ACCEPTED MANUSCRIPT
ACCEPTEIPT
Hype spec al imaging applied o he analysis o Goya
pain ings in he Museum o Za agoza (Spain)
F. Daniel(1), A. Mounie (1), J. Pé ez-A an egui(2), C. Pa dos(2), N. P ie o-Taboada(3),
S. Fdez-O iz de Vallejuelo(3) and K. Cas o(3)
(1) IRAMAT-CRP2A UMR 5060 CNRS-Uni e si é Bo deaux Mon aigne. 33607 Pessac (F ance)
(2) Ins i u o Uni e si a io de in es igación en Ciencias Ambien ales de A agón (IUCA) and Dep. Química
Analí ica. Uni e sidad de Za agoza. 50009 Za agoza (Spain)
(3) Depa men o Analy ical Chemis y, Facul y o Science and Technology, Uni e si y o he Basque Coun y
(UPV/EHU). 48940 Leioa (Spain)
Abs ac : In ecen yea s, esea ch ca ied ou mos ly on pain ings has p o ed ha imaging
spec oscopy echniques can be used e ec i ely o ma e ial iden i ica ion and mapping on
wo ks o a . Rega ding he analysis o pain ings, he hype spec al imaging p o ides he
associa ed e lec ance spec um o e e y poin o he image, which can be compa ed wi h
hose o a e e ence da abase. In his wo k, we e alua ed he analy ical sui abili y o wo
di e en hype spec al imaging sys ems o Goya’s pain ings kep in he Museum o Za agoza
(Spain). The i s one used a "push-b oom" sys em, whe eas he second one was a "mi o -
scanning" sys em o s udy educed a eas. P ocessing he hype spec al imaging spec a
p o ided alse colou in a ed images ha showed es o ed zones o he pain ing and helped o
guide he choice o he pigmen analysis o signi ican zones o he o iginal pain ing. Thei
di e se na u e, and he p esence o complex mix u es, a nishes and o ganic binde s, made
he analysis o he pigmen s di icul , hus, he con i ma ion by po able Raman Spec oscopy
and hand-held X- ay luo escence was needed. Finally, hanks o he good esponse o he
s udied hype spec al sys ems, Goya’s pain ings we e mapped and he e o e he main
pigmen s used we e iden i ied by imaging e lec ion spec oscopy. In his sense, he
a ailabili y o a ela i ely quick p ocedu e would allow a la ge numbe o Goya’s pain ings
in he museum o be su eyed, inc easing he possibili y o ob aining signi ican esul s and
p o iding a chance o ex ensi e compa isons, which a e ele an om he poin o iew o
a his o y issues.
Keywo ds: hype spec al imaging, Raman, EDXRF, in-si u, non-in asi e, pigmen s
This is he accep ed manusc ip o he a icle ha appea ed in inal o m in Mic ochemical Jou nal 126 : 113-120 (2016)), which
has been published in inal o m a h ps://doi.o g/10.1016/j.mic oc.2015.11.044. © 2015 Else ie unde CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/)
ACCEPTED MANUSCRIPT
1. In oduc ion
A s udies and he conse a ion o pain ings equi es knowledge o he ma e ials used, such
as colo an s, binde s, and p epa a o y laye s. This in o ma ion can also p o ide insigh in o
he a is 's wo king me hods. In his sense, in ecen yea s, esea ch ca ied ou mos ly on
pain ings has p o ed ha imaging spec oscopy echniques can be used e ec i ely o
ma e ial iden i ica ion and mapping on wo ks o a [1]. Mo eo e , he ecen ad ances in he
spec o-imaging ield [2], as well as he demand o non-in asi e echniques o he s udy o
cul u al he i age ma e ials, o e a a ou able en i onmen o he de elopmen o hese
me hods.
The e m "hype spec al imaging" was c ea ed by Alexande Goe z in 1985 [3]. I can be
conside ed as he combina ion o digi al imaging wi h e lec ance spec oscopy (applied
success ully o conse a ion h ough he use o UV–Vis–NIR spec opho ome e s) and ib e
op ic e lec ance spec oscopy (FORS). Recen ly, a s a e o he a su ey [2] conce ning
hype spec al imaging ed ew he his o y o hese echniques and applica ions. Spec al
imaging sys ems p o ide a powe ul ool o non-in asi e, non-con ac iden i ica ion and
cha ac e iza ion o pigmen s, inks, subs a es and ea men s o a e ac s, allowing comple ely
non-des uc i e analyses o esea ch and p ese a ion in a chaeology [4], can as pain ings
[1,5], illumina ions [6,7,8], wall pain ings [9], d awings [10], a chi es [11], manusc ip s [12]
and o conse a ion/ es o a ion pu poses [13,14].
As ega ds ed-g een-blue image y (RGB), in a ed alse-colou (IRFC) and ul a iole (UV)
luo escence image y, HSI expand no only he ange o possible applica ions: iden i ying
pigmen s, moni o ing colou changes, colou ma ching, bu also he documen a ion o IR
e lec ance and induced UV luo escence. Image y and e lec ance spec a in he IR ange
ACCEPTED MANUSCRIPT
allows he iden i ica ion o pigmen s, binde s and di e se ma e ials [15,16]. The p ope ies o
IR imaging we e used, o example, o ind he p epa a o y d awing and o iden i y pigmen s
[10]. The combina ion o HSI and spec o luo escence is an impo an ool o he non
in asi e s udy o o ganic ma e ials [17]. Indeed, since he 1920s, pho og aphic documen a ion
o luo escence emi ed unde UV exposu e has been an impo an diagnos ic me hod o
s udying his o ical and a is ic objec s. Since many ino ganic and o ganic subs ances exhibi a
cha ac e is ic emi ed luo escence unde UV illumina ion, luo escence spec oscopy,
heo e ically, can also be applied o di e en ia e ma e ials such as esins, pigmen s and
o ganic binde s [18].
