Light is a finicky guest. It flies into a cavity but doesn't stay: part of the pulse inevitably reflects and carries away quantum information. Classical single-pass capture schemes hit this limit no matter how much you improve control. Reflection was considered a loss, but the authors of a new study suggest looking at it differently.
An interpreter gets distracted by every rustle in the hall, and a one-hour meeting turns into two. In quantum optics, the interpreter's role is played by the cavity: a chamber that confines light. A photon enters, but not all its energy manages to transfer into the cavity mode, a specific type of field oscillation. Part of the pulse reflects and escapes. The authors of the paper, posted on arXiv, propose not to fight reflection but to use it as a resource.
The method is called 'throw-and-catch.' After the first pass, the reflected pulse is delayed by Δt and directed back into the cavity through a second port. The second pass compensates for the losses of the first. For square and exponentially decaying pulses, the optimal coupling control is derived analytically, achieving perfect capture.
Previously, it was believed that perfect capture of an arbitrary pulse requires infinite coupling strength. The new scheme shows that finite coupling is sufficient if you use two passes. With internal losses $κ_i$, the two-pass method eliminates the factor $e^{-O(γ/κ_{max})}$ from the efficiency, leaving only the dependence on the ratio of losses to the pulse decay rate. Numerical examples show that for $κ_{max}$ = 3γ and $κ_i$ = 0.001γ, the two-pass scheme achieves efficiency close to the fundamental limit, whereas the single-pass scheme loses a noticeable fraction.
This is an echo caught twice. The first shout bounces off the wall, but the second throw returns it to the cavity, where it finally settles. For quantum networks, this approach means more reliable transfer of states between nodes without loss of fidelity. Already today, the method can improve quantum repeaters, memories, and frequency converters in optical and microwave systems. In the long run, two-pass capture could become a standard component of the quantum internet.
The threshold $κ_{max}$ ≥ 2 ln 2 for perfect capture of a square pulse means that the coupling strength must be about 1.386 times the square of the pulse amplitude. This is the minimum requirement; below it, even two passes won't save you.