Exploring the stability of fe–ni alloy nanoparticles exsolved from double-layered perovskites for dry reforming of methane

catalysts Article Exploring the Stability of Fe–Ni Alloy Nanoparticles Exsolved from Double-Layered Perovskites for Dry Reforming of Methane Alfonso J. Carrillo * and JoséManuel Serra *   Citation: Carrillo, A.J.; Serra, J.M. Exploring the Stability of Fe–Ni Alloy Nanoparticles Exsolved from DoubleLayered Perovskites for Dry Reforming of Methane. Catalysts 2021,11, 741. https://doi.org/10.3390/catal11060741 Academic Editor: Kalliopi Kousi Received: 19 May 2021 Accepted: 15 June 2021 Published: 16 June 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). Instituto de Tecnología Química, Universitat Politècnica de València-CSIC, Avda. Los Naranjos s/n, 46022 Valencia, Spain *Correspondence: [email protected].es (A.J.C.); [email protected].es (J.M.S.) Abstract: Exsolution is emerging as a promising route for the creation of nanoparticles that remain anchored to the oxide support, imparting remarkable stability in high temperature chemical processes such as dry reforming of methane. This process takes place at temperatures around 850 ◦ C, which causes sintering-related issues in catalysts prepared using conventional impregnation methods, which could be overcome by using exsolution functionalized oxides. In this work, FeNi 3 alloy nanoparticles exsolved from Sr 2 Fe x Ni 1-x MoO 6 - δ double-layered perovskites were evaluated as a dry reforming catalyst, paying special attention to structure–activity relationships. Our results indicate that increasing the Ni content favors the nanoparticle dispersion, eventually leading to increased CO 2 and CH 4 conversions. The exsolved nanoparticles presented remarkable nanoparticle size (ca. 30 nm) stability after the 10 h treatment, although the formation of some phase segregations over the course of the reaction caused a minor decrease in the nanoparticle population. Overall, the results presented here serve as materials processing guidelines that could find further potential use in the design of more efficient (electro)catalysts in other fuel production or energy conversion technologies. Keywords: exsolution; alloy nanoparticle; dry methane reforming; double-layer perovskite 1. Introduction A myriad of chemical processes for the production of fuels, pharmaceuticals, fertilizers, or for environmental remediation rely on the use of a metal catalyst dispersed onto oxide supports [ 1 ]. These metallic nanoparticles represent the active sites in the heterogeneous reactions and are conventionally deposited on oxide supports (e.g., Al 2 O 3 , ZrO 2 , CeO 2 ) using impregnation methods. However, in some chemical processes operating at high temperatures, metal nanoparticles created by impregnation can grow due to sintering related processes, resulting in an increase in the particle size and a decrease in the catalyst dispersion [ 2 ], which will eventually negatively affect their catalytic activity (Figure 1). This is, for example, the case of the dry reforming of methane (DMR) reaction (Equation (1)). CH4+CO2→2CO +2H2∆Hr=247 kJ mol−1(1) DMR is a promising technology for the conversion of CH 4 and CO 2 into syngas, which could be further transformed into liquid fuels through Fischer–Tropsch units. However, DMR faces two major issues related to (i) carbon depositions and (ii) metallic nanoparticle sintering, the latter being caused by the elevated process temperatures (600-900 ◦C) [3]. In order to overcome these challenges, several alternatives have been pursued to obtain more stable metallic catalysts and supports. One method that has attracted attention in the past few years is nanoparticle exsolution due to the high stability that it confers to the metallic catalyst and the strong interaction with the support [ 2 , 4 – 6 ]. Exsolution consists of the creation of nanoparticles by the migration, under a reductive atmosphere, of cations contained in the bulk of the oxide support that nucleate and grow in the surface. Since the nanoparticles grow from inside, in contrast to the deposited particles, they remain anchored to the oxide surface, which provides high stability against sintering and Catalysts 2021,11, 741. https://doi.org/10.3390/catal11060741 https://www.mdpi.com/journal/catalysts Catalysts 2021,11, 741 2 of 14 coking ( Figure 1 ) [ 7 ]. Most works have been focused on nanoparticle exsolution on perovskites [ 2 , 4 , 5 ], although lately, exsolution has also been observed on other oxides like in fluorite structures [ 8 , 9 ]. Perovskites, of the ABO 3 formula (with A = La, Sr, Ca, Ba, and B, a transition metal cation) have been widely explored as catalysts for DMR based on their high thermal stability [ 10 ]. Therefore, it became apparent that exsolution-decorated metallic catalysts based on perovskites could find a remarkable niche of application in DMR [ 10 , 11 ] or other high-temperature thermocatalytic processes such as the chemical looping reforming of methane [ 12 – 15 ]. To our knowledge, the first work devoted to DMR using nanoparticle exsolution was by Zubenko et al., who developed Re–Ni–Fe alloy nanoparticles exsolved from LaNi 0.2 Re 0.2 Fe 0.6 O 3+δ -La 3 ReO 8 oxide with x < 0.2, with ca. 100% CH 4 and CO 2 conversion at 900 ◦ C [ 16 ]. One of the additional advantages of exsolution is the ability to create alloy nanoparticles in a facile manner by just adjusting the B-site composition in the perovskite host [ 16 ]. As illustrated by Zubenko et al., alloy exsolution can result in high catalytic activity. Other works that have explored bimetallic exsolution for DMR synthesized Fe–Ni, Co–Fe, and Co–Ni alloys exsolved from (La 0.75 Sr 0.25 )(Cr 0.5 Fe 0.35 Ni 0.15 )O 3−δ [ 17 ], Pr 0.5 Ba 0.5 Mn 0.85 Co 0.15 O 3−δ (infiltrated with Fe) [ 18 ], Pr 0.5 Ba 0.5 Mn 0.85 Co 0.05 Ni 0.1 O 3−δ [ 19 ], and La(Co0.1Ni0.9)0.5Fe0.5O3[20], respectively. The reported CO2conversion values using exsolved alloys range from 81% for Fe–Ni [ 17 ] to 30% for Co–Fe [ 18 ], indicating a better