Sum frequency generation (SFG) vibrational spectroscopy is used to characterize intermediate species of hydrogenation reactions on the surface of platinum nanoparticle catalysts. In contrast to other spectroscopy techniques which operate in ultra-high vacuum or probe surface species after reaction, SFG collects information under normal conditions as the reaction is taking place. Several systems have been studied previously using SFG on single crystals, notably alkene hydrogenation on Pt(111). In this thesis, many aspects of SFG experiments on colloidal nanoparticles are explored for the first time. To address spectral interference by the capping agent (PVP), three procedures are proposed: UV cleaning, H2 induced disordering and calcination (core-shell nanoparticles). UV cleaning and calcination physically destroy organic capping while disordering reduces SFG signal through a reversible structural change by PVP.

Disordering is a dynamic and powerful way to perform SFG experiments on colloidal nanoparticles. Simply adding H2 (a reactant) to Pt-PVP reduces signal of PVP by >90%, making it possible to resolve intermediates. The selection rules of SFG indicate intensity may be reduced by a decrease in concentration or an increase in disorder. As H2 dissociates on Pt, the bond between PVP and Pt is weakened and PVP is able to rearrange into an unstructured geometry. Disordering makes size-controlled SFG studies possible for the first time because with PVP left intact, the nanoparticles are much less likely to aggregate. As evidence that a chemisorbed reaction intermediate is can be monitored, cyclohexene hydrogenation is performed. When PVP is disordered and the particles are "solvent-cleaned," the distinct C-H stretch of adsorbed 1,4-cyclohexadiene (1,4-CHD) appears at 2765 cm-1. In addition, the mechanism of hydrogenation of 1,3-butadiene (1,3-BD) at 75 °C is characterized using the disordering procedure. For larger Pt-PVP nanoparticles (4.6-6.7 nm), one major pathway is preferred (2-buten-1-yl). However, using smaller nanoparticles (0.9-1.8 nm), two major pathways exist (1-buten-4-yl and 2-buten-1-yl) and coincide with preference for full hydrogenation.

Growth in the aliphatic stretch range following H2 dosing shows UV treatment is not able to produce "naked" nanoparticles. Although outer layers removed quickly, persistent C fragments remain near the Pt surface. The properties of residual fragments of PVP following UV treatment are very mystifying. PVP is known to block active sites on Pt and reduce turnover frequency for ethylene hydrogenation. (After three hours of UV treatment, ethylene hydrogenation increases tenfold.) However, it is shown that for methanol oxidation, residual fragments become impermeable and actually reduce the turnover rate. Under oxidation conditions, the regular capped Pt-PVP nanoparticles are more active than the same particles following PVP removal with UV treatment. As further evidence of this effect, cyclohexene hydrogenation is performed with different level of H2. Residual fragments are permeable with H2 and the 1,4-CHD intermediate appears in the SFG spectrum. However, once H2 is removed, chemisorbed 1,4-CHD completely disappears in favor of molecular (physisorbed) cyclohexene (not bonded to Pt). Even for a strongly bound dehydrogenated intermediate like 1,4-CHD, access to Pt is only possible with H2. Under varying amounts of UV treatment (PVP removal), the mechanism of cyclohexene hydrogenation varies widely among 1,4-CHD, 1,3-cyclohexadiene and pi-allyl. Removing PVP does not simply open equivalent sites but shifts the reaction mechanism in this case.

Using Stöber encapsulation, a variety of PVP-capped catalysts can be coated in a thin porous SiO2 shell (5-10 nm) and used for in situ characterization. Surrounding the catalyst with SiO2 allows PVP to be thoroughly removed without compromising morphology. It is shown that alloying Pt with Sn (with SiO2 shell) allows a new reaction mechanism of CO oxidation to take place which is not limited by surface O2 concentration (as on pure Pt). In addition, 1,3-BD hydrogenation on 4 nm Pt@SiO2, Pd@SiO2 and Rh@SiO2 is studied with SFG. Previous work shows Pd makes only partially hydrogenated products while Pt makes all products. SFG and kinetic experiments indicate multiple pathways (1-butene-4-yl, butan-1,3-diyl and 2-buten-1-yl) are possible on Pt@SiO2 while only one (2-buten-1-yl) is favored on Pd@SiO2. This work is the first SFG reaction selectivity study on core-shell nanoparticles. Using catalysts inside SiO2 shells in dense 2-D films, it may become possible to assess industrial 3-D catalysts at a molecular level.