Surface Reconstruction of Decalcified CaMnO₃ Perovskites for Low‐Temperature NO Oxidation

Abstract

Effective oxide‐based catalysts with the potential to replace noble metal catalysts remain elusive for NO‐to‐NO₂ oxidation, which is essential for low‐temperature NO _x abatement. Herein, we report a time‐controlled selective decalcification strategy to reconstruct dense CaMnO₃ perovskite into Mn‐rich porous oxides for NO oxidation. Acid etching progressively removed Ca from the Ca—O—Mn framework, forming MnO₂‐like porous structures with well‐defined [MnO₆] units and Mn—O—Mn bridges. The optimized catalyst, denoted as CMO‐24 h, decreased the temperature required for 50% NO conversion from 378 to 181 °C and increased the maximum NO conversion from 54% to 92%. The promoted activity arose from an optimized Mn—O—Mn lattice‐oxygen microenvironment within the reconstructed porous framework. Mn enrichment and pore formation increased the accessibility of Mn‐rich surface domains, while the connected Mn—O—Mn network facilitated Mn redox cycling among Mn²⁺/Mn³⁺/Mn⁴⁺ at low temperatures. Meanwhile, enriched surface‐accessible lattice oxygen and an expanded surface/subsurface lattice‐oxygen reservoir provided reactive oxygen for the oxidation of adsorbed NO‐derived species. Temperature‐programed desorption and in situ diffuse reflectance infrared Fourier transform spectroscopy (with NO or NO + O₂ as probes) further revealed that CMO‐24 h enhanced NO _x adsorption/desorption and accelerated the evolution of nitrosyl/nitrite‐type species toward nitrate‐ and NO₂‐related species.

Time‐controlled decalcification reconstructs dense CaMnO₃ into Mn‐rich porous oxides with well‐defined [MnO₆] units and redox‐flexible Mn—O—Mn bridges, promoting intersite electron transfer, lattice‐oxygen accessibility, and rapid NO _x intermediate turnover for efficient low‐temperature NO oxidation.

1. Introduction

Catalytic oxidation of NO to NO₂ is important for improving the abatement efficiency of nitrogen oxides (NO _x ) at low temperatures from both stationary and mobile emission sources [1, 2]. Pt‐based diesel oxidation catalysts exhibit high activity for NO oxidation, but their practical application is constrained by cost, limited resource availability, thermal sintering, and susceptibility to poisoning [3, 4, 5]. Developing earth‐abundant non‐noble‐metal catalysts is therefore essential for resource‐efficient NO _x control. Among various transition‐metal oxides [6, 7, 8, 9, 10, 11, 12], Mn‐based perovskites (with a formula of ABO₃), such as LaMnO₃, SrMnO₃, and CaMnO₃, are promising for NO oxidation because their Mn—O frameworks can accommodate variable Mn oxidation states and oxygen non‐stoichiometry [13, 14]. However, pristine perovskites are usually obtained by high‐temperature calcination, which often produces dense particles with a low surface area and limited exposure of active Mn—O sites, thereby restricting their activities in gas–solid reactions [15]. Selective A‐site leaching offers an effective route to reconstruct such dense frameworks. Perovskite‐derived catalysts obtained by acid treatments have shown an increased surface area, MnO₂‐like phase formation, and enhanced NO oxidation activity [15, 16]. However, it remains a challenge to optimize the reconstruction degree because aggressive acid treatments may disrupt the parent framework after the formation of MnO₂‐like domains.

CaMnO₃ is particularly suitable for selective decalcification because its Ca‐containing A‐site sublattice can be preferentially leached under acidic conditions, while the remaining Mn—O framework can be reorganized into Mn‐rich oxide domains [14]. Selective Ca removal is expected to open the compact framework, enrich surface Mn species, and expose more accessible Mn—O surface sites [17, 18]. However, decalcification doesn’t proceed as a monotonic activation process. Inadequate A‐site dissolution may leave insufficient porosity and active Mn—O exposure, whereas over‐etching can compromise the reconstructed surface layer and thus weakens the favorable surface chemistry [15, 19]. Therefore, identifying an optimal reconstruction window that balances Ca leaching, Mn enrichment, pore generation, and structural integrity is crucial for designing CaMnO₃‐derived catalysts. Beyond this structural balance, the catalytic consequences of decalcification should be understood in terms of the reconstructed Mn—O oxygen microenvironment. Recent studies have shown that NO oxidation over Mn‐based oxides is governed not only by surface metal sites but also by lattice oxygen participation, Mn—O electronic structure, and NO _x intermediate evolution [13, 20]. Zheng et al. demonstrated cooperative lattice oxygen redox on SmMn₂O₅ mullite surfaces, where neighboring lattice oxygen atoms participate in nitrate formation, oxygen vacancy replenishment, and NO₂ release [21]. Exsolved Mn₂O₃/SmMn₂O₅ interfaces were further shown to activate lattice oxygen through enhanced Mn—O covalency and reversible Mn valence changes, thereby promoting NO₂ desorption via a cooperative Mars–van Krevelen mechanism [20]. These findings suggest that an effective Mn‐based NO oxidation catalyst should not only expose Mn sites but also construct a redox‐active Mn—O environment that enables a continuous oxygen supply, Mn redox cycling, and efficient NO _x intermediate transformation.

