High-time resolution and size-segregated elemental composition in high-intensity pyrotechnic exposures

(1)

ContentslistsavailableatSciVerseScienceDirect

Journal

of

Hazardous

Materials

jo u rn al h om epa g e : w w w . e l s e v i e r . c o m / l o c a t e / j h a z m a t

High-time

resolution

and

size-segregated

elemental

composition

in

high-intensity

pyrotechnic

exposures

Javier

Crespo

a,∗

,

Eduardo

Yubero

a

_,

_Jose

_F.

_Nicolás

a

_,

_Franco

_Lucarelli

b

_,

_Silvia

_Nava

b

_,

Massimo

Chiari

b

_,

_Giulia

_Calzolai

b

a_Laboratorio_de_{Contaminación}_Atmosférica_(LCA),_Área_de_Física_Aplicada,_Universidad_Miguel_Hernández,_03202,_Elche,_Spain b_Department_of_Physics_and_Astronomy_and_INFN,_University_of_Florence,_Sesto_Fiorentino,_50019,_Florence,_Italy

h

i

g

h

l

i

g

h

t

s

Streaker/Dekati/PIXEaresuitabletostudyhigh-intensityshort-livedﬁreworkevents. Fireworksrelatedelementsshowed>80%elementalmassinthesubmicrometricregion. Resuspendedsoildustisanimportantsourceofatmosphericparticlesintheseshows.

a

r

t

i

c

l

e

i

n

f

o

Articlehistory: Received30March2012

Receivedinrevisedform5September2012 Accepted9September2012

Available online 19 September 2012 Keywords: Pyrotechnicemissions Airpollution Metalairborne PIXE Aerosolsize-distribution

a

b

s

t

r

a

c

t

TypicaloffestivalsinEasternSpain,mascletàsarehigh-intensitypyrotechniceventswherethousands offirecrackersareburntinanintense,rapidepisodethatgeneratesshort-livedheavyaerosolclouds. Hightemporalresolutionandsizedistributioncharacterisationofaerosolcomponentswereperformed toevaluatetheeffectsofthebrief(<30min)andacuteexposureonthespectatorspresent.Veryhigh con-centrationsoffireworkspecificelements,especiallyinthefinefraction,werereachedduringmascletàs, withvaluesofabout500␮g/m3_for_K_and₃₀₀_␮g/m3_for_Cl._Sr,_Al,_Mg,_Ba,_Cu,_Co,_Zn,_and_Pb_{concentration}

increasefactorsofmorethan100(1000forSrandBa)wereobservedinthefinefractionwithrespect tobackgroundlevels.Crustaloriginelements,likeCa,Fe,Si,Ti,alsoshowedanimportantconcentration rise(∼10timesabovebackgroundlevels)butthisisduetodustresuspensionbypyrotechnicexplosions. Thecrustalcomponentsaremainlyinthecoarsemode(>90%elementalmass),between2and3␮m. Mostfireworkrelatedmetalsareconcentratedinthesubmicrometricregion(>80%)withatrimodalsize distribution.Thismaybeinterestingtoepidemiologistsgiventhetoxiceffectsthatsuchfine,metal-rich particlescanhaveonhumanhealth.

1. Introduction

Inthelastfewyearsseveralresearchstudieshavebeen pub-lished in literature reporting on urban airborne particles data related toworldwide popularpyrotechnic events, suchas New Year’sEve[1],nationalfestivities[2,3],sportingchampionship cel-ebrations [4],and light festivals [5,6]. Most of these published studieswerecarriedoutfrompre-eventtopost-eventpyrotechnic performancesinordertoevaluateurbanbackgroundairquality variationsusually withinanhourlyor dailytime resolution.As ageneralconclusion, ﬁreworkdisplaysgiverisetolarge(upto hundredsof␮g/m3₎_but_transitory_(up_to_hours_or_days)_increases

∗ Correspondingauthorat:LaboratoriodeContaminaciónAtmosférica(LCA),Área deFísicaAplicada,UniversidadMiguelHernández,AvenidadelaUniversidadS/N, 03202,Elche,Spain.Tel.:+34966658326;fax:+34966658397.

E-mailaddress:jcrespo@umh.es(J.Crespo).

ofurbanatmospheric particulatemattermeanlevels,especially metalliferousparticles(K,Mg,Ba,Cu,Sr,Al,Pb)[4].Forexample, intheIndianDiwalifestival,24h-averageconcentrationofPM10

increasedupto5.7timeswithrespecttoanormalday[7,8];during theLanternfestivalinChinatheriseofPM2.5wasover6timesthe

usualdailylevel[9];intheSummersolsticefeastinGirona(Spain) PM2.5dailylocalbackgrounddoubled,from13␮g/m3to25␮g/m3,

multiplyingbyafactorrangingfrom2to86theconcentrationsof metalsandmetalloidstypicallyemittedfromﬁreworks[10].

Duringfireworkdisplays,allsizerangesofatmospheric parti-clesrisebuttheincreaseisparticularlysignificantforfineparticles [4].Ascanbefoundincitedliterature,0.4–2.0␮msizebinisthe characteristicintervalrelatedtoatmosphericaerosoloriginating frompyrotechnicexhibitions[1,2,11].Thisintervalcorrespondsto breathableairborneparticlefractionandthereforeitisexpected, takingintoaccountthereportedhighlevelsofatmosphericmetal, metalloid and other pollutants, that substantial human health effects can be induced by fireworks. Consequently, numerous 0304-3894/$–seefrontmatter © 2012 Elsevier B.V. All rights reserved.

(2)

J.Crespoetal./JournalofHazardousMaterials241–242 (2012) 82–91 83

Fig.1. Locationmapandpictureofsamplingsite.Whitearrowsindicatenorthdirection.

papershavebeenpublisheddescribingthehazardousimpactthat bothlong-termandshort-termﬁreworkeventscanhaveonhuman health[12–16].

Mostofthestudieshavebeendesigned,asmentionedbefore, tomeasuredailyurbanbackgroundlevelsofairbornefine parti-clesbymeansofstandardPMsamplers,belonging,inmanycases, topublicmonitoringnetworks,normallyatconsiderabledistances (>1km)fromsiteswherefireworkdisplaysarelaunched[1,17,18]. Eventhoughitisverycommonthatpeopleattendingsuchanevent canbeengulfedbysmokeplumesarisingfrompyrotechnic dis-plays,theimpactofthebriefandacuteexposuretotheseplumeson spectatorsstandingclosetoanoutdoorshort-livedfireworkevent isscarcelyknown[19,20].

Thispaperisfocussedonaerosolsgeneratedbyhigh-intensity pyrotechnicevents,calledmascletàs.Theseeventstakeplace dur-ingannualfestivalsinnumerouscitiesintheSpanishsouth-eastern Mediterranean region (Valencian Region) (Fig. 1). Unlike usual fireworks,mascletàsareintended largelytostimulate the audi-torysystemandbodyvibrationthroughthestrongrhythmicnoise sequenceproducedbytheburningofatypeofbangerscalled mas-clets.Themascletisapowerfulsoundfirecrackerthatcanbeburst bothatgroundlevelandatlowaltitudelevel(1.5–2.0m).When itburns,aloudintensitydetonation,lighteffects andabundant smokegenerationareproducedasafunctionofitscomposition. Thousandsofmascletsarelinkedbyawickformingaline(traca)2m abovegroundlevel,inrectangularorcircularconnectedstructures, andaresuccessivelyburstathighspeed.Furthermore,during mas-cletàs,hundredsoffireworkshellsarealsoshotintotheairusing verticalcannons(mortarsorlaunchtubes).Theyproducepowerful explosionsat20–40mhighthatareusuallyaccompaniedbybursts ofwhitelightandcolours.

