⚛️ quantum physics

Error Correction in a Distributed Quantum Computer

This paper presents the first experimental demonstration of distributed quantum error detection and correction by generating entanglement between two separate trapped-ion processors to perform remote syndrome measurements, successfully detecting phase-flip errors and actively correcting arbitrary single-qubit Pauli errors on a distributed logical qubit.

Original authors: E. M. Ainley, A. Agrawal, T. Araki, A. R. Martínez, D. Main, E. Malinowski, J. A. Blackmore, S. Chen, P. Drmota, M. Mallweger, D. P. Nadlinger, R. Srinivas, S. C. Benjamin, G. Araneda, D. M. Lucas

Published 2026-09-14

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: E. M. Ainley, A. Agrawal, T. Araki, A. R. Martínez, D. Main, E. Malinowski, J. A. Blackmore, S. Chen, P. Drmota, M. Mallweger, D. P. Nadlinger, R. Srinivas, S. C. Benjamin, G. Araneda, D. M. Lucas

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

To build a computer that can solve problems beyond the reach of today's machines, scientists are turning to a strange and powerful form of physics called quantum mechanics. These future machines, known as quantum computers, rely on tiny units of information called qubits. Unlike the bits in a standard computer that are either zero or one, qubits can exist in a delicate state of being both at once. This allows them to process vast amounts of information simultaneously. However, this delicate nature is also their greatest weakness. The slightest disturbance from heat, vibration, or stray electromagnetic fields can cause a qubit to lose its information, a problem known as an error. To build a useful machine, researchers must find a way to protect these fragile states. The solution lies in spreading the information out across many physical qubits, creating a single, robust logical unit. If one part of the group makes a mistake, the others can detect it and fix it without destroying the information. This process is called error correction, and it is the essential key to unlocking the full potential of quantum computing.

For years, researchers have been able to perform this error correction within a single, isolated device. But as the need for more powerful computers grows, a single machine will not be enough. The next step is to link many smaller devices together to form a massive, distributed network. This approach, known as modular architecture, allows scientists to scale up by adding more processors rather than trying to fit everything into one giant chip. The challenge is that these separate processors must talk to each other to perform the error correction. They need to check the status of qubits in one machine against qubits in another, miles away, without physically moving the particles. Until now, this specific task—measuring the relationship between distant qubits to catch errors—had never been successfully demonstrated in a real experiment.

A team of researchers at the University of Oxford has now achieved this milestone. They built a small-scale quantum network consisting of two separate processors, placed about two meters apart in their laboratory. Each processor contained trapped ions, which are individual atoms suspended in a vacuum by electric fields. These ions served as the qubits. To connect the two machines, the scientists used light. They coaxed the ions to emit photons, or particles of light, and then guided these photons to a central point where they interfered with one another. This interaction created a special link called entanglement between the two distant processors. In this state, the two processors became a single, shared system, even though they were physically separated.

With this link established, the team tested whether they could use it to perform error detection. They encoded a single piece of information across the two processors, effectively creating a logical qubit that lived in both places at once. They then introduced a specific type of error, a "phase flip," which is a common way for quantum information to go wrong. By measuring the relationship between the distant qubits using their entangled link, the researchers could tell if an error had occurred. When the measurement indicated a mistake, they immediately stopped the process, reset the system, and tried again. By repeating this cycle, they were able to filter out the errors and keep the correct information intact. They found that this method successfully suppressed the errors, proving that they could protect a logical qubit spread across two different machines.

The experiment went a step further by not just detecting errors, but actively fixing them. The researchers prepared a specific, highly entangled state between the two processors, known as a Bell state. They then introduced random noise to scramble the information. Using the same remote link, they measured the system to see exactly what kind of mistake had happened. Based on the result of that measurement, they sent a classical signal back to the processors to apply a precise correction. This happened in real time. The system detected the error, communicated the finding, and corrected the state before the information was lost. The result was that the quality of the entangled state remained high, even as the noise around it increased. Without this active correction, the state would have quickly degraded.

This work demonstrates that the fundamental building blocks for a large-scale, fault-tolerant quantum computer can operate across a network. The researchers showed that it is possible to generate the necessary links between processors, measure the status of distant qubits, and use that information to correct mistakes. While the current system is small and the error rates are not yet low enough for a full-scale computer, the experiment proves the concept works. The main obstacles remaining are technical, such as improving the speed of the connections and the quality of the gates used to manipulate the ions, rather than any fundamental flaw in the idea. By combining local quantum logic with remote connections, this research opens a clear path toward building quantum computers that are large enough to solve the world's most complex problems.

