Perovskite-Type Oxides as Exsolution Catalysts in CO₂ Utilization

Definition

Perovskite-type oxides (ABO₃) are a highly versatile class of materials. They are compositionally flexible, as their constituents can be chosen from a wide range of elements across the periodic table with a vast number of possible combinations. This flexibility enables the tuning of the materials’ properties by doping the A- and/or B-sites of the base structure, facilitating the application-oriented design of materials. The ability to undergo exsolution under reductive conditions makes perovskite-type oxides particularly well-suited for catalytic applications. Exsolution is a process during which B-site elements migrate to the surface of the material where they form anchored and finely dispersed nanoparticles that are crucially important for obtaining a good catalytic performance, while the perovskite base provides a stable support. Recently, exsolution catalysts have been investigated as possible materials for CO₂ utilization reactions like reverse water–gas shift reactions or methane dry reforming.

1. Introduction

An overwhelming majority of published studies on the topic agree that current global warming is a consequence of human-caused (anthropogenic) climate change [1]. The level of the greenhouse gas CO₂ in the atmosphere—which is the main cause of anthropogenic climate change—rose sharply between the beginning of the Industrial Era (approximately 280 ppm) and today (more than 400 ppm), thus straining the carbon budget (i.e., the amount of carbon that can be released in the atmosphere while still limiting global warming to a given temperature difference) [2]. Therefore, efforts are being undertaken to reduce CO₂ to achieve net zero emissions (e.g., the European Green Deal [3] aims for climate neutrality of the European Union by 2050).

Closing the carbon cycle via CO₂ conversion is a possible way towards meeting this goal. This means no longer treating emitted CO₂ as a waste gas but as a readily available carbon source for base materials or fuels. A schematic depiction of such a closed carbon cycle in industries is shown in Figure 1. Flue gas from industries and power generation is captured and converted. The products can be used as feedstock in the chemical industry or as fuels which can be stored for later use. CO₂ conversion can be driven in different ways. In a 2019 review, De Luna et al. discussed the question of how CO₂ electrolysis (i.e., electrochemical processes) could be a pathway to replacing the chemicals and fuels produced petrochemically [4]. In the authors’ own work, we investigate a different approach to achieve CO₂ conversion. Instead of electrolysis, seemingly straightforward reactions like the reverse water–gas shift reaction (rWGS, Equation (1)) or the dry reforming of methane (DRM, Equation (2)) are used to directly transform CO₂ into CO or syngas (a mixture of CO and H₂), respectively.

$${\mathrm{C}\mathrm{O}}_{2\left(\mathrm{g}\right)}+{\mathrm{H}}_{2\left(\mathrm{g}\right)} \rightleftharpoons {\mathrm{C}\mathrm{O}}_{\left(\mathrm{g}\right)}+{{\mathrm{H}}_2\mathrm{O}}_{\left(\mathrm{g}\right)}\hspace{1em}(\mathsf{\Delta }{H}_r^{298}=+42 {\mathrm{k}\mathrm{J} \mathrm{m}\mathrm{o}\mathrm{l}}^{-1})$$

(1)

$${\mathrm{C}\mathrm{O}}_{2\left(\mathrm{g}\right)}+{\mathrm{C}\mathrm{H}}_{4\left(\mathrm{g}\right)} \rightleftharpoons 2 {\mathrm{C}\mathrm{O}}_{\left(\mathrm{g}\right)}+{2 \mathrm{H}}_{2\left(\mathrm{g}\right)}\hspace{1em}(\mathsf{\Delta }{H}_r^{298}=+247 {\mathrm{k}\mathrm{J} \mathrm{m}\mathrm{o}\mathrm{l}}^{-1})$$

(2)

Due to its stability, reactions using CO₂ as a raw material necessitate the use of a catalyst for activation. Furthermore, CO₂ activation processes demand large amounts of energy. This opens up energy storage as a second potential application of CO₂ conversion (apart from the reduction of emissions), as electrical energy (ideally renewably generated with solar panels, wind turbines, geothermal plants…) is converted into chemical energy of fuels that can be stored more easily.

