Tailoring the mesopore structure of HZSM-5to control product distribution in the conversion of propanal
Xinli Zhu,Lance L.Lobban,Richard G.Mallinson,Daniel E.Resasco *
Center for Biomass Refining,School of Chemical,Biological,and Materials Engineering,The University of Oklahoma,Norman,OK 73019,USA
a r t i c l e i n f o Article history:
Received 23September 2009Revid 14December 2009Accepted 2February 2010
Available online 6March 2010Keywords:Propanal HZSM-5Desilication
Biomass conversion Bio-oil upgrading
a b s t r a c t
Conversion of propanal to gasoline-range molecules was investigated over a ries of HZSM-5catalysts with controlled mesoporosity generated by desilication.Characterization of the structure of
the solid by powder X-ray diffraction (XRD),scanning electronic microscopy (SEM),ammonia and isopropylamine temperature programmed desorption (TPD),and n -butane diffusivity measurements confirmed the development of various degrees of mesoporosity in the zeolites.This structural modification ems to have little influence on Brønsted acid density.The catalyst stability was improved upon desilication due to an increa in coke tolerance.The product distribution of the propanal conversion was found to vary with the verity of the desilication.Increasing the extent of desilication gradually reduced the aro-matization and cracking reactions,due to a reduction in the fraction of micropores and in the diffusion path length.Mildly desilicated samples were found to exhibit the best stability on stream and inhibited coke formation.
Ó2010Elvier Inc.All rights rerved.
1.Introduction
Conversion of lignocellulosic biomass into liquid hydrocarbon fuels provides a CO 2neutral energy production route,which poten-tially can reduce the dependency on fossil fuels [1–3].In the ther-mochemical route (e.g.fast pyrolysis),the molecular structure of biomass is broken down into smaller fragments that subquently undergo further conversion in the vapor and liquid phas con-densing
into a complex product termed bio-oil.Some of the con-stituents of this product are larger than the desirable fuel range,others are shorter,but all of them contain significant amounts of oxygen.The chemically unstable,highly viscous,corrosive,and low-heating value liquid product includes acids,aldehydes,ke-tones,phenolic compounds,sugars,and dehydrosugars [4,5].Deoxygenation of the larger oxygenated molecules (guaiacols,vanillins,cresols,catechol,etc.)is being extensively investigated [6–10];and hydrogenation,hydrogenolysis,and decarbonylation are potential reaction pathways to improve the quality of the heavy molecules.By contrast,the short oxygenates (e.g.,alde-hydes,acids,ketones)need to be condend into larger molecules to become uful fuel components.Under hydrotreating conditions for refining the complete bio-oil,short oxygenates are converted to light hydrocarbons of low value while consuming substantial hydrogen.Alternative strategies that avoid discarding short oxy-genates should be considered since they constitute a significant
fraction of the product [4,5].Due to the highly complex nature of the bio-oil,understanding the reaction pathways for each kind of compound conversion is highly desirable for catalyst and process screening.Therefore,the study of model compounds is the first step in simplifying the complexity of the problem [11–14].While our next studies will include more complex mixtures,in this work,propional
dehyde (propanal)has been lected as a model com-pound to investigate the conversion of short aldehydes into gaso-line-range molecules.
The study of propanal conversion is also relevant to the utiliza-tion of glycerol,a major by-product of bio-diel production.Glyc-erol is readily converted into acrolein by dehydration [15,16].Subquent hydrogenation produces propanal [17].Thus,conver-sion of propanal may also reprent a potential approach for the conversion of bio-diel by-products to gasoline-range fuels.
Previous studies on the conversion of small oxygenates (meth-anol,ethanol,etc.)to hydrocarbons (alkene/alkane,aromatics)over zeolites have addresd propanal conversion briefly.It has been re-ported that propanal can yield aromatics in higher lectivity than acetone and much higher than other C 3oxygenates (alcohol,acid,ester)[18–20].However,it was also found that propanal caus a rapid catalyst deactivation [19].
Zeolites are widely ud in hydrocarbon conversion due to their high density of strong acid sites and their well-defined micropo-rous channel structure that enable shape lective reactions inside the pore channels.However,transport of both reactants and prod-ucts in and out of the micropores may be limited by diffusion.Con-ventional mesoporous materials such as MCM-41and SBA-15have superior diffusion properties but lower thermal/hydrothermal
0021-9517/$-e front matter Ó2010Elvier Inc.All rights rerved.doi:10.1016/j.jcat.2010.02.004
*Corresponding author.Fax:+14053255813.
