betaFeedback

WorksPhysics and Astronomy

Stable dissipative topological solitons in media with even-order dispersion

Abstract

The formation and dynamics of dissipative solitons are governed by the interplay of dispersion and nonlinearity, alongside the gain-loss balance. While it is known that dispersion orders beyond the conventional second order support stable solitons, the possibility of the creation of dissipative topological solitons in this setting was not explored. Here, we reveal a class of stable dissipative solitons enabled by the pure-high-even-order dispersion (PHEOD), with dispersion orders m=4, 6, 8, and 10. The resulting dissipative topological solitons (DTSs) and moving topological solitons (MTSs) are dark-type structures featuring a zero-amplitude point that separates two CW backgrounds with opposite signs. In particular, the PHEOD solitons feature oscillatory tails, which are not present in the conventional dark states. The linear-stability analysis confirms full robustness of the PHEOD-supported DTSs and MTSs. We also explore collision dynamics of MTSs, whose repulsive interactions are inelastic, leading either to rebound, accompanied by emission of dispersive radiation, or complete fusion. These findings establish PHEOD as a means for the generation of new species of dissipative solitons, and broaden the variety of nonlinear modes available to experimental studies in nonlinear optics.

© 2026 Optica Publishing Group

Introduction

Dissipative solitons are localized structures that emerge in driven nonconservative systems far from equilibrium. Unlike conservative solitons, whose formation relies on the balance between nonlinear effects and dispersion or diffraction, dissipative solitons rely upon the additional balance between gain and loss, which is necessary to compensate energy dissipation and sustain the localized modes under the action of the drive [1, 2]. The inclusion of the gain-loss balance makes dissipative solitons fixed-point (attractor) solutions of the corresponding complex Ginzburg-Landau equation (CGLE), in contrast to the standard nonlinear Schrödinger equation (NLSE) for conservative solitons [3], which exists in continuous families, rather than isolated attractors.

In nonlinear optics, CGLEs model a wide range of systems, including passively mode-locked lasers [4] and optical parametric oscillators [5, 6, 7, 8, 9]. Among these models, two generic types are directly forced CGLEs, alias Lugiato-Lefever equations [10], and parametrically-driven damped NLSEs, in which parametric forcing compensates linear losses [11, 12, 13, 14, 15, 16, 17, 18].

A broader class of localized nonlinear waves emerges when the dispersion is represented by a single pure-high-even-order dispersion (PHEOD) term, rather than the conventional second-order one. In this case, solitons arise from the balance between the nonlinear self-focusing and PHEOD, with pure-quartic solitons (PQSs) corresponding to the lowest-order member of the family. Dispersion engineering has provided experimental realization of PQSs in photonic-crystal waveguides and mode-locked laser systems employing intracavity spectral pulse shaping [19, 20, 21]. More recently, experiments have extended this concept to sixth-, eighth-, and tenth-order dispersion, producing the respective solitons [22]. PHEOD solitons have also been shown to form stable bound states, revealing increasingly complex localized-wave dynamics beyond the PQS regime [23, 24]. An important feature of this hierarchy is its specific energy scaling: for dispersion order mm, the soliton energy EE varies with the pulse duration τ0\tau_0 as E∝τ0−(m−1)E \propto\tau_0^{-(m-1)}. This characteristic enables the PHEOD solitons to break the energy constraint inherent to conventional second-order-dispersion solitons.

The study of dissipative higher-order solitons has extended from PQSs to a broader class of PHEOD setups. PQSs exhibit relatively flat spectra and, as mentioned above, the favorable energy scaling, making them attractive for high-power ultrashort-pulse generation [25, 26, 27]. Dissipative PQSs have subsequently been demonstrated in microring resonators within the Lugiato–Lefever framework [28, 29], and their robustness against the Raman self-frequency shift has been explored [30, 31, 32]. Beyond the quartic dispersion, dissipative PHEOD solitons have been realized through the loss compensation provided by the ac or parametric drive [33, 34]. These advances establish PHEOD as a viable platform for the creation of dissipative solitons.

In this work, we investigate dissipative topological solitons supported by PHEOD. These states feature localized phase defects connecting two asymptotic continuous-wave (CW) backgrounds with a phase difference of π\pi [35]. The term “topological soliton” refers to this kink-like structure, following the terminology adopted in optics [36, 37, 38]. The linear-stability analysis and direct numerical simulations show that the topological PHEOD solitons remain completely stable throughout the parameter regions examined, indicating their strong intrinsic robustness. In addition to representing a new type of stable dissipative solitons, these modes may be potential candidates for optical telecommunications and data-processing schemes [39], as well as high-Q sources of soliton frequency combs [40].

The model and methods

We consider one-dimensional dissipative solitons in parametrically driven dispersive cavities with the Kerr nonlinearity [16, 17]. The evolution of the intracavity field envelope A(ts,T)A(t_s,T) obeys NLSE containing the PHEOD term (we address the cases of m=4,6,8,10m=4, 6, 8, 10) [22, 23]. We establish a necessary theoretical model incorporating the PHEOD, Kerr nonlinearity, parametric driving and loss. The equation is written as

itR∂A∂ts+imβmm!∂mA∂Tm+κ∣A∣2A+ΔA=ρA∗−iΓA,(1)i t_R \frac{\partial A}{\partial t_s} + i^m \frac{\beta_m}{m!}\frac{\partial^m A}{\partial T^m} + \kappa|A|^2 A + \Delta A = \rho A^* - i\Gamma A, \tag*{(1)}

where the fast time TT denotes the retarded time in the reference frame moving at the group velocity, tst_s is the slow-time variable, and tRt_R represents the cavity round-trip time. Furthermore, βm\beta_m is the mmth-order dispersion coefficient, κ\kappa is the nonlinearity coefficient, Δ\Delta denotes the cavity detuning, ρ\rho represents the strength of the parametric drive, and Γ\Gamma is the effective loss coefficient.