Peopleattending theseeventsare directlyenvelopedbythe aerosolcloudswhichareproduced,soourmainobjectiveinthis paperistocharacterise,asbestaspossible,theairborneparticles originatedbysuchnear-groundlevel,short-term,highintensity outdoorﬁreworkepisodes.Forthispurpose,wecarriedoutastudy toconsidertheparticularcharacteristicsoftheseepisodes.Indoing so,we determined elemental composition during mascletàs by meansofhightimeresolutionanalysisandparticlesizesegregation samplers.

Animportant methodological difference exists betweenour studyand themajorityofthestudiesfoundintheliterature.In

thesestudiesthesamplingpointislocatedfarawayfromthe fire-workslaunchingzone.Asaconsequencethefireworkaerosolsare founddispersedanddilutedintheatmosphericairmass. Further-more,thelocalmeteorologicalconditionscanchangetheobtained ambientPMconcentrationsindependentlyofthetypeoffirework event.Inourcase,thesamplingpointislocatedveryclosetothe launchingzone,sotheemissionaerosol cloudismeasured. Itis importanttoremarkthatthisdifferentstrategyaffectsthe extrap-olationoftheresults.ItisnotthesametoconsideremissionPM concentrationsthaninmissionconcentrationwhenhuman expo-sitionarebeingcompared.Anestimationofthemetallicparticles doseinhaledbythepeopleassistingtothepyrotechniceventshas beencalculated.

2. Experimental

2.1. Samplingsite

Alicante (38◦2043N, 0◦2859W) is a medium-size city (∼350,000 inhabitants)inSESpain(Fig.1).The cityliesonthe shoresoftheMediterraneanSea,extendedfromnorthtosouthin aslopingterraintowardsthesea(4%),surroundedbyanumberof hillsandelevations.Alicantehasanarid,Mediterraneanclimate withmildtemperaturesallyearround(annualaverageof17.8◦C) andscarcerain(336mm/year)mostlyconcentratedintheautumn period.Themeandailytemperaturesvaryfrom19◦Cto30◦Cin summerandfrom7◦Cto17◦Cinwinter.Predominantwind direc-tionisfromthenorth-westinwinterandfromthesouth-eastin summer,withatypicalregimeofMediterraneanSeabreezesinthe lattercase.

Duringthe2007JunefestivalofAlicanteHoguerasdeSantJoan (SaintJohn’s Bonﬁres)thepopularsonorousand visualdisplays ofmascletàsoccurredin theafternoons(at2p.m.)from19thto 24th,withhighattendance.Differenttypesofexplosivedevices werestrategicallysituatednexttoandaroundadowntowncity square(∼10,000m2₎_(Fig.₁₎_and_ignited_under_computer_control.

Thousandsofmascletsandﬁrecrackershells,terrestrial(75%)and aerial(25%),wereburntinahigh-intensity(125kggunpowder), short-time(_∼8min)loudepisode(average_∼80dBA).Intheﬁnal

minutenolessthan500powerfulbangersweredetonatedwith noise peaksupto 120dBA and withan earthquakeeffect. Asa

(3)

Fig.2. WhitesmokecloudtypicallygeneratedduringAlicanteFestivalMascletà.

consequence,thepeople(∼20,000)concentrated(50mfaraway)in thestreetsaroundmascletàsquareandonnumerousoverlooking balconiesfromthesurroundingapartmentsblockswereengulfed bydensewhiteaerosolclouds andgasses producedthroughout theevent(Fig.2).DespitethedailyafternoonMediterraneanSea breezesthatfavourcontaminantsdispersion,veryintenseshort pollutionspikeswereexpected.Thesepollutionspikesareof spe-cialconcerntoindividualsinattendancesincemassiveincreases ofsuspendedparticles(upto1500␮g/m3 _1-h_PM

2.5)havebeen

reportedinanurbanenvironmentalexposurestudyduringa ﬁre-workepisode[19].

2.2. Samplingandanalysis

Toassessthechemicalenrichmentofaerosolcloudsthatengulf individualsinattendance,westudieda)theﬁneandcoarse par-ticlehourlyelementalcompositionandb)theparticleelemental sizedistributionofaerosolemittedonthe19thofJunemascletà display.Forthispurposeweinstalledthesamplerinstrumentson abalconysituatedataheightof15mfromgroundlevelandwithin 50mfromthepyrotechnicsiteundertheinﬂuenceoftheplume. Thesamplingswerecarriedoutatdifferenttemporalresolution dependingoninstrumentandsoughtresults.

2.2.1. Fineandcoarseparticleelementalcomposition:streaker Theaerosolwascollectedduringoneweek(June18–24)bya “streaker” sampler (P.I.X.E. InternationalCorporation) [21].This device is designed to separate the fine (<2.5␮m aerodynamic diameter)andthecoarse(2.5–10␮m)fractionsoftheaerosol.A paraffin-coated Kapton foilis used asan impaction surface for coarseparticlesandaNucleporefilterasafineparticlecollector. Thetwocollectingplatesarepairedonacartridgethatrotatesat aconstantspeedforaweek:acircularcontinuousdepositionof particulatematteratbothstagesisproduced.Therotationspeed duringsampling(1.2mm/h),thepumpingorificewidth(1.2mm) andthebeamsizenormallyusedforthesubsequentPIXE (Particle-InducedX-rayEmission)analysisaresuchthatanoverallresolution ofaboutone houris obtainedontheelementalcomposition of airparticulate[22,23].Howeveritmustbetakenintoaccountthat wehadtochangethe1.2mmnozzlewitha2mmonetoprevent

ﬁltercloggingduringtheintensemascletàevent;therefore,since therotationspeedofthesamplercouldnot bechanged,alittle smoothingofthe1hconcentrationtimetrendsshouldbeexpected. PIXEanalyseswereperformedwith3MeV protons fromthe 3MV Tandetronaccelerator of theLABEC laboratoryofINFN in Florence,withtheexternalbeamset-upextensivelydescribed else-where[24,25].Thebeamscannedthestreakinstepscorresponding to1hofaerosolsampling;eachspotwasirradiatedforabout180s withabeamintensityrangingfrom20nAto30nA.PIXEspectra wereﬁttedusingtheGUPIXsoftwarepackage[26]andelemental concentrationswereobtainedviaacalibrationcurvefromasetof thinmonoorbi-elementalstandardsofknownsurfacedensityby Micromatter[27].Theexperimentaluncertaintiesontheelemental concentrations(ng/m3₎_measured_by_PIXE_are_given_by_the_sum_of

independentuncertaintiesonthefluxmeasurement(about2–3%), oncertifiedstandardsampleconcentrations(5%)andonpeakareas. TheuncertaintyonpeakareasincludestheX-rayscounting statis-ticsandallthefittinguncertainties(backgroundsubtraction,peak overlaps,etc.).Thecountingstatisticsuncertaintymayvaryfrom fewpercentupto20–30%ormorewhenconcentrationsapproach minimumdetectionlimits(MDLs).

Detectionlimitswereabout10ng/m3_for_elements_from_Na_to

Vand1ng/m3_(or_below)_for_elements_from_Cr_to_Pb._The_following

elementsweredetected:Na,Mg,Al,Si,S,Cl,K,Ca,Ti,V,Cr,Mn,Fe, Co,Ni,Cu,Zn,As,Br,Rb,Sr,Ba,andPb.

