Evaluation of grinding methods for pellets preparation aiming at the analysis of plant materials by laser induced breakdown spectrometry

ContentslistsavailableatScienceDirect

Talanta

j ourna l h o me p a g e : w w w . e l s e v i e r . c o m / l o c a t e / t a l a n t a

Evaluation

of

grinding

methods

for

pellets

preparation

aiming

at

the

analysis

of

plant

materials

by

laser

induced

breakdown

spectrometry

Marcos

da

Silva

Gomes

,

Dário

Santos

Junior

,

Lidiane

Cristina

Nunes

,

Gabriel

Gustinelli

Arantes

de

Carvalho

_,

_Flávio

_de

_Oliveira

_Leme

_,

_Francisco

_José

_Krug

a_{UniversidadeFederaldeSãoCarlos,DepartamentodeQuímica,RodoviaWashingtonLuizkm235,SãoCarlos/SP,Brazil}

b_{UniversidadedeSãoPaulo,CentrodeEnergiaNuclearnaAgricultura(CENA),LaboratóriodeQuímicaAnalítica,CaixaPostal96,CEP13416-000,Piracicaba/SP,Brazil} c_{UniversidadeFederaldeSãoPaulo,DepartamentodeCiênciasExatasedaTerra,RuaProfessorArthurRiedel275,Diadema/SP,Brazil}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory: Received13April2011

Receivedinrevisedform22June2011 Accepted27June2011

Available online 3 July 2011

Keywords: LIBS

Macronutrients Micronutrients Plantmaterials Cryogenicgrinding Ballmilling

a

b

s

t

r

a

c

t

Ithasbeendemonstratedthatlaserinducedbreakdownspectrometry(LIBS)canbeusedasanalternative

methodforthedeterminationofmacro(P,K,Ca,Mg)andmicronutrients(B,Fe,Cu,Mn,Zn)inpelletsof

plantmaterials.However,informationisrequiredregardingthesamplepreparationforplantanalysisby

LIBS.Inthiswork,methodsinvolvingcryogenicgrindingandplanetaryballmillingwereevaluatedfor

leavescomminutionbeforepelletspreparation.Theparticlesizeswereassociatedtochemicalsample

propertiessuchasfiberandcellulosecontents,aswellastopelletsporosityanddensity.Thepelletswere

ablatedat30differentsitesbyapplying25laserpulsespersite(Nd:YAG@1064nm,5ns,10Hz,25Jcm−2_).

Theplasmaemissioncollectedbylenseswasdirectedthroughanopticalfibertowardsahighresolution

echellespectrometerequippedwithanICCD.Delaytimeandintegrationtimegatewerefixedat2.0and

4.5␮s,respectively.Experimentscarriedoutwithpelletsofsugarcane,orangetreeandsoyleavesshowed

asignificanteffectoftheplantspeciesforchoosingthemostappropriategrindingconditions.Byusing

ballmillingwithagatematerials,20mingrindingfororangetreeandsoy,and60minforsugarcaneleaves

ledtoparticlesizedistributionsgenerallylowerthan75␮m.Cryogenicgrindingyieldedsimilarparticle

sizedistributionsafter10minfororangetree,20minforsoyand30minforsugarcaneleaves.Therewas

upto50%emissionsignalenhancementonLIBSmeasurementsformostelementsbyimprovingparticle

sizedistributionandconsequentlythepelletporosity.

© 2011 Elsevier B.V.

1. Introduction

Thedeterminationofmacro(P,K,CaandMg)and micronutri-ents(B,Fe,Cu,MnandZn)inplantleavesisofkeyimportanceto evaluatethenutritionalstatusofcropsofeconomicinterestandis commonlyusedforthedetectionofnutritionaldeficiencies,that maylimittheproductionand/orqualityofe.g.fruits,vegetables andcereals[1].

In general,thedirect determination ofessential elementsin plant materialsis carried out in leavesproperly collected, and requires,atleast,threesamplepreparationsteps,namely clean-ing (washing), drying and homogenization [2]. Depending on themethodchosensuchasslurrysamplinginductively coupled plasmaopticalemissionspectrometry(ICPOES)[3],slurry sam-plinggraphitefurnaceatomicabsorptionspectrometry[4],solid samplinggraphite furnaceatomic absorptionspectrometry (SS-GFAAS)[5],X-ray fluorescence spectrometry [6], laser ablation

∗_{Correspondingauthor.Tel.:+551934294774;fax:+551934294774.}

E-mailaddress:[email protected](F.J.Krug).

inductivelycoupledplasmamassspectrometry(LA–ICP-MS)[7], instrumentalneutronactivationanalysis(INAA)[8]aswellasLIBS

[9,10],acomminutionstepisalsoneeded.

