Spectroscopic investigation of Roman decorated plasters by combining FT-IR, micro-Raman and UV-Raman analyses

Spectroscopic

investigation

of

Roman

decorated

plasters

by

combining

FT-IR,

micro-Raman

and

UV-Raman

analyses

V.

Crupi

a,

*,

V.

Allodi

b

,

C.

Bottari

a

,

F.

D

_’Amico

c

,

G.

Galli

d

,

A.

Gessini

c

,

M.F.La

Russa

e

,

F.

Longo

a

,

D.

Majolino

a

,

G.

Mariotto

b

,

C.

Masciovecchio

c

,

A.

Pezzino

f

,

B.

Rossi

c

,

S.A.

Ruffolo

e

,

V.

Venuti

a

a

DipartimentodiScienzeMatematicheeInformatiche,ScienzeFisicheeScienzedellaTerra,UniversitàdegliStudidiMessina,VialeFerdinandoStagno d’Alcontres31,98166Messina,Italy

b_{DipartimentodiInformatica,UniversitàdegliStudidiVerona,StradaleGrazie15,37134Verona,Italy} c_{Elettra—SincrotroneTrieste,StradaStatale14km163.5,AreaScience70Park,34149Trieste,Italy} d

SoprintendenzaSpecialeperiBeniArcheologicidiRoma,VilladeiQuintili,viaAppiaNuova1092,00197Roma,Italy

e

DipartimentodiBiologia,EcologiaeScienzedellaTerra(DiBEST),UniversitàdegliStudidellaCalabria,ViaPietroBucci,87036ArcavacatadiRende,Cs,Italy

f

DipartimentodiScienzeBiologiche,GeologicheeAmbientali—SezionediScienzedellaTerra,UniversitàdegliStudidiCatania,CorsoItalia57,95129Catania, Italy

ARTICLE INFO

Articlehistory: Received5October2015

Receivedinrevisedform22January2016 Accepted26January2016

Availableonline29January2016

Keywords: Archaeometry

Romandecoratedplasters Pigments

IRabsorbance Ramanmapping

ABSTRACT

Inthiswork,thecomplementaryuseofFouriertransforminfrared(FT-IR)spectroscopy,conventional micro-RamanspectroscopyandUV-Ramanscatteringprovedsuccessfulforthecharacterizationofbulk minerals and of a varietyof pigments fromdecorated finishing layers of plasters from a Roman archaeologicalsiteknownasVilladeiQuintili,amonumentalresidencelocatedinthesouth-easternpart ofRome(Italy).

Theusedmulti-techniqueapproachprovidedinsightsonthepictorialtechnique,givinginformation thatcouldbeusefulforproperrestoration.

ItisworthunderliningthatthepresentstudyrepresentsthefirstattemptofcarryingoutUVresonance Ramanmeasurementsforanalysingculturalheritagematerials.

ã2016ElsevierB.V.Allrightsreserved.

1.Introduction

Theknowledgeofbulkandpigmentminerals,their composi-tionandmethodsofutilizationisveryimportantfor understand-ingthehistoryofa workof artandtheresolution ofproblems relatedtoconservation,restoration,datingandauthorattribution

[1,2].Aswellknown,whendealingwitharchaeologicalmaterials, eventhefragmentssampledfromtheartwork,whateveritis,can bepreciousandthepreservationoftheirintegrityisoftenrequired. Hence,theuseofnon-destructive orat mostmicro-destructive analyticaltechniquessuchasthevibrationalspectroscopies,have been extensively used for the identification of archaeological source materials [3,4], especially ancient pigments [5–8]. In particular,thecombinedapplicationofFouriertransforminfrared (FT-IR),Ramanandmicro-Ramanspectroscopies[9–11]hasshown toyieldvaluablecomplementary[9]informationfortheanalysisof pigmentsonwallpaintings.Aswellknown,infact,althoughboth

FT-IRandRamanspectroscopyprovideinformationabout charac-teristicvibrationallevelsofmolecularsystems[12–15], neverthe-less,accordingtoselectionrules,sometransitionsmayturnout IR-allowedandRaman-forbidden,orviceversa.Furthermore,someof themaredetectedinbothIRandRamanprofilesatcoincidental frequenciesand,otherscanbeobservedinneitherofthesetwo kindsofspectra.Again,ifthemoleculehasalowsymmetry,most ofthevibrationalmodesappearinbothitsIRandRamanspectra, but,usually,withverydifferentintensities.Itoccursbecausethe polarizabilitychange(whichprovidestherequiredselectionrule forRamanscattering)associatedtoaspecificvibrationdiffersfrom thedipolemomentchange(i.e.:therequiredselectionruleforIR absorption)associatedtothesamevibration[16,17].The simulta-neous use of thetwo techniques, then, will behelpful for the completeidentificationofthevariouspossiblematerials character-ing the pigments, such as the minerals, the oxides and the sulphides,asexamples.Finally,theincorporationofmicro-Raman spectroscopyintothestudyallowsustoachieve moredetailed results withhigh spatial resolution by means of a rapid, non-destructivespectroscopictool.

