Reverse Water Gas Shift by Chemical Looping with Iron-Substituted Hexaaluminate Catalysts

: The Fe-substituted Ba-hexaaluminates (BaFeHAl) are active catalysts for reverse water-gas shift (RWGS) reaction conducted in chemical looping mode. Increasing of the degree of substitution of Al 3 + for Fe 3 + ions in co-precipitated BaHAl from 60% (BaFeHAl) to 100% (BaFe-hexaferrite, BaFeHF), growing its surface area from 5 to 30 m 2 / g, and promotion with potassium increased the CO capacity in isothermal RWGS-CL runs at 350–450 ◦ C, where the hexaaluminate / hexaferrite structure is stable. Increasing H 2 -reduction temperature converts BaFeHAl to a thermally stable BaFeHF modiﬁcation that contains additional Ba-O-Fe bridges in its structure, reinforcing the connection between alternatively stacked spinel blocks. This material displayed the highest CO capacity of 400 µ mol / g at isothermal RWGS-CL run conducted at 550 ◦ C due to increased concentration of oxygen vacancies reﬂected by greater surface Fe 2 + / Fe 3 + ratio detected by XPS. The results demonstrate direct connection between CO capacity measured in RWGS-CL experiments and calculated CO 2 conversion.


Introduction
CO 2 can be converted to syngas, a key intermediate in production of green chemicals and fuels [1,2], by reverse water-gas shift (RWGS). Catalysts for this reaction are Cu, Ce, Ni, Fe-based oxides as well as supported noble metals (Pt, Rh), and multicomponent metal oxides [1][2][3][4][5][6]. The reaction is reversible and endothermic, thus it yields relatively low conversion at equilibrium at <600 • C. The shortcoming of those catalysts is significant methanation activity reducing the yield of carbon oxide at high temperatures [6].
Chemical looping (CL) combined two separate steps: hydrogen reduction of metal oxide catalyst followed by CO 2 oxidation of the catalyst to produce CO (Scheme 1). Each one of the two catalytic steps is not thermodynamically limited, thus complete conversion of CO 2 can be accomplished at relatively low temperature. Furthermore, since CO and H 2 are not present simultaneously on the catalyst surface, no methanation is expected. Metal oxide catalytic materials suitable for RWGS-CL should have the ability to eliminate oxygen from materials surface for efficient splitting of C-O bond of CO 2 during adsorption on O-vacancies at surface of the catalyst. It is important to maximize the surface area and surface concentration of suitable O-vacancies governed by cationic environment of precursor oxygen ions.
Iron oxides display natural abundance and high re-oxidation capacity with CO 2 over a wide range of operating conditions [7][8][9]. However, pure iron oxides tend to deactivate rapidly. The major factor for deactivation of pure iron oxide materials is sintering [10,11]. To overcome this problem, iron oxides are often modified with other oxide materials such as CeO 2 , MgO, TiO 2 , ZrO 2 , or deposited on SiO 2 or agreement with magnetoplumbite structure displaying the unit cell parameters a and c similar to that reported in the literature [53], ICDD card 79-1742. The iron and aluminum atoms occupy the same positions in the unit cells of BaFeHAl and BaFeHF materials (Figure 1b,c) with different bonds length determined by ionic radii of corresponding ions. According to Rietveld refinement, part of iron ions' positions in the unit cell of hexaaluminate structure BaFe12O19 [50,53,54] in our co-precipitated materials remain not occupied living cationic vacancies. Therefore, the formula of these materials followed from XRD data differs from this classical composition containing 11 (Al + Fe) atoms for BaFeHAl and 11 Fe atoms for BaFeHF with less oxygens for keeping the crystals' electroneutrality.  Table 1 depicts the chemical compositions and texture parameters of BaFeHAl and BaFeHF tested in RWGS by chemical looping. The texture of co-precipitated (CP) and carbon-templated (CT) materials. With 60% and 100% degree of Al 3+ to Fe 3+ exchange, it was also characterized after modification with potassium (6% wt) and after High Energy Ball Milling (HEM). Carbon-templating and HEM increased the surface area of BaFeHAl material from 5 m 2 g −1 to 20 and 30 m 2 g −1 , respectively, without change of phase and chemical composition. In the case of BaFeHF application of HEM, the surface area increased from 2 m 2 g −1 to 18 m 2 g −1 with no change in structure and composition. Carbon templating decreases the platelets size of primary nanocrystals of BaFeHAl from 0.3-0.7 (co-a)  Table 1 depicts the chemical compositions and texture parameters of BaFeHAl and BaFeHF tested in RWGS by chemical looping. The texture of co-precipitated (CP) and carbon-templated (CT) materials. With 60% and 100% degree of Al 3+ to Fe 3+ exchange, it was also characterized after modification with potassium (6% wt) and after High Energy Ball Milling (HEM). Carbon-templating and HEM increased the surface area of BaFeHAl material from 5 m 2 g −1 to 20 and 30 m 2 g −1 , respectively, without change of phase and chemical composition. In the case of BaFeHF application of HEM, the surface area increased from 2 m 2 g −1 to 18 m 2 g −1 with no change in structure and composition. Carbon templating decreases the platelets size of primary nanocrystals of BaFeHAl from 0.3-0.7 (co-precipitation) to 0.1-0.4 µm forming additional mesopores between nanocrystals aggregates [50]. This is reflected by the appearance of small mesopores with radius of <70 Å (Figure 2b). The HEM treatment further decreased the size of co-precipitated BaFeHAl primary crystals due to grinding to 0.03-0.1 µm (Figure 3c,d), strongly increasing the pore volume at wide range of pore radii ( Figure 2).

