Archaeological and Anthropological Sciences (2020) 12:3

https://doi.org/10.1007/s12520-019-00953-x

ORIGINAL PAPER

Trace elements content and cause of color in ancient treated

carnelian and its natural counterpart from SE Asia
Seriwat Saminpanya 1 & Chatree Saiyasombat 2 & Narong Chanlek 2 & Nirawat Thammajak 3 & Ekkasit Sirisurawong 2 &
Rattanavalee Viriyasunsakun 1 & Phusuda Kingkanlaya 1 & Patcharin Rakponramuang 1

Received: 16 March 2019 / Accepted: 16 December 2019

# Springer-Verlag GmbH Germany, part of Springer Nature 2020

Abstract
Iron Age treated carnelian beads and their natural counterparts from SE Asia have been discussed in terms of attribution of
transition and lighter elements in their matrix. They were investigated by using synchrotron techniques (XRF: X-ray fluores-
cence, XANES: X-ray absorption near edge structure, and XPS: X-ray photoelectron spectroscopy) and SEM-EDS: scanning
electron microscope-energy dispersive spectroscopy. Ferrihydrite or ferrihydrite + goethite have been found as the causes of
color. The treated stones were subjected to the sugar + sulfuric acid treatment of about 1800-year BP. This work has fulfilled
information about the treatment in the ancient chalcedony to the archeomineralogy archive.

Keywords Chalcedony . Carnelian . Sugar . Treatment . Sulfuric acid . Carbon

Introduction The answer is that they have long been putative to be exported
from India (Lamb 1965; Bellwood 1976). However, the potential
Chalcedony, fibrous polycrystalline quartz (Frondel 1962; sources could be originated from India, Sri Lanka, and Thailand
Klein and Hurlbut Jr 2007) has been subdivided into several (Theunissen et al. 2000). Several works throw new light on the
varieties e.g. carnelian, sard, onyx, and agate. (Frondel 1962; relations between India and SE Asia in the first millennium BC
Webster 1995). They have been treated and used all over the (e.g. Bellina 2002, 2003; Ramli et al. 2009; Carter 2015; Bellina
world since prehistoric and ancient times. For example, car- 2016). Carter and Dussubieux (2016) used the geochemical sig-
nelian beads from the tombs of the Iron Age, ~ 3rd centuries in nature to prove the sources of the Iron Age carnelian beads from
Sa Kaeo province, eastern Thailand, are introduced here to be SE Asia, and they concluded that most samples were made from
investigated (Fig. 1a–b). However, the provenance of the an- raw materials from the Deccan Traps, northwest India. Our sam-
cient carnelian beads in Thailand and Southeast Asia is still ples are found to have been treated into black or dark brownish
unclear. Several researchers attempted to prove their origin black to imitate onyx or black onyx. There are some evidences
and trade (e.g. Insoll et al. 2000; Theunissen et al. 2000). including when a cylinder bead was cut in the transverse section

* Seriwat Saminpanya Phusuda Kingkanlaya

seriwat@g.swu.ac.th izups-us@live.com

Chatree Saiyasombat Patcharin Rakponramuang

chatree@slri.or.th Ppatcharin1995@gmail.com

Narong Chanlek 1
Department of General Science, Faculty of Science, Srinkharinwirot
narong@slri.or.th
University, 114 Sukhumvit 23 Road, Khlong Toey Nua, Watthana,
Nirawat Thammajak Bangkok 10110, Thailand
n.thammajak@gmail.com 2
Synchrotron Light Research Institute, 111 University Avenue,
Ekkasit Sirisurawong Muang District, Nakhon Ratchasima 30000, Thailand
ekkasit@slri.or.th 3
Department of Chemistry, Faculty of Science, Suranaree Univerity of
Rattanavalee Viriyasunsakun Technology, 111 University Avenue, Suranaree Sub-District, Muang
Jeweliiz.pby@hotmail.com Nakhon Ratchasima District, Nakhon Ratchasima 30000, Thailand
 3 Page 2 of 11 Archaeol Anthropol Sci (2020) 12:3