For both rWGS and DRM, a large number of diverse catalytically active materials have been extensively researched, for instance, supported catalysts based on noble metals [5,6] or Ni [7]. However, there are drawbacks for these catalysts, e.g., Ni-based catalysts are vulnerable to deactivation via carbon deposition [8] and catalysts using noble metals, while not as susceptible to coking [9], are expensive. Additionally, high temperatures are required when performing rWGS or DRM, which can cause nanoparticles to sinter together, thus reducing the catalytically active surface. This means that suitable catalysts for large-scale applications need to be stable at high temperatures, resist coking, and contain catalytically active elements that are available at the surface (e.g., as nanoparticles, which should not be prone to sintering in order to maintain their activity). A promising class of materials for use as such CO₂ conversion catalysts are perovskite-type oxides due to their stability at high temperatures, their compositional flexibility, and their ability to undergo exsolution:

Perovskite-type oxides share the general formula ABO₃, where A and B are two cations of different sizes that are arranged as shown in Figure 2a. The larger A-site cations make up the corners of the unit cell; the smaller B-site cations fill the center of the cell and are octahedrally coordinated by oxygen atoms. Depending on the differences in the atomic radii of the cations, cubic, orthorhombic, or rhombohedral cells can occur. A- and B-site elements can vary over a wide range of the periodic table, paving the way to a vast number of possible materials. Moreover, both sites can be doped, which not only introduces an even larger number of possible materials, but also facilitates a material design approach, as the materials’ properties can be tuned with the specific applications in mind [10]. Aside from catalysis, applications of perovskite-type oxides include, for instance, electrodes used in electrochemistry or fuel cells [11,12,13].

Exsolution is a striking property of perovskite-type oxides that causes their particular potential as catalysts [14], especially with respect to the requirements of finely dispersed and sinter-resistant nanoparticles. Exsolution is an extensively studied process [15,16,17], during which the surface of the perovskite is decorated by nanoparticles made up of B-site elements (schematically shown in Figure 2b): Under reducing conditions, B-site cations move to the surface of the material, where they form finely dispersed and well anchored nanoparticles. The make-up (oxidic vs. metallic composition) of these nanoparticles can be tuned by the composition of the perovskite (influenced by doping) [10] as well as the exsolution conditions [17], thus allowing even more control of the specific catalytic properties of the material.

Figure 2. “(a) The perovskite structure and (b) a schematic depiction of the exsolution process. In a wet hydrogen atmosphere and under heat treatment, additional oxygen vacancies are created, and dopant atoms reduce, resulting in their migration from the bulk to the surface, where they form nanoparticles.” Reprinted from [18] under CC-BY 4.0.

Figure 2. “(a) The perovskite structure and (b) a schematic depiction of the exsolution process. In a wet hydrogen atmosphere and under heat treatment, additional oxygen vacancies are created, and dopant atoms reduce, resulting in their migration from the bulk to the surface, where they form nanoparticles.” Reprinted from [18] under CC-BY 4.0.

This entry aims to showcase how perovskite-type oxides as catalysts in CO2 conversion processes are investigated to assess their suitability.

2. Materials—Synthesis and Characterization

As mentioned before, CO₂ conversion processes impose a set of requirements (high temperature stability, coking resistance, nanoparticle/active site stability…) on catalyst materials. A deeper understanding of catalysts in general and detailed insights into the structure–property relations necessary to tune a catalyst as well as the intensive characterization and controlled synthesis of precursor materials, are crucial. A wide array of different methods is used to gain these insights.

2.1. Synthesis of Perovskites

Common methods used to prepare powder samples of perovskite-type oxides are solid-state [19] and sol-gel synthesis (e.g., Pechini synthesis [10]) as well as flame-assisted spray pyrolysis [20,21].