E-mail address:mallinson@ou.edu (R.G.Mallinson),resasco@ou.edu (D.E.Resasco).
Journal of Catalysis 271(2010)
88–98
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage:www.el v i e r.c o m /l o c a t e /j c a
t
stability and weaker acidity.To improve the performance of zeo-lites by enhancing transport,a post-synthesis method called desi-lication has been ud[21–25].This method lectively removes silica from the zeolite crystals,generating mesopores.The resul-tant material has a hierarchical pore structure with pores of vary-ing dimensions and with a shortened diffusion path length for reactants and products[26,27],as well as improved accessibility for large molecules[28–30].Desilicated zeolites have been investi-gated in veral reactions,including cumene cracking,methanol to propylene,methanol to gasoline,hydroxylation of benzene to phe-nol,methane aromatization,and hexene conversion[29–35].In most cas,improved activity,stability,and lectivity have been reported.In general,the obrved improvement has been ascribed to enhanced diffusion due to the generation of mesopore channels. In this work,four zeolite samples with varying degree of desilica-ti
on have been characterized for changes in texture,structure, acidity,and diffusivity,and then ud in the conversion of propanal with the objective of investigating the effects of mesopore genera-tion on the conversion and lectivity toward gasoline-range prod-ucts.H-ZSM-5has been chon as the basis for this study due to its well-known activity for the conversion of short oxygenates to aromatics.
2.Experimental
2.1.Zeolite synthesis and desilication
景The parent zeolite was synthesized hydrothermally using so-dium aluminate(Aldrich)dissolved in deionized water as the Al source to which tetrapropylammonium hydroxide(Fluka,20%) TPAOH)wasfirst added under stirring as the structure directing agent,silica gel(Ludox,40%)was then added dropwi while stir-ring,as the Si source.The resultant gel composition was 150SiO2:1.0Al2O3:8TPAOH:1600H2O(Si/Al=75).After stirring at 700rpm for10h at room temperature,the gel was transferred to a Teflon-lined autoclave,where the zeolite crystallized at180°C over5days with stirring at60rpm.The solid product was recov-ered byfiltration,washed,dried at110°C,andfinally calcined in air at550°C for6h to remove the template.
Four desilicated samples(DS1,DS2,DS3,and DS4)were pre-pared by varying the basicity of the alkali
ne solution,treatment temperature,and time,resulting in different extents of silica re-moval.To achieve the desired level of desilication,the parent sam-ple(P)was treated in the following desilication baths,keeping in all cas a ratio of30mL solution per gram of zeolite and using 350rpm stirring rate.DS1:0.45M Na2CO3(Aldrich)solution for 30h at75°C;DS2:0.2M NaOH(Aldrich)solution at65°C for 30min;DS3:at80°C for30min,and DS4:at80°C for4h.After each treatment,the desilication was stopped by quenching the sample in an ice-water bath.The resultant samples were thenfil-tered,washed,and dried at110°C overnight.After this step,the samples were further suspended in deionized water at80°C for 2h to remove any amorphous silica and alkaline metals remaining in the solid.Finally,the samples were againfiltered,washed,and dried at100°C overnight.
The H-form of the zeolite was obtained by repeating three times the ion exchange of the Na-form of the zeolite with1M NH4NO3 (Aldrich)solution(10mL/g)at80°C for10h.After each exchange, the samples werefiltered,washed,and dried,andfinally calcined inflowing air at550°C for4h.
2.2.Catalyst characterization
Powder X-ray diffraction(XRD)patterns of the zeolite samples were recorded using a Bruker D8Discover diffractometer,with a Cu K a radiation source(k=1.54056Å).High resolution scanning elec
tronic microscopy(SEM)obrvations were performed on gold-coated samples in a Jeol JSM-880electron microscope equipped with X-ray elemental analyzer.Nitrogen adsorption measurements[36]were performed in an Autosorb-1analyzer (Quantachrome)at liquid nitrogen temperature after outgassing the samples under vacuum at300°C for5h.The micropore volume (V micro)and micropore surface area(S micro)were derived by the t-plot method[37]using the adsorption data of0.2<p/p0<0.6. The mesopore size distribution was obtained from the BJH method [38]applied to adsorption branch[39]data for p/p0>0.35.Finally, the total pore volume was determined at p/p0=0.99.
