3.1. Catalysts Characterization
Figure 1 presented SEM images of TiO2-NS and TiO2-NTO and HRTEM images of TiMn15Ce30-NS and TiMn15Ce30-NTO. As shown in Fig. 1a, TiO2-NS presented a nanosheet-like morphology with 70 ~ 160 nm in length, and TiO2-NTO presented a truncated bipyramid with the edge length of the upper and lower planes of the bipyramid was about 30 ~ 100 nm (Fig. 1b). HRTEM image of TiMn15Ce30-NS in Fig. 1c showed the lattice spacing of 0.24 and 0.35 nm which could be assigned to the interplanar distances of the {001} plane and {101} plane of anatase phase TiO2 (Wang et al. 2020b).
Combined the SEM and HRTEM analysis, we presented schematic diagrams of TiMn15Ce30-NS in Fig. 1e, and the flat top and bottom surfaces should be the {001} planes based on the symmetries of anatase TiO2 (Li et al. 2019b), indicating the preferential exposed {001} plane of TiMn15Ce30-NS catalyst. Similarly, the lattice fringes of 0.36 nm corresponded to the {101} plane for the TiO2 in TiMn15Ce30-NTO (Fig. 1d) (Leng et al. 2015) and we concluded that TiMn15Ce30-NTO preferentially exposed {101} plane which could be presented as the schematic diagrams in Fig. 1f.
Besides, the nanosheet-like and truncated bipyramid morphology were mainly retained in TiMn15Ce30-NS (Fig. S1a) and TiMn15Ce30-NTO (Fig. S1b) and uniformly distributed particles were also observed on the surface due to the addition of Mn and Ce. Energy spectrum scanning of TiMn15Ce30-NS revealed that four elements including Mn, Ce, O and Ti were detected on the surface of the catalyst (Fig. 2), demonstrating that Mn and Ce were successfully supported on TiO2 nanosheet. The Mn and Ce were dispersed evenly on the surface of the carrier, which is conducive to improving the catalytic activity (Jiao et al. 2019).
The XRD patterns of anatase TiO2-NS, TiO2-NTO, TiMn15Ce30-NS and TiMn15Ce30-NTO were presented in Fig. 3, and the results showed that the peaks of TiO2-NTO and TiO2-NS were consistent with the anatase phase TiO2 (JCPDS 21-1272) (Gui et al. 2009). After loading Mn, the corresponding peaks of TiMn15-NS did not change significantly (Fig. S2). With the loading of Ce in TiMn15-NS, the peak at 28.9o attributing to MnO2 (110) (JCPDS 24-0735) crystal plane was observed (Cao et al. 2017). The average crystal sizes of TiO2-NS, TiO2-NTO, TiMn15Ce30-NS and TiMn15Ce30-NTO were calculated by Scherrer equation based on the (101) diffraction peak of TiO2 (Table S1). Noticeably, the grain size was 43.8 nm in TiO2-NS, and the value reduced to 34.2 nm in TiMn15Ce30-NS, demonstrating the crystal size of TiO2 was refined by introducing Mn and Ce. While the counterpart data in TiMn15Ce30-NTO increased to 40.0 nm from 37.9 nm in TiO2-NTO. It is well-known that smaller grain size endows more abundant surface area or interface area, which might provide more defect sites as active sites for catalytic reactions, promoting catalytic activity (Zhao et al. 2020b).
Deng et al. (Deng et al. 2016) have proved that the preferential exposure of the (001) crystal plane of TiO2 crystal was related to the broad peak of the (004) diffusion and the narrower peak of (200). As can be seen in Table S1, the full width at half maximum (FWHM) of TiMn15Ce30-NS (004) (0.312o) peak was larger than that of TiMn15Ce30-NTO (0.192o), while its FWHM of (200o) peak (0.223o) was smaller than the corresponding data of 0.227o for TiMn15Ce30-NTO. This phenomenon further proved that TiMn15Ce30-NS preferentially exposed {001} crystal plane as obtained in Fig. 1.
Figure 4 depicted the N2 adsorption/desorption isothermal of pure TiO2, TiMn15-NS and TiMn15Ce30-NS. TiO2-NS showed I-type isotherm (IUPAC) (Rahman et al. 2019) and its curve increased slightly in the relative pressure ranging from 0.0 to 0.2 ascribed to micropores filling (Abebe et al. 2018), implying that the pore structure of TiO2-NS was mainly microporous. The TiMn15-NS was presented as a type IV isotherm and it displayed an H3-type hysteresis loop at relative pressure from 0.75 to 1.0, besides, the rapid rise during the relative pressure from 0.85 to 1.0 might be attributed to the condensation of adsorbate gas (Abebe et al. 2018). After the introduction of Ce, the isotherm of TiMn15Ce30-NS retained type IV, and the H3-type hysteresis loop appeared at P/P0 of 0.7 ~ 1.0.
