Anisotropic impedance surfaces have been demonstrated to be useful for a variety of applications ranging from antennas, to surface wave guiding, to control of scattering. To increase their anisotropy requires elongated unit cells which have reduced symmetry and thus are not easily arranged into arbitrary patterns.

We'll discuss the limitations of existing patterning techniques and explore options for generating anisotropic impedance surfaces with arbitrary spatial variation. A new patterning technique, called a point-shifting method combined with a Voronoi cell generation technique, will be explained that allows a wide range of anisotropic impedance profiles. This approach can be used to produce patterns which include highly elongated cells with varying orientation, and cells which can smoothly transition between square, rectangular, hexagonal, and other shapes with a wide range of aspect ratios. Various applications fabricated by point sifting method will be introduced. Also, for validating designed surface impedances we'll discuss a new method to extract surface impedances from arbitrarily shaped patch cells using the moment of inertia equations.

Photon spin has received great interest in the recent decades for many applications such as encoding quantum information and spin-filtering. However, very little is known about controlling the direction and properties of the spin. It was recently found that surface waves with evanescent tails possess an inherent in-plane transverse spin which is dependent on the propagation direction.

In this dissertation, we investigate different 1D, 2D and 3D designs that support strong spin-dependent propagation. Starting with a 1D C-shaped waveguide, we show that the spin-density can be enhanced through dipole-to-dipole coupling resulting in highly directional wave propagation. We then show spin-dependent wave splitting in 2D metasurface by engineering the equifrequency contours. We demonstrate the possibility of steering the surface wave along curved paths. We also introduce a new type of surface wave called a chiral surface wave which has two transverse spins, an in-plane one that is inherent to any surface wave and an out-of-plane spin which is enforced by the design due to strong $x$-to-$y$ coupling and broken rotational symmetry. We show that the two transverse spins are locked to the momentum providing a highly confined spin-dependent propagation. Similar chiral modes can be induced in 3D structures by introducing screw dislocation defect in a diamond photonic crystal.

Our study opens a new direction for enhancing and controlling the spin properties of electromagnetic waves through engineering the symmetry of shapes in 1D, 2D and 3D. This provides an additional degree of freedom to control the propagation direction as well as the transverse spin of electromagnetic waves.

Nonlinear active metamaterial surfaces are constructed of planar periodic engineered structures in the sub-wavelength (lambda/4) scale on which the nonlinear circuit components have been populated. Unusual electromagnetic properties of the metamaterials derived by the resonant behavior of the constitutive unit cells have produced remarkable effects such as negative index of refraction, cloaking, and an electromagnetic band gap due to high impedance, while the implementations are restricted in bandwidth and polarization.

The added nonlinearity from the nonlinear components can give the degree of freedom to achieve the unique and useful functionalities which could not be realized with linear and passive metamaterials. This thesis studies the theory, characterization, and capability of nonlinear active metasurfaces. The primary application of the invented metasurfaces is focused on exploring a new type of microwave absorbing structure, and the other unique electromagnetic properties.

The state-of-the-art nonlinear circuits deployed on the metasurface offer the adaptive capabilities to tune the inherited electromagnetic properties. First, the theoretical limitations of the linear lossy coating are addressed to establish the necessity for the need for advanced absorbing structures. Subsequent chapters introduce the invented nonlinear metasurfaces with the different tunabilities, specifically the switchable metasurfaces that selectively absorb high power signals to avoid destructive interference to sensitive electronic devices. The self-tuning metasurfaces adaptively tune the resonance frequencies to match the frequencies of the incident waves for broadband absorbing bandwidth. The reconfigurable impedance surface has octave tunability maintaining the artificial magnetic conductor property to support an extreme broadband antenna system, and the omni-directional metamaterial surface to response all-directional incoming waves to minimize unexpected scattering effects in oblique angles.

This dissertation discusses how generative-type artificial neural network (ANNs) enables the real-time inverse design process of reconfigurable EM megastructures. Specifically, we demonstrate two designs: 1. a 2-D beamformer with rotatable dielectric slabs and 2. a conformal EM coating responding to free-form design goals on reflection pattern in a dynamic environment, the latter being an ultimate ambition of EM scattering control and nearly impossible to realize with conventional methods. These two examples demonstrate the superiority of ANN methods in dealing with high-nonlinear EM design problems requiring fast responses. Furthermore, the proposed data-driven free-form inverse-design approach can be accommodated for other science/engineering tasks.

There has long been speculation of inter-cellular communication through weak electromagnetic (EM) fields playing a functional role in biological systems. The predominant theory for how a cell could ever generate electric fields is mechanical vibration of charged or polar biomolecules such as cell membranes or microtubules. The challenge to this theory is explaining how high-frequency vibrations would not be over-damped by surrounding biological media. As many of these suspected resonators are too large for atomistic molecular dynamics simulations, accurately modeling biological resonators remains an ongoing challenge. Starting with energy transfer expressions from coupled-mode theory, we derive expressions for the minimum quality factor required to sustain communication for both near- and far-field interactions. Next, we analyze microtubules as candidate ``antennas", using a custom electromagnetic simulation technique to assess the field strength from vibrating microtubules. Building off the difficulties of simulating the microtubule in conventional molecular simulation techniques, we develop our own method of simulating electrodynamic interactions in biological systems and implement it for a model of linear DNA. Lastly, we discuss experimental work with our collaborators in the biological sciences. We developed a series of band pass filter frequency selective surfaces to narrow down the frequency of signaling in barrier experiments, then followed up with listening and stimulation type experiments in the biologically active frequency spectrum. Overall, our modeling suggests that any potential biological EM signaling could not overcome the damping from surrounding viscous media, at least based on identified biological resonances. Biological resonators contain insufficient charge or dipole moments to produce significant fields to overcome background thermal energy. Electrodynamic interactions seem to require electric fields or velocities unrealistic in biological systems. Our experimental results are mixed in their findings, giving indirect evidence of biological signaling, but providing no direct measurement of EM signals.

Conversion of photons to electrical energy has a wide variety of applications including imaging, solar energy harvesting, and infrared detection. The coupling of electromagnetic radiation to free electron oscillations at a metal interface results in interesting properties including enhanced electric fields tightly confined to the surface. Taking advantage of this nonlinear light-matter interaction, this work presents a variety of resonant surfaces optimized for combining electrical and photonic excitations in order to liberate electrons in a vacuum-channel device for applications ranging from enhanced photoemission, optical rectification, and infrared photodetection.