Research article

Received: 18 November 2015 Revised: 18 April 2016 Accepted: 20 April 2016 Published online in Wiley Online Library

(wileyonlinelibrary.com) DOI 10.1002/jrs.4956

Raman spectroscopy of green minerals and

reaction products with an application in
Cultural Heritage research
A. Coccato,a* D. Bersani,b A. Coudray,c J. Sanyova,c L. Moensd
and P. Vandenabeelea
Raman spectroscopy is a powerful technique for the characterization of materials and is of valuable use in archaeometrical
research in general. Green compounds of natural or synthetic origin are found in many research areas, ranging from mineralogy,
to pigment identification, to corrosion studies. However, a detailed and comprehensive database of spectra and references is still
missing in the literature. This paper provides both, a literature review and downloadable Raman spectra of reference products, to
the researcher dealing with green materials in cultural heritage. Moreover, it tackles nomenclature issues. The collected spectra
are discussed in relation to the preliminary/commercial identification of the material itself and to the published data. Practical as-
pects regarding the laser wavelength selection are also discussed with regards to the comparison to published reference spectra.
The range of studied green materials is wide and encompasses Cu containing compounds (natural and synthetic, more or less
known as pigments or degradation products, including polymorphs of the same formula), Fe based (green earths and synthetic
organic pigments), modern Cr and Co green pigments.
This approach is illustrated by analysing a cross section of a green zone of the Early Netherlandish panel painting ‘Ghent
Altarpiece’ by the Van Eyck brothers. Copyright © 2016 John Wiley & Sons, Ltd.
Additional supporting information may be found in the online version of this article at the publisher’s web site.

Keywords: Raman spectroscopy; green earths; copper containing green pigments; copper resinate; synthetic organic pigments

Introduction
It is common knowledge for the (amateur) artist that green paint
can be obtained by mixing the primary colours yellow and blue
(subtractive synthesis). This was known to artists and artisans since
millennia, as we can now observe in many examples from antiquity
to modern times.[1–8] Moreover, a huge variety of green com-
pounds suitable for the use as pigments is available and is now
identified in cultural heritage objects, in contrast to the supposedly
restricted variety of green pigments existing according to tradi-
tional sources (mainly malachite and verdigris).[5,9]
The observed variety of materials encompasses both natural and
synthetic compounds (containing Cu, Fe, Cr or Co), as well as
mixtures (for example: chrome green is a mixture of chrome yellow
and Prussian blue, while chrome oxide green is a green material of
formula Cr2O3).
The focus in this paper is given to green minerals and to reaction
compounds of importance for cultural heritage studies, which
covers natural and synthetic pigments, as well as corrosion
products.
Nomenclature problems have to be faced, as the correspon-
dence between a traditional name and a specific chemical compo-
sition is not always certain, especially when referring to ancient
sources (for example the term ‘chrysocolla’ had a different meaning
to Pliny the Elder in comparison to a 21st century geologist).[8,10]
Natural Fe containing green pigments (green earths) are found in
many local deposits, which makes these pigments less likely to be
traded over long distances. Glauconite, celadonite and other green
earth minerals have been used as pigments.[10,11]
Copper based green pigments such as malachite, verdigris and
copper resinate are well known as painting materials,[5,9] while
sulphates, chlorides and other Cu salts were mainly investigated
in relation to corrosion processes.[12–16] However, more detailed
investigation of works of art revealed that a huge variety of
Cu salts were also used as pigments (sulphates, phosphates,
chlorides).[5,17–23] This fact implies that the identification of Cu in a
green area cannot be straightforwardly interpreted as related to
the presence of the traditional pigments malachite or verdigris,
but a deeper investigation is required.[5] For example, the simulta-
* Correspondence to: Alessia Coccato, Ghent University, Department of
Archaeology, Sint-Pietersnieuwstraat 35, B-9000 Ghent, Belgium.
E-mail: Raman@UGent.be
a Ghent University, Department of Archaeology, Sint-Pietersnieuwstraat 35, B-9000,
Ghent, Belgium
b Università degli Studi di Parma, Dipartimento di Fisica e Scienze della Terra, Parco
Area delle Scienze 7/a, 43124, Parma, Italy
c Institut Royal du Patrimoine Artistique (IRPA-KIK, Belgium), 1, Parc du
Cinquantenaire, 1000, Brussels, Belgium
d Ghent University, Department of Analytical Chemistry, Krijgslaan 281, S12, B-9000,
Ghent, Belgium

J. Raman Spectrosc. (2016) Copyright © 2016 John Wiley & Sons, Ltd.
 A. Coccato et al.

neous detection of Cu and Cl could be related, at the same time, to Table 1. Overview of green pigments and degradation compounds of
the natural mineral atacamite or to the synthetic copper interest for archaeometry
phtalocyanine green, which is a modern synthetic pigment.[24] Cu containing green pigments
Moreover, this elemental association can be related to the degrada-
tion of the blue pigment azurite in the presence of chloride ions or Cu2(CO3)(OH)2 Malachite
to verdigris, in case it was prepared with additives such as salt, (Cu,Al)2H2Si2O5(OH)4 · zH2O Chrysocolla
honey or urine.[25] Cux(PO4)y(OH)z Cornetite, libethenite
Raman spectroscopy plays a prominent role in this field of pseudomalachite,
archaeometrical research, as its capability in discriminating among reichenbachite, ludjibaite
different molecular structures might help in identifying green pig- CuSO4 · yCu(OH)2 · zH2O Antlerite, brochantite,
ments. However, some limitations have to be taken into account, posnjakite, langite
like for example the weak Raman scattering properties of some of Cu2Cl(OH)3 Atacamite, botallackite,
the investigated materials in relationship to a specific laser wave- clinoatacamite
length, or the overlap of the main band of anatase (TiO2, Fe containing green pigments
3+
143 cm 1) to the ν(O―Al―O) stretching vibration of kaolinite ((K,Na)(Fe ,Al,Mg)2(Si,Al)4O10(OH)2) Glauconite
3+ 2+
(Al2Si2O5(OH)4, 143 cm 1), which is of importance in investigating (K[(Al,Fe ),(Fe ,Mg)](AlSi3,Si4)O10(OH)2) Celadonite
clay pigments.[26] Therefore, using a combination of sensitive ana- Synthetic green pigments
lytical techniques seems to be of extreme importance, in combina- xCu(CH3COO2) · yCu(OH)2 · zH2O Verdigris
tion with having access to an extended, accessible Raman Copper resinate
2+
database. (Ca,Cu )3(SiO3)3 Egyptian green
C32H3Cl13CuN8 to C32HCl15CuN8 PG 7
C32Br6Cl10CuN8 PG 36
Experimental C30H18FeN3O6Na PG 8
Cr2O3 Chrome oxide green
Instrumentation Cr2O3 · 2H2O Viridian
Raman spectra of the green materials were recorded with both Co2TiO4 Cobalt green
lasers of a Bruker Optics ‘Senterra’ dispersive Raman spectrometer,
i.e. with laser wavelengths of 532 and 785 nm. Spectra were
recorded in the ‘high resolution’ mode (3–5 cm 1) of the instru- provides an overview of the materials of interest for cultural
ment. Full range spectra (60–3700 (532-nm laser) and 80–3500 heritage investigations, and Table 2 a list of the studied samples,
(785-nm laser) cm 1) were recorded using a 20× objective, giving including details about their origin and preliminary identification.
a spotsize of ca. 10 μm. The laser power was adjusted so that no The experimental conditions and the Raman identification of the
degradation was caused on the materials, and it was always 4 mW materials (including some modifications to the preliminary label-
and 1.4 mW at the sample for the 785-nm and the 532-nm laser ling) are also given in Table 2.
respectively. If a lower laser power was used, it is mentioned. The In addition to these reference materials, a real sample coming
acquisition time and accumulations were defined for each case, in from the Ghent Altarpiece panel painting was investigated with
order to obtain good signal-to-noise ratios. At least three spectra Raman spectroscopy in order to establish the potentialities of this
per sample were acquired, on different positions. The system uses technique in the detection of copper resinate in cross sections.
a thermo-electrically cooled CCD detector, operating at 65 °C. All
the measurement parameters and instrumental conditions are con-
trolled by the OPUS software.
For the analysis of copper resinate and of a real sample con-
Results and discussion
taining this pigment also a 514-nm green laser was used. The All the reference spectra (without any correction) can be requested
Raman spectra were acquired with a Renishaw InVia multiple laser at http://www.analchem.ugent.be/RAMAN/.
Raman spectrometer with a Peltier cooled ( 70 °C) NIR enhanced The results are presented separately for natural copper con-
deep depletion CCD detector. The laser power was kept below taining minerals of interest for cultural heritage, natural iron based
300 μW. Embedded stratigraphic samples can be analysed at green pigments, which are the green earths glauconite and
magnifications of 5× to 100× in the direct-coupled Leica DMLM celadonite, and finally synthetic pigments (containing Cu, Cr, Co,
microscope with enclosure. To minimize the high reflection, a set Fe). As regards the copper containing materials, the discussion will
of polarizers is added to the microscope when viewing the cross be split according to the anionic species (carbonates, silicates,
sections under white light. phosphates, sulphates, chlorides). An overview of the detected
Data processing was performed by using GRAMS 8.0 software Raman bands, including their relative intensity[27] with both 785
(Thermo). All the spectra presented in the figures, as well as the and 532-nm laser excitation and of the identified materials in all
downloadable files, are without baseline correction. the studied samples can be found in Table 3.

Samples
Green reference materials of natural and synthetic origin were
purchased by Kremer Pigmente (Aichstetten, Germany) and were
available in the laboratories (copper resinates) or on the mineral
market, from collectors. Most of the Cu containing minerals of
relevance for cultural heritage belongs to the third group. Table 1
Cu containing materials
Malachite (Cu2(CO3)(OH)2), together with the artificial green pig-
ment verdigris (xCu(CH3COO2) · yCu(OH)2 · zH2O), is known as a
painting material since antiquity.[5,42] The blue copper-containing
pigment azurite (Cu3(CO3)2(OH)2) is also known to degrade into

wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd. J. Raman Spectrosc. (2016)
 Table 2. Overview of the studied samples listed according to their commercial identification. Provenance and measurement conditions are given, for both used lasers. Finally, the identified materials on the
basis of the recorded vibrational signature are listed
Denomination Description and Measurement conditions (number of points * New identification
provenance seconds, accumulations) based on Raman spectra
785 nm 532 nm

Cu carbonates Malachite 10300 Kremer – natural, standard — 3 * 30s, 30acc Malachite Cu2(CO3)(OH)2

