Raman Microspectroscopy of Garnets from SFibulae from the Archaeological Site Lajh (Slovenia)

Abstract

Garnets (19 pieces) of Late Antique Sfibulae from the archaeological site at LajhKranj (Slovenia) were analysed with Raman microspectroscopy to obtain their mineral characteristic, including inclusion assemblage. Most garnets were determined as almandines Type I of pyralspite solid solution series; however, three garnets showed a higher Mg, Mn and Ca contents and were determined as almandines Type II. Most significant Raman bands were determined in the range of 169–173 cm−1 (T(X2+)), 346–352 cm−1 (R(SiO4)), 557–559 cm−1 (ν2), 633–637 cm−1 (ν4), 917–919 cm−1 (ν1), and 1042–1045 cm−1 (ν3). Shifting of certain Raman bands toward higher frequencies was the result of an increase of the Mg content in the garnet composition, which also indicates the presence of pyrope end member in solid garnet solutions. Inclusions of apatite, quartz, mica, magnetite, ilmenite, as well as inclusions with pleochroic or radiation halo and tension fissures (zircon), were found in most of the garnets. Rutile and sillimanite were found only in garnets with the highest pyrope content. Spherical inclusions were also observed in two garnets, which may indicate the presence of melt or gas residues. The determined inclusion assemblage indicates the formation of garnets during medium- to high-grade metamorphism of amphibolite or granulite facies. According to earlier investigations of the garnets from Late Antique jewellery, the investigated garnets are believed to originate from India.

Full text

Minerals 2020, 10, 325; doi:10.3390/min10040325 www.mdpi.com/journal/minerals
Article
Raman Microspectroscopy of Garnets from S-Fibulae
from the Archaeological Site Lajh (Slovenia)
Saša Kos 1, Matej Dolenec 2, Judita Lux 3 and Sabina Dolenec 4,*
1 Geological Survey of Slovenia, 1000 Ljubljana, Slovenia; sasa.kos@geo-zs.si
2 Department of Geology, Faculty of Natural Sciences and Engineering, University of Ljubljana, 1000
Ljubljana, Slovenia; matej.dolenec@ntf.uni-lj.si
3 Institute for the Protection of Cultural Heritage of Slovenia, 4000 Kranj, Slovenia; judita.lux@zvkds.si
4 Slovenian National Building and Civil Engineering Institute, 1000 Ljubljana, Slovenia
* Correspondence: sabina.dolenec@zag.si
Received: 31 January 2020; Accepted: 2 April 2020; Published: 4 April 2020
Abstract: Garnets (19 pieces) of Late Antique S-fibulae from the archaeological site at Lajh-Kranj
(Slovenia) were analysed with Raman microspectroscopy to obtain their mineral characteristic,
including inclusion assemblage. Most garnets were determined as almandines Type I of pyralspite
solid solution series; however, three garnets showed a higher Mg, Mn and Ca contents and were
determined as almandines Type II. Most significant Raman bands were determined in the range of
169–173 cm−1 (T(X2+)), 346–352 cm−1 (R(SiO4)), 557–559 cm−1 (ν2), 633–637 cm−1 (ν4), 917–919 cm−1 (ν1),
and 1042–1045 cm−1 (ν3). Shifting of certain Raman bands toward higher frequencies was the result
of an increase of the Mg content in the garnet composition, which also indicates the presence of
pyrope end member in solid garnet solutions. Inclusions of apatite, quartz, mica, magnetite,
ilmenite, as well as inclusions with pleochroic or radiation halo and tension fissures (zircon), were
found in most of the garnets. Rutile and sillimanite were found only in garnets with the highest
pyrope content. Spherical inclusions were also observed in two garnets, which may indicate the
presence of melt or gas residues. The determined inclusion assemblage indicates the formation of
garnets during medium- to high-grade metamorphism of amphibolite or granulite facies. According
to earlier investigations of the garnets from Late Antique jewellery, the investigated garnets are
believed to originate from India.
Keywords: garnets; inclusions; S-fibulae; Late Antiquity; provenance; Raman microspectroscopy;
XRF spectroscopy
1. Introduction
The production of ornamented metal jewellery decorated with red garnets has a long history in
both European and Eastern cultures, dating back to the 3rd century BC. Garnet-inlaid metalwork
made in the “cloisonné” technique, became widely spread among Migration Period Germanic tribes
during the Late Antique to Early Middle Ages (4th to 8th century AD), although, the manufacture of
cloisonné jewellery supposedly originated from the East in the Black Sea region [1,2]. The
investigation of geographical and geological origin of garnets using chemical and mineralogical
characterisation is and remains the subject of many studies, from the perspective of understanding
the deposits of minerals and trade routes that linked them to the rest of the world, since large amounts
of gemstones have been used in jewellery production [2–10].
The determination of the structure, composition and physical properties of minerals from
archaeological artefacts is difficult since they should not be compromised. In this case,
non-destructive analytical methods can be used. Chemical analysis is obtained by various analytical

Minerals 2020, 10, 325 2 of 21
techniques (particle-induced X-ray emission/particle-induced gamma-ray emission—PIXE/PIGE,
scanning electron microscopy with energy dispersive Xray spectroscopy—SEM/EDS, X-ray
fluorescence (XRF) spectroscopy, etc.), although they are sometimes insufficient to define the nature
of the crystal and some are even destructive, expensive and time-consuming. The crystal structure
can be determined by complementary techniques such as Raman spectroscopy, of which advantages
are that it is rapid, requires little or no sample preparation and above all it is non-contact and non-
destructive [11–13].
Garnets are a complex group of nesosilicate minerals with the general chemical formula
X3Y2(SiO4)3, where X sites are filled by either Fe2+, Mg2+, Ca2+ or Mn2+ ions and Y sites by Al3+, Fe3+, Cr3+,
Ti3+ or V3+ ions [14]. Also, Si4+ is commonly substituted with Al3+, Fe3+ and Ti3+ [11]. The crystal
structure of garnets is composed of the SiO4 tetrahedron and YO6 octahedron, which are bonded by
an oxygen atom and thus form distorted dodecahedral XO8 cells [15]. Due to similar ionic radii, ionic
charges and chemical affinity of present metallic cations, cation exchanges are common in garnets,
thereby forming solid solutions of end members. Natural silicate garnets prefer to be either non-calcic
or calcic [16], and therefore most frequently form a group called pyralspites in which end members
all have Al3+ in Y position, and a group of ugrandites in which end members all have Ca2+ in X
position, respectively. End members of the pyralspite group are pyrope (X = Mg2+), almandine (X =
Fe2+) and spessartine (X = Mn2+), and of the ugrandite group are grossular (Y = Al3+), andradite (Y =
Fe3+) and uvarovite (Y = Cr3+) [17,18]. Cation exchange and chemical composition of garnets not only
dependent on the composition of the source rocks, but also on the temperature and pressure of the
formation environment [15]. They are commonly found in different metamorphic rocks and rarely as
primary minerals in igneous rocks. Due to the close connection to formation conditions, garnets are
a useful indicator of metamorphic facies [19].
Vibrational spectroscopy, including Raman spectroscopy, is useful for distinguishing between
different solid solutions as the spectra change as a function of the garnet composition [18]. Although
Raman spectra of garnets show similar frequencies of Raman bands for each end member of
individual solid solution series (i.e., ugrandites and pyralspites), we can easily distinguish between
the spectra of solid solutions due to different cation arrangement [20–23]. The most significant
difference in the Raman spectra between the ugrandite serie and pyralspite serie lies in the fact that
the relatively strong Raman bands occur at ~360–370 cm−1, ~513–540 cm−1 and ~870–885 cm−1 for
ugrandites and ~342–368 cm−1, ~555–560 cm−1 and ~905–920 for pyralspites [20]. Raman bands of
garnets are assigned to vibrational modes of Si–O stretching (ν1 and ν3), Si–O bending (ν2 and ν4),
rotation and translation modes of SiO4 tetrahedron, and the translation mode of the X2+ cation in
dodecahedron sites of the garnet crystal lattice [18]. Raman shifts for all members within either of the
garnet series are mainly attributed to the atomic mass, atomic structure, ionic radius and polarisations
of cations in the X polyhedra for pyralspites and in the Y octahedra for the ugrandites, which affect
the volume of crystal unit-cell of garnet and force constants of bonds [20,24]. The silicate garnet
system pyrope–almandine–spessartine–grossular–andradite–uvarovite shows extensive homovalent
substitutional solid solution over two structural sites and complete compositional variation between
pyralspite species and ugrandite species has been documented [25]. Thus, the shift of a Raman band
due to a chemical substitution is a useful indicator of chemical composition and instead of observing
Raman vibrations of these specific elements, one can conveniently observe the shifting of the
numerous types of Si–O vibrations within or between the tetrahedral sites [21–23]. For instance,
Raman mode frequencies of garnets along the almandine-pyrope solid solution series show the
vibrational spectra change with the composition, where the Si–O stretching, Si–O bending, and the
rotation of the SiO4 tetrahedron (R(SiO4)) mode frequencies decrease linearly, but the translational
mode frequencies of the SiO4 tetrahedron (T(SiO4)) increase with the almandine content [26].
Inclusions are yet another useful tool in interpreting the geological environment in which
garnets are formed, although this study area is still under development [27]. Inclusions are either
minerals, melts or fluids that can be trapped or formed in various phases of mineral formation [28].
Minerals can be residual reactants or products of the chemical reaction that are responsible for the
host formation or phases that were not involved in these reactions [29]. Fluid and melt inclusions

