Operando Soft X-ray Absorption of LaMn_1–xCo_xO₃ Perovskites for CO Oxidation

Abstract

We employed operando soft X-ray absorption spectroscopy (XAS) to monitor the changes in the valence states and spin properties of LaMn_1–xCo_xO₃ catalysts subjected to a mixture of CO and O₂ at ambient pressure. Guided by simulations based on charge transfer multiplet theory, we quantitatively analyze the Mn and Co 2p XAS as well as the oxygen K-edge XAS spectra during the reaction process. The Mn sites are particularly sensitive to the catalytic reaction, displaying dynamics in their oxidation state. When Co doping is introduced (x ≤ 0.5), Mn oxidizes from Mn²⁺ to Mn³⁺ and Mn⁴⁺, while Co largely maintains a valence state of Co²⁺. In the case of LaCoO₃, we identify high-spin and low-spin Co³⁺ species combined with Co²⁺. Our investigation underscores the importance to consider the spin and valence states of catalyst materials under operando conditions.

1. Introduction

The release of CO into the atmosphere has significant adverse impacts on both human health and the environment.¹ It is imperative to develop an efficient catalyst to mitigate CO emissions from fuel vehicles. Historically, Pt-based catalysts were the first to demonstrate a high level activity of CO oxidation at temperatures below 200 °C,^2,3 but high cost, low abundance and inferior thermal stability limit their widespread applications.⁴ Over the last decades, La-based perovskites (LaTMO₃, TM = transition metal) have garnered attention regarding thermal catalytic reactions of CO oxidation, NO_x reduction, and hydrocarbon oxidation.^5,6 In particular, LaCoO₃ stands out as one of the potential catalysts for CO oxidation at moderate temperatures due to its ability to adjust its morphology, size and electronic structure.^7,8 Lu et al.⁹ reported a mesoporous LaCoO₃ catalyst that showed 100% CO conversion at a temperature of ∼130 °C, where the Mn incorporation could optimize the catalytic activity and thermal stability. LaMn_1–xCo_xO₃ polycrystalline samples have been studied by ex-situ X-ray spectroscopy, suggesting a divalent Co²⁺ ion and a Mn⁴⁺–Mn³⁺ double-exchange at low concentrations of Co (15–20%), while at x = 0.5, the systems were found to contain a combination of Co²⁺–Mn⁴⁺.^10,11

CO oxidation (2CO + O₂ → 2CO₂), a prototypical reaction, is important to understand the heterogeneous catalysis surface mechanism. CO₂ cannot be formed when gaseous CO directly interacts with oxygen adsorbed at the surface, i.e., the Eley–Rideal mechanism. The reactants have to be coadsorbed onto the surface of the catalysts, resulting in the oxidation of CO, i.e., the Langmuir–Hinshelwood mechanism.¹² The kinetic steps of CO oxidation are as follows, where (i) is an elementary step and (ii) and (iii) are combined reactions:¹³