J. Raman Spectrosc. (2016)

10310 Kremer – natural, extra fine — 3 * 30s, 30acc Malachite Cu2(CO3)(OH)2
Cu silicates Chrysocolla 1035 Kremer – natural — 3 * 40s, 30acc Chrysocolla (Cu,Al)2H2Si2O5(OH)4 · zH2O,
malachite Cu2(CO3)(OH)2
573 Reppia (NW Italy) 300 s, 70acc 3 * 40s, 40acc Allophane (Al2O3(SiO2)1.3–2.0 · 2.5–3.0H2O)
145 Libiola (NW Italy) 3 * 300 s, 30acc 3 * 200 s, 30acc Brochantite Cu4(SO4)(OH)6
Cu phosphates Libethenite 1160 Libethen (Slovakia) 120 s, 70acc 3 * 60s, 80acc Libethenite Cu2(PO4)(OH)
Cu sulphates Brochantite 555 (Morocco) — 3* 150 s, 15acc, Brochantite Cu4(SO4)(OH)6,
3* 60s, 80acc antlerite Cu3(SO4)(OH)4
Brochantite, pyromorophite Baccu Locci, (Sardinia, Italy) 6 * 300 s, 10acc 3 * 30s, 10acc, Brochantite Cu4(SO4)(OH)6,
— 3 * 60s, 40acc pyromorphite (Pb5(PO4)3Cl)
Brochantite, serpierite Corchia (NW Italy) — 3 * 30s, 10acc, Brochantite Cu4(SO4)(OH)6,
— 3 * 60s, 10acc serpierite(Ca(Cu,Zn)4(OH)6(SO4)2 · 3H2O)
Langite 456 (Voltaggio, NW Italy) 2 * 120 s, 10acc 3 * 30s, 10acc Langite Cu4(SO4)(OH)6,
3 * 300 s, 30acc 3 * 120 s, 40acc brochantite Cu4(SO4)(OH)6
Cu chlorides Atacamite 103900 Kremer — 3 * 30s, 20acc Malachite Cu2(CO3)(OH)2
1320 Monte Fucinaia (central Italy) — 3 * 30s, 15acc Atacamite Cu2Cl(OH)3
Raman spectroscopy of green compounds in Cultural Heritage research

Paratacamite 1192 Piombino (central Italy) — 3 * 30s, 20acc Clinoatacamite Cu2Cl(OH)3

Green earths Glauconite/Celadonite 17410 Kremer (Cyprus) – blue green 3 * 100 s, 60acc 3 * 40s, 20acc Celadonite
3+
((K,Na)(Fe ,Al,Mg)2(Si,Al)4O10(OH)2)/ 17400 Kremer (Cyprus) – green 3* 60s, 30acc 3 * 60s, 80acc Celadonite, anhydrite
3+ 2+

Copyright © 2016 John Wiley & Sons, Ltd.

(K[(Al,Fe ),(Fe ,Mg)](AlSi3,Si4)O10(OH)2) 40830 Kremer (France) – natural 3 * 30s, 30acc 3 * 60s, 30acc PG 7, bismutoferrite, gypsum,
magnesite, calcite, quartz
40821 Kremer (Verona, NE Italy) – 3 * 20s, 20acc 3 * 20s, 20acc PG 7, green earth, calcite
natural
40800 Kremer – natural 3 * 30s, 40acc 3 * 30s, 20acc Calcite, kaolinite, carbon
(amorphous and diamondlike),
celadonite, rhodonite or fowlerite
11110 Kremer (Russia) – natural 3 * 60s, 100acc 3 * 200 s, 50acc Green earth (unspecified), calcite,
quartz, kaolinite, dolomite, anhydrite
41700 Kremer – ground and purged 3 * 40s, 30acc 3 * 30s, 30acc PG 7, green earth, norbergite, calcite,
green earth, enhanced with viridian magnesite, gypsum
Celadonite 11100 Kremer (Bavaria) – natural 3 * 30s, 30acc 3 * 30s, 20acc Glauconite, calcite
40810 Kremer (Bohemia) – natural 3 * 120 s, 20acc 3 * 60s, 60acc Glauconite
11000 Kremer (Verona, NE Italy) – 3 * 30s, 30acc 3 * 20s, 30acc Anatase, calcite, goethite
worse quality of the historical pigment
110005 Kremer (Verona, NE Italy) – 3 * 60s, 10acc 3 * 60s, 80acc Anatase, calcite, celadonite, altered
worse quality of the historical pigment green earth

(Continues)

wileyonlinelibrary.com/journal/jrs
 A. Coccato et al.

green compounds.[43] Historical sources are found in Cyprus,

Copper acetates Cux(CH3COO)y(OH)z · nH2O

Macedonia, Spain and Armenia.[8]

oxide hydrated Cr2O3 and Cr2O3 · 2H2O

PG 7 C32H3Cl13CuN8 to C32HCl15CuN8
Moreover, green Cu containing salts are found as degradation
products of copper or bronze artifacts,[13,15,44–46] or of Cu con-
based on Raman spectra

Chromium oxide and chromium

taining pigments.[24,47,48] The formation of chlorides, sulphates,
New identification

phosphates and mixed salts is related to a variety of environmental

and intrinsic factors, so that the presence of specific secondary

PG 50 C30H18FeN3O6Na
Cobalt titanate Co2TiO4
Tridymite (Ca,Cu)3Si3O9

Chromium oxide Cr2O3

Chromium oxide Cr2O3

PG 36 C32Br6Cl10CuN8
minerals is related to local conditions leading to degradation and
to the stability of the compounds.[13,49,50]
Copper resinates
Copper resinates

These observations are crucial to a correct approach to the

Fluorescence
interpretation of analytical results of green pigments in works of
art, as the detection of Cu in a green area has no straightforward
implications on the identification of the actual pigment. In all these
circumstances, it is necessary to couple the elemental technique
with a molecular one (see for example the combination of PIXE with
3 * 120 s, 30 acc
Measurement conditions (number of points *

3 * 120 s, 10acc

3 * 120 s, 40acc
3 * 100 s, 60acc

3 * 150 s, 60acc
3 * 60s, 100acc

3 * 60s, 100acc
3 * 60s, 15 acc

3 * 30s, 30acc
3 * 30s, 30acc
3 * 40s, 60acc
3 * 60s, 60acc

Raman spectroscopy for the study of green pigments in

532 nm

manuscripts[5]).
seconds, accumulations)

Carbonates
The green basic carbonate malachite is described in treatises
3 * 200 s, 100acc (0.4 mW)

already in the Roman times,[42] and the term ‘chrysocolla’ seems

3 * 300 s, 30acc (0.4 mW)

3 * 300 s, 30acc (0.4 mW)

to have been used as a synonym for that[51] (see the next paragraph
for more details). It is also widely found in works of art from various
785 nm

3 * 150 s, 30 acc

3 * 120 s, 10acc
3 * 10s, 100acc

cultures, from the Egyptians[52] to 19th century artists.[53] A list of

3 * 60s, 10 acc
—
—

—
3 * 30s, 30acc

findings is reported in literature,[51] as well as of its Raman

identification.[3,54–57]
Malachite is typically present in Cu ore deposits (Hungary, East
Central Europe, France[51]; Cyprus, Macedonia, Spain, Armenia[8]),
and it is often associated with azurite, chrysocolla and cuprite.[51]
Malachite is also found as a corrosion product on bronze.[49,50,58]
Differences in the Raman spectra of malachites are attributable
to different formation processes (mineral vs corrosion product).[49]
CR-KIK2 (1973: resinate + oil)
Description and

However, the main bands still allow for the identification of the
CR-KIK1 (1981: verdigris +
provenance

monoclinic carbonate[3,8,49,53,59,60] (Fig. 1, spectrum a). The most

Abies pectinata resin)

intense and diagnostic bands are the CO3 stretching modes at

1493, 1462 and 1368 cm 1 (asymmetric ν3[59,61]), the doublet at
44100 Kremer
44450 Kremer

10064 Kremer

1097–1059 cm 1 (doubly degenerated ν1) and the ν4 and ν2 at ca.

1220 Kremer

720 and 750 cm 1 and 820 cm 1 respectively.[61] Below 600 cm 1,

the lattice mode at 435 cm 1 is very intense, but also other bands
44204
44200

PG 36
4425

PG 7

PG 8

are observable (see Table 3).[61] The OH stretching and bending

vibrations are found at 3380, 3382 and 1638 cm 1, in good agree-
ment with published data.[59,61] All these bands are observed with
the 532-nm laser only, as absorption effects of red light hamper
the recording of a Raman spectrum with the 785-nm laser.[62]
Denomination

Cu containing wollastonite

Silicates
Cu salts of abietic acid

The mineralogical term chrysocolla ((Cu,Al)2H2Si2O5(OH)4 · zH2O)

Cr oxide hydrated

now indicates a series of hydrated copper silicates, while in the past

it has been used as a synonym for malachite.[8] It is currently
believed that chrysocolla is a hydrogel containing Si and Cu,
Verdigris

Cr oxide

associated with copper oxides and carbonates.[63]

PG 50

Three samples described as chrysocolla were available: the

commercial pigment 1035 Kremer and two samples from Northern
Italy (Libiola and Reppia mines).
Copper phtalocyanine
Copper phtalocyanine
Table 2. (Continued)

The Kremer sample 1035 (532 nm, Fig. 1, spectrum b) showed

mainly the bands of malachite, but after careful observation of
Fe azo complex
Egyptian green

the spectrum and subtraction of a malachite reference spectrum

Cu acetates
Cu resinate

(after normalization to the most intense malachite band at

Co green
Cr green

Viridian

180 cm 1), additional bands attributable to chrysocolla were identi-

fied (Fig. 1, spectrum c; Table 3). The bands above 3600 cm 1 and at
1460 cm 1 are attributed respectively to the symmetric ν(OH)

wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd. J. Raman Spectrosc. (2016)
Raman spectroscopy of green compounds in Cultural Heritage research

Table 3. Overview of the Raman band positions on all the identified green materials of interest for cultural heritage, with both the used laser wave-
lengths. The relative intensity values are given according to.[27] As explained in the text, no intensity values are given for the green earths (*)
1
Identified materials Band positions in cm
785 nm 532 nm

