Promoted Catalytic Properties of Acetone over Cerium-Modified Mullite Catalyst YMn 2 O 5

Mullite catalysts have become one of the most widely studied catalysts due to their highly stable structure and unique coordination with oxygen. In this work, Ce-modified mullite-type oxides Y 1-x Ce x Mn 2 O 5 have been prepared by sol-gel method to explore their Ce doping amount-dependent catalytic performance for acetone elimination. Experimental results confirm that Y 0.9 Ce 0.1 Mn 2 O 5 had optimum acetone oxidation activity, completely achieving 100% acetone conversion at 120 ° C under the reaction conditions of acetone concentration = 1000 ppm, 20 vol% O 2 /N 2 and WHSV = 36000 mL g –1 h –1 . This excellent catalytic activity comes from its larger specific surface area and higher Mn 4+ /Mn 3+ molar ratio. XRD and TEM results show that YMn 2 O 5 and CeO 2 phases form a multiphase oxide and interfacial structure. XPS results show that the content of doped CeO 2 mainly affects the surface adsorbed oxygen (Oads) and Mn 4+ content of the catalyst. Manganese species with higher chemical states are indeed more favorable for oxidation reactions on manganese-based catalysts. In addition, the reduction temperature of mixed oxides shifts to the lower temperature region, indicating that manganese and cerium oxides are more reducible, where the mobility of oxygen species is greatly enhanced. Y 0.9 Ce 0.1 Mn 2 O 5 also exhibits strong long-term stability and has good resistance to acetone elimination, showing excellent potential in eliminating acetone.


INTRODUCTION
Volatile organic compounds (VOCs), as an air pollutant, mainly come from industrial emissions, road transportation and stationary combustion (Yin et al., 2015;Liang et al., 2017;McDonald et al., 2018).They are not only toxic and carcinogenic, but also pollute the environment by generating secondary air pollutants through photochemical reactions (Colman Lerner et al., 2012;Li et al., 2018).For example, it reacts with nitrogen oxides and sulfur oxides in the environment to generate ozone and secondary aerosols.Acetone with three carbon atoms is one of the most abundant oxygen-containing VOCs in the atmosphere.As a solvent or intermediate, it is widely used in industries such as pharmaceuticals, plastics and printing (Mellouki et al., 2015).In addition, the bond energy of C=O is about 745 kJ mol -1 , which is difficult to decompose.There is an urgent need for a technology to remove acetone in the air.Among the currently known adsorption, condensation, photolysis and non-thermal plasma, catalytic oxidation is one of the most feasible strategies due to its high efficiency and economy (Yang et al., 2011;Belaissaoui et al., 2016;Mustafa et al., 2018;Zhang et al., 2019b;Zhao et al., 2019a;Lee et al., 2020;Zhang et al., 2020b).It has a lower operating temperature, higher energy efficiency, and wider applicability.Two categories of representative catalysts have been formed for catalytic oxidation of VOC, mainly including supported noble metals and transition-metal-based catalysts (Kamal et al., 2016;Liu et al., 2016;Cheng et al., 2019;Liu et al., 2019).Although noble metal-based catalysts have excellent catalytic performance, their high cost and vulnerability still limit their extensive application (Liu et al., 2016;Cheng et al., 2019).It must be a feasible and challenging approach to design transition metal oxide catalysts with activity comparable to that of noble metals.
At present, many transition metal oxide catalysts have been used in the catalytic oxidation of acetone.However, due to the problem of catalytic activity and stability, it is still a very challenging task to design transition metal oxide catalysts with the same activity and stability as precious metals.In general, mullite materials are inherently more stable.Two ligand fields (octahedron and pyramid) with Mn-Mn as the center are interlinked by oxygen atoms to form Mn-Mn dimers.A-site elements can be located in the space surrounded by the central ligand unit, forming stable crystal structures with different symmetries.Compared with traditional binary manganese oxides such as MnO2, studies have found that ternary manganese oxide AMn2O5 (A = lanthanide) shows excellent catalytic activity and thermal stability in the catalytic oxidation of various organic pollutants (Chen et al., 2018;Zhang et al., 2019c;Dong et al., 2020).Chen et al. (2020) found that compared with perovskite SmMnO3, spinel Mn3O4, and bixbyite Mn2O3, mullite SmMnMn2O5 has obvious advantages in the catalytic oxidation of ethanol and toluene in terms of catalytic activity, hydrothermal stability and durability.In essence, catalysis is an electron transfer process controlled by the surface electronic structure, so adjusting the p-d orbital hybridization intensity of Mn and O in the two ligand units can effectively change the bond strength between Mn and absorbed O*, ensuring the overall activity of the catalyst to improve.For mullite materials, the catalytic-related electronic structure can be adjusted by changing the A or B elements, thus obtaining greater freedom to optimize the catalytic behavior (Shen et al., 2022;Wen et al., 2022).Although mullite catalysts have been widely used in the catalytic oxidation of various VOCs, there are few studies on the oxidation of acetone with mullite catalysts, especially YMn2O5.
The synthesis of multiphase mullite oxide YMn2O5 and the effect of cerium doping on multiphase mullite catalyst have been systematically studied.A series of Y1-XCexMn2O5 catalysts (x = 0.1, 0.2, 0.4 and 0.6) were synthesized by sol-gel method.The catalytic activity of Ce modified samples for the oxidation of acetone was improved.The sample Y0.9Ce0.1Mn2O5with x = 0.1 showed the best catalytic performance and lower activation energy.Through XRD and TEM characterization, all samples showed similar mullite, CeO2 composite crystal phase and irregular morphology.XPS characterization showed that the content of oxygen and Mn 4+ adsorbed on the catalyst surface increased after the introduction of Ce.The N2 adsorption desorption and H2-TPR characterization tests show that the Catalyst doped with Ce has higher pore volume and specific surface area, which can provide more surface-active sites for the adsorption and oxidation of acetone, thus obtaining higher catalytic activity.In addition, the interaction between manganese and cerium oxide greatly improves the fluidity of oxygen.