*Correspondingauthor.

E-mailaddress:vcrupi@unime.it(V.Crupi).

http://dx.doi.org/10.1016/j.vibspec.2016.01.009 0924-2031/ã2016ElsevierB.V.Allrightsreserved.

ContentslistsavailableatScienceDirect

Vibrational

Spectroscopy

In the present work, we carried out an archaeometric characterization,atmolecularlevel,byusingFT-IRspectroscopy, micro-Raman and UV-Raman spectroscopy, of the bulk and of differentpaintedsurfacesofplasters,sampledfromvariousareas ofanimportantRomanmonumentalcomplex,knownasVilladei Quintili(Rome,Italy),datedbacktothefirsthalfofthe2ndcentury CE [18].In the frameof a wider researchproject aimedat the reconstructionofthisattractivearchaeologicalsite,uptonowlittle explored,thestudyofthesearchaeologicalfindingsthroughthe investigationofanychangesinthecompositionofthematerialsis assessed in order to understand variations in the artistic methodology, styles, or production processes. The obtainable informationcanstronglycontribute,ontheonehandtothestudies ofthearchaeologistsforthefullknowledgeofthehistoryofthe site,andontheotherhandtorestorationaims.

Thin-sectionsofsampleshavebeenalreadyclassifiedfromthe petrographicpointofviewintothreedifferentgroupsbasedonthe type/sorting of the aggregate and on the main features of the cluster[19]. Preliminary measurementsperformed byscanning electronmicroscopycoupledwithenergy-dispersiveX-ray spec-trometry(SEM-EDS)revealedthepresenceoftwolayers,identified as the preparation layer and the paintedlayer. Moreover, two portableinstruments(XRFandRaman)havebeenalreadyusedin ordertoachieveinformation,innon-invasiveway,onthesurfaceof some of these samples [20]. In particular, a handheld energy-dispersiveX-ray fluorescence(hXRF) analyserwas employedin order to obtain, starting from the elemental composition, an indirectidentificationofpigmentsthroughthedetectionoftheir key elements. For example, in the case of some plasters with reddishlayer,thesimultaneouspresenceofHgandSallowedusto hypothesizetheuseofcinnabar(HgS).However,XRFisinsensitive tothechemicalstateand/orthemolecularenvironmentinwhich the elements are present. Consequently, it cannot provide informationofthespecifickindofpigmentpresentinthesample. For this reason, the analysis was integrated with conventional Ramanmeasurements performedwitha portable spectrometer (pRaman)workinginthevisiblelightregion.Thecollecteddata evidenced the typical peaks of cinnabar for those samples for whichXRFrevealedthesimultaneouspresenceofHgandS[20]. Here,weperformedtheanalysisofpigmentswiththeuseofa micro-Ramanset-upworkinginthevisiblelightrange,inorderto obtainanincreasedsignal-to-noiseratioandareductioninofthe sampleareasdowntoafew

m

m2_.

Finally,wehavecarriedout,forthefirsttimeinthecultural heritage filed,complementary resonanceRaman measurements employingtheUVradiationavailableattheIUVSbeamlineatthe Elettra facility [21]. The use of UV radiation instead of the conventionalvisibleonnear-IRonestakesseveraladvantages:(a) itincreasesoforderofmagnitudetheRamansignalcomingfrom moleculargroups having

p

-

p

* electronic transitions in theUV range,e.g.suchasaromaticringsand/oroxygencomposites;(b)it allowstogetasignificantreductionoftheunwantedfluorescence background,thuspermittingamoreaccuratespectrallineshape analysis;(c)itallowstoanalysealsoexcitationmodesoccurringin the low-energy region, which is not easily accessible to a conventionalRamanset-up.

2.Experimental

2.1.Materialsandsampling

AdetaileddescriptionofthearchaeologicalsiteoftheVilladei QuintilihasbeenalreadyreportedinRef[19];werecallhereonly themainpoints.