DMR performance of the former. Interestingly, previous reports on bimetallic Fe–Ni catalysts over La 2 O 3 showed that this catalyst has no activity to DMR due to encapsulation of Ni by a LaFeO 3 phase [ 21 ]. Besides the work of Papargyriou et al., for DMR, several authors have explored Fe–Ni alloy exsolution, especially for solid oxide fuel cells and electrolyzers [ 22 – 25 ]. From these works, it can be noted that the double layer perovskite family Sr 2 Fe 1-x Mo x O 6 has been found to have excellent exsolution properties, in which Ni incorporation on the B-site allows for exsolving Fe–Ni alloy nanoparticles after a reduction in H 2 . Despite the promise, these oxides have not been tested for DMR yet. For this reason, in this work, we performed a systematic analysis of the intrinsic (B-site composition) and extrinsic factors (time and temperature) to elucidate their effect on the nanoparticle size and dispersion, and, eventually, the structure–activity relationships in DMR. Catalysts 2021, 11, x FOR PEER REVIEW 2 of 16 Figure 1. Schematic depicting the morphological evolution differences of dispersed metallic nanoparticles on oxide supports when prepared by impregnation or exsolution methods. In order to overcome these challenges, several alternatives have been pursued to obtain more stable metallic catalysts and supports. One method that has attracted attention in the past few years is nanoparticle exsolution due to the high stability that it confers to the metallic catalyst and the strong interaction with the support [2,4–6]. Exsolution consists of the creation of nanoparticles by the migration, under a reductive atmosphere, of cations contained in the bulk of the oxide support that nucleate and grow in the surface. Since the nanoparticles grow from inside, in contrast to the deposited particles, they remain anchored to the oxide surface, which provides high stability against sintering and coking (Figure 1) [7]. Most works have been focused on nanoparticle exsolution on perovskites [2,4,5], although lately, exsolution has also been observed on other oxides like in fluorite structures [8,9]. Perovskites, of the ABO 3 formula (with A = La, Sr, Ca, Ba, and B, a transition metal cation) have been widely explored as catalysts for DMR based on their high thermal stability [10]. Therefore, it became apparent that exsolution-decorated metallic catalysts based on perovskites could find a remarkable niche of application in DMR [10,11] or other high-temperature thermocatalytic processes such as the chemical looping reforming of methane [12–15]. To our knowledge, the first work devoted to DMR using nanoparticle exsolution was by Zubenko et al., who developed Re–Ni–Fe alloy nanoparticles exsolved from LaNi 0.2 Re 0.2 Fe 0.6 O 3+δ -La 3 ReO 8 oxide with x < 0.2, with ca. 100% CH 4 and CO 2 conversion at 900 °C [16]. One of the additional advantages of exsolution is the ability to create alloy nanoparticles in a facile manner by just adjusting the B-site composition in the perovskite host [16]. As illustrated by Zubenko et al., alloy exsolution can result in high catalytic activity. Other works that have explored bimetallic exsolution for DMR synthesized Fe–Ni, Co–Fe, and Co–Ni alloys exsolved from (La 0.75 Sr 0.25 )(Cr 0.5 Fe 0.35 Ni 0.15 )O 3−δ [17], Pr 0.5 Ba 0.5 Mn 0.85 Co 0.15 O 3−δ (infiltrated with Fe) [18], Pr 0.5 Ba 0.5 Mn 0.85 Co 0.05 Ni 0.1 O 3−δ [19], and La(Co 0.1 Ni 0.9 ) 0.5 Fe 0.5 O 3 [20], respectively. The reported CO 2 conversion values using exsolved alloys range from 81% for Fe–Ni [17] to 30% for Co–Fe [18], indicating a better DMR performance of the former. Interestingly, previous reports on bimetallic Fe–Ni catalysts over La 2 O 3 showed that this catalyst has no activity to DMR due to encapsulation of Ni by a LaFeO 3 phase [21]. Besides the work of Papargyriou et al., for DMR, several authors have explored Fe–Ni alloy exsolution, especially for solid oxide fuel cells and electrolyzers [22–25]. From these works, it can be noted that the double layer perovskite family Sr 2 Fe 1-x Mo x O 6 has been found to have excellent exsolution properties, in which Ni incorporation on the B-site allows for exsolving Fe–Ni alloy nanoparticles after a reduction in H 2 . Despite the promise, these oxides have not been tested for DMR yet. For this reason, in this work, we performed a systematic analysis of the intrinsic (B-site composition) and extrinsic factors (time and temperature) to elucidate Figure 1. Schematic depicting the morphological evolution differences of dispersed metallic nanoparticles on oxide supports when prepared by impregnation or exsolution methods. 2. Results 2.1. Understanding Fe–Ni Alloy Exsolution on Sr2FexNi1-xMoO6-δ A series of four samples with Sr 2 Fe x Ni 1-x MoO 6-δ composition (x = 0.35–1, see Table 1) were synthesized via the Pechini method varying the Fe/Ni content. The X-ray patterns depicted in Figure 2a show that all the samples exhibited the tetragonal double-layer perovskite (ICSD 150701, space group, I4/m) as the main phase after the calcination treatment at 1100 ◦ C. However, the presence of a secondary impurity phase corresponding to SrMoO 4 was observed for the four samples, in agreement with previous reports in Catalysts 2021,11, 741 3 of 14 the literature [ 26 ], caused by the high concentration of Mo above the solid solubility [ 27 ]. The presence of the impurity decreased with increasing amounts of Ni from 33.4 wt% for SFMO to 9.3 wt% for SFNM35, as analyzed by Rietveld refinement (see example in Figure S1). The increasing presence of Ni stabilized the double-layered perovskite due to its lower oxidation state (+2), which facilitated the presence of Mo in +6 while maintaining charge neutrality. In order to drive the exsolution of metallic nanoparticles, the as-synthesized perovskite powders were treated with a 5% H 2 atmosphere at 900 ◦ C for 2 h. Figure 2b depicts the XRD patterns of the four samples after such treatment. It can be seen that the amount of SrMoO 4 impurities decreased after the treatment