Comparative Study of α- and β-MnO₂ on Methyl Mercaptan Decomposition: The Role of Oxygen Vacancies

Abstract

As a representative sulfur-containing volatile organic compounds (S-VOCs), CH₃SH has attracted widespread attention due to its adverse environmental and health risks. The performance of Mn-based catalysts and the effect of their crystal structure on the CH₃SH catalytic reaction have yet to be systematically investigated. In this paper, two different crystalline phases of tunneled MnO₂ (α-MnO₂ and β-MnO₂) with the similar nanorod morphology were used to remove CH₃SH, and their physicochemical properties were comprehensively studied using high-resolution transmission electron microscope (HRTEM) and electron paramagnetic resonance (EPR), H₂-TPR, O₂-TPD, Raman, and X-ray photoelectron spectroscopy (XPS) analysis. For the first time, we report that the specific reaction rate for α-MnO₂ (0.029 mol g⁻¹ h⁻¹) was approximately 4.1 times higher than that of β-MnO₂ (0.007 mol g⁻¹ h⁻¹). The as-synthesized α-MnO₂ exhibited higher CH₃SH catalytic activity towards CH₃SH than that of β-MnO₂, which can be ascribed to the additional oxygen vacancies, stronger surface oxygen migration ability, and better redox properties from α-MnO₂. The oxygen vacancies on the catalyst surface provided the main active sites for the chemisorption of CH₃SH, and the subsequent electron transfer led to the decomposition of CH₃SH. The lattice oxygen on catalysts could be released during the reaction and thus participated in the further oxidation of sulfur-containing species. CH₃SSCH₃, S⁰, SO₃²⁻, and SO₄²⁻ were identified as the main products of CH₃SH conversion. This work offers a new understanding of the interface interaction mechanism between Mn-based catalysts and S-VOCs.

1. Introduction

As a particular class of volatile organic compounds (VOCs), sulfur-containing volatile organic compounds (S-VOCs) can be converted into sulfate aerosols in the atmosphere through complex physicochemical reactions [1]. They can also react indirectly with NO_x through photochemistry reactions, which are the crucial precursors for forming PM _2.5 and O₃. Methyl mercaptan (CH₃SH), a representative S-VOC, is considered as an important air odor pollutant, which is harmful to the ecosystem and human health owing to its severe toxicity and low olfactory threshold [2,3,4]. In previous studies, various methods have been employed to eliminate CH₃SH, such as adsorption [5,6], biodegradation [7], photocatalytic oxidation [8], and catalytic oxidation [9,10]. However, these remediation technologies suffer from secondary pollution because of incomplete removal and high cost. Until now, catalytic decomposition has been regarded as the most promising strategy for removing S-VOCs due to its high purification efficiency, energy-saving nature, lack of additional additives (O₂, H₂, O_3, etc.), and less secondary pollution [11,12].

Metal-based catalysts are widely used for the removal of VOCs on account of their superior catalytic performance. Among the various transition metal oxides, MnO₂ is of great interest because of its low cost, low toxicity, environmental friendliness, and wide natural distribution [13,14]. Meanwhile, MnO₂ has been extensively studied in heterogeneous catalysis due to its unique physicochemical properties (i.e., multivalent, reactive oxygen species, and polycrystalline nature) and is recognized as one of the most active catalysts for VOCs removal among transition metal oxides [15,16,17]. Nanostructured MnO₂ possesses a rich structural flexibility, which adopts various crystallographic forms such as α-, β-, δ-, and γ-MnO₂, depending upon the size of the tunnel [18,19]. These polymorphs of MnO₂ include a one-dimensional chain-like tunnel (α-, β-, and γ-MnO₂) and two-dimensional layer structures (δ-MnO₂) based on different linkage ways of the basic octahedral molecular sieves [20,21].