The Roman archaeological site of Villa dei Quintili is a monumentalresidencelocatedinthesouth-easternpartofRome,

Italy(Fig.1(a)).Itwasbuiltinthefirsthalfofthesecondcenturyby thefamilyoftheQuintilibrothers,onthepromontoryformedby the_‘CapodiBove_’lava_flow,eruptedduringthevolcanicactivityof theAlbanHills[22].Themonumentalcomplex,whichbelongedto several emperors,consistsofmanystructuresexpandedaround variouscentresandcharacterizedbydifferentbuildingtechniques ascribable to nine construction phases ranging from 98–138CE (Trajan-Hadrian) up to Middle and Modern Ages [23]. The archaeological importanceof theresidenceis mainlyrelatedto the high quality of the building and decorative materials, particularlyfinemarblefromaroundtheworld.Theornamental elements of the floors generally consist of opus sectile and polychromemosaics,whereasmarble,frescoes,andmosaicswere usedforthewalls.

The samples under investigation are seven fragments of plasterscomingfromanextensivesamplingperformedonseveral edificesofthemonumentalcomplexofVilladeiQuintili(Fig.1(b)). Some images of the analysed archaeological fragments are reportedinFig.2.

Thesamplingwascarriedoutatfullrespectoftheintegrityof themonumentaccordingtotheItalianregulations[24].withthe

Fig.1.SatelliteviewofVilladeiQuintili(a)anddetailsofthesampledareas(b). Notes:A=representationarea;B=privateresidences;C=BasisVillae;D= Frigidar-ium; E=Calidarium;F=Ludus-Viridarium; L=Tepidarium; R=Hippodrome; T= Arcades.

assistanceofarchaeologistsfromtheArchaeological Superinten-denceof Rome, in order totakesamples representative of the differentconstructionphases.Thelistoftheanalysedsamplesis reported in Table 1, together with a brief description of their macroscopicfeatures,theindicationofthesamplingarea,ofthe previously performed analyses [20], and of the methods of investigationappliedinthiswork.

2.2.FT-IRspectroscopymeasurements

Fourier transforminfraredmeasurementswereperformedin the “micro-destructive” transmission configuration on all the

samplesbymeansofaBOMEMDA8FT-IRspectrometer,equipped withaGlobaraslightsource,aKBrbeamsplitter,andaDTGS/MIR detector.Spectrawerecollectedfrom500to4000cm 1_,usinga

resolutionof4cm 1_{.Thesampleswerepreparedinpellets,using}

small quantities (2mg) of sample dispersed in 200mg of powderedCsI,thatis,transparentintheinvestigatedIRfrequency range.32repetitive scanswereautomaticallyaddedtoobtaina good signal-to-noise ratio and spectral reproducibility. The experimental absorbanceprofileswerecompared withthoseof standard minerals and clays [25] and with data reported in differentsources[26,27]forareliableassignmentofthebands. 2.3.Micro-Ramanspectroscopymeasurements

Allmicro-Ramanmeasurementswerecarriedoutonsamples deposited onaglass slidein airand at roomtemperature.The correspondingspectrawereobtainedinbackscatteringgeometry andbyusingtwodifferentexperimentalsetupstobetterexplore differentspectralfeaturesofthesamples.

2.3.1.Ramanmappingmeasurements

Raman maps were acquired by using a software-controlled motorizedXY stagecoupledtoa Ramansystem (HORIBAJobin Yvon LabRam HR800) consisting of an 80cm focal length spectrograph using a 600 grooves/mm grating and a charge-coupleddevice(CCD)detectorcryogenicallycooled.TheRaman spectrain thewavenumber rangebetween200 and 1200cm 1

werecollectedbyusinganexcitingradiationat632.8nm(He-Ne laser,powerattheoutput20mW).Thelaserwasfocusedonto thesamplesurfacewithaspotareaofabout4

m

m2_throughthe

80objective(NA=0.9)ofthemicroprobe setup.Theelastically scattered radiation was filtered by using a notchfilter. In this configuration, the resolution was about 1cm 1_{/pixel. Raman}

mappingwasachievedbycollectingsets ofspectrasequentially recordedstep-by-stepfromsamplepoints,3

m

mawayfromeach other, within rectangular regions of interest. Repeated Raman mapswerecarriedoutfromthesamesampleinordertoassessthe analysisreproducibility.

