Probing phase transitions in active systems with cluster tomography

Phase transitions in active and other non-equilibrium systems can be challenging to study using the standard tools of statistical physics. However, phase transitions can often be described in terms of the percolation properties of appropriate geometric clusters, and cluster formation is a common feature of many active systems. Using motility-induced phase separation as a paradigmatic example, we show how to efficiently locate and characterize first and second-order phase transitions in active systems by studying one-dimensional cross-sections of emerging clusters. This approach—known as cluster tomography—enables us to measure some critical exponents without explicitly introducing system-specific order parameters or mapping to an equilibrium system. Cluster tomography is an easy-to-implement yet powerful approach for studying higher-order correlations in non-equilibrium systems, making it a useful addition to the standard toolbox of statistical physics.