Conversion of wooden structures into porous SiC with shape memory synthesis

Conversion

of

wooden

structures

into

porous

SiC

with

shape

memory

synthesis

Rajnish

Dhiman

,

Victor

Petrunin

,

Kuldeep

Rana

,

Per

Morgen

*

a_{DepartmentofPhysicsandChemistry,UniversityofSouthernDenmark,Campusvej55,DK-5230OdenseM,Denmark}

b_{DepartmentofChemistry,BilkentUniversity,Bilkent,06800Ankara,Turkey}

Received6May2011;accepted23May2011

Availableonline27May2011

Abstract

Synthesisof structuredsiliconcarbidematerialscanbeaccomplishedusingwoodenmaterialsas thecarbonsource,withvarious silicon impregnationtechniques.WehaveexploredthelowcostsynthesisofSiCbyimpregnationofcarbonfromwoodwithSiOgasathightemperatures, whichlargelyretainsthestructureofthestartingwood(shapememorysynthesis).Suitablystructured,porousSiCcouldprovetobeanimportant typeofcatalystsupportmaterial.Shapememorysynthesis(SMS)hasearlierbeentriedonhighsurfaceareacarbonmaterials.Herewehavemade anextensivestudyofSMSoncarbonstructuresobtainedfromdifferenttypesofwood.

#_{2011ElsevierLtdandTechnaGroupS.r.l.Allrightsreserved.}

Keywords:B.Electronmicroscopy;B.X-raymethods;Carbonpreforms;MesoporousSiCstructures;Shapememorysynthesis

1. Introduction

Siliconcarbideembodiessomeveryusefulproperties like low mass density, high-thermal conductivity (350– 490Wm1K1) [1], strong resistance towards oxidation, andhighmechanicalstrengthandhardness,whichmakesit a potential material for many industrial applications [2–4]. Crystalline or amorphous silicon carbide is a semiconductor material, which has some unique electrical and electronic qualities [1,5] facilitating applications at high temperatures, highfrequencies,highfields,andhighpowers[6,7].Materials with such properties and which, additionally, can be easily formed with meso- or nanoporosity, should also be strongly solicitedforapplicationsincatalysis,gasreformingreactions, sensors,biotechnology,fuelcells,forhydrogenstorage,andas hydrogensources[8].

Nowadays synthesizing ceramics and composites from biologicalmaterials[9,10]isattractivebecauseitusesnaturally abundant resources such as wood. Plants are also used as starting materials, and the resulting products, under certain conditions, nearly retain the shape of the original solid

frameworkoftheplants.Moltensiliconimpregnation[11–15]

andsiliconvaporimpregnation[16]havebeenemployedearlier to convert charcoal obtained from wood into SiC. But this processdoesnotretainanyporosityofthewoodenmaterialasit fillstheporeswithmoltensilicon.Silica,fromimpregnationof thewoodwithtetraethylorthosilicate(TEOS),hasbeenusedas silicon precursor [9,17], preserving the porosity. Vogli et al.

[18] converted oak wood to SiC replicas by a reactionwith gaseoussiliconmonoxide.Thisparticulartypeofprocessingof carbonwithsiliconmonoxideiscommonlyknownas ‘‘shape memory synthesis’’. Vogli et al. [18] at present is the only reportedworksynthesizingSiCusingshapememorysynthesis withcarbonobtainedfromwood.Inthepresentwork,wehave extensivelyusedtheshapememorysynthesismethodwiththe carbonstructuresobtainedfromdifferentkindofwoodsforthe synthesisofporousSiCelements.Wereportadetailedstudyof themechanismsandthestructuralandcompositionalproperties of the products, such as crystal structure, composition, and specificsurfaceareas.

The shape memory synthesis method has already been shownfeasibleforproducinghighsurfaceareaSiC structures withvariouskindsofcarbonelementssuchasactivatedcarbon, carbonnanotubesetc.[2,19,20].Alocalizedchemicalreaction takesplacebetweenSiOvaporsandC-atomseverywhereonthe carbonframeworkthatresultsinliberationofCOgasandthe

www.elsevier.com/locate/ceramint

Available online at www.sciencedirect.com

CeramicsInternational37(2011)3281–3289

*Correspondingauthor.Tel.:+4565503529;fax:+4565503529.

E-mailaddress:per@ifk.sdu.dk(P.Morgen).

0272-8842/$36.00#_{2011ElsevierLtdandTechnaGroupS.r.l.Allrightsreserved.}

formationofanearly1:1replicaoftheentirestructureaspure SiC.Duetothe hightemperaturesused, theprocess tendsto producecrystallinematerials.