Fu he mo e, spec al imaging has he ad an age o being a apid and ela i ely low-cos
solu ion o he examina ion o la ge a eas. I is s ill oo li le used o he s udy o easel
pain ings. This is because some limi a ions o he me hod make i s applica ion di icul : 1)
when he componen s o a pain ing (pigmen , binde s) a e al e ed, 2) i he unde lying
ma e ials o complex mix u es o pigmen s cause spec al in e e ence, 3) when specula
e lec ance occu s due o a nish, 4) when ligh s a e placed a an unsui able angle, and 5) i
he e e ence spec al lib a y is incomple e [19]. In spi e o hese di icul ies, HSI is an
exci ing new ield o a chaeome y and conse a ion science and he inc ease in new
de elopmen s o s udies in cul u al he i age objec s con ibu e o i s imp o emen . Mo eo e ,
in he pa icula cases men ioned, he use o complemen a y echniques sol es he p oblems
ound. In his sense, po able and hand-held Raman spec oscopy has demons a ed i s
e sa ili y and adap abili y, as well as a e y p omising pe o mance, as can be seen in
scien i ic s udies ca ied ou on ock pain ings in ca es [20,21], mu al pain ings [22] o
aluable collec ions in museums [23]. Toge he wi h his, po able X- ay luo escence
ACCEPTED MANUSCRIPT
spec oscopy (EDXRF) has become a e y aluable echnique in he ield o cul u al he i age
[24,25], o example, in he s udies o po e y, ca e a , b onzes, wall pain ings, e c.
Taking all o hese in o accoun , o s udy he pigmen s used by Goya on can as be ween 1758
and 1790 inside an on-going p ojec , wo di e en po able HSI sys ems based on e lec ion
spec oscopy we e conside ed. Thus, he aim o his wo k was o e alua e he analy ical
sui abili y o hype spec al imaging o iden i ying o iginal pigmen s, ma e ials added in
es o a ions, and he pain e 's echnical s yle.
2. Ma e ials and me hods
2.1. Goya pain ings
F ancisco de Goya y Lucien es (Fuende odos, Za agoza, Spain, 1746 – Bo deaux, F ance,
1828) was an a is con inually aining and p og essing. Consequen ly, he ied di e en
gen es, and he ook an in e es in he di e en s yles coexis ing in he pe iod when he was
wo king in o de o c ea e his own s yle [26,27]. He was also in e es ed in e y di e se
echniques and o ma s: mu al pain ing, oil pain ing wi h di e en dimensions, eng a ing
using a ious echniques, e c. He was a mas e o oil pain ing, c ea ing inc edible e ec s o
ligh , and also o he use o colou ed backg ounds (o en eddish), ha he could lea e isible
o anspa en .
The i s pain ing s udied, 'Po ai o a young man wi h ha ', belongs o he collec ion o he
Royal Academy o Fine A s o San Luis, Za agoza. I appea s egis e ed in he ca alogue o
1828, as 'Po ai o F ancisco Bayeu by his own hand' [28]. Al hough, he mos ecen
s udies on he wo ks by F ancisco Bayeu (Za agoza, 1734 - Mad id, 1795) p e e o conside
his pain ing as a Bayeu sel -po ai , pain ed a ound 1758 [29, 30], he esea che s a ibu ed
ACCEPTED MANUSCRIPT
i o F ancisco de Goya, because o i s ine quali y and because bo h a is s we e e y close.
Bayeu was Goya's b o he -in-law and also his eache be ween 1759 and 1763. The a is ’s
wo k is a ske ch, a p elimina y s udy whe e he p epa a ion backg ound is s ill isible, a ed
laye made wi h och e, which is e y cha ac e is ic o he A agonese pain e s a ha ime, and
which is usually p esen in many o Goya’s pain ings. I is a e y p ecise d awing; he pain e
used he chia oscu o, lea ing pa o he ace in he shadow, a e y exp essi e and na u al
isage. The man wea s a ha on a headsca ied in he A agonese s yle.
The second pain ing s udied was 'Po ai o Juan Ma in de Goicoechea y Gala za' [31].
Goya pain ed his po ai in 1790. He p obably d ew some ske ches om li e in Za agoza and
he inished i la e in Mad id, whe e he li ed a ha ime. Juan Ma in was a iend o Goya's,
an Enligh enmen man, ade and businessman. In 1789 he was awa ded he C oss o he
Royal O de o Cha les III, and his was p obably he eason why he asked his iend Goya o
he po ai . On a black backg ound, we can see his uppe -body image, in h ee-qua e
posi ion, and looking ahead. He wea s a b own d ess coa whe e he medal is ou lined. The
colou s o he ace and he igh hand eally s and ou , and he ex u es o he d ess-coa el e
and o he lace in he colla and cu we e e y well done.
2.2. Hype spec al imaging
The hype spec al imaging (HSI) was done on he Goya pain ings wi h wo di e en HSI
sys ems. Rega ding he analysis o pain ings, o e e y image poin he HSI supplies he
associa ed e lec ance spec um, which can be compa ed wi h hose o a e e ence da abase. In
a hype spec al came a, he ligh e lec ed om one line segmen o he a ge sample eaches
he en ance sli in he spec og aph. This ligh is dispe sed in o di e en wa eleng hs by he
dispe sing elemen in he spec og aph. This dispe sed ligh is cap u ed by he CCD senso as
ACCEPTED MANUSCRIPT
a wo dimensional image; he spa ial posi ion and he spec al dimension. Mo ing he a ge
objec o he spec al came a in a di ec ion pe pendicula o a measu ed line, he HSI sys em
can cap u e he images o di e en lines on he a ge a ea. By synch onizing he mo emen
be ween came a and a ge objec and he acquisi ion o he came a, one can cap u e di e en
lines o he objec and, hus, gene a e he da acube wi h he second spa ial dimension. The
acquisi ion o e lec ance spec a equi es a scanning sys em and o his wo me hods a e
possible.