Cu containing materials
Malachite — 3382w, 3380vw, 1638vw, 1493 m, 1462w (shoulder),
(Fig. 1, spectrum a) 1368vw-w, 1097–1059w, 820vw, 750w, 720w, 599w,
535w-m, 435 m, 351w, 267 m, 216 m, 180vs, 169 s,
154vs 144 s, 118 m
Chrysocolla — 3630vs, 2750 m-s, 1460vw, 1300w-m, 1073 m, 988 s-vs,
(Fig. 1, spectra b and c) 973 m, 792w, 678 m, 464w-m, 400 m, 340 m, 209vw
Libethenite 975 s, 300vs 3470w, 1125vw, 1069w, 1051w, 1020s, 1010 (shoulder),
(Fig. 2, spectrum a) 975vs, 862vw, 815vw, 650vw, 628w, 585vw, 558w,
466w, 450 (shoulder), 300 m, 280vw, 250vw, 228 m,
195 m, 154 m, 75 m
Brochantite 973vs, 725w, 620w, 421w, 3588vw-w, 3567vw-w, 3403vw-w, 3371vw-w, 3262vw,
(Fig. 2, spectrum b) 390w, 240vw-w, 195w-m, 1129vw, 1105vw, 1097vw, 1076vw, 973vs, 910vw-w,
168vw, 156w-m, 137w-m, 871vw, 776vw, 728vw, 620w, 609w, 595w, 507w, 483w,
123w, 117w, 92w 448w, 424w, 387w-m, 366vw-w, 319w, 262vw, 243vw-w,
227vw, 198w, 188vw, 177vw, 170vw, 157vw, 150vw,
139w, 132vw, 125vw, 119vw, 105vw, 90w-m
Langite 1094vw, 970vs, 605w, 428 m, 3587w, 3578w, 3399vw, 3370vw, 1156vw, 1127vw, 1093vw,
(Fig. 2, spectra c and d) 239 m, 130w 970vs, 611vw-w, 510vw, 487vw, 450 (shoulder), 432 m, 242w
Antlerite — 3581w, 3489vw-w, 1171vw-w, 1122vw, 1076w, 988vs,
(Fig. 2, spectrum e) 784vw, 748vw, 501w, 484w, 470 (shoulder), 442w, 416 m,
341vw, 334vw, 296vw, 267w, 248vw-w, 230vw, 216vw, 173vw,
149vw, 143vw, 124w, 108w, 97w, 82vw, 70vw
Serpierite — 3616vw, 3570vw, 1168vw, 1132w, 1115(shoulder), 1085vw,
(Figure S1, spectrum a) 991vs, 651vw, 605vw, 474w (broad, shoulder), 445w, 426w,
415w, 338vw, 244vw-w, 218vw
Pyromorphite — 3252vw, 2662vw, 2534vw, 2422vw, 2041w, 1501vw, 1529vw,
1150vw, 1013vw, 971vw, 944 m, 918vs, 819 m, 576vw, 553vw,
411w, 391w, 344vw, 324vw, 178w-m, 157w-m, 106vs, 89 s-vs
Connellite — 3550vw-w, 984 m-s, 585w, 524w, 446w, 404vs, 350w, 262vw,
(Figure S1, spectrum b) 236vw, 192w, 184w, 132w-m
Chalcophyllite 981vs, 854vs (asymmetric); 3600–3400w, 979 s, 829vs (broad, asymmetric), 613vw, 495 m,
(Figure S1, spectrum c) 464, 208, 128 (quartz) 454 m, 393w-m, 370 (shoulder), 332vw, 305vw, 272vw, 219vw,
177vw, 118w-m
Spangolite 968vs, 615w, 520 m, 410w, 968vs, 615w, 520 s, 410 m, 168w-m
(Figure S1, spectrum d) 168w
Atacamite — 3436 m to vs, 3346 m to vs, 3328 w to m 977w, 911 m, 846vw,
(Fig. 3, spectra a, b and c) 822w, 591vw, 511 m, 452vw, 362–357vw-w, 263w, 237vw,
216vw, 149 m, 138 m, 119vs, 106w, 65w
Clinoatacamite — 3445vs, 3355vs, 3311 m, 967 m, 929 m, 897 m, 867vw, 803w,
(Fig. 3, spectrum d) 575vw, 514 s, 444w, 424w, 364 m, 266vw, 166vw, 142 s, 119vs
Allophane 859vs 3420vs (broad), 2942vw, 1638vw, 1357vw,
1103w, 982vw, 858 m,
720vw, 502w-m, 396w, 364w
Fe containing materials (*)
Bavarian green earth Calcite (1087, 711) Glauconite (3561, 597, 264)[8,11,26]
(11100 Kremer; Fig. 4, spectrum a)
Bohemian green earth Glauconite (450)[11,26]
(40810 Kremer)
Cyprian blue green earth Celadonite (3560)[8,26]
(17410 Kremer; Fig. 4, spectrum b)
Cyprian green earth Celadonite
(17400 Kremer) (<300, 966, 3560)[8,11,26]

(Continues)

J. Raman Spectrosc. (2016) Copyright © 2016 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/jrs
 A. Coccato et al.

Table 3. (Continued)

1
Identified materials Band positions in cm

785 nm 532 nm
[28]
Anhydrite (1017), unidentified
compounds (1210, 1290)
Green earth from France, light PG7[29]
(40830 Kremer; Fig. 4, spectrum c) Bismutoferrite (1536, 1290, Gypsum (1136, 1008, 493, 414), magnesite (1764, 1094, 741,
1219, 695, 440, 347, 332, 145)[30] 328, 212), calcite (1086, 712, 283, 156), quartz (463),[28,31]
water molecules (3403, 3498),[28] unassigned (2874, 2828,
2220, 2190, 1614, 670)
Green earth from Verona PG 7[29]
(40821 Kremer) Unspecified green earth Calcite (1086, 712, 283, 156)[31,32]
(694, 349, 332, 294, 147)[11]
Green earth light Calcite (1087),[31,32] kaolinite
(40800 Kremer) (395, 426, 469, 509)[33]
Amorphous carbon and Celadonite (609, 550, 467, 341, 253, 164)[11]
diamondlike structures
(1600, 1400, 1270),[34]
unspecified green earth
(673, 604, 547 and 465),[11]
kaolinite hydroxyl bending
(913),[33] rhodonite or fowlerite
(508, 568, 714)[35]
Green earth from Russia Unspecified green earth[11] Calcite (1084, 288)[32]
(11110 Kremer) Unspecified green earth, probably glauconite (698, 398,
356, 266 and 147)[8];
Quartz (464),[36] Si―O―Si stretch of kaolinite (633).[33]
Calcite and dolomite (1084 and 1096).[37] Anhydrite
(1126 and 1113)
Green earth from Verona Anatase (636, 510, 398, 148),[38]
(11000 Kremer) calcite (1088)[39]
Some bands can be attributed to Goethite (515, 480, 395, 298, 215)[40]
khandite minerals: 797 (hydroxyl
vibration of surface Al―OH)
and 280[33]
Unidentified bands, broad and
overlapping: (1232, 1293, 1419,
1523, 1690)
Green earth from Verona Anatase (147, 395, 516, 637). Calcite
(bulk) (110005 Kremer) (1087, 280, 355, 713)[31,32]
Green earths (celadonite) signature
is also visible, even if the spectrum
of TiO2 is overwhelming (587, 703).[11]
Celadonite (959).[11] Altered green earth (589).[11]
[29]
Verona green earth PG 7
(41700 Kremer) Unspecified green earth Calcite (1086, 711, 282)[31,32,37] and gypsum (1134, 1008, 415)[28]
(694 and 349).[11] Norbergite
MgFeCa(SiO4)F(OH) (979 and 955).[41]
Carbonates: calcite (1086), magnesite
(1091 and 738),[37] gypsum (1006).[28]
Synthetic origin
Verdigris 948 m, 320 m, 180vs 2941vw, 2700vw, 1670vw, 1531vw, 1441vw-w, 1421w,
(Fig. 5, spectrum a) 1365vw, 1095vw, 1057vw, 951 s, 937w, 878vw, 700w, 686w
(shoulder), 633vw, 551vw, 321vs, 301 (shoulder) 267vw,
253vw, 232w, 225w, 215w, 187w, 179vw, 144vw, 125vw, 107 m
Cu resinate 2797 (broad), 1610vs, 1445w, 1300w-m, 1201w-m, 708 m,
(Fig. 5, spectrum b) 435w, 229w, 124w-m

(Continues)

wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd. J. Raman Spectrosc. (2016)
Raman spectroscopy of green compounds in Cultural Heritage research
Table 3. (Continued)
Identified materials
785 nm
Cu resinate 514 nm: 2933vs, 2873 m (shoulder),
(Fig. 7, spectrum c) 1619 m, 1461w, 1201w, 707vw-w
Cu resinate — 2933vs (broad), 2873 m (shoulder), 1610 m (broad, asymmetric),
(Fig. 59, spectrum d)
Trydimite 1075 w-m (broad), 789w, 433 m,
(Fig. 5, spectra c and d) 355vs, 302 s-vs, 210 m
Co titanate 709vs, 520 m, 460 m, 330vw, 236vw,
(Fig. 5, spectrum e) 175 s-vs, 150w (shoulder),
119w, 80vw
Cr oxide 613 m, 555vs, 351 m, 299w
(Figure S2, spectrum a)
Cr oxide hydrated — 624vw, 583w, 552vw-w, 487vs, 266w
(Figure S2, spectra b, c and d)
PG 7 1535vs, 1508vw (shoulder), 1387vw,
(Fig. 4, spectrum d) 1334 m, 1280 m, 1207 m, 1082vw-w,
977vw, 951vw, 814vw, 770 m, 740 s,
689vs, 640vw, 345w, 331vw,
290vw-w, 223vw, 197vw, 165vw,
147vw, 99 m
PG 36 1523 m-s, 1427vw, 1368vw, 1315 m,
(Fig. 6, spectra a, b and c) 1265 m, 1186w, 1053vw, 957vw-w,
748vs, 662 m, 571vw, 534vw,
514vw, 479vw, 360vw (shoulder),
330w, 276vw, 258vw, 223w, 184vw,
155vw, 92vw
PG 8 1590w, 1550vw-w, 1511w-m, 1469w,
(Fig. 6, spectrum d) 1444w, 1416vw, 1355–1345 m,
1324w, 1302vw, 1254w-m, 1147w,
1076w-m, 1056 m, 1021 m, 878 m,
752vs, 668 s, 633 s, 614 m, 533 s,
494 m, 470 s, 447 s, 380 m, 351 m,
313 m, 223w, 196w, 151w, 119w
Figure 1. Raman spectra of selected Cu containing carbonates and silicates
(532 nm). a: malachite (Kremer 10300). b: chrysocolla (Kremer 1035). c:
subtraction of the spectrum of malachite from that of chrysocolla. The
bands of the latter are more clearly visible.
J. Raman Spectrosc. (2016) Copyright © 2016 John Wiley & Sons, Ltd.
1
Band positions in cm
532 nm
1446w-m, 1407w, 1301vw-w, 1200w, 1072vw, 1051vw, 978vw,
933vw, 887vw, 834vw, 774vw, 708w, 614vw (broad), 441vw-w,
367vw, 312vw, 221vw, 119 m, 75vw
1075w-m (broad), 789w, 550vw, 434 m, 355vs, m-s, 210 m, 141w,
112w
710 vs, 645w, 618w, 520w-m, 460w-m, 380vw, 330w, 236vw,
175w, 150w-m, 115vw
597w-m, 542vs, 391vw, 340w, 304w
3059vw, 2810vw, 2632vw, 2580vw, 1551 m (shoulder), 1532 m,
1498 m, 1475w, 1434vw, 1379w, 1330w, 1277w-m, 1192w,
1079w, 972vw, 814w, 769vw, 686vs, 642vw, 514vw, 506vw,
462vw, 368vw, 346vw, 332vw, 233vw-w, 163vw, 146vw
2597vw, 1530 m, 1489 m, 1426vw, 1376 m, 1317w (shoulder),
1272 m, 1186w, 1166w, 1057w, 966vw, 914vw, 771w, 661vs,
568vw, 533vw, 452vw, 326vw, 222vw, 167vw-w, 115vw-w
1589w, 1511w, 1468w, 1446w, 1417w, 1357 m, 1322w,
1208w, 1085–1077 m, 1056 m, 1021 m, 879vs, 752vs,
734 m (shoulder), 688w, 671w-m, 636 s, 617 s, 536 m, 495w,
465 m, 441 m, 405w, 381w-m, 354w, 314 m, 285w, 222w,
151 m, 110vs
stretching and to the δ(OH) bending modes, while silicate related
bands are those at ca. 680 cm 1 (ν4 SiO3) and between 500 and
300 cm 1 (ν2 SiO3).[63] These bands are in good agreement with
published values.[63,64] Also, the bands at 1073, 988 and 973 cm 1
could possibly be related to antlerite.[5]
The sample labelled as chrysocolla from Libiola mine, a classical
locality for Italian chrysocolla, gave very weak Raman spectra with
both lasers, and was finally attributed to brochantite as the main
observed band is that at 973 cm 1 with both lasers.[14,55,65–67]
Finally, the sample from the Reppia mine, labelled as chrysocolla-
allophane, was studied on many different points, but it always gave
the same spectrum with the 532-nm laser, with bands at 3420
(asymmetric and very broad), 2942, 1638, 1357, 1103, 982, 858,
720, 502, 396 (shoulder) and 364 cm 1. These bands do not
correspond to any of the published spectra of chrysocolla,[63,64]
while they show better correspondence to allophane (Al2O3
(SiO2)1.3–2.0 · 2.5–3.0H2O). This is not surprising because Cu-rich blue
allophane and chrysocolla are both present in the Reppia mine and
they are usually found mixed together; it is impossible to discrimi-
nate between them by simple observation, even by microscope.
One of the most prominent bands, at 858 cm 1, seems related to
wileyonlinelibrary.com/journal/jrs

