Synthesis, characterization, and wear and friction properties of variably structured SiC/Si elements made from wood by molten Si impregnation

(1)

Availableonlineatwww.sciencedirect.com

JournaloftheEuropeanCeramicSociety32(2012)1105–1116

Synthesis,

characterization,

and

wear

and

friction

properties

of

variably

structured

SiC/Si

elements

made

from

wood

by

molten

Si

impregnation

Rajnish

Dhiman

a

,

Kuldeep

Rana

b

,

Erman

Bengu

b

,

Per

Morgen

a,∗

a_Department_of_Physics_and_Chemistry,_University_of_Southern_Denmark,_Campusvej_55,_5230,_Denmark b_Department_of_Chemistry,_Bilkent_University,_Bilkent,₀₆₈₀₀_Ankara,_Turkey

Received28July2011;receivedinrevisedform15November2011;accepted21November2011 Availableonline16December2011

Abstract

Wehavesynthesizedpre-shapedSiC/Siceramicmaterialelementsfromcharcoal(obtainedfromwood)byimpregnationwithmoltensilicon,which takesplaceinatwo-stageprocess.Inthefirstprocess,aporousstructureofconnectedmicro-crystalsof␤-SiCisformed,while,inthesecond process,moltenSitotallyorpartlyinfiltratestheremainingopenregions.Thisprocessformsadensematerialwithcubic(␤-)SiCcrystallites,of whichthemajorityisimbeddedinamorphousSi.Thesynthesisofpreshaped“sprocket”elementsdemonstratesthatdesiredshapesofsuchadense SiC/Sicompositeceramicmaterialcanbeachieved,thussuggestingnewindustrialapplications.

Thestructureandcompositionofnumerousas-synthesizedsampleswerecharacterizedindetailbyusingawiderangeoftechniques.Wearand frictionpropertieswerealsoinvestigated,withpolishedsamples.Thepropertiesfoundforthepresentsamplesareverypromisingforabrasive applicationsandfornewgenerationbrakesystems.

Keywords:SiC;Precursors-organic;Composites;Electronmicroscopy;Structuralapplications

1. Introduction

Silicon carbide is among the most sought ceramic and semiconducting materials for industrial applications because of its low mass density, high-thermal conductivity (350–490Wm−1K−1),1 strong resistance towards oxidation, andhighmechanicalstrength.Siliconcarbideisalsoawideband gap(2.39–3.33eV)2semiconductingmaterialwithhighelectron mobility,highbreakdownelectricfieldstrength(3–5MV/cm), initscrystallineforms,1aswellasahighoxidationresistance, andetch-resistantproperties,makingit eminentlysuitable for electronicandopticalapplications athightemperatures, high frequencies,andhighpowers.3Suchqualitiesmakeit applica-ble withdifferent functions inintegratedbiomaterials and in light-weight/high-strengthstructures.4

During the last two decades, biological materials5–8 have been favoured as raw materials for synthesizing engineering ceramicsandcomposites.Naturalmaterialslikericehulls,9 cot-tonfibers10andwoodsfromdifferentkindoftrees11–20havethus

∗_{Corresponding}_author._Tel.:₊₄₅₆₅₅₀_3529;_fax:₊₄₅₆₆₁₅_8780.

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

beenusedasthestartingmaterial.Theadvantageofusingwood asstartingmaterialisitstendencytoretaintheoriginalshape. ThusSiCcomponentshavebeensuccessfullysynthesizedfrom woodusingmethodssuchasmoltensiliconimpregnation,10–16 silicon vapour impregnation,17 carbothermal reduction after tetraethyl orthosilicate (TEOS) impregnation,6,18 and shape memorysynthesisusingSiOvapour impregnation.19,20In the moltensiliconimpregnationmethod,moltensiliconisforcedto penetrateintotheporechannelsystemofpreviouslycarburized woodbycapillaryforcesandspontaneouswetting,undergoing alocalsolid–liquidreactiontoform␤-SiC.Thismethodnearly retainsthe shapeof the startingwood asacrystallized struc-ture,inadditiontothefillingorcappingofporesandvoidswith silicon.Thus,thesampleproducedinthiswayisactuallya more-or-lessdense,compositestructureofsiliconandSiC.Itisvery importanttounderstandthesynthesismechanismoftheseSiC/Si ceramics.Despitealotofworkalreadyreportedonmolten sili-conimpregnationofcarbonobtainedfromwood,thereareonly afewattempts13,21reportedstudyingthedetailsofthe forma-tion mechanismofSiC inthe ceramicbutnoconclusionsare givenfor theentirecompositesystem(SiC/Si)andtheroleof thesiliconforthepropertiesoftheresultingmaterial.Itisalso ofinteresttostudythewearandfrictionpropertiesofpreformed

(2)

1106 R.Dhimanetal./JournaloftheEuropeanCeramicSociety32(2012)1105–1116

SiC/Sistructurestotesttheirapplicabilityforfrictionand abra-sionpurposes,asthesewouldnotrequirepost-synthesisforming orshapingprocesses.Therearenumerousstudiesonwearand frictionpropertiesof conventionallymade SiC ceramics.22–29 However,veryfewreportshavefocusedonthewearproperties ofSiC/Sicompositessynthesizedfromnaturalproducts.30

Inthiswork,thesiliconimpregnationmethodhasbeenused to fabricate “sprocket shaped” SiC/Si structures from wood. Thus it is realized, how this low cost synthesis of variably structuredSiC/Sisampleswithpredefinedshapescouldmake themusableinanyshapes,asinhighlyabrasiveelementsfor theautomotiveindustry,i.e.inadvancedbrakesystems. Simi-larly,theycouldbeseenasalternativestoaluminawhenused forhightemperatureapplications,whereespeciallytheirhigh thermal conductivity could be explored. Thissynthesis route couldalso beseenas animportant methodformaking struc-tural(predefinedshapes)partsofSiCcomposite materialsfor nuclearfusionreactorsandinheatengines.31–33Wehave care-fullyexplored thereaction mechanismsandcannow confirm that the synthesis of SiC/Si ceramic is a two-stage process (from a pure carbon preform), where we can point out the important role of silicon for the ceramic material formation, andfor its finalproperties. Wehaveanalysed manyproducts formedfromwoodenpreformsbySi-impregnation,usingX-ray diffraction(XRD)forcrystalstructuredetermination,scanning electronmicroscopy(SEM)for imagingthesurface morphol-ogy,X-rayphotoelectronspectroscopy(XPS)fordetermining the physicaland chemicalsurface compositions, Auger elec-tronspectroscopywithdepthprofilingfor determinationofin depthphysicalandchemicalcompositions,nanoindentationfor evaluationofhardnessproperties,andspecificsurfacearea mea-surements.Aseriesofexperimentshavefinally beenmadeto studytribologicalpropertieslikefrictionandwearofpolished surfacesofthesesamples.

