Mercury is a pollutant of grave concern with well documented neurological and developmental health impacts. Better sensing methodology would improve detection and control of mercury and thus reduce its health burden. Gold nanoparticles provide a sensing medium with potential advantages in sensitivity, selectivity, robustness, and cost over established techniques. Mercury readily adsorbs onto the surface of the gold changing the localized surface plasmon resonance which is measured as a shift in the peak optical absorbance wavelength. This shift is dependent on the mercury concentration and predictable with classical electromagnetism.

This work investigates some of the fundamental relationships driving sensor response. The effects of mass transfer and surface kinetics on mercury/gold nanoparticle adsorption are determined with analytical models and experimental results based on impinging flow geometry. To decouple mass transfer and surface kinetics adsorption, electrical analogy models are constructed and fit to the experimental data. The models can account for variations in flow conditions and surface coatings on the nanoparticles. These models are generalizable to other systems.

Results from these fundamental investigations are used to improve and extend sensor performance. The time response or collection efficiency is optimized depending on system requirements. Using the knowledge gained, the applicability of gold nanoparticle mercury sensors is extended to a fiber optic based system and aqueous detection. Nanorods deposited on the surface of a fiber optic cable have a linear response with concentration and are able to detect mercury down to 1.0 μg/m3. The modification of an established oxidation/reduction scheme for use with the sensor allows for the detection of ionic and organic mercury from water samples which ordinarily would not be reactive with gold nanoparticles. The aqueous sensor was able to detect mercury below the EPA's drinking water limit.