2.2.2. Particleelementalsizedistribution:Dekati

Particulatematter sampleswerecollectedfor both the19th Junemascletàperiod(∼50min)andduringanon-festivalday (nor-malsummerday,NSD),1st_July_(daily_average)._We_used_a_cascade

smalldepositarealowpressureimpactor(SDI,Dekati®_)._The

sam-plerclassifies airborneparticles, from30nmto10␮m,into 12 size-fractionsmalldepositareastages,withcut-offaerodynamic diameters(␮m)of8.50,4.08,2.68,1.66,1.06,0.796,0.591,0.343, 0.231,0.153,0.086,and0.045.Theparticlesarecollectedon25mm Kaptonfoilswithaflowrateof11L/min.Inordertoreduceparticle bounce-offeffects,thefilmswerecoatedwithapiezon.

Impactor samples were chemically analysed by PIXE at the LABECLaboratory[28].Fieldblankcorrectionwasappliedtoallthe datasets.TherelativePIXEuncertainties(duetocountingstatistics

(4)

Table1

Elemental1-hconcentration(ng/m3₎_averaged_for_the_whole_festival_week_(aver),_during_(mas)_and_out_(bck)_of_the_mascletàs_events_for_the_ﬁne_and_coarse_fraction._The

ﬁne/coarseratioduringandoutofthemascletàseventsisalsoreported.Theconcentrationduringmascletàshasbeencalculatedastheaverageofthepeakvaluesreached throughoutthesixepisodes.Thebackgroundconcentrationreferstothemeasurementscarriedoutattimesduringtheweekwhenneithermascletàspeaksnorotherﬁesta eventstookplace.

Element Fine Coarse f/c

aver mas bck aver mas bck mas bck

Na 1020 1440 1250 2100 1400 3700 1.0 0.3 Mg 710 7060 130 380 2300 380 3.1 0.3 Al 1960 22500 80 1770 31800 190 0.7 0.4 Si 1890 3000 430 710 3690 460 0.8 0.9 S 8670 80900 1450 950 12500 400 6.5 3.6 Cl 18700 216000 120 42300 13300 3420 16.2 0.04 K 33600 384000 250 2380 43200 240 8.9 1.0 Ca 1560 6800 390 3310 22600 1630 0.3 0.2 Ti 57 610 <7a ₄₇ ₄₃₀ ₁₉ _1.4 _– V 36 190 22 26 54 11 3.4 2.0 Cr 16 150 7 9 31 8 4.9 0.9 Mn 23 65 5 9 45 4 1.4 1.1 Fe 250 850 106 360 1900 280 0.4 0.4 Co 105 1700 <1a ₁₀₀ ₂₀₀₀ ₅ _0.8 _– Ni 8 18 7 2 8 2 2.3 3.9 Cu 120 2200 12 17 180 10 12.4 1.3 Zn 50 362 8 18 80 9 4.6 0.9 As 20 80 <1a ₃ ₁₀ _<2a _8.0 _0.9 Br 9 36 5 3 9 3 3.9 1.8 Rb 22 63 3 5 19 2 3.4 1.4 Sr 410 5750 4 40 780 10 7.4 0.4 Ba 1600 7430 <20a ₂₂₀ ₇₅₀ _<40a _9.9 _– Pb 160 1270 8 12 100 5 12.8 1.7

a_{Concentrations}_under_experimental_MDL_values.

andspectrumdeconvolution)typicallyrangedfromabout1–10%, asexplainedinpreviousparagraph.Theaerosolsamplingin gen-eralwasexpectedtointroduceanadditionalrelativeuncertainty ofabout10%[29].Thetwocontributionsweresquaresummedto givetheoveralluncertainties.

TherawelementalsizedistributiondataoftheindividualSDI sampleswereconvertedintosmoothdistributionsusingthe inver-sioncodeMICRON[30]andlognormalcurveswerethenﬁttedtothe latter,sothattheelementalmassconcentration,geometricmean aerodynamicdiameter(GMAD)andgeometricstandarddeviation (GSD)ofthedifferentcontributingmodeswereobtained.

2.3. Additionalmeasurements

Meteorological data (temperature, relative humidity, wind speed and direction) for the study period were obtainedfrom theenvironmentalmonitoringstationofAlicanteUniversity.The stationissituated3kmnorthofthemascletàsiteinaclearand well-ventilatedarea.Seabreezes,duringfestivalweek,blewmainly fromSE,between127◦and141◦.

3. Resultsanddiscussion

3.1. Elementalconcentrationwithone-hourtemporalresolution 24-hconcentrationresolutionisaperiodoftimemuchlonger thanthedurationofthedisplayssocorrespondingdatabasecoming fromdailysample analysisit isnotadequatetoassessthe fire-workoriginofdifferentelements,especiallythosethatcouldalso comefromsourcesotherthanfireworks(e.g.Cufromtraffic). Fur-thermore,theindividualexposureperiodofspectatorsattending short-termoutdoordisplaysismuchshorterthan24h.Forthese reasons,1-hresolutionsamplingismoresuitablebecauseitallows evaluatingwithoutanydoubtthefireworkoriginofthedifferent elementsanditmakespossibletoquantifytherealexposureof spectatorstoshort-livedfireworksplume.

The elemental hourly temporal evolution, both in the fine (<2.5␮m)andcoarse(2.5–10␮m)fraction,wascarriedoutusing theStreakerdevice.InTable1,theelemental1-hconcentration averagedduringandoutofthemascletàseventsandoverthewhole festivalweekforthefineandcoarsefractionaresummarised.The fine/coarseratioisalsoreported.Theconcentrationduring mascle-tàshasbeencalculatedastheaverageofthepeakvaluesreached throughoutthesixepisodes.Thebackgroundconcentrationrefers tothemeasurementscarriedoutalongtheweekinperiodswhen mascletàpeaksdidnot takeplaceandshouldrepresentthecity “normal”concentrations.Theconcentrationswefoundarehigher thanthosefoundinotherstudies[4,10,17]andformanyelements similartotheonesfoundbyKuhlteshtaetal.andSarkaretal.for Diwalifestival[6,18]andWangetal.forLanternFestival[5](our datawereaveragedinthesametimescalesusedinthecitedarticle). AscanbeobservedinTable1,duringthemascletàeventsthe massconcentrationsofmostof theanalysed elementsshowan extraordinaryincrease relative tobackground levels(especially forfinefraction).Potassiumandchlorineexhibitthehighest lev-els(384␮g/m3_and₂₁₆_␮g/m3_,_{respectively)}_and_relative_increases

sincebackgroundconcentrationsaremultipliedbyafactorhigher than1500duringtheseevents.Theconcentrationsofother ele-mentsrelatedtopyrotechnicdisplays,likeSr,Al,Mg,Ba,Cu,Co, Zn,andPb,arealsoextremelyhighduringthemascletàs.Intheﬁne fraction,increasefactorswithrespecttobackgroundlevelsrange fromaround50forMgandZn,tomorethan1000forSrandBa.