SiC possess the properties required for a good catalyst material.Agoodcatalyticsupportmaterialhasahighspecific surfacearea,highmechanical strength,highthermal stability and high conductivity. Alumina, silica and a carbon-based materials [2,21,22], are very often used as catalytic support materials,butagainstthecriteriapresentedaboveforchoicesof catalyticmaterials,theyarehavingsomedrawbacks[2].Thus, Moeneetal.[22,23]andLedouxandco-workers[2,19–21,24– 26] have already discussed the large potential in SiC as a supportmaterial. Moeneetal.[22,23]have proventhathigh surfacearea(30m2/g)SiCshowsanexcellentthermalstability undernon-oxidativeconditions.TheyhavethusmeasuredSiC reactivityagainstthoseofsilicaandalumina[23].Duetothe highlyporoushoneycombstructure(asshowninSEMpictures) ofourSiCsamples,andtheirhighspecificsurfaceareainthe range of 11–21m2/g, it should be easy to use the incipient wetnessimpregnationmethodtopreparethe catalystonsuch SiCsupports.Thus,theSiC samplessynthesizedinthiswork from carbon preforms obtained from wood are potentially valuablecatalyst supports. These SiC samples contain many stackingfaultsandstructuralimperfections,asshownwith X-raydiffraction(XRD)andRamantechniques.TheSiCsamples madehereareeasilycrushedintopowderformbymillingand arethereforewellsuitedforproducingmixtureswithtransition metalsinmechanicalalloyingprocesses.Theycouldtherefore beveryusefulforhighvelocityoxygenfuel(HVOF)thermal sprayapplicationstodepositwear-andcorrosionresistiveSiC coatings.Ahighamountofdefectsiscommonlyconsideredto resultinanenhancedchemicalsurfaceactivity[27,28],making iteasiertomanufacturethemetalmatrixcompositesrequired forthisapplication[29]intheformofmetal-oxideand metal-carbidecompoundsofhighstrengths.

2. Experimental

Wooden samples were carved into rectangular blocks of differentsizesroughlyrangingfrom10mmto15mminlength, 7mmto10mm inbreadthand5mmto8mm inheight,and theirinitialmassesweremeasured.Differentkindsofwoodlike

Indianpine(Pinussp.),Indianmango(Mangiferaindica),Silk cotton tree (Bombax ceiba), Indian blackberry (Syzygium cumini), Cutch tree (Acacia catechu), Danish beech (Fagus sylvatica)wereused.Thesamplesweredriedatatemperature of708Cfor2handthenheatedataslowrateof18C/minupto 5008C. During this heating, polyaromatic hydrocarbon polymers likecellulose, hemicelluloseandlignin completely decomposetocarbon.After5008C,thesampleswereheatedto 12008Cattheheatingrateof 58C/minandkeptatthe same temperaturefor6h forfurthercrystallizationandpurification ofthecarbonstructures.Alltheprocessstepswerecarriedout inan argonflow inordertoprevent thewood fromburning. After this pyrolysis process, solidified carbon skeleton structures (also termed as carbon preforms) were obtained with the same shape as in the original wood but somewhat contractedfromtheoriginaldimensionsofthestartingwoodas shown inTable 1.The wood samples were characterized by thermogravimetricanalysis(TGA)inanargonatmospherewith aflowrateof50ml/mintomonitorthechangeinweightwith variationoftemperature.Thecarbonpreformswere character-izedbySEM, XRD,andRamanmicroscopy.

Thesecarbon‘‘preform’’materialswerethenplacednextto auniform1:1molarmixtureofsiliconandsilicapowderinan aluminacrucibleinsideanaluminatubelinedtubularfurnace. The gasinducedreaction processisdone withatemperature programinwhichthemainpartofthereactionhappensinthe temperaturerange of 1400–14508C inan argonflow, witha flowrateofaround250–300ml/min,for 12–15h.

The resulting SiC samples are characterized by X-ray diffraction (XRD) todetect the Si:C stoichiometryand their crystalstructures,usingaSiemensdiffractometerD5000.The samples were also analysed for their topographical micro-structurebyusingaLEO435VPscanningelectronmicroscope (SEM), and the composition was checked by the energy dispersive X-ray technique (EDX) witha RONTEC detector attached to the same SEM. The confirmation of the silicon carbidestructureisalsodonewithaRamanmicroscopefrom Dilor,usingthe514nmAr-ionlaser line.Foranalysesofthe surfacecompositions,X-rayphotoelectronspectroscopy(XPS) is donewith aSPECSPHOIBOS1 100 system. The surface areaofthesampleswascheckedwithaSurfaceAreameterfrom Quantachrome.

Table1

Changesinthedimensionsofthewoodaftercarburization.

Samplename Lossinlength/% Lossinbreadth/% Lossinheight/% Lossinvolume/% Lossinweight/%

IP_3_8_CB 32.9 23.6 30.5 64.4 75.3 IM_3_8_CB 21.1 32.5 26.0 60.6 76.5 IS_3_8_CB 23.7 32.1 27.9 62.7 77.0 IBB_3_8_CB 29.4 37.5 22.8 65.9 79.5 IM_3_8_Si 21.6 26.8 35.2 62.8 76.8 IS_3_8_Si 21.9 34.7 33.9 66.3 79.1 USL1 21.3 22.7 46.1 67.2 77.3 USM1 21.6 37.9 35.2 68.4 78.1 US1 21.4 34.0 32.6 65.0 76.3 US2 22.0 36.1 36.7 68.5 75.5

3. Results anddiscussion

3.1. Conversionofwoodtocarbonpreforms

The wooden structures were pyrolyzed in an argon atmosphere toobtain the solid carbonpreforms, whichwere retained for further processing. Some shrinkage in the dimensions of the wood, and a weight loss, are observed duringthe pyrolysis of wood,as reportedearlier [11,30–33].