To obtain a dimensionless form of the governing equation, we introduce the following scaled variables and dimensionless parameters:

t=∣Δ∣tRts,T0=(∣βm∣m!∣Δ∣)1/m,τ=TT0−Vt,(2)t = \frac{|\Delta|}{t_R}t_s,\qquad T_0 = \left(\frac{|\beta_m|}{m!|\Delta|}\right)^{1/m},\qquad\tau= \frac{T}{T_0} - Vt, \tag*{(2)}

and

ψ=∣κ∣∣Δ∣A,h=ρ∣Δ∣,γ=Γ∣Δ∣.(3)\psi= \sqrt{\frac{|\kappa|}{|\Delta|}}A,\qquad h = \frac{\rho}{|\Delta|},\qquad\gamma= \frac{\Gamma}{|\Delta|}. \tag*{(3)}

With these definitions, the governing equation can be recast into the normalized parametrically driven NLSE

i∂ψ∂t−iV∂ψ∂τ−(−1)m/2∂mψ∂τm−∣ψ∣2ψ+ψ=hψ∗−iγψ,(4)i\frac{\partial\psi}{\partial t}-iV\frac{\partial\psi}{\partial\tau}-(-1)^{m/2}\frac{\partial^{m}\psi}{\partial\tau^{m}}-\lvert\psi\rvert^{2}\psi+\psi=h\psi^{*}-i\gamma\psi, \tag*{(4)}

where the scaled moving coordinate is used to accommodate moving solitons with velocity VV. Parameters γ≥0\gamma\ge0 and h>0h > 0 denote, respectively, the normalized cavity loss and parametric-drive amplitude, with ∗* standing for the complex conjugate. It is well known that parametrically driven systems with the second-order dispersion support bright dissipative solitons, a part of which may be stable [16, 17].

It is worth noting that, under the sign convention adopted here, all the considered cases correspond to negative high-even-order dispersion coefficients, βm<0\beta_m < 0, and the self-defocusing Kerr nonlinearity in Eq. (1). In Eq. (4), the alternating signs arise from the differential representation of the even-order dispersion, and do not indicate a change between anomalous and normal dispersion. Thus, all the cases share the same relative dispersion–nonlinearity sign supporting dark modes.

To obtain stationary profiles in the comoving frame, we seek solutions of Eq. (4) in the form of

ψ(τ)=ψr(τ)+iψi(τ),(5)\psi(\tau)=\psi_r(\tau)+i\psi_i(\tau), \tag*{(5)}

where ψr\psi_r and ψi\psi_i denote the real and imaginary components of the stationary field. Substitution of Eq. (5) into Eq. (4) yields the coupled equations:

Vdψidτ−(−1)m/2dmψrdτm−∣ψr+iψi∣2ψr+ψr−hψr−γψi=0,(6)V\frac{d\psi_i}{d\tau}-(-1)^{m/2}\frac{d^m\psi_r}{d\tau^m}-\lvert\psi_r+i\psi_i\rvert^2\psi_r+\psi_r-h\psi_r-\gamma\psi_i=0, \tag*{(6)}
−Vdψrdτ−(−1)m/2dmψidτm−∣ψr+iψi∣2ψi+ψi+hψi+γψr=0.(7)-V\frac{d\psi_r}{d\tau}-(-1)^{m/2}\frac{d^m\psi_i}{d\tau^m}-\lvert\psi_r+i\psi_i\rvert^2\psi_i+\psi_i+h\psi_i+\gamma\psi_r=0. \tag*{(7)}

The stationary soliton profiles are obtained numerically by solving Eqs. (6) and (7) using a Newton–conjugate-gradient method [41, 42]. This approach allows both quiescent and moving localized states to be computed within the same stationary formulation, with V=0V=0 corresponding to the quiescent limit.

To assess the linear stability of the resulting solutions, we introduce a small perturbation according to

ψpert(τ,t)=[u(τ)+iv(τ)]eδt,(8)\psi_{\mathrm{pert}}(\tau,t)=[u(\tau)+iv(\tau)]e^{\delta t}, \tag*{(8)}

where u(τ)u(\tau) and v(τ)v(\tau) are the real and imaginary components of the perturbation eigenmode, respectively, and δ=δR+iδI\delta=\delta_R+i\delta_I is the associated eigenvalue. The real part δR=Re⁡(δ)\delta_R=\operatorname{Re}(\delta) determines the perturbation growth rate. Substitution of Eq. (8) into Eq. (4) and linearization about the stationary solution yield the eigenvalue problem

\begin{pmatrix} L_{11} & L_{12} \\ L_{21} & L_{22} \end{pmatrix} \begin{pmatrix} u \\ v \end{pmatrix} =\delta\begin{pmatrix} u \\ v \end{pmatrix}, \tag*{(9)}\end{pmatrix}

where

L11=2ψrψi+Vddτ−γ,(10)L_{11}=2\psi_r\psi_i+V\frac{d}{d\tau}-\gamma, \tag*{(10)}
L12=+(−1)m/2dmdτm+(ψr2+3ψi2)−1−h,(11)L_{12}=+(-1)^{m/2}\frac{d^m}{d\tau^m}+(\psi_r^2+3\psi_i^2)-1-h, \tag*{(11)}
L21=−(−1)m/2dmdτm−(3ψr2+ψi2)+1−h,(12)L_{21}=-(-1)^{m/2}\frac{d^m}{d\tau^m}-(3\psi_r^2+\psi_i^2)+1-h, \tag*{(12)}
L22=−2ψrψi+Vddτ−γ.(13)L_{22}=-2\psi_r\psi_i+V\frac{d}{d\tau}-\gamma. \tag*{(13)}

A localized solution is unstable if the linearized operator possesses at least one eigenvalue with δR>0\delta_R > 0; otherwise, it is stable within the linear approximation. We compute the eigenvalue spectrum using a Fourier-collocation discretization [42], which transforms Eq. (9) into a finite-dimensional matrix eigenvalue problem for the Fourier coefficients of the perturbation modes uu and vv.

