Light-Off in Plasmon-Mediated Photocatalysis

In plasmon-mediated photocatalysis it is of critical importance to differentiate light-induced catalytic reaction rate enhancement channels, which include near-field effects, direct hot carrier injection, and photothermal catalyst heating. In particular, the discrimination of photothermal and hot electron channels is experimentally challenging, and their role is under keen debate. Here we demonstrate using the example of CO oxidation over nanofabricated neat Pd and Au50Pd50 alloy catalysts, how photothermal rate enhancement differs by up to 3 orders of magnitude for the same photon flux, and how this effect is controlled solely by the position of catalyst operation along the light-off curve measured in the dark. This highlights that small fluctuations in reactor temperature or temperature gradients across a sample may dramatically impact global and local photothermal rate enhancement, respectively, and thus control both the balance between different rate enhancement mechanisms and the way strategies to efficiently distinguish between them should be devised.

ports were used to illuminate the sample with a mercury xenon arc light source (Fig. S12) Figure S6. Unscaled Light absorption efficiency in Pd and Au x Pd 100-x alloy nanodisks for x = 10,20,30,40,50. Same data as in Fig. 3 in the main text but without scaling the experimentally measured absorption efficiency spectra to the FDTD-simulated spectra.     S4) and excludes therefore IR-radiation, which is absorbed in the IR filter and UV-radiation, which is blocked by the quartz and glass walls of the reactor tube and pocket, respectively. S10

Microkinetic model
A first principles informed microkinetic model was constructed to investigate the different convergence profiles obtained for the pure palladium and palladium-gold alloy catalysts. Here we outline the first principles calculations and present the kinetic model.

First principles calculations
Four systems were constructed to model the palladium and palladium-gold systems. Pd (111) was used as the model for the pure Pd catalyst. For  Table S1 and Table S2. As seen in Table S3, the model with Pd as a top layer behaves similarly to the pure Pd model, while the mixed alloy has a slightly lower barrier, and the Au top layer was found to have no barrier after taking the maximum shown above. These findings agree with previous DFT results. 10 The free energies of each state were computed by adding entropic contributions from translational, rotational, and vibrational degrees of freedom to the activation energy (see Table   S3), with adsorbate energies calculated in the harmonic limit and gases using the ideal gas S11 approximation. 11 For the empty surfaces, the Au top layer was found to be the most favorable configuration, followed by the mixed top layer, with the Pd top layer least favored (Figure ).
Here, * refers to a free site and the type is either Pd or Au depending on the system. The adsorption rate constants were taken from collision theory, where is the site area (estimated from the van der Waals radius of the atom), is the mass of the adsorbing molecule, is the Boltzmann constant and is the temperature. The B desorption processes were assumed to be in thermal equilibrium, where is the binding energy (see Table S3) and is the gas constant. The oxidation rxn reaction was assumed to have Langmuir-Hinshelwood kinetics, with Arrhenius rate constant, Here, is the Planck constant and is the activation free energy. site fraction was calculated as . The system kinetics were described by three * = 1 -CO -O equations: That is, the reactor was modelled as a continuously stirred tank reactor (CSTR). The differential equations were solved in Python using the SciPy LSODA integrator 12 and the BDF method.
Relative and absolute tolerances of and were specified.  (   Table S4). 0.08 Total system pressure (bar) 1.0 The scaling constant, , which describes the conversion of the site-based rate to a pressurebased rate accounting for the number of catalyst sites in the reactor volume, was chosen to roughly predict the correct maximum conversion observed experimentally for the Pd and PdAu systems. This parameter was necessarily chosen separately for both systems, as it is challenging to obtain an accurate estimate for the number of exposed surface sites in each case.
With a suitable choice of the number of active sites, the Pd(111) model was found to provide good agreement with the experimental profile and light-off behavior of the pure palladium system ( Figure S14, teal line). The reaction light-off occurs as the surface coverage is shifting from CO-dominated to mixed CO and O coverage ( Figure S15). For the alloy model with a Pd top layer, the profiles obtained were very similar to the pure Pd case (Figure S14 and Figure   S15, blue lines), reflecting the similar energetics for this system. For the alloy model with the In general, the models tested here indicate that inclusion of gold in the catalyst surface would tend to reduce both the adsorbate interaction with the surface and the reaction barrier, leading to low surface coverage -thus low conversion -but with the lower barrier reducing the kinetic limitations that are responsible for the light-off profile in the Pd dominated systems. S18