Herein, we report a time‐controlled selective decalcification strategy to reconstruct dense CaMnO₃ (CMO) perovskite into Mn‐rich porous oxides for low‐temperature NO oxidation. CMO was used as a structurally programable precursor, and the HNO₃ etching time was regulated to tune the reconstruction degree. The etching‐time‐dependent NO oxidation performance of the resulting CMO‐derived catalysts was subsequently evaluated and compared with the commercial γ‐MnO₂ as a reference. The morphology, elemental distribution, textural properties, phase evolution, nanoscale crystallinity, and local Mn—O lattice structure are examined by scanning electron microscopy coupled with energy‐dispersive X‐ray spectroscopy (SEM‐EDS), N₂ physisorption, powder X‐ray diffraction (XRD), Raman spectroscopy, and aberration‐corrected high‐resolution transmission electron microscopy (HRTEM), selected‐area electron diffraction (SAED), and fast Fourier transform (FFT) analyses. Surface chemical states and defect‐related electronic features are further probed by X‐ray photoelectron spectroscopy (XPS) and electron paramagnetic resonance (EPR), respectively. The reducibility, oxygen‐release behavior, NO _x adsorption/desorption properties, and surface NO _x intermediate evolution are investigated by hydrogen temperature‐programed reduction (H₂‐TPR), oxygen temperature‐programed desorption (O₂‐TPD), NO temperature‐programed desorption (NO‐TPD), and in situ diffuse reflectance infrared Fourier transform spectroscopy (in situ DRIFTS). This enables a systematic assessment of the relationships among selective Ca removal, Mn—O framework reconstruction, oxygen‐transfer behavior, NO _x intermediate turnover, and NO oxidation performance.

2. Results and Discussion

2.1. Optimization of CMO Decalcification for NO Oxidation Catalysis

The NO oxidation performance of pristine CMO and acid‐etched CMO‐derived catalysts was evaluated, and the results are shown in Figure 1a. Acid etching markedly shifts the NO conversion profiles toward lower temperatures, indicating a pronounced enhancement of low‐temperature activity. All catalysts exhibit a volcano‐shaped temperature dependence, with NO conversion increasing at low temperatures and decreasing at higher temperatures because of the thermodynamic limitation of NO₂ formation [22]. Neither intraparticle nor external mass‐transfer limitations were significant under the employed conditions (Figure S1). The enhanced light‐off behavior is reflected by the T ₅₀ values (see Figure 1b and Table S1), defined as the temperatures required to reach 50% NO conversion [23, 24]. The T ₅₀ decreases sharply from 378.2 °C for pristine CMO to 278.8, 225.0, and 180.6 °C for CMO‐6 h, CMO‐12 h, and CMO‐24 h, respectively. Further prolonging the etching time to 72 h slightly increases the T ₅₀ to 189.6 °C, indicating that excessive acid treatment does not further improve the low‐temperature light‐off performance. A similar trend is observed for both the maximum NO conversion and the average NO conversion in the 150–300 °C range (see Figure 1c and Table S1). The maximum NO conversion increases from 53.7% for pristine CMO to 70.0%, 82.2%, and 91.8% for CMO‐6 h, CMO‐12 h, and CMO‐24 h, respectively, before decreasing slightly to 89.2% for CMO‐72 h. The average NO conversion in the 150–300 °C range follows the same tendency, increasing from 10.3% for pristine CMO to 69.8% for CMO‐24 h and then decreasing to 65.8% for CMO‐72 h. Notably, CMO‐24 h also outperforms commercial γ‐MnO₂ under identical reaction conditions, highlighting the effectiveness of time‐controlled acid etching in promoting NO oxidation.

FIGURE 1.

NO oxidation performance of pristine CMO, acid‐etched CMO‐derived catalysts, and commercial γ‐MnO₂. (a) NO conversion as a function of reaction temperature; (b) T ₅₀ value as a function of etching time; (c) maximum NO conversion and average NO conversion in the 150–300 °C temperature range as functions of etching time; and (d) r _Mn as a function of etching time at 250 °C.

The etching‐time‐dependent activity therefore reveals a distinct optimal window rather than a monotonic enhancement. Insufficient etching provides limited activation of the parent CMO catalyst, whereas excessive etching leads to a slight deterioration in catalytic performance. This trend is also supported by the Mn‐normalized apparent NO oxidation rate (r _Mn; see Figure 1d; the calculation details are provided in Section S1) at 250 °C [25], which reaches the highest value for CMO‐24 h, showing the highest apparent Mn‐normalized activity under the standard reaction conditions. Additional measurements under differential conditions (X _NO < 15%) showed that CMO‐24 h exhibited markedly higher mass‐ and Mn‐normalized apparent rates than CMO, whereas its BET‐area‐normalized rate (r _BET) was lower (Tables S2 and S3), highlighting the important contribution of increased surface accessibility after decalcification. At comparable BET surface areas and Mn contents, CMO‐24 h still showed a slightly higher r _BET than CMO‐72 h, indicating an additional contribution from surface reactivity. The reconstructed samples also exhibited lower apparent activation energies than CMO (Figure S2). Overall, these catalytic results confirm that CMO‐24 h exhibits the most effective NO catalytic oxidation activity of all tested samples.

The feed tolerance of CMO‐24 h was further evaluated at 275 °C (see Figures S3 and S4). Introducing 3 vol% H₂O slightly and reversibly promoted NO oxidation, with stable performance maintained for 24 h, whereas 8 ppm SO₂ decreased the NO conversion from approximately 87% to 75% and caused further deactivation during H₂O/SO₂ co‐feeding. Post‐reaction characterization associated the persistent SO₂‐induced activity loss with sulfate accumulation and perturbation of the surface Mn–O redox environment (see Figure S5); our previous work further showed that trace Pd modification of acid‐treated CaMnO₃ suppresses sulfate accumulation and improves SO₂ resistance, providing a feasible route for further optimization [14].

