KEYWORDS: Data modeling, Sensors, Point spread functions, James Webb Space Telescope, Calibration, Data analysis, Physics, Equipment, Visibility, Tunable filters
The Aperture Masking Interferometer (AMI) on board the James Webb Space Telescope (JWST) has a unique place in observational astronomy as the first imaging interferometer in space, promising highly-precise observations resistant to optical aberrations. While the optical system and Point-Spread Function (PSF) are very stable, the infrared detectors on board suffer from a series of non-linearities – primarily charge migration or the “brighter-fatter effect” that, while challenging for other observing modes, are ruinous to the visibility calibration of the AMI mode. Local nonlinear effects produced cannot be straightforwardly corrected in the Fourier domain. Efforts using the existing pipelines have delivered some improvements, but outcomes remain far from the theoretical photon-noise limit of the instrument. This manuscript presents initial work using a fundamentally different approach: the joint implementation of a differentiable physics model of the optics, and a machine-learned Effective Detector Model (EDM), using dLux. These are trained together end-to-end, by gradient descent using the full ensemble of point-source reference targets so far observed by AMI. We infer highly-precise metrology of the AMI and NIRISS optical systems, a preliminary EDM which restores commissioning data to near-ideal precision, and illustrate initial and final residual noise floors representing the present state of this ongoing project.
The James Webb Space Telescope Aperture Masking Interferometer provides NIRISS with its highest angular resolution imaging mode, an ultra-stable non-redundant masking Fizeau interferometer. Until recently, the precision of its interferometric visibilities has been limited to ~ 1% by systematic uncertainties in its optical state and detector noise properties. Using a data-driven calibration of AMI with a differentiable forwards model, this can be improved by more than an order of magnitude, uniquely enabling high angular resolution science not possible from the ground. We will discuss the pipeline and observing strategies required to achieve this, illustrated with science highlights enabled this way from the first two years of AMI data, and generalizations of this approach to kernel phase interferometry.
Charge migration in infrared detectors such as in JWST leads to a 'brighter-fatter effect', where photoelectrons from bright pixels spill to nearby faint pixels and blur the pixel response function at its finest spatial scales - a limiting noise floor for high angular resolution astronomy. We demonstrate an effective forwards model: a nonlinear convolution predicting the effect on every pixel as a polynomial of the pixels in its neighbourhood, learning the coefficients by gradient descent together with a differentiable model of the point spread function. We apply this to the JWST/NIRISS Aperture Masking Interferometer, inferring an accurate model for the BFE in NIRISS; overcoming the main barrier to precise interferometric observations with JWST; and illustrating a simple path to high-quality BFE calibration in other JWST instruments and infrared detectors in general.
The TOLIMAN space mission confronts the challenge of detecting Earth analogues in the immediate solar neighbourhood by using novel astrometric techniques. This bespoke, low-cost mission will employ a novel optical and signal encoding system, enabling high-precision measurements that typically require larger instruments. Targeting the Alpha Centauri system, TOLIMAN will utilise an innovative diffractive pupil to mitigate the limitations of a relatively modest satellite and payload infrastructure to make measurements at the extreme precisions required. In this work, we describe the design and manufacturing of the pupil, which employs liquid crystal technologies and substrates with low coefficients of thermal expansion, with the goal of making measurements resistant to inevitable optical distortions and aberrations.
The TOLIMAN space telescope is purpose-built to probe our stellar neighbourhood for potentially habitable Earth-like exoplanets. Our novel diffractive pupil design will allow TOLIMAN to detect extremely subtle changes in the positions of stars in a binary system, down to the microarcsecond scale. One of the many challenging factors in the detection of this diminutive astrometric signal is instability in the telescope pointing, known as jitter.
