Optical properties of gold nanoshells on monodisperse silica cores: Experiment and simulations

Gold nanoshells (GNSs) on silica cores are widely used in various biomedical applications that need the spectral tunability and controlled absorption/scattering ratio. However, the plasmonic quality of experimental extinction spectra of GNS colloids differs from that predicted by Mie theory. In this work, we fabricated highly monodisperse silica nano spheres to use them further as cores for synthesis of silica/gold nanoshells. Four GNS samples with 1I6-nm core and gold shell thickness ranging from 16 to 34 nm (116/16, 18, 25, 34) were additionally separated in glycerol gradient solutions to obtain fractions with dominant percentage of single particles or aggregates of various sizes. The separated samples demonstrated extinction spectra with a high extinction maximum to minimum ratio about 3. Optical properties of GNS monomers and aggregates with fixed and random orientations were calculated by Mie theory for polydisperse GNSs, by a generalized multiparticle Mie (GMM) theory for aggregates of separated GNSs, and by the finite difference time-domain (FDTD) method for aggregates of overlapped GNSs. The extinction spectra of upper fractions from 116/25 and 116/34 samples are shown to be well described by Mie theory for GNSs with polydisperse shell thickness. However, for as prepared 116/16 sample this approach fails because of strong near infrared (NIR) contribution from GNS dimers and trimers. The formation of such aggregates is due to coupling of silica cores at early stages of nanoshell synthesis, thus leading to peanut structures with overlapped gold shells. We suggested TEM-based ensemble model with single particles and small dimer and trimer aggregates, which gives satisfactory agreement between measured and FDTD simulated spectra in the vis-NIR region. Thus, the proposed synthetic technology produces high quality gold nanoshells, which remarkable optical properties are in good agreement with electromagnetic simulations based on TEM data. (C) 2016 Elsevier Ltd. All rights reserved.