Catalyst Preparation
The mullite catalyst AMn2O5 was prepared by the Sol-gel method.First, a certain amount of Y(NO3)3•6H2O and Mn(NO3)3•4H2O were put into deionized water under magnetic stirring until completely dissolved.Citric acid in excess of 10% of the cation equivalent was added to the above solution and the pH of the solution was brought to 1.After that, the mixed solution was evaporated with vigorous stirring at 90°C until viscous to form a sol, which was then dried at 110°C overnight.The xerogel was ground and calcined at 300°C for 1 h, then was heated upto 800°C at 1°C min -1 for 5 h.Finally, the powder Y0.5Ce0.5Mn2O5catalyst was obtained.The comparative catalysts Y1-xCexMn2O5 (x = 0.1, 0.2, 0.4, 0.6) was prepared by controlling the molar ratio of manganese element to cerium element, and named as YCxMO (x = 0.1, 0.2, 0.4, 0.6).

Catalyst Characterization
The crystal phase characterized by the powder X-ray diffraction (XRD) technique was accomplished on a Bruker D8 Advance with CuKα radiation in the 2θ range from 10-90° under the conditions of 40 kV and 40 mA.The morphological characteristics of catalysts were taken on scanning electron microscopy (SEM) by using Phillips XL-30 FEG instrument.The microstructures of all catalysts were recorded on high-resolution transmission electron microscopy (HRTEM) by using JEOL JEM-2100F instrument operating at an accelerating voltage of 200 kV.X-ray photoelectron spectroscopy (XPS) analysis was conducted on Thermo ESCALAB 250 electron spectrometer with AlKαX radiation.Binding energies (BE) were calibrated using the C(1s) peak of contaminant carbon (BE = 284.6 eV) as the standard.Raman spectra were conducted by using HORIBA LabRAM HR Evolution with the excitation wavelength of 532 nm at room temperature.N2 adsorption-desorption isotherms for all catalysts were performed on Quant achrome Autosorb-iQ-AG at the temperature of -196°C, from which the pore volume (VP), pore diameter (DP) and surface area (SBET) were obtained by using the classic BJH and BET methods from the desorption branches.H2 temperature programmed reduction (H2-TPR) and O2 temperature programmed desorption (O2-TPD) experiments were measured by applying Quanta chrome Autosorb-iQ-C instrument to quantify the consumed amount of H2 in the temperature range of 100-300°C and the released amount of O2 in the temperature range of 50-300°C through using a TCD detector.50 mg of catalyst was initially pretreated in a protective gas at 300°C for 50 min, then cooled to 50°C for 20 min, and finally heated up to 900°C with a heating ramp of 10°C min -1 in 10% H2/Ar.

