Zebao Rui,Chunyan Chen,Yubing Lu,Hongbing Ji*
Department of chemical engineering,School of Chemistry&chemical engineering,and The Key Lab of Low-carbon Chemistry&Energy Conservation of Guangdong Province,Sun Yat-sen University,Guangzhou 510275,China
Anodic Alumina Supported Pt Catalyst for Total Oxidation of Trace Toluene☆
Zebao Rui,Chunyan Chen,Yubing Lu,Hongbing Ji*
Department of chemical engineering,School of Chemistry&chemical engineering,and The Key Lab of Low-carbon Chemistry&Energy Conservation of Guangdong Province,Sun Yat-sen University,Guangzhou 510275,China
A R T I C L E I N F o
Article history:
Received 25 December 2013
Received in revised form 13 January 2014
Accepted 20 January 2014
Available online 19 June 2014
Anodic alumina
Pt
Electrochemical anodization
Catalytic combustion
Toluene
Featuring an assembly of identical pores,through-pore anodic alumina(AAO)makes an idealmonolith-like catalyst support for volatile organic compound(VOC)combustion.This work employs the oxidation of toluene as a model reaction to investigate the applicability of AAOsupported Pt catalysts in VOCcatalytic combustion.In order to modify the microstructure of AAO,some AAOsamples were exposed to hot water treatment(HWT)firstly.Results show that the optimum HWT time is 18 h.Pt/HWT18 gives a toluene conversion of95%at 200°C,which is comparable to the initial activity of commercialγ-Al2O3particle supported Pt catalyst.Considering its confinement effect for the supported metal and its monolith-like compactunit,AAO support offers potential applications in VOC catalytic combustion.
?2014 The Chemical Industry and Engineering Society of China,and Chemical Industry Press.Allrights reserved.
Volatile organic compounds(VOCs)are the most concerned air pollutants which create great environmental problems and do damages to human health.Catalytic oxidation is an effective way to reduce the emission of VOCs due to its high removal efficiency and the absence of secondary pollution[1].In this way,supported precious metals such as Pt and Pd are well established as high activity catalysts for the oxidation of different VOCs[1-5].In selecting a catalyst,one needs also to consider other factors such as reactor design.The reactors with particle packed beds are easy to construct,but they cause a high-pressure drop and do not allow reducing particle size to obtain high specific surface area.Furthermore,these powdered or granular catalysts cannot be formed in desired shapes and cannot be readily used in indoor equipment such as air-conditioners or air-cleaners.Hence,a structured catalyst is favorable because of its low pressure drop and forming property [6].Especially,honeycomb monoliths have become the standard catalyst shape in most applications of environmental catalysis,after their successful commercial application to the control of automotive exhausts and to the reduction of nitrogen oxides.However,fluid flow in a small and long channel of a typical monolith is undoubtedly laminar,which will result in poor mass transfer and heat transfer[4].Additionally,the possible loss of catalyst washing-coated film by erosion would therefore limit its use to particulate-free atmospheres[7].
Recently,anodic oxidation technology was applied to produce monolith-like catalyst support[3-5,8-15].The anodic film formed on the surface of base metal(aluminum,titanium and stainless steel)can be directly used as the support together with the base metals[3-5],or can be isolated from the base metal and taken as the support alone [9-15].Due to the pores of anodic alumina(AAO)are unbranched, regular,controllable and nearly parallel,AAO shows advantage for mass and heat transfer.Furthermore,the functionality of a heterogeneous catalyst depends critically on its structure over a range of length scales[4,10].At the length scale of 10 nm-1000 nm,the dimension and topology of the catalyst pore structure can influence reagent fl ow, the sequencing of catalytic active sites,and the contact time between reagents and catalyst[4,10].Finally,the active site,especially novel metals,can be dispersed uniformly onto the AAO,and its confinement effect should be an important reason for the activity and stability of AAO supported catalysts[12,13],which is a key issue for catalytic combustion of VOCs.Hence,AAO is a highly promising support structure for the application in catalytic combustion.The applicability of AAO nanolith in VOC catalytic combustion was also demonstrated in a theoretical study by our group[14].