in H 2 , whereas for SFNM50 and SFNM35, the appearance of minor impurities of a Ruddlesden–Popper-like phase Sr 3 MoO 7 was observed. Nevertheless, the double-layered perovskite phase purity increased for the four samples upon hydrogen treatment (Table 1). For instance, SFNM65 purity increased from 83.5 to 86.7%. Importantly, after the 2 h treatment, a negligible phase transition from the double layer perovskite structure to a Ruddlesden–Popper could be ascertained, which was observed previously with other similar compositions (e.g., Sr 2 Fe 1.35 Mo 0.45 Co 0.2 O 6-δ ) [ 28 ]. The differences between both works might arise from the different concentration of Fe and Mo on the B-site. In order to provide more insights on the crystallographic changes that occur upon exsolution, we conducted an in situ XRD analysis by heating the SFNM50 sample up to 800 ◦ C under a 5% H 2 atmosphere. Figure 2 c shows the evolution of the ( 112 ) reflection located at ca. 32 ◦ (room temperature). Upon heating in the H 2 reductive atmosphere, the main reflection gradually shifted toward lower 2 θ values. This behavior can be ascribed to the lattice expansion originated by the formation of oxygen vacancies upon reduction and confirms the rise in cell volume as determined from the ex-situ XRD analyses (Table 1). Unfortunately, we were not able to detect the appearance of metal nanoparticles with the current setup, although, as it will be discussed later, these conditions led to metallic nanoparticle exsolution ( Figure 3 ). Figure 2d collects the temperature evolution of the cell volume and lattice parameters as determined by Rietveld refinement. Cell volume for SFNM50 increased from ca. 244 Å 3 at 100 ◦ C to ca. 251 Å 3 at 800 ◦ C. Both aand c-axes from the tetragonal double-layer perovskite lattice elongated upon heating up in the H2atmosphere. Now we turn to the morphological analyses of the materials after the exsolution treatment. Figure 3depicts the SEM images of the four Sr 2 Fe x Ni 1-x MoO 6-δ samples after 2 h exposure to a 5% H 2 atmosphere at 900 ◦ C. As observed in the SEM images, all the samples exhibited the presence of well-dispersed nanoparticles covering the surface of the porous perovskite except for the sample without Ni, SFMO. Based on this fact, it seems that the presence of Ni on the B-site boosts the metallic nanoparticle exsolution. In the literature, there is a certain discrepancy about the possible exsolution of Fe from SFMO materials. On one hand, Lv and coauthors stated that Sr 2 Fe 1.5 MoO 6-δ was stable under reductive environments at high temperatures without evidence of metallic Fe nanoparticle exsolution [ 28 ]. They obtained the segregation energy of Fe through DFT calculations, which was 0.91 eV, whereas that of Co was − 0.64 eV. These results show that Co exsolution is more thermodynamically favored with respect to Fe in SFMO-like double-layered perovskites. On the other hand, Chen et al. obtained a metallic Fe exsolution from Sr 2 Fe 1.5+x Mo 0.5 O 6-δ (x = 0–0.1) [ 29 ]. In that work, the authors increased the concentration of Fe on the B-site to promote the exsolution of Fe nanoparticles of about 95 nm after 3 h exsolution at 850 ◦ C. Thus, it seems that without Fe excess, as was the case of the present work, Fe exsolution will unlikely happen due to thermodynamic unfavorability, which can be altered by compositional modifications. In the case of the three samples with the presence of Ni on the B-site, nanoparticle exsolution was observed on the surface of the oxide. The nanoparticles were well dispersed over the double-layered perovskite, without evidence of preferential location for the exsolution. Regarding the size, most of the nanoparticles presented diameters in the 20–50 nm range. A more thorough analysis of the nanoparticle size and distribution, and its dependence with exsolution conditions will be covered in the following sections. Catalysts 2021,11, 741 4 of 14 Table 1. Label, formula, and crystallographic data of the four Sr2FexNi1-xMoO6-δsamples before and after exsolution. Label Formula As Prepared After Exsolution 2 h, 900 ◦C, 5% H2 Crystal Phase (Space Group) Cell Vol. (Å3)Phase Purity % Crystal Phase (Space Group) Cell Vol. (Å3)Phase Purity % SFMO Sr2FeMoO6-δTetragonal (I4/m) 244.2 66.6 Tetragonal (I4/m) 244.9 69.1 SFNM65 Sr2Fe0.65Ni0.35MoO6-δTetragonal (I4/m) 244.4 83.5 Tetragonal (I4/m) 244.6 86.7 SFNM50 Sr2Fe0.5Ni0.5MoO6-δTetragonal (I4/m) 243.9 86.6 Tetragonal (I4/m) 244.4 90.6 SFNM35 Sr2Fe0.35Ni0.65MoO6-δTetragonal (I4/m) 243.8 90.7 Tetragonal (I4/m) 244.5 92.2 Catalysts 2021, 11, x FOR PEER REVIEW 4 of 16 Table 1. Label, formula, and crystallographic data of the four Sr 2 Fe x Ni 1-x MoO 6-δ samples before and after exsolution. Label Formula As Prepared After Exsolution 2 h, 900 °C, 5% H 2 Crystal Phase (Space Group) Cell Vol. (Å 3 ) Phase Purity % Crystal Phase (Space Group) Cell Vol. (Å 3 ) Phase Purity % SFMO Sr 2 FeMoO 6-δ Tetragonal (I4/m) 244.2 66.6 Tetragonal (I4/m) 244.9 69.1 SFNM65 Sr 2 Fe 0.65 Ni 0.35 MoO 6-δ Tetragonal (I4/m) 244.4 83.5 Tetragonal (I4/m) 244.6 86.7 SFNM50 Sr 2 Fe 0.5 Ni 0.5 MoO 6-δ Tetragonal (I4/m) 243.9 86.6 Tetragonal (I4/m) 244.4 90.6 SFNM35 Sr 2 Fe 0.35 Ni 0.65 MoO 6-δ Tetragonal (I4/m) 243.8 90.7 Tetragonal (I4/m) 244.5 92.2 Figure 2. (a) XRD diffractograms of the Sr 2 Fe x Ni 1-x MoO 6-δ materials as-synthesized (calcination at 1100 °C in air). (b) XRD diffractograms of the Sr 2 Fe x Ni 1-x MoO 6-δ materials after the exsolution treatment (5% H 2 , 900 °C, 2 h). (c) In situ XRD experiment performed to SFNM50, heating up to 800 °C under 5% H 2 . The selected zone depicts the most intense reflection of the tetragonal double layer perovskite. (d) Tetragonal lattice parameters (a,c) and lattice volume variation extracted from the in situ XRD tests for the SFNM50 sample. The schematic inside the figure depicts the tetragonal double layer perovskite crystalline structure (green, red, purple, and brown spheres are Sr, O, Mo, and Fe/Ni atoms, respectively). Figure 2. ( a ) XRD