Thermochemical splitting of CO 2 using solution combustion synthesized lanthanum – strontium – manganese perovskites

Redox reactivity of La (1-x) Sr x MnO 3 (LSM) perovskites towards a solar thermochemical CO 2 splitting (CS) cycle is investigated. The LSM perovskites are synthesized via a solution combustion synthesis (SCS) method using glycine as the reducing agent. Multiple analytical techniques are used for the structural characterization of the LSM perovskites. Thermogravimetric thermal reduction (TR) and CS cycles (in three sets: one, three and ten cycles) are conducted to estimate the amounts of O 2 released ( n O 2 ) and CO produced ( n CO ) by each LSM perovskite. Higher n O 2 by each LSM perovskite, as compared to the n CO during the first cycle. The n O 2 is decreased, and the re-oxidation capacity of each LSM perovskite is improved from cycle one to three. In terms of the average n O 2 and n CO from cycle 2 to cycle 10, the La 0.60 Sr 0.41 Mn 0.99 O 2.993 (214.8 μ mol of O 2 /g ⋅ cycle) and La 0.30 Sr 0.70 Mn 0.99 O 2.982 perovskites (342.1 μ mol of CO/g ⋅ cycle) are observed to have the uppermost redox reactivity. The redox reactivity of all the LSM perovskites (except for La 0.88 Sr 0.11 Mn 1.00 O 2.980 ) is recorded to be higher than that of the widely studied CeO 2 material.


Introduction
Meeting the ever-increasing worldwide demand for energy by using solar fuels produced via H 2 O (WS)/CO 2 splitting (CS) reactions is one of the feasible approaches for harnessing the renewable and profusely available solar energy [1,2]. This process exploits the concentrated solar power to drive the high-temperature thermal reduction (TR) of metal oxides (MOs) [3]. Effectively, by applying this technology, the storage of solar energy in the form of chemical energy is possible [4,5]. This stored chemical energy is preferred as it can be transported and stockpiled for a long time without any degradation. Recently, Marxer et al. [6] developed a first of its kind pilot-scale set up for the production of 700 L of solar syngas via 291 stable redox cycles.
According to the studies so far, the process efficiency of the thermochemical WS/CS process heavily depends on the redox properties of the MOs [7]. The desirable characteristics of a good MO include high H 2 /CO yield, faster TR and re-oxidation (RO) kinetics, lower cycle time, higher thermal stability over multiple cycles, elevated O 2 diffusion rates from the surface to the bulk of the MO, and a smaller temperature difference between the TR and RO steps. The redox materials investigated for both WS and CS reactions include volatile MOs such as zinc oxide [8][9][10] and tin oxide [11,12], and non-volatile MOs for instance, iron oxide [13][14][15], ferrites [16][17][18][19], doped ceria [20][21][22][23], and hercynite [24,25]. Among these, ceria and doped ceria appears to be a beneficial choice, as these oxides possess the anticipated material properties.
In recent years, perovskite-based oxides [26][27][28] were investigated for thermochemical WS/CS reactions with an assumption that they will outperform ceria. Among these perovskites, the La-Sr-Mn-based (LSM) perovskites were inspected for the production of both H 2 and CO via WS and CS cycles. For the TR temperature range of 1523 to 1923 K, the La 0.7 Sr 0.3 MnO 3 (LSM30) and La 0.6 Sr 0.4 MnO 3 (LSM40) perovskites showed higher TR yield as compared to ceria [29]. According to Yang et al. [30], the La 1-x Sr x MnO 3 materials (x = 0 to 0.5) represented a higher TR yield as a function of an increase in the atomic concentration of Sr. Demont and Abanades [31] synthesized and tested the La 0.35 Sr 0.65 MnO 3-δ , La 0.5 Sr 0.5 MnO 3-δ , La 0.65 Sr 0.35 MnO 3-δ , and La 0.8 Sr 0.2 MnO 3-δ towards the CS reactions (in one cycle) and reported La 0.5 Sr 0.5 MnO 3-δ as the best choice for producing a maximum amount of O 2 (n O2 = 195 μmol/g⋅cycle) and CO (n CO = 242 μmol/g⋅cycle). Dey and Rao [32] studied the La 1-x Sr x MnO 3 (x = 0.3, 0.4, and 0.5) towards splitting of CO 2 at isothermal operating conditions and observed high CO production (n CO = 134 μmol/g⋅cycle) by La 0.5 Sr 0.5 MnO 3 at 1400 • C (partial pressure of O 2 = 10 -5 atm and partial pressure of CO 2 = 1 atm). Dey and Rao [32] further reported that the La 0.5 Sr 0.5 MnO 3 was capable of producing ~three times higher CO than the CeO 2 at 1500 • C. As per the several studies, the rise in the Sr content was responsible for an increase in the reduction extent and decreased in the re-oxidation yield [29,31,[33][34][35][36][37][38].
As per the literature review, selected LSM compositions were investigated for solar thermochemical fuel production. It is essential to find the finest combination of LSM perovskite, which is capable of producing the maximum amount of solar fuel in multiple thermochemical cycles. In this regard, this work concentrates on the exploration of the TR and RO behavior of the LSM perovskites, i.e., La (1-x) Sr x MnO 3 (where x = 0.1 to 0.9) in multiple CS cycles (Fig. 1). This study will assist in improving the understanding and rational design for the application of LSM perovskites for WS/CS reactions and in comparing them with the CeO 2 .