AccordingtoMarkert[2],comminutionofsamplematerialsis oneofthemostimportantstepsintheoverallanalyticalprocedure. Comminutionusuallyreferstothegrindingofsolidsamplesmade upoflargeparticlesintoapowderconsistingofsmallparticles.If appropriate,itimprovesanalytemicrohomogeneity,andprevents segregationofthematerialasaresultofwideparticlesize distribu-tionwithinthesample.Plantmaterialsarenothomogeneousand thedifferentconcentrationsofanalytesintheleavesmayimpair quantitativedirectmicroanalysis.Examplesofsuchvariabilityin plantleavescanbefoundinacomprehensivereview[11].

Indeed,thechoiceofthegrindingmethodgenerallydepends onthefollowingparameters[2]:(i)overallquantityandnumber ofsamplesofthematerialtobehomogenized;(ii)particlesizeof theoriginalsample;(iii)finenessofthematerialwhen ground; (iv)physico-chemicalpropertiesofthesampletobecomminuted andthegrindingequipment(possibilityofsamplecontamination and/oranalytevolatilization);and(v)hardnessofthematerialto behomogenized.

0039-9140© 2011 Elsevier B.V. doi:10.1016/j.talanta.2011.06.069

Open access under the Elsevier OA license.

85 (2011) 1744–1750 1745

RossbachandZeiller[12]pointedoutthatfinalparticle-size dis-tributioniscriticalforthepreparationofnaturalmatrixreference materialssuitable formicroanalytical techniques.Thematerials shouldexhibitalowmaximum(≤50␮m)andanarrowrange.On theotherhand,Kurfürst[13]andZeisler[14]suggestedthat sam-plecanbeconsideredhomogeneouswhenitsparticlesaresmaller than10␮m.Thus,grindingproceduresforsamplecomminution shouldbeeffectiveforobtainingsmallerparticlesizes,especially becausethesamplemassanalyzedbymicroanalyticaltechniques usuallyvariesfrom0.1to10mg[15].

Bothcryogenicgrinding[4,16–18]andballmilling[2,7]canbe effectivetoreduceparticlesize aimingatmicroanalytical tech-niques.Samplepreparationinvolvingcryogenicgrindingofplant materialsforpelletspresentationtoLIBSisdiscussedelsewhere

[10,19].

Arroyoetal.[7]investigatedthesoilcomminutionwitha plan-etaryballmillingpriortopelletpreparationforfurtheranalysisby LA–ICP-MS,demonstratingthattheparticlesizeofgroundmaterial wasgenerallysmallerthan1␮mafter20mingrindingintungsten carbidejars.Thishomogenizationprocedureimprovedcohesionof soilpelletswithouttheadditionofbindingagents.Furthermore,a representativesamplingatthemicro-scaleandappropriated repro-ducibilitywereobserved.

LIBSanalysisofpelletsofplantmaterialsaimingatthe determi-nationofmacroandmicronutrientshavebeenrecentlydescribed

[9,10,20–22].InLIBS,thelaser/sampleinteractiondependsstrongly onthelaserparametersandonthephysicalandchemical charac-teristicsofthesample[23,24].Inthecaseofgroundmaterials,large differencesinparticlesizedistributionandparticlecomposition amongsamplesmayalsoproducedifferencesinthelaser/sample interactionthusaffectingthequalityofresults[7,9,19].Forthesake ofinformation,insomespecialapplicationsLIBScanbeusedto investigatetheelementalaccumulationindifferentlayersand/or partsofplantleaveswithoutgrinding(Galiováetal.[25,26]and Kaiseretal.[27,28]),but thisisoutof thescopeofthepresent paper.

Theaimofthisworkwastoevaluatedsamplepreparation pro-ceduresforthecomminutionofleavespriortopelletpreparation forthedeterminationofmacro(P,K,CaandMg)andmicronutrients (Fe,Cu,Mn,ZnandB)byLIBS.Cryogenicgrindingandballmilling wereinvestigatedforsugarcane(Saccharumofficinarum), orange tree(Citrussinensis), andsoy(Glycinemax)leavescomminution. Thecorrelationsofsampleproperties(e.g.fiberandcellulose con-tents),particlesizedistribution,pelletporosity,and LIBSresults (i.e.emissionsignalintensityandrepeatability)wereinvestigated.