The i s one (called HSI-sys em 1 in his pape ) is a "push-b oom" sys em whe e he
scanning is ob ained by means o a ansla ion ail. The hype spec al came a is placed on a
ansla ion ail, ixed on wo ipods (Fig. 1). The came a can be placed e ically o
ho izon ally. The sys em can be used in ixed con igu a ion o an applica ion in he
labo a o y o in mobile con igu a ion o in-si u s udies in museums o mu al pain ings. The
hype spec al CCD came a (HS-XX-V10E), de eloped by SPECIM (Finland), has a 1600 x
840 pixel esolu ion, a spec al esolu ion o 2.8 nm and a wa eleng h ange be ween 400 o
1000 nm. The ocal leng h is 23 mm. The ail allowing he ho izon al displacemen o he
came a sys em is 1.30 m long and is con olled by he IDAQ so wa e. The pain ings we e
illumina ed by wo halogen lamps a 45°, 60 cm away om he objec . Fo he s udy o he
pain ings, he expe imen al condi ions a e: wo king dis ance o 114 cm, scan speed o 13
mm/s and exposu e ime ( ime o which he CCD a ay is exposed o ligh ) o 25 ms. This
“exposu e ime” is a cha ac e is ic o he CCD senso . I e e s o he speci ied ime in e al
whe e he s o ed cha ge is accumula ed wi hin each CCD pho odiode. The du a ion o a scan
depends o he size o he a wo k. Fo he Goya pain ings he du a ion o he scan a ied
om 30” o 1 minu e. A he speci ied wo king dis ance, he halogen ligh ing deli e s 1000
lux on o he su ace o he pain ing, so he illumina ion le el was below he 50 lux.h
ACCEPTED MANUSCRIPT
ecommended o agile objec s [32]. Dec easing he illumina ion by doing binning (adjacen
pixels o be combined) would lead o a dec ease o he esolu ion. Spec al IDAQ so wa e
p o ides da a acquisi ion, s o age and wa eleng h calib a ion.
The second sys em (“whiskb oom”, called HSI-sys em 2 in his pape ) uses "a mi o -
scanning" echnique mo e adap ed o he s udy o educed a eas. The came a, ligh s and
so wa e a e he same as desc ibed abo e. The hype spec al imaging VNIR sys em
(SPECIM, Finland) is composed o he V10E imSpec o spec og aph (CCD came a, 30 ps,
1600 x 1200 pixels, 400-1000 nm) and he mi o scanne (op imised o 380 o 1700 nm,
ield o iew: 80 x 30 deg ees –depending on he o e objec i e lens–, and scan speed be ween
0.01 and 25 deg ee/s). The spec al came a has a sli o 30 μm (spec al esolu ion 2.8 nm,
nume ical ape u e F/2.4), and i uses a Xenoplan 1.4/17 lens (Schneide , K euznach,
Ge many) wi h a ocal leng h o 17.6 mm. The scan speed o he mi o scanne was also
op imized in o de o ob ain good esol ed images bu a oiding long- ime exposu es o he
halogen lamps. Scan angles o 25 and 45 deg ees we e used, depending on he size o he
pain ing.
Wi h bo h HSI sys ems, calib a ed di use e lec ance a ge s (Spec alon, by Labsphe e,
USA) we e used o calib a e he esul ing spec a. The acquisi ion o he da a was ela i ely
as (one minu e). The ea men o he da acube was pe o med wi h ENVI 5.1 + IDL
so wa e. The ea men o he da acube gi es spec al images and/o iden i ica ion o he
pigmen s by compa ison o he e lec ance spec a. One o he main p oblems wi h pigmen
iden i ica ion using spec al e lec ance is he lack o da abases o e e ence pigmen and
pain . We ha e ou own e e ence pigmen da abase (>150 e e ence pigmen s) mos ly
cha ac e is ic o medie al imes (bu no only) as ou spec al lib a y was buil p incipally o
ACCEPTED MANUSCRIPT
he applica ions o HSI on illumina ions [9]. Howe e , we ha e o poin ou ha mos o hese
pigmen s we e used in Goya’s ime and we ha e added mode n pigmen s o ou da abase. In
addi ion, he CNR-IFAC open-access on-line da abase o e lec ance spec a is a ailable, wi h
he 270 nm-1700 nm spec a o a ious pigmen s in common binding media ound in Wes e n
Eu opean pain ings [33].
2.2.1. RGB and IRFC images
Hype spec al imaging allows us o ob ain images in he isible and nea in a ed (NIR)
anges. The HSI sys ems can also iden i y pigmen s and map hei dis ibu ion in a pain ing
by p ocessing he e lec ance spec a. In mul ispec al imaging he RGB and IRFC use
b oadband il e s. In Hype spec al imaging, speci ic wa eleng hs a e chosen o achie e his.
