Thermochemical Properties of High Entropy Oxides Used as Redox-Active Materials in Two-Step Solar Fuel Production Cycles

Abstract

The main challenges and obstacles to the development of hydrogen/carbon monoxide production from the splitting of water/carbon dioxide through two-step solar thermochemical cycles are strongly related to material concerns. Ineed, ceria is the main benchmark redox material used in such processes because it provides very good oxidation reaction kinetics, reactions reversibility and thermal cycling stability. This is at the expense of a low reduction yield (non-stoichiometry δ in CeO_2-δ) at relatively high temperatures (≥1400 °C), which requires operation at low oxygen partial pressures during the reduction step. Hence, the specific fuel output per mass of redox material, i.e., the amount of H₂/CO produced per cycle, remains low, thereby limiting the overall solar-to-fuel conversion efficiency. Perovskites offer larger amounts of fuel produced per cycle but the reaction kinetics are slow. This study addresses the thermochemical investigation of a new class of metal oxides, namely high entropy oxides (HEOs), with the aim of improving the specific amount of fuel generated per cycle with good kinetic rates. Different formulations of high entropy oxides were investigated and compared using thermogravimetric analysis to evaluate their redox activity and ability to split CO₂ during thermochemical cycles. Among the different formulations tested, five HEOs yielded CO with a maximum specific fuel output of 154 µmol/g per cycle. These materials’ performances exceeded the production yields of ceria under similar conditions but are still far from the production yields reached with lanthanum–manganese perovskites. This new class of materials, however, opens a wide path for research into new formulations of redox-active catalysts comparing favorably with the ceria redox performance for solar thermochemical synthetic fuel production.

1. Introduction

Hydrogen is an energy vector that will play an important role in the future worldwide energy mix, as evidenced in the recent public policies of different countries [1,2,3]. Current production paths cannot be considered as sustainable solutions, because most of the hydrogen produced arises from the steam methane reforming process which emits large amounts of CO₂. Hence, hydrogen must be produced from renewable energy sources and new efficient processes have to be deployed. Converting solar energy into hydrogen and synthetic fuels is a promising pathway. Among the different possibilities, hydrogen can be produced from water electrolysis with electricity coming from photovoltaic modules or from photochemical processes, however, these two options face limited theoretical energy conversion (photons to hydrogen) efficiencies. Alternatively, the thermochemical conversion of solar energy into hydrogen or synthetic fuels offers a higher theoretical conversion efficiency by directly using the heat generated by concentrated solar radiations in high-temperature thermochemical reactions [4]. Such a thermochemical process offers a thermodynamically favorable pathway for fuel production because it exploits the entire solar spectrum as the high-temperature process heat source to perform the thermochemical conversion. Two-step thermochemical cycles for splitting H₂O and CO₂ were extensively investigated for the last two decades, resulting in the development of advanced processes [5,6] and complete techno-economic analyses [7,8]. This production path first consists of the thermal activation of metal oxides (M_xO_y). During this first step, the metal oxide is thermally reduced (totally or partially) by releasing oxygen (Equation (1)). Then, during the second step, the reduced oxide reacts with steam (or CO₂) to be regenerated and produce hydrogen (or CO) (Equation (2)).

M_xO_y → M_xO_y-δ + δ/2O₂

(1)

M_xO_y-δ + δH₂O (or CO₂) → M_xO_y + δ H₂ (or CO)

(2)

Two-step thermochemical cycles were initially proposed to reduce the temperature of the water-splitting reaction down to ~1400 °C, compared to direct thermolysis which requires at least 2500–3000 °C. Furthermore, such cycles do not require any material consumption (except water or CO₂ as reactants) because the oxide is entirely recycled (it thus acts as a redox catalyst), and they avoid any gas separation problems because O₂ and H₂ (or CO) are released in separate steps. The solar energy input is only necessary during the first reduction step to activate the intermediate redox material (creation of oxygen vacancies reflected by the sub-stoichiometry δ) because it is an endothermal reaction, whereas the second oxidation step is exothermal. Hence, the higher the temperature of the reduction step, the higher the reduction yield; conversely, the lower the temperature of the oxidation step, the higher the H₂ or CO production. A high theoretical conversion efficiency (solar-to-fuel) can be reached because this process involves a direct conversion of concentrated solar heat into fuel without any intermediate conversion limitations linked to the solar electricity production required for the electrolysis process.

