How Many Secret Details Could a Systematic Multi-Analytical Study Reveal About the Mysterious Fresco Trionfo Della Morte?

heritage

Article How Many Secret Details Could a Systematic Multi-Analytical Study Reveal About the Mysterious Fresco Trionfo della Morte?

Maria Francesca Alberghina 1, Salvatore Schiavone 1, Caterina Greco 2, Maria Luisa Saladino 3,* , Francesco Armetta 3, Vincenzo Renda 4 and Eugenio Caponetti 3,5

1 S.T.Art-Test di S. Schiavone & C. S.A.S, Via Stovigliai 88, I-93015 Niscemi (CL), Italy 2 Museo Archeologico Regionale “Antonino Salinas”, Via Bara all’Olivella 24, I-90133 Palermo, Italy 3 Dipartimento Scienze e Tecnologie Biologiche, Chimiche e Farmaceutiche - STEBICEF and INSTM UdR—Palermo, Università di Palermo, Viale delle Scienze pad.17, I-90128 Palermo, Italy 4 CNR – Istituto per I Processi Chimico-Fisici (IPCF), V.le F. Stagno D’Alcontres, 37, I-98158 Messina, Italy 5 LaborArtis C.R. Diagnostica s.r.l., Viale delle Scienze Ed.16, I-90128 Palermo, Italy * Correspondence: [email protected]

Received: 30 June 2019; Accepted: 3 August 2019; Published: 7 August 2019

Abstract: The “Trionfo della morte” is a detached fresco painting dated at the half of the XV century. Its history is strictly connected with the history of Palermo and it is considered a symbol of the late Gothic period. Some small areas of the fresco were analyzed using a combination of non-invasive techniques and hand-held instrumentations (multispectral imaging analysis, X-ray ﬂuorescence (XRF), and IR spectroscopy). The characterization of the nature of pigments used in its realization and restoration works was performed and some indications about its conservation state were obtained. More interestingly, some hidden details were revealed on the mysterious painting. They constitute additional evidence of the preciousness of the fresco.

Keywords: multispectral analysis; XRF; portable instruments; wall painting; Trionfo della morte

1. Introduction The "Trionfo della morte" (Triumph of Death), as shown in Figure1, is an imposing and unique fresco painting (600 642 cm) and is one of the greatest masterpieces of all time, with a shocking × representative power [1]. An unknown painter realized the fresco indicatively between 1440 and 1450. In the mid-fifteenth century, Palermo was Hispanic and under the reign of Alfonso V d’Aragona known as the Magnanimous, when the spread of epidemics, famines, and the Black Plague changed the perception of death, until the formation of a new expressive sensitivity, which led to the birth of vast literature and an intense figurative development on his iconography. The “Trionfo della morte” thus became the most significant and emblematic artistic expression of the late Gothic period in Sicily. The fresco was realized in the courtyard of Palazzo Sclafani, built in 1330 at the behest of Count Matteo Sclafani, near the Palazzo dei Normanni, the Royal Palace. After Alfonso V d’Aragona’s death and the long struggle for his succession, in 1400 the palace was confiscated and assigned to a noble Spanish family, who however, returned to Spain abandoning it to a progressive degradation. Thirty years later, the palace was chosen to be the first public hospital of Palermo and its courtyard has been decorated with sculptures, paintings, and frescoes, which in some way could accompany and alleviate, spiritually, the guests in their sufferings.

Heritage 2019, 2, 2370–2383; doi:10.3390/heritage2030145 www.mdpi.com/journal/heritage Heritage 2019, 2 2371 Heritage 2019, 2 FOR PEER REVIEW 2

Figure 1.1. PhotoPhoto ofofTrionfo Trionfo della della Morte Morte, XV, XV century, century, Gallery Gallery of Art of forArt the for Sicilian the Sicilian region region Abatellis Abatellis Palace, Palace,Palermo Palermo (Italy). (Italy).

The large fresco remained in the courtyard of the hospital for five five centuries, and it remained in almost perfect condition until the outbreak of the Second World War.War. The bombingbombing ofof Palermo Palermo in in 1943 1943 caused caused the collapsethe collapse of the of vaulted the vaulted roof and roof other and serious other damageserious damageto Palazzo to Sclafani;Palazzo Sclafani; therefore, therefore, in order to in preserve order to thepreserve masterpiece the masterpiece it was decided it was to decided detach it to from detach the wallit from and the move wall it and to a move safer place.it to a Consideringsafer place. Considering the size, the frescothe size, was the divided fresco intowas fourdivided parts into in order four partsto detach in order it from to thedetach wall it more from easily. the wall It was more saved easi atly. Sala It was delle saved Lapidi at ofSala Palazzo delle PretorioLapidi of until Palazzo 1954 whenPretorio it wasuntil moved 1954 towhen the Regionalit was moved Gallery to of Sicilythe Regional in Palazzo Gallery Abatellis of Sicily (Palermo, in Palazzo Italy) where Abatellis even (Palermo,now it is exposed Italy) where in the even new now museum it is exposed layout designed in the new by themuseum Venetian layout architect designed Carlo by Scarpa. the Venetian architectThe Carlo fresco Scarpa. was restored by the Istituto Centrale del Restauro (ICR), under the care of Cesare Brandi. A furtherThe fresco restoration was inrestored the 1980s by wasthe carriedIstituto outCentrale by restorers del Restauro Carlo (ICR) Giantomassi, under the and care Donatella of Cesare Zari Brandi.under the A supervisionfurther restoration of Giovanni in the Urbani 1980s of thewas ICR. carried The restorationout by restorers provided Carlo for theGiantomassi cleaning of and the paintingsDonatella andZari the under application the supervision of new supports. of Giovanni Urbani of the ICR. The restoration provided for the cleaningToday, theof the“Trionfo paintings della and morte” the applicationis considered of new as an supports. “iconographic encyclopedia”, a medieval allegoryToday, containing the “Trionfo the essence della morte” of everything: is considered life and as death, an "iconographic beauty, old age, encyclopedia", ease and poverty, a medieval sadness, allegorybitterness, containing amazement, the disdain, essence care,of everything: compassion, life love, and and death, hope. beauty, old age, ease and poverty, sadness,Considering bitterness, the amazement, uniqueness, disdai the peculiarity,n, care, compassion, and the vicissitudes love, and hope. of this painting, it was deeply studiedConsidering by a historical the uniqueness, point of view, the peculiarity, but, to our and knowledge, the vicissitudes no scientific of this investigationpainting, it was about deeply the studiedmaterials by constituting a historical the point painting of view, was but, ever to performed. our knowledge, Therefore, no scientific we thought investigation that a scientific about study the materialscould be crucialconstituting to reveal the painting invisible was details, ever hiddenperforme byd. theTherefore, artist or we by thought time [2 ,that3] to a givescientific additional study couldinformation be crucial about to its reveal conservation invisible state. details, hidden by the artist or by time [2-3] to give additional information about its conservation state.