InFig.3a–cweshowtheconcentrations(␮g/m3₎_of_(a)_K,_Cl,

S(finefraction);(b)Al,Ba,Sr(finefraction);and(c)Ca(fineand coarsefraction)registeredduringthewholeHoguerasweek.Ascan beseenfromthefigures,thehightimeresolutionallowsustofollow therapidevolutionofthepollutantconcentrationduetothe short-timedailyevents.Allthereportedelementshadverysharppeaks onthetimeofmascletàsburning.Theyalsoshowedaquickdecrease aftertheeventsduetohighdispersioncapabilityoftheatmosphere favouredbyseabreezeandthermalconvectiveprocesses.Asimilar timepatternisalsopresentedbyMg,Si,Ti,V,Cr,Mn,Fe,Co,Cu,Zn, Br,Rb,Ba,andPb.Mostoftheseelementswerealreadyidentified

(5)

Fig.3.HourlytimetrendsoffineK,Cl,S(a);fineAl,Sr,Ba(b);fineandcoarseCa(c).ThelackofdataduringsomehoursonJune24thwasduetoapowerblack-out.

inliteratureastracersofﬁreworkemissions.However,itisclear thatthedifferentnatureofmascletàsrespecttostandardﬁreworks mayvaryenormouslytheratioofmetalsintheheterogeneousand denseaerosoltheyproduce.

Fig.3ashowsthetemporalpatternofK,ClandS concentra-tionsintheﬁnefraction.Levelsmeasuredinsidetheplumeinour studywereroughlyofthesameorderofmagnitudeasthosefound byDrewnicketal.andJolyetal.[1,19]andfarhigherthanthose foundbyDutcheretal.[11]whichreferstoindoorparticulate.These elementsoriginatefromgunpowdercombustion[11],stillthe prin-cipalcomponentofactualpyrotechnicdevices.Pyrotechnicprocess basicallyconsistsinanoxidation–reductionreactionbetweenan oxygensource(oxidant),typicallypotassiumnitrateorperchlorate, andsomereducingagents(fuel),classicallycharcoalandsulphur

butalsoothercomponentslikebinders,metalpowdersorsmoke generatingagentsaddedtotheaforementionedsubstancesinorder toreachthedesiredpyrotechnicperformances[11,31].

Dutcher et al.[11] reported thatgunpowder, typically com-posed of74.0% KNO3, 10.4%S, and 15.6% C bymass, showsan

ambientairK/Smassratioafterﬁreworks,practicallyequivalent thetheoreticallydeterminedvalue,2.76.Incontrast,themeasured ambientaverage K/Smass ratio foundin this study wasmuch higher,about5.0,notfarthatobtainedbyCroteauetal.(5.73)ina close-monitoredpyrotechnicevent[20].Furthermore,theambient K/Clmassratioobtained(between1.1and3.1alongmascletàweek) indicatesthatmascletàepisodesreleasedintotheatmospherelarge quantitiesofchlorinetogetherwithpotassiumintheﬁnefraction. Theseresultsseemrelatedtotheuseofpotassiumperchlorate,a

(6)

Table2

Geometricmeanaerodynamicdiameter(GMAD,␮m),associatedgeometricstandarddeviation(GSD,␮m)andrelativemodeconcentration(RMC)ofthedetectedmodesfor investigatedelementsoftheMascletàsamples.

Element Accumulationmode Intermediatemode Coarsemode

GMAD GSD RMC GMAD GSD RMC GMAD GSD RMC

Mg 0.22 1.87 0.62 0.85 1.84 0.31 3.356 1.53 0.08 Al 0.21 1.68 0.18 2.10 2.11 0.82 Si 0.14 1.88 0.13 2.13 2.27 0.87 S 0.19 1.50 0.41 0.67 1.62 0.49 2.978 1.53 0.10 Cl 0.19 1.47 0.44 0.61 1.53 0.53 2.378 1.26 0.03 K 0.19 1.51 0.44 0.65 1.60 0.50 2.75 1.45 0.07 Ca 0.16 1.72 0.05 2.49 2.13 0.95 Ti 2.32 2.17 1.00 V 0.34 2.75 1.00 Cr 0.50 3.35 1.00 Mn 0.32 2.63 0.69 2.15 2.43 0.31 Fe 0.284 1.40 0.14 1.13 1.46 0.42 3.39 1.41 0.44 Co 0.319 1.18 0.06 1.03 1.37 0.43 3.24 1.47 0.51 Cu 0.24 1.63 0.60 0.64 1.45 0.37 2.9 1.35 0.04 Zn 0.30 2.36 0.95 3.34 1.47 0.05 As 0.34 2.23 0.96 3.46 1.73 0.04 Rb 0.404 2.52 1.00 Sr 0.24 1.65 0.61 0.69 1.41 0.37 2.719 1.25 0.02 Ba 0.197 1.61 0.58 0.604 1.57 0.38 2.955 1.38 0.05

morepowerfuloxidiser,whenintensedaylightﬂashesand

pro-nouncednoiseeffectsarerequired,asoccurredinmascletàs[32].

So,KClO4appearstobeamajorsourceoftheKfoundinmascletà

PMfractions.Itiscommontheuseofanexcessofoxidisersinthe compositionoftheﬁreworksmixtureinordertoassurea com-pletecombustion.Inthiscase,someofthepotassiumperchlorate couldremainwithoutreactingintoaerosolcloudafterthe explo-sions.Thatwouldimplyanadditionalharmfuleffectforthepeople attendingthemascletàsgiven thehightoxicity of ClO4− ionon

humans[33]

Moreover,takingintoaccounttheelevatedchlorine concentra-tionfoundduringtheeventsandtheurbanpollutionbackground ofAlicante,organicchloridesperhapsshouldalsobeformedafter mascletàsincreasingstillfurthertheharmfuleffects.

In addition, it wasfound that the K/S mass ratio is almost inversely proportional to K/Cl mass ratio, (Pearson correlation, R2₌_0.96)_indicating_a_competition_for_potassium_between_S_and

Clasa functionofavailability oftheseelementsresultingfrom chemicalcompositionofdailyemployedpyrotechnicdevices.

Inthecoarsefractiontheconcentrationsobtainedforthese ele-mentswerefarlowerthanthosefoundinfinefraction(average fine/coarseratiosduringmascletàwere8.9;16.2;6.5forK,Cland S,respectively).Sconcentrationsofcoarsefractionwerehigher thaninthefinefraction,withanaverageCl/Smassratioof1.2as comparedto2.6forthefinefraction.Ithastobenotedthatin back-groundconditions,Clwasmainlyinthecoarsefractionandshowed atimepatternverysimilartothatofNa,pointingtoacommon ori-ginwhichissea-saltaerosolsource.Furthermore,theaverageK/S ratiowas,inthiscase,closeto2.8,indicatingalowerinfluenceof mascletàgeneratedClinthecoarsefraction.Inaddition,aweak correlation(R2_<_0.4)_between_K/S_and_K/Cl_mass_ratios_was_found

whichalsosuggestsdifferentsourcesfortheseelements.

In Fig.3b thetemporal evolution of other elements related withcrackersand ﬁreworksactivity,namelyAl,Sr,and Ba,are presented.Aluminiumisusedasacommonconstituentforfuel, aloneortogetherwithmagnesium(whichindeediscorrelatedto Alformascletàepisodes:R2_∼_0.80)_as_the_alloy_magnalium_(50:50

MgAl)toproduce sparksandﬂash effects[34].Duringmascletà periodstheconcentrationsofAlreachedvaluesintheﬁnefraction ofabout35␮g/m3_and₄₅_␮g/m3 _in_the_coarse_one,_two_order_of

magnitudehigherthanbackgroundvalues(<0.2␮g/m3_in_all_cases).

Bariumand strontium salts are used ascolour agents, impart-ingaredorgreencolour,respectively,toﬁreworkdisplays.They

arealsoimportant forstabilisingﬁreworksandbarium nitrates orchloratesasoxidisertoo[4,10].SrandBa,elementswithstill lower backgroundconcentrations (<10ng/m3_), _reached

concen-trationpeaksduringmascletàsupto16and10␮g/m3_in_the_ﬁne

fraction,respectively,andabout1.5␮g/m3_in_the_coarse_fraction.