However, a scientific barrier is still blocking the competitiveness of such processes, delaying deployment at an industrial scale. This barrier is related to material science and more precisely to the catalytic redox activity of reactive materials. Indeed, the solar fuel production cost is strongly dependent on the metal oxide performance that determines both the fuel production capacity (in mol per gram of oxide) and the reaction kinetics (fuel production rates in mol per gram per second). Explicitly, the higher the material reactivity in both redox steps is, the higher the fuel production output obtained per cycle; and the faster the reaction kinetics, the higher the number of cycles performed per day. Both aspects affect the global fuel production capacity of the process. As the material is cycled several times per day during the overall lifetime of the production plant, the material redox performance is crucial. Current state-of-the-art materials that appear to be the most suitable for solar-driven thermochemical hydrogen and fuel production are nickel ferrites (NiFe₂O₄) [9,10,11,12], cerium oxide (CeO₂) [13,14,15,16,17,18,19,20,21] and lanthanum manganite perovskites ((La,Sr)MnO₃) [22,23,24,25,26,27,28]. However, these materials suffer, respectively, from deactivation during cycling which is largely due to sintering [29], a low amount of fuel produced per cycle [30,31] and slow oxidation kinetics [32,33]. This is the reason why intense research efforts to discover a new class of materials that could overcome this scientific barrier are currently being carried out worldwide [34,35,36].

This study is dealing with high entropy oxides (HEOs) [37], a new family of materials discovered in 2015 which has gained significant interest in recent years. The definition of a high entropy oxide relates to a multicationic oxide with a random and homogenous cation distribution that is entropy stabilized in a single sublattice. As key criteria for being classified as a HEO, four or five cations must be mixed, the equimolarity must be fulfilled to maximize the configurational entropy and the parent oxides must exhibit different crystal structures with low solubility in each other. In such a manner, some of the mixed cations can adopt an original configuration that cannot be obtained with the existing simple oxides. This may lead to original redox properties because the electronic configuration differs. Moreover, given that the synthesis of HEO induces reversible high entropy variation, it could be interesting to integrate it in the reduction step of a thermochemical cycle. As such, the reduction temperature could be decreased by maintaining a sufficiently high reduction enthalpy for a good reactivity during the oxidation step. The configurational disorder of such compounds provides a new path to design and discover several new crystalline phases of materials. This novel type of compound presents a high versatility and the possibility to design materials with engineered properties for specific applications. Some new advances were discovered in the field of energy storage [38,39], electrochemistry [40], optics [41,42], thermal material [43] or catalytic activity [44]. The redox properties of several of these compounds were also investigated. Zhai et al. [45] first studied iron oxide-related HEOs during thermochemical cycles. Gao et al. [46,47] investigated the same compounds during microwave-assisted thermochemical cycles. Pianassola et al. [48] and Djenadic et al. [49] first synthesized a new HEO with a Ce₂O₃-related structure. These compounds could be interesting for thermochemical H₂/CO production as long as the authors demonstrated that Ce⁴⁺ cations are reducible in such structures. Sarkar et al. [50] as well as De Santiago et al. [51] studied the thermal cycling of several rare-earth high entropy perovskites. Furthermore, a promising high oxygen storage capacity of rocksalt HEO (Mg,Co,Ni,Cu,Zn)O was reported [52,53]. Therefore, a series of specific HEO materials formulations [45,46,47,48,49,50,51,52,53,54] were selected in this work as potentially attractive redox-active candidates for solar fuel production and their performances for two-step thermochemical redox cycles were evaluated, compared, and discussed.

