📝 SummarXiv

Abstract

We introduce and analyze a class of heat engines composed of interacting units, in which the thermal reservoir is associated with the neighborhood surrounding each unit. These systems can be mapped onto stochastic opinion models and are characterized by collective behavior at low temperatures, as well as different types of phase transitions, marked by spontaneous symmetry breaking and classifications that depend on topology and neighborhood structure. For the case of contact with two thermal baths, equivalent to each unit having $k = 4$ nearest neighbors, the system can be tuned to operate at maximum power without sacrificing efficiency or increasing dissipation. These quantities are related by the general expression $ \beta_2 {\cal P} \eta_c =-J_4 \eta \sigma$ when the worksource stems from different interaction energies. The heat engine placed in contact with more than three reservoirs is more revealing, showing that the intermediate thermal reservoir can be conveniently adjusted to achieve the desired compromise between power, efficiency, and dissipation. The influence of lattice topology (regular and random-regular networks), as well as the distinct ratios between the temperatures of the thermal baths, has also been investigated.