The acid properties of the various zeolites were characterized by temperature programmed desorption of ammonia(NH3-TPD) and isopropylamine(IPA-TPD),using d.quartz reactor. Before each experiment,the zeolite sample(100mg for NH3-TPD and50mg for IPA-TPD)was pretreated for0.5h inflowing He (30mL/min)at600°C to eliminate any adsorbed water.Then,the temperature was reduced to100°C,and the sample was expod to NH3(2%NH3/He,30mL/min,30min)or to IPA(4l L/pul,10 puls,3min/pul).After exposure to the respective adsorbate, Heflowed for0.5h to remove weakly adsorbed NH3or IPA.To start the TPD,the temperature was incread to650°C at a heating rate of10°C/min.The evolution of desorbed species
was continuously monitored by a Cirrus mass spectrometer(MKS)recording the fol-lowing signals m/z=17and16(NH3),18(H2O),44(IPA),and41 (propylene).The density of acid sites was quantified by calibrating the MS signals using the average of105-mL-puls of2%NH3/He.
The changes in diffusivity upon desilication were evaluated by nding an n-butane pul using the same system ud for TPD. The zeolite samples(100mg,40–60mesh)were pretreated in flowing He(30mL/min)at400°C for0.5h to remove water.Then, the temperature was reduced to90°C,and a5mL n-butane pul (10%n-butane/He)was nt underflowing He(30mL/min).The butane concentration was monitored by following the m/z=43sig-nal in a mass spectrometer(MS)and compared to puls over non-porous blanks.
The amounts of coke deposits were quantified by Temperature Programmed Oxidation(TPO)by passing a2%O2/He stream over a20mg spent catalyst sample,using a linear heating rate of 10°C/min.The signals of H2O(m/z=18),CO2(m/z=44),and CO (m/z=28)were continuously monitored by MS.Quantification was calibrated on the basis of the signals from100l L CO2and CO puls inflowing He.
2.3.Catalytic measurements
The catalytic performance of the different samples was exam-ined in a quartz reactor(d.)at atmospheric pressure. The catalyst sample(10–400mg,40–60mesh)was packed in the reactor between two layers of quartz wool.The thermocouple was affixed to the external wall of the reactor clo to the catalyst bed.The temperature of the catalyst bed was incread to400°C using a rate of10°C/min and held at400°C for0.5h inflowing H2(35mL/min)before reaction.Liquid propanal(from Aldrich) was fed using a syringe pump(kd scientific)equipped with a nee-dle at a rate of0.12mL/min.The liquid was completely vaporized in the line before entering the reactor.All lines were kept at300°C to avoid condensation of reactant or products.The products were analyzed online using a gas chromatograph(GC6890,Agilent) equipped with aflame ionization detector(FID)and a60m Inno-wax capillary column.The effluent was trapped in methanol using an ice-water bath and analyzed using a QP2010s GC–MS(Shima-dzu)with an Innowax column.Quantification of products was done by combination of GC–MS analysis and injection of known amounts of standard compounds.The space time(W/F)is defined
X.Zhu et al./Journal of Catalysis271(2010)88–9889
as the ratio of catalyst mass(g)to propanal massflow rate(g/h) with a carrier gasflow rate of30mL/min.The range of W/F ud in this study was0.1–4.0h.The propanal conversion and product
yield were calculated bad on carbon atoms.
3.Results
3.1.Catalyst characterization of desilicated samples
3.1.1.Porous structure
As shown in Table1,different levels of desilication were ob-tained by treating the ZSM-5zeolite with various solutions that ranged from a weakly basic solution(Na2CO3)to a strongly basic solution(NaOH)that is very effective in removing silica species from the zeolite crystal even at low temperatures and short times. As a result,the weight loss gradually incread from DS1(22%)to DS4(52%)as the verity of the treatment incread.It has been suggested that negatively charged AlOÀ
4
protects against OHÀat-tack,whereas the Si A O A Si bond is more easily attacked by OHÀ[21–25].In agreement with previous studies,the calculated Si/Al ratios that result from assuming that only silica is removed are consistent with the elemental analysis data obtained in the SEM by EDX.
