Introduction

                The NO oxidation reaction (NO + ½ O₂ ↔ NO₂) on platinum is the first step in the NOx storage and reduction process (NSR), a technology currently being commercialized for emissions reduction in diesel engines. In previous studies, such as the work by Mulla et al. [1], the turnover rate (TOR) was shown to be a strong function of platinum particle size with higher TOR occurring on catalysts with lower dispersions indicating this reaction may be sensitive to the structure of the Pt surface. To better understand this apparent structure sensitivity, the kinetics on Pt(111) and Pt(110) single crystals at the same conditions used on supported catalysts were determined.  The surface coverage of oxygen and nitrogen species as well as the oxidation state of Pt under simulated reaction conditions was studied by in-situ X-Ray Photoelectron Spectroscopy (XPS).  

 

Materials and Methods

                The kinetic study was performed in a custom atmospheric pressure batch reactor system combined with an ultrahigh vacuum (UHV) system which allows for in-vacuuo sample transfer. A FTIR spectrometer with a gas cell was used to measure NO and NO₂ concentrations in the batch reactor. The Pt single crystals were heated by passing electrical current through them. A K-type thermocouple was spot welded to the side of the crystal for temperature measurement. The UHV chamber contains a sputter gun for sample cleaning, mass spectrometer, low energy electron diffraction (LEED), and a cylindrical mirror analyzer (CMA) with electron gun for Auger electron spectroscopy (AES). The crystals were prepared by repeated argon sputtering, prolonged exposure to 1x10^-5 O₂ at 500°C and annealing to 800°C until no impurities such as Si, S, Ca, C, and O were seen by AES. The Pt crystal structures were verified by LEED after final annealing. After reaction, AES was used to verify the absence of impurities and LEED was used to check for any surface reconstruction which occurred during reaction. The TOR on Pt was determined using numerical methods in order to account for the homogenous gas phase reaction as well as approach to equilibrium. Kinetic measurements without a catalyst were used to verify the homogenous gas phase kinetic model. Stainless steel foil similar in size to the crystals was used in a series of blank experiments to verify that any activity beyond the gas phase reaction was originating only from the Pt. Accurate measurements of the rate of heat transfer from the Pt to the gas phase was used to rule out external mass transfer limitations using the Chilton-Colburn relationships. The resulting data set allowed us to accurately determine apparent activation energies and reaction orders for NO,  NO₂, and O₂ using a simple power rate law.

 

In-situ X-Ray Photoelectron Spectroscopy (XPS) experiments were performed at the Ambient Pressure Photoemission Endstation, Beamline 9.3.2.1, at the Advanced Light Source (ALS), Lawrence Berkeley National Labs (LBL). It contains a sputter gun for sample cleaning, a temperature controlled sample stage, and a high pressure photoemission spectrometer (HPPES) which consists of a differentially pumped electrostatic lens system combined with a PHI hemispherical electron analyzer [2]. The HPPES system allows pressures up to a few Torr in the analysis chamber during XP spectra acquisition. Pt single crystals were heated using a ceramic button heater in contact with the back side of the sample. The temperature was measured using a K-type thermocouple spot welded to the side of the crystal. The Pt surface was cleaned in accordance with procedures used in the kinetic study which yield a clean, well annealed surface. Sample cleanliness was checked by XPS throughout the experiments. 

 

Surface coverage of oxygen was estimated by using the O 1s/Pt 4p_3/2 and O 1s/Pt 4f intensity ratios. Oxygen coverage was calibrated and verified using various conditions that yield known oxygen coverages which have been well established in previous studies found in the literature. The incident X-ray energy was varied for each excitation such that all photoelectrons had a kinetic energy of approximately 260eV in order to maintain a constant inelastic mean free path. The binding energy scale was calibrated by setting the Pt 4f_7/2 peak location to 70.9 eV since the Fermi edge was not measurable. The Pt 4f doublet and Pt 4p_3/2 line shapes were fit using a hybrid Doniach–Sunjic / summed Gauss-Lorenzian function. The Pt 4f_5/2:4f_7/2 area ratio was constrained to 0.75. The O 1s line shape was fit using a summed Gauss-Lorenzian function. A Shirley background was used for the Pt 4f doublet and a linear background was used for the O 1s / Pt 4p_3/2 peaks. The adjustable parameters for each line shape were optimized using the entire data set.