2. Results and Discussion

The list of HEO materials investigated for the thermochemical redox cycle as well as the synthesis and characterization methods are reported in the Materials and Methods section (Section 3).

2.1. Iron Oxide-Based HEO (Related to HEO 1 and HEO 2)

Fe_0.25Mg_0.25Co_0.25Ni_0.25O_x (HEO 1) has been already characterized for thermochemical redox cycling by Zhai et al. [45] and Gao et al. [46]. In their investigations, they highlight a reversible phase swing during cycling between a rocksalt structure (reduced phase) and a spinel structure (oxidized phase). Both structures co-exist but the ratio between them changes during the thermal cycling.

The XRD patterns of the synthesized materials are presented in Figure 1. Both HEO 1 (Fe_0.25Mg_0.25Co_0.25Ni_0.25O_x) and HEO 2 ((Fe_0.25Mg_0.25Co_0.25Ni_0.25)Zr_0.6O_x) materials calcined at 1000 °C in air possess a dual phase of spinel and rocksalt with a predominance of the spinel phase. HEO 2, which also contains Zr, displays a monoclinic ZrO₂ phase after annealing at 1000 °C. The heating at 1300 °C in air does not change the ratio between the two phases (spinel and rocksalt) but the width of the diffraction peaks decreases, thereby highlighting a crystallite growth. In HEO 2, the ZrO₂ phase changes from monoclinic to tetragonal. The XRD analysis after two complete thermochemical cycles is also shown in Figure 1. The rocksalt phase peak intensity increases for both HEO 1 and HEO 2. As the rocksalt phase corresponds to the reduced phase, this means that the CO₂-splitting step under these conditions does not allow the full reoxidation of these oxides.

Figure 2 presents the temperature-programmed XRD analyses of HEO 1 during two consecutive cycles, thus under the real atmosphere conditions including reduction in inert gas followed by oxidation in CO₂. Diffractograms were recorded at different temperatures during the reduction step and the oxidation step. The temperature profile of the analysis is presented in Section 3. The predominant phase swing between spinel and rocksalt is observed (Figure 2b). During the reduction step, the relative intensity of the peak associated with the spinel structure (2θ = 35°) decreases compared to the one of the rocksalt phase (2θ = 37°) (Figure 2b). This phenomenon is reversible during the different steps of the thermochemical cycles. A peak shift toward low angles is also observed for all phases during heating. This peak shift is reversible during cooling.

The EDS analysis of HEO 2 is presented in Figure 3. The atomic distribution is not perfectly homogeneous, and it can be noticeably observed that iron atoms are not well distributed after annealing at 1300 °C (Figure 3a) and after cycling (Figure 3c). Two different oxides are thus mixed in this material in good agreement with XRD analysis: an iron-rich phase (likely associated with the spinel phase) and a (Co,Ni,Mg)O oxide phase corresponding to the rocksalt phase. After thermochemical cycling, the cationic distribution is still heterogeneous. In contrast to previously published investigations on Fe_0.25Mg_0.25Co_0.25Ni_0.25O_x [45,46,47], the entropy stabilization is not evidenced in the present study. Because a phase swing is observed like in the mentioned publications, the rocksalt phase and the spinel phase should result from the migration of iron during cycling.

Thermogravimetric analysis during two consecutive CO₂-splitting cycles is presented in Figure 4 and the quantitative production yields are reported in

Table 1. Summary of O₂ release and CO production yields during two consecutive thermochemical cycles with the considered high entropy oxides.