This work demonstrates the capability of mitigating the blurring effects of telescope jitter through a forward modelling approach and a new precise optical positioning system. We utilise ∂Lux – a cutting-edge differentiable optical simulation framework built in Jax by our team at the University of Sydney – to model the effects of telescope jitter on the final image. The demanding stability requirements have also inspired innovative engineering approaches, including the design of a piezo-driven tip/tilt system. This methodology enables us to recover the crucial astrometric parameters despite telescope pointing instability, offering TOLIMAN the unique opportunity to observe exoplanetary signatures with unprecedented precision.
The TOLIMAN mission will fly a low-cost space telescope designed and led from the University of Sydney. Its primary science targets an audacious outcome in planetary astrophysics: an exhaustive search for temperateorbit rocky planets around either star in the Alpha Centauri AB binary, our nearest neighbour star system. By performing narrow-angle astrometric monitoring of the binary at extreme precision, any exoplanets betray their presence by gravitationally, engraving a tell-tale perturbation on the orbit. Recovery of this challenging signal, only of order micro-arcseconds of deflection, is normally thought to require a large (meter-class) instrument. By implementing significant innovations optical and signal encoding architecture, the TOLIMAN space telescope aims to recover such signals with a telescope aperture of only a 12.5cm. Here we describe the key features of the mission: its optics, signal encoding and the 16U CubeSat spacecraft bus in which the science payload is housed - all of which are now under construction. With science operations forecast on a timescale of a year, TOLIMAN aims to determine if the Sun’s nearest neighbour hosts a potential planetary stepping stone into the galaxy. Success would lay down a visionary challenge for futuristic high speed probe technologies capable of traversing the interstellar voids.
The sensitivity limits of space telescopes are imposed by uncalibrated errors in the point spread function, photon-noise, background light, and detector sensitivity. These are typically calibrated with specialized wavefront sensor hardware and with flat fields obtained on the ground or with calibration sources, but these leave vulnerabilities to residual time-varying or non-common path aberrations and variations in the detector conditions. It is, therefore, desirable to infer these from science data alone, facing the prohibitively high dimensional problems of phase retrieval and pixel-level calibration. We introduce a new Python package for physical optics simulation, ∂ Lux, which uses the machine learning framework Jax to achieve graphics processing unit acceleration and automatic differentiation (autodiff), and apply this to simulating astronomical imaging. In this first of a series of papers, we show that gradient descent enabled by autodiff can be used to simultaneously perform phase retrieval and calibration of detector sensitivity, scaling efficiently to inferring millions of parameters. This new framework enables high dimensional optimization and inference in data analysis and hardware design in astronomy and beyond, which we explore in subsequent papers in this series.
∂Lux is a newly developed optical modelling framework deigned to harness the tools underpinning the modern machine learning revolution and directly apply them to optics. Both neural networks and optical systems map an input vector to some output vector employing a series of intermediary linear transformations and nonlinear matrix operations. This isomorphism allows for optical models to be directly constructed within existing automatic differentiation libraries. ∂Lux exploits this relationship harnessing automatic differentiation libraries to create a naively end-to-end fully differentiable optical modelling framework. This may precipitate a paradigm shift in the power and utility of these optical models, opening the possibility to entirely novel algorithms and approaches. This manuscript explores some of the many ways to harness the potential of these codes, particularly focusing on the application example provided by the Toliman space telescope mission.
Although discovery technologies are now populating exoplanet catalogs into the thousands, contemporary astronomy is poorly equipped to find the most compelling exoplanetary real-estate: earth-analog systems within our immediate solar neighbourhood. The TOLIMAN space telescope program aims to develop low-cost, agile mission concepts dedicated to astrometric detection of exoplanets within 10PC, and in particularly targeting the Alpha Cen system. It accomplishes this by deploying an innovative optical and signal encoding architecture that targets the most promising technique for this critical stellar sample: high precision astrometric monitoring. Two pathfinder missions, the first a cubesat slated for 2021 launch, and the second a 10cm space telescope under development at the University of Sydney. We will present an overview of the family of missions and the novel technologies underlying the signal detection strategy.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.