Table1showsthepercentagelossindimensions,volumeand weightofsomeoftheselectedsamples.Itisclearlyindicated that wood loses around 20–35% of its length, breadth and height, which makes its volume loss 60–70%. Careful observation of the percentage loss of any of the linear dimensions indicates three categories.The first categoryhas lengthlossesofaround32–35%;the secondhas25–27%and thethirdaround20–22%ineachofthesamples.Thissuggests thatthereissomepreferreddirectionofshrinkagealongaxial, tangentialorradialdirectionsasobservedinearlierworks[11].

Table1showsaweightlossofaround75–80%.Theshrinkage andweightlossescanbeattributedtotheremovalofvolatile components likepolyaromatic hydrocarbon polymers (cellu-lose,hemicelluloseandlignin),whichareknownconstituents ofwood.

3.2. Thermogravimetricanalysis(TGA)

The mass change during the conversion of wood to pure carbon was investigated using thermogravimetric analysis (TGA) as shown in Fig. 1. This measurement shows a significant change in the mass from 2508C to 3508C. The entireTGcurveshowsfourstages.Thefirststageisfromroom temperature to 708C, the second from 200–2508C to 250– 5008Casthe third,andthefourthoneisseenabove6008C. Thelossofmoisturefromthesampleisresponsibleforthemass changeinthefirststage.Thechangefrom200to2508C,ata slowerrate,isduetotheremovalofphysicallyadsorbedwater

andotheradsorbedatmosphericgaseslikenitrogen,oxygenetc. inthe sample. The majormass lossoccurs inthe third stage from2508Cto5008C.Asreportedintheliterature[11,34],this can be understood as the decomposition of hemicellulose, cellulose andlignin,whichare theprime softconstituentsof wood,withrespectivedecompositiontemperaturesof280,340 and 4008C [35].This means thatwood started degrading at 2508C. The major masschange hashappenedbelow 4008C andonlyaslightlossisobservedafterwards.After6008Cthe carbon structure starts rearranging [35] from broken –C–C– chainsofbiopolymeraromaticcarbonstructurestocrystallites ofthecarbonpreform.Thisprocessisresponsiblefortheslow mass change in the fourth stage of the TG curve. There is roughlya75–80%masslossfromthestartingwood,whichis different from the theoretical value of 50% of volatile components in wood and this extra loss may be due to the presence of moisture, adsorbedwater andothergases,which are escapingmainlyinthe firstandsecondphasesofthe TG curve.FromTable1itisclearthatthepercentagemasslossof different samples varies, which is explained by the factthat differentwoodshavedifferentlignintocelluloseratiosandalso different cellularstructures.

3.3. Characterization ofthe carbonpreforms

Scanning electron microscopy (SEM) images of two different carbon preforms obtained from Danish beech and IndianpineareshowninFig.2(aandb),respectively.Flakesof carbon arebrokenoff withthe purpose ofshowing the inner structureofthecarbonpreforms.Thesefiguresclearlyshowthe presenceofcapillariesinsidethesolidframeworkofthecarbon preform,formingahoneycomblikestructure.Differentkinds of woods have different shapes and sizes of pores and capillaries.Danishbeechwoodhastwokindsofporestructure withdifferent sizes asshownin Fig.2(a). Biggerpores have sizesintherangeof20-25mmandsmallerporeshavesizesin therangeof5–8mm.Fig.2(b)showsaSEMimageofIndian pine.IthasrelativelybiggerporesthantheDanishbeechwood. Thissamplewas cutatan obliqueangletothe poreaxes.Its structureisawell-orderedhoneycombstructure.Theporesize incaseofIndianmango,silkcottontree,andIndianblackberry isapproximatelyidenticalandisintherangeof5–10mm(as suggested by Figs. 5(a–c), which are of their SiC replicas). Theseopenstructuresintheinitialcarbonpreformhelpsinthe conversionofcarbontoSiCbyfacilitatingSiOvaporstopass intoandthroughtheporesandtoreactwiththecarbon.These pores are also responsible for increasing the porosity and surfaceareaandreducingthedensityoftheprocessedsamples.

Fig.3(a)showsan X-raydiffractionpatternof the carbon preform. The peaks at 2u=23.68 and 43.88 are indexed as (002)and(101)peaksof 2H-graphite.Fromthe shapeand intensityofthepeaks,anamorphousnatureofthecarbonasa wholeis evidentalong withthe presence of somesmallsize graphitic crystallites. The presence of these small size crystallites isalso confirmedby the presence of aG-bandin the Raman spectrum of this sample; see Fig. 3(b). The crystallinityisduetorearrangementofbroken–C–C–chainsof

Fig.1.TGAofawoodsampleheateduptoatemperatureof10008Cinthe

biopolymeraromaticcarbonstructures,whichhappens above 6008Cas explainedinthediscussion ofthe TGAresults.

Fig.3(b)showsthefirstorderRamanspectrumofthecarbon preformrecordedinthe region between900 and1900cm1. Thepeaksappearingat1590and1325cm1areattributedtoG andD-bandscorrespondingtopeaksincrystallinegraphiteat 1582and1350cm1,respectively[36,37].Thestrongbandat 1590cm1 corresponds to one of the E2g modes (mode G)

representingthe movementof twoneighboringcarbonatoms oppositetoeachotherinagraphenesheet.The1325cm1band is assigned to the D-mode, which is normally not a Raman active mode in graphitic carbon. Its occurrence here is attributedtodefectsanddisorders.TheFWHMoftheG-band isaround99cm1,whichis much higher thanthe width for highly oriented pyrolytic graphite (HOPG) (15–18cm1 range). The Raman spectrum thus shows that the carbon preform contains amorphous carbon and some small size crystallites,inagreementwiththeXRDofthematerialshown inFig.3(a).

