Exact Charge, Current, and Velocity Fields of Interacting Korteweg–de Vries Solitons
Thibault Bonnemain, François Copie, Pierre Suret, Stéphane Randoux · 2026
Read Exact Charge, Current, and Velocity Fields of Interacting Korteweg–de Vries Solitons by Thibault Bonnemain, François Copie, Pierre Suret, Stéphane Randoux (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2606-02043.txt.
Introduction
Bonnemain and co-authors derive exact fields for KdV solitons in interaction. Solitons act as both waves and particles in such systems. Mid interaction, they deform and spread out. Telling single solitons apart then gets hard. A field-based view at small scale resolves that issue here.
Exact fields of charge, current, and speed are built in space-time. They come from the inverse scattering transform for KdV. Single solitons can be tracked through full interactions. Paths and deformation are measured past the far-field regime. N interacting solitons obey N continuity equations. Interactions are encoded in initial links between them. Those links are then carried forward in time.
Catalog record
| CREATOR | Thibault Bonnemain, François Copie, Pierre Suret, Stéphane Randoux |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2606.02043
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Exact Charge, Current, and Velocity Fields of Interacting Korteweg–de Vries Solitons Thibault Bonnemain, François Copie, Pierre Suret, Stéphane Randoux Abstract Solitons in integrable systems exhibit a dual wave–particle character, yet their identification as individual objects becomes ambiguous during interactions, where they deform and delocalize. We develop a microscopic, field-based description that resolves this issue by introducing exact space–time fields of charge, current, and velocity derived from the inverse scattering transform (IST) of the Korteweg–de Vries (KdV) equation. This framework enables individual KdV solitons to be tracked throughout interactions, providing a quantitative description of their trajectories and deformation beyond the asymptotic regime. We show that the dynamics of N N interacting solitons can be formulated in terms of N N independent continuity equations, in which interactions are encoded in initial correlations that are subsequently propagated. From this microscopic viewpoint, effective velocities and hydrodynamic behavior emerge upon coarse graining, recovering kinetic theory of soliton gases and Generalized Hydrodynamics as scaling limits. Our results establish a direct connection between the wave-based IST formalism and particle-like emergent descriptions, offering a unified framework for soliton dynamics across scales. Self-localized nonlinear wave packets known as solitons, sustained by an exact compensation between dispersion and nonlinearity, emerge as universal excitations across a wide range of physical platforms, from optics and quantum gases to fluids, metamaterials, and biological systems 58 ; 5 ; 67 ; 19 ; 49 ; 57 ; 17 ; 40 . Solitons exhibit a dual nature in nonlinear physics, behaving both as particle-like entities and as coherent nonlinear waves. On the one hand, they display remarkable robustness, asymptotically recovering their shape, amplitude, and velocity after interactions, as first revealed in numerical simulations of the Korteweg–de Vries (KdV) equation by Zabusky and Kruskal 71 . On the other hand, their wave nature is captured by the inverse scattering transform (IST), introduced by Gardner et al. 35 , which provides an exact solution method for a broad class of integrable nonlinear dispersive equations 72 ; 56 ; 69 ; 2 . A hallmark of soliton interactions is their elastic character: collisions preserve all intrinsic parameters and, asymptotically, result only in finite position shifts 57 ; 17 . Within the IST framework, solitons correspond to nonlinear normal modes associated with discrete eigenvalues that remain invariant in time 53 ; 52 ; 24 . However, during interactions, solitons deform and delocalize, obscuring their position, extent, and individual contribution to the wavefield 33 ; 70 ; 54 ; 6 ; 3 ; 4 . This raises a fundamental question: how can one define and track solitons as physical entities beyond the asymptotic (well-separated) regime? This issue becomes particularly acute in many-body settings, where solitons form coherent structures such as dispersive shock waves or statistical ensembles known as soliton gases (SGs) 50 ; 26 ; 73 ; 25 ; 59 ; 66 ; 31 . Effective descriptions based on kinetic theory 73 ; 27 ; 28 ; 30 and, more recently, Generalized Hydrodynamics (GHD) 7 ; 12 ; 22 ; 63 ; 8 ; 47 ; 65 successfully capture large-scale dynamics by treating solitons as interacting quasiparticles 61 ; 13 ; 32 ; 48 , thereby providing a framework to address fundamental questions related to, e.g., modulational instability or wave-mean flow interaction 36 ; 14 . Yet, these approaches rely on coarse-grained descriptions in which the underlying wave structure is implicit, leaving unclear what microscopic information is lost and how it affects dynamics at intermediate