2. Experimental

2.1. LIBSinstrumentation

ExperimentswerecarriedoutwithaQ-switchedNd:YAGlaser (Brilliant,Quantel,France)at1064nm,generating5nspulsesup to365±3mJ,in a6mm beamdiameterwithquality factorM2

lowerthan2,at10Hzrepetitionrate.Thelaserpulsewasfocused onthesamplepelletbyaplane-convexlenswith2.54cm diam-eterand20cmfocallength(Newport,USA).Thelens-to-sample distance(LTSD)and thepulseenergy wereadjustedat 17.5cm and110mJ,respectively,leadingto25Jcm−2_{atthesamplesurface}

[20].

Individualtestsamples(i.e.15mmpelletdiameter)wereplaced inamanuallycontrolledtwo-axistranslationstagethatwasmoved intheplaneorthogonaltothelaserpropagationdirection.Argon flowingat0.5Lmin−1_{wascontinuouslyfedintotheablation}

atmo-spherebyoneentranceinletpositionedinthesampleholderand theflow-ratewascontrolledbyarotameter.Theplasmaemitted

radiationwascollectedbyusingafusedsilicalens(i.e.80mmfocal length) and collimatedintoa spectrometeroptical fiber(1.5m, 600␮mcore)matchingitsnumericalaperturebyusingalensof 50mmfocallength(LLAInstrumentsGmbH,Germany).The opti-calaxisofthecollectingsystemwasapproximately25◦ _fromthe laseraxis.

A model ESA 3000 spectrometer (LLA Instruments GmbH,

Germany)equippedwithechelleopticsandfocallengthof25cm withnumerical apertureof 1:10,and a 24.5mm×24.5mmflat imageplanewasused.Thiswasselectedasacompromisebetween resolution in thewavelength range of 200–780nmand resolv-ingpowerrangingfrom10to20pm.Thelineardispersionper pixelrangedfrom5pmat200nmto19pmat780nm.The

detec-tor is an ICCD camera, comprised of a Kodak KAF 1001 CCD

array of 1024×1024 pixels full frame (24␮m×24␮m) and a microchannelplateimageintensifierof25mmdiametercoupledto aUV-enhancedphotocathode.Theimagesignalsweredigitalized indynamicrange of16bitsandfurtherprocessedbyan indus-trialcomputer.Thedarkcurrentof theICCDwasautomatically subtractedfromthemeasuredspectraldata.Thedelaytime, inte-grationtimegateandthenumberofaccumulatedpulseswerefixed at2.0␮s,4.5␮sand25,respectively[20].

2.2. Samplepre-treatmentandgrindingprocedures

Thecollectionofleaveswascarriedouttakingintoaccountthe agriculturalrecommendationforplantdiagnosis.Leavesoforange tree,soyandsugarcanewerewashedseparatelywithrunningtap waterandfurtherrinsedtwicewithdistilledwaterandthreetimes withultrapurewater[29].Forsugarcaneleavesthecentralveinwas removedasrecommended[29].Forsoyandorangetreeleaves,the wholedriedleaveswereused.Afterwashing,samplesweredried, chopped,andoven-driedtoconstantmassat60◦_{C,andgroundby} usingaknifemill(Marconi,Brazil)withoutletapertureof600␮m. Theknifegroundsamplewasusedinordertofacilitatethegrinding processincryogenicorballmill.

Cryogenicgrindingwasperformedin acryogenic mill(Spex model6800,USA)withaself-containerliquidnitrogenbath.The pre-coolingtimewas5minandgrindingtimeswereintherange from10to50min(5–25cyclesof2mingrinding).Aftereach grind-ing cycle,themagnetic field wasturned off for 1minto allow samplere-cooling.Fourlaboratorysamplesof2gcanbe indepen-dentlyandsimultaneouslygroundandhomogenized.

Planetaryballmillingwascarriedoutin aRetschmodel PM 400mill(Germany)whichwasfurnishedwithfourgrindingagate jars (250mL; Retsch, Germany) containing 10 or 20g of sam-ple(<600␮m)with100agateballs(10mmdiameter). Grinding wasperformedat400rpm,during5minclockwise/5min counter-clockwisewith10-sstopbeforechangingtherotationdirection. Grindingtimesfrom20to60minweretestedfororangetreeand soyleaves,andfrom60to120minforsugarcaneleaves,basedon preliminaryexperiments.

