arXiv:2610.08445
Quantifying the fidelity of a quantum memory
Victor Barizien, Sophie Egelhaaf, Angelo Gelmini Rodriguez, Louis Nicolas, Théo Sanchez Mejia, Mikael Afzelius, Pavel Sekatski, Nicolas Brunner
Abstract
Quantum memories are key ingredients for future quantum networks. Developing methods for their certification is therefore an important, yet challenging, problem. In this work, we present a general framework for certifying quantum memories and quantifying their performance in an operational and efficient manner, taking advantage of developments in the theory of quantum channels. We start by giving a pedagogical introduction to these concepts and tools and illustrate their natural application to quantum memories. Specifically, we quantify the fidelity of a quantum memory via the notion of average channel fidelity, which can be estimated precisely in an experiment via a finite set of probe states forming a so-called $2$-design. We then present a number of witnessing techniques for certifying this quantity, and show that it gives access to other important properties of the channel, such as its dimensionality and its capacity to transmit quantum information. We consider mostly a simple prepare-and-measure scenario, but discuss also the case of an entanglement-based setup, as well as a semi-device-independent approach. Finally, we apply these methods to a solid-state quantum memory experiment, certifying an average channel fidelity of 98.6%. Our certification methods can be readily applied to other quantum devices, such as fibers, transducers and gates.
AI Summary
The paper addresses how to certify quantum-memory performance using operationally comparable quantities and experimentally accessible measurements. It uses average channel fidelity $F_{\rm avg}$ to the identity as the figure of merit, relating it to Choi fidelity through $(d+1)F_{\rm avg}=dF_{\rm CJ}+1$. A pedagogical treatment explains how probe states forming a projective $2$-design determine $F_{\rm avg}$ without process tomography. For qubits, six Pauli eigenstates suffice; an entanglement-based implementation uses three Pauli correlators. Reduced-data witnesses include the four-state bound $F_{\rm avg}\geq(4T_4-1)/3$ and a two-state bound computed by semidefinite programming. Higher-dimensional constructions include product probes, two mutually unbiased bases, and a minimal set of $d$ suitably connected, linearly independent states whose perfect preservation certifies the identity channel. General measurement statistics are handled through optimization over compatible Choi states. Fidelity bounds also certify channel Schmidt number and lower-bound quantum capacity via coherent information. A semi-device-independent qubit protocol trusts preparations but not output measurements, certifying fidelity after an extraction channel through a steering-based self-testing argument. Application to a solid-state time-bin memory gives conditional average fidelities of $98.6\%$ and $98.5\%$ at storage times of $10$ and $125\,\mu\mathrm{s}$, with efficiencies approximately $16.5\%$ and $3.65\%$. Reduced-state witnesses generally underestimate fidelity, especially the minimal construction away from perfect performance. Experimental analysis assumes input-state-independent efficiency; semi-device-independent certification additionally uses fair sampling. The probabilistic-memory treatment separates efficiency from postselected fidelity, and non-entanglement-breaking certification alone does not guarantee positive quantum capacity.
Error found: The abstract's 98.6% experimental fidelity is conditional on successful retrieval, whereas the unconditional average input-output overlap is approximately 16.3% at the reported 16.5% efficiency. Table 1's positive quantum-capacity lower bounds are computed from post-selected fidelities and do not apply per use of the actual lossy memory, whose unassisted quantum capacity is zero under the stated input-independent-success assumption. The claimed equality between global Choi fidelity and the six-state product-probe statistic is false for multiple qubits and must instead be a lower-bound inequality.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 12+8 pages ; 6+2 figures