All-dielectric metamaterials offer a potential low-loss alternative to plasmonic metamaterials at optical frequencies. Here, we take advantage of the low absorption loss as well as the simple unit cell geometry to demonstrate large-scale (centimeter-sized) all-dielectric metamaterial perfect reflectors made from silicon cylinder resonators. These perfect reflectors, operating in the telecommunications band, were fabricated using self-assembly based nanosphere lithography. In spite of the disorder originating from the self-assembly process, the average reflectance of the metamaterial perfect reflectors is 99.7% at 1530 nm, surpassing the reflectance of metallic mirrors. Moreover, the spectral separation of the electric and magnetic resonances can be chosen to achieve the required reflection bandwidth while maintaining a high tolerance to disorder. The scalability of this design could lead to new avenues of manipulating light for low-loss and large-area photonic applications.

All-dielectric metamaterials utilizing Mie resonances in high-permittivity dielectric resonators offer a low-loss alternative to plasmonic metamaterials. Here we present the demonstration of a single-negative all-dielectric metamaterial, comprised of a single layer of cylindrical silicon resonators on a silicon-on-insulator substrate, that possesses peak reflectance over 99% and an average reflectance over 98% across a 200 nm wide bandwidth in the short-wavelength infrared region. The study is also extended to disordered metamaterials, demonstrating a correlation between the degree of disorder and the reduction in reflectance. It is shown that near-unity reflection is preserved as long as resonator interaction is avoided. Realization of near-unity reflection from disordered metamaterials opens the door to large-area implementations using low-cost self-assembly based fabrication techniques.

No abstract

We present full-band structure calculations of temperature-and wavelength-dependent two-photon absorption coefficients and free-carrier absorption cross sections in GaAs, InP, and 0.92 eV-band gap Ga 64 In 36 As and InP 60 As 40 alloys. The calculated coefficient decreases with increasing wavelength and band gap but increases with temperature. Using detailed band structure analysis, we identify various contributions to the free-carrier absorption in GaAs and InP. Although the free-carrier absorption is found to arise predominantly from hole absorption, we show that direct absorption by excited electrons is possible, leading to an enhanced free-carrier absorption coefficient. This excited state absorption could be exploited to modulate the transmission of light at communication wavelengths ͑of 1.33 or 1.55 m͒ with, for example, the more commonly available 0.8 m diode laser. We further show that the high-intensity transmission calculated with our values of nonlinear parameters in GaAs agrees very well with the measured values.

We present the results of full band structure calculations of Fermi levels, intrinsic carrier densities, and one-photon absorption coefficients in undoped HgCdTe alloys. The full band structure is used in the calculation of majority and minority carrier densities and the minority carrier lifetimes limited by radiative, Auger-1, and Auger-7 mechanisms in both n-and p-doped alloys. The lifetimes we predicted differ substantially from those calculated with widely used analytical expressions, except in cases where the hole density is small. This difference originates from the significantly non-parabolic and anisotropic valence bands, which are of increasing significance as the hole density increases. From a comparison of the calculated and measured lifetimes, we deduce that the lifetimes at low temperatures are limited by the ShockleyRead-Hall (SRH) recombination. We have generalized the SRH expression to include Fermi-Dirac statistics, but we still treat the density, energy level, and cross section as adjustable parameters. We find that the calculated radiative and Auger recombination lifetimes, as well as SRH lifetimes, can fit the measured lifetimes using traps located (a) near the conduction band edge in n-HgCdTe and (b) near the valence band edge for p-HgCdTe. In addition, the movement of Fermi level with respect to the trap level explains the observed temperature-dependence of the lifetimes. We conclude that there is considerable room for improvement in HgCdTe material quality.