diffractograms of the Sr 2 Fe x Ni 1-x MoO 6-δ materials as-synthesized (calcination at 1100 ◦ C in air). ( b ) XRD diffractograms of the Sr 2 Fe x Ni 1-x MoO 6-δ materials after the exsolution treatment (5% H 2 , 900 ◦ C, 2 h). ( c ) In situ XRD experiment performed to SFNM50, heating up to 800 ◦ C under 5% H 2 . The selected zone depicts the most intense reflection of the tetragonal double layer perovskite. ( d ) Tetragonal lattice parameters ( a , c ) and lattice volume variation extracted from the in situ XRD tests for the SFNM50 sample. The schematic inside the figure depicts the tetragonal double layer perovskite crystalline structure (green, red, purple, and brown spheres are Sr, O, Mo, and Fe/Ni atoms, respectively). Next, we analyzed SFNM65 samples with TEM in order to better understand the nature of the exsolved nanoparticles. Figure 4a shows the HRTEM images of the SFNM65 sample after 2 h exsolution under 5% H 2 at 900 ◦ C. The presence of exsolved nanoparticles of about 20 nm size were observed, which were partly submerged into the perovskite backbone. This characteristic anchoring, typical of nanoparticle exsolution, avoids sintering Catalysts 2021,11, 741 5 of 14 with neighboring particles due to the high degree of attachment of the nanoparticles to the oxide surface. Figure 4b–f shows the HAADF-STEM image and EDX mapping of the five elements that compose the double-layer perovskite. These images provide important insights on the composition of the exsolved nanoparticles. As can be seen in the EDX mapping in Figure 4c,d, the nanoparticles were composed of both Ni and Fe, thus the reductive treatment performed in this work resulted in the exsolution of the Ni–Fe alloy nanoparticles from Sr 2 Fe x Ni 1-x MoO 6-δ materials while maintaining the double-layer perovskite structure. As observed in Figure 4c–g, the presence of the initial five elements that composed the perovskite was detected in the backbone, whereas only Fe and Ni were observed in the exsolved nanoparticles. Fe–Ni alloy exsolution has been previously reported in Sr–Fe–Mo based perovskites [ 22 , 30 – 32 ], titanates [ 25 ], and chromites [ 17 , 33 ]. In most of these cases, the composition of the exsolved alloy was FeNi 3 . For instance, in the work of Du et al., they obtained FeNi 3 exsolution from Sr 2 FeMo 0.65 Ni 0.35 O 6−δ after 10 h of treatment in pure H 2 [ 22 ]. Increased exposure times and higher H 2 concentrations, as performed by Du et al. compared to our work, commonly resulted in larger particle sizes [ 34 , 35 ]. Formation of larger particles could eventually ease their detection by XRD, and, thus the identification of the actual alloy composition of the exsolved nanoparticles. For that reason, we performed a 20 h exsolution treatment at 900 ◦ C in a 5% H 2 atmosphere. Figure 5a depicts a comparison of the XRD patterns for SFNM65 as-synthesized, and after 2 and 20 h exsolution. The exsolution time increase lead to several changes in the crystal structure of the SFNM65 material. First, the impurities of SrMoO 4 disappeared. Second, there was a remarkable phase transformation, resulting in the formation of a Ruddlesden–Popper phase Sr 3 FeMoO 7 (ICSD 152243). This result is in line with previous reports by Du et al. (FeNi 3 exsolution from Sr 2 FeMo 0.65 Ni 0.35 O 6−δ ) [ 22 ] and Lv et al. (CoFe exsolution on Sr 2 (Fe 1.35 Mo 0.45 Co 0.2 ) 1-z O 6−δ ) [ 28 ], in which the Ruddlesden–Popper phase appeared after 10 h and 4 h treatment, respectively. Third, the presence of a Fe–Ni alloy with composition FeNi 3 (cubic, ICSD 5116, space group, Pm-3m) was detected, with the main reflections appearing at 2 θ = 44.0 and 51.3 ◦ . For the sake of comparison, the main XRD reflections of Fe and Ni metal are plotted in Figure 5a, which did not correspond to any of the peaks present in the SFNM65 material after 20 h exsolution treatment. The detection of the alloy phase was possible by the larger particle size obtained after using a more prolonged treatment (20 h instead of 2 h). As observed in Figure 5b, the longer exsolution treatment time resulted in larger nanoparticles, with a particle size of 65 nm, although particles as large as 110 nm were also observed, which facilitated the detection of the exsolved alloy phase by XRD. Although the goal of this treatment was to determine the composition of the exsolved alloy nanoparticles, it also provided information about the effects of prolonged exsolution treatment on the particle size, and eventually on the perovskite backbone stability. Here, it is shown that the particle size can be modified by exposing the materials to more prolonged times, however, it results in larger particles, and, especially, in a phase transition of the perovskite backbone, which might have detrimental effects on the catalytic activity. Our results, however, are in contrast with Wang et al., who did not observe particle size enlargement when increasing the treatment time from 3 to 40 h in humidified H 2 at 800 ◦ C using Sr 2 Fe 1.3 Ni 0.2 Mo 0.5 O 6-δ sintered bars (exsolution of 26 nm Fe–Ni alloys) [ 23 ]. Probable causes from the discrepancies of both works can arise from differences in the materials’ composition and processing. The use of dense sintered bars could affect the mechanism of nanoparticle growth, since it has been shown that smaller grain sizes could favor the exsolution extent [36]. Catalysts 2021,11, 741 6 of 14 Catalysts 2021, 11, x FOR PEER REVIEW 5 of 16 Now we turn to the morphological analyses of the materials after the exsolution treatment. Figure 3 depicts the SEM images of the four Sr 2 Fe x Ni 1-x MoO 6-δ samples after 2 h exposure to a 5% H 2 atmosphere at 900 °C. As observed in the SEM images, all the samples exhibited the presence of well-dispersed nanoparticles covering the surface of the porous perovskite except for the sample without Ni, SFMO. Based on this fact, it seems that the presence of Ni on the B-site boosts the metallic nanoparticle exsolution. In the literature, there is a certain discrepancy about the