Lattice Surgery on a 17-Qubit Superconducting Processor
This article describes a quantum error correcting breakthrough using a 17-qubit superconducting processor for lattice surgery. Researchers devised a split approach that separated a surface-code logical qubit into two entangled repetition-code qubits. Using a fault-tolerant circuit for bit-flip faults, the team improved logical observable values over non-encoded systems. This experiment validates the functional building pieces needed for advanced gates in planar quantum architectures. The study benchmarks performance utilizing logical process tomography to provide fault-tolerant, scalable quantum computing. Discovery marks a change from state preservation to active manipulation of protected quantum information.
The Logical Entanglement Challenge
Scientists must find ways to protect fragile quantum information to achieve universal quantum computation. Despite past research showing that single logical qubits can keep their state, the researchers found it difficult to execute gates between them. Many hardware designs for qubits have fixed local connections, making conventional “transversal” entangling gates expensive to build.
Lattice surgery is a promising alternative for fault-tolerant gate operations with a planar qubit arrangement. The code lattice is deformed by “merge” and “split” operations to measure observables over many logical qubits, which are the building blocks for CNOT gates and magic state distillation.
The Lattice Split Method
Ilya Besedin and Michael Kerschbaum led the ETH Zurich experiment using 17 flux-tunable transmon qubits. This hardware was designed to build a rotating distance-three surface algorithm that uses local stabilizer measurements to find physical issues. The experiment's goal was an X-type lattice split on a surface-code logical qubit.
The split procedure splits one logical degree of freedom into two independent degrees, according to the sources. Researchers converted a distance-three surface code into two bit-flip repetition codes that operated independently. To predict this action, the researchers used a Pauli-frame update in post-processing to flip logical outcomes based on mid-circuit data qubit readouts.
Overcoming Bit-Flip Errors
Bit-flip fault tolerance distinguishes this work. Z-type stabilizer measurements throughout the circuit, including the split and readout phases, insulated the logical observables from any physical failure, according to the study. The researchers used a minimum-weight perfect matching (MWPM) decoder to find the most likely error paths in noisy data.
This error correction has major effects. From 0.189 to 0.730, the logical ZZ observable improved for the team. A near-ideal 0.998 was achieved when the experiment was repeated in error detection mode, excluding runs when a syndrome measurement was wrong. This performance was far better than a “non-encoded” circuit without redundant qubits for error correction.
Benchmarking, Process Tomography
Researchers moved beyond cardinal states to define the split procedure. State injection was used to create arbitrary logical states on the Bloch sphere with a changeable polar angle. They could see how different initial states affected the entangled logical qubits.
Researchers employed logical quantum process tomography to recreate the split operation Pauli transfer matrix. This research found that one logical qubit had an unintentional coherent phase rotation of approximately 0.11π. An AC-Stark effect from a nearby data qubit's mid-circuit measurement causes this alteration. Despite this minor coherent error, experimental results matched numerical calculations qualitatively.
Toward Scalability
Even though the protocol is mostly fault-tolerant for bit-flip failures, the researchers say it gives a conceptual blueprint for larger systems. According to the sources, future research using larger surface codes will be able to tolerate bit-flip and phase-flip errors simultaneously.
This work's characterisation tools, such as logical process tomography, are essential for assessing logical processes' performance in the noisy intermediate-scale quantum era. By integrating these functional building blocks on superconducting circuits, the ETH Zurich team has paved the way for ubiquitous, error-corrected quantum computers. The study's findings, including 2.2% average two-qubit gate errors and energy relaxation times (T1) from 24 to 78 μs, offer a viable foundation for future lattice surgery research.
