MPI-DS Colloquium: Particle clustering in turbulence: From volcanic ash to stochastic thermodynamics

Turbulent flows are often regarded as efficient mixers, yet inertial particles suspended in turbulence frequently develop strong spatial clustering. This phenomenon plays a central role in a broad range of geophysical and environmental systems, including volcanic ash transport, atmospheric aerosols, cloud microphysics, and sediment dynamics. In this talk, I will first present some recent results on particle clustering in turbulent flows, combining direct numerical simulations with laboratory experiments involving volcanic ash particles. I will then discuss a non-equilibrium statistical framework for particle dynamics based on a coarse-grained description of turbulent velocity fluctuations. Particle velocity increments across scales are described as a Markov process governed by a Fokker–Planck equation. This formulation enables the definition of stochastic trajectory entropy and fluctuation re-lations for inertial particles in turbulence. I will discuss how, in a coarse-grained descrip-tion, unresolved turbulent fluctuations appear as effective heat, or equivalently as the mismatch between energy transfer and useful work along particle trajectories. From this perspective, entropy production quantifies the irreversibility of turbulent transport and provides a natural bridge between multiscale turbulence and particle organization. Building on this idea, I will show recent results indicating that entropy production is strongly corre-lated with preferential concentration and clustering. Different classes of particles self-organize into states that maximize trajectory entropy, suggesting that clustering in turbu-lence can be interpreted as an entropy-selected non-equilibrium process.