3.4. Determinationofmassand density

After carburization of the wooden samples the resulting carbonpreformstructuresweretreatedwiththeshapememory synthesisprocedure.

In the shape memory synthesis,the SiO vaporsreactwith carbontoformSiCasaproduct,whileCO(gas)isliberatedas

byproduct. This reaction occurs in the temperature range of 1400–14508C.Siliconmonooxidevaporsareobtainedbythe reaction of silicon powder with silica powder at the same temperatureasshown:

SiðsolidÞþ SiO2ðsolidÞ!2SiOðgasÞ

Then,

SiOðgasÞ þ 2CðsolidÞfi _{SiCðsolidÞ þCOðgasÞ}

Sincethisreactionisreversibleinnature,thetubeisflushed withstreamingargongastoremovetheCOgassothereaction proceedsintheforwarddirection.Inthisreaction,SiOvapors react randomlywith carbonatoms toform SiC andCO gas, such that exactly half of the carbon atoms are replaced by silicon atoms, which may leave the shape of the preform unchanged. The expected mass (M) and density (D) of the samplesarenowcalculatedwiththisassumptionusingEqs.(2) and(3): MSiC ¼ Mcarbon 2 þ Mcarbon 2  28:0855 12   (2) DSiC¼ Dcarbon 2 þ Dcarbon 2  28:0855 12   (3) From Table 2, we can see how the experimental and calculatedvaluesareingoodagreementwitheachother.Small deviationsinthedensityandmassesofboththecalculatedand

Fig.2. SEMimagesoftwocarbonpreformsobtainedfrom(a)Danishbeechand(b)Indianpinewood.

theexperimentalvaluescanbeattributedtoerrorsinmeasuring thevolume,thepresenceofheavyelementslikecalciumfrom the carbon preform, and the presence of the adsorbed atmospheric gases likenitrogen and oxygen in the samples, as confirmedbythe EDX,XPS,andAugerresults.

Thesamplesretainthesameamountofporosityasispresent in the startingcarbon preform andthey are alsoobserved to haveaspecificsurfaceareaof 11–21m2/g.

3.5. X-raydiffractionanalysis

The crystallinity and polymorphism of the samples were checked with X-ray diffraction. The phase analysis of the diffractionpatternswasdoneusingtheX’pertHighScoreplusTM software.TheRietveldrefinement[38,39]techniquewasusedto obtainaccuratevaluesoflatticeparametersandphasespresentin thecrystalstructures.Thesamplesarefoundtocontainmixtures of a-SiC and b-SiC forms. The software calculates the percentage of different polymorphs in the samples. Table 3

summarizesthepercentagedistributionofdifferentpolymorphs ofSiCinsomeofthesynthesizedsamples.

ThediffractionpatternsofthesamplesareshowninFig.4. ConsideringthebackgroundintheXRDpatterns,wecansay thatthesesamples are lesscrystallinethan samples,we have

Ta b le 2 Cal culated an d ex per imenta l v alues of mass and de nsity of the samples synt hesized by SMS. Sam ple name Mass of carb on/g D ensity of carb on/g cm  3 Calcluated mass of SiC/g Experi men tal mass of Si C/g De viatio n from calcul ated v alue of mass/% Exper imenta l v alue of density /g cm  3 Calc ulated v alue of densi ty/g cm  3 De viation fro m calc ulated v alue of densi ty/% IM _21_7 0.20 49 0.29 0.34 22 0.33 32 2.6 0.46 7 0.48 0 2.6 IM _28_7 0.15 49 0.33 0.25 87 0.25 09 3.0 0.55 8 0.54 9  1.6 IM _3_8_S i 0.06 96 0.31 0.11 62 0.12 00  3.2 0.52 6 0.51 3  2.5 IM _18_11-1 0.11 61 0.33 0.19 39 0.18 18 6.3 0.53 0.55 7 4.9 IP_3 _8_Si 0.17 25 0.45 0.28 81 0.28 16 2.3 0.72 0.75 3 4.4 IS_3 _8_Si 0.03 54 0.25 0.05 91 0.05 8 1.9 0.38 0.41 5 8.5 IBB _3_8_S i 0.06 87 0.38 0.11 47 0.11 75  2.4 0.64 6 0.63 8  1.4 IBB _28_7 0.14 07 0.44 0.23 50 0.23 14 1.5 0.72 4 0.74 3 2.5 D B_18_1 1-1 0.12 17 0.35 0.20 33 0.20 05 1.6 0.09 3 0.09 6 3.0 D B_28_7 0.21 53 0.34 0.35 96 0.33 41 7.1 0.55 6 0.57 5 3.2 Table3

PercentageconcentrationofdifferentpolymorphsofSiCfromXRD.