the in plane δ(Si―OH) bending.[68] Another Raman band, related to
silicatic species but not expected in the spectrum of allophane, is
here observed at 1103 cm 1. This can be interpreted as the
stretching of disilicatic units.[69] As regards the presence of water,
the δ(OH) bending is observed at 1638 cm 1,[59,61] while the
stretching modes seem to correspond to those of bonded
OH.[70] The Raman spectrum recorded with the red laser,
although of lower quality, showed a band at 859 cm 1, which
corresponds to what observed with the 532-nm excitation
(Si―OH bending[68]).
Phosphates
The copper phosphates of interest for archaeology and art history
are cornetite (Cu3(PO4)(OH)3), libethenite (Cu2(PO4)(OH)) and the
polymorphs ludijbaite, pseudomalachite and reichenbachite (Cu5
(PO4)2(OH)4).[71–73] These bright green copper containing minerals
have been identified in cultural heritage objects.[17,52] Other Cu
containing phosphates are sampleite (NaCaCu5(PO4)4Cl · 5H2O)[74]
and veszelyite ((Cu,Zn)2ZnPO4(OH)3 · 2(H2O)).[75]
Raman spectra of libethenite acquired with different laser wave-
lengths are published in literature (780 nm,[64] 633 nm,[71,72]
532 nm,[64] 514 nm[64,76]), and show good agreement with our sam-
ple measured with the 532-nm laser (Fig. 2, spectrum a). The PO34
vibrations are depending on orientational effects and on which
crystal face is exposed to the laser.[77] The most diagnostic bands
are the phosphate ν1 symmetric stretching at 975 cm 1, the ν3
antisymmetric stretching at 1069 cm 1. The ν4 bending modes
are found at 650, 628, 585 and 558 cm 1, while the ν2 is at
450 cm 1.[72,77] Lattice modes are dominated by a feature at ca.
300 cm 1, which is always clear, independently on the orientation
of the sample.[71,72,76] Other bands are observed below 300 cm 1
(see Table 3).[71] The ν(OH) stretching band is observed at ca.
3470 cm 1, while the deformation bands are at 815 and
862 cm 1.[71,72]
It seems that the band at ca. 865 cm 1 is only present in
natural samples of libethenite, and might be orientation
dependent.[64,76,77]
The Raman spectrum of libethenite acquired with the 785-nm
laser shows only weak bands at 975 and 300 cm 1.
Figure 2. Raman spectra of selected Cu containing phosphates and
sulphates. a: Libethenite (Libethen, Slovakia; 532 nm). b: Brochantite (Baccu
Locci, Sardinia, Italy; 532 nm). c: Langite (Voltaggio, NW Italy; 785 nm). d:
Langite (Voltaggio, NW Italy; 532 nm) e: Antlerite (Morocco; 532 nm).
Spectra c and d were recorded on a sample described as langite.
wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd.
A. Coccato et al.
Sulphates
Copper sulphates and mixed salts are of significance for pigment
identification,[55,78] pigment degradation studies[48,79] and for
corrosion studies of copper and bronze artefacts,[14,16,66] as well as
for geological and mineralogical applications.[16] Green copper
sulphates were also identified in Netherlandish paintings from the
period 1520–1530[20–22,80] and on 16th century Portuguese–Flemish
paintings.[9] It seems that brochantite has been intentionally used
as a pigment.
Antlerite (Cu3(SO4)(OH)4), brochantite (Cu4(SO4)(OH)6), posnjakite
(Cu4(SO4)(OH)6 · H2O) and langite (Cu4(SO4)(OH)6 · 2H2O) are
some of the most interesting basic copper sulphates from
those points of view. Others are chalcantite (CuSO4 · 5H2O),[66]
serpierite (Ca(Cu,Zn)4(OH)6(SO4)2 · 3H2O),[81,82] connellite (Cu36(SO4)
(OH)62Cl8 · 6H2O)[58,83] and chalcophyllite (Cu18Al2(AsO4)4(SO4)3
(OH)24 · 36H2O).[84]
Brochantite and antlerite seem to be more stable than the other
sulphates.[66] The sulphate symmetric stretching vibration is
centred between 970 and 990 cm 1.[16,66] Brochantite Raman spec-
trum (532-nm excitation, Fig. 2, spectrum b) shows multiple bands
in the OH stretching region, two at ca. 3588 and 3567 cm 1, and
some broader ones at 3403, 3371 and 3262 cm 1. In one of the
measured samples, these bands were sharper and those at ca.
3400–3370 cm 1 were resolvable into four components (3411,
3402, 3383 and 3372 cm 1).[14,55,65–67] The ν(SO4) stretching
vibrations are observed at 972–973 cm 1 (ν1), 1129, 1105, 1097
and 1076 cm 1 (ν3) and 620, 609, 595 cm 1 (ν4).[14,55,65–67] The OH
librations are relatively weak and correspond to published values:
910, 871, 776 and 728 cm 1.[14,55,65–67] The Cu―O vibrations and
lattice modes are numerous and in good agreement with literature
(see Table 3).[14,55,65–67] The most intense of these bands are those
at ca. 480, 390 and 140 cm 1.
The Raman spectrum of brochantite recorded with the red laser
(785 nm, Fig. 2, spectrum d) shows most of the bands detected with
the green one (see Table 3).
The reference sample from Voltaggio (456, NW Italy) showed
both green brochantite and bluish crystals of langite. The Raman
spectrum of langite (532 nm) shows good agreement with pub-
lished values.[5,64,66] The ν(OH) stretching vibrations are observed
at 3587, 3578, 3399 and 3370 cm 1 while the bending modes are
weak at 770 cm 1; the very strong ν(SO4) symmetric stretch is
found at 970 cm 1. Other bands are those at 1156, 1127, 1093,
611, 510, 487, 450 (shoulder), 432 (strong) and 242 cm 1. The
Raman spectrum collected with the red laser shows bands at 970,
428 and 239 cm 1. (Fig. 2, spectra c and d; Table 3).
Among the available reference samples, more sulphates were
detected and identified. Antlerite was present together with
brochantite (the first as light green aggregates, the second as dark
green elongated crystals): with the 532-nm excitation, the OH
bands are sharper and in number of two in the stretching region
(3581 and 3489 cm 1)[5,16,66] and two in the deformation modes
(784 and 748 cm 1).[16] The sulphate symmetric stretching band is
found at 988 cm 1 and the antisymmetric ones at 1171, 1122 and
1076 cm 1. The sulphate bending region (510 to 410 cm 1) is
complex (see Table 3).[16,66] Lattice modes are weaker and reported
between 341 and 70 cm 1[16,66] (Fig. 2, spectrum e; Table 3). To the
best of our knowledge, we are the first to report the Raman bands
of antlerite below 125 cm 1.
Serpierite (Figure S1, spectrum a) was also identified together
with brochantite in a sample from Corchia (PR, Northern Italy).
The recorded band positions (Table 3) are in good agreement with
J. Raman Spectrosc. (2016)
Raman spectroscopy of green compounds in Cultural Heritage research
the spectrum presented in RRUFF,[64] and with some published
data.[81,82]
Pyromorphite (Pb5(PO4)3Cl) was also identified in some light
yellowish green crystals occurring together with brochantite in
one sample. The recorded spectrum corresponds to published
data[64,85] as regards the region 1120–150 cm 1 (see Table 3). Some
features have been recorded outside this range: very weak ones at
1150, 1501 and 1529 cm 1, broad bands at 3252, 2662, 2534, 2422
and 2041 cm 1 and two strong features at 106 and 89 cm 1, which
are for the first time here reported.
Connellite (Cu36(SO4)(OH)62Cl8 · 6H2O) is recognizable by the
bands at 984, 585, 446, 404, 350, 262, 236, 192, 184 (shoulder)
and 132 cm 1[58,64,83] (Table 3). As the mineral contains OH units,
stretching bands are expected. In all the spectra of connellite that
we recorded, a broad band at ca. 3550 cm 1 is observed[83] (Figure
S1, spectrum b). Its formation on copper containing artefacts can be
described as bronze disease.[58,83]
Chalchophyllite (Cu18Al2(AsO4)4(SO4)3(OH)24 · 36H2O) is another
emerald green copper mineral, containing sulphate, arsenate and
hydroxyl anions. The stretching bands related to these species are
found at 979, 843 (broad and asymmetric) and 3400–3600 cm 1
respectively.[84] The bending vibrations and lattice modes are in
good agreement with published data;[64,84] however, symmetry
lowering produces multiple features in the region 620–110 cm 1
(Figure S1, spectrum c; Table 3). When the sample was measured
with the red laser, the signal of quartz (464, 208 and 128 cm 1[36])
was recorded, in addition to (broad, asymmetric) bands at 981
and 854 cm 1, corresponding respectively to sulphate and arsenate
ions vibrations.[84]
Spangolite (Cu6Al(SO4)(OH)12Cl · 3(H2O)) Raman bands are
observed at 968, 615, 520, 410 and 168 cm 1 for both lasers
(Figure S1, spectrum d). They correspond to published spectra of
the RRUFF database,[64] which are the only available spectra
recorded so far.
Chlorides
Atacamite (Cu2(OH)3Cl) is the orthorombic polymorph of
hydroxychlorides, while clinoatacamite and botallackite are
monoclinic and anatacamite is triclinic.[12,13,15,45,86] Botallackite is
the most unstable polymorph[15]; paratacamite is trigonal (Cu,Zn,
Ni)2(OH)3Cl. Many issues still need to be faced, as the term
paratacamite has sometimes been used as a synonym for
clinoatacamite[15,45,87,88] even in some X-Ray diffraction databases
leading to bad naming of mineralogical samples.
It seems that the ν(OH) stretching region is effective for discri-
minating among the different polymorphs.
The expected bands of atacamite are related to the ν(OH)
stretching region (3200–3600 cm 1), the δ(OH) or δ(Cu―OH)
bending region (800–1000 cm 1) and the low wavenumber region
(below 600 cm 1) containing the O―Cu―O and Cl―Cu―Cl
modes.[15,45] The most significant band for these compounds is
the one around 510 cm 1, which is attributed to ν(CuO)
stretching.[45] However, strong orientational effects are observed[15]
(Fig. 3, spectra a, b and c).
The Kremer pigment atacamite (103900) was identified as
malachite on the basis of the Raman spectrum recorded with the
green laser.
The spectrum recorded on our atacamite from Monte Fucinaia
(Italy) sample shows bands at 3436, 3346 and 3330 cm 1 (this last
one only detected in a sample from Atacama[45]); the other bands,
as reported in Table 3, are in agreement with literature.[12,15,45]
J. Raman Spectrosc. (2016) Copyright © 2016 John Wiley & Sons, Ltd.
Figure 3. Raman spectra of selected Cu chlorides. a, b, c: atacamite (Monte
Fucinaia, central Western Italy; 532 nm). d: clinoatacamite (Piombino, central
Western Italy; 532 nm). The spectrum d was collected on a sample labelled
as paratacamite. For more details about the nomenclature issue, see the text.
The sample labelled as paratacamite, on the other hand, did not
correspond to the Zn, Ni-rich copper hydroxychloride, but to
clinoatacamite on the basis of the bands reported in Table 3
[13,15,45]
(Fig. 3, spectrum d). Such labelling mistakes are not unusual,
because of the confusion in the past on the classification of copper
hydroxychlorides. Moreover, this spectrum showed broad features
at ca. 2960–2850 and 1450, 1300 cm 1. These features are attrib-
uted to organic compounds, such as those characteristic of human
skin. It seems logic that handling of the sample is responsible for
the detection of phospholipidic materials: symmetric and antisym-
metric ν(CH2) at 2850 and 2930 cm 1 respectively; symmetric and
antisymmetric ν(CH3) at 2870 and 2960 cm 1 respectively; δ(CH2)
scissoring and twisting at 1450 and 1300 cm 1 respectively.[89]
Fe containing materials
Green earths are widely used in space and time.[10,11] Local deposits
of clay materials of appropriate green shade are easily found, and it
is likely that local sources were exploited (Verona, NE Italy; Cyprus;
Spain[8,90]). The two minerals composing the so-called green earths
are celadonite (K[(Al,Fe3+),(Fe2+,Mg)](AlSi3,Si4)O10(OH)2) and glauco-
nite ((K,Na)(Fe3+,Al,Mg)2(Si,Al)4O10(OH)2). The chemistry of these
micas is similar; however, they are formed, respectively, in
metamorphic/volcanic and sedimentary rocks.[8,53,90,91] Depending
on the origin, they have different molecular structures, and hence
different Raman spectra.[39] This also means that the two minerals
do not occur together.[52] However, the associated minerals are
common (quartz, clays, calcite, iron oxides, feldspars, anatase) and
are therefore not useful for discrimination.[8,10] Additional green
clay-like materials can be referred to as green earths.[10,11] The
green colour of green earths is explained by the simultaneous pres-
ence of divalent and trivalent iron.[4,26,90] Powder X-ray diffraction
(XRD) is not always successfully discriminating between the two
silicates,[10] while Raman and IR spectroscopies can yield good
results,[4,8,10,11] on the condition that the sample is not degraded.[91]
It is reported in literature that green earths are not very Raman
active when excited with a 1064-nm laser,[26,36] although both red
and green lasers are successfully used to characterize and identify
these silicates, even if they are relatively poor Raman
scatterers.[4,11,26,39,53,92,93] It is important to note that Raman studies
of phyllosilicates are affected by the natural complexity of the
wileyonlinelibrary.com/journal/jrs

