Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 121 (2014) 415–419

Contents lists available at ScienceDirect

Spectrochimica Acta Part A: Molecular and

Biomolecular Spectroscopy
journal homepage: www.elsevier.com/locate/saa

Analytical Raman spectroscopy in a forensic art context:

The non-destructive discrimination of genuine and fake lapis lazuli
Esam M.A. Ali ⇑, Howell G.M. Edwards
Division of Chemical and Forensic Sciences, University of Bradford, Bradford BD7 1DP, UK

h i g h l i g h t s g r a p h i c a l a b s t r a c t

 Raman spectroscopy has been used to

differentiate between genuine and
fake lapis lazuli.
 The spectra were readily acquired in-
situ without alteration or destruction
of the sample.
 The results offer a special role for
portable Raman spectroscopy in
forensic art analysis.

a r t i c l e i n f o a b s t r a c t

Article history: The differentiation between genuine and fake lapis lazuli specimens using Raman spectroscopy is
Received 14 June 2013 assessed using laboratory and portable instrumentation operating at two longer wavelengths of excita-
Received in revised form 29 September 2013 tion in the near-infrared, namely 1064 and 785 nm. In spite of the differences between the spectra
Accepted 4 November 2013
excited here in the near infrared and those reported in the literature using visible excitation, it is clear
Available online 11 November 2013
that Raman spectroscopy at longer wavelengths can provide a means of differentiating between the fakes
studied here and genuine lapis lazuli. The Raman spectra obtained from portable instrumentation can
Keywords:
also achieve this result, which will be relevant for the verification of specimens which cannot be removed
Raman spectroscopy
Lapis lazuli
from collections and for the identification of genuine lapis lazuli inlays in, for example, complex jewellery
Forensic art analysis and furniture. The non-destructive and non-contact character of the technique offers a special role for
Fake specimens portable Raman spectroscopy in forensic art analysis.
Ó 2013 Elsevier B.V. All rights reserved.

Introduction in their visual appearance due to variations in the ratio of the

Lapis lazuli (Lat. Lapis lauzulus, blue stone), also known as
lazurite, is a blue semi-precious gem of the sodalite group which
also includes jade and opal. It is a rock and not a crystalline mineral
such as ruby, diamond and emerald. As such, it does not have a
precise chemical composition, although the pigment responsible
for its intense blue colour has been identified as ultramarine, of
formula (Na,Ca)8 (AlSiO4)6 (S,SO4,Cl)2; lazurite specimens can differ
⇑ Corresponding author. Tel.: +44 0788347177.
E-mail address: esmaroes@yahoo.co.uk (E.M.A. Ali).
yellow and blue ions, S2 and S3 respectively, which have low-ly-
ing electronic states at 390 nm and 610 nm in the indigo and red
visible regions of the electromagnetic spectrum. Particle size of
the chromophores and their distribution in the sodalite mineral
matrix also affects the depth of colour of the lapis lazuli. Lapis
lazuli, a mineralized limestone, occurs alongside other blue miner-
als such as hauyne [(Na,Ca)4 8 (AlSiO4)6(SO4)1 2], sodalite [Na8
(AlSiO4)6Cl2], noselite [Na8(AlSiO4)6 SO4], and associated minerals
such as calcite, CaCO3, pyrite, FeS2, diopside, CaMgSi2O6, and
zircon, ZrSiO4.
Lapis lazuli is known in many ancient cultures and highly prized
for its intense blue colour. The association of lapis lazuli with

