A strong intrinsic signal is advantageous over labeling for optical detection of nanoparticles. Intense scattering and absorption by the surface plasmon resonance, which exceeds molecular cross sections, provides a direct method for visualizing noble metal nanoparticles. While two-photon luminescence in gold nanoparticles yields a strong signal, one-photon luminescence is generally regarded to be much weaker and has seldom been employed for optical nanoparticle detection. In this article we investigated one-photon luminescence of gold nanospheres and nanorods using single particle spectroscopy with excitation at 514 and 633 nm. We characterized the polarization dependence, determined the quantum yield, and present a mechanism describing one-photon luminescence. Our results suggest fast interconversion between surface plasmons and hot electron–hole pairs and show that the luminescence occurs via emission by a surface plasmon. Using the information obtained from the single particle studies, we were able to successfully employ one-photon luminescence for correlation spectroscopy measurements and to correctly interpret auto- and cross-correlation functions, which were used to determine the hydrodynamic sizes of several gold nanoparticle samples and to extract rotational dynamics of nanorods. Because of the difference in size dependence for one-photon luminescence compared to scattering, luminescence correlation spectroscopy of metal nanoparticles is advantageous as it is not as strongly affected by the presence of larger nanoparticles or aggregates. This was verified by measuring luminescence as well as scattering correlation traces for a mixture of nanoparticles containing 98% 57 nm and 2% 96 nm gold nanospheres.

In 1908, Gustav Mie solved Maxwell's equations to account for the absorption and scattering of spherical plasmonic particles. Since then much attention has been devoted to the size dependent optical properties of metallic nanoparticles. However, ensemble measurements of colloidal solutions generally only yield the total extinction cross sections of the nanoparticles. Here, we show how Mie's prediction on the size dependence of the surface absorption and scattering can be probed separately for the same gold nanoparticle by using two single particle spectroscopy techniques, (1) dark-field scattering and (2) photothermal imaging, which selectively only measure scattering and absorption, respectively. Combining the optical measurements with correlated scanning electron microscopy furthermore allowed us to measure the size of the spherical gold nanoparticles, which ranged from 43 to 274 nm in diameter. We found that even though the trend predicted by Mie theory is followed well by the experimental data over a large range of nanoparticle diameters, for small size variations changes in scattering and absorption intensities are dominated by factors other than those considered by Mie theory. In particular, spectral shifts of the plasmon resonance due to deviations from a spherical particle shape alone cannot explain the observed variation in absorption and scattering intensities.

No abstract

We have evaluated the effect of varying three key parameters for Fluorescence Correlation Spectroscopy analysis, first in the context of a one species/one environment system, and then in a complex system composed of two species, or conversely, two environments. We establish experimentally appropriate settings for the (1) minimum lag time, (2) maximum lag time, and (3) averaging times over which an autocorrelation is carried out, as a function of expected diffusion decay time for a particular solute, and show that use of appropriate settings plays a critical role in recovering accurate and reliable decay times and resulting diffusion constants. Both experimental and simulated data were used to show that for a complex binary system, to extract accurate diffusion constants for both species, decay times must be bounded by adequate minimum and maximum lag times as dictated by the fast and slow diffusing species, respectively. We also demonstrate that even when constraints on experimental conditions do not permit achieving the necessary lag time limits for both of the species in a binary system, the accuracy of the recovered diffusion constant for the one species whose autocorrelation function is fully time-resolved is unaffected by uncertainty in fitting introduced by the presence of the second species.