chemical structure (variation in ions and ionic groups), the small
size of the mineral grains (<2 μm), the low degree of crystallinity
and possibly from the contamination from organic molecules. Also,
some resonance effects may occur in clay materials for specific laser
wavelengths.[26,91,93] For all these reasons, and for the simultaneous
presence of other minerals together with glauconite and
celadonite, Table 3 reports only band positions, without indication
of relative intensity.
The Raman bands of glauconite and celadonite are reported in
literature; however, it is sometimes hard to differentiate between
the two green earths,[4,39] because of similarity in the molecular
structure.[8] It seems that the most useful bands for discrimination
are the ones in the lower wavenumber region (100–300 cm 1,
MoO6 octahedra vibrations): celadonite presents a doublet at
174–202 cm 1 and a band at ca. 280 cm 1, while glauconite has a
single band at 170 cm 1 and presents a downshift in the band at
ca. 270 cm 1,[8,11] as a consequence of increasing Al content.[8,11,92]
Also, the ν(OH) stretching region seems diagnostic, as glauconite
shows a weak broad feature, while celadonite has resolved bands,
both in the Raman[11] and FT-IR spectra.[4,8,11]
Among the studied reference samples it was possible to identify
both glauconite and celadonite. However, in some cases the dis-
tinction was not feasible.
Glauconite
Glauconite samples are the green earths from Bavaria, Bohemia and
Russia (11 100, 40 810 and 11 110 Kremer respectively). The bands
which allowed the assignment to the Al-rich green earth glauconite
are at ca. 700, 544, 450, 380, 270 and 180 cm 1. The latter band is
actually broad and reported between 188 and 200 cm 1.[11]
Resonance effects are described by Ospitali et al.,[11] so that it is
not surprising that spectra recorded with the green laser are dom-
inated by the features at 700, 450 and ca. 265 cm 1 (Fig. 4, spec-
trum a), while with the red laser the most prominent features are
the 590 cm 1 and 384–389 cm 1 ones.[11] Moreover, a variety of
spectra can be ascribed to glauconite,[8,11,26] and it is not surprising
that not all the recorded spectra are identical. Moreover, the
absence of sharp bands of ν(OH) supports the glauconite identifica-
tion, both in Raman and FT-IR spectra.[8]
Figure 4. Raman spectra of green earth pigments. a: green earth from
Bavaria, identified as glauconite (Kremer 11100, 532 nm). b: celadonite
from Cyprus (Kremer 17410, 532 nm). c: green earth from France (Kremer
40830, 532 nm). d: PG 7 (532 nm).
wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd.
A. Coccato et al.
Celadonite
The Raman spectrum of celadonite is expected to show narrower
bands than the glauconite spectrum, as a result of greater order,
as the substitution of Al for Si is limited.[11] Celadonite containing
samples are the two green earths originating from Cyprus (blue
green and green earth, Kremer pigments 17410 and 17400
respectively. See Fig. 4, spectrum b). These spectra show a doublet
at 170–200 cm 1, and bands at 270, 394–400 cm 1. The higher
wavenumber band (550 cm 1) is affected by resonance effects.[11]
Cyprus green earth (celadonite is geologically associated to basaltic
rocks) was used in works of art since antiquity.[52] In these cases, the
bands at 272–276 cm 1 and the sharp feature at 3560 support the
identification.[8,11,26] Moreover, in the case of the Cyprian green
earth, a feature at 970 cm 1 is also visible, which is only present
in celadonite.[11] A good agreement is observed with the spectra
of celadonite from Cyprus previously published.[11,26]
Other pigments, described as green earths by the supplier, did
not necessarily contain green earths only. It seems relevant to high-
light that names provided by suppliers usually refer to the hue or
the colour and not automatically to the mineral composition of
the material. In some cases such as green earth from France
(40830 Kremer—Fig. 4, spectrum c; Table 3), and green earth from
Verona (40821 Kremer, Table 3), synthetic organic pigments were
detected such as PG 7 (Fig. 4, spectrum d). In some other cases,
the signature of anatase overwhelmed the signal from the other
components. Indeed, anatase is a very good Raman scatterer, while
in general the Raman intensity of silicates is weaker. Anatase was
detected together with goethite, calcite and clay minerals
(halloysite, dickite) in powdered green earth from Verona (11000
Kremer, Table 3). The bulk Verona green earth sample yielded a
complex Raman signature including anatase, calcite, haematite
and silicates. The green earth mineral does possibly contain
celadonite, as indicated by the presence of OH stretching vibrations
at 3606 and 3566 cm 1.
Synthetic origin
Some copper containing green pigments are also artificial products,
prepared according to various recipes.[52,94,95] All the treatises
require the exposition of copper plates to different acidic media
(urine, acetic acid, etc.) with a variety of additives (salt, honey,
etc.) to produce blue–green compounds that are now named
verdigris.[52,94] This material could be dissolved in resin to produce
the so-called copper resinate, appreciated for the use as a glaze.[94]
Nomenclature issues have been found: in fact a variety of names
have been used to identify the reaction product of metallic copper
and acetic acid, such as «viridis cupris», «aerugo» or «viride
salsum».[52,96] Raman spectroscopy proved successful in identifying
these compounds, with some significant differences related to the
excitation wavelength used.[97] The best overall results seem to be
obtained with a 785-nm laser (detection of both verdigris and
copper resinate, even when mixed with linseed oil).[97]
Verdigris
Seen the variety of reagents that have been listed in recipes from
the past as necessary to produce verdigris, it is clear that a variety
of (similar) chemical compounds have to be expected in the final
pigment. Different blue–green copper acetates have actually been
synthesized following the traditional recipes, and charac-
terized.[52,95,97,98] Raman spectroscopy has shown good results in
this sense,[95–98] with the successful identification of different
J. Raman Spectrosc. (2016)
Raman spectroscopy of green compounds in Cultural Heritage research
copper acetates, Cux(CH3COO)y(OH)z · nH2O[98] and other com-
pounds (cuprite, atacamite, etc.[95,96]). These latter materials seem
to disappear after recrystallization/purification of the pigment, as
they are converted into acetates.[96] Reference values for different
acetates can be found in literature[39,62,95,99]: it seems that verdigris
is always a mixture of compounds,[95,96] not necessarily related
to the pigment synthesis.[95] The most characteristic bands
which are used for successful identification of this copper
pigment are related to methyl groups (stretch: 2930–3020 cm 1;
bend: 1350–1450 cm 1; rock: 1050–1060 cm 1) and to
carboxylates (antisymmetric and symmetric stretch: 1540–1650 and
1260–1450 cm 1); the ν(C―C) stretching vibration is observed at
930–940 cm 1; the δ(O―C―O) deformations at 600–680 cm 1.[98]
In case OH groups are present, bands appear at 3050–
3650 cm 1.[98] Slightly different band positions and assignments
can be found in.[39,97] Also, the results from the study of the
pigments themselves are not straightforwardly applicable to real
samples, as the bands of binding media overlap with the organic
part of these greens,[97] and binding media alteration because of
ageing can also have an effect.[100] The best results seem to be
obtained both with the 532 and 785-nm lasers, for the identifica-
tion of the pigment, even mixed with linseed oil.[97]
In the present study, Raman spectra of verdigris (44450 Kremer)
were recorded both with the 785 and 532-nm lasers. In both cases
the identification was successful, even if the spectrum recorded
with the red laser is noisier and shows only a few diagnostic bands
(948, 320, 180 cm 1[97]). On the other hand, with the green laser
(Fig. 5, spectrum a) it is possible to better understand the chemistry
of copper acetates present in the sample. When comparing our
spectrum with published values, the greatest similarity is observed
with samples B2 Cu(CH3COO)2 · Cu(OH)2 · 5H2O and C Cu(CH3COO)
[95]
2 · 2(Cu(OH)2) published by Chaplin et al., with specific bands at
1057 and 2695 cm 1 pointing to each of the two compounds
respectively.
Copper resinate
Copper resinate is a transparent green glazing material obtained by
mixing resinous materials with verdigris.[97] Copper salts of abietic
acid are considered diagnostic for a positive identification of this
material.[101,102] Such a pigment has been identified in artworks
Figure 5. Raman spectra of Cu containing artificial green pigments. a:
verdigris (Kremer 44450, 532 nm). b: Cu resinate Kremer 1220 (532 nm)
3 × 60 s. c: Egyptian green (Kremer 11100, 532 nm). d: Tridymite SiO2
R090042. e: Co green (Kremer 4410, 532 nm).
J. Raman Spectrosc. (2016) Copyright © 2016 John Wiley & Sons, Ltd.
from different cultures and time periods.[23,52,94] However, it has
to be noted that copper resinate can also be a product of interac-
tion of copper pigments, such as verdigris, with organic binding
media.[97] Therefore, it is difficult to conclude if the identified
copper compound was prepared as copper resinate before its
application, or if it results from the alteration of a Cu containing
painting layer. The 1220 Kremer sample was brittle and naturally
aged in our laboratories (Fig. 5, spectrum b). Two more naturally
aged reference materials prepared in KIK-IRPA laboratories were
also studied. The sample KIK-CR1 was prepared in 1981 using the
resin from Abies pectinata, while KIK-CR2 (1973) is a mixture of
copper resinate and oil. Also these samples were brittle, but the
second one was sticky, probably as a result of the oil component.
Because of the glaze-nature of this pigment, its identification is
difficult. As regards Raman spectroscopy, the main issue is
fluorescence.[97] In comparison to verdigris, the Raman spectrum
of copper resinate is showing bands of ν(C―C) stretch (1600–
1700 cm 1).[97] The increase in the ratio 1612/1648 cm 1 can sug-
gest a higher aromatic unsaturation or slower oxidation of abietic
acid.[97]
The best laser wavelengths for studying copper resinate, accor-
ding to a recently published systematic study,[97] are 785 and
830 nm. A Raman spectrum attributable to a copper containing
organo-complex from an illuminated manuscript has been
recorded with 633-nm source.[100]
No results could be obtained with the 785-nm laser, while green
lasers (532 nm and 514 nm) provided better results. However, the
spectra recorded on sample CR-KIK2 were overwhelmed by fluores-
cence (probably related to the oily component) and no bands could
be observed. The band positions are in good agreement with pub-
lished values[97] for both 532 and 830-nm lasers (Fig. 5, spectrum b).
Differences in acquisition time and accumulations have to be
pointed out and could explain the absence of some Raman
bands[97] and the change in the Raman profile of some bands.
Moreover, different instruments, having different setups, will give
slightly different results in term of Raman spectra. These aspects
will be further discussed in the next paragraphs. A very strong,
broad, asymmetric band with maximum at 2933 cm 1 is attributed
to the antisymmetric stretch of ν(CH2) and ν(CH3) units (with the
symmetric stretching visible at ca. 2870 cm 1); the ν(COO) stretching
vibration is also intense, broad and asymmetric, with a maximum at
1614 cm 1, while its in plane bending is at 709 cm 1. Bands related
to CH3 vibrations are observed at 1445 and 1074 cm 1. ν(C―C)
stretching and δ(C―H) wagging modes are observed respectively
at ca. 970 and ca. 770 cm 1. Some features attributable to copper
acetates are those at 1410 and 120 cm 1.[97] Lattice vibrations in-
volving Cu might explain the band at 437 cm 1, which is found as
well in malachite.[39,49,59] The band at 202 cm 1[97,100] is however
missing from our spectra, unless it is corresponding to the one at
220 cm 1, therefore displaying a significant Raman shift which
could be explained by compositional differences or by ageing.
Egyptian green
The synthetic pigment Egyptian green was obtained from different
proportions of the same ingredients used for producing Egyptian
blue.[52,103–105] This latter is a calcium copper silicate of formula
CaCuSi4O10, while the green material has been defined as copper
containing wollastonite (Ca,Cu)3Si3O9.[52,103–105] The insertion of
ca. 2% Cu in the formula, however, does not allow for distinction
between the Raman spectra of wollastonite and cupro-
wollastonite.[103] There are no ancient recipes available for this
green pigment production.[103]
wileyonlinelibrary.com/journal/jrs