2.2. Decalcification‐Induced Porous Mn—O—Mn Lattice Reconstruction

The structural reconstruction of CMO during acid etching treatment was investigated. As shown in Figure 2a,b, pristine CMO consists of compact microspheres assembled from densely packed primary particles. After acid etching, the spherical morphology is largely preserved, whereas the surface becomes progressively roughened and loosened. CMO‐6 h and CMO‐12 h show evident surface corrosion (see Figure S6), while CMO‐24 h develops a highly open sheet‐like porous texture (see Figure 2c,d), indicating effective opening and reorganization of the initially dense framework. Further extending the etching time to 72 h causes partial fragmentation and local collapse of the reconstructed architecture rather than further pore development (see Figure 2e,f). Elemental analyses confirm that this morphological reconstruction is accompanied by substantial Ca removal and Mn enrichment (see Figures S7–S11). The Ca content decreases from 24.0 wt% for pristine CMO to 0.9 wt% for CMO‐24 h, whereas the Mn content increases from 34.0 to 60.1 wt% (see Table S4). A similar surface trend is observed by XPS, with the surface Ca concentration decreasing from 20.3 to 0.3 at% and the Mn concentration increasing from 13.3 to 33.9 at% after 24 h of etching (see Table S4). Consistently, the survey XPS spectra show that the Ca 2s and Ca 2p signals nearly disappear for CMO‐24 h and CMO‐72 h, confirming the near‐complete depletion of surface Ca species after optimized acid etching (see Figure S12). N₂ adsorption–desorption measurements further support the formation of accessible porosity (see Figure S13 and Table S4). The BET surface area increases from 5.2 m²·g⁻¹ for pristine CMO to 238.4 m²·g⁻¹ for CMO‐24 h, accompanied by an increase in pore volume from 0.006 to 0.463 cm³·g⁻¹. As summarized in Figure S14, surface area, pore volume, and apparent Ca removal degree (calculation details are provided in Section S2) increase rapidly up to 24 h and then approach stagnation. For CMO‐72 h, only marginal additional Ca removal is achieved, whereas both surface area and pore volume slightly decrease, indicating that excessive etching mainly perturbs the reconstructed porous skeleton rather than generating additional accessible porosity.

FIGURE 2.

Representative SEM images of (a,b) pristine CMO, (c,d) CMO‐24 h, and (e,f) CMO‐72 h, (g) XRD patterns, and (h) Raman spectra of pristine CMO and acid‐etched CMO‐derived catalysts.

XRD was used to monitor the phase evolution after acid etching (see Figure 2g). Pristine CMO shows well‐defined diffraction peaks indexed to crystalline CaMnO₃ perovskite, with characteristic reflections at 2θ ≈ 23.8°, 33.9°, 48.8°, 60.8°, and 71.5°, corresponding to the (101), (121), (202), (321), and (400) planes, respectively [26, 27]. With increasing etching time, these CaMnO₃ reflections gradually weaken, indicating progressive disruption of the long‐range Ca—O—Mn perovskite framework. Meanwhile, new reflections assigned to γ‐MnO₂ become increasingly evident, including peaks at 2θ ≈36.8°, 42.1°, 55.3°, and 66.1°, corresponding to the (131), (300), (160), and (421) planes, respectively [28]. For CMO‐24 h, the CaMnO₃ reflections are largely attenuated, whereas the γ‐MnO₂‐related features become dominant, demonstrating the formation of MnO₂‐like crystalline domains. Further etching to 72 h does not generate additional crystalline phases, but the γ‐MnO₂‐like reflections become less defined, suggesting a partial loss of structural order under excessive etching. These results show that acid etching drives a time‐dependent transformation from dense CaMnO₃ perovskite to Mn‐rich γ‐MnO₂‐like oxides. Aberration‐corrected HRTEM, FFT, and SAED analyses further reveal locally ordered γ‐MnO₂‐like nanocrystalline domains in CMO‐24 h (see Figure S15).

Raman spectroscopy was further employed to probe the local Mn—O framework reconstruction (see Figure 2h). The band in the ca. 564–636 cm⁻¹ region can be assigned to the Mn—O stretching vibrations of [MnO₆] octahedral units in MnO₂‐like domains [29, 30]. Pristine CMO exhibits a broad Mn—O vibration band centered at 564 cm⁻¹, reflecting Mn—O units embedded in the compact Ca—O—Mn framework. Upon acid etching, this band becomes more pronounced and shifts toward the characteristic vibration region of MnO₂‐like [MnO₆] octahedral units, indicating progressive reconstruction toward a [MnO₆]‐dominated local structure. The Mn—O Raman shift reflects the vibrational frequency and force constant of Mn—O bonds, while the full width at half maximum (FWHM) is sensitive to the distribution and homogeneity of local Mn—O coordination environments [31, 32]. CMO‐24 h exhibits the highest Raman shift at 636 cm⁻¹ and the smallest FWHM among the etched samples (see Table S5), suggesting the most well‐defined [MnO₆] octahedral configuration with a relatively higher Mn—O bond force constant. In contrast, the redshift to 631 cm⁻¹ and peak broadening observed for CMO‐72 h indicate partial relaxation and increased heterogeneity of the reconstructed Mn—O framework. Moreover, low‐frequency bands observed at 252, 244, and 247 cm⁻¹ for CMO‐12 h, CMO‐24 h, and CMO‐72 h, respectively, can be assigned to Mn—O—Mn‐related lattice vibrations of linked [MnO₆] octahedra, mainly involving inter‐octahedral bending or deformation [33]. Based on the harmonic oscillator approximation, the lowest Mn—O—Mn shift of CMO‐24 h indicates the smallest effective bridge stiffness and thus the most compliant Mn—O—Mn network, which may favor intersite electronic communication and lattice‐oxygen migration during NO oxidation [34, 35]. The correlations between T ₅₀ and the Raman shift of Mn—O and Mn—O—Mn vibrations further support that CMO‐24 h possesses an optimized local lattice structure, featuring well‐defined [MnO₆] units connected by redox‐flexible Mn—O—Mn bridges (see Figure S16).