Name	First Reduction, O2 Release (µmol/g)	First Oxidation, CO Production (µmol/g)	Second Reduction, O2 Release (µmol/g)	Second Oxidation, CO Production (µmol/g)
HEO 1	185	154	111	116
HEO 2	174	131	78	131
HEO 3	259	23	233 (1000 °C)	0 (600 °C)
HEO 4	143	99	79	90
HEO 5	128	81	62	79
HEO 6	276	87	63	83
HEO 7	407	44	153	20
CeO2	50	102	53	105
LSMMg	254	246	203	249

. The first reduction step at 1300 °C yields 185 and 174 µmol/g of oxygen for HEO 1 and HEO 2, respectively. During the first CO₂-splitting step, HEO 1 shows two distinct reaction stages, a first rapid mass increase followed by a slower mass increase. This must correspond to different oxidation mechanisms that lead to 154 µmol of CO produced per gram of oxide. HEO 2 produces 131 µmol of CO per gram with a steady mass increase. These amounts do not correspond to a full re-oxidation of HEO 1 and HEO 2 but it can be observed that the mass uptakes are not fully completed when the CO₂ injection is stopped. Kinetics are indeed too slow to fully re-oxidize both materials. During the second cycle, the reduction step yields 111 and 78 µmol of O₂ per gram of oxide for HEO 1 and HEO 2, respectively. HEO 1 offers a higher reduction yield than HEO 2, however, during the re-oxidation step, HEO 2 produces slightly more CO, with 116 and 131 µmol of CO per gram for HEO 1 and HEO 2, respectively. The addition of zirconium in the material (HEO 2) enables stabilizing the CO production by avoiding a partial deactivation of the material during cycling compared to HEO 1. Comparing the CO production with state-of-the-art best-known materials (CeO₂ and La_0.5Sr_0.5Mn_0.9Mg_0.1O₃) under similar experimental conditions, both HEO 1 and HEO 2 do not overpass their performance in terms of kinetics or overall CO production (cf. Table 1).

Compared to the values reported in the literature, the performances measured for HEO 1 and HEO 2 are far below those obtained with Fe_0.25Mg_0.25Co_0.25Ni_0.25O_x [45] (450 µmol of H₂ per gram). However, the experimental conditions strongly differ and are much more favorable in [39], with a H₂:H₂O ratio of 10⁻³, a re-oxidation temperature of 800 °C and 5 h for reduction and oxidation steps. Even higher performances were obtained from microwave-assisted thermochemical cycles (2830 µmol and 4840 µmol of H₂ per gram with HEO 1 and HEO 2, respectively) [46,47].

2.2. Ceria-Based HEO (Related to HEO 3)

The HEO 3 formula is (Ce_0.2La_0.2Pr_0.2Sm_0.2Y_0.2)₂O₃, and this compound has the specificity to integrate Ce⁴⁺ cations in a typical Ce³⁺ oxide structure, which is a C-type cubic structure [48,49]. It has been reported that Ce⁴⁺ is reducible in this structure; thus, this study deals with the possibility of reoxidizing it by splitting CO₂ during redox cycles, which has never been investigated before. The XRD pattern of HEO 3 is presented in Figure 5. A single structure corresponding to the C-type cubic structure is observed after calcination at 1000 °C. This result differs from those obtained for solid-state synthesised materials as the single structure is only observed after annealing at 1500 °C in this case [48].

The lattice parameter (a) was calculated using the Bragg Equation (3).

$$2.d_{hkl}·sin\theta =n.\mathsf{\lambda }$$

(3)

Given that the lattice is cubic, the relation between the inter-reticular distance and the lattice parameter is given by (4).

$$a=d_{hkl}·\sqrt{h^2+k^2+l^2}$$

(4)

By combining these two equations,

$$a=\frac{\lambda }{2.\sin \theta }·\sqrt{h^2+k^2+l^2}$$

(5)

Values of the lattice parameter are reported in

Table 2. Lattice parameter of HEO 3.

Heat Treatment	Lattice Parameter, a (Å)
1000 °C-1 h-in air	10.966 (±0.006)
1400 °C-1 h-in air	10.929 (±0.006)
Post cycling	10.990 (±0.006)

. The calculated values (a = 10.966 Å after annealing at 1000 °C) fit those found in the literature (a = 10.956 Å [49]). After a heat treatment at 1400 °C, the lattice parameter decreases.