2. Experimental

Wooden samples were carved into three different shapes, cylindricaldiscs,sprocketshapes,andrectangularblocksof dif-ferentsizesroughly rangingfrom 10mmto15mm inlength, 7mmto10mminbreadthand5mmto8mminheight. Sam-plesweremadewithdifferentkindofwoodssuchasIndianpine (Pinussp.),Indianmango(Mangiferaindica),Silkcottontree (Bombaxceiba), Indian blackberry(Syzygium cumini), Cutch tree(Acaciacatechu),Danishbeech(Fagussylvatica),etc.Most oftheresultsreportedhereareobtainedwithDanishbeechwood. Samplenames showthe typeof woodused tosynthesizethe sample.Thesprocketshapedsampleswerecarefullycutexactly perpendiculartoanaxialdirectioninthewoodinordertoachieve discsofuniformthicknessafterpyrolysis.Differentgeometries havebeenusedtodemonstratethattheendproductceramiccan beobtainedindesiredshapes,thusprovingtheusefulnessofthe presentsynthesisroute.Sincelengthcontractionduring pyrol-ysis of woodisorientation dependent,the contractionsalong radial-,axial-,andtangentialdirectionsaredifferent.13,20,34–36 Pyrolysisofwoodisdoneinthreestages.Duringthefirststage, sampleswereheatedto70◦Cfor2htoremovethe moisture.

Sampleswerethenheatedto500◦Cataslowheatingrate of 1◦C/mininthesecondstage.Thisstageinvolvesthe decompo-sitionofthepolyaromatichydrocarbonpolymerslikecellulose, hemicelluloseandlignin,toformapurecarbonstructure.13,20,36 The thirdstageisacrystallizationprocesswheresamplesare heatedto1200◦Catahigherheatingrateof5◦C/minfor6h. Thisheatingleadstocrystallizationandpurificationofthe car-bonstructures.Chancesofoxidationandburningofthewood wereminimizedbycarryingoutalltheprocessstepsinanargon flow. Thepyrolysisofthe woodsresultedinthe formationof porouscarbonstructures(carbonpreforms)withanetworkof capillariesas showninSEMimages.Severalcarbonpreforms wereplacedinanaluminacrucibletogetherwithsiliconpowder andheateduptoatemperatureof1500◦C±25◦C(tomeltSi incontactwiththecarbonshapes),insideanaluminatubeina tubularfurnacefor12h.

The finished sampleswerecharacterizedbyX-ray diffrac-tion (XRD) tocheck theSi:C stoichiometryandtheir crystal structures by using aSiemensDiffractometer D5000 andthe XRDresultswereanalysedwiththeX’pertHighScorePlusTM software.Thetopographicalmicrostructurewasimagedwitha SEM(ZeissLEO435VP),andthecompositionwascheckedby theenergydispersiveX-ray(EDX)techniquewithaRÖNTEC detectorattachedtothesameSEM.TheconfirmationoftheSiC structurewasalsodonewithaRamanmicroscopefromDilor, usingthe514nmAr-ionlaserline.Foranalysesofthesurface compositions, AESwasusedwitha1␮mspotelectronbeam excitationinaPerkinElmer560system(withabout1nmdepth ofinformation)andXPSinaSPECSPHOIBOS100system® (witha2mmdiameterareaofstudy,andabout7nmdepthof information).Hardnessmeasurementsofthesampleswere car-riedoutbynanoindentationusingatriboIndentersystemfrom Hysitron.Thespecificsurfaceareaofthesampleswasobtained withasurfaceareaanalyserfromQuantachrome.Rectangular blocksareusedforalltheabove-mentionedcharacterizations. Further,dryweartestingofsynthesizedSiC/Sisamplesof cylin-dricalshapeshasbeenconductedbyusingahightemperature ball-on-disctribometer(CSM-THT).Thesamplesusedforthis werefirstgroundwithSiCpapersofdifferentgritsizeandfinally polishedusingdiamondslurryonfelt.Duringtheball-on-disc weartesting,samplesweretestedbyslidingagainstcommercial alumina balls of 6mm diameter. Thesetests were conducted foratotaltravelleddistanceof50mundernormalatmospheric conditionsatroomtemperaturewitharelativehumidityof55% usingloadsof1,2,4and5Newtons(N).

3. Resultsanddiscussion

3.1. ConﬁrmationofthenatureoftheSiC/Sicomposite 3.1.1. X-raydiffractionanalysis

The XRDpatternsconfirm theformationof cubic (␤-SiC) withsomeadditionally un-reactedsiliconasshowninFig.1. Table 1lists thepercentageof differentphasespresent inthe samples. The diffraction patterns of the samples show major peaksat2θ=28.4◦,35.5◦,41.3◦,47.2◦,59.8◦,and71.6◦,which

(3)

Table1

PercentageconcentrationofdifferentpolymorphsofSiC.Samplenameshows thetypeofwoodusedtosynthesizethesample.

Sample Cubic/% Si/% SiO2/%

Beechwood(1) 87.6 12 0.4 Beech-wood(1)-oxidized 94.7 4.5 0.8 Beechwood(2) 89.4 10.6

Beechwood(3) 96.8 3.2

Cutchtree 94.5 5.5

correspondstodiffractionfromtheSi(111),SiC(111),SiC (200),Si(220),SiC(220)andSiC(311)planes.

Theaveragegrainsizeofthesamplesisfoundtobe93nm andhasbeencalculatedbyusingaWilliamson–Hallplot.37The Rietveldrefinement38,37techniquewasusedtoobtainaccurate valuesoflatticeparametersofcrystalsandphasespresentinthe crystalsamples.The oxidationof sampleBeechwood(1) for 3hat700◦Cinairleadstodecreaseinpercentageofun-reacted siliconfrom12% to4.5%(see Fig.1andTable1).Thusthe previouslyun-reactedsiliconhasnowformedamorphousSiO2,

shownasadecreaseinintensityoftheSipeakat28.3◦,while theSiCpeaksincreaseinintensity.

Table 1 summarizes the percentage distribution of cubic SiC,un-reactedSi andSiO2insomeofthesynthesized

sam-ples.ThecomparisonofXRDresultswithexperimentalmass measurements isdoneinasubsection discussingthe ceramic formation.TheXRDpatternssuggestthatthesamplesmadeby siliconimpregnationaremorecrystallinethanforshapememory synthesis.20

3.1.2. SEMimages

TheSEMimagesoftheSiC/Sisamplesandthestarting car-bonpreformsareshowninFig.2.Fig.2(a)showsthetopsurface ofacarbonpreform,whichwascutperpendiculartoanaxial directioninthestartingDanishbeechwood.Thisimage indi-catestwodifferentkindsof poresorcapillariesinthestarting carbonstructureoriginatingfromvesselsandtracheidchannels,

Fig.1.XRDpatternsofthesamplessynthesizedfrombeechwood,andcutch tree.XRDpatternfromindividualcrystallitesisalsoshown(explainedlater). Samplenameshowsthetypeofwoodusedtosynthesizethesample.