i

ii

iii

Bonn et al. investigated the thermal excitation and picosecond (ps) laser excitation for CO oxidation on the O-covered Ru(0001) surface in vacuum,¹² and showed that ps-laser excitation of CO on O/Ru(0001) can give oxidation while thermal excitation cannot. Traditionally, the thermal reaction on a catalyst’s surface is driven by phonons to overcome the activation barrier, but the picosecond laser excitation reaction dynamics involve hot substrate electrons activation. Knowledge about the electronic parameters of catalysts at working conditions, such as the charge-transfer, the oxidation state, and the covalence, are important descriptors for the development of new catalysts.⁶ Recently, a number of studies on perovskites have found correlations between the electronic structure and catalytic activity in both electrocatalysis^14,15 and thermochemical catalysis applications.¹⁶ Mueller et al.¹⁷ demonstrated that the adsorbed molecules near the Fermi level modify the electronic structure and covalency of the catalyst surface. The details of covalent bonding can be probed by oxygen K-edge X-ray absorption spectroscopy (XAS).¹⁸ In combination with X-ray emission spectroscopy (XES), the charge-transfer energy (Δ_ct) can be derived. Δ_ct is defined as the (many-body) excitation energy between the 3dⁿ and 3dⁿ⁺¹L configurations, where L is a hole in the ligand valence band.¹⁹ Volcano-type plots were established in heterogeneous catalysis, where details regarding the spin state and orbital occupations in LaCoO₃ are complex.^6,14−16 In a LaCoO₃ single crystal, Co³⁺ is low-spin (LS) at low temperature and gradually increases in high-spin (HS) character.²⁰ The variation of the Co spin and valence state for oxidation or reduction involved in reactant interaction in catalysis depends on the reaction conditions, such as temperature and pressure.²¹ LaMnO₃, mainly octahedrally coordinated with six oxygens, is an antiferromagnetic insulator below the Neel temperature (T_N ≈ 140 K), in which the Mn³⁺ ions have a half-filled e_g orbital and undergo a Jahn–Teller distortion.²²⁻²⁴ Mixed valence states can be formed, for example, by La and/or transition metal substitution, interface engineering, and/or nonstoichiometric oxygen. With Co incorporation, the LaMn_1–xCo_xO₃ compounds can form different combinations of oxidation states of Co and Mn, including Co²⁺, Co³⁺, Mn²⁺, Mn³⁺, and Mn⁴⁺. Among them, the Co³⁺ site can be activated by temperature, magnetic field, pressure, and lattice strain.²⁵⁻²⁷

Operando characterization is vital to capture the complexity of a catalyst under working conditions. Hard X-ray XAS is commonly employed in operando catalysis studies. Transition metal K-edge XAS provides bulk information, where the near-edge structure (XANES) gives electronic structure information and the extended fine structure (EXAFS) provides geometric information. Operando soft X-ray XAS at the transition metal 2p (L_2,3 edges) and oxygen 1s (K-edge) edge provide a different view on catalyst materials with a number of advantages:

• 1)Operando soft X-ray XAS detected with electron yield detection has an approximately 4 nm probing depth and as such is (near) surface-sensitive.
• 2)The lifetime broadening of soft X-ray XAS is ∼200 meV, which implies that the spectral features have an at least 5 times higher spectral resolution. This much improved intrinsic resolution reveals a larger detail and accuracy regarding the valence state and ground state symmetry, also due to the fact that the L_2,3 edges directly probe the important 3d states.
• 3)The metal L_2,3 edges can be combined on the same beamline with the carbon and oxygen K-edge spectra, allowing the tracking of both catalyst and reactants.

In this work, we study LaMn_1–xCo_xO₃ catalysts for the CO + O₂ reaction mechanism at ambient pressure using soft X-ray absorption spectroscopy. The TM L-edges and O K-edges have been measured by electron yield under operando conditions. Conducting operando soft XAS experiments at 1 bar gas phase reactions are not easy due to the relatively short penetration depth of soft X-rays.^28,29 With the help of SiN_x windows, we separate the vacuum and operando environments, allowing electron yield detection by XAS.²⁸⁻³² Using charge transfer multiplet theory simulations, we quantitatively identify the dynamics of the valence and spin states. The results are dependent on the reaction conditions and different from previous in situ studies at low pressure (0.37 kPa)⁶ and from vacuum conditions.

2. Experimental Methods

Synthesis of Perovskite Co-Doped LaMnO₃ Catalysts

The perovskite Co-doped LaMnO₃ samples were synthesized via a sol–gel method.³³⁻³⁵ Stoichiometric amounts of metal nitrate salts (La(NO₃)₃·6H₂O, Co(NO₃)₂·6H₂O, and Mn(NO₃)₂·4H₂O) and citric acid (∼5 times the used metal nitrate amounts) were dissolved in 250 mL of deionized water. The resultant solution was heated at 80 °C under stirring to form a gel, and at 150 °C was treated for 12 h to decompose, forming a solid. The solid was decomposed at ∼400 °C for 5 h to remove the organic components and further calcined at 900 °C for 5 h in a ceramic reactor with a ramp rate of 8 °C min^–1 to yield the LaMnO₃, LaCoO₃, LaMn_0.5Co_0.5O₃, LaMn_0.75Co_0.25O₃, and LaMn_0.25Co_0.75O₃ perovskite nanoparticles.

