Role of Iron on the Structure and Stability of Ni3.2Fe/Al2O3 during Dynamic CO2 Methanation for P2X Applications

An energy scenario, mainly based on renewables, requires efficient and flexible Power‐to‐X (P2X) storage technologies, including the methanation of CO2. As active Ni0 surface sites of monometallic nickel‐based catalysts are prone to surface oxidation under hydrogen‐deficient conditions, we investigated iron as “protective” dopant. A combined operando X‐ray absorption spectroscopy and X‐ray diffraction setup with quantitative on‐line product analysis was used to unravel the structure of Ni and Fe in an alloyed Ni−Fe/Al2O3 catalyst during dynamically driven methanation of CO2. We observed that Fe protects Ni from oxidation and is itself more dynamic in the oxidation and reduction process. Hence, such “sacrificial” or “protective” dopants added in order to preserve the catalytic activity under dynamic reaction conditions may not only be of high relevance with respect to fine‐tuning of catalysts for future industrial P2X applications but certainly also of general interest.


Role of Iron on the Structure and Stability of Ni 3.2 Fe/Al 2 O 3 during Dynamic CO 2 Methanation for P2X Applications
Marc-André Serrer, [a, b] Kai F. Kalz, [a, b] Erisa Saraçi, [a, b] Henning Lichtenberg, [a, b] and Jan-Dierk Grunwaldt* [a, b] An energy scenario, mainly based on renewables, requires efficient and flexible Power-to-X (P2X) storage technologies, including the methanation of CO 2 . As active Ni 0 surface sites of monometallic nickel-based catalysts are prone to surface oxidation under hydrogen-deficient conditions, we investigated iron as "protective" dopant. A combined operando X-ray absorption spectroscopy and X-ray diffraction setup with quantitative on-line product analysis was used to unravel the structure of Ni and Fe in an alloyed NiÀ Fe/Al 2 O 3 catalyst during dynamically driven methanation of CO 2 . We observed that Fe protects Ni from oxidation and is itself more dynamic in the oxidation and reduction process. Hence, such "sacrificial" or "protective" dopants added in order to preserve the catalytic activity under dynamic reaction conditions may not only be of high relevance with respect to fine-tuning of catalysts for future industrial P2X applications but certainly also of general interest.
Renewable energy sources, such as wind and solar, provide a sustainable solution to the ever-growing demand for energy. [1] To ensure an overall grid stability, their (seasonal) fluctuations must be balanced. [2] P2X storage technologies represent one option to overcome this problem. Energy can be stored e. g. as Fischer-Tropsch products, methanol or synthetic natural gas (SNG). [3] For SNG, the existing and long-ranging gas grid can be used for low-cost storage and distribution. Within the "powerto-gas" process chain, excess renewable electric power is catalytically converted to methane, a chemical energy carrier (Sabatier reaction, [Equation (1)]). [4] Renewable H 2 from electrochemical water splitting, [5] and CO 2 , e. g. from biomass gasification or industrial exhaust gases, [6] can be used as reactants for this catalytic process. The production of hydrogen via electrolysis in a renewable energy scenario depends directly on weather-related fluctuations, especially when produced in decentralized plants with small H 2buffer tanks. [1b,2] As these fluctuations are transferred to the reactor bed and the catalytic system, efficient and stable catalysts that can withstand the highly dynamic and demanding P2X conditions are required. Catalysts based on nickel are commonly used and have been extensively studied under stationary reaction conditions, providing satisfactory catalytic activity at low cost. [7] Recently, the dynamics of these systems have received growing attention in research. [8] It was concluded that during short-term H 2 -dropouts the preservation of the catalytically active Ni 0 species and especially the removal of oxygen from the catalytic surface were crucial for maintaining the catalytic activity. One possible approach to protect Ni 0 from forming surface oxygen species is the addition of a second "sacrificial" or "protective" metal with a higher affinity to oxygen, such as iron. Under stationary reaction conditions, we recently found that by addition of iron the catalyst exhibited improved long-term stability and superior CO 2 conversion compared to a monometallic nickel catalyst. [9] This additionally motivates to characterize the role of iron in detail in such a bimetallic NiÀ Fe catalyst, especially under dynamic reaction conditions, as present in P2X applications. By combining X-ray absorption spectroscopy (XAS) and X-ray diffraction (XRD) as powerful operando tools with quantitative product analysis in a single experiment, we aimed at revealing the role of iron in this system. To our knowledge, the role of a "sacrificial" or "protective" metal in order to preserve the catalytically active species has not yet been investigated with in-depth operando studies under dynamic reaction conditions for bimetallic NiÀ Fe/ Al 2 O 3 catalysts in the methanation of CO 2 .