Among the MnO₂ polymorphs, α-MnO₂ has one of the largest tunnel sizes (4.6 Å) consisting of 1D (1 × 1) and (2 × 2) channels, composed of double chains of edge-sharing [MnO₆] octahedra, filled with alkali metal cations, NH₄⁺ or H₃O⁺, inside the 2 × 2 tunnels to stabilize the structure [16]. The pyrolusite-type β-MnO₂ has a rutile-type structure with infinite [MnO₆] octahedral chains that share opposing edges. Each chain is connected to four similar chain corners, forming the smallest tunnel structure (2.3 Å) of MnO₂ polymorphs, consisting of 1D (1 × 1) and (1 × 1) channels [22,23]. It is generally accepted that catalysts with different crystal structures exhibit different catalytic efficiencies and reaction mechanisms for the reactants. For instance, Hayashi et al. evaluated the aerobic catalytic performance of six types of MnO₂ (α-, β-, γ-, δ, λ, ε -phases) for the transformation of 5-hydroxymethylfurfural to 2,5-furandicarboxylic acid and concluded the best activity of β-MnO₂ [19]. Chen et al. found that the α- and γ-MnO₂ presented higher benzene oxidation activity than β- and δ-MnO₂, whereas δ-MnO₂ displayed the best in formaldehyde oxidation among all of the MnO₂ materials [20]. This may be due to the varying oxygen species in different types of MnO₂, which play distinct roles in the catalytic oxidation of formaldehyde and benzene. In addition, surface defects of manganese-based catalysts are regarded as an important determinant of their catalytic activity, and their formation and driving of catalytic reactions are often related to their surface oxygen species [24,25,26]. Yang et al. investigated the phase-activity relationship of MnO₂ toward toluene catalytic oxidation. They proposed that the excellent catalytic performance of δ-MnO₂ may be associated with the rich oxygen vacancy and the strong mobility of oxygen species [27]. Tian et al. prepared α-, β-, and ε- MnO₂ for CO oxidation and found that β-MnO₂ possesses the lowest energies for oxygen vacancy generation as well as excellent redox properties, thus exhibiting the best CO oxidation activity [13]. These studies highlight the importance of oxygen vacancies in VOC removal and illustrate that the concentration of oxygen vacancies in different crystal structures tends to dominate their catalytic activity. Therefore, it is necessary to understand the relationship between the catalytic activity of manganese oxide and its surface structure to provide a standard for the further modification of MnOx or other metal oxide catalysts. However, noticeable differences are presented in the surface morphology and crystal structure of different crystalline MnO₂, which makes it difficult to clarify the contribution of oxygen vacancies to the catalytic reaction. For example, δ-MnO₂ is a layered structure, whereas α-, β- and γ-MnO₂ are common tunneling structures. Among them, γ-MnO₂ is a spherical structure, and α- and β-MnO₂ are similar nanorod-like structures. The differences in crystal structure and surface morphology can inherently lead to variations in catalyst surface properties, which can obscure the critical role of oxygen vacancies. Therefore, it is essential to reveal the effect of surface vacancies on catalytic reactions based on the same morphology. In addition, oxygen vacancy-mediated catalytic reactions may be accompanied by the migration and release of oxygen species and the generation of new oxygen vacancies, thus the transformation of these active surface species in catalytic reactions and their contribution to the removal of VOCs need further clarification.

Herein, we compared the removal efficiency of CH₃SH by α-MnO₂ and β-MnO₂ with similar surface morphology but different crystal structures. Their physicochemical properties were subsequently characterized by various analysis techniques. The number of surface low valence Mn, oxygen vacancies and redox properties were studied regarding high-resolution transmission electron microscope (HRTEM) and electron paramagnetic resonance (EPR), X-ray photoelectron spectroscopy (XPS), H₂-TPR and O₂-TPD. The changes of catalyst surface species before and after the reaction were characterized by XPS, and the variation of intermediate species of CH₃SH during the reaction were also monitored.

2. Materials and Methods

2.1. Chemical Reagents

α-MnO₂ and β-MnO₂ were synthesized through the hydrothermal synthesis method according to previous research [28]. Potassium permanganate (KMnO₄, Chengdu Colon Chemicals Co., LTD, Chengdu, China), hydrated manganese sulfate (MnSO₄·H₂O, Aladdin Reagent Co., LTD, Shanghai, China) and ammonium persulfate ((NH₄)₂S₂O₈, Aladdin Reagent Co., LTD, Shanghai, China) were used without further purification.

Synthesis of α-MnO₂: 0.1 M KMnO₄ and 0.05 M MnSO₄·H₂O were dissolved in 70 mL deionized water and stirred for 30 min. The resulting solution was transferred to a 100 mL Teflon-lined autoclave and maintained at 160 °C for 12 h. After cooling to room temperature, the precipitate was centrifuged and washed with distilled water (700–1000 mL) three times. Finally, the precipitate was dried at 80 °C for 4 h and calcination at 360 °C for 2 h.

