Single-Mode Dispersive Wave Dynamics

• Single-mode dispersive waves are optical excitations confined to one mode, with their spectral and temporal dynamics governed by frequency-dependent propagation constants.
• They enable efficient soliton-induced resonant radiation in silicon wire waveguides and microresonators, facilitating applications like supercontinuum generation and optical frequency division.
• Dispersion engineering and precise phase-matching conditions allow controlled tuning of dispersive wave frequencies, which is critical for advanced metrology and integrated photonic systems.

Searching arXiv for recent and foundational papers on single-mode dispersive waves and related dispersive-wave theory. A single-mode dispersive wave is a dispersive optical or wave-dynamical excitation whose evolution is confined to one guided, cavity, or otherwise selected mode while its spectral and temporal behavior is governed by the frequency dependence of the propagation constant. In linear narrowband theory, the term denotes pulse propagation in a single spatial mode of a linear, homogeneous, isotropic dielectric, where $k(\omega)=n(\omega)\omega/c$ determines phase velocity, group velocity, pulse broadening, and chirp. In nonlinear guided-wave and microresonator systems, it denotes resonant radiation emitted by a soliton into a single guided or cavity mode when a higher-order-dispersion or mode-hybridization phase-matching condition is satisfied. The concept therefore spans linear envelope propagation, soliton-induced resonant radiation in single-mode silicon wire waveguides, and the limiting microcomb regime in which dispersive-wave power is concentrated into one cavity mode (Mansuripur, 2020, Leo et al., 2014, Yi et al., 2016, Ji et al., 2024).

1. Linear single-mode dispersive-wave framework

In a linear, homogeneous, isotropic medium with frequency-dependent constitutive parameters $\varepsilon(\omega)$ and $\mu(\omega)$, a monochromatic plane wave $$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$ obeys

$$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$

The phase velocity and group velocity are

$$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$

For a narrowband wave packet centered at $\omega_0$, the Taylor expansion

$$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$

isolates the first two dispersive corrections beyond the carrier. In delayed time $\tau=t-z/v_g$, the complex envelope acquires a quadratic spectral phase $i\tfrac12 k''_0 z(\omega-\omega_0)^2$, which produces pulse broadening and chirp. For an initially Gaussian spectrum $\varepsilon(\omega)$0, the temporal width evolves as

$\varepsilon(\omega)$1

and the chirp parameter is

$\varepsilon(\omega)$2

This framework establishes the basic meaning of a single-mode dispersive wave: the field remains in one mode, but different spectral components within that mode acquire different phase delays, so the envelope broadens and develops a time-dependent instantaneous frequency (Mansuripur, 2020).

The same formalism also clarifies a frequent ambiguity. “Single-mode” here refers to modal confinement, not to monochromaticity. A single-mode dispersive wave can still have finite bandwidth, nontrivial chirp, and strong temporal reshaping. This suggests that the defining feature is the absence of intermodal dynamics, rather than the absence of dispersion.

2. Dispersion engineering in single-mode silicon wire waveguides

A concrete guided-wave realization is provided by silicon-on-insulator wires with silicon thickness $\varepsilon(\omega)$3 nm and widths $\varepsilon(\omega)$4 nm or $\varepsilon(\omega)$5 nm, designed so that only the fundamental quasi-TE mode is guided around $\varepsilon(\omega)$6 nm, with all higher modes cut off for $\varepsilon(\omega)$7 nm. A full-vectorial mode-solver yields the propagation constant $\varepsilon(\omega)$8 of this fundamental mode, and expansion around $\varepsilon(\omega)$9 defines $\mu(\omega)$0. For the $\mu(\omega)$1 nm-wide wire at $\mu(\omega)$2 nm, the reported coefficients are approximately

$\mu(\omega)$3

with dispersion terms retained up to $\mu(\omega)$4 in the propagation model (Leo et al., 2014).

The anomalous sign of $\mu(\omega)$5 places the pump in the anomalous-GVD regime, while the cross-section ensures single-mode operation. In this setting, pulse propagation is modeled by the generalized nonlinear Schrödinger equation

$\mu(\omega)$6

The model includes $\mu(\omega)$7 dB/cm linear loss, free-carrier absorption $\mu(\omega)$8 with $\mu(\omega)$9, free-carrier dispersion through $$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$0, a complex Kerr coefficient $$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$1 scaled by $$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$2, a weak Raman fraction $$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$3, and carrier dynamics

$$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$4

In this geometry, “single-mode dispersive wave” does not mean a cavity mode selected from a comb grid. It means that soliton dynamics and resonant radiation occur within the fundamental quasi-TE guided mode of a dispersion-engineered wire (Leo et al., 2014).