of 18
Catalysts 2020, 10, x FOR PEER REVIEW 5 of 18 precipitation) to 0.1-0.4 µm forming additional mesopores between nanocrystals aggregates [50]. This is reflected by the appearance of small mesopores with radius of <70 Å (Figure 2b). The HEM treatment further decreased the size of co-precipitated BaFeHAl primary crystals due to grinding to 0.03-0.1 µm (Figure 3c,d), strongly increasing the pore volume at wide range of pore radii ( Figure 2 precipitation) to 0.1-0.4 µm forming additional mesopores between nanocrystals aggregates [50]. This is reflected by the appearance of small mesopores with radius of <70 Å (Figure 2b). The HEM treatment further decreased the size of co-precipitated BaFeHAl primary crystals due to grinding to 0.03-0.1 µm (Figure 3c,d), strongly increasing the pore volume at wide range of pore radii ( Figure 2). . Figure 3. SEM micrographs of BaFeHF and BaFeHAl before (a,c) and after HEM treatment (b,d), respectively.
Similar effect was observed after HEM treatment of BaFeHF (Figures 2 and 3a,b). HEM treatment decreased the average crystal size of both co-precipitated BaFeHAl and BaFeHF materials to the level detectable from the widening of XRD reflections, 30 and 40 nm, respectively. Promotion with potassium decreased the surface area of all materials due to aggregation of primary crystals to less porous agglomerates [50]. This brings the increase of pore diameter due to the predominant disappearance of small pores (Figure 2a,c).

Factors Affecting the Performance of Fe-Substituted Ba-Hexaaluminates
RWGS includes two consecutive steps: H 2 -reduction and CO 2 oxidation of catalyst at elevated temperatures, where thermal stability of the structure of selected phases is critical for testing their performance. The BaFeHAl and BaFeHF materials are very stable at oxidative conditions being prepared by annealing at temperatures of >1000 • C. However, they may decompose into different phases in reductive H 2 -atmosphere [49]. The spectra of water evolution during Temperature Programmed Reduction (TPR) of as-prepared BaFe-hexaxaluminate and BaFe-hexaferrite are shown in Figure 4. After low-temperature peaks corresponding to removal of adsorbed water, the spectra contain groups of high-temperature peaks corresponding to water evolution due to H 2 -reduction of corresponding bulk phases. BaFeHAl is more stable to reduction compared with BaFeHF: high-temperature water evolution starts at 450 and 300 • C, respectively. In addition, the intensity of low-temperature reduction peak centered at~400 • C in case of BaFeHF is significantly higher compared with that for BaFeHAl centered at 550 • C. This may be a result of deep deoxygenation of the structure of both materials in hydrogen atmosphere at temperatures >300-350 • C for BaFeHF and >450-550 • C for BaFeHAl, and their conversion to other oxygen-depleted phases. The phase compositions of as-prepared BaFeHAl and BaFeHF catalysts after treatment in 10% H 2 -Ar flow in a tubular reactor at 35 cm 3 min −1 gram −1 for 20 min are shown in Table 2. From these data it follows that at H 2 -reduction step of RWGS-CL cycle the BaFeHAl phase is stable up to 450 • C and the BaFeHF phase up to 350 • C. Similar effect was observed after HEM treatment of BaFeHF (Figures 2 and 3a,b). HEM treatment decreased the average crystal size of both co-precipitated BaFeHAl and BaFeHF materials to the level detectable from the widening of XRD reflections, 30 and 40 nm, respectively. Promotion with potassium decreased the surface area of all materials due to aggregation of primary crystals to less porous agglomerates [50]. This brings the increase of pore diameter due to the predominant disappearance of small pores (Figure 2a,c).

Factors Affecting the Performance of Fe-Substituted Ba-Hexaaluminates
RWGS includes two consecutive steps: H2-reduction and CO2 oxidation of catalyst at elevated temperatures, where thermal stability of the structure of selected phases is critical for testing their performance. The BaFeHAl and BaFeHF materials are very stable at oxidative conditions being prepared by annealing at temperatures of >1000 °C. However, they may decompose into different phases in reductive H2-atmosphere [49]. The spectra of water evolution during Temperature Programmed Reduction (TPR) of as-prepared BaFe-hexaxaluminate and BaFe-hexaferrite are shown in Figure 4. After low-temperature peaks corresponding to removal of adsorbed water, the spectra contain groups of high-temperature peaks corresponding to water evolution due to H2-reduction of corresponding bulk phases. BaFeHAl is more stable to reduction compared with BaFeHF: hightemperature water evolution starts at 450 and 300 °C, respectively. In addition, the intensity of lowtemperature reduction peak centered at ~400 °C in case of BaFeHF is significantly higher compared with that for BaFeHAl centered at 550 °C. This may be a result of deep deoxygenation of the structure of both materials in hydrogen atmosphere at temperatures >300-350 °C for BaFeHF and >450-550 °C for BaFeHAl, and their conversion to other oxygen-depleted phases. The phase compositions of asprepared BaFeHAl and BaFeHF catalysts after treatment in 10%H2-Ar flow in a tubular reactor at 35 cm 3 min −1 gram −1 for 20 min are shown in Table 2. From these data it follows that at H2-reduction step of RWGS-CL cycle the BaFeHAl phase is stable up to 450 °C and the BaFeHF phase up to 350 °C.    