Heritage 2019, 2 2372 Heritage 2019, 2 FOR PEER REVIEW 3

InIn this this paper, paper, results results obtained obtained by by the the application application of ofmulti-analytical multi-analytical non-invasive non-invasive investigations investigations carriedcarried out out on on a few a few areas areas of of“Trionfo“Trionfo della della morte” morte” areare reported. reported. TheThe diagnostic diagnostic campaign campaign was was carried carried out out by following by following a well-consolidated a well-consolidated approach approach applied applied to paintingsto paintings starting starting from from the thedocumentation documentation step step by bymultispectral multispectral analysis analysis and and followed followed by by the the spectroscopicspectroscopic investigation investigation [4-9]. [4–9 Multispectral]. Multispectral imaging imaging (in (in UV, UV, visible visible and and infrared infrared range, range, and and false false colorcolor infrared, infrared, FCIR), FCIR), X-ray X-ray fluorescence fluorescence (XRF), (XRF), and and Fourier Fourier transform transform infrared infrared (FT-IR) (FT-IR) spectroscopy spectroscopy werewere applied applied using using hand-heldhand-held instrumentsinstruments in in order order to investigateto investigate the pigmentsthe pigments belonging belonging to diff erentto differentfigures figures and to identify and to identify the materials the materials and methods and me usedthods inthe used original in the realizationoriginal realization and in the and restoration. in the restoration.Some photos, Some taken photos, during taken thecampaign, during the are campai reportedgn, inare Figure reported2. The in results Figure here 2. reportedThe results constitute here reporteda preliminary constitute phase a preliminary of the project phase study of the proposed project tostudy characterize proposedthe to characterize materials and the the materials painting andtechnique the painting of the technique whole fresco of the surface, whole among fresco the surface, most importantamong the in most the panorama important of in mural the panorama painting of ofthe mural XV painting century inof Sicily.the XV century in Sicily.

Figure 2. Photos of the in situ diagnostic campaign. (left) multispectral imaging acquisition; (right-up) Figure 2. Photos of the in situ diagnostic campaign. left) multispectral imaging acquisition; right-up) XRF spectra acquisition; (right-down) IR spectra acquisition. XRF spectra acquisition; right-down) IR spectra acquisition. 2. Instrumentations and Methods 2. Instrumentations and Methods The multispectral imaging acquisition was carried out using a digital CCD cooled scientific camera (CHROMAThe multispectral C4-DSP series,imaging model acquisition C250ME was by DTAcarried srl) out with using a KAF8300ME a digital CCD sensor, cooled equipped scientific with cameraan internal (CHROMA wheel C4-DSP containing series, interference model C250ME filters by (350, DTA 400, srl) 450, with 540, a KAF8300ME 600, 750, 1000 sensor, nm bandsequipped filter, with50 nman bandwidth).internal wheel The containing digital scientific interference camera filt usesers a(350, Cooled 400, CCD, 450, 6540, Megapixel 600, 750, (3326 1000 nm2504 bands points) × filter,effective 50 nm (pixels bandwidth). with side The of 5.4digitalµm). scientific The acquired camera visible uses anda Cooled infrared CCD, images 6 Megapixel were also (3326 processed × 2504 to points)obtain effective infrared (pixels false color with information side of 5.4 (FCIR)µm). Th fore aacquired preliminary visible mapping and infrared of different images pictorial were layersalso processedand for discriminatingto obtain infrared original false or color restoration informat surfaces.ion (FCIR) for a preliminary mapping of different pictorialXRF layers spectra and werefor discriminating acquired with original a Tracer or III restoration SD Bruker surfaces. AXS instrument having a rhodium anode (VoltageXRF spectra 40 kV and were current acquired 11 µ A)with as sourcea Tracer and III a SD Silicon Bruker Drift AXS XFlash instrument®as a detector. having Each a rhodium spectrum anodewas acquired(Voltage for40 kV 30 sand in an current area of 11 3 mmµA)2 asby source placing and the a instrument Silicon Drift in XFlash® contact with as a thedetector. surface Each of the 2 spectrumpainting. was The acquired collected for spectra 30 s werein ananalyzed area of 3 using mm theby softwareplacing the ARTAX instrument 7. In all in acquired contact spectra,with the the surfacesignals of of the argon painting. (Ar), rhodium The collected (Rh), and spectra palladium were (Pd)analyzed elements using are the present software due toARTAX the environment 7. In all acquiredand the spectra, setup of the the signals used device. of argon (Ar), rhodium (Rh), and palladium (Pd) elements are present due to Reflectancethe environment FT-IR and spectra the setup were acquiredof the used using device. the portable Bruker ALPHA FT-IR Spectrometer equippedReflectance with FT-IR an External spectra Reflection were acquired QuickSnapTM using the portable module. Bruker Spectra ALPHA were acquired FT-IR Spectrometer between 3500 equipped with1 an External Reflection QuickSnapT1 M module. Spectra were acquired between 3500 and 360 cm− with a resolution of 4 cm− , and averaging 60 scans for each measurement, in an area of -1 -1 and5 mm3602 cm, by placingwith a resolution the instrument of 4 cm in a, and contactless averaging manner 60 scans with for the each painting’s measurement, surface in with an area a working of 5 2 mmdistance, by placing of about the 15instrument mm. All in reflectance a contactless spectra manner were with processed the painting’s using OPUS surface 7.5 with software; a working every distance of about 15 mm. All reflectance spectra were processed using OPUS 7.5 software; every spectrum was converted to pseudo-absorbance (log R) units in order to compare the spectrum with a database of reflectance reference spectra [10].