Catalytic Activity Evaluation
The acetone oxidation was carried out in a fixed-bed quartz tube (φ = 8 mm, length = 605 mm) requiring 50 mg catalyst (40-60 mesh) for each experiment.Each temperature-programmed oxidation (TPO) reaction was conducted from 100°C to 300°C by a resistance furnace equipped with a K-thermocouple.The reactant gases, 1000 ppm acetone balanced with air, were continuously passed through the quartz tube with a flow rate of 30 mL min -1 .The calculated weight hourly space velocity (WHSV) of catalytic oxidation was approximately 36,000 mL g -1 h -1 .For studying the effect of WHSV on catalytic activity, WHSVs of about 18,000 mL g -1 h -1 and 60,000 mL g -1 h -1 were acquired by changing the total streams of feed gases to 15 mL min -1 and 50 mL min -1 , respectively.We also investigated the catalytic performance in the presence of 5.0 vol% H2O.The concentrations of reactants and products were detected by gas chromatography (GC) equipped with a hydrogen flame ionization detector (FID) and a thermal conductivity detector (TCD) from Techcomp GC7900.The catalytic performances of all YCxMO catalysts were evaluated by using T50 and T90 representing the reaction temperatures required to achieve acetone conversion (ηacetone) of 50% and 90%, respectively.And the ηacetone was calculated according the Eq. ( 1): where, Cin and Cout represent acetone concentrations before and after the oxidation reaction, respectively.The kinetics measurements were carried out in the fixed-bed reactor for acetone oxidation.The reaction rate (mol g -1 s -1 ) of acetone conversion was calculated as the Eq. ( 2): where, ηacetone is the acetone conversion, Q is the acetone flow rate (mL min -1 ) and M is the weight (g) of the catalyst.
A good linear Arrhenius plots for the oxidation of acetone could be obtained by assuming firstorder kinetics with respect to acetone and a zero-order kinetics with respect to oxygen.The activation energy (Ea) is calculated from the slope of the Arrhenius-type plot.The reaction rate r can also be expressed by the Eq. ( 4): After taking the logarithm: 2 ln ln ln ln The activation energy (Ea) can be obtained from the slope of the resulting linear plot of lnr versus T -1 .