Minerals 2020, 10, 325 3 of 21
provide information about material transfers in the Earth system [27]. Melt, fluid or gas phases could
also be observed in minerals, which can be trapped or preserved during (re-)crystallisation and are
mostly associated with the material transfer in the Earth system, from shallow mineralization to
re-fertilization of the mantle by subduction [27].
The characteristics of inclusions can be identified by Raman microspectroscopy. In garnets, solid
inclusions are mostly found in almandines, pyrope-almandines and pyropes, although various
inclusions or textures may be observed in other types of garnets as presented in Table 1, depending
on the environment of formation and origin [30,31].
Table 1. Commonly recognized solid inclusions in garnets [30,31].
Group Composition of an End Member Inclusions
Pyralspite
Pyrope Mg3Al2(SiO4)3 Apatite, ilmenite, rutile, undetermined
fibrous minerals
Pyrope-almandine (Fe, Mg)3Al2(SiO4)3 Zircon, monazite, apatite, rutile, mica,
quartz, often very pure
Almandine Fe3Al2(SiO4)3 Zircon, rutile, mica, hornblende, apatite,
spinel, quartz
Spessartine Mn3Al2(SiO4)3 Rarely present, growth lines – dark
wavy feathers
Ugrandite
Andradite Ca3Fe2(SiO4)3 Chryzotile
Grossular Ca3Al2(SiO4)3
Mostly no inclusions are found
Tsavorite – zircon, fibrous minerals
Hessonite – zircon, diopside, treacle and
heat wave textures
Uvarovite Ca3Cr2(SiO4)3 Not determined
By determining the chemical composition and, in rare cases, even the mineral inclusions of
Merovingian Kingdom (mid-5th–7th century AD); gemstones from France, Belgium and South
Germany; and five different types of garnets and their probable geographical-geological origin have
been possible to be identified to date (e.g., [2–9]). Two types of almandine, group of garnets with the
intermediate composition of pyrope-almandine (also in terms of gemmology—rhodolite or
pyraldine, e.g., [5]) and two types of pyrope were determined. Differences in the garnet composition
are shown in Table 2.
Table 2. Determined groups of garnets from Late Antique–Early Middle age jewellery, according to
the definition of Calligaro et al. (2002) [4] and Gilg et al. (2010) [8], based on the chemical composition
of garnets.
Calligaro et
al. [4] Type I Type II - Type III Type IV Type V
Gilg et al. [8] Group B Group A Group C[10] Group X Group D Group E
Chemical
Composition
of Garnets
Almandines
poor in Mg,
Ca, Mn, also
Cr and Y
Almandines
rich in Mg,
Mn, also Cr
and Y
Almandines
rich in Ca
and Mg
Pyrope-
almandines
(rhodolite)
Pyropes
poor in
Cr
Pyropes
rich in
Cr
Almandine (Type I and II) garnets supposedly originate from India, with several possible
deposits in Rajasthan, Andhra Pradesh and Odisha, whereas rhodolites (Type III) originate from Sri
Lanka. Pyropes (Type V) originate from the Czech Republic, but also Portuguese and/or East African
pyropes (Type IV) can be found [5–7]. Another variety of almandine garnet with high Ca content
(Group C) was found mainly in Scandinavian jewellery [8,10]. Depending on the time of the
occurrence, rhodolites and Type II almandines were common until the 5th century AD, while type I
almandines prevailed since the 6th century AD. In the 7th century AD, significant changes in trading
routes occurred on the East, which contributed to the use of Bohemian pyropes (Type V garnets) [8].

Minerals 2020, 10, 325 4 of 21
Optical microscopy and Raman spectroscopy used in inclusion analysis show that mostly
apatite, monazite, zircon, uraninite, xenomorphic Fe-chlorite, and rarely rutile, were present in Type
I almandines. On the other hand, for type II almandines, xenomorphic quartz crystals, ilmenite,
zircon, uraninite, monazite, rutile needles and sillimanite were identified [4,8,31,32]. The latter might
indicate the formation in high metamorphic rocks [31,32]. Concerning the geological origin of garnets
and their deposits, they are all vastly characterised by various metamorphic rocks of different ages
(Precambrian to younger). The Indian deposits are geologically characterised by various rocks of
amphibolite to granulite facies (e.g., different types of schists, gneiss and amphibolites) formed
during medium- to high-grade metamorphism, which can provide almandines and pyropes of
various chemical compositions [33,34]. The island of Sri Lanka is formed of various Precambrian
metamorphic rocks (e.g., granulites, gneiss, schists, and quartzite), while intact garnet crystals can be
often found in alluvial deposits originating from these rocks [35]. The geological variety of Sri Lanka
also greatly contributed to the chemical diversity of almandine–pyrope garnets found on the
jewellery. Deposits of pyropes in Bohemia are mainly found in alluvial plains south of the Bohemian
massive, although the formation of minerals took place in garnet peridotites [36]; while in Portugal,
garnets could be excavated from Palaeozoic gneisses, gabbro or peridotites in the vicinity of Lisbon,
where pyropes can still be found [37]. The East African deposits are rich in many gemstones,
including spessartines, pyropes and rhodolites; due to the high complexity of the rocks found in the
eastern area, many of the deposits remain strongly active [38].
The archaeological site at Lajh in Kranj is one of the most important archaeological cemeteries
in Slovenia (Figure 1) dating back to the Late Antiquity (5th to 6th century AD). More than 700 graves
have been excavated so far [39]; some of them are rich in metallic artefacts. These small finds are
related mainly to the Romans and Germanic tribes of the Lombards and Eastern Goths [40,41].
Ornamented metal jewellery found in the cemetery is especially characteristic of Germanic females.
Among other finds, variously fashioned gilded silver and bronze brooches were found, which were
decorated with gemstones in the cloisonné technique or more rarely, with a coloured glass [39]. A
special type of brooches from this site are S-shaped fibulae, which were mostly found among
Romans, Lombardic and other Germanic tribes between the 5th to 7th century AD [42,43].
Figure 1. Archaeological site Lajh in city of Kranj in Slovenia.
Several pieces of jewellery and belt buckles from this area have been investigated in earlier
works of Kramar et al. (2011) [44], Šmit et al. (2014) [45] and Nemeček et al. (2016) [46] using various

Minerals 2020, 10, 325 5 of 21
analytical techniques (XRF spectroscopy, Raman spectroscopy, PIXE/PIGE). On the studied artefacts,
Type I almandines dominated, with several examples of Type II almandines and a few rhodolites
(Type III). One of three garnet types appears either alone or as a combination of two types on an
individual piece of jewellery. All three types of garnets are rarely found together [40]. The
investigated garnets of Type I and II correspond to Indian deposits, while Type III rhodolites
correspond to the deposits of Sri Lanka, as recognised in previous works [2–8]. Whether the different
types of garnets were used intentionally on jewellery is still unknown [45].
The aim of this study was to investigate garnets of Late Antique jewellery (mid-6th century AD)
from Slovenian archaeological site Lajh (Kranj) by means of Raman microspectroscopy in order to
determine the type of garnets and inclusions present, as well as to discuss their potential provenance.
The data were supported by X-ray fluorescence (XRF) spectroscopy.
2. Materials and Methods
2.1. Artefacts
A total of 19 garnets from six S-fibulae were analysed in this study, as shown in Figure 2. Fibulae
were excavated in 2007 during the archaeological excavation of the Lajh cemetery in Kranj (Slovenia).
Fibulae are dated to the mid-6th century AD, according to their specific type Schwechat–Pallersdorf
(NA246, NA247, NA843, NA806) [42]. Milavec (2007) [42] described fibulae similar to NA892 as
Schwechat–Pallersdorf type alike, which is supposedly a South German derivate of this type. The
type of S-fibulae NA401 has not yet been determined. Garnets of the fibulae were polished to
approximately 3-mm-thick inlays and attached in metal cell work in the cloisonné technique. Apart
from the fibulae NA806 with five, all other fibulae were inlaid with three garnets (on fibula NA892
one of the three garnets is missing). Each garnet was numbered and abbreviated with the letter G
(e.g., NA246 G1).
Figure 2. Investigated S-fibulae from the archaeological site at Lajh (Kranj, Slovenia) with labelled
garnets.
2.2. Study of Garnets.