This work employs the oxidation of toluene as a modelreaction to investigate the applicability of AAO supported Pt catalysts in catalytic combustion of VOCs.Toluene,chosen probe molecule,is a commonly used solvent in chemical and processing industries and presents an importanttoxic pollutant[16].The through-pore AAOmembrane was produced by a two step anodization method[17].In order to increase the BET-area and modify the microstructure of the support,some AAO samples were firstly exposed to hot water treatment(HWT x,x refers to the treating time).Results show that Pt/HWT18 performs the best with the temperature at which 90%toluene conversion is achieved(T90)lowerthan 200°C,which shows a comparable initial activity with Pt/γ-Al2O3. Considering the fast thermal diffusion along the nanopores of AAO[18], its confinement effect for the supported metal[12,13],and its monolith like compactness,the AAO support offers potential application in VOC catalytic combustion.
2.1.Preparation ofcatalysts
The AAO structure growing method follows the well-established procedures[17,19]with some modifications.High-purity aluminum sheets(99.99%)with a thickness of 0.4 mm were used as the starting materials.The as-received aluminum sheets were firstly calcined in air at500°C for 4 h.Then,the aluminum sheets were degreased in acetone, ethanol and deionized water for 3 min of ultrasonic cleaning,separately. Prior to anodizing,the sheets were electro polished at5°C and 8.5 A in a 1:8 by volume mixture of perchloric acid to ethanol.Subsequently,anodization was carried out in 0.3 mol·L-1oxalic acid at5°C under a constant voltage of 50 V.After the first anodization for 10 h,the AAO membrane was removed into a 1:1 by volume mixture of phosphoric acid(1.8%)and chromic acid(6%)at 85°C and dipped for 4 h.Subsequently,the second anodization was conducted for 24 h under the same conditions as the first anodization.To obtain a through-pore membrane,the sample was then anodically oxidized in a solution of HClO4(72%,by mass)and(CH3CO)2(98%)(volume ratio is 1:1)at a voltage of60 V for 3 s.After this,the porous alumina film was separated from the substrate Alsheets immediately.Then,the as-prepared AAO membranes were immersed in 6%(by mass)phosphoric acid at 30°C for 35 min to widen the pores.Next,the AAOmembranes were cleaned with deionized water and dried at 120°C in air overnight.Finally,the obtained AAO membranes were annealed at 900°C for 6 h in air for use.In order to increase the BET-area of the support,some AAO samples were firstly calcined at 350°C for 1 h,and then exposed to HWT at ca. 100°C for some hours,referred as HWT x.It is should be noted that the heating of the solution should be controlled carefully to avoid intensive boiling and obtain a stable HWT process.HWT x was then calcined at 900°C for 6 h for use.
Pt nanoparticles were dispersed onto AAO orγ-Al2O3by a simple ultrasonic-aided incipient wetness impregnation method[13]with H2PtCl6·6H2O(Alfa Aesar)as the Ptprecursor.γ-Al2O3supportwasobtained by calcining commercialγ-Al2O3powder(Sinopharm Chemical Reagent Co.,China)at 900°C for 6 h.The samples were dried at 100°C overnight,calcined at 300°C for 6 h,and then washed with water to remove Cl-and dried for characterization.According to the inductively coupled plasma(ICP,IRIS(HR),Thermo JarrellAsh Corporation,US)element analysis,the amount of platinum supported onγ-Al2O3powder and HWT x was similar(ca.0.25%,by mass).The catalyst was reduced with a NaBH4solution before reaction.In a typical NaBH4reduction process,a NaBH4solution as reducing agent was quickly and thoroughly mixed with the catalyst(NaBH4/Pt=20,molar ratio) under supersonic vibration.After reduction,the suspension was dried in air at120°C for 6 h without further high temperature calcinations.
2.2.Characterization of catalysts
The morphologies were observed on a Quanta 400F Scanning electron microscope(SEM)ata 20 kV accelerating voltage.The phase purity and crystal structure of the catalysts were examined by X-ray diffraction (XRD,D-MAX 2200 VPC,Rigaku,Japan),using monochromatic Cu Kαradiation.The Brunauer-Emmett-Teller(BET)surface area,pore volume, and pore size distribution of the samples were measured with a Micromeritics ASAP 2020 instrument using adsorption of N2at 77 K. The density functional theory(DFT)model was employed to evaluate the pore distribution,cumulative pore volume and total area in pores. Prior to adsorption analysis,the catalysts were degassed in a fl owing N2at 300°C for 3 h.CO chemisorption was performed using a Micromeritics ASAP 2020C automated system.Before CO chemisorption, around 0.2 g of fresh calcined catalyst was evacuated to 10-6mm Hg at 110°C for 30 min;then it was reduced under fl owing H2at 200°C for 30 min.The catalyst was evacuated again at 200°C for 30 min to desorb any H2.The chemisorption analysis was carried out at 35°C.A CO/Pt average stoichiometry of 1 has been assumed for the calculation of dispersion.