Table 1 summarized the specific surface area and average pore diameter of catalysts. The average pore size of TiO2-NS was 1.41 nm, and the value was 7.83 nm in TiMn15-NS, indicating that TiMn15-NS mainly possessed mesoporous. Moreover, the specific surface area of TiMn15-NS was 32.0 m2/g comparing to the 19.8 m2/g in TiO2-NS attributed to the loading of Mn. The specific surface area of TiMn15Ce30-NS did not change significantly after further loading Ce, while TiMn15Ce30-NS presented a hierarchical porous structure with a relatively wide pore distribution at 8.78, 17.63, and 36.51 nm ranging from 2 ~ 86 nm (Fig. 3), indicating the coexistence of mesoporous and macroporous. Studies have reported that the mesoporous structure could provide relatively higher internal specific surface area (Liu &He 2010), and the macroporous possessed relatively small diffusion resistance, which is conducive to the mass transfer process (Arandiyan et al. 2018). Thereby, we could reasonably deduce that the wide range of hierarchical porous structure of TiMn15Ce30-NS may be favorable for o-DCB diffusion and adsorption on TiMn15Ce-NS, promoting the catalytic oxidation activity.
Table 1
Surface areas, total pore volumes and average pore diameters of TiO2-NS, TiMn15-NS and TiMn15Ce30-NS.
Samples | SBET [a] (m2/g) | Vp [a] (cm3/g) | Dp [a] (nm) |
TiO2-NS | 19.8 | 0.05 | 1.41 |
TiMn15-NS | 32.0 | 0.06 | 7.83 |
TiMn15Ce30-NS | 32.0 | 0.07 | 6.83 |
[a] Structure parameters of the catalysts calculated via N2 isotherm; SBET: BET surface area; Vp: total pore volume; Dp: average pore diameter. |
The reduction performances of TiO2-NS, TiMn15-NS and TiMn15Ce30-NS were studied through H2-TPR with the results reflected in Fig. 5. The reduction temperature of pristine TiO2-NS was 658 oC. After loading Mn on the TiO2-NS, the reduction temperature of TiO2 in TiMn15-NS decreased to 636 oC. In addition, two new reduction peaks centered at 265 oC assigned to the reduction of MnO2 to Mn2O3 and 335 oC attributed to the transformation of Mn2O3 to Mn3O4 were observed (Chen et al. 2018). The peak at 447 oC can be attributed to the reduction of Mn3O4 to MnO (Zhao et al. 2019). For TiMn15Ce30-NS, the temperature of MnO2 to Mn2O3 and Mn2O3 to Mn3O4 decreased to 236 oC and 297 oC, respectively. The peak attributed to Mn3O4 reduction to MnO was also reduced to 396 oC on the TiMn15Ce30-NS curve (Wu et al. 2019). These results revealed that the addition of Ce significantly affected the reduction temperatures of MnOx, proving that the loading of Ce contributed to the electron transfer between MnOx and CeOx and promoted the migration of surface oxygen or lattice oxygen, (Geng et al. 2020) consequently improving the reduction performance of MnOx (Wang et al. 2013). Note that the peak at 466 oC corresponded to the reduction of CeO2 on the catalyst surface to Ce2O3 (Zhao et al. 2019).
XPS was used to characterize the surface properties of the TiMn15-NS and TiMn15Ce30-NS. Fig. 6a showed the Mn 2p spectrum of the catalysts, the peaks observed at 641.8 eV and 653.3 eV corresponded to Mn3+, and the peaks at 643.1 eV and 654.7 eV were attributed to Mn4+ (Ferrel-Álvarez et al. 2018). it is widely established that the higher ratio of Mn4+ can introduce more available oxygen species (Wang et al. 2013) and facilitate the redox cycling during reaction processes, which could promote catalytic activity (Geng et al. 2020). The relative surface Mn4+ ratio calculated by Mn4+/(Mn3++Mn4+) was displayed in Table 2, the surface Mn4+ content on TiMn15-NS was 65.8%., the value increased to 76.7% with introducing Ce. The phenomenon further elucidated that the addition of Ce was favorable for the electron transformation between Ce and Mn through the equilibrium of “\({\text{Mn}}^{\text{3+}}\text{+ }{\text{Ce}}^{\text{4+}}\text{↔}{\text{ Mn}}^{\text{4+}}\text{+}{\text{ Ce}}^{\text{3+}}\)” (Wang et al. 2020a).
