Study of the diffusion of small molecules on catalyst surfaces is of broad general interest, and there have been numerous investigations of surface CO diffusion on Pt under ultrahigh vacuum (UHV) or gas phase conditions. 1-7 Both diffusion coefficients (D CO ) as well as activation energies (E d ) for diffusion have been measured and are of importance in the context of, among other topics, CO hydrogenation in fuel synthesis 8 and CO oxidation in heterogeneous catalysis. 9 The latter topic is also of interest in the context of fuel cell catalysis, but there has been no direct experimental determination of D CO in an electrochemical environment due to problems associated with the presence of the electrolyte. 10 Fortunately, however, NMR methods are not plagued by these problems, [11][12][13] and in this paper, we report the first direct determination of the diffusion constants of CO on Pt in a liquid electrochemical environment, together with the activation energy for diffusion, using the techniques of electrochemical NMR (EC-NMR) 11-14 and selective spin inversion NMR. 7 To determine diffusion constants, we used the "S-shape" pulse sequence developed by Becerra et al. 7 The S-shape pulse sequence (Figure 1) exploits the fact that CO molecules adsorbed on a Pt nanoparticle can have different 13 C resonance frequencies, depending on the angle of CO's unique molecular axis with respect to the external magnetic field. A part of the magnetization is selectively inverted by the first two pulses, and the 13 C spins are then allowed to diffuse to different regions of the nanoparticle during the evolution period T ev . Motion of a CO molecule due to surface diffusion alters the 13 C spin's Larmor frequency (ω), and experimentally, the amplitude M + (T ev ) of the noninverted part of the spectrum is measured for various values of T ev . If only T 1 processes are involved, M + (T ev ) grows back to its equilibrium value independent of T ev , but when CO molecules diffuse, a mixing of inverted and noninverted parts of the nuclear magnetization occurs, leading to an initial decrease in M + (T ev ), which then grows back to its equilibrium value with increasing T ev .To calculate the diffusion rate, we follow the time evolution of a normalized signal amplitude, A + , defined by the following equation, at various T ev :where ∆ is the line width and ω 0 is the frequency where M + (T ev ) has its maximum; λ n and A n are the coefficients from a Fourier series solution that are determined by boundary and initial conditions. 7 D ω (the diffusion coefficient in the frequency domain) is obtained as the sole fitting parameter to eq 2, and D ω can then be converted to D CO , the diffusion constant, using the relation, D CO ) (π 2 d 2 /32Ω 2 )D ω , where d is the average particle diameter, and Ω is the upper bound for diffusion in the frequency domain, as reported elsewhere. 7 NMR measurements were carried out on 13 CO adsorbed onto electrochemically cleaned platinum black in 0.5 M D 2 SO 4 using a home-built 8.47 T NMR spectrometer. 14...