Metallic nanostructures exhibit a strong plasmon resonance at a wavelength whose value is sensitive to the charge density in the nanostructure, its size, shape, interparticle coupling, and the dielectric properties of its surrounding medium. Here we use UV-visible transmission and reflectance spectroscopy to track the shifts of the plasmon resonance in an array of gold nanoparticles buried under metal-oxide layers of varying thickness produced using atomic layer deposition (ALD) and then coated with bulk layers of one of three metals: aluminum, silver, or gold. A significant shift in the plasmon resonance was observed and a precise value of ωp, the plasmon frequency of the gold comprising the nanoparticles, was determined by modeling the composite of gold nanoparticles and metal-oxide layer as an optically homogeneous film of core-shell particles bounded by two substrates: one of quartz and the other being one of the aforementioned metals, then using a Maxwell-Garnett effective medium expression to extract ωp for the gold nanoparticles before and after coating with the bulk metals. Under illumination, the change in the charge density of the gold nanoparticles per particle determined from the change in the values of ωp is found to be some 50-fold greater than what traditional electrostatic contact electrification models compute based on the work function difference of the two conductive materials. Moreover, when using bulk gold as the capping layer, which should have resulted in a negligible charge exchange between the gold nanoparticles and the bulk gold, a significant charge transfer from the bulk gold layer to the nanoparticles was observed as with the other metals. We explain these observations in terms of the "plasmoelectric effect", recently described by Atwater and co-workers, in which the gold nanoparticles modify their charge density to allow their resonant wavelength to match that of the incident light, thereby achieving, a lower value of the chemical potential due to the entropy increase resulting from the conversion of the plasmon's energy to heat. We conclude that even the act of registering the spectrum of nanoparticles is at times sufficient to alter their charge densities and hence their UV-visible spectra.
Read full abstract