Table 2
Surface atomic ratios of Mn 2p, O1s in TiMn15-NS and TiMn15Ce30-NS.
Samples | Surface atomic ratio[a] (%) |
| Mn4+/(Mn3++Mn4+) | Oα/(Oα+ Oβ) |
TiMn15-NS | 65.8 | 34.1 |
TiMn15Ce30-NS | 76.7 | 46.4 |
[a] Surface atomic ratios were calculated based on the peak areas of binding energy. |
The O1s peaks can be deconvoluted into two parts (Fig. 6b): the binding energies at 528.1 eV - 531.2 eV corresponding to lattice oxygen O2− (denoted as Oβ), and the one at 528.1 eV - 534.8 eV ascribing to the surface oxygen species (denoted as Oα), such as O− from the defect-oxide or OH− ascribed to the hydroxyl-like group (Chen et al. 2017). And the relative concentration of surface oxygen species was calculated by Oα/(Oα+ Oβ) as listed in Table 2. The area ratio of Oα/(Oα+ Oβ) in TiMn15Ce30-NS (46.4%) was higher than that in TiMn15-NS (34.1%), indicating that the addition of Ce increased the concentration of reactive oxygen species.
3.2. Catalytic performance of the prepared catalysts in o-DCB oxidation
The conversion of o-DCB as a function of temperature over TiO2-NTO, TiO2-NS, TiMn15-NS and TiMn15-NTO were shown in Fig. 7. It can be seen that the activity of TiO2-NS was higher than that of TiO2-NTO in the entire temperature range. For example, the o-DCB conversion rate of TiO2-NS was 38.2% at 350 oC, which was much higher than 15.6% of TiO2-NTO, suggesting that the crystal plane of TiO2 significantly affected the activity of o-DCB oxidation, and the fact that TiO2-NS with preferentially exposed {001} facet exhibited good catalytic activity could be explained by the abundant amount of unsaturated coordinated Ti atoms on the {001} crystal plane favoured high chemical activity, favouring the o-DCB adsorption and oxidation (Liu et al. 2016). The above result is in agreement with our previous study (Wen et al. 2021). TiO2-NS doped with different amount of Mn or Ce were investigated. As illustrated in Fig. S3, the TiO2-NS with 15 mol% of Mn/(Ti+Mn) exhibited the catalytic performance for o-DCB oxidation. Therefore, TiMn15-NS and TiMn15-NTO were synthesized and investigated. As obtained in Fig. 7, TiMn15-NS still presented superior activity with the T50% of 234 oC than that of 266 oC in TiMn15-NTO.
Moreover, the oxidation activity of TiMn15-NS was further improved when doping with 30 mol% of Ce in TiMn15Ce30-NS as suggested in Fig. S3. Similarly, TiMn15Ce30-NS also exhibited better oxidation activity than TiMn15Ce30-NTO (Fig. 8). For instance, the T50% of TiMn15Ce30-NS was 156 oC, and the value was 213 oC for TiMn15Ce30-NTO. These results further confirmed the influence of the crystal plane of TiO2 on o-DCB decomposition, which be closely associated with the preferentially exposed {001} crystal plane of TiMn15Ce30-NS (Wen et al. 2021). In addition, based on the results in Fig. 7 and Fig. 8, it showed that introduction of CeOx on both TiMn15-NS and TiMn15-NTO significantly increased the activity for o-DCB oxidation. Especially for the TiMn15Ce30-NS, it exhibited excellent low temperature catalytic activity than TiMn15Ce30-NTO. Given the results mentioned above, the excellent catalytic activity of TiMn15Ce30 for o-DCB decomposition could be explained as follows. On the one hand, TiMnCe-NS with preferential exposed {001} crystal plane has more oxygen vacancies (Li et al. 2019b), which could enhance the activation of the active oxygen species and provide more adsorption sites for gas molecules (Liu et al. 2009). On another aspect, the strong interaction between MnOx and CeOx occurred in TiMnCe-NS, (Wang et al. 2020a) and hence the low-temperature reducibility of MnOx was improved (Wang et al. 2013). In addition, TiMn15Ce30-NS possessed a higher content of Mn4+ and Oα than those of TiMn15-NS according to the XPS results due to the addition of Ce, which facilitated redox cycles between Mn3+ and Ce4+, enhancing the migration of oxygen species (Wang et al. 2020a). Noteworthy, the current TiMn15Ce30-NS exhibited a significantly superior catalytic performance with the T90% of 238 oC compared with the TiFe5Ca70-S (322 oC) under the same space velocity condition, (Wen et al. 2021) suggesting that the catalytic activity of MnCeOx/TiO2 for decomposing o-DCB could be further improved.