Atomic Structure Characterization

Powder X-ray diffraction data were recorded at room temperature on a Bruker AXS D2 Phaser diffractometer using Co Kα radiation (λ = 1.790 Å) at 30 kV and 10 mA with 2° min^–1 and steps of 0.01°.

Operando Reactor

The experiments were performed at the APE-HE beamline of the Elettra Synchrotron in Trieste, with proposal no. 20205379. The beamline was equipped with an operando cell for gas phase reactions between a few mbar to 1 bar pressure.³⁰⁻³² All spectra were collected in total electron yield (TEY) mode by measuring the drain current and applying a bias of 60 V between the sample and the nano-Si₃N₄ membrane separating the sample environment from the UHV chamber hosting the reaction cell. The Agilent 490 Micro Gas Chromatography system was employed to monitor the catalytical products occurring in the reactor during the operando XAS measurements.^31,32

Operando Soft XAS Measurements

The samples in the form of powders were mounted on titanium sample holders by mechanical compression and introduced in the operando reactor cell. The operando reactor was equipped with a gas line composed of three gas flowmeters, and the experiments were performed with a total flow of 50 mL·min^–1 at 1 bar of total pressure, unless stated otherwise. The measurement procedure is as follows: the first step of the experiment consisted of a thermal treatment in a He (100%) atmosphere in order to remove the surface contaminants (step 1). In case of the LaMnO₃, LaMn_0.5Co_0.5O₃, LaMn_0.75Co_0.25O₃, and LaCoO₃ samples, the temperature of the samples was increased from room temperature to 350 °C under He. After cooling down to room temperature (step 2), except in the case of LaCoO₃ which was cooled to 200 °C, similarly to previous experiments,³⁶ the samples were heated again to the target temperature at a rate of 5 °C min^–1 under a 12% CO plus O₂ gas mixture (V_CO:V_O2 = 1) (step 3). At the same time, the gas chromatograms measured on the exhaust gas coming out from the reaction cell were acquired in order to analyze the reaction products. The measurements were performed after waiting for 20 min in the target temperature. Then, keeping the samples at the maximum temperature, the 6% CO/6% O₂ mixture was removed and substituted with a 12% CO gas (step 4). The Co L_2,3, Mn L_2,3, La M_4,5, and oxygen K-edge XAS have been recorded with ∼20 cycles of fast scan cycles with a photon energy step of 0.05 eV. The energy resolution was ∼0.3 eV full width at half-maximum (fwhm).

3. Theoretical Simulations

The 2p XAS spectra of Mn and Co have been calculated with charge transfer multiplet theory based on a cluster model of MO₆ (M is Co or Mn and O is the ligand oxygen) using the Quanty program.³⁷⁻⁴² We used the Anderson impurity model with two configurations on the basis of on-site Coulomb repulsion interaction U_dd, ligand to metal change transfer Δ, symmetry-dependent hopping (V_t2g and V_eg), and the core-hole potential (Q_2p). We only consider two configurations because the third configuration does not visibly affect the 2p XAS spectra. The electron–electron interactions are parametrized with Slater integral parameters F_dd², F_dd⁴, F_pd², G_pd¹, and G_pd³ and are calculated from first-principles with the Cowan code.⁴³ The configuration weights are calculated with CTM4XAS.⁴⁴ The energy diagrams and ground-state projections are carried out with CTM4DOC.⁴⁵ The temperature-dependent XAS calculations used a Boltzmann distribution over all low-lying configurations.⁴² Given that the sample is in powder form, all theoretical XAS calculations consider isotropic spectra summed over the different polarizations. The calculated photon energy is adjusted to match the experimental spectra.

4. Results

We start by showing the XAS spectra and data analysis of the four samples one by one. In the Discussion, we compare the main results from the samples in relation to their catalytic activity.

4.1. Heating of LaMnO₃ in Helium Atmosphere

Figure 1 shows the pretreatment heating (step 1) of the LaMnO₃ nanoparticles in 1 bar He atmosphere from room temperature (r.t. is 25 °C) to 350 °C. Figure 1a shows temperature-dependent Mn 2p XAS spectra, in which the Mn²⁺ character increases (indicated with an orange rectangle). Due to the surface carbon contaminants, we observe a reduction of the catalyst, where we note that electron yield probes the top 4 nm of the sample. Because of this surface reduction, we have pretreated all catalysts samples by heating in He. In case of the soft X-ray L-edges, the white line is much larger than the edge jump. In addition, there is background from the beamline plus potential saturation effects. This makes it usually unreliable to normalize the XAS edges to their edge jump.