Synthesis of β-MnO₂: 0.14 M MnSO₄·H₂O and 0.14 M (NH₄)₂S₂O₈ were dissolved in 70 mL deionized water and stirred for 30 min. The resulting solution was transferred to a 100 mL Teflon-lined autoclave and maintained at 140 °C for 12 h. After cooling to room temperature, the precipitate was centrifuged and washed with distilled water (700–1000 mL) three times. Finally, the precipitate was dried at 80 °C for 4 h and calcination at 360 °C for 2 h.

2.2. Catalyst Characterization

The refined test of X-ray powder diffraction (XRD) of the products was performed using a Bragg-Brentano-type powder diffractometer (Nihongo TTRIII, Tokyo City, Japan, operated at 40 kV and 200 mA, Cu Kα radiation, λ = 0.15418 nm). To investigate the Brunauer-Emmett-Teller (BET) surface areas, average pore diameters, and total pore volumes of the samples, N₂ adsorption-desorption isotherms were determined using a NOVA 4200e Surface Area and Pore Size Analyzer. Electron paramagnetic resonance (EPR) signals were carried out on a Bruker A300 spectrometer (Saarbrucken, Germany) at 25 °C. XPS profiles were obtained with a Thermo Scientific K-Alpha spectrometer (Waltham, MA, USA). The binding energy (BE) values were calibrated using the C 1 s peak at 284.8 eV. The Raman spectra were recorded using a 514 nm laser excitation source with an integration time of 3 s and 30 accumulations (Raman, BX41, HOEIBA Scientific, Paris, France). Scanning electron microscopy (SEM, VEGA3SBH, Brno, Czech Republic) and high-resolution transmission electron microscopy (HRTEM, Talos F200X, Thermo Scientific, Waltham, MA, USA) were used to observe catalyst morphologies.

Hydrogen temperature-programmed reduction (H₂-TPR) and oxygen temperature- programmed desorption (O₂-TPD) experiments were performed on a FULI II 7970 gas chromatograph (Fuli Analytical Instrument Inc., Hangzhou, China) with a thermal conductivity detector (TCD). In H₂-TPR experiments, 50 mg of the sample was placed in a quartz tube and pretreated in a gas flow of 10% H₂/Ar (30 mL min⁻¹) at 100 °C for 30 min to remove impurities. After the pretreatment process, the sample was reduced by 10% H₂/Ar (30 mL min⁻¹) from 100 to 800 °C with a heating rate of 10 °C/min. For O₂-TPD analysis, 50 mg of sample was loaded on the quartz tube, heated to 105 °C and pretreated with He (30 mL min⁻¹) for 30 min to remove surface adsorbed water, followed by cooling to 30 °C. Subsequently, the sample was adsorbed by 10% O₂/He (30 mL min⁻¹) at room temperature for 60 min, and then He (30 mL min⁻¹) was used to purge the sample for 30 min to remove physically adsorbed O₂ and stabilize the baseline. Subsequently, the temperature was ramped from 30 to 850 °C at 10 °C mL min⁻¹.

2.3. Catalyst Activity Evaluation

The catalytic performance for the CH₃SH decomposition was investigated in a fixed-bed quartz tube reactor (i.d. = 6 mm). 200 mg samples with the size of 40–60 meshes were loaded into the reactor. The reaction temperature was controlled and maintained for about 1 h at each designated temperature. The inlet CH₃SH concentration was set at 5000 ppm, and the total flow rate was maintained at 30 mL min⁻¹. The concentration of CH₃SH was recorded by GC-9790 (FULI, China) equipped with a flame ionization detector (FID) and flame photometric detector (FPD), and the CH₃SH conversion ratio was calculated as follows:

$${\mathrm{CH}}_3\mathrm{SH} \mathrm{conversion}=\frac{C_{in}-C_{out}}{C_{in}}\times 100\%$$

$C_{in}$ represents the inlet concentration of CH₃SH and $C_{out}$ is the outlet concentration of CH₃SH.

The reaction rates of CH₃SH decomposition were determined in the kinetic regime at a CH₃SH conversion lower than 20% at different temperatures; The reaction rate (r_CH3SH; mol g⁻¹ h⁻¹) for CH₃SH decomposition was calculated according to the following equations:

$$r_{C{H}_{3}SH}=\frac{C_{C{H}_{3}SH}\times X_{C{H}_{3}SH}\times F}{m_{cat}}$$

$$r_{norm}=\frac{C_{C{H}_{3}SH}\times X_{C{H}_{3}SH}\times F}{m_{cat}\times S_{BET}}$$

where the $C_{C{H}_{3}SH}$ represents the initial methyl mercaptan concentration, F (mol·h⁻¹) represents the total flow rate, $X_{C{H}_{3}SH}$ denotes CH₃SH conversion, and $S_{BET}$ (m²·g⁻¹) represents the specific surface area of catalysts.