Minerals 2020, 10, 325 6 of 21
The Raman spectra of the garnets were obtained using a LabRAM HR800 spectrometer equipped
with a high-stability BX 40 optical microscope (Horiba JobinYvon, Villeneuve d'Ascq, France). The
Raman spectrometer had a grating with 600 grooves per mm and an air-cooled CCD detector.
Measurements were performed in the LabSpec acquisition software program (software version
5.25.15, Horiba JobinYvon, Villeneuve d'Ascq, France) using a 785 nm laser excitation line with an
output power of 31.4 mW and an Olympus LMPLFL N 50x/0.50 objective, at a spectral resolution of
about 1 cm−1. On each garnet, at least three measuring points were selected, with the acquisition time
of 60 s. Spectra were acquired in the range of 80–2000 cm1. All spectra were also analysed in LabSpec
5 Raman spectroscopic data processing software. Baseline corrections or normalisation of spectra
were not undertaken. The classical method of multivariate statistics (hierarchical cluster analysis)
was applied for checking and evaluating the variability in Raman spectra of studied garnets. Cluster
analysis is a typical method that aims to identify similarity patterns (clusters) in a data set. The
similarity was detected and demonstrated between the features (bands yielded and shifting of bands
in Raman spectra) that characterize the garnets. In general, cluster analysis follows several steps:
calculation of the similarity distances; in the present study, the squared Euclidean distances were
chosen as similarity measure; linking of the objects to clusters using the Ward’s linking method; and
graphical representation of the clustering by a hierarchical tree diagram. All statistical calculations
were performed with the software package STATISTICA (software version 13.3, TIBCO Software Inc.,
Munich, Germany). The data used for the cluster analysis are averages of three measurements of
Raman mode frequencies for each garnet in the fibulae.
Garnets from the S-fibulae were analysed with the handheld energy dispersive X-ray
fluorescence spectroscopy (ED-XRF) Niton instrument GOLDD XL3t 900S-He (0.1 mA; 50 kV)
(ThermoFisher Scientific, Billerica, MA, USA) with a 3 mm X-ray spot to determine their chemical
composition. The analyses were performed using the Cu/Zn Mining mode with continuous helium
purge for better detection of light elements (Mg, Si, Al, S and P), which are included in the
manufacturer’s software. The reference materials SiO2, NIST-88b and NIST-1d were used at the
beginning and end of the measurements to calibrate light elements. The pre-calibration with 25
international standards was performed with the same specifications and preferences as the
measurements. Calibration curves were generated for all elements of interest (Si, Al, Fe, Mg, Ca, Mn).
Two measurements were made on each garnet of the fibulae, with an acquisition time of 210 s for
each measurement. The element assignments were defined using the NITON Data Transfer (version
NDT_REL_8.0.1 software, Thermo NITON Analyser LLC, Billerica, MA, USA). The peak intensities
for the ED-XRF spectra were given as Counts per Second (cps).
2.3. Study of Inclusions
Inclusions in garnets were determined by optical microscopy and Raman microspectroscopy.
The microscopic examination of the inclusions in garnets was performed with the Olympus BX-60
optical microscope (Olympus, Tokyo, Japan). One garnet (NA401 G3) was no longer attached to the
metal of the fibula that enabled to observe the inclusions in transmitted light. The rest of the garnets
were observed in reflected light at both parallel and crossed polars. The microphotographs of
inclusions were taken with the Olympus SC50 camera and Olympus Stream software.
The Raman spectra of inclusions were obtained with the same Raman microspectrometer and
software settings as for garnet analysis, except for adjusting the acquisition time, which varied
between 10 and 60 s. The Raman spectra of inclusions were often masked by a strong signal of garnet,
except for the minerals with a strong Raman scattering (e.g., rutile, quartz).
3. Results
3.1. Garnets
Frequencies of bands in Raman spectra and assigned Raman vibration modes are contained in
Table 3.

Minerals 2020, 10, 325 7 of 21
Table 3. Raman bands (cm−1) for the investigated garnets, assigned to site motion and mode of vibration.
Assignment
(mode)
Site
motion
[22]
Band
S-fibulae
NA246
NA247
NA401
NA843
NA806
NA892
G1
G2
G3
G1
G2
G3
G1
G2
G3
G1
G2
G3
G1
G2
G3
G4
G5
G1
G2
Translation
of X cation T(X2+) I 170.0 170.2 170.0 169.6 169.6 168.9 170.5 169.6 170.3 170.5 169.7 171.0 170.3 172.5 170.1 170.0 170.8 169.5 171.0
Translation
of SiO4 T(SiO4) II 214.4 214.2 213.5 215.0 215.0 215.0 214.2 215.6 214.5 213.9 215.2 214.0 214.3 213.2 215.5 214.4 213.7 215.9 214.2
Rotation of
SiO4
R(SiO4) III 316.2 315.7 315.8 315.9 316.0 315.6 316.4 315.9 316.1 316.5 316.6 316.6 316.3 317.1 315.9 315.9 316.6 315.7 316.3
T(SiO4) IV 330.8 330.8 331.5 330.8 330.9 330.3 332.3 331.9 333.0 332.9 331.9 - 331.9 - 330.6 330.8 333.6 330.3 333.5
R(SiO4) V 348.0 347.8 347.4 346.8 347.0 346.1 347.7 346.7 347.6 347.7 346.9 349.4 347.7 351.0 347.0 347.7 349.6 346.5 348.0
R(SiO4) VI 373.2 373.0 373.2 373.1 372.9 372.9 373.2 373.0 373.3 374.1 373.4 373.6 373.8 372.6 373.6 372.9 372.8 373.0 373.2
Si–O
bending
ν2 VII 481.0 479.7 479.7 479.7 479.5 478.9 480.3 479.8 480.1 480.4 479.7 481.2 480.1 482.0 480.2 480.6 481.0 479.6 480.4
ν4 VIII 501.8 501.8 501.4 501.4 501.1 500.9 501.9 501.0 501.8 501.9 501.3 502.7 502.1 503.7 501.6 502.2 502.7 501.0 502.0
ν2 IX 557.7 557.9 557.7 557.8 557.6 556.9 558.0 557.5 558.0 558.1 557.6 558.4 558.0 559.0 558.3 558.0 558.1 557.7 558.3
ν4 X 585.4 585.6 585.3 584.7 585.2 584.5 585.8 584.5 585.6 585.7 585.2 585.9 585.8 587.0 585.2 585.1 585.8 584.9 585.9
ν4 XI 600.2 600.4 599.7 599.3 600.2 599.0 601.1 599.8 599.6 600.6 600.0 602.0 600.9 602.6 600.3 601.5 601.7 596.0 599.0
ν4 XII 634.8 634.8 634.1 633.8 633.5 633.1 634.8 633.7 634.7 635.1 634.0 635.6 634.4 636.9 634.2 634.5 635.7 634.0 635.3
Si–O
stretching
ν3 XIII 866.4 865.4 866.1 866.0 865.8 866.0 865.6 865.3 865.7 866.4 866.3 864.3 865.5 864.6 866.5 864.9 862.9 865.6 866.7
ν1 XIV 919.3 918.8 918.8 919.2 919.2 918.5 918.9 918.4 918.8 919.1 918.7 918.3 919.2 919.0 919.2 918.3 917.4 918.4 919.5
ν3 XV 1044.0 1042.6 1042.3 1043.4 1042.6 1042.5 1043.7 1041.3 1043.7 1043.3 1042.7
1044.0 1043.6 1045.3 1042.0 1042.5 1043.0 1042.3 1043.7

