This Ph.D. thesis concerns a very active area of current research in theoretical physics, which is non-equilibrium statistical mechanics. More specifically, the focus will be on the study of the non-equilibrium dynamics and fluctuations in isolated many body systems driven out of equilibrium. We shall consider primarily quantum systems, albeit many of the theoretical techniques that will be used can be applied to both classical and quantum statistical systems. The novel contribution of this work is two-fold. First, we have developed new analytical techniques to study exactly the large-deviation statistics of physical quantities relevant for the description of systems driven out of equilibrium, like the work done in a quantum quench or the time-integrated current characterizing transport of matter, energy or charge etc, discussed in Part I of the thesis. Second, we have singled out a class of experimentally relevant nondisordered quantum systems showing a dramatically slow approach to thermal equilibrium and we have provided quantitative predictions for this behavior, which is discussed in Part II.
The irreversible entropy increase described by the second law of thermodynamics is fundamentally tied to thermalization and the emergence of equilibrium. In the first part of our work (Ref: arXiv.2503.04152), we constructed an isolated gas system model and numerically demonstrated irreversible growth of entanglement entropy caused by erasure of spread non-equilibrium state information. Here, we mathematically prove that for a typical macroscopic system in a non-equilibrium state
∣ϕ0⟩, the quantum state
∣ϕ0′⟩=O^(t)∣ϕ0⟩ will inevitably evolve toward equilibrium. Our work demonstrates that the second law of thermodynamics, and consequently the ergodic hypothesis in statistical physics, can be understood and proven from a quantum information perspective. From this perspective, the second law can be stated as: In typical macroscopic physical systems, the spreading and erasure of non-equilibrium information is inevitable.