The results of RWGS-CL presented in Figure 5 show that the CO capacity of this material increases proportionally to its surface area both in as-prepared and K-promoted forms. K-promoted BaFeHAl displayed significantly higher CO capacity compared with as-prepared material in spite of lower surface area. Specific CO capacity (SC) of K-promoted catalysts normalized to the unit of surface area (µmol m −2 ) increased by a factor of 1.8-3.2. This effect is related to higher concentration of oxygen vacancies in reduced Fe-materials due to reductive action of potassium [50,56]. The efficiency of BaFeHAl and BaFeHF catalysts with similar surface areas was compared in isothermal RWGS-CL runs at 350 • C. Data presented in Figure 6 show that BaFeHF display higher CO capacity, and thus is more efficient than BaFeHAl. The SC is 3.5 and 2.6 µmol m −2 for BaFeHF and BaFeHAl, respectively. The ratio SC BaFeHF /SC BaFeHAl is close to the ratio of iron content in these materials ( Table 1), reflecting that the active sites for RWGS reaction are anionic vacancies formed at the catalysts surface at H 2 -reduction step due to the conversion of Fe 3+ to Fe 2+ [50,56]. Their surface concentration should be proportional to the iron content taking in account the similar structure of both materials. Promoting BaFeHF with K increased its CO capacity ( Figure 6) and SC value to 8.7 µmol m −2 . BaO -----4 Catalysts 2020, 10, x FOR PEER REVIEW 7 of 18 Fe3O4  -9  41  55  80  41  FeO  -3  5  13  11  49  BaCO3  --4  6  9  6  BaO  -----4 The results of RWGS-CL presented in Figure 5 show that the CO capacity of this material increases proportionally to its surface area both in as-prepared and K-promoted forms. K-promoted BaFeHAl displayed significantly higher CO capacity compared with as-prepared material in spite of lower surface area. Specific CO capacity (SC) of K-promoted catalysts normalized to the unit of surface area (µmol m −2 ) increased by a factor of 1.8-3.2. This effect is related to higher concentration of oxygen vacancies in reduced Fe-materials due to reductive action of potassium [50,56]. The efficiency of BaFeHAl and BaFeHF catalysts with similar surface areas was compared in isothermal RWGS-CL runs at 350 °C. Data presented in Figure 6 show that BaFeHF display higher CO capacity, and thus is more efficient than BaFeHAl. The SC is 3.5 and 2.6 µmol m −2 for BaFeHF and BaFeHAl, respectively. The ratio SCBaFeHF/SCBaFeHAl is close to the ratio of iron content in these materials ( Table  1), reflecting that the active sites for RWGS reaction are anionic vacancies formed at the catalysts surface at H2-reduction step due to the conversion of Fe 3+ to Fe 2+ [50,56]. Their surface concentration should be proportional to the iron content taking in account the similar structure of both materials. Promoting BaFeHF with K increased its CO capacity ( Figure 6) and SC value to 8.7 µmol m −2 .  Increasing of H2-reduction temperature (Tr) of BaFeHF catalyst beyond 350 °C leads to decomposition of this phase converting it to iron oxides, barium oxide, and carbonate ( Table 2). Bacompounds are not active in the redox cycle. Presented in Figure 7 are the results of non-isothermal testing of 6%KBaFeHF-18 keeping the reduction temperature at 350 °C that avoids catalysts reductive Increasing of H 2 -reduction temperature (T r ) of BaFeHF catalyst beyond 350 • C leads to decomposition of this phase converting it to iron oxides, barium oxide, and carbonate (Table 2). Ba-compounds are not active in the redox cycle. Presented in Figure 7 are the results of non-isothermal testing of 6%KBaFeHF-18 keeping the reduction temperature at 350 • C that avoids catalysts reductive decomposition. Increasing CO 2 oxidation temperature increased CO capacity from 68 ( Figure 6) to 114 (µmol g −1 ) (Figure 7a), depicted by the shape of CO production peaks (Figure 7b). XRD analysis indicates that the structure of 6%K/BaFeHF did not change after the CO 2 oxidation step at 550-650 • C. It means that increasing oxidation temperature involves additional Fe 2+ ions with lower oxidation ability to the CO 2 oxidation process. Increasing of H2-reduction temperature (Tr) of BaFeHF catalyst beyond 350 °C leads to decomposition of this phase converting it to iron oxides, barium oxide, and carbonate (Table 2). Bacompounds are not active in the redox cycle. Presented in Figure 7 are the results of non-isothermal testing of 6%KBaFeHF-18 keeping the reduction temperature at 350 °C that avoids catalysts reductive decomposition. Increasing CO2 oxidation temperature increased CO capacity from 68 ( Figure 6) to 114 (µmol g −1 ) (Figure 7a), depicted by the shape of CO production peaks (Figure 7b). XRD analysis indicates that the structure of 6%K/BaFeHF did not change after the CO2 oxidation step at 550-650 °C . It means that increasing oxidation temperature involves additional Fe 2+ ions with lower oxidation ability to the CO2 oxidation process. The decomposition of BaFeHAl catalyst during H2-reduction step of the RWGS-CL cycle at temperatures of >450 °C yields novel thermostable BaFeHFr phase as a main component with the depletion of BaCO3, Al2O3, and Fe3O4 phases ( Table 2). Since BaFeHF is a more efficient catalyst, the effect of reduction temperature on CO capacity of 6%K/BaFeHAl was studied at the temperature range of 350-600 °C where the catalyst contained BaFeHAl or BaFeHFr phases, keeping constant the CO2 oxidation temperature at 450 °C. The results of this non-isothermal testing series are shown in Figure 8.  (Table 2). Since BaFeHF is a more efficient catalyst, the effect of reduction temperature on CO capacity of 6%K/BaFeHAl was studied at the temperature range of 350-600 • C where the catalyst contained BaFeHAl or BaFeHF r phases, keeping constant the CO 2 oxidation temperature at 450 • C. The results of this non-isothermal testing series are shown in Figure 8. The CO capacity of K-promoted BaFeHAl reached a maximum at 550 °C, increasing by a factor of 1.75 compared to the value at 450 °C ( Figure 8a). This is illustrated by changes of the shape of CO production peaks recorded at 450 °C in CO2 oxidation step (Figure 8b). No changes of catalysts phase composition were detected by XRD after CO2 oxidation step.