Fig. 1 a Map of Thailand

showing localities of samples. b
Two upper rows, ancient bead
samples (from left to right, NPW1
to NPW9 and KP1 to KP9) and
bottom row, natural samples
(from left to right, HDL1 to
HDL9. c Certain samples
embedded in the epoxy resin,
attached to a glass slide (27 ×
46 mm) and cut in cross-section.
The color of NPW8 is not natural.
d A closer look for NPW8 under
transmitted light of an optical
microscope, seen as the carbon
concentration in cracks (arrows)

and it is seen as uneven colors, it is darker at the rim and becom- Materials and methods
ing gradually lighter to the core (Fig. 1c). Such black stones have
been known to be treated since ancient times but no one had Twenty-seven samples of carnelian (from three localities:
revealed until 1819 that the sugar-acid treatment process for Nong Phak Waen, Khok Plai, and Huai Di Lert) were collected
black agate was from Italy (Nassau 1994). Our stone was more (Fig. 1). Eighteen samples of the first two localities are the an-
likely to be treated by this method. Generally, colors of agate and cient beads exhumed from the tombs in Sa Kaeo province,
carnelian can be enhanced by heat treatment and other methods Thailand. Their appearance includes orange and dark brown to
(Francis Jr 2002; Kenoyer 2003; McClure 2009; Yazdi et al. black with the shapes of a sphere, truncated hexagonal bi-
2016). They can be treated to obtain different colors e.g. red by
Fe-nitrate + heating; brown by sugar + strong heating; black by
sugar + sulfuric acid (Farrington and Laufer 1927). Several
works dealt with ancient and modern chalcedonies (e.g. Insoll
et al. 2004; Law et al. 2007; Hatipoğlu et al. 2010; Gliozzo et al.
2014). Although Kenoyer et al. (1991) and Domański et al.
(2009) can prove the heat treatment on modern and ancient
artifacts. Kenoyer et al. (1991) mentioned the technology of
heating carnelian beads from India. However, the treatment in-
formation on ancient materials is still lacking, especially for the
elemental interaction to the matrix after treatment even the causes
of color for the carnelian itself. In this paper, we have revealed the
mechanism and technology of the treatment and explained the
chemical ions affected and unaffected by the treatment. The find-
ing of the cause color of carnelian will be addressed. This study
has unveiled the technology of chalcedony treatment during the
Iron Age for people understanding of how to improve gemstone
Fig. 2 Trace elements from XRF analyses for analyzed samples of NPW,
appearance. Lastly, the findings possibly mark another timeline KP, and HDL with a logarithmic scale on the intensity axis, and a linear
of gemstone enhancement in prehistoric times. scale on the energy axis
 Archaeol Anthropol Sci

Table 1 Relative contents (atomic %) of each element found in the samples by XRF and fitted by PyMCA