Heritage 2019, 2 FOR PEER REVIEW 4

3. Results and Discussion The application of the multispectral non-invasive investigation to the two pictorial areas of the “Trionfo della morte” fresco allowed the discovery of some iconographic details no longer visible to the naked eye and to map the precious “a secco” details. Some repainting areas are well evident because of the use of “rigatino” pictorial integration techniques. Nevertheless, thanks to different spectral behaviors, the original layers are immediately distinguishable from pictorial integrations not distinguishable to the naked eye, not even to a close view. Some of the multispectral images obtained by the multispectral investigation (visible, FCIR, and IR reflectography) are reported in Figures 3–8. The first analyzed area, corresponding to the group of women in the foreground at the bottom of the fresco, as shown in Figures 3–5, confirmed the importance of a systematic imaging analysis on this work of art. It is in fact clearly visible the preciousness and complexity of the iconographic painted subjects that become again readable thanks to the pictorial surface observation in the single filtered spectral bands. HeritageIn2019 some, 2 cases, the FCIR images provided preliminary information about the pigments on 2373the basis of IR spectral behavior, and, thanks to the comparison with FCIR images, the XRF results were extended within the whole painted area characterized by the same spectral color. For example, as spectrumobservable was in convertedFigure 5, the to pseudo-absorbance yellow tone of the (log red R) dress units of in the order woman to compare suggests the spectrumthe presence with of a databasecinnabar ofused reflectance in the mixture reference due spectra to the [10 ].typical spectral response of this red pigment in FCIR 3.imaging Results analysis and Discussion [11–13]. The details of the two women's collars and the headgear of the central figure are more noticeableThe application in the filter of visible the multispectral range, which non-invasive became transparent investigation in infrared, to the twoas shown pictorial in Figures areas of 3–5, the “andTrionfo consequently della morte” identifiable fresco allowed as the“a secco discovery” details. of some On iconographicthe contrary, detailsthe fresco no longer surface visible is not to thetransparent naked eye to the and IR to spectral map the range precious due “toa the secco high” details. IR absorption Some repainting of calcium areas carbonate are well matrix evident [9]. becauseUnder a ofUV the lamp use ofa “yellowishrigatino” pictorialfluorescence integration in correspondence techniques. to Nevertheless, the “a secco thanks” decorations to different was spectralobserved. behaviors, This evidence the original could layersbe associated are immediately to a resinous distinguishable oleo material from used pictorial as binder integrations for pictorial not distinguishable“a secco” layers [14,15]. to the naked eye, not even to a close view. SomeMoreover, of the IR multispectral reflectography images of the obtained central byfemale the multispectral figure highlights investigation the breast (visible, contour, FCIR, hidden and IRby reflectography)a white-collar decoration are reported in visible in Figures range,3–8 .as shown in Figures 3 and 5b.

Figure 3.3. Infrared reflectographyreflectography image acquiredacquired with a 1000 nm filterfilter on a particular of the female Heritagefigurefigure 2019 group., 2 FOR PEER REVIEW 5

FigureFigure 4.4. VisibleVisible range range images images acquired acquired with with (a a)) 450 450 nm, nm, ( bb)) 540540 nm,nm, (c)c) 600 nm filtersfilters onon thethe detaildetail ofof thethe femalefemale figurefigure group. group. The first analyzed area, corresponding to the group of women in the foreground at the bottom of the fresco, as shown in Figures3–5, confirmed the importance of a systematic imaging analysis on this work of art. It is in fact clearly visible the preciousness and complexity of the iconographic painted subjects that become again readable thanks to the pictorial surface observation in the single filtered spectral bands.

Figure 5. Comparison between a) the false color infrared (FCIR) image and b) the IR reflectography (1000 nm) on the detail of the female figure group. The characteristic spectral color provided by FCIR acquisition suggests the typology of the used pigments for original or restoration areas. For example, the green area of the dress could be typical of copper-based green pigment (malachite) and the red area could be attributable to chromium green constituting the localized integration in pictorial layers.

Also, in the second investigated area, as shown in Figures 6–8, the multispectral imaging provided a clearer legibility of iconographic detail and allows us to provide preliminary information about the executive techniques of pictorial layers. For example, the details of the “a secco” arrows, almost transparent to the IR reflectography, allow us to see the underlying drapery, as shown in Figure 6. The beard “a secco” details of the bishops become much more appreciable with a 450 nm filter, as shown in Figure 7c. Moreover, the contours of the "giornata" frescoes are more evident along the profile of the bishop’s face in the IR range, as shown in the detail in Figure 7d. Finally, the decorations of the red hat, no longer appreciable to the naked eye, appear readable in the IR image thanks to the IR transparency of the red pigment, as shown in Figure 8.

Heritage 2019, 2 FOR PEER REVIEW 5

HeritageFigure2019, 24. Visible range images acquired with a) 450 nm, b) 540 nm, c) 600 nm filters on the detail of 2374 the female figure group.

Figure 5. Comparison between (a)a) the the false false color color infrared (FCIR) image andand (b)b) the the IR IR reflectography reﬂectography (1000 nm) on the detail of the female female figure ﬁgure group. The The characteristic characteristic spectral color color provided provided by FCIR FCIR acquisition suggests the typology of the used pigments for original original or or restoration restoration areas. areas. For For example, example, the green areaarea ofof the the dress dress could could be be typical typical of of copper-based copper-based green green pigment pigment (malachite) (malachite) and and the redthe areared Heritageareacould 2019 could, be 2 FOR attributable be PEERattributable REVIEW to chromium to chromium green green constituting constituting the localized the localized integration integration in pictorial in pictorial layers. layers. 6

Figure 6. InfraredInfrared reflectography reﬂectography image image ac acquiredquired with with a a 1000 1000 nm nm filter ﬁlter on a particular of the lower central part of frescos.