A na u al colou image is ob ained by aking h ee isible bands. Bo h HSI sys ems gene a ed
da a which could be ea ed by selec ing h ee wa eleng hs o ob ain a ed-g een-blue (RGB)
image (R: 670 nm; G: 550 nm; B: 448 nm) om a ma ix o e lec ance spec a (400 o 1000
nm) associa ed wi h each pixel o he image. IRFC is an a bi a y cons uc ion which in he
same analy ical condi ions allows us o compa e a pain ing o e e ence pigmen s. The e a e a
a ie y o ways o combine wo o mo e spec al bands in o a pseudo colou image by
selec ing some o he o he bands (IR, UV, e c.) and assigning hem o one o he isible ligh
colou bands ( ed, g een, o blue). The e o e, an in a ed alse-colou (IRFC) image can also
be displayed by selec ing an in a ed band ins ead o he ed band, and selec ing wo o he s in
he isible ange o he alse g een and blue bands (R: 900 nm; G: 650 nm; B: 550 nm). The
IRFC image ypically dis inguishes di e en pigmen s ha seem he same o he naked eye.
Since he 1960’s, in a ed e lec og aphy is a echnique ha has been used o look h ough he
pain laye s. When he longe wa eleng hs o in a ed adia ion pene a e he pain laye s, he
uppe laye s appea anspa en . The deg ee o pene a ion depends on he hickness o he
ACCEPTED MANUSCRIPT
pain , he ype o pain used and he leng h o he wa e o in a ed adia ion. The con as in
he abso p ion o a ious ma e ials e eals laye s o he pain ing no isible o he naked eye,
such as he unde d awings and changes in he pain laye s. Fo example, he compa ison o
he IRFC image o he e e ence pigmen s wi h hose o he pain ing allows us o hypo hesize
abou he na u e o he pigmen s [34]. Some imes, mo e ypes o image y, such as UV
luo escence [35,36], can help o emo e unce ain ies abou he in e p e a ion o IRFC
images. A be e iden i ica ion o mos o he pain ing componen s is done by compa ison o
he e lec ance spec a wi h ib e op ics e lec ance spec ome y (FORS) da abases [33,37]
which allows he pigmen mapping (dis ibu ion o he endmembe s on he su ace o he
pain ing).
2.2.2. Pigmen mapping
By a p ocedu e named “classi ica ion”, he so wa e compa es he e lec ance o he e e ence
spec a o he da abase wi h hose (endmembe s) ob ained om he pain ing. The e a e
di e en ways o ge he spa ial dis ibu ion o he endmembe s/pigmen s. This can be done
by algo i hms. One o he mos common spec al mapping me hods is Spec al Angle Mappe
(SAM) [38]. This me hod conside s each spec um as a ec o in an n-dimensional sca e plo
(o n-D space). The ma hema ical ea men compu es an angle be ween he e e ence and he
obse ed spec um. Smalle angles ep esen close ma ches o he e e ence spec um
Endmembe spec a used by SAM can come om ou spec al lib a y. The numbe o pixels
displayed as a speci ic class is a unc ion o he h eshold used o gene a e he classi ica ion.
Jus because a gi en pixel is classi ied as a speci ic pigmen does no mean i is ac ually ha
pigmen . SAM is a simila i y measu e, no an iden i ie . Tha is why poin me hods a e used
o con i m he iden i ica ion o he endmembe s mapped. In his pape , he classi ica ion o
he HSI da a spec a was done using he SAM algo i hm (angle 0.4) in o de o ob ain he
ACCEPTED MANUSCRIPT
In he po ai o “Juan Ma in de Goicoechea”, a pain ing by Goya, he ange o pigmen s
was la ge han o he “Po ai o a young man wi h ha ”. The RGB and IRFC images (Fig.
5a and b) did no show isible es o a ions.
The classi ica ion o HS da acube spec a ga e a mapping o he di e en iden i ied pigmen s
(Fig. 5c). In Fig. 5d, each colou code co esponds o a pigmen : da k ed o e milion mixed
wi h ed och e and lead whi e; pink (cheek) o e milion; ligh blue and iole o a mix u e o
a ed lake wi h e milion; da k g ey ( ibbon o he medal) and black o a non iden i ied
pigmen ; g een (hai ) o calci e.
The HSI mapping showed ha he endmembe ha maps o he man le and ace had a
e lec ance ansi ion edge a 630 nm. Ve milion (HgS) has a ansi ion edge posi ion a ound
610 nm [37]. The shi o his ansi ion edge a 630 is p obably due o a ed lake which has
ansi ion edges a highe wa eleng hs and he classi ica ion sugges ed his. The e lec ance
spec um was well desc ibed by e milion mixed wi h lead whi e. EDXRF analysis a a si e
wi hin he mapped a ea con i med he p esence o me cu y, which is consis en wi h he
e milion assignmen , and lead.
The whi e pigmen s o he pain ing we e desc ibed by he wo endmembe s ha we e ound.
The i s one was lead whi e, consis en wi h he e lec ion spec al analysis, mixed wi h
e milion. The EDXRF measu emen ound a la ge amoun o lead. The second whi e
pigmen (calci e) was localized in he hai o Juan Ma in.
P ussian blue (Fe4[Fe(CN)6]3) was used o he da ke blue o he medal ibbon and de ec ed
by Raman bands obse ed a 268 (m) and 2153 (s) cm-1 (see Fig. 3b). This pigmen , al eady
usual a ha ime, had indeed been disco e ed by 1706. Despi e hese esul s which ag eed
ACCEPTED MANUSCRIPT
wi h he EDXRF and Raman esul s, he endmembe s and associa ed map showed ha he
e lec ance spec a om he same zone did no co espond o he usual shape o P ussian blue,
since he spec al assignmen o his pigmen is made based on he low e lec ance in he ed.
The in ensi y and maximum e lec ance a ied om 440 o 460 nm, depending o he dilu ion
wi h lead whi e pigmen [49]. I is possible ha he p esence o a nish o he mix u e wi h
ano he pigmen modi ied he shape o he P ussian blue e lec ance spec a.
