ¹S₀-³P₂ magnetic quadrupole transition in strontium by performing Doppler-free spectroscopy in a Stark-shift-cancelled "magic" optical lattice. We obtain the magic condition in a 1064 nm lattice by adjusting the lattice polarization to tune the vector light shift of the ³P₂ state. Besides the absolute transition frequency, we measure the vector polarizability and the isotope shift between ⁸⁷Sr and ⁸⁸Sr. As a proof-of-principle, we demonstrate first results aimed at isolating a single layer of an optical lattice in a magnetic field gradient in preparation for applications in quantum simulation and quantum computation schemes.   

High-Fidelity State Detection of Alkali Atoms in Optical Tweezers

State detection often limits the performance of alkali atom in optical tweezer platforms, which are widely used for studies in areas such as quantum computing, many-body physics, and quantum chemistry. Typical detection schemes use state-dependent atom loss, but this imposes a vacuum-dependent upper bound on readout fidelity, slows repetition rate, and complicates algorithms with mid-circuit measurement. An alternate atom retaining method is to collect state-dependent fluorescence photons. Until now, (without cavity enhancement) this technique has only been demonstrated to ≥ 1.2(2)% readout error (Fuhrmanek et al., 2011).

In work towards Rydberg-dressing quantum computation, we demonstrate hyperfine state readout of individual Cesium atoms in optical tweezers with ≤ 0.7(2)% infidelity and low loss. After further study of relevant atom-photon interaction physics, we obtain this result via adaptive measurement and careful choice of detection parameters to minimize depumping and heating. Our techniques impose no stringent vacuum or optical requirements and can be readily adopted to improve detection on similar systems.