Fig.1shows that even though a significant amount of silica was removed from the zeolite,the XRD patterns show that the MFI structure is prerved.However,a clo examination of the characteristic peaks of MFI structure in the2h range of23–25°(e the int in Fig.1),revealed a slight shift in the peak posi-tions to lower angles for the desilicated samples.This shift may be interpreted as a slight expansion of the unit cell of the zeolites, which could be ascribed to the lective removal of Si.This result is in agreement with tho of Ohayon et al.[40],who reported that the micropores were slightly enlarged by the desilication-stabilization process.It must be noted that in the most heavily desilicated sample(DS4),the diffraction peaks appear to shift back to higher angles.This reverd structural modification could be explained by a previously reported‘healing effect’[22,33] undergone by the zeolite after prolonged treatment that allows some of the dissolved Si species to re-inrt back into the zeolite framework.
The SEM micrographs of the parent sample(P)show well-crys-tallized particles of$6l m(Fig.2a).The particles are twinned crys-tals(Fig.2b),and the well-defined surfaces are rather smooth with few defects(Fig.2c).Desilication ems to break up some of the particles into smaller fragments(Fig.2d).The surface becomes rougher even for tho particles that prerve their original shape and size(Fig.2e).The higher-magnification image(Fig.2f)shows evidence of the prence of etched channels on the surface of the particle,indicating the development of mesoporosity.
Fig.3a and b shows the N2adsorption–desorption isotherms and the results of the BJH analysis for both the parent and desilicat-ed HZSM-5samples.The incread adsorption in the range p/p0> 0.5and the appearance of hysteresis loops in the desorption branch at p/p0of$0.42of the desilicated samples indicate the develop-ment of mesopores.From the analysis of the data,it can be inferred that the treatment in weak ba(Na2CO3)for prolonged times (DS1)results in a relatively wide distribution of pore sizes,cen-tered at12.0nm but with a rather small overall pore volume.By contrast,the treatment with the strong ba(NaOH)at low tem-peratures(sample DS2)results in a relatively narrower pore size distribution centered at7.3nm and a larger overall pore volume. Increasing the verity of the NaOH treatment by using higher temperatures and longer times(DS3and DS4),results in a gradual increa in both pore size and overall pore volume,with a peak center at10.7and13.9nm,respectively.The results from the N2 adsorption are summarized in Table1.They confirm that the frac-tion of mesoporosity significantly increas at the expen of the microporosity with increasing desilication verity.The results of the effect of temperature and time of alkaline treatment are in good agreement with tho reported by Groen et al.[25,39,41].It is noted that the V micro is slightly incread for DS1,possibly as a result of the prolonged treatment time with the weakly basic Na2CO3.This long treatment may lead not only to the removal of Si but also to re-incorporation of some of the dissolved species into the zeolite structure,further creating microporosity.
难忘的人作文
3.1.2.Acidity
The effects of desilication on the acidity of the HZSM-5zeolites were studied by TPD of adsorbed NH3and IPA,as shown in Figs.4 and5,respectively.For sample P,two distinct desorption peaks are obrved at180°C and385°C in the NH3-TPD.The are usually ascribed to NH3desorption from weak and strong acid sites,
Table1
Weight loss,final Si/Al ratio,specific area(S BET),and pore volume of parent(P)and desilicated(DS)zeolite samples.
腿肿脚肿怎么治疗Sample W loss(wt.%)Si/Al a Si/Al b S BET(m2/g)V total(cm3/g)V micro(cm3/g)V meso c(cm3/g)S micro(m2/g)S meso c(m2/g)V meso/V micro
P––763920.2170.1610.056359330.35 DS12259624140.2970.1710.126369490.74 DS22854524290.3520.1510.20133594 1.33 DS34839424530.4330.1380.296310143 2.14 DS45236374410.4920.1380.354311130 2.56
a Estimated from weight loss(W
loss
)assuming that only silica was removed.
b Estimated by SEM elemental analysis of veral particles.
c V
meso =V totalÀV micro,S meso=S BETÀS micro.