Chemicals
a., ≥97.0%), and Sr(NO 3 ) 2 (ACS reagent, ≥99.0%) were obtained from Sigma Aldrich, USA, and used as received. De-ionized water (ultrapure Type 1 from Direct-Q system, Millipore, France) was consumed for the preparation of the solution containing metal precursors and glycine. Ultra-high pure Ar and CO 2 -Ar (50:50) were obtained from Buzwair Gases, Qatar.

Synthesis of LSM perovskites
Synthesis of the LSM perovskites was carried out via a glycine based solution combustion synthesis (SCS) method. In a typical recipe, calculated amounts of precursors (to prepare 1 g of the LSM perovskites) [39] were dissolved in 50 ml of deionized water along with the glycine (fuel to precursor ratio = 1). As-prepared solution was then pre-heated up to 300 • C (exposed to air) by using a temperature-controlled hot plate. A viscous gel comprised of the precursors and glycine was obtained after complete evaporation of the water. The combustion reaction was propagated as the temperature of the gel was increased up to the auto-ignition temperature. As-synthesized powder of the LSM perovskites was crushed using pestle and mortar and further calcined up to 1000 • C in the air for 4 h using a muffle furnace. Table 1 reports the abbreviations assigned to each LSM perovskite derived via the SCS method. By using the powder X-ray diffractometer (PXRD, PANalytical XPert MPD/DY636), the phase composition of the LSM perovskites was determined. The scanning electron microscope (SEM, Nova Nano 450, FEI, 200kx) equipped with the energy-dispersive X-ray spectroscopy (EDS) was utilized to identify the elemental composition and material morphology.

CO 2 splitting experiments
Multiple thermochemical experiments were conducted by using a high-temperature SETSYS Evolution TGA (Setaram Instruments, France) ( Fig. 2). Details allied with the TGA set-up were already reported elsewhere [40]. A platinum pan was placed inside the furnace (surrounded by an alumina tube) to support the reference and the sample alumina (100 µl) crucibles. The flowrates of the gases were monitored and controlled by using mass flow controllers, and the temperature was regulated by using a Pt-Rh type-B thermocouple. Thermogravimetric CS experiments were carried out by using approximately 50 mg of the LSM perovskite powder. The drift in the mass of the LSM perovskite during the TR (at 1400 • C for 60 min, 100 ml/min Ar), and CS (1000 • C for 30 min, 100 ml/min of 50% CO 2 /Ar mixture) steps were recorded by using the Calisto software. Based on the variation in the mass during both steps and following equations, the amounts O 2 released (n O2 ) and CO produced (n O2 ) by each LSM perovskite were estimated.
In the above equations, the Δm loss and Δm gain represents the loss and gain in the mass of the LSM perovskites during the TR and CS steps. The molecular weights of the O and O 2 are represented as M O and M O2 . Likewise, the amount of LSM perovskite used during the TGA experiments is presented as M LSM . TGA blank runs (performed by using empty crucible) were subtracted from the TGA actual experiments (conducted by using the LSM perovskites) to avoid the effect of the thermal buoyancy.

