Symphotic devices: volumetric electromagnetic metamaterials for information processing

Abstract

The design of optical and electromagnetic devices has been an active topic of research for centuries. Despite this, designing devices that perform many functions concurrently and with high efficiency is still a major challenge to this day. The ability to construct optics encoding a large number of functionalities at near-perfect efficiency would have profound repercussions in many fields of basic science, physics, and engineering, from optical and quantum computing to memory storage, scientific instrument design, computer vision, and sensing.In recent years, the advent of metamaterials and the progress of fabrication techniques has made it possible to construct artificial media with exquisite, subwavelength control over their scattering properties. This vast design space has yet to be exploited fully, because the development of specialized design techniques has lagged behind the availability of new materials and properties.I contribute to the field of electromagnetic design by exploring what limitations there are on the functionalities that may be achieved with electromagnetic devices, and how the characteristics of metamaterials can be used to formulate a new design method to encode a large number of functions within the same object. This design technique, called the symphotic method, combines the principles of metamaterial, variational, and adjoint-state design to yield multifunctional devices that encode tens of operations at efficiencies exceeding 90\%. In this dissertation, after proving a fundamental information-theoretical limit on the capabilities of electromagnetic scattering devices, I derive the symphotic method, demonstrate its use, corroborate analytical results with experimental evidence, and present examples of application-oriented devices, such as detectors for autonomous vehicles, multifunctional superresolving microscopes, and portable compressive spectrophotometers.