Results and discussion

To characterize the structure of the soliton solutions, we first analyze a pair of the underlying CW states, which Eq. (4) produces as

ψ±=A±eiΘ±,(14)\psi_{\pm}=A_{\pm}e^{i\Theta_{\pm}}, \tag*{(14)}

with

A±2=1∓h2−γ2,Θ±=12arccos⁡(±1−γ2h2),(15)A_{\pm}^{2}=1\mp\sqrt{h^{2}-\gamma^{2}},\quad\Theta_{\pm}=\frac{1}{2}\arccos\left(\pm\sqrt{1-\frac{\gamma^{2}}{h^{2}}}\right), \tag*{(15)}

under the condition of h2≥γ2h^{2}\geq\gamma^{2}. To test the CW states against plane-wave perturbations, we introduce perturbation amplitudes (u,v)(u,v) with real wavenumber kk and complex growth rate δ\delta as

ψCW(τ,t)=ψ±+(ueδt+ikτ+iveδt−ikτ).(16)\psi_{\mathrm{CW}}(\tau,t)=\psi_{\pm}+\left(ue^{\delta t+ik\tau}+ive^{\delta t-ik\tau}\right). \tag*{(16)}

The linearization about ψ±\psi_{\pm} with respect to (u,v)(u,v) yields the dispersion relation

δ=−γ+iVk±iZ,(17)\delta=-\gamma+iVk\pm i\sqrt{Z}, \tag*{(17)}
Z≡(1−2A±2−km)2+A±2(A±2−2)−h2,(18)Z\equiv\left(1-2A_{\pm}^{2}-k^{m}\right)^{2}+A_{\pm}^{2}\left(A_{\pm}^{2}-2\right)-h^{2}, \tag*{(18)}

the CW state being stable provided that condition Re⁡(δ)≤0\operatorname{Re}(\delta)\leq0 holds for all real kk. Because Eq. (15) yields A+2<1A_{+}^{2}<1, the CW state ψ+\psi_{+} is always unstable, while ψ−\psi_{-}, which has A−2>1A_{-}^{2}>1, is stable.

A related analysis predicts the shape of tails of stationary solutions existing on top of the CW states, such as the dissipative topological solitons (DTSs) considered below. In this case, the ansatz is taken as ψ=ψ±+ψ1(τ)\psi=\psi_{\pm}+\psi_{1}(\tau), where ψ1(τ)\psi_{1}(\tau) is the small tail component (cf. the stationary version of Eq. (16), with δ=0\delta=0). Splitting ψ−\psi_{-} and ψ1(τ)\psi_{1}(\tau) into real and imaginary parts, substituting this in Eqs. (6) and (7), linearizing the equations with respect to ψ1(τ)\psi_{1}(\tau), and finally looking for the solutions as (ψ1)r,(ψ1)i∼exp⁡(λτ)(\psi_{1})_{r},(\psi_{1})_{i}\sim\exp(\lambda\tau) leads to the equation that determines the tail’s damping rate λ\lambda:

λ2m+2(−1)m/2(2A±2−1)λm+V2λ2−2γVλ+3A±4−4A±2−2hA±2cos⁡(2Θ±)+1+γ2−h2=0,(19)\lambda^{2m}+2(-1)^{m/2}\left(2A_{\pm}^{2}-1\right)\lambda^{m}+V^{2}\lambda^{2}-2\gamma V\lambda+3A_{\pm}^{4}-4A_{\pm}^{2}-2hA_{\pm}^{2}\cos(2\Theta_{\pm})+1+\gamma^{2}-h^{2}=0, \tag*{(19)}

where A±A_{\pm} and Θ±\Theta_{\pm} are defined as per Eq. (15). Although functions (ψ1)r(\psi_{1})_{r} and (ψ1)i(\psi_{1})_{i} are defined as real variables, complex solutions of Eq. (19) for λ\lambda make sense, predicting the possibility of oscillatory decaying tails (cf. [29]), as shown below in Figure 1.

Energy of stationary DTS solutions and characteristic temporal profiles

Figure 1. The energy (22) of stationary DTS solutions, for varying values of the loss coefficient γ\gamma (attached as labels to the plotted branches) and PHEOD order mm, versus the parametric-drive’s strength hh. Panels (a2_2)–(e2_2) exhibit characteristic profiles of ∣ψ(τ)∣|\psi(\tau)|, produced, at fixed h=2h=2, for different values of γ\gamma. The insets present the comparison between the oscillatory decaying tails, as predicted by Eq. (19) and produced by the numerical DTS solutions (dotted and solid lines, respectively).

Stable dissipative topological solitons (DTSs)

We begin by considering the quiescent DTSs described by Eqs. (6) and (7) in the stationary limit, V=0V=0. In the conventional quadratic-dispersion case (m=2m=2), this solution recovers the familiar dark-soliton state [14]:

ψ(τ)m=2=iAtanh⁡(A2τ)eiΘ,(20)\psi(\tau)_{m=2}=iA\tanh\left(\frac{A}{\sqrt{2}}\tau\right)e^{i\Theta}, \tag*{(20)}
A2=1+h2−γ2,Θ=12arcsin⁡(γh),(21)A^{2}=1+\sqrt{h^{2}-\gamma^{2}},\quad\Theta=\frac{1}{2}\arcsin\left(\frac{\gamma}{h}\right), \tag*{(21)}

which exist under the above-mentioned condition, h2>γ2h^2 > \gamma^2.