possible exsolution of Fe from SFMO materials. On one hand, Lv and coauthors stated that Sr 2 Fe 1.5 MoO 6-δ was stable under reductive environments at high temperatures without evidence of metallic Fe nanoparticle exsolution [28]. They obtained the segregation energy of Fe through DFT calculations, which was 0.91 eV, whereas that of Co was −0.64 eV. These results show that Co exsolution is more thermodynamically favored with respect to Fe in SFMO-like double-layered perovskites. On the other hand, Chen et al. obtained a metallic Fe exsolution from Sr 2 Fe 1.5+x Mo 0.5 O 6-δ (x = 0–0.1) [29]. In that work, the authors increased the concentration of Fe on the B-site to promote the exsolution of Fe nanoparticles of about 95 nm after 3 h exsolution at 850 °C. Thus, it seems that without Fe excess, as was the case of the present work, Fe exsolution will unlikely happen due to thermodynamic unfavorability, which can be altered by compositional modifications. In the case of the three samples with the presence of Ni on the B-site, nanoparticle exsolution was observed on the surface of the oxide. The nanoparticles were well dispersed over the double-layered perovskite, without evidence of preferential location for the exsolution. Regarding the size, most of the nanoparticles presented diameters in the 20–50 nm range. A more thorough analysis of the nanoparticle size and distribution, and its dependence with exsolution conditions will be covered in the following sections. Figure 3. SEM micrographs of the Sr 2 Fe x Ni 1-x MoO 6-δ materials after the exsolution process carried out at 900 °C for 2 h under a 5% H 2 atmosphere. Next, we analyzed SFNM65 samples with TEM in order to better understand the nature of the exsolved nanoparticles. Figure 4a shows the HRTEM images of the SFNM65 sample after 2 h exsolution under 5% H 2 at 900 °C. The presence of exsolved nanoparticles of about 20 nm size were observed, which were partly submerged into the perovskite backbone. This characteristic anchoring, typical of nanoparticle exsolution, avoids sintering with neighboring particles due to the high degree of attachment of the nanoparticles to the oxide surface. Figure 4b–f shows the HAADF-STEM image and EDX mapping of Figure 3. SEM micrographs of the Sr 2 Fe x Ni 1-x MoO 6-δ materials after the exsolution process carried out at 900 ◦C for 2 h under a 5% H2atmosphere. Catalysts 2021, 11, x FOR PEER REVIEW 7 of 16 . Figure 4. (a) TEM image of the SFNM65 material after 2 h exsolution at 900 °C under 5% H 2 , depicting the exsolved nanoparticles. (b) HAADF-STEM image of SFNM65 (c–g) EDX mapping images for Ni, Fe, Mo, Sr, and O, respectively, showing that the exsolved nanoparticles are just composed by Fe and Ni. Figure 5. (a) Comparative of the XRD patterns of the SFNM65 material as-synthesized, and after 2 h and 20 h exsolution at 900 °C in a 5% H 2 atmosphere. Crystal structure of a Sr 3 FeMoO 7 Ruddlesden–Popper phase. Green, brown, and red spheres represent Sr, Fe/Mo, and oxygen atoms, respectively. Structure drawn with VESTA [37]. (b) SEM image of the SFNM65 material after 20 h exsolution. Fe and Ni reflections are shown in (a) for the sake of comparison. An alternative way to tune the particle size in exsolution is with temperature [12,34,35,38]. In order to explore this aspect, we treated the double layer perovskite materials at 700 and 800 °C and compared them with the original exsolution temperature, 900 °C. As observed in Figure 6a, there were remarkable differences on the exsolved nanoparticle sizes when lowering the temperature. Namely, the exsolved nanoparticle sizes (d 50 ) were 15.6 ± 4.6, 24.4 ± 5.4, and 31.0 ± 8.4 nm at 600, 700, and 900 °C, respectively. The monotonic increase in the particle size with temperature observed in this work was in line with previous reports that indicate, based on the nucleation theory, that increasing the temperature particularly affects the particle growth [12]. It can also be observed that lowering the temperature (700 °C) resulted in a more uniform particle size distribution, whereas at 900 °C, the distribution was much wider (Figure 6b). On the other hand, we did not observe a clear trend when analyzing the effect of temperature in the exsolved particle population (Figure 6c). The treatment at 700 °C resulted in a particle population of 84 µm −2 , which decreased at 800 °C (59 µm −2 ) and reached a maximum value at 900 °C (88 µm −2 ). In other perovskite oxide systems, especially the widely studied Ni exsolution on titanates, the increase in temperature leads to an increase in particle size and a decrease in nanoparticle dispersion [12,34]. Interestingly, in the Sr 2 Fe x Ni 1-x MoO 6-δ system studied Figure 4. ( a ) TEM image of the SFNM65 material after 2 h exsolution at 900 ◦ C under 5% H 2 , depicting the exsolved nanoparticles. ( b ) HAADF-STEM image of SFNM65 ( c – g ) EDX mapping images for Ni, Fe, Mo, Sr, and O, respectively, showing that the exsolved nanoparticles are just composed by Fe and Ni. Catalysts 2021, 11, x FOR PEER REVIEW 7 of 16 . Figure 4. (a) TEM image of the SFNM65 material after 2 h exsolution at 900 °C under 5% H 2 , depicting the exsolved nanoparticles. (b) HAADF-STEM image of SFNM65 (c–g) EDX mapping images for Ni, Fe, Mo, Sr, and O, respectively, showing that the exsolved nanoparticles are just composed by Fe and Ni. Figure 5. (a) Comparative of the XRD patterns of the SFNM65 material as-synthesized, and after 2 h and 20 h exsolution at 900 °C in a 5% H 2 atmosphere. Crystal structure of a Sr 3 FeMoO 7 Ruddlesden–Popper phase. Green, brown, and red spheres represent Sr, Fe/Mo, and oxygen atoms, respectively. Structure drawn with VESTA [37]. (b) SEM image of the SFNM65 material after 20 h exsolution. Fe and Ni reflections are shown in (a) for the sake of comparison. An alternative way to tune the particle size in exsolution is with temperature [12,34,35,38]. In order to explore this aspect, we treated the double layer perovskite materials at 700 and 800 °C and compared them with the original