TheconcentrationofSrandBafoundin thegenerated mascletà plumeisnotconstantforeachfestivaldaybecause,asalready men-tioned,thecompositionofthereleasedparticulatematterdepends onspeciﬁctypeofpyrotechnicdeviceused.Forexample,duringthe 23rdJunemascletà,therewereasmallcontributiontoﬁnefraction fromSrandnegligiblefromBa,whileinthenextdaySrandBa concentrationsincreasedupto16and8␮g/m3_,_{respectively.}

Similarpatterns,withpeakseverydayat2p.m.,areshownby Mg(maximumconcentrationsfortheﬁneandthecoarsefraction, respectively,10and4␮g/m3_),_V_(0.4_and_0.06_␮g/m3_),_Mn_(0.1_and

0.07␮g/m3_),_Cu₍₇_and_0.3_␮g/m3_),_Zn_(0.7_and_0.1_␮g/m3_),_Br_(0.1

and0.01␮g/m3_), _Rb_(0.08_and_0.02_␮g/m3_)._A_few_elements_are

presentonlyinsomemascletàevents:Cr(maximum concentra-tionsfor theﬁneandthecoarsefraction,respectively,0.16 and 0.05␮g/m3_),_As_(0.17_and_0.01_␮g/m3_),_Pb₍₅_and_0.2_␮_g/m3_),_and

Co(1.8and6␮g/m3_).

Itisshouldbestressedthattheaveragemassconcentrationof allthereportedelements(exceptS,ClandK)intheﬁnefraction duringmascletàwasabout17_␮g/m3 _and_reached_values_as_high

as34␮g/m3_,_to_be_compared_to_a_value_of_0.03_␮g/m3_during_the

backgroundperiod.InFig.3c,thetemporalevolutionofCa con-centrationsinbothfractions,fineandcoarse,isshown.Ahigher increaseinthecoarsefractionthaninthefineoneduringtheevents isevident.Thistemporalpatternisverysimilarformostofthe ele-mentsofcrustaloriginlikeFe,Si,Ti,etc.Thefine/coarseratios(see Table1)forSiwere0.8and0.9respectivelyformascletàand back-groundperiod,forCa0.2and0.3,forFe0.4andforTi0.4(Tiwas belowmdlinnormaldaysinthefinefraction),whileformostof theotherelementsincreasesofatleastafactor2(uptomorethan 10)inthemascletàrespecttothebackgroundperiod.Insharp con-trasttowhatseenfortheelementsreportedinFig.3aandb,the ratiobetweentheconcentrationofthese4crustalelementsduring themascletàandduringthebackgroundperiodwasnearly con-stant(forTithedifferenceisduetotwodayswhentherewassome contributionbyfireworks,seebelow).Finally,theenrichment fac-tors(EF),calculatedusingSiasreferenceelementandthecrustal compositiongivenbyMasonandMoore[35],arecloseto1,both duringthemascletàandthebackgroundperiod,exceptCawitha

(7)

Fig.4.Mass-sizedistributionforMascletàand“Normalsummerday”fortypicalgunpowderbulkelements(K,S),forapyrotechnicoriginelement(Cu)andforacrustal originelement(Ca).Thehistogramgivestherawsizedistribution,theinverteddistributionisindicatedbythesquaresymbols,andthesmoothdashedlinesrepresentthe lognormalﬁts.

higherEF,butitisknownthatAlicantesoilisCaenriched[36,37]. Forexample,EFwere:Al1.2,Ca21,Ti2.1,Fe2.8Sr13.2during normalperiod;duringmascletàweobtainAl28,Ca31,Ti5.9,Fe 2.4,Sr126.Soitisevidentthedifferencebetweenelementslike Fe,CaorTi(withtheproblemdiscussedabove)andotherelements producedalsobyﬁreworkslikeAlorSr.Therefore,the concentra-tionincreaseseemsnottobeduetothecombustionproductsbutto theresuspensionofcrustaloriginmaterialdepositedontheground

wheretheexplosionstakeplace.Aconﬁrmofthishypothesiscould comefromfurtherstudiesontheofU/Raratioinisolatedparticles byAMSasreportedbySteinhauserandMusilek[38].

Anestimateoftheresuspendedsoildustconcentrationduring mascletàscanbecalculatedconsideringthecrustalelementsas oxides[39].SincewehavediscussedabovethatAl,KandSrare producedbyﬁreworks(indeedtheyshowedhighenrichment fac-tors),onlythenaturalfractionofAl,KandSr(theirconcentration

(8)

J.Crespoetal./JournalofHazardousMaterials241–242 (2012) 82–91 89 dividedbytheirenrichmentfactor)wasincludedinthecrustal

mat-tercalculation.Thesumoftheﬁneandcoarseresuspendedsoil dustconcentrationvariedfrom44␮g/m3_to₁₃₀_␮g/m3_,_while_the

backgroundconcentrationisnormallylessthan10␮g/m3_.

3.2. Elementalmass-sizedistribution

Inordertomeasuretheaerosolelementalmass-size distribu-tionsduringmascletàandtocomparethemwiththoseona“normal summerday”(NSD),12-stagesDekaticascadeimpactorwasused. ThefollowingPIXEanalysisprovidedrawconcentration-sizedata (massperstageandperm3₎_for_the_elements_Mg,_AI,_Si,_S,_Cl,_K,_Ca,

Ti,V,Cr,Mn,Fe,Co,Cu,Zn,As,Br,Sr,BaandPb.Fromtheobtained resultsweobservesigniﬁcantchangesbothinconcentrationsas wellasinmodaldistributionsofinvestigatedelements.

Toobtainmorequantitativeresults, therawsizedistribution datawereconvertedintosmoothsizedistributionsbymeansof theinversioncodeMICRON[40]torevealthemodalstructureof theaerosolmoreaccuratelythanwouldbethecaseifusingthe discretedistributionsonly.Lognormalcurveswerethenﬁttedto theinvertedsmoothdistributions,sothattheelementalmass con-centration,geometricmeanaerodynamicdiameter(GMAD)and geometricstandarddeviation(GSD)ofthedifferentcontributing modeswereobtained.FortheMascletàsamples,theseparameters fortheﬁreworkrelatedelementsarereportedinTable2.

Fig.4showstherawdata,theMICRONinversionsandthe log-normalﬁtsforK,S,CuandCabothinthemascletàandintheNSD samples.ItcanbeclearlyseenfromFig.4thatmodemass concen-trationsincreasedingeneralthroughoutmascletàs,dependingon elementandmode.ForK,S,andCudifferenttypesofdistribution proﬁles,onmascletàsandonNSD,werealsoobtained.

Onmascletàthedistributionfortypicalgunpowderbulk ele-ments, K, S (Fig. 4), is trimodal, with a coarse mode (with GMAD2.4–3␮m,GSD1.3–1.5␮mandRMC<10%),andtwo sep-aratesubmicrometer-sizedsubmodes,onewithGMAD0.19_␮m, GSD∼1.5␮m and RMC∼40% and the intermediate one with GMAD∼0.65␮m, GSD∼1.6␮m and RMC∼50%. Therefore, all theseelementspresentedhigherconcentrationsinthe submicrom-etersizebins.Correlationcoefﬁcientsbetweendifferentpairsof these elementsfor each diameter bin have been calculated to quantifythe correspondencesamong theirelemental mass-size distributions.ThedeterminationcoefﬁcientsR2_between_K_and_Cl

andbetweenKandSduringmascletàswere0.98and0.99, respec-tively.Agoodcorrelationwasobservedforallthesizebins,with moreClassociatedtoKinthesubmicrometerfractionandmoreS relatedtoKinthecoarseone.