thebiggerporeshaving asizeof20–25␮mwhilethesmaller poreshavesizesintherangeof5–8␮m.Fig.2(b)showsthe inter-nalcapillary systemofthe carbonpreformwhen itiscleaved alongtheaxialdirection.Fig.2(c)and(d)areimagesofthetop surfaceof thesamples afterconversiontotheSiC/Siceramic at different magnifications, corresponding to the preforms in Fig.2(a).SiCandsiliconregionshavealsobeenmarkedinthe images.Thetopsurfaceofthesesamplesappearsashavingthe individualSiCcrystalsembeddedinsidesilicon.Fig.2(c)and(d) indicatessignificantredistributionduringthesolidphase reac-tionofcarbonwithsilicon.Alltheporeshavevanishednearthe topsurface.Fig.2(e)and(f)showstheimagesofthesamplefrom inside.Thebiggerpores,whichappearaschannelsinthestarting carbon(Fig.2(b))arestillpreservedaftertheSiC/Siformationas shown(labelledaschannels)inFig.2(e),buttheyarenotuniform asbeforetheonsetoftheprocess,asSiCcrystalshaveexpanded duringgrowthtomakeitnonuniform.Thesmallerporesofthe startingcarbonhavetotallyvanishedafterthereaction dueto SiCcrystalliteformation(CompareFig.2(aandb)with(c–f)). The starting carbonframework has been redistributed during thesolidphasereactionandtheSiCcrystalliteshavenucleated everywhere, whichisevident fromthe Figs. 2(c–f).Fig.2(f) showstheindividualSiCcrystalsformedinsidetheSiC/Si sam-ple. Comparison of Fig. 2(a) and (b) with (c–f) proves that thecapillariespullthemoltensiliconinsideinthesolidphase contact reaction toform SiC/Si. The presence of Si andSiC crystallitesinthecompositeisexplainedfurtherinSection3.2.

3.1.3. XPSanalysis

XPSspectraoftwosamplesmadefromDanishbeechwood are showninFig.3 andcomparedwithastandardSiC refer-ence sample (6H-SiC wafer).Electrons with kineticenergies from200eVto1260eVhavebeenrecorded,coveringtherange ofbindingenergiesfrom1053eVto−7eV,astheyareexcited withanMgX-raysource.Inthefiguresthekineticenergieshave beenconvertedtobindingenergies(relevantonlyfor the pho-toelectronpeaks).The sampledareahasadiameter of 2mm. Allthespectra showthepresenceof carbon,oxygenand sili-con.Thereferencesample(a6H-SiCwafer)alsoshowsAr(2p) andN(1s)photoelectronpeaks.Thissamplehadbeensputtered withargonforashorttimetoremovethecontaminationonthe topsurface,beforetheXPS measurements.Smallamountsof calcium (≈0.5% concentration) are seenin the two samples, comingfromthestartingwood.39

Theprocessedsamplesareoxidizedduetothehandlingin airafterthesynthesis.Thesilicontocarbonratiointhesetwo samplesareequal,butdifferentfromtheratioofthereference sample,whichisduetodifferentamountsofoxideandadsorbed carbonspeciesatthesurfacesofthesamples.Thegeneralshape ofthewidescanXPSspectravarieswiththedensityandstructure of thesamples.Theinelasticlower-energybackgroundbelow the sharppeaks dependsonthesample densityand morphol-ogy.The6H-SiCsamplehastheflattestbackgroundbelowthe O(1s)(720eV)andOKLLpeaks(500eV)(i.e.atlowerkinetic

energies),whilethesilicon-impregnatedsamplesshowahigher slopeof the backgroundinthisregion.Different background slopesbelowtheoxygensignals(lowerkineticenergies)indicate

(4)

Fig.2.SEMimagesoftwocarbonpreforms(a)and(b)fromDanishbeechwood;andSiC/Sisamples(c)–(f).(a)isperpendiculartoaxialdirection;(b)isalongthe axialdirection.(c)and(d)arethetopsurfaceofsample,SiC/Sireplicaof(a)atthedifferentmagnifications.(e)and(f)areimagesfrominsidethesample,(e)isthe SiC/Sireplicaof(b).(f)showsSiCcrystalsimbeddedinsilicon.

differences in the in-depth distributions of oxygen in the samples40 _and _effects _of _roughness. _Higher _slopes _indicate deeperpenetrationofoxygen.

The reported concentrations of the impregnated samples seemtoindicate,whencomparedtothe(mildlysputtered) ref-erencesample,thattheamountofcarboninthetoplayersofthe impregnated samplesisslightlyhigherthanforthisreference sample.Webelievethatmostofthisexcesscarbon(compared toSi)comesfromtheambientduringhandlingofthesamples, andavariabledegreeofoxidation,whichtendstolowertheSi concentrationatthesurface.Wethusdonotdetectanyexcess ofSioverSiCwiththistechnique,evenaveragingoveranarea of2mm.Itisthereforeinterestingtocheck,withabetter reso-lution,anddeeperinthesamples,howtheirlocalcompositions change.ThisisdonewithAugerelectronspectroscopy(AES).

3.1.4. Augeranalysis

Inordertobetteraccessthelocalconcentrationofthe sam-plesbeneaththeoxidizedsurface,andcontaminationlayers,we

used electronbeamexcitedAESincombinationwithAr+_-ion

sputtering.

Fig. 4 shows the result of such asputterprofile measure-ment.Thisanalysisis donewithAES excitedwithafocused electron beam (1␮m) of 5keV energy, focused on a SiC crystal-likehumponthesurface,andthesurfaceisrepeatedly bombardedforperiodsof10minuteswithAr-ionstoremovethe upperlayers,andthenanalysed.Each10minsputtering inter-valwouldremove10 ˚Aofauniform,flatsampleascalculated bytheSRIMcode41,42forthepresentexperimentalconditions. Fig.4(a)showspeaksat92eV(SiL23VV),272eV(CKLL),290eV

(CaLMM),and505eV(OKLL).From thesputterprofile itcan

clearlybeseenthattheintensityoftheoxygenpeakisreduced withthesputteringtimewhiletheintensityofthecarbonand sil-iconpeaksareincreasing.Fromthespectrawecanalsoseethe presenceofcalcium,whichisanessentialnutritionalcomponent ofbeechwood.39Fig.4(b)showsthecorrespondingatomic per-centageofSi,CandOwithdepth.Fromthemeasuredvalues,as shownsystematicallyinFig.4(b),wejudgethedeviationofthe concentrationsfromsmoothcurvestobelessthan2.5%,which

(5)

Fig.3.XPSspectraoftwosamplesquotedas1&2(bothsynthesizedfrom Danishbeechwood)inthefigure,andthe6H-SiCreferencesamplequotedas 3.Theinsettableinthefigureshowsthepercentageconcentrationsofthese samplesasderivedbytheCasaXPSTM_software.

westateastheexperimentalaccuracyofthemethodasappliedto thissample.Wecanclearlyseethegradualapproachtoanearly 1:1stoichiometryofsiliconandcarbonindepthswellbelowthe oxidized surface,butwithsomelocal fluctuations, whichare probablyduetothenon-uniformcharacterofthesample.