The results are evident for two effects: one is the improved thermal stability of BaFeHFr obtained from BaFeHAl compared with direct co-precipitation, and the other is activation of 6%K/BaFeHAl. The latter is related to the formation of BaFeHFr phase with a higher activity. BaCO3 and Al2O3 cannot participate in redox cycles, and Fe3O4 formed together with BaFeHFr (Table 2) has low activity in RWGS-CL [12] and its content is small. The 6%K/BaFeHAl was characterized by XRD, N2-adsorption, The CO capacity of K-promoted BaFeHAl reached a maximum at 550 • C, increasing by a factor of 1.75 compared to the value at 450 • C (Figure 8a). This is illustrated by changes of the shape of CO production peaks recorded at 450 • C in CO 2 oxidation step (Figure 8b). No changes of catalysts phase composition were detected by XRD after CO 2 oxidation step. The results are evident for two effects: one is the improved thermal stability of BaFeHF r obtained from BaFeHAl compared with direct co-precipitation, and the other is activation of 6%K/BaFeHAl. The latter is related to the formation of BaFeHF r phase with a higher activity. BaCO 3 and Al 2 O 3 cannot participate in redox cycles, and Fe 3 O 4 formed together with BaFeHF r ( Table 2) has low activity in RWGS-CL [12] and its content is small. The 6%K/BaFeHAl was characterized by XRD, N 2 -adsorption, and XPS after reduction at different temperatures. Figure 9 compares the XRD patterns of 6%K/BaFeHF obtained by co-precipitation with the XRD patterns of BaFeHF component of 6%K/BaFeHAl reduced at 550 • C. The latter was derived from the experimental XRD data by deconvolution using Rietveld program that separates reflections related to BaFeHF phase from that characteristic of BaCO 3 and Al 2 O 3 . The XRD patterns of BaFeHF r after reduction of 6%K/BaFeHAl at 500, 550, and 600 • C ( Table 2) were identical. The CO capacity of K-promoted BaFeHAl reached a maximum at 550 °C, increasing by a factor of 1.75 compared to the value at 450 °C (Figure 8a). This is illustrated by changes of the shape of CO production peaks recorded at 450 °C in CO2 oxidation step (Figure 8b). No changes of catalysts phase composition were detected by XRD after CO2 oxidation step.
The results are evident for two effects: one is the improved thermal stability of BaFeHFr obtained from BaFeHAl compared with direct co-precipitation, and the other is activation of 6%K/BaFeHAl. The latter is related to the formation of BaFeHFr phase with a higher activity. BaCO3 and Al2O3 cannot participate in redox cycles, and Fe3O4 formed together with BaFeHFr (Table 2) has low activity in RWGS-CL [12] and its content is small. The 6%K/BaFeHAl was characterized by XRD, N2-adsorption, and XPS after reduction at different temperatures. Figure 9 compares the XRD patterns of 6%K/BaFeHF obtained by co-precipitation with the XRD patterns of BaFeHF component of 6%K/BaFeHAl reduced at 550 °C. The latter was derived from the experimental XRD data by deconvolution using Rietveld program that separates reflections related to BaFeHF phase from that characteristic of BaCO3 and Al2O3. The XRD patterns of BaFeHFr after reduction of 6%K/BaFeHAl at 500, 550, and 600 °C ( Table 2) were identical.  Comparison of XRD patterns of BaFeHF phases obtained by co-precipitation and reduction of BaFeHAl ( Figure 9) shows significant differences although both structures relate to the same space group P63/mmc (194). BaFeHF r displays two reflections at low angles of 2θ = 7.6 and 15.2 • corresponding to planes (002) and (004), respectively, not found in BaFeHF. The relative intensities of XRD peaks are significantly different. In the triad of most intensive peaks corresponding to planes (110), (107), and (114), the relation between intensities is I 110 /I 107 /I 114 = 50/90/100 for 6%K/BaFeHF and 40/100/65 for BaFeHF r . In addition, the XRD reflections related to BaFeHF r are shifted to higher angles compared with BaFeHF. This yields significant changes of the unit cell parameters from a = 5.902 Å, c = 23.235 Å (BaFeHF) to a = 5.801 Å, c = 23.361 Å (BaFeHF r ).
Analysis of these variances of XRD patterns using Rietveld method showed that the atomic distributions in the unit cells of BaFeHF and BaFeHF r are also different (Tables S1 and S2). Though the nomenclature of atoms and their positions in the structure is similar in both cases, the coordinates of all atoms, besides Ba and Fe 1 , are meaningfully dissimilar, as may be clearly seen along the c axis (Z coordinates). This explains the appearance of reflections (002) and (004) for BaFeHF r . The differences of intensities of XRD patterns between BaFeHF co-pre . and BaFeHF r is mainly a result of shifting atomic positions in the hexaferrite structure. Presented in Figure 10 are the oxygen environments of Fe atoms in the framework of both materials. The color of tetrahedra and octahedra corresponds to the color of the groups of Fe atoms in the unit cell as denoted in Figure 1b,c. In both structures the Ba atoms are located at the main points of the lattice symmetry P63/m (Z Ba = 1/4 and 3/4). This means that Ba atoms are the force centers of stability for this type of crystals. In case of BaFeHF co-pre . (Figure 10a) the yellow-orange zone in the unit cell center between Ba atoms is represented by yellow octahedra including two atoms Fe 1 and orange tetrahedra including four atoms Fe 2 . Located at the top and bottom of this zone are zones of gray octahedra including six atoms Fe 5 in each gray zone. Bonding between Ba-atoms and Fe-atoms belonging to the gray zones occurs through brown octahedra including Fe 3 atoms (two atoms at every side). The two atoms Fe 4 denoted by green color are located in Ba-plane m with