We present a systematic experimental test of whether curvature- or geometry-mediated mechanisms can induce long-range entanglement in open quantum systems beyond direct Hamiltonian couplings. Using the minimal nontrivial topology—a three-node linear chain with nearest-neighbor interactions only—we find no evidence for robust long-range entanglement under hub-localized noise on real IBM superconducting quantum hardware.
Summary
This work presents a systematic experimental test of whether curvature- or geometry-mediated mechanisms can induce long-range entanglement in open quantum systems beyond direct Hamiltonian couplings.
Using the minimal nontrivial topology—a three-node linear chain (A–B–C) with nearest-neighbor interactions only—we probe multiple physically meaningful noise regimes on real IBM superconducting quantum hardware. Across ideal simulation, noisy simulation, and hardware execution, we find no evidence for robust long-range A–C entanglement under hub-localized noise. A small, bounded A–C signal appears only under dephasing-dominant noise and remains far below nearest-neighbor entanglement.
These results place quantitative constraints on strong “curvature-channel” claims in minimal geometries and clarify that any geometry-mediated entanglement effects, if present, are not operative at the smallest nontrivial scale.
What was tested
- A three-qubit linear chain (A–B–C) with XX-like couplings on A–B and B–C only
- No direct A–C coupling
- Pairwise entanglement measured via full Pauli-basis tomography and negativity
- Multiple hub-noise regimes applied to the intermediate node B:
- Idle gate injection (unitary control)
- Delay-based decoherence (time-under-T₁/T₂)
- Dephasing-dominant Pauli-Z noise
- Experiments were carried out using Qiskit Aer and IBM Quantum hardware via Qiskit Runtime
Key results
- Nearest-neighbor entanglement (A–B and B–C) behaves as expected under all noise regimes
- Physical decoherence introduced via delay produces controlled monotonic degradation without inducing long-range entanglement
- Dephasing-dominant noise produces a small, bounded A–C correlation near the tomography noise floor
- No regime exhibits A–C dominance or a robust long-range entanglement channel
Interpretation
This study falsifies the strong form of curvature-mediated long-range entanglement in minimal three-node chains. The results indicate that geometry- or curvature-based entanglement mechanisms, if they exist, require additional structural ingredients—such as redundant paths, feedback, or network-level topology—to manifest at measurable strength.
Rather than supporting shortcut or topology-bypassing entanglement, the data reinforce that geometric effects in open quantum systems are emergent and collective, not pointwise.
Why this matters
Negative results of this kind are essential for bounding speculative extensions of open-system dynamics and for preventing misinterpretation of noise-assisted effects in near-term quantum hardware. This work provides a clean experimental baseline that clarifies where geometry does not matter, thereby sharpening the conditions under which it might.
Reproducibility
All experiments were executed using publicly available tools and required less than ten minutes of total QPU runtime on IBM Quantum’s standard free-access allocation. Circuit definitions, raw measurement counts, and analysis scripts are publicly available.
Code & Data
Implementation details and numerical results are available upon request from the CURV Institute.