90X.Zhu et al./Journal of Catalysis271(2010)88–98
respectively.With increasing desilication verity from DS1to DS4,it is obrved that the peak ascribed to strong acid sites gradually los its intensity and shifts to lower temperatures.The changes are accompanied by a gradual increa in the intensity at interme-diate temperatures ($250°C),but little change in the peak as-cribed to weak acid sites.The changes either indicate that desilication converts some of the strong Brønsted acid sites into sites of weaker acidity or modifies the accessibility of the sites due to the partial removal of silica.
The advantage of the IPA-TPD method is that it can be ud to lectively quantify Brønsted acid sites that catalyze the conver-sion of IPA into propylene and NH 3[42–45].For the undissociated IPA (Fig.5a),upon desilication,the intensity of the peak at $190°C becomes smaller than that for the parent sample;at the same time,a small peak at higher temperatures ($250°C)gradually develops十二生肖的来历
with increasing desilication verity.The former peak could be ascribed to desorption of the undissociated amine from weak acid sites of Si A OH.The reduction in intensity is probably due to the removal of internal Si A OH sites (silanol surface defects).This explanation is in good agreement with previous IR measure-ments that showed that internal Si A OH sites were removed u
pon desilication [25,33,46].A new peak develops at higher tempera-tures with increasing desilication verity.Desorption from the sites occurs at relatively high desorption temperatures,but they are unable to catalyze the IPA decomposition [42–45].Infrared measurements of pyridine adsorption have shown an increa in the density of Lewis acid sites due to dealumination upon vere desilication [29,31,46].Accordingly,this desorption peak may be associated with the prence of tho sites.In all samples,the Brønsted sites produced propylene desorption peaks centered at 351°C (Fig.5b),indicating that the density (and possibly
strength)
Fig.2.SEM micrographs of parent zeolite sample P (a–c)and desilicated sample DS2(d–f).
X.Zhu et al./Journal of Catalysis 271(2010)88–9891
of Brønsted acid sites is not significantly changed despite the high desilication verity.The NH3peaks appear slightly later than tho of propylene(Fig.5c)due to re-adsorption/desorption,as previously indicated[42–45].The NH3-TPD and IPA-TPD results are summarized in Table2.The Brønsted acid density estimated from the TPD measurements is in good agreement with the nomi-n
al density of HZSM-5corresponding to a Si/Al ratio , 0.22mmol/g).Both techniques show that the total acid density and Brønsted acid density remain largely unchanged upon desilica-tion.A slight increa for mild desilication followed by a slight de-crea with increasing desilication verity ems to be detected by both techniques,but the changes are very small.
3.1.3.Diffusivity
Fig.S1(in Supplementary content)shows the evolution profiles resulting from nding an n-butane pul through the reactor with and without a zeolite bed.The prence of the zeolite significantly changes the shape and width of the obrved peak as a result of the combination of adsorption and diffusion through the zeolite bed.It is obrved that,upon mild desilication,the elution peak initially shifts to longer retention time and becomes wider(compare P and DS1),but it shifts back to shorter times and becomes narrower as the degree of desilication increas(compare DS2and DS4).The apparent diffusivity derived from applying the dispersion model for deviations from plugflow to the data[47]are summarized in Table3.Details of thefitting method are included in the Supple-mentary content.It is en that the apparent diffusivityfirst de-creas slightly(DS1)but then increas gradually(DS4). Incread diffusivities upon desilication have been previously re-ported[26,29].But,at the same time,decreas in diffusivity have also been reported[30].While incread diffusivities can
be ex-pected from enhanced mesoporosity,decreas have been ascribed to enhanced Al concentration in the desilicated zeolite that affects adsorption and may lower diffusivity.
However,the adsorption effects may not be as determinant of the overall zeolite performance as the changes in diffusion path length.For example,Gobin et al.[48]have shown that in MFI crys-tals(3–5l m),the rate-determining step of the overall transport is intracrystalline diffusion.The situation ems to be the same in the prent study,in which relatively large crystallite sizes have been ud and very minor changes in acid density have been obrved after desilication.In fact,at the larger extents of desilication(sam-ples DS2to DS4)significant increas in apparent diffusivity are en,and they correspond well with the increa in mesoporous volume accompanied by a shortening in the diffusion path length.
Table2
Total acid density(A total)and Brønsted acid density(A B).
Sample P DS1DS2DS3DS4
A total(mmol/g)0.260.290.290.270.27
A B(mmol/g)0.200.230.240.220.22 A total was derived from NH3-TPD,and A
B was derived from and IPA-TPD.Table3
Apparent diffusivity(D)for n-butane.