Figure 1.

Pretreatment heating study of LaMnO₃ in 1 bar He gas. (a) Mn L_2,3 XAS spectra from room temperature to 350 °C and after cooling down to 200 °C. (b) Charge transfer multiplet simulations of Mn³⁺ (D_4h) and Mn²⁺(O_h) 2p XAS in comparison with the experiments. (c) The components of Mn³⁺ and Mn²⁺ over the measurements. (d) Temperature-dependent average oxidation state. All spectra have been measured with total electron yield and are normalized from 0 to 1.0, indicated with “normalization (to) arbitrary units (a.u.)”.

Figure 1b shows charge transfer multiplet simulations of high-spin Mn³⁺ (3d⁴) 2p XAS in tetragonal (D_4h) symmetry and high-spin Mn²⁺ (3d⁵) 2p XAS in octahedral (O_h) symmetry. The related parameters for the calculations are explained in the Supporting Information (Table S2). Figure 1c shows the population of Mn³⁺ and Mn²⁺ over the measured temperature range, where we assume that the r.t. and 70 °C spectra are pure Mn³⁺ and the spectrum after cooling down to 200 °C is pure Mn²⁺. This is in agreement with the simulations in Figure 1b. We note that the Mn²⁺ spectrum shows saturation effects (indicated with the green rectangle in Figure 1b), which have also been observed in previous measurements;⁴⁶ more details of the spectral fitting are given in the Supporting Information (Figure S3). Figure 1d shows the temperature-dependent average oxidation state graphically.

4.2. Operando CO Oxidation of LaMnO₃

Figure 2a shows a scheme of the operando reactor, where the XAS spectrum is measured by TEY drain current.³⁰Figure 2b shows the Mn L_2,3-edge of LaMnO₃ under the working condition of 6% CO/6% O₂ in He at 1 bar total pressure. Spectrum ① is measured at 200 °C and was reoxidized to Mn³⁺. The spectral shape does not change (significantly), indicating that the oxidation state remains constant at Mn³⁺. It is known that LaMnO₃ easily forms defects to become LaMnO_3+δ, resulting in up to ∼5% Mn²⁺ during the heating/operando process. Figure 2c shows the oxygen K-edge XAS, where the spectra are a combination of the oxide catalyst and the gas phase oxygen species, both adsorbed and in the gas phase. Under the operando conditions of CO + O₂ and CO, the oxygen K-edges are completely dominated by the gas phase spectra that are given in Figure 2c. Next to the CO and the O₂ peaks, we also observe the CO₂ peaks at 300 °C, indicating that the CO oxidation reaction is running. The oxygen K-edge of LaMnO₃ contains oxygen p-character of the empty metal states, respectively, Mn 3d (528 to 531 eV), La 5d (532 to 539 eV), and Mn 4p (540 to 548 eV), as can also be shown with Density-Functional-Theory-based calculations.¹⁸

Figure 2.

Operando Mn 2p and O 1s XAS in the CO oxidation using LaMnO₃ catalyst. (a) The scheme of an operando reactor. (b) Operando Mn 2p XAS. (c) Operando O 1s XAS normalized from 0 to 1. The abbreviations rt and a.r. are room temperature and after reaction, respectively; the gray rectangle in the bottom panel indicates the La 5d–O 2p band. Labels ① to ⑦ indicate the temperature and gas conditions. If no gas conditions are indicated, the measurements are performed under helium. For clarity, we have also indicated each gas condition with the color of the line: black for helium, red for CO + O₂, and blue for CO. Details are given in the Experimental Methods Section.