3.1.5. Ramananalysis

Ramanspectroscopycanbeusedtodistinguishtheindividual polymorphictypesofSiC43,44anditalsoconfirmsthe forma-tionof␤-SiCinthesamples.The␤-SiChasazincblendelike structurewiththesmallestunitcellofalltheSiCpolymorphs. Bulk␤-SiChastwoopticalmodesattheΓ pointofthe Bril-louinzone, atransverse optic (TO)mode at796cm−1 and a longitudinaloptic(LO)modeat972cm−1.44

TheXRDresultsshowedaround90%␤-SiCasamajorphase andtherestasun-reactedsilicon.TheRamanspectraarealso

Fig.5.Ramanspectraofthesamples(allsynthesizedfromDanishbeechwood).

Table2

Peak positions in theRaman spectra of samplesand suggestions for the polymorphs.

Sample Peakpositions/cm−1 (±4cm−1)

Suggestionsfor polymorph/cm−143,44

Beechwood(1)-oxidized 795.6 3C(796and972) Beechwood(2) 799.5 3C(796and972) Beechwood(3)-oxidized 798.3,971.5 3C(796and972) Beechwood(4)-polished 791.3,965.4 3C(796and972) Beechwood(5) 796.1,967.3 3C(796and972)

showing peaksof␤-SiCandsilicon,whichthusconfirms the presenceofSiCandsiliconinthesamples.SincetheRamanlines ofSiCsamplesareratherbroad,theresolutionofRaman micro-scopewas setto±4cm−1.Fromthe positionsofthe phonon peaksshowninFig.5andsummarizedinTable2,itcanclearly beconcludedthatallpeaksarefallinginthevicinityof796cm−1 and972cm−1,whichareknownas TOandLOmodesfor ␤-SiC,respectively.Thefullwidthathalfmaximum(FWHM)of thesepeaksisintherangeof22–26cm−1,whichsuggeststhat thesamplesarehavingstackingfaults.

Fig.4.(a)SputterprofileofaSiCcrystallitehumponthesurfaceofasamplemadefromDanishbeechwood.(b)RelativepercentageconcentrationofSi,CandO withthedepth,fromtheAugerspectra.

(6)

Fig.6.SEMimagesofindividualcrystallites(a)and(b)atdifferentmagnifications(insidethesamplewhichwasbrokenonpurposeandtakenoutafter6hprocessing). (c)SEMand(d)opticalmicroscopeimageofapolishedsamplesurface.SamplemadefromDanishBeechwood.

3.2. Ceramicformation

Inthepresentmethod,carbonpreformsareimpregnatedwith moltensiliconandcapillaryactionpullstheliquidsiliconinto thepores.Alldetailsofthepyrolysisprocess(seealsoabove)to obtainthecarbonpreformsthemselvesandcharacterizationby XRD,SEMandRaman,includingTGAofwoodhavealready beengiven.20

Afterbeingdrawnintothecarbonstructures,Sireactswith carbonatatemperatureof1500±25◦C,formingacompact,low surfaceareaSiC/Sicompositematerial.Siliconmeltsat1410◦C andtheliquidentersintothe poresundergoingasolid–liquid reactiontoformSiCasfollows:

C(solid)+Si(liquid)1500−→ SiC◦C (solid)

3.2.1. Roleofsilicon

The synthesis of SiC/Si ceramics from carbon preforms involvesatwo-stepprocess.Inthefirststepmoltensilicon infil-trates the capillaries of the carbon element to react withthe carbontoformcrystallineSiCand,inourcasethisprocessruns for morethan 5h toprocess allthe carbon. The second step involvesfurtherinfiltrationwithmoltenSi untilsomekindof saturation.Moltensiliconisreadilyavailable13 inthe capillar-iesinthebeginning,butwehaveobservedthatitspenetration becomesveryslowattheend.Theconversionofcarbontosilicon wasverywellexplainedbyZollfranketal.,21andsomefurther detailsaregiveninthesupportmaterial.Thesecond–or last – stepofceramicformation,whichinvolvesfurtherpenetration intovoidsbythesiliconmelt,isverycrucialandisresponsible forkeeping thewholeceramic togetherasasinglesolidunit.

Almostalloftheun-reactedsiliconisamorphousinnatureas observedfromtheXRDpatternofthepolishedsample,shown inFig.1.

Wehavecarriedoutthesiliconimpregnationat1500◦Cunder theargonflowandhavecheckedthesampleafterthefirststage. Fig.6(a)and(b)showstheSEMimagesoftheindividual crys-tallites insidethesample havingdimensions rangingfrom15 to40␮m.Theseimageshavebeentakenfrominsidethe sam-pleandthissamplewastakenout“prematurely”after6h.Thus onlythefirststageofreactionhasbeencompleted(incaseof Fig.6(aandb)),whichhasonlyresultedintheformationofSiC micro-crystals.Thus,Figs.6(a)and(b)aresnapshotsofa sam-ple afterthefirststageofthereaction.TheXRD(Fig.1)and Raman (Fig.7c)of crystallitesconfirmthat thesamples after completionofthisstepareintheformofpurecrystallineSiC. Theseindividualcrystallitescanberemovedfromtheexposed surfacebygentleabrasion.ThusthepureSiCcrystalsformed afterthefirststagearenotheldstronglytogether,butthefurther additionofsilicon(duringthesecondstage)holdsthemstrongly together,formingadenseunit.

IfinfiltrationofSiintothevoids(Fig.6(a)and(b))between theindividualcrystallitescontinuestofillthemupcompletely, it would result in SiC/Si ceramics like the ones shown in Fig.2(c)–(e),wheretheyappearasverydenseunits.Fig.6(c)and (d)showstherespectiveSEMandopticalmicroscopeimageson thesurfaceofapolishedsamplewiththismaximumdensity.The crystallineSiCinclusionsappearasislandsinthematrixof sil-icon.TheSiCappearsdarkgreywhilesiliconappearsaslight greyintheSEMimage(Fig.6(c)).Similarly,SiandSiCappear withdifferentyellowbrightnessinopticalmicroscopeimages (Fig.6(d)).WehavealsoproducedapowderconsistingofSiC micro-crystalsfromsimilarsamplessimplybyetchingawaythe

(7)

Fig.7.OpticalmicroscopeimagesandtheircorrespondingRamanspectrum(fromthemarkedplaces)ofapolishedsamplein(a)and(b),andSiCcrystallitein(c). (a)aSiCregion,(b)aSiregion,and(c)aSiCcrystallite.SamplemadefromDanishbeechwood.

amorphousSiwithamixture ofHFandHNO3.Itprovesthat

thismethodcanalsobeusedtoproduceSiCmicro-crystals,and itindirectlyprovestheroleofamorphoussiliconasthebinder inthesesamples.

Fig.7(a)and(b)showstheRamanspectra takenona pol-ishedsample at the places markedin the opticalmicroscope images onthe leftside inthe figure.The opticalmicroscope imageinFig.7(a)showsadarkgreyregionandthe correspond-ingRamanpeakindicates␤-SiC.Similarly,Fig.7(b)showsthe siliconregionandthesiliconpeakinRaman.Fig.7(c)shows theRamanspectrumoveroneoftheSiCcrystallitesshownin theopticalmicroscopeimageandasdiscussedpreviouslyin ref-erencetoFig.6(a)and(b).TheresultsofRamanstudiesofthe polishedsampleshavealsobeenconfirmedbytheEDX mea-surementsasshowninFig.8.TheEDXhasbeendoneattwo differentareas,atSiandSiCrichregions.Theatomic percent-agesofrespectiveareasare tabulated asthe insertsinFig.8, whichsuggestthepresenceofSiandSiCphasesincomposite.

