Enhanced catalytic performance of Cu-and/or Mn-loaded Fe-Sep catalysts for the oxidation of CO and ethyl acetate☆

Lisha Liu ,Yong Song ,Zhidan Fu ,Qing Ye ,*,Shuiyuan Cheng ,Tianfang Kang ,Hongxing Dai*

1 Key Laboratory of Beijing on Regional Air Pollution Control,Department of Environmental Science,Beijing University of Technology,Beijing 100124,China

2 Beijing Key Laboratory for Green Catalysis and Separation,Key Laboratory of Beijing on Regional Air Pollution Control,Key Laboratory of Advanced Functional Materials,Education Ministry of China,Laboratory of Catalysis Chemistry and Nanoscience,Department of Chemistry and Chemical Engineering,College of Environmental and Energy Engineering,Beijing University of Technology,Beijing 100124,China

1.Introduction

Volatile organic compounds(VOCs,e.g.,toluene and ethylacetate)and CO are environmental pollutants.Most of VOCs are toxic,malodorous,mutagenic,and carcinogenic in nature,and the precursors of ozone and smog[1].Hence,it is highly desired to remove CO and VOCs[1,2].Catalytic combustion is one of the most effective pathways for the removal of VOCs.Supported precious metals have been reported to be active in oxidation reactions[3].Although precious metal catalysts(e.g.,Au,Pt,and Pd)exhibit satisfactory catalytic activities at low temperatures[4-6],their applications are limited due to the high cost,susceptibly poisoning tendency,and sintering.Therefore,there is a need to develop highly active and stable catalysts.Compared with supported noble metal catalysts,the transition-metal(Cu,Ni,Co,and Mn)oxide catalysts have the advantages of easy availability and low cost,among which the Cu-and Mn-based oxides are the potential candidates for catalytic oxidation[7,8].Gongetal.[9]investigated the copper-manganese oxide for CO oxidation,and pointed out that a synergistic effect could exist in the amorphous Cu-Mn composite,resulting in the improvement in activity,as compared with the individual oxide.

Generally speaking,the activity of a catalyst depends upon the active component and physical property of the support used,such as crystal structure,chemical composition,surface area,and thermal stability.Therefore,it is critical to develop a suitable support for the Cu-Mnbased catalysts in order to enhance their catalytic performance.In one of our previous works[10],much attention was paid on the selection of an appropriate support with good hydrothermal stability.Sepiolite is a cheap non-layered mineral clay built by two tetrahedral silica sheets with a central magnesia sheet(a 2:1 layered structure),and its ideal formula is[Si12Mg8O30(OH)4(H2O)4]·8H2O[11].Compared to other types of silicate minerals,sepiolite exhibits a higher adsorbent capacity and better stability.Such excellent behaviors render it to be useful in adsorption and catalysis[10,11].For example,Bautistaet al.[12]used sepiolite as a support to generate the Ni/Cu-sepiolite catalysts for the liquid-phase selective hydrogenation of fatty acid ethyl esters.Linet al.[13]claimed that the sepiolite-supported Co and Fe catalysts showed high performance in the high temperature gas flow treatment of dye.These results indicate that sepiolite could be used as support in a catalyst.

In this paper,the Fe-exchanged sepiolite was modified by the Cu-Mn mixed oxides and the obtained materials were used to catalyze the oxidation of CO and ethyl acetate.Structural features of the composites were systematically characterized.It was found that the synergic action between the Cu-Mn species and Fe-exchanged sepiolite gave rise to better low-temperature reducibility and oxygen mobility,which was responsible for the improvement in catalytic activity of CO and ethyl acetate oxidation.The most significant finding of the present work is the development of low cost catalyst as a replacement of supported noble metals.

2.Experimental

2.1.Catalyst preparation

2.1.1.Iron-exchange sepiolite

The natural sepiolite sample used in this study was obtained from Hunan Hongyan Sepiolite Company,China.Its chemical composition was 75.0 wt%SiO2,8.05 wt%Al2O3,8.71 wt%MgO,3.54 wt%Fe2O3,1.17 wt%K2O,0.51 wt%CaO,and other trace compounds,which was determined using the X-ray fluorescence(XRF)spectroscopic technique.The original sepiolite was purified according to the procedures described elsewhere[14].The support was prepared using the sodium-leached sepiolite.The leaching treatment was carried out in a 2 wt%sepiolite aqueous solution by adding Na2CO3under the sodium/sepiolite mass ratio of 4%at 60°C for 2 h.The sodium-treated sepiolite was in turn aged at room temperature for 24 h,filtered,washed with deionized water,and dried at 80°C for 24 h.The obtained sodium-treated sepiolite support was denoted as Sep.

The Fe-Sep was preparedviaan ion exchange route using Sep as support and iron nitrate as iron source.3.0 g of Sep was dispersed in 300 ml of deionized water and the slurry was stirred for24 h.The ion-exchange solution was prepared by dropwise adding a NaOH solution(0.2 mol·L-1)into a Fe(NO3)3aqueous suspension(0.1 mol·L-1)under stirring at room temperature for complete mixing.Then,the solution was aged at 70°C for 24 h.The obtained solution was dropwise added to the vigorously stirred Sep-containing suspension until the Fe/gSepratio was achieved 10 mmol·g-1.The resulting solution was stirred at 70 °C for 24 h,washed,dried at 100 °C for 12 h,and calcined at 400°C for 2 h.The obtained sample was denoted as Fe-Sep.