PelletsofgroundleaveswerepreparedinaSpexmodel3624B X-Pressbytransferringapproximately0.5gofpowderedmaterialtoa 15mmdiesetandapplying8.0toncm−2_{during5min.Theresulting}

pelletswereapproximately2mmthickand15mmdiameter.Thirty differentsitesonthepelletsurfacewereanalyzed.Eachtestportion wasthemassremovedfromeachsiteofthepelletafter25laser shotsat25Jcm−2_{.Thebackgroundsignal,inthesurroundingof}

1746 85 (2011) 1744–1750

2.3. Particlessizeandpelletscharacterization

Thedeterminationofparticlesizedistributionofthe commin-uted leaves was performedby low angle laser light scattering usinganLS13320TornadoDryPowderSystem(Beckman Coul-ter,USA),accordingtotheISOguide13320-1:1999.Pelletsporosity wasdeterminedbyusingamercuryporosimeter(Aminco5000psi, USA). Pycnometric density of pellets was determined using a heliumpycnometer(Accupyc1330,MicromeriticsInstrument).

Themorphologicalcharacteristicsoftheanalyzedpellets(i.e. craters)wereevaluatedusingaLEOscanningelectronmicroscope (Stereoscan 440, UnitedKingdom).Pelletswerecovered witha thin Ptlayerduring80susinga Bal-Tec equipment(MED 020, UnitedKingdom)andthemicrographswereobtainedby apply-ingan electronaccelerationvoltageof 10kV usinga secondary electronsdetector.

Thecratervolumeonthepelletsurfaceafterlaserablationwas determinedwithaTaylorHobsonPrecisionperfilometer (Form-tracerSVC525,UnitedKingdom).

2.4. Determinationoflignin,celluloseandfibercontents

Thefiber,ligninandcellulosecontentsoftheorangetree,soy, andsugarcanegroundleaves(<600␮m)weredeterminedbyusing theAOACofficialmethod[31].Theaciddetergentfiber(ADF)was determinedastheresidueremainingafterdigestionwithH2SO4

and20gL−1_{cetyltrimethylammoniumbromide.Thefiberresidues}

werepredominantly cellulose and lignin.The ADF contentwas determined gravimetricallyas the residue remaining after acid detergentextraction,andthelignincontentwasdetermined gravi-metricallyaftertheADFresidueextractionwith72%(v/v)H2SO4

andashed.Cellulosecontentsweredeterminedbysubtractingthe pre-ashedligninvaluefromtheADFvalues.

2.5. ICPOESanalysisofdigests

Inordertoverifythepotentialcontaminationoccurrence dur-ing the comminution procedures with planetary ball mill (i.e. agatedevices)orcryogenicmill(polycarbonatetubesandstainless steeldevices),allgroundmaterialsweremicrowave-assistedacid digestedintriplicateandanalyzedbyradiallyviewedICPOES(Vista RL,Varian, Australia)withtheoperationalconditions described elsewhere[9,10].

3. Resultsanddiscussion

IthasbeenfrequentlymentionedthatLIBSisa spectrochem-icaltechnique witha minimalor no samplepreparation steps. However,inviewoftheinhomogeneouscharacterofpowdered materials,severaldrawbackscouldbeexpectedforquantitative analysisofplantmaterials[9].Microanalyticalmethodsbasedon direct solid samplinggenerally require grindingprocedures for samplehomogenizationpriortotheelementsdetermination[30]. Inthecaseofpelletanalysisassistedbylaserablation,thesmall par-ticlesmayincreasethecohesionofsamplepelletsimprovingthe reproducibilityofmeasurements.Thecohesionofaggregated par-ticlesaffectsprecisionbecausethemorecompactandmechanically resistantthepellet,themorereproduciblethelaser–sample inter-action[7].Indeed,ithasbeenstressedthatparticlesizedistribution isthemostimportantfactorrelatedtopelletspresentationtoLIBS thataffectslaserablationefficiencyandmeasurementprecision[9]. Themicroheterogeneityofelementsisanintrinsic characteris-ticofnaturalsamples[32],andthedirectanalysisofplantmaterials bymicroanalyticalmethodscanresultinhighcoefficientsof vari-ationofmeasurementsduetotheinhomogeneousdistributionof elementsinthematrix.Atourbestknowledge,twostrategieswere

50 40 30 20 10 0 12500 25000 37500 50000

Peak area (a. u.)

Grinding time (min)

B Mn P Mg

Orange tree leaves

(a)

50 40 30 20 10 0 12500 25000 37500 50000

Peak area (a. u.)

