Journal of Archaeological Science 50 (2014) 219e226

Contents lists available at ScienceDirect

Journal of Archaeological Science

journal homepage: http://www.elsevier.com/locate/jas

Review

Goldwork in Ancient Egypt: workshop practices at Qurneh in the 2nd

Intermediate Period
Lore G. Troalen a, *, Jim Tate a, Maria Filomena Guerra b
a
National Museums Scotland, Collections Services Department, National Museums Collection Centre, 242 West Granton Road, Edinburgh, EH5 1JA, UK
b
ArchAm e UMR 8096 CNRS e Universit
e Paris 1 Panth

eon-Sorbonne, Maison Arch

eologie & Ethnologie, 21 All

ee de l'Universit

e, 92023,
Nanterre Cedex, France

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

Article history: Described by Petrie as ‘the largest group of goldwork that had left Egypt’, the jewellery from the intact
Received 24 March 2014 burial of an adult and child discovered at Qurneh in 1908 is the most important group of gold objects
Received in revised form excavated in Egypt dating from the 2nd Intermediate Period (c. 1800e1550 BC). This unique collection
26 June 2014
has been studied using several non-invasive analytical techniques (mPIXE, PIGE, XRF, and SEM-EDS),
Accepted 10 July 2014
Available online 19 July 2014
while calculation of the effective penetration depth values allowed the degree of surface enrichment
to be assessed. The most recent results in respect of gold-working techniques are discussed and related
to published work on the techniques used in Egypt in the same era and the subsequent era. The data
Keywords:
Qurneh
showed, the coexistence, in a single grave, of jewellery with different levels of wear and colours of gold.
Gold alloys The extensive use of hard soldering by the addition of copper to the gold-based alloys was also revealed.
Solder All the objects presented PGE inclusions implying the use of alluvial gold and/or recycling of ancient
Polychromy alloys made with this type of gold.
Egypt Crown Copyright © 2014 Published by Elsevier Ltd. All rights reserved.
Recycling

1. Introduction been investigated in recent years (Eremin et al., 2000; Manley et al.,
2002; Tate et al., 2009; Troalen et al., 2009).
The most important group of gold jewellery attributed to the The detailed description of the Qurneh jewellery items can be
2nd Intermediate Period (c. 1800e1550 BC) forms part of the burial found in Petrie's excavation report (Petrie, 1909) and also in several
discovered intact by Sir William Flinders Petrie in 1908 at Qurneh,1 subsequent publications (Eremin et al., 2000; Roehrig, 2007; Tate
near Thebes (Petrie, 1909). The burial, which is dated to the 16th et al., 2009; Troalen et al., 2009). The adult individual wore a
Century BC on stylistic grounds as well as by radiocarbon dating necklace made of 1699 gold rings2 (4.5 mm external diameter)
(Eremin et al., 2000), was centred on a large anthropoid rishi-coffin strung together to form four decorative strands,3 two penannular
(Miniaci, 2011), painted dark blue and gilded and containing the gold earrings or hair-rings, four gold bangles, an electrum girdle
mummified remains of a young adult female surrounded by a wide consisting of 26 semi-circular so-called ‘wallet beads’ (10 mm
range of grave goods, including a large group of gold jewellery diameter) spaced by two threads of 6 barrel beads (4 mm length)4
items, see Fig. 1 (Petrie, 1909; Eremin et al., 2000; Tate et al., 2009). and an electrum button (the latter today missing). The child wore
Above the foot of the woman's coffin was a simple chest-shaped
coffin containing the remains of a young child also buried with
2
several jewellery items (Petrie, 1909; Eremin et al., 2000). The Petrie measured and weighed the rings of the 4 strands of the necklace and
richness of the burial was so exceptional that Petrie, when it was counted a total of 1653 rings (394; 416; 422; 421) (Petrie, 1909). When the necklace
was disassembled for conservation at NMS in 2006 the number counted was 1699
brought back to Britain, described it as ‘the largest group of goldwork (397, 410, 415, 477) (Tate et al., 2009).
that had left Egypt’ (Petrie, 1932). The burial is today part of the 3
The Qurneh necklace has often been described as the earliest example of she-
extensive National Museums' Scotland (NMS) Ancient Egyptian biu-type collar, although this is could be an erroneous description. Shebiu-type
collection and several aspects of the mummies and coffins have necklaces are made of large, thick lentoid-beads, tied around the neck (Roehrig,
2007 p. 19).
4
Petrie noted in his description of the girdle that: ‘… the spaces between these
(‘wallet beads’) had two threads of six beads each, and in one case a space of seven
* Corresponding author. beads' (Petrie, 1909). There is a total of 316 barrel beads, of which 5 have fallen
E-mail address: l.troalen@nms.ac.uk (L.G. Troalen). inside two ‘wallet beads’ and are now only visible by X-Radiography (Troalen et al.,
1
Other possible spellings include Qurna, or Gourna. 2009).

http://dx.doi.org/10.1016/j.jas.2014.07.010
0305-4403/Crown Copyright © 2014 Published by Elsevier Ltd. All rights reserved.
220 L.G. Troalen et al. / Journal of Archaeological Science 50 (2014) 219e226

2013) and the subsequent era (Gale and Stos-Gale, 1981;

Schorsch, 2001; Lilyquist, 2003). As part of the study the limitation
of the non-invasive analysis of ancient gold is discussed, following
calculation of the effective penetration depth values of the different
techniques; this confirms the feasibility of comparing the analysis
of objects from other collections by using different non-invasive X-
ray based techniques.