Solid-state and sol-gel methods use precursors (oxides, carbonates, or nitrates are widely used in perovskite synthesis) that are either dispersed in small amounts of volatile solvents like acetone (in case of the solid-state route [19]) or dissolved, for instance, in water or nitric acid (sol-gel [10]). Then, the precursors are mixed in the desired stoichiometric ratio. Additionally, complexing agents such as citric acid are added during sol-gel processes. In both synthesis routes, the liquid is evaporated. During the solid-state method, the remaining slurry is calcined directly at high temperatures and finely ground down (e.g., via ball milling). In sol-gel synthesis, the gel that remains after evaporation is heated until self-ignition, which leads to a sponge-like structure that is milled to a fine powder which, in turn, is calcined as well. Additional grinding and calcination steps might be necessary to achieve phase purity [10,19].

The low surface area of the resulting powders is a major drawback of the above-mentioned methods, which is especially problematic for catalytic materials. The reason for this is sintering, which happens during calcination steps. A workaround to the manufacture of powders with high surface areas is liquid-feed flame-assisted spray pyrolysis. The precursors (usually salts with organic anions like ethyl hexanoate) are dissolved as well, mixed in the proper ratio, and fed via a pressurized nozzle into a CH₄/O₂ flame where nanocrystalline powders are formed, which are collected and used for characterization [21].

Alternative methods of synthesizing perovskite oxides include, but are not limited to, the co-precipitation method, hydrothermal synthesis, and microwave synthesis. These methods are only briefly described here; however, there is already extensive literature discussing this topic [22]. During co-precipitation, all cations are mixed and precipitated simultaneously either due to the addition of a reagent or thermal decomposition [23]. However, precise control of the procedure is a necessity to obtain a phase-pure product. In hydrothermal synthesis, the reaction is performed in an aqueous solution but at elevated temperatures and pressures [24]. As an alternative heating source, microwaves can be used to very quickly heat an aqueous solution of dissolved precursors.

2.2. X-ray Diffraction

Ex situ X-ray Diffraction (XRD) under ambient conditions can be used to check the phase purity after synthesis or additional calcination steps, as the presence of impurities causes additional reflexes to appear [18].

Moreover, the emergence of nanoparticles after exsolution experiments (in the form of additional oxidic or metallic phases) as well as the continued stability of the perovskite lattice after (high temperature) catalytic tests (see Section 3.1) can be verified [15]. Additionally, some common deactivation products (e.g., carbonates, graphite) can be observed via XRD.

XRD measurements can also be performed in situ or operando to track structural changes during reactions (see Section 3.2).

2.3. Micrsocopic Characterization—SEM and TEM

Scanning Electron Microscopy (SEM) and Transmission Electron Microscopy (TEM) are methods that are routinely used to characterize the surfaces of catalyst materials as well as nanoparticles after exsolution [25].

SEM images provide information not only about the presence of nanoparticles (i.e., whether or not exsolution occurred, for example, see the left side of Figure 3 [18]) and their sizes, shapes, and population density, coupled with Energy Dispersive X-ray Spectroscopy (EDX) [26], it can also give elemental information. The image shown on the right-hand side of Figure 3 [18], for instance, reveals iron accumulation in the locations of the nanoparticles, lending credibility to the conclusion that exsolution of iron particles has occurred.

Information about the anchorage as well as geometric data (orientation of the exposed face, distances between lattice planes…) of the nanoparticles can be extracted from TEM images. The latter requires a high resolution though. The elemental make-up of the particles can be studied from TEM+EDX as well [25]. Additionally, environmental TEM can be used to visualize the exsolution process on an atomic scale [27]. In this study, Neagu et al. observed the formation of Ni metal nanoparticles from La_0.43Ca_0.37Ti_0.94Ni_0.06O₃ in situ using environmental TEM, as seen in Figure 4. They confirmed that the particles grow while maintaining their initial nucleation positions. During the growth, the perovskite lattice rises around the particle, leading to anchoring of the particles into the lattice [27].

