Dissipation and control in microscopic nonequilibrium systems

Quantifying the flow of energy, entropy, and information within and through nonequilibrium systems remains a central challenge in understanding the microscopic physics of biological systems. Over the past two and a half decades, parallel developments in the fields of theoretical stochastic thermodynamics and single-molecule experiments have made tremendous steps towards this end, advancing our understanding of the fundamental physical limitations and constraints faced by biological systems in vivo. Central in this focus are molecular machines: nanoscale protein complexes which interconvert between different forms of energy to perform useful functions to the cell. While single-molecule experiments on molecular machines have predicted impressively high efficiencies, much is still unknown about their performance in vivo. In this thesis we build upon these primitives, largely by making use of near-equilibrium phenomenological models to simplify and make tractable the problem of quantifying dissipation in molecular machines and predicting the operational modes which are imperative to minimizing their dissipation. By exploring the relevance of near-equilibrium models in the experimental investigation of a DNA hairpin, we find that such an approach can provide utility in understanding the strategies to reduce dissipation in nonequilibrium processes. However, single-molecule manipulations are significantly separated from the in vivo dynamics of molecular machines, and thus for the remainder of the thesis we expand upon this approach in various ways, generalizing the existing theoretical framework to more closely parallel the dynamics of molecular machines. By incorporating the inter-system feedback present in molecular machines, we find that familiar intuitions about how excess work and entropy production are related break down. Finally, we derive a phenomenological expression for the energy flows communicated within the components of a mechanochemical molecular machine. Ultimately, our analysis shows that intersystem feedback can lead to nonvanishing energy flows which are the manifestation of a Maxwell demon in the molecular machine itself.