Overall, acid etching transforms dense CaMnO₃ into Mn‐rich porous oxides through selective Ca removal, framework opening, γ‐MnO₂‐like phase formation, and local Mn—O—Mn lattice reconstruction. Among the etched samples, CMO‐24 h achieves the optimal structural balance, featuring Mn enrichment, accessible porosity, well‐defined [MnO₆] units, and redox‐flexible Mn—O—Mn bridges. The increased accessibility of Mn‐rich domains, together with the optimized Mn—O—Mn lattice framework, provides the structural basis for enhanced lattice‐oxygen accessibility and improved low‐temperature NO oxidation activity.

2.3. Redox Behaviors of Surface Mn Species

The near‐surface Ca depletion and Mn‐rich lattice reconstruction established above are expected to modify the chemical environment of surface Mn and its redox behaviors. The Mn 2p spectra were therefore analyzed to examine the surface Mn valence states of pristine CMO and acid‐etched CMO‐derived catalysts (see Figure 3a). The Mn 2p _3/2 envelopes can be mainly deconvoluted into Mn²⁺, Mn³⁺, and Mn⁴⁺ components, together with a high binding energy (BE) satellite feature [36]. Pristine CMO already contains mixed‐valence Mn species, whereas acid etching markedly changes the surface Mn²⁺/Mn³⁺/Mn⁴⁺ distribution (see Table S6). The Mn³⁺ fraction decreases from 50.1% for CMO to 40.9% for CMO‐24 h and 37.8% for CMO‐72 h, accompanied by increased Mn⁴⁺ and low‐valence Mn²⁺‐related contributions. This evolution is consistent with valence reorganization during the reconstruction of the Ca–O–Mn framework into Mn‐rich oxide domains. In addition, the Mn³⁺‐ and Mn⁴⁺‐related components of CMO‐24 h and CMO‐72 h shift slightly toward lower BEs, indicating a modified electronic environment around Mn sites after deep decalcification [37]. Together with the Raman evidence for Mn—O—Mn bridges linking [MnO₆] units, this negative shift suggests enhanced intersite electronic communication within the reconstructed Mn—O framework.

FIGURE 3.

Surface chemical states of pristine CMO and acid‐etched CMO‐derived catalysts revealed by XPS: (a) Mn 2p, (b) Mn 3s, (c) EPR spectra, and (d) H₂‐TPR profiles of pristine CMO and acid‐etched CMO‐derived catalysts.

To further describe the distribution of surface Mn valence states, the Mn valence‐state distribution entropy [38] was calculated from the relative Mn²⁺, Mn³⁺, and Mn⁴⁺ fractions (see Table S6; the calculation details are provided in Section S3). CMO‐24 h exhibits an entropy value close to unity, corresponding to the most even Mn²⁺/Mn³⁺/Mn⁴⁺ distribution among the tested samples. This feature is consistent with the lowest Mn—O—Mn Raman shift of CMO‐24 h, suggesting that its redox‐flexible inter‐octahedral bridge network favors Mn valence redistribution and electronic communication within the reconstructed Mn–O framework.

The average Mn oxidation state was further evaluated by Mn 3s exchange splitting, which is commonly used to estimate the average oxidation state (AOS) of Mn (see Figure 3b) [39]. The splitting energy, ΔE _3s , increases from 4.62 eV for pristine CMO to 4.84, 4.95, and 5.17 eV for CMO‐6 h, CMO‐12 h, and CMO‐24 h, respectively, followed by a slight decrease to 5.06 eV for CMO‐72 h. According to the empirical relationship AOS = 8.956–1.126ΔE _3s  [39], the calculated Mn AOS decreases from 3.75 for CMO to 3.51, 3.38, and 3.13 for CMO‐6 h, CMO‐12 h, and CMO‐24 h, respectively, and then slightly increases to 3.26 for CMO‐72 h (see Table S6). The positive correlation between Mn 3s‐derived AOS and T ₅₀ indicates that a lower average Mn valence in this series is closely associated with improved low‐temperature light‐off behavior (see Figure S17). Combined with the Raman evidence for Mn—O—Mn bridge formation, the lowered AOS of CMO‐24 h suggests enhanced intersite charge redistribution among connected [MnO₆] units, forming a redox‐flexible Mn—O environment that is more relevant to NO oxidation than a high average Mn oxidation state alone.

EPR spectroscopy was further employed to probe Mn‐associated electronic states and magnetic interactions (see Figure 3c). Pristine CMO exhibits a broad resonance centered at g ≈ 1.99, which can be associated with exchange‐interacting Mn species in the compact Ca—O—Mn framework [40, 41]. This broad signal weakens after acid etching, consistent with disruption of the original long‐range CMO perovskite lattice framework. In contrast, CMO‐24 h exhibits a well‐resolved six‐line hyperfine signal centered at g ≈ 2.01, which is characteristic of isolated or weakly interacting Mn centers with hyperfine coupling to the ⁵⁵Mn nucleus (I = 5/2) [42]. Combined with the decreased Mn 3s‐derived AOS, this EPR feature supports the formation of defect‐associated low‐valence Mn centers within the reconstructed Mn—O framework [43]. These results suggest that selective decalcification weakens the compact long‐range Mn coupling of pristine CMO, while generating locally redox‐active Mn sites and a more flexible Mn–O network that can facilitate redox cycling during NO oxidation. Low‐temperature EPR at 100 K further resolves a g = 2.003 signal that is markedly stronger for CMO‐24 h than for CMO, consistent with an increased abundance of oxygen‐vacancy‐related paramagnetic centers after selective decalcification (Figure S18).