2.3.2.UV-Ramanspectroscopymeasurements

UVMicroRamanscatteringmeasurementswerecarriedoutat theBL10.2-IUVSbeamlineattheElettraSynchrotronlaboratoryin Triestebyexploitingtheexperimentalset-updescribedinRef[20]. VerticallypolarizedVVRamanspectrawereexcitedat266nm.The inelastic scattered signal was collected in a back-scattering geometrybymeansof areflective40objective.Atriple stage spectrometer(Trivista,PrincetonInstruments)andaback-thinned CCD detectorwereusedtoget theRamanspectra in the200– 1200cm 1 _{wavenumberrange.Theexperimentalresolutionwas}

setto5cm 1_{inordertoensureenoughresolvingpowerand} count-Fig.2.RepresentativeimagesofI2(a)andMD1(b)fragments.

Table1

Listofinvestigatedfragmentsofplasters,togetherwiththeirmacroscopicfeatures,theindicationofprovenancearea,ofthepreviouslyperformedanalyses,andofthe methodsofinvestigationappliedhere.

Fragment Macroscopicfeatures Provenance

area

Previouslyperformed analyses

Presentperformedanalyses

I1A Reddishfinishinglayer R54 hXRF,pRaman IR

I1B Fragmentofplasterwithreddishfinishinglayer R54 hXRF IR,Micro-Raman I2 Fragmentofplasterwith3finishinglayers:reddish-brown,greenandlight

yellow

R6 hXRF IR,Micro-Raman, UV-Raman

I3 Fragmentofplasterwithtworeddish-brownfinishinglayers A5-A6 hXRF IR,Micro-Raman I4A Fragmentofplasterwithwhitishfinishinglayer BasisVillae hXRF,pRaman IR

I4B Fragmentofplasterwithreddishfinishinglayer BasisVillae hXRF,pRaman IR,UV-Raman MD1 Fragmentofplasterwithreddishfinishinglayer E2 – IR,Micro-Raman

rates.Nobaselinecorrectionwasperformedtoachievethefinal spectra.

3.Resultsanddiscussion

3.1.FT-IRspectroscopymeasurements

FT-IRanalysis,performedforallsamplesinthe500–4000cm 1

range, has been aimed at qualitative identification of the mineralogical composition of the bulk of the findings, and to the determination of the molecular nature of the pigmenting agents.

InFig.3 theFT-IRspectrumof thebulkofthesampleI1Bis reportedasanexample.

In spite of thecomplexity of thevibrational features in the investigatedrange,wehavebeenabletoassignspectratospecific materials.Inparticular,inFig.4,weplotted,asexamples,theIR spectra ofthefindings I2,I3,MD1,and I4Arelative todifferent sampling areas,amplifiedin the 500–2000cm 1 _{region, where}

mostofthevibrationalcontributionsofthedifferentmineralogical phasesfall,togetherwiththecorrespondingcentre-frequenciesof themainvibrationalbands.

These IR spectra revealed the most relevant differences between the bulk and the corresponding painted surface, so hereinafterwewillfocusthediscussiononlyonthesesamples.

Table2summarizestheobtainedresults.

AsfarassampleI2isconcerned(Fig.4(a)),theFT-IRprofileof thebulkisdominatedbytheabsorptionbandsofcalcite(CaCO3),

centred,inparticular,at714cm 1_,874cm 1_,1440cm 1_and

1797cm 1.Furthermore,abroadbandisclearlyevidentinthe 950–1268cm 1 _{rangethat canbeascribedtothepresence of}

feldspars.ConsideringthehighH2Ocontent,assuggestedbythe

observation,forallsamples,between3000and3800cm 1_of

theO Hstretchingvibration(seeFig.3),wecanhypothesizethat thepresentfeldspariskaolinite(Al2O3

2SiO2

2(H2O)),productofa

slow and complex hydrothermal alteration of feldspars, of feldspathoids and of other aluminiferous silicates, present as essential components in many rocks, mainly of granitic and gneissicotype.TheFT-IRprofileoftheredpigmentclearlyrevealed thepresenceofironoxidespigmentbasedcontributionscentredat 544 and 910cm 1_{. Furthermore, the large band between}

950 and 1268cm 1 _{includes,otherthan thecontribution of}

feldspar,alsothoseofironoxidecentredat1011and1052cm 1. ItisinterestingtonotehowverysimilartheFT-IRspectraofgreen

Fig.3.FT-IRspectrum,inthe4000–500cm 1

range,ofthebulkofI1Bsample.

Fig.4. FT-IRspectrum,inthe2000–500cm 1

range,oftheI2(a),I3(b),MD1(c),andI4A(d)samples,relativetothedifferentanalysedzones.Darkline:bulk,redline:red pigment,greenline:greenpigment,greyline:whitepigment.Themaincentre-frequenciesarealsoindicated.(Forinterpretationofthereferencestocolourinthisfigure legend,thereaderisreferredtothewebversionofthisarticle.)

andredpigmentsappear.Consequently,wecanguessthat,with goodprobability,thesameagentsacttocolourthesampleboth greenandred.