small deviation from Z axis at half occupancy. These atoms do not have bulk oxygen environment. The character of oxygen environment of Fe atoms in the lattice of BaFeHF r is identical to that in BaFeHF co-pre . (Figure 10b). However, due to shifting of atomic positions in the compressed lattice, the octahedral-tetrahedral oxygen environment in yellow-orange and gray zones is distorted with their expansion along c axis. The most important difference is a significant shift of the positions of green Fe 4 atoms along c axis relative to the plane m of Ba-atoms located perpendicular to the c axis at distances of 1 4 and 3 4 from the origin. This may be clearly seen in Figure 10 according to positions of green spheres representing Fe 4 atoms in unit cells. Due to this shift near these Fe 4 atoms, the bulk oxygen surrounding appears evident for the formation of additional Ba-O-Fe bridges in the structure of BaFeHF r , reinforcing the connection between alternatively stacked spinel blocks. This may explain the higher thermostability of BaFeHF r phase formed by reductive decomposition of BaFeHAl compared with that obtained by co-precipitation. The original valence state of iron ions substituting Al 3+ ions in hexaaluminate structure is Fe 3+ . Therefore, the concentration of anionic vacancies originated from reduction of these ions at the catalyst surface to Fe 2+ is proportional to the ratio of Fe 2+ /Fe 3+ that may be measured by XPS. Shown in Figure 11 are the XPS spectra of the material 6%K/BaFeHAl reduced at 350-550 °C. The existence of iron ions at different valence state and environment follows from deconvolution of Fe2p3/2, core spectra in three signals. Peaks at lowest energy of 709.3-710.0 eV reflect the existence of Fe 2+ ions in the surface layer [50]. The Fe 3+ ions at different oxygen environment in the hexaaluminate lattice are represented in XPS spectra by peaks of higher BE of 710.4-711.6 and 712-714 eV [50]. The values of surface Fe 2+ /Fe 3+ ions ratio were calculated based on their surface concentrations derived from deconvoluted XPS spectra. They are shown in Figure 11b as a function of reduction temperature of 6%BaFeHAl catalyst. The rate of increase of Fe 2+ /Fe 3+ ratio in the range 450-600 °C (9.0 × 10 −3 °C −1 ) is about three times higher than at 350-450 °C (3.5 × 10 −3 °C −1 ), corresponding to deeper reduction of iron in 6%K/BaFeHAl. This may be attributed to higher reducibility of Fe 3+ ions at the surface of BaFeHF phase, yielding more Fe 2+ ions related with oxygen vacancies that play a role of active sites at the CO2 oxidation step. Shown in Figure 11 are the XPS spectra of the material 6%K/BaFeHAl reduced at 350-550 • C. The existence of iron ions at different valence state and environment follows from deconvolution of Fe2p3/2, core spectra in three signals. Peaks at lowest energy of 709.3-710.0 eV reflect the existence of Fe 2+ ions in the surface layer [50]. The Fe 3+ ions at different oxygen environment in the hexaaluminate lattice are represented in XPS spectra by peaks of higher BE of 710.4-711.6 and 712-714 eV [50]. The values of surface Fe 2+ /Fe 3+ ions ratio were calculated based on their surface concentrations derived from deconvoluted XPS spectra. They are shown in Figure 11b as a function of reduction temperature of 6%BaFeHAl catalyst. The rate of increase of Fe 2+ /Fe 3+ ratio in the range 450-600 • C (9.0 × 10 −3 • C −1 ) is about three times higher than at 350-450 • C (3.5 × 10 −3 • C −1 ), corresponding to deeper reduction of iron in 6%K/BaFeHAl. This may be attributed to higher reducibility of Fe 3+ ions at the surface of BaFeHF phase, yielding more Fe 2+ ions related with oxygen vacancies that play a role of active sites at the CO 2 oxidation step.
The existence of iron ions at different valence state and environment follows from deconvolution of Fe2p3/2, core spectra in three signals. Peaks at lowest energy of 709.3-710.0 eV reflect the existence of Fe 2+ ions in the surface layer [50]. The Fe 3+ ions at different oxygen environment in the hexaaluminate lattice are represented in XPS spectra by peaks of higher BE of 710.4-711.6 and 712-714 eV [50]. The values of surface Fe 2+ /Fe 3+ ions ratio were calculated based on their surface concentrations derived from deconvoluted XPS spectra. They are shown in Figure 11b as a function of reduction temperature of 6%BaFeHAl catalyst. The rate of increase of Fe 2+ /Fe 3+ ratio in the range 450-600 °C (9.0 × 10 −3 °C −1 ) is about three times higher than at 350-450 °C (3.5 × 10 −3 °C −1 ), corresponding to deeper reduction of iron in 6%K/BaFeHAl. This may be attributed to higher reducibility of Fe 3+ ions at the surface of BaFeHF phase, yielding more Fe 2+ ions related with oxygen vacancies that play a role of active sites at the CO2 oxidation step. a) Figure 11. XPS spectra of 6%BaFeHAl-30 catalyst after reduction at 350, 450, and 550 °C Fe 2+ , F 3+ Oh, Fe 3+ Th (a); effect of reduction temperature on the surface Fe 2+ /Fe 3+ ions ratio calculated based on XPS spectra deconvolution (b). Since the surface area of reduced 6%K/BaFeHAl catalyst remains about the same at the range of 350-550 • C (Figure 12), the CO capacity of 6%K/BaFeHAl catalyst strongly increased at 500-550 • C after complete reductive transformation of BaFeHAl phase to BaFeHF r . Further increase of the reduction temperature to 600 • C did not change the phase composition of the catalyst ( Table 2) and crystal size of BaFeHF r , BaCO 3 , Al 2 O 3 , and Fe 3 O 4 phases. However, the distortion of nanocrystals aggregates at this temperature reduced the pore volume and decreased the surface area by a factor of~2 ( Figure 12) that caused catalyst deactivation (Figure 8).