Technical Summary: Error Correction in a Distributed Quantum Computer

Problem Statement
Scaling quantum computers to the fault-tolerant regime requires managing large numbers of physical qubits while minimizing the overhead of error-correcting codes. While high-rate quantum low-density parity-check (qLDPC) codes offer superior encoding efficiency compared to surface codes, they often require non-local connectivity that exceeds the capabilities of individual, monolithic processors. Modular architectures, utilizing photonic interconnects to distribute qubits across multiple processors, address this scalability challenge. However, a critical gap exists: the ability to perform stabiliser measurements between remote processors has not been experimentally demonstrated. Without this capability, extracting error syndromes for data qubits distributed across a network remains theoretical, preventing the realization of fault-tolerant distributed quantum computing (DQC).

Methodology
The authors utilize a two-node trapped-ion quantum network, with nodes separated by approximately 2 meters. Each node contains a mixed-species ion trap housing an $^{88}\text{Sr}^+$ ion (serving as a network qubit, $Q_N$ ) and a $^{43}\text{Ca}^+$ ion (providing a long-lived data qubit, $Q_D$ , and an auxiliary qubit, $Q_X$ ).

The core experimental primitive is remote stabiliser measurement, mediated by entanglement between network qubits:

Remote Entanglement: A Bell pair is generated between the $Q_N$ ions of the two modules via photonic interconnects (polarisation-entangled photon interference).
Syndrome Extraction: This shared Bell pair acts as a distributed ancilla. Each module performs a local CNOT gate (mediated by $Q_X$ ) between its network qubit and data qubit. This maps the joint parity of the remote data qubits onto the network qubits.
Measurement and Feedforward: The network qubits are measured mid-circuit. The outcomes are exchanged via a classical link in real time to determine the syndrome. Based on the syndrome, feedforward operations (physical or virtual rotations) are applied to correct errors.

The authors demonstrate two distinct protocols:

Distributed Error Detection: Using a distributed [[2, 1, 1]] repetition code ( $|0_L\rangle = |++\rangle, |1_L\rangle = |--\rangle$ ), they repeatedly measure the $X_1X_2$ stabiliser. If an error is detected (odd parity), the run is aborted and restarted; if no error is detected, the state is preserved for subsequent rounds.
Active Error Correction: They prepare a distributed Bell state $|\Phi^+\rangle$ and subject it to injected depolarising noise. They perform sequential remote measurements of both $X_1X_2$ and $Z_1Z_2$ stabilisers. Upon detecting bit-flip or phase-flip errors, they apply real-time corrective rotations ( $U_X$ or $U_Z$ ) to the data qubits.

Key Contributions

First Experimental Demonstration: This work reports the first experimental realization of distributed quantum error detection and correction across separate quantum processors.
Remote Syndrome Extraction: The paper establishes remote syndrome extraction as a functional primitive, demonstrating that entanglement between network qubits can mediate stabiliser measurements on spatially separated data qubits without disturbing the encoded state.
Active Correction: The authors successfully implement active feedforward correction for arbitrary single-qubit Pauli errors on a distributed entangled state, completing the full cycle of syndrome extraction, classical communication, and conditional correction.

Results

Error Detection: In the repetition code experiment, repeated remote stabiliser measurements successfully filtered out phase-flip errors. Conditioning on even-parity outcomes increased the probability of the state remaining in the code space ( $\langle X_1X_2 \rangle$ ) without measurably increasing the logical error rate. At zero applied error, the system achieved a target outcome probability of $96.4(6)\%$ , corresponding to a total error of $3.6(6)\%$ .
Error Correction: For the distributed Bell state, the application of feedforward corrections based on remote syndromes maintained the fidelity to $|\Phi^+\rangle$ approximately constant across a range of injected depolarising noise strengths ( $\lambda$ ). In contrast, the uncorrected state showed a monotonic decrease in fidelity as noise increased.
Error Budget: The dominant error sources were identified as technical imperfections in local mixed-species entangling gates and the fidelity of the remote entanglement link, rather than fundamental protocol limitations.

Significance and Claims
The authors claim this work provides an experimental foundation for quantum error correction across modular quantum architectures. By combining remote entanglement generation, local quantum logic, mid-circuit measurement, and real-time feedforward, the experiment demonstrates that all essential ingredients for DQC can operate coherently across a network.

The paper positions these results as a "proof of principle" for remote stabiliser measurements. It asserts that while current fidelities are not yet sufficient for typical fault-tolerant implementations, the limitations are technical (specifically gate fidelities and network link quality) rather than fundamental. The work establishes a route toward full fault-tolerant distributed QEC, where arbitrary logical quantum information can be protected across a network, and suggests that the same principles are applicable to other modular platforms (e.g., neutral atoms, superconducting circuits) capable of generating remote entanglement.

More like this