From the all the points in this discussion, it can now be concludedthatsiliconfillsthevoidsbetweentheindividual crys-tallitesandtherebyhelpsinholdingtheSiCcrystallitestogether

toproduceadensepieceofasolidmaterial.Thusmolten sili-conplaystheroleofbinderofindividualcrystallites.Itprevents themfromfallingapartandcreatesadensestructure.

3.2.2. Determinationofmassanddensity

Inthisanalysisthecarbonatomsaresupposedtoreactwith anequal numberof Siatomstoform SiC.Thedensityof the resultantproductcanthereforebecalculatedbyusingasimple formula: Dcomposite= Dcarbon+Dcarbon×28.0855 12 × 1+ MSi MSiC , (1) whereDcompositeisthecalculatedvalueofdensityofthe

sam-ple,Dcarbonisthedensityofthecarbonpreformofaparticular

sample,andMSiandMSiCarethemassesofSiandSiCinthe

composite.Thederivationofthisformulaandotherequivalent formulaearegiveninsupportingmaterial.

The term (Dcarbon+Dcarbon×(28.0855/12)) is the

theo-retical value of the density of SiC when only oneadditional atom of silicon is added to the carbon atoms of the

(8)

Fig.8.SEMimageandEDXatthemarkedplacesshowingSiandSiCregioninapolishedsample.SamplemadefromDanishbeechwood.

earlier framework. The additional term (Dcarbon+Dcarbon×

(28.0855/12))(MSi/MSiC)isduetothepresenceofun-reacted

silicon inside the sample. The mass of the un-reacted sili-con has been estimated by taking the difference of mass of final composite and the theoretical mass, which is given as (Mcarbon+Mcarbon×(28.0855/12)).

Thedifferencebetweenexperimentalvaluesandcalculated valuesisbetween0and8.1%.Thesedeviations,althoughmostly small,maycomefromerrorsinthemeasurementofthevolume of the sample andfrom the presence of otherelements such ascalciumintheoriginalwoodasseenfromEDX,XPS,and Augerresults.Thereisalwaysapresenceofoxygen,mostlikely adsorbedonthe surfaceof thesamples,butnotaccountedfor inthemodel.Inthismodelwehavenotconsideredanyexcess carbonbecausewehavenotfoundanysignsofsuchunreacted carbon(lessthan0.8%)intheTGAofSiC/Si,asshowninthe supportingmaterial.

Thecomparisonofthepercentageofsilicondeducedfrom thechange inmassofthe samples(showninTable3),which isaround35–40%,andfromXRDanalysis(showninTable1), whichisshowing5–12%,suggeststhat mostofthe siliconis presentinitsamorphousform.

FromTable3itisconcludedthatallofthecarbonpreform materialhasbeen convertedtocrystalline SiC.Un-reacted Si fillsandtakesuptheemptyspaceinthesamplesandaccounts fortheadditional40%ofthematerial(bymass).Ifandwhen thereisgoodagreementbetweenexperimentalandcalculated valuesin the table,the useof Eq. (1) for the analysis seems justified,andthemodelpresentedhereisthenvalid.

Afterthecompletionofthereactionprocess,mostofthe big-ger pores in the starting wooden framework must havebeen

filledwithsiliconorhavebeencappedfromthetop,resultingin ≈0m2g−1specificsurfaceareas(discussedlater).Therearestill somevoidsoropenchannelsinsidethestructuralunitasshown intheSEMimageinFig.2(d),whichleadstoasmallervaluefor thedensityoftheSiC/Sicomposite(giveninTable3)thanthe densityofbulksilicon(2.32gcm−3)orbulkSiC(3.2gcm−3). Thepresentvaluesareapproximatelysimilartothoseobtained by Greilet al.14 The startingcarbon structure iswitha hon-eycomb like structure of the bigger size pores, togetherwith a distributionof small size pores as shownin Fig.2(a). The smallerporesvanishquicklyduringthereactionofsiliconwith carbontoformSiCcrystalliteswhilethebiggerpores (capillar-ies)maystill remaininexistenceinthe formofnon uniform voids(seeFig.2(d)),nothavingbeencompletelyfilledbythe moltensilicon.Thus,overallitcanbesaidthatSiimpregnation retainsthestartingbulkstructureofthecarbonpreformwhile itsmicrostructure(smallerpores)islostduringtheformationof SiCcrystallites.Fig.9shows,asanillustrationoftheprocedures usedhere,howacompact,hardandlightweightsprocketshaped structure of SiC/Sican be formed froma preshapedwooden structure,whenimpregnatedwithmoltensilicon.

3.3. Hardnessandspeciﬁcsurfaceareaanalyses

The hardness at various positions of one of the samples (synthesizedfromDanishbeechwood)wasmeasuredbynano indentationwithaBerkovichdiamondindenter.TheXRD,SEM, EDXandRamanmeasurementssuggesteddifferentregionsof SiC andSionthesamples,solocalvariationsinhardnessare expected,whichwasourreasonforusingnanoindentationthat also provides informationabout elasticmodulus. The sample

(9)

Table3

Calculatedandexperimentalvaluesofmassanddensityofthesamplessynthesizedbysiliconimpregnation. Samplename Massof

carbon/g Densityof carbon/gcm−3 Theoretical massof SiC/g Experimental massof composite/g Un-reacted silicon/% Experimental valueofdensityof composite/gcm−3 Calculatedvalue ofdensityof composite/gcm−3 Deviationfrom calculatedvalue ofdensity/% Beechwood(1) 0.56 0.4 1.87 2.82 33.7 2 1.999 0.2 Beechwood(2) 0.25 0.37 0.83 1.37 39.4 2.08 2.079 ≈0 Beechwood(3) 0.26 0.32 0.86 1.48 41.9 1.84 1.835 ≈0 Beechwood(4) 0.25 0.37 0.83 1.14 27.5 1.78 1.696 4.8 Beechwood(5) 0.2 0.37 0.67 1.09 38.5 2.12 1.986 6.5 Beechwood(6) 0.29 0.34 0.96 1.59 39.8 2.06 1.904 8.1 Beechwood(7) 0.10 0.38 0.32 0.58 44.6 2.26 2.261 0.1 Beechwood(8) 0.07 0.36 0.24 0.46 47.3 2.31 2.299 0.5

usedwasflat,denseandpolishedwithafinishdownto500␮m. TheOliver andPhaarmethod45 isused toevaluate the hard-nessvaluesusingtheloadingandunloadingcurves,whenload displacementcurvesarerecordedunderload-controlledmode. Theloaddisplacementcurvesshowahysteresis-likebehaviour withelastic–plasticloadingfollowedbyanelastic unloading. Theknowledgeofthecontactareabetweentheindenterandthe sampleisveryimportant.Itisknownthattheplasticproperties of the sample underobservation can change the contactarea andhenceresultinwrongvalues.46Thishappensmainlywhen a“pileup”or “sinkingin” of thedisplaced materialhappens aroundtheindentationspot.Theratioofhc/hmax45indicatesthe

tendencyof the material topile up,wherehmax isthe

maxi-mumdepthofpenetrationduringtheindentationtestandhcis

thefinaldisplacementafterthecompleteunloadingofthe con-tact.Iftheratiohc/hmax<0.7thenthecontactareagivenbythe