1386-1425/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.saa.2013.11.005
416 E.M.A. Ali, H.G.M. Edwards / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 121 (2014) 415–419
clusters of golden coloured pyrites, which was often mistaken for
gold, was much appreciated despite the rarity of the material.
The most important source in the ancient world being Afghanistan
(particularly the Badakhasan mines), although others are now
known in Chile, China and Russia. Pliny the Elder referred to lapis
lazuli as ‘‘a fragment of the starry firmament’’ and it has long been
associated with peace of mind and tranquillity. A synthetic form of
lazurite, known as ultramarine blue, was first synthesised in 1828
and marked especially its wider use in art, which hitherto repre-
sented a very expensive outlay for illustrators of early manuscripts
and paintings [1,2]. The earliest use of lapis lazuli in beads and as
pendants can be attributed to the ancient Egyptian pre-dynastic
culture around 3500 BC and has been found as a blue pigment
on sarcophagi dating from at least 1800 BC and also more rarely
in Roman wall-paintings [3,4]. Generally, lapis lazuli pigment
was prepared from finely ground rock minerals, sometimes with
the addition of fine quartz river sand to assist in the pulverizing
procedure, and therefore contains analytical signatures of other
components of the rock matrix. In mediaeval paintings and fres-
coes, the use of expensive lapis lazuli was hierarchical and reserved
for the robes of only the most important figures, such as Christ and
the Virgin Mary in religious scenes.
The Raman spectrum of lapis lazuli has been reported by several
authors [5–9] and has been used to characterise this material in a
variety of works of art in historical and archaeological contexts.
The S3 radical anion primarily responsible for the blue colour in
lapis lazuli has three Raman-active fundamental vibrational bands
at 549 (symmetric stretching), 585 (asymmetric stretching) and
259 (deformation) cm 1; resonance enhancement of the Raman
spectrum is affected by laser excitation using green or red radia-
tion between 515 and 647 nm [10] and overtone bands of the very
strong intensity symmetric stretching vibration have been
reported out to a wavenumber shift of about 3000 cm 1. In
contrast, the Raman spectrum of the S2 radical anion also found
in lapis lazuli, which is yellow in colour, has an absorption maxi-
mum at 390 nm and this is not resonance enhanced with green
or red excitation; excitation at 442 nm does effect a resonance
enhancement of this species, and affords a series of overtones
out to 2900 cm 1. Previously, it has been noted that the Raman
spectrum of lazurite does exhibit signatures also of other compo-
nents of the mineral matrix, such as pyrites and calcite.
The Raman spectroscopic analysis of lapis lazuli specimens
from several locations has been reported and this database has
afforded an opportunity to assign these specimens to a geograph-
ical source, which offers an alternative and supporting method
for establishing provenance [11]. Favaro et al. [12] have recently
proposed an analytical methodology for the provenancing of lapis
lazuli pigments in works of art and for the discrimination between
specimens of a natural or synthetic origin based upon the charac-
terisation of lazurite mineral phases in stones from Afghan,
Siberian and Chilean sources, on the purified natural pigment
and on synthetic ultramarine. Infrared spectroscopy was found to
be unsuitable for distinguishing between materials from geograph-
ical sources but could differentiate between synthetic and natural
materials. In contrast, scanning electron microscopy and energy
dispersive analysis (SEM/EDAX) and Xray diffraction (XRD)
techniques in combination with chemometric analysis were able
to show the presence of specific mineral phases associated with
lapis lazuli and was more successful in determination of geograph-
ical provenancing issues but were found to be not so effective in
the identification of synthetic specimens.
As perhaps expected for such a valuable commodity, lapis la-
zuli has also been faked: a process that involves the dyeing of
cheaper, more porous minerals such as howlite, a calcium boro-
silicate, Ca2B5SiO9(OH)5, which in various forms simulates lapis
lazuli and turquoise and has also then been termed turquenite.
Recently, the appearance of a commercial synthetic alternative
to lazurite based on a resin formulation which has been coloured
in simulation of the genuine article has generated much interest