2.1.2.Fe-exchanged sepiolite-supported copper and manganese mixed oxides

The CuxMny/Fe-Sep samples were prepared by a co-precipitation method with a Cu loading of 5 wt%,as described elsewhere[8].Firstly,the desired amounts of Cu(CH3COO)2·4H2O and/or Mn(CH3COO)2·4H2O were dissolved in 80 ml of deionized water at80°C under stirring for 30 min.Secondly,the certain amount of Fe-Sep was added into the above solution and thoroughly stirred at room temperature for 30 min.Thirdly,an appropriate amount of Na2CO3aqueous solution(0.05 mol·L-1)was dropwise added until the pH of the solution reached ca.8.3,followed by aging at 80°C for 12 h.Finally,the mixture was washed with deionized water,dried at 120°C for 24 h,and calcined at 450°C for 2 h.The obtained products are denoted as CuxMny/Fe-Sep,whereas the“x”and“y”representthe Cu/Mn molar ratio(x/y=0:2,1:0,1:0.5,1:2 and 1:4)in the samples,respectively.

2.2.Catalyst characterization

The crystal phases of the support and catalysts were measured by the powder X-ray diffraction(XRD)technique on a Bruker D8 advance powder diffractometer using CuKαradiation(λ=0.15406 nm)at a scan ramp of 0.02(°)·s-1.BET surface areas of the samples were determined using N2adsorption-desorption at-196°C(JW-BK132F).Prior to the BET measurements,each sample was degassed at 200°C for 6 h under vacuum conditions.X-ray photoelectron spectroscopic(XPS)spectra of the samples were determined on an ESCALAB 250 apparatus,using Al Kα radiation as excitation source.The signal from adventitious carbon(binding energy=284.6 eV)was used for binding energy calibration.The chemical analysis of the samples was carried out on the energy-dispersive X-ray fluorescence(EDXRF)spectrometer(Panalitical Magix,PW2403)with a rhodium tube as radiation source.H2temperature-programmed reduction(H2-TPR)profiles of the samples were obtained on a chemical adsorption analyzer(PCA-1200).In each test,about 100 mg of the sample was placed in a U-shaped quartz tube,and then pretreated in a flow (30 ml·min-1) of 5 vol%H2+95 vol%N2at a ramp of 10 °C·min-1from room temperature to 900°C.The CuO sample(Aldrich,99.995%)was used as a calibrated standard sample.

Oxygen temperature-programmed desorption(O2-TPD)experiments were carried out on a PCA-1200 apparatus.About 100 mg of the sample was pretreated in pure helium(20 mL·min-1)from room temperature to 350°C for 1 h and subsequently cooled to room temperature.Anhydrous O2(10 vol%O2/He,20 mL·min-1)was adsorbed on the sample at room temperature for 1 h.After O2adsorption,the sample was flushed with pure helium(20 mL·min-1)for 1 h at room temperature.The temperature was increased linearly at a ramp of 10 °C·min-1from 30 to 900 °C under a helium flow of 20 mL·min-1.

2.3.Catalytic activity measurement

Catalytic activity measurements of the samples for the oxidation of CO and ethyl acetate were carried out in a continuous flow fixed-bed quartz tubular microreactor at atmospheric pressure.A certain amount of catalyst(100 mg)and equal amount of quartz sands were well mixed and put into the microreactor.In the case of CO oxidation,the feed gas composition was 1.0 vol%CO and air(balance),the total flow rate was 50 ml·min-1,and the corresponding space velocity(SV)was 15000 ml·g-1·h-1.In the case of ethyl acetate(which was chosen as the representative VOCs)oxidation,a N2flow of 2.53 ml·min-1was passed through an ethyl acetate-containing tubular saturator at 30°C,and then mixed with an air flow,thus giving a total flow rate of 200 ml·min-1and a SV of 20000 ml·g-1·h-1.In the feed gas mixture,the concentration of ethyl acetate was 2000 ppm.The reaction products were analyzed online by a Shimadzu GC-14C gas chromatograph equipped with a thermal conductivity detector(TCD)for CO analysis or by a Techcomp GC-7900 gas chromatograph equipped with a flame ion detector(FID)for ethyl acetate concentration analysis.The conversion of CO or ethyl acetate was calculated according to the changes of CO or ethyl acetate concentrations in the inlet and outlet gas mixtures:

Fig.1.XRDpatterns of(a)Sep,(b)Fe-Sep,(c)Mn2/Fe-Sep,(d)Cu1/Fe-Sep,(e)Cu1Mn0.5/Fe-Sep,(f)Cu1Mn2/Fe-Sep,(g)Cu1Mn4/Fe-Sep,and(h)Cu1Mn2/Sep.

3.Results and Discussion

3.1.Crystal composition

The XRD patterns of the Sep and Cu x Mn y/Fe-Sep samples are shown in Fig.1.Obviously,the main characteristic signals in each sample were observed at 2θ=20.6°,28.0°,35.2°,40.4°,and 50.2°,which were the characteristic diffraction peaks of sepiolite[10].This result suggests that the sepiolite structure was present in all of the samples.In the Cu x Mn y/Fe-Sep samples,there were additional diffraction peaks at 2θ =33.5°and 35.5°due to the(104)and(110)planes of Fe2O3(JCPDS PDF#84-0307)[15],respectively.This result demonstrates the existence of Fe2O3in these samples.Meanwhile,no additional diffraction peaks due to the phases of copper oxides were observed in the Cu x Mn y/Fe-Sep samples,indicating that the Cu species were highly dispersed on the surface of Fe-Sep.However,the Mn2/Fe-Sep exhibited a diffraction peak at2θ=54.1°(Fig.1c)assignable to the(211)plane of MnO2(JCPDS PDF#82-2169).Moreover,there was a diffraction signal at 2θ=29.8°due to a cubic CuMn2O4phase(JCPDS PDF#70-0260)in the Cu x Mn y/Fe-Sep samples(Fig.1e-h).A similar result was reported by Fang et al.[16]who prepared copper and manganese mixed oxides supported on TiO2after calcination at 400°C.