HDL1 HDL1Or HDL1W HDL2 HDL4 HDL5 HDL6 HDL9 KP1 KP2 KP4 KP5 KP7 NPW1 NPW4 NPW5 NPW7 NPW8 Average
(2020) 12:3

Ca 3.3 1.0 1.7 1.1 5.2 0.7 0.5 0.9 2.2 1.6 2.1 2.7 3.0 5.6 4.3 1.0 2.0 5.8 2.5
Sc 0.2 0.2 0.3 0.1 0.1 0.1 0.1 0.1 0.2 0.1 0.2 0.2 0.2 0.7 0.1 0.1 0.1 0.1 0.2
Ti 1.1 1.7 1.9 1.6 0.7 0.1 0.1 0.0 1.2 0.4 2.4 1.4 1.0 1.0 0.9 0.5 0.9 0.6 1.0
V 0.1 0.1 0.0 0.0 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.1 0.3 0.0 0.1 0.1 0.1 0.1
Cr 0.5 0.2 0.6 0.2 0.7 0.2 0.1 0.2 0.4 0.4 0.2 0.5 0.3 0.2 0.4 0.1 0.5 0.5 0.3
Mn 5.4 2.1 5.8 1.5 5.2 1.7 1.3 3.6 4.2 4.5 2.2 4.8 4.1 4.6 3.0 1.4 9.2 5.6 3.9
Fe 42.2 73.3 50.4 83.0 49.9 83.1 84.5 73.7 48.2 50.9 70.4 42.1 47.9 27.3 51.3 80.3 53.6 33.3 58.1
Co 0.1 0.3 0.2 0.4 0.2 0.3 0.3 0.3 0.2 0.2 0.3 0.1 0.1 0.1 0.2 0.3 0.3 0.1 0.2
Ni 35.9 17.2 31.6 6.5 33.7 12.4 11.8 19.1 28.7 30.2 16.5 35.1 36.7 37.2 23.0 8.4 23.6 33.8 24.5
Cu 4.6 1.4 2.8 2.5 2.1 0.7 0.8 1.0 11.0 9.4 2.6 9.2 3.6 20.5 12.9 4.9 4.0 13.8 6.0
Zn 6.6 2.3 4.6 3.2 2.0 0.7 0.5 0.9 3.4 2.0 2.7 3.7 2.9 2.4 3.4 1.4 4.8 3.4 2.8
I 0.1 0.1 0.1 0.0 0.1 0.1 0.1 0.1 0.2 0.1 0.1 0.2 0.2 0.1 0.1 0.0 0.1 0.1 0.1
Cs 0.0 0.0
Ba 0.5 1.5 0.5 2.6 1.3
La 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
Pm 0.0 0.0 0.1 0.1 0.1
Gd 0.1 0.3 0.0 0.1
Total 100.1 99.9 100.0 100.1 100.0 100.1 100.1 99.9 100.0 99.8 99.8 100.1 100.1 100.0 100.1 100.1 100.0 99.8 101.1
Page 3 of 11
3
 3 Page 4 of 11 Archaeol Anthropol Sci (2020) 12:3

Fe/10
HDL

KP

NPW

Mn Cu
Fig. 3 Plot of Fe/10-Mn-Cu of the samples, HDL contains a higher
content of Fe and NPW contains higher Cu

pyramid or truncated cone, and cylinder. The last locality is a

natural geological occurrence, Huai Di Lert, Lop Buri province,
Fig. 5 Representative graphs of three Fe groups comparing to standards
Thailand and they were collected for comparison.
Synchrotrons applied for other minerals in the archeological
major silicon content was not analyzed due to the limitation of
objects have been reported e.g. Farges (2005) and Chalmin et al.
the excited X-ray power of the experiment. A highly sensitive
(2006). In this work, we used techniques of X-ray fluorescence
Canberra 19-element Ge solid-state detector measured the
(XRF), X-ray absorption near edge structure (XANES), and X-
fluorescence photons emitted from the samples after excited
ray photoelectron spectroscopy (XPS) for our samples. The XRF
by 10 keVand 14 keV X-ray beam. The PyMCA program was
and XANES were accomplished at the beamline 1.1 W
used to calculate spectral weights of trace element contents
(BL1.1 W) and the XPS at the beamline 5.3 (BL5.3) of the
semi-quantitatively (Solé et al. 2007).
Synchrotron Light Reseach Institute (SLRI), Thailand. The
The XANES probed the main elements Cu, Ni, Fe, and Mn
SLRI storage ring is operated at 1.2 GeVof electron energy with
which were postulated to be related to the stone treatment. The
an electron beam current between 80 and 140 mA.
K-edge XANES spectra for these elements were recorded in
The 2.2-Tesla multipole wiggler is installed at BL1.1 W to
fluorescence yield mode, using the same Ge detector as for the
produce sufficient photon flux in the region from 4 to 18 keV
XRF experiments. A standard foil for each element was used
region. The energy of X-rays was selected by the double-
to calibrate X-ray energies by setting the absorption edges of
crystal X-ray monochromator equipped with Si (111) crystals.
Cu, Ni, Fe, and Mn to 8979 eV, 8333 eV, 7112 eV, and
The XRF analyzed trace elements contents in the samples in
6539 eV, respectively. Furthermore, the standard compounds
the area of the cut and polished surfaces of the samples. The
CuO, Cu2O, CuSO 4, NiO, NiO2, FeO, Fe2O 3, goethite
FeO(OH), ferrihydrite Fe3+10O14(OH)2), MnO, Mn3O4,
Mn2O3, and MnO2 were measured for spectral comparison
in a transmission mode by using ionization chambers. At least
two spectra from each sample were merged to assure precision
and to gain a high signal-to-noise level. NPW8 sample was
additionally measured from the black to the white areas.
At BL5.3, the XPS measured the binding energies of ligh-
ter elements: C, N, O, Mg, Si, Ca, and Fe which were expected
to be affected by the treatment mechanism. This used a PHI
5000 Versa Probe II Scanning XPS (ULVAC-PHI) equipped
with Al Kα radiation source (1486.8 eV). To eliminate surface
contamination from sample preparation processes, the sam-
ples were sputtered inside the XPS preparation chamber with
3 keV Ar ions for 60 s before measurements. All binding
energies were calibrated with the C 1s (C-C bond) peak at a
Fig. 4 The Cu XANES spectra of some samples compared to those of the binding energy of 284.8 eV.
standards. The spectra of NPW4, NPW8, and KP4 could be a mixture of SEM-EDS (scanning electron microscope-energy disper-
Cu2O and CuSO4. In contrast, NPW5 and NPW7 could be a mixture of sive spectroscopy), FEI QUANTA 450 was employed for
Cu2O and CuO
Archaeol Anthropol Sci (2020) 12:3 Page 5 of 11 3

