Plasmons are collective electron oscillations in a metal which can be excited by laser light. They are associated with very strong, localized electric fields which can be used as nanoscopic flashlights to investigate naturally weak physical processes. But generating plasmons efficiently relies on precise sample fabrication which has so far been out of reach. The authors take a step toward solving the problem by using atomic layer deposition on etched silicon nanopillars to make arrays of silver disks with controllable gaps between them.
Localized plasmons can be generated in gaps between grains of metal, like silver or gold. The optimal gap size is one to two nanometers, with the field enhancement dropping drastically as the gap increases. Generating these gaps currently involves making many samples, searching them for effective areas, and then performing experiments using those areas only. It’s not cost effective or marketable, and it barely suffices for research purposes.
Most recipes for plasmonic samples involve top-down fabrication. They begin with a smooth, thin layer of metal on a dielectric substrate. Bits of metal are then removed using chemical, optical, electron-beam or ion-beam lithography. But the metal grains can be large compared to the required fabrication tolerances, and the etching techniques don’t have the resolution needed. Ion-beam etching, which gives the best resolution, deposits ions into the sample that can ruin its performance.
Caldwell et al. use top-down fabrication to etch nanopillars from bulk silicon, which is a well controlled process. In a bottom-up process, silver is then deposited on the pillars a single atomic layer at a time until the desired thickness is achieved, just as an artist uses many brush strokes to form a figure. The result is reminiscent of a block of houses after a snowfall: pads of silver form an array, with nearly identical gaps between them.
The authors demonstrate that the gap sizes and circle diameters are fairly reproducible, and they use a Raman-active molecule to test the efficiency of the devices. Raman spectroscopy is a very weak process that can be greatly enhanced by localized plasmon activity. The results show that the enhancement achieved is about ten times better than for similar methods, indicating that the localized plasmons generated by the array are quite strong, and that a large percentage of the gaps contribute to the effect. They repeat the experiment for arrays of circles with different diameters and gaps, and present general trends of the Raman enhancement. The arrays that perform best are those with silver pads that are resonant with the incoming light. Resonance, like that found in musical instruments or lasers, allows the electrons in the metal to efficiently absorb incident laser light, converting it into near-field energy and ultimately driving the Raman process.
Since the plasmonic sample can be generated over large areas reproducibly, the authors believe that this technology should be useful for single molecule detection, fluorescence sensors, and other applications in which localized plasmons are useful.
You must log in to add comments.