Minerals 2020, 10, 325 8 of 21
The Raman spectra of investigated garnets agree well with Raman spectra of pyralspites and the
dominant end member almandine [20–22,26], and can be grouped into three distinct energy regions:
low frequency regions between (169374 cm−1); medium energy bands (479–637 cm−1); and high
energy peaks (862–1046 cm−1) (Figure 3).
Figure 3. Raman spectrum of almandine garnet (NA247 G1).
Among the expected 25 Raman active modes (3A1g + 8Eg + 14T2g) of garnet group minerals [20],
a total of 15 vibration modes were observed in the Raman spectra obtained that could be grouped
into four main regions [22]. The two Raman bands at the range 120–280 cm−1 are assigned to
translation modes, i.e., to the translation motion of Fe2+ (band I) [47] and translation motions of SiO4
(band II) [22,48]. At the range of 280–450 cm−1, four Raman bands were observed (bands IIIVI)
dominated by rotation motions of SiO4 [22]. Six Raman bands (bands VIIXII) were located in the
range 450–750 cm−1 assigned to Si–O bending (symmetric ν2 or asymmetric ν4), while three Raman
bands (bands XIIIXV) yielded at the range 950 to 1100 cm−1 and were assigned to stretching motions
of the Si–O band, either symmetric (ν1) or asymmetric stretch (ν3).
As furthermore seen from Figure 4, multivariate analysis based on Raman mode frequencies
showed that the 19 garnets are clustered into three major distinct groups. This reflects the results of
X-ray fluorescence spectroscopy, where in relation to major chemical element contents, three groups
of the garnets were recognized (Figure 5). According to ED-XRF spectra, all garnets consisted of Fe,
Si and Al, confirming almandine composition of the pyralspite solid solution serie. Magnesium,
indicating the presence of pyrope component, was also detected, followed by minor Ca (grossular
component) and Mn (spessartine component) contents. However, some differences in Fe, Mg, Mn
and Ca contents were observed for the garnets, which affect the shifting of Raman bands towards
higher or lower frequencies. Hence, group II includes seven garnets (NA247 G1-G3, NA401 G2,
NA806 G3, NA843 G2 and NA892 G1) and is characterised with the highest Fe contents, whereas Mg,
Ca and Mn contents are the lowest among the studied garnets. The lowest values of most of the
Raman mode frequencies were observed for this group. On the other hand, group III consisted of
three garnets (NA843 G3, NA806 G2 and G5) and is characterised with the highest Mg content as well
as Ca and Mn contents and shifting of certain vibrational modes to higher frequencies (e.g., SiO
stretching ν3, SiO bending ν2 and ν4, and rotation of SiO4 modes). However, the majority of analysed
garnets belong to Group I, which contains nine garnets (NA246 G1-G3, NA401 G1 and G2, NA806
G1 and G4, NA843 G1 and NA892 G2) and is the middle range between Groups II and III.

Minerals 2020, 10, 325 9 of 21
Figure 4. Three diagrams for 19 garnets of the S-fibulae based on Ward’s method and squared
Euclidean distance using Raman mode frequencies.
Figure 5. Average intensity ED-XRF spectrum of three-garnet groups. Marked are the K/Kβ1 peak
of Mg, Ca, Mn and Fe.
The comparison of Raman spectra of representative members of each group is shown in Figure
6, in which NA247 G3 is assigned to Group I, NA246 G2 to Group II and NA806 G2 to Group III.

Minerals 2020, 10, 325 10 of 21
Figure 6. Raman spectra of almandine garnets, representative for each group.
The majority of mode frequencies decrease linearly with the almandine contents (i.e., decrease
with ionic radii Mg2+(0.89) > Fe2(0.92) Å) in all four regions of the Raman spectra. As seen from the Figure
6, this is observed as shifting of certain Raman bands to the left for garnets of group III to group II.
Namely, most Raman bands of the pyralspite group can show strong overlap, especially in the range
of SiO bending and rotation modes R(SiO4)). For spessartine, shifts to lower frequencies of SiO
stretching mode bands at 905–915 cm−1 and 1030–1035 cm−1 are indicative, due to the presence of a
slightly larger Mn2+ ion in the octahedral coordination, with respect to the Fe2+ ions in almandine
[11,20]. On the contrary, in pyrope, the presence of smaller Mg2+ cation in octahedral coordination
[18] affects the shifts of several vibrational modes to higher frequencies, especially ~363 cm−1, ~510
cm−1, ~640 cm−1 and ~1055 cm−1 [20,43]. In addition, Raman bands at ~170 cm−1 (band I) and ~330 cm−1
(band VI) showed a tendency to diminish towards pyrope composition as also recognised by Pinet
and Smith (1994) [22]. Anyway, some bands, for instance, Raman band at 214 cm−1 (band II) assigned
to the translational mode of the SiO4 tetrahedron, showed the opposite trend with increasing the
mode frequencies with almandine contents. An inverse correlation with ionic radii Mg2+(0.89) < Fe2+(0.92)
< Mn2+(0.96) Å for this frequency mode was reported also by other authors [22,26].
Figure 7 shows trends in the selected frequency distribution from the four ranges of the Raman
spectra assigned to the translation mode of the X2+ cation (band I), rotation of the SiO4 tetrahedron
(band V) and vibration modes of Si–O bending (ν4—band XII) and SiO stretching (ν3—band XV).
Raman shifts of vibration modes were attributed to the changes in the end member composition,
which was confirmed by XRF spectroscopy (Figure 5). The lowest observed frequency modes in the
ranges of 169–170 cm−1, 346–347 cm−1 and 633–634 cm−1 are the result of higher Fe contents for the
garnets of the Group II. On the other hand, the increase of Mg content affected the shifts of the Raman
bands in the ranges of 171–173 cm−1, 350–352 cm−1 and 635–637 cm−1 to higher frequencies in three
investigated garnets of Group III (NA806 G2 and G5, NA843 G3), respectively (Figure 7a–c). The
proposed trend could not be observed completely in the distribution of frequencies 1041–1045 cm−1
(Figure 7d) assigned to Si–O stretching (ν3). With the exception of NA247 G2, the frequencies for the
garnets of group II were shifted to lower values, whereas for the garnet of group III, NA806 G2, the
band for this mode was shifted to higher frequencies, which corresponds to increased Mg content
within the group. Rather low frequency was observed for the garnets NA806 G5 and NA892 G1,
which could be explained with slightly lower values of Mg and enhanced contents of Mn compared
to garnet NA806 G2 within group III.

Minerals 2020, 10, 325 11 of 21
(a) (b)
(c) (d)
Figure 7. Selected Raman mode frequencies change (a) Raman shifts for translation mode of the X2+
cation in the range of 169–173 cm−1 (band I); (b) Raman shifts for tetrahedron rotation in the range
346–352 cm−1 (band V); (c) Raman shifts for Si–O bending mode in the range of 633–637 cm−1 (ν4) (band
XII); (d) Raman shifts for Si–O stretching mode in the range of 1041–1045 cm1 (ν3) (band XV).
3.2. Inclusions
As observed by optical microscopy, garnets from S-fibulae contained several inclusions that
were additionally examined by Raman microspectroscopy. Most of the studied garnets contained
apatite, quartz and minerals with radiation and pleochroic halo (e.g., zircon). Some also contained
mica (muscovite, phlogopite) and opaque minerals (ilmenite). Rutile was observed only in two
garnets, while sillimanite fibres only appeared in one garnet. In two of the garnets, unidentified
spherical inclusions were observed, which could be fluid or gas inclusions. The inclusions
determined in garnets are listed in Table 4. Figures 8–16 show characteristic microphotographs of
mineral inclusions and their associated Raman spectra.
Table 4. Observed inclusions in garnets from S-fibulae, regarding a type of inclusion, their mineralogy
and affiliation to certain garnets.
Type of Inclusion Mineralogy and Chemical Formula Artefact and Garnet
Quartz SiO2 NA401 G3, NA806 G2, G4
and G5
Transparent crystal Apatite Ca5(PO4)3(Cl/F/OH)
All garnets of fibulae NA246,
NA247, NA401, NA806 G1
and G4
Mica (muscovite KAl2(AlSi3O10)(OH)2 or
phlogopite KMg3(AlSi3O10)(OH)2)
NA401 G3, NA806 G2, G4
and G5
Crystal with radiation
halo, pleochroic halo or
tension fissures
Zircon Zr(SiO4)
NA246 G2, NA247 G2,
NA401 G2, NA806 G2, 3 and
5, NA843 G2
Transparent fibres Sillimanite Al2(SiO4)O NA806 G2
Opaque clusters Magnetite (Fe2+,Fe3+)2O4, ilmenite Fe2+TiO3 NA247 G1 and 2, NA401 G2