3. Resonant radiation and high-order soliton fission

In the silicon-wire setting, dispersive-wave emission is the resonant radiation emitted by a soliton when its nonlinear propagation constant matches that of a linear wave at another frequency. For a soliton of peak power $$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$5, the nonlinear shift is

$$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$6

The phase-matching condition is

$$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$7

which is equivalently written, with $$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$8, as

$$E(z,t)=\mathrm{Re}\{E_0 e^{i[k(\omega)z-\omega t]}\}$$9

Higher-order dispersion $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$0 provides the phase-mismatch compensation that allows energy transfer from the anomalous-GVD soliton to the normal-GVD dispersive wave, while $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$1 shifts the matching point (Leo et al., 2014).

The triggering mechanism is high-order soliton fission. The soliton order is defined by $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$2, with $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$3 and $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$4. For $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$5 fs pulses at $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$6 W, the reported values are $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$7 for the $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$8 nm wire and $$k^2(\omega)=\omega^2\mu(\omega)\varepsilon(\omega)/c^2, \qquad k(\omega)=n(\omega)\omega/c.$$9 for the $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$0 nm wire. The high-order soliton compresses over a distance $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$1 and then fissions into $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$2 fundamental solitons under the action of higher-order dispersion and nonlinear losses. At the point of strongest compression, the instantaneous bandwidth is largest, and the phase-matching condition launches sharp dispersive-wave sidebands in the normal-dispersion region (Leo et al., 2014).

Experimentally, as the on-chip peak power is raised from $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$3 W to $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$4 W, narrow dispersive-wave peaks appear near $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$5 nm in the $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$6 nm wire and near $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$7 nm in the $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$8 nm wire. Further spectral broadening leads to a $$v_p(\omega)=\omega/k(\omega)=c/n(\omega), \qquad v_g(\omega)=\frac{1}{dk/d\omega} =\frac{c}{n(\omega)+\omega\,dn/d\omega}.$$9 nm-wide supercontinuum for the $\omega_0$0 nm guide. Using the phase-matching equation with soliton peak power $\omega_0$1 W extracted from simulations, the predicted dispersive-wave wavelengths are $\omega_0$2 nm for $\omega_0$3 nm and $\omega_0$4 nm for $\omega_0$5 nm, in close accord with both measurements and numerics (Leo et al., 2014).

A common misconception is that the narrow sideband is an incidental spectral artifact. In the reported silicon-wire regime, it is instead a quantitatively predicted resonant feature tied to higher-order dispersion and soliton compression.

4. Single-mode dispersive waves in dissipative Kerr soliton microresonators

In Kerr microresonators, dissipative Kerr solitons are described by the Lugiato–Lefever equation or by coupled-mode equations for discrete resonator modes. In normalized form one reported version is

$\omega_0$6

while a mean-field formulation writes

$\omega_0$7

The cold-cavity eigenfrequencies are expanded as

$\omega_0$8

and the integrated dispersion is

$\omega_0$9

Phase matching of the soliton tail to a cavity resonance occurs when the zero-phase-mismatch condition

$$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$0

is satisfied (Yi et al., 2016, Ji et al., 2024).

The usual resonator dispersive-wave limit involves an ensemble of optical modes. The single-mode limit arises when exactly one integer solution exists in the comb spectral window, or when one particular mode $$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$1 is brought into near-exact resonance, for example via an avoided-mode crossing with a second transverse family. In that limit, essentially all the soliton’s radiation is funneled into that single mode (Yi et al., 2016).

This usage differs from the single-mode silicon-wire case. In the wire, the entire nonlinear evolution remains in one guided spatial mode; in the microresonator, the comb spans many longitudinal modes, but the dispersive-wave enhancement is concentrated into one selected cavity mode. The distinction is terminological rather than contradictory.

5. Back-action, bistability, and frequency agility

Once the dispersive wave is concentrated into a single cavity mode, it exerts nonlinear back-action on the soliton. For the lower hybrid mode at $$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$2, the intracavity amplitude $$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$3 obeys

$$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$4

and in steady state its power is

$$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$5

The recoil generated by the single-mode dispersive wave is

$$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$6

while the Raman shift is

$$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$7

The total shift $$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$8 enters the repetition rate as

$$k(\omega)\approx k_0+k'_0(\omega-\omega_0)+\tfrac12 k''_0(\omega-\omega_0)^2$$9

These equations form a closed, nonlinear system that can exhibit bistability. In the reported silica whispering-gallery resonator, with $\tau=t-z/v_g$0, $\tau=t-z/v_g$1 GHz, hybridization near $\tau=t-z/v_g$2, crossing detuning $\tau=t-z/v_g$3 MHz, and coupling rate $\tau=t-z/v_g$4 MHz, the optical spectrum exhibits a pronounced single comb line at $\tau=t-z/v_g$5 that can appear or disappear abruptly as $\tau=t-z/v_g$6 is varied. The measured dispersive-wave power and soliton spectral-center shift show clear bistability, and on the upper hysteresis branch $\tau=t-z/v_g$7 exhibits a stationary quiet point with $\tau=t-z/v_g$8 (Yi et al., 2016).