It is known that differences in the production process of this
pigment (firing temperature, atmosphere, cooling rate[105]) influ-
ence the resulting material; however, a silicatic enriched compound
is expected as a final product.
The Raman spectra recorded on the commercial sample of Egyp-
tian green (10064, Kremer, Fig. 5, spectrum c), however, show differ-
ent compounds, such as monoclinic tridymite (bands at 210, 302,
355, 433, 789 and 1075 cm 1,[106] Fig. 5, spectrum d; Table 3) and
Cu-containing materials. Tridymite is reported to be the dominant
crystalline phase in recently produced pigment produced at tem-
peratures higher than 950 °C.[104] The weak bands at 406, 391,
253, 142 and 110 cm 1 may be assigned respectively to ν(Cu―Cl)
stretching,[45] symmetric stretching ν(O―Cu―O),[107] bending
δ(O―Cu―O)[107] and bending δ(O―Cu).[45,107]
Raman spectra of wollastonite are published mainly from a
mineralogical point of view,[108] and correspondences are observed
with archaeological samples from Egypt.[103]
Pigments based on cobalt or chromium
Cobalt containing greens are described as zincates or titanates,
commercially sold as pigment green 50 (PG 50).[52,109] The studied
sample (44100 Kremer, Fig. 5, spectrum e) is described as cobalt
titanate green spinel; however, this name may apply to a variety
of compositions and structures.[109] Cobalt titanate green pigments
show shifts in the main Raman band position, which is the one at
ca. 700 cm 1. This band shows as well broadening. This depends
on the presence of other metallic ions in the lattice.[109] This band
is attributed to the symmetric stretching of CoO6 octahedra in
CoTiO3, which suggests that it might be assigned to the same
octahedral vibration in Co2TiO4 green pigments.[109] Other weaker
bands are reported in Table 3, which are in reasonable agreement
with published values of reference cobalt greens,[109] considering
the already mentioned band shifts. The observed band positions
at 1135, 982, 620 and 460 cm 1 suggest the presence of an
extender of the family of sulphates (possibly epsomite
MgSO4 · 7H2O[65]). It is known that such additives have been used
for the production of modern pigments, but this finding is in
contrast with what previously identified in cobalt green (44100
Kremer), which was barite (BaSO4).[109] A cobalt titanate green
pigment has been recently identified in a 20th century painting.[109]
Chromium oxide greens were produced starting from the 19th
century and following various recipes.[52,110] The end product is
however chromium (III) oxide, anhydrous or dihydrated (viridian
Cr2O3 · 2H2O).[52,110] ‘Chrome green’ was also used as a name for
the mixture of Prussian blue and chrome yellow,[110–113] while
viridian was often mixed with barium sulphate, which negatively
affects the saturation of the pigment but makes it easier to disperse
it in binding media.[110] Notable occurrences of chrome oxide
greens can be found listed in literature;[110] however, Raman spec-
troscopic identification is more limited.[114]
The structure of Cr2O3 corresponds to the mineral eskolaite, and
Raman spectra of chromium oxide are published both in relation to
geology and archaeometry.[39,40,110,114–118]
The most intense Raman band of Cr2O3 (Figure S2, spectrum a) is
the A1g mode at ca. 550 cm 1,[119] and weaker bands are observed
at 597, 391, 340 and 304 cm 1 with the 532-nm laser for both
chrome oxide greens studied (44200 and 44204 Kremer; Table 3).
These band positions are in good agreement with published
data.[47,117] With the 785-nm one, bands are observed at 613, 555,
351 and 299 cm 1.[47,114] It seems important to highlight that the
position of the main Raman band of chromium oxide shifts accor-
ding to the used laser (532 nm: 542 cm 1; 785 nm: 555 cm 1).
wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd.
A. Coccato et al.
The term viridian refers to the hydrated form Cr2O3 · 2H2O.[47,52,53]
It seems that the pigment viridian (4425 Kremer) is a mixture of
anhydrous and hydrated chromium oxides, as the recorded Raman
bands (532 nm, Figure S2, spectra b, c and d) at 583, 552, 487 and
266 cm 1 are attributed to the hydrated oxide, and those at 543,
345, 307 and 266 cm 1 to the anhydrous one.[116]
Synthetic organic pigments
Green synthetic organic pigments are based on a variety of
molecular structures, such as phtalocyanines, azo and monoazo
pigments. The pigments studied during this research are the
polycyclic phtalocyanine greens PG 7 and PG 36 (Cu containing,
chlorinated and brominated, respectively) and PG 8 (Fe containing,
azo metal complex).[29] The detection of such pigments in archaeo-
logical objects allows the identification of recent restoration treat-
ments or forgeries,[8,120] as these pigments were synthesized
starting from the beginning of the 20th century.[52]
The Raman spectra of the phtalocyanine greens PG 7 and PG 36
were successfully recorded both with the red and the green lasers,
showing different relative intensities for some of the bands (see
Table 3). The PG 7 spectrum (Fig. 4, spectrum d) is in good agree-
ment with published values,[29,121] even if the spectra are recorded
with a different source (488 and 514 nm respectively). In fact, some
bands present a downshift of ca. 5 cm 1 when compared with
those references,[29,121] and some bands reported for the 488-nm
laser excitation are also visible in our 532-nm excited spectrum
(642, 506, 368, 346, 332, 163 and 146 cm 1[121]) and some in the
785-nm spectrum (740, 345, 331, 288, 165, and 147 cm 1[121,122]).
The band at 740 cm 1 is the most intense in the Raman spectrum
acquired with the 785-nm laser in this study, while another
published spectrum displays a very strong 1560 cm 1 band.[122]
PG 36 is a copper phtalocyanine containing chlorine and
bromine as well. Our spectra show generally a good agreement
with published values.[29,122] Again, different lasers give slightly dif-
ferent spectra (Fig. 6, spectra a and b). The band positions and in-
tensities recorded with both lasers are reported in Table 3 and are
in good agreement with published values.[29,122] A downshift in
some bands is observed when comparing different green lasers,
as the published values (514 nm) are 1539 and 1498 cm 1, and
the observed ones (532 nm) are 1527 and 1490 cm 1. PB 15 is also
detected by the bands at 1528, 950 and 748 cm 1.[29]
Figure 6. Raman spectra of synthetic organic green pigments. a: PG 36
(532 nm). b: PG 36 (785 nm). c: reference spectrum of PG 36 (785 nm[122]).
d: PG 8 (785 nm).
J. Raman Spectrosc. (2016)
Raman spectroscopy of green compounds in Cultural Heritage research
PG 8 is based on Fe and azo groups. Our sample is in good agree-
ment with the only reference spectrum published so far (Fig. 6,
spectrum d), with a 785-nm laser,[122] while some differences are
observed with the 532-nm excited spectrum. To the best of our
knowledge, we are the first ones to report the Raman spectrum of
PG 8 with a 532-nm excitation (Table 3).
Raman spectroscopic identification of copper resinate in a cross
section
It has been reported in literature that the behaviour of copper
resinate as a pure pigment and in real samples, together with
oleo-resinous binders and other components, is different.[97] During
this research we had the opportunity to investigate a sample from
the Ghent Altarpiece. This masterpiece of Early Netherlandish art
(Jan and Hubert van Eyck, second quarter of 15th century[123]) is cur-
rently being restored and investigated for conservation purposes. A
sample taken from a green area of panel XIII (Archangel and
Prophet Zacharias) during the conservation campaign of years
1951–1952 was re-investigated. The stratigraphy of the sample is
complex, and shows the superposition of green layers on top of a
greyish imprimatura. The lowest layer (layer 4) shows a bluish green
hue, the above one (layer 5) is light green. Finally, a semitransparent
green layer (layer 6) is applied before a yellow glacis. Copper
resinate has been identified in this sample, as well as in six others
taken from the panels IX, X, XII and XX of the Ghent Altarpiece.
The identification was performed in 1979 by Kockaert in the KIK-
IRPA laboratories, using microchemistry. Copper resinate was
identified also in other 27 samples from various Netherlandish
paintings.[124]
In this study, micro Raman analysis was carried out in order to
verify the presence of copper resinate in the cross section and to
evaluate if any difference could be observed between spectra
recorded with different green lasers as excitation sources (514
and 532 nm, see Fig. 7). For this last purpose, care was taken in
order to acquire the spectra in the same spots with the different
instruments. The results obtained with the two instruments proved
comparable, even if the quality of the recorded spectra is not the
same. Differences in acquisition time and accumulations have to
be pointed out and could explain the observation of additional
Figure 7. Raman spectra of a Ghent Altarpiece cross section. a. C10.020,
layer 5 (532 nm) 10 × 150 s. b. C10.020, layer 5 (514 nm) 10 × 40 s. c. Cu
resinate (Kremer 1220, 514 nm). d. Cu resinate Kremer 1220 (532 nm)
3 × 60 × 100 s. The bands marked with dashed lines correspond to lead tin
1
yellow type I (195 and 128 cm ).
J. Raman Spectrosc. (2016) Copyright © 2016 John Wiley & Sons, Ltd.
Raman bands and some changes in the band profile, see for exam-
ple: 3 × 60 s[97] (Fig. 5, spectrum b) versus 60 × 100 s (Fig. 7, spectrum
d). Also, different Raman spectrometers will affect the results, as a
consequence of the different setups (Fig. 7: spectra a and b are from
the sample, spectra c and d from the reference material). It is there-
fore of extreme importance to be careful when comparing spectra
with published data. We are the first ones to report the Raman
spectrum of copper resinate with a 514-nm laser excitation (Fig. 7
spectrum c)
In layer 4, indigo (1706, 1584–1570, 1364, 1310, 1252, 1148,
598 cm 1) and lead tin yellow type I (129 and 196 cm 1) were iden-
tified. Once again, the mixture of yellow and blue to obtain green is
noted. The Raman spectrum of indigo seems better excited with
the 514-nm laser.
Layers 5 and 6 contain both copper resinate (identified by the
features at ca. 2900 cm 1 and that at ca. 1610 cm 1). It is important
to mention that the ν(CH2) and ν(CH3) broad stretching band at
2900 cm 1 can be related to the oily binding medium, and that
the most diagnostic feature of copper resinate is that of ν(CC)
stretching vibrations at ca. 1610 cm 1.[97] In some spectra,
additional bands at 1437, 1305 and 700 cm 1, also referable to Cu
resinate δ(CH2) bending and ν(CC) stretching modes, were detected
with both lasers (Fig. 7, spectra a and b.). In addition to copper
resinate, in layer 5 also lead tin yellow type I was identified by the
bands at 195 and 128 cm 1[39] (Fig. 7, spectra a and b).
In layers 4 and 5 an intense broad band centred at 549 cm 1 was
recorded with both instruments. The assignment of this band to
lazurite was considered at first; however, various aspects contradict
this hypothesis. First, the lack of the medium intensity band at
1096 cm 1 does not allow for a reasonable attribution to this blue
pigment.[39] Also, considering the high price and symbolic value
of this blue pigment, especially during the Middle Ages, its use in
the green area of a character of secondary importance (compared
to the Virgin Mary, whose vest was traditionally painted with lapis
lazuli), seems suspicious.[52] This band could not be assigned.
Moreover, it is important to note that the detection of Cu resinate
mixed with drying oils was never published so far with green lasers.
Both the 514 and 532-nm lasers were considered not suitable for
the identification of pure copper resinate or mixed with oleoresin-
ous binding media.[97]
Conclusions
The present study highlighted some remarkable aspects regarding
the study of green materials of interest for the investigation of cul-
tural heritage.
An extensive literature survey has been performed regarding the
potentialities of Raman spectroscopy in identifying different green
pigments in a variety of objects, as well as to find published spectra
for comparison purposes. By doing this, it appeared clear that not
all of the necessary reference spectra were available, and that the
low wavenumber (below ca. 200 cm 1) and ν(OH) stretching (above
3400 cm 1) regions are not always considered, although they pro-
vide many interesting features useful for spectral identification.
Therefore, the extended range spectra have been discussed in this
paper, and all the spectra collected on reference materials during
this research can be requested online at http://www.analchem.
ugent.be/RAMAN/.
By means of Raman spectroscopy, several commercial green pig-
ments have been fully characterized. It was possible to define which
wileyonlinelibrary.com/journal/jrs

