PerspectiveMaterials Science

Optical Metamaterials—More Bulky and Less Lossy

See allHide authors and affiliations

Science  17 Dec 2010:
Vol. 330, Issue 6011, pp. 1633-1634
DOI: 10.1126/science.1198858

Usually, investigators in materials science have asked: “What properties does a certain new material or structure have?” Now, the inverse problem arises: “I want to achieve certain—possibly unheard-of—material properties. How should the corresponding micro- or nanostructure look?” Examples could be: efficiently blocking acoustic noise due to a highway from a nearby village by a tailored wall, concentrating electromagnetic energy into as-tight-as-possible spaces, or avoiding reflections from a material's surface. The underlying common scheme is wave physics. Material properties that were otherwise unachievable, e.g., negative refraction and cloaking, may eventually be designed into optical metamaterials and photonic crystals. Both require tailoring of the properties (i.e., phase velocity and impedance) of an electromagnetic wave moving through the substance at the local level. In photonic crystals, the phase velocity of an electromagnetic wave moving through the crystal is controlled by tuning the photonic band structure; the impedance is determined by the electromagnetic field distributions throughout the material. In metamaterials, this amounts to tailoring the effective electric permittivity and magnetic permeability. In either case, introducing resonances is the key to controlling the local wave properties. The recent development of advanced fabrication techniques being applied to metamaterials and photonic crystals may lead to realization of such designer materials.

In photonic crystals, the focus has traditionally been on many wavelength-scale building blocks arranged on a dielectric lattice (generalized Bragg resonances) that produce collective resonances in response to an electromagnetic field, whereas researchers working with metamaterials started with densely packed, ideally noninteracting resonances of individual subwavelength-size building blocks (generalized Mie resonances). Mie resonances are especially pronounced for metals. In general, both Braggand Mie-type resonances are advantageous.

In 2007, optical metamaterials (1) were not actually “materials” but were rather metafilms or metasurfaces—planar monolayers of meta-atoms. Today, various truly three-dimensional (3D) optical metamaterial structures have been reported (see the figure) (25). The larger the 3D metal-based metamaterial structures become and the more the “bulk” is approached, the more obvious it becomes that losses must be reduced. Otherwise, a large chunk of metamaterial will simply be black and opaque, rendering it useless. At optical frequencies, most of the losses result from the constituent metal. Thus, avoiding sharp edges for the current flow and/or avoiding nearby resonances can eliminate part of the losses. Superconductors may appear as a remedy for resistive losses, but their small superconducting band-gap frequency inherently prohibits operation at optical or even visible frequencies.

Bulking up.

Metamaterials operating at optical frequencies have developed from planar monolayers of meta-atoms (A) to truly 3D nanostructures (25) (B). Incorporating gain materials (yellow in top box) led to loss reduction (810). (C) Blueprints for isotropic metamaterials based on bottom-up self-assembly (13) or direct laser writing (14) do exist. If all of these aspects could be combined, one of the early dreams of the field, namely, visible 3D isotropic zero-loss negative-index metamaterials, might eventually become reality.


A different approach to reduce losses is to incorporate active gain materials to counteract the losses and achieve negative phase velocities of light. The fact that we are unable to change the past strongly connects losses and phase velocities of light. Combined with the condition of stability, this imposes rather stringent fundamental constraints on what is possible—regardless of the particular design (6, 7). As in any laser, the system becomes intrinsically unstable for macroscopic metamaterial structures and for long times even if the gain exceeds the loss only at a single frequency. Thus, the effective loss can only be zero at a single frequency, for which the phase velocity of light may be negative. For all other frequencies, the losses remain finite. Spontaneous and stimulated emissions will occur and must be considered as intense sources of “noise” if the artificial material shall be used to guide, focus, and/or influence external light. This aspect might force a retraction from the exact gain-equals-loss condition, leaving behind losses that can still be orders of magnitude lower compared to those in the absence of gain material.

Recent experiments on gain have returned to metafilms to avoid tackling both the bulk and the loss challenges simultaneously. Only partial loss compensation has been obtained in experiments with quantum dots (8) and single quantum wells (9), whereas it was possible to compensate for losses completely in the visible range via dyes embedded in epoxy resin and pumped by picosecond pulses (10). The latter experiment and corresponding theory (11, 12) have raised hopes, but it is not clear that this success can be translated to macroscopic 3D photonic metamaterials under steady-state conditions.

Three crucial aspects have already been demonstrated experimentally: (i) operation in the visible spectrum, (ii) truly 3D optical metamaterials, and (iii) loss-free metamaterials. Another crucial aspect for many applications that has only been achieved theoretically so far is isotropy (13, 14). Realizing the ideal negative-index metamaterials, i.e., a 3D isotropic zero-loss negative-index metamaterials operating at visible wavelengths, requires the combination of all these aspects. It is unlikely that compensating loss via optical pumping will meet real-world requirements, whereas injecting an electrical current into a semiconductor through two wires coming out of a chunk of metamaterial is more practical. Once all of this has been achieved, the next frontier toward commercialization is bringing the fabrication cost down, e.g., with nanochemistry-based bottom-up self-assembly approaches (13). Another interesting though unexplored avenue is replacing the lithographically defined meta-atoms by tailored large molecules. Ideally, these molecules could be assembled to 3D molecular crystals.

References and Notes

  1. We thank M. Decker for preparing the figure. Supported by the European Union FET project PHOME (contract 213390); by Ames Laboratory, U.S. Department of Energy (Basic Energy Sciences), under contract DE-AC02-07CH11358; by Office of Naval Research and Air Force Office of Scientific Research–Multidisciplinary University Research Initiative grants (C.M.S.); and by projects CFN A1.5 and METAMAT (M.W.).

Navigate This Article