Researchers at the Max-Planck-Institut für Quantenoptik and the Center for Hybrid Quantum Networks (Hy-Q) have demonstrated entanglement between atomic qubits and photons. The team encoded information into photons within a graph state using single-atom emitters, observing an efficiency gain in retrieving quantum correlations for single-qubit measurements. This work addresses a longstanding challenge, as previous experiments attempting this approach failed to show improved information retrieval.
Atomic-Photonic Entanglement with 50% Signal Loss
The experiment, using a graph state encoding scheme, achieved entanglement despite a 48% average loss of photons from generation to detection, a threshold previously considered prohibitive for this type of system. This success hinges on distributing quantum information across multiple physical qubits, anticipating that some will inevitably be lost during transmission or measurement. The team employed two single-atom emitters within an optical cavity to create the entangled photons, a configuration that allowed for precise control and manipulation of the quantum states.
This improvement is not merely theoretical. The researchers measured a logical loss, the ultimate failure rate of retrieving the encoded information, of 0.461, compared to a physical loss of 0.481, indicating a genuine reduction in information decay. The ability to tolerate such high loss rates is critical for building practical quantum communication networks, where photons are often transmitted over long distances through lossy channels like optical fiber or the atmosphere.
The researchers specifically designed their experiment to avoid post-selection, a technique where only successful events are recorded, which can artificially inflate performance metrics. Instead, all generated photon combinations were accepted, providing a more realistic assessment of the system’s resilience. “An important feature of our graph code generation is that, in contrast to previous implementations of quantum error correction of photon loss, it does not rely on postselection,” the researchers state, emphasizing the rigor of their approach.
The graph state used in the experiment is not arbitrary; its structure is specifically chosen to enable error correction through a process of “deforming” logical operators by multiplying them with stabilizers. This allows the system to recover information even when some qubits are lost, by effectively shifting the entanglement to remaining qubits. For instance, if photons 1, 2, and 4 are lost, the encoded information can be recovered by measuring qubits 3 and 5 in specific ways, as defined by the equation Z = .
This resilience is not limited to a single loss pattern. The system is designed to handle multiple combinations of lost photons, offering a degree of robustness not seen in earlier implementations. Despite the success in maintaining entanglement, technical imperfections in the graph state generation introduce additional errors, preventing a complete increase in the probability of obtaining the correct logical measurement outcome.
The researchers analyzed various loss configurations, finding that for both X and Z measurements, there are 2 configurations with 3 lost photons, 8 with 2 lost photons, 5 with 1 lost photon, and 1 with no photon loss. This distribution of loss patterns was factored into their calculations of logical loss, providing a comprehensive assessment of the system’s performance.
The team’s calculations reveal that the logical loss probability is consistent across both X and Z measurements, further validating the effectiveness of their error correction scheme. The physical loss was extracted by calculating the ratio of undetected photons to the number of generation attempts, regardless of the photon’s position within the graph state.
For the X basis, they measured a physical loss of 0.479, and a logical loss of 0.458, while for Z, they measured 0.483 and 0.464 respectively. These results demonstrate a clear, measurable reduction in information loss through the use of the graph code and the distribution of quantum information. Lowering the encoding errors remains a key focus for future work, promising even more robust and efficient quantum information processing.
Hexagonal Graph State Encoding of Logical Qubits
Distributing qubit information across multiple photons allows for continued operation even with substantial signal loss. This approach bypasses the inability to clone qubits by encoding logical information across an entangled set, mitigating the impact of losing individual physical qubits. The experiment, designed to test information retrieval efficiency, achieved a total physical loss in photon generation, manipulation and detection below 50%, entering a regime where logical loss can be less than physical loss.
The experimental setup began with the generation of a three-qubit linear cluster state, followed by the removal of one photonic qubit via an X measurement, and then repeated emissions and rotations to create a seven-qubit state. A final fusion step produced the six-qubit hexagonal graph state, where two atoms redundantly encode a single computational qubit used as the encoding qubit for the graph code. This process highlights the deliberate construction of entanglement to protect information.
The resulting system allows for accessing the measurement outcome of the encoding qubit through the use of stabilizers and logical operators of the graph code, conditioned on successful fusion gates. Analysis of the efficiency budget revealed that the experimental data surpassed the logical loss observed with single-photon encoding, as illustrated in a comparison of theoretical curves and experimental points. The green and blue points representing logical Z and X measurements, respectively, consistently fell below the theoretical loss curve, indicating a demonstrable advantage in information preservation.
The team defined the abscissa for one data point using the physical loss of the most efficient photon, providing a conservative assessment of the system’s performance. All error bars presented represent one-sigma confidence intervals, providing a measure of the uncertainty in the measurements.
Stabilizer Framework Enables Loss Recovery
The stabilizer framework can be efficiently employed to detect and correct qubit errors, including physical loss of the qubit carrier. The researchers explored various loss configurations, calculating the probability of failing to recover logical Pauli information. Their analysis revealed that the intersection point between physical and logical loss occurs at a physical loss value of 0.5 demonstrating the framework’s effectiveness in mitigating signal degradation. The team accounted for these differences in their analysis, ensuring accurate assessment of the overall system performance.
Further refinement of the encoding process, specifically reducing the introduction of errors during the initial state preparation, represents a logical next step toward realizing fully loss-tolerant photonic quantum-information processing. This approach is key for building practical quantum networks capable of transmitting and processing information reliably.
Demonstrated Entanglement Fidelity Exceeds 70.5%
Although technical errors reduce the probability to obtain the correct result, they achieve average fidelities as high as without and with error detection, demonstrating entanglement between the atomic and encoded photonic qubits. Entanglement is arguably the most counterintuitive concept of quantum physics.
It describes the observation that outcomes of measurements performed on a set of particles can be more strongly correlated than allowed by classical physics. Over the past few decades, experiments using pairs of entangled particles have not only explored fundamental quantum phenomena such as nonlocality, but have also established entanglement as a resource for new applications within the second quantum revolution.
Nevertheless, the generation, characterization and control of entanglement in multi-qubit systems remains a central challenge. In this context, graphs have emerged as a powerful language for visualizing a certain class of multipartite entangled states: graph states. This formalism provides a pictorial representation of their entanglement structure. A key feature is their description in terms of a set of multi-qubit operators known as stabilizers.
Graph states are therefore a powerful resource for quantum information processing based on photons, for which loss is the dominant limitation. However, despite the tremendous progress achieved with photonic platforms, an experimental demonstration that quantum information can survive photon loss has so far been missing.
Source: https://arxiv.org/abs/2611.21753




See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.
