Boosting Benzene’s Ozone Catalytic Oxidation at Mild Temperatures over Highly Dispersed Ag-Doped Mn₃O₄

Abstract

Transition metal oxides show high activity while still facing the challenges of low mineralization and poor durability in the ozone catalytic oxidation (OCO) of volatile organic compounds (VOCs). Improving the oxygen mobility and low-temperature reducibility of transition metal oxides was found to be an effective way to address the above challenges. Here, highly dispersed Ag was added to Mn₃O₄ via the co-precipitation oxalate route, and the obtained Ag/Mn₃O₄ exhibited higher mineralization and stability in benzene catalytic ozonation at room temperature. Compared to Mn₃O₄, the concentration of CO₂ formed from benzene oxidation over Ag/Mn₃O₄ was significantly increased, from 585.4 ppm to 810.9 ppm, while CO generation was greatly suppressed to only one tenth of its original value (194 ppm vs. 19 ppm). In addition, Ag/Mn₃O₄ exhibited higher catalytic stability than Mn₃O₄. The introduction of Ag obviously improved the oxygen mobility and low-temperature reducibility of Mn₃O₄. Moreover, the highly dispersed Ag also promoted the activity of surface oxygen species and the chemisorption of benzene on Mn₃O₄. The above physicochemical properties contributed to the excellent catalytic performance and durability of Ag/Mn₃O₄. This research could shed light on the improvement in VOC mineralization via ozone catalytic oxidation.

1. Introduction

Volatile organic compounds (VOCs), typically emitted from paint shops, oil refineries, etc., are considered precursors of photochemical smog and haze and could cause serious harm to both humans and the eco-environment [1,2]. VOCs not only have a high potential to form ozone and particulate matter (PM_2.5) but also exhibit toxicity and carcinogenicity [3,4]. Therefore, the governments in many countries have implemented strict measures and extremely low emission limits to control VOC pollution. Currently, there are various treatment methods for VOCs, such as adsorption, catalytic oxidation, combustion, and so on [5,6]. Among the numerous VOC removal technologies, ozone catalytic oxidation (OCO) exhibited several advantages. For example, during the OCO process, O₃ can be decomposed into oxygen atoms and further transformed into a variety of powerful reactive oxygen species (ROS), which can precisely and effectively react with the VOCs at near-room temperature, thus ensuring energy efficiency and safety [7,8]. Consequently, OCO has been extensively investigated for the removal of various types of VOCs, such as formaldehyde, methyl mercaptan, dichloroethane, aromatic VOCs, etc. [9,10,11,12].

Catalysts play an important role in the OCO of VOCs. In general, noble metal catalysts such as Pt, Pd, and Rh always show excellent performance, but the high price and the disadvantage of easy poisoning limit their practical application [13]. Compared to noble metal catalysts, transition metal oxides (Mn, Co, Ni, etc.) show several advantages, such as a preferential reducibility, abundant oxygen species, a high anti-poisoning ability, and so on. Moreover, the cost of these transition metal oxides is also lower than that of noble metals [14]. Therefore, transition metal oxides have been widely applied in the OCO of VOCs.

Among the various transition metal oxides, MnO_x usually exerts multiple manganese oxidation states, resulting in fruit oxygen vacancies as well as strong low-temperature reducibility, which are key factors for the redox steps of VOC ozonation [15,16]. Compared to other MnO_x such as Mn₂O₃ and MnO₂, Mn₃O₄ was reported to facilitate the reactions between VOC molecules and ozone to form surface-adsorbed oxygen atomic species, which could be more easily and deeply oxidized [17]. However, Mn₃O₄ usually suffered from the incomplete oxidation of VOC molecules, low ozone utilization, and severe instability, which prevented its further application in OCO technology [18]. As a result, further research should be conducted to improve both the catalytic activity and stability of Mn₃O₄.

