Universal Bright-Bright Integrated Soliton Molecule via Parametric Binding

Dissipative Kerr solitons (DKSs) have emerged as the preferred solution for on-chip integrated optical frequency comb (OFC) generation in metrology. A multi-pumped DKS enables either all-optical trapping in the Kerr-induced synchronization regime, or a multi-component OFC with \greg{a locked repetition rate yet with constant frequency offsets between the components} in the multi-color DKS regime. The multi-color DKS regime is of particular interest since nonlinear mixing between the DKS and the secondary pumped component generates idler waves at different frequencies that are useful for spectral extension of the DKS comb. Here, we explore multi-color idler generation at frequencies in which the resonator free spectral range matches that at the DKS. We demonstrate theoretically and experimentally that without phase matching, the idler forms a bright pulse fundamentally bound to the bright DKS through parametric interaction, despite occurring in normal dispersion. Our work can enable new applications in metrology and spectroscopy of quantum systems toward visible wavelengths, as the parametric nature of our bright-bright state eliminates dependence on dispersion regime or visible wavelength pumping.

💡 Research Summary

The paper investigates a novel bound state in micro‑ring resonators that host dissipative Kerr solitons (DKSs) when driven by two independent pumps: a primary pump that generates the soliton comb and a secondary pump that is deliberately detuned from the soliton’s phase‑matching condition. Conventional multi‑pump DKS studies have focused on two regimes. In the Kerr‑induced synchronization (KIS) regime the secondary pump locks to the soliton, while in the multi‑color DKS regime the secondary pump shares the soliton’s group velocity but has a different phase velocity, leading to bright‑dark bound states when the secondary pump is phase‑matched to the soliton’s integrated dispersion D_int(μ).

The authors go beyond this picture by showing that even when the secondary pump is far off‑phase‑matching (|ϖ⁻| ≫ D_int(μ) for all mode numbers μ), a parametric four‑wave mixing interaction between the soliton field a₀ and the secondary pump field a⁻ can generate an idler field a⁺ at a frequency offset –ϖ⁻ from the soliton comb. Using a multi‑color Lugiato‑Lefever equation (mLLE) they derive coupled equations for the three “colors” (a₀, a⁻, a⁺). The idler’s evolution contains a term –iγ a₀² a⁻* that is purely parametric: it is proportional to the product of the soliton intensity squared and the complex conjugate of the secondary pump. Because this term is only significant where the soliton intensity is high, the idler field is generated exclusively within the temporal envelope of the soliton. Consequently, the idler inherits the soliton’s temporal shape (a sech‑type bright pulse) regardless of the local group‑velocity dispersion at the idler’s central mode μ⁺.

Numerical simulations performed with the open‑source pyLLE package reproduce the experimental device’s integrated dispersion profile, which features two zero‑crossings and a region of normal dispersion at the target idler wavelength. By sweeping the secondary pump detuning, the authors observe a clear transition: when –ϖ⁻ lies within the range of D_int(μ) (phase‑matched case) the idler spectrum consists of discrete synthetic dispersive‑wave peaks and the temporal profile is essentially continuous. When –ϖ⁻ exceeds D_int(μ) for all μ (phase‑mismatched case) the synthetic wave peaks disappear, the idler spectrum broadens into a smooth sech² envelope, and the temporal profile collapses into a bright pulse that is perfectly overlapped with the original DKS. Importantly, this bright‑bright bound state appears even though the idler’s mode resides in a normal‑dispersion regime, confirming that the parametric driving dominates over dispersion‑induced effects.

Experimentally, the authors implement an octave‑spanning Si₃N₄ microring (FSR ≈ 100 GHz) pumped at ~1550 nm to generate a DKS. A secondary pump, offset by roughly one FSR, is tuned across a range of detunings. Optical spectra recorded on a high‑resolution OSA reveal the predicted transition from discrete dispersive‑wave sidebands to a broadband bright idler centered near the visible band (~780 nm). Time‑domain measurements using a frequency‑resolved optical gating (FROG) setup confirm that the idler pulse is temporally locked to the soliton and retains the same repetition rate.

The key insights can be summarized as follows:

1. Parametric Master‑Slave Binding – The idler is not an independent soliton; it is a parametric replica (slave) of the primary DKS (master). Its existence relies on the nonlinear term a₀² a⁻* rather than on a balance of gain/loss and dispersion that defines a true dissipative soliton.

2. Dispersion‑Independent Bright Pulse – Because the idler is generated only where the soliton field is strong, its temporal profile is bright regardless of the sign of the group‑velocity dispersion at μ⁺. This lifts the usual restriction that bright solitons require anomalous dispersion.

3. Phase‑Mismatched Operation – Counter‑intuitively, deliberately avoiding phase matching between the secondary pump and the soliton enables the bright‑bright bound state. The secondary pump acts merely as a seed for the parametric process; its own power is largely depleted outside the soliton region.

4. Frequency‑Comb Extension to Visible – The mechanism provides a deterministic way to translate an octave‑spanning soliton comb into the visible regime without requiring separate visible‑wavelength pumps or engineered anomalous dispersion at those wavelengths.

5. Metrological Implications – Since the idler shares the soliton’s repetition rate and group velocity, the multi‑color comb remains phase‑coherent across a vastly extended spectral range, which is valuable for dual‑color optical clocks, frequency division, and spectroscopy of quantum systems that have transitions in the visible.

Overall, the work establishes a new class of “bright‑bright” soliton molecules that are bound through parametric four‑wave mixing rather than through mutual XPM or direct soliton‑soliton interaction. This expands the design space for integrated photonic frequency combs, allowing engineers to target arbitrary wavelengths—including those in normal dispersion—while preserving the low‑power, compact, and chip‑scale advantages of dissipative Kerr soliton platforms.