We studied the antiphase self-pulsation in a globally coupled three-mode laser operating in different optical spectrum configurations. We observed locking of modal pulsation frequencies, quasiperiodicity, clustering behaviors, and chaos, resulting from the nonlinear interaction among modes. The robustness of [p:q:r] three-frequency locking states and quasiperiodic oscillations against residual noise has been examined by using joint time-frequency analysis of long-term experimental time series. Two sharply antithetical types of switching behaviors among different dynamic states were observed during temporal evolutions; noise-driven switching and self-induced switching, which manifests itself in chaotic itinerancy. The modal interplay behind observed behaviors was studied by using the statistical dynamic quantity of the information circulation. Well-organized information flows among modes, which correspond to the number of degeneracies of modal pulsation frequencies, were found to be established in accordance with the inherent antiphase dynamics. Observed locking behaviors, quasiperiodic motions, and chaotic itinerancy were reproduced by numerical simulation of the model equations.

We investigated chaotic dynamics in a microchip three-mode solid-state laser subjected to frequency-shifted optical feedback. When the frequency shift was tuned to harmonic frequencies of the relaxation oscillation, a bifurcation from a periodic sustained relaxation oscillation ("soft-mode") state to a chaotic spiking ("hard-mode") state via a chaotic itinerancy was observed as the feedback intensity was increased. Dynamic characterizations of modal interplay and self-induced switching between the soft- and hard-mode chaotic states over times (i.e., chaotic itinerancy) were carried out by the information circulation analysis and joint time-frequency analysis of long-term experimental time series. Drastic changes in information transfer rates among oscillating modes and occasional frequency locking among periodicities of two chaotic states associated with switchings were identified in chaotic itinerancy. Essential dynamical behaviors were reproduced by numerical simulation.

No abstract