Fig. 6 Fe XANES spectra of

NPW8 in different areas, from
black to white (P1 to P5)

inspecting a variation of the elements in the area of the

polished surface of the samples. The elemental mapping and
spot analyses were undertaken in high vacuum mode with an
accelerating voltage of 20 keV. Then, the EDS data was ana-
lyzed by Aztec software version 3.3.
A contemporary shell bead collected from Ra Boek Kham
~ 1.5 km away to the southwest of Khok Plai was dated by the
DirectAMS, radiocarbon dating service, USA and yielded the
age of 1811 ± 27 year BP. The detail of the experiment proto-
col and the shell bead is published elsewhere. This age is used
to correlate with this research because an archeological age of
carnelian samples cannot be obtained directly. The ancient
precious organic and inorganic materials in this area are found
in the same Holocene sediment and the depth of their tombs is
in the same elevation.

Fig. 7 a Fe pre-edge XANES spectra of position P1, P3, and P5. b In the
inset, the experimental Fe(III) pre-edge peak (black dot) and peak fitting
using Gaussian peaks (green lines) and their accumulated fitted peaks (red
lines) representing the transitions to localized electronic states. The gray
ellipses enclose the values of Fe(III) pre-edge peak integrated areas versus
their centroid positions, measured on tetra-, penta-, and hexa-coordinated
Fe(III) reference compounds [labelled (4), (5), and (6), respectively (Giuli Fig. 8 The unsmooth spectra of Mn in the samples indicates the low
et al. 2011)]. P1 is penta-coordinated Fe(III) and P5 is hexa-coordinated content and the K-edge energy is higher than the Mn foil standard
Fe(III) which suggests the Mn should have the valency > 0
 3 Page 6 of 11 Archaeol Anthropol Sci (2020) 12:3

Table 2 Relative atomic concentration of the elements observed on the

NPW8

Relative atomic concentration (%)

Si C Ca N O Mg

NPW8 (black area 1) 18.84 36.51 0.73 3.54 39.65 0.72

NPW8 (black area 2) 18.06 38.89 0.87 3.84 37.46 0.87
NPW8 (white area 1) 16.70 40.60 0.98 4.23 36.39 1.09
NPW8 (white area 2) 13.06 52.87 1.00 3.79 28.43 0.84

Results and discussion

XRF
Fig. 9 XPS survey scan of black and white areas of NPW8 showing Mg,
O, N, Si, Ca, and C in the sample
Trace element contents in Fig. 2 and Table 1 indicate that the
top five elements Fe, Ni, Cu, Mn, and Zn enriched the sam-
ples. Only Cu, Fe, and Mn absorption edges were measured
for concentration, and they were believed to contribute to the

Fig. 10 High-resolution XPS spectra of Si 2p, C 1s, Ca 2p, N 1s, O 1s, and Mg 1s indicate no difference between black and white areas of NPW8
Archaeol Anthropol Sci (2020) 12:3 Page 7 of 11 3