Our findings for DTS families are summarized in Figure 1(a$_Figure 1$)–Figure 1(e$_Figure 1$), by means of dependencies of the complementary power (similar to the total power of bright solitons),

N=∫−τ0+τ0(∣ψCW∣2−∣ψ∣2) dτ,with τ0=32,(22)N = \int_{-\tau_0}^{+\tau_0} \left(|\psi_{\mathrm{CW}}|^2 - |\psi|^2\right)\,\mathrm{d}\tau,\quad\text{with }\tau_0=32, \tag*{(22)}

on the drive’s strength hh, for several fixed values of the loss coefficients values γ\gamma (starting from γ=0\gamma=0), and dispersion orders m=2,4,6,8,10m=2,4,6,8,10. The linear stability analysis confirms that all DTS solutions are stable.

Typical examples of their temporal profiles, found at fixed h=2h=2 and various values of γ\gamma, are displayed in Figure 1(a$_Figure 2$)–Figure 1(e$_Figure 2, clearly exhibiting oscillatory decaying tails, which is consistent with previous results [22]. As mentioned above, the oscillatory decay corresponds to complex roots of Eq. (19), λ≡λr+iλi\lambda\equiv\lambda_r+\mathrm{i}\lambda_i. In the case of V=0V=0, Eq. (19) amounts to the factorized form:

DTS solutions of Eqs. (6) and (7) for $m=4$ and $m=10$

Figure 2. DTS solutions of Eqs. (6) and (7), with V=0V=0, varying with the drive’s strength hh and PHEOD order mm. Panels (a1)–(b1) and (a2)–(b2) trace the real and imaginary field components, ψr\psi_{r} and ψi\psi_{i}, as hh changes at fixed γ=1\gamma=1 for m=4m=4 and m=10m=10, respectively. Panels (a3)–(b3) and (a4)–(b4) display, respectively, the corresponding field’s modulus profiles and phase distributions of the DTSs.

[λm−2(−1)m2+1(1+h2−γ2)][λm−2(−1)m2+1(h2−γ2)]=0.(23)\left[\lambda^m-2(-1)^{\frac{m}{2}+1}\left(1+\sqrt{h^2-\gamma^2}\right)\right]\left[\lambda^m-2(-1)^{\frac{m}{2}+1}\left(\sqrt{h^2-\gamma^2}\right)\right]=0. \tag*{(23)}

The oscillatory decaying tails of the complex solutions of Eqs. (6) and (7) are determined by eigenvalues λ\lambda in the form of

ψ=ψ−+k1eλrτcos⁡(λiτ+ϕ1)+ik2eλrτcos⁡(λiτ+ϕ2),(24)\psi=\psi_-+k_1e^{\lambda_r\tau}\cos(\lambda_i\tau+\phi_1)+\mathrm{i}k_2e^{\lambda_r\tau}\cos(\lambda_i\tau+\phi_2), \tag*{(24)}

where the constants k1k_1, k2k_2, ϕ1\phi_1, and ϕ2\phi_2 are fitting parameters. The magnified view for 1<τ<151<\tau<15, displayed in insets in Figure 1(a$_Figure 2$)–Figure 1(e$_Figure 2 demonstrates full agreement of the analytical prediction (24) with the numerical solution.

Furthermore, recent experiments have demonstrated the feasibility of engineering pure high-even-order dispersion in fiber lasers using the intracavity programmable spectral shaping [23]. Reported net cavity dispersion coefficients include β4=−2.66 ps4\beta_{4}=-2.66\,\mathrm{ps}^{4}, β6=−2.128 ps6\beta_{6}=-2.128\,\mathrm{ps}^{6}, β8=−2.128 ps8\beta_{8}=-2.128\,\mathrm{ps}^{8}, and β10=−2.128 ps10\beta_{10}=-2.128\,\mathrm{ps}^{10}. Using these dispersion magnitudes as benchmarks and taking a round-trip phase detuning as Δ=0.01 rad\Delta=0.01\,\mathrm{rad}, the normalization T0=[∣βm∣/(m!∣Δ∣)]1/mT_{0}=\left[|\beta_{m}|/(m!|\Delta|)\right]^{1/m} gives characteristic time scales ≈1.82\approx1.82, 0.816, 0.519, and 0.377 ps for m=4m=4, 6, 8, and 10, respectively. For the stationary dark solitons with h=2h=2 and γ=1\gamma=1, the corresponding dimensional widths are therefore ≈3.5\approx3.5, 1.7, 1.2, and 0.87 ps, respectively. These estimates illustrate the physical temporal scales of the predicted dark modes. Note that the reported solitons are domain walls connecting two opposite-phase CW backgrounds in an effectively infinite temporal domain. An isolated wall therefore does not satisfy ordinary periodic boundary conditions for the complex optical field. In a periodic cavity, a reversed-phase domain can instead be bounded by a pair of oppositely oriented walls, allowing the field to recover its initial complex amplitude after one round trip. When sufficiently separated, the individual walls may locally approximate the isolated solutions studied here, although the existence and stability of the complete periodic configuration require special verification. A promising implementation is a phase-stabilized, parametrically driven Kerr fiber cavity incorporating the programmable intracavity dispersion. Reported experiments have demonstrated these driving and dispersion-engineering capabilities, providing a basis for exploring the proposed dark-domain-wall regime.

Structural features of quiescent DTSs exhibit a pronounced dependence on the parametric-drive’s strength hh. For a fixed loss constant γ=1\gamma=1, Figure 2(aFigure 1–Figure 2(bFigure 1 and 2(a2)–2(b2) present the variation of the real and imaginary field components (ψr\psi_{r} and ψi\psi_{i}) as functions of hh. The corresponding field’s modulus profiles are displayed in Figure 2(aFigure 3–Figure 2(bFigure 3, which connect to the CW background AN≡±1+h2−γ2A_{\mathrm{N}}\equiv\pm\sqrt{1+\sqrt{h^{2}-\gamma^{2}}} at τ→±∞\tau\to\pm\infty, provided that the above-mentioned condition, h2>γ2h^{2}>\gamma^{2}, holds. Phase distributions of DTSs are plotted in Figure 2(aFigure 4–Figure 2(bFigure 4, revealing that each DTS is formed by a pair of half-infinite CW states, with the π\pi phase shift between them, making it possible to identify the DTS as a topological soliton.