No. Samplename Cubic 2H 4H 6H 27R 29R

1 DB_18_11_9-2 37.5 62.5 2 IM_3_8_CB 43.4 56.6 3 DU_18_11-1 38.6 61.4 4 IBB_28_7 40.3 59.7 5 IBB_3_8_CB 25.2 73.1 1.7 6 IS_3_8_Si 59.4 40.6 7 IM_28_7 40.4 52.7 6.9 8 IP_3_8_SI 38.3 61.7 9 IBB_28_7 40.3 59.7

madewiththeliquidsiliconimpregnationmethodinaparallel study.The presentsamplesthuslargelyretainthestructureof theoriginalwood,i.e.anisotropicandamorphousandresultsin a higher XRD background in comparison to the samples synthesized by liquid silicon impregnation. The diffraction patternsofsamplesshowthemajorpeaksat2u=35.58,59.88 and71.68,which correspondto diffractionfrom SiC (111), SiC(220)andSiC(311)planes.Smallshouldersat2u=338 and37.58arealsofoundwhichcorrespondstoa-SiC.Fromthe resultsinTable3,wecanclaimthattheSMSalwaysproducesa mixtureofa-SiC(2H,4H,6H,27R,29Retc.)andb-SiC(3C). Thegrainsizeofthesamplesvariesintherangeof43–68nm andthestrainvariesfrom0.038to0.1%,ascalculatedbyusing theWilliamson–Hallplot[40]from theXRDdata.

3.6. Surfaceanalysisofthe samples 3.6.1. SEMandEDXanalyses

TheSEMimagesofthesamplesareshowninFig.5,which after comparison with Fig. 2 suggests that they retain the morphologyofthestartingcarbonpreforms.Porestructureand poresizeof the samplessynthesized fromIndian blackberry, andsilkcottontreeisalmostthesameasinFig.5(aandb).The sampleswerecutperpendiculartotheporeaxesinFig.5(aand b).Fig.5(c)showstheSEMimageoftheIndianmango.This samplewascutparalleltotheporeaxestoshowthecapillary arrangement. Comparison of the SEM image of the carbon preformobtainedfromDanishbeechwoodinFig.2(b)withthe SiC replica (Fig. 5(d)) confirms that the shape memory synthesis method preserves the initial shape, size and morphology of the carbonpreform. Siliconcarbide whiskers havealsobeen observedoverthe topsurface ofthe samples,

whicharesimilartotheonesobservedin[41],wheretetraethyl orthosilicatehasbeen usedas thesilicaprecursor.

TheEDXanalysisshowsSi,C,O,andCa.Calciumwasalso found in the carbon preform of beech wood, taken up as nutrition [42]. A 6H–SiC wafer reference sample is used to calibratetheEDXsystem.Thecalculatedatomicpercentages of the synthesized sample were 39.50.4% for silicon, 45.60.4% for carbon, 10.80.3%for oxygen and 4.00.2%forcalcium.

Only2.5%oxygenisdetectedonthe6H–SiCwafer,while thesynthesizedsampleshows10.8%.Thissuggeststhatoxygen is more prone to adsorb in the porous framework of the synthesized samples at high temperatures. Mapping of elements with EDX shows homogeneous distributions of siliconandcarbonoverthewholesample.

3.6.2. XPSanalysis

XPSspectraofdifferentsamplesareshowninFig.6.They are recordedfrom 200eV to1260eVkineticenergy,excited with an MgX-ray source (Mg Ka).They all show peaks of

carbon,oxygenandsiliconalongwithsomeotherpeakssuchas (Ar 2p) and (N 1s) seen for the 6H–SiC wafer acting as a reference sample. This reference sample had initially been sputtered with argon for 30min. There are also traces of calcium(0.5%concentration)inthe samples,whichcomes fromtheinitialwoodenprecursor[42].Allsampleshavebeen analysed forthe presenceof carbon,silicon andoxygen.The samplesgetoxidizedduetothehandlinginairformanydays afterthesynthesis.Thehandlingofthereferencesampleinthe open air has also oxidized the 6H–SiC wafer, to a similar degree. The ratio of Si:C in the different samples scatter somewhat. The surface region (as measuredby XPS) of the

Fig.5.ScanningelectronmicroscopyoftheSiCsamplesmadefrom(a)Indianblackberry,(b)silkcottontree,(c)Indianmango,and(d)Danishbeechwood

samplesiscarbonrich exceptfor onesamplewhichisrather showingapproximately1:1ratioofSi:C.Thegeneralshapeof thewidescanXPSspectravarieswiththedensityandstructure of the samples. The inelastic background regions show a dependence on density andmorphology. Thus, the reference 6H–SiCsamplehastheflattestbackgroundbelowtheO1sand OKLLpeaks(at720eV and500eV),whiletheleastdenseand mostporoussamplesmadebyshapememorysynthesisshow the highest background slopes. Different background slopes below the oxygen signals (lower kinetic energies) indicate differences in the in-depth distributions of oxygen in the samples.Higherslopesindicatedeeperpenetrationofoxygen. FromtheXPSresultsinFig.6,weseethatthetopsurfaceofthe SiCsamplesisoxidizedas(amorphous)SiO2.Thesynthesized

samplesareallthetimekeptintheopenaftertheirsynthesisin the non-reactive argon atmosphere (for comparison, a differentlypreparedstandardreferencesamplewithanominal 1:1 stoichiometrygives aSi:Csurface concentration ratioof 56%:44%,usingtheCasaXPSTMsoftware).

Thus all samples studied here are definitely enriched in carbonatthesurface,asreceived.Thismightbeduetocarbon contamination during handling, insertion and storage of samplesinultrahighvacuumconditions.