Water is an inevitable component in the actual flue gas, and the presence of water has an important effect on the catalytic oxidation for PCAHs (Ma et al. 2016). As shown in Fig. 8, in the presence of water (0.5 vol.% H2O), the activity of TiMn15Ce30-NS in the low-temperature region (< 200 oC) was inhibited to some extent, which was most likely ascribed to the competition adsorption between the H2O and o-DCB on active sites of the catalyst surface (Hetrick et al. 2011). When the temperature was higher than 200 oC, the o-DCB conversion increased with the total conversion of o-DCB was still obtained at 300 oC. Moreover, the overall activity of TiMn15Ce30-NS with T50% of 179 oC and T90% of 240 oC was still clearly better than that of TiMn15Ce30-NTO with T50% of 205 oC and T90% of 246 oC under humid conditions. The above results suggested that the crystal plane also affected the catalytic performance of o-DCB oxidation in humid condition, (Wen et al. 2021) which might be explained by that the nanosheet morphology TiO2 {001} crystal surface allowed more hydroxyl groups to combine with the ortho Ti sites to generate more acidic sites (Wen et al. 2018) which could facilitate the adsorption of o-DCB (Li et al. 2021, Sun et al. 2016) and favor H2O dissociation more easily compared with TiO2 {101} crystal surface (Vittadini et al. 1998), thus reducing the negative effect of competitive adsorption of H2O and o-DCB.
The lifetime of TiMn15-NS, and TiMn15Ce30-NS for the o-DCB oxidation were evaluated at 350 oC (Fig. 9). TiMn15Ce30-NS catalyst exhibited excellent stability with o-DCB conversion retained at 98.9% for 12 h. However, for TiMn15-NS, the corresponding conversion decreased to 83.5%. The higher stability of TiMn15Ce30-NS could be also explained by its nanosheet structure with preferentially exposed {001} facet and the strong interaction among CeO2, MnO2 and TiO2-NS support, thus reducing the adsorption of Cl− species on the active site of MnOx and CeOx, which is advantageous for chlorination species removal from the catalyst surface and maintain a relatively stable activity (Wang et al. 2014).
In brief, TiO2 nanosheet supported MnCeOx/TiO2 (TiMn15Ce30-NS) showed excellent low temperature catalytic activity, enhanced water resistance, and good stability for o-DCB oxidation, which was significantly better than most of the previous reported Mn-based catalysts (Table 3).
Table 3
The states of the arts of Mn-based catalysts for PCAHs oxidation.
Catalysts | Pollutant | Reaction conditions | T50/T90 (oC) | Ref. |
TiMn15Ce30-NS | o-DCB | 50 ppm, 10% O2/N2, 36000 mL·g−1·h−1 | 156/238 | This work |
MnOx-D | o-DCB | 1000 ppm, 20% O2/N2, 15000 mL·g−1·h−1 | 238/321 | (Ling et al. 2020) |
CeMnOx-rod | o-DCB | 1000 ppm, 20% O2/N2, 15000 mL·g−1·h−1 | 293/346 | (Zhao et al. 2019) |
CeMn30 | o-DCB | 1000 ppm, 20% O2/N2, 15000 mL·g−1·h−1 | 291/347 | (Yang et al. 2019) |
5.0%Ce/OMS | o-DCB | 1000 ppm, 60000 mL·g−1·h−1 | >300/>335 | (Chen et al. 2019) |
Mn/Ce (0.50) | Chlorobenzene | 1000 ppm, 20% O2/N2, 15000 h−1 | 165/231 | (Hu et al. 2018) |
15CM/TS-1.5 | o-DCB | 500 ppm, 20% O2/N2, 15000 mL·g−1·h−1 | >300/>360 | (Zhao et al. 2018) |
MCZ-0.67 | Chlorobenzene | 1000 ppm, 21% O2/N2, 20000 h−1 | >175/226 | (Zhu et al. 2019) |
Mn-Ce-Zr ternary mixed oxides | o-DCB | 1000 ppm, 21% O2/N2, 30000 h−1 | >320/>390 | (Long et al. 2019) |
Mn-Ce-Fe | o-DCB | 1000 ppm, 12000 mL·g−1·h−1 | 253/380 | (Tang et al. 2016) |
30Cu/MnOx | Chlorobenzene | 1000 ppm, 20% O2/N2, 15000 h−1 | >150/290 | (Chen et al. 2020) |
MnOx/Co3O4-4h | Chlorobenzene | 1000 ppm, 20% O2/N2, 60000 mL·g−1·h−1 | >275/>330 | (Hu et al. 2021) |
[email protected] | Chlorobenzene | 1000 ppm, 20% O2/N2, 15000 mL·g−1·h−1 | 341/392 | (Zhao et al. 2020a) |
3.3. In situ FT-IR analysis of TiMn15Ce30-NS