Minerals 2020, 10, 325 12 of 21
Opaque needles Rutile α-TiO2 NA806 G5,
NA843 G3
Brownish plate
Biotite
K(Fe2+/Mg)2(Al/Fe3+/Mg)([Si/Al]Si2O10)(OH/F)2
or Xenotime (Y/Yb)(PO4)
NA892 G2
Spherical inclusions unidentified (melt, fluid or gas) NA843 G1, NA892 G2
Apatite was determined in garnets from fibulae labelled NA246, NA247, NA401, NA806 (G1 and
G4) and NA892 (G1). As shown in Figure 8b, strong characteristic Raman bands of apatite yielded at
272 cm−1, 432 cm−1, 592 cm−1, 965 cm−1 and 1083 cm−1 [49]. Apatite mainly occurred as elongated
xenomorphic crystals, from 50 to 300 μm in size (Figure 8a). However, in garnet G3 from fibula
NA401, it occurred as idiomorphic crystals in a size of approximately 300 μm.
(a) (b)
Figure 8. Inclusions in garnet G3 of fibula NA401. (a) Idiomorphic apatite inclusion. Reflected light,
parallel polars; (b) Raman spectrum of apatite in almandine garnet.
Garnet G3 from fibula NA401 also contained several xenomorphic grains of quartz in a size of
up to 20 μm (Figure 9a). Characteristic Raman bands at 128, 208 and 465 cm−1 confirmed the presence
of -quartz, however, some spectra also yielded minor bands at 394 cm−1 [50] (Figure 9b). Quartz
inclusions were also identified in other garnets (NA806 G2, G4 and G5).
(a) (b)
Figure 9. Inclusions in garnet G3 of fibula NA401. (a) Xenomorphic inclusion of quartz. Transmitted
light, parallel polars; (b) Raman spectrum of quartz in almandine garnet.
Opaque grains mostly occurred in groups as tiny elongated crystals (20–50 μm) (Figure 10a).
These inclusions were identified in fibulae NA247 (G1 and G2) and NA401 G2. Raman bands
indicating the presence of iron oxide minerals were observed in several spectra, e.g., in Figure 10b,
where besides quartz (black asterisk), a broad band at ~680 cm−1 is also visible (red asterisk), which
could be assigned to ilmenite [50–52]. Namely, ilmenite flakes with quartz inclusions are commonly
found in garnets [31].

Minerals 2020, 10, 325 13 of 21
(a) (b)
Figure 10. Inclusions in garnet G1 of fibula NA247 and garnet G2 of fibula NA401. (a) A group of
elongated inclusions of Fe oxides (i.e., ilmenite) in NA401 G2. Reflected light, parallel polars; (b)
Raman spectrum of quartz with bands at 130, 465 cm1 (black asterisk) and ilmenite at ~680 cm−1 (red
asterisk) in almandine garnet.
Thin intersecting rutile needles were visible in two garnets, NA806 G5 and NA843 G2 (Figure
11a). In the latter, the Raman spectrum of rutile (Figure 11b) was obtained with broad Raman bands
at 448 and 609 cm−1 and a weak band at 143 cm−1.
(a) (b)
Figure 11. Inclusions in garnet G3 of fibula NA843. (a) Needles of rutile. Reflected light, parallel
polars; (b) Raman spectrum of rutile in almandine garnet.
Anhedral rounded inclusions with pleochroic (brownish) or radiation halo and inclusions
damaged with tension fissures are common in most of the investigated garnets (Figure 12a–d). Their
size varies from 20–200 μm. They occurred in fibulae NA246 (G2), NA247 (G2), NA401 (G2), NA806
(G2, G3 and G5) and NA843 (G2). Inclusions with tension fissures are minerals, in which a
deformation of the crystal lattice occurs due to changes in the environmental conditions (pressure,
temperature) during the growth of the host crystal or in subsequent processes, that influence changes
in the structure of the host crystal [53]. Pleochroic or radiation halos are the result of radiation damage
within the host crystal structure, caused by the process known as metamictization [18]. This process
is characteristic for crystals containing radioactive elements in their crystal structure and can result
in radiation damage and crystalline structure changes to an amorphous state, while the outer crystal
shape is retained [18]. Such minerals are zircon, monazite, xenotime or apatite, which accommodate
thorium or uranium into the crystal structure [54].

Minerals 2020, 10, 325 14 of 21
(a)
(b)
(c)
(d)
Figure 12. Inclusions with radiation deformations. Reflected light, parallel polars. (a) Inclusion with
a pleochroic halo in garnet G2 of fibula NA843; (b) several inclusions with pleochroic halo in garnet
G2 of fibula NA806; (c) small inclusions with pleochroic halo and probably tension fissures in garnet
G2 of fibula NA247; (d) light radiation halo around inclusion in garnet G2 of fibula NA246.
Raman spectra of some deformed inclusions showed strong luminescence, which masked most
bands (Figure 13). A weak Raman band at ~1014–1017 cm−1 and a stronger band at ~1398 cm−1, often
accompanied by a broad band at 1280 cm−1, can be indicative of zircon. Namely, as suggested by some
authors (i.e., Kloprogge (2017) [18], Nasdala et al. (2002, 2018) [55,56]), the crystal structure of zircon
inclusion may be damaged due to metamictization or pressurization in their hosting minerals
resulting in decreasing of intensity and broadening of the strong Raman band characteristic of zircon
at ~1008 cm−1. Moreover, an increase in temperature and pressure will cause a shift and lower the
intensity of the Raman band to 1014–1022 cm−1 [57]. However, luminescence and presence of strong
Raman bands in the range of 1200–1400 cm−1 are related to amorphous SiO2, which appears due to
changes in inclusions after heath treatment [57].
(a) (b)

Minerals 2020, 10, 325 15 of 21
Figure 13. Inclusion in garnet G2 of fibula NA806. (a) Inclusion deformed by tension fissures.
Reflected light, parallel polars; (b) Raman spectrum of inclusion with strong luminescence and Raman
bands at ~1014 cm−1, ~1286 cm−1 and ~1399 cm−1 indicating presence of zircon in almandine garnet.
Raman spectra of particular xenomorphic transparent crystals of 20 to 200 um in size that
corresponded to mica were observed in several garnets (NA401 G3, NA806 G2, G4 and G5 (Figure
14). Namely, two types of phyllosilicates were recognised. Raman bands at ~100 and ~700 cm−1 are
indicative of muscovite type (Figure 14b, below), whereas at ~670–680 cm−1 of phlogopite type (Figure
14b, above) [58]. According to Schönig et al. (2018) [59], bands determined at 190–200 cm−1 can be
present for both types of mica, however, a strong band of ~270 cm−1 is characteristic only in
muscovite–paragonite type.
(a) (b)
Figure 14. Xenomorphic grains in garnet G2 of fibula NA806. (a) Xenomorphic inclusions of mica (i.e.,
muscovite or/and phlogopite). Reflected light, parallel polars; (b) Raman spectrum of muscovite (Ms,
spectrum below)) and Raman spectrum of phlogopite (Phl, spectrum above) in almandine garnet.
Among the investigated garnets, only in garnet G5 of fibula NA806 fibre-like inclusions were
present, which might be attributed to sillimanite (Figure 15a). Similar curved fibres of sillimanite
were also determined in almandines by Calligaro et al. (2002) [4], Schmetzer et al. (2014) [31] and
Horváth and Bendó (2011) [32]. Due to strong luminescence, only a Raman band at 235 cm−1 (black
asterisk) is present besides common almandine bands (171 cm−1, 312 cm−1, 349 cm−1, 558 cm−1 and 918
cm−1) (Figure 15b). For sillimanite, bands at ~235 cm−1, 310 cm−1, 456 cm−1, 870 cm−1 and 965 cm−1 are
characteristic [50,60]. Raman bands at 235 and 312 cm−1 (for analysed almandine garnets, this band is
observed at slightly higher values 315–317 cm−1) could indicate the presence of sillimanite.
(a) (b)
Figure 15. Inclusions in garnet G5 of fibulae NA806 and G2 of fibulae NA892. Reflected light, parallel
polars. (a) Inclusions of sillimanite in garnet G5 of fibula NA806; (b) Raman spectrum of fibre-like
inclusion in almandine, with band at 235 cm−1, indicative for sillimanite (black asterisk).