In a study by Kim and co-workers [28], the reduction time and temperature of La_0.6Ca_0.2Ni_0.25Ti_0.75O₃ were varied. It was found that at reduction temperatures of 800 °C and below (in this case 600 °C, 700 °C, and 800 °C), only a small difference in the size of the formed nanoparticles occurred when increasing the reduction time from 3 h to 24 h. All exsolved particles were spherical at these temperatures, and the size increased with an increasing temperature. Going to an even higher temperature of 900 °C showed the expected trend when reducing for 3 h. However, when the reduction time was increased to 24 h, faceting of the nanoparticles occurred, forming triangular and pyramidal shapes with larger sizes. At 1000 °C, the faceted nanoparticles had already formed after 3 h of reduction but doubled in size after 24 h. A comparison of the catalytic activity between the spherical and faceted nanoparticles for DRM revealed that the faceted nanoparticles were not only more active but also exhibited greater stability. This was also found in the comparison of smaller particles exsolved in situ versus bigger particles exsolved during a reductive pre-treatment in the authors’ own work [29]. Interestingly for systems where Ni nanoparticles were deposited on SiO₂, an inverse trend (bigger nanoparticles lead to decreased catalytic activity) was found [30].

In a study by Gao et al. [31], analyzing the energetics of exsolution, four distinct steps in the process were proposed: diffusion, reduction, nucleation, and particle growth. As grain boundaries are preferential pathways for diffusion, it becomes clear why they are also preferred nucleation sites for nanoparticles.

2.4. X-ray Photoelectron Spectroscopy

X-ray Photoelectron Spectroscopy (XPS) is a surface sensitive method that can be used to characterize surface species in more detail, for instance, the nature of exsolved nanoparticles (oxidic vs. metallic) or the oxidation states of surface atoms before and after a reaction [32].

In situ XPS measurements are particularly useful to elucidate the reaction mechanisms even further by observing changes to surface species under the reaction conditions and to monitor the formation of suspected additional surface species, like carbonaceous deposits on the surface (see Section 3.3).

2.5. Theoretical Modeling

Electronic structure calculations are another powerful tool to further understand the properties of the materials under investigation as well as to predict novel candidates for given applications.

Density Functional Theory (DFT) can be utilized to study the effects of doping on the electronic properties (for instance, changes in conductivity caused by A-site doping [33]) or on the atomic structure (changes in lattice parameters, atomic distances, and angles), which are of particular interest, as exsolution requires atoms to diffuse through the bulk to the surface [18]. Thermodynamic quantities, like enthalpies of formation [34], energy barriers during diffusion [35], defect formation energies [36], or energetics of adsorption processes (especially relevant for heterogeneously catalyzed gas phase reactions) [37], can be calculated as well. Moreover, DFT studies are a useful tool for the proposal of possible mechanisms and/or reaction pathways, for instance, Huang et al. investigated the effect of exsolved Ni nanoparticles on La-based perovskite surfaces on water–gas shift reactions [38].

As mentioned before, perovskite-type oxides span a vast compositional space due to the large number of viable compositions with possible doping on top. High-throughput DFT calculations are needed to screen this space. In a recent effort to create a dataset of multinary perovskite-type oxides, Bare et al. [39] generated almost 70,000 perovskite-type oxides and optimized their structures to assess their stability (by predicting the formation and decomposition energies). The resulting test set can be used for data mining or the training of machine learning models to find potential new catalyst materials without having to synthesize and manually test a huge number of candidates.

3. Catalytic Tests and In Situ/Operando Measurements

So far, the characterization methods presented only address ex situ measurements (meaning the measurement conditions do not represent exsolution or reaction conditions). While they still give valuable information by allowing comparisons between the initial pristine state of a catalyst before any further treatment and the states after exsolution has been triggered or reactions have been performed, the analysis of the materials during reactions or exsolution and at higher temperatures or pressures is crucial to obtain a more detailed picture of the catalyst’s properties. To achieve this, in situ or operando measurements have to be performed, for instance, to assess the catalytic performance (by tracking the conversion of CO₂ or the yield of the respective product), to monitor the stability of the perovskite lattice, to track structural changes due to exsolution (including the emergence of newly exsolved nanoparticles), or to observe the possible formation of additional phases (e.g., carbonaceous species) that might affect the catalytic activity [40].