The turnover frequency (TOF, h⁻¹) was calculated for different crystal types based on oxygen vacancy concentration for MnO₂, and indicates the number of reactions of methyl mercaptan at each active site per unit of time, thus TOF was obtained using the following equation:

$$TOF\left(h^{-1}\right)=\frac{C_{C{H}_{3}SH}\times X_{C{H}_{3}SH}\times F}{\frac{m_{Mn{O}_2}}{M_{Mn{O}_2}}\times \left(M{n}^{2+}+M{n}^{3+}\right)}$$

$M_{Mn{O}_2}$ (mol·g⁻¹) is the molar mass of $Mn{O}_2$, and $M{n}^{2+}+M{n}^{3+}$ derived from XPS data, which represent the concentration of the oxygen vacancies of MnO₂ deduced from the obtained XPS spectra.

3. Results and Discussion

3.1. Structure and Morphology

XRD was used to determine the crystal structure of the prepared material. The XRD patterns of as-prepared MnO₂ with various crystal types are shown in Figure 1. The diffraction peaks located at ~12.7°, ~18.1°, ~28.8°, and ~37.5° can be assigned to α-MnO₂ (JCPDS card no. 44-0141) (Figure 1A), and the peaks at ~28.7°, ~37.2°, ~42.7° and ~56.4° can be ascribed to β-MnO₂ (JCPDS card no. 24-0735) (Figure 1B) [29]. There was no apparent crystal transformation on these samples after 360 °C calcination. The sharp and strait peaks of β-MnO₂ could indicate its great crystallization and large grain size; α-MnO₂ presented wide bands with relatively lower crystallinity and smaller grain sizes. The above results indicate the successful obtaining of the two kinds of MnO₂ with specific crystal phases.

The morphologies of MnO₂ samples were characterized by a scanning electron microscopy (SEM), a transmission electron microscope (TEM), and a high-resolution TEM (HRTEM). As can be seen in Figure 2A, α-MnO₂ showed a stacking-nanorod structure with an average length of about 330 nm. β-MnO₂ (Figure 2F) also showed a typical rod shape with a diameter near 50 nm and a length of ~1.2 μm. TEM showed consistent results with the SEM that α-MnO₂ (Figure 2B) and β-MnO₂ (Figure 2G) had similar nanorod-like structures as previously reported [30]. The well-identified periodic lattice fringes of 6.9 Å can be clearly observed in Figure 2C, corresponding to the interplanar distance of (110) facet of α-MnO₂. Figure 2H exhibited the lattice fringes of 3.1 Å, which match the interplanar distance of the (110) facet of β-MnO₂ well. Compared with the β-MnO₂ samples, α-MnO₂ showed more blurry lattice fringes, representing poor crystallinity, which also agrees well with the XRD patterns. In addition, the presence of defects was further demonstrated using the inverse Fast Fourier Transform (FFT) pattern (Figure 2E,J). Significantly more lattice distortion can be clearly observed on the surface of α-MnO₂ (Figure 2D) (highlighted by red ovals), thus leading to more defects than β-MnO₂ [10]. Besides, severe blurring of the lattice fringes was also detected on α-MnO₂ than β-MnO₂. Lattice distortion can be caused by nearby point defects. Simultaneously, a defect layer will be formed once the defect concentration is high enough, resulting in a blurry lattice fringe in the HRTEM images [31]. Hence, the intrinsic defective structure of α-MnO₂ was confirmed. Oxygen vacancies, as an important point defect in catalysts, play a prominent role in the catalytic reaction process, and the high oxygen vacancy concentration will result in a blurry lattice fringe, which can be reflected in the HRTEM images [32]. As shown in Figure 3, the EPR signal corresponding to g = 2.003 can be attributed to oxygen vacancies, and its signal intensity can represent the number of oxygen vacancies [27]. Therefore, more oxygen vacancies on α-MnO₂ than β-MnO₂ can be confirmed based on EPR, consistent with HRTEM analysis.