Catalysts 2020, 10, x FOR PEER REVIEW 12 of 18 Since the surface area of reduced 6%K/BaFeHAl catalyst remains about the same at the range of 350-550 °C (Figure 12), the CO capacity of 6%K/BaFeHAl catalyst strongly increased at 500-550 °C after complete reductive transformation of BaFeHAl phase to BaFeHFr. Further increase of the reduction temperature to 600 °C did not change the phase composition of the catalyst ( Table 2) and crystal size of BaFeHFr, BaCO3, Al2O3, and Fe3O4 phases. However, the distortion of nanocrystals aggregates at this temperature reduced the pore volume and decreased the surface area by a factor of ~2 ( Figure 12) that caused catalyst deactivation (Figure 8). Reducing 6%K/baFeHAl-30 at 550 °C increased CO capacity (Figure 8a). Thus, this catalyst was tested at this reduction temperature and oxidation temperature range of 450-550 °C where the material is resistant to sintering. CO capacity depicted in Figure 13a, based on data in Figure 13b, indicate that with increasing of oxidation temperature, the CO capacity increased from 350 to 400 (µmol/g) (Tr = 550 °C) and from 208 to 248 (µmol g −1 ) at Tr = 450 °C. According to XRD and N2adsorption data, the structure, phase composition, and crystals size of 6%K/BaFeHAl-30 reduced at 550 °C did not change after these experiments. Therefore, the amount of available active sites (oxygen vacancies) controlled by the reduction did not change during oxidation, indicating that higher oxidation temperature involves additional Fe 2+ ions with lower oxidation ability. Reducing 6%K/baFeHAl-30 at 550 • C increased CO capacity (Figure 8a). Thus, this catalyst was tested at this reduction temperature and oxidation temperature range of 450-550 • C where the material is resistant to sintering. CO capacity depicted in Figure 13a, based on data in Figure 13b, indicate that with increasing of oxidation temperature, the CO capacity increased from 350 to 400 (µmol/g) (T r = 550 • C) and from 208 to 248 (µmol g −1 ) at T r = 450 • C. According to XRD and N 2 -adsorption data, the structure, phase composition, and crystals size of 6%K/BaFeHAl-30 reduced at 550 • C did not change after these experiments. Therefore, the amount of available active sites (oxygen vacancies) controlled by the reduction did not change during oxidation, indicating that higher oxidation temperature involves additional Fe 2+ ions with lower oxidation ability.
Reducing 6%K/baFeHAl-30 at 550 °C increased CO capacity (Figure 8a). Thus, this catalyst was tested at this reduction temperature and oxidation temperature range of 450-550 °C where the material is resistant to sintering. CO capacity depicted in Figure 13a, based on data in Figure 13b, indicate that with increasing of oxidation temperature, the CO capacity increased from 350 to 400 (µmol/g) (Tr = 550 °C) and from 208 to 248 (µmol g −1 ) at Tr = 450 °C. According to XRD and N2adsorption data, the structure, phase composition, and crystals size of 6%K/BaFeHAl-30 reduced at 550 °C did not change after these experiments. Therefore, the amount of available active sites (oxygen vacancies) controlled by the reduction did not change during oxidation, indicating that higher oxidation temperature involves additional Fe 2+ ions with lower oxidation ability. At fixed CO2 concentration in the inlet gas at the CO2 oxidation step of RWGS-CL cycle, the catalysts CO2 capacity is connected with CO2 conversion at given weight hour space velocity of CO2 (WHSVCO2) according to Equation (1). This equation describes well the CO2 conversions calculated according to evolved CO amounts measured at WHSVCO2 of 0.4-0.9 h −1 with 6%K/BaFeHAl-30 catalyst at optimal isothermal RWGS-CL conditions. These data fit well to the CO2 conversions calculated according to Equation (1) assuming the constant CO capacity of 400 µmol g −1 (Figure 14). At fixed CO 2 concentration in the inlet gas at the CO 2 oxidation step of RWGS-CL cycle, the catalysts CO 2 capacity is connected with CO 2 conversion at given weight hour space velocity of CO 2 (WHSV CO2 ) according to Equation (1). This equation describes well the CO 2 conversions calculated according to evolved CO amounts measured at WHSV CO2 of 0.4-0.9 h −1 with 6%K/BaFeHAl-30 catalyst at optimal isothermal RWGS-CL conditions. These data fit well to the CO 2 conversions calculated according to Equation (1) assuming the constant CO capacity of 400 µmol g −1 (Figure 14). The maximal CO 2 conversion of 14.9% was obtained at lowest tested WHSV CO2 = 0.4 h −1 . For further increase of CO 2 conversion, the CO capacity of BaFe-hexaaluminate catalyst should be improved more. The reported data assume that this challenge may be achieved by improvement of hexaaluminate catalyst-further increasing the materials surface area and application of K-promoted high-temperature modification of BaFeHF as a pure phase not diluted with Fe-, Al-oxides, and Ba-carbonate.
Catalysts 2020, 10, x FOR PEER REVIEW 13 of 18 The maximal CO2 conversion of 14.9% was obtained at lowest tested WHSVCO2 = 0.4 h −1 . For further increase of CO2 conversion, the CO capacity of BaFe-hexaaluminate catalyst should be improved more. The reported data assume that this challenge may be achieved by improvement of hexaaluminate catalyst-further increasing the materials surface area and application of K-promoted high-temperature modification of BaFeHF as a pure phase not diluted with Fe-, Al-oxides, and Bacarbonate.