Acetone Oxidation Activity of Doped Ce Samples
The catalytic performance of YCxMO was investigated for the acetone oxidation reaction as shown in Fig. 1(b).The acetone conversion were plotted at temperatures from 100°C to 250°C under the reaction conditions of acetone concentration = 1000 ppm and WHSV = 36 000 mL g -1 h -1 .Before the catalytic activity evaluation, a blank experiment (Fig. 1(a)) was conducted and no  appreciable acetone oxidation was detected without the existence of catalysts in the examined temperature range.Fig. 1(a) shows the dependence of the catalytic oxidation performance of acetone on the gas hourly space velocity (GHSV) on the sample YC0.1MO.The T50 and T90 values at GHSV = 18,000 mL g -1 h -1 were 167.6°C and 180°C, which were 5.4°C and 5°C lower than that at 36,000 mL g -1 h -1 , respectively.When the GHSV increased to 60,000 mL g -1 h -1 , T50 and T90 were 182.1°C and 209.4°C, which were higher than the values at 36000 mL g -1 h -1 .A higher space velocity led to a short residence time of the injection gas.The reaction gas was desorbed from the catalyst before it completely reacted with the catalyst, which reduced the catalytic efficiency.This shows that a larger GHSV may reduce the catalytic effect of the catalyst, and a higher reaction temperature is required under the same conversion rate.As shown in Fig. 1(b), the acetone conversion of all catalysts increased with the increase of reaction temperature.Meanwhile, reaction temperatures T50, T90 and T100 when the acetone conversion reached 10%, 50%, and 90%, respectively, were used as the evaluation standard for acetone oxidation.It is worth to mention that only CO2 and H2O as reaction products were detected.Using YC0.1MO, YC0.2MO, YC0.4MO and YC0.6MO, the T90 values were 185, 190, 206.5 and 211.5°C, respectively.In the presence of YC0.1MO, both the initial combustion temperature and the total oxidation temperature were lower.Obviously, YC0.1MOwas considered as the optimal catalyst for acetone oxidation in this study.CO2 selectivity was also tested to assess the inherent catalytic performance of the catalyst (Fig. 1(c)).The presence of CO was not detected in the accompanying chromatograph, Since the only product of acetone conversion is CO2, the selectivity of CO2 is consistent with the conversion of acetone.Acetone was sustained for 48 hours at 220°C to measure the durability of the catalyst (Fig. 1(d)).The catalytic activities of YC0.1MO and YMnO remain at 96% and 94% after 48 hours, which also indicates that the addition of Ce improves the thermal stability of the catalyst.The various catalysts used in recent references are given in Table 1.As can be seen from the table, the T90 temperature of YC0.1MO is generally 10-180°C lower than that of this kind of catalyst, and it shows excellent activity in the process of acetone catalytic oxidation.

Physical Properties
Fig. 2(a) shows the XRD pattern of Y1-xCexMn2O5 catalyst.The diffraction pattern of pure YMn2O5 catalyst mainly shows strong and sharp peaks located at 15.5°, 29.1°, 32.1°, 34.7° and 43°.This is the main feature of the mullite structure with Mn octahedra and cones as ligand units (JCPDS No. 34-667) (Zhang et al., 2019c).The addition of Ce broadened the diffraction peaks of YMn2O5, which means that the crystal size of mullite decreased.Its lower peak intensity indicates that Ce doping inhibited the growth of the mullite crystal phase.As x increased from 0.1 to 0.6, a gradually intensifying peak at 28° associated with CeO2 with a fluorite structure was detected (Zhu et al., 2016).With the addition of cerium, the diffraction peak at 16.1° belonging to the mullite phase weakened and broadened, and the diffraction peak at 47.5° belonging to CeO2 increased.It is shown that the mullite structure can be changed by the incorporation of Ce and the formation of CeO2 phase.For pure YMO, the main diffraction peak intensity is 29.1°, but the peak intensity becomes higher after Ce doping.It is indicated that the diffraction peak at 29.1° should include the diffraction peaks of other substances other than mullite.Referring to standard CeO2 diffraction peak card, its main peak is located at 29.3°, being overlapped with one of the main peaks of YMO.The synthesized YCMO catalyst should contain two phases, namely mullite and ceria.It is worth noting that the Ce sources doped for x = 0.1 and 0.2 are relatively small and highly dispersed because of the weaker ceria diffraction peaks.Fig. 2(b) shows the Raman spectra in the range 185-900 cm -1 for Y1-xCexMn2O5 catalyst.There are three main peaks at 225 cm -1 , 615 cm −1 and 670 cm −1 on Y1-xCexMn2O5, which are assigned to the Y 3+ ions, the Mn-O stretching vibrations of pyramidal MnO5 and octahedral MnO6 on YMn2O5, respectively.As can be seen from the atlas, the high vibration intensity and the shift of the Raman peak position on the YCe0.1MnOcatalyst indicates that the content of Mn-O increases relatively and the presence of surface stresses after a small amount of Ce doping.
Scanning electron microscopy was used to observe the morphology of catalyst doping and sintering effects.The pure phase of mullite YMO consists of nanoparticles with relatively uniform morphology, but the incorporation of Ce significantly changed the morphology of the catalyst, as shown in Fig. 3. Overall, the four samples showed similar morphological structures, and the samples under the electron microscope showed a bulky structure.Further local magnification of the block's surface showed that it is built up of finer particles, where the particles were irregular, which is consistent with SEM observation results.The finer particle composition means that the incorporation of cerium dioxide greatly inhibits crystal growth.Aggregation between particles became slightly more pronounced with increasing A-site cation defects.The packing of fine particles reduced the pore size and increased the pore volume, which was consistent with the BET results below.TEM was used to further study the microstructure of YC0.1MO more intuitively, as shown in Fig. 4. The fringe distance of 0.583-0.589nm is attributed to the lattice spacing of the (001) plane of YMn2O5.The smaller lattice plane distance of YMn2O5 (001) in YCMO is mainly because the lattice constant of cerium oxide is smaller than that of mullite.After doping, cerium oxide will inevitably generate inbound strain on the mullite surface, thus reducing the lattice fringe distance of the YMn2O5 (001) surface.The addition of Ce reduced the crystallinity of the catalyst.Due to the smaller size, the shape of the YCMO sample became irregular, which was consistent with the SEM observation.The existence of interphase stress greatly changed the morphology of the material.In addition, the lattice mismatch between mullite and ceria will inevitably lead to oxygen defects, which is conducive to the efficient progress of the catalytic reaction.
The nitrogen adsorption-desorption isotherms and pore size distributions of all modified mullite catalysts are shown in Fig. 5. Structural properties such as specific surface area and average pore size are listed in Table 2. Type IV isotherms with H3-type hysteresis loops can be observed for all catalysts, suggesting the presence of mesopores due to non-uniform size or shape of catalyst particles (Tang et al., 2015).In addition, the PSD (Particle Size Distribution) curve is shown in Fig. 5(b).The particle size was basically concentrated in the 10-30 nm range, with YC0.1MO having the highest average pore size (29.04 nm).The results of BET analysis showed that the introduction of Ce increased the specific surface area of the catalyst from 29.40 m 2 g -1 to 34.2 m 2 g -1 , and the pore volume increased from 0.21 cm 3 g -1 to 0.274 cm 3 g -1 .On the contrary, the pore size decreased.For catalysts, higher pore volume and surface area can provide more surface-active sites for    17.9 21.9 0.127 acetone adsorption and oxidation, leading to higher catalytic activity.However, the lack of linear correlation between catalytic performance and specific surface area may not be the main influencing factor.

