When
Title:
Abstract:
Our future astronomical investigations need to capture the faintest objects with unparalleled fidelity to study their structure and composition. One such example is the direct imaging of Earth-like exoplanets in reflected light. However, these sources are millions to billions of times dimmer and separated by fractions of an arcsecond from their host star. Conventional imaging is insufficient in this regime as diffracted light from the host star completely overwhelms any retrievable signal from an exoplanet. Thus, coronagraphs have been developed to enable high-contrast imaging. These advanced instruments reject on-axis starlight using a combination of pupil and focal plane masks. However, they are incredibly sensitive to residual wavefront errors (WFEs) which create focal plane speckles that are still orders of magnitude brighter than any exoplanet. High-order wavefront sensing and control (HOWFSC) can be used to iteratively solve for a deformable mirror (DM) shape which destructively interferes starlight with these speckles---creating a region of high-contrast known as the "dark hole". However, this dark hole must be maintained over long periods of time to collect enough signal before an exoplanet can be detected at a high level of confidence. In addition, residual WFEs are not static, they vary over time due to various factors on-orbit. Changing thermal gradients and gravity vectors bring the system out of alignment and introduce slowly-varying, low-order WFEs. Meanwhile, pointing errors cause the wavefront at a particular field to sample different regions of the telescope's optical surface errors over time (i.e. beamwalk). Therefore, a suite of wavefront sensing and control (WFSC) solutions is required at multiple levels within the telescope and the coronagraph is vital for maintaining a stable, well-corrected wavefront over a long period of time.
In this thesis, various methods of wavefront sensing, control, and shaping have been developed to help combat these issues. We begin at the telescope front-end, where a multi-field phase retrieval (MFPR) algorithm was developed using a combination of field diversity and algorithmic differentiation. This work demonstrated MFPR as a potential WFSC solution for telescope commissioning and science operations on-orbit using an end-to-end model of a 3-meter space telescope concept. We showed that our MFPR algorithm was primarily limited by guide star signal-to-noise ratio (SNR) and was able to retrieve field-dependent, low-order WFEs with single-nanometer accuracy over the full telescope field-of-view. We also discuss current limitations and potential improvements to the algorithm in future work. Next, we move into the coronagraph itself, where a modulated self-coherent camera (mSCC) was demonstrated on the Space Coronagraph Optical Bench (SCoOB) for HOWFSC. We present the design and fabrication process of a custom Lyot stop for the mSCC. We demonstrated that the mSCC was able to achieve sub-1x10-7 dark hole contrast while the baseline implicit electric field conjugation algorithm (iEFC) used on SCoOB was able to achieve sub-1x10-8 contrast. We discuss current limitations of the mSCC on SCoOB and how pointing instability at the time of the experiments had an outsized effect on the mSCC compared to iEFC. Finally, we introduce pinwheel segmentation---a method of segmenting circular apertures using curved segment edges leading to a "quasi-Airy" point spread function (PSF) which emulates the standard Airy pattern created by a monolithic circular aperture. We created physical optics models for a monolithic circular aperture, 7-segment pinwheel, and 37-segment pinwheel. Using these models, we demonstrated how instrument design and science data post-processing with pinwheel apertures can achieve similar results when compared to a monolithic circular aperture. Specifically, we showed how optimized APLC and PAPLC pupil apodizers could achieve similar contrast and throughput curves for all three apertures when assuming an ideal, monochromatic system. In addition, presented similar levels of centroid and flux error when finding and fitting sources in crowded star field images simulated using all three apertures.
Committee:
Dr. Ewan Douglas (Chair)
Dr. Daewook Kim
Dr. Jared Males
Dr. Kyle Van Gorkom