ForNSD,completelydifferentmass-sizedistributionpatternfor Kwasfound.Itshowedrelativehighconcentrationsinthecoarse modebutwithfarlowerintensitywithrespecttomascletàevent. Sulphurpresented adifferentbehaviour,ascanbealsoseen in Fig.4.ThedistributionproﬁlesbothonNSDandduringmascletàs weretrimodal,withanequivalentGMADinallcases.Asimilar con-centrationincreaseofabouttwoordersofmagnitudeisproduced approximatelyinallsizesbins.

Mostoftheotherﬁreworkrelatedelements(Mg,Cr,Cu(Fig.4), Zn,AsSr,Pb,Rb,Ba),arecharacterisedbyatrimodalsize distri-bution,withtheaccumulationmode(GMAD∼0.2–0.3␮m)asthe dominantonewithRMCbetween50%and100%.Forsomeofthose elements(Mg,Cu(Fig.4),Zn,Pb,andBa)alsotheintermediatemode (GMAD∼0.6–0.8␮m)isimportant(RMC30–50%).Infact,Ba,Sr,Cu andPb,speciﬁcindicatorsofpyrotechnicevents,raisedtheir accu-mulationmodeconcentrationsfrom<10ng/m3_up_to₁_␮g/m3_(Pb,

Cu)and10␮g/m3_(Ba,_Sr)._It_has_to_be_noted_the_sharp_difference_of

thepatternofNSDSr,whoselowconcentrationlevelsarelimited exclusivelytothecoarsemode. Otherelementsshowedsimilar behaviourbutwitharelativelowerconcentrationingeneraland

particularlyinthecoarsemode.AnthropogeniccompoundslikeV, Zn,Ni,Asarefoundinthisgroupwithlevelsabovethatofnormal anthropogenicsourcesapproximatelyintwoordersofmagnitude. Mgshowsasimilarpattern,butwitharelativelower concentra-tioningeneralandparticularlyinthecoarsemode.Anexceptionis constitutedbyCo,whosemassisconcentratedintwomodeswith GSD∼1and3.2␮m.

InFig.4themassdistributionforatypicalcrustalelement,Ca, is shown.During mascletàscrustalelements(Ca, Mg, Si,Mn, Ti andFe)showtwomainmodes:withGMADof2–3␮m(containing morethan90%oftheelementalmass)andwithGMADofabout 0.14–0.3␮m (containing the remaining elemental mass, lower than10%).TypicalGSDsofthemodeswere∼2and1.3–1.8␮m, respectively.Theirsizedistributionpatternsremainquitesimilar onNSD,butwitharemarkableconcentrationdecrease.Thisfact, togetherwiththestrongcorrelationbetweenthetypicalcrustal elementsdistributions(i.e.,Si,Ca,Ti,andFeresultedincorrelation coefﬁcientsabove0.9)andtheenrichmentfactors[35]forallthe sizebinscloseto1(similartothevalues derivedforthewhole coarsesizefractionin thestreakersamples) hint toa common mainsourcefortheseelements.Theincreaseintheconcentrationis probablyduetoresuspensionofsoilmaterial,consequenceof mas-sivegroundexplosionsthroughoutthepyrotechnicevent,already notedfromthestreakerdata.

SomedifferencesforsomecrustalelementslikeAl,Mgwere observed.Wehaveseenintheprevioussectionthattheyalso orig-inatedfromﬁreworks;togetherwithacoarsemodesimilartothe onepresentin NSD,a strongaccumulationmode(maximumat about0.3␮m)waspresent,duetothepyrotechnicevent.

4. Conclusions

Thispapershowsanexperimentalproceduretostudythetime evolutionandsizedistributionofaerosolelementsemittedintothe atmospherebyhigh-intensitypyrotechnicevents.Theprocedure wascarriedoutduringtheMascletàs,dailyeventstypicaloffestivals ineasternSpain.Thecouplingof1-hresolutionsamplingfollowed byPIXEanalysisofmascletàcollectedaerosolshaveproventobe apowerfulmethodologytomonitorthetemporaltrendofaerosol elementalconcentration,bothintheﬁneandthecoarsefraction. Mascletàscombustiongeneratesheavyaerosolcloudswhich con-tainacomplexmixtureofdifferentmetalsandothercomponents. Extremelyhighconcentrationsintheﬁnefractionwithvaluesof about384␮g/m3_for_K_and₂₁₆_␮g/m3_for_Cl_were_reached_during

mascletàs.Increase factors ofmore than 1500relating to back-groundlevelswerefoundfortheseﬁreworkspeciﬁcelements.

OtherelementsrelatedtopyrotechnicdisplayslikeAl,Mg,Cu, Co,Zn, andPbalsopresenteda largerise,withincreasefactors above100,mainlyinthefinefractionincomparisontotheirnormal values.IntheparticularcaseofSrandBa,factorsupto∼1000were observed.Theconcentrationsfoundingeneratedmascletàplumes werenotconstantduetovariabilityofgunpowdercompositionand pyrotechnicdevicesusedtoattainspecificfireworkeffects.

Afewelementswerepresentonlyinsomemascletàevents:Cr, As,PbandCo.Itistobenotedthatthesemetalsaredangerous elementsbecauseoftheirtoxicity,especiallyAsandPb [41,42], andtheiruseinpyrotechnicdevicesareforbiddenbylawinthe EuropeanUnion.Nevertheless,asstatedinthebibliography[3], pyrotechnicmaterialscanbeimportedfromcountrieswithmore permissivelegislationresultinginsporadicatmospherictoxic dis-chargeasweobservedinourstudy.Inourcase,insomemascletàs, AsandPbreachedconcentrationsofupto0.17and5␮g/m3

respec-tively,twoandthreeordersofmagnitudeabovebackgroundlevels. Suchvery highconcentrations inashorttime ina placewhere thousandsofpeoplearegatheredmaybeofconcern,particularly

(9)

forpeoplewhosufferfromchronicrespiratoryhealthproblems, cardiovasculardisease,etc.

Ambientconcentrationincreasewasalsoobservedformostof theelementsof crustal origin,likeCa, Fe,Si, Ti.This isdue to theresuspension,producedbypyrotechnicdeviceexplosions,of materialalreadydepositedontheground.Suchanimportant con-tributionoftheresuspendedsoildust(∼10timesabovebackground levels)hastobetakenintoaccountasasourceofparticulatematter inthesetypesofevents.

We also compared the aerosol elemental size distribution relatedtomascletàeventswiththecorrespondingelementalsize distributionofa normal summerday(NSD).While asexpected thecrustalcomponentsaremainlyinthecoarsemodewithtwo mainmodes,allthefireworkrelatedelements(exceptCo)aremore concentratedinthesubmicrometricregion(>80%elementalmass) withatrimodalsizedistributionformascletàevents.This infor-mationmaybeusefultoepidemiologistsgiventheharmfuleffects ofmetallicparticlesonhumanhealth[43–46].Withregardstoa NSD,fireworkassociatedelementsweregenerallyevidentin rel-ativehighconcentrationsinthecoarsemodebutwithfarlower intensity.Thesizedistributionpatternsremainedquitesimilarfor crustalelementsbutwitharemarkableconcentrationdecreasethat confirmsmascletàresuspensionprocesses.

