Realistic quantum systems interact with their environment and, as a consequence, may lose their quantum properties. This phenomenon is known as decoherence, and it keeps the many oddities of quantum mechanics at the level of elementary particles. But while doing so, decoherence constitutes one of the biggest hindrances to efficient technologies fueled by quantum mechanics. Hence, it is essential to understand the different mechanisms of decoherence and how to control them.

Recently, reservoir engineering, i.e., manipulating the environmental degrees of freedom and their initial correlations, has attracted a lot of attention as a means to control decoherence. Reservoir engineering allows, e.g., to restore information previously leaked into environment back to open quantum systems—a phenomenon often associated with memory and non-Markovianity.

In this Thesis, we study decoherence and reservoir engineering in the context of linear optical systems, where the polarization degree of freedom of single photons is the open quantum system. We begin with a short introduction to the very basics of quantum theory, from which we gradually proceed to the dynamics of open quantum systems.

The rest of the Thesis is dedicated to the main results of Publications I–VII. We derive the decoherence functions of a biphoton system and show how to control them independently of each other. Using the same methods, we can even reverse the direction of decoherence. This allows us to realize quantum teleportation without the resource qubits being entangled, which we demonstrate also experimentally.

We also consider decoherence occurring in interferometric setups, revealing the interesting effects of which-path-information in Mach-Zehnder interference and photon bunching in Hong-Ou-Mandel interference. Monitoring the open-system dynamics in these setups allows us to estimate different parameters outside the interferometers’ more common working region. As for the interferometric region, we present numerical results implying the possibility of breaking the so-called quantum Cramér-Rao bound, a fundamental lower bound for the sensitivity of parameter estimation.

Finally, we consider parameter estimation from the opposite point of view, i.e., when the decoherence model is not known and we cannot monitor it. We implement our alternative protocol in two experiments and apply the results in snapshot verification of non-Markovianity—a task typically requiring monitoring the open-system dynamics.