Surface Chemical State
The chemical states of Mn and O elements on the surface of several modified mullite catalysts were investigated by XPS method.The measured spectra of Y1-xCexMn2O5 catalyst were shown in Fig. 6, in which the Mn 2p, Mn 3s and O 1s signals were calibrated by C 1s (284.8 eV).According to previous experimental studies, the XPS spectra of Mn 2p3/2 of YMO and YCMO can be deconvoluted into two main peaks (Biesinger et al., 2011;Weng et al., 2019).The main peak at about 641.9 eV belongs to Mn 3+ , the main peak at about 642.3 eV belongs to Mn 4+ , and the peak at 640.7 eV belongs to Mn 2+ (Zeng et al., 2021).The Mn 2p3/2 and Mn 2p1/2 peaks of the samples shifted to high binding energies after Ce doping, indicating strong interactions between Ce species and mullite components (Liao et al., 2013).The percentage of Mn 3+ and Mn 4+ on the catalyst surface in the Mn 2p region can be calculated from the area of the sub-peaks, and the results are integrated in Table 3.It can be seen that Mn 4+ and Mn 3+ in YC0.1MO account for 49.47 at% and 36.58 at% of the total Mn ions, respectively, which are higher than those of pure YMn2O5.The ratio (1.15) of Mn 4+ /Mn 3+ in the Mn 2p region of YC0.1MO was higher than that of pure mullite phase of 1.09.This phenomenon indicates that Ce doping can increase the average valence of Mn ions in mullite and lead to more Mn 4+ , corresponding to more oxygen vacancies.The abundant oxygen vacancies are beneficial to the replenishment and rapid transfer of gas-phase oxygen, which will greatly improve the catalytic activity of the catalyst.In addition, it can be seen from Table 3 that the size (4.81 eV) of the Mn 3s doublet splitting (∆Es) of YC0.1MO is slightly lower than that of YC0.2MO (4.92 eV) and YC0.4MO (5.07 eV).This means that there is a higher proportion of Mn 4+ on the surface of YC0.1MO (Yang et al., 2015), which is consistent with the results of the Mn 2p2/3 XPS results.
The O 1s XPS spectra of several catalysts were used to further study the effect of Ce doping on the surface oxygen species of mullite, as shown in Fig. 6.The asymmetric O 1s spectra of all catalysts can be fitted to three distinct peaks at 529.2-529.6 eV, 530.9-531.4eV and 532.1-532.8eV, which  are assigned to surface lattice oxygen (Olatt), adsorbed oxygen (Olab) and carbonate (CO3 2-)/hydroxide (OH-) (Jiang et al., 2018;Zhao et al., 2020).The adsorbed oxygen includes O -, O 2-, O2 2-, etc.
Compared with pure YMO, the O 1s peaks of YCxMO are all slightly shifted to the direction of higher binding energy.It can be observed that the most important change in Ce doping is to adjust the ratio of various oxygen species on the catalyst surface.For example, for YC0.1MO, the proportion of lattice oxygen species Olatt dropped from 69% to 65.9%.Meanwhile, the proportion of adsorbed oxygen species Oads increased from 27% to 29%, respectively.Apparently, the addition of Ce promoted the oxygen mobility of the catalyst and converts Olatt to more active Oads.