Ve milion and lead whi e we e also ound by Raman spec oscopy. The i s pigmen was
iden i ied mainly in he lips and cheeks a eas o he lesh colou , mixed wi h lead whi e and,
some imes, wi h calcium ca bona e (CaCO3, main Raman band a 1082 cm-1). Lead whi e was
also iden i ied in he whi e a eas o he hai and shi . Rega ding he EDXRF analysis, he
p esence o Hg in he lips and cheeks co obo a ed he iden i ica ion o e milion by Raman
spec oscopy.
In addi ion, he p esence o i on in he ed a eas o he jacke indica ed, p obably, he use o
i on oxides such as ed ea h. In he case o he black backg ound, he p esence o i on was
also no able. Howe e , in his case, a highe p esence o coppe oge he wi h calcium was
de e mined. EDXRF elemen al analysis pe o med on he jacke sugges ed he p esence o
i on ea hs and whi e pigmen s (lead whi e).
As in he “Po ai o a young man wi h ha ”, he ace was pain ed wi h a mix u e o
e milion and lead whi e, as shown by EDXRF analysis (Pb, Hg, S), in di e se p opo ions.
The hai was pain ed wi h calci e. I on was also de ec ed in he ace. Re lec ance spec a in
he same zones showed he p esence o och e.
ACCEPTED MANUSCRIPT
In his a wo k, as in he p e ious one, cadmium was iden i ied in all he measu ed poin s.
Da a analysis o he EDXRF esul s e ealed once again he co ela ion be ween lead and a
g oup o elemen s, as in he case o he “Po ai o a young man wi h ha ”, in which
cadmium was included. This g oup o elemen s was simila in bo h pain ings wi h mainly he
p esence o Cd, S , Z , Cl and S. Finally, i on and manganese we e also co ela ed in his
a wo k.
4. Conclusions
All he analysed pigmen s co esponded o hose used a he ime o Goya and his has al eady
been es ablished in p e ious publica ions [45]. Howe e , he hype spec al imaging allowed
us o de e mine he non- es o ed zones and o make hypo heses abou he na u e o he
pigmen s used by compa ing he e lec ance spec a wi h hose o ou e e ence da abase. The
wo hype spec al sys ems ga e compa able esul s, om he “Po ai o a young man wi h
ha ” (1758) almos monoch ome o he “Po ai o Juan Ma in de Goicoechea y Gala za”
(1790), whe e he pale e was widened. In bo h pain ings, he pain e used a ed p epa a o y
laye ( ed och e o bu n ea hs), which p o es he widesp ead use o his echnique among
he A agonese pain e s o ha pe iod. Mo eo e , e milion mo e o less li up o da kened
wi h lead whi e o ca bon based pigmen s, was also used in bo h po ai s. Besides hese mo e
common pigmen s, on he po ai o Juan Ma in de Goicoechea, P ussian blue was
iden i ied.
Howe e , he mix u e o pigmen s and he p esence o a nish made some in e p e a ions
di icul . Thus, addi ional me hods such as Raman spec ome y and X- ay Fluo escence we e
necessa y o con i m he hypo heses. Wi h hype spec al imaging he e a e some unce ain ies
in he in e p e a ion o he e lec ance spec a. Ano he di icul y occu s when supe icial non-
ACCEPTED MANUSCRIPT
homogenei y and composi ional complexi y o he pic o ial ma e ial make he analy ical
iden i ica ion mo e complex. Besides, some imes, he mix u e o pigmen s modi ies he shape
o he e lec ance spec a. I is clea ha new algo i hms ha e o be de eloped o he mapping
o complex mix u es o ma e ials.
Conce ning he di e ences be ween bo h HSI acili ies, he HSI-sys em 1 gi e high quali y
images allowing ele an pos ea men (image, mapping) and in e p e a ion o he da a. The
HSI-sys em 2 showed some ad an ages ela ed o mobili y (low weigh and dimensions) and
o scanning small a eas, bu achie ing uni o mi y o ligh ing was mo e di icul because he
ligh sys em had a ixed posi ion, especially when la ge a eas we e s udied.
Raman spec oscopy also p esen s some ea u es ha make he in e p e a ion o he spec a
di icul , which is he case o luo escence, and ha is p obably due o he p esence o a nish
and/o p o ec i e laye s. In addi ion, EDXRF spec a we e o ally condi ioned by he p esence
o a high amoun o lead.
In conclusion, hype spec al imaging can be applied o agile pain ings, wi hou damage, o
he loca ion and iden i ica ion o ma e ials. Mo eo e , his sys em allows he highligh ing o
he complexi y o he d awing and he ligh s and he di e si y o he pigmen s used. The
analysis ga e a global ision o he dis ibu ion o pigmen s and hei possible changes, as
well as epo ing on he s a e o he objec (losses o pic o ial ma e ial) o chemical changes.
Acknowledgemen s
This s udy was ca ied ou wi hin he amewo k o a 2013-14 p ojec inanced by he
Communau é de T a ail des Py énées - Comunidad de T abajo de los Pi ineos, wi h he
pa ne ship o h ee egions: A agon (IUCA-Uni e sidad de Za agoza), Basque Coun y
ACCEPTED MANUSCRIPT
(Uni e sidad del País Vasco, CTP2012-P10) and he Aqui aine egion (IRAMAT-CRPAA,
CNRS-Uni e si y Bo deaux Mon aigne).Thanks also o he Museum o Za agoza, and
specially o Ma isa A guis and Ca men Gallego om he Pain ing Depa men . N. P ie o-
Taboada is g a e ul o he Uni e si y o he Basque Coun y (UPV/EHU) o he pos -doc o al
con ac .