EDS analysis mappings are shown in Figure 6. The atomic cartography confirms the randomized distribution of cations. The atomic ratio of the different elements is in good agreement with the desired composition (

Table 3. Elemental composition of HEO 3 measured by EDS.

Elements	% (Atomic)
	After Annealing at 1400 °C	After Cycling
O	58.98	59.85
Y	7.89	8.09
La	8.34	7.89
Ce	7.95	7.56
Pr	7.71	7.33
Sm	9.13	9.28
Total %	100	100

): the corresponding composition is (Ce_0.20La_0.21Pr_0.19Sm_0.23Y_0.19)₂O_2.95 after annealing at 1400 °C and (Ce_0.19La_0.19Pr_0.18Sm_0.23Y_0.20)₂O_2.99 after cycling.

The thermogravimetric analyses of CO₂-splitting cycles are presented in Figure 7. In Figure 7a, HEO 3 was heated until 1400 °C and CO₂ was injected at 1050 °C. During the reduction step, the mass loss reaches 0.84%, which corresponds to an oxygen amount of 259 µmol/g. However, the re-oxidation yield is very low with only 23 µmol of CO produced at 1050 °C. Because the reduction in HEO 3 starts at a low temperature (500 °C), during the re-oxidation step at 1050 °C, competition between the reduction and oxidation reactions may occur, explaining the low amount of CO produced. Another thermochemical cycle was performed at a lower temperature (1000 °C for reduction and 600 °C for re-oxidation) to observe whether the CO production could be improved (Figure 7b). The reduction yield at 1000 °C is still high with 233 µmol of O₂ per gram, however, the reduced material does not react with CO₂ at 600 °C. Although HEO 3 appears to be easily reducible, its reduced counterpart cannot be re-oxidized with CO₂ in these conditions. Further investigations such as XPS analysis could help understand why the reduced material is not reactive with CO₂ even though Ce³⁺ is present in the reduced phase. Charvin et al. [55] studied mixed cerium oxides with a pyrochlore structure in which Ce³⁺ was also not directly reactive with water and required an additional activation step (3-step cycle) to re-oxidize Ce³⁺, showing that Ce³⁺ can be less reactive in a different structure, in comparison with non-stoichiometric ceria. The calculated lattice parameter is larger after cycling (Table 3), which confirms that the reduced species still remain after cycling because reduced cations have a larger ionic radius than oxidized ones. HEO 3 keeps its chemical composition (Table 3) and structure (Figure 5) after cycling without any phase modification.

Figure 8 presents the high-temperature XRD diffractograms recorded during two consecutive cycles. The temperature profile of the in-situ analysis is presented in the Materials and Methods section (Section 3). This analysis confirms that no phase change happens during cycling. A peak shift to lower angles is observed when the material is heated up to 1400 °C but it does not correspond to an effect of material reduction because it is reversible during cooling whereas the TGA analysis highlights that no re-oxidation occurs.

2.3. Perovskites-Based HEO (Related to HEO 4, HEO 5 and HEO 6)

HEO 4, HEO 5 and HEO 6 are perovskite-structured materials related to high entropy oxides. The desired chemical compositions are, respectively, (Gd_0.2La_0.2Nd_0.2Sm_0.2Y_0.2)(Co_0.2Cr_0.2Fe_0.2Mn_0.2Ni_0.2)O₃, (La_0.5Sr_0.5)(Mn_0.2Ce_0.2Ni_0.2Mg_0.2Cr_0.2)O₃ and (La_0.8Sr_0.2)(Mn_0.2Fe_0.2Co_0.4Al_0.2)O_3-δ. XRD analyses of the compounds are shown in Figure 9 after different heat treatments and after thermochemical cycling.