InthecaseofI3fragment(Fig.4(b)),theFT-IRspectrumofthe bulk presents, as for I2 sample, the same absorption bands of calcite.Theabsenceoffeldsparscanbeascribedtothedifferent samplingareaof thetwo findings. In thespectrum of thered pigmentwereveal,onceagain,thepresenceofthecharacteristic peaksofthebulkwith,inaddition,thoseofpigmentsbasedoniron oxides,centredat542,910,1011,and1052cm 1_(theselatter

two,inparticular,giverisetothebroadbandobservedbetween 950and1268cm 1_).

Calcite is always present in the FT-IR profile of the bulk of MD1sample (Fig.4(c)).Theobservation ofthespectruminthe wholeanalysedwavenumberrangesuggestsusthatthesampleis strongly hydrated, as appears evident from the intense O H stretching band between 3000 and 3800cm 1_{. The band}

between950and1268cm 1_{isinpartattributabletokaolinite}

thatcontributes,inparticular,withpeakscentredat1002and 1118cm 1_{.Inthespectrumoftheredpigmentitispossibleto}

recognize,similarlytothesampleI2,thecharacteristic contribu-tionsofthepigmentbasedonironoxide,at550and913cm 1. Thebroadbandbetween950and1268cm 1_{isdue,otherthan}

tofeldspars, also tothe presence of iron oxide, with peaks at 1011and1052cm 1_.

Finally,insampleI4A(Fig.4(d)),vibrationalbandsofcalcitecan be noticed for the bulk, and for the white pigment as well. Nevertheless, in the case of the white surface layer, the contributionofcalcite(peakat_876cm 1_{)appearsmuchmore}

intense with respect to the broad band observed between 950and1268cm 1_{duetofeldspars.Thisoccurrenceindicates}

calciteasmainconstituent ofthewhitepigment,in agreement withourpreviousportableRamanmeasurements.It isprobably producedbythecalciumoxide(CaO)alsousedforthepreparatory layer.

To conclude, the predominant presence of Ca in all the investigated samples can be attributed to the fact that IR absorptiondoesnotallowtodistinguishthesurfacelayerfrom theceramic bulkthat inevitablyremainedasaresultof micro-sampling.Nevertheless,itisalsotruethat,aswellknown,Roman frescoesweregenerallymadebyapplyingpigmentsondryplaster, whichreacted withtheatmospheric CO2 producinganexternal

layerofCaCO3.Amixturebasedonwaterandlimewasoftenused

forfixingpigmentsandgettingthedesiredshade.

3.2.Micro-Ramanspectroscopymeasurements

Micro-Raman analysis has been performed on thefinishing layerofthosefragmentsreportedinTable1,andwediscusshere those spectra for which fluorescence didnot mask meaningful information.Again,theresultsobtainedfortheanalysedfragments arereportedinTable2.

In Fig.5,someopticalmicroscopy images,taken ontheI1B fragment, as detected by Raman mapping, (a) and the related typicalspectra(b),areshown.

Several clearlycontrasted peaks are observed in the region between200and700cm 1_{,associatedtohematiteandmagnetite}

(redarrow)[28–30].thussuggestingthepresenceofredochre,an inorganic substance which was known and largely used since antiquity.Thespectracollectedontheseveralinvestigatedareas appear to be more similar to each other, therefore we can hypothesizethatmagnetiteiswidelydistributedinthepigment layer.

Itisworthnotingthat,althoughI1AandI1Barefragmentsof plasters both collected from the R54 area, different pigments (cinnabar[20]andredochre)arerevealed.Thislatterresultcould beindicativeofadifferentmanufacturingtechniquerelatedtothe occurrencethatthemonumentalcomplexofVilladeiQuintiliwas theresidenceofemperorsbelongingtodifferenttimes,ormore probablyrelatedtothevariousrestorationworksincurred.

Table2

ResultsofIR,micro-RamanandUV-Ramananalysesperformedontheinvestigatedfragmentsofplasters.

