Picophotonics -- Subatomic Optical Localization Beyond Thermal Fluctuations

TL;DR

This paper demonstrates a novel optical measurement technique that achieves sub-Brownian localization of nanostructures using deep learning and structured light, surpassing thermal fluctuation limits with unprecedented precision.

Contribution

It introduces a deep learning-based method for optical localization of nanostructures with sub-Brownian accuracy, bridging a resolution gap in nanometrology.

Findings

Achieved nanowire position measurement with 92 pm precision.

Demonstrated localization beyond thermal fluctuation limits.

Used structured light scattering and deep learning for non-invasive metrology.

Abstract

Despite recent tremendous progress in optical imaging and metrology, the resolution gap between atomic scale transmission electron microscopy and optical techniques has not been closed. Is optical imaging and metrology of nanostructures exhibiting Brownian motion possible with resolution beyond thermal fluctuations? Here we report on an experiment in which the average position of a nanowire with a thermal oscillation amplitude of ~150 pm is resolved in single-shot measurements with precision of 92 pm using light at a wavelength of {\lambda} = 488 nm, providing the first example of such sub-Brownian metrology with ~{\lambda}/5,300 precision. To localize the nanowire, we employ a deep learning analysis of the scattering of topologically structured light, which is highly sensitive to the nanowire's position. As a non-invasive optical metrology with sub-Brownian absolute errors, down to a fraction of the typical size of an atom (Si: 220 pm diameter), it opens the exciting field of picophotonics.