It is widely accepted that the addition of another suitable metal could always promote the catalytic activity and durability of single-metal oxides, which would be attributed to the “electronic effect” and “geometric effect” [19,20]. Thus, many transition metals have been introduced to modify the Mn₃O₄ in ozone decomposition, such as Cu, Fe, Ce, and so on [21,22]. For example, Fu et al. found that the partial substitution of Mn by Fe in Mn₃O₄ could result in abundant oxygen species and active sites of Mn³⁺, thus greatly promoting ozone decomposition [23]. Of all the transition metals, Ag has distinctive characteristics. It has been reported that Ag-containing catalysts usually have quantities of active sites with high activity in ozone decomposition or the OCO of VOCs [24]. In addition, the application of Ag could also promote the chemisorption of VOCs, which is a crucial step in the OCO process [10]. Therefore, Ag was introduced into manganese oxides, and the synthesized Ag-Mn catalysts were applied to ozone decomposition. Xie et al. argued that the deposition of Ag on Mn₃O₄ nanosheets noticeably improved the reducibility and generated more highly active oxygen vacancies; thus, the obtained Mn₃O₄@Ag exhibited excellent and stable catalytic performance in O₃ decomposition [25]. Moreover, Liu et al. reported that the OCO of VOCs involves ozone decomposition and the oxidation of VOC molecules [7]. Therefore, the higher catalytic performance of Ag/Mn₃O₄ in ozone decomposition also indicated its potential in the OCO of VOCs.

In this paper, highly dispersed Ag was added to Mn₃O₄ via the co-precipitation oxalate route, and the obtained Ag/Mn₃O₄ catalyst was then applied to the OCO of benzene. The introduction of Ag greatly improved benzene mineralization, with much less by-products, and the catalytic stability of Mn₃O₄ at room temperature, which was attributed to the promoted oxygen mobility, enhanced low-temperature reducibility, activated surface oxygen species, and enhanced chemisorption of benzene. This work provides us with a new way to improve VOC mineralization with low by-products and catalytic durability in the OCO of VOCs.

2. Results and Discussions

2.1. Crystal Phase and Textures of Mn₃O₄-Based Catalysts

The XRD patterns and N₂ adsorption/desorption isotherms and pore size distributions are shown in Figure 1. From Figure 1a, the Mn₃O₄, Ag/Mn₃O₄ somewhat showed the same XRD patterns as standard cubic Mn₃O₄ (PDF # 18-0803) according to the analysis results of Jade 6.5, indicating that Ag doping did not change the crystal phase of manganese oxide. In addition, the diffraction peaks assigned to Ag or Ag₂O could hardly be observed, indicating the high dispersion of Ag in Mn₃O₄ [26].

The N₂ adsorption/desorption isotherms of the two catalysts in Figure 1b are both type IV with H3 hysteresis loops within a wide relative pressure range of 0.4–1.0, implying the development of mesoporous structures after calcination [27]. Notably, Mn₃O₄ showed a specific surface area of 385.2 m²·g⁻¹, while Ag doping resulted only in a slight decrease in the specific surface area of Ag/Mn₃O₄ (see

Table 1. Summary of BJH parameters of Mn₃O₄ and Ag/Mn₃O_4.

Sample	Specific Surface Area/m2·g−1	Pore Volume/cm3·g−1	Average Pore Size/nm
Mn3O4	385.2	0.40	4.2
Ag/Mn3O4	372.4	0.44	4.8

). The results were beneficial to the adsorption of ozone and benzene, and the higher specific surface area would also provide more active sites to promote the OCO of benzene.

Figure 2 shows the high-resolution TEM (HRTEM) image and high-angle annular dark-field STEM images of Ag/Mn₃O₄, respectively. As shown in Figure 2b, the interplanar spacings were 0.299 nm and 0.253 nm, which were very close to the (2 2 0) and (3 1 1) crystal planes of Mn₃O₄ (PDF # 13-0162), respectively. Furthermore, the dispersion of Ag was uniform, without any agglomeration, as shown in Figure 2b,d, which was in somewhat of an agreement with the XRD results in Table 1. Summary of BJH parameters of Mn₃O₄ and Ag/Mn₃O_4.

2.2. Surface Chemical Properties of the Obtained Catalysts

The surface chemical compositions, apparent element concentration and valence, and oxygen species of the two samples were determined by the XPS technique. The Mn 2p, Ag 3d, and O 1s spectra of Mn₃O₄ and Ag/Mn₃O₄ are shown in Figure 3, and the quantitative analysis results are summarized in

Table 2. Summary of the XPS analysis results.