Minerals 2020, 10, 325 16 of 21
Furthermore, brown plate-like inclusions that occurred in garnets (NA246 G2, NA843 G3,
NA892 G2) could indicate the presence of biotite (Figure 16a), whereas according to Horváth and
Bendó (2011) [32], similar inclusions were attributed to the phosphate mineral xenotime.
In addition, garnet G1 of fibula NA843 and garnet G2 of fibula NA892 showed the presence of
transparent spherical grains, 100–400 μm in size, which most probably represent fluid or gas
inclusions (Figure 16b).
(a) (b)
Figure 16. (a) Brown plate in garnet G2 of fibula NA89.2 Reflected light, parallel polars; (b) Small
spherical inclusions in garnet G2 of fibula NA892. Reflected light, parallel polars.
4. Discussion
The Raman spectra of garnets obtained in this study all show similar Raman bands indicating
almandine as the dominant end member in solid solution. However, shifts of the Raman bands to
higher frequencies, e.g., modes assigned to the translation mode of X2+, the rotation mode of SiO4
tetrahedron and the Si–O stretching mode, indicate the presence of higher pyrope content in the solid
solution for the three garnets. The correlation between the composition of the natural garnets and
changes in Raman vibrations was determined in the study of Henderson (2009) [11,22,26]. In the case
of almandine garnets, Henderson (2009) [11] observed shifts to higher frequencies (the Si–O bending
(ν3) at 10441047 cm−1, the stretching (ν4) 636641 cm-1 and the rotation (R(SiO4)) modes at 350357
cm−1) with an increase in the pyrope content, although when spessartine content increased in
composition, this end member interfered with the linear trend of the increase in all mode frequencies.
Similarly, Kuang et al. (2019) [26] observed a linear increase in the frequencies of Si–O stretching and
bending, the rotation mode of the tetrahedron (R(SiO4)) and the translation of the X2+ mode (T(X2+))
with increasing pyrope content and vice versa for the translation of the tetrahedron (T(SiO4)) when
studying synthetic pyrope-almandine garnets, which is consistent to a certain extent (considering the
vibration modes) with our results.
According to the observed trends in Raman shifts of vibrational modes in the obtained Raman
spectra, the presence of three groups of almandines on the examined S-fibulae was determined, each
of which differs in its pyrope content, while spessartine and grossular components were also present
in minor amounts. According to similar Fe, Mg, Ca and Mn contents, we classified 16 garnets from
the S-fibulae from Groups I and II as Type I almandines (almandines poor in Mg, Ca, Mn), which
showed a shifting of the Raman bands to lower frequencies compared to the three garnets of Group
III. Garnets from this group were classified as Type II (almandines rich in Mg, Mn) and characterised
with shifting of the Raman bands of certain vibration modes (i.e., T(X2+), Si–O stretching and R(SiO4))
to higher frequencies due to increased Mg contents. In comparison to the study of several Slovenian
brooches by Šmit et al. (2014) [45], Type I almandine also predominates in S-fibulae and can be
combined with Type II almandines, which are also present in minor numbers. Almandines Type II
may point to older artefacts of the mid-5th to 6th century AD, according to Šmit et al. (2014) [45].
Concerning the geographical provenance, as Bimson et al. (1982) [6] first suggested almandines, both
of Type I and II, originate from Indian deposits, most likely from Rajasthan or NW India.

Minerals 2020, 10, 325 17 of 21
In addition, several solid inclusions in garnets from Late Antique jewellery have been observed
in works by Calligaro et al. (2002) [4], Horváth and Bendó (2011) [32], which have been systematically
classified as Type I and Type II almandine. Apatite, quartz, Fe-oxide minerals, mica and deformed
crystals with pleochroic halos, radiation halos or tension fissures can be observed in both garnet types
of the examined S-fibulae. In the present study, sillimanite and rutile needles were found only in
Type II almandine, which was also reported by Horváth and Bendó (2011) [32] as being more
common for Type II almandines
As far as geological formation is concerned, observed mineral inclusions—apatite, quartz, mica
and zircon are common in almandines and can form in different geological environments. They are
more common in felsic lithologies, although they also occur in mafic rocks [59]. Sillimanite and rutile
are considered important for geological provenance studies as they indicate a smaller range of
metamorphic facies [59]. Both are usually associated with medium- to high-grade metamorphism in
which rutile is an abundant accessory mineral. According to Force (1980) [61] and Zack et al. (2004)
[62], the increasing pressure of the upper amphibolite facies favours rutile formation, but when
subjected to low-grade formation, it breaks down to form other titanium bearing phases. Sillimanite
is restricted to the medium pressure part of the amphibolite facies and is mostly restricted to granulite
facies [63]. Spherical inclusions observed in two garnets of the S-fibulae may indicate the presence of
either gas or melt residue. Melt droplets may be entrapped at any phase of garnet recrystallization,
especially under rapid heating conditions and disequilibrium melting [64,65]. According to Ferrero
et al. (2018) [27], garnet is the most common host for melt inclusions because it is the most widespread
peritectic phase resulting from incongruent melting in the continental crust. Since garnets were
processed into thin inlays to fit the cloisonné technique of jewellery, some of the inclusions were
probably excluded from the host mineral, leaving an inclusion arrangement in the garnets studied
incomplete.
As described in the introduction, metamorphic rocks are spatially distributed in the region of
NW Indian (Rajasthan), from amphibolite facies of medium-grade metamorphism in the west to
granulite facies of high-grade metamorphism in the east, which support the theory of the geological
formation of inclusion assemblage in garnets and their probable deposition in the Rajasthan area [33].
Similar rock formations can also be found in parts of East India (Garibpet deposit), where according
to Schmetzer et al. (2017) [31], garnets were also excavated in Late Antique times and traded to the
west. However, these almandine garnets, which the author has examined, were mainly used and
traded as beads.
5. Conclusions
The studied garnets from the S-fibulae of the Slovenian archaeological site Lajh-Kranj (mid-6th
century AD) were identified as almandines, determined as Type I and Type II, according to their
Raman spectra, chemical characteristics and inclusion assemblage.
Most significant Raman bands were determined in the range of translation mode of X2+ at 169–
173 cm−1 (T(X2+)), rotation mode of SiO4 tetrahedron at 346–352 cm−1 (R(SiO4)), Si–O bending mode at
557–559 cm−1 (ν2) and 633–637 cm−1 (ν4), and Si–O stretching mode at 917–919 cm−1 (ν1) and 1042–1045
cm−1 (ν3), which are indicative for almandines. Shifting of Raman bands to higher frequencies were
observed for the garnets with enhanced Mg contents (e.g., for translation mode of X2+ (T(X2+)) at ~172
cm−1, rotation mode of SiO4 tetrahedron at ~350 cm−1 (R(SiO4)), Si–O bending mode at ~636 cm−1 (ν4),
and Si–O stretching mode at ~1045 cm−1 (ν3)), especially evident for almandines of Type II. Garnets of
both types show similar inclusion assemblage with apatite, quartz, crystals with radiation or tension
damages (e.g., zircon), ilmenite and mica. Rutile and sillimanite were only found in almandines of
Type II. Melt or gas inclusions were also observed in two garnets, although their chemistry cannot be
identified.
Specific mineral inclusion assemblage, with emphasis on rutile and sillimanite, imply the origin
of garnets in metamorphic rocks of amphibolite or granulite facies. Almandine garnets from the
studied S-fibulae are consistent with other studied garnets of Late Antique artefacts found in Slovenia
and also in a wider European area of 6th century AD and originating from Indian deposits.

