The self-assembly of monodisperse inorganic nanoparticles into highly ordered arrays (superlattices) represents an exciting route to materials and devices with new functions. It allows programming their properties by varying the size, shape, and composition of the nanoparticles, as well as the packing order of the assemblies. While substantial progress has been achieved in the fabrication of superlattice materials made of nanospheres, limited advances have been made in growing similar materials with anisotropic building blocks, which is particularly true for free-standing two-dimensional superlattices. In this paper, we report the controlled growth of free-standing, large-area, monolayered gold-nanorod superlattice sheets by polymer ligands in an entropy-driven interfacial self-assembly process. Furthermore, we experimentally characterize the plasmonic properties of horizontally aligned sheets (H-sheets) and vertically aligned sheets (V-sheets) and show that observed features can be well described using a theoretical model based on the discrete-dipole approximation. Our polymer-ligand-based strategy may be extended to other anisotropic plasmonic building blocks, offering a robust and inexpensive avenue to plasmonic nanosheets for various applications in nanophotonic devices and sensors.

Propagation characteristics of surface plasmon-polaritons (SPPs) in linear chains of metallic nanospheres (LCMNs) can be found from a dispersion equation, by assuming that either frequency or wave number of a SPP is real. In this paper, we present a comparative study of SPP modes corresponding to the two types of complex solutions for an infinitely long LCMN embedded in a gain medium. We show that even though gain predominantly affects the SPP dispersion obtained with real frequency, both solutions result in the same dispersion and attenuation of SPP modes, when Ohmic losses are almost compensated by gain. In this regime, an analytic expression for the propagation length of SPPs exists, and the SPPs' dispersion is determined by a real equation. We also demonstrate that for a given amount of gain (below the amplification limit of ∼1000 cm −1 ), transversely polarized SPPs attenuate slower than longitudinally polarized SPPs and are, therefore, preferable for the purpose of energy transfer in gain-supplied LCMNs. The transmission windows for SPP modes of different polarizations do not overlap each other, which facilitates realization of LCMN-based plasmonic filters. Our results may prove useful in design optimization of all-optical chips for power-efficient optical supercomputers.

The energy transport properties of plasmonic waveguides can be analyzed by solving the dispersion relation for surface plasmon-polaritons (SPPs). We use this approach to derive an approximate analytical expression for SPP propagation length when the waveguide is composed of linearly arranged metallic nanoparticles, while assuming that metal losses are small or partially compensated by gain. Applied to metal-dielectric (composite) nanospheres, the obtained expression allows us to optimize the performance of the waveguide and arrive at a number of practical design rules. Specifically, we show that SPP attenuation can be minimized at a certain interparticle distance for transverse modes, but gradually grows for both longitudinal and transverse modes with the increase of particle separation. We also show that the two basic methods of supplying gain to the system, i.e., embedding the particles into a gain medium or having a metal-gain composition for the particles, do not perform equally well and the former method is more efficient, but the way the two methods affect depends on the polarization of SPPs. To investigate the role of the nanoparticles' arrangement in determining SPP characteristics, we follow a purely numerical approach and consider a two-segment bent waveguide as an example. Analyzing the waveguide's transmission shows that it behaves in an oscillatory manner with respect to the angle between the two segments and is therefore higher for certain angles than for the others. This suggests that, in the design of waveguides with bends, careful attention needs to be paid in order to avoid bend angles that yield low transmission and to choose angles that give maximum transmission.

We study noise transfer from pump to signal in silicon Raman amplifiers, with particular emphasis on the regimes of strong cumulative free-carrier absorption and heavy pump depletion. We calculate the relative intensity noise (RIN) transfers in copumped and counterpumped amplifiers and provide intuitive explanations for RIN peculiarities. We show that noise transfer at low frequencies may be suppressed by carefully choosing the pump intensity, effective free-carrier lifetime, or amplifier length, but only at the expense of a rise in noise at high frequencies. © 2010 Optical Society of America OCIS codes: 190.4360, 230.0230, 230.4480, 250.4390. Despite several experimental demonstrations of net Raman gain in silicon-on-insulator (SOI) waveguides [1][2][3], the problem of efficient Raman amplification in silicon is far from being fully solved. The main reason for this is the need to minimize nonlinear optical losses, caused by two-photon absorption (TPA) and free-carrier absorption (FCA), while avoiding excessive signal noise. The noise performance is especially important for silicon Raman amplifiers (SRAs), because they operate at high pump powers that result in enhanced transfer of the relative intensity noise (RIN) from pump to the signal. Unlike the challenge of nonlinear losses, which has been much investigated in recent years [4][5][6], the physics of RIN transfer in SRAs is poorly understood. The results of a few recent theoretical studies on this phenomenon [7,8] are not widely applicable, for they are based on strong simplifying assumptions, which may not be met in practice. Specifically, in order to estimate the impact of RIN transfer on the noise figure in SRAs, Sang et al.[7] assumed instantaneous FCA and employed the undepleted-pump approximation. While the first assumption is valid for low-frequency noise components, the second holds only when the signal power remains much smaller than the pump power all along the amplifier. In our recent work [8], we abandoned the undepleted-pump approximation and analytically calculated the low-frequency RIN in copumped SRAs. Although the results allowed us to qualitatively predict the possibility of zero noise transfer under certain conditions, a precise quantitative assessment of this phenomenon is still required. In this Letter, we present a general study of the RIN-transfer problem in SRAs. In our work, we consider both the depletion of the pump and the cumulative nature of FCA, draw an intuitive picture of the physics behind the noise transfer, and present guidelines for RIN minimization.The starting point for our study is the set of partial differential equations that describe the interaction of two cw fields (pump and signal) propagating through an SOI waveguide of constant cross section. These equations relate the pump intensity I p ðz; tÞ, the signal intensity I s ðz; tÞ, and the density Nðz; tÞ of free carriers as [8,9]

Finite-difference time-domain (FDTD) simulations of any electromagnetic problem require truncation of an often-unbounded physical region by an electromagnetically bounded region by deploying an artificial construct known as the perfectly matched layer (PML). As it is not possible to construct a universal PML that is non-reflective for different materials, PMLs that are tailored to a specific problem are required. For example, depending on the number of dispersive materials being truncated at the boundaries of a simulation region, an FDTD code may contain multiple sets of update equations for PML implementations. However, such an approach is prone to introducing coding errors. It also makes it extremely difficult to maintain and upgrade an existing FDTD code. In this paper, we solve this problem by developing a new, unified PML algorithm that can effectively truncate all types of linearly dispersive materials. The unification of the algorithm is achieved by employing a general form of the medium permittivity that includes three types of dielectric response functions, known as the Debye, Lorentz, and Drude response functions, as particular cases. We demonstrate the versatility and flexibility of the new formulation by implementing a single FDTD code to simulate absorption of electromagnetic pulse inside a medium that is adjacent to dispersive materials described by different dispersion models. The proposed algorithm can also be used for simulations of optical phenomena in metamaterials and materials exhibiting negative refractive indices.

We analyze the two different types of complex solutions to the dispersion equation of surface plasmon-polaritons residing in a linear chain of metallic nanoparticles in the presence of gain in the host medium.

No abstract

Gain-assisted propagation of surface plasmon-polaritons along chains of metallic nanoparticles is studied with the coupled-dipole method, the validity of which is justified using finite-difference time-domain (FDTD) simulations.