Re e ences
[1] J.K. Delaney, J.G. Zeibel, M. Thou y, R. Li le on, K.M. Mo ales, M. Palme , E.R. de la
Rie, Visible and in a ed e lec ance imaging spec oscopy o pain ings: pigmen mapping and
imp o ed in a ed e lec og aphy, P oc. SPIE, O3A: Op ics o A s, A chi ec u e, and
A chaeology II. 7391 (2009) 7391 03.
[2] H. Liang, Ad ances in mul ispec al and hype spec al imaging o a chaeology and a
conse a ion, Appl. Phys. A. 106 (2012) 309-323.
[3] J.S. MacDonald, S.L. Us in, M.E. Schaepman, The con ibu ions o D . Alexande F.H.
Goe z o imaging spec ome y, Remo e Sens. En i on. 113 (2009) S2-S4.
[4] C. Bassani, R.M. Ca alli, R. Go edo, A. Palombo, S. Pascucci, S. Pigna i, Speci ic
spec al bands o di e en land co e con ex s o imp o e he e iciency o emo e sensing
a chaeological p ospec ion: The A pi case s udy, J. Cul . He i . 10, Supplemen 1 (2009) e41-
e48.
[5] D. Lau, C. Villis, S. Fu man, M. Li e , Mul ispec al and hype spec al image analysis o
elemen al and mic o-Raman maps o c oss-sec ions om a 16 h cen u y pain ing, Anal. Chim.
Ac a. 610 (2008) 15-24.
[6] A. Mounie , G. Le Bou don, C. Aupe i , C. Belin, L. Se an , S. Laza e, Y. Le ais, F.
Daniel, Hype spec al imaging, spec o luo ime y, FORS and XRF o he non-in asi e
s udy o medie al minia u es ma e ials, He i . Sci. 2 (2014) 1-12.
[7] A. Mounie , F. Daniel, Hype spec al imaging o he s udy o wo mediae al I alian
minia u es (XIII h c.) in he MARCADE Collec ion, S ud. Conse . (2015) (in p ess).
[8] J.K. Delaney, L.D. Glinsman, M. Facini, M. Thou y, M. Palme , E.R.d.l. Rie, Use o
imaging spec oscopy, ibe op ic e lec ance spec oscopy, and X- ay luo escence o map
and iden i y pigmen s in illumina ed manusc ip s, S ud. Conse . 59 (2014) 91-101.
[9] H. Liang, A. Lucian, R. Lange, C.S. Cheung, B. Su, Remo e spec al imaging wi h
simul aneous ex ac ion o 3D opog aphy o his o ical wall pain ings, ISPRS J.
Pho og amm. Remo e Sens. 95 (2014) 13-22.
ACCEPTED MANUSCRIPT
[10] M. A as, E. Clou is, C. Collins, D. Gol z, C. Majzels, J.R. Mans ield, H.H. Man sch,
Nea -in a ed spec oscopic imaging in a conse a ion: in es iga ion o d awing
cons i uen s, J. Cul . He i . 4 (2003) 127-136.
[11] R. Padoan, M.E. Klein, G. de B uin, B.J. Aalde ink, T.A.G. S eeme s, Moni o ing Aging
P ocesses o A chi al Documen s by Means o Quan i a i e Hype spec al Imaging: A Pa o
he Hype spec al P ojec a he Na ionaal A chie (Na ional A chi es o he Ne he lands),
The Book and Pape G oup Annual. 28 (2009) 63-72.
[12] E. Ma engo, M. Man edi, O. Ze bina i, E. Robo i, E. Mazzucco, F. Gose i, G.
Bea man, F. F ance, P. Sho , De elopmen o a echnique based on mul i-spec al imaging o
moni o ing he conse a ion o cul u al he i age objec s, Anal. Chim. Ac a. 706 (2011) 229-
237.
[13] V. Papadakis, A. Loukai i, P. Pouli, A spec al imaging me hodology o de e mining on-
line he op imum cleaning le el o s onewo k, J. Cul . He i . 11 (2010) 325-328.
[14] S. Joo Kim, F. Deng, M.S. B own, Visual enhancemen o old documen s wi h
hype spec al imaging, Pa e n Recogni . 44 (2011) 1461-1469.
[15] P. Riccia di, J.K. Delaney, Combining isible and in a ed imaging spec oscopy wi h
si e speci ic, in-si u echniques o ma e ial iden i ica ion and mapping, Re is a de his o ia da
a e. 1 (2011) 253-261.
[16] J.R. Mans ield, M. A as, C. Majzels, E. Clou is, C. Collins, H.H. Man sch, Nea in a ed
spec oscopic e lec ance imaging: a new ool in a conse a ion, Vib. Spec osc. 28 (2002)
59-66.
[17] M. Romani, F. Colao, R. Fan oni, M. Guiso, M.L. San a elli, Hype spec al Fluo escence
o O ganic Pigmen Cha ac e iza ion in Con empo a y A wo k, J. Appl. Lase Spec osc. 1
(2014) 29-36.
[18] A. Mounie , C. Belin, F. Daniel, Spec o luo ime ic s udy o he ageing o mix ions used
in he gildings o mediae al wall pain ings, En i on. Sci. Pollu . Res. 18 (2011) 772-782.
[19] M. Kubik, Chap e 5 Hype spec al Imaging: A New Technique o he Non-In asi e
S udy o A wo ks, in: D. C eagh, D. B adley (Eds.), Physical Techniques in he S udy o A ,
A chaeology and Cul u al He i age, 1s Edi ion ed., Else ie Science, The Ne he lands, 2007,
pp. 199-259.