in the lapidary and gemstone communities; this substitute is
considered so good that the visual identification of genuine lapis
lazuli is no longer appropriate nor successfully accomplished.
The accepted differentiation that now applies commercially be-
tween genuine and simulated specimens of lapis lazuli is depen-
dent upon several chemically and thermally destructive tests.
The first of these is an acid test, whereby strong hydrochloric
acid is applied to the artefact, causing effervescence in the cal-
cite component of the genuine article during which the pyrites
component exudes a sulphurous gas but this procedure also re-
sults in a surface pitting, hence destroying a portion of the sur-
face. Secondly, the application of a blowtorch will cause
combustion damage to a resin simulate but this has also been
known to cause undesirable surface discolouration and fissure
damage to genuine specimens of lapis lazuli. Thirdly, poorer
quality lapis lazuli artefacts have been found treated with a
blue-dyed oil or resin which masks cracks and surface blemishes
and for which the recommended detection experiment involves
the application of acetone solvent, which again can cause detri-
mental effects to polished stone surfaces and which certainly
would not be applicable to many types of artefact. Clearly, these
analytical protocols are not acceptable in the context of speci-
men preservation and the minimisation of sample damage that
is now a de rigueur requirement for a contemporary forensic
analytical application. A further variant of fake lapis lazuli is
the so-called Swiss/German lapis, which is a type of jasper (chal-
cedony, an opaque silica) that has been stained blue by use of
potassium ferrocyanide and ferrous sulfate.
The approach adopted here, although still vibrationally spectro-
scopic in context is rather different in that the discrimination
between genuine and fake specimens of lapis lazuli is studied with
portable instrumentation that necessarily uses several different
wavelengths of excitation that would not normally be adopted
.This introduces a novel capability of forensic provenancing with
the added advantage of in-field characterisation and obviates the
need to remove specimens for subsequent laboratory analysis. This
will be especially relevant for the interrogation of large artefacts
and specimens in museum collections, in gemmological scenarios
and especially for artefacts which have an inlay of suspected lapis
lazuli where it is not feasible to separate the components, such as
inlaid furniture and complex jewellery in mounts. In recent years,
the application of portable Raman spectroscopic instrumentation
for the field identification of minerals has been described in several
important papers, particularly those of Jehlicka et al. [13–21] who
carried out some comprehensive and comparative studies on
miniaturised equipment at altitude and in hot and cold desert
conditions. The rather limited laser excitation wavelengths
currently available (specifically 532, 785 and 1064 nm), the
reduced wavenumber shift range (typically 200 to 2000 cm 1)
and lower spectral resolutions operable were examined and as-
sessed in comparison with the data acquired from laboratory
instrumentation undertaken with specimens removed from the
field. Major factors in this comparison and in others using portable
instrumentation for the in situ studies of art works and archaeolog-
ical artefacts such as wall paintings and frescoes [22] are the high
level of background light insolation, the positioning stability of the
spectrometer whilst acquisition of the spectral data is in progress
and the instrumental wavenumber calibration of the recorded
spectral data. A comprehensive review of portable and mobile
Raman spectroscopic instrumentation, along with a critical survey
of its advantages and disadvantages, in the fields of mineralogy,
forensic science and art and archaeology has recently been
undertaken [23].
 E.M.A. Ali, H.G.M. Edwards / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 121 (2014) 415–419
Experimental
Samples
A genuine specimen of lapis lazuli gemstone (Fig. 1) from the
Flor de Los Andes mine, Ovalle, Chile, donated by the Sajuna KkaLa
Chilean export company, Santiago, Chile, was selected specifically
for the increased importance of this geological source during the
recent and ongoing war in Afghanistan that has severely curtailed
supplies of this material from the Middle East.
A fake specimen from Pierre Gilson et Cie (Fig. 2A), who
manufacture lapis lazuli simulates of very high quality, was also