Evolution tests and linear-stability spectra for quiescent DTSs

Figure 3. Evolution tests and linear-stability spectra for quiescent DTSs. For m=4m=4, h=2h=2, and γ=0.2\gamma=0.2, panel (a) shows the evolution of a DTS subjected to 10% random-amplitude noise; panel (b) tracks ∣ψ(τ=0)∣|\psi(\tau=0)|; panel (c) gives the corresponding stability spectrum; and panel (d) compares the perturbed input with the profile at t=1000t=1000. Panels (e)–(h) present the analogous results for m=8m=8, h=2h=2, and γ=0.6\gamma=0.6.

MTS existence curves and representative modulus profiles

Figure 4. MTS existence curves in the conservative limit, γ=0\gamma= 0. Panels (a1)–(a4), corresponding to m=4,6,8m = 4, 6, 8, and 1010, plot the momentum PP [Eq. (26)] against velocity VV for several values of hh. Dashed vertical lines mark the two critical velocities Vcr1V_{\rm cr1} and Vcr2V_{\rm cr2} on the upper (P>0P > 0) branches. Panels (b1)–(b4) show representative modulus profiles at h=1/15h = 1/15.

Topological solitons are robust states with a localized core, whose temporal profiles are governed by the interplay of the PHEOD and optical nonlinearity. To assess their stability, we performed direct propagation simulations with perturbed initial conditions, ψ(τ,0)=ψ0(τ)[1+ϵr(τ)]\psi(\tau,0)=\psi_{0}(\tau)[1+\epsilon r(\tau)], where ψ0\psi_{0} is the stationary soliton profile, ϵ=0.1\epsilon=0.1, and r(τ)r(\tau) are independent real random numbers uniformly distributed over [0,1][0,1] at numerical-grid points. This perturbation introduces multiplicative amplitude noise with a relative amplitude of 10%, without imposing independent phase noise. Although the random factors are real, the additive field perturbation is, generally, complex because ψ0\psi_0 is complex, its real and imaginary components being multiplied by the same random factor. The propagation simulations, presented in Figure 3, provide direct numerical evidence of the robustness of the DTSs against the imposed amplitude perturbations. Panels 3(a) and 3(e) show that the solitons retain their shapes in the course of long evolution. The central amplitude ∣ψ(τ=0)∣|\psi(\tau=0)|, plotted in panels 3(b) and 3(f), quickly settles after the introduction of the perturbation. The spectra in panels 3(c) and 3(g), obtained from Eq. (9), are consistent with this behavior. The comparison of the initial and final profiles, displayed in panels 3(d) and 3(h), corroborate that perturbation-induced oscillations gradually decay, without destroying the DTS.

Stable moving topological solitons (MTSs)

The parametric-drive and PHEOD terms break the Galilean invariance of the NLSE [43], which makes the construction of moving solitons a nontrivial problem. To produce moving topological solitons solutions of Eq. (4), we here consider the lossless case (γ=0\gamma=0). In the case of the quadratic dispersion, Eqs. (6) and (7) for this conservative system give rise to the well-known family of dark-soliton solutions [14],

ψ(τ)m=2=−iAtanh⁡(2hτ)±1−3h sech⁡(2hτ),(25)\psi(\tau)_{m=2}=-iA\tanh\left(\sqrt{2h}\tau\right)\pm\sqrt{1-3h}\,\operatorname{sech}\left(\sqrt{2h}\tau\right), \tag*{(25)}

whose existence is restricted to 0<h<1/30<h<1/3.

Numerical solutions of Eqs. (6) and (7) yield the PHEOD-supported MTS families summarized in Figure 4. Panels 4(a1)–4(a4) represent these families, plotting the momentum as a function of VV. In the conservative limit considered here, both PP and the norm NN are conserved.

P=i2∫−∞+∞(dψ∗dτψ−dψdτψ∗)dτ,(26)P = \frac{i}{2}\int_{-\infty}^{+\infty}\left(\frac{d\psi^*}{d\tau}\psi-\frac{d\psi}{d\tau}\psi^*\right)d\tau, \tag*{(26)}

The existence diagrams for MTSs exhibit odd symmetry with respect to the velocity VV, revealing that MTSs exist only within a range ∣V∣≤∣Vcr2∣|V| \le|V_{\rm cr2}|, where Vcr2V_{\rm cr2} is the critical velocity. Furthermore, each upper branch (P>0P > 0) features another critical velocity, Vcr1V_{\rm cr1}, which defines the maximum speed of the retrograde motion of the MTSs on that branch. Table 1 reveals the variation of the critical velocities. Specifically, Vcr2V_{\mathrm{cr2}} increases with the parametric-gain strength hh, and decreases as the PHEOD order mm rises. Further, representative field-modulus profiles of MTSs are presented in Figure 4(bFigure 1–Figure 4(bFigure 4. Figure 4(bFigure 1 and Figure 4(bFigure 2} show stationary MTSs at zero velocity (V=0V=0), while Figure 4(bFigure 3 and Figure 4(bFigure 4} represent MTSs propagating at the critical velocities Vcr1V_{\mathrm{cr1}} and Vcr2V_{\mathrm{cr2}}, respectively. The linear-stability analysis further confirms that all these MTS solutions are stable.