Performedanalyses Analyzedfragment Bulk Redlayer Greenlayer Whitelayer

IR I2 Calcite

Kaolinite

Ironoxides Ironoxides –

I3 Calcite Ironoxides – –

MD1 Calcite

Kaolinite

Ironoxides – –

I4A Calcite – – Calcite

Redlayer Greenlayer Whitelayer

Micro-Raman I1B – Hematite

Magnetite – – I2 – Hematite Magnetite Hematite Magnetite – Redlayer Greenlayer Whitelayer

UV-Raman I2 – Calcite Calcite –

I4B – Cinnabar

Calcite

– –

Fig.5.OpticalmicroscopyimageofthesampleregioninvestigatedbyRaman mapping(a)andarelatedtypicalspectrum(b)ofI1Bsample,fragmentofplaster withreddishfinishinglayer.Theredarrowindicatesthemaincontributionof magnetite.(Forinterpretationofthereferencestocolourinthisfigurelegend,the readerisreferredtothewebversionofthisarticle.)

InFig.6,weshowtheopticalmicroscopyimageofasample region((a), top panel) together withtwo Ramanmaps of this region((a),middle and bottompanels) andtwo related typical spectra (b) recorded from reddish-brown finishing layer of I2sample.

Again,Fig.7reportstheopticalmicroscopyimageofasample region((a), top panel) together withtwo Ramanmaps of this region((a),middleandbottompanels)andtworelatedtypical(b), this time recorded fromgreenish and light yellowish finishing layersofI2fragment.

Inbothcases,thepigmentedlayershowstwodifferentregions: oneischaracterizedbythepresenceofochre(redline),constituted by hematite and magnetite (red arrow), widely distributed as indicatedbythehomogeneityofthespectrain alltheanalysed area,and theotherregionshowsthepresenceof calcite(green line), with its typical fingerprints occurring at 283, 714 and, especially,1088cm 1_[31].

For this sample, micro-Raman spectral mapping, otherthan confirmingtheresultsobtainedbyFT-IRmeasurements,furnished useful information on the composition of the ochre. We can

associatethepresenceofCaCO3tothetechniqueofRomanfrescoes

orconsideritasaccessorymineralthatcontributestoaparticular nuanceofredochre.

Finally,thenoticeablesimilarityofthesetwospectraconfirms thehypothesisfortheuseofthesameagentsfor obtainingthe wantedcolouration.

3.3.UV-Ramanspectroscopymeasurements

As first attempt in the cultural heritage field, UV-Raman measurementshavebeencarriedoutforsamplesI2andI4B,inthe aim tocompare them with, micro-Raman and portable Raman results,respectively.

UV-RamanspectrumoftheredpigmentofI2(seeFig.8(a)and

Table 2)sample gaveclearevidenceof calcitepeakscentredat 283,713and1088cm 1_(mainpeak).

Asexpected, inthis casethesepeaks arestronglyenhanced, sinceweareworkinginresonantconditionswiththefunctional groupsofCaCO3,whichoccurjustatUVwavelengths.Moreover,

forthesamefragment,thegreenpigmentexhibitedanUV-Raman spectrum (not reported here) very similar to that of the red pigment,inagreementwithwhathasalreadybeenhypothesized accordingtotheanalysisoftheFT-IRandmicro-Ramanspectra.

Concerning I4B fragment (Fig. 8(b), the well evident peak centred at 262cm 1 _{can be ascribed to cinnabar [28–31],}

confirming the portable-Raman results. In addition, the most prominentpeakat1080cm 1_{testifiesonceagainthepresenceof}

calcite already associated to the painting technique of Roman frescoes.

Theagreementoftheobtainedresultswiththosecomingfrom conventional Raman methodologies encourages us on possible futureapplicationofUV-Ramantechniquesintheculturalheritage field. We would like to further stress that, in principle, the tunabilityofthesynchrotronradiationsource,offeredbyIUVS set-up, will allow us to work in resonant conditions between the excitationwavelengthand thewavelengthof UV-absorptionfor specificfunctionalgroupsoftheinvestigatedsamples,whichisnot accessibletoconventionallaserRamansystems.Finally,itcould alsobeconsideredasanalternativespectroscopicapproachwith

Fig.7.Opticalmicroscopyimageofasampleregion((a),toppanel)togetherwith twoRamanmapsofthisregion((a),middleandbottompanels)andtworelated typical (b) collected from greenish and light yellowish finishing layers of I2fragmentofplaster.Redline:ochre,Greenline:calcite.Theredarrowindicates themaincontributionofmagnetite.(Forinterpretationofthereferencestocolour inthisfigurelegend,thereaderisreferredtothewebversionofthisarticle.)

