Tailoring microcombs with inverse-designed, meta-dispersion microresonators

Nonlinear wave mixing in optical microresonators offers new perspectives to generate compact optical-frequency microcombs, which enable an ever-growing number of applications. Microcombs exhibit a spectral profile that is primarily determined by their microresonator's dispersion. One example is the sech(2) spectrum of dissipative Kerr solitons under anomalous group-velocity dispersion. Here we introduce an inverse-design approach to spectrally shape microcombs, by optimizing an arbitrary meta-dispersion in a resonator. By incorporating the system's governing equation into a genetic algorithm, we are able to efficiently identify a dispersion profile that produces a microcomb closely matching a user-defined target spectrum, such as spectrally flat combs or near-Gaussian pulses. We show a concrete implementation of these intricate optimized dispersion profiles, using selective bidirectional-mode hybridization in photonic-crystal resonators. Moreover, we fabricate and explore several microcomb generators with such flexible 'meta' dispersion control. Their dispersion is not only controlled by the waveguide composing the resonator, but also by a corrugation inside the resonator, which geometrically controls the spectral distribution of the bidirectional coupling in the resonator. This approach provides programmable mode-by-mode frequency splitting and thus greatly increases the design space for controlling the nonlinear dynamics of optical states such as Kerr solitons. This work reports an inverse design approach that can spectrally shape Kerr microcombs by imprinting a nanophotonic dispersion filter to a microresonator to engineer solitonic frequency-comb states in the resonator with an optimization algorithm.

Automated summary

Nonlinear wave mixing in optical microresonators allows for the generation of compact optical-frequency microcombs with tailored spectral profiles. An inverse-design approach is introduced to shape microcombs by optimizing arbitrary meta-dispersions in resonators. By incorporating a genetic algorithm, dispersion profiles can be efficiently identified to closely match a user-defined target spectrum. This work demonstrates controlled dispersion in microcomb generators using bidirectional-mode hybridization in photonic-crystal resonators.