The oxygen K-edges that have been measured in helium atmosphere also show peaks related to (adsorbed) O₂ possibly from air absorbed into the nanoparticle of the powder sample. The presence of the gas phase spectra make it difficult to determine the exact oxygen K edge spectra of the sample surface. In principle, one could subtract the spectra of O₂ and CO, but in practice, this creates too much uncertainty to reliably determine potential small changes. As far as we could determine, no reliable visible changes can be determined in the oxygen K-edge of the surface. Because of this, we focus on the metal spectra in the remainder. The oxygen K-edge spectra of all other samples are given in the Supporting Information.

4.3. Operando CO Oxidation of LaMn_0.5Co_0.5O₃

Figure 3 shows Co and Mn 2p XAS of the LaMn_0.5Co_0.5O₃ nanoparticles. The Co 2p XAS (Figure 3a) shows only minor changes during heating and under operando CO + O₂ catalytic conditions. Figure 3b shows crystal field multiplet calculations of Co²⁺. The 3d⁷ high-spin ⁴T₁ ground state is changing with temperature due to the increased occupation of the (3d spin–orbit split) ⁴T₁ state. Apart from this temperature-induced effect, the Co L_2,3-edge does not change, indicating that the Co sites remain high-spin ⁴T₁ under all conditions. Minor spectral variations can be due to small (averaged) symmetry distortions that can slightly change the spectral shape.

Figure 3.

Operando Co and Mn 2p XAS of LaMn_0.5Co_0.5O₃ in the CO oxidation. (a) Co 2p XAS under different temperatures with or without CO + O₂ (1 bar). (b) Multiplet ligand field simulations of Co²⁺ 2p XAS. (c) Mn 2p XAS under different temperatures with or without CO + O₂; the two bottom spectra are a Mn 2p XAS of LiMn₂O₄ reproduced with permission from ref (47), Copyright (1994, Elsevier), and a Mn⁴⁺ simulation. α indicates the position of the main peak of Mn²⁺, β indicates the first peak of Mn⁴⁺, and γ indicates the main peak of Mn³⁺. (d) Components of Mn²⁺, Mn³⁺, and Mn⁴⁺. Labels 1 to 11 indicate the temperature and gas conditions. If no gas conditions are indicated, the measurements are performed under helium. For clarity, we have also indicated each gas condition with the color of the line: black for helium, red for CO + O₂, and blue for CO. Details are given in the Experimental Methods Section.

Figure 3c shows the Mn 2p XAS spectra. The room temperature spectrum can be reproduced from an octahedral Mn⁴⁺ site. From r.t. to 345 °C and then cooling to 50 °C, Mn⁴⁺ gradually reduces to low oxidation state (Mn³⁺ and Mn²⁺). We performed linear combination fitting using Mn²⁺, Mn³⁺, and Mn⁴⁺ 2p XAS reference spectra to determine the components (Figure 3d). The details are given in the Supporting Information Figures S4–S5. Heating in helium slowly increases the amount of Mn²⁺ (conditions 1 to 4).The measurements in CO + O₂ up to 160 °C reoxidize the system to Mn⁴⁺. Operando measurements up to 212 °C create a combination of Mn³⁺ and Mn⁴⁺. Switching to pure CO + He reduces Mn again to partial Mn²⁺. In conclusion, we observe that under operando reaction conditions, Co remains 2+ and the amount of Mn⁴⁺ increases.

4.4. Operando CO Oxidation of LaMn_0.75Co_0.25O₃

Figure 4a shows the Co 2p XAS spectra of LaMn_0.75Co_0.25O₃, which can be interpreted as Co²⁺, similar to LaMn_0.5Co_0.5O₃. The detailed spectral shape is different from LaMn_0.5Co_0.5O₃ (Figure 3a) due to a symmetry reduction of the octahedral Co site to a tetragonal symmetry, which is likely caused by the dominance of the Mn³⁺ sites being Jahn–Teller ions. The tetragonal symmetry has been simulated in Figure 4b (more details in the Supporting Information Figure S6). The related parameters of D_s and D_t effectively cause a broadening of the spectral shape. Figure 4c shows the Mn 2p XAS, in which the spectral variations indicate a combination of Mn⁴⁺, Mn³⁺, and Mn²⁺ valences. The details of the fitting are given in the Supporting Information Figures S7–S8. The heating effects are sensitive to the Mn²⁺–Mn⁴⁺ components (Figure 4d). Operando CO + O₂ conditions at both 175 and 260 °C increase the Mn⁴⁺ and reduce the Mn²⁺ components.

