Self-propelled particles at sufficient density undergo motility-induced phase separation — spontaneous clustering into dense and sparse regions without any attractive interaction. The mechanism is purely kinematic: particles slow where they're crowded, accumulate where they slow, and the feedback loop drives macroscopic phase separation. But many real active systems also exhibit chemotaxis — directed motion along chemical gradients that the particles themselves generate through consumption or secretion. Zhao, Yu, Kosmrlj, and Datta show that chemotaxis doesn't just modify MIPS; it can suppress it entirely, replacing static phase separation with traveling waves and spiral patterns.
The analysis reveals four bifurcation types governing the transition between stationary and oscillatory states: pitchfork, saddle-node, infinite period, and supercritical Hopf. Closed-form expressions for pattern amplitude and wave velocity match simulations closely. The framework extends to mixtures — particles with different chemotactic sensitivities coexisting in the same medium — where heterogeneous response further enriches the pattern zoo.
The structural principle: adding a directional bias to a system prone to clustering doesn't just shift the clustering threshold — it can replace the instability entirely with a qualitatively different one. Chemical sensing converts a static accumulation problem into a dynamic propagation problem. The same density fluctuation that seeds phase separation becomes the front of a traveling wave when particles can read their own chemical wake. Perception changes the topology of instability.
(arXiv:2603.02595)