(a)

(b)

Fig. 8.UV-Ramanspectrum, inthe 1130–200cm 1

range, recordedfrom the reddishfinishinglayerofI2(a)andI4B(b)samples,relativetothedifferentanalysed zones.(Forinterpretationofthereferencestocolourinthisfigurelegend,thereader isreferredtothewebversionofthisarticle.)

Fig.6.Opticalmicroscopyimageofasampleregion((a),toppanel)togetherwith twoRamanmapsofthisregion((a),middleandbottompanels)andtworelated typical(b)recordedfromreddish-brownfinishinglayerofI2fragmentofplaster. Redline:ochre,Greenline:calcite.Theredarrowindicatesthemaincontributionof magnetite.(Forinterpretationofthereferencestocolourinthisfigurelegend,the readerisreferredtothewebversionofthisarticle.)

respecttosynchrotron radiationX-ray absorptionspectroscopy (SR-XAS) techniques, recently successfully exploited in some culturalheritagestudies,inwhichthenon-destructiveidenti fica-tion of the main pigmenting agents of coloured coatings was indirectlyachievedbymeansofstructuralinformation.

4.Conclusions

TheplastermakingtechniqueusedintheRomanarchaeological siteofVilladeiQuintiliwallpaintingsishereexploredbya multi-methodological characterization of seven fragments of plasters withcoloured_finishinglayers.

TheFT-IR profilesofthebulkofalltheinvestigatedsamples revealedastronghydrationandaredominatedbythepresenceof calcite. In addition, kaolinite is recognized for I2 and MD1fragments.

Concerningthedecoratedlayers,thecombineduseofFT-IRand micro-Raman spectroscopies allowed, as main result, for the unambiguous identification of red ochre as red pigment. This occurrencesuggeststheuseoflocallyavailable materialsinthe paintingsofVilladeiQuintili.

Finally,weattemptedforthefirsttimetotestandvalidatethe UV-Raman technique as a non-destructive powerful tool for obtaining a detailed compositional characterization of the decoratedsurfacesof artworksofhistorical-artistic interest.For thetwosamplesinvestigated,theresultsareinagreementwith thoseobtainedbyconventionalRamantechniques,encouragingus onpossible future application of UV-Raman techniques in the culturalheritagefieldwhenevertheconventionalRamanset-upis notabletodetectRamanscattering.

Acknowledgments

TheauthorsaregratefultoDr.RitaParis,ArchaeologistDirector ofSoprintendenzaSpecialeperiBeni ArcheologicidiRoma, for allowingarchaeologicalsamplestobetaken.

References

[1]F.Bardelli,G.Barone,V.Crupi,F.Longo,G.Maisano,D.Majolino,P.Mazzoleni, V.Venuti,IronspeciationinancientAtticpotterypigments:anon-destructive SR-XASinvestigation,J.SynchrotronRadiat.19(2012)782–788.

[2]G.Barone,V.Crupi,F.Longo,D.Majolino,P.Mazzoleni,G.Spagnolo,V.Venuti, E.Aquilia,Potentialityofnon-destructiveXRFanalysisforthedeterminationof CorinthianBamphoraeprovenance,X-raySpectrom.40(2011)333–337. [3]V.C.Farmer,InfraredSpectraofMinerals,MineralogicalSociety,London,1974. [4]G.A.Hope,R.Woods,C.G.Munce,Ramanmicroprobemineralidentification,

Miner.Eng.14(2001)1565–1577.

[5]H.G.M.Edwards,E.M.Newton,J.Russ,Ramanspectroscopicanalysisof pigmentsandsubstratainprehistoricrockart,J.Mol.Struct.550–551(2000) 245–256.

[6]M.Bouchard,D.C.Smith,Catalogueof45referenceRamanspectraofminerals concerningresearchinarthistoryorarchaeology,especiallyoncorroded

metalsandcolouredglass,Spectrochim.ActaPartA:Mol.Biomol.Spectrosc. 59(2003)2247–2266.

[7]P.Vandenabeele,H.G.M.Edwards,L.Moens,AdecadeofRamanspectroscopy inartandarchaeology,Chem.Rev.107(2007)675–686.

[8]K.Eremin,J.Stenger,J.Huang,A.Aspuru-Guzik,T.Betley,L.Vogt,I.Kassal,S. Speakman,N.Khandekar,ExaminationofpigmentsonThaimanuscripts:the firstidentificationofcoppercitrate,J.RamanSpectrosc.39(2008)1057–1065. [9]M.C.Edreira,M.J.Feliu,C.Fernandez-Lorenzo,J.Martın,Spectroscopicanalysis ofromanwallpaintingsfromCasadelMitreoinEmeritaAugusta,Merida, Spain,Talanta59(2003)1117–1139.