Therefore, we prepared two model catalysts, consisting of 17 wt % Ni 3.2 Fe/Al 2 O 3 and 17 wt % Ni/Al 2 O 3 by precipitation with urea [9] with similar particle size for good comparability. In order to obtain information about the structural changes of amorphous as well as crystalline phases, we performed alternatingly operando XAS-and XRD experiments at the beamline BM31 (ESRF) using a micro quartz capillary setup [10] and a micro-GC for quantitative on-line gas analysis (details in ESI, Figure S1). Prior to the catalytic experiments, the catalysts were activated by temperature-programmed reduction (TPR) with H 2 . The experiments were performed at 350°C and atmospheric pressure. Experimental details including further time-resolved data are given in the ESI.
As an extreme case for dynamic reaction conditions during P2X applications, we simulated H 2 dropouts for 70 minutes in the reactant feed gas mixture during the methanation of CO 2 . When we performed the experiment with the monometallic Ni/ Al 2 O 3 catalyst, we achieved a CO 2 conversion of 28 % (79 % selectivity to CH 4 , see Figure 1b) under stationary reaction conditions (350°C, 1 bar). The only detected by-product was CO. Since no differences between the NiÀ K XANES spectra ( Figure 1a, for time-resolved see Figure S2 in the ESI) recorded after TPR (À ) and during methanation (---) were observed, the catalyst remained stable during CO 2 methanation. When we simulated hydrogen-lean conditions resulting from a H 2 -dropout ( *** ), the shift of the pre-edge feature (Ni 0 at 8333 eV) and the increasing whiteline at 8350 eV in the NiÀ K XANES spectra in Figure 1a indicate that nickel is in the oxidation state 2 + . According to linear combination analysis (LCA),~96 % of the Ni 0 species were oxidized to Ni 2 + . As it is difficult to distinguish between oxygen and carbon species by EXAFS, we simultaneously followed the formation of NiO or NiCO 3 species by XRD. NiO (2Θ = 11.7°and 19.5°, cf. Figure S2d) was the only detectable oxidized nickel phase, thus, a significant formation of carbonates during the H 2 dropout using technical grade CO 2 / N 2 could be excluded. During the subsequent methanation step,~35 % of Ni 0 were recovered (~76 % Ni 0 before H 2 dropout) resulting in a CO 2 conversion of 11 % (~28 % before H 2 dropout). Since only minor sintering occurred (crystallite size estimated by Scherrer equation: 3.4 nm to 3.9 nm) and the decrease in Ni 0 species directly correlates to the loss in activity, we can conclude that the active species for the conventional Ni-based catalyst is the reduced Ni 0 . The oxygen species either originating from CO 2 dissociation, [8c,11] or from oxygen impurities in the technical grade gas feed might be responsible for the formation of the NiO species and are therefore critical for monometallic nickel catalysts.
To protect the active Ni 0 species from oxidation by surface oxygen species, we modified the catalyst by adding a second metal which, compared to Ni, thermodynamically favors the formation of an oxide species. In this case iron was used, as bimetallic NiÀ Fe/Al 2 O 3 catalysts recently showed promising activity and long-term stability in the stationary methanation of CO 2 . [9,12] In this study, the prepared 17 wt % Ni 3.2 Fe/Al 2 O 3 catalyst reached 56 % CO 2 conversion (97 % selectivity to CH 4 , see Figure 2d) during methanation of CO 2 at 350°C at 1 bar, which was significantly higher compared to the monometallic nickel catalyst (28 %, see Figure 1b). After 60 minutes of CO 2 methanation (full data set, cf. ESI in Figure S3), a H 2 -dropout was applied for 70 minutes using the same gases, setup and reaction conditions as for the monometallic Ni catalyst. During the H 2 dropout, the increasing whiteline intensities in the NiÀ K and FeÀ K XANES spectra (FeÀ K measured in fluorescence mode) in Figure 2 and Figure S4 demonstrated a slight oxidation of Ni 0 to Ni 2 + (~7 % to 11 %) and of Fe 0 to Fe 2 + (~33 % to 47 %). This moderate oxidation is also visible in the operando XRD pattern (Figure 2c), where the intensity of the Ni (200) reflection at 2Θ = 15.95°slightly decreased during the H 2 dropout. A slight shift of the Ni(200) reflection was observed, which might represent the formation of a Ni-richer alloy, as some Fe migrated to the surface under formation of Fe 2 + . [13] We can assume that mainly FeO was formed, similar to the oxidation of the monometallic nickel catalyst due to oxygen species present during the H 2 dropout. The preferential oxidation of iron suggests that the less-noble metal acts as a "protective" element