 

Results and Discussion

                Preliminary results from the kinetic study on Pt(111) and Pt(110) show that the NO oxidation kinetics, e.g. activity, apparent activation energies and reaction orders, are relatively independent of the structure of the Pt surface. The reaction orders for NO and O₂ are equal to +1 and equal to -1 for NO₂, indicating the product NO₂ inhibits the forward rate of reaction. The TOR on the single crystals show the highest rates observed for the NO oxidation reaction reported in literature on supported Pt. For example, the observed rate on Pt(111) at 250°C, 159ppm NO, 40ppm NO₂, 10% O₂ and N₂ balance, the TOR is 0.23±0.03 s^-1 approximately 260 times greater than a Pt/Al₂O₃ monolithic catalyst with an average particle size of 2.3 nm reported by Mulla et al. [1]. Measured apparent activation energies and reaction orders on the single crystals are the same with in experimental error to those measured by Mulla on supported Pt. Preliminary measurements on supported Pt in our lab with only large Pt particles show TOR similar to the single crystals. It was expected that since the Pt(111) surface represents the dominant surface orientation of Pt under the limit of large particle size, and the rates increase with particle size, it is expected that the TOR on a Pt(111) single crystal represents the highest TOR possible on a platinum catalyst without support effects. Interestingly, the Pt(110) also showed similar kinetic behavior which seems to contradict this hypothesis. However, current experiments are also indicating that each crystal face begins to deviate from the reported kinetics, e.g. deactivate, under different combinations of temperature, NO:NO₂ ratio, and time which may help explain the observed deactivation of smaller Pt particles which are thought to deactivate by oxidation.     

 

In Situ XPS was performed in order to determine the amount and chemical identity of Pt-bound oxygen under reaction conditions for each crystal face. As of now, only studies on Pt(111) have been completed. To determine the maximum oxygen coverage under reaction conditions, the Pt(111) was exposed to 0.275 torr NO₂ (362ppm NO₂) at  250°C which yielded an atomic oxygen coverage of 0.7±0.1 ML. The error represents the standard deviation of five repeat measurements. No surface nitrogen species and Pt oxide formation were observed during the experiment which lasted about an hour. In a similar experiment with O₂, an oxygen coverage of 0.5 ML was observed. Since NO₂ is a much more effective oxidizer than O₂, we believe that the maximum oxygen coverage on a Pt(111) single crystal is 0.75 ML even with the addition of O₂ up to 1 bar at 250°C. In a series of experiments with NO and NO₂, we show that the ratio of NO:NO₂ plays a critical role in determining the coverage of oxygen on the surface and therefore the availability of empty sites for adsorption of O₂, the proposed rate determining step in our L-H mechanism. The experiments show why the product NO₂ effectively inhibits the forward reaction rate by poisoning the surface with oxygen.

 

According to Mulla and co-workers [1], metallic Pt is the active state of Pt in the NO oxidation reaction and the Pt oxide is inactive. Based on the kinetics and in-situ XPS results presented here along with the work of Mulla, we believe that large platinum particles resist oxide formation due to the closed-packed structure and remain active whereas small particles which have a more open structure on average are more easily oxidized and become inactive. This work using Pt single crystals helps to explain these results, which show that all three faces are equally active under the right conditions, however experiments are also indicating that each crystal face begins to deactivate under different combinations of temperature, NO:NO₂ ratio, and time which may help explain the observed deactivation of smaller Pt particles.

 

References

1. .Mulla, S.S., Chen, N., Cumaranatunge L, Blau, G.E., Zemlyanov, D.Y., Delgass, W.N.,

Ribeiro, F.H., J. Catal. 241, 389 (2006)

 

2. Ogletree, D.F., et al., A differentially pumped electrostatic lens system for photoemission studies in the millibar range. Review of Scientific Instruments, 2002. 73(11): p. 3872-3877.