Grinding time (min)

B Mn P Mg

Soy leaves

(b)

50 40 30 20 10 0 8750 17500 26250 35000

Peak area (a. u.)

Grinding time (min)

B Mn P Mg

Sugarcane leaves

(c)

Fig.1.BI249.772nm,MnII257.610nm,PI213.618nmandMgI277.669nm emis-sionsignalintensities(peakarea)frompelletspreparedwith(a)orangetree,(b) soyand(c)sugarcaneleavesafter10to50minofcryogenicgrinding.Errorbars correspondto±oneestimatedstandarddeviation(n=30).

evaluatedforLIBSanalysisofpowderedplantmaterials.Onewas basedonthefixationofthepowderedmaterialonaflatsurface usingadouble-sidedtape[33,34]andtheotherbypressingthe groundleavestoformapellet[10].Althoughapparentlysimple, thefirstalternativerequirescarefulmanipulationfortheuniform depositionofthecomminutedsampleontothedouble-sidedtape forgettingreproduciblelayers.

85 (2011) 1744–1750 1747

Table1

Fiber,ligninandcellulosecontents(%m/m)inorangetree,soyandsugarcaneleaves. Uncertaintiesrepresentedby±onestandarddeviation(n=3).

Plantleaf Fiber Lignin Cellulose

Orangetree 26±0.3 7±0.2 19±0.1

Sugarcane 41±0.5 5±0.3 35±0.7

Soy 51±0.9 11±0.4 37±0.5

andK.However,significantstatisticaldifferencesinemission sig-nalintensitieswereobservedforsoyandsugarcanesamplesafter 10mingrinding.Itwasnecessary20and30mingrindingtoreach themaximumemissionsignalintensitiesforpelletspreparedwith particlesofsoy(meanparticlesize15␮m)andsugarcaneleaves

(meanparticlesize18␮m),respectively(Fig.1bandc).

Thefiber, cellulose and lignincontents in plantleaveswere determinedinordertoinvestigatethecorrelationbetween mate-rialcompositionanddurationofthegrindingstep(Table1).The dataobtainedindicatethatthehigherthefiberandcellulose con-tentsinthesamplesthehigherthegrindingtimenecessarytoreach themaximumemissionsignalintensitiesofanalytesinpellets ana-lyzedbyLIBS(Fig.1).However,it wasnot possibletoestablish correlationswiththelignincontentandgrindingtimesbecause orangetree,soyandsugarcaneleaveshaveanarrowrangeoflignin content(7–11%).AccordingtoVanSoest[35],woodtissueswitha lowlignin-to-celluloseratiotendtobendratherthanbreak,while withahighlignin-to-celluloseratiotendtobreakratherthanto bend.Consequently,thelesslignifiedtissuestendtogrindinlong pieceswhilemorelignifiedtissuestendtofragmentintoshorter pieces.Duringthegrindingprocesstheligninbecomesselectively distributedamongthelargerparticles[35].Inthepresentwork,the lignin-to-celluloseratioswere0.37,0.30and0.14fororangetree, soyandsugarcaneleaves,respectively,suggestingthatsugarcane groundleavessamplesaremoredifficulttocomminutingthanthe orangetreeandsoyleaves.

Initially,theplanetaryballmillingexperiments with250mL agatejarswere carried outby using50 grindingagate ballsof 10mm,asrecommendedbythemanufacturer.Inthiscondition, itwasobservedbyvisualinspectionthatcomminutionwas effec-tiveonlyfororangetreeandsoyleaves.Forsugarcaneleaves,100 grindingballswerenecessary.Inthisway,100grindingballswere selectedforthecomminutionofallsamples.

Fig.2ashowsthatatleast20mingrindingwithballmillwere neededinordertoreachthemaximumemissionsignal intensi-tiesfromtheelementsinthepelletspreparedwithparticlesof orangetreeleaves(meanparticlesize15␮m).Forpelletsofsoy leaves(meanparticlesize20␮m),nosignificantdifferenceswere observedin emission signalintensitiesfor B,Mn, Pand Mg by changingthegrindingtimefrom20to60min(Fig.2b).Ingeneral, theremainingaforementionedelementsdidnot present differ-ences at 95%confidence level in emission signalintensities by increasingthegrindingtime.Significantdifferencesonparticlesize distributionwereobservedforsugarcaneleavesprocessedby dif-ferentgrindingtimesusingballmill.Itmustbementioned that with20mingrindingitwasnotpossibletoobtainpelletswith ade-quatecohesion.Resultsindicatedthatatleast60minwererequired forappropriatesugarcaneleavescomminution(e.g.meanparticle sizeintheorderof15␮m).It isimportanttopoint outthatby increasingthegrindingtimefrom60to120min,themeanparticle sizewasreduced from15 to8␮m. However,nosignificant dif-ferenceswereobservedinemissionsignalintensitiesobtainedby LIBS(Fig.2c).