Catalyst	Mn3+		Mn4+		Ag+		Oα		Oβ + Oγ
	BE/eV	%	BE/eV	%	BE/eV	% §1	BE/eV	%	BE/eV	%
Mn3O4	641.8	79.81	643.3	20.19	-	-	529.8	66.1	531.2	33.9
Ag/Mn3O4	641.5	75.70	643.0	24.30	367.4	5.01	529.5	57.0	530.6	43.0

^§1: The ratio of surface Ag⁺ to the total surface metal ions.

.

As seen in Figure 3a, the asymmetric Mn 2p _3/2 spectra of the two catalysts can be decomposed into two symmetric peaks ranging from 641.5 to 641.8 eV and from 643.0 to 643.3 eV, which should be attributed to the surface Mn³⁺ and Mn⁴⁺, respectively [28,29]. As shown in Table 2, Ag doping resulted in a higher proportion of surface Mn⁴⁺ compared to Mn₃O₄, and the result was favorable for the formation of oxygen vacancies due to the charge compensation principle [30]. Furthermore, the higher-valence manganese ions were reported to be in the form of Mn⁴⁺-O_surf. (surface oxygen species) Lewis acid–base pairs, which could efficiently activate O₃ [14]. Notably, the Ag 3d _5/2 spectra in Figure 3b show only one peak, centered at 367.4 eV, which could correspond to Ag⁺ [31]. As shown in Table 2, the surface ratio of Ag⁺ to Mnⁿ⁺ was very close to the designed value, indicating that the co-precipitation method via the oxalate route could make the Ag uniformly dispersed in both the bulk and apparent phases. Ag or other noble metal nanoparticles with higher surface energy often agglomerate and form large-sized particles as a result of Ostwald ripening [32]. Considering the analysis results of XRD, HAADF-STEM, and XPS, it can be concluded that Ag doping through the oxalate route could avoid the formation of Ag agglomeration.

The O 1s XPS spectra of Mn₃O₄ and Ag/Mn₃O₄ shown in Figure 3c could be fitted with two main symmetrical peaks, and the peaks centered at 529.5–529.8 eV should be surface lattice oxygen species (denoted as O_α), while the peaks at 530.6–531.2 eV are assigned as surface adsorbed oxygen species or oxygen vacancies (denoted as O_α) [33,34]. In addition, the O 1s spectra of Ag/Mn₃O₄ also showed another peak centered at about 533.7 eV, which was widely accepted as surface-adsorbed OH/oxy-carbonate species [35]. Noticeably, the total amount of surface-adsorbed oxygen species/vacancies (including both O_β and O_γ) of Ag/Mn₃O₄ was larger than that of Mn₃O₄ displayed in Table 2, indicating that Ag doping not only made the surface oxygen species more active but also increased the ratio of more active adsorbed oxygen species. The above analysis results of the O 1s XPS spectra are somewhat consistent with those of the Mn 2p XPS spectra. Many researchers have proven that surface adsorbed oxygen species/vacancies and surface OH/oxy-carbonate species are beneficial for both O₃ activation and VOC catalytic oxidation [36,37]. Thus, we can expect a satisfactory catalytic performance of Ag/Mn₃O₄.

2.3. Evolution of Oxygen Species and Reducibility of the Gained Catalysts

The evolution of the surface oxygen species of the two catalysts was investigated by the O₂-TPD technique, and the results are shown in Figure 4a. In general, the desorption of adsorbed oxygen species occurred at temperatures ranging from 200 to 400 °C, while the surface lattice oxygen species usually desorbed at temperatures ranging from 400 to 600 °C [38,39]. For the Mn₃O₄ sample, the peak centered at 276 and 498 °C should represent the desorption of adsorbed oxygen species and lattice oxygen species, respectively. In comparison, the desorption peak of adsorbed oxygen species from Ag/Mn₃O₄ at 325 °C, with an obvious shoulder at 241 °C, indicated that the introduction of Ag into Mn₃O₄ resulted in surface oxygen species with a higher reactivity. It is widely accepted that active surface oxygen species would not only generate unsaturated chemical bonds but also lead to unbalanced charges, which would ultimately make the decomposition of ozone as well as the redox reactions occur more easily [40,41]. Since the surface oxygen species of Ag/Mn₃O₄ showed a higher activity than that of Mn₃O₄, the former sample would exert a better catalytic performance in the OCO of benzene.