Apreliminaryapproach toquantifytheelementalmassdose receivedbyindividualsduetosmokeplumeduringmascletàswas attempted.ThecalculationwasmadeforPb.Thiselementwas cho-senasitisoneofthemetalswhichappearingreatestconcentration duringmascletàs(despiteitsprohibitioninpyrotechnicuse). More-over,giventhatitisoneofthemetalswhoseharmfuleffectshave beenmoststudiedduringrecentdecades[43,44],theresultofthe calculationcanbecomparedwithbibliographicdata.

Dosecanbecalculatedas:DPb=CPbmas×IR×ED×DF

where CPbmas=mascletà ﬁne fraction Pb concentration (␮gPbairm−3); IR=human inhalation rate (airm3_min−1_);

ED=exposure duration (minmascletà−1); DF=deposition factor (asafractionofPbmassinhaled)[47].

Since the mascletà episode lasted approximately 8min, we assumed that CPbmas, at least (as a minimum), was equal to ﬁnefractionPb1-hconcentration,i.e.∼1.3␮gPbairm−3(Table1). Althoughnaturalbreathingpatternshavevariableinhalationrates, weusedtheaveragebreathingpatterninadultsatrest,typically 8.5×10−3airm3_min−1_[47]._Moreover,_there_were_several_factors,

likethelengthoftimepeoplespentatthemascletàsiteorthe dis-tancetolaunchingsite,thatcausedvariationinEDofspectators. Itcouldbeestimatedminimum8minmascletà−1.Inaddition,as statedbyKinandHu[47],inhealthyadultsDFvariedfrom0.12 to0.25andfrom0.26to0.68foraerosolparticleswith1␮mand 3␮mindiameter,respectively,dependingonthebreathing pat-ternapplied.Weusedthelowest1␮mDFvalue(0.12)toobtain a minimumPb dose, DPb∼= 10ngPbindividual−1mascletá−1.It is

importanttostressthat thisresultis onlyanestimation ofthe actualdoseintermsofmassofPbintheaerosolﬁnefraction(1␮m) depositedintheairwayperpersonandmascletà.Calculationscan beappliedtoothermetallicelementsanalysedin the pyrotech-nicaerosol.Additionalharmfuleffectswouldbeexpectedifoverall metallicelementaldoseisconsidered.Thenumberofpeoplethat aretypicallyexposedtothesehighmetallicPMconcentrationlevels duringthemascletàscanbeestimatedbytakingintoaccountthe characteristicsofthesquarefromwheretheﬁreworksarelaunched andthepositionofthepeoplearoundit.Thedirectionofthebreeze waslargelyfromtheSEsector(120–140◦).Undertheseconditions, thepyrotechnicplumeimpactsononlyathirdofthetotalsurfaceof thesquare(∼0.3×10−2km2_)._If_we_assume_that_there_are₃

spec-tatorspersquaremetre,anaverageof9000spectatorswouldbe

directlyaffectedbytheimpactoftheaerosolplumeduringeach mascletà.Thiswouldrepresent2–3%oftheentireAlicante popula-tion,althoughitshouldbenotedthattensofthousandsoftourists visitthecityduringLasHoguerasdeSanJuan.

Acknowledgements

WethanktheAirQualitySurveillanceNetworkoftheValencian CommunityRegionalGovernmentforsupplyingdata,andP. Nord-stromandG.Escribanofortheirassistanceinthiswork.Weare deeplyindebtedtoDr.WillyMaenhautforhiskindcollaboration fortheuseofthecodeMICRON.

This study was partially funded by the Spanish Ministry ofEnvironment undertheresearchprogramConsolider-Ingenio 2010: GRACCIE Project, CSD2007-00067, by the Spanish Min-istryof ScienceandInnovation,I+D+iProgram: PASSEProject, CGL2009-08036,bytheItalian MinistryofEducation,University andResearch,underthePRIN2008grant,andbyINFNunderthe NUMENproject.

References

[1] F.Drewnick,S.S.Hings,J.Curtius,G.Eerdekens,J.Williams,Measurementof ﬁneparticulateandgas-phasespeciesduringtheNewYear’sﬁreworks2005 inMainz,Germany,Atmos.Environ.40(2006)4316–4327.

[2]D.Y.Liu, D. Rutherford, M.Kinsey,K.A.Prather, Real-timemonitoringof pyrotechnicallyderivedaerosolparticlesinthetroposphere,Anal.Chem.69 (1997)1808–1814.

[3]T.Moreno,X.Querol,A.Alastuey,M.C.Minguillon,J.Pey,S.Rodriguez,J.V. Miro,C.Felis,W.Gibbons,Recreationalatmosphericpollutionepisodes: inhal-ablemetalliferousparticlesfromﬁreworkdisplays,Atmos.Environ.41(2007) 913–922.

[4]R.Vecchi,V.Bernardoni,D.Cricchio,A.D’Alessandro,P.Fermo,F.Lucarelli,S. Nava,A.Piazzalunga,G.Valli,Theimpactofﬁreworksonairborneparticles, Atmos.Environ.42(2008)1121–1132.

[5] Y.Wang,G.Zhuang,C.Xu,Z.An,Theairpollutioncausedbytheburning ofﬁreworksduringthelanternfestivalinBeijing,Atmos.Environ.41(2007) 417–431.

[6] U.C.Kulshrestha,T.N.Rao,S.Azhaguvel,M.J.Kulshrestha,Emissionsand accu-mulationofmetalsintheatmosphereduetocrackersandsparklesduring DiwalifestivalinIndia,Atmos.Environ.38(2004)4421–4425.

[7]S.C.Barman,R.Singh,M.P.S.Negi,S.K.Bhargava,AmbientairqualityofLucknow city(India)duringuseofﬁreworksonDiwalifestival,Environ.Monit.Assess. 137(2008)495–504.

[8]K.Ravindra,S.Mor,C.P.Kaushik,Short-termvariationinairqualityassociated withﬁreworkevents:acasestudy,J.Environ.Monit.5(2003)260–264. [9] X.Wang,X.Bi,G.Sheng,J.Fu,HospitalindoorPM10/PM2.5andassociatedtrace

elementsinHuangzhou,China,Sci.TotalEnviron.366(2006)124–135. [10]T.Moreno,X.Querol,A.Alastuey,F.Amato,J.Pey,M.Pandolﬁ,N.Kuenzli,L.

Bouso,M.Rivera,W.Gibbons,Effectofﬁreworkseventsonurbanbackground tracemetalaerosolconcentrations:isthecocktailworththeshow?J.Hazard. Mater.183(2010)945–949.

[11]D.D.Dutcher,K.D.Perry,T.A.Cahill,S.A.Copeland,Effectsofindoorpyrotechnic displaysontheairqualityinthehoustonastrodome,J.AirWasteManage.Assoc. 49(1999)156–160.

[12]K.Hirai,Y.Yamazaki,K.Okada,S.Furuta,K.Kubo,Acuteeosinophilic pneumo-niaassociatedwithsmokefromﬁreworks,InternalMed.39(2000)401–403. [13]R.MerrillSmith,V.D.Dinh,Changesinforcedexpiratoryﬂowduetoair

pollu-tionfromﬁreworks.Preliminaryreport,Environ.Res.9(1975)321–331. [14]W.Bach,L.Dickinson,B.Weiner,G.Costello,Someadversehealtheffectsdue

toairpollutionfromﬁreworks,HawaiiMed.J.31(1972)459–465.

[15]K.J.Godri,D.C.Green,G.W.Fuller,M.Dall’Osto,D.C.Beddows,F.J.Kelly,R.M. Harrison,I.S.Mudway,Particulateoxidativeburdenassociatedwithﬁrework activity,Environ.Sci.Technol.44(21)(2010)8295–8301.