The redox behavior of the Mn—O framework was further examined by H₂‐TPR (Figure 3d). All samples exhibit two main Mn‐related reduction events. The low‐temperature peak α is generally associated with the initial reduction of surface or near‐surface Mn—O species, including MnO₂‐like Mn⁴⁺ → Mn³⁺ reduction, whereas the high‐temperature peak β corresponds to deeper reduction of MnO _x species toward Mn²⁺‐containing phases [44]. Pristine CMO exhibits peak α and peak β at 389 and 508 °C, respectively, indicating that Mn—O species embedded in the compact Ca—O—Mn framework require higher thermal energy for reduction compared with the etched samples. After acid etching, both reduction peaks shift markedly to lower temperatures. The peak α temperature (T _α ) decreases to 310, 290, and 251 °C for CMO‐6 h, CMO‐12 h, and CMO‐24 h, respectively, and remains nearly unchanged for CMO‐72 h at 254 °C. Meanwhile, peak β shifts from 508 °C for pristine CMO to 374 °C for CMO‐24 h. These pronounced downward shifts demonstrate that decalcification‐induced Mn—O reconstruction substantially enhances low‐temperature reducibility and facilitates Mn redox cycling [15, 45]. The strong correlation between T _α and T ₅₀ further confirms the close relationship between low‐temperature reducibility and NO oxidation light‐off behavior (Figure S19).

Overall, the Mn 2p, Mn 3s, EPR, and H₂‐TPR results demonstrate that selective decalcification reconstructs the surface Mn chemical environment from a compact perovskite‐like Mn–O framework into a more redox‐responsive Mn‐rich oxide network. Among the etched catalysts, CMO‐24 h combines an even Mn²⁺/Mn³⁺/Mn⁴⁺ distribution, a low average Mn oxidation state, defect‐associated Mn centers, and markedly improved low‐temperature reducibility. These features provide the Mn‐site basis for efficient redox cycling during low‐temperature NO oxidation.

2.4. Surface Oxygen Species and Accessible Lattice‐Oxygen Reservoir

The reconstructed Mn—O framework is also expected to affect the surface oxygen environment and the availability of lattice oxygen. The O 1s spectra were therefore analyzed to distinguish surface oxygen species on pristine CMO and acid‐etched CMO‐derived catalysts (see Figure 4a). The components at 529.4–529.8, 530.8–531.3, and 532–533 eV are assigned to surface lattice oxygen (O_latt), adsorbed/defect‐related oxygen species (O_ads), and adsorbed H₂O or weakly bound oxygenated species, respectively [46]. After acid etching, both O_latt and O_ads shift slightly toward higher BEs, indicating that the local oxygen environment and Mn—O electronic polarization are modified by framework reconstruction [20, 46, 47]. More importantly, the relative abundance of surface lattice oxygen changes systematically with etching time. The O_latt/O_ads ratio increases from 2.25 for pristine CMO to 2.44, 2.64, and 2.94 for CMO‐6 h, CMO‐12 h, and CMO‐24 h, respectively, and then slightly decreases to 2.83 for CMO‐72 h (see Figure 4a). Considering the concurrent Mn enrichment, porous reconstruction, and formation of well‐defined [MnO₆] units linked by redox‐flexible Mn—O—Mn bridges, this trend suggests that CMO‐24 h exposes the largest relative population of surface‐accessible Mn—O lattice oxygen sites. The inverse correlation between O_latt/O_ads and T ₅₀ further supports the important role of surface‐accessible lattice oxygen in low‐temperature NO oxidation (see Figure 4b).

FIGURE 4.

(a) XPS‐O 1s spectra of pristine CMO and acid‐etched CMO‐derived catalysts and (b) the relationship between O_latt/O_ads ratio and T ₅₀; (c) O₂‐TPD profiles of pristine CMO and CMO‐24 h and (d) the integrated area of Peak a and Peak b.

O₂‐TPD was further used to evaluate the oxygen‐release behavior of pristine CMO and the optimized CMO‐24 h (see Figure 4c). For Mn‐based oxides, oxygen desorption below 400 °C can be assigned to weakly bound surface‐adsorbed or activated oxygen species, whereas oxygen release in the 400–700 °C region can be associated with more accessible surface/subsurface lattice oxygen; oxygen release above 700 °C can be attributed to strongly bound bulk lattice oxygen [48, 49]. The low‐temperature activated‐oxygen peak appears at nearly identical temperature for CMO and CMO‐24 h (376.6 vs. 376.2 °C), but its integrated area increases from 64.1 to 774.1 a.u. after selective decalcification (see Figure 4d). This result indicates a substantial increase in thermally releasable surface oxygen species on the reconstructed Mn‐rich porous oxide. Moreover, the lattice‐oxygen release region is more strongly affected by selective decalcification. Pristine CMO shows a surface/subsurface lattice‐oxygen peak at 584.1 °C with an area of 123.6 a.u., together with a strongly bound bulk‐lattice‐oxygen feature at 774.6 °C. In contrast, CMO‐24 h exhibits an intense surface/subsurface lattice‐oxygen peak at 627.8 °C, while the high‐temperature lattice‐oxygen feature shifts markedly from 774.6 to 687.1 °C, moving from the strongly bound bulk‐oxygen region into the more accessible lattice‐oxygen region (see Figure 4c). Consequently, the quantified lattice‐oxygen area in the 400–700 °C region increases to 6454.2 a.u. for CMO‐24 h, approximately 52.2 times that of pristine CMO (see Figure 4d). These results demonstrate that selective decalcification not only increases the abundance of labile surface oxygen but, more importantly, exposes and activates a much larger reservoir of surface/subsurface lattice oxygen.

Taken together, the O 1s XPS and O₂‐TPD results reveal that CMO‐24 h possesses a more accessible and thermally responsive oxygen environment than that of pristine CMO. The reconstructed Mn—O lattice provides enriched surface‐accessible lattice oxygen, more abundant labile surface oxygen species, and an expanded surface/subsurface lattice‐oxygen reservoir. These oxygen‐related features, coupled with the redox‐responsive Mn species discussed above, establish a favorable Mn–O oxygen microenvironment for lattice‐oxygen‐assisted NO oxidation.