Fig. 11 SEM-EDS for NPW8 at

the area of black rim. a Electron
image. b, c, d The mapping
images showing the distribution
of Si, O, and C, respectively and
the C disperses all over the area. e
The result of spot analysis which
was unable to detect a low
concentration of C at the point. f
The result of mapping analysis
shows very high carbon content,
22.1 wt%

color of the samples. The samples from HDL contain a higher NPW4, NPW8, and KP4 samples could be a mixture of Cu2O
content of Fe but the samples from KP and NPW contain and CuSO4. In contrast, Cu in NPW5 and NPW7 could be a
higher Cu (Fig. 3), suggesting that the studied samples are mixture of Cu2O and CuO. These point out that Cu does not
not from the same provenance. The ancient samples could affect the colors of the samples.
have been imported from outside the country (cf. Carter and
Dussubieux 2016), possibly, India. Iron

XANES Fe XANES results can be separated into 3 groups (Fig. 5 and

Fig. A.2 in the supplementary section) ranging from the low-
Copper est to the highest major peaks at ~ 7133 eV; group I is the
untreated sample, HDL1 white, group II includes HDL1 or-
The Cu XANES spectra reveal that coppers in all samples are ange, HDL5, HDL9, and NPW8, and group III contains
the mixture of Cu1+ and Cu2+ except naturally untreated sam- NPW1, NPW4, NPW5, NPW7, KP1, and KP4. The last group
ples. NPW4, NPW8, and KP4 have similar XANES features gives the spectra matching to that of ferrihydrite, a poorly
and have the highest absorption edge comparing to NPW5 crystalline oxide/hydroxide (Cudennec and Lecerf 2006; Das
(Fig. 4). NPW7 has the lowest absorption edge, indicating that et al. 2011). However, the shape of peaks KP4, NPW1, and
it has the highest Cu1+ among all studied samples. The natural NPW5 are more likely to be a mixture of ferrihydrite and
untreated samples, HDL (Fig. A.1, in supplementary section), goethite, especially KP4 with the superior goethite.
show a noisy data caused by a very low Cu content in the Transformation of ferrihydrite to goethite can exist
matrix. Therefore, no conclusion is made for these untreated (Cudennec and Lecerf 2006; Das et al. 2011; Jones et al.
samples. Linear combination fitting (LCF) suggests that Cu in 2017). This is possible to concomitantly find them in a single
 3 Page 8 of 11 Archaeol Anthropol Sci (2020) 12:3

Fig. 12 SEM-EDS for NPW8 at

the area of white core. a Electron
image. b, c, d are the mapping
images showing the distribution
of Si, O, and C, respectively and
the C disperses all over the area. e
The result of spot analysis which
was unable to detect a low
concentration of C at the point. f
The result of mapping analysis
shows very high carbon content,
15.2 wt%

sample. Fritsch and Rossman (1988) reported that carnelian is color in chalcedony. However, Fe in the black area of NPW8
colored by hydrous Fe-oxides or submicroscopic to micro- is surprisingly lower than in the white area.
scopic inclusions of hydrous Fe-oxides. Therefore, the present The Fe3+ in the black samples i.e. NPW5 and NPW8 re-
study has clarified the issue. The lower height of the major ceived an electron from the surrounding environment, proba-
peak is associated with how iron behaves in the matrix and the bly by oxidation condition, then it transformed to Fe2+. This
order of site where iron atoms are located. The iron in untreat- process might have been associated with H2SO4, and this is
ed HDL samples should locate in less order location. The related to CuSO4 which is possibly part of NPW8 as sug-
treated samples (NPW) might have been subjected to a heating gested by LCF. The source of Cu found in the sample is still
process. This caused the structure of the matrix to be more unclear or it might be from an inherited natural structure in the
rigid and the iron location distribution to be narrower or caus- chalcedony matrix. If the samples were treated with H2SO4,
ing the iron atoms to even interact with the matrix during the then this acid reacted with the Cu in the matrix.
process. According to the edge step, the orange area of un- The treated sample, NPW8, was cut to show the different
treated samples (HDL) has a higher value of Fe. This suggests color zones (black and white zones) and the areas were then
that higher iron content which is normal for Fe causing orange analyzed to see the different valences of Fe (Fig. 6). The
Archaeol Anthropol Sci (2020) 12:3 Page 9 of 11 3