Transient reaction processes characterized by in situ DRIFT spectra were performed to identify various reactive species and deduce a possible mechanism of the reaction. The DRIFT spectra collected at different time intervals during the adsorption and oxidation of o-DCB over TiMn15Ce30-NS at 300 oC were shown in Fig. 10a. It can be seen that from 1 min to 30 min, no obvious change was observed in the range of 1000 ~ 2500 cm−1. The weak peak of 1283 cm−1 (C-O stretching vibration) was assigned to the phenols (Lichtenberger 2004), the band at 1410 cm−1 was associated with the formation of the o-benzoquinone (Lichtenberger 2004). It is widely accepted that the initial stage of o-DCB oxidation mainly includes the dissociative adsorption of o-DCB and nucleophilic substitution reaction. The lone pair electrons of the Cl atom are captured by the Lewis acid sites, causing a nucleophilic attack on the position of chlorine and breaking the C-Cl bonds, then o-DCB adsorbed on the active sites of transition metal oxides, generating phenolics via the reaction with surface oxygen (Lichtenberger 2004, Ma et al. 2013). Moreover, the o-DCB molecules can react at the Brønsted acid sites and produce HCl during the oxidation process according to Taralunga (Taralunga et al. 2006). Consequently, the characteristic peak of 1283 cm−1 indicated that o-DCB was dechlorinated to form phenolics on the surface of TiMn15Ce30-NS, and the band at 1410 cm−1 implied that the phenolics were further transformed into o-benzoquinone (Lichtenberger 2004). The band at 1374 cm−1 corresponded to the symmetric -COO- stretching vibration of the surface formates species (Krishnamoorthy et al. 2000). The peaks at 1360 cm−1 and 1540 cm−1 were ascribed to the formation of carboxylates (acetates type) (Krishnamoorthy et al. 2000, Ma et al. 2016). Hence, it can be deduced that the o-benzoquinone was oxidized to formates and acetate over the TiMn15Ce30-NS. The weak peaks at 2320 cm−1 and 2355 cm−1 were assigned to the CO2 (Silva et al. 2018), indicating that the intermediate formates and acetic species were further converted to CO2 by deep oxidation. Besides, in addition to the intermediates mentioned above, the o-DCB catalyzed oxidation process is usually accompanied by the formation of H2O, which can be explained from our previous studies (Ma et al. 2019).
Figure 10b showed the in situ DRIFT spectra over TiMn15Ce30-NS collected at 100, 200, and 300 oC after 20 min reaction. As can be seen, the intensity of formates (1374 cm−1), acetates (1360 cm−1), and surface carbonates (1540 cm−1) were significantly enhanced when increasing the reaction temperature to 200 oC from 100 oC, demonstrating that the dechlorination, ring-opening and catalytic oxidation reactions took place over TiMn15Ce30-NS at 200 oC during o-DCB decomposition. When increasing the temperature to 300 ℃, the intensity of formates, acetates, and surface carbonates all weakened obviously, meanwhile the peaks of CO2 at 2320 cm−1 and 2355 cm−1 slightly increased. Therefore, it can be deduced that increasing temperature in a certain range (particularly from 200 to 300 oC) could promote the deep oxidation of o-DCB to CO2 over TiMn15Ce30-NS.
Based on the mentioned above, the catalytic oxidation mechanism of o-DCB over TiMn15Ce30-NS can be elucidated in Fig. 11. As reported in previous reports (Mahmood et al. 2019), TiO2 {001} crystal planes endow abundant 5-fold under-coordinated Ti atoms and 2-fold coordinated oxygen sites which were proved to be the adsorption active sites, we hence deduce that the o-DCB was prone to adsorbed on the {001} facet of TiMn15Ce30-NS. Subsequently, the adsorbed o-DCB transformed into phenolic species and then o-benzoquinone through nucleophilic substitution. Afterward, formates and acetate species were formed via a ring-opening reaction of o-benzoquinone, and these species were finally oxidized to CO2 and H2O.