First Principle Simulator of a Stochastically Varying Image Plane for Photon-counting High Contrast Applications

Abstract

Optical and near-infrared Microwave Kinetic Inductance Detectors, or MKIDs are low-temperature detectors with inherent spectral resolution that are able to instantly register individual photons with potentially no false counts or readout noise. These properties make MKIDs transformative for exoplanet direct imaging by enabling photon-statistics-based planet-discrimination techniques as well as performing conventional noise-subtraction techniques on shorter timescales. These detectors are in the process of rapid development, and as such, the full extent of their performance enhancing potential has not yet be quantified.MKID Exoplanet Direct Imaging Simulator, or MEDIS, is a general-purpose end-to-end numerical simulator for high-contrast observations with MKIDs. The simulator exploits current optical propagation libraries and augments them with a new MKIDs simulation module to provide a pragmatic model of many of the degradation effects present during the detection process. We use MEDIS to demonstrate how changes in various MKID properties affect the contrast-separation performance when conventional differential imaging techniques are applied to low-flux, short duration observations. We show that to improve performance at close separations will require increasing the maximum count rate or pixel sampling when there is high residual flux after the coronagraph. We predict that taking pixel yield from the value achieved by current instruments of 80% and increasing it to 100% would result in an improvement in contrast of a factor of ∼ 4 at 3λ/D and ∼ 8 at 6λ/D. Achieving better contrast performance in this low flux regime would then require exploiting the information encoded in the photon arrival time statistics.

Probing Photon Statistics in Adaptive Optics Images with SCExAO/MEC*

“…The most prevalent nonanalytic SSD techniques overlap with machinelearning-based approaches for detecting patterns in time-series data such as long short-term memory recurrent neural networks (Hochreiter & Schmidhuber 1997). To facilitate further development of machine-learning-based SSD, the MKID Exoplanet Direct Imaging Simulator (MEDIS) can be used to generate more realistic simulated MKID data sets (Dodkins et al 2020).…”

Speckle Space–Time Covariance in High-contrast Imaging

“…To create the simulated data set, we use a simulator specifically designed for high-contrast imaging with next-generation detectors, such as MKIDs, called MEDIS (the MKID Exoplanet Direct Imaging Simulator), the first end-to-end simulator for high contrast imaging instruments with photon counting detectors (Dodkins 2018;Dodkins et al 2020).…”

The MKID Exoplanet Camera for Subaru SCExAO

“…Simulations with the MEDIS package (Dodkins et al 2020) have been performed to estimate the impact of dead pixels on the raw contrast (before post processing with ADI/SDI/SSD/ etc.) for exoplanet direct imaging.…”

“…The MKID Pipeline will also be undergoing continual improvements which will allow us to more accurately determine the energy of each incident photon and to better remove false counts. For additional details on how future array improvements will affect the performance of MEC, see Dodkins et al (2020).…”