2.5. NO Adsorption and Surface NO _x Intermediate Evolution

NO‐TPD was performed after NO/N₂ adsorption to evaluate the adsorption and activation behavior of NO‐derived species over pristine CMO and CMO‐24 h. In general, low‐temperature NO desorption below 150 °C is associated with physically adsorbed or weakly bound NO _x species, whereas medium‐temperature desorption at 150–250 °C and high‐temperature desorption above 250 °C can be assigned to chemisorbed NO _x species and the decomposition of more strongly bound nitrite/nitrate species [50]. As shown in Figure 5a, pristine CMO exhibits weak NO desorption features at 131, 204, and 348 °C, with a total NO desorption amount of 4.33 μmol·g⁻¹ (see Figure S20 and Table S7). In contrast, CMO‐24 h shows much stronger NO desorption peaks at 109, 180, and 293 °C, and the total NO desorption amount increases markedly to 27.18 μmol·g⁻¹. The enhanced NO release indicates that the reconstructed porous Mn—O—Mn framework provides more accessible sites for NO adsorption. Meanwhile, the shift of the main NO desorption features toward lower temperatures suggests that adsorbed NO _x species become more labile on the reconstructed Mn—O surface, facilitating their desorption or subsequent transformation during NO oxidation [20, 36].

FIGURE 5.

Temperature‐programed desorption of NO _x species over pristine CMO and CMO‐24 h. (a) NO and (b) NO₂ desorption profiles after NO/N₂ adsorption.

The simultaneously monitored NO₂ desorption signal further reveals the oxidative activation of adsorbed NO species (see Figure 5b). Pristine CMO shows only a weak NO₂ release peak at 97 °C, with a total NO₂ desorption amount of 0.27 μmol·g⁻¹. By contrast, CMO‐24 h exhibits a much stronger low‐temperature NO₂ peak at 98 °C and an additional broad feature centered at 280 °C, increasing the total NO₂ desorption amount to 2.86 μmol·g⁻¹ (see Figure S20 and Table S7). Since the adsorption step was conducted under NO/N₂, the enhanced NO₂ release over CMO‐24 h indicates that adsorbed NO‐derived species can be more effectively oxidized by surface/lattice oxygen during temperature‐programed desorption. This behavior is consistent with the expanded surface/subsurface lattice‐oxygen reservoir and improved Mn—O redox properties discussed above.

The surface NO _x intermediates were further identified by in situ DRIFTS during sequential NO adsorption and oxidation at 150 °C, with the main IR band assignments summarized in Table S8. Under NO + N₂, CMO‐24 h rapidly develops NO _x ‐related bands at 1680, 1630, 1590, 1543, 1480, 1288, and 1025 cm⁻¹ (see Figure 6a) [15, 36, 51, 52, 53]. The nitrate‐related bands at 1543, 1480, and especially 1288 cm⁻¹ indicate that part of the adsorbed NO can be oxidized on CMO‐24 h even in the absence of gaseous O₂. This behavior suggests the participation of pre‐existing surface‐accessible lattice oxygen associated with the reconstructed Mn—O domains in the initial NO oxidation step, in agreement with the low‐temperature NO₂ release observed in NO‐TPD. By contrast, pristine CMO shows weaker nitrate‐related bands but more pronounced nitrosyl‐, nitrite‐, and nitro‐type intermediates under the same NO + N₂ atmosphere (see Figure S21a), with the 1335 cm⁻¹ band serving as a representative nitrite/nitro‐related signal [36]. This comparison indicates that NO adsorption on pristine CMO is less efficiently coupled with lattice‐oxygen participation, leading to the accumulation of lower‐oxidation‐state NO _x intermediates rather than their rapid conversion into nitrate‐type species.

FIGURE 6.

In situ DRIFTS analysis of surface NO _x evolution during NO oxidation. DRIFTS spectra of CMO‐24 h at 150 °C under sequential (a) NO + N₂, (b) NO + O₂ + N₂, and (c) O₂+ N₂ atmospheres; (d) time‐dependent evolution of representative NO _x ‐related bands at 1288 cm⁻¹ over CMO‐24 h and 1335 cm⁻¹ over pristine CMO during the same gas‐switching sequence.

Upon switching to NO + O₂ + N₂, the nitrate‐related bands over CMO‐24 h are further intensified (see Figure 6b), particularly the band at 1288 cm⁻¹, together with those at 1543 and 1480 cm⁻¹. Their time‐dependent growth is further confirmed in Figure S22. This nitrate evolution indicates that gaseous O₂ promotes the further oxidation and turnover of adsorbed NO _x intermediates, most likely by replenishing oxygen vacancies and restoring reactive lattice oxygen on the reconstructed Mn—O surface [20, 21]. This behavior is consistent with the improved low‐temperature redox properties evidenced by H₂‐TPR (see Figure 3d) and the enlarged labile lattice‐oxygen reservoir revealed by O₂‐TPD (see Figure 4c). In contrast, pristine CMO exhibits a more dispersed spectral evolution after O₂ introduction (see Figure S21b), with nitro/nitrite‐, NO₂ ⁻‐, and NO⁻‐related bands coexisting rather than evolving predominantly toward nitrate‐type species. Although O₂ can oxidize part of the adsorbed NO _x species on pristine CMO, the transformation among surface intermediates is slower and less directed, consistent with its weaker NO₂ release (see Figure 5) and lower NO oxidation activity (see Figure 1).

The subsequent O₂+ N₂ purge provides further insight into the stability and evolution of accumulated surface NO _x species (see Figures 6c and S21c). For CMO‐24 h, the nitrate‐related band at 1288 cm⁻¹ gradually decreases after NO is removed from the feed, indicating that nitrate‐type species formed on the reconstructed Mn—O surface can be further transformed or removed under oxidative conditions rather than remaining as inert spectator deposits. Such dynamic NO _x intermediate evolution helps maintain accessible Mn—O sites during continuous NO oxidation. By contrast, pristine CMO retains multiple residual NO _x bands during the O₂+ N₂ purge, suggesting slower conversion and removal of nitrite/nitro‐ and nitrate‐related intermediates on the dense CMO surface. These persistent intermediates may partially block active Mn—O sites and suppress continuous NO oxidation.