3.2.N2 adsorption-desorption isotherm and surface area

Fig.2 shows the N2adsorption-desorption isotherms of the Sep and CuxMny/Fe-Sep samples.All of the samples exhibited a N2adsorptiondesorption isotherm of type I in the low relative pressure range,suggesting the existence of a micropore structure.In the region of high relative pressure,the adsorption-desorption isotherm displayed type IV,indicating the presence of mesopores.Moreover,there was a type H3 hysteresis loop,implying the formation of slit-shaped pores in these samples.

Fig.2.N2 adsorption-desorption isotherms of(a)Sep,(b)Fe-Sep,(c)Mn2/Fe-Sep,(d)Cu1/Fe-Sep,(e)Cu1Mn0.5/Fe-Sep,(f)Cu1Mn2/Fe-Sep,(g)Cu1Mn4/Fe-Sep,and(h)Cu1Mn2/Sep.

BET surface areas of the Sep and CuxMny/Fe-Sep samples are summarized in Table 1.It can be observed that the Fe-Sep sample possessed a surface area(89.3 m2·g-1)much higher than that(60 m2·g-1)of the raw Sep sample.As compared with the Fe-Sep sample,however,surface areas of the CuxMny/Fe-Sep samples decreased to 71.2-87.9 m2·g-1,probably due to the partial blocking of the pores by the Cu and/or Mn oxide species[17].The surface area decreased in the sequence of Fe-Sep＞Mn2/Fe-sep＞Cu1/Fe-Sep＞Cu1Mn2/Fe-Sep＞Cu1Mn0.5/Fe-Sep＞Cu1Mn4/Fe-Sep＞Cu1Mn2/Sep,which was different from the order in catalytic activity(shown below).This result indicates that the BET surface area did not solely determine the catalytic performance of the samples for CO oxidation.

The SEM technique was used to determine the morphologies of the Fe-Sep,Mn2/Fe-Sep,Cu1/Fe-Sep,and Cu1Mn2/Fe-Sep samples,as shown in Fig.3A-D.It can be seen that all of the samples exhibited a fibrous morphology with a smooth surface,which was typical sepiolite clay.There was no noticeable morphological difference between the natural sepiolite and the supported samples,although surface areas of the resulting catalysts were different after loading of Cu and/or Mn on the Fe-modified sepiolite.

3.3.XPS characterization

To investigate the surface element compositions and metal oxidation states of the samples,the Mn 2p,Cu 2p,and O 1s XPS spectra and quantitative analysis results of the samples are shown in Fig.4 and Table 2,respectively.The Mn 2p XPS spectrum(Fig.4A)of each sample displays two main asymmetric peaks at ca.642.0 and 654.0 eV,which corresponded to the Mn 2p3/2and Mn 2p1/2,respectively.The Mn 2p3/2signal could be decomposed into two components at binding energy(BE)=ca.641.5 and ca.643.5 eV,assignable to the surface Mn3+and Mn4+species[18],respectively.This result indicates that each of the CuxMny/Fe-Sep samples contained a mixed-valence manganese species.The Mn4+/Mn3+atomic ratios of the samples are summarized in Table 2.Obviously,the Mn4+/Mn3+atomic ratio was remarkably in fluenced by the surface Mn/Cu molar ratio,and it decreased in the order of Cu1Mn2/Fe-Sep(0.86)＞Cu1Mn4/Fe-Sep(0.70)＞Cu1Mn0.5/Fe-Sep(0.62)＞Mn2/Fe-Sep(0.48),with the highest surface Mn4+/Mn3+atomic ratio being achieved on the Cu1Mn2/Fe-Sep(0.86)sample.In other words,the Cu1Mn2/Fe-Sep sample possessed the highest surface Mn4+concentration.As compared with the Cu1Mn2/Fe-Sep sample,the Fe-free Cu1Mn2/Sep exhibited a lower surface Mn4+/Mn3+atomic ratio(0.68).Moreover,the Mn2/Fe-Sep showed the lowest surface Mn4+/Mn3+atomic ratio(0.48),as compared with those of the CuxMny/Fe-Sep samples.Therefore,we conclude that the strong interaction between the Cu or Mn species and the Fe-Sep support also probably exerted a remarkable influence on surface Mn4+/Mn3+atomic ratio,which might be one of the main reasons for the enhanced catalytic activity.

The Cu 2p3/2XPS spectra of the samples are presented in Fig.4B.For each sample,the XPS spectrum exhibited two asymmetric signals thatcould be decomposed into three components:the one at BE=933.2-934.6 eV and the other at BE=935.6-936.2 eV as well as a satellite at BE=943.4-944.0 eV.According to the literature[19,20],the peak at a BE higher than 933.1 eV was one important XPS characteristic of CuO clusters and isolated Cu2+species.Moreover,the isolated Cu2+species could be distinguished from the copper oxide(CuO)by the BEs since the BE of the isolated Cu2+species was higher than that of the copper oxide(CuO).Hence,the peak at BE=933.2-934.6 eV could be assigned to the CuO cluster species,whereas the peak at BE=935.6-936.2 eV corresponded to the isolated Cu2+species,and the satellite at BE=943.4-944.0 eV was indicative of the Cu2+species.As shown in Table 2,the Cu2+/CuO atomic ratio was also remarkably influenced by the surface Mn/Cu molar ratio,and it decreased in the order of Cu1Mn2/Fe-Sep(2.03)＞Cu1Mn4/Fe-Sep(1.77)＞Cu1Mn0.5/Fe-Sep(1.51)＞Cu1/Fe-Sep(0.79).Such a change trend was almost similar to that in surface Mn4+/Mn3+atomic ratio.Obviously,the surface Cu2+/CuO atomic ratio(2.03)on Cu1Mn2/Fe-Sep was the highest,whereas that(0.79)on Cu1/Fe-Sep was the lowest,i.e.,the Cu1Mn2/Fe-Sep sample contained the highest amount of isolated Cu2+species and the Cu1/Fe-Sep sample possessed the lowest amount of isolated Cu2+species.It means that most of the copper species in the CuxMny/Fe-Sep samples were present as lattice Cu2+,but only a fraction of the copper species was weakly bound CuO clusters.This could be due to the interaction between Cu and Mn species in the Cu-Mn-loaded samples[10],as confirmed by the results of H2-TPR investigations(shown below).Moreover,it is obvious that compared with the Cu2+/CuO molar ratio(1.62)on Fe-free Cu1Mn2/Sep,that(2.03)of the Cu1Mn2/Fe-Sep sample was higher,suggesting that a stronger interaction would occur between the Cu or Mn species and the Fe-Sep support in the Cu1Mn2/Fe-Sep sample[10].