Vcr1(P)V_{\rm cr1}(P), Vcr2(P)V_{\rm cr2}(P) \ hh1/201/151/10
mm
4-0.420(2.085), 1.081(4.778)-0.369(2.023), 1.167(4.924)-0.287(1.901), 1.312(5.244)
6-0.436(2.045), 0.862(4.930)-0.388(1.980), 0.941(5.068)-0.310(1.877), 1.077(5.415)
8-0.451(2.034), 0.772(5.292)-0.404(2.009), 0.842(5.260)-0.327(1.899), 0.965(5.427)
10-0.464(1.945), 0.721(5.190)-0.417(1.932), 0.787(5.243)-0.339(1.891), 0.902(5.292)

Table 1. The critical velocities Vcr1V_{\rm cr1} and Vcr2V_{\rm cr2} of the upper MTS branches (P>0P > 0) in Figs. (4)(a1-a4), for different values of the parametric-gain strength hh and PHEOD orders mm.

Figure 5 summarizes how the MTS structure changes with momentum PP and PHEOD order mm. At a fixed parametric-gain strength hh, the evolution of the real and imaginary components (ψr\psi_r and ψi\psi_i) with PP is shown in Figure 5(aFigure 1–Figure 5(bFigure 1 and Figure 5(aFigure 2–Figure 5(bFigure 2. The associated field-modulus profiles are displayed in Figure 5(aFigure 3–Figure 5(bFigure 3, which asymptotically merge with the CW background (±1+h\pm\sqrt{1+h}). Distinct from DTSs, the oscillatory tails of MTSs are modulated both by PHEOD order mm and soliton’s momentum PP, with the modulation amplitude rising with mm and PP. Furthermore, as PP approaches its maximum value, corresponding to the critical velocity Vcr2V_{\mathrm{cr2}}, the region around the central point τ=0\tau=0 no longer retains the central dip, inherent to conventional dark solitons, with a bubble-like structural feature emerging instead. The corresponding phase distributions of the MTSs are presented in Figure 5(aFigure 4–Figure 5(bFigure 4, confirming that the MTSs arise from two half-infinite CW states separated by the π\pi phase shift.

Dependence of MTS profiles on momentum and dispersion order

Figure 5. Dependence of MTS profiles on momentum PP and dispersion order mm at fixed h=1/15h=1/15. For m=4m=4 and m=10m=10, panels (a1), (b1) and (a2), (b2) display ψr\psi_r and ψi\psi_i, respectively. The associated modulus and phase profiles are presented in panels (a3)–(b3) and (a4)–(b4).

The typical dynamical behavior of the MTSs is displayed in Figure 6. The trajectories in Figure 6(aFigure 1, Figure 6(bFigure 1, and Figure 6(cFigure 1} show that the MTS profiles remain essentially unchanged up to t=1000t=1000. The comparison between Figure 6(aFigure 1 and Figure 6(bFigure 1} demonstrates that two MTSs with opposite velocities undergo counterpropagating evolution. As illustrated in 6(c1), the value of ∣V∣|V| determines the drift speed of each soliton. The spectra in Figure 6(aFigure 2, Figure 6(bFigure 2, and Figure 6(cFigure 2, obtained from Eq. (9), are consistent with the stability for all three examples. Figure 6(aFigure 3, Figure 6(bFigure 3, and Figure 6(cFigure 3} compare the initial shapes of the perturbed soliton and final ones, produced by the extended evolution, further validating the robustness of the MTSs.

Evolution and linear-stability spectra of MTSs for $m=4$ and $h=1/15$

Figure 6. The evolution and linear-stability spectra of MTSs for m=4m=4 and h=1/15h=1/15. The stable evolution of the MTSs subjected to 10% random-amplitude noise for different soliton’s velocities and momenta: (a1_1) V=0.05V=0.05 and P=2.916P=2.916; (b1_1) V=−0.05V=-0.05 and P=−2.916P=-2.916; (c1_1) V=−0.2V=-0.2 and P=0.776P=0.776. Panels (a2_2), (b2_2), and (c2_2) give the associated stability spectra, while panels (a3_3), (b3_3), and (c3_3) compare the perturbed input profiles with those obtained at t=1000t=1000.

Finally, to explore interactions between MTSs, we construct the initial field profile for the head-on collision, as ψ(τ,t=0)=ψ1(τ+τ1)+ψ2(τ−τ2)\psi(\tau,t=0)=\psi_1(\tau+\tau_1)+\psi_2(\tau-\tau_2), where the two solitons have opposite velocities, V1=−V2V_1=-V_2. Figure 7(aFigure 1} illustrates the repulsive soliton-soliton interaction, featuring rebound and emergence of pronounced amplitude oscillations, alongside emission of dispersive waves, as seen in Figure 7(bFigure 1. Figure 7(aFigure 2 and Figure 7(aFigure 3} present another repulsive regime, in which the collision triggers conspicuous deformation of the bouncing solitons, accompanied by the emission of weak dispersive-wave radiation. These phenomena reveal the inelastic character of

Collisions between stable MTSs

Figure 7. Collisions between stable MTSs, with parameters (a1_1) V1=−V2=0.4V_1=-V_2=0.4 and P=1.717P=1.717; (a2_2) V1=−V2=0.2V_1=-V_2=0.2 and P=2.680P=2.680; (a3_3) V1=−V2=0.24V_1=-V_2=0.24 and P=3.310P=3.310; (a4_4) V1=−V2=0.3V_1=-V_2=0.3 and P=2.481P=2.481. Panels (b1_1), (b2_2), (b3_3), and (b4_4) display the initial MTS profiles and final ones, observed at t=500t=500. Other parameters are τ1=τ2=48\tau_1=\tau_2=48, m=10m=10, and h=1/15h=1/15.

the MTS collisions, as corroborated by the corresponding field profiles in Figure 7(bFigure 2 and Figure 7(bFigure 3. Figure 7(a4) depicts a different outcome of the collision, when the two MTSs undergo complete fusion, along with intense emission of dispersive waves, as seen in Figure 7(bFigure 4.