Figure 4.

Operando Co and Mn 2p XAS of LaMn_0.75Co_0.25O₃ in the CO oxidation. (a) Co 2p XAS under different conditions: temperature and CO + O₂ gas. (b) The multiplet ligand field simulation of Co²⁺ 2p XAS by using a tetragonal symmetry distortion at 300 K. (c) Operando Mn 2p XAS under different conditions. (d) The occupations of Mn²⁺, Mn³⁺, and Mn⁴⁺ over the reaction process. Labels 1 to 10 indicate the temperature and gas conditions. If no gas conditions are indicated, the measurements are performed under helium. For clarity, we have also indicated each gas condition with the color of the line: black for helium, red for CO + O₂, and blue for CO. Details are given in the Experimental Methods Section.

4.5. Operando CO Oxidation of LaCoO₃

Figure 5a shows the operando Co L_2,3 XAS spectra of LaCoO₃. The labels “A”, “B”, and “C” indicate Co²⁺ features. Figure 5b shows the r.t. Co 2p XAS linear combination Co³⁺ high-spin/low-spin fitting. The treatment of LaCoO₃ was different from the other three samples in the sense that the sample was not cooled to room temperature after the initial heat treatment in helium. Because Co³⁺ is known to exist in both high-spin and low-spin in LaCoO₃ system, including a temperature-dependence in single crystal LaCoO₃,^20,48 we used pure low-spin Co³⁺ of LaCoO₃ (20 K)²⁰ and high-spin Co³⁺ of Sr₂CoO₃Cl⁴⁹ for the linear combination fitting at room temperature. Co 2p XAS shows a 34% high-spin Co³⁺ occupation, 66% low-spin Co³⁺, and no Co²⁺ component. We note that the LaCoO₃ nanoparticle is a bit different from single crystals as measured by Haverkort et al.²⁰ and Tomiyasu et al.⁴⁸ In order to quantitatively estimate the occupations of high-spin and low-spin and the presence of Co²⁺ above r.t., we focus on the L₂-edge and: (i) determine the spin ratio by linear combination from reference low-spin Co³⁺ of LaCoO₃ and high-spin Co³⁺ of Sr₂CoO₃Cl; (ii) assume that the same temperature has the same low-spin/high-spin ratio under operando conditions; (iii) the temperature-dependent high-spin Co²⁺ was used as the Co²⁺ 2p spectra reference taken from Figure 3a to minimize thermal effects. All L₂-edge fittings are given in Supporting Information Figures S10–S13, resulting in the numbers given in Figure 5c. Both operando 200 and 300 °C spectra show similar results for the presence of Co²⁺.

Figure 5.

(a) The operando Co 2p XAS of LaCoO₃ for CO oxidation. The reaction process is from bottom to top, with reference r.t. LaCoO₃ (gray) single crystal reproduced with permission from ref (20). Copyright (2006, American Physical Society). (b) r.t. Co 2p XAS fitting by pure low-spin Co³⁺ of LaCoO₃ (20 K) reproduced with permission from ref (20). Copyright (2006, American Physical Society) + high-spin Co³⁺ of Sr₂CoO₃Cl reproduced with permission from ref (49). Copyright (2009, American Physical Society). (c) The population of high-spin and low-spin Co³⁺ and Co²⁺ during the reaction process from panel (a). Labels 1 to 10 indicate the temperature and gas conditions. If no gas conditions are indicated, the measurements are performed under helium. For clarity, we have also indicated each gas condition with the color of the line: black for helium, red for CO + O₂, and blue for CO. Details are given in the Experimental Methods Section.

5. Discussion

Table 1 shows the main observations regarding the averages valences of Mn and Co, where (cond.6) indicates the spectrum number 6. The symmetry (O_h vs D_4h) and spin state (% HS) of the four measured LaMn_1–xCo_xO₃ catalysts for the CO oxidation reaction are provided. The activity, measured as the CO₂ production, is normalized to 1.0 for the highest activity of LaMn_0.5Co_0.5O₃. We observe CO₂ being formed from the oxygen K-edge at 300 °C (Figure 2c and Figure S15–S16) and the CO₂ products in Figure S18 (Supporting Information), which confirms the operando CO oxidation reaction.