Ontheotherhand,bytakenintoaccountthepresenceofsome elementsofinterestinthecomponentsofthegrindingmaterials, experimentswerealsocarriedtoevaluatethepossibilityof sys-tematicerrorsduetocontamination.Agatematerialiscomposed

0 12500 25000 37500 50000

60

Peak area (a. u.)

Grinding time (min)

B Mn P Mg

Orange tree leaves

20

(a)

0 23750 47500 71250 95000

60

Peak area (a. u.)

Grinding time (min)

B Mn P Mg

20

Soy leaves

(b)

0 8750 17500 26250 35000

120

Peak area (a. u.)

Grinding time (min)

B Mn P Mg

60

Sugarcane leaves

(c)

Fig.2.BI249.772nm,MnII257.610nm,PI213.618nmandMgI277.669nm emis-sionsignalintensities(peakarea)frompelletspreparedwith(a)orangetree,(b)soy and(c)sugarcaneleavesafter20,60,and120minofplanetaryballgrinding.Error barscorrespondto±oneestimatedstandarddeviation(n=30).

1748 85 (2011) 1744–1750

Table2

Particlesizeparametersaftercomminutionoforangetree,soyandsugarcaneleavesbyballandcryogenicgrindings.

Sample Planetaryballmilling Cryogenicgrinding

Grindingtime(min) Meansize(␮m) d95(␮m) Grindingtime(min) Meansize(␮m) d95(␮m)

Orangetree 20 15 38 10 20 64

Soy 20 20 75 20 15 57

Sugarcane 60 15 54 30 18 62

d95=95%ofcumulativeparticlessmallerthan75␮m.

diesetwasinvestigatedusinghigh-puritycellulose. Theresults indicatedthattherewasnoperceptiblecontamination.

The reproducibility of laser–sample interaction and, conse-quently,therepeatabilityofmeasurementsdependonparticlesize usedforpelletpreparation.Table2shows thatthemean parti-clesizediameterwassmallerthan20␮mand95%ofcumulative particles(d95)weresmallerthan75␮m,afterplanetaryballand cryogeniccomminutionprocedures.LIBSanalysisofthepellets pro-ducedwiththeseparticlesresultedincoefficientsofvariationof measurementsintherangefrom5to20%(site-to-siteprecision,

n=30sites).Althoughthemeanparticlesizecouldbesmallerthan 10␮m,byincreasingthegrindingtime upto120minwithball mill,nosignificantimprovementswereobservedinemissionsignal intensitiesfromtheinvestigatedelements,aswellasin measure-mentprecision(site-to-siterepeatability).

TheparticlesizedistributionandSEMimagesofcraters pro-ducedbylaserablationinpelletspreparedwithparticlesoforange tree,soy and sugarcaneleaves afterball milling and cryogenic grindingareshowninFigs.3–5,respectively.

TheSEMimagesand perfilometricanalysisshowedthat pel-letspreparedwithparticlesobtainedbybothgrindingprocedures presentedsimilarcratermorphologies,althoughslightdifferences

werefoundamongplantspecies.Itwasobservedthatleaveswith lowerfibercontentsuchasfromorangetree(Table1)presented homogeneouscraterswithlowerborderdeformities(Fig.3).Itmust bementionedthatcraterdiameterandablationdepthperpulse dependprimarilyonlaserirradiance[36],butthisparameterwas keptconstantalongthiswork.Ingeneral,laserablationofpellets preparedwithparticlesizeslowerthan75␮mdidnotpresentsigns ofcrackandnobinderadditionwasnecessary.Themassablated ineachsamplingsitewasestimatedinviewofpelletdensityand cratervolumeobtainedbyperfilometricanalysis.Underthis con-dition,approximately100␮gwereablatedpersite(25laserpulses persite,n=30sites).