Fig. 13 SEM-EDS for NPW7 at

the color zoning areas. a Electron
image; b, c, d The mapping
images showing the distribution
of Si, O, and C, respectively and
the C distributes all over the area.
e The result of spot analysis
which was unable to detect a low
concentration of C

results show that the Fe in black and white areas cannot be each measured position. The Gaussian fitting (Fig. 7b) is used
differentiated by only the absorption edges of the sample. The to determine the centroid position and peak area of the pre-
pre-edge Fe XANES spectra in Fig. 7a show the difference in edge peak cf. Giuli et al. (2011). P1 is found to be penta-

Fig. 14 SEM-EDS for HDL2. a

Electron image. b The result of
spot analysis which was unable to
detect a low concentration of C or
the sample is possibly C-free
 3 Page 10 of 11 Archaeol Anthropol Sci (2020) 12:3

coordinated Fe(III) while P5 is hexa-coordinated Fe(III). result contradicts the data from XPS, mentioned above. This
These results suggest that the Fe oxidation state does not play can be explained that the latter technique provides only infor-
a major role in the black and the white colors in the treated mation from the surface of the sample.
samples. The treatment to make a black color may be relevant
to iron symmetry where the iron situates in the matrix and this
requires further research.
Conclusions
Nickel and manganese
The synchrotron (XANES) reveals that the causes of natural
Ni has been reported to be the cause of color in green chalce- orange color in carnelian include ferrihydrite or ferrihydrite +
dony (chrysoprase) as inclusions of Ni-compounds in the form goethite inclusions. Synchrotron and SEM-EDS help explain
of colloidal particles scattered throughout the silica matrix that the ancient carnelian samples were treated with the sugar
(Sachanbiński et al. 2001; Shigley et al. 2009). However, the and sulfuric acid as still in use at the present time of black
studied samples are mainly in orange and the Fe-oxide is a color treatment. The cause of black color in ancient treated
dominant cause of their natural color. Moreover, the Ni con- carnelian beads was in the process developed by sugar + sul-
tent in the samples is very low and its spectra are hard to be furic acid as a reaction which leaves the carbon in cavities and
differentiated. Apart from that, an average oxidation state of porosity of the stones as the following equation:
Mn is more than 0, and Mn occupies in the matrix structure by C 12 H 22 O 11 (sugar) + H 2 SO 4 (sulfuric acid) → 12C
linking with other ions, probably oxygen. It is slightly differ- (carbon) + 11H2O (water) + mixture water and acid
ent for peak M1 position (Fig. 8) in Mn XANES spectra of This reaction allows H2SO4 to bond with other cations e.g.
black and white zones in the sample NPW8, which may link the Cu in the chalcedony matrix. Then free carbon will be
between the Mn and the treatment in the sample. However, concentrated in the cracks/fractures in the chalcedony host
there is the noise of Mn XANES spectra which prevents any which causes the stone to become black or very dark brown.
solid conclusion. Other elements e.g. Fe, Mn, and Ni do not play a role in the
black color treatment process at all. For untreated carnelian,
XPS copper exists in the forms of Cu2O (Cu+) and CuO (Cu2+), and
then after treatment, they become a derivative of the treated
The XPS measurements were performed on two points in the process by two reactions below to form the CuSO4, an end
different color zones (black and white) of NPW8. The XPS product in a treated carnelian.
survey spectra (Fig. 9) reveal the presence of Si, C, Ca, N, O, (i) Copper oxide reacts with sulfuric acid:
and Mg in both black and white areas. Fe, Cu, and Ni spectra Cu2O + H2SO4 → Cu + CuSO4 + H2O
were unable to be detected due to a low atomic concentration. and
High-resolution XPS spectra of Si 2p, C 1s, Ca 2p, N 1s, O 1s, (ii) Copper(I) oxide reacts with sulfuric acid:
and Mg 1s peaks and their binding energies are in Fig. 10. The CuO + H2SO4 → CuSO4 + H2O.
calculated relative atomic concentration of these elements is The ancient carnelians from the tombs in the regions of e.g.
shown in Table 2. The result indicates no difference in the Thailand and Cambodia have been revealed their treatment
chemical state of Si, C, Ca, N, O, and Mg between the black technology and marked a gemstone treatment of the Iron Age.
and the white areas of NPW8. The ancient treatment does not
Acknowledgements We are in debt with the staff of SLRI including
change any chemical/oxidation states of these elements. Anuchit Ruangvittayanon and Thumrongsak Witchanantakul for SEM
However, the relative atomic concentration of C in the white analysis, Rachanon Klondon, Pornthep Pongchalee, Krongthong
area is slightly higher than that in the black area. Kamonsuangkasem, and Prae Chirawatkul for helping on the machine
operation and data analyses. Thanks also go to Prasertsri Sriwilai who
supported this research and to Pitthayut Saminpanya for proofreading of
SEM-EDS the manuscript.