3.7. Ramananalysis

Ramanspectroscopyisoneofthemostsensitivetechniques fordistinguishingbetweenthedifferentpolymorphictypesof SiC[43,44].Siliconcarbideisknowntohavealargenumberof

polymorphs,ofwhichb-SiChasazincblendekindofstructure withthesmallestunitcellofalltheSiCpolymorphs.Bulk b-SiChastwoopticalmodesattheGpointoftheBrillouinzone,a transverse optic (TO) mode at 796cm1 and a longitudinal optic(LO)modeat972cm1[44].Thesemodeshavedifferent energies, which in ‘‘higher’’ polytypes can be used to distinguish the different polymorphs. We worked with the most commonly used 514nm laser line for the Raman measurements and then compared our spectra with those in theliterature[43]regardingboththepositionofRamanpeaks andtheirintensitydistributions.

Fig. 7 shows the Raman spectra of the samples; the corresponding peak positions are given in Table 4. The full widthsat halfmaximum(FWHM) for theseRamanlinesare ratherbroad,intherangeof35–48cm1,whichsuggeststhat they are havinglargenumber of stacking faults,as expected fromthelocalpoly-anisotropicnatureofthestartingwood.Due to this large FWHM we cannot determine the exact peak positionofeachsampleandcannotclaimtheexactpolymorph typewithcertainty,butsimpleassociation,inthetable,makes somesuggestionsobvious.InTable4,wecanseethatthereare differences in the positions of the peaks and from the comparison with literature [43,44] it is suggested that these can be from 2H or 4H–SiC. These assignments are also supportedbycomparisonwiththeXRDresultsfortheamounts of differentpolymorphsinthe samplesas showninTable3.

Fig.6. XPSspectraofdifferentsamplesalongwiththe6H–SiC reference

sample.Theinsettableshowsthepercentageconcentrationsoftherespective

samplesascalculatedbytheCasaXPSTMsoftware.

Fig.7. Ramanspectraofthesamples.

Table4

PeakpositionsintheRamanspectraofthesamplesandrespectivesuggestionsforthedifferentpolymorphs.

Samplename Peakpositions/cm1(4cm1) Suggestionsfor

polymorph/cm1[43,44] IP_28_7 770.2,886.3,930.2 2Hor4H IP_18_11_2 770.7,861.5,903.4,935.9 2Hor4H DB_18_11-1 777.8,896.7,956.8 4H(776&964) IM_18_11_2 766.9,897,934 2H(764&968) IM_21_7 763.6,874.7,917.5 2H(764&968)

Thus from the aboveresults we can say that the samples producedbytheshapememorysynthesismethodareformedin differentphaseswithrespectivepercentagesobtainedfromthe XRDanalysis. Thehigher FWHM of thepeaks indicates the higheramountof stackingfaults andstructuralimperfections thatmayresultinenhancedsurfaceactivities.

3.8. Specificsurfacearea

Thespecificsurfaceareaofthesampleshavebeenmeasured using the BET adsorption isotherm method. Initially, the samples are evacuated and out gassed at high temperature (3508C). After cooling them down to liquid nitrogen temperature, nitrogen gas at variable pressures is used to obtainsixpointsontheBETadsorptionisotherm,fromwhich thespecificsurfaceareaiscalculated.Thespecificsurfaceareas ofthedifferentsamplesvaryintherangeof11.82–21.16m2/g.

Table5 shows thespecific surfacearea of the samples made from different woods. This table suggests that the specific surfaceareaofsamplesmadefromallthesedifferentkindsof woodisalmostcomparable.Thisisfollowedbythefactthatthe densityofcarbonobtainedfromthesedifferentwoodsisalso nearlyequal(varying from0.29 to0.35gcm3)as shownin

Table2.The ‘‘nitrogen’’ porevolumeof one of the samples synthesizedfromDanishbeechwoodis0.14ml/cm3whilepore volume determination using water until external wetting is perceived,shows a porevolume of 1.1ml/g. Thus it can be concludedthatSiC ceramicelements synthesized fromthese particularkindsofwoodsappearwithapproximatelyidentical surface areas and cellular structures. With SMS the silicon atomsarethusreplacinghalfofthecarbonatomsandleaving theinitialcarbonframeworkintactinshape.Thismethodgives ahigherspecificsurfaceareathanwhenthesamesamplesare synthesizedbyimpregnationwithliquidsilicon.

4. Conclusions

Porous SiC elements are successfully synthesized with naturalwoodasthestartingmaterialusingtheshapememory synthesis procedure. The SiC structures obtained contain mixturesofa-SiCandb-SiC.Thedensityoftheporoussamples is roughlyaround 6 timesless than the density of bulk SiC (3.21gm/cm3),showingthattheypreservetheporousstructure. Thesamples madebyshape memorysynthesishave obvious applicationsascatalysissupportmaterialsduetotheirporous natureandrelativelyhighspecificsurfaceareas(11–21m2/g). Sincethe SMSprocessed materialshave structural

imperfec-tions,whichenhancetheirsurfaceactivity,andmakeiteasyto crushthemintopowderform,theymayalsobeusedinSiC– transition metal matrix alloys,made by mechanical alloying during millingof amixtureof SiC powderandmetal. These alloyed particlesare predictedtobe highly usefulfor HVOF thermalsprayapplications,andinthiswaybeusedtodeposit high endurance, wear- and corrosion resistant SiC surface coatings.