2. Analytical techniques

Several non-invasive techniques were used in order to gain in-

formation on the morphology of the objects and their elemental
composition with no surface cleaning or polishing. These tech-
niques allowed high spatial resolution with elemental mapping for
the study of the areas around the joints, and were undertaken in air
allowing the analysis of the larger and more complex-shaped ob-
©
Fig. 1. The Qurneh adult jewellery set, National Museums Scotland. jects. Measurements were also made using portable equipment to
analyse the original gilding on the rishi coffin.
The objects were examined visually under an Olympus SZX12
two ivory bangles, several strings of faience beads around the waist stereo-microscope equipped with an Olympus DP70 digital camera
and both ankles, a necklace made of 215 small gold rings (<1.5 mm and by X-radiography using a 320 kV Pantak system at NMS.
external diameter) strung together, and two asymmetric rings Elemental composition of the alloys was determined using: (1) an
made of three-and-a-half gold rings soldered together, which have Oxford Instruments ED 2000 air-path X-ray fluorescence spec-
been interpreted as earrings from their location in the burial. trometer (XRF), with Rh target X-ray tube collimated to a point of
From visual examination the workshop practices seem to be the about 2
1.5 mm, coupled to a Si(Li) detector; (2) mPIXE (Particle
same for all the objects, with the use of sheets or strips of gold Induced X-ray Emission) and PIGE (Particle Induced g-ray Emis-
which were then hammered, stamped-died or rolled, sometimes sion) with a proton beam of 3 MeV, with an analytical spot of
chased, with the different parts joined together. However the 50 mm, Si(Li) and Ge(Li) detectors at the AGLAE accelerator of the
jewellery items showed variable levels of wear, indicating different C2RMF; (3) a Niton XL3t portable XRF system with a ‘GOLDD’ de-
amount of usage. The gold items belonging to the young adult fe- tector, set to “Precious Metals” mode; (4) CamScan 2500 SEM with
male can be separated into three groups according to their level of a Noran Vantage Energy Dispersive X-ray analysis (EDS) system, at
wear: the necklace and the two penannular earrings are virtually the analytical working distance of 35.0 mm, 300 s measurement at
un-used; the four bangles are slightly used; while the girdle shows electron beam energies of 20 and 25 kV. A range of Au/Ag/Cu
very intensive wear. The jewellery items belonging to the child also standards were used for calibration in all experiments and the
show marks of wear; these are extensive for the necklace beads and inter-instrument compatibility can be found in a previous publi-
less so for the two earrings. cation (Troalen et al., 2009).
In the jewellery of the adult the wear-marks indicate that the
bangles were certainly worn, presumably during the life-time of 2.1. Surface analysis of ancient gold
the woman, in contrast to the girdle, which seems to be a much
older piece. Interestingly, on the girdle beads the wear-marks show The main issue with non-invasive surface analysis of ancient
that the present construction is the original one, as we observed gold is the phenomenon of surface enrichment, due to the deple-
some deformation of the holes in the wallet beads where they sit tion of copper and silver through either deliberate surface treat-
against the barrel beads (S.I. 1). In the case of the child, the earrings ment or from corrosion during burial (Scott, 1983; Rapson, 1996). It
do not show much signs of wear prior to deposit, but these can is accepted however that for a binary goldesilver alloy above
hardly have belonged to the young child,5 while the necklace is 64.6 wt% gold/35.4 wt% silver, corrosion is limited (Scott, 1983).
clearly a re-use of gold beads from different gold items. It is inter- Usually this corrosion layer is in the range of a few mm and some
esting to note that several of the beads exhibit a slight border, studies have shown that chemical-induced surface depletion is
suggesting that their form was adjusted slightly to match an usually less than 10 mm (Grimwade, 1999). Such surface enrichment
adjacent round stone or paste/glass bead, such as a spacer (S.I. 1). can particularly affect the composition determined by analytical
Petrie's report only records numerous faience beads around the methods based on X-ray fluorescence spectra, and in some cases
ankles and waist of the child rather than at the neck (Petrie, 1909), the difference between surface and sub-surface composition can
but the adjustments might relate to an earlier use of the beads. reach 8e10 wt% gold by SEM-EDS analysis (Mongiatti et al., 2010).
This paper presents a comprehensive study of all the gold Depth profiles of gold (Au), silver (Ag) and copper (Cu) depend
jewellery items and the coffin gilding from the Qurneh burial and on the energy and type of the ionising source, the composition of
discusses the diversity in the apparent colours of the gold items, the the alloys being analysed, and the spectral lines used for analysis. In
possible origin of the gold and the joining techniques. The results order to compare the measurements obtained by PIGE, mPIXE, XRF
obtained are discussed in respect of gold-working techniques used and EDS analysis, we calculated the density r (g cm3) and mass
in Egypt, and compared to the few published jewellery items that absorption coefficients m expressed in cm2 g1 for four Au/Ag/Cu
can be attributed to the 2nd Intermediate Period (Miniaci et al., alloys close in composition to those of the Qurneh items (Table 1).
The effective penetration depth values, corresponding to the
thickness in mm from which 95% of the detected X-rays are pro-
5
duced, were calculated for Au, Ag and Cu at these experimental
Closer examination revealed that the child's earrings do not match as a pair.
Both earrings were found to fasten together, as would be expected from a clasp for a
conditions (Table 1). The penetration for PIGE depends on the en-
necklace. It could be that this clasp was recycled as earrings, and deposited next to ergy of the incident protons (28 mm for 3 MeV), and since the
the ears of the child. emitted gamma-rays are hardly attenuated by the matrix the
 L.G. Troalen et al. / Journal of Archaeological Science 50 (2014) 219e226 221

