Ph.D Thesis

Ph.D StudentRosenblatt Gilad
SubjectExotic Wave Phenomena at Interfaces of Negative Complex
Electromagnetic Media
DepartmentDepartment of Electrical and Computer Engineering
Supervisor PROF. Meir Orenstein
Full Thesis textFull thesis text - English Version


Light changes its character at the vicinity of nearby matter. It no longer appears as a transverse wave, and may exhibit an arbitrarily short wavelength and slow speed at a given frequency. This light is called the near-field, and its localized nature makes its manipulation a formidable challenge.

Phenomena that allow such manipulation in a tractable manner are vital to applications seeking optical control at the nanoscale. This research concerns such exotic phenomena, those at interfaces of materials with negative electromagnetic responses.

Among these are Surface Plasmon Polaritons, collective oscillations of light and charged plasma at the surface of conductors, appearing as slowly propagating surface waves of short wavelength.

We resolved the physical mechanism by which these waves attain slower speeds. Surface plasmons have a peculiar bi-directional power flow: backward in the conductor and forward outside. We proved that it comes about by a circular motion of power within each optical cycle. The elongated path light takes translates to reduced net propagation speeds, and dictates a slow-light dispersion.

We also studied the interaction of adjacent surface plasmons in layered structures, and derived simple rules to design their seemingly complex dispersion. Surface plasmons interact via either weakly or strongly nonuniform fields, yielding opposite properties. By tuning the reliance of collective oscillations on each interaction type, we attained predefined field symmetries, polarizations, and dispersion. Subsequently, we predicted the formation of Dirac points for plasmonic waves - a phenomenon not yet observed.

Left-handed media are hypothetical electromagnetic media with negative permittivity and permeability. Amazingly, a slab of such a medium would form optical images at absolute fidelity. This slab perfect lens has been pursued ever since left-handed behavior was demonstrated by metamaterials: structures designed to mimic electromagnetic media not found in nature. But it requires lossless materials, and metamaterials have intrinsic Ohmic loss, which severely degrades lensing performance. It therefore became accepted that intrinsic loss fundamentally limits perfect lensing.

Yet we proved that perfect lensing can be done by passive and lossy left-handed media. For instance, by an interface between a dielectric on the source side, and a lossy left-handed medium on the image side. Despite loss, the interface refocuses light emitted by a source onto a perfect image. We proved this analytically by several techniques, including the rigorous non-approximate field derivation for localized excitations.

The irreversible transfer of electromagnetic power to Ohmic loss dictates that not all of the power emitted by the source arrives at the image, but image fidelity remains absolute. Moreover, the power that does arrive must drain there. Hence, to emulate perfect lensing with metamaterials, power outlets must be built into the lens design, possibly as detection elements.

 We found that loss only limits perfect lensing for lens designs in which it translates to resonant feedback. The asymptotically uniform transmission perfect lensing requires is cut-off by such feedback, which generates resonances. Perfect lenses must therefore employ geometries that avoid feedback.

Our findings provide a way to carry perfect lensing into the realm of application, with existing left-handed metamaterials.