scales. In this Letter, we develop a microscopic, field-based description of interacting solitons directly rooted in the IST spectral framework. Building on the association of the local phase-space density of KdV solitons with the squared eigenfunctions of the associated spectral problem 32 , we construct exact space–time fields for soliton charge, current, and velocity. This representation provides a physically transparent picture in which individual solitons remain well defined even during strong interactions, beyond the standard isolated sech-profile description. Within this framework, effective velocities and local contributions to conserved quantities are naturally defined without invoking a collision-rate ansatz, and kinetic and hydrodynamic theories emerge as scaling limits. Our results establish an explicit bridge between the wave-based IST formalism and particle-based emergent descriptions, providing a unified view of soliton dynamics across scales. Figure 1: Numerical simulation showing the interaction between 2 (upper panel) and 8 (lower panel) KdV solitons. (a) The field u(x,t) is computed from Eq. ( 3 ) while the eigenfunctions \psi_{m}(x,t) are determined by solving the Marchenko system ( 4 ) using arbitrary numerical precision. (b), (e), (i), (l) Charge densities Q_{m}(x,t)=\psi_{m}^{2}(x,t) associated with solitons with eigenvalues \eta_{m} . (c), (f), (j), (m) Associated currents J_{m}(x,t) computed using Eq. ( 8 ). (d), (g), (k), (n) Associated velocity fields v_{m}(x,t) computed using the definition v_{m}=J_{m}/Q_{m} or equivalently the Darboux-deleted relation given by Eq. ( 13 ). In the upper panel, \eta_{1}=1 , \eta_{2}=1/\sqrt{2} . In the lower panel, \eta_{m}=[1.1,0.7,0.6,0.5,0.4,0.3,0.2,0.1] . IST formalism and basic ideas— Because of its universal physical relevance, the KdV equation u_{t}+3(u^{2})_{x}+u_{xxx}=0\;, (1) is seen as a paradigmatic example of nonlinear, dispersive, integrable partial differential equation featuring solitons 1 ; 55 ; 10 ; 60 . The KdV initial value problem is commonly addressed via the IST, by solving the auxiliary linear problem \psi_{xx}+(u+\lambda)\psi=0\;,\quad\psi_{t}+\psi_{xxx}-3(u+\lambda)\psi_{x}=0\;. (2) If the KdV wavefield u is purely solitonic, it can be reconstructed in terms of N eigenfunctions \psi_{m} and the associated invariant eigenvalues \lambda_{m}=-\eta_{m}^{2} 34 u(x,t)=4\sum_{m=1}^{N}\eta_{m}\psi_{m}^{2}(x,t)\;. (3) All the eigenfunctions are real with normalisation \int_{\mathbb{R}}{\rm d}x\,\psi_{m}^{2}(x,t)=1 , and are related to one another through the Marchenko system of linear equations \mathcal{A}\,\bm{\psi}(x,t)=\bm{a}(x,t)\;. (4) Here \bm{\psi}(x,t):=\big(\psi_{1}(x,t),\dots,\psi_{N}(x,t)\big)^{T} and \bm{a}=\left(c_{1}(t)\exp(-\eta_{1}x),\ \cdots,\ c_{n}(t)\exp(-\eta_{N}x)\right)^{T} , while \mathcal{A}_{mn}(x,t):=\delta_{mn}+\frac{c_{m}(t)c_{n}(t)}{\eta_{m}+\eta_{n}}\,e^{-(\eta_{m}+\eta_{n})x}\;, (5) where we introduced the norming constants c_{m}(t)=c_{m}(0)\exp\left(4\eta_{m}^{3}t\right) , which encode the initial condition. A direct consequence of the decomposition ( 3 ) is the identification of solitons with the squared eigenfunctions Q_{m}:=\psi_{m}^{2} , which encode both their position and extension as they deform through interactions. This identification naturally reframes solitons as densities of (conserved) charges in motion. In this sense, the Q_{m} provide a microscopic realization of the quasiparticle densities underlying GHD 22 , where the distribution of eigenvalues plays the role of a momentum distribution. This perspective is also consistent with recent rigorous approaches to the GHD of the Toda lattice 3 ; 4 . Dynamics of the squared eigenfunctions— To study the dynamics of interacting solitons, we study that of the squared eigenfunctions. Multiplying the first equation of ( 2 ) by 6\psi_{x} , the second by 2\psi , adding the results and using the identity (\psi^{2})_{xxx}=3(\psi_{x})^{2}_{x}+2\psi\psi_{xxx} , we have 34 \partial_{t}Q_{m}+6u\partial_{x}Q_{m}+\partial_{xxx}Q_{m}=0\;, (6) sometimes called “interacting KdV equations” 70 . We propose a new interpretation of these equations and show they can be put in more convenient forms which do not involve the wavefield u . Because of their normalisation, the Q_{m} ’s are, by definition, conserved densities, so that there exists N associated currents, J_{m} ’s, defined by \partial_{t}Q_{m}+\partial_{x}J_{m}=0\;, (7) with J_{m}=\int_{-\infty}^{x}{\rm d}x\left[6u\partial_{x}Q_{m}+\partial_{xxx}Q_{m}\right] . To explicitly compute these currents, recall the auxiliary problem ( 2 ) which implies 6u\partial_{x}Q_{m}=-6\partial_{x}\left[(\partial_{x}\psi_{m})^{2}+\lambda Q\right] , hence J_{m}=-6\left(\partial_{x}\sqrt{Q_{m}}\right)^{2}+6\eta_{m}^{2}Q_{m}+\partial_{xx}Q_{m}\;, (8) meaning we may define a velocity field v_{m} that only depends on the associated Q_{m} v_{m}:=\frac{J_{m}}{Q_{m}}=-\frac{3}{2}\left(\partial_{x}\log Q_{m}\right)^{2}+\frac{\partial_{xx}Q_{m}}{Q_{m}}+6\eta_{m}^{2}\;. (9) This explicit relation between v_{m} and Q_{m} is a