The BET surface areas (S_BET), and pore volumes of the two catalysts are shown in Figure 4A,B. It is reported that the different structures assembled by MnO₆ octahedra in MnO₂ will affect the related surface areas and pore volumes. β-MnO₂ presented relatively low specific surface areas (12.76 m² g⁻¹) and pore volumes (0.06 cm³ g⁻¹), whereas α-MnO₂ showed higher specific surface areas (34.59 m² g⁻¹) and pore volumes (0.13 cm³ g⁻¹). Moreover, the nitrogen adsorption-desorption isotherms of α-MnO₂ and β-MnO₂ displayed a type IV curve with H₃-type hysteresis loops, indicating that both samples were mesoporous structures [33].

3.2. Catalytic Performance

In order to explore the activity of two catalysts on sulfur-containing volatile organic pollutants (S-VOCs), methyl mercaptan (CH₃SH) was chosen as the model S-VOCs, and the catalytic activities of α-MnO₂ and β-MnO₂ are shown in Figure 5A. Two MnO₂ samples exhibited significantly different catalytic performance in CH₃SH catalytic reaction. α-MnO₂ (74%) exhibited significantly better catalyst activity than β-MnO₂ (3%) at 30 °C. The decreases of CH₃SH conversion for α-MnO₂ at 50 °C may be due to the desorption of CH₃SH on the catalyst. As the temperature increased, the conversion of CH₃SH reached 100% at 100 °C with both catalysts. Furthermore, the reaction rates of α-MnO₂ and β-MnO₂ at 50 °C were calculated based on the activity experiments. As shown in Figure 5B, α-MnO₂ showed the CH₃SH reaction rate of 2.9 × 10⁻² mol g⁻¹ h⁻¹, this being ~4.1 times higher than the rates measured for and β-MnO₂ at 50 °C, which was consistent with the results for the catalytic activity.

It is well known that the specific surface area plays a critical role in catalytic reactions. To eliminate its influence, the reaction rates with surface area normalization were calculated at different temperatures based on the data from the activity experiments. The results of the normalized reaction rates (r_norm, mol m⁻² h⁻¹) of CH₃SH decomposition are shown in Figure 6A. The normalized reaction rates for α-MnO₂ were obviously higher than those of β-MnO₂ at 30, 50, 60 and 80 °C, which suggested that reactivity was not governed by the specific surface area. Turnover frequency (TOF) is essential for studying the intrinsic reactivity of catalysts. In this work, the TOF (h⁻¹) was calculated based on oxygen vacancy concentration, and the TOF value for the CH₃SH catalytic decomposition was conducted at 50 °C with 0.01 g of catalyst and was calculated within a low CH₃SH conversion (1 h of reaction, below 15.0%). As displayed in Figure 6B, the α-MnO₂ showed the highest TOF value of 0.14 h⁻¹, which was 1.8 times as that of the β-MnO₂ (0.08 h⁻¹), indicating that α-MnO₂ has better catalytic performance for CH₃SH.

3.3. Redox Capacity and Oxygen Species

To evaluate the reduction behaviors of MnO₂ samples, H₂-temperature-programmed reduction (TPR) was performed (Figure 7A). For α-MnO₂, the peaks at 289 and 309 °C corresponded to the reduction of Mn⁴⁺ → Mn³⁺ and Mn³⁺ → Mn²⁺, respectively, and the peaks around 291 and 317 °C for β-MnO₂ were attributed to Mn⁴⁺ → Mn³⁺ and Mn³⁺ → Mn²⁺, respectively [20,34]. The reduction temperature of α-MnO₂ was lower than that of β-MnO₂, indicating that the reduction of α-MnO₂ is relatively faster. More importantly, the more remarkable reduction ability of α-MnO₂ means easier deoxygenation during hydrogen treatment, suggesting that oxygen migration is more likely to occur on its surface. Therefore, α-MnO₂ features stronger oxygen species mobility than β-MnO₂.

O₂-TPD was conducted further to explore the oxygen species of the MnO₂ catalysts. Figure 7B shows three desorption peaks related to oxygen species that can be observed on the MnO₂. The low-temperature peak below 400 °C was ascribed to the chemisorbed active oxygen species on the surface (O⁻ and O₂⁻) [35]. The desorption peaks at 400–650 °C and 700–850 °C were related to the release of subsurface and bulk lattice oxygen species (O²⁻), respectively [36,37]. The desorption of surface oxygen at low temperature (<400 °C) plays the primary role as reactive oxygen species participating in the catalytic reaction [38]. Moreover, the lower temperature of the surface oxygen desorption peak means better low-temperature mobility of oxygen species. As depicted in Figure 7B, α-MnO₂ showed a lower temperature at 146 °C of the surface oxygen desorption peak than β-MnO₂ at 367 °C, indicating the better low-temperature mobility of oxygen species, which is in agreement with H₂-TPR.