In some cases, the FCIR images provided preliminary information about the pigments on the basis of IR spectral behavior, and, thanks to the comparison with FCIR images, the XRF results were extended within the whole painted area characterized by the same spectral color. For example, as observable in Figure5, the yellow tone of the red dress of the woman suggests the presence of cinnabar used in the mixture due to the typical spectral response of this red pigment in FCIR imaging analysis [11–13].

Figure 7. Comparison between a) the false color infrared (FCIR) image and b) the IR reflectography (1000 nm) on the detail of the bishop. c) Details of the beard “a secco” of the bishops; d) the IR reflectography zoomed images highlight the path of “giornata” frescos along the contour of the bishop’s figure (face and hand).

Heritage 2019, 2 FOR PEER REVIEW 6

HeritageFigure2019, 26. Infrared reflectography image acquired with a 1000 nm filter on a particular of the lower 2375 central part of frescos.

Figure 7. Comparison between (a)a) the the false false color infrared (FCIR) image andand (b)b) the the IR IR reflectography reflectography (1000 nm) onon thethe detaildetail ofof thethe bishop.bishop. (c)c) Details Details of of the beard “a secco” of thethe bishops;bishops; (d)d) the IR reflectographyreflectography zoomedzoomed images images highlight highlight the the path path of “ giornataof “giornata” frescos” frescos along thealong contour the contour of the bishop’s of the Heritagebishop’sfigure 2019,(face 2 FORfigure and PEER (face hand). REVIEW and hand). 7

Figure 8. Comparison between (up)up) the the visible visible image do (downwn) )the the IR IR reflectography reﬂectography (1000 (1000 nm) nm) on the detail of the red hat. The hidden decorations of the red hat are readable in the IR image.

The detailsXRF and of the IR twospectra women’s were collars acquired and thein headgearsitu to investigate of the central the ﬁgure nature are of more pigments, noticeable both in originalthe ﬁlter and visible used range, in restoration which became in order transparent to provide in infrared, preliminary as shown information in Figures on3 –the5, andunknown consequently artist's palette and, then, on the typical materials and executive techniques of the XV century Sicilian artistic production still not systematically studied from an archaeometric point of view. The points for spectra acquisition were chosen on the basis of the multispectral investigation, distinguishing the different materials thanks to the different spectral behavior. The maps of the points analyzed by XRF are reported in Figure 9. Measuring points’ numbering, color hatch together with a short description of the analyzed points and the identified elements are reported in Table 1. Some of the XRF spectra are reported in Figures 10–12.

Figure 9. Location of the X-ray Fluorescence (XRF) measuring points in some area of the painting..

Heritage 2019, 2 FOR PEER REVIEW 7

Heritage 2019, 2 2376 identifiable as “a secco” details. On the contrary, the fresco surface is not transparent to the IR spectral range due to the high IR absorption of calcium carbonate matrix [9]. Under a UV lamp a yellowish fluorescence in correspondence to the “a secco” decorations was observed. This evidence could be associated to a resinous oleo material used as binder for pictorial “a secco” layers [14,15]. Moreover, IR reflectography of the central female figure highlights the breast contour, hidden by a white-collar decoration in visible range, as shown in Figures3 and5b. Also, in the second investigated area, as shown in Figures6–8, the multispectral imaging provided a clearer legibility of iconographic detail and allows us to provide preliminary information about the executive techniques of pictorial layers. For example, the details of the “a secco” arrows, almost transparent to the IR reflectography, allow us to see the underlying drapery, as shown in Figure6. The beard “ a secco” details of the bishops become much more appreciable with a 450 nm filter, as shown in Figure7c. Moreover, the contours of the “giornata” frescoes are more evident along the profile of the bishop’s face in the IR range, as shown in the detail in Figure7d. Figure 8. Comparison between up) the visible image down) the IR reflectography (1000 nm) on the Finally, the decorations of the red hat, no longer appreciable to the naked eye, appear readable in detail of the red hat. The hidden decorations of the red hat are readable in the IR image. the IR image thanks to the IR transparency of the red pigment, as shown in Figure8. The XRFXRF andand IR IR spectra spectra were were acquired acquired in situ in tositu investigate to investigate the nature the nature of pigments, of pigments, both original both originaland used and in restoration used in restoration in order to in provide order to preliminary provide preliminary information information on the unknown on the artist’s unknown palette artist's and, palettethen, on and, the then, typical on materials the typical and ma executiveterials and techniques executive oftechniques the XV century of the XV Sicilian century artistic Sicilian production artistic productionstill not systematically still not systematically studied from anstudied archaeometric from an pointarchaeometric of view. The point points of forview. spectra The acquisitionpoints for werespectra chosen acquisition on the were basis chosen of the multispectralon the basis of investigation, the multispectral distinguishing investigation, the didistinguishingfferent materials the differentthanks to materials the different thanks spectral to the behavior.different spectral The maps behavior. of the pointsThe maps analyzed of the points by XRF analyzed are reported by XRF in areFigure reported9. Measuring in Figure points’ 9. Measuring numbering, points’ color numberin hatch togetherg, color with hatch a shorttogether description with a short of the description analyzed pointsof the analyzed and the identified points and elements the identified are reported elements in Tableare reported1. Some in of Table the XRF 1. Some spectra of the are XRF reported spectra in areFigures reported 10–12 in. Figures 10–12.

FigureFigure 9. 9. LocationLocation of of the the X-ray X-ray Fluorescence Fluorescence (XRF) (XRF) measur measuringing points points in in some some area area of the painting.. painting.

Heritage 2019, 2 2377

Table 1. Measuring points’ numbering, color hatch together with a short description of the analyzed point and the identiﬁed chemical elements. The abundant elements are reported in bold.

