Supplementary Information

Controlling the CO oxidation rate over Pt/TiO2 catalysts by defect engineering of the TiO2 support

Y. P. Gavin Chua1,2, G. T. Kasun Kalhara Gunasooriya2, Mark Saeys2# and Edmund G. Seebauer1*

1 Department of Chemical & Biomolecular Engineering, University of Illinois - Urbana Champaign, 114 Roger Adams Laboratory, 600 S. Mathews Ave, Urbana, IL 61801, USA

2 Department of Chemical & Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117576


XPS and UPS data for a Pt sputtering time of 5 seconds.

Figure SI1. XPS scans of (a) Ti 2p, (b) O 1s and (c) Pt 4f for TiO2 thin films with different carrier concentrations for a Pt sputtering time of 5 s (0.89 μg Pt/cm2 or 1.8 ML Pt). When compared to the XPS spectra obtained for a shorter 1 s Pt sputtering time (Figure 2), a small 0.3 eV shift in the Pt 4f7/2 BE to 71.1 eV is detected, consistent with the development of bulk Pt metallic character. Compared to Pt-free TiO2, the Ti 2p3/2 peak shifts from 458.4 to 459.0 eV, while the O 1s peak shifts from 529.6 to 530.3 eV. This can be attributed to charge depletion near the surface of the TiO2 film, since a more positive potential increases the BE of the core electrons.


Figure SI2. UPS scan for 1.8 ML (0.89 μg Pt/cm2) Pt deposited on anatase TiO2 thin films with different carrier concentrations (Nd). The bulk Pt spectrum is included for reference. The Fermi edge is at 0 eV. The UPS spectra suggest that, within the equipment detection limit, the Pt valence structure is independent of the TiO2 carrier concentration. This is consistent with the screening in larger Pt particles.


Evaluation of possible mass transfer limitations in the experiments.

In order to evaluate the possibility of mass transfer limitations under the experimental conditions we simulated the experiments using a 3D non-steady state diffusion-reaction transport model as implemented in COMSOL Multiphysics®.

The vacuum batch reactor was simulated as a spherical chamber with a radius of 3.1 cm. A 1 cm2 square plate at the center of the reactor was used to model the catalyst sample. In the simulations, the reactor was filled with 1200 Pa CO and 13 Pa O2, at a reaction temperature of 623 K. As the boundary conditions, zero flux was specified at the sphere boundary and a first-order rate equation (eq. 2) was used for the planar model catalyst surface. A binary diffusivity of 45 cm2/s was used for all species, consistent with the low pressure and the tempreature of 623 K. The computational domain of the reactor was discretized with fine meshing and the evolution of CO2 was monitored as a function of time.

For diffusive transport of the reactants and product, Fick’s time-dependent second law is used:

where,

D = diffusion coefficient of molecules

c = concentration

t = time

As shown in Figure SI3(a), for the conditions in our experiment, mass transfer limitations are small, and the CO2 evolution at the detector follows the CO2 formation rate on the planar catalyst surface. When the rate coefficient is increased by a factor 50, the CO2 concentration profile begins to show obvious gradients, indicative of mass transfer limited conditions (Fig. SI3(b)). When the binary diffusion coefficients are decreased to 0.45 cm2/s, a value typical value under atmospheric pressure (Bird, Stewart, Lightfoot, Transport Phenomena, Wiley, 2007), mass transfer limitations also become obvious (Fig. SI3(c)), as expected for a turnover frequency of ~400 s-1 (manuscript). The higher diffusion coefficients at the reduced pressure hence allowed measuring such high rates.

Figure SI3. CO2 concentration as function of the distance from the planar catalyst surface and for the geometry of the vacuum batch reactor during the first 5 seconds of the reaction. (a) CO2 concentration profile for the experimental conditions (t=1–5 s). No strong concentration gradients are observed, suggesting negligible mass transfer limitations. (b) CO2 concentration profile when the reaction rate coefficient is increased by a factor 50. (t=1–5 s). For these conditions, strong concentration gradients are found. (c) CO2 concentration profile when the binary diffusion coefficients are reduced by a factor 100. (t=1–5 s) Again, significant concentration gradients develop.

1