Acknowledgements

ThisworkissupportedbytheDanishMinistryforResearch andInnovation,throughitsprogramforSustainableEnergyand TheEnvironment,withthegranttoR.Dhiman,2104-05-0073. Theauthorsaregratefulforadviceandtechnicalsupportfrom E.Skou,S.Tougaard,T.Warner,P.B.Hansen,D.Kyrping,and T.SørensenatSDU;toN.DamMadsen,andP.Hald,fromThe University of Aarhus, to Jens Rafaelsen from Aalborg University,andtoS.VikramSingh fromThe RISØNational Laboratory,Roskilde,Denmark.

References

[1] Z. Li, Z. Zhang, A. Meng, J. Guo, Large-area highly-oriented SiC

nanowirearrays:synthesis,Raman,andphotoluminescence properties,

J.Phys.Chem.B110(2006)22382–22386.

[2] J.M.Nihut,R.Vieira,L.Pesant,J.P.Tessonnier,N.Keller,G.Ehret,C.P.

Huu,M.J. Ledoux,Synthesisandcatalyticusesofcarbonand silicon

carbidenanostructures,Catal.Today76(2002)11–32.

[3] J.A.Powell,L.G.Matus,M.A.Kuczmarski,Growthandcharacterization

ofcubicSiCsingle-crystalfilmsonSi,J.Electrochem.Soc.134(1987)

1558–1565.

[4] W.J.Choyke,G.Pensk,PhysicalpropertiesofSiC,Mater.Res.Soc.Bull.

22(1997)25–29.

[5] Z.Ju,X.Ma,N.Fan,P.Li,L.Xu,Y.Qian,High-yieldsynthesisof

single-crystalline3C–SiCnanowiresbyafacileautoclaveroute,Mater.Lett.61

(2007)3913–3915.

[6] A.Fissel,B.Schroter,W.Richter,LowtemperaturegrowthofSiCthin

filmsonSiand6H–SiCbysolidsourcemolecularbeamepitaxy,Appl.

Phys.Lett.66(1995)3182–3184.

[7] Z.C.Feng,A.J.Mascarenhas,W.J.Choyke,J.A.Powell,Ramanscattering

studiesofchemicalvapordepositedcubicSiCfilmsof(100)Si,J.Appl.

Phys.64(1998)3176–3186.

[8] G.Mpourmpakis,G.E.Froudakis,SiCnanotubes:anovelmaterialfor

hydrogenstorage,NanoLett.6(2006)1581–1583.

[9] A.V.Srinivasan,G.K. Haritos, F.L.Hedberg, Biomimetics:advancing

man-madematerialsthroughguidancefromnature,Appl.Mech.Rev.44

(1991)463–481.

[10] Y. Shin, C. Wang, G.J. Exarhos, Synthesis of SiC ceramics by the

carbothermal reduction of mineralizedwoodwith silica, Adv. Mater.

17(2005)73–77.

[11] P. Greil, L. Thomas,A. Kaindl, Biomorphic cellular silicon carbide

ceramicsfromwood.I.Processingandmicrostructure,J.Euro.Ceram.

Soc.18(1998)1961–1973.

[12] P. Greil, L. Thomas,A. Kaindl, Biomorphic cellular silicon carbide

ceramicsfromwood.II.Mechanicalproperties,J.Euro.Ceram.Soc.8

(1998)1975–1983.

[13] T.Ota,M.Takahashi,T.Hibi,M.Ozawa,S.Suzuki,Y.Hikichi,

Biomi-meticprocessforproducingSiCfromwood,J.Am.Ceram.Soc.78(1995)

3409–3411.

[14] O.P.Chakrabarti,H.S.Maiti,R.Majumdar,Si–SiCceramicsfromplant

precursor,J.Mater.Sci.39(2004)4715–4717.

Table5

Specificsurfaceareaofthesamples.

Sr.No. Samplename Typeofstartingwood Surfacearea(m2/g)

1 IM_3_8_CB Indianmango 11.82

2 IM_7_10 Indianmango 17.01

3 IP_28_7 Indianpine 18.74

4 DB_18_11_1 Danishbeach 19.66

[15] D.W.Shin, S.S. Park,Y.H. Choa, K.Niihara, Silicon/silicon carbide

compositesfabricatedbyinfiltrationofasiliconmeltintocharcoal,J.

Am.Ceram.Soc.82(1999)3251–3253.

[16] E.Vogli,H.Sieber,P.Greil,BiomorphicSiC–ceramicpreparedby

Si-vaporphaseinfiltrationofwood,J.Euro.Ceram.Soc.22(2002)2663–

2668.

[17] J.Locs,L.B.Cimdina,A.Zhurish,D.Loca,Optimizedvacuum/pressure

solimpregnationprocessingofwoodforthesynthesisofporous,

biomor-phicSiCceramics,J.Euro.Ceram.Soc.29(2009)1513–1519.

[18] E. Vogli, J. Mukerji, C. Hoffman, R. Kladny, H. Sieber, P. Greil,

Conversion of oakto cellular silicon carbide ceramicby gas-phase

reactionwithsiliconmonoxide,J.Am.Ceram.Soc.84(2001)1236–

1240.

[19] N.Keller,O.Reiff,V.Keller,M.J.Ledoux,Highsurfacearea

submic-rometer-sizedb-SiCparticlesgrownbyshapememorysynthesismethod,

DiamondRelat.Mater.14(2005)1353–1360.

