Quantum Entanglement Falters in Large Systems

Quantum entanglement, a phenomenon where particles remain interconnected regardless of distance, is a cornerstone of quantum computing. In a study using IBM's 53-qubit Rochester quantum computer, researchers tested how well entanglement holds up in systems of increasing size. This matters because entanglement enables quantum computers to solve problems that are intractable for classical machines, such as simulating complex molecules or optimizing large systems. reveal that while entanglement is robust for small groups of qubits, it becomes unstable in larger setups due to environmental noise, posing s for scaling up quantum technologies.

The key is that entanglement in the IBM quantum computer performs well for systems with up to four qubits but degrades significantly for five or more qubits. Entanglement was assessed using Mermin's inequalities, a mathematical test that distinguishes quantum behavior from classical limits. For two to four qubits, most measurements showed entanglement properties, even when they did not exceed the classical boundary. However, for five to seven qubits, the entanglement was largely lost, with resembling classical, non-entangled systems.

To conduct the experiment, the researchers used GHZ-like states, which are specially prepared quantum states designed to maximize entanglement. They implemented these states on the IBM Rochester quantum computer, which has a hexagonal layout of qubits. For each test, they created circuits involving Hadamard gates to put qubits into superposition and CNOT gates to entangle them. Measurements were taken using Pauli-X and Pauli-Y gates, with each experiment repeated five times using 1024 shots to ensure statistical reliability. The setup allowed them to probe different qubit connectivities, from pairs to chains of seven qubits.

, Detailed in Figures 3 and 4 of the paper, show that normalized Mermin's polynomial values for two to four qubits often clustered along the expected quantum axis, indicating entanglement. For instance, in Figure 3, many data points for two to four qubits exceeded the local realism limit, represented by a red line at unity. Standard deviations were smaller for these cases, suggesting consistent performance. In contrast, for five to seven qubits, the data points scattered widely, with larger standard deviations, and few combinations violated the classical limit. Orthogonal measurements in Figure 4 further confirmed this: for smaller qubit numbers, aligned with quantum predictions, while for larger systems, they became isotropic and independent, like classical outcomes.

This research has real-world for the development of quantum computers. It suggests that current noisy intermediate-scale quantum (NISQ) devices, like the IBM Rochester, are suitable for tasks involving small entangled systems, such as simple simulations or basic algorithms. However, applications requiring large-scale entanglement, such as advanced cryptography or complex optimization, may face limitations until noise is reduced. The study highlights the importance of qubit connectivity and environmental stability, indicating that not all qubit arrangements perform equally well under noise.

Limitations of the study include the influence of NISQ noise, which reduces entanglement strength, as analyzed in Supplement C using an entangled parameter. The researchers note that for larger qubit numbers, only specific connectivities maintained entanglement, and standard deviations increased significantly, reflecting environmental fluctuations. This means that are specific to the IBM Rochester system and its current noise levels, and further work is needed to generalize these to other quantum platforms or improved conditions.