|Ph.D Student||Bunin Guy|
|Subject||Fluctuations away from Equilibrium|
|Department||Department of Physics||Supervisor||Professor Yariv Kafri|
|Full Thesis text|
Equilibrium statistical mechanics offers a theoretical framework that has been invaluable to understanding phenomena in practically every area of physics. Yet much of the dynamics of complex systems lie beyond the scope of this theory. In this Thesis we study how quantities, such as energy or number of particles, fluctuate during dynamical processes. Processes studied include thermally isolated systems driven by external fields, transport of heat through a system, and driven open systems. We focus on two regimes. In the first, a system is composed of sub-systems, where the energy transfer between the subsystems is slow, compared with internal relaxation times within the subsystems. Using this time-scale separation, universal relations between the average and variance of the energy transfer are derived. These relations are independent of the details of the sub-systems' internal structure and the coupling between subsystems.
We also study in detail the dynamics of the energy density field in a heat transport setting, where heat flows through a system. The currents induce generic long-range correlations even above any phase transitions. The analog of the free-energy, known as the large-deviation function, is then non-local and very hard to compute. We develop numerical methods to evaluate this quantity in any model of diffusive transport. We show that this quantity can be understood using simple approximations and toy models, which capture both qualitative and quantitative aspects of its behavior. We show, for the first time for boundary-driven diffusive systems, that the large deviation function can exhibit singular features, whose structure is analogous to phase transitions.