More bonding properties were discussed through Raman spectra (Figure 8A). The peaks at 348 and 640 cm⁻¹ corresponded to the Mn-O bending and the stretching vibration, respectively [39]. Significantly weaker and broader Raman peaks at around 640 cm⁻¹ were detected for α-MnO₂ than β-MnO₂, suggesting lower crystallinity and more defects due to the lattice distortion [40]. To evaluate the strength of the Mn-O bond, the bond force constant (k) was calculated from Hooke’s law [41,42] using the following equation: ω =$\frac{1}{2\Pi c}\sqrt{\frac{k}{\mu }}$, where ω is the Raman shift (cm⁻¹), c is light velocity, and μ is the effective mass of the Mn-O bond. The calculated Mn-O force constant (k) is shown in the inset of Figure 8B. Thus, the Mn-O bond force constant of α-MnO₂ (293 N/m) was smaller than that of β-MnO₂ (296 N/m), implying the weaker Mn-O bond. The weaker Mn-O bond means easier migration of O and easier redox of Mn during the reaction, which is beneficial for catalytic reactions [25].

3.4. Identification of the Role of Oxygen Vacancies in CH₃SH Degradation

The surface elemental composition and chemical state of these MnO₂ samples were identified by XPS. The XPS spectra of Mn 2p_3/2 of the samples are shown in Figure 9. The peaks corresponding to binding energies at 642.7, 641.7 and 640.6 eV can be attributed to Mn⁴⁺, Mn³⁺ and Mn²⁺, respectively [43,44]. It is noteworthy that the binding energy corresponding to different valences of Mn were slightly different in both MnO₂ samples, indicating that crystal phase structure has a certain effect on the electron density of the MnO₂ surface, which is related to the degree of charge imbalance, oxygen vacancies, as well as the relative content of Mn²⁺, Mn³⁺ and Mn⁴⁺ [45]. Specifically, the oxygen vacancy will be generated to maintain electrostatic balance with the increasing Mn²⁺ and Mn³⁺ proportion, and the proportion of low-valence Mn is generally regarded as an indicator of surface oxygen vacancies [44]. As shown in

Table 1. Mn 3s Mn 2p_3/2 results before and after the reaction of B-MnO₂ and C-MnO₂.

Catalysts	Mn2+ (%)	Mn3+ (%)	Mn4+ (%)	Mn2+ + Mn3+/ Mn4+ (%)	AOS 1
α-MnO2-Fresh	11.31	30.31	58.38	0.71	3.57
β-MnO2-Fresh	2.71	35.04	62.26	0.6	3.75
α-MnO2-Spent	25.19	43.77	31.04	2.22	2.83
β-MnO2-Spent	16.39	46.99	36.62	1.99	3.25

¹ AOS = 8.956 − 1.126 × ΔE.

and Figure 9A, the proportion of the low valence Mn (Mn²⁺ + Mn³⁺) showed α-MnO₂ (41.62%) > β-MnO₂ (37.74%). Besides, the average oxidation state (AOS) of MnO₂ was calculated according to the formula of AOS = 8.956 − 1.126 ΔE [46], which was based on the size of Mn 3s multiple splitting (ΔE) in Mn 3s XPS spectra (Figure 10A). The AOS values of the Mn element in α-MnO₂ and β-MnO₂ were calculated to be 3.57 and 3.75, respectively. In previous studies, lower AOS of MnO₂ was also able to indicate more surface oxygen vacancies [47]. Therefore, it can be inferred that α-MnO₂ has a greater surface oxygen vacancy density than β-MnO₂ (in agreement with HETEM and EPR).