exsolution temperature, 900 °C. As observed in Figure 6a, there were remarkable differences on the exsolved nanoparticle sizes when lowering the temperature. Namely, the exsolved nanoparticle sizes (d 50 ) were 15.6 ± 4.6, 24.4 ± 5.4, and 31.0 ± 8.4 nm at 600, 700, and 900 °C, respectively. The monotonic increase in the particle size with temperature observed in this work was in line with previous reports that indicate, based on the nucleation theory, that increasing the temperature particularly affects the particle growth [12]. It can also be observed that lowering the temperature (700 °C) resulted in a more uniform particle size distribution, whereas at 900 °C, the distribution was much wider (Figure 6b). On the other hand, we did not observe a clear trend when analyzing the effect of temperature in the exsolved particle population (Figure 6c). The treatment at 700 °C resulted in a particle population of 84 µm −2 , which decreased at 800 °C (59 µm −2 ) and reached a maximum value at 900 °C (88 µm −2 ). In other perovskite oxide systems, especially the widely studied Ni exsolution on titanates, the increase in temperature leads to an increase in particle size and a decrease in nanoparticle dispersion [12,34]. Interestingly, in the Sr 2 Fe x Ni 1-x MoO 6-δ system studied Figure 5. ( a ) Comparative of the XRD patterns of the SFNM65 material as-synthesized, and after 2 h and 20 h exsolution at 900 ◦ C in a 5% H 2 atmosphere. Crystal structure of a Sr 3 FeMoO 7 Ruddlesden–Popper phase. Green, brown, and red spheres represent Sr, Fe/Mo, and oxygen atoms, respectively. Structure drawn with VESTA [ 37 ]. ( b ) SEM image of the SFNM65 material after 20 h exsolution. Fe and Ni reflections are shown in (a) for the sake of comparison. Catalysts 2021,11, 741 7 of 14 An alternative way to tune the particle size in exsolution is with temperature [12,34,35,38] . In order to explore this aspect, we treated the double layer perovskite materials at 700 and 800 ◦ C and compared them with the original exsolution temperature, 900 ◦ C. As observed in Figure 6a, there were remarkable differences on the exsolved nanoparticle sizes when lowering the temperature. Namely, the exsolved nanoparticle sizes (d 50 ) were 15.6 ± 4.6, 24.4 ± 5.4, and 31.0 ± 8.4 nm at 600, 700, and 900 ◦ C, respectively. The monotonic increase in the particle size with temperature observed in this work was in line with previous reports that indicate, based on the nucleation theory, that increasing the temperature particularly affects the particle growth [ 12 ]. It can also be observed that lowering the temperature (700 ◦ C) resulted in a more uniform particle size distribution, whereas at 900 ◦ C, the distribution was much wider (Figure 6b). On the other hand, we did not observe a clear trend when analyzing the effect of temperature in the exsolved particle population (Figure 6c). The treatment at 700 ◦ C resulted in a particle population of 84 µ m −2 , which decreased at 800 ◦ C (59 µ m −2 ) and reached a maximum value at 900 ◦ C (88 µ m −2 ). In other perovskite oxide systems, especially the widely studied Ni exsolution on titanates, the increase in temperature leads to an increase in particle size and a decrease in nanoparticle dispersion [ 12 , 34 ]. Interestingly, in the Sr 2 Fe x Ni 1-x MoO 6-δ system studied here, the standard treatment performed initially (900 ◦ C, 2 h) led to the highest degree of nanoparticle dispersion (also observed in SFNM35 and SFNM65, Figure 3). Catalysts 2021, 11, x FOR PEER REVIEW 8 of 16 here, the standard treatment performed initially (900 °C, 2 h) led to the highest degree of nanoparticle dispersion (also observed in SFNM35 and SFNM65, Figure 3). In summary, throughout this section, the effect that extrinsic (temperature, time) and intrinsic (B-site composition) factors have on the nanoparticle exsolution in Sr 2 Fe x Ni 1x MoO 6-δ double-layer perovskites was studied. We confirmed nanoparticle exsolution of an alloy composition of FeNi 3 . Additionally, in the absence of Ni (i.e., SFMO sample), metallic exsolution was not observed. Increasing exsolution time and temperature led to larger particles, and, in the case of the former, also to a major phase transformation into a Ruddlesden–Popper phase. Figure 6. Influence of exsolution temperature on the particle size and population. (a) SEM images of SFNM50 material after exsolution at 700, 800, and 900 °C during 2 h in a 5% H 2 atmosphere. (b) Particle size distribution histograms. (c) Particle size and population versus temperature. 2.2. Dry Reforming Tests on Fe–Ni Alloy Exsolved Nanoparticles In the previous section, we showed that by appropriately selecting the reductive treatment conditions, it is possible to tune the exsolved nanoparticle size and dispersion. Regarding the time, the shortest treatment of 2 h led to the optimum results in terms of particle size and perovskite-backbone stability. In the case of temperature, the smallest particles were obtained at 700 °C (ca. 15 nm), whereas the highest dispersion was observed at 900 °C (87 µm −2 , d 50 ~ 31 nm). Although materials with smaller nanoparticles will be ideally preferred for the catalysis, the reaction process temperature adds some restrictions to the exsolution temperature. Even if exsolution is known to alleviate particle sintering, it is not advisable to perform nanoparticle exsolution at temperatures lower than the reaction temperature, since the particles could eventually grow, even if anchored, during the reaction, resulting in variations in the catalytic activity. For the catalytic test, the temperature of choice was 850 °C, since it is the optimal temperature for dry reforming of methane [3]. Due to the endothermic nature of the reaction, lowering the process temperature will decrease the equilibrium conversion [8]. Thus, the materials used for the dry reforming test were treated at 900 °C for 2 h in a 5% H 2 atmosphere in order to drive the exsolution (Figure 3). The dry reforming catalytic tests were performed in a fixed bed quartz