Time‐dependent tracking of representative IR bands further highlights the different surface reaction behaviors of pristine CMO and CMO‐24 h (see Figure 6d). Over CMO‐24 h, the nitrate‐type band at 1288 cm⁻¹ increases rapidly during NO + N₂ exposure, continues to grow after O₂ introduction, and then gradually decreases during the O₂+ N₂ purge, reflecting efficient NO capture, lattice‐oxygen‐assisted nitrate formation, and subsequent removal of NO _x intermediates. In contrast, the 1335 cm⁻¹ band over pristine CMO, assigned to nitrite/nitro‐related intermediates, shows a slower and less responsive evolution. These results reveal a shift in the dominant surface pathway from nitrite/nitro accumulation on pristine CMO to more directed nitrate‐type intermediate formation and turnover on CMO‐24 h. This pathway shift is consistent with the much higher NO₂ release observed in NO‐TPD and accounts for the superior low‐temperature NO oxidation activity of CMO‐24 h.

2.6. Proposed Mechanism for Decalcification‐Promoted NO Oxidation

Based on the above results, a decalcification‐promoted NO oxidation mechanism is proposed in Figure 7. For pristine CMO, the dense Ca—O—Mn perovskite framework restricts the exposure of accessible Mn—O sites and limits low‐temperature oxygen transfer. NO adsorption is therefore weakly coupled with lattice‐oxygen participation, leading to the accumulation of nitrosyl‐, nitrite‐, and nitro‐type intermediates. Their further oxidation into nitrate‐ and NO₂‐related species is restricted, resulting in sluggish NO _x transformation and limited NO oxidation activity.

FIGURE 7.

Proposed NO oxidation mechanism over pristine CMO and CMO‐24 h.

Acid etching fundamentally changes the surface reaction environment. Selective Ca removal opens the compact Ca—O—Mn framework and induces Mn‐rich porous reconstruction. At the optimized etching time of 24 h, CMO‐24 h develops well‐defined [MnO₆] units connected by redox‐flexible Mn—O—Mn bridges. These bridges link neighboring Mn centers and provide pathways for intersite electron transfer and charge redistribution, as supported by the redistributed Mn²⁺/Mn³⁺/Mn⁴⁺ valence population, the highest Mn valence‐state distribution entropy, and the lowest Mn 3s‐derived average Mn oxidation state. The correlation between Mn AOS and T ₅₀ further indicates that low‐temperature NO oxidation is favored by a redox‐flexible mixed‐valence Mn–O environment. This electronic structure is coupled with enhanced lattice‐oxygen availability and mobility. CMO‐24 h exhibits the highest O_latt/O_ads ratio, improved low‐temperature reducibility, and a substantially expanded surface/subsurface lattice‐oxygen reservoir. These features enable adsorbed NO to be efficiently oxidized by surface‐accessible lattice oxygen, while gaseous O₂ replenishes consumed oxygen species and sustains Mn—O redox cycling.

Accordingly, the dominant surface pathway shifts from nitrite/nitro accumulation on pristine CMO to more directed nitrate‐type intermediate formation and turnover on CMO‐24 h. NO‐TPD confirms stronger NO uptake and higher NO₂ release over CMO‐24 h, while in situ DRIFTS shows rapid nitrate formation, O₂‐promoted intensification, and subsequent removal of nitrate‐related species. Excessive etching to 72 h partially disturbs the reconstructed porous framework and weakens the balance among Mn—O—Mn connectivity, lattice‐oxygen accessibility, and redox mobility. Therefore, the superior activity of CMO‐24 h originates from the coupled optimization of reconstructed Mn‐rich Mn—O domains with redox‐flexible Mn centers, accessible lattice oxygen, and oxygen‐defect sites, together with dynamic NO _x intermediate turnover, enabling efficient and sustained low‐temperature NO‐to‐NO₂ conversion.

3. Conclusions

In this work, we developed a time‐controlled selective decalcification strategy to reconstruct dense CaMnO₃ perovskite into Mn‐rich porous oxides for low‐temperature NO oxidation. Acid etching progressively removed Ca from the Ca—O—Mn framework and induced the formation of porous γ‐MnO₂‐like structures. The optimized CMO‐24 h catalyst exhibited the best catalytic performance, lowering T ₅₀ from 378.2 to 180.6 °C and increasing the maximum NO conversion from 53.7% to 91.8%. Moreover, 3 vol% H₂O slightly promoted NO oxidation, with stable activity maintained over 24 h. The superior activity of CMO‐24 h is closely associated with increased accessibility of reconstructed Mn‐rich Mn—O domains and an optimized Mn—O—Mn lattice microenvironment. Structural and spectroscopic results show that CMO‐24 h possesses accessible porosity, well‐defined [MnO₆] units, redox‐flexible Mn—O—Mn bridges, a highly even Mn²⁺/Mn³⁺/Mn⁴⁺ valence distribution, and the lowest Mn 3s‐derived average oxidation state. These features indicate efficient charge redistribution and enhanced Mn redox flexibility within the reconstructed Mn—O framework. Meanwhile, the highest O_latt/O_ads ratio, enhanced low‐temperature reducibility, and expanded surface/subsurface lattice‐oxygen reservoir confirm improved lattice‐oxygen accessibility and mobility.