in relation to nickel. In addition, according to DFT calculations, [14] iron can migrate out of a NiÀ Fe alloy to the surface to form FeO, if a monolayer of oxygen is present. This explains the hint for dealloying we observed during hydrogen-lean conditions resulting in a larger amount of surface oxygen species in the catalyst bed. Hence, we observe that iron protected the reduced Ni 0 species from major oxidation by FeO formation, as illustrated in Figure 3 and assumed in the concept of the "protective" less-noble metal Figure 1. Operando NiÀ K edge XANES spectra and k 2 -weighted FT-EXAFS spectra after activation and after the experiment (a), simultaneously measured catalytic performance (b) of Ni/Al 2 O 3 during methanation of CO 2 before (I), during (II) and after (III) a simulated H 2 -dropout. Reaction conditions: T = 350°C, p = 1 bar, 30 mL min À 1 total flow; Methanation: H 2 /CO 2 /N 2 = 4/1/5, H 2 -dropout: CO 2 /N 2 = 1/9. dopant. After 70 minutes of H 2 dropout conditions, the catalyst was re-exposed to methanation conditions. Surprisingly, in the bimetallic Ni x Fe y /Al 2 O 3 catalyst both nickel and iron regained the initial oxidation states, as depicted by the XAS data shown in Figure 2a and Figure 2b. However, the dealloying observed during the H 2 dropout might not be entirely reversible, as the Ni(200) reflection in Figure 2c seemed not to fully return to its initial position. Nevertheless, and in contrast to the monometallic Ni catalyst (Figure 1b), the activity of the bimetallic NiÀ Fe catalyst was immediately regained after switching to methana-tion conditions, reaching 58 % CO 2 conversion (98 % selectivity to CH 4 ), as demonstrated in Figure 2d. The slightly higher activity might be related to the increased amount of iron sites at the surface of the catalyst after the H 2 dropout which might change the mechanism of CO 2 activation and its reaction intermediates, as observed during DRIFTS studies on NiÀ Fe catalysts. [15] In conclusion, the comparison of a monometallic Ni/Al 2 O 3 and a bimetallic Ni 3.2 Fe/Al 2 O 3 model catalyst under dynamic reaction conditions by using an advanced combination of operando XAS and XRD with quantitative on-line product analysis elucidated the important role of iron. During the simulated H 2 -dropout the monometallic nickel catalyst was prone to form surface oxygen species resulting in an irreversible formation of NiO leading to catalyst deactivation during the subsequent methanation step. This provides evidence that Ni 0 sites e. g. at edges/corners of the particles, that are prone to oxidation, are protected or assisted in the re-reduction. By introducing the concept of adding a less-noble "sacrificial" or "protective" metal, such as iron, we were able to preserve the catalytically active species and revealed in detail the role of iron by using advanced X-ray based techniques: The active Ni 0 species was protected from oxidation during the simulated H 2 dropout by preferential formation of FeO (Figure 3). Surprisingly, both metals were completely reduced again in the subsequent methanation step. This resulted in a fully recovered catalytic activity. The use and the understanding of the role of less-noble metals as dopants in order to preserve the active Figure 2. Simultaneously recorded operando NiÀ K edge XANES spectra, as well as k 2 -weighted FT-EXAFS spectra after activation and after the experiment (a), FeÀ K fluorescence spectra (b), XRD with λ = 0.4943 Å (c) and catalytic performance (d) of NiÀ Fe/Al 2 O 3 before (I), during (II) and after (III) a simulated H 2dropout; Ni (#), NiO (*), γ-Al 2 O 3 (♦). Reaction conditions: T = 350°C, p = 1 bar, 30 mL min À 1 total flow; Methanation: H 2 /CO 2 /N 2 = 4/1/5, H 2 -dropout: CO 2 /N 2 = 1/9. catalytic species from deactivation due to oxidation under transient reaction conditions represents a major step forward in the design of future P2X catalysts and may be used as a concept in other applications as well. Our results furthermore show that it is very important to establish such structure-activity relationships by a) using complementary synchrotron-based operando techniques such as XAS and XRD to characterize both amorphous and crystalline phases and b) combining them in one experiment with quantitative product analysis, since the structure depends on the reaction conditions. Further in-depth operando spectroscopic studies would be helpful to deepen the investigation both with respect to the role of iron in the CO 2 activation mechanism and of less noble promotors as sacrificial metal dopants to generalize the concept.