A later Si$\tau=t-z/v_g$9N$i\tfrac12 k''_0 z(\omega-\omega_0)^2$0 implementation uses a “three-coupled-ring” geometry with three concentric waveguide rings whose radii differ by $i\tfrac12 k''_0 z(\omega-\omega_0)^2$1, $i\tfrac12 k''_0 z(\omega-\omega_0)^2$2, and $i\tfrac12 k''_0 z(\omega-\omega_0)^2$3. Differential thermal tuning adjusts $i\tfrac12 k''_0 z(\omega-\omega_0)^2$4, which controls the wavelength $i\tfrac12 k''_0 z(\omega-\omega_0)^2$5 of the middle band’s avoided crossing, and $i\tfrac12 k''_0 z(\omega-\omega_0)^2$6, which controls the local curvature near $i\tfrac12 k''_0 z(\omega-\omega_0)^2$7. By choosing $i\tfrac12 k''_0 z(\omega-\omega_0)^2$8 so that the middle band has a single narrow region of normal-to-anomalous dispersion crossing, only one phase-matching condition is met, yielding a spectrally isolated dispersive wave (Ji et al., 2024).

Three mechanisms enable continuous tuning of the dispersive-wave center frequency: pump-cavity detuning $i\tfrac12 k''_0 z(\omega-\omega_0)^2$9, thermal tuning of the coupled-ring dispersion, and soliton recoil. Reported performance includes fine tuning of $\varepsilon(\omega)$00 nm through detuning, up to $\varepsilon(\omega)$01 nm shift of the short-wavelength dispersive wave around $\varepsilon(\omega)$02 nm and up to $\varepsilon(\omega)$03 nm shift of the long-wavelength dispersive wave around $\varepsilon(\omega)$04 nm by tuning ring A by $\varepsilon(\omega)$05, an additional $\varepsilon(\omega)$06 MHz recoil-induced shift, and an aggregate tuning range of $\varepsilon(\omega)$07 nm $\varepsilon(\omega)$08 THz). The tuning sensitivity via thermal control alone is on the order of $\varepsilon(\omega)$09 nm per $\varepsilon(\omega)$10, and the detuning sensitivity is on the order of $\varepsilon(\omega)$11 MHz pump-cavity detuning per GHz of dispersive-wave frequency shift (Ji et al., 2024).

6. Metrological role and broader extensions

The most developed application in the supplied record is two-point optical frequency division. In that scheme, two comb lines $\varepsilon(\omega)$12 and $\varepsilon(\omega)$13, one near the pump and one at the single-mode dispersive wave, are tightly phase-locked to two modes of an ultrastable Fabry–Pérot cavity separated by $\varepsilon(\omega)$14 THz, and the divided output is a $\varepsilon(\omega)$15 GHz microwave. In the reported microcomb system, the engineered single-mode dispersive wave yields a beat-note signal-to-noise ratio exceeding $\varepsilon(\omega)$16 dB in a $\varepsilon(\omega)$17 kHz resolution bandwidth, which is a $\varepsilon(\omega)$18 dB improvement over the case without dispersive-wave formation. The $\varepsilon(\omega)$19 GHz repetition tone reaches $\varepsilon(\omega)$20 dBc/Hz at $\varepsilon(\omega)$21 Hz offset, $\varepsilon(\omega)$22 dBc/Hz at $\varepsilon(\omega)$23 kHz, and $\varepsilon(\omega)$24 dBc/Hz at $\varepsilon(\omega)$25 kHz, with residual comb-servo noise contributing $\varepsilon(\omega)$26 dBc/Hz at $\varepsilon(\omega)$27 kHz offset. The reference cavity is specified as a vacuum-free, UHV-quality fused-silica rod of length $\varepsilon(\omega)$28 mm, with FSR $\varepsilon(\omega)$29 GHz, linewidth $\varepsilon(\omega)$30 kHz, and $\varepsilon(\omega)$31 (Ji et al., 2024).

This metrological use underscores an important point: a single-mode dispersive wave is not only a nonlinear-spectral byproduct. It can function as a deliberately engineered spectral endpoint whose high power, spectral isolation, and tunability are operationally central.

The term also has a distinct extension outside nonlinear optics. In a beam-plasma weak-turbulence setting, a “single-mode” dispersive-wave model has been formulated in which one continuous branch,

$\varepsilon(\omega)$32

is chosen to reproduce both low-frequency and high-frequency behavior, thereby replacing separate Langmuir and ion-sound mode sums in the kinetic equations. In that usage, the single mode exhibits both low and high frequency regions, which ultimately play the roles of ion-sound and Langmuir modes, respectively, and numerical experiments reproduce rapid growth of the forward branch, 3-wave decay, formation of a superthermal tail, and similar wave-particle energy partition as in the two-mode case (Irumé et al., 2024).

This broader usage suggests that “single-mode dispersive wave” is not tied to one experimental platform. Across optical dielectrics, integrated waveguides, Kerr microresonators, and plasma kinetic models, the unifying structure is a one-mode description in which dispersion, phase matching, and nonlinear or kinetic back-action determine the observable dynamics.