‘green earth’ was present and to identify impurities and additives,
thanks to the combined use of different excitation wavelengths.
In the case of Cu containing materials, most of the suggested
identifications were confirmed, but in some cases a new interpreta-
tion was suggested on the basis of non-destructive analytical
techniques.
The traditional pigment copper resinate has been investigated
both as reference materials and in a real sample, with different
green lasers, which both proved effective in detecting it, even if
the measurement conditions and instrumental parameters may
affect the quality of the recorded spectra.
The study of modern pigments also allowed the collection of
reference spectra with different wavelengths of some less usual
materials, which increases the possibilities of successful identifica-
tion of pigments in forthcoming studies.
The successful identification of copper resinate in cross sections
from masterpieces was achieved thanks to the comparison of the
recorded spectra with those of reference materials. However, this
comparison needs to be careful, as differences in the acquisition
of the Raman spectra affect the result of the measurements. In this
case, green lasers of different wavelength (532 and 514 nm) proved
effective in detecting the synthetic pigment copper resinate in a
cross section from an aged oil painting.
Acknowledgements
The authors thank Ghent University for its financial support through
the concerted research actions (GOA) programme. Thanks also go
to the Gieskes & Strijbis Foundation for the financial support of
the scientific research on the Ghent Altarpiece materials and
techniques. Sylvia Lycke, Debbie Lauwers and Cecile Glaude are
also acknowledged for their support during the project.
References
[1] A. Jurado-López, O. Demko, R. J. H. Clark, D. Jacobs, J. Raman Spectrosc.
2004, 35, 119.
[2] R. J. H. Clark, R. R. Hark, N. Salvadó, S. Butí, T. Pradell, J. Raman
Spectrosc. 2010, 41, 1418.
[3] N. Buzgar, A. Buzatu, A. Apopei, V. Cotiugă, Vib. Spectrosc. 2014, 72,
142.
[4] L. M. Moretto, E. F. Orsega, G. A. Mazzocchin, J. Cult. Herit. 2011, 12,
384.
[5] B. Gilbert, S. Denoël, G. Weber, D. Allart, Analyst 2003, 128, 1213.
[6] A. Lo Monaco, E. Mattei, C. Pelosi, M. Santancini, J. Cult. Herit. 2013, 14,
537.
[7] H. G. M. Edwards, R. Wolstenholme, D. S. Wilkinson, C. Brooke,
M. Pepper, Anal. Bioanal. Chem. 2007, 387, 2255.
[8] I. Aliatis, D. Bersani, E. Campani, Spectrochim. Acta Part A Mol. Biomol.
Spectrosc. 2009, 73, 532.
[9] S. Valadas, R. Freire, Microsc. Microanal. 2015, 21, 518.
[10] R. Feller, Artists’ Pigments: Volume 1: A Handbook of their History and
Characteristics, National Gallery of Art and Archetype Publications
Ltd., 1987.
[11] F. Ospitali, D. Bersani, G. Di Lonardo, P. P. Lottici, J. Raman Spectrosc.
2008, 39, 1066.
[12] V. Hayez, V. Costa, J. Guillaume, H. Terryn, A. Hubin, Analyst 2005, 130,
550.
[13] P. Ropret, T. Kosec, J. Raman Spectrosc. 2012, 43, 1578.
[14] C. Chiavari, K. Rahmouni, H. Takenouti, Electrochimica 2007, 52, 7760.
[15] G. Bertolotti, D. Bersani, P. Lottici, Anal. Bioanal. Chem. 2012, 402, 1451.
[16] W. Martens, R. Frost, J. T. Kloprogge, P. A. Williams, J. Raman Spectrosc.
2003, 34, 145.
[17] A. Derbyshire, R. Withnall, J. Raman Spectrosc. 1999, 30, 185.
[18] D. A. Scott, Stud. Conserv. 2000, 45, 39.
[19] K. Eremin, J. Stenger, M. Li Green, J. Raman Spectrosc. 2006, 37, 1119.
wileyonlinelibrary.com/journal/jrs Copyright © 2016 John Wiley & Sons, Ltd.
A. Coccato et al.
[20] J. R. J. Asperen de Boer, Examen scientifique des peintures du groupe
Jan van Scorel, Musee Central d’Utrecht and Musee de la Chartreuse,
Douai, 1977.
[21] J. Sanyova, S. Saverwyns, Sci. Artis 2006, 3, 259.
[22] J. Klipa, A. Pokorny, J. Sanyova, Prepr. Symp. ALMA 24th–25th
Novemb. 2010, 2010, 189.
[23] H. Kühn, Artists’ Pigments: A Handbook of Their History and
Characteristics, National Gallery of Art, Oxford University Press, New
York vol. 2, 1993 p, 131.
[24] M. Irazola, M. Olivares, K. Castro, M. Maguregui, I. Martínez-Arkarazo,
J. M. Madariaga, J. Raman Spectrosc. 2012, 43, 1676.
[25] A. Dominguez-Vidal, M. J. de la Torre-López, M. J. Campos-Suñol,
R. Rubio-Domene, M. J. Ayora-Cañada, J. Raman Spectrosc. 2014, 45, 1006.
[26] V. Košařová, D. Hradil, I. Nemec, P. Bezdicka, V. Kanicky, J. Raman
Spectrosc. 2013, 44, 1570.
[27] P. Vandenabeele, L. Moens, J. Raman Spectrosc. 2012, 43, 1545.
[28] S. V. Sharma, A. K. Misra, S. M. Clegg, J. E. Barefield, C. Wiens, T. Acosta,
Philos. Trans. R. Soc. A 2010, 368, 3167.
[29] N. C. Scherrer, Z. Stefan, D. Francoise, F. Annette, K. Renate,
Spectrochim. Acta. A. Mol. Biomol. Spectrosc. 2009, 73, 505.
[30] R. Frost, S. Bahfenne, J. Čejka, J. Raman Spectrosc. 2010, 41, 814.
[31] P. Gillet, C. Biellmann, B. Reynard, P. McMillan, Phys. Chem. Miner.
1993, 20, 1.
[32] S. Gunasekaran, G. Anbalagan, S. Pandi, J. Raman Spectrosc. 2006, 37,
892.
[33] R. Frost, Clays Clay Miner. 1995, 43, 191.
[34] A. Coccato, J. Jehlicka, L. Moens, P. Vandenabeele, J. Raman Spectrosc.
2015, 46, 1003.
[35] A. Buzatu, N. Buzgar, Geologie 2010, LVI, 1.
[36] L. Burgio, R. J. Clark, Spectrochim. Acta Part A Mol. Biomol. Spectrosc.
2001, 57, 1491.
[37] H. Edwards, S. Villar, Spectrochim. Acta Part A Mol. Biomol. Spectrosc.
2005, 61, 2273.
[38] P. Falaras, A. H.-L. Goff, Sol. energy Mater. Sol. Cells 2000, 64, 167.
[39] I. M. Bell, R. J. H. Clark, P. J. Gibbs, Spectrochim. Acta Part A Mol. Biomol.
Spectrosc. 1997, 53, 2159.
[40] M. Bouchard, D. C. Smith, Spectrochim. Acta Part A Mol. Biomol.
Spectrosc. 2003, 59, 2247.
[41] R. Frost, S. Palmer, J. Bouzaid, J. Reddy, J. Raman Spectrosc. 2007, 38,
68.
[42] Pliny the Elder, Natural History, Loeb Class, Harvard University Press,
Cambridge, 1938.
[43] L. Dei, A. Ahle, P. Baglioni, D. Dini, E. Ferroni, Stud. Conserv. 1998, 43,
80.
[44] T. Kosec, P. Ropret, A. Legat, J. Raman Spectrosc. 2012, 43, 1587.
[45] R. L. Frost, W. Martens, J. T. Kloprogge, P. A. Williams, J. Raman
Spectrosc. 2002, 33, 801.
[46] A. G. Shen, X. H. Wang, W. Xie, J. Shen, H. Y. Li, Z. A. Liu, J. M. Hu,
J. Raman Spectrosc. 2006, 37, 230.
[47] K. Castro, M. Pérez-Alonso, M. D. Rodríguez-Laso, L. A. Fernández,
J. M. Madariaga, Anal. Bioanal. Chem. 2005, 382, 248.
[48] K. Castro, A. Sarmiento, I. Martinez-Arkarazo, J. M. Madariaga,
L. A. Fernandez, Anal. Chem. 2008, 80, 410.
[49] B. Yu, J. Fang, E. Huang, J. Raman Spectrosc. 2013, 44, 630.
[50] L. McCann, K. Trentelman, T. Possley, B. Golding, J. Raman Spectrosc.
1999, 30, 121.
[51] R. Gettens, E. Fitzhugh, Stud. Conserv. 1974, 19, 2.
[52] N. Eastaugh, V. Walsh, T. Chaplin, R. Siddall, Pigment Compendium,
Routledge, Oxford, 2004 p, 2008.
[53] A. M. Correia, R. J. H. Clark, M. I. M. Ribeiro, M. L. T. S. Duarte, J. Raman
Spectrosc. 2007, 38, 1390.