Catalysts Preparation
A series of Ba-Fe-hexaaluminates with Fe to Al exchange degree of 60% denoted as BaFeHAl was prepared by hot co-precipitation strategy as described in [52,53]. The solution of Ba(NO3)2 (Alfa Aesar, Yehud, Israel) and Fe(NO3)3·9H2O (Fisher Chemicals, Yehud, Israel) in hot distilled water after addition of Al(NO3)3 . 9H2O (Riedel de Haën, Or Yehuda, Israel) was acidified to pH~1 with diluted HNO3 (Gadot, Netanya, Israel). This solution was poured under vigorous stirring in excess of (NH4)2CO3 solution heated at 60 °C. The mixed hydroxides were precipitated at pH 7.5-8.0. After aging of obtained slurry at 60 °C for 3 h, filtering, washing with distilled water and in air at 110 °C overnight, the material was calcined in air at 450 °C for 2 h with heating rate of 2 °C min −1 . Finally,

Catalysts Preparation
A series of Ba-Fe-hexaaluminates with Fe to Al exchange degree of 60% denoted as BaFeHAl was prepared by hot co-precipitation strategy as described in [52,53]. The solution of Ba(NO 3 ) 2 (Alfa Aesar, Yehud, Israel) and Fe(NO 3 ) 3 ·9H 2 O (Fisher Chemicals, Yehud, Israel) in hot distilled water after addition of Al(NO 3 ) 3 ·9H 2 O (Riedel de Haën, Or Yehuda, Israel) was acidified to pH~1 with diluted HNO 3 (Gadot, Netanya, Israel). This solution was poured under vigorous stirring in excess of (NH 4 ) 2 CO 3 solution heated at 60 • C. The mixed hydroxides were precipitated at pH 7.5-8.0. After aging of obtained slurry at 60 • C for 3 h, filtering, washing with distilled water and in air at 110 • C overnight, the material was calcined in air at 450 • C for 2 h with heating rate of 2 • C min −1 . Finally, the catalyst was calcined in air at 1200 • C for 3h with heating rate of 2 • C min −1 . Potassium was deposited at the calcined catalyst by incipient wetness impregnation with aqueous solution of K 2 CO 3 (the K 2 CO 3 solution of suitable concentration was added to the catalyst powder at an amount corresponding to the water capacity of the material leaving the material dry) with overnight drying at 110 • C and ending calcination at 450 • C.
The preparation of 100% substituted BaFe-hexaaluminate (hexaferrite-HF) with formula Ba 0.82 Fe 10.74 O 18.1 was conducted by regular hot co-precipitation strategy using a different calcination procedure [53]. The dried samples were ground and then calcined by intermediate steps at 500, 700, 900, 1000, 1100, and 1200 • C (heating rate, 2 • C min −1 ; hold at each step, 10 h). This series was denoted as BaFeHF. Then a series of HF catalysts containing 6%K was prepared by incipient wetness, called 6%K/BaFeHF.
The BaFe-hexaaluminate catalysts with increased surface area and substitution of Al with 60% iron was prepared by carbon templating (CT) [46,47]. A hot (60 • C) acidic aqueous solution of Ba-Fe-Al-nitrates was added under vigorous stirring to the carbon black (Alfa Aesar, Yehud, Israel, 75 m 2 g −1 , bulk density 80-120 g L −1 ) previously hydrophilized by treatment with nitric acid. The amounts of solution and carbon black were selected to yield 45 wt% Ba-Fe-hexaaluminate (calculated as sum of oxides) and 55 wt% carbon in the dried precipitate. A large excess of aqueous (NH 4 ) 2 CO 3 solution heated to 60 • C was poured to the obtained slurry heated to 60 • C under vigorous stirring. Then the slurry was aged for 3 h, filtered, washed and dried by freeze-drying method (Instrument Christ Beta 1-8) for 72 h. The dried precursors were calcined in air at 450 • C for 2 h (2 • C min −1 ) and then at 1000 • C (2 • C min −1 ) for 8 h. The catalyst was designated as BaFeHAl-20. A K-promoted catalyst was then prepared by deposition of 6 wt% potassium as described above to this CT material designated as 6%K/BaFe-HAl-20.
The carbon templating (CT) strategy [46,47] was applied for preparation of BaFeHAl material with Fe-substitution degree of 60% and increased surface area. The carbon black (Alfa Aesar, Yehud, Israel 75 m 2 g −1 , bulk density 80-120 g L −1 ) hydrophilized by treatment with nitric acid at an amount corresponding to the 55 wt% in dried precipitate was added under vigorous stirring to a hot (60 • C) acidified (pH = 1) aqueous solution of Ba-Fe-Al-nitrates. An excess of aqueous solution of ammonium carbonate was poured under stirring to the obtained slurry at 60 • C. After aging of obtained slurry at 60 • C for 3 h, filtering, washing with distilled water and freeze-drying (Instrument Christ Beta 1-8) for 72 h, the material was calcined in air at 450 • C for 2 h with heating rate of 2 • C min −1 . Finally, the catalyst was calcined in air at 1000 • C for 8h with heating rate of 2 • C min −1 . The CT-catalyst was designated as BaFeHAl-20. Potassium was deposited at the calcined catalyst by incipient wetness impregnation with aqueous solution of K 2 CO 3 with overnight drying at 110 • C and ending calcination at 450 • C. This CT-catalyst was designated as 6%K/BaFeHAl-20.

Catalysts Characterization
The N 2 adsorption-desorption isotherms were recorded for all catalysts after outgassing under vacuum at 250 • C for 2 h using NOVA 3200e (Quantachrome, Anton Paar QuantaTec Inc., Boynton Beach, Florida, USA) instrument. The surface area, pore size, and volume of the catalysts were calculated from these isotherms using conventional BET (Brunauer-Emmett-Teller) [57] and BJH (Barrett-Joyner-Halenda) [58] methods. The catalysts chemical composition was measured by EDS (Energy Dispersive Spectroscopy) method using Quanta-200, SEM-EDAX (FEI Co., Hillsboro, OR, USA) instrument. The Panalytical Empyrean Powder Diffractometer (Cambridge, UK) equipped with position-sensitive detector X'Celerator fitted with a graphite monochromator, at 40 kV and 30 mA, was used for collecting of catalysts XRD patterns. They were analyzed with software developed by Crystal Logic Inc. (Los Angeles, CA, USA). The SBDE ZDS computer search/match program coupled with the ICDD database was used for phases identification. The Rietveld refinement of the XRD profile implementing the DBWS-9807 program (Atlanta, GA, USA) was applied for calculating of phases content in catalytic materials.