2.3.Testing of catalysts
The catalytic tests were performed using a continuous-fl ow quartz fixed-bed reactor(i.d.=7 mm)under atmospheric pressure,as schematized in Fig.1.N2from a cylinder fl owed through a saturator filled with toluene solution and then mixed with the gas mixture of N2and O2from another cylinder.The molar ratio of N2/O2in the final inlet gas was 4.The concentration of toluene was around 300μg·g-1,which could be controlled by controlling the temperature of the saturator and the flow rate of N2through the saturator.Activity test experiments were performed at a constant GHSV(gas hourly space velocity,de fined as the total flow rate of gas at STP per unit mass of catalyst)of 30,000 ml·h-1·g-1.The temperature was raised by steps of about 25°C or 50°C from 150 to 300°C to obtain curves of toluene conversion as a function of temperature(light-offcurves).Ateach temperature,we measured 3 to 5 data with around 15 min per datum.The acquired data show less than 10%difference with each other.Toluene conversion was calculated by measuring the toluene removal by GC1(FID,GC 7890II, Shanghai Tianmei Scienti fi c Instrument Co.,China)and CO2production by GC2(GC 2060,FID,Tengzhou Lunan Analysis Instrument Co.,China), which was equipped with a nickelcatalyst-based methanizer to enable FID to detect CO2.Toluene conversions were calculated by
where nCO2and ntolueneare the molar fl ow(mol·s-1)of CO2and C7H8at the outlet of the reactor,respectively.
3.1.Catalysts characterization
Figs.2 and 3 are the top and cross-sectional views of as-prepared AAO support and HWT x samples.To obtain a good crystallinity and effectively remove electrolyte impurities incorporated into the oxide material during anodization,a high calcination temperature of900°C was chosen in this work.As shown in Figs.2(a)and 3(a),no destruction of the AAO structure has been observed,and AAO is of highly-ordered and vertically oriented tube array structure even after a high temperature annealing(900°C).High annealing temperature(as high as 1300°C)was also frequently used in the literature to accompany the crystallization and the crystalline transformation of AAO without changing the architecture of AAO[20].
The HWT technology was applied to increase the BET-area of the support,as employed by Wang et al.[5,21].The HWT time was set to 0,2,3,6,12,18 and 24 h,respectively.Figs.2 and 3 reveal considerable differences in the nanotube morphology for the as-prepared AAO and the HWT x samples after the calcination at 900°C.With an increase in HWT time,numerous cracks in the nanotube walls appear and the nanotube walls become rough.However,the monolith-like structure still keeps for these samples,as revealed by the cross section images in Fig.3.In the HWT process,the alumina in the anodic film reacts with hot water to form hydrated alumina(i.e.Al2O3+n H2O→Al2O3· n H2O).The hydrate water in hydrated alumina willbe lost in the subsequent high-temperature calcination(>673 K),and in the meanwhile hydrated alumina rearranges in toγ-alumina[5,21].In the process ofrearrangement,a micropore structure with smaller pore size appears in the location where hydrate water loses,which may cause a remarkable increase in BET surface area.However,in the HWT,an effect of sealing pore will simultaneously happen due to the increased volume of hydrated alumina,in particular atthe top partof the pore structure where the reaction of alumina layer with hot water is the rapidest[21],as indicted inFig.2.It should be noted that the as-prepared AAO has a thickness of~65μm.During HWT,no obvious change of the membrane thickness was observed.
Fig.1.Schematic of apparatus for catalytic evaluation.
Fig.2.Surface morphology of AAO treated with different HWT time.
Fig.3.Cross section morphology of AAO treated with different HWT time.
