L34 · Driven nonlinear ring

Microresonator Kerr Comb

When can a laser-driven microring keep a short pulse circling inside it? A pulse can persist when dispersion and the intensity-dependent phase shift balance, and a continuous pump replaces the energy lost each round trip. The pump frequency must also suit the heated cavity. A surviving prepared pulse does not guarantee that a frequency scan will form one.

Move the pump tuning dial and compare a surviving ring pulse with normal dispersion or no Kerr response. Inspect the captured envelope and comb spectrum.

Slow time / field decay time $\text{—}$

Current cold-cavity detuning $\text{—}$

Effective detuning $\text{—}$

Thermal resonance shift $\text{—}$

Mean normalized intracavity power $\text{—}$

Non-pump mode power fraction $\text{—}$

Periodic field cells $\text{—}$

Captured spectral modes $\text{—}$

Illustrative mode spacing $\text{—}$

Captured peak intensity $\text{—}$

Captured localized peak count $\text{—}$

Relative instantaneous PDE residual $\text{—}$

Outer-third spectral energy fraction $\text{—}$

Captured field description $\text{—}$

Captured slow time $\text{—}$

Explore the mechanism

When can a laser-driven microring keep a short pulse circling inside it?

Normalized pump amplitude

F

2

Initial cold-cavity detuning

\alpha_0

3.5

Signed second-order dispersion

\beta

-0.02

Kerr strength

\gamma

1

Thermal shift per mean power

\chi

0

Cold-detuning scan speed

v

0.03

Scan end detuning

\alpha_{\rm end}

4.4

Start: perturbation / one / two / pattern

s

1

Detail: ring / pump / spectrum

j

0

Every selection integrates sixty field-decay times and captures the result. Replay carries that frozen envelope around the enlarged ring in six seconds; it adds no PDE steps and resolves neither optical phase nor round-trip flight time. Ring-marker size and glow use a fixed intensity reference of ten; pump and weak-drop transport use fixed logarithmic display references. The weak-drop signal represents sampled intracavity intensity; the separate through-port dump has no calculated interference signal.

A prepared pulse tests survival, and a blue-to-red scan tests formation. Neither one guarantees a stationary soliton. Before reading meaning into a step, compare the pulse, the spectrum, the actual power history and the residual.

The model is the normalized high-Q mean-field Lugiato–Lefever equation with a periodic boundary, second-order dispersion and an instantaneous scalar Kerr response. Negative dispersion is anomalous. Slow time is measured in field-decay times and detuning in field half-linewidths; detuning is positive when the pump sits below the cold resonance. A single thermal pole, chosen for teaching, lowers the resonance frequency: the effective detuning is the cold detuning minus the thermal shift, and the shift relaxes over ten field-decay times. The thermal coefficients are teaching values, not fitted to a material. The 100 GHz mode spacing only sets the fast-time scale, and normalized power is not in watts. The single and double presets deliberately insert approximate sech pulses on a continuous-wave background. The pattern and scan presets start instead from tiny, specified, repeatable perturbations, and a scan never inserts a pulse. Spectra are FFTs of the evolved complex field; no sech envelope is imposed on them. A peak count and a small instantaneous PDE residual are evidence of a localized stationary state, but they do not prove that the soliton is stable or that an experiment would be quiet. An ideal weak drop port samples the intracavity field with flat coupling; the plotted pump line is not a through-port interference measurement. Raman response, higher-order dispersion, avoided mode crossings, polarization, quantum noise and real heat diffusion are left out. The ring size and the intensity beads are schematic; the beads show the co-moving envelope, not photon paths. Before a record is shown, its spectral tails on the finite grid are checked.

Start with the essentials

Focus question: When can a pumped ring hold on to a localized pulse?
One-sentence intuition: Dispersion and the Kerr phase shift shape the pulse; continuous pumping replaces the energy that leaks away.

Core mathematical model

Mean-field equation around the ring

\begin{aligned}\partial_\tau\psi={}&-(1+i\alpha_{\rm eff})\psi+F\\&+i\gamma|\psi|^2\psi-i\tfrac{\beta}{2}\partial_\theta^2\psi\end{aligned}

In this sign convention, anomalous dispersion is negative. Because the field wraps around the ring, the equation follows many longitudinal modes and their interactions at once.

Single-pole heating model

\begin{aligned}\dot H&=(\chi\langle|\psi|^2\rangle-H)/10\\\alpha_{\rm eff}&=\alpha_{\rm cold}-H\end{aligned}

Heating lowers the resonance frequency. With the sign convention used here, that shrinks a positive, red detuning.

Steady continuous-wave response

F^2=P\left[1+(\alpha_{\rm eff}-\gamma P)^2\right]

At one pump level the response can take several values, and these give the possible backgrounds. The field around the ring still has to be evolved in full.

Growth of modulation sidebands

\begin{aligned}\Omega_m&=-\alpha_{\rm eff}+2\gamma P+\tfrac12\beta m^2\\\lambda_m&=-1+\sqrt{(\gamma P)^2-\Omega_m^2}\end{aligned}

When the growth rate has a positive real part, sidebands can grow out of the small specified perturbation.

An approximate starting pulse

\begin{aligned}\psi&\simeq\psi_{\rm cw}+A\operatorname{sech}(b\theta)\\A&=\sqrt{2\alpha}\,e^{i\varphi}\\b&=\sqrt{2\alpha/|\beta|}\end{aligned}

This shape only seeds the survival experiment. The solver then either sustains the pulse or destroys it.

The same field in time and frequency

\begin{aligned}a_m&=\frac1{2\pi}\int_{-\pi}^{\pi}\psi e^{-im\theta}\,d\theta\\\sum_m|a_m|^2&=\langle|\psi|^2\rangle\end{aligned}

Parseval's theorem ties the recorded pulse to its discrete spectrum: the comb is the Fourier transform of that same field.

Common difficulties

A comb means a soliton

Typical misconceptionAny set of evenly spaced sidebands proves there is a quiet solitary pulse.

Better mental modelPeriodic patterns also produce combs. Look at the envelope in time, the residual and how the field evolves over a long run.

The prepared pulse formed by itself

Typical misconceptionThe prepared pulse appeared on its own during the detuning scan.

Better mental modelPrepared pulses only test survival. A scan starts from a slightly perturbed CW field and never inserts a pulse.

Run the experiment

01
Test survival
Compare Prepared single pulse and Prepared double pulse. Each selection integrates sixty decay times and captures the actual field.
What to observe: Count the localized peaks, check that the residual stays small, and look at the Fourier spectrum computed from the field.
02
Take away what holds the pulse
Repeat with normal dispersion, or with the Kerr strength set to zero.
What to observe: The same starting pulse can decay into a continuous field.
03
Form a pulse by scanning
Open Advanced controls, choose Blue-to-red scan start and use Scan to target & capture. Then vary the target, scan speed and heating.
What to observe: The computed power history may show several pulses, steps, or the comb disappearing completely.

Microresonator Kerr Comb

A comb emerges from one circulating field