The first part of this thesis focuses on verifying the quantum nonequilibrium work relation in the presence of decoherence. The nonequilibrium work relation is a generalization of the second law of thermodynamics that links nonequilibrium work measurements to equilibrium free energies via an equality. Despite being well established for classical systems, a quantum work relation is conceptually difficult to construct for systems that interact with their environment. We argue that for a quantum system which undergoes decoherence but not dissipation, these conceptual difficulties do not arise and the work relation can be proven similarly to the case of an isolated system. This result is accompanied by an experimental demonstration using trapped ions.

The second part of this thesis examines the relationship between quantum work and coherence by constructing analogous quantities in classical physics. It has recently been shown that quantum coherence can function as a resource for work extraction. Furthermore, it has been suggested that this property could be a truly quantum aspect of thermodynamics with no classical analog. We examine this assertion within the framework of classical Hamiltonian mechanics and canonical quantization. For classical states we define a so called non-uniformity measure and show that it is a resource for work extraction similar to quantum coherence. Additionally, we show that work extracted from non-uniformity and coherence agree in the classical limit. This calls into question the idea that coherence qualitatively separates classical and quantum thermodynamics.

The final part of this thesis explores the connection between decoherence and adiabatic (quasistatic) driving. This topic is inspired by an experiment where it was seen that strong dephasing suppressed energy level transitions. Using a perturbative method we investigate this mechanism in the regime of small to moderate decoherence rate and ask if decoherence can help suppress energy transitions when compared with an adiabatic process without decoherence. We find that strategies that include decoherence are inferior to those where decoherence is absent.