Additionally, Mn²⁺-O and Mn³⁺-O bonds are weaker than Mn⁴⁺-O [48]. Thus, the higher ratio of low valence Mn (Mn²⁺ + Mn³⁺) endows MnO₂ with a larger proportion of weaker M-O bonds on its surface, meaning the easier release of O to participate in the reaction. Figure 9A showed that α-MnO₂ had more low valence Mn (Mn²⁺ + Mn³⁺) than β-MnO₂, which also implies the easier release of surface oxygen species, consistent with the results of H₂-TPR and O₂-TPD. Comparison of the Mn 2p_3/2 spectra before and after the reaction (Figure 9B) of MnO₂ with CH₃SH showed that the valence state of Mn in both samples changed obviously, suggesting the electron transfer during the reaction. For both samples, Mn⁴⁺ decreased, and Mn²⁺ and Mn³⁺ increased, proving that the high-valent Mn (IV) was reduced by gaining electrons during the reaction. After the reaction, the Mn²⁺+Mn³⁺/Mn⁴⁺ of α-MnO₂ increased by 1.51 and that of β-MnO₂ by 1.39, and AOS decreased by 0.74 for α-MnO₂ and that of β-MnO₂ decreased by 0.5 (Figure 10B), testifying that α-MnO₂ was reduced to a greater extent by gaining more electrons than β-MnO₂, which can well match H₂-TPR results. It is noteworthy that higher AOS usually indicates a stronger electron-gaining ability of the catalyst because of the presence of more high-valent atoms, however, α-MnO₂ exhibited a stronger electron-gaining ability in the reaction with CH₃SH, suggesting that oxygen vacancies play a more important role in catalyzing CH₃SH comparing to the high-valent Mn. This may explain the fact that chemisorption is the rate-limiting step for electron transfer, and more surface oxygen vacancies provide more surface adsorption sites for CH₃SH.

In addition, the reaction between CH₃SH and the catalyst could change the electronic environment of the catalyst. During the reaction, as the ratio of Mn²⁺ and Mn³⁺ increased, weaker Mn-O bonds were continuously formed and broken, leading to deoxygenation and further generation of oxygen vacancies to maintain electrostatic equilibrium, which may provide new sites for the reaction, and these desorbed oxygen species may favor the catalytic oxidation of CH₃SH as well.

The XPS spectra of O 1s of the samples are shown in Figure 11. Peaks with binding energies at 529–529.8, 530.9–532 and 533 eV in the XPS spectra of O 1s of the MnO₂ samples (Figure 11) can be attributed to lattice oxygen (O_latt) and surface adsorption oxygen (O_ads), and surface hydroxyl oxygen (O_adsO-H), respectively [49,50,51]. The molar ratio of O_ads/O_latt is shown in

Table 2. O 1s results before and after the reaction of B-MnO₂ and C-MnO₂.

Catalysts	Oads	OH2O	Olatt	Oads/Olatt
α-MnO2-Fresh	30.22	23.53	46.25	0.65
β-MnO2-Fresh	20.68	14.10	65.22	0.32
α-MnO2-Spent	55.20	18.39	26.41	2.09
β-MnO2-Spent	39.78	20.34	39.87	1.00

and follows the order of α-MnO₂ (0.65) > β-MnO₂(0.35). O_ads was generally considered the most reactive oxygen species in the catalytic reaction and capable of participating in the catalytic oxidation of VOCs in previous reports [52,53]. The oxygen species changes before and after the reaction are shown in Figure 11B and Table 2. The O_latt and O_ads for both two materials decreased and increased, respectively, suggesting the migration of O_latt to form O_ads during the reduction of Mn. Obviously, α-MnO₂ formed more O_ads after the reaction, which corresponds to its greater degree of reduction, also implying that α-MnO₂ has a stronger catalytic capacity.

3.5. Product Detection during the Reaction

The main gas phase products were monitored quantitatively to better understand the reaction process of CH₃SH over two different catalysts. As displayed in Figure 12, the decomposition of CH₃SH at different temperatures corresponded to the production of CH₃SSCH₃. Meanwhile, the concentration of CH₃SH during the reaction showed an excellent correlation with the concentration of CH₃SSCH₃, indicating that CH₃SSCH₃ was the main gas-phase product. At 150 °C, CH₃SH was completely decomposed for both MnO₂ catalysts, consistent with the thermodynamic theory that catalytic reactions proceed easier at higher temperatures. The yield of CH₃SSCH₃ gradually decreased when T > 100 °C, which may be due to the further catalytic oxidation of CH₃SSCH₃ at higher temperatures.

