A Minimal Active-Particle Realization of Non-Hermitian Chern Bulk-Boundary Correspondence

Tong Zhu, Zhigang Zheng · 2026

Read A Minimal Active-Particle Realization of Non-Hermitian Chern Bulk-Boundary Correspondence by Tong Zhu, Zhigang Zheng (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-2606-24926.txt.

Introduction

Tong Zhu and Zhigang Zheng show that a simple swarm of active particles can realise a topological principle from non Hermitian physics. Their model blends two classics: Vicsek flocking, where particles align headings, and Kuramoto synchronisation, where phases lock together. A phase lag in the local alignment frustrates the flock, driving bulk instabilities at a finite wavelength. Under collision boundaries, a robust one way flow appears along the edge.

A nonlinear saturation ansatz organises the explanation. The linearised hydrodynamics selects the unstable wavelength and the spectral topology. Particle dynamics then saturates the chosen mode. The continuum spectrum supports Chern numbers of plus and minus two, fixed by the dispersion matrix's spin structure. Strip spectral flow predicts chiral edge propagation from those numbers. Particle simulations agree in the nontrivial sectors. Zhu and Zheng present the model as a minimal active matter home for Chern bulk boundary physics.

Catalog record

CREATORTong Zhu, Zhigang Zheng
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE11 KB
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2606.24926