Oxidation-Reduction Property
H2-TPR was used to evaluate the reduction performance of Y1-xCexMnO3 catalyst, and its distribution is shown in Fig. 7.The reduction curves of the four catalysts all showed three distinct reduction regions in the range of 100-300, 300-500, and 500-700°C.The two peaks at 388°C and 484°C indicated the stepwise reduction of manganese ions: Mn 4+ was reduced to Mn 3+ and then to Mn 2+ (Liu et al., 2012).Compared with the reduction characteristics of pure YMO, the H2-TPR curve of YC0.1MO in Fig. 5 showed two reduction peaks shifted to low temperature at 230°C and 419°C.This may be caused by the interaction of manganese and cerium oxides, where the mobility of oxygen species was greatly enhanced.In addition, the characteristic peak located around 647°C was attributed to the reduction of the remaining deep bulk Mn 3+ .It is well known that the reduction of CeO2 usually occurs at 350-550°C and 800°C (Francesco et al., 2007;Du et al., 2017).Although XRD and TEM analysis have confirmed the presence of CeO2 in the YCMO catalyst, the reduction peak of CeO2 could not be detected.This may be because the consumption of H2 was less in the low cerium concentration state, or its reduction site overlapped with the reduction process of manganese ions.YC0.4MO and YC0.6MO showed the highest reduction temperature because of the large metal oxide crystallites on the catalyst surface.This may lead to agglomeration between particles, which was not conducive to the progress of the catalytic reaction.The SEM image side confirmed this.O2-TPD experiment was further employed to characterize the active oxygen species on the catalyst surface.The results were shown in Fig. 8 and three main oxygen desorption peaks of the YC0.1MO catalyst were detected.The first region was below 210°C, which was attributed to the release of physisorbed oxygen (O2) and weakly chemisorbed oxygen (O2 -or O2 2-).The second region in the range 210-470°C was assigned to surface lattice oxygen species (Olatt).The last region at 470-650°C was related to the bulk oxygen (Obulk) produced by the valence transition of manganese oxides during high temperature treatment, which was closer to the surface lattice oxygen (Mo et al., 2020).Compared with pure mullite, Ce doping significantly increased the proportion of oxygen species adsorbed below 250°C.The lower the temperature was, the easier it was for these reactive oxygen species to participate in the oxidation of acetone at lower temperatures.This was because the introduction of Ce led to a large number of oxygen vacancies, which were good sites for oxygen adsorption and activation during the oxidation reaction.The presence of oxygen vacancies on the surface of YC0.1MO catalyst was beneficial to improve its catalytic activity towards acetone.