Figure 13 shows the changes of S 2p before and after the reaction, which was used to detect the solid-phase intermediates in the reaction process. No S species were detected on two catalysts before the reaction. In contrast, significant amounts of S species were detected on both samples after the reaction, indicating that some sulfur-containing products were adsorbed on the catalyst surface. As shown in Figure 13B, three peaks at 163.2, 167.9, and 169.2 eV corresponding to S⁰, SO₃²⁻ (S⁴⁺), and SO₄²⁻ (S⁶⁺), respectively [30,54], which all showed higher valence than S^2- from CH₃SH, indicating the oxidation of S during the reaction. It is worth noting that the catalytic experiments in CH₃SH were performed under a nitrogen atmosphere, so it can be concluded that the O in the S-O species was mainly derived from the O_ads of MnO₂. This illustrated the O_ads involvement in the catalytic reaction of CH₃SH. SO₃²⁻ and SO₄²⁻ were mainly retained on the manganese dioxide surface in the form of MnSO₃ and MnSO₄, implying that chemisorption was a prerequisite for the decomposition of CH₃SH on MnO₂. Notably, reacted α-MnO₂ showed a higher proportion of SO₄²⁻ (S⁶⁺) (26.31%) than β-MnO₂ (12.56%) (

Table 3. S 2p results before and after the reaction of B-MnO₂ and C-MnO₂.

Catalysts	S0	S4+	S6+
α-MnO2-Fresh	0	0	0
β-MnO2-Fresh	0	0	0
α-MnO2-Spent	12.96	60.73	26.31
β-MnO2-Spent	13.38	74.06	12,56

and Figure 13B), suggesting a greater degree of S oxidation, which corresponds to a greater reduction of Mn⁴⁺ after the reaction (Table 1 and Figure 9B). Furthermore, SO₄²⁻ requires more oxygen to be coordinated with S than SO₃²⁻, so a higher proportion of SO₄²⁻ production requires more O_ads to participate in the reaction. Correspondingly, the H₂-TPR, O₂-TPD, and XPS analysis demonstrated more O_ads and better surface oxygen mobility for α-MnO₂ than β-MnO₂.

Based on the above experimental and characterization analysis, the catalytic mechanism of CH₃SH by α-MnO₂ and β-MnO₂ can be inferred in Figure 14. CH₃SH was first chemisorbed on the MnO₂ surface and subsequently underwent a single electron transfer to form CH₃S·, and then the two CH₃S· were coupled to form CH₃SSCH₃. Based on the formation of S-S bonds, it is speculated that the single electron transfer occurs on S, suggesting that the chemisorption may be through the formation of Mn-S bonds. Moreover, more lattice oxygen was released during the reduction of Mn, which was involved in the further catalytic oxidation of S-containing species to produce SO₃²⁻ and SO₄²⁻, and may further form new oxygen vacancies to support more active sites. Although β-MnO₂ enjoys a higher AOS based on the proportion of high valence Mn, α-MnO₂ showed better catalytic activity due to more oxygen vacancies and stronger oxygen mobility.

4. Conclusions

In this paper, MnO₂ catalysts with similar morphology but different crystal structures (α-MnO₂ and β-MnO₂) were successfully prepared, and the effects of the physicochemical properties on the catalytic activities were systematically investigated. Both Mn-based catalysts showed significant removal of CH₃SH at 150 °C, achieving complete conversion, whereas α-MnO₂ exhibited significantly better catalytic activity compared to β-MnO₂ at a lower temperature (T < 100 °C) under a GHSV of 9000 mL g⁻¹ h⁻¹. Coupled with O₂-TPD, H₂-TPR, Raman spectra, XPS, EPR, and HRTEM, it was demonstrated that α-MnO₂ has more oxygen vacancies, stronger surface oxygen migration ability, and better redox properties, which can be favorable for CH₃SH decomposition. The readily released lattice oxygen during the reaction promoted further oxidative decomposition of S-containing species. The intermediate products of the solid and gas phases were determined as CH₃SSCH₃ and the S⁰, SO₃²⁻, and SO₄²⁻, respectively. The catalytic mechanism was further proposed as the oxygen vacancies on MnO₂ provided active sites for the adsorption of CH₃SH, facilitating the electron transfer of MnO₂ with CH₃SH, and the oxygen species derived from the Mn surface were further involved in the CH₃SH catalytic oxidation. The findings of this study are essential for broadening the application of Mn-based catalysts in the removal of S-VOCs and providing new insights into the mechanism of interfacial reactions between VOCs and metal-based catalysts.