UC Santa Barbara

Since the invention of the electronic integrated circuit and planar process manufacturing, computers and other miniaturized electronic devices have become integral to many aspects of human life and nearly every industrial undertaking. However, because electrons cannot travel efficiently over large distances, it is necessary to use electromagnetic waves for these devices to engage with the wider world. While radio waves and microwaves offer up to many GHz of bandwidth, some applications, such as data transmission, benefit from the greater bandwidth available at optical frequencies, while other applications, such as spectroscopy, require particular frequencies in the hundreds of THz. To facilitate these applications at the scale of integrated circuit manufacturing, light is produced with semiconductor lasers and manipulated by various secondary opto-electronic elements.In recent years, thin film lithium niobate (TFLN) has become a favorite platform for planar-process manufactured optical elements, especially modulators and frequency doublers. However, lasers are generally manufactured separately and combined with lithium niobate via hybrid integration techniques, adding considerable drag to an otherwise streamlined manufacturing process. In particular, short wavelength lasers have not previously been heterogeneously manufactured on TFLN. This thesis demonstrates flexible heterogeneous integration of several types of GaAs-based semiconductor lasers on TFLN, enabling high throughput integrated manufacturing of electrically driven TFLN systems. Coupling is achieved with on-chip facets and adiabatic tapers. The platform is then used to produce spectroscopy-quality tunable lasers, which incorporate a wide range of passive components. (In an appendix, the coupler designs and the flexibility of the heterogeneous procedure are further validated by producing heterogeneous lasers on low-confinement silicon nitride.) Among other software systems, the designs are supported by eigenmode computations via EMode Photonix, and extensive supporting python scripts are described and presented. Finally, preliminary results are reported for the application of fully integrated frequency doubling on TFLN.