provided by Sajuna KkaLa, gemmological experts in lapis lazuli.
The Gilson specimens have proved very difficult analytically to
differentiate from the real lapis lazuli hitherto, generating a
comment in the literature [24] that they have the same spectro-
chemical and XRD patterns as natural lapis lazuli, yet have a lower
specific gravity and a strong reaction exhibited by these materials
with hydrochloric acid seem inappropriate for the behaviour of
genuine lapis lazuli. From several formulations it is believed that
the Gilson synthetic samples contain synthetic ultramarine, zinc
phosphate, pyrites, resins and dyes. It will be a challenge, therefore,
to determine if non-destructive Raman spectroscopic analysis at
long wavelengths can distinguish between the commercial high
quality Gilson sample provided and genuine Chilean mined lapis
lazuli.
A third type of lapis lazuli specimen has also been provided
from an unknown source purchased by the authors in a bazaar in
Istanbul, comprising several bright blue polished stones with
visible white inclusions (Fig. 2B). Specimens of sodalite, pyrites,
howlite and azurite minerals from a verified geological collection
were used as standards for the definitive provision of comparative
Raman spectra of associated minerals.
Raman spectroscopic instrumentation
Two approaches were adopted in this Raman spectroscopic
experimental study: the examination of the lapis lazuli samples
using laboratory-based bench instrumentation with laser excita-
tion at 785 and 1064 nm and portable and hand-held units with
laser excitation at the same wavelengths. The major thrust of this
work was the assessment of the viability of Raman spectroscopy
for the detection of fake lapis lazuli specimens, especially address-
ing the specimen of Gilson manufacture that hitherto seems to
have defeated analytical attempts to characterise it instrumentally
Fig. 1. Genuine lapis lazuli.
417
Fig. 2. Fake lapis lazuli; (A) fake specimen from Pierre Gilson et Cie and (B) fake
lapis pebbles.
and which therefore presents a problem in authentication. A
positive result in this respect will be a vindication for the Raman
spectroscopic technique in a forensic art context, especially if this
can be accomplished using portable instrumentation that could be
applied both for the interrogation of lapis lazuli inlaid in furniture
or artefacts or mounted in jewellery, and for the rapid, non-
destructive analysis of unverified specimens presented for sale
which would suffer from the potentially serious damage incurred
using alternative but currently accepted thermal or wet chemical
methods.
Laboratory bench-top instrumentation comprised a Bruker
FT-Raman IFS 66 system with FRA 106 Raman module delivering
excitation at 1064 nm from a Nd3+/YAG laser with an operating
spectral resolution of 4 cm 1 over a spectral wavenumber range
of 100–4000 cm 1 from a spectral footprint of 100 lm, and Reni-
shaw InVia confocal Raman spectrometer operating with a diode
laser at 785 nm with a spectral resolution of 2 cm 1 over a spectral
wavenumber range of 200–3500 cm 1 from a spectral footprint
provided by a 50 lens objective of approximately 5 lm. Portable
instrumentation comprised three commercial systems, namely a
DeltaNu Inspector Raman unit and a Renishaw RX210 (RIAS) unit
operating at 785 nm each with spectral footprints of approxi-
mately 10 lm and a DeltaNu advantage prototype unit operating
at 1064 nm with a spectral footprint of about 8 lm. The major
operational differences between the bench-top and portable
instruments can be summarised as follows: the portable units
are able to record Raman spectral data significantly more rapidly
than their fixed bench-top analogues and in an ex-laboratory situ-
ation but this is achieved with only a limited wavenumber range
accessible, which is usually about 200–2000 cm 1. A successful
observation of differentiation between genuine and fake specimens
using portable instrumentation comparable to that which can be
achieved in the laboratory will realistically support the future
adoption of this instrumentation for the field analysis of lapis lazuli
specimens.
Results and discussion
Raman spectra obtained from the genuine and fake lapis lazuli
specimens are dependent upon the excitation wavelength used;
for example, the Raman spectrum of Chilean lapis lazuli at
1064 nm is very different from those cited in the literature, where
the dominant spectral feature is a strong band at 549 cm 1 arising
from the S–S stretching mode of the S3 radical anion chromophore
