In this thesis I describe three independent projects that advance the development of broadband quantum cryptography. While each project pertains to a different part of the QKD chain, together they provide key developments in implementing QKD at bit rates that are practical for use in the modern telecommunications infrastructure.

The first project comprises the bulk of the thesis and involves developing a novel source of correlated photon pairs for use in free-space QKD. This source is based on a birefringent semiconductor optical waveguide as a Kerr medium. We demonstrate the feasibility of using birefringent phase-matched four-wave mixing to generate correlated photon pairs. We further propose that, by reversing the process and pumping with conjugate wavelengths, one can use the same effect to produce entangled photon pairs with the same device. These pairs can then be used for QKD to realize the most secure and efficient quantum cryptographic data links.

The second project examines the implications of operating a BB84 QKD protocol at clock rates that are faster than the recovery time of the constituent detectors. We show that operating such systems under conventional protocols results in a security violation that allows an eavesdropper to learn significant information about the key and present a modification to the BB84 protocol that maintains key security at fast transmission rates. This modification to the protocol will become vital to QKD viability as links become faster and clock rates go into the tens of gigahertz. We also demonstrate, rather counterintuitively, that there exists an optimal transmission rate for a QKD system that exceeds the inverse of an individual detector's dead time.

The final project describes a new design for a free-space QKD link that centers around faster silicon detectors. These detectors have a peak quantum efficiency in the visible range, requiring that the system operate at a wavelength that is more susceptible to solar interference. To mitigate this effect, the link is designed around a Fraunhofer line in the solar spectrum where the background solar light levels are reduced by up to 90%. By implementing this system, we expect at least a two-fold increase in the secret key rate, coming ever closer to the goal of a 10 Mb/s QKD system compatible with first-generation ethernet technology.