Introduction
In the pursuit of practical quantum advantage, two primary challenges dominate the field: scaling the number of entangled qubits and developing robust methods for information processing. While the circuit model is the most common paradigm, Measurement-Based Quantum Computation (MBQC) offers a compelling alternative. In MBQC, computation proceeds through sequential measurements on a highly entangled resource known as a cluster state.
However, generating large-scale cluster states is technically demanding. Cumulative noise from gates, environmental decoherence, and readout errors typically degrade the "genuine entanglement" required for computation as the system size increases. This research demonstrates a significant breakthrough using the Zuchongzhi 3.1 superconducting processor to generate and verify 95-qubit one-dimensional (1D) and 72-qubit two-dimensional (2D) cluster states, marking a new milestone in quantum scalability.
Hardware and Methodology
The experiments were conducted on the Zuchongzhi 3.1 processor, a 105-qubit superconducting system. Achieving high-fidelity entanglement at this scale required several hardware and algorithmic innovations:
- Material Enhancements: The team utilized tantalum thin films for qubit fabrication, which extended the median energy relaxation time ($T_1$) to approximately 50 µs, a 1.6-fold improvement over previous iterations.
- Readout Optimization: By integrating Traveling-Wave Parametric Amplifiers (TWPA) and optimized Purcell filters, the researchers achieved a median readout fidelity of 99.35%.
- Parallel Gate Control: To entangle nearly 100 qubits efficiently, the team employed dynamically optimized parallel Controlled-Z (CZ) gates. This approach minimized crosstalk and frequency collisions that typically plague large-scale superconducting arrays.
- Error Mitigation: Two sophisticated readout error mitigation schemes were tested: a Tensor Product (TP) method for independent errors and a Continuous-Time Markov Process (CTMP) model to account for correlated noise between neighboring qubits. Interestingly, the researchers found that correlated noise was relatively weak, making the simpler TP method highly effective.
Large-Scale Entanglement Generation
The primary achievement of this work is the certification of genuine multipartite entanglement (GME) in record-breaking system sizes. The team generated:
- A 95-qubit 1D cluster state with a fidelity of $0.5603 \pm 0.0084$.
- A 72-qubit 2D cluster state with a fidelity of $0.5519 \pm 0.0054$.
In both cases, the fidelities exceeded the critical 0.5 threshold required to certify GME. To estimate these fidelities, the researchers performed thousands of "stabilizer" measurements, sampling 3,490 stabilizers for the 95-qubit state alone. This data suggests a "Quantum Moore’s Law," where the number of entangled qubits in superconducting systems is doubling every one to two years.
Topological Verification and Teleportation
Beyond state preparation, the study explored the topological properties of these cluster states. Cluster states are known to possess symmetry-protected topological (SPT) phases. The researchers used quantum teleportation as a direct probe for these phases. By treating the 1D cluster state as a "quantum wire," they demonstrated high-fidelity teleportation across the entire 95-qubit chain.
The experiments confirmed that this teleportation is robust against perturbations that preserve the underlying $Z_2 \times Z_2$ symmetry of the state. Conversely, when symmetry-breaking perturbations were introduced, the teleportation fidelity oscillated and eventually failed. This experimental validation confirms that the generated states possess the specific topological protection necessary for stable MBQC.
Conclusion and Future Outlook
While this work represents a major leap forward, several challenges remain for practical MBQC. Although the fidelities achieved are sufficient to certify entanglement, they must be significantly higher to support complex, multi-step quantum algorithms. Furthermore, true MBQC requires active feedforward—the ability to adjust future measurement bases in real-time based on previous measurement outcomes. Implementing this in superconducting hardware remains a significant engineering hurdle. Future research will likely focus on transitioning from error mitigation to full quantum error correction to enable deeper and more complex quantum simulations.
Summary updated