[54] L. Appolonia, D. Vaudan, V. Chatel, M. Aceto, P. Mirti, Anal. Bioanal.
Chem. 2009, 395, 2005.
[55] L. Burgio, R. J. H. Clark, R. R. Hark, Proc. Natl. Acad. Sci. U. S. A. 2010, 107,
5726.
[56] J. Van Pevenage, D. Lauwers, D. Herremans, E. Verhaeven,
B. Vekemans, W. De Clercq, L. Vincze, L. Moens, P. Vandenabeele,
Anal. Meth. 2014, 6, 387.
[57] D. Lauwers, V. Cattersel, L. Vandamme, A. Van Eester, K. De Langhe,
L. Moens, P. Vandenabeele, J. Raman Spectrosc. 2014, 45, 1266.
[58] F. Stranges, M. La Russa, A. Oliva, G. Galli, J. Archaeol. 2014, 2014,
312981.
[59] R. Frost, W. Martens, J. Raman Spectrosc. 2002, 33, 252.
[60] B. Guineau, Stud. Conserv. 1984, 29, 35.
[61] N. Buzgar, A. Apopei, Geologie 2009, LV, 2.
J. Raman Spectrosc. (2016)
Raman spectroscopy of green compounds in Cultural Heritage research
[62] N. De Laet, S. Lycke, J. Van Pevenage, L. Moens, P. Vandenabeele, Eur.
J. Mineral. 2013, 25, 855.
[63] R. L. Frost, Y. Xi, Vib. Spectrosc. 2013, 64, 33.
[64] R. Downs, 2006. [Online]. Available: https://www.zotero.org/jeleibold/
items/itemKey/TD7TUW5D/itemPage/2. [Accessed: 02-Apr-2015].
[65] P. Makreski, G. Jovanovski, S. Dimitrovska, Vib. Spectrosc. 2005, 22, 25.
[66] V. Hayez, J. Guillaume, A. Hubin, H. Terryn, J. Raman Spectrosc. 2004,
35, 732.
[67] M. Schmidt, H. Lutz, Phys. Chem. Miner. 1993, 20, 27.
[68] B. Creton, D. Bougeard, K. Smirnov, J. Guilment, O. Poncelet, J. Phys.
Chem. C 2008, 112, 358.
[69] J. Phillips, Phys. Rev. B 1985, 32, 5350.
[70] Q. Sun, Vib. Spectrosc. 2009, 51, 213.
[71] S. Kharbish, P. Andráš, J. Luptáková, S. Milovská, Spectrochim. Acta. A.
Mol. Biomol. Spectrosc. 2014, 130, 152.
[72] R. L. Frost, P. A. Williams, W. Martens, J. T. Kloprogge, J. Raman
Spectrosc. 2002, 33, 260.
[73] M. M. Naumova, S. A. Pisareva, G. O. Nechiporenko, Stud. Conserv.
1990, 35, 81.
[74] M. Fabrizi, H. Ganiaris, S. Tarling, D. A. Scott, Stud. Conserv. 1989, 34, 45.
[75] R. Garcia Moreno, F. Mathis, V. Mazel, M. Dubus, T. Calligaro, D. Strivay,
Archaeometry 2008, 50, 658.
[76] A. Belik, H. Koo, M. Whangbo, N. Tsujii, P. Naumov,
E. Takayama-Muromachi, Inorg. Chem. 2007, 46, 8684.
[77] R. Frost, J. Sejkora, E. Keeffe, J. Raman Spectrosc. 2010, 41, 202.
[78] K. Castro, A. Sarmiento, M. Pérez-Alonso, J. M. Madariaga, E. Princi,
S. Vicini, E. Pedemonte, M. D. Rodríguez-Laso, TrAC Trends Anal.
Chem. 2007, 26, 347.
[79] M. Pérez-Alonso, K. Castro, J. M. Madariaga, Anal. Chim. Acta 2006,
571, 121.
[80] M. Spring, Natl. Gall. Tech. Bull. 2000, 21, 20.
[81] L. Zaharia, Stud. UBB, Geol. 2012, XLVIII, 77.
[82] R. Frost, P. Williams, W. Martens, Am. Mineral. 2004, 89, 1130.
[83] R. L. Frost, P. A. Williams, W. Martens, J. T. Kloprogge, J. Raman
Spectrosc. 2002, 33, 752.
[84] R. Frost, S. Palmer, E. Keeffe, J. Raman Spectrosc. 2010, 41, 1769.
[85] G. Bartholomäi, W. E. Klee, Spectrochim. Acta Part A Mol. Spectrosc.
1978, 34, 831.
[86] C. Engelbrekt, P. Malcho, J. Andersen, J. Nanoparticle Res. 2014, 16,
2562.
[87] J. Jambor, Can. Mineral. 1996, 34, 61.
[88] M. Fleet, Acta Crystallogr. Sect. B 1975, 31, 183.
[89] E. E. Lawson, A. N. C. Anigbogu, A. C. Williams, B. W. Barry,
H. G. M. Edwards, Spectrochim. Acta. A. Mol. Biomol. Spectrosc. 1998,
54, 543.
[90] D. Hradil, T. Grygar, J. Hradilova’, B. Petr, Appl. Clay Sci. 2003, 22, 223.
[91] O. Cristini, C. Kinowski, S. Turrell, J. Raman Spectrosc. 2010, 41, 1410.
[92] A. Tlili, D. Smith, J.-M. Beny, H. Boyer, Mineral. Mag. 1989, 53, 165.
[93] A. Wang, J. Freeman, B. Jolliff, J. Raman Spectrosc. 2015, 46, 829.
[94] H. Kühn, Stud. Conserv. 1970, 15, 12.
[95] T. D. Chaplin, R. J. H. Clark, D. A. Scott, J. Raman Spectrosc. 2006, 37, 223.
[96] J. De la Roja, V. Baonza, M. S. Andrés, Spectrochim. Acta Part A Mol.
Biomol. Spectrosc. 2007, 68, 1120.
[97] C. Conti, J. Striova, I. Aliatis, E. Possenti, G. Massonnet, C. Muehlethaler,
T. Poli, M. Positano, J. Raman Spectrosc. 2014, 45, 1186.
J. Raman Spectrosc. (2016) Copyright © 2016 John Wiley & Sons, Ltd.
[98] M. San Andrés, J. M. de la Roja, V. G. Baonza, N. Sancho, J. Raman
Spectrosc. 2010, 41, 1468.
[99] D. C. Pereira, D. L. A. De Faria, V. R. L. Constantino, J. Braz. Chem. Soc.
2006, 17, 1651.
[100] I. Nastova, O. Grupče, B. Minčeva-Šukarova, S. Turan, M. Yaygingol,
M. Ozcatal, V. Martinovska, Z. Jakovlevska-Spirovska, J. Raman
Spectrosc. 2012, 43, 1729.
[101] M. P. Colombini, G. Lanterna, A. Mairani, M. Matteini, F. Modugno,
M. Rizzi, Ann. Chim. 2001, 91, 749.
[102] M. Abdel-Ghani, H. G. M. Edwards, B. Stern, R. Janaway, Spectrochim.
Acta. A. Mol. Biomol. Spectrosc. 2009, 73, 566.
[103] S. Pagès-Camagna, S. Colinart, C. Coupry, J. Raman Spectrosc. 1999, 30,
313.
[104] P. Bianchetti, F. Talarico, M. G. Vigliano, M. F. Ali, J. Cult. Herit. 2000, 1,
179.
[105] S. Pagès-Camagna, S. Colinart, Archaeometry 2003, 45, 637.
[106] K. Kihara, T. Hirose, K. Shinoda, J. Mineral. Petrol. Sci. 2005, 100, 91.
[107] X.-D. Liu, M. Hagihala, X.-G. Zheng, Q.-X. Guo, Vib. Spectrosc. 2011, 56,
177.
[108] E. Huang, C. Chen, T. Huang, E. Lin, J. Xu, Am. Mineral. 2000, 85, 473.
[109] F. Casadio, A. Bezúr, I. Fiedler, K. Muir, T. Trad, S. Maccagnola, J. Raman
Spectrosc. 2012, 43, 1761.
[110] E. Fitzhugh, Artists’ Pigments: A Handbook of Their History and
Characteristics, 3rd ed., vol. 3, National Gallery of Art, Oxford
University Press, New York, 1997.
[111] V. Desnica, K. Furic, B. Hochleitner, M. Mantler, Spectrochim. Acta Part B
At. Spectrosc. 2003, 58, 681.
[112] M. Abdel-Ghani, H. G. M. Edwards, R. Janaway, B. Stern, Vib. Spectrosc.
2008, 48, 69.
[113] F. Rosi, C. Miliani, I. Borgia, B. Brunetti, A. Sgamellotti, J. Raman
Spectrosc. 2004, 35, 610.
[114] P. Ropret, S. A. Centeno, P. Bukovec, Spectrochim. Acta. A. Mol. Biomol.
Spectrosc. 2008, 69, 486.
[115] M. Bouchard, Spectrochim. Acta Part A Mol. Biomol. Spectrosc. 2007, 68,
1101.
[116] J. Maslar, W. Hurst, W. Bowers, Appl. Surf. Sci. 2001, 180, 102.
[117] E. von Aderkas, M. Barsan, Spectrochim. Acta Part A-Molecular Biomol.
Spectrosc. 2010, 77, 954.
[118] F. Hardcastle, I. Wachs, J. Mol. Catal. 1988, 46, 173.
[119] V. Sammelselg, A. Tarre, J. Lu, J. Aarik, A. Niilisk, T. Uustare, I. Netšipailo,
R. Rammula, R. Pärna, A. Rosental, Surf. Coatings Technol. 2010, 204, 2015.
[120] L. Burgio, R. J. H. Clark, J. Raman Spectrosc. 2000, 31, 395.
[121] F. Schulte, K.-W. Brzezinka, K. Lutzenberger, H. Stege, U. Panne,
J. Raman Spectrosc. 2008, 39, 1455.
[122] W. Fremout, S. Saverwyns, J. Raman Spectrosc. 2012, 43, 1536.
[123] P. Coremans, Les Primitifs Flamands III. Contribution a l’etude des
Primitifs Flamands 2. L’Agneau Mystique au Laboratoire, De Sikkel,
Antwerp, 1953.
[124] L. Kockaert, Stud. Conserv. 1979, 24, 69.
Supporting information
Additional supporting information may be found in the online ver-
sion of this article at the publisher’s web site.
wileyonlinelibrary.com/journal/jrs