Results and discussion
The phase composition of the SCS synthesized LSM perovskites were identified by performing the PXRD analysis. The PXRD peaks, presented in Fig. 3, shows nominally phase pure LSM perovskites with no evidence of any impurities such as La-, Mn-, Sr-based individual oxides, or La, Sr, Mn metals. As the crystal ionic radii of Sr (132 pm) is higher than that of La (117.2 pm), the increase in the atomic concentration of the Sr resulted in a shift in the PXRD peaks towards higher 2θ angle. This observation provided further confirmation of the successful synthesis of LSM perovskites via the SCS method. The PXRD peaks reported in Fig. 3 matched very well with the PXRD findings reported in previous studies [41]. By employing the Scherrer formula, the average crystallite size of all LSM perovskites was estimated to be in the range of 50 to 70 nm.
In addition to the PXRD, by using the SEM/EDS instrument, the elemental composition of each LSM perovskite was verified. The EDS results were observed to be consistent with the findings reported by the PXRD analysis. The EDS patterns associated with the LSM20, LSM40, LSM60, and LSM80 perovskites (exemplified) are presented in Fig. 4. Besides, the atomic concentrations of La, Sr, and Mn, and the chemical composition of each LSM perovskite identified by using the EDS patterns are reported in Table 1.
In order to examine the microstructural morphology of the LSM perovskites, SEM analysis was conducted. The representative images obtained for LSM20, LSM40, LSM60, and LSM80 perovskites are shown in Fig. 5. The drift in the La and Sr atomic concentrations had an insignificant effect on the LSM morphology. However, as the images were taken at different locations (for each LSM), we cannot wholly neglect the chances of having dissimilar/disordered images. The SEM analysis showed that all the LSM perovskites possess a porous  morphology.
We have first tested the redox performance of the SCS synthesized LSM perovskites in one thermochemical CS cycle. In thermochemical cycles, the mass of a MO decreases during the TR step due to the release of lattice O 2 . In contrast, due to the gain of O 2 , the mass of the MO increases during the re-oxidation step. The mass of each LSM perovskite was decreased as a function of the increase in the TR temperature (T H ) and the reaction time during the TR step (Fig. 6). In terms of the mass loss, the LSM50 perovskite reached its plateau after attaining the TR temperature of 1400 • C. In contrast, the remaining LSM perovskites continued to lose weight throughout the entire TR step. By considering that the TR starts at 800 • C, the kinetics was quickest for LSM90 and the slowest for the LSM10 perovskite. Furthermore, from 800 • C to 1400 • C (dwell time equal to 60 min), the %Δm loss was highest for LSM90 perovskite (~6.8%) and lowest in the case of the LSM10 perovskite (~0.65%).
The TGA profiles obtained in case of each LSM perovskite (during the TR step) were translated into the n O2 by using Eq. (1). The numbers reported in Fig. 7 indicate that the n O2 by the LSM90 perovskite was the greatest as compared to other LSM perovskites. For instance, the n O2 by LSM90 perovskite (1476.6 μmol/g) was higher by 1298.7, 1243.5, 1003.5, 932.0, 885.9, 866.7, 828.3, and 675.0 μmol/g when compared to the LSM10, LSM20, LSM30, LSM40, LSM50, LSM60, LSM70, and LSM80 perovskites, respectively. The inclusion of Sr +2 as a partial substitute for La +3 on A-site (from x = 0.1 to 0.9) results into a deviation in the oxidation state of Mn from +3 to +4. Because of this drift, the LSM perovskites (with higher Sr content) seems to be favorable towards TR reaction as compared to the LSM perovskites with the lower Sr content.
The thermally decomposed LSM perovskites were further examined towards CS step at 1000 • C for 30 min. Fig. 8 represents the TGA profiles associated with the first CS step. As mentioned earlier, it was expected that the mass of the MO would rise due to the re-oxidation. As per the expectation, as shown in Fig. 8, the mass of each LSM perovskite was increased during the CS step. Close inspection of these TGA profiles indicates that the LSM70 and LSM50 perovskites exhibited the quickest RO rates. At the same time, the LSM90 perovskite showed the slowest RO kinetics. Obtained results further show that the LSM70 perovskite attained the maximum mass gain, and the LSM90 perovskite indicated the lowest increase in the mass during the first CS step. With the help of the obtained TGA profiles and Eq. (2), the calculated n CO by each LSM perovskite is presented in Fig. 9. The LSM70 perovskite displayed the highest CO production (424.0 μmol/g), and the LSM90 displayed the lowest n CO (64.7 μmol/g) in cycle 1. The LSM40, LSM60, and LSM80 perovskites indicated approximately identical n CO in the range of 285.7 to 309.5 μmol/g. Likewise, the LSM20 and LSM30 perovskites produced ~162.0 μmol of CO/g in cycle 1. The authors would like to admit that it is difficult to know the exact reason for such results as the TGA set-up only provides the information related to the mass variations.