HEO 4 presents a single-phase structure after annealing at 1000 °C, corresponding to an orthorhombic perovskite related to SmCrO₃. The entropy-stabilized phase has been obtained at 1000 °C. The atomic composition derived from EDS corresponds to (Gd_0.20La_0.22Nd_0.22Sm_0.33Y_0.20)(Co_0.18Cr_0.20Fe_0.21Mn_0.20Ni_0.21)O_2.73, in good agreement with the desired composition. The EDS analysis reveals a homogeneous cation distribution at 1400 °C, except for an isolated yttrium-rich particle (Figure 10a) that was confirmed by XRD at 1400 °C (Figure 9). Indeed, at 1400 °C, a secondary phase appears on the XRD pattern and, after cycling, several additional phases are observed with an increased peak intensity. EDS after cycling shows a heterogeneous composition, in which three phases are observed. The SEM micrographs (Figure 10b) reveal dark particles that are composed of Co, Ni and O (EDS), bright particles of yttrium oxide and grey particles corresponding to the desired perovskite phase. The XRD analysis reveals that the new phases observed after cycling correspond to the NiO-type phase (NiCoO) and Y₂O₃-type phase (Y,Sm,Gd)₂O₃.

HEO 5 should also have a single crystallographic structure corresponding to a cubic perovskite. Unfortunately, a multi-phase composition is revealed by XRD analysis whatever the annealing temperature. The mix is composed of phases related to CeO₂, MgO, Sr₂MnO₄ and the cubic perovskite LaMnO₃. Even after annealing at 1400 °C for 3 h, the single phase is not formed. The non-homogeneity of the material was confirmed by EDS, with the detection of four different phases/areas including CeO₂, a (Mg,Ni) oxide, a Sr-rich phase associated with Mn and O, and a (La,Mn)O phase (Figure 10).

The XRD analysis of HEO 6 reveals a single structure after annealing at 1000 °C in air. The crystallographic structure is related to the LaMnO₃ perovskite. After annealing at 1300 °C, diffraction peaks are split, which means that a second perovskite phase with a different chemical composition appears after annealing.

The thermograms of two consecutive CO₂-splitting cycles with HEO 4, HEO 5 and HEO 6 are presented in Figure 11. The TGA of La_0.5Sr_0.5Mn_0.9Mg_0.1O₃ (LSMMg) is added to compare with the best-in-the-class perovskite. The performances of HEO 4, HEO 5, and HEO 6 do not overpass LSMMg with production yields of 143, 128 and 276 µmol/g of O₂ during the first reduction step at 1400 °C, and 79, 62 and 63 µmol/g during the second reduction step, respectively (Table 1). The CO production of HEO 4, HEO 5 and HEO 6 reaches 99 µmol/g, 81 µmol/g and 87 µmol/g during the first cycle, and 90 µmol/g, 79 µmol/g and 83 µmol/g during the second cycle, respectively. These CO amounts are stable during cycling but still much lower than those measured with LSMMg. Such low performances are explained by the multi-phase that appears during cycling in HEO 4 and HEO 5. Indeed, the NiO-type and Y₂O₃-type phases are not reactive during thermochemical CO₂-splitting cycles. Only the orthorhombic phase might react with CO₂ in HEO 4, and only CeO₂ and the cubic perovskite might react in HEO 5. HEO 6 shows a high reduction yield during the first heating at 1400 °C such as LSMMg, but the re-oxidation steps are associated with slow kinetics that lead to a poor CO production yield.

2.4. Ferrite-Based HEO (Related to HEO 7)

The chemical composition of HEO 7 is (Mg_0.2Co_0.2Ni_0.2Cu_0.2Zn_0.2)Fe₂O₄. The crystallographic structure is related to classical ferrites as shown in Figure 12. The calculated lattice parameter is a = 8.386(6) Å at 1000 °C and a = 8.388(6) Å at 1300 °C.