Activation Energy
In the presence of sufficient oxygen, the total oxidation of acetone conforms to the first-order reaction mechanism for the concentration of acetone.This also has been mentioned in the catalytic activity evaluation.Based on the activity data in Fig. 1(a) and previous work, the apparent activation energy (Ea) of all catalysts was calculated when the acetone conversion rate was less than 20%.This was due to the fact that the diffusion control, thermal effects and the dependence of the reaction rate on the products of CO2 and H2O can be ignored under such conditions (Zhao et al., 2016;Sun et al., 2019).As shown in Fig. 9, the Ea values of all catalysts increased in the following order 60.71 kJ mol -1 (YC0.1MO)< 63.44 kJ mol -1 (YMO-SG) < 67.53 kJ mol -1 (SMO-SG) < 125.67 kJ mol -1 (SM-MP).YC0.1MO had lower activation energy barrier and therefore had better low temperature catalytic performance.In addition, the reason why it had such excellent performance was also related to other factors.

Mechanism of Acetone Oxidation over Mullite Catalyst
The different catalytic mechanisms are essentially the difference between dominant species in the interconversion process of oxygen species.According to previous literature, the MVK mechanism plays a dominant role in the catalytic oxidation of acetone.According to the above XPS and H2-TPR analysis, Ce doping can effectively improve the adsorption of oxygen concentration and low-temperature reducibility of mullite catalysts.Combined with the above analysis and references (Zhang et al., 2020a), the oxidation path of acetone on YCMO can be proposed (Fig. 10).The catalytic cycle mainly includes the following processes: gaseous acetone interacts with surface lattice oxygen and is adsorbed on the catalyst, and the oxygen vacancies generated during this  process are replaced by gaseous oxygen and disappear.Finally, the adsorbed acetone reacts with the reactive oxygen species to produce both CO2 and H2O.

CONCLUSIONS
In conclusion, mullite catalyst YMn2O5 doped with different Ce (x = 0.1, 0.2, 0.4, 0.6) was prepared successfully, and its catalytic performance for acetone oxidation was investigated.The results show that when x = 0.1, the YC0.1MO catalyst has the best acetone oxidation activity, and the T50 and T90 temperatures are 172°C and 185°C.SEM and TEM showed that the crystal structure of the catalyst was distorted, the crystallinity of mullite decreased, and the corresponding impurities CeO2 and Mn2O3 increased when the Ce doping amount was more than 0.1.The nitrogen adsorption-desorption curve shows that Ce doping can change the specific surface area, and the catalyst has the largest specific surface area (34.2 m 2 g -1 ) when x = 0.1.XPS analysis showed that the Mn 4+ /Mn 3+ ratio on the surface of YC0.1MO mullite catalyst was 1.15 and the Olatt/Oads ratio was 2.27.With the increase of Ce doping content, the mole ratio of Mn 4+ /Mn 3+ decreased and the ratio of Olatt/Oads increased.Considering that the higher the concentration of manganese ions and adsorbent oxygen, the more favorable the oxidation reaction of the manganese-based catalyst is, this may be one of the reasons for the higher acetone oxidation activity of YC0.1MnO catalyst.H2-TPR analysis showed that the reduction temperature of the mixed oxide shifted to the lower temperature region compared with that of the pure mullite.This indicates that the oxides of manganese and cerium are easier to reduce and the mobility of oxygen is greatly improved.The YC0.1MO catalyst also maintained high catalytic activity for at least 48 hours in the activity stability test at 200°C.This study provides a new idea for the rational design of efficient VOC removal catalysts. of hierarchical MnOx/TiO2 composite nanofibers for complete oxidation of low-concentration acetone.J. Hazard. Mater. 337, 105-114. https://doi.org/10.1016/j.jhazmat.2017.03.053

Table 1 .
A catalyst for the catalytic oxidation of acetone.

Table 2 .
Specific surface area and average pore size of Y1-xCexMn2O5.

Table 3 .
Specific surface area and elemental analysis of modified mullite catalysts.