Oliver–Phaarmethodisknowntomatchverywellwiththetrue contactarea,andifhc/hmax>0.7,thenitmayleadtolargeerrors

incontactarea.47Inthismethodthecontactdepthisestimated fromtheloaddisplacementdatausingthefollowingequation:

hc=hmax−εPmax

S , (2)

wherePmax is the maximum indentation load, S is the

stiff-nessobtainedfromtheunloadingcurveasS=(dP/dh)_P=Pmax

andεisaconstant,whichdependsupontheindentergeometry; empiricalstudieshaveshownthatε≈0.75.

Thehardness,H,andelasticmodulus,E,arecalculatedusing thefollowingformulae:

H= Pmax S , (3) and E= 1 β √ π 2 S √ A, (4)

whereβisaconstant,dependingonthegeometryoftheindenter. Itsvaluefor the Berkovichindenteris1.034.The areaof the impression,A,isdeterminedfromtheindentershapefunction atthecontactdepthhcandisproperlycalibrated.

Thehardnessvaluesobtainedforthesamplesvarybetween 10.7 and31.4GPa. From Table 4, givingthe results, we can group our values in three sets. The first five values in this tablehavehardnessvaluesfrom26.5to31.4GPa(witha stan-darddeviation(SD)of2.30GPa),andelasticityvaluesranging from236.9to332.1GPa(SDof39.8GPa),whichmatchesthe respectivevaluesforcompactbulkSiC.48Thissetofvalues con-stitutesafirstcategory.Thefollowingcategoryhastwovalues 10.7and12GPaforhardnessandelasticityvaluesof115.8and 121.9GPa,which,inturn,clearlymatchvaluesforsilicon.49

Thethirdcategoryhasvaluesvaryingfrom17.1to22.3GPa (SD of 2.32GPa) for hardness and 169.9–261.9GPa (SD of 35.5GPa) for the elasticity values.Thiscategory most likely includes erroneous values due to an overestimation of area becauseofpileupduringtheindentation,sincethehc/hmaxratio

(10)

Table4

HardnessandYoung’smodulusvaluesmeasuredindifferentpointsononesample,madefromDanishbeechwood.

S.no. hc/nm Area/nm2 Load/microN F=hc/hmax Hardness/GPa E/GPa

1 73.9 232736.3 7302.54 0.66 31.4 253.2 2 51.9 131368 3504.8 0.66 26.7 236.9 3 75.7 241581.1 6979.9 0.68 28.9 259.5 4 90 317848.4 8433.58 0.69 26.5 235.2 5 67.5 201649 6251.45 0.71 31 332.1 6 75.2 238709 2858.9 0.71 12 121.9 7 65.8 207725 2229.4 0.71 10.7 115.8 8 98.9 371146.5 6333.86 0.72 17.1 169.9 9 93.5 338434.7 6009.6 0.73 17.8 194.7 10 79.3 259817 4797.1 0.74 18.5 213.6 11 121.6 528854.5 8596.06 0.75 16.3 185 12 94.1 342117 7625.35 0.75 22.3 261.9

exceeds0.7,asdiscussedabove,oritmaybeduetothe averag-ingofthecontributionsfrombothSiandSiC.Otherwise,these valuesmayhavebeen influenced bythe non-uniformsurface ofthepolishedsample(seeFig.2(f))usedfortheindentation measurements.

TheBETadsorptionisothermmethodisusedtomeasurethe specificsurfaceareasofthesamples.Initially,thesamplesare evacuatedanddegassedat350◦C.Aftercoolingthemdownto liquidnitrogentemperature,nitrogengasatvariablepressures is usedto obtain six pointsonthe BET adsorption isotherm, fromwhichthespecificsurfaceareaiscalculated.Thespecific surfaceareasofthethreesampleshavebeenmeasuredandtheir respectivevaluesare foundtobe 0.33,0.00and0.06m2g−1. With thesevalues, we cansay that these samples are almost totally non-porous. This is understandable assuming that the moltensilicongoesintotheporouscarbonframeworkandfills theholesuptoalargeextent,andcapstherest,andthusmay notleaveanypassagefornitrogentogoinsideandadsorb.The correspondingSEMpicturesintheFigs.2(c)–(f)showthe pres-enceofuniformlyimbeddedcrystallites,indicatingtheabsence ofirregularitiesanddefects,whichcouldotherwisehaveacted asadsorptionsitesforthenitrogengas.

3.4. Wearandfrictionproperties

Inthisstudy,preformedSiC/Sistructures(synthesizedfrom Danishbeechwood)withflatfaces(asshowninFig.9)were usedforexperimentstotestthewearandfrictionproperties.The wearrate(Wr)wascalculatedbyusingthefollowingequation:

Wr =Wν

Fl, (5)

whereWνisthewearvolume,Fisthenormalload(inN)applied onflatsurfaceandlistheslidingdistance.

Figs. 10(a)and(b) shows the SEMmicrographs fromthe weartrackofatestedSiC/Sicompositeunderdrysliding con-ditions.Theweartrackshowsasmooth,near-polishedsurface withsomeweardebrisonthetrack.Experimentaldataof fric-tionmeasurements(asshowninthesupportmaterial)indicates an increase inthe coefficientof friction(COF) in the begin-ning.Itbecomesconstantafterwards,whichmaybeattributed toremovaloftheinitialsurfaceroughnessofthematerial.From

theSEMimagesinFig.10(a)and(b),themainwearmechanism couldbeattributedtomicro-crackingandpullingofthegrains inthecompositematerial,whichresultsincavityformationand, later,fillingofthesecavitiesbytheweardebris.25Fig.10(c)also showstheeffectofnormalloadonCOF(μ)atroomtemperature (25◦C).Thisfigureindicatesthatthereisnosignificant varia-tion intheCOFwithnormal loadsupto2N.However,COF increases toanaveragevalue of 0.75for loadsof 4 and5N. TheseresultsarecomparablewiththoseofDongetal.,26where ␣-SiCballswereusedonSiCceramicsandμwasfoundtobe 0.67 onaverage.Sangetal.,29 havealsoreportedsimilar val-uesofCOFforSi/SiCcomposite.Ourvaluesareslightlyhigher anditmightbeduetotheroughsurfaceandthepresenceoffree SiC-crystallites,whichgetontopofthesurfaceduetoploughing whenthesampleispolished.ItissuggestedthattheSiC compo-nentintheSiC/Sicompositesleadstohigherμ,accompaniedby higherwearduetoitsabrasivenature,whilethepresenceofSi inthecompositematrixhasalargeimpactonthemorphology andadhesivefriction,whichactuallyalsoincreasesμ.28 Sili-concanthusreactwithoxygeninthealuminaballsandcause formationofhardSiO2particles,whichareexcellentabrasives.

