Séminaire

Mardi 27 Janvier 2026 à 14h00.

Phase transitions in and out of equilibrium: on-lattice 2D melting and active particles with memory

Alexis Poncet
(Laboratoire de Physique de l ENS de Lyon)

Salle de séminaires Lippmann

Invité(e) par
Cécile Cottin-Bizonne

présentera en 1 heure :

''
Phase transitions are ubiquitous in physical systems, they are well known at thermal equilibrium but also prevalent in far-from-equilibrium contexts. In this talk, I will present two recent projects that give new insights on two paradigmatic transitions: the equilibrium melting of two-dimensional spheres, and the non-equilibrium phase separation of active particles.

First, I will discuss the melting of 2D spheres, which typically follows a two-step scenario involving an intermediate hexatic phase. However, the nature of the hexatic-liquid transition, continuous (Kosterlitz-Thouless type) or discontinuous, depends on the interactions. To explore this crossover, we introduce a minimal on-lattice model: a clock model with vacancies, which reproduces the full range of 2D melting scenarios as a function of the crystal field controling vacancy density. By combining Monte Carlo simulations and recent tensor network renormalization techniques, we map the complete phase diagram, track the evolution of the second transition, and pinpoint the tricritical point where the Kosterlitz-Thouless line terminates. This unusual tricritical point should be relevant for realistic systems.

In the second part, I will focus on the dynamics of active particles in a viscoelastic medium, where memory effects dominate. These effects are modeled through a memory kernel acting on individual particles. While standard active particles undergo motility-induced phase separation (MIPS) at high density and activity, we investigate how memory alters this behavior. Our findings reveal that memory typically hinders MIPS, leads to a reentrant phase transition and gives rise to a distinct low-activity phase separation that differs from conventional MIPS. We rationalize these results by analyzing single-particle trajectories, offering insights into the underlying mechanisms.

''