Table 1BET surface areas,reduction peak temperatures,and H2 consumption of the Cu x Mn y/Fe-Sep samples

Fig.3.SEM images of(A)Fe-Sep,(B)Mn2/Fe-Sep,(C)Cu1/Fe-Sep,and(D)Cu1Mn2/Fe-Sep.

Fig.4.(A)Cu 2p,(B)Mn 2p,and(C)O 1s XPS spectra of(a)Sep,(b)Fe-Sep,(c)Mn2/Fe-Sep,(d)Cu1/Fe-Sep,(e)Cu1Mn0.5/Fe-Sep,(f)Cu1Mn2/Fe-Sep,(g)Cu1Mn4/Fe-Sep,and(h)Cu1Mn2/Sep.

Table 2Binding energies(eV)and surface atomic ratios determined by quantitative XPS analysis of the Cu x Mn y/Fe-Sep samples

The O1s XPS spectra of the samples are illustrated in Fig.4C.Only one signal at BE=532.8 eV of the raw Sep(Fig.3C(a))was observed,which was consistent with the O1s signal at BE=532.5±0.2 eV due to the Si-O bond in SiO2[21].It is worth noting that the weight content of SiO2was as high as 75.0%in the raw Sep sample(XRF result).Therefore,the signal at BE=532.8 eV was ascribable to the surface lattice oxygen on SiO2of the raw Sep sample.On all of the Cu-and/or Mn-loaded samples,there were two asymmetric signals that could be decomposed into three components:the one at BE=529.7-529.8 eV was ascribed to the surface lattice oxygen(Olatt)species of Cu and/or Mn oxides,whereas the one at BE=530.5-530.8 eVwas due to the surface lattice oxygen species of sepiolite and surface adsorbed oxygen(Oads,e.g.,,or O-formed in the oxide defects)species[22].The formation of surface oxygen species was due to the presence of surface oxygen vacancies on CuxMny/Fe-Sep,which implies that there might be the coexistence of Mn3+and Mn4+species in/on the CuxMny/Fe-Sep samples.Such a deduction was supported by the Mn 2p XPS results of these samples.Due to the overlapping of surface lattice oxygen species of sepiolite and surface adsorbed oxygen species on the CuxMny/Fe-Sep samples,it is hence difficult to isolate them accurately.Compared with the Fe-free Cu1Mn2/Sep sample,CuxMny/Fe-Sep exhibited a lower BE of oxygen,implying that the existence of the interaction between the Cu or Mn species and the Fe-Sep support might be favorable for the mobility of oxygen species.

3.4. O2-TPD characterization

Fig.5.O2-TPD profiles of(a)Fe-Sep,(b)Mn2/Fe-Sep,(c)Cu1/Fe-Sep,(d)Cu1Mn0.5/Fe-Sep,(e)Cu1Mn2/Fe-Sep,(f)Cu1Mn4/Fe-Sep,and(g)Cu1Mn2/Sep.

O2-TPD experiments were carried out to gain further insights into the nature of surface oxygen species of the samples.Fig.5 shows the O2-TPD pro files of the samples.For the Fe-Sep and Cu1/Fe-Sep samples,no obvious oxygen desorption was detected in the temperature range of 100-1000°C(Fig.5a and c).For the Mn2/Fe-Sep and MnxCuy/Fe-Sep samples,there were two prominent desorption peaks:the first desorption peak at a medium temperature(570-660°C)came from the lattice layer close to the surface,whereas the desorption peak at a high temperature(790-920°C)was due to the evolution of the bulk lattice oxygen from the sample[22].As compared with the Mn2/Fe-Sep sample(Fig.5b),the CuxMny/Fe-Sep samples showed distinctive desorption at lower temperatures(Fig.5d-f).Additionally,the Mn/Cu molar ratio remarkably affected the desorption temperature of oxygen species on the sample.Generally,more attention should be paid on the first desorption peak since it could be desorbed at relatively low temperatures and was likely to participate in the oxidation of reactants[23].The temperature for the low-temperature desorption of oxygen species increased in the order of Cu1Mn2/Fe-Sep＜Cu1Mn4/Fe-Sep＜Cu1Mn0.5/Fe-Sep＜Mn2/Fe-Sep.Evidently,the temperature(570°C)for the release of oxygen species from the Cu1Mn2/Fe-Sep sample was the lowest among all of the samples,suggesting that the Cu1Mn2/Fe-Sep sample possessed the highest mobility of oxygen species.

To investigate the effect of Fe doping on the oxygen species of the Cu-and Mn-loaded samples,we prepared the Fe-free Cu1Mn2/Sep sample.Compared with the temperature(620°C)for desorption of oxygen species from the Cu1Mn2/Sep sample(Fig.5g),the temperature(570°C)for desorption of oxygen species from the Fe-free Cu1Mn2/Fe-Sep sample was lower.Hence,it is likely to have a strong interaction between the Cu or Mn species and the Fe-Sep support,which might facilitate the oxygen mobility on/in the CuxMny/Fe-Sep samples.