Althoughnotshown,thesmallertheparticlesthesmallerthe pelletporosity,andthiscouldbeanotherfactorcontributingfor bettercrateruniformity.Inthiswork,itwasobservedthat poros-ityvariesbyvaryingtheplantspeciesandthegrindingmethod.In spiteofsimilarparticlesizeprofilesaftercomminution,onlysoy leavesbybothcryogenicgrindingandballmillingaswellas sug-arcaneleaveswithballmilling,presentedapproximatelythesame pelletporosity(16%).Ontheotherhand,upto31%porositywas foundinpelletsofsugarcaneleavesandorangetreeleavesafter cryogenicgrinding.Thisshowswhyprecautionisrecommended

85 (2011) 1744–1750 1749

Fig.4.Particlesizedistributionandscanningelectronmicroscopicimagesofcratersproducedbylaserablation(25Jcm−2_{,25pulses,10Hz)inpelletsofsoyleaves.Samples} groundfor20minbyusingtheplanetaryballmill(aandb)and20minbycryogenicgrinding(candd).

1750 85 (2011) 1744–1750

forLIBScalibrationwithdifferentplantspecies,underthe experi-mentalconditionsoutlinedinthiscontribution.

4. Conclusions

Particlesizeisanintrinsicparameterthatinfluencesemission signalintensitiesinLIBS.Smallparticlesreducepelletporosityand mayinfluencetheablationprocess andupto50%emission sig-nalenhancementonLIBSmeasurementswereobservedformost elements.Bothplanetaryballmillingandcryogenicgrinding pre-sented good performances for comminuting knife mill ground leaves, allowingsimultaneous grinding and homogenization of fourindependentsamples,andcanberecommendedforpellets preparationforLIBS.Thecomminutiontimeforgettingsimilarand appropriateparticlesizedistributionsdependsonthegrinding pro-cess,ontheplantspecieandonfiberandcellulosecontents.Data obtainedherewithsuggestthatmatrixeffectscanbeminimizedby usingparticlessmallerthan75␮m,andbothcryogenicgrinding andplanetaryballmillingfitforthisintendedpurpose.

Acknowledgments

FinancialsupportandfellowshipsfromFundac¸ãodeAmparoà PesquisadoEstadodeSãoPaulo(FAPESP04/15965-2, 2010/16379-0, 2010/17158-8) and Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq140879/2008-0, 140926/2009-7, 305913/2009-3,578728/2008-7).AuthorsarethankfultoProf. Elis-abeteA.deNadaiFernandesforplanetaryballmillingfacilities,Dra. AnaRitadeA.NogueiraforICPOESanalysis,Dr.GilbertoB.Souza forthedeterminationofcellulose,lignin,andfibercontents,andto Prof.EliasA.G.Zagattoforsuggestionsandlanguageimprovements.

References

[1]K.Mengel,E.A.Kirkby,PrinciplesofPlantNutrition,fifthed.,KluwerAcademic Publishers,Dordrecht,2001.

[2]B.Markert,Sci.TotalEnviron.176(1995)45–61.

[3]D.Baralkiewicz,A.Kanecka-Hanc,H.Gramowska,Cent.Eur.J.Chem.5(2007) 1148–1157.

[4] C.Engelsen,G.Wibetoe,FreseniusJ.Anal.Chem.366(2000)494–503. [5] M.G.R.Vale, N.Oleszczuk, W.N.L.Santos,Appl.Spectrosc. Rev.41(2006)

377–400.

[6]W.E.Stephens,A.Calder,Anal.Chim.Acta527(2004)89–96.

[7] L.Arroyo,T.Trejos,P.R.Gardinali,J.R.Almirall,Spectrochim.ActaPartB64 (2009)16–25.

[8]M.Rossbach,P.Ostapczuk,H.Emons,FreseniusJ.Anal.Chem.360(1998) 380–383.

[9]L.C.Trevizan,D.SantosJr.,R.E.Samad,N.D.VieiraJr.,L.C.Nunes,I.A.Rufini,F.J. Krug,Spectrochim.ActaPartB64(2009)369–377.

[10]L.C.Trevizan,D.SantosJr.,R.E.Samad,N.D.VieiraJr.,C.S.Nomura,L.C.Nunes, I.A.Rufini,F.J.Krug,Spectrochim.ActaPartB63(2008)1151–1158. [11]B.Kratochvil,D.Wallace,J.K.Taylor,Anal.Chem.56(1984)113R–129R. [12]M.Rossbach,E.Zeiller,Anal.Bioanal.Chem.377(2003)334–339.

[13]U.Kurfürst,SolidSampleAnalysis:DirectandSlurrySamplingusingGF-AAS andETV-ICP,firsted.,Springer,Berlin,1998.

[14]R.Zeisler,FreseniusJ.Anal.Chem.360(1998)376–379.

