UC Santa Barbara

Detection of ultraviolet to near-infrared light is useful for a variety of applications from dark matter searches to biological imaging and astronomy. The performance of these detectors often limits the achievable science goals for an application, so improvements to detector technologies can be transformative. This dissertation focuses on these detector enhancements, emphasizing the requirements for one particular application, exoplanet direct imaging. However, the work done here remains broadly applicable to fields needing highly sensitive sensors in this wavelength range.

Finding and studying the properties of exoplanets orbiting distant stars cantell us much about solar system dynamics and the formation of our own solar system. With sufficiently precise measurements we might also discover the presence of water or even biological processes on these planets. To achieve this goal, we need astronomical instrumentation capable of separating an exoplanet's light from its host star by directly imaging it. For many exoplanets, though, we receive at our telescope only roughly one photon for every billion from the star. Even worse, these two light sources often partially overlap because of the relative closeness of exoplanets to their stars and the diffraction limit set by the finite size of our telescope optics. Therefore, the extreme contrast ratio between the brightness of the star and planet sets the performance requirements for this kind of instrument.

Carrying out this measurement with the traditional semiconductor based sensors can be difficult. They detect the intensity of light at each pixel and introduce excess noise into the system. However, superconducting sensors avoid this limitation because of their extra sensitivity. Each individual incident photon can be resolved making them essentially perfect photon counting detectors, which enables the detection of exoplanets with more challenging contrast ratios than detectable with conventional detectors. The system noise in a superconducting sensor, instead, determines how accurately the photon energy can be measured through the size of the detector response. Because of this extra energy measurement capability these detectors do not need the complicated optical systems typically used to measure an exoplanet's atmospheric spectrum. The accuracy at which we can resolve these planetary spectra determines how much we can learn about a planet and, as such, is the principal metric for the performance of these devices.

The Microwave Kinetic Inductance Detector (MKID) is unique among other superconduct­ing technologies because it allows for the simple readout of tens to hundreds of thousands of pixels, which is a requirement for when the exact location of an exoplanet is unknown. This dissertation focuses on improving the spectral resolving power of MKIDs to make them the superior option for ultraviolet to near­-infrared measurements and, in particular, for exoplanet direct imaging. Chapters 1 and 2 discuss the significance of MKIDs as astronomical detectors and the relevant physics needed to understand their operation. In the following chapters, four separate areas where I have contributed to the advancement of MKIDs are laid out. Chapter 3 covers improvements to the data analysis. A new sensor material for MKIDs is characterized in chapter 4. Chapter 5 shows how the readout scheme can be improved to lower the system noise. Finally, in chapter 6 improvements are made to the detector geometry. With the ad­ vancements considered in this dissertation, both the spectral range and resolving power of MKIDs have been increased by a factor of 3.