Waveguiding by subwavelength arrays of active quantum dots

Guiding light on a nanometer scale without extensive loss is one of the major obstacles to ultra-high-density photonic integrated circuits. Due to the diffraction limit, the width of a passive waveguide must be on the order of the transmitted wavelength or larger to achieve confinement. This restriction necessitates micrometer-wide structures for most optical components.

Within the last decade, a number of researchers have made gold- and silver-nanoparticle¹ and gold layered-slot² wave-guides, which rely on the plasmonic behavior and negative dielectric property of metals to transfer an optical signal. Photonic-crystal-based devices, ³ whose loss is less than that of plasmonics, may also confine light within a space of several hundred nanometers in a matrix of holes that provides a photonic band gap. However, loss is not actively mediated in any of these cases, and confinement of the field to the actual device footprint and possible integrated-fabrication methods are not optimal.

Our approach to reducing loss and shrinking the structure is to form a waveguide from gain-enabled quantum dots (QDs) that are deposited via self-assembly in a subwavelength template.⁴ The operation, depicted in Figure 1(a), uses pump light to excite electrons from the valence to the conduction band in the QDs. A light signal introduced at one edge via a fiber probe propagates along the device to promote recombination of electron-hole pairs, resulting in stimulated emission. The photons cascade one after the other and travel downstream, leading to increased throughput. Transmission is tempered only by the coupling efficiency between adjacent QDs, and is detected at the output edge by another fiber probe.

We have used two different self-assembly chemistries to fabricate such nanophotonic QD waveguides.⁵ The first exploits complementary strands of DNA as programmable elements to deposit QDs in areas defined by the DNA sequence. The second method is a two-layer self-assembly process that enables rapid prototyping and leads to higher QD packing density on the substrate. We check assembly of the device with fluorescence and atomic force microscopy. The fluorescence micrograph in Figure 1(b) shows 500nm waveguides, both isolated and in pairs with 200 and 500nm spacing.

We tested the device, in the configuration described above, with both pump and signal lasers. To determine the net contribution of the waveguide, we ramped the pump light to vary the absorption or gain of the QDs while the signal light was toggled on and off. Subsequently, the sample was moved out from underneath the test probes, and we ran control tests on the bare substrate. After performing data analysis, we found that the output power from the waveguide, relative to the substrate transmission, increased with pump power. The curves in Figure 2 demonstrate the expected trend for both straight and 90°-bend waveguides with improved throughput due to the QDs.

To further qualify our waveguide for integration with existing optoelectronics, we are measuring transmission through structures of varying length to find loss as a function of distance. We will also investigate crosstalk between adjacent devices. Comparing the results with conventional dielectric waveguides will test whether the near-field energy transfer between quantum dots in our device reduces the crosstalk to neighboring wave-guides and increases the overall direct transmission. The characterization of the device will demonstrate its viability as a cornerstone for nanoscale photonic integrated circuits.