Sample P DS1DS2DS3DS4 D(Âl0À10m2/s) 5.46 4.93 5.617.939.44
92X.Zhu et al./Journal of Catalysis271(2010)88–98
The slight decrea in apparent diffusivity for the mildly desilicat-ed DS1sample might be ascribed to the increa in microporous volume combined with the minor increa in accessibility of adsor-bate to the micropore structure.
三亚必去的几个景点顺序3.2.Catalytic activity
3.2.1.Effect of time on stream
Fig.6shows the evolution of propanal conversion with time on stream.It can be en that after a few hours under identical reac-tion conditions,the propanal conversion obtained on the desilicat-ed samples is significantly higher than on the original sample P. The order of conversion after a few hours on stream is DS1%DS2>DS3%DS4>P.To distinguish the effects of catalyst stability from level of activity,sample P was run at a higher W/F (0.625h,dotted line in Fig.6)so for thefirst couple of hours the le-vel of conversion was similar to that of DS1.However,after that, the conversion dropped much more rapidly than over DS1.There-fore,it appears that the desilicated samples exhibit not only a high-er activity,but also a higher stability.
Fig.7shows the variation of product yields as a function of con-version for varying time on stream at constant W/version decreasing as the catalyst deactivates.A significant change in prod-uct distribution upon desilication is clearly evident.While the yield of aromatics greatly incread with conversion for all samples,an important difference is obrved for the production of C4-9al-kane/alkene(mainly C5+)products compared to C3.While the light hydrocarbon fractions(C1-2and C3)follow a similar trend to that of aromatics,that is,increasing with propanal conversion,the C4-9 alkane/alkene fraction exhibits a maximum at intermediate prop-anal conversions.Products in the C4-9fraction are converted to C1-3and aromatics as propanal conversion increas.With increasing desilication verity,the yields of C1-2,C3,and aromat-ics decrea while the yield of C4-9alkane/alkene increas.Table4 shows the product distribution for all samples at a propanal con-version of$90%.It is evident that desilication caus a decrea in the fraction of C1-2hydrocarbons,toluene,and p-xylene,but an increa in C4-9alkane/alkene and C10+aromatics.A noticeable difference in the product distribution obrved in this study com-pared to studies conducted at lower temperatures is the lack of products containing oxygen.Even at the lowest conversion levels, no oxygenates other than unconverted propanal were obrved in significant amounts.
3.2.2.Effect of varying space time(W/F)
The product distribution as a function of W/F was compared for the parent sample(P),the mildly desilicated sample(DS1),and the highly desilicated sample(DS4).Fig.8shows the change in prop-anal conversion with increasing W/F,which is higher for DS1and DS4than for P over the entire W/F range.As discusd later,the higher activity of DS1may be associated with rapid transport of propanal into the micropore channels.Fig.9shows the variation
Table4
Product distribution of propanal conversion over parent and desilicated HZSM-5with
a propanal conversion of90%.
Product yield(mol.%)P DS1DS2DS3DS4DS4-Si a
Non-aromatics
C1-2 6.8 4.8 4.0 3.7 2.6 2.8
C322.022.021.020.218.622.1
C4-921.823.030.228.934.934.4
Aromatics
Benzene0.40.6 1.00.9 1.00.9
Toluene8.68.7 6.3 6.3 5.3 5.7
C8aromatics
Ethylbenzene 1.8 2.0 1.4 1.4 1.2 1.3
m-Xylene12.810.3 6.87.0 5.5 5.8
p-Xylene 1.0 2.0 2.8 2.3 2.5 3.0
o-Xylene0.30.80.80.70.70.8
C9aromatics
Propyl-benzene0.80.80.80.80.80.7施工员实习日记
Methyl-ethyl-benzene7.67.87.27.47.07.3
Trimethyl-benzene 1.6 2.7 3.3 2.2 2.1 2.4
C10+aromatics 5.3 6.5 5.87.17.7 2.2丝带玫瑰花
P
aromatics40.242.236.236.133.830.3
a DS4surface was deposited with SiO
2
to passivate the strong acid sites on the
surface.Reaction conditions:T=400°C,W/F=0.5h.
X.Zhu et al./Journal of Catalysis271(2010)88–9893手机维修工具