Furthermore,Sibyitselfcouldsticktothealuminaball,causing anincreaseinthefrictionforces.

Fig.10(d)showsthe variationof thecoefficientoffriction withtemperatureataconstantnormalloadof4N.Thefigure showsaverysmall increaseinthecoefficientof frictionwith temperature.Thereisastepof0.073inμfromroomtemperature to100◦Candafterthatthereisslightincreaseofapproximately 0.014inμwithevery200◦Criseintemperature.Thissetof observationsliespartlyinregionIIandpartlyinregionIIIof thetransitiondiagramsuggestedbyDongetal.26TheregionII startsfromroomtemperatureto250◦Capproximatelyat nor-mal loadslessthan10NandregionIIIisconstitutedbyCOF valuesmeasuredfor thetemperaturerangeof 250–1000◦Cat normalloadslessthan10N.Thisregionisgovernedby tribo-oxidation reactions.Takadoumet al.,27 _have_also _studied _the frictioncoefficientandpropertiesofSiC/Sisystemsandfound

μvaluesvaryingintherangeof0.4–0.85withincreasingsliding distanceatarelativehumidityof50%.Theμvaluesfoundhere arethusslightlyhigherthanreportedbyothers.29,30

ThewearratecalculatedbyusingEq.(5)atnormalloadof 4NinasampleoftheSiC/Sicompositeis6.6×10−4mm3/Nm

(11)

Fig.10.(a)and(b)SEMimagesofweartrackstakenfromdifferentregionsandatdifferentmagnificationsofsametrack.(c)Effectofnormalloadoncoefficientof frictionatroomtemperature(25◦C).(d)Effectoftemperatureoncoefficientoffrictionataconstantnormalloadof4N.SamplesmadefromDanishbeechwood.

atroomtemperature.Thisvalueishigherthanthereported val-uesforsimilarmaterials(inrangeof10−6)30andiscomparable toresultsreportedbySangetal.29In theliterature,thelower wearrateshavebeenexplainedbytheformationofcarbondebris duringslidingandpresenceofcarbonintheEDXspectrum col-lectedfromtheweartrack.However,wehavefoundanegligible amountof carboninRaman spectra collectedfrom the wear track.

4. Conclusions

All the samples made in the present work with silicon impregnationof aporouscharcoalskeletoncomingfrom nat-ural wooden samples havea composite SiC/Si character and verylowspecificsurfaceareas.Thisisaresultofthecharacter of thereaction process.Thusthe formationof SiC/Si ceram-icsfromporouscharcoalstructuresobtainedfromnaturalwood takesplaceintwostages.Thefirststageinvolvesthereaction ofsiliconwithcarbontoproducecrystallineSiCandthesecond stagecontinueswithinfiltrationofthestructuresbymoltenSi untilsaturation.Thesecondstageisresponsibleforbindingthe individualSiCcrystallitesinamatrixofsilicontoconstitutea singleanddenseunit.

TheSiCintheceramicisinthe␤-SiCformasconfirmedby XRDandRamanstudies.Thedensityoftheresultingmaterial variesintherangeof2.0–2.3gm/cm−3.Thisisroughlyaround 3.5timesdenserthanthesamplesmadewiththeshapememory synthesismethod,20fromsimilarwoodenpreforms,which pre-servestheopenskeletonswithoutfillingthepores.Thepresent

samplesarefoundtobeverydenseandhardwithhardnessvalues intherange26.5–31.4GPa.

The wear results indicateda wear rate higher than earlier reportedforSiCelements,whichisbelievedtobemostlydue totheadhesivefrictionofamorphousSi.Coefficientoffriction measurementsonthesecompositesindicatedarelativelystable valueovertimewithrespecttochangingtemperaturesandloads. Overall,itcanbesaidthatthesesamples,whichcanbemade indesiredshapes,arequitestablewithrespecttoloadand tem-peraturevariations,thusqualifyingthemforwearandabrasive applications,forcuttingtoolapplications,andfornew genera-tionbrakematerialsforautomobilesandaircrafts.Theycanalso beusedasrefractorymaterials.Artificiallyshapedelementsof thehard,compact,andlightweight␤-SiC/Sicomposite mate-rialhavebeenfabricatedasshowninFig.9,whichmightbeof interestfortheproductionofmorecomplexshapesofSiCthan possiblethroughsinteringandinjectionmolding.

Acknowledgements

TheDanishMinistry forResearchandInnovation,through itsprogramforSustainableEnergyandTheEnvironment,has fundedthisworkwithagranttoR.Dhiman,2104-05-0073.The authors aregrateful for advice andtechnicalsupportfrom E. Skou, S.Tougaard,T.Warner,P.B.Hansen,D.Kyrping,and T. Sørensen atSDU; to N.Dam Madsen, and P. Hald, from The University of Aarhus, S. Vikram Singh from The RISØ National Laboratory,Roskilde,DenmarkandtoR.Bala from HPUShimla,India.

(12)

AppendixA. Supplementarydata

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/ j.jeurceramsoc.2011.11.029.

References

1. LiZ,ZhangZ,MengA,GuoJ.Large-areahighly-orientedSiCnanowire arrays:synthesis,Raman,andphotoluminescenceproperties.JPhysChem B2006;110:22382–6.

2.JuZ,MaX,FanN,LiP,XuL,QianY.High-yieldsynthesisof single-crystalline 3C–SiC nanowires by a facile autoclave route.Mater Lett

2007;61:3913–5.

3.FisselA,SchroterB,RichterW.Low-temperaturegrowthofSiCthinfilms onSiand6H–SiCbysolid-sourcemolecularbeamepitaxy.ApplPhysLett

1995;66:3182–4.

4.Choyke WJ,PenskG.Physicalproperties ofSiC.Mater ResSoc Bull

1997;22:25–9.

5.Srinivasan AV, Haritos GK, Biomimetics Hedberg FL. Advancing man-made materials through guidance from nature. Appl Mech Rev

1991;44:463–81.

6.ShinY,WangC,ExarhosGJ.SynthesisofSiCceramicsbythecarbothermal reductionofmineralizedwoodwithsilica.AdvMater2005;17:73–7. 7.Calvert P. Biomimetic ceramics and composites. Mater Res Soc Bull

1992;17:37–40.

8.Dhiman R, Johnson E, Morgen P. Growth of SiC nanowhiskers from wooden precursors, separation, and characterization. Ceram Int 2010, doi:10.1016/j.ceramint.2011.06.001.

9.NayakBB,MohantyBC,SinghSK.Synthesisofsiliconcarbidefromrice huskinadcarcplasmareactor.JAmCeramSoc1996;79:1197–200. 10.AmirthanG,UdayakumarA,PrasadVVB,BalasubramanianM.Synthesis

andcharacterizationofSi/SiCceramicspreparedusingcottonfabric.Ceram Int2009;35:967–73.