In addition,the surface morphology transformation of AAOis related with its specific surface area and pore size and volume.From Table 1,it can be found that AAO prepared without HWT has the lowest BET-area of52.0 m2·g-1and pore volume of0.14 cm3·g-1but the largest pore size of11.0 nm.The BET-area and pore volume of the support increase while the pore size decreases with the HWT time and they show reverse trend when HWT time exceeds 18 h.HTW18 has the largest BET-area of 127.2 m2·g-1and pore volume of 0.24 cm3·g-1among these HWT x samples.
After loading Pt particles into the support and the subsequent NaBH4solution reduction,the BET-area and pore volume of the catalysts decrease a lot.The BET-area and pore volume decrease from 306.7 m2·g-1and 0.45 cm3·g-1forγ-Al2O3to 272.7 m2·g-1and 0.41 cm3·g-1for Pt/γ-Al2O3,and they are from 127.2 m2·g-1and 0.24 cm3·g-1for HWT18 to 83.2 m2·g-1and 0.16 cm3·g-1forPt/HWT18.This is probably due to the blockage of the small pores by Pt and Na+.The average Pt particle size and Pt dispersion calculated by CO chemisorptions are also listed in Table 1.It is observed that the high BET-area surface brings the high dispersion of platinum and small average Ptparticle size.The Pt dispersion and particle size of Pt/γ-Al2O3are 23.5%and 4.8 nm,while they are only 10.4%and 10.8 nm for Pt/AAO. After HWT for 18 h,Pt dispersion increases to 19.5%and Pt particle size decreases to 5.8 nm over Pt/HWT18.This result indicates that the blockage by Pt and Na+is more serious for the sample with small pores.For example,although HWT24 and HWT18 have the same BET-area,the BET-area and Pt dispersion over Pt/HWT18 are much larger than those for Pt/HWT24.XRD patterns ofγ-Al2O3,Pt/γ-Al2O3,Pt/AAO and Pt/HWT18 are shown in Fig.4.All samples are structured with theγ-Al2O3phase,which indicates that HWT does not change the surface phase of AAO.The alumina in the AAO with HWT was boehmite. The calcination after HWT made the boehmite film lose its hydrate water and rearrange intoγ-alumina[21].The characteristic peaks of PtO2or Ptare too weak to be detected during the XRD characterizations due to the low Pt content in the catalysts[22].
Table 1 Specific surface area,average pore size,pore volume,Pt dispersion and particle sizes of the different samples
Fig.5.The activity of Pt/AAO with different HWT time and Pt/γ-Al2O3for trace toluene combustion.■Pt/γ-Al2O3;●Pt/AAO;▼Pt/HWT6;△Pt/HWT12;▽Pt/HWT18;◆Pt/ HWT24.
3.2.Catalytic activity
The catalytic combustion of toluene was used to evaluate the catalytic activity of the developed catalysts.The obtained results are expressed in Fig.5.As presented,Pt/HWT18 performs the best among all the Pt/ HWT x catalysts evaluated with a toluene conversion of 95%at 200°C, which is comparable to the initial activity of Pt/γ-Al2O3.It also can be found that the activities of Pt/HWT6,Pt/HWT12 and Pt/HWT24 are lower than that of Pt/AAO,even though their BET-areas are much enlarged after HWT,and higher Pt dispersions can also be expected.
It is well known that a high dispersion of noble metal particles over the support,resulting in a high metal exposure surface area,is beneficial to the heterogeneous catalysis reaction.Among tested catalysts,Pt/γ-Al2O3holds the largest BET-area and Pt dispersion among these catalyst exhibits the highest activity.However,toluene combustion over supported noble catalysts can also be influenced by both support structures[2,3]and noble metal particle size[23,24].It was reported that toluene combustion reaction over supported Pt catalyst was structure sensitive[23,24],namely,changes in a specific activity as a function of noble metal particle size were attributed to morphological effects rather than to the chemicalstate of Pt.On large Ptcrystallites,oxygen was held on Ptatoms with a lower bond strength followed by that on small Pt crystallites,which explained the structure sensitivity found for toluene oxidation[23].Additionally,it was suggested that effective toluene diffusion near the active site of the catalysts was important to toluene combustion[2].Kameyama et al [3]showed that the catalyst with a bigger pore performed better for toluene combustion than that with a smaller one[3].Thus,although the BET-area and Pt dispersion of Pt/HWT18 are lower than those of Pt/γ-Al2O3,Pt/HWT18 shows comparable high toluene oxidation activity with respect to Pt/γ-Al2O3due to the relative larger Pt particle size and pore size on Pt/HWT18.When the differences in the BET-area and Pt dispersion among the samples are not too large,the effect of Pt particle size and pore size becomes dominant,which should be the reason for the superior activity of Pt/AAO to Pt/HWT6,Pt/HWT12 and Pt/HWT24.In all,an appropriate combination of active noble component(including dispersion,particle size and chemicalstate etc.)and support oxide(including BET-area,pore size and pore volume)is important for an efficient VOC combustion catalyst.