Point Color Identiﬁed Chemical Elements 1 white vest S, Ca, Fe, Sr, Pb 2 grey vest K, Ca, Ti, Fe, Sr, Pb 3 incarnate friar Ca, Fe, Sr 4 yellow sleeve Ca, Fe, Sr, Sn, Pb, Au 5 green next to yellow K, Ca, Cr, Mn, Fe, Sr, Pb, Sn, Au 6 orange Si, K, Ca, Ti, Fe, Sr, Pb 7 dark green Si, P, K, Ca, Cr, Mn, Fe, Cu, Sr, Pb 8 brown beard Si, K, Ca, Mn, Fe, Sr, Pb 9 blue hat K, Ca, Cr, Mn, Fe, Cu, Sr, Pb, Ba 10 yellow hat Si, S, K, Ca, Ti, Cr, Fe, Sr, Au 11 white turban S, Ca, Fe, Sr 12 black turban Si, S, K, Ca, Fe, Sr 13 hand of bishop Si, S, Ca, Fe, Sr 14 blue mantle Si, S, K, Ca, Cr, Mn, Fe, Cu, Sr, Ba, Pb 15 yellow hat Si, P, S, K, Ca, Ti, Cr, Fe, Sr, Pb 16 green water Si, P, K, Ca, Cr, Mn, Fe, Sr 17 pink mantle Si, P, S, Ca, Fe, Sr 18 green mantle Si, S, K, Ca, Cr, Fe, Sr 20 grey light Si, S, Ca, Fe, Sr 22 red right pendant Si, S, K, Ca, Cr, Mn, Fe, Sr, Ba, Hg, Pb 24 red hat Si, S, K, Ca, Fe, Sr, Hg, Pb 25 red hat Si, S, K, Ca, Fe, Sr, Hg, Pb 26 red low pendant Si, S, K, Ca, Cr, Mn, Fe, Sr, Hg, Pb 27 red right low pendant Si, S, K, Ca, Ti, Cr, Mn, Fe, Sr, Hg, Pb, Ba 28 integration red hat Si, S, Ca, Cr, Mn, Fe, Sr, Pb 33 pink mantle Si, P, S, K, Ca, Cr, Fe, Sr, Pb 34 grey decoration Si, P, S, K, Ca, Fe, Sr, Pb 35 red mantle S, K, Ca, Cr, Fe, Sr, Hg, Pb 36 red stole Si, P, S, K, Ca, Mn, Fe, Sr, Pb 37 red mantle S, K, Ca, Fe, Sr, Hg, Pb 40 red left ﬁgure Si, S, K, Ca, Cr, Mn, Fe, Sr

In analyzing the spectra, it should be taken into account that, due to the penetration depth of X-rays, the signals are related both to the pigments and the wall support. All XRF spectra show the signals of calcium (Ca), ascribed to the presence of calcite (CaCO ) and/or gypsum (CaSO 2H O). 3 4· 2 The presence of sulfur (S) and strontium (Sr) in high quantities suggests the presence of gypsum as well as celestite (SrSO4), a mineral usually found associated with gypsum. More precise information could be obtained by using a portable X-ray diffractometer for the investigation of minerals. The presence of gypsum can be an alteration product of calcium carbonate or a component deliberately used by the artist [16,17]. The presence of strontium was connected to the last reason. On the other hand, the celestite in Sicily is often present in volcanic deposits from Mount Etna, but also can be commonly present in gypsum derived from seawater. The widespread presence of lead (Pb) in some analyzed areas, in smaller quantities than calcium, is indicative of the presence of white lead (basic lead carbonate (PbCO ) Pb(OH) ) used in the realization 3 2· 2 of the details “a secco”. It is well known that the fresco techniques are not compatible with the presence of white lead [9] because of its reactivity in the basic environment. Its presence is thus evidence of the “a secco” details (confirmed in some cases by the IR transparency obtained in the corresponding IR reflectography) probably applied by using an oil as binder, as verified under UV lamp examination. Heritage 2019, 2 2378 Heritage 2019, 2 FOR PEER REVIEW 9 Heritage 2019,, 2 FORFOR PEERPEER REVIEWREVIEW 9

Figure 10. XRF spectra of some white and brown analyzed points/colors. FigureFigure 10.10. XRF spectra of some whitewhite andand brownbrown analyzedanalyzed pointspoints/colors./colors.

Figure 11. XRF spectra of some analyzed blue and red points/colors. FigureFigure 11.11. XRF spectra ofof somesome analyzedanalyzed blueblue andand redred pointspoints/colors./colors.

. .. Figure 12. XRF spectra of some analyzed green and yellow points/colors. Figure 12. XRF spectra of some analyzed green and yellow points/colors. Figure 12. XRF spectra of some analyzed green and yellow points/colors.

Heritage 2019, 2 FOR PEER REVIEW 10

In analyzing the spectra, it should be taken into account that, due to the penetration depth of X-rays, the signals are related both to the pigments and the wall support. All XRF spectra show the signals of calcium (Ca), ascribed to the presence of calcite (CaCO3) and/or gypsum (CaSO4·2H2O). The presence of sulfur (S) and strontium (Sr) in high quantities suggests the presence of gypsum as well as celestite (SrSO4), a mineral usually found associated with gypsum. More precise information could be obtained by using a portable X-ray diffractometer for the investigation of minerals. The presence of gypsum can be an alteration product of calcium carbonate or a component deliberately used by the artist [16,17]. The presence of strontium was connected to the last reason. On the other hand, the celestite in Sicily is often present in volcanic deposits from Mount Etna, but also can be commonly present in gypsum derived from seawater. The widespread presence of lead (Pb) in some analyzed areas, in smaller quantities than calcium, is indicative of the presence of white lead (basic lead carbonate (PbCO3)2·Pb(OH)2) used in the realization of the details “a secco”. It is well known that the fresco techniques are not compatible with the presence of white lead [9] because of its reactivity in the basic environment. Its presence is thus evidence of the “a secco” details (confirmed in some cases by the IR transparency obtained in the Heritage 2019, 2 2379 corresponding IR reflectography) probably applied by using an oil as binder, as verified under UV lamp examination. OnOn the the other other hand, hand, the the P1 P1 and and P11 P11 spectra spectra of of white white mainly mainly show show the the signals signals of of Ca Ca together together with with ssmallmall signals ofof lead lead (Pb), (Pb), indicating indicating that that the whitethe white is calcium is calcium carbonate. carbonate.In addition, In addition, the IR spectrum the IR spectrumof the white of showsthe white a sequence shows of a peaks sequence matching of withpeaks calcite matching and gypsum. with calcite The pseudo-absorbance and gypsum. The IR pseudo-absorbancespectrum obtained byIR takingspectrum the negativeobtained logarithm by taking of the the negative reﬂectance logarithm data acquired of the (-log reflectance R) is reported data acquiredin Figure (-log 13. R) is reported in Figure 13.