11.ShinDW,ParkSS,ChoaYH,NiiharaK.Si/SiCcompositesfabricatedby infiltrationofaSimeltintocharcoal.JAmCeramSoc1999;82:3251–3. 12.SinghM.Environmentalconsciousceramics(Ecoceramics).CeramEngSci

Proc2000;21:39–44.

13.GreilP,ThomasL,KaindlA.Biomorphiccellularsiliconcarbide ceram-ics from wood. I. Processing and microstructure. J Eur Ceram Soc

1998;18:1961–73.

14.GreilP,ThomasL,KaindlA.Biomorphiccellularsiliconcarbideceramics fromwood.II.Mechanicalproperties.JEurCeramSoc1998;18:1975–83. 15.OtaT,TakahashiM,HibiT,OzawaM,SuzukiS,HikichiY.Biomimetic processforproducingSiCfromwood.JAmCeramSoc1995;78:3409–11. 16.ChakrabartiOP,MaitiHS,MajumdarR.Si–SiCceramicsfromplant

pre-cursor.JMaterSci2004;39:4715–7.

17.VogliE,SieberH,GreilP.BiomorphicSiC-ceramicpreparedbySi-vapor phaseinfiltrationofwood.JEurCeramSoc2002;22:2663–8.

18.LocsJ,Berzina-CimdinaL,ZhurishA,LocaD.Optimizedvacuum/pressure solimpregnationprocessingofwoodforthesynthesisofporous,biomorphic SiCceramics.JEurCeramSoc2009;29:1513–9.

19.VogliE,MukerjiJ,HoffmanC,KladnyR,SieberH,GreilP.Conversion ofoaktocellularsiliconcarbideceramicbygas-phasereactionwithsilicon monoxide.JAmCeramSoc2001;84:1236–40.

20.DhimanR,PetruninV,RanaK,MorgenP.Conversionofwooden struc-tures into porousSiC with shapememory synthesis. Ceram Int 2010, doi:10.1016/j.ceramint.2011.05.124.

21.ZollfrankC,SieberH.Microstructureevolutionandreactionmechanismof biomorphousSiSiCceramics.JAmCeramSoc2005;88:51–8.

22.StadlerZ,KrnelK,KosmacT.Frictionbehaviorofsinteredmetallicbrake padsonaC/C–SiCcompositebrakedisc.JEurCeramSoc2007;27:1411–7.

23.LópezOB,OrtizAL,GuiberteauF,PadtureNP.Microstructuraldesignof sliding-wear-resistantliquid-phase-sinteredSiC:anoverview.JEurCeram Soc2007;27:3351–7.

24.Zum Gahr KH, Blattner R, Hwang DH, Pöhlmann K. Micro- and macro-tribological propertiesof SiCceramics in slidingcontact.Wear

2001;250:299–310.

25.WascheR,KlaffkeD,TroczynskiT.TribologicalperformanceofSiCand TiB2againstSiCandAl2O3atlowslidingspeeds.Wear2004;256:695–704.

26. DongX,JahanmirS,IvesLK.Weartransitiondiagramforsiliconcarbide.

TribolInt1995;28:559–72.

27.TakadoumJ,ZsigaZ,RhoumaMB,CarmesCR.Correlationbetween fric-tioncoefficientandwearmechanismofSiC/SiCsystem.JMaterSciLett

1994;13:474–6.

28.FouquetS,RollinM,PaillerR,BourratX.Tribologicalbehaviorof com-positesmadeofcarbonfibresandceramicmatrixinSi–Csystem.Wear

2008;264:850–6.

29.SangK,JinZ.UnlubricatedwearofSi/SiCanditscompositewithnickel Si/SiC–Ni.TribolInt2001;34:315–9.

30.AmirthanG,BalasubramanianM.Reciprocatingslidingwearstudieson Si/SiCceramiccomposites.Wear2011,doi:10.1016/j.wear.2011.04.006. 31.NovakS,RadeK,KonigK,BoccacciniAR.Electrophoreticdepositionin

theproductionofSiC/SiCcompositesforfusionreactorapplications.JEur CeramSoc2008;28:2801–7.

32. MurogaT,GasparottoM,ZinkleSJ.Overviewofmaterialsresearchfor fusionreactors.FusionEngDes2002;61–62:13–25.

33. TavassoliAAF.Presentlimitsandimprovementsofstructuralmaterialsfor fusionreactors—areview.JNuclMater2002;302:73–8.

34.ByrneCE,NagleDC.Carbonizationofwoodforadvancedmaterials appli-cations.Carbon1997;35:259–66.

35.ParisO,ZollfrankC,ZicklerGA.Decompositionand carbonisationof woodbiopolymers—amicrostructuralstudyofsoftwoodpyrolysis.Carbon

2005;43:53–66.

36.ShafizadehF,SekiguchiY.Developmentofaromaticityincellulosicchars.

Carbon1983;21:511–6.

37.WilliamsonGK,HallWH.X-raylinebroadeningfromfiledaluminiumand wolfram.ActaMetall1953;1:22–31.

38.ShafferPTB.Areviewofthestructureofsiliconcarbide.ActaCrystallogr B1969;25:477–88.

39.FrommJ.Woodformationoftreesinrelationtopotassiumandcalcium nutrition.TreePhysiol2010;30:1140–7.

40.TougaardS.EnergylossinXPS:fundamentalprocessesandapplications forquantification,non-destructivedepthprofilingand3Dimaging.JElectr SpectroscRelatPhenom178–179:128–53.

41.ZieglerJF.SRIM-2003.NuclInstrMethB2004;219–220:1027–36. 42. Ziegler JF. SRIM—the stopping and range of ions in matter.

<http://www.srim.org>;2009.

43.NakashimaS,HigashihiraM,MaedaK.Ramanscatteringcharacterization ofpolytypeinsiliconcarbideceramics:comparisonwithX-raydiffraction.

JAmCeramSoc2003;86:823–9.

44.BechelanyM,BrioudeA,CornuD,FerroG,MieleP.ARamanspectroscopy studyofindividualSiCnanowires.AdvFunctMater2007;17:939–43. 45.OliverWC,PharrGM.Animprovedtechniquefordetermininghardness

andelastic-modulususingloadanddisplacementsensingindentation exper-iments.JMaterRes1992;7:1564–83.

46.BolshakovA,PharrGM. Influencesof pile-uponthemeasurementof mechanicalpropertiesbyloadanddepthsensingindentationtechniques.

JMaterRes1998;13:1049–58.

47.PharrGM.Measurementofmechanicalpropertiesbyultra-lowload inden-tation.MaterSciEngA1998;253:15–9.

48.SanghoonS,JangJ,PharrGM.Extractionofflowpropertiesof single-crystalsiliconcarbidebynanoindentationandfinite-elementsimulation.

ActaMater2008;56:3824–32.

49.VandeperreLJ,GiulianiF,LloydSJ,CleggWJ.Thehardnessofsiliconand germanium.ActaMater2007;55:6307–15.