The overall analysis of the 1st thermochemical cycle indicate that the n CO produced by the LSM perovskites were lower than the n O2 . The probable reasons for such trends are a) all the LSM perovskites were freshly prepared and never underwent any thermal cycling before performing the first TR step, and b) all the LSM perovskites have not reached their thermal stability during cycle 1. It was highly essential to know if the LSM perovskites show a similar trend in additional cycles. Hence, these redox materials were further scrutinized by performing three consecutive cycles (Fig. 10).     The results obtained during three consecutive cycles indicate a reduction in the n O2 by all the LSM perovskites. In contrast, an upsurge in the n CO /n O2 ratio was noticed for all the LSM perovskites from cycle 1 to cycle 3. The probable reason for these results is the transformation of the redox reactivity of the selected LSM perovskites from an unstable zone (cycle 1) to a more stable zone (cycle 3). This probably happened as the LSM perovskites underwent three thermal cyclings. After the 3rd cycle, it is believed that some of the LSM perovskites have improved their thermal stability and redox reactivity towards the TR and CS reactions.
To determine the worthiest combination of the LSM perovskite, which can attain a stable production of CO for a longer duration, this study was further extended towards performing ten successive thermochemical cycles. Fig. 11 presents the TGA profiles obtained (experimental time ~22 h) during ten thermochemical CS cycles. As each LSM perovskite has behaved differently towards the multiple TR and CS steps, the TGA profiles reported for all LSM perovskites (Fig. 11) look different than each other. To avoid the misrepresentation of the data, the average n O2 , n CO , and n CO /n O2 ratio was hereafter calculated by excluding the first cycle (which is deemed as the most unstable zone). Similar to the previous sections, the TR (Fig. 12) and CS (Fig. 13) capacity of all the LSM perovskites, from cycle 2 to cycle 10, estimated by using Eqs. (1) and (2).
According to the data reported in Fig. 13, the LSM20, LSM50, LSM60, and LSM70 perovskites exhibited stable production of CO from cycle 2 to cycle 10. Remaining perovskites showed varying amounts of CO production and especially in case of the LSM10 the n CO decreased with an increase in the number of cycles. LSM perovskites can be organized, based on the average n CO from cycle 2 to cycle 10, as: LSM70 The LSM perovskites were further compared with each other based on their average n CO /n O2 ratio from cycle 2 to cycle 10. As per the obtained data, the average n CO /n O2 ratio in the case of LSM60 and LSM80   perovskites was the highest as compared to the other LSM perovskites. For most of the LSM perovskites (except for LSM40 and LSM90) the average n CO /n O2 ratio was higher than 1.6. The average n CO /n O2 ratio > 1.6 indicates that the re-oxidation capacity of the LSM perovskites is reduced, and further work needs to be done for its improvement. A possible option is to improve the ion mobility of the LSM perovskites via the inclusion of one or more suitable dopants in the crystal structure. Overall, the LSM perovskites, based on their average n CO /n O2 ratio, can be organized as: Table 3 reports the comparison between the LSM perovskites and CeO 2 (experiments conducted at identical operating conditions). The data listed indicate that, except for LSM10 perovskite, all the LSM perovskites produced higher n O2 and n CO than CeO 2 . In contrast, CeO 2 exhibited higher average n CO /n O2 ratio than most of the LSM perovskites. The elevated average n CO /n O2 ratio indicates that the RO Table 2 LSM perovskites: n O2 , n CO , and n CO /n O2 ratio in three cycles (TR at 1400 • C for 60 min, CS at 1000 • C for 30 min, and heating/cooling rate = 25 • C/min

