(560fl) A Temporal Analysis of Products (TAP) Study of Passive NOx Adsorption (PNA) on 1% Pd/SSZ-13

Most engine exhaust after-treatment catalytic processes such as selective catalytic reduction (SCR) are effective at temperatures above 200°C. Temperatures below 200°C are considered to be the cold start period when the majority of NO_x and hydrocarbons slip through the after-treatment system. To meet strict NO_x emission standards, trapping NO_x upstream of the SCR catalyst during the cold start period and releasing the NO_x at higher temperatures to enable its reduction is a potential way to lower NO_x emissions further. A material with demonstrated capability for this purpose is Pd-exchanged zeolites, such as ZSM-5, BEA, and SSZ-13. While the PNA concept has been demonstrated in a group of recent studies, the nature and role of the various active sites determining their performance and NO_x absorption chemistry is inadequately understood despite intensive research efforts.¹ In this investigation, we characterize the active sites of a model passive NO_x adsorber (PNA), Pd/SSZ-13 using Temporal Analysis of Products (TAP). The data are used to further develop a working mechanism comprising reaction steps that describe the NO_x trapping and release over a range of temperatures. The TAP data are interpreted with the help of help of flow reactor and computational methods.

The NO_x trapping and release properties on the 1% Pd/SSZ-13 were tested using feeds with different compositions. The experimental results showed the potential of this material to be an effective PNA at low temperatures. Literature shows multiple Pd species co-existing in the catalyst: (i) atomically dispersed or ‘naked’ Pd in the cationic sites in the pores of the zeolites, and (ii) PdO_x particles on the external surfaces. Forms of active Pd sites present in the catalyst are referred to as Pd^II, Pd^I, Pd^II-OH and Pd^IIO.

The 1% Pd/SSZ-13 catalyst was prepared using an incipient wetness impregnation method. Powder catalysts of 250-400 µm were used for experiments. A generation-1 TAP reactor² was used for conducting transient experiments. Controlled pulse experiments were carried out over 5 mg of the catalyst at a defined state to unravel the reaction mechanism. The products from the reaction were measured by a high resolution 100C UTI mass spectrometer. The catalyst was pretreated with 1500 pulses (10¹⁵ molecules/pulse) of O₂/Ar (1:1) mixture at 500°C before each experiment. Experiments were conducted at temperatures between 100°C and 450°C with mixtures of isotopic and non-isotopic NO, with and without O₂ and CO under dry and wet conditions to study the participation of different Pd species and the effects of H₂O, CO on NO_x uptake and release. The interpretation of experimental data was supported through periodic, spin-polarized density functional theory (DFT) calculations using the Vienna ab-initio simulation package (VASP) with the BEEF-vdW exchange correlation functional. Different active site motifs of Pd were modeled in bulk unit cells of SSZ-13 with one or two framework Al atoms.

A DFT investigation on the stability of the above-mentioned sites and the NO adsorption energies on them suggest that NO binds strongly on Pd^I sites, the formation of which can be explained through the two step site modification mechanism as in Eq (1).

2 Z [Pd^IIOH] → Z [Pd^I – O –Pd^I] Z + H₂O → 2 ZPd^I (1)

H₂-temperature programmed reduction was performed in the TAP reactor and the H₂ consumption at temperatures below 300°C is indicative of the amount of cationic Pd present in the catalyst. It was estimated to be ~85% of the total Pd present in the catalyst. A ¹⁶O₂ pretreated catalyst pulsed with a ¹⁵N¹⁸O and ¹⁶O₂ resulting in the production of ¹⁵N¹⁶O¹⁸O and ¹⁵N¹⁶O₂ is consistent with the formation of the Pd⁺ sites from two adjacent Pd²⁺OH sites:

¹⁵N¹⁸O + 2 Pd^II -¹⁶OH ↔ ¹⁵N¹⁶O¹⁸O + 2 Pd^I + H₂¹⁶O (2)

¹⁵N¹⁸O + Pd^{II 16}O ↔ ¹⁵N¹⁶O¹⁸O + Pd^II (3)

The two step modification mechanism for the formation of Pd⁺ sites was confirmed by additional isotopic experiments with C¹⁸O on ¹⁶O₂ pretreated catalyst by monitoring the production of C¹⁶O¹⁸O.

The presence of ¹⁸O₂ and N₂ in the product outlet provide evidence for possible NO dissociation. A more detailed insight into the reaction mechanism is provided by the observation of oxygen scrambling between the feed ¹⁸O and the catalytic ¹⁶O.

¹⁵N¹⁸O + ZPd^II ↔ ¹⁵N¹⁸O- ZPd^II (4)

2¹⁵N¹⁸O- ZPd^II ↔ ¹⁵N₂ + 2 ZPd^II-¹⁸O (5)

2 ZPd^II-¹⁸O ↔ 2 ZPd^II + ¹⁸O₂ (6)

Pd^II-¹⁸O + Pd^II-¹⁶O ↔ 2Pd^II + ¹⁸O¹⁶O (7)

The effects of O₂, H₂O and CO in the feed were also studied for PNA performance of Pd/SSZ-13. The substantial reduction of NO_x storage in the presence of H₂O was primarily attributed to the loss of NO uptake on Brønsted acid sites, which are easily poisoned by H₂O. NO uptake on Pd^I is only mildly impeded by H₂O, as suggested by DFT results. Considering that NO_x uptake over Pd zeolites is enhanced in the presence of CO,³ our results suggest that the oxidation or reduction potential of the gas phase alters the relative availability of active storage sites. In particular, the strongly exothermic nature of CO oxidation renders the reduction from Pd^IIOH to Pd^I thermodynamically favorable.

This study aims to identify the various active sites for passive NO_x adsorption using Pd/SSZ-13 catalyst, which is the most promising material to date. An atomic-level understanding of the ad-/desorption and reaction steps on these active sites will greatly accelerate the design and development of improved PNA materials.

References

¹ Y. Zheng, L. Kovarik. M. H. Engelhard, Y. Wang, F. Gao and J. Szanyi (2017). The Journal of Physical Chemistry

² J. Gleaves, G. Yanblonsky, X. Zheng, R. Fushimi, P. L. Mills (2010). Journal of Molecular Catalysis A: Chemical

³ A. Vu., J. Y. Luo, J. H. Li; W. S. Epling (2017); Catalysis Letters