Coronagraphic Planet Finding with Energy Resolving Detectors

The primary goal of this research was the development and construction of a Microwave Kinetic Inductance Detector (MKID) based integral field spectrograph for optical/near-IR high contrast astronomy applications. At the beginning of the grant period (Fall 2011), optical MKIDs had been demonstrated at the telescope for the first time just a few weeks prior with the commissioning of ARCONS at Palomar observatory. Over the course of the grant the research was essentially split into two branches: one continuing the development of optical/near-IR MKIDs in general, and the other designing the camera that would eventually be known as DARKNESS. These branches then merged in the second half of the grant period when we applied all the lessons learned from our experience with ARCONS to fabricating the first MKID arrays for DARKNESS. In the final year development of ARCONS arrays was discontinued entirely in favor of implementing any design/material changes directly into DARKNESS detectors. Given this duality, the synopsis of research undertaken will be divided into sections along these branches. The Array Camera for Optical to Near-IR Spectrophotometry (ARCONS) was commissioned at Palomar for the first time in August 2011. During this run ARCONS housed a 1,024-pixel MKID array optimized for observations across a 400-1100 nm bandwidth. After that first commissioning run the MKIDs underwent a significant redesign to mitigate the effect of photons interacting with the silicon substrate, and this design has been the base pixel architecture for our optical MKIDs moving forward. ARCONS’ first 2024 pixel array was subsequently commissioned at Lick observatory in 2012, and returned to Palomar with upgraded 2024 pixel arrays in Fall/Winter 2012, 2013, and 2014. Prior to all of these runs I was tasked with creating and simulating new pixel designs, typically incorporating minor changes with the aim of optimizing detector sensitivity and energy resolution, then empirically determining the formulae we used to procedurally generate the thousands of pixels required for the arrays. The methods we developed to rapidly design new arrays allowed us to integrate new ideas into our detector design with each observing run, and were easily expanded to the generation of larger arrays required for DARKNESS. Unfortunately one conclusion from the optical MKID design tweaking was that our energy resolution was being degraded by factors beyond the pixel geometry and other avenues of research would be required to improve it. As part of my visiting technologist experience at JPL I worked with Peter Day’s lab on the integration of their superconducting travelling-wave parametric amplifier with our MKIDs. Initial results from this experiment were very promising, demonstrating that amplifier noise from the HEMT amplifiers used in ARCONS was contributing significantly to our poor energy resolution. However, the MKIDs available for this experiment had too low energy resolution to begin with. While they did show improved energy resolution when paired with the para-amp, they did not surpass the energy resolution achieved by previous pixels being read out by HEMT. This project has stalled slightly while we produce new pixel geometries that incorporate developments from other work (outside the scope of this grant) that will help us to fully probe the amplifiers’ contribution to resonator noise compared to other known dominant sources (such as two level system noise). It is worth mentioning that all development of the optical MKID software pipeline took place during this grant period and involved many members of the Mazin lab. While all members had some involvement in every module of the pipeline, the modules I contributed to most were those for performing spectral response calibration using spectrophotometric standards and for performing short cadence photometry for long time-stream targets. In the course of this development countless data products were produced to both analyze science targets and characterize the instrument on sky. The culmination of this development was a working (though constantly improving) software package for analyzing optical MKID data. All of the experience and software generated in this pipeline development will be directly inherited by DARKNESS and all future optical MKID instruments. Work on DARKNESS began in the first year of this grant with simulations of the Project 1640 coronagraph operating at I and J-band to predict the level of coronagraphic suppression at those wavelengths. Through these simulations, and applying analytical estimates for post processing using the Dark Speckle technique, we found that MKIDs could conceivably improve P1640’s contrast ratios by greater than an order of magnitude at J-band and simultaneously access I-band for the first time. Based on these estimates we proposed for instrumentation funding and received a NSF ATI grant to construct DARKNESS in 2013. Design of the cryostat proceeded rapidly (led by Ben Mazin) while I simulated new MKID designs optimized for a redder bandwidth and simulated DARKNESS’s optics in Zemax. DARKNESS houses a 10,000 pixel MKID array, optimized for 700-1400 nm (covering I and J band) with a plate scale that Nyquist samples the diffraction limit behind P1640. The cryostat was delivered in Fall 2014 and I have been responsible for verification of the cryostat and detectors. As of the end of this grant a prototype array (D-1) has been demonstrated in the lab, showing that the desired detector sensitivity was achieved for DARKNESS’s bandwidth, but energy resolution was lower than expected. This array would suffice for science operations, but we expect our recent improvements in MKID fabrication (see work of NSTRF fellow Paul Szypryt) will generate higher quality arrays before DARKNESS is scheduled for commissioning. The final two years of my Ph.D. dissertation work will be the final in-lab characterization of DARKNESS, telescope commissioning, and first science.