3.5.H2-TPR characterization

Fig.6.H2-TPR profiles of(a)Fe-Sep,(b)Mn2/Fe-Sep,(c)Cu1/Fe-Sep,(d)Cu1Mn0.5/Fe-Sep,(e)Cu1Mn2/Fe-Sep,(f)Cu1Mn4/Fe-Sep,and(g)Cu1Mn2/Sep.

Fig.7.Catalytic activity of the CuxMny/Fe-Sep samples for the oxidation of(A)CO at SV=60000 ml·g-1·h-1 and(B)ethyl acetate at SV=240000 ml·g-1·h-1.

H2-TPR is an effective technique to evaluate the reducibility of a catalyst.Fig.6 illustrates the H2-TPR profiles of the samples,and the peak positions and quantitative analysis results are summarized in Table 1.For the Fe-Sep sample(Fig.6a),there were three asymmetrical reduction peaks at 450,520,and 780°C,respectively.As suggested by Oliveiraet al.[24],these peaks could be attributed to the sequential reduction of Fe2O3→Fe3O4or even Fe0,Fe3O4→FeO,and FeO→Fe0,respectively.A noteworthy phenomenon was that the H2consumption(0.97 mmol·g-1)of the third peak was higher than the total H2consumption of the first peak(0.32 mmol·g-1)and the second peak(0.53 mmol·g-1).This result indicates that the reduction of the third peak(780°C)might include the simultaneous reduction of Fe2O3→Fe0and FeO→Fe0.For the Cu1/Fe-Sep(Fig.6c)and Mn2/Fe-Sep(Fig.6b)samples,there were several reduction peaks at 420-440,550-600,650-720,and 780°C.According to the literature[25],the lower-temperature reduction peaks at 420 and 550°C for the Cu1/Fe-Sep sample were assigned to the reduction of isolated Cu2+species and highly dispersed surface oxygenated CuO clusters,respectively.Notably,the Cu2+/CuOmolar ratio was close to 0.8,which was confirmed by the result of the Cu 2p XPS study.For the Mn2/Fe-Sep sample,the lower-temperature reduction peaks at 420 and 600°C were due to the reduction of Mn3+and Mn4+species[9],respectively.According to the above result of the Fe-Sep catalyst and that reported by Putluruet al.[26],the iron species were not easily reduced below 500°C.Therefore,it is reasonable to assign the peaks at 650 and 780°C for the Cu1/Fe-Sep sample and at 720°C for the Mn2/Fe-Sep sample to the reduction of Fe2O3to Fe3O4and Fe3O4to Fe,respectively,in which the temperatures were obviously lower than those for the Fe-Sep sample,possibly due to the synergistic effect between the Fe and Mn or Cu species.

For the CuxMny/Fe-Sep samples(Fig.6(d-f)),the TPR profile shows H2consumption peaks and reduction temperature distinctively different from those for the Cu1/Fe-Sep and Mn2/Fe-Sep samples,there were four peaks at 250,280,340,and 650°C,which are denoted as peaksα,β,γandλ,respectively.According to the XRD analysis results and the above discussion as well as the literature work[9],the α peak could be assigned to the reduction of the Cu species,and the β peak could be attributed to the redox interaction between the dispersed Cu and Mn species in the CuxMny/Fe-Sep samples.The γ peak at 340 °C could be due to reduction of the Mn species,and the λ peak was the reduction of the Fe species.Importantly,it can be found that,compared with the Mn2/Fe-Sep or Cu1/Fe-Sep sample,the CuxMny/Fe-Sep samples shifted the reduction temperatures to lower values and showed a remarkable increase in H2consumption of the α,β,and γ peaks,further confirming the existence of a synergistic effect between the Cu and Mn species.The low-temperature peaks α and β were active and could improve the catalytic performance of the samples,therefore more attention was paid to these peaks.The low-temperature reducibility increased in the order of Cu1Mn2/Fe-Sep(220°C)＜ Cu1Mn4/Fe-Sep(240°C) ＜Cu1Mn0.5/Fe-Sep(270°C)＜Mn2/Fe-Sep(420°C)＜Cu1/Fe-Sep(440 °C) ＜ Fe-Sep(460 °C).Especially,compared with the other samples,the α peak for the Cu1Mn2/Fe-Sep sample exhibited the lowest reduction temperature(220°C),i.e.,the Cu1Mn2/Fe-Sep catalyst possessed the best low-temperature reducibility,which might result from the interaction of the manganese or copper and iron species in this sample[27].

Moreover,the reduction temperature for the Cu1Mn2/Fe-Sep sample(220°C)shifted to a lower temperature,as compared with the reduction temperature(260°C)for the Fe-free Cu1Mn2/Sep sample(Fig.6g).Hence,it is likely to have a strong interaction between the Cu or Mn species and the Fe-Sep support,facilitating the reduction of the CuxMny/Fe-Sep samples.

3.6.Catalytic performance

Fig.7A and B shows the catalytic activities of the CuxMny/Fe-Sep samples for the oxidation of CO and ethyl acetate,respectively.It is convenient to use theTx(temperatures at ax%conversion)to evaluate the catalytic activity,as summarized in Table 3.The blank experiments(only quartz sands)exhibited that there were no significant conversions of CO and ethyl acetate below 200 and 400°C,respectively.In other words,obvious homogeneous reactions of CO and ethyl acetate with oxygen did not take place under the adopted reaction conditions.