[10]L.M.Moretto,E.F.Orsega,G.A.Mazzocchin,Spectroscopicmethodsforthe analysisofceladoniteandglauconiteinRomangreenwallpaintings,J.Cult. Herit.12(2011)384–391.

[11]I.Aliatis,D.Bersani,E.Campani,A.Casoli,P.P.Lottici,S.Mantovana,I.G.Marino, F.Ospitali,GreenpigmentsofthePompeianartists’palette:Spectrochim,Acta PartA:Mol.Biomol.Spectrosc.73(2009)532–538.

[12]R.J.H.Clark,S.Mirabaud,Identificationofthepigmentsonasixteenthcentury PersianbookofpoetrybyRamanmicroscopy,J.RamanSpectrosc.37(2006) 235–239.

[13]B.Schrader,InfraredandRamanSpectroscopy:MethodsandApplications, VCHpublishers,Weinheim,1995.

[14]J.R.Ferraro,K.Nakamoto,IntroductoryRamanSpectroscopy,AcademicPress, Boston,1994.

[15]E.Smith,G.Dent,ModernRamanSpectroscopy—APracticalApproach,J.Wiley &Sons,Hoboken,2005.

[16]I.N.Levine,MolecularSpectroscopy,JohnWileyandSons,NewYork,1975. [17]R.S.Czernuszewicz,T.G.Spiro,IRramanandresonanceramanspectroscopy,

in:E.I.Solomon,A.B.P.Lever(Eds.),InorganicElectronicStructureand Spectroscopy;VolumeI:Methodology,JohnWileyandSons,NewYork,1999, pp.353–441.

[18]A.Rotondi,FormaUrbis-ItinerariNascostiDiRomaAntica,ANNOXVII(2), Roma,2012.

[19]C.M.Belfiore,G.V.Fichera,M.F.LaRussa,A.Pezzino,S.A.Ruffolo,G.Galli,D. Barca,AmultidisciplinaryapproachforthearchaeometricstudyofPozzolanic aggregatesinRomanmortars:thecaseofVilladeiQuintili(RomeItaly), Archaeometry57(2015)269–296.

[20]V.Crupi,G.Galli,M.F.LaRussa,F.Longo,G.Maisano,D.Majolino,M.Malagodi, A.Pezzino,M.Ricca,B.Rossi,S.A.Ruffolo,V.Venuti,Multi-technique investigationofromandecoratedplastersfromvilladeiquintili(Rome,Italy), Appl.Surf.Sci.349(2015)924–930.

[21]F.D’Amico,M.Saito,F.Bencivenga,M.Marsi,A.Gessini,G.Camisasca,E. Principi,R.Cucini,S.DiFonzo,A.Battistoni,Nucl.Instrum.MethodsPhys.Res., Sect.A703(2013)33–37.

[22]A.Rotondi,LaVilladeiQuintilieisuoiantichiproprietari,Forma Urbis-ItinerarinascostidiRomaAnticaANNOXVII(2)(2012)6–8.

[23]G.V.Fichera,D.Barca,C.M.Belfiore,M.F.LaRussa,A.Pezzino,Rendiconti OnlineSocietàGeologicaItaliana21(2012)659–660.

[24]NORMAL3/80.Materialilapidei:campionamentoeconservazionedei campioni,1980,Roma.

[25]SadtlerDatabaseforFT-IR,BioRadLaboratories.

[26]W.P.Griffith,AdvancesinRamanandinfraredspectroscopyofminerals,in:R.J. H.Clark,R.E.Hester(Eds.),SpectroscopyofInorganic-BasedMaterials,John Wiley&Sons,Chichester,UK,1987,pp.119–186.

[27]G.E.DeBenedetto,R.Laviano,L.Sabbatini,P.G.Zambonin,Infrared spectroscopyinthemineralogicalcharacterizationofancientpottery,J.Cult. Herit.3(2002)177–186.

[28]I.M.Bell,R.J.H.Clark,P.J.Gibbs,RamanSpectroscopicLibraryofNaturaland SyntheticPigments,http://www.chem.ucl.ac.uk/resources/raman/. [29]RRUFFProject,2010.DepartmentofGeosciences,UniversityofArizona,

Tucson,USA.http://rruff.info/.

[30]HandbookofRamanspectra,http://ens-lyon.fr/LST/Raman.

[31]S.P.S.Porto,J.A.Giordmaine,T.C.Damen,DepolarizationofRamanscatteringin calcite,Phys.Rev.147(1966)608–611.