reactor. The exsolution-treated samples were heated in an Ar atmosphere up to 850 °C, the temperature at which the dry reforming reaction was performed for 10 h. The main goals of these catalytic tests were two-fold. First, to evaluate the catalytic activity of the four synthesized samples to ascertain the possible structure–activity relationships derived from the different compositions of the B-site sublattice (i.e., different Fe/Ni ratios). Second, to determine the stability of the exsolved nanoparticles after 10 h under dry reforming conditions at 850 °C. Here, we would like to note that our experiments were carried out with dilute streams of CH 4 and CO 2 to facilitate the screening of our materials. Ideally, pure streams of CH 4 and CO 2 will be used in industrial operation. Figure 6. Influence of exsolution temperature on the particle size and population. ( a ) SEM images of SFNM50 material after exsolution at 700, 800, and 900 ◦ C during 2 h in a 5% H 2 atmosphere. ( b ) Particle size distribution histograms. ( c ) Particle size and population versus temperature. In summary, throughout this section, the effect that extrinsic (temperature, time) and intrinsic (B-site composition) factors have on the nanoparticle exsolution in Sr 2 Fe x Ni 1-x MoO 6-δ double-layer perovskites was studied. We confirmed nanoparticle exsolution of an alloy composition of FeNi 3 . Additionally, in the absence of Ni (i.e., SFMO sample), metallic exsolution was not observed. Increasing exsolution time and temperature led to larger particles, and, in the case of the former, also to a major phase transformation into a Ruddlesden–Popper phase. 2.2. Dry Reforming Tests on Fe–Ni Alloy Exsolved Nanoparticles In the previous section, we showed that by appropriately selecting the reductive treatment conditions, it is possible to tune the exsolved nanoparticle size and dispersion. Regarding the time, the shortest treatment of 2 h led to the optimum results in terms of particle size and perovskite-backbone stability. In the case of temperature, the smallest particles were obtained at 700 ◦ C (ca. 15 nm), whereas the highest dispersion was observed at 900 ◦ C (87 µ m −2 ,d 50 ~ 31 nm). Although materials with smaller nanoparticles will be ideally preferred for the catalysis, the reaction process temperature adds some restrictions to the exsolution temperature. Even if exsolution is known to alleviate particle sintering, it is not advisable to perform nanoparticle exsolution at temperatures lower than the reaction Catalysts 2021,11, 741 8 of 14 temperature, since the particles could eventually grow, even if anchored, during the reaction, resulting in variations in the catalytic activity. For the catalytic test, the temperature of choice was 850 ◦ C, since it is the optimal temperature for dry reforming of methane [ 3 ]. Due to the endothermic nature of the reaction, lowering the process temperature will decrease the equilibrium conversion [ 8 ]. Thus, the materials used for the dry reforming test were treated at 900 ◦ Cfor2hina5%H 2 atmosphere in order to drive the exsolution (Figure 3). The dry reforming catalytic tests were performed in a fixed bed quartz reactor. The exsolution-treated samples were heated in an Ar atmosphere up to 850 ◦ C, the temperature at which the dry reforming reaction was performed for 10 h. The main goals of these catalytic tests were two-fold. First, to evaluate the catalytic activity of the four synthesized samples to ascertain the possible structure–activity relationships derived from the different compositions of the B-site sublattice (i.e., different Fe/Ni ratios). Second, to determine the stability of the exsolved nanoparticles after 10 h under dry reforming conditions at 850 ◦ C. Here, we would like to note that our experiments were carried out with dilute streams of CH 4 and CO 2 to facilitate the screening of our materials. Ideally, pure streams of CH 4 and CO2will be used in industrial operation. Figure 7a,b show the CO 2 and CH 4 conversion (X CO2 and X CH4 ) versus time for the four samples, respectively. Regarding the X CO2 , it can be observed that, overall, for the four samples, it was stable during the 10 h test excluding the first injection. From the four samples tested, SFNM50 showed the highest decay in X CO2 going from 72% (second injection) to ca. 70% (10 h). For SFMO and SFM35, the CO 2 conversion slightly increased during the first 2 h of reaction. SFNM35 and SFNM50 showed the highest CO 2 conversion values, which progressively decay with lowering the Ni content. In Figure 7c, the X CO2 after 10 h is shown and plotted versus the Ni content. The X CO2 values with decreasing Ni content in Sr 2 Fe x Ni 1-x MoO 6-δ were 72, 70, 64, and 45% for x = 0.65, 0.5, 0.35, and 0, respectively. A similar trend was observed for the CH 4 conversion in Figure 7b, although the overall conversion values were lower than for CO 2 . After 10 h, the X CH4 values with decreasing Ni content were 47, 43, 37, and 26% for x = 0.65, 0.5, 0.35, and 0, respectively. In terms of stability, the SFNM50 exhibited a progressive decay on its X CH4 values, which limits its further applicability, which went from 52 to 43% in 10 h on stream. For all the samples tested, CO 2 conversion was higher than CH 4 conversion. This fact resulted in H 2 /CO ratios below the stoichiometric value (Figure 7c), which according to Equation (1), should be equal to 1. This is caused by the competing reverse water-gas shift reaction (Equation (2)). Regarding the H 2 /CO ratios, the trends are analogous to the CO 2 and CH 4 conversion, being the highest value achieved (H 2 /CO = 0.49) for the sample with the highest Ni content, SNM35 (Figure 7c). H 2 (S H2 ) and CO selectivity (S CO ) are depicted in Figure S2 (Supplementary Materials). The values, taken at the end of the reaction, show that both parameters increase with increasing Ni content. Namely, SFMO showed ca. 20% selectivity toward CO production, whereas for SFNM35 (the highest Ni content), S CO was around 52%. As expected