Table 3Comparison of specific reaction rates(r cat)for CO oxidation at 100 °C and ethyl acetate oxidation at 180 °C,and T50 and T100 over the Cu x Mn y/Fe-Sep samples and the related samples reported in the literature

As shown in Fig.7,the Fe-Sep sample exhibited the lowest catalytic activity and only 30 and 8%conversions of CO and ethyl acetate oxidation were achieved at 250°C,respectively.Compared with the Fe-Sep sample,a distinct decrease in reaction temperature was observed after the doping of Cu and/or Mn to Fe-Sep.For example,over the Cu1/Fe-Sep or Mn2/Fe-Sep samples,theT50andT90were 210-250 and 230-250 °C for CO oxidation and 235-250 and 300-320 °C for ethyl acetate oxidation,respectively.Moreover,as compared with the Cu1/Fe-Sep and Mn2/Fe-Sep samples,the CO and ethyl acetate conversions significantly increased over the CuxMny/Fe-Sep samples.This result indicates that there was a synergistic effect between the Cu and Mn species in the CuxMny/Fe-Sep samples,leading to significantly enhanced activity for CO and ethyl acetate oxidation,especially at low temperatures.Furthermore,catalytic activity of the CuxMny/Fe-Sep samples depended on the Mn/Cu molar ratio.The most active catalyst was Cu1Mn2/Fe-Sep,over which theT50andT90were 110 and 140°C for CO oxidation and 170 and 210°C for ethyl acetate oxidation,respectively.In order to determine the important role of the Fe species in the samples,the Fe-free Cu1Mn2/Sep sample was prepared.The Cu1Mn2/Sep sample showed a lower catalytic activity than the Cu1Mn2/Fe-Sep sample,with theT50andT90being 130 and 160 °C for CO oxidation and 180 and 240 °C for ethyl acetate oxidation over the former sample,respectively.Therefore,the Fe species were beneficial for enhancing catalytic activity of the Cu-Mn-loaded samples for CO and ethyl acetate oxidation.

It is noticeable that catalytic activity of a sample is usually associated with several factors,including surface area,reducibility,and oxygen species.It has been generally accepted that surface area of a catalyst plays an important role in the oxidation of CO and ethyl acetate,and a higher surface area would show a better catalytic activity.In the present study,however,it seems that the Cu1Mn2/Sep sample with the lowest surface area(71.0 m2·g-1)showed higher activity than the Mn2/Fe-Sep sample with the highest surface area(89.3 m2·g-1),indicating that the surface area was not the main factor influencing the activity.The co-loading of Cu and Mn resulted in a significant enhancement in CO or ethyl acetate conversion,demonstrating that the Cu-Mn species exerted a dominant influence on catalytic activity.

According to one of our previous studies[4],the specific reaction rate(rcat)can reflect the inherent catalytic activity.Thercatis defined as the molar amount of CO or ethyl acetate converted per gram catalyst and per second(mol·(g cat)-1·s-1).Thercatvalues at100 and 180 °C for CO and ethyl acetate oxidation were summarized in Table 3.Thercatvalues for CO oxidation at 100°C(2.9× 10-7-4.4× 10-7mol CO·(g cat)-1·s-1)over Fe-Sep,Cu1/Fe-Sep,and Mn2/Fe-Sep were the same order of magnitude as those(7.2× 10-8-1.09× 10-6mol CO·(g cat)-1·s-1)at 100 °C over MnxOy[28],but lower than those(6.8 ×10-7-1.5 × 10-6mol CO·(g cat)-1·s-1)at 100 °C over MnOx[29].Co-loading of Mn and Cu to Fe-Sep significantly increased the catalytic activity.Thercatvalues(9.9× 10-7-4.4× 10-6mol CO·(g cat)-1·s-1)for CO oxidation at 100 °C over CuxMny/Fe-Sep were much higher than those(3.1 × 10-7mol CO·(g cat)-1·s-1)at 100°C over PdFeMn/Cord[30].Thercatdecreased in the order of Cu1Mn2/Fe-Sep＞Cu1Mn4/Fe-Sep＞Cu1Mn0.5/Fe-Sep＞Mn2/Fe-Sep＞Cu1/Fe-Sep＞Fe-Sep,in good agreement with the change trend in activity.Moreover,one can see that for the combustion of ethyl acetate,thercatvalues over Fe-Sep,Cu1/Fe-Sep and Mn2/Fe-Sep were 1.4 × 10-7-2.8 × 10-7mol ethyl acetate·(g cat)-1·s-1,and those over CuxMny/Fe-Sep were 5.5×10-7-1.9×10-6mol ethyl acetate·(g cat)-1·s-1.All of thesercatvalues were much higher than those(2.3 × 10-8-9.1 × 10-8mol ethyl acetate·(g cat)-1·s-1)over CuOx-CeO2/Al2O3and CuMn2O4/Al2O3[31].Obviously,thercatvalues for ethyl acetate oxidation decreased in the sequence of Cu1Mn2/Fe-Sep＞Cu1Mn4/Fe-Sep＞Cu1Mn0.5/Fe-Sep＞Mn2/Fe-Sep＞Cu1/Fe-Sep＞Fe-Sep,which was also in good agreement with the change trend in catalytic activity.

Generally speaking,the total oxidation of CO and ethyl acetate over transition-metal oxides follows the redox Mars-van Krevelen mechanism.Liuet al.[32]proposed an active assembly formed by the redox couple of Cu2+/Cu+,Mn4+/Mn3+,and O2-species for CO oxidation.The key step of this mechanism was the supply of active oxygen species by the readily reducible oxide and the lattice oxygen:

where Vorepresents an oxygen vacancy.

The highly reducible bridged oxygen in the Cu2+-O2--Mn4+entity can first react with CO to produce CO2and a partially reduced Cu+-Vo-Mn3+couple.Then,the Cu+-Vo-Mn3+entities can be rapidly reoxidized by O2in the feed gas.During the reaction,CO molecules are adsorbed and activated into COadson the active metal sites(Cu2+-O2--Mn4+),whereas O2molecules are adsorbed and activated into Oadson oxygen vacancies at the catalyst surface(2Cu+-Vo-Mn3++O2→2Cu2+-O2--Mn4+).Finally,the adsorbed COadsand O2adsare easily transformed into CO2.Therefore,performance of the catalyst may be associated with its reducibility and oxygen mobility and adsorption of CO and O2.Moreover,it is well known that the oxidation of organic compounds also proceedsviaa Mars-van Krevelen mechanism[33],in which the key step is the supply of active oxygen species by the readily reducible oxide and the lattice oxygen.Since the Mars-van Krevelen mechanism involves in the exchange of oxygen between the lattice and gas phase and the oxidation of feed gas,oxygen mobility and reducibility of the catalyst are also important in catalytic oxidation of organic compounds.