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

A Minimal Active-Particle Realization of Non-Hermitian Chern Bulk-Boundary Correspondence Tong Zhu, Zhigang Zheng Abstract We show that a minimal frustrated Vicsek–Kuramoto active-particle model realizes a non-Hermitian Chern bulk-boundary correspondence. A Sakaguchi-type phase lag in the local heading alignment generates finite-wavenumber bulk instabilities and, under collision boundaries, robust one-way boundary flow. The organizing principle is a nonlinear saturation ansatz: the linearized hydrodynamic operator selects the unstable wavelength and spectral topology, while nonlinear particle dynamics saturates the selected mode. The isotropic continuum spectrum compactifies the wave-number plane and supports spectral projectors with Chern numbers C = ± 2 C=\pm 2 , fixed by the spin structure of the dispersion matrix. Strip spectral flow then predicts chiral edge propagation, in agreement with particle simulations in the nontrivial sectors. Introduction.— Active matter converts free energy input into collective motion, producing flocking, clustering, vortices, and ordered nonequilibrium patterns, the famous example is Vicsek model 9 ; 8 ; 13 ; 12 . Meanwhile, a complementary route to collective order is synchronization: Kuramoto-type interactions lock phases through local coupling 5 ; 11 ; 1 . Kuramoto–Vicsek models that combine self-propelled motion with phase-like orientation dynamics therefore provide an idealized simple setting in which synchronization, pattern formation, and nonequilibrium steady transport can meet 2 . We study a frustrated Vicsek–Kuramoto model in which self-propelled particles align their heading angles with a local Sakaguchi-type phase lag 10 . Based on simulations previous studies on systems under periodic boundary have identified an onset of pattern with finite range of the wave-number near \alpha=\pi/2 ; a hexagonal lattice pattern emerges from the swarming particles and a narrow orientation-modulated hyper-uniform regime appears 6 ; 7 . These previous systems have an inherent toric topology; to dissect its role from those of local interactions among particles played in the bulk spectra, one naturally replaces the periodic boundary with a reflecting boundary representing elastic collisions. The linearized continuum operator is non-Hermitian, and its isotropy makes the compactified wave-number plane a natural base manifold for spectral projectors. In the line-gapped regimes, the relevant projectors carry Chern numbers C=\pm 2 . The corresponding strip spectrum contains edge branches whose group velocities determine the observed one-way boundary flow. Thus the saturated particle dynamics provides a minimal active-matter realization of non-Hermitian Chern bulk-boundary correspondence. Particle model.— The microscopic dynamics are \dot{\mathbf{x}}_{i}=v\mathbf{e}(\theta_{i}),\qquad\mathbf{e}(\theta_{i})=(\cos\theta_{i},\sin\theta_{i}), (1) and \dot{\theta}_{i}=\omega_{i}+\frac{K}{|\mathcal{N}_{i}|}\sum_{j\in\mathcal{N}_{i}}[\sin(\theta_{j}-\theta_{i}+\alpha)-\sin\alpha], (2) for i=1,2,\dots,N . Here K is the numerical neighborhood-averaged coupling, \omega_{i} is set to 0 for each i , \mathcal{N}_{i} is the interaction neighborhood, and \alpha is the frustration angle. The subtraction of \sin\alpha removes the trivial drift of an exactly aligned neighborhood. In the continuum equations below, the coefficient \lambda denotes the effective strength after replacing the discrete average by an unnormalized kernel convolution; for a typical neighborhood population N^{\prime} , \lambda\simeq K/N^{\prime} up to kernel normalization 14 . The periodic-boundary baseline is summarized in Fig. 1 . Below the finite-wavenumber threshold \alpha=\pi/2 , the system is a synchronized state. Above threshold, the dominant linear mode selects a finite wavelength and most parameters produce dynamic swarming lattices; hyperuniformity occurs only in a narrow post-onset window where the phase-locking ansatz gives an effective vortex scale incompatible with the linearly selected wavelength 6 ; 7 ; 14 . Representative circular collision-boundary states are shown in Fig. 2 . Figure 1: Periodic-boundary phase diagram in the (K,d_{0}) plane for L=7,N=2000,K=20,d_{0}=1.55 and v=3 , reproduced from the published swarming-lattice analysis 6 . The diagram provides the bulk baseline for the collision-boundary response: below the finite-wavenumber threshold the periodic system is synchronized, whereas above threshold the linear spectrum selects a finite wavelength. (a) \alpha=0.4\pi (b) \alpha=0.6\pi (c) \alpha=\pi Figure 2: Representative circular collision-boundary states for L=7 , N=2000 , K=20.75 , and d_{0}=1 . (a) At \alpha=0.4\pi , the skin effect is clean: particles are tightly localized on the circular boundary. (b) At \alpha=0.6\pi , the patterned state rotates as a whole, while the boundary flow runs opposite to the bulk rotation. (c) At \alpha=\pi , the Chern sector is trivial, but the circular boundary supports a visible boundary lattice; supplementary movies resolve the two counter-propagating boundary lattice streams, i.e. boundary flows cancel out. Continuum spectrum.