Figure 13. IR spectra of (left) (left) white white (C (C = calcite,calcite, G G == gypsum)gypsum) and and of of (right) (right )blue blue (A (A == azurite).azurite).

TheThe absorbance absorbance and and derivative derivative peaks peaks were were identified, identified, by by comparison comparison with with a a series series of of pseudo-absorbancepseudo-absorbance spectra spectra obtained obtained on on standards, standards, as asgypsum gypsum and and calcite calcite and and are arehighlighted highlighted in the in figurethe figure (calcite (calcite C, gypsum C, gypsum G). G).Regarding Regarding calcite, calcite, in inreflection reflection mode, mode, the the lowest lowest absorption absorption due due to to 1 1 symmetricsymmetric carbonate carbonate stretching around 10001000 cmcm−-1 showsshows a a typical typical derivative derivative appearance; at 1800 cmcm−-1 aa very very sharp sharp and and characteristic characteristic absorption absorption due due to to a a νν1 1++ ν4ν combination4 combination is ispresent. present. Calcite Calcite has has also also a 1 1 signaturea signature and and a avery very wide wide band band at at 2512 2512 cm cm-1−, ,with with a a shoulder shoulder at at 2594 2594 cm -1− duedue to to νν11 ++ νν33 modemode coupling.coupling. Gypsum Gypsum presents presents signals signals which which are are overall overall less less traceable traceable in in respect respect to to the the calcite calcite signals. signals. 1 TheThe lowest lowest signal, signal, under under 1000 1000 cm -1−, ,can can be be attributed attributed to to bending bending mode mode vibrations. vibrations. Due Due to to the the higher higher absorptionabsorption index, νν33 antisymmetricantisymmetric stretching stretching of sulphateof sulphate is a ffisected affected by reststrahlen by reststrahlen, having, having its minima its 1 1 minimaaround 1160around cm −1160. The cm weak-1. The peak weak emerging peak emerging near 2200 near cm 2200− is duecm-1to is adue combination to a combination band of band bending of 1 bendingand vibration and vibration modes ofmodes H2O( ofν 2H+2Oν (Lν).2 + The νL). lastThe band last band in the in NIRthe NIR zone, zone, around around 5140 5140 cm− cm, can-1, can be beattributed attributed to theto the mixing mixing of vibrationalof vibrational modes modes connected connected to OH to bondsOH bonds of hydration of hydration water water (ν1/ν3 (OHν1/ν3+ OHν2 OH). + ν2 OH). The presenceThe presence of gypsum of gypsum could could be due be due to a to sulphating a sulphating process, process, common common degradation degradation in this in thistypology typology of painting. of painting. The The presence presence of silicon of silicon (Si) and (Si) potassium and potassium (K), characteristic (K), characteristic of silicates, of silicates, can be candue be to due some to impuritiessome impurities of the of chromophores the chromophores or aggregate or aggregate in the in support the support layers. layers. These These signals signals can canbe relatedbe related to both to both the the pigments pigments and and the the wall wall support. support. Traces Traces of titaniumof titanium (Ti) (Ti) were were recognized recognized in P2in and P6 indicating the possible use of titanium dioxide as white pigment in some areas of pictorial repainting. The high signals of iron (Fe) present in most of the spectra indicate the extensive use of iron oxide-based pigments, ochres, or earths which can be in several colors and are not expensive. Some considerations about the nature of the pigments are presented and discussed for each color. Black and grey: Due to the absence of signals in the XRF, black (P12) pigment can be probably constituted by carbon black which is not identifiable by the XRF technique. The grey (P20) was obtained by mixing the black pigment with an appropriate amount of calcium carbonate to obtain the desired tonality. The vest (points P2 and P34) was realized by using white lead. Yellows: The signals of iron (Fe) are present in both XRF spectra of the P10 and P15 indicating the use of yellow ochres. The low counts of the chromium signal could be attributable to the presence of the modern pigment chrome yellow (PbCrO4) used for little retouching during the previous restorations. The signals of lead (Pb) and tin (Sn) are present in the P4 spectrum of the sleeve indicating the presence of the lead-tin yellow (Pb2SnO4 or PbSnO3). This yellow pictorial area is characterized by a totally different IR spectral behavior (light grey tone) with respect the previous analyzed area (P10 and P15) which returned a greater absorption in IR (dark grey tones), as shown in Figures6 and7. This evidence Heritage 2019, 2 2380 confirms the advantages of the integrated use of multispectral imaging and spectroscopic techniques, extending the result of chemical identification on a single point to extensive areas characterized by the same spectral behaviors. Furthermore, the obtained results underline the variety of the artist’s palette (more materials, such as yellow ochres and lead-tin yellow for the yellow layers) to obtain different tonal effects for the same color. Finally, the small peak of gold (Au) in P4 and P10 spectra, can be related to some decoration in the drapery, confirming the preciousness of the fresco and providing important evidence of the original feature. Green: The signals of chromium (Cr) in the XRF spectra of the P5, P7, P16, and P18 indicate the use of chromium (III) oxide, used for pictorial integration as confirmed by multispectral findings. However, the signal of phosphorus (P) is present in the spectra P7, P16, and P18. Generally, the presence of this element in fresco surfaces is attributable to the trace of the organic chemical compounds applied over time as protective identification. A further phase of the research project will involve a GS-MS analysis to exactly identify the original or restoration materials overlapped on these precious surfaces. The high Cu signal intensity revealed that the original dark green layers (P7), according to the spectral behavior in FCIR as shown in Figure5, is identifiable as malachite [Cu 2(CO3) (OH)2] or verdigris [copper acetate] [11–13], not unambiguously discriminable through the non-destructive techniques used to date for the study of the here investigated fresco. Brown: The presence of iron (Fe) and manganese (Mn) in the P8 spectrum indicative of umber (oxide Fe O H O hydrated iron with manganese dioxide MnO ) and clay silicates like kaolin (Al O 2 3· 2 2 2 3· 2SiO 2H O) or burnt umber (anhydrous iron oxide Fe O and manganese dioxide MnO ). 2· 2 2 3 2 Flesh tone: The intensive XRF signals of calcium (Ca), iron (Fe), and strontium (Sr) in P3, P13, P17, and P33 spectra indicate a combination of calcium carbonate and red ochres to reach the desiderated tonality. Blue: The XRF spectra of the P9 and P14 show the signals of copper (Cu). In addition, the pseudo-reflectance spectrum reported in Figure 13 shows a sequence of peaks attributable to the azurite (Cu3(CO3)2(OH)2, letter A in the spectrum). In reflection mode, the strong absorption due to 1 the carbonate antisymmetric stretching (ν3) in the region 1450–1420 cm− is inverted by reststrahlen effects. In particular, azurite, in a different way in respect to other types of carbonates like cerussite and malachite, presents a fine structure of two inverted peaks for this vibrational mode. The second fine 1 structure occurs around 2800–2300 cm− , whose possible attribution is due to coupling of ν1 + ν3 or 1 2ν2 + ν4 modes. Moreover, the three distinct contributions between 2500 and 2600 cm− and the third 1 fine structure at a higher wavenumber, in the NIR range between 4200 and 4400 cm− , referable to both combination ν + δ (OH) and overtone 3ν3, are characteristic of azurite [18]. The high signals of iron (Fe) together with the small signals of manganese (Mn), silicon (Si), and sulfur (S) in the XRF spectrum indicate the presence of primer realized “on bolo” used for the spreading of the azurite. The low signals of chromium (Cr) and barium (Ba), in the spectrum can be attributed to some repainting made on the surface layer to give tonal homogeneity to the original blue layer. The presence of a veiling retouching layer was evidenced by the FCIR which returned a reddish spectral response superimposed on the typical azurite response in false color infrared, as shown in Figure7. Red: The investigation about red colors has been performed mainly in the red hat at the bottom figure where some blackening is present, and the reflectometry revealed some hidden decorations, as shown in Figure8. The signals of iron (Fe) are present indicating the use of red ochre. The signals of lead (Pb) are also present in the P22–P40 spectra indicating the presence of lead oxide, generally known as minium. However, the signals of mercury (Hg) are present in the P24, P25, P26, P27, and P35 spectra. Results are in agreement with the IR reflectography by which the transparency of cinnabar at 1000 nm highlights the lost pendants (both at left and right). The blackening of red can be explained as the well know degradation of cinnabar [19,20]. The P27 spectrum revealed the presence of zinc (Zn) and chromium (Cr) attributable to zinc white (ZnO) and chrome red (PbCrO4 PbO), respectively, · added to surface during the pictorial retouches of this area. Heritage 2019, 2 2381