Conclusion

This work aims to produce dissipative dark-soliton families, sustained by the interplay of the PHEOD (pure-high-even-order dispersion) and parametric drive, extending the setting beyond the conventional second-order-dispersion regime. Two new families of stable topological solitons are found, namely, quiescent DTSs (dissipative topological solitons) and MTSs (moving topological solitons), for PHEOD orders m=4m = 4, 6, 8, and 10. The topological solitons are combined with the zero-amplitude point, which connects two semi-infinite CW backgrounds with the π\pi phase shift between them. PHEOD gives rise to pronounced oscillatory tails, distinguishing these states from conventional dark solitons. The systematic linear-stability analysis and direct simulations corroborate the stability of both DTSs and MTSs. Pairwise collisions between the MTSs are inelastic, featuring two possibilities: the inter-soliton rebound or complete fusion, both accompanied by the emission of radiation waves.

These results demonstrate that PHEOD can do more than smoothly distort conventional dissipative soliton profiles: it provides the mechanism for generating topological solitons with specific internal structures and interaction dynamics. These findings establish PHEOD as a basis for engineering robust dissipative topological solitons, both quiescent and moving ones. As an extension of the analysis, it may be interesting to construct nonlinear states in systems exhibiting competition between dispersion terms of different orders and different signs (normal/anomalous).

Funding. National Natural Science Foundation of China (Grant No. 62675244); Fundamental Research Program of Shanxi Province (Grant No. 202303021211185); Israel Science Foundation (Grant No. 1695/22); ANID (Chile) through FONDECYT (Grant No. 1260401).

Disclosures. The authors declare no conflicts of interest.

Data availability. Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.

