Role of Electronic Factor in Soot Oxidation Process Over Tunnelled and Layered Potassium Iron Oxide Catalysts

Several iron containing oxides were used in the study: α-Fe₂O₃, K₂Fe₂₂O₃₄ as well as promoted and unpromoted KFeO₂. α-Fe₂O₃ was obtained from Merck and used without further processing. The samples of K₂Fe₂₂O₃₄ ferrite were prepared by solid-state reaction of a stoichiometric mixture of K₂CO₃ with α-Fe₂O₃ at 1200 °C as described previously [12]. The KFeO₂ sample was synthesized from K₂CO₃ with α-Fe₂O₃ according to the procedure described in [13]. In brief, the mixtures of finely ground powders were placed in a porcelain crucible and heated with the rate of 5 °C/min in static air up to the final temperatures of 800 °C for KFeO₂ and 1200 °C for K₂Fe₂₂O₃₄. For 0.5–4 wt% Ce-doped KFeO₂ samples, the appropriate fraction of α-Fe₂O₃ powder was replaced by CeO₂. Four CeO₂ types of precursors were used for KFeO₂ promotion: micrometric commercial cerium(IV) oxide (Merck), nanometric CeO₂ obtained by sol–gel method, solid ammonium cerium(IV) nitrate (Merck) forming an oxide upon calcination and ammonium cerium(IV) nitrate water solution with subsequent impregnation of KFeO₂, drying in 100 °C for 1 h and calcination in 400 °C for 4 h. All such obtained samples were mixed with carbon black, Printex 80 (Degussa AG) with catalyst to soot ratios: 1/1, 4/1, 8/1. Both “loose” than a “tight” conditions [14] were tested in preliminary experiments, but since the results were better reproducible in the latter conditions the “tight contact” was applied in the comparative tests. Catalytic activity of the samples was determined by temperature-programmed oxidation (TPO) tests, carried out with the catalyst/soot mixtures as described above. The amount of sample loaded into the reactor varied from 50 to 100 mg. The soot combustion experiments were carried using a heating rate of 10 °C/min, in the flow of 5 % O₂ in He. The gas phase composition (O₂, CO, CO₂, H₂O) was monitored by the QMS spectrometer (SRS RGA200). The ignition temperatures were determined with the use of the Arrhenius-like plot (lnX vs. 1/T), which gives rise to a sharp turning point defining the start of the oxidation process in a non-arbitrary way in contrast to more subjective yet commonly use method of T _5 % or T _10 %.

The samples were characterized by X-ray diffraction (XRD) (X’pert Pro Philips) for phase composition, SEM (Philips XL 30 ESEM) for grain size morphology, N₂-BET (Quantasorb Jr) and XPS (Prevac/VG SCIENTA R3000) for surface area and composition, and Kelvin Probe (McAllister KP6500) for evaluation of electronic properties. The contact potential difference (V _CPD) measurements were carried out by the dynamic condenser method of Kelvin. To standardize the catalyst surface the measurements were carried out in vacuum of 10⁻⁷ mbar after annealing the sample to 400 °C. The work function values were obtained from a simple relation e⋅V _CPD = Φ _reference – Φ _sample, using a standard stainless steel plate as a reference electrode (d = 3 mm, Φ _reference = 4.1 eV).

Typical XRD patterns and the corresponding SEM images of the obtained ferrites are shown in Fig. 1. The diffraction revealed that the obtained samples of KFeO₂ (indexed within Pbca-JCPDS-ICDD-39-0892 databases as reference) and K₂Fe₂₂O₃₄ (P6 ₃ /mmc-JCPDS-ICDD 31-1034) were virtually monophasic. The K₂Fe₂₂O₃₄ has the hexagonal layered structure and can be considered to be built up from spinel blocks of γ-Fe₂O₃ and separated by conducting potassium layers. The KFeO₂ exhibits a tetragonal tunneled structure, that is not congruent with parent γ-Fe₂O₃. Since this ferrite has no layer structure, K, Fe, and O constituents are spatially more uniformly distributed (Fig. 1).

