Two-axis twisting by two-photon exchange: rate, depth, and the self-Kerr constraint

Abstract:The coherent exchange of pairs of quanta between two bosonic modes is equivalent, up to a numerical factor, to the generator of two-axis twisting, which attains Heisenberg scaling of the squeezing; we determine how much of that equivalence survives in a realistic nonlinear medium. In any such medium the exchange term is accompanied by a local Kerr nonlinearity, and within a sector of fixed particle number the Hamiltonian depends on a single dimensionless ratio of the effective Kerr strength to the exchange strength. We derive two criteria governing this ratio and show that they are independent. The first is the rate at which squeezing is generated: the exponential rate follows in closed form, vanishes exactly when the two strengths are equal in magnitude, and recovers as the square root of the detuning from that point. The second is the depth to which squeezing can be driven: at fixed particle number the attainable depth is a function of the anisotropy of the mean-field saddle alone, reaches the full two-axis-twisting value only for the isotropic saddle, and is independent of the rate, parameter sets sharing the same anisotropy but differing in rate by a sizeable factor giving depths that agree to every printed digit. The closed-form Gaussian dynamics reveals a pitfall: extracting rates from short-time data overestimates them by more than a factor of two unless the saddle is isotropic. An isotropic cubic nonlinearity imposes the value of the ratio where the rate vanishes, so realizing two-axis twisting by pair exchange requires detuning it; a moderate mismatch already recovers almost all of the optimal depth. We also explain why the constraint has not surfaced before: at two quanta the self-Kerr shifts the two coupled states equally, so the two twisting Hamiltonians are rigorously indistinguishable, the degeneracy being lifted only from four quanta.

arXiv:2610.15774 [quant-ph]

(or arXiv:2610.15774v1 [quant-ph] for this version)

https://doi.org/10.48550/arXiv.2610.15774 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.