from the H 2 /CO ratios lower than 1, S CO values were higher than SH2. For instance, H2selectivity for SFNM35 was ca. 32%. H2+CO2↔CO +H2O(2) To check the stability of the exsolved nanoparticles, the four samples were analyzed by SEM after the 10 h dry reforming reaction test (Figure 8a). It can be observed that the three samples that presented nanoparticle exsolution before the reaction (those with Ni on B-site) also exhibited exsolved nanoparticles after the 10 h of the dry reforming test at 850 ◦ C. Through image analyses of the materials after the reaction, we observed that the particle size values were in the range of the values obtained before the reaction. This result emphasizes the robustness of nanoparticles created by the exsolution method. For instance, the SFNM50 material showed exsolved nanoparticles of ~31 nm before ( Figure 6c ) and after the reaction (Figure 7c), maintaining the initial values. SFNM35 and SFNM65 presented nanoparticle sizes of ~29 and 26 nm after the reaction, which in both cases were lower than SFNM50. The exsolved nanoparticle dispersion after the reaction is shown in Figure 7c. The observed trend indicates that the amount of exsolved Catalysts 2021,11, 741 9 of 14 nanoparticles increased with the Ni content. Namely, the nanoparticle population was 51, 70, and 85 µ m −2 for SFNM65, SFNM50, and SFNM35, respectively. Interestingly, SFNM50 showed a nanoparticle population of 88 µ m −2 before the reaction (Figure 6c), which decreased to 70 µ m −2 after the reaction. This decay on the nanoparticle dispersion correlates well with the decay in the catalytic activity by SFNM50, especially observed for the CH 4 conversion (Figure 7b). In addition, the SFNM50 image after the reaction depicted in Figure 8revealed the appearance of a major phase segregation that was not observed before the reaction (Figure 3). The formation of this additional phase (characterized by its darker color) could be the responsible for the decay on the nanoparticle population after reaction. As observed in the SEM image (Figure 8), this segregated phase did not exhibit nanoparticle exsolution. XRD was utilized to confirm the nature of this phase. Figure 8b shows the XRD patterns of the four samples after the dry reforming reaction tests. The main impurity corresponds to SrMoO 4 , which was present on the as-synthesized materials (Figure 2a), disappeared after the exsolution treatment (Figure 2b), and regenerated during the 10 h dry reforming test. In addition, we identified the presence of SrCO 3 signatures at 2 θ = 25–26 ◦ (Figure 8b) as a consequence of the reaction of Sr 2+ with the CO 2 used for the dry reforming tests. Carbonation of A-site cations in perovskites (e.g., Sr, Ca, Ba) is known to be a major issue in several CO 2 valorization processes [ 39 , 40 ], although for most of the samples, its presence was negligible. The material that was more affected by Sr carbonation was SFNM50. The concentration of impurity phases after the reaction was determined by Rietveld refinement, indicating a 7.9% of SrCO3in SFNM50 material after the reaction. Furthermore, it showed 51.6% of SrMoO 4 , making SFNM50 the sample in which the double perovskite phase, which supports the exsolved nanoparticles, was more degraded after the dry reforming test. This fact had strong implications on the exsolved nanoparticle dispersion, since SrMoO 4 and SrCO 3 were not expected to allocate Ni and/or Fe cations in their crystal lattices, thus, limiting the amount of available surface with the presence of alloy nanoparticles. In particular, the exsolved nanoparticle dispersion of SFNM50 decreased from 88 particles µ m −2 (Figure 6c) to 70 particles µ m −2 (Figure 7c). The combination of both facts, the appearance of the SrMoO 4 segregated phase and the decrease in the nanoparticle dispersion, might be responsible for the progressive decay in the SFNM50 CH 4 conversion observed in Figure 7b. In the case of SFNM35, such decay was more attenuated, which could correlate well with a lower amount of detected phase segregation after reaction. Namely, SFNM35 exhibited 3.9% and 44.3% for the SrCO 3 and SrMoO 4 impurity phases, respectively. Interestingly, samples SFMO and SFM65, which exhibited lower CH 4 and CO 2 conversion values, presented minor contributions of the phase segregation after the reaction. For instance, SFMO showed a negligible presence of SrCO 3 (0.1%) and 27.8% of SrMoO 4 , with the latter in the concentration range detected before the reaction (Table 1). Next, we conducted Raman spectroscopy to analyze the possible formation of carbon deposits during the 10 h reaction test. Figure 8c shows the Raman spectra of the four samples after the reaction, depicting the 1000–2500 cm −1 wavenumber region, in which modes related to carbon species are expected to appear. Namely, the D and G bands were located at 1368 and 1560 cm −1 , respectively [ 41 ]. Only a minor contribution of the mode at around 1560 cm −1 (ascribed to the G band) was observed for the samples SFNM50 and SFNM65, whereas no modes related to carbon deposition were observed for SFNM35, confirming the high resistance to coking obtained with the exsolution method [ 7 ]. With these postmortem analyses, we can conclude that the morphological changes suffered by the materials were mostly related with phase segregations occurring in the host perovskite matrix, whereas the exsolved nanoparticles exhibited high stability, which were responsible for an increase in the catalytic activity. It was observed that a higher Ni/Fe ratio favors a higher population of FeNi 3 nanoparticles, ascribed to the thermodynamically favorability for Ni cations to exsolve over Fe cations, resulting in these Ni rich metallic alloys. Nonetheless, even if the nanoparticles retain its initial particle size and level of anchoring, the presence of phase segregations alter the nanoparticle population if, as was the case here, the segregations do not allocate exsolved nanoparticles. This fact caused a decrease in the CH 4 conversion for