The reducibility, which could be detected by the H₂-TPR technique, played an important role in the catalytic oxidation of VOCs, and the H₂-TPR curves of Mn₃O₄ and Ag/Mn₃O₄ are shown in Figure 4b. The reduction of Mnⁿ⁺ by H₂ usually followed the order of Mn⁴⁺→Mn³⁺→Mn²⁺ [42,43]. For sample Mn₃O₄, the first reduction peak at 275 °C with a shoulder at 180 °C should represent the reduction of the Mn³⁺ located in the lattice sites to Mn²⁺, and the latter peak, centered at 431 °C, could be assigned to the final reduction of Mn³⁺ to Mn²⁺ [44]. With Ag doping, the reduction peaks of Ag/Mn₃O₄ were shifted to lower temperatures compared to those of Mn₃O₄. In other words, the Ag/Mn₃O₄ catalyst in this research exerted a much better low-temperature reducibility, bringing a positive effect to the subsequent OCO of benzene.

2.4. Temperature-Programmed Surface Reaction on the Obtained Catalysts

To illustrate the activity of surface oxygen species and the adsorption/desorption behavior of benzene on the as-prepared catalysts, the C₆H₆-TPSR technique was introduced, and the profiles are shown in Figure 5.

As shown in Figure 5, the desorption of benzene from Ag/Mn₃O₄ had a larger peak than that of Mn₃O₄, which was inconsistent with their pore volumes in Table 1. The CO₂ was probably formed from the reaction between chemisorbed benzene and adsorbed surface oxygen species, considering both the C₆H₆-TPSR process and the O₂-TPD results [45,46]. Apparently, sample Ag/Mn₃O₄ showed a higher CO₂ desorption but a lower peak temperature than Mn₃O₄ under the N₂ atmosphere, indicating that the surface oxygen species of the former catalyst was more active than that of the latter. In addition, the introduction of Ag also provided more active sites, which was beneficial for the chemisorption of benzene [47]. Researchers have demonstrated that active surface oxygen species could easily react with VOC molecules and that the formed oxygen vacancies would be favorable for ozone decomposition, which, in turn, would be beneficial for benzene ozonation [48].

2.5. Catalytic Performance of Benzene OCO

The performance of benzene OCO over the two obtained catalysts is shown in Figure 6, and some important related data are summarized in

Table 3. Summary of benzene OCO with two samples at 25 °C and 70 °C, respectively.

Sample	25 °C				70 °C
	C6H6 rem./% §2	CO2 /ppm	CO /ppm	Res. O3/ppm	C6H6 rem./% §3	CO2 /ppm	CO /ppm	Res. O3/ppm
Mn3O4	91.8	585.4	194.7	270.2	99.8	1046.4	301.4	0
Ag/Mn3O4	97.4	810.9	19.2	67.2	99.8	1165.4	157.5	0

^§2 The value was measured after the OCO of benzene lasted for 170 min at 25 °C. ^§3 The benzene removal efficiency was detected after the OCO of benzene lasted for 70 min at 70 °C.

.

From Figure 6a,c, it can be seen that the benzene removal efficiencies over Mn₃O₄ and Ag/Mn₃O₄ rapidly increased from 0 to nearly 100% but then gradually decreased to 91.8% and 97.4% at 25 °C after 170 min (Table 3), respectively. This is because benzene can hardly be completely mineralized by ozone at low temperatures, and some inorganic intermediates would be formed during this process on the surface active sites, leading to the gradual deactivation of the catalysts [49]. According to Einaga et al., some intermediates such as weakly bound formic acid and strongly bound formates (carboxylic acid) may be formed and accumulated on Mn-based catalysts during the OCO of benzene, and some of the above intermediates could be further oxidized to CO₂ or CO at relatively higher temperatures [11]. With the consumption of these intermediates, the activity of the catalyst could be recovered, and this would be consistent with the catalytic performance of the two samples in Figure 6. Noticeably, Ag doping improved the catalytic activity and stability of Mn₃O₄, which might have been due to the active oxygen species on Ag/Mn₃O₄ which could easily react with the intermediates at low temperatures, and the results were in fair agreement with the O₂-TPD and XPS data. Consequently, more CO₂ and much less CO as a by-product were generated when Ag/Mn₃O₄ was applied (see Figure 6d), which might have been mainly due to more active oxygen species, according to the O₂-TPD and C₆H₆-TPSR results. In Table 3, we can clearly see that the CO₂ generation at 25 °C was about 585.4 ppm and 810.9 ppm with Mn₃O₄ and Ag/Mn₃O₄, respectively, while CO production was suppressed to only one tenth of the original value when the latter catalyst was used (19.2 ppm with Ag/Mn₃O₄ vs. 194.7 ppm with Mn₃O₄). That is to say, introducing Ag is beneficial for the complete mineralization of benzene, while also reducing the formation of by-products.