— Let f(\mathbf{x},\theta,t) be the one-particle probability-density field. Its lowest angular moments define \displaystyle\rho \displaystyle=\int_{0}^{2\pi}f\,d\theta, (3) \displaystyle p_{x} \displaystyle=\int_{0}^{2\pi}f\cos\theta\,d\theta,\qquad p_{y}=\int_{0}^{2\pi}f\sin\theta\,d\theta. After angular Fourier truncation and fast-mode closure, linearizing the hydrodynamic equations around (\rho,\mathbf{p})=(\rho_{0},0) gives \partial_{t}\widehat{U}(\mathbf{q},t)=L(\mathbf{q})\widehat{U}(\mathbf{q},t),\quad\widehat{U}=(\widehat{\delta\rho},\widehat{\delta p_{x}},\widehat{\delta p_{y}})^{T}, (4) with L(\mathbf{q})=\begin{pmatrix}0&-ivq_{x}&-ivq_{y}\\ -\frac{iv}{2}q_{x}&a(q)&b(q)\\ -\frac{iv}{2}q_{y}&-b(q)&a(q)\end{pmatrix},\quad q=|\mathbf{q}|, (5) a(q)=\frac{\lambda\rho_{0}}{2}\widehat{G}(q)\cos\alpha, (6) and b(q)=-\omega+\lambda\rho_{0}\widehat{G}(0)\sin\alpha-\frac{\lambda\rho_{0}}{2}\widehat{G}(q)\sin\alpha+\frac{v^{2}q^{2}}{4D_{0}}. (7) For a disk-like interaction kernel, \widehat{G}(q)=2\pi d_{0}J_{1}(qd_{0})/q . The continuum reduction, the derivation of Eqs. ( 5 )–( 7 ), and the detailed spectral-topological analysis used below are given in Ref. 14 . Since L(\mathbf{q}) is generally non-Hermitian, its eigenvalues \sigma_{j} are complex; \operatorname{Re}\sigma_{j} is the linear growth rate and \operatorname{Im}\sigma_{j} is the oscillation frequency. Figure 3 (a)–(c) shows the key spectral change. At \alpha=0.4\pi , the dominant growth is at zero wavenumber. At \alpha=0.6\pi , the real part selects a finite wavenumber, consistent with patterned bulk states. At \alpha=\pi , finite-wavenumber growth reappears, but the imaginary line gap closes, so the corresponding sector is topologically trivial. (a) \alpha=0.4\pi (b) \alpha=0.6\pi (c) \alpha=\pi (d) mixed-band Chern (e) single-band Chern Figure 3: Linear spectrum and Chern platforms. (a)–(c) Representative one-direction dispersions at \alpha=0.4\pi , 0.6\pi , and \pi ; isotropy makes a single direction representative of the wave-number plane. The comparison shows the zero-wavenumber to finite-wavenumber transition from 0.4\pi to 0.6\pi , and the imaginary line-gap closure at \pi . (d,e) Mixed-band and single-band Chern computations for v=3 , \omega=0 . Nontrivial branches lock to C=\pm 2 over broad parameter intervals, while the full three-band projector is trivial. Chern number from isotropy and spin.— For a separated band or band cluster, define the Riesz projector P(\mathbf{q})=\frac{1}{2\pi i}\oint_{\Gamma}[zI-L(\mathbf{q})]^{-1}\,dz, (8) where \Gamma encloses only the target spectrum 4 . The first Chern number is C(P)=\frac{1}{2\pi i}\int_{\mathcal{M}}\mathrm{Tr}\left(P\,dP\wedge dP\right), (9) with \mathcal{M} the compactified wave-number plane. Numerically, we also evaluate the same invariant using a biorthogonal Fukui–Hatsugai–Suzuki construction 3 . The integer value can be understood analytically. Rotational covariance gives L(q,\phi)=R(\phi)L(q,0)R^{-1}(\phi), (10) where R rotates the polarization components. The angular generator has spin weights 0,+1,-1 on the density and circular-polarization sectors. If the target projector connects opposite circular-polarization sectors between the compactification endpoints, the Chern number is the spin-weight difference, |C|=|(+1)-(-1)|=2. (11) Figure 4: Strip spectral flow at \alpha=0.5\pi . Colored points are left- and right-edge branches; gray points are bulk-like strip modes. The edge branches cross the line gap with a definite orientation, giving the boundary chirality associated with the nontrivial Chern sectors. The computed platforms in Fig. 3 (d,e) match this endpoint argument. Boundary spectral flow.— To connect the bulk invariant to an edge response, one fixes the tangential wavenumber k_{y} and constructs a strip operator from the Fourier blocks of L(k_{x},k_{y}) . The strip spectrum contains bulk-like modes and boundary-localized modes. For an edge branch \sigma(k_{y})=\gamma(k_{y})+i\Omega(k_{y}), (12) the oscillatory phase gives group velocity v_{g}=-\partial_{k_{y}}\Omega(k_{y}). (13) The edge modes in Fig. 4 have positive group velocity with the plotted orientation. When mapped onto a closed two-dimensional collision boundary, this predicts counterclockwise edge circulation, as observed in the particle simulations. This comparison is made at the level of linear selection and nonlinear saturation: the topology belongs

Read the complete text