responsible for the deep blue colour of the lapis lazuli. With 1064
and 785 nm near infrared excitation this band is not observed and,
in fact, the Raman spectrum is very different at these wavelengths
of excitation (Fig. 3A and B). The origin of these spectral differences
418 E.M.A. Ali, H.G.M. Edwards / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 121 (2014) 415–419
Fig. 3. (A) Genuine lapis; FT_1064 and (B) genuine lapis; Renishaw Invia Micro-
scope 785.
Fig. 4. (A) Gilson; FT_1064 and (B) Gilson; Renishaw Invia Microscope 785.
is not yet known but it is likely that some bands could possibly be
attributable to spectral artefacts. The spectrum of genuine Chilean
lapis lazuli at 785 nm exhibits an electronic Raman spectrum
which itself seems to be a characteristic one which the fake spec-
imens do not possess, so in itself this could be potentially used as a
diagnostic confirmation of the genuine specimen in discrimination
between authentic and the fake samples.
A major objective in the present study centres on the ability of
Raman spectroscopy to differentiate between the Gilson synthetic
form of substitute lapis lazuli and the genuine mineralogical spec-
imen. In Fig. 4A and B the Raman spectra of the Gilson specimen at
1064 and 785 nm excitation wavelengths can be seen; it is clear
that at both wavelengths of excitation used here the spectra of
the fake lapis lazuli is different from that of the genuine article.
The presence of the calcite inclusions in the Gilson sample is noted
by the bands at 1086, 712 and 283 cm 1, but the most significant
spectral signatures are those that arise from the organic compo-
nents of the synthetic sample. Here, bands at 1531, 1451, 1342,
1142, 747, 714 and 680 cm 1 can be correlated with a resin and
dyestuff, which lend both colour and strength to the Gilson mate-
rial - these bands can be matched with those of methylene blue
[22], and others can be assigned to the C–C and C–O bands of an
organic carrier component. This result is a vindication for the
non-destructive application of Raman spectroscopy in the detec-
tion of the use of fake lapis lazuli in jewellery or artefacts without
resorting to more destructive wet chemical methods. Of special
interest in this forensic art context is the use of the portable Raman
instrumentation which here demonstrates the capability of in-field
or on-site application, with the added advantage of a rapid
Fig. 5. (A) Gilson 1064 portable and (B) Gilson 785 portable RIAS.
Fig. 6. (A) Fake lapis pebbles; FT_1064 and (B) fake lapis pebbles; Renishaw Invia
Microscope 785.
non-destructive method since spectra can be recorded typically
in 10 s. Corresponding Raman spectra of the Gilson sample
recorded with the portable instruments operating at both excita-
tion wavelengths of 785 and 1064 nm are very similar to each
other and to those recorded with the laboratory bench equipment
and examples are shown in Fig. 5A and B.
Examination of the other fake specimen of lapis lazuli studied
here, which was formed into rough blue pebbles with exposed
white inclusions, again resulted in a successful outcome for the
Raman spectroscopic method; Fig. 6A and B show the spectra of
a blue pebble recorded with 1064 and 785 nm excitation, from
which we can conclude that the material is definitely not genuine
lapis lazuli and we may conclude that it is comprised of a blue
pigment coating upon a dolomitic calcite substrate with bands at
1445, 1100, 309, 183 cm 1. The organic bands are characteristic
of methylene blue (1543, 1339, 1139, 748, 682 cm 1) [22].
The spectra of the substrate obtained from a pigment-free
‘‘inclusion’’ area are given in Fig. 7. Fig. 7 A has been obtained with
a Fourier-transform bench top instrument equipped with an infra-
red laser operating at 1064 nm wavelength and its corresponding
spectrum obtained in 10s from a portable instrument with
785 nm excitation is given in Fig. 7B. Spectra of several geological
mineral standards are shown in Fig. 8 – namely azurite, howlite,
sodalite and pyrite – were also examined in this study, comprising
inclusions that are normally found associated with genuine lapis
lazuli such as pyrites and calcite and minerals that have been
known to be used in the manufacture of fake specimens. The
spectra of these geological minerals were successfully recorded
 E.M.A. Ali, H.G.M. Edwards / Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 121 (2014) 415–419
Fig. 7. (A) Fake lapis pebbles core; FT_1064 and (B) Fake lapis pebbles core;
portable RIAS 785.
Fig. 8. Spectra recorded using a benchtop Renishaw Invia Microscope, 785 nm. (A)