When the temperature was increased to 70 °C, the benzene conversion rates recovered rapidly to nearly 100%, and no deactivation was observed in the last 70 min (Figure 6a,c). Moreover, the elevated temperature also favored the production of CO₂ and CO, which showed a sharp increase in the first 30 min and then decreased to stable, normal values. The sharp peaks for both CO₂ and CO were attributed to the oxidation of some organic intermediates simulated in the catalytic oxidation of benzene at 25 °C [44]. The benzene mineralization percentage for both samples at 70 °C in the last 60 min was very close to 100%, according to the data in Table 3, but Ag/Mn₃O₄ also showed higher CO₂ production and lower CO generation than Mn₃O₄. Meanwhile, nearly no O₃ could be detected in the outlet, further demonstrating that the two samples exerted a satisfying catalytic performance at 70 °C.

Briefly, the Ag doping of Mn₃O₄ improved mineralization and durability in the OCO of benzene at mild temperatures. CO₂ production was promoted while CO generation was noticeably suppressed when Ag/Mn₃O₄ was applied. Considering the above analysis results of the catalysts’ characterization, it can be concluded that the highly dispersed Ag in Mn₃O₄ resulted in more active oxygen species, which improved the oxygen mobility and low-temperature reducibility of Mn₃O₄. Moreover, the active oxygen species were also beneficial for the decomposition and activation of ozone, generating reactive oxygen species which could easily and rapidly react with the benzene molecules or intermediates on Ag/Mn₃O₄. As a result, the inorganic intermediates accumulated on the surface of Mn₃O₄ were quickly and deeply oxidized with Ag doping. The above factors can likely be attributed to the preferred mineralization and catalytic durability of Ag/Mn₃O₄ in the OCO of benzene.

3. Experimental

3.1. Catalyst Preparation

Three reagents were used in this research without further purification. The manganese nitrate solution (50 wt%, AR) and ammonium oxalate monohydrate (AR) were purchased from Sinopharm Chemical Reagent Co., Ltd., Shanghai, China. Silver nitrate (AR) was purchased from Lingfeng Chemical Reagent Co., Ltd., Shanghai, China.

The MnO_x and AgMn bimetallic catalysts were synthesized by the (co)precipitation method. First, 3.58 g of manganese nitrate solution and 1.42 g of ammonium oxalate monohydrate were diluted in 100 mL of deionized water. Then, the manganese nitrate solution was added drop by drop to the oxalic acid solution while stirring, and the mixed solution was stirred for another 30 min to ensure the complete precipitation of manganese ions. Then, the mixed solution was filtered and washed three times with deionized water and once with absolute alcohol. The filtered sediment was then dried at 80 °C overnight. Finally, the dried sediment was calcined at 250 °C for 5 h under static air, with a heating rate of 2 °C·min⁻¹. After calcination, the resulting catalyst was Mn₃O₄. The AgMn bimetallic catalyst with a Ag/Mn molar ratio of 1:19 was obtained by the co-precipitation method. The molar ratio of metal ions to oxalate ions was the same as that used in the preparation of Mn₃O₄, and it was designated as Ag/Mn₃O₄.