[15] M.Stoeppler,W.R.Wolf,P.J.Jenks,ReferenceMaterialsforChemicalAnalysis: Certification,AvailabilityandProperUsage,firsted.,Wiley-VCH,Weinheim, 2001.

[16]D.Santos,F.Barbosa,A.C.Tomazelli,F.J.Krug,J.A.Nóbrega,M.A.Z.Arruda,Anal. Bioanal.Chem.373(2002)183–189.

[17]W.P.C.Santos,V.Hatje,L.N.Lima,S.V.Trignano,F.Barros,J.T.Castro,M.G.A. Korn,Microchem.J.89(2008)123–130.

[18]D.Koglin,F.Backhaus,J.D.Schladot,Chemosphere34(1997)2041–2047. [19] M.Pouzar,T. ˇCernohorsk ´y,M.Pr ˚uˇsova,P.Prokopˇcáková,A.Krejˇcová,J.Anal.At.

Spectrom.24(2009)953–957.

[20] L.C.Nunes,J.W.B.Braga,L.C.Trevizan,P.F.Souza,G.G.A.Carvalho,D.Santos Junior,R.J.Poppi,F.J.Krug,J.Anal.At.Spectrom.25(2010)1453–1460. [21]J.W.B.Braga,L.C.Trevizan,L.C.Nunes,I.A.Rufini,D.SantosJr.,F.J.Krug,

Spec-trochim.ActaPartB65(2010)66–74.

[22]L.C.Nunes,G.A.daSilva,L.C.Trevizan,D.Santos,R.J.Poppi,F.J.Krug, Spec-trochim.ActaPartB64(2009)565–572.

[23]A.W.Miziolek,V. Palleschi, I. Schechter, Laser-InducedBreakdown Spec-troscopy(LIBS):FundamentalsandApplications,firsted.,CambridgeUniversity Press,NewYork,2006.

[24]R.E.Russo,X.L. Mao,C. Liu,J. Gonzalez,J. Anal.At.Spectrom.19(2004) 1084–1089.

[25]M.Galiová,J.Kaiser,K.Novotn ´y,O.Samek,L.Reale,R.Malina,K.Páleníková, M.Liˇska,V.Cudek,V.Kanick ´y,V.Otruba,A.Poma,A.Tucci,Spectrochim.Acta PartB62(2007)1597–1605.

[26]M.Galiová,J.Kaiser,K.Novotn ´y,J.Novotny,T.Vaculovic,M.Liˇska,R.Malina, K.Stejskal,V.Adam,R.Kizek,Appl.Phys.A:Mater.Sci.Process.93(2008) 917–922.

[27]J.Kaiser,O.Samek,L.Reale,M.Liˇska,R.Malina,A.Ritucci,A.Poma,A.Tucci, F.Flora,A.Lai,L.Mancini,G.Tromba,F.Zanini,A.Faenov,T.Pikuz,G.Cinque, Microsc.Res.Tech.70(2007)147–153.

[28] J.Kaiser,M.Galiová,K.Novotn ´y,R. ˇCervenka,L.Reale,J.Novotn ´y,M.Liˇska,O. Samek,V.Kanick ´y,A.Hrdliˇcka,K.Stejskal,V.Adam,R.Kizek,Spectrochim.Acta PartB64(2009)67–73.

[29]E.Malavolta,G.C.Vitti,S.A.d.Oliveira,Avaliac¸ãodoEstadoNutricionaldas Plantas,seconded.,Potafos,Piracicaba,1997.

[30] G.G.A.deCarvalho,L.C.Nunes,P.FlorenciodeSouza,F.J.Krug,T.C.Alegre,D. SantosJr.,J.Anal.At.Spectrom.25(2010)803–809.

[31]P.Cunniff,OfficialMethodsofAnalysisofAOACInternational,Arlington,1995. [32] P.M.Gy,SamplingofHeterogeneousandDynamicMaterialSystems:Theories ofHeterogeneity,Sampling,andHomogenizing,Elsevier,Amsterdam,1992. [33]Q.Sun,M.Tran,B.W.Smith,J.D.Winefordner,Can.J.Anal.Sci.Spec.44(1999)

164–170.

[34] T.Ctvrtnickova,M.P.Mateo,A.Ya ˜nez,G.Nicolas,Spectrochim.ActaPartB65 (2010)734–737.

[35]P.J.VanSoest,NutritionalEcologyoftheRuminant,seconded.,Cornell Univer-sityPress,NewYork,1994.