4. Conclusions In this paper, the results of an in situ diagnostic campaign carried out on a few small areas of the famous detached painting “Trionfo della morte” saved at the Gallery of Art for the Sicilian region Abatellis Palace in Palermo (Italy) are reported. Thanks to the non-invasive investigations, such as multispectral analysis, XRF, and IR spectroscopies carried out step-by-step with a critical and reasoned approach, it was possible to identify the nature of pigments to map the precious “a secco” details, to obtain some indications about: 1) The conservation state, in particular linked to the thinning (“a secco” layers) or chromatic alteration (blackening of cinnabar red layers) of some pictorial details which now prevents the appreciating of some iconographies; 2) The areas of previous restorations, in particular the integrations performed mimetically and therefore not directly distinguishable to the naked eye; 3) Some interesting hidden details, appreciable only through single spectral bands in the visible (450 nm, for on surface “a secco” layers) or in the infrared (1000 nm, for underlying layers) spectral range. The identified original pigments were red, brown, red and yellow ochres, umber, cinnabar, azurite (overlapped on iron oxide-based pigment), green earth, minium, lead-tin yellow, lead white, and gold. In particular, a systematic XRF mapping of the trace of gilding could give back the preciousness of the original appearance of the decorations. Some of the several hidden details discovered in the two figures of the women and along the profile of the bishop’s face as well as in the red hat of a cardinal are very interesting and constitute additional evidence of the preciousness of the fresco, rich in details despite the fact that for a long time it was placed outside. Nevertheless, the disappearing of these details is due to time and to the nature of materials used together during the various restorations. Moreover, the possibility of retracing the conservative history through the localization of the different pictorial integration methods permits the highlighting of the complex history of the past restorations. It has to be underlined that only a few small areas were investigated. Even though the total surface of the investigated areas represents a small part of the whole fresco, this study gives a significant contribution to the deepening knowledge of this masterpiece, and it can be useful for an informed conservation and restoration. The finds confirmed the use of a typical pigment palette available in the XV century and also of the technical devices generally used by fresco painters from other areas of Italy. A complete diagnostic study will allow the widening of knowledge and to establish more precise comparisons based on objective data and used raw materials with contemporary schools not only in the Sicilian regional territory but also in the central Italian ones. In this way, a comparative archaeometric study could propose solutions for identifying the local or foreign artist, currently unknown, that in the XV century realized in Palermo, that this work famous all over the world. Moreover, this study can represent a very effective example to demonstrate how useful and productive a multi-analytical scientific approach by using complementary spectroscopic techniques can be. In fact, in some cases, the nature of each pigment can be defined by only integrating the results coming from each technique. In addition, it aims to the dissemination and valorization of cultural heritage and, in our opinion, it constitutes a base and an incentive to perform a complete scientific study on the whole fresco. NOTE. The “Trionfo della morte” has been focused on by a historical, artistic, as well as a scientific point of view, during the exhibition Arte è Scienza 2017 promoted by AIAr [21], which involved simultaneously 13 sites in Italy. This work of art has proven to represent a very effective example for demonstrating how a multidisciplinary approach aimed at the dissemination and valorization of cultural heritage can be useful and productive. The visitor participated in a (always) new discovery that provided the interpretation of the analytical data, accompanying them in an experience that is very different from the traditional museum visit. Heritage 2019, 2 2382