The thermogravimetric analysis of two consecutive CO₂-splitting cycles with HEO 7 is presented in Figure 13. During the first reduction at 1300 °C, a strong mass decrease is observed corresponding to an O₂ release of 407 µmol/g (Table 1). Nevertheless, the reduced HEO 7 does not significantly react with CO₂ during the oxidation step with only 44 µmol/g of CO. During the second cycle, the material still continues to release oxygen at 1300 °C (153 µmol/g) but a very small mass increase occurs during the oxidation step (20 µmol/g of CO).

3. Materials and Methods

3.1. Materials Synthesis

Table 4. List of HEO materials investigated for thermochemical redox cycles.

Name	Compositions	Structure	References
HEO 1	Fe0.25Mg0.25Co0.25Ni0.25Ox	Rocksalt/spinel	[45,46]
HEO 2	(Fe0.25Mg0.25Co0.25Ni0.25)Zr0.6Ox	Rocksalt/spinel	[47]
HEO 3	(Ce0.2La0.2Pr0.2Sm0.2Y0.2)2O3	Cubic type C	[48,49]
HEO 4	(Gd0.2La0.2Nd0.2Sm0.2Y0.2)(Co0.2Cr0.2Fe0.2Mn0.2Ni0.2)O3	Perovskite	[50]
HEO 5	(La0.5Sr0.5)(Mn0.2Ce0.2Ni0.2Mg0.2Cr0.2)O3	Perovskite
HEO 6	(La0.8Sr0.2)(Mn0.2Fe0.2Co0.4Al0.2)O3-d	Perovskite	[51]
HEO 7	(Mg0.2Co0.2Ni0.2Cu0.2Zn0.2)Fe2O4	Spinel	[54]

lists the different materials investigated in this study. Compositions of the high entropy oxides were either chosen from literature data related to an investigation of their reducibility, or because they incorporate reducible Ce⁴⁺ or Fe³⁺ cations in an original configuration.

These materials (#HEO 1–7) were synthesized by using the modified Pechini method described elsewhere [56]. Briefly, metallic nitrate precursors were dissolved in distilled water under stirring during 10 min. Citric acid was then added to the solution under stirring with a molar concentration twice that of the cations concentration (n_CA= 2 × n_cations). Ethylene glycol was then poured into the solution with a molar concentration 4-fold that of the cations concentration (n_EG= 4 × n_cations). The solution was stirred and heated at 90 °C until gelation (brownish color). The gel was then heated in a furnace at 250 °C for two hours; the resulting powder was crushed in an agate mortar and calcined in air at 1000 °C during 1 h. To prevent any sintering effect during the thermal analysis, the oxides were then calcined in air during one hour at either 1300 °C or 1400 °C, corresponding to the maximum temperature of the TGA analysis.

3.2. Materials Characterization

The obtained powders were first characterized by X-ray diffraction (XRD) using a Panalytical X’PERT PRO diffractometer with the Cu Kα radiation (αCu = 0.15406 nm, angular range = 20–80°, 2θ, tube current 20 mA, potential 40 kV). The diffraction patterns were compared with those from the literature and lattice parameters were calculated.

High-temperature X-ray diffractometer (HT XRD- Empyrean Panalytical with Anton-Paar HTK1 chamber) equipped with Cu Kα radiation was used for the in-situ investigation of phase reactivity/transitions during redox cycling (tube current 20 mA, potential 45 kV, angular range = 20–80°, 2θ, angle variation 0.02626°/s). The temperature profile of the in-situ analysis is presented in Figure 14. The powder was heated up to 1300 or 1400 °C under nitrogen flow for the reduction step and cooled down to 800 °C for the oxidation step with CO₂ injection. Two consecutive cycles were performed for a total of 17 diffractograms.

Chemical elementary cartography was realized by energy dispersive X-ray spectroscopy (EDS) analysis to both estimate the chemical composition and observe the cations distribution (surface mapping) in the material. Analyses were performed using a Zeiss Sigma 300. The accelerating voltage was 15 keV. All these characterizations allowed to validate the synthesis of HEO with a single crystallographic structure and a randomized cationic distribution.