3.1. Catalytic Tests

Catalytic tests are the method of choice for assessing the suitability and performance of any material as a catalyst for a reaction. Not only the catalytic activity itself is of interest, the suppression of side reactions (either via modifying the material itself [41] or by choosing proper reaction conditions [29]) to increase the selectivity, the long-term stability of the material (in terms of a reproducible and continued performance) to ensure the viability as an industry catalyst with extended usability [42], or, more specific to exsolution catalysts, the reversibility of nanoparticle formation which enables the regeneration of the catalytically active surface [43], are examples of factors that have to be investigated intensively to understand materials well enough to design a catalyst.

In Figure 5 and

Table 1. “BET surface areas, average specific activities calculated for CO production during rWGS at 600 °C and WHSVs of the respective measurements for all investigated perovskite materials” [44]. Accompanying table to Figure 5. Reprinted from [44] under CC-BY 4.0.

Catalyst	Surface Area m2 g−1	Specific Activity 10−6 mol m−2 s−1	WHSV L g−1 h−1
La0.9Ca0.1FeO3-δ	3.8	5.7	24.6
La0.6Ca0.4FeO3-δ	2.8	5.9	34.1
Nd0.9Ca0.1FeO3-δ	2.2	11.3	32.7
Nd0.6Ca0.4FeO3-δ	1.5	6.6	20.5
Nd0.6Ca0.4Fe0.9Ni0.1O3-δ	1.6	18.0	36.0
Nd0.6Ca0.4Fe0.9Co0.1O3-δ	1.2	27.2	32.1
commercial LSF	5.7	4.8	28.8

, the specific activities with respect to the rWGS of various newly synthesized mixed iron-based perovskites are compared (with the well-known La_1-xSr_xFeO_3-δ (LSF) as a reference) at different temperatures [44]. Multiple effects of the elemental composition and doping are visible. Switching the main A-site constituent from La to Nd increases the specific activity, while the effect of the doping concentration on the A-site follows no clear-cut trend; in case of lanthanum, increasing the Ca content appears to be advantageous, while the reverse is true for neodymium. Moreover, introducing B-site doping raises the activity even further with both Ni and Co beneficially affecting the catalyst’s performance (Ni more so than Co).

3.2. In Situ XRD

The structural properties of catalysts play important roles for both the actual catalytic activity of the material as well as the stability. Therefore, it is necessary to investigate candidate materials with respect to their stability under realistic conditions (e.g., high temperatures during reactions [45] or under reaction atmospheres). Furthermore, changes in the material (both in bulk, for instance (partial) decomposition, or at the surface, like exsolution or the formation of impurity phases [29,44,46]) caused by reaction conditions need to be studied as well, especially if catalyst design is the goal. In situ XRD allows us to address these questions.

Figure 6 shows a series of in situ XRD measurements performed on a mixed perovskite with Ca-doping on the A- and Co-doping at the B-site at elevated temperatures in a reducing atmosphere (humid H₂) [18]. Over the whole temperature range, the perovskite lattice is stable (perovskite reflexes marked with P are present at all temperatures and hardly exhibit any changes in intensity). Additional peaks start to appear from 575 °C with the onset of nanoparticle exsolution. A Fe/Co bcc phase emerges, which becomes more pronounced with an increasing temperature. Moreover, unwanted CaO begins to appear at around 37° above 600 °C, probably due to A-site Ca²⁺ migrating to the surface where CaO is formed, and, in turn, catalytic activity is reduced due to a loss in the active surface. In recent literature, it was shown that exsolved nanoparticles exhibit outstanding stability under reducing conditions. Zang et al. [47] performed stability tests over 200 h and observed that the nanoparticle size did not change significantly after an initial forming phase.