[20] N.Keller,V.Keller,F.Garin,M.J.Ledoux,AnewTiO2–b-SiCmaterial

foruseasphotocatalyst,Mater.Lett.58(2004)970–974.

[21] N.Keller,C.P.Huu,S.Roy,M.J.Ledoux,Influenceofthepreparation

conditions on the synthesis of high surface area SiC for use as a

heterogeneouscatalystsupport,J.Mater.Sci.34(1999)3189–3202.

[22] R.Moene,M.Makkee,J.A.Moulijn,Highsurfaceareasiliconcarbideas

catalystsupportcharacterizationandstability,Appl.Catal.A:Gen.167

(1998)321–330.

[23] R.Moene, E.P.A.M.Tijsen,M.Makkee,J.A.Moulijn,Synthesisand

thermalstabilityofNi,Cu,Co,andMocatalystsbasedonhighsurface

areasiliconcarbide,Appl.Catal.A:Gen.184(1999)127–141.

[24] C.P.Huu,C.Bouchya,T.Dintzer,G.Ehret,C.Estournes,M.J.Ledoux,

Highsurfaceareasiliconcarbidedopedwithzirconiumforuseascatalyst

support.Preparation, characterizationand catalytic application, Appl.

Catal.A:Gen.180(1999)385–397.

[25] M.Benaissa,P.H.Cuong,J.Werckman,C.Crouzet,M.J.Ledoux,A

sub-nanometerstructuralstudyofPt–RhcatalystssupportedonCedopedSiC,

Catal.Today23(1995)283–298.

[26] P.H.Cuong,S.Marin,M.J.Ledoux,M.Wiebel,G.Ehret,M.Benaissa,E.

Peschiera,Synthesisandcharacterizationofplatinum–rhodiumsupported

onSiCandSiCdopedwithcerium:catalyticactivityfortheautomobile

exhaustreactions,Appl.Catal.BEnviron.4(1995)45–63.

[27] C.Jia,Z.Li,Z.Xie,AresearchondetonationguncoatingwithFe–SiC

compositepowdersmechanicallyactivated,Mater.Sci.Eng.A263(1999)

96–100.

[28] H. Ferkel,Propertiesof copper reinforced by laser-generatedAl2O3

-nanoparticles,Nanostruct.Mater.11(1999)595–602.

[29] J.Pelleg,ReactionsinthematrixandinterfaceoftheFe–SiCmetalmatrix

compositesystem,Mater.Sci.Eng.A269(1999)225–241.

[30] C.E.Byrne,D.C.Nagle,Carbonizationofwoodforadvancedmaterials

applications,Carbon35(1997)259–266.

[31] T. Hata, Microstructural analysis ofbiocarbon from wood, Cellulose

Commun.6(1999)127–133.

[32] O.Paris,C.Zollfrank,G.A.Zickler,Decompositionandcarbonisationof

wood biopolymers – a microstructural study of softwood pyrolysis,

Carbon43(2005)53–66.

[33] G.A.Zickler,W.Wagmaier,S.S.Funari,M.Burghammer,O.Paris,Insitu

X-raydiffractioninvestigationofthermaldecompositionofwood

cellu-lose,J.Anal.Appl.Pyrolysis80(2007)134–140.

[34] P. Ehrburger, L. Lahaye,E. Wozniak, Effect of carbonization onthe

porosityofbeechwood,Carbon20(1982)433–439.

[35] F.Shafizadeh, Y.Sekiguchi,Developmentof aromaticityin cellulosic

chars,Carbon21(1983)511–516.

[36] G.Maurin,F.Henn,B.Simon,J.F.Colomer,J.B.Nagy,Lithiumdopingof

multiwalledcarbonnanotubesproducedbycatalyticdecomposition,Nano

Lett.1(2001)75–79.

[37] H.Hiura,T.W.Ebbesen,K.Tanigaki,H.Takahashi,Ramanstudiesof

carbonnanotubes,Chem.Phys.Lett.202(1993)509–512.

[38] C.L.Burdick,E.A.Owen,Theatomicstructureofcarborundum

deter-minedbyX-rays,J.Am.Chem.Soc.40(1918)1749–1759.

[39] P.T.B.Shaffer,Areviewofthestructureofsiliconcarbide,Acta

Crystal-logr.B25(1969)477–488.

[40] G.K.Williamson,W.H.Hall,X-raylinebroadeningfromfiledaluminium

andwolfram,ActaMetall.1(1953)22–31.

[41] R.Dhiman,P.Morgen,GrowthofSiCnano-whiskersonpowderedSiC,

nanotechnologicalbasisforadvancedsensors,NATOScienceforPeace

and Security Series B: Physics and Biophysics (2011) 281–285. ,

doi:10.1007/978-94-007-0903-4_31, Part8.

[42] J.Fromm,Woodformationoftreesinrelationtopotassiumandcalcium

nutrition,TreePhysiol.30(2010)1140–1147.

[43] S.Nakashima,M.Higashihira,K.Maeda,Ramanscattering

characteri-zationofpolytypeinsiliconcarbideceramics:comparisonwithX-ray

diffraction,J.Am.Ceram.Soc.86(2003)823–829.

[44] M. Bechelany,A. Brioude,D.Cornu, G. Ferro,P. Miele,A Raman

spectroscopystudyofindividualSiCnanowires,Adv.Funct.Mater.17