3.3. Reactivity Testing during Thermochemical Cycles

Metal oxide redox pairs are intended to be used ultimately in a solar reactor for CO₂/H₂O splitting to produce CO/H₂. In such a device, the material is placed in a cavity receiver (e.g., as powder or foam) in which the concentrated solar beams are focussed at the reactor aperture [19]. Because of technological concerns (reactor design and high-temperature materials’ resistance and stability), the maximum temperature should not exceed approximately 1400–1500 °C. In a typical solar thermochemical cycle, the oxide is heated up to ~1400 °C under argon flow (reduction) and then cooled down at below 1100 °C with H₂O or CO₂ injection for the splitting step (temperature-swing cycle). Several successive cycles must be operated during a day, which implies that the redox materials’ microstructure and morphology have to be stable to warrant a repeatable thermochemical performance (i.e., no reactivity decrease during the cycling and fast kinetics during the splitting step). To investigate the materials’ activity under representative conditions comparable to those found in solar experimental reactors, a thermogravimetric analysis was carried out using similar temperatures and atmosphere compositions.

Once the powders were fully characterized, their thermochemical activity during two-step CO₂-splitting cycles was investigated. The thermogravimetric analysis (TGA) was used to measure the mass variations (amount of oxygen exchanged) associated with the reduction and the oxidation steps of thermochemical cycles. TGA was performed using a SETARAM Setsys Evo 1750 apparatus in a controlled atmosphere. Approximately 100 mg of material (exact mass determined with a high precision balance) was introduced in a platinum crucible hung to the microbalance with platinum suspensions and placed inside the furnace chamber. Then, residual air in the furnace chamber was eliminated by preliminary pumping to operate in an inert atmosphere and to improve the reduction yield thanks to the low oxygen partial pressure. The sample was heated in argon flow (20 mL/min) at a heating rate of 20 °C/min up to the selected temperature set-point and the mass variation was registered continuously. The successive steps of the thermochemical cycles were realized at different temperatures (typically 1400 °C or 1300 °C dwelled for 45 min during the reduction step and 1050 °C dwelled for one hour during the CO₂-splitting step) and CO₂ was injected during the oxidation step (50% CO₂ in Ar with a CO₂ flowrate of 10 mL/min). The mass variations associated with the temperature-programmed cycles correspond to the oxygen release during reduction and the oxygen uptake during CO₂-splitting, which allow the calculation of both the reduction yield and the CO production yield.

4. Conclusions

High entropy oxides (HEOs) open a new route for renewable synthetic fuel production through solar-driven thermochemical cycles. This study compares the thermochemical performance and behavior of a series of HEO formulations used as redox catalysts during two-step redox cycles. Different formulations of HEO were synthesized and characterized based on various structural configurations and cationic substitutions in the crystal lattice of the materials. Furthermore, their redox activity was investigated by focusing on their ability to perform CO₂ splitting during thermochemical cycles while quantifying the fuel production yields, rates and performance stability. Compared to previous values reported in the literature, HEO 1 and HEO 2 related to iron oxides do not exhibit an equivalent performance but they offer fuel production yields higher than those of ceria under similar conditions. Other HEO formulations were considered in this work and tested for the first time for thermochemical cycles application. Unfortunately, relatively small amounts of CO were produced per cycle for the other tested materials and experimental conditions. The ceria-related high entropy oxide (HEO 3) and the ferrite-related high entropy oxide (HEO 7) do not produce CO during thermochemical cycling, while the high entropy perovskites (HEOs 4, 5, and 6) do not exhibit higher performances than LSMMg. The HEO phase formation is strongly dependent on the heating rate or quenching. The reversible multi-phase to single-phase change could be integrated in a redox cycle and the high entropy variation between the different steps could permit the decrease in the working temperatures. The wide compositional space offered by HEO enables the design of new redox-active materials devoted to H₂O- or CO₂-splitting cycles using concentrated solar energy and should lead to abundant and high-impact research in this field.