References

  1. [1]N. Akhmediev and A. Ankiewicz, eds., Dissipative Solitons (Springer, 2005).
  2. [2]N. Akhmediev and A. Ankiewicz, eds., Dissipative Solitons: From Optics to Biology and Medicine (Springer, 2008).
  3. [3]M. F. S. Ferreira, ed., Dissipative Optical Solitons (Springer, 2022).
  4. [4]P. Grelu and N. Akhmediev, “Dissipative solitons for mode-locked lasers,” Nat. Photonics 6, 84–92 (2012).
  5. [5]A. Mecozzi, W. L. Kath, P. Kumar, and C. G. Goedde, “Long-term storage of a soliton bit stream by use of phase-sensitive amplification,” Opt. Lett. 19, 2050–2052 (1994).
  6. [6]S. Longhi, “Ultrashort-pulse generation in degenerate optical parametric oscillators,” Opt. Lett. 20, 695–697 (1995).pubmed.ncbi.nlm.nih.gov/19859300
  7. [7]S. Longhi, “Stable multipulse states in a nonlinear dispersive cavity with parametric gain,” Phys. Rev. E 53, 5520–5522 (1996).pubmed.ncbi.nlm.nih.gov/9964894
  8. [8]W. H. Renninger, A. Chong, and F. W. Wise, “Dissipative solitons in normal-dispersion fiber lasers,” Phys. Rev. A 77, 023814 (2008).DOI
  9. [9]J. Yelo-Sarrión, F. Leo, and S.-P. Gorza, “Dissipative solitons in mode-locked parity–time-symmetric lasers,” Nat. Photonics 20, 310–316 (2026).
  10. [10]L. A. Lugiato and R. Lefever, “Spatial dissipative structures in passive optical systems,” Phys. Rev. Lett. 58, 2209–2211 (1987).pubmed.ncbi.nlm.nih.gov/10034681
  11. [11]I. V. Barashenkov, M. M. Bogdan, and V. I. Korobov, “Stability diagram of the phase-locked solitons in the parametrically driven, damped nonlinear Schrödinger equation,” Europhys. Lett. 15, 113–118 (1991).DOI
  12. [12]N. V. Alexeeva, I. V. Barashenkov, and D. E. Pelinovsky, “Dynamics of the parametrically driven NLS solitons beyond the onset of the oscillatory instability,” Nonlinearity 12, 103–140 (1999).
  13. [13]I. V. Barashenkov and E. V. Zemlyanaya, “Stable complexes of parametrically driven, damped nonlinear Schrödinger solitons,” Phys. Rev. Lett. 83, 2568–2571 (1999).DOI
  14. [14]I. V. Barashenkov, S. R. Woodford, and E. V. Zemlyanaya, “Parametrically driven dark solitons,” Phys. Rev. Lett. 90, 054103 (2003).
  15. [15]N. Dror and B. A. Malomed, “Spontaneous symmetry breaking in coupled parametrically driven waveguides,” Phys. Rev. E 79, 016605 (2009).pubmed.ncbi.nlm.nih.gov/19257158
  16. [16]N. Engelbert, F. De Lucia, P. Parra-Rivas, C. Mas Arabí, P.-J. Sazio, S.-P. Gorza, and F. Leo, “Parametrically driven Kerr cavity solitons,” Nat. Photonics 15, 857–861 (2021).
  17. [17]G. Moille, M. Leonhardt, D. Paligora, N. Engelbert, F. Leo, J. Fatome, K. Srinivasan, and M. Erkin̈talo, “Parametrically driven pure-Kerr temporal solitons in a chip-integrated microcavity,” Nat. Photonics 18, 617–624 (2024).arxiv.org/abs/2306.04078
  18. [18]D. Wang, R. Li, D. Laroze, B. A. Malomed, and P. Li, “Quiescent and traveling solitons in the fractional parametrically driven damped nonlinear Schrödinger equation,” Phys. Rev. E 114, 014223 (2026).
  19. [19]A. Blanco-Redondo, C. M. de Sterke, J. E. Sipe, T. F. Krauss, B. J. Eggleton, and C. Husko, “Pure-quartic solitons,” Nat. Commun. 7, 10427 (2016).
  20. [20]A. F. J. Runge, D. D. Hudson, K. K. K. Tam, C. M. de Sterke, and A. Blanco-Redondo, “The pure-quartic soliton laser,” Nat. Photonics 14, 492–497 (2020).arxiv.org/abs/1910.10314
  21. [21]Z. Wang, Y. Lin, Y. Song, Z. Deng, J. Liu, and C. Zhang, “Revealing the dynamics of multiple solitons and soliton pulsations in a quartic-dispersion fiber laser,” Laser Photonics Rev. 20, e01851 (2026).DOI
  22. [22]A. F. J. Runge, Y. L. Qiang, T. J. Alexander, M. Z. Rafat, D. D. Hudson, A. Blanco-Redondo, and C. M. de Sterke, “Infinite hierarchy of solitons: Interaction of Kerr nonlinearity with even orders of dispersion,” Phys. Rev. Res. 3, 013166 (2021).DOI
  23. [23]Y. Han, B. Gao, H. Wen, C. Ma, J. Huo, Y. Li, L. Zhou, Q. Li, G. Wu, and L. Liu, “Pure-high-even-order dispersion bound solitons complexes in ultra-fast fiber lasers,” Light Sci. Appl. 13, 101 (2024).DOI
  24. [24]Q. Li, F. Qiao, H. Wen, Y. Li, L. Zhou, G. Wu, L. Liu, and B. Gao, “The bound states of pure-high-even-order dispersion solitons,” Chaos Solitons Fractals 204, 117742 (2026).DOI
  25. [25]Z. C. Qian, M. Liu, A. P. Luo, Z. C. Luo, and W. C. Xu, “Dissipative pure-quartic soliton fiber laser,” Opt. Express 30, 22066–22073 (2022).DOI
  26. [26]Z. L. Wu, M. Liu, Y. X. Gao, Z. X. Zhang, M. Luo, Y. Hu, T. J. Li, A. P. Luo, W. C. Xu, and Z. C. Luo, “Spectral sidebands of dissipative soliton in a positive fourth-order-dispersion fiber laser,” Opt. Express 32, 47882–47892 (2024).DOI
  27. [27]Z. L. Wu, Z. X. Zhang, T. J. Li, M. Luo, M. Liu, A. P. Luo, W. C. Xu, and Z. C. Luo, “Pulsating dynamics of dissipative pure-quartic soliton in a fiber laser,” Opt. Express 33, 10129–10139 (2025).DOI
  28. [28]H. Taheri and A. B. Matsko, “Quartic dissipative solitons in optical Kerr cavities,” Opt. Lett. 44, 3086–3089 (2019).DOI
  29. [29]P. Parra-Rivas, S. Hetzel, Y. V. Kartashov, P. Fernández de Córdoba, J. A. Cenejnero, A. Aceves, and C. Milián, “Quartic Kerr cavity combs: Bright and dark solitons,” Opt. Lett. 47, 2438–2441 (2022).
  30. [30]K. Liu, S. Yao, and C. Yang, “Raman pure quartic solitons in Kerr microresonators,” Opt. Lett. 46, 993–996 (2021).
  31. [31]S. Mei, M. Liu, H. Huang, T. He, Z. Lu, and W. Zhao, “Route to pure-quartic solitons in the mid-infrared,” Opt. Lett. 50, 2848–2851 (2025).
  32. [32]H. Zhu, M. Liu, H. Huang, J. Wang, S. Mei, and W. Zhao, “Raman-assisted pure quartic platicons in microresonators,” Opt. Lett. 50, 3357–3360 (2025).
  33. [33]C. Silvestri, Y. L. Qiang, K. Panda, J. Widjaja, S. Coen, C. M. de Sterke, and A. F. J. Runge, “Pure high-order dispersion dissipative Kerr solitons in optical cavities,” Opt. Lett. 50, 4262–4265 (2025).
  34. [34]P. Li, L. Xing, D. Wang, D. Mihalache, D. Laroze, and B. A. Malomed, “Parametrically driven pure-quartic solitons,” Opt. Lett. 51, 3217–3220 (2026).arxiv.org/abs/2605.15591
  35. [35]N. D. Mermin, “The topological theory of defects in ordered media,” Rev. Mod. Phys. 51, 591–648 (1979).DOI
  36. [36]N. Manton and P. Sutcliffe, Topological Solitons, Cambridge Monographs on Mathematical Physics (Cambridge University Press, Cambridge, 2004).DOI
  37. [37]S. Trillo, M. Haelterman, and A. Sheppard, “Stable topological spatial solitons in optical parametric oscillators,” Opt. Lett. 22, 970–972 (1997).pubmed.ncbi.nlm.nih.gov/18185722
  38. [38]G. L. Oppo, A. J. Scroggie, and W. J. Firth, “From domain walls to localized structures in degenerate optical parametric oscillators,” J. Opt. B: Quantum Semiclassical Opt. 1, 133–138 (1999).DOI
  39. [39]M. Gilles, P.-Y. Bonny, J. Garnier, A. Picozzi, M. Guasoni, and J. Fatome, “Polarization domain walls in optical fibres as topological bits for data transmission,” Nat. Photonics 11, 102–107 (2017).DOI
  40. [40]N. Engelbert, R. M. Gray, L. Ledezma, R. Sekine, T. Zacharias, R. Ramesh, B. K. Gutierrez, P. Parra-Rivas, and A. Marandi, “Topological soliton frequency comb in nanophotonic lithium niobate,” Nature 652, 76–81 (2026).
  41. [41]J. Yang, “Newton-conjugate-gradient methods for solitary wave computations,” J. Comput. Phys. 228, 7007–7024 (2009).DOI
  42. [42]J. Yang, Nonlinear Waves in Integrable and Nonintegrable Systems (SIAM, 2010).DOI
  43. [43]J. Widjaja, E. Kobakhidze, T. R. Cartwright, J. P. Lourdesamy, A. F. J. Runge, T. J. Alexander, and C. M. de Sterke, “Absence of Galilean invariance for pure-quartic solitons,” Phys. Rev. A 104, 043526 (2021).DOI

Paper details

vlnc:W_a9czjuaj
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 11

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents