arXiv:2610.04111
Exchange-controlled quantum beats and entanglement in an exciton--bimodal-cavity system
Cesar E. P. Villegas, Victor P. Brasil, Pedro A. S. Contri, Kevin Lizárraga, Jonas Maziero
Abstract
We theoretically demonstrate tunable population transfer and entanglement generation in a two-exciton system coupled to a bimodal optical cavity, including an exciton-exciton exchange interaction. By using dressed-state transformations and Löwdin partitioning, we derive effective Hamiltonians that identify three dynamical time scales: local polaritonic oscillations, long-period dressed-state beats, and direct excitonic exchange. We apply the framework to graphene nanoribbon excitons, using excitonic energies, transition dipoles, and lifetimes obtained from first-principles calculations to parameterize the model. In the small- and moderate-exchange regimes, the dynamics become multiscale, and the beat channel sustains sizable quantum entanglement over long dissipative time windows. In the large-exchange regime, the excitonic Hamiltonian dominates and drives nearly complete population transfer between excitonic states over many coherent cycles. The resulting analytical expressions provide direct design rules connecting the beat period, transfer amplitude, and coherent-cycle number to the exchange interaction, cavity detuning, and light--matter coupling. Our results establish a controllable route for engineering long-lived entanglement and coherent excitonic operations in exciton--bimodal-cavity systems.
AI Summary
The paper examines how exciton–exciton exchange controls population transfer and pairwise entanglement between two excitonic transitions and two cavity modes. It studies a rotating-wave Hamiltonian in the four-dimensional single-excitation sector, using dressed-state transformations and Löwdin partitioning to derive effective two-level descriptions, and compares their predictions with numerical Hamiltonian and Lindblad evolution. Three dynamical scales are identified: local exciton–photon oscillations, a slower channel coupling the second cavity mode to the upper dressed exciton, and direct excitonic exchange. The exchange splitting is $\Omega_X=\sqrt{\Delta_X^2+4J^2}$, while the slow-channel splitting is $\Omega_B=\sqrt{(\omega_2-\varepsilon_+)^2+4g_2^2\cos^2\theta}$. These expressions relate oscillation periods to exchange, detuning and light–matter coupling. For graphene-nanoribbon parameters, the exciton energies are 1.30 and 1.62 eV and cavity couplings are 5.32 and 1.88 GHz. Representative calculations show approximately 80- and 120-ps oscillations at $J=3$ meV, and approximately 8-fs excitonic oscillations at $J=0.21$ eV. Dissipative simulations exhibit population exchange over several hundred picoseconds and slowly decaying concurrence plateaus. Strong-exchange calculations with selected initial states show nearly complete excitonic population oscillations. The material connection uses first-principles exciton energies, dipoles and lifetimes, but treats $J$ phenomenologically rather than calculating it microscopically. The effective local/slow-channel decomposition requires weak residual coupling and spectral separation and breaks down when exchange dominates. The proposed exchange dynamics supplies an ingredient for coherent operations, not a specified quantum gate.
Central claim may not hold: The stated strong-exchange examples permit only 63.3% transfer from |X_2⟩ in Fig. 3(a) and an approximately 0.080–0.900 population range in Fig. 4(c), weaker than the advertised nearly complete transfer. The central unified amplitude formula fails to preserve inner products at the same mixing order as its retained interference terms, invalidating its claimed general-initial-state description. Both the printed cavity quality-factor relation and the listed linewidths are inconsistent with the photon-loss convention and numerical quality factor used in the paper. The coherent-cycle estimate omits a factor of two in the pure-dephasing contribution implied by the explicitly defined Lindblad equation. The claimed connection between slower beats and longer entanglement lifetime is not established by the selected concurrence plateau or by the beat-period engineering maps.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.