The spot analyses by EDS reveal that the sample, NPW8, was Funding information This project was funded by the Faculty of Science,
treated by carbon (Figs. 11 and 12), while the orange ones SWU through grant (grant number 179/60).
were not treated i.e. no carbon was detected (Figs. 13 and
14). Although, the EDS performed the elemental mappings
in the general area on the polished sample surface. However,
it could not be differentiated the distribution of the element
References
between the black and the white areas of the treated sample. Bellina B (2002) Le port protohistorique de Khao Sam Kaeo en Thaïlande
However, the results show a higher C content (22.1 wt%) in péninsulaire. Bulletin de l’Ecole française d’Extrême-Orient 89:
the black area than that of the white area (15.2 wt%). This 329–343
Archaeol Anthropol Sci (2020) 12:3 Page 11 of 11 3

Bellina B (2003) Beads, social change and interaction between India and sub-Saharan Africa: the application of UV-LA-ICP-MS to carnelian
sSouth-east Asia. Antiquity 77:285–297 from Gujarat, India, and West Africa. J Archaeol Sci 31:1161–1173
Bellina B (2016) The inception of the trans-national processes between Jones AM, Collins RN, Waite TD (2017) Redox characterization of the
the Indian Ocean and the South China Sea from an early city-state on Fe(II)-catalyzed transformation of ferrihydrite to goethite. Geochim
the Thai-Malay peninsula (4th-2nd c. BCE). In: Boussac M-F, Salles Cosmochim Acta 218:257–272
J-F, Yon J-B (eds) The ports of the ancient Indian Ocean. Primus Kenoyer JM (2003) The technology of stone beads. In: Lankton J (ed) A
Books, Delhi, pp 481–510 bead timeline volume 1: prehistory to 1200 CE, a resource for iden-
Bellwood P (1976) Archaeological research in Minahasa and the Talaud tification, classification and dating. The Bead Society of Greater
Islands, northeastern Indonesia. Asian perspectives: the bulletin of Washington, Washington D.C., pp 14–19
the Far-Eastern Prehistory Association Asian Perspectives 19:241– Kenoyer JM, Vidale M, Bhan KK (1991) Contemporary stone bead mak-
288 ing in Khambhat, India: patterns of craft specialization and organi-
Carter AK (2015) Beads, exchange networks and emerging complexity: a zation of production as reflected in the archaeological record. World
case study from Cambodia and Thailand (500 BCE-CE 500). Camb Archaeol 23:44–63
Archaeol J 25:733–757 Klein C, Hurlbut CS Jr (2007) The 23rd edition of manual of mineral
Carter AK, Dussubieux L (2016) Geologic provenience analysis of agate science. John Wiley & Sons, Inc., New York, 716 pp
and carnelian beads using laser ablation-inductively coupled
Lamb A (1965) Some observations on stone and glass beads in early
plasma-mass spectrometry (LA-ICP-MS): a case study from Iron
South-east Asia. Journal of the Malaysian Branch of the Royal
Age Cambodia and Thailand. J Archaeol Sci Rep 6:321–331
Asiatic Society 38:87–124
Chalmin E, Farges F, Vignaud C, Susini J, Menu M, Brown GE Jr (2006)
Discovery of unusual minerals in Paleolithic black pigments from Law, R., Carter, A., Bhan, K., Malik, A., Glascock, M.D., 2007. INAA of
Lascaux (France) and Ekain (Spain). In: Hedman P, Pianetta P (eds) agate sources and artifacts from the Indus, Helmand, and Thailand
X-ray absorption fine structure XAFS13. American Institute of regions. Proceedings of the 19th International Conference of the