[20] A. He nanz, J.M. Ga i a-Vallejo, J.F. Ruiz-López, S. Ma in, Á. Ma o o-Valien e, R. de
Balbín-Beh mann, M. Menéndez, J.J. Alcolea-González, Spec oscopy o Palaeoli hic ock
pain ings om he Ti o Bus illo and El Buxu Ca es, As u ias, Spain, J. Raman Spec osc. 43
(2012) 1644-1650.
[21] M. Oli a es, K. Cas o, M.S. Co chón, D. Gá a e, X. Mu elaga, A. Sa mien o, N.
E xeba ia, Non-in asi e po able ins umen a ion o s udy Palaeoli hic ock pain ings: he
case o La Peña Ca e in San Roman de Candamo (As u ias, Spain), J. A cheol. Sci. 40 (2013)
1354-1360.
ACCEPTED MANUSCRIPT
[22] M. Magu egui, U. Knuu inen, I. Ma ínez-A ka azo, A. Giakoumaki, K. Cas o, J.M.
Mada iaga, Field Raman analysis o diagnose he conse a ion s a e o exca a ed walls and
wall pain ings in he a chaeological si e o Pompeii (I aly), J. Raman Spec osc. 43 (2012)
1747-1753.
[23] P. Colomban, A. Tou nié, M.C. Caggiani, C. Pa is, Pigmen s and enamelling/gilding
echnology o Mamluk mosque lamps and bo le, J. Raman Spec osc. 43 (2012) 1975-1984.
[24] G. Simsek, F. Casadio, P. Colomban, L. Bello -Gu le , K.T. Fabe , G. Zelleke, V.
Milande, E. Moine , On-Si e Iden i ica ion o Ea ly BÖTTGER Red S onewa e Made a
Meissen Using Po able XRF: 1, Body Analysis, J. Am. Ce am. Soc. 97 (2014) 2745-2754.
[25] L. Beck, H. Rousseliè e, J. Cas aing, A. Du an, M. Lebon, B. Moigna d, F. Plassa d,
Fi s use o po able sys em coupling X- ay di ac ion and X- ay luo escence o in-si u
analysis o p ehis o ic ock a , Talan a. 129 (2014) 459-464.
[26] V. Bozal, Goya, 1ºEd. ed., An onio Machado, Mad id, 2011.
[27] N. Glendinning, A e, Ideología y O iginalidad En La Ob a De Goya, 1ºEd. ed.,
Ediciones Uni e sidad de Salamanca, Salamanca, 2007.
[28] F. To alba So iano, A. Ansón Na a o, Realidad e Imagen: Goya, 1746-1828 : Museo
De Za agoza, 3 De Oc ub e-1 De Diciemb e, 1996, ELECTA, Mad id, 1996.
[29] A. Ansón Na a o, I. Gu ié ez Pas o , J.M. de la Mano, F ancisco Bayeu y Sus
Discípulos, Cajalón, Za agoza, 2007.
[30] A. Ansón Na a o, Los Bayeu, una amilia de a is as de la ilus ación, Za agoza, Caja
Inmaculada, 2012.
[31] F. Bel án Llo is, G.M. Bo ás Gualis, La Memo ia De Goya (1828-1978), Gobie no de
A agón, Za agoza, 2008.
[32] Michalski, S., The Ligh ing Decision. Fab ic o an Exhibi ion, P ep in s o Tex ile
Symposium 97. O awa: Canadian Conse a ion Ins i u e, 1997, pp. 97-104.
[33] Ins i u e o Applied Physics (IFAC), Fibe Op ics Re lec ance Spec a (FORS) o
Pic o ial Ma e ials in he 270-1700 nm ange, h p:// o s.i ac.cn .i /index.php (2011).
[34] T. Moon, M.R. Schilling, S. Thi ke le, A no e on he use o alse-colou in a ed
pho og aphy in conse a ion. S ud. Conse . 37 (1992) 42-52.
[35] A. Mounie , S. Laza e, F. Daniel, A new po able LEDµSF de ice o agile a wo ks
analyses: applica ions on medie al pigmen s, Techna 2015: Non-des uc i e and
mic oanaly ical echniques in a and cul u al he i age. (2015) O6.
[36] A. Cosen ino, Iden i ica ion o pigmen s by mul ispec al imaging; a lowcha me hod,
He i . Sci. 2 (2014) 1-12.
ACCEPTED MANUSCRIPT
[37] A. Cosen ino, FORS spec al da abase o his o ical pigmen s in di e en binde s, e-
conse a ion Jou nal. 2 (2014) 57-68.
[38] P. Riccia di, J. K. Delaney, M. Facini, L. Glinsman, Use o imaging spec oscopy and in
si u analy ical me hods o he cha ac e iza ion o he ma e ials and echniques o 15 h cen u y
illumina ed manusc ip s, Jou nal o Ame ican Ins i u e o Conse a ion, 52, 1, (2013) 13-29.
[39] T. Ca ale i, A. Gio agnoli, M. Ne o, Pigmen s and Mix u es Iden i ica ion by Visible
Re lec ance Spec oscopy, P ocedia Chem. 8 (2013) 45-54.
[40] M. Pé ez-Alonso, K. Cas o, J.M. Mada iaga, Vib a ional Spec oscopic Techniques o
he Analysis o A e ac s wi h His o ical, A is ic and A chaeological Value, Cu . Anal.
Chem. 2 (2006) 89-100.
[41] K. Cas o, M. Pé ez-Alonso, M.D. Rod íguez-Laso, L.Á. Fe nández, J.M. Mada iaga,
On-line FT-Raman and dispe si e Raman spec a da abase o a is s' ma e ials (e-VISART
da abase), Anal. Bioanal. Chem. 382 (2005) 248-258.
