Kernel-Phase Interferometry for Super-Resolution Detection of Faint Companions

Filling out the dearth of detections between directly imaged and radial velocity planets will test theories of planet formation across the full range of semi-major axes, connecting formation of close to wide separation gas giants, and also substellar companions. Direct detection of close-in companions is notoriously difficult: coronagraphs and point spread function (PSF) subtraction techniques are significantly limited in separation and contrast. Non- redundant aperture masking interferometry (NRM or AMI) can be used to detect companions well inside the PSF of a diffraction limited image, though the technique is severely flux-limited since the mask discards ~95% of the gathered light. Kernel-phase analysis applies similar interferometric techniques to an unobscured diffraction limited image, simulating the full telescope aperture as an interferometer composed of a grid of subapertures. I have developed a new faint companion detection pipeline which analyzes kernel-phases utilizing Bayesian model comparison. I break open the black box of interferometry by demonstrating the use of this pipeline on archival HST/NICMOS images of nearby brown dwarfs. I refine astrometry of previously known companions and search for new companions, in order to constrain formation models at au scales. I also present contrast curves to demonstrate the strength of this technique at separations inaccessible to classical imaging techniques. Using this method, it is possible to detect companions down to flux ratios of $\sim10^2$—reaching the planetary-mass regime for young targets—at half the classical $\lambda/D$ diffraction limit while using a fraction of the telescope time as NRM. Since JWST will be able to perform NRM and unobscured imaging, further development and characterization of kernel-phase analysis will allow efficient use of competitive JWST time.