Table 1. Main Observations of the Catalytic Activity, the Average Valences of ⟨Mn⟩ and ⟨Co⟩, and the Symmetry and the Spin State of Mn and Co Ions of LaMn_1–xCo_xO₃ Samples.

Sample	⟨Mn⟩ Helium	⟨Mn⟩ Operando	⟨Co⟩ Helium	⟨Co⟩ Operando	Activity (normalized)
LaMnO3	3.00	2.97	–	–	0.2
LaMn0.75Co0.25O3	3.43 (cond.6)	3.67 (cond.7)	2.0 (D4h)	2.0	0.3
LaMn0.5Co0.5O3	3.64 (cond.7)	3.92 (cond.9)	2.0 (Oh)	2.0	1.0
LaCoO3	–	–	3.0 (35% HS)	2.65 (63% HS)	0.2

In LaMnO₃, the valence state of Mn is 3+ under helium, and it does not change much under reaction conditions. Both LaMn_0.5Co_0.5O₃ and LaMn_0.75Co_0.25O₃ catalysts show mixtures of the Mn valences, where Mn⁴⁺ increases and Mn²⁺ decreases during the operando CO + O₂ oxidation, indicating that Mn⁴⁺ is the active site. In LaMn_0.75Co_0.25O₃, the average valence increases to 3.67, and in LaMn_0.5Co_0.5O₃, it increases to 3.92, indicating that a larger amount of Mn⁴⁺ increases the activity. Cobalt remains 2+ in LaMn_0.5Co_0.5O₃ and LaMn_0.75Co_0.25O₃. In LaMn_0.75Co_0.25O₃, the Co²⁺ site has tetragonal (D_4h) symmetry, likely induced by the dominance of the LaMnO₃ structure. The Co²⁺ site is octahedral in LaMn_0.5Co_0.5O₃, similar to the LaCoO₃ structure. In LaCoO₃, we observe fluctuations between (high-spin) Co²⁺ and mixed high-spin/low-spin Co³⁺. We describe the increase in Co²⁺ under reaction conditions to the compensation of the valence increase of Mn. Cobalt thus allows manganese to reach a higher average valence, which drastically increases the activity.

Unfortunately, analysis of the oxygen K-edge was hampered by the background of the gas phase signal. Future efforts to better separate the oxygen surface signal from the gas phase signal will improve this situation and drastically improve the options to also use the full strength of the oxygen K-edge spectra.

6. Conclusions

We have investigated operando soft X-ray Mn and Co 2p XAS as well as O 1s XAS spectra of LaMn_1–xCo_xO₃ nanoparticles for CO oxidation. Based on charge transfer multiplet calculations, we quantitatively identified the valence and spin states of Co and Mn. LaMn_0.5Co_0.5O₃ has the highest activity attributed to octahedral Co²⁺ combining with (almost pure) Mn⁴⁺. LaMn_0.75Co_0.25O₃ has the second highest activity, which is caused by the combination of its LaMnO₃ structure (shown by the tetragonal Co²⁺ sites) and its high average valence (under reaction conditions) of 3.67. LaMnO₃ and LaCoO₃ are much less active due to the absence of Mn⁴⁺ (or Co⁴⁺) sites. We have shown that operando soft X-ray XAS are very effective to study the electronic structure of the (near) surface states due to the combination of ∼4 nm probing depth and the sharp soft X-ray XAS spectra into the 3d states, which are very sensitive to details of the electronic structure.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acscatal.4c03259.

• Experimental procedure scheme, XRD, Co–Mn 2p XAS fitting, Co²⁺ 2p XAS calculation in D_4h distortion and HF value effects, Mn²⁺ 2p XAS simulation at 20–1000 K, operando O 1s XAS, XAS simulation parameters in charge transfer multiplets (PDF)

Author Contributions

Investigation–visualization: Q.C, M.G, and F.d.G. Synchrotron: Q.C, M.G, L.B, M.P, S.M, and P.T. Writing–editing: Q.C and F.d.G. Discussions: all (co)authors.

The authors declare no competing financial interest.