Minerals 2020, 10, 325 18 of 21
Author Contributions: All authors have read and agree to the published version of the manuscript. J.L. provided
the samples; S.K., M.D. and S.D. methodology; S.K. conducted the measurements; S.K., M.D. and S.D. data
processing; S.D. writing—original draft preparation; M.D., J.L. and S.D. writing—review and editing; M.D.
funding acquisition; S.K. and S.D. wrote the paper.
Funding: This work was financially supported by the Slovenian Research Agency Programme Groups P2-0273
and P1-0195.
Acknowledgements: We wish to thank mag. Nataša Nemeček, senior conservator-restorer, for providing us
with the artefacts from National Museum of Slovenia archives.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Adams, N. The Garnet Millennium: The Role of Seal Stones in Garnet Studies. In Gems of Heaven: Recent
Research on Engraved Gemstones in Late Antiquity c. AD 200-600. Research Publication; Adams, N., Entwistle,
C., Eds.; British Museum: London, UK, 2011; Volume 177, pp. 10–24.
2. Scukin, M.; Bazan, I. L’origine du style cloisonné de l’époque des Grandes Migrations. In La Noblesse et Les
Chefs Barbares du IIIe au VIIe Siècle: Mémoires Publiés par l’Association Française d’Archéologie Mérovingienne V;
Vallet, F., Kazinski, M., Eds.; Association Française D’archéologie Mérovingienne: Société des amis du
Musée des Antiquités Nationales: Rouen, France, 1993; pp. 63–69.
3. Mathis, F.; Vrielynck, O.; Laclavetine, K.; Chêne, G.; Strivay, D. Study of the provenance of Belgian
Merovingian garnets by PIXE at IPNAS cyclotron. Nucl. Instrum. Methods Phys. Res. B 2008, 266, 2348–2352.
4. Calligaro, T.; Colinart, S.; Poirot, J.-P.; Sudres, C. Combined external-beam PIXE and μ-Raman
characterisation of garnets used in Merovingian jewellery. Nucl. Instrum. Methods Phys. Res. B 2002, 189,
320–327.
5. Farges, F. Mineralogy of Louvres Merovingian garnet cloisonné jewelry: Origins of the gems of the first
kings of France. Am. Mineral. 1998, 83, 323–330.
6. Bimson, M.; La Neice, S.; Leese, M. The characterisation of mounted garnets. Archaeometry 1982, 24, 51–58.
7. Quast, D.; Schüssler, U. Mineralogische Untersuchungen zur Herkunft der Granate merowingerzeitlicher
Cloisonnéarbeiten. Germania 2000, 78, 75–96.
8. Gilg, H.A.; Gast, N.; Calligaro, T. Vom Karfunkelstein. In Karfunkelstein und Seide: Neue Schätze aus Bayerns
Frühzeit; Wamser, L., Ed.; Friedrich Pustet Verlag: München, Germany, 2010; pp. 87–100.
9. Bugoi, R.; Oanta-Marghitu, R.; Calligaro, T. IBA investigations of loose garnets from Pietroasa, Apahida
and Cluj-Someseni treasures (5th century AD). Nucl. Inst. Methods Phys. Res. Sect. B 2015, 371, 401–406.
10. Thoresen, L. Archaeogemmology and Ancient Literary Sources on Gems and their Origins. In Gemstones in
the First Millennium AD: Mines, Trade, Workshops and Symbolism; Greiff, S., Hilgner, A., Quast., D., Eds.;
Römisch Germanisches Zentralmuseum: Mainz, Germany, 2017; pp. 155–218.
11. Henderson, R.R. Determining Chemical Composition of the Silicate Garnets Using Raman Spectroscopy.
Master’s Thesis, The Univeristy of Arizona, Tuscon, AZ, USA, 2009.
12. Dubessy, J.; Caumon, M.-C.; Rull Pérez, F. Raman Spectroscopy Applied to Earth Sciences and Cultural Heritage:
University Textbook; Dubessy, J., Caumon, M.-C., Rull Pérez, F., Eds.; European Mineralogical Union:
London, UK, 2012.
13. Smith, G.D.; Clark, J.H.R. Raman microscopy in archaeological science. J. Archaeol. Sci. 2004, 31, 1137–1160.
14. Stockton, C.M.; Mason, D.V. A Proposed New Classification for Gem-Quality Garnets. Gems. Gemol. 1985,
21, 205–218.
15. Will, T.M. Thermodynamics of solid solutions. In Phase Equilibria in Metamorphic Rocks: Thermodynamic
Background and Petrological Applications; Springer: Berlin, Germany, 1998; pp. 5–17.
16. Merli, M.; Callegari, A.; Cannllo, E.; Caucia, F.; Leona, M.; Oberti, R.; Ungaretti, L. Crystal-chemical
complexity in natural garnets: Structural constraints on chemical variability. Eur. J. Mineral. 1995, 7, 1239–
1249.
17. Grew, E.; Locock, A.; Mills, S.J.; Galuskina, I.; Galuskin, E.; Hålenius, U. IMA Report Nomenclature of the
garnet supergroup. Am. Mineral. 2013, 98, 785–811.
18. Kloprogge, J.T. Infrared and Raman spectroscopy of minerals and inorganic materials. In Encyclopedia of
Spectroscopy and Spectrometry, 3rd ed.; Lindon, J.C., Tranter, G.E., Koppenaal, D.W., Eds.; Elsevier:
Amsterdam, The Netherlands, 2017; Volume 1; pp. 274–276.

Minerals 2020, 10, 325 19 of 21
19. Krippner, A.; Meinhold, G.; Morton, A.C.; Eynatten, H. Evaluation of garnet discrimination diagrams using
geochemical data of garnets derived from various host rocks. Sediment. Geol. 2014, 306, 36–52.
20. Mingsheng, P.; Mao, H.K.; Dien, L.; Chao, E.C.T. Raman spectroscopy of garnet-group minerals. Chin. J.
Geochem. 1994, 13, 176–183.
21. Pinet, M.; Smith, D.C. Raman microspectrometry ofgarnets X3Y2Si3O12: 1. The natural calcic series
uvarovite–grossular–andradite. Schweiz. Mineral. Petrogr. Mitt. 1993, 73, 21–40.
22. Pinet, M.; Smith, D.C. Raman microspectrometry of garnets X3Y2Z3O12 2. The natural aluminium series
pyrope-almandine-spessartine. Schweiz. Mineral. Petrogr. Mitt. 1994, 74, 161–179.
23. Smith, D.C. The RAMANITA© method for non-destructive and in situ semi-quantitative chemical analysis
of mineral solid-solutions by multidimensional calibration of Raman wavenumber shifts. Spectrochim. Acta
A Mol. Biomol. Spectrosc. 2005, 61, 2299–2314.
24. Gilg, H.A.; Gast. N. Determination of titanium content in pyrope by Raman spectroscopy. J. Raman
Spectrosc. 2016, 47, 486–491.
25. Geiger, C.A. A tale of two garnets: The role of solid solution in the development toward a modern
mineralogy. Am. Mineral. 2016, 101, 1735–1749.
26. Kuang, Y.; Xu, J.; Li, B.; Ye, Z.; Huang, S.; Chen, W.; Zhang, D.; Zhou, W.; Ma, M. Crystal-Chemical
Properties of Synthetic Almandine-Pyrope Solid Solution by X-Ray Single-Crystal Diffraction and Raman
Spectroscopy. Crystals 2019, 9, 541.
27. Ferrero, S., Angel, R.J. Micropetrology: Are Inclusions Grains of Truth? J. Petrol. 2018, 59, 1671–1700.
28. Gems-inclusions all about inclusions in gemstones. Inclusions types—By time of entrapment. Available
online: https://www.gems-inclusions.com/inclusions-types/by-type-of-trapping/for-solid-inclusions-2/
(accessed on 20 August 2019).
29. Yardley, B.W.D.; Mackenzie, W.S.; Guilford, C. Atlas of metamorphic rocks and their textures. Terra Nova
1991, 3, 217–218.
30. Gem-A, The gemmological association of Great Britain. News & Publication: The source of garnets found
at the Arikamedu archaeological site in South India. Available online: https://gem-a.com/news-
publications/news-blogs/journal-digests/journal/the-source-of-garnets-found-at-the-arikamedu-
archaeological-site-in-south-india (accessed on 20 August 2019).
31. Schmetzer, K.; Gilg, H.A.; Schussler, U.; Panjikar, J.; Calligaro, T.; Perin, P. The Linkage between garnets
found in India at the Arikamedu archaeological site and their source at the Garibpet deposit. J. Gemm. 2017,
35, 598–627.
32. Horváth, E.; Bendő, Z. Provenance study on a collection of loose garnets from a gepidic period grave in
Northeast Hungary. Archeometriai Muh. 2011, 8, 17–32.
33. Kataria, P. Book review: Geology of Rajasthan: Status and perspective. J. Geol. Soc. India 2000, 55, 452–453.
34. Rameshchandra Phani, P. Mineral Resources of Telangana State, India: The Way Forward. Int. J. Innov. Res.
Sci. Eng. Technol. 2014, 3, 15450–15459.
35. Suomen Geologinen Seura—Geologiska Sällskapet i Finland, The Geological Society of Finland. Articles in
press, 1980, 52. Available online: http://www.geologinenseura.fi/bulletin/Volume52/index.html (accessed
on 20 August 2019).
36. Maharaj, D. Chemical Classification of Gem Garnets. Master’s Thesis, University of Pretoria, Pretoria,
South African, November 2015.
37. Min-dat.org—Portugal. Available online: https://www.mindat.org/loc-14425.html (accessed on 27 January
2020).
38. Caincross, B. Geology of East Africa. In Minerals & Gemstones of East Africa; Struik Nature: Cape Town,
South Africa, 2019.
39. Lux, J.; Ravnik, J. Poskus rekonstrukcije obsega poznoantičnega grobišča Lajh v Kranju. Varst. Spomenikov
2008, 43–69.
40. Stare, V.; Stopar, B.; Žgur, A.; Goričan, A.; Čenčič, L.; Habič, S.; Petru, P.; Vinski, Z.; Kiszely, I. Kranj:
Nekropola iz Časa Preseljevanja Ljudstev; Narodni Muzej: Ljubljana, Slovenia, 1980.
41. Inštitut za arheologijo ZRC SAZU. Strani za študente arheologije zgodnjega srednjega veka na FF v
Ljubljani; Poznoantično obdobje. Available online: http://iza.zrc-sazu.si/ff/slavko/pa_grobisca.html
(accessed on 10 December 2019).
42. Milavec, T. Prispevek h kronologiji S-fibul v Sloveniji. Arheol. Vestn. 2007, 58, 333–355.
43. Gregorietti, G. Jewelry through the Ages; Hamlyn: London, UK, 1970; p. 139.