NO‐TPD and in situ DRIFTS further reveal that CMO‐24 h promotes NO adsorption, lattice‐oxygen‐assisted oxidation, nitrate‐type intermediate formation, and dynamic turnover of NO _x intermediates, whereas pristine CMO tends to accumulate less reactive nitrosyl‐, nitrite‐, and nitro‐type species. These findings establish a mechanistic link among selective Ca removal, Mn‐rich Mn—O—Mn reconstruction, mixed‐valence redox flexibility, lattice‐oxygen activation, and low‐temperature NO‐to‐NO₂ conversion. This work highlights controlled A‐site removal as an effective route to transform dense perovskite precursors into efficient, earth‐abundant catalysts for sustainable NO _x abatement.

4. Materials and Methods

4.1. Catalyst Synthesis

CaMnO₃ perovskite was synthesized by a two‐step co‐precipitation method. In a typical synthesis, stoichiometric amounts of CaCO₃ (0.1 mol) and MnCO₃ (0.1 mol) were dissolved in 300 mL of HNO₃ solution (1.5 mol·L⁻¹). Subsequently, 500 mL of aqueous (NH₄)₂CO₃ solution (1.0 mol·L⁻¹) was added dropwise to the acidic solution under continuous stirring using a peristaltic pump. The molar ratio of total metal ions, namely, Ca²⁺ + Mn²⁺, to CO₃ ²⁻ was maintained at 1:2.5. After stirring for 30 min, the resulting suspension was centrifuged, washed thoroughly with deionized water and ethanol, and dried under vacuum at 80 °C overnight. The dried precursor was then calcined in air at 900 °C for 5 h with a heating rate of 5 °C·min⁻¹ to obtain pristine CaMnO₃, denoted as CMO.

For selective A‐site etching, 1.0 g of CMO was dispersed in 50 mL of HNO₃ solution (3.0 mol·L⁻¹) and stirred at room temperature (25 °C) for 6, 12, 24, or 72 h. The etched solids were collected by filtration, washed with deionized water until the filtrate reached approximately neutral pH, and dried at 80 °C for 20 h. The obtained samples were denoted as CMO‐6 h, CMO‐12 h, CMO‐24 h, and CMO‐72 h, respectively. Commercial γ‐MnO₂ purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. was used as a reference catalyst.

4.2. Catalyst Characterization

The morphology, phase structure, and textural properties of the catalysts were characterized by SEM equipped with EDS, N₂ physisorption, XRD, and Raman spectroscopy, respectively. Surface elemental compositions and chemical states were examined by XPS, while defect‐related electronic states were probed by EPR. The reducibility, oxygen‐release behavior, and NO _x adsorption/desorption properties were evaluated by H₂‐TPR, O₂‐TPD, and NO‐TPD, respectively. The nanoscale structure and local crystallinity of CMO‐24 h were further examined by aberration‐corrected HRTEM, SAED, and FFT analysis. In situ DRIFTS was performed to investigate NO adsorption, NO oxidation, and the evolution of surface NO _x intermediates. Detailed characterization procedures and data‐processing methods are provided in Section S4 of the Supporting Information.

4.3. Catalytic Activity

NO oxidation activity was evaluated in a fixed‐bed quartz tubular microreactor with an inner diameter of 5.0 mm. For each test, 100 mg of catalyst granules with a particle size of 40–60 mesh was thoroughly mixed with 400 mg of quartz sand and loaded into the reactor. The feed gas contained 500 ppm NO and 10 vol% O₂, with N₂ as the balance gas. The total flow rate was 400 mL·min⁻¹, corresponding to a weight hourly space velocity (WHSV) of 240,000 mL·g⁻¹·h⁻¹ based on the catalyst mass. The reaction temperature was increased from 150 to 450 °C. The outlet NO concentration was continuously monitored using a chemiluminescence NO analyzer (Model 42i‐HL, Thermo Fisher Scientific, USA). At each temperature, the reactor was held for ca. 30 min until a steady outlet NO concentration was reached, thereby minimizing the contribution of transient adsorption to the calculated NO conversion. For the feed‐tolerance tests, CMO‐24 h was evaluated at 275 °C under the same base feed, with 3 vol% H₂O and/or 8 ppm SO₂ introduced as indicated. A separate 24 h stability test was conducted under continuous 3 vol% H₂O co‐feeding. The NO conversion was calculated according to Equation (1), where C _in and C _out are the inlet and outlet NO concentrations, respectively.

$$X_{\mathrm{NO}}=\frac{C_{\mathrm{in}}-C_{\mathrm{out}}}{C_{\mathrm{in}}}\times 100\mathrm{\%}$$ (1)

Funding

This study was supported by the National Key Research and Development Program of China (2024YFC3908700), National Natural Science Foundation of China (22376068), Guangdong Provincial Department of Science and Technology (2025A0505020008, 2024A1515013097) and Shenzhen Science and Technology Program (ZDCYKCX20250901093901003). We thank Prof. Daiqi Ye for fruitful discussions.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Experimental details for catalyst characterization and spectroscopic measurements; differential kinetic, normalized‐rate, Arrhenius, and mass‐transfer sensitivity data; H₂O/SO₂ tolerance and post‐reaction characterization; calculation procedures for the apparent Mn‐normalized NO oxidation rate, apparent Ca removal degree, and Mn valence‐state distribution entropy; SEM, HRTEM, FFT, SAED, XPS, N₂ physisorption, Raman, low‐temperature EPR, H₂‐TPR, NO‐TPD, and in situ DRIFTS data; time‐resolved DRIFTS band‐evolution profiles; catalytic, elemental, textural, Raman, XPS, and NO‐TPD parameters; and IR band assignments.

Zhou Yongxin, Zheng Yuwei, Liu Huibin, Xu Xin, Ji Xiang, Xia Bolin, Xiao Hailin, Chen Peirong, Surface Reconstruction of Decalcified CaMnO₃ Perovskites for Low‐Temperature NO Oxidation, ChemSusChem 2026, 19, e71058. 10.1002/cssc.71058

Data Availability Statement

The data that support the findings of this study are available in the supporting information of this article.