Allthese results show that AAO with an optimal HWT time,holding the special monolith-like structure,is beneficial for VOC combustion, and provides a new platform for the future VOC combustion catalyst development.However,in order to put the AAO based catalysts into practice,more experimental work needs to be done.Firstly,the resistance of the AAO based catalyst to sintering upon mild steam aging and their resistance to sulfur containing compounds need to be explored.Secondly,some other precious metal,such as,Pd and Rh,may also need to be supported on AAO and explored for the VOC oxidation under a specific condition.For example,it was reported that under reducing condition,supported Pt catalyst was more active,while under the oxygen-rich combustion condition,supported Pd catalyst performed better[25].Finally,the catalytic performance of a combustion catalyst also depends on the chemical nature of the support[26].It was reported that reducible metaloxide supported precious metal catalyst,such as ceria,was more active to the VOC combustion[26].Thus, the catalytic performance of AAO based catalysts may be further optimized by proper selection of another metaloxide loading.
Fig.4.XRD patterns ofγ-Al2O3,Pt/γ-Al2O3,Pt/AAO and Pt/HWT18.
Through-pore anodic alumina(AAO)consisting of an assembly of identical pores was applied as a monolith-like catalyst support for VOC combustion.Hot water treatment(HWT)displays a significant effect on both the surface area and pore size of AAO support.Both the surface area and pore size show an effect on the dispersion of Pt particle and the performance of Pt/AAO catalysts.The optimal HWT time is 18 h.Pt/HWT18 gives a toluene conversion of95%at 200°C,which is comparable to the initial activity of commercialγ-Al2O3particlesupported Pt catalyst.Considering the fast thermal diffusion along the nanopores of AAO and its confinement effect for the supported metal, and its monolith-like compact unit,AAO support offers potential applications in VOC catalytic combustion.
[1]S.H.Taylor,C.S.Heneghan,G.J.Hutehings,I.D.Hudson,The activity and mechanism of uranium oxide catalysts for the oxidative destruction of volatile organic compounds,Catal.Today 59(2000)249-259.
[2]J.Fang,X.Chen,Q.Xia,H.Xi,Z.Li,Effect of relative humidity on catalytic combustion of toluene over copper based catalysts with different supports,Chin.J.Chem.Eng.17 (5)(2009)767-772.
[3]I.E.Sungkono,H.Kameyama,T.Koya,Development of catalytic combustion technology of VOC materials by anodic oxidation catalyst,Appl.Surf.Sci.121/122(1997) 425-428.
[4]L.F.Wang,T.P.Tran,D.V.Vo,M.Sakurai,H.Kameyama,Design of novelPt-structured catalyst on anodic aluminum support for VOC's catalytic combustion,Appl.Catal.A Gen.350(2008)150-156.
[5]L.F.Wang,M.Sakurai,H.Kameyama,Study ofcatalytic decomposition of formaldehyde on Pt/TiO2alumite catalyst at ambient temperature,J.Hazard.Mater.167 (2009)399-405.
[6]G.Groppi,E.Tronconi,Honeycomb supports with high thermal conductivity for gas/solid chemical processes,Catal.Today 105(2005)297-304.
[7]A.Janbey,W.Clark,E.Noordally,S.Grimes,S.Tahir,Noble metalcatalysts for methane removal,Chemosphere 52(2003)1041-1046.
[8]M.Chen,Y.Ma,G.F.Li,X.M.Zheng,Support effect,thermal stability,and structure feature of toluene combustion catalyst,Catal.Commun.9(2008)990-994.
[9]P.C.Stair,C.Marshall,G.Xiong,H.Feng,M.J.Pellin,J.W.Elam,L.Curtiss,L.Iton,H. Kung,M.Kung,H.-H.Wang,Novel,uniform nanostructured catalytic membranes, Top.Catal.39(2006)181-186.