The SEM images of catalytic materials were recorded using Quanta-200, SEM-EDAX, FEI Co. instrument (Hillsboro, OR, USA). TPR measurements were performed using Chemisorption Analyzer Autochem II 2920 instrument (Micrometrics Norcross, Georgia, USA) equipped with TCD detector. TPR was done in 10%H 2 /Ar flow of 5 mL min −1 . The X-ray photoelectron spectrometer ESCALAB 250 apparatus (Hamamatsu City, Japan) working at ultrahigh vacuum (1 × 10 −9 bar) with an Al Kα X-ray source and a monochromator was applied for collecting of XPS spectra. Fitting a sum of the single component lines to the experimental data by means of a non-linear least-square curve was used for identification of the spectral components of Fe signals. EX05 argon gun system performed controlled removal of surface layers. Cleaning the surface from adsorbed species before recording the XPS spectra was done using the EX05 argon gun system.

Catalysts Testing
The reduction-oxidation steps of CO 2 conversion experiments were conducted in a quartz U-tube reactor at catalyst loadings of 0.10-0.22 g using a Chemisorption Analyzer AutoChem II 2920 instrument Norcross, Georgia, USA, Micrometrics Co. equipped with mass spectrometer Cirrus 2, MKS detector. The catalyst powder (fraction 25-180 µm) was fixed at the isothermal section of the reactor between glass-wool plugs. The inlet CO 2 concentration at oxidation steps in all testing experiments was 5% vol. The detected CO MS signals (m/z = 28) intensities were calibrated with CO/He mixtures of corresponding varied compositions. The RWGS-CL cycles were conducted at different temperatures in the range of 350-600 • C. The sample was first reduced in 10%H 2 /Ar flow of 15 mL min −1 for 20 min. This step was followed by a 20 min He flushing (50 mL min −1 ) and then a catalyst oxidation step (CO 2 reduction to CO) was carried out in 5% CO 2 /He flow of 15 mL min −1 for 20 min. The system was flushed with He again for 20 min, and the cycle was repeated. First, the catalyst was stabilized in three consecutive redox cycles followed by additional three cycles of stable operation. CO peaks recorded in RWGS-CL isothermal cycles # 4-6 at 450 • C with 6%K/BaFeHAl catalyst are shown in Figure 15.
Catalysts 2020, 10, x FOR PEER REVIEW 15 of 18 15 mL min −1 for 20 min. This step was followed by a 20 min He flushing (50 mL min −1 ) and then a catalyst oxidation step (CO2 reduction to CO) was carried out in 5% CO2/He flow of 15 mL min −1 for 20 min. The system was flushed with He again for 20 min, and the cycle was repeated. First, the catalyst was stabilized in three consecutive redox cycles followed by additional three cycles of stable operation. CO peaks recorded in RWGS-CL isothermal cycles # 4-6 at 450 °C with 6%K/BaFeHAl catalyst are shown in Figure 15. The catalysts performance was characterized by two parameters: CO capacity (COcap.), and CO2 conversion (XCO2). Integration of calibrated CO peaks areas recorded in the oxidation step of the cycle yielded amount of µmols of CO/g.cat, denoted as catalysts CO capacity at selected testing conditions. The average CO2 conversion corresponding to the oxidation part of the RWGS-CL cycler was calculated according to Equation (1): (1) The catalysts performance was characterized by two parameters: CO capacity (CO cap. ), and CO 2 conversion (X CO2 ). Integration of calibrated CO peaks areas recorded in the oxidation step of the cycle yielded amount of µmols of CO/g.cat, denoted as catalysts CO capacity at selected testing conditions. The average CO 2 conversion corresponding to the oxidation part of the RWGS-CL cycler was calculated according to Equation (1): where, M CO -amount of CO formed during oxidation step of the RWGS-CL cycle (mmol), M CO2 -amount of CO 2 fed to reactor during oxidation step of the cycle (mmol), WHSV CO2 (g CO 2 gcat −1 h −1 ), t 0 -the length of the CO 2 -oxidation cycle (h).

Conclusions
The Fe-substituted Ba-hexaaluminates are active catalysts for RWGS reaction conducted in chemical looping mode. Increasing of the degree of substitution of Al 3+ for Fe 3+ ions in co-precipitated Ba-hexaaluminate from 60% to 100%, increased its surface area, and promotion with potassium increased the CO capacity in isothermal RWGS-CL runs at temperatures 350-450 • C where the hexaaluminate structure is stable. At higher temperatures, the fully Fe-substituted hexaaluminate-6%K/Ba-hexaferrite-undergo reductive decomposition at the H 2 -reduction step of RWGS-CL cycle, yielding Fe-oxides, Ba-oxide, and Ba-carbonate. However, the partially Fe-substituted 6%K/BaFe-hexaaluminate after H 2 -reduction at >450 • C is transformed to a thermally stable modification of Ba-hexaferrite that contains additional Ba-O-Fe bridges in its structure, reinforcing the connection between alternatively stacked spinel blocks. The structure and surface area of this modification of Ba-hexaferrite promoted with potassium are stable against H 2 reduction up to 550 • C. This allows getting higher CO 2 capacity in isothermal tests at higher temperatures. The stable 6%K/BaFe-hexaferrite derived from 6%K/BaFe-hexaaluminate display higher concentration of surface oxygen vacancies compared with BaFe-hexaaluminate reflected by greater Fe 2+ /Fe 3+ ratio according to XPS. Conducting of RWGS-CL cycles in isothermal mode at 550 • C where hexaferrite structure and texture is stable, further increases the materials CO capacity. It was demonstrated the direct connection between CO capacity measured in RWGS-CL experiments and calculated CO 2 conversion.