Azurite, (B) Howlite, (C) Pyrite and (D) Sodalite.
Fig. 9. Spectra recorded using a portable RIAS instrument operating at 785 nm. (A)
Azurite, (B) Howlite, (C) Pyrite and (D) Sodalite.
using bench-top instruments operating at 1064 and 785 nm
wavelengths. Howlite is noteworthy as a substrate for artificially
pigmented fake lapis lazuli, whereas azurite and sodalite are
naturally occurring blue coloured minerals that are sometimes
passed off as poorer quality lapis lazuli. From the Raman spectra
recorded here, we can definitively state that neither of the fake
specimens presented for analysis contain azurite, howlite or
419
sodalite as components, but of course it is perfectly reasonable to
suggest that fake specimens from other sources could contain
these materials as components and then the Raman spectroscopic
identification of their presence would be possible from a mineral
database. We can confirm that the portable instrumentations
operating at 1064, 785 nm wavelengths also were effective in the
identification of these minerals (Fig. 9).
Conclusions
From the present study we can confirm that Raman spectros-
copy has facilitated the identification of genuine and fake lapis
lazuli specimens; it is especially important to note that unambigu-
ous evidence is presented for the characterisation of the Gilson
synthetic lapis lazuli, used commercially as a high quality replace-
ment for the genuine material, which has been claimed hitherto to
be indistinguishable analytically from the genuine article, unless
destructive wet chemical test reactions are carried out on the
specimens. Of equal importance is the future role of portable
Raman spectroscopic instrumentation for the interrogation of lapis
lazuli artefacts, which has not been reported before in this forensic
context.
Acknowledgement
We are grateful for the donation of genuine and fake speci-
mens of lapis lazuli to Dr. Daniel Llovera Alvarez of the Sajuna
Kkala Company in Santiago, Chile, experts in Chilean lapis lazuli.
References
[1] J. Plesters, in: A. Roy (Ed.), Artists’ Pigments, Oxford University Press, New
York, 1993.
[2] R.D. Harley, Artists’ Pigments, Butterworths, London, 1970. Chapter 1600–
1835.
[3] H.G.M. Edwards, S.E. JorgeVillar, K.A. Eremin, J. Raman Spectrosc. 35 (2004)
786–795.
[4] H. Bearat, M. Fuchs, M.M.D. Paunier (Eds.), Roman wall painting: materials,
techniques, analysis and conservation, Proceedings of the International
Workshop on Roman Wall painting, Fribourg, March 1996, Institute of
Mineralogy and Petrography, Fribourg University, Perolles, Fribourg,
Switzerland, 1997.
[5] M. Bicchieri, M. Nardone, P.A. Russo, A. Sodo, M. Corsi, G. Cristoforetti, V.
Palleschi, A. Salvetti, E. Tognoni, Spectrochim. Acta Part B 56 (2001) 915–
922.
[6] I.M. Bell, R.J.H. Clark, P.J. Gibbs, Spectrochim. Acta Part A 53 (1997) 2159–
2179.
[7] R.J.H. Clark, J. Mol. Struct. 347 (1995) 417–428.
[8] R.J.H. Clark, T.J. Dines, M. Kurmoo, Inorg. Chem. 22 (1983) 2766–2772.
[9] V. Desnica, K. Furic, M. Schreiner, E-Preservation Sci. 1 (2004) 15–21.
[10] R.J.H. Clark, M.L. Curri, C. Laganara, Spectrochim. Acta Part A 53 (1997) 597–603.
[11] C.M. Schmidt, M.S. Walton, K. Trentelman, Anal. Chem. 81 (2009) 8513–8518.
[12] M. Favaro, A. Guastoni, F. Marini, S. Bianchin, A. Gambirasi, Anal. Bioanal.
Chem. 402 (2012) 2195–2208.
[13] P. Vitek, E.M.A. Ali, H.G.M. Edwards, J. Jehlicka, R. Cox, K Page, Spectrochim.
Acta Part A 86 (2012) 320–327.
[14] P. Vitek, J. Jehlicka, H.G.M. Edwards, Appl. Spectrosc. 67 (2013) 767–778.
[15] P. Vandenabeele, J. Jehlicka, H.G.M. Edwards, R. Cox, Spectrochim. Acta Part A
80 (2011) 32–35.
[16] J. Jehlicka, P. Vandenabeele, H.G.M. Edwards, A. Culka, Spectrochim. Acta Part
A 80 (2011) 36–40.
[17] J. Jehlicka, P. Vandenabeele, H.G.M. Edwards, A. Culka, T. Capoun, Anal. Bioanal.
Chem. 3979 (2010) 2753–2760.
[18] J. Jehlicka, H.G.M. Edwards, A. Culka, Phil. Trans. Roy. Soc. London A 368 (2010)
3127–3136.
[19] J. Jehlicka, P. Vandenabeele, H.G.M. Edwards, J. Raman Spectros. 41 (2010)
440–444.
[20] J. Jehlicka, H.G.M. Edwards, P. Vandenabeele, M.D. Hargreaves, T. Capoun, J.
Raman Spectros. 40 (2009) 1645–1651.
[21] J. Jehlicka, H.G.M. Edwards, P. Vandenabeele, M.D. Hargreaves, T. Capoun,
Spectrochim. Acta Part A 73 (2009) 410–419.
[22] P. Vandenabeele, H.G.M. Edwards, Analytical Archaeometry: Selected Topics,
Royal Society of Chemictry Publishing, Cambridge, 2012.
[23] P. Vandenabeele, H.G.M. Edwards, J. Jehlicka, Chem. Soc. Rev. to be published
and under final review, October 2013.
[24] K. Nassau, Gems and Gemology 15 (1976) 194–199.