As discussed above,the temperatures of reduction peaks in H2-TPR profiles and of desorption peaks in O2-TPD profiles were closely related to the reducibility and mobility of oxygen species of the samples.As shown in Fig.7 and Table 3,the catalytic activity for CO and ethyl acetate oxidation decreased in the order of Cu1Mn2/Fe-Sep＞Cu1Mn4/Fe-Sep＞Cu1Mn0.5/Fe-Sep＞Mn2/Fe-Sep＞Cu1/Fe-Sep＞Fe-Sep,which was in good accordance with those in reducibility and oxygen mobility of the samples.Obviously,the Cu1Mn2/Fe-Sep sample performed the best in the oxidation of CO and ethyl acetate,coinciding with its best reducibility and highest oxygen mobility.That is to say,the higher reducibility and mobility gave rise to a higher catalytic activity.Therefore,we conclude that catalytic performance of the CuxMny/Fe-Sep samples for CO and ethyl acetate oxidation was associated with the strong interaction between the Cu or Mn species and the Fe-Sep support,good low-temperature reducibility,and high oxygen mobility.

4.Conclusions

The CuxMny/Fe-Sep catalysts with different Mn/Cu molar ratios could be prepared using the co-precipitation method.Catalytic activities of the Cu-Mn-loaded samples were higher than those of the Cu or Mn-loaded sample for the oxidation of CO and ethyl acetate,and the catalytic activity decreased in the order of Cu1Mn2/Fe-Sep＞Cu1Mn4/Fe-Sep＞Cu1Mn0.5/Fe-Sep＞Mn2/Fe-Sep＞Cu1/Fe-Sep＞Fe-Sep.The Cu1Mn2/Fe-Sep catalyst exhibited the highest specific reaction rate and the lowestT50andT90of 4.4 × 10-6mmol·g-1·s-1,110,and 140 °C for CO oxidation,and 1.9 × 10-6mmol·g-1·s-1,170,and 210°C for ethyl acetate oxidation,respectively.According to the characterization results,we conclude that the excellent catalytic performance of Cu1Mn2/Fe-Sep was associated with the strong interaction between the Cu or Mn species and the Fe-Sep support,good low-temperature reducibility,and high mobility of chemisorbed oxygen species.In addition,the most significant finding of the present work is the possible development of low-cost catalysts as a replacement of noble metalbased catalysts.

[1]Z.B.Rui,C.Y.Chen,Y.B.Lu,H.B.Ji,Anodic alumina supported Pt catalyst for total oxidation of trace toluene,Chin.J.Chem.Eng.22(2014)882-887.

[2]T.O.Wubah,M.M.Du,W.B.Osei,D.H.Sun,J.L.Huang,Q.B.Li,Facile synthesis of porous Pd nano flowers with excellent catalytic activity towards CO oxidation,Chin.J.Chem.Eng.23(2015)1907-1915.

[3]N.An,P.Wu,S.Li,M.Jia,W.Zhang,Catalytic oxidation of formaldehyde over Pt/Fe2O3catalysts prepared by different method,Appl.Surf.Sci.285(2013)805-809.

[4]Q.Ye,J.Zhao,F.Huo,D.Wang,Nanosized Au supported on three-dimensionally ordered mesoporous β-MnO2:Highly active catalysts for the low-temperature oxidation of carbon monoxide,benzene,and toluene,Microporous Mesoporous Mater.172(2013)20-29.

[5]A.Manasilp,E.Gulari,Selective CO oxidation over Pt/alumina catalysts for fuel cell applications,Appl.Catal.B Environ.37(2002)17-25.

[6]M.F.Garc,A.M.N.Arias,A.I.Juez,A.B.Hungr,J.A.Anderson,New Pd/CexZr1-xO2/Al2O3three-way catalysts prepared by microemulsion:Part 1.Characterization and catalytic behavior for CO oxidation,Appl.Catal.B Environ.31(2001)39-50.

[7]E.C.Njagi,C.Chen,H.Genuino,H.Galindo,H.Huang,Total oxidation of CO at ambient temperature using copper manganese oxide catalysts prepared by a redox method,Appl.Catal.B Environ.99(2010)103-110.

[8]J.Li,P.Zhu,R.Zhou,Effect of the preparation method on the performance of CuOMnOx-CeO2catalysts for selective oxidation of CO in H2-rich streams,J.Power Sources196(2011)9590-9598.

[9]Y.Gong,H.Chen,Y.Chen,X.Cui,Y.Zhu,X.Zhou,J.Shi,Cu/Mn co-loaded mesoporous ZrO2-TiO2composite and its CO catalytic oxidation property,Microporous Mesoporous Mater.173(2013)112-120.

[10]Q.Ye,L.Yan,H.Wang,S.Cheng,D.Wang,T.Kang,H.Dai,Enhanced catalytic performance of rare earth-doped Cu/H-Sep for the selective catalytic reduction of NO with C3H6,Appl.Catal.A Gen.431-432(2012)42-48.

[11]F.Li,Y.Dai,M.Gong,T.Yu,X.Chen,Synthesis,characterization of magneticsepiolite supported with TiO2,and the photocatalytic performance over Cr(VI)and 2,4-dichlorophenol co-existed wastewater,J.Alloys Compd.638(2015)435-442.

[12]F.M.Bautista,J.M.Campelo,A.Garcia,R.Guardeño,D.Luna,J.M.Marinas,Influence of Ni/Cu alloying on sepiolite-supported nickel catalysts in the liquidphase selective hydrogenation of fatty acid ethyl esters,J.Mol.Catal.A104(1996)229-235.