Author Contributions: Methodology, M.F.A. and M.L.S.; data analysis, S.S., V.R. and F.A.; data curation, S.S. and F.A.; writing—review and editing, M.L.S. and M.F.A.; supervision, E.C., C.G. Funding: This work is part of the project “Development and Application of Innovative Materials and processes for the diagnosis and restoration of Cultural Heritage—DELIAS”—PON03PE 00214 2 (Programma Operativo Nazionale Ricerca e Competitività 2007–2013). Acknowledgments: Thanks to the Director of Galleria Interdisciplinare Regionale della Sicilia di Palazzo Abatellis, Palermo (Italy) for having allowed to carry out the investigations. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Abbate, V.; Cordaro, E.M. Il Trionfo della morte di Palermo. In L’opera, le vicende conservative, il restauro; Sellerio Editore: Palermo, Italy, 1989; ISBN 10: 8876810404. 2. Luciani, G.; Pelosi, C.; Agresti, G.; Monaco, A.L. How to reveal the invisible the fundamental role of diagnostics for religious painting investigation. Eur. J. Sci. Theol. 2019, 15, 209–220. 3. Mazzeo, R. Analytical Chemistry for Cultural Heritage; Springer: Berlin, Germany, 2017; ISSN 2367-4067. 4. Nardo, V.M.; Renda, V.; Anastasio, G.; Caponetti, E.; Saladino, M.L.; Vasi, C.S.; Ponterio, R.C. A combination of portable non-invasive techniques to study on reverse glass paintings at Mistretta museum. Microchem. J. 2019, 146, 640–644. [CrossRef] 5. Saladino, M.L.; Ridolfi, S.; Carocci, I.; Martino, D.C.; Lombardo, R.; Spinella, A.; Traina, G.; Caponetti, E. A multianalytical non-invasive and micro-invasive approach to canvas oil paintings. General considerations from a specific case. Microchem. J. 2017, 133, 607–613. [CrossRef] 6. Saladino, M.L.; Ridolfi, S.; Carocci, I.; Chirco, G.; Caramanna, S.; Caponetti, E. A multi-disciplinary investigation of the “Tavolette fuori posto” of the main hall wooden ceiling of the “Steri” (Palermo, Italy). Microchem. J. 2016, 126, 132–137. [CrossRef] 7. Giorgi, L.; Nevin, A.; Comelli, D.; Frizzi, T.; Alberti, R.; Zendri, E.; Piccolo, M.; Izzo, F.C. In-Situ technical study of modern paintings—Part 2: Imaging and spectroscopic analysis of zinc white in paintings from 1889 to 1940 by Alessandro Milesi (1856–1945). Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 2019, 219, 504–508. [CrossRef][PubMed] 8. Izzo, F.C.; Capogrosso, V.; Gironda, M.; Alberti, R.; Mazzei, C.; Nodari, L.; Gambirasi, A.; Zendri, E.; Nevin, A. Multianalytica non-invasive study of modern yellow paints from postwar italian paintings from the International Gallery of Modern Art Cà Pesaro, Venice. X-ray Spectrom. 2015, 44, 296–304. [CrossRef] 9. Frosinini, C. Progetto Giotto. In Tecnica artistica e stato di conservazione delle pitture murali nelle Cappelle Peruzzi e Bardi a Santa Croce; Edifir Edizioni: Firenze, Italy, 2017. 10. Renda, V.; Saladino, M.L.; Caramanna, S.; Chirco, G.; Castello, A.F.; Conti, C.; Marco, A.; Sali, D.; Caponetti, E. Investigation on a low environmental impact solvent mixture applied to a wooden painted slab. Int. J. Conserv. Sci. 2016, 7, 789–796. 11. Cosentino, A. Identification of pigments by multispectral imaging; a flowchart method. Herit. Sci. 2014, 2, 8. [CrossRef] 12. Moon, T.; Schilling, M.R.; Thirkettle, S. A note on the use of false-color infrared photography in conservation. Stud. Conserv. 1992, 37, 42–52. [CrossRef] 13. Mairinger, F. The Infrared Examination of Paintings. In Radiation in Art and archeometry; Creagh, D.C., Bradley, D.A., Eds.; Elsevier: Amsterdam, The Netherland, 2000; pp. 40–55. 14. De La Rie, E.R. Fluorescence of paint and varnish layers (part I). Stud. Conserv. 1982, 27, 1–7. 15. De La Rie, E.R. Fluorescence of paint and varnish layers (part III). Stud. Conserv. 1982, 27, 102–108. 16. Rosell, L.; Ortí, F.; Kasprzyk, A.; Playà, E.; Peryt, T.M. Strontium geochemistry of Miocene primary gypsum: Messinian of Southeastern Spain and Sicily and Badenian of Poland. J. Sediment. Res. 1998, 68, 63–79. [CrossRef] 17. Franceschi, E.; Locardi, F. Strontium, a new marker of the origin of gypsum in cultural heritage? J. Cult. Herit. 2014, 15, 522–527. [CrossRef] 18. Miliani, C.; Rosi, F.; Daveri, A.; Brunetti, B.G. Reflection infrared spectroscopy for the non-invasive in situ study of artists’ pigments. Appl. Phys. A 2012, 106, 295–307. [CrossRef] 19. McCormack, J.K. The darkening of cinnabar in sunlight. Miner. Depos. 2000, 35, 796–798. [CrossRef] Heritage 2019, 2 2383

20. Radepont, M.; Coquinot, Y.; Janssens, K.; Ezrati, J.-J.; de Nolf, W.; Cotte, M. Thermodynamic and experimental study of the degradation of the red pigment mercury sulﬁde. J. Anal. At. Spectrom. 2015, 30, 599. [CrossRef] 21. Available online: http://www.associazioneaiar.com (accessed on 6 August 2019).

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).