The introduction of ceria (at the level of 2 %) did not lead to any appreciable changes in the diffraction pattern. The SEM images of the bare ferrites reveal different morphologies of the KFeO₂ and K₂Fe₂₂O₃₄ grains (Fig. 1). Whereas for the KFeO₂ crystallites of 3–5 μm are interconnected forming coarse blunt aggregates of about 10 μm, in the case of K₂Fe₂₂O₃₄ the hexagonal well-shaped crystallites of 0.5–2 μm in diameter and thickness of ~0.2 μm were observed, and resembled that of a single K⁺-β-ferrite with stoichiometric composition [12]. The EDX analysis showed that the distributions of K and Fe are strongly correlated as expected for the ferrite phases, whereas ceria segregates at the surface in the form of spherical smaller nanocrystals with the size in the range of 10–500 nm, depending on the applied preparation method. The XPS surface composition of the investigated samples indicated that within the surface layer only the constituent elements were found with the diagnostic binding energies for KFeO₂ (BE_Fe2p3/2 = 709.7–710.0 eV, BE_K2p3/2 = 292.2–292.4 eV, BE_O1s = 530.4–530.8 eV) and K₂Fe₂₂O₃₄ (BE_Fe2p3/2 = 710.2–710.4 eV, BE_K2p3/2 = 294.4–294.5 eV, BE_O1s 529.4–529.5 eV). Small concentration of CeO₂ combined with the overlapping Fe LMM Auger signal made it impossible to determine the Ce 3d binding energies.

The comparison of the soot combustion activity of the several investigated catalysts as a function of temperature is presented in Fig. 2. The synergetic effect of nanostructuration of iron oxide by potassium into tunneled KFeO₂ and layered K₂Fe₂₂O₃₄ and the surface doping with ceria can be clearly noticed. The transformation of iron oxide into ferrites by potassium addition lowers the light off temperature by about 70 and 100 °C for K₂Fe₂₂O₃₄ and KFeO₂, respectively. Yet, the reactivity of the most active phase KFeO₂ can be further enhanced by ~30 °C upon the 2 % CeO₂ addition (Fig. 3, green points). In order to check the stability of the best material the 10 consecutive tests of soot oxidation were performed for the series of CeO₂–KFeO₂ catalysts (at each time a new portion of soot was added to the same catalyst at the same ratio of 8/1). The results revealed that the sustainable activity was maintained within the experimental limit (the standard deviation of T _ig < 8.5 °C).

It is worth mentioning that for the whole series of the investigated samples a strong correlation of the catalytic activity gauged by both the ignition (T _ig) and 10 % conversion (T _10 %) temperatures of soot combustion with the work function value exists (Fig. 3a). The observed correlations reveal that the electronic effect (availability of electrons for oxygen activation) governs the soot combustion process. It is valid with the same spectacular trend either in the case of nanostructuration of iron oxide by potassium (gray points) and for subsequent surface modification of the best performing KFeO₂ ferrite by 2 % of CeO₂ (green points). Such level of doping was found to be optimal, as it can be inferred from the Fig. 3b. The characteristic Topping-like behavior confirms that the observed changes of T _ig can actually account for by the associated work function changes. As a result, those findings show that the unique, concise parameter—the catalyst work function—plays the key role in the soot combustion process. The observations of a prime importance of catalyst work function modification for the activity was also observed for other redox reactions [15, 16].

As mentioned, the electronic promotion effect on soot combustion can be explained in terms of enhanced energetic accessibility of the electrons for activation of the dioxygen molecule by the catalyst. Indeed, the facilitation of the electron transfer from the catalyst surface to O₂ molecules gives rise to the formation of strongly reactive surface oxygen species, such as O _x ⁻ radicals (O⁻ and O₂ ⁻), which are known as efficient soot oxidation agents taken into account in kinetic modeling of soot oxidation [17]. It is well known that the work function controls the nature of surface oxygen speciation. Two main scenarios related to dioxygen chemisorption include singly ionized molecular (O₂ ⁻) and atomic (O⁻) species as revealed elsewhere [18]. Nonetheless, at this stage the established quantitative work function—combustion temperature relationship can be used as a guidelining principle for the optimization of the catalyst for soot combustion, and searching for new promising washcoats of DPF filters.