[13]X.Lin,F.Jian,M.Chen,H.Zhi,C.Su,Co and Fe-catalysts supported on sepiolite:Effects of preparation conditions on their catalytic behaviors in high temperature gas flow treatment of dye,Environ.Sci.Pollut.Res.23(2016)1-8.

[14]M.R.Weir,E.Rutinduka,C.Detellier,C.Y.Feng,Q.Wang,T.Matsuura,R.Mao,Fabrication,characterization and preliminary testing of all-inorganic ultra filtration membranes composed entirely of a naturally occurring sepiolite clay mineral,J.Membr.Sci.182(2001)41-50.

[15]H.Guo,H.Zhang,F.Peng,H.Yang,L.Xiong,C.Wang,C.Huang,X.Chen,L.Ma,Effects of Cu/Fe ratio on structure and performance of attapulgite supported CuFeCo-based catalyst for mixed alcohols synthesis from syngas,Appl.Catal.A Gen.503(2015)51-61.

[16]D.Fang,J.Xie,D.Mei,Y.Zhang,F.He,X.Liu,Y.Li,Effect of CuMn2O4spinel in Cu-Mn oxide catalysts on selective catalytic reduction of NOxwith NH3at low temperature,RSC Adv.4(2014)25540-25551.

[17]V.Belessi,D.Lambropoulou,I.Konstantinou,A.Katsoulidis,P.Pomonis,D.Petridis,T.Albanis,Structure and photocatalytic performance of TiO2/clay nanocomposites for the degradation of dimethachlor,Appl.Catal.B Environ.73(2007)292-299.

[18]G.E.Muilenbergy,Handbook of X-ray Photoelectron Spectroscopy,Perkin Elmer Corporation,Minnesota,1979.

[19]J.Wang,Z.Liu,G.Feng,L.Chang,W.Bao,In situ synthesis of CuSAPO-34/cordierite and its selective catalytic reduction of nitrogen oxides in vehicle exhaust:The effect of HF,Fuel109(2013)101-109.

[20]Y.Dong,X.Niu,Y.Zhu,F.Yuan,H.Fu,One-pot synthesis and characterization of Cu-SBA-16 mesoporous molecular sieves as an excellent catalyst for phenol hydroxylation,Catal.Lett.141(2010)242-250.

[21]M.N.Rahaman,Y.Boiteux,L.C.Dejonghe,Surface characterization of silicon nitride and silicon carbide powders,Am.Ceram.Soc.Bull.65(1986)1171-1176.

[22]L.Yu,G.Diao,F.Ye,M.Sun,J.Zhou,Y.Li,Y.Liu,Promoting effect of Ce in Ce/OMS-2 catalyst for catalytic combustion of dimethyl ether,Catal.Lett.141(2011)111-119.

[23]C.Liu,Q.Liu,L.Bai,A.Dong,G.Liu,S.Wen,Structure and catalytic performances of nanocrystalline Co3O4catalysts for low temperature CO oxidation prepared by dry and wet synthetic routes,J.Mol.Catal.A370(2013)1-6.

[24]L.Oliveira,R.Rios,J.D.Fabris,K.Sapag,V.K.Garg,R.M.Lago,Clay-iron oxide magnetic composites for the adsorption of contaminants in water,Carbon40(2002)2177-2183.

[25]M.P.Pachamuthu,V.V.Srinivasan,R.Maheswari,K.Shanthi,A.Ramanathan,The impact of the copper source on the synthesis of meso-structured CuTUD-1:A promising catalyst for phenol hydroxylation,Catal.Sci.Technol.3(2013)3335-3342.

[26]S.R.Putluru,L.Schill,A.D.Jensen,B.Siret,F.Tabaries,R.Fehrmann,Mn/TiO2and Mn-Fe/TiO2catalysts synthesized by deposition precipitation—promising for selective catalytic reduction of NO with NH3at low temperatures,Appl.Catal.B Environ.165(2015)628-635.

[27]M.Ferrandon,J.Carno,S.Jaras,E.Bjornbom,Total oxidation catalysts based on manganese or copper oxides and platinum or palladium I:Characterization,Appl.Catal.A Gen.180(1999)141-151.

[28]S.A.C.Carabineiro,S.S.T.Bastos,J.J.M.Orfao,M.F.R.Pereira,J.J.Delgado,J.L.Figueiredo,Carbon monoxide oxidation catalysed by exotemplated manganese oxides,Catal.Lett.134(2010)217-227.

[29]L.C.Wang,Q.Liu,X.Huang,Y.M.Liu,Y.Cao,K.N.Fan,Gold nanoparticles supported on manganese oxides for low-temperature CO oxidation,Appl.Catal.B Environ.88(2009)204-212.

[30]Q.Huang,X.Yan,B.Li,X.Xu,Y.Chen,S.Zhu,S.Shen,Activity and stability of Pd/MMnOx(M=Co,Ni,Fe and Cu)supported on cordierite as CO oxidation catalysts,J.Ind.Eng.Chem.19(2013)438-443.

[31]P.O.Larsson,A.Andersson,Oxides of copper,ceria promoted copper,manganese and copper manganese on Al2O3for the combustion of CO,ethyl acetate and ethanol,Appl.Catal.B Environ.24(2000)175-192.

[32]X.Liu,Z.Jin,J.Lu,X.Wang,M.Luo,Highly active CuO/OMS-2 catalysts for lowtemperature CO oxidation,Chem.Eng.J.162(2010)151-157.

[33]H.Sun,Z.Liu,S.Chen,X.Quan,The role of lattice oxygen on the activity and selectivity of the OMS-2 catalyst for the total oxidation of toluene,Chem.Eng.J.270(2015)58-65.