[10]H.Feng,J.W.Elam,J.A.Libera,M.J.Pellin,P.C.Stair,Catalytic nanoliths,Chem.Eng.Sci. 64(2009)560-567.
[11]H.Feng,J.W.Elam,J.A.Libera,M.J.Pellin,P.C.Stair,Oxidative dehydrogenation of cyclohexane over alumina-supported vanadium oxide nanoliths,J.Catal.269 (2010)421-431.
[12]D.Y.Feng,Z.B.Rui,H.B.Ji,Monolithic-like TiO2nanotube supported Ru catalyst for activation of CH4and CO2to syngas,Catal.Commun.12(2011)1269-1273.
[13]D.Y.Feng,Z.B.Rui,Y.B.Lu,H.B.Ji,A simple method to decorate TiO2nanotube arrays with controllable quantity of metal nanoparticles,Chem.Eng.J.179(2012)363-371.
[14]Z.B.Rui,Y.B.Lu,H.B.Ji,Simulation ofVOCs oxidation in a catalytic nanolith,RSC Adv. 3(2013)1103-1111.
[15]Z.B.Rui,D.Y.Feng,H.Y.Chen,H.B.Ji,Evaluation of TiO2nanotube supported Ru catalyst for syngas production,Catal.Today 216(2013)178-184.
[16]F.G.Shahna,F.Golbabaei,J.Hamedi,H.Mahjub,H.R.Darabi,S.J.Shahtaheri,Treatment of benzene,toluene and xylene contaminated air in a bioactive foam emulsion reactor,Chin.J.Chem.Eng.18(1)(2010)113-121.
[17]H.Masuda,K.Fukuda,Ordered metalnanohole arrays made by a two-step replication of honeycomb structures of anodic alumina,Science 268(1995)1466-1468.
[18]D.-A.Borca-Tasciuc,G.Chen,Anisotropic thermal properties of nanochanneled alumina templates,J.Appl.Phys.97(2005)084303.
[19]J.H.Yuan,F.Y.He,D.C.Sun,X.H.Xia,A simple method for preparation of throughhole porous anodic alumina membrane,Chem.Mater.16(2004)1841-1844.
[20]Y.Chang,Z.Ling,Y.Liu,X.Hu,Y.Li,A simple method for fabrication of highly ordered porousα-alumina ceramic membranes,J.Mater.Chem.22(2012) 7445-7448.
[21]Y.Guo,M.Sakurai,H.Kameyama,A.Matsuyama,Y.Kudoh,Preparation of alumite support and preliminary activity investigation for NO removal in SCR-HC over alumite catalyst,J.Chem.Eng.Jpn 36(2003)1470-1479.
[22]P.Claus,S.Schimpf,R.Schodel,P.Kraak,W.M?rke,D.H?nicke,Hydrogenation of crotonaldehyde on Pt/TiO2catalysts:in fl uence of the phase composition of titania on activity and intramolecular selectivity,Appl.Catal.A Gen.165(1997)429-441.
[23]N.Radic,B.Grbic,A.Terlecki-Baricevic,Kinetics of deep oxidation of n-hexane and toluene over Pt/Al2O3catalysts Platinum crystallite size effect,Appl.Catal.B Environ. 50(2004)153-159.
[24]B.Grbic,N.Radic,A.Terlecki-Baricevic,Kinetics of deep oxidation of n-hexane and toluene over Pt/Al2O3catalysts oxidation of mixture,Appl.Catal.B Environ.50 (2004)161-166.
[25]P.Gélin,M.Primet,Complete oxidation of methane at low temperature over noble metalbased catalysts:a review,Appl.Catal.B Environ.39(2002)1-37.
[26]S.M.Saqer,D.I.Kondarides,X.E.Verykios,Catalytic activity of supported platinum and metaloxide catalysts for toluene oxidation,Top.Catal.52(2009)517-527.
☆Supported by the National Natural Science Foundation of China(21106189, 21036009),the Natural Science Foundation(S2011040001767)and the Foundation from the Educational Commission(LYM11004)of Guangdong Province.
*Corresponding author.
E-mailaddress:jihb@mail.sysu.edu.cn(H.Ji).
Chinese Journal of Chemical Engineering2014年8期