Minerals 2020, 10, 325 20 of 21
44. Kramar, S.; Dolenec, M.; Lux, J. Characterisation of 6th century fibulae with gem materials (cementary Lajh
in Kranj, Slovenia) by means of XRF and micro-Raman spectroscopy. In Proceedings of the 6th International
Congress on the Application of Raman Spectroscopy in Art and Archaeology, Parma, Italy, 5–8 September
2011; Timeo: Bologna, Italy, 2011; pp. 172–173.
45. Šmit, Ž.; Fajfar, H.; Jeršek, M.; Knific, T.; Lux, J. Analysis of garnets from the archaeological sites in Slovenia.
Nucl. Inst. Methods Phys. Res. Sect. B 2014, 328, 89–94.
46. Nemeček, N.; Kramar, S.; Podobnik, T. Interdisciplinarni pristop—Konserviranje-restavriranje,
naravoslovne preiskave in arheološka interpretacija dveh zaponk z grobišča Lajh. In Konservator—
Restavrator: Povzetki Strokovnega Srečanja 2016; Nemeček, N., Ed.; Skupnost Muzejev Slovenije, Društvo
Restavratorjev Slovenije: Ljubljana, Slovenia, 2016; p. 55.
47. Sibi, N.; Subodh, G. Structural and microstructural correlations of physical properties in natural
almandine-pyrope solid solution: Al70Py29. J. Electron. Mater. 2017, 46, 6947–6956.
48. Kolesov, B.; Geiger, A.C. Raman spectra of silicate garnets. Phys. Chem. Miner. 1998, 25, 142–151.
49. Scholz, R.; Frost, R.L.; Xi, Y.; Graça, L.M.; Lagoeiro, L.; López, A. Vibrational spectroscopic characterization
of the phosphate mineral phosphophyllite—Zn2Fe(PO4)2·4H2O, from Hagendorf Süd, Germany and in
comparison with other zinc phosphates. J. Mol. Struct. 2013, 1039, 22–27.
50. Lafuente, B.; Downs, R.T.; Yang, H.; Stone, N. The power of databases: The RRUFF project. In Highlights in
Mineralogical Crystallography; Armbruster, T., Danisi, R.M., Eds.; W. De Gruyter: Berlin, Germany, 2015; pp.
1–30.
51. Tan, W.; Wang, C.; He, H.; Xing, C.; Liang, X.; Dong, H. Magnetite-rutile symplectite derived from ilmenite-
hematite solid solution in the Xinjie Fe-Ti oxide-bearing, mafic-ultramafic layered intrusion (SW China).
Am. Mineral. 2015, 100, 2348–2351.
52. Tan, W.; He, H.; Wang, C.; Dong, H.; Liang, X.; Zhu, J. Magnetite exsolution in ilmenite from the Fe-Ti
oxide gabbro in the Xinjie intrusion (SW China) and sources of unusually strong remnant magnetization.
Am. Mineral. 2016, 101, 2759–2767.
53. Eppler, W.F. The Origin of Healing Fissures in Gemstones. J. Gemm. 1959, 7, 40–66.
54. Alex Strekeisen. Plutonic Rocks, Pleochroic halo. Available online:
http://www.alexstrekeisen.it/english/pluto/ pleochroichalo.php (accessed on 20 August 2019).
55. Nasdala, L.; Wenzel, M.; Vavra, G.: Irmer, G.; Wenzel, T.; Kober, B. Metamictisation of natural zircon:
Accumulation versus thermal annealing of radioactivity-induced damage. Contrib. Mineral. Petrol. 2002,
143, 767–768.
56. Nasdala, L.; Akhmadaliev, S.; Artac, A.; Chanmuang, N.C.; Habler, G.; Lenz, C. Irradiation effects in
monazite-(Ce) and zircon: Raman and photoluminescence study of Au-irradiated FIB foils. Phys. Chem.
Miner. 2018, 45, 855–871.
57. United ID RAMAN LAB. Applications/Solutions: Gemstones. Raman Spectroscopic Inspection and
Analysis of Zircon Inclusion in Corundum—Effect of Heat Treatment on Zircon Inclusion. Available
online:
https://static1.squarespace.com/static/55e5983ee4b043c85db4fd4c/t/5d1e1fd511031d0001ec94a3/156225532
2783/Raman+Spectroscopic+Inspection+and+Analysis+of+Zircon+Inclusion+in+Corundum.pdf (accessed
on 21 March 2020).
58. Tlili, A.; Smith, D.C.; Beny, J.M.; Boyer, H. A Raman microprobe study of natural micas. Mineral. Mag. 1989,
53, 165–179.
59. Schönig, J.; Meinhold, G.; von Eynatten, H.; Lünsdorf, N.K. Provenance information recorded by mineral
inclusions in detrital garnet. Sediment. Geol. 2018, 376, 32–49.
60. Calligaro, T. Probing Works of Art with Photons and Charged Particles. In Spectroscopy of Emerging
Materials. NATO Science Series II: Mathematics, Physics and Chemistry, Volume 165; Faulques, E.C., Perry,
D.L., Yeremenko, A.V., Eds; Springer: Dordrecht, Germany, 2004.
61. Force, E.R. The provenance of rutile. J. Sediment. Res. 1980, 50, 485–488.
62. Zack, T.; von Eynatten, H.; Kronz, A. Rutile geochemistry and its potential use in quantitative provenance
studies. Sediment. Geol. 2004, 171, 37–58.
63. Bucher, K.; Frey, M. Petrogenesis of Metamorphic Rocks; Bucher, K., Frey, M. Eds.; Springer-Verlag:
Berlin/Heidelberg, Germany, 2002.
64. Cesare, B.; Salvioli Mariani, E.; Venturelli, G. Crustal anatexis and melt extraction in the restitic xenoliths
at El Hoyazo (SE Spain). Mineral. Mag. 1997, 61, 15–27.

Minerals 2020, 10, 325 21 of 21
65. Acosta-Vigil, A.; Buick, I.; Hermann, J.; Cesare, B.; Rubatto, D.; London, D.; Morgan, G.B. Mechanisms of
crustal anatexis: A geochemical study of partially melted metapelitic enclaves and host dacite, SE Spain. J.
Petrol. 2010, 51, 785–821.
© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).