3.3. In Situ XPS

As mentioned before, XPS is a method that allows us to study the surface of a given material closely. Changes in the oxidation states of the constituent atoms of the materials accompany the formation of metallic nanoparticles during exsolution. XPS is particularly well-suited to the monitoring of these changes at a quantitative level. Additionally, surface impurities (carbon species) or deposits of inactive solid phases (e.g., oxides of A-site constituents already mentioned in the previous section) can be observed as well.

Figure 7 illustrates how XPS spectra can be used to follow changes to a perovskite surface (and possibly present nanoparticles) under reducing conditions [17]: After an initial oxidizing step in oxygen, only oxidic iron—blue and green contributions to the main peak—is observed (top spectrum), which corresponds to bulk iron cations. By switching to a reducing atmosphere (wet hydrogen), notable exsolution can occur by the shoulder that begins to develop on the right-hand side of the iron peak, which is caused by the emergence of metallic iron (red contribution to the fit). Going to higher temperatures gradually increases the amount of metallic iron (second and third spectra from the top). Adding a negative bias leads to more strongly reducing conditions which, in turn, causes even more metallic iron to appear.

The deposition of (often) unwanted species on the surface (e.g., carbonaceous species during catalytic reactions—a process called coking that leads to deactivation of the catalyst) can be followed by XPS as well. Figure 8 shows the temperature-dependent formation of carbon species on an iron-based perovskite used as a DRM catalyst [29]. At moderate temperatures, small amounts of carbon on the surface could be observed, which disappeared at higher temperatures, thus highlighting the coking resistance of perovskite-type oxides.

4. Conclusions

Perovskite-type oxides are very promising materials for catalysis in general, due to their vast compositional flexibility (which is further enlarged by the possibility of doping on both cationic lattice sites) and their exsolution capabilities. The perovskite lattice forms the oxidic backbone (which, in many cases, is stable up to very high temperatures), while certain constituent atoms migrate to the surface to form nanoparticles under reducing conditions. Choosing the proper elemental make-up of the perovskite (for instance, via the doping of highly catalytically active elements) leads to a perovskite surface that is decorated with nanoparticles that can serve as active sites during catalytic reactions. For the design of an optimal catalyst, several aspects must be considered:

(1)

The stability of the material: To achieve the synthesis of a phase-pure catalyst, the thermodynamic requirements for forming a stable phase have to be fulfilled. Databases like the one from Bare et al. [39] can be used as guidelines.

(2)

A-site composition: Even though materials on the A-site have not been as deeply studied as materials on the B-site, several design principles are common. Lanthanoids such as La, Nd, Pr, or Sm are often used as major components on the A-site. Their ionic radii make them ideally suited elements to use together with transition metals on the B-site. Additionally, divalent ions such as Ca, Sr, or Ba are often doped on the A-site. This promotes the formation of oxygen vacancies and/or the oxidation of other elements in the material, thus enabling exsolution.

(3)

B-site composition: The B-site is often made up of one or more transition metals. The main concern in choosing the B-site is the nature of the nanoparticles that one wants to exsolve. Often, the B-site is made from a base element such as Fe or Ti, which is then doped with the desired exsolving element (e.g., Cu, Co, Ni, …). One has to keep in mind that the element to be exsolved should be more easily reducible than the base element.

Currently, these materials are being investigated with respect to their suitability in CO₂ utilization reactions like reverse water–gas shift or the dry reforming of methane, with very encouraging results already published in literature.

A deeper understanding of perovskite oxides can be obtained by a variety of methods like (in situ/operando) X-ray Diffraction, (in situ) X-ray Photoelectron Spectroscopy, and Scanning and Transmission Electron Microscopy. Moreover, their behaviors under reaction conditions (stability, side reactions…) need to be well understood in order to design catalysts that make use of the beneficial properties that many perovskites share (such as stability at high temperatures, resilience against coking, sinter-resistant exsolved nanoparticles).