Self-cooling, blue-detuned dissipative Kerr microresonator soliton comb

TL;DR

This work tackles TRN-limited phase noise in microresonator soliton combs by enabling self-cooling through blue-detuned DKS generation via pump-assisted AMX in a coupled-ring resonator. By locally shifting dispersion with AMX, the pump detuning can sit on the blue side while maintaining a positive comb detuning, producing a self-cooled, high-efficiency blue-detuned DKS. The authors report up to 14.5 dB reduction in $f_{\rm rep}$ phase noise and a pump-to-comb conversion efficiency near 37%, with robust long-term stability. The approach offers a simplified, thermally robust, power-efficient path for chip-scale microcombs applicable to mm/THz systems and LiDAR, without requiring additional cooling lasers.

Abstract

Dissipative Kerr solitons (DKSs) generated in high-Q microresonators driven by continuous-wave (CW) lasers provide chip-scale optical frequency combs composed of mutually coherent CW lines. However, their small mode volume makes them highly susceptible to thermal fluctuations, and the resulting thermo-refractive noise (TRN) perturbs the repetition rate $f_{\rm rep}$. Here, we experimentally demonstrate a blue-detuned DKS in a coupled-ring microresonator. By employing avoided-mode-crossing (AMX)-induced dispersion engineering at the pump mode, DKSs are generated even when the pump laser is tuned to the higher-frequency (blue) side of the resonance. In this regime, the pump laser not only seeds DKS formation but also serves as a cooling laser for the thermally sensitive pumped mode. We observe a self-cooling effect that reduces the phase noise of $f_{\rm rep}$ by up to 14.5 dB, while achieving a pump-to-comb conversion efficiency as high as 37 %. These results establish blue-detuned DKSs as a thermally robust and power-efficient solution for integrated microcomb systems, eliminating the need for auxiliary lasers.

Figures (6)

• Figure 1: (a) Concept of the pump-AMX configuration. Red and blue resonances show the resonances of the main and auxiliary rings, respectively. The purple line shows the frequency of the pump CW laser. At the pump mode, AMX is introduced, and the resonances are repulsed by the amount of $\varepsilon$, modifying the detuning relation as $\alpha_{\rm comb} = \alpha_{\rm pump} + \varepsilon$, which allows $\alpha_{\rm comb} > \sqrt{3}$ while $\alpha_{\rm pump} < 0$, enabling a blue-detuned DKS. For the blue-detuned DKS, the self-cooling effect is expected as shown in the bottom figure. (b) Schematic of the coupled-ring microresonator. The main ring generates the DKS, and the auxiliary ring controls AMX via a microheater. (c) Measured resonance-frequency shifts near $193.97~\mathrm{THz}$ versus heater power. Colored curves show the coupled resonances, dotted lines indicate the uncoupled resonances calculated using Eq. (\ref{['eq:AMX']}).
• Figure 2: Schematic of the experimental setup. Two independent CW lasers were used for the pump and probe. Circ.: optical circulator, PD: photodetector, OSC: oscilloscope, NF: optical notch filter, ESA: electrical spectrum analyzer, OSA: optical spectrum analyzer.
• Figure 3: Dependence of blue-detuned DKS generation on $\varepsilon$. (a) Measured $\varepsilon$ as a function of heater power. Black circles labeled ii to vii mark the $\varepsilon$ values where comb-power evolution, RF and optical spectra are analyzed. The green highlighted region indicates the range where blue-detuned DKSs without breathing oscillations are accessed.(b) Comb power excluding the pump versus detuning. Dotted lines indicate where the scan direction is reversed. The pump frequency is scanned from blue to red, then from red to blue. State i corresponds to $\varepsilon = 0$. (c) RF noise spectra and (d) optical spectra at the detunings marked by green lines in (b) for states i to vi.
• Figure 4: (a) Comb power variation with $\varepsilon \approx 9.5$. At the dashed line, the direction of the frequency scan is reversed from blue-to-red to red-to-blue. The regions highlighted in red, blue, and green indicate the non-periodic breathing state, periodic breathing state, and blue-detuned DKS, respectively. Lines i through iv mark the detunings at which the RF spectra and transmission are measured, as shown in (b) and (c). (b) RF noise spectra of the combs measured at points i through iv in (a). (c) Transmission spectra around the pumped resonance, measured at points i through iv in (a). The green squares highlight the beat signals between the pump CW laser and the probe laser. (d) Optical spectra of the blue-detuned DKS (red) and the non-periodic breathing state (blue).
• Figure 5: (a) Single-sideband phase-noise power spectral density (PSD) of $f_{\rm rep}$ for blue-detuned DKSs (red and blue). Labels iii and iv correspond to those in Fig. 4. The phase noise of $f_{\rm rep}$ for a red-detuned DKS is also shown (gray). (b) Phase noise of $f_{\rm rep}$ for blue-detuned DKSs at different $P_{\rm pump}$. The phase-noise floor set by the intensity noise of the pump CW laser is also shown (black). (c) Transmission spectra around the pumped resonance in the self-cooling state for each $P_{\rm pump}$. The sharp spectral lines are beat signals between the probe CW laser and the pump CW laser. (d) Pump-to-comb power-conversion efficiency in the self-cooling state for each $P_{\rm pump}$. (e) Effective cooling power $P_{\rm cooling}$ (blue) as a function of pump power, calculated from the data in (c) and (d). The red line represents $P_{\rm pump}$ with unit slope for comparison.