Physics, Stanford, pp 220–222 European Association of South Asian Archaeology Ravenna, Italy,
Cudennec Y, Lecerf A (2006) The transformation of ferrihydrite into 2-6 July 2007
goethite or hematite, revisited. J Solid State Chem 179:716–722 McClure S (2009) A new chalcedony treatment. Gems & Gemology 45:
Das S, Hendry MJ, Essilfie-Dughan J (2011) Transformation of two-line 288–289
ferrihydrite to goethite and hematite as a function of pH and tem- Nassau K (1994) Gemstone Enhancement, 2nd edn. Butterworth-
perature. Environmental Science & Technology 45:268–275 Heinemann, Wiltshire, 272 pp
Domański M, Webb J, Glaisher R, Gurba J, Libera J, Zakościelna A Ramli Z, Shuhaimi NH, Rahman NA (2009) Beads trade in peninsula
(2009) Heat treatment of polish flints. J Archaeol Sci 36:1400–1408 Malaysia: based on archaeological evidences. Eur J Soc Sci 10:585–
Farges F (2005) Ab initio and experimental pre-edge investigations of the 593
Mn K-edge XANES in oxide-type materials. Phys Rev B 71:155109 Sachanbiński M, Janeczek J, Platonov A, Rietmeijer FJM (2001) The
Farrington, O.C., Laufer, B., 1927. Agate: physical properties and origin, origin of colour of chrysoprase from Szklary (Poland) and Sarykul
archaeology and folk-lore. Geology Leaflet 8 (Field Museum of Boldy (Kazakhstan). Neues Jahrbuch für Mineralogie,
natural history). Chicago: Field Museum Press Abhandlungen 177:61–76
Francis P Jr (2002) Asia’s maritime bead trade. 300 B.C. to the present. Shigley JE, Laurs BM, Renfro ND (2009) Chrysoprase and prase opal
University of Hawai’i Press, Honolulu, 320 pp from Haneti, Central Tanzania. Gems & Gemology 45:271–279
Fritsch E, Rossman GR (1988) An update on color in gems. Part 3: colors Solé VA, Papillon E, Cotte M, Walter P, Susini J (2007) A multiplatform
caused by band gaps and physical phenomena. Gems & Gemology code for the analysis of energy-dispersive X-ray fluorescence spec-
24:81–102 tra. Spectrochim Acta B At Spectrosc 62:63–68
Frondel, C., 1962. Dana’s system of mineralogy, 3. Silica Minerals. New Theunissen R, Grave P, Bailey G (2000) Doubts on diffusion: challenging
York: John Wiley & Sons, 334 pp. the assumed Indian origin of Iron Age agate and carnelian beads in
Giuli G, Paris E, Cicconi MR, Hess K-U, Dingwell DB, Fehr KT, Valenti Southeast Asia. World Archaeol 32:84–105
P, Eeckhout SG (2011) XAS determination of the Fe local environ- Webster R (1995) Gems: their sources, descriptions and identification, 5th
ment and oxidation state in phonolite glasses. Am Mineral 96:631– edn. Butterworth-Heinemann, Revised by Peter G. Read. Cornwall,
636 1072 pp
Gliozzo E, Mattingly DJ, Cole F, Artioli G (2014) In the footsteps of
Yazdi M, Lotfi R, Masoudi F, Pak NM (2016) Chemical treatment effects
Pliny: tracing the sources of Garamantian carnelian from Fazzan,
of blue, yellow and green colors with heating methods on agates of
south-west Libya. J Archaeol Sci 52:218–241
Cheshme Shoor area, Qom, Iran. J Miner Mater Charact Eng 04:
Hatipoğlu M, Tuncer Y, Kibar R, Çetin A, Karali T, Can N (2010)
210–217
Thermal properties of gem-quality moganite-rich blue chalcedony.
Phys B Condens Matter 405:4627–4633
Insoll T, Polya DA, Bhan K, Irving D, Jarvis K (2004) Towards an Publisher’s note Springer Nature remains neutral with regard to jurisdic-
understanding of the carnelian bead trade from Western India to tional claims in published maps and institutional affiliations.