Table 3
Comparison between the results recorded in this study and the findings reported by Bhosale and Takalkar [40], Dey and Rao [32] and Yang et al. [30]. 1400 -203.8 -potential of CeO 2 is superior to the LSM perovskites. We believe that the CS temperature (1000 • C) and re-oxidation time (30 min) used were appropriate for the CeO 2 material to re-oxidized completely. However, the experimental conditions employed were insufficient for the LSM perovskites to regain their oxidized state. In addition to CeO 2 , the LSM perovskites investigated in this study were also compared with LSM perovskites studied by Dey and Rao [32] and Yang et al. [30]. Dey and Rao [32] tested LSM30, LSM40, and LSM50 perovskites for the thermochemical splitting of CO 2 at isothermal experimental conditions (1673 K). The amounts of O 2 released and CO produced by the LSM30, LSM40, and LSM50 perovskites investigated by us were considerably higher than the similar perovskites examined by Dey and Rao [32]. The probable reason for this lower amount of fuel production reported by Dey and Rao [32] is the lesser time (15 min) permitted for the TR step.

Materials
Yang et al. [30] examined LSM10, LSM20, LSM30, and LSM40 perovskites for the thermochemical splitting of H 2 O (TR at 1400 • C and reoxidation at 800 • C). As the aim of the presented investigation was to split the CO 2 , only the results associated with the TR step (O 2 release) were compared. As the time allowed for the TR step by us and by Yang et al. [30] were comparable (especially for LSM30 and LSM40), the amount of O 2 released by LSM10, LSM20, LSM30, and LSM40 in this study was approximately identical to the TR ability of the similar perovskites reported by Yang et al. [30]. Based on the comparison with published literature, presently, we are investigating the effect of temperature and dwell time allied with both TR and CS steps to understand the long-term redox reactivity of the LSM perovskites.

Summary and conclusions
In this investigation, La (1-x) Sr x MnO 3 (where x = 0.1 to 0.9) i.e., LSM perovskites were examined towards thermochemical CS cycles. A solution combustion synthesis approach was utilized for the synthesis of LSM perovskites. Derived LSM perovskites were further characterized using PXRD, EDS, and SEM techniques. Formation of nominally phase pure LSM perovskites with no evidence of any impurities such as La-, Mn-, Srbased individual oxides or La, Sr, Mn metals was confirmed from the PXRD and EDS analysis. The average crystallite size of all LSM perovskites was estimated to be in the range of 50 to 70 nm. The SEM analysis verified the insignificant effect of the variation in the La and Sr atomic concentrations on the porous morphology of the LSM perovskites. The CS ability of each LSM perovskite was examined by performing one, three, and ten thermochemical cycles (at a fixed T H = 1400 • C and T L = 1000 • C). Obtained results indicate that the rise in the Sr molar concentration is favorable to improve the TR yield of the LSM perovskites. As the number of cycles increased from two to ten, most of the LSM perovskites reached their thermal stability and produced roughly stable amounts of O 2 and CO. The long term thermal cycling (from cycle 2 to cycle 10) shows that the LSM40 perovskite was the best choice in terms of n O2 = 214.8 μmol/g⋅cycle, whereas LSM70 showed the highest activity towards CS reaction (n CO = 342.1 μmol/g⋅cycle). In terms of the re-oxidation ability (n CO /n O2 ratio), the LSM60 was observed to be the most promising choice as it is capable of attaining the highest n CO /n O2 ratio = 1.96. All the LSM perovskites (except LSM10) exhibited higher amount of O 2 release and CO production when compared with the ceria material. The improved fuel production capacity of the LSM perovskite will result into a higher solar-to-fuel-energy conversion efficiency.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.