[42] I.M. Bell, R.J.H. Cla k, P.J. Gibbs, Raman spec oscopic lib a y o na u al and syn he ic
pigmen s (p e- ≈ 1850 AD), Spec ochim. Ac a A. 53 (1997) 2159-2179.
[43] R.T. Downs, The RRUFF P ojec : an in eg a ed s udy o he chemis y, c ys allog aphy,
Raman and in a ed spec oscopy o mine als, in: Anonymous P og am and Abs ac s o he
19 h Gene al Mee ing o he In e na ional Mine alogical Associa ion in Kobe, Japan, , 2006,
pp. O03--13.
[44] R. B uque as, T. An elo, Á. A eaga, P. Bo ego, M. Bueso, M. del Egido, A. Gabaldón,
M. Gómez, C. Ma ín de Hijas, C. Vega, D. Juanes, Es udio écnico de Fe nando VII a caballo
de F ancisco de Goya, Bienes Cul . 8 (2008) 117-132.
[45] J.H. Hubbell, S.M. Sel ze , Tables o X- ay mass a enua ion coe icien s and mass
ene gy-abso p ion coe icien s 1 keV o 20 MeV o elemen s Z= 1 o 92 and 48 addi ional
subs ances o dosime ic in e es , The Na ional Ins i u e o S anda ds and Technology (NIST).
NISTIR--5632 (1995).
[46] M. G ä e, E. Donne , R.N. Collins, E. Lombi, Specia ion o me al(loid)s in
en i onmen al samples by X- ay abso p ion spec oscopy: A c i ical e iew, Anal. Chim.
Ac a. 822 (2014) 1-22.
[47] N. Eas augh, V. Walsh, T. Chaplin, R. Siddall, The Pigmen Compendium: A Dic iona y
o His o ical Pigmen s, Bu e wo h Heinemann, I aly, 2008.
[48] J.R. Ba ne , S. Mille , E. Pea ce, Colou and a : A b ie his o y o pigmen s, Op . Lase .
Technol. 38 (2006) 445-453.
[49] J. Ki by, D. Saunde s, Fading and Colou Change o P ussian Blue: Me hods o
Manu ac u e and he In luence o Ex ende s, Na ional Galle y Technical Bulle in. 25 (2004)
73-99.
ACCEPTED MANUSCRIPT
Illus a ions
FIG. 1.- The di e en HSI sys ems and below hei co esponding In a ed False Colou Image (IRFC) o
“Po ai o a young man wi h ha ” (1758) a) HSI Sys em 2, b) HSI sys em 1.
FIG. 2. a) Visible colou image o “Po ai o a young man wi h ha ” (1758) wi h zone o Raman analyses; b)
False colou nea -in a ed luminescence image (IRFC) (550, 650, and 900 nm); c) A plo o he e lec ance
spec a o he end membe s and d) hei associa ed spa ial dis ibu ions in he pain ing. Blue ( ed lips): e milion;
Black: da k zones whe e he pigmen s a e di icul o iden i y because he e lec ance spec um has a ela i ely
la p o ile and low in ensi y; Red and g een colou code co esponds o a mix u e o e milion and lead whi e a
di e en concen a ions; Yellow zone ep esen s he dis ibu ion o ed och e (p epa a ion laye ); Viole and da k
g een ( es o a ions): e milion + ed och e.
FIG. 3. a) EDXRF spec um o a black poin o he “Po ai o a young man wi h ha ” whe e he high p esence
o lead is obse ed. b) Pu e e milion pigmen Raman spec um ob ained om he “Po ai o a young man wi h
ha ” pain ing and he Raman spec um o he P ussian Blue iden i ied in he medal ibbon o he “Juan Ma in de
Goicoechea” a wo k.
FIG. 4. Re lec ance spec a on he cheek o he “Po ai o a young man wi h ha ”, 1758. The con ibu ion a 450
nm co esponds o he lead whi e mixed wi h he e milion. F om he bo om up he di e en spec a show low
concen a ions o lead whi e in he mix u e o high concen a ions.
FIG. 5.a) Visible colou image (RGB) o “Po ai o Juan Ma in de Goicoechea” (1790) wi h zone o Raman
analyses; b) False colou nea -in a ed luminescence image (IRFC) (550, 650, and 900 nm); c) A plo o he
e lec ance spec a o he end membe s and d) hei associa ed spa ial dis ibu ions in he pain ing. Da k ed:
e milion + ed och e + lead whi e; pink (cheek): e milion; ligh blue and iole : mix u e ed lake + e milion;
da k g ey ( ibbon o he medal): non iden i ied; black: non iden i ied; g een (hai ): calci e.
Table 1. Pale es o he “po ai o a young man wi h ha ” and “Don Juan de Goicoechea”, x = colo an
iden i ied by he echnique ( o XRF: indica ion o key-elemen s); s. = colo an sugges ed by he echnique; n.i. =
colo an no iden i ied by he echnique.
Hue
“Young man”
“Don Juan”
HSI e lec ance
spec a
XRF
Raman
Red
Ve milion + lead whi e
Ve milion + ed lake
x
x
450 a, 610b
630b
x
x
x
x
Red Och e
Red Och e
x
615a, 740a
x
n.i.
Whi e
Lead whi e
Lead whi e
s.
_c
x
x
Calci e
Calci e
s.
_c
s.
x
Black
Ca bon black
Ca bon black
n.i.
_c
n.i.
x
Blue
P ussian blue
s.
460a
n.i.
x
aAppa en abso bance maximum. bIn lec ion poin . cNo clea spec al ea u e de ec able.
ACCEPTED MANUSCRIPT
Figu e 1
