Plasmonics and metamaterials: looking beyond gold and silver

The science of light has outgrown conventional optics and expanded its frontiers into lasers, optical communications, and beyond. Plasmonics is one such frontier that enables the squeezing of light beyond the diffraction limit (about half a wavelength), thereby allowing many novel applications such as nanoantennas, sub-wavelength waveguides, and nanolasers.¹ The advances in plasmonics have been greatly augmented by two revolutionary fields, metamaterials and transformation optics.² While the latter is a paradigm that establishes principles and techniques for focusing and channeling light with unprecedented control, the former is a technology that enables the practical implementation of these ideas. Metamaterials are made of natural materials such as metals and dielectrics that are patterned to sizes much smaller than the operating wavelength, and arranged in special geometries to yield an artificial material with extraordinary responses to electromagnetic excitation. These engineered metamaterials enable a host of novel functions such as optical magnetism, a negative index of refraction, optical cloaks, and sub-diffraction imaging.³

Taken together, plasmonics, metamaterials, and transformation optics provide a new toolset to control the flow of light at the nanoscale. However, the effective use of these tools in any real applications or technologies would not be possible without corresponding advancements in nanofabrication and materials technology. Nanofabrication has now matured to such an extent that many metamaterial geometries can be produced with relative ease. However, a more fundamental hurdle is finding the right materials to enable effective, real-world applications of plasmonics and metamaterials principles.⁴

Metamaterials are commonly composed of metallic and dielectric constituents (see Figure 1). In the optical range, dielectrics can be loss-free and exhibit a positive dielectric constant (permittivity) in the 1–10 range. However, metallic components are invariably accompanied by losses that can be severely detrimental to a metamaterial device's performance. Owing to their relatively lower losses, noble metals such as gold and silver are conventionally used as metallic components in metamaterial devices. However, the losses even in these metals are still too high for efficient optical-range applications. Another problem with metals is that their negative permittivity magnitudes are too large and cannot be tuned. For the best metamaterial designs, the metal constituent's permittivity should be negative and of the same order as the dielectric constituent. In reality, however, metals have permittivity values on the order of −100 in the optical range. Thus, conventional metals are not suitable for many metamaterial applications. As a result, we must turn to alternative plasmonic materials to replace metals.⁵

Figure 1. Metamaterials are made out of natural materials such as metals and dielectrics that are patterned to sizes much smaller than the operating wavelength and arranged in special geometries to provide extreme control over their response to light. The metallic components often limit performance owing to undesirable properties such as high losses. Unconventional metals such as semiconductor-based materials offer better alternatives to their conventional counterparts. The bottom panel shows various plasmonic materials located in parameter space. An ideal plasmonic material would provide zero interband losses, the highest possible mobility, and a carrier concentration of 10^21-22cm ³. Sb: Antimony. Bi: Bismuth. Si: Silicon. Ge: Germanium. III-V: III-V compounds. ZnO: Zinc oxide. In₃O₂: Indium oxide. Ag: Silver. Cu: Copper. Au: Gold.

Materials that are less metallic than noble metals (they have lower concentrations of free carriers) are good candidates for many metamaterial and plasmonic applications. In the near-IR range, for example, heavily doped oxide semiconductors are good candidates (see Figure 1).⁴ These materials possess lower carrier concentration and result in weaker interaction with light. The weaker interaction reduces loss, but still provides an overall metallic property.⁶

Heavily doped oxide semiconductors or transparent-conducting oxides (TCOs) are good candidates to replace metals in the near-IR region. The optical losses of TCOs can be more than four times smaller than those of conventional metals. Moreover, unlike metals, TCOs allow control of their optical properties through doping. Another important difference is that TCOs are compatible with standard silicon nanofabrication processes, while gold and silver are prohibited in industrial semiconductor cleanrooms. Notably, TCOs can be processed by conventional lithography and etching. Also, TCO deposition can be achieved by techniques such as laser ablation, sputtering, and solution processing. Therefore, TCOs are ideal substitutes for metals in near-IR metamaterial designs: see Figure 2(a).⁷ Recently, the operation of a high-performance, negative-refracting slab made out of zinc oxide and aluminum-doped zinc oxide (a transparent conducting oxide) was demonstrated in the near-IR range: see Figure 2(b).⁸ The performance of this metamaterial slab was about three orders of magnitude higher than that of conventional metal-based devices.

Figure 2. Hyperbolic metamaterials are special anisotropic metamaterials that enable applications such as sub-diffraction imaging and negative refraction. They can be fabricated by stacking sub-wavelength-thick, alternating layers of a metal and a dielectric. (a) Structures based on silver and gold show performance nearly three orders smaller than that of oxide-semiconductor or titanium nitride-based devices.¹⁰(b) Negative refraction in a hyperbolic metamaterial made of aluminum-doped zinc oxide (AZO), an unconventional metal. ε_k, ε_⊥: Permittivity of the lamellar-layer stack parallel and perpendicular to the layers, respectively. GZO: Gallium zinc oxide. AlN: Aluminum nitride. TiN: Titanium nitride. Al₂O₃: Aluminum oxide. TM: Transverse metallic mode. TE: Transverse electric mode. θ: Angle of incidence. k_inc: Wavevector of the incident light.

In the visible range, however, TCOs cannot be used because their carrier concentrations are too low to provide metallic properties in this range. Instead, other materials such as intermetallics and ceramics are useful. In particular, transition-metal nitrides can be good substitutes for metals in the visible frequencies.⁹ Unlike novel metals, these nitrides are compatible with CMOS fabrication, and metamaterial devices based on them can perform comparably to or better than conventional metal-based devices.¹⁰

Advances in materials have always accelerated technological progress. An outstanding example in recent decades is silicon technology. Developments in plasmonics and metamaterials could be similarly transformative. In a step toward addressing this important goal, we found that alternative plasmonic materials or unconventional metals are superior to conventional metals in the optical range. Specifically, we showed that oxide semiconductors are better than metal for hyperbolic metamaterial applications and could prove equally advantageous for other uses, such as cloaking and light funnels. Metal nitrides represent an additional promising technology. Our future work will focus on fabricating plasmonic and metamaterial devices using new material components and demonstrating their efficacy.