Characterisation of SemiconductorLasers for use as a Pump Source for Microresonators

Abstract

Optical frequency combs (OFCs) were first developed in the 1990s, initially as a means to count the cycles from atomic clocks. They have since proved very useful as a tool for metrology, sensing and frequency synthesis. John L. Hall and Theodor W. H?nsch would go on to be awarded the 2005 Nobel Prize in Physics for their contributions to the development of OFCs. These first combs were generated in optical fibre by a train of pulses from mode-locked femtosecond lasers. These systems were initially quite bulky, often taking up the entirety of a lab bench, and were therefore unsuitable for practical applications. Further research was therefore necessary in order to miniaturise these systems such that the light source as well as the medium for comb generation could be integrated together on a single chip. An important advancement in this regard has been the generation of OFCs in microresonators. These are structures which have dimensions on the micrometre scale, which is a step towards chip-scale operation. Our research group has designed and tested microring resonators of varying geometries and made from different materials such as AlN and Si3N4 in order to generate OFCs. We have successfully generated octave-spanning frequency combs as well as Kerr-soliton frequency combs within these microresonators. A tunable semiconductor laser (TSL) amplified by an erbium-doped fibre amplifier (EDFA), is used as a pump source. Both of these are bench-top pieces of equipment and are quite bulky, so our attention has now turned towards miniaturising this setup. Another project that is being researched in our group is the slotted lasers project. This is where we have developed arrays of semiconductor lasers that are tunable over the C-band and O-band for use in optical communications, particularly dense wavelength division multiplexing (DWDM). These lasers have micrometre dimensions which are fabricated with surface slots for wavelength discrimination rather than a buried heterostructure. This makes the fabrication process much simpler and efficient. They are perfect for photonic integration but have significantly broader linewidths than the lasers typically used to pump microresonators as well as lower output power. If it can be shown that soliton comb generation is possible using such devices, they would present excellent candidates as an alternative pump source for our microresonators that would aid in development of a chip scale system. This work brings together both of these projects and demonstrates a path towards integrating a microresonator with one of our slotted lasers on the same chip. Firstly, three slotted laser arrays are characterised with each array having a different slot design. The wavelength tunability, side-mode suppression ratio (SMSR), linewidth and output power are all investigated in order to determine which slot design results in a performance best suited for use as a pump source. A slotted laser with wavelength near one of the resonance modes of a microresonator was then selected as a pump source. Despite the broader linewidth of the slotted laser compared with the TSL, it was shown that an octave-spanning Kerr soliton microcomb could be generated using the slotted laser by sweeping the current supplied in order to red-shift the wavelength. Turn-key generation of the soliton microcomb was then demonstrated by increasing the current supplied to the slotted laser in a single step. The soliton comb was reliably produced using both methods, however the soliton existence range was found to be narrower for the slotted laser compared with the TSL and the overall noise was found to be greater when the slotted laser was used. Finally, to demonstrate the generation of microwave frequencies from an OFC, a commercial packaged device based on slots was used to generate a dual comb consisting of a soliton comb and a primary comb. By beating the primary comb lines with the soliton comb lines, microwave frequencies were generated. Once again, the noise of the signals generated by the laser was considerably greater than that of signals produced with the TSL. Successfully generating soliton microcombs with a slotted laser is a promising step for miniaturising the system. However, the EDFA was still found to be necessary when pumping with the slot laser. Future work will need to be done in both improving the output power and linewidth of the slotted lasers. Improving the quality (Q) factor of the resonators would also reduce the required input power from the pump laser.