[16]J.M.Becker,S.Iskandrian,J.Conkling,Fatalandnearfatalasthmainchildren exposedtoﬁreworks,Ann.Allergy.Asthma.Immunol.85(2000)512–513. [17]R.Camilleri,A.J.Vella,EffectofﬁreworksonambientairqualityinMalta,Atmos.

Environ.44(2010)4521–4527.

[18]S.Sarkar,P.S.Khillare,D.S.Jyethi,A.Hasan,M.Parween,Chemicalspeciation ofrespirablesuspendedparticulatematterduringamajorﬁreworkfestivalin India,J.Hazard.Mater.184(2010)321–330.

[19]A.Joly,A.Smargiassi,T.Kosatsky,M.Fournier,E.Dabek-Zlotorzynska,V.Celo, D.Mathieu,R.Servranckx,R.D’amours,A.Malo,J.Brook,Characterization ofparticulateexposureduringﬁreworksdisplays,Atmos.Environ.44(2010) 4325–4329.

[20] G.Croteau,R.Dills,M.Beaudreau,M.Davis,Emissionfactorsandexposures fromground-levelpyrotechnics,Atmos.Environ.44(2010)3295–3303. [21]P.I.X.E.InternationalCorporationWebsite:http://pixeintl.com/Streaker.asp

(10)

J.Crespoetal./JournalofHazardousMaterials241–242 (2012) 82–91 91 [22]A.D’Alessandro,F.Lucarelli,P.A.Mandò,G.Marcazzan,S.Nava,P.Prati,G.Valli,

R.Vecchi,A.Zucchiatti,Hourlyelementalcompositionandsources identiﬁca-tionoftheﬁneandcoarsePM10particulatematterinfourItaliantowns,J. AerosolSci.34(2003)243–259.

[23]M.Chiari,P.DelCarmine,I.GarciaOrellana,F.Lucarelli,S.Nava,L.Paperetti, Hourlyelementalcompositionandsourceidentiﬁcationofﬁneandcoarse PM10inanItalianurbanareastressedbymanyindustrialactivities,Nucl. Instrum.MethodsPhys.Res.Sect.B249(2006)584–587.

[24]G.Calzolai,M.Chiari,I.García-Orellana,F.Lucarelli,A.Migliori,S.Nava,F. Tac-cetti,ThenewexternalbeamfacilityforenvironmentalstudiesattheTandetron acceleratorofLABEC,Nucl.Instrum.MethodsPhys.Res.Sect.B249(2006) 928–931.

[25]G.Calzolai,M.Chiari,F.Lucarelli,S.Nava,S.Portarena,Protoninduced␥-ray emissionyieldsfortheanalysisoflightelementsinaerosolsamplesinan externalbeamset-up,Nucl.Instrum.MethodsPhys.Res.Sect.B268(2010) 1540–1545.

[26]J.L.Campbell,N.I.Boyd,N.Grassi,P.Bonnick,J.A.Maxwell,TheGuelphPIXE softwarepackageIV,Nucl.Instrum.MethodsPhys.Res.Sect.B268(2010) 3356–3363.

[27] Micromatter(AAPS),http://www.micromatter.com/xrf.php

[28] V.Bernardoni,E.Cuccia,G.Calzolai,M.Chiari,F.Lucarelli,D.Massabò,S.Nava, P.Prati,G.Valli,R.Vecchi,ED-XRFset-upforsize-segregatedaerosolsamples analysis,X-RaySpectrom.40(2011)79–87.

[29]I.Salma,R.Ocskay,N.Raes,W.Maenhaut,Finestructureofmasssize distribu-tionsinanurbanenvironment,Atmos.Environ.39(2005)5363–5374. [30]W.Maenhaut,R.Hillamo,T.Makela,J.L.Jaffrezo,M.H.Bergin,C.I.Davidson,

AnewcascadeimpactorforaerosolsamplingwithsubsequentPIXEanalysis, Nucl.Instrum.MethodsPhys.Res.Sect.B109/110(1996)482–487. [31]G. Steinhauser, T. Klapötke,“Green” pyrotechnics:a chemists’ challenge,

Angew.Chem.Int.Ed.47(2008)3330–3347.

[32]M.S.Russell,TheChemistryofFireworks,TheRoyalSocietyofChemistry, Cambridge,2000.

[33]R.Wilkin,D.Fine,N.Burnett,Perchloratebehaviorinamunicipallakefollowing ﬁreworksdisplays,Environ.Sci.Technol.41(2007)3966–3971.

[34]R.Lancaster,R.Butler,J.Lancaster,T.Shimizu,T.Smith,Fireworks:Principles andPractice,ChemicalPublishingCompany,NewYork,1998.

[35]B.Mason,C.B.Moore,PrinciplesofGeochemistry,fourthed.,JohnWiley,New York,1982.

[36]N.Galindo,E.Yubero,J.Nicolás,J.Crespo,C.Pastor,A.Carratalá,M. Santa-catalina,Water-solubleionsmeasuredinﬁneparticulatematternexttocement works,Atmos.Environ.45(2011)2043–2049.

[37]J.Nicolás,M.Chiari,J.Crespo,N.Galindo,F.Lucarelli,S.Nava,E.Yubero, Assess-mentofpotentialsourceregionsofPM2.5componentsatasouthwestern Mediterraneansite,Tellus63B(2011)96–106.

[38]G.Steinhauser,A.Musilek,Dopyrotechnicscontainradium?Environ.Res.Lett. 4(2009)034006.

[39]F.Lucarelli,S.Nava,G.Calzolai,M.Chiari,R.Udisti,F.Marino,IsPIXEstillauseful techniquefortheanalysisofatmosphericaerosols?TheLABECexperience, X-RaySpectrom.40(2011)162–167.

[40] J.K.Wolfenbarger,J.H.Seinfeld,Inversionofaerosolsizedistributiondata,J. AerosolSci.21(1990)227–247.

[41] C.Abernathy,Y.Liu,D.Longfellow,H.Aposhian,B.Beck,B.Fowler,R.Goyer,R. Menzer,T.Rossman,C.Thompson,M.Waalkes,Arsenic:healtheffects, mech-anismsofactions,andresearchissues,Environ.HealthPerspect.107(1999) 593–597.

[42] D.Juberg,C.Kleiman,S.Kwon,PositionpaperoftheAmericanCouncilon Sci-enceandHealth:LeadandHumanHealth,Ecotoxicol.Environ.Saf.38(1997) 162–180.

[43]M.Kampa,E.Castanas,Humanhealtheffectsofairpollution,Environ.Pollut. 151(2008)362–367.

[44]L. Järup,Hazardsofheavy metalcontamination,Br.Med.Bull.68(2003) 167–182.

[45]F.Schaumann,P.J.A.Borm,A.Herbrich,J.Knoch,M.Pitz,R.P.F.Schins,B.Luettig, J.M.Hohlfeld,J.Heinrich,N.Krug,Metal-richambientparticles(PM2.5)cause airwayinﬂammationinhealthysubjects,Am.J.Respir.Crit.CareMed.170 (2004)898–903.

[46]A. Rohr, R. Wyzga, Attributing health effects to individual particu-latematterconstituents,Atmos.Environ.(2012),http://dx.doi.org/10.1016/ j.atmosenv.2012.07.036.

[47]C.Kim,S.Hu,Totalrespiratorytractdepositionofﬁnemicrometer-sized par-ticlesinhealthyadults:empiricalequationsforsexandbreathingpattern,J. Appl.Physiol.101(2006)401–412.