3.2. Characterizations

The XRD patterns of the obtained catalysts were recorded using a Rigaku RINT 2200 system (Rigaku Co. Ltd., Sevenoaks, UK) with a Cu Kα radiation, and the 2θ ranged from 10 to 80°. In addition, N₂ adsorption/desorption isotherms were performed using a Micromeritics TriStar II, and the sample was degassed at 200 °C for 1 h prior to measurement. High-resolution transmission electron microscopy (JEM-ARM200F, JEOL) was used to determine the microstructures and interplanar spacings of the as-prepared samples. Oxygen temperature-programmed desorption was carried out using a BELCAT-30 flow-type catalyst analyzer (BEL JAPAN, Inc., Haradanaka, Toyonaka-shi, Japan); both obtained catalysts were pretreated under Ar flow at 250 °C for 1 h before testing. In the O₂-TPD test, the atmosphere was changed to 5% O₂/N₂ when the pretreated catalyst was cooled to r.t. The Ar flow was turned on again when the adsorption of oxygen on the pretreated catalyst reached dynamic equilibrium. Finally, the data were recorded at a temperature ranging from r.t. to 600 °C. The pretreatment of the catalyst during the H₂-TPR test was the same as that of O₂-TPD. The H₂-TPR signal was marked down when the Ar flow was replaced by 5% H₂/N₂, and the operating temperature also ranged from r.t. to 600 °C. X-ray photoelectron spectroscopy (XPS, AXIS UltraDLD, Kratos group, Spring Valley, NV, USA) was used to determine the metal oxidation states, chemical composition, and surface oxygen species of the as-prepared catalysts. In addition, a C₆H₆ temperature-programmed surface reaction (C₆H₆-TPSR) was performed using FTIR (Perkin-Elmer, Spectrum 100, Waltham, MA, USA), with a gas cell at a resolution of 4 cm⁻¹.

3.3. Catalytic Activity Measurements

The catalytic performance test of the synthesized catalysts (40–60 mesh) was carried out in a U-type quartz tube fixed-bed reactor with an inner diameter of 8 mm. In each test, 100 mg of the selected sample was used and heated at 250 °C in simulated air for 1 h. The temperature was maintained at 25 °C for some time and then increased to 70 °C until the end of the test. The weight hourly space velocity (WHSV) in this research was 60,000 mL·g_cat⁻¹·h⁻¹, and the reaction air flow was 100 mL·min⁻¹, with an initial benzene concentration of 200 ppm. The ozone was generated by a UV lamp, and the initial concentration was 2200 ppm. The compositions (C₆H₆, CO, CO₂, etc.) of the downstream flow were analyzed online using an FTIR spectrometer (Perkin-Elmer, Waltham, MA, USA, resolution of 4 cm⁻¹) equipped with a gas cell (2 m path length), while the residual ozone was measured using an ozone analyzer (EG-600, Ebara Jitsugyo, Japan). Prior to the tests, the potential compositions were calibrated with standard C₆H₆, CO, CO₂, etc. In addition, the benzene removal efficiency, CO₂ selectivity, and CO_x (CO₂ + CO) selectivity were calculated using the following equations [44]:

$${\eta }_{C_6{H}_6}={\displaystyle \frac{200-C_{C_6{H}_6}}{200}}\times 100\%$$

(1)

$$S_{C{O}_2}={\displaystyle \frac{C_{C{O}_2}}{6\times \left(200-C_{C_6{H}_6}\right)}}\times 100\%$$

(2)

$$S_{C{O}_x}={\displaystyle \frac{C_{C{O}_2}+C_{CO}}{6\times \left(200-C_{C_6{H}_6}\right)}}\times 100\%$$

(3)

where C_C6H6, C_CO2, and C_CO were the initial concentrations of benzene, CO₂, and CO, respectively, after the reaction became stabilized.

4. Conclusions

Ag/Mn₃O₄ was synthesized by the co-precipitation oxalate route and applied to benzene OCO. Ag doping greatly improved crucial physicochemical properties of Mn₃O₄, such as better low-temperature reducibility, oxygen mobility, and surface active oxygen species, and promoted the chemisorption of benzene. As a result, Ag/Mn₃O₄ exhibited an excellent catalytic performance in the OCO of benzene at mild temperatures. For example, CO₂ generation was 585.4 ppm with Mn₃O₄, while it increased to 810.9 ppm when Ag/Mn₃O₄ was used. Meanwhile, CO generation was strongly suppressed to one tenth of its original value (194 ppm vs. 19 ppm). The addition of Ag species to Mn₃O₄ could lead to higher benzene mineralization during the OCO process. Moreover, Ag/Mn₃O₄ also exerted an improved catalytic durability, indicating that the surface intermediates oxidized faster than those on Mn₃O₄.