Table 1
Calculated density r (g cm3) and mass absorption coefficient m expressed in cm2 g1 for Au/Ag/Cu alloys with different compositions (wt%). Density values were calculated
with CASINO 2.42 module and compared to theoretical values (Kraut and Stern, 2000). The given m values were calculated using the GUCSA module of GUPIX software
(Maxwell et al., 1989). The effective penetration depth values represent the thickness in mm from which 95% of the detected X-rays are produced for XRF, mPIXE and SEM-EDS
analysis. For PIGE analysis, the proton range in the different type of alloys was calculated in mm. Exciting radiation: XRF e Rh tube, 150 s: 35 kV, 1000 mA, 0.125 mm Rh filter and
then 300 s: 50 kV, 1000 mA, 0.5 mm Cu filter; mPIXE and PIGE e 3 MeV protons with 75 mm Cu filter for low energy Si(Li) detector; SEM-EDS - 20 kV accelerating voltage
(Trouslard and Tirira on Pyrrole software; Maxwell et al., 1989; Drouin et al., 2007).

Proton range Effective penetration depth

Mass absorption PIGE mPIXE XRF SEM-EDS

coefficient m PYRROLE (mm) GUYLS from GUPIX (mm) GUCSA from GUPIX (mm) CASINO 2.42 (mm)
(cm2 g1)

Au/Ag/Cu Density Au Ag Cu Au/Ag/Cu Au Ag Cu Au Ag Cu Au Ag Cu

(wt%) (g cm3) La Ka Ka La Ka Ka La Ka Ka La La Ka

95.8/4/0.2 18.63 127 58 204 28.00 7.40 11.09 5.07 12.68 27.76 7.89 0.50 0.50 0.50
86/12/2 17.17 129 54 202 28.50 7.86 12.11 5.50 13.54 32.36 8.65 0.45 0.45 0.45
68/30/2 15.14 130 46 203 29.10 8.77 13.98 6.15 15.24 43.08 9.76 0.60 0.60 0.50
50/48/2 13.54 130 37 205 29.60 9.64 15.82 6.77 17.04 59.88 10.81 0.60 0.60 0.60

Table 2
Comparing the compositional surface analysis of different types of alloys occurring in the Qurneh burial using XRF, mPIXE, PIGE and SEM-EDS (mean values calculated from 5 to
6 measurements). PIGE were calculated using the g-ray lines at 279, 309 and 152 keV for the measurement of gold, silver and copper, respectively, with normalisation of the
dose to the Standard 6917 from CLAL-France: 75Au-17Ag-8Cu (Guerra and Calligaro, 2004).

Composition in wt%

PIGE XRF mPIXE SEM-EDSa

Qurneh [NMS number] Au Ag Cu Au Ag Cu Au Ag Cu Au Ag Cu

279 309 152 La Ka Ka La Ka Ka La La Ka

Adult's penannular earring [A.1909.527.18] 94.0 5.6 0.4 95.4 4.3 0.3 95.6 4.0 0.4 97 3 <0.2
Adult's bangle [A.1911.527.16] 88.1 11.6 0.3 88.0 11.9 0.1 88.1 11.6 0.3 Not analysed
Adult's necklace, Ring 3 [A.1909.527.19] 86.1 12.3 1.6 86.3 12.1 1.6 88.0 10.0 2.0 92/86 7/12 1/2
Child's earring [A.1909.527.4] 83.5 14.8 1.7 81.1 15.8 3.1 84.6 13.6 2.9 82e89 8e14 3e4
Child's necklace, Ring bead A [A.1909.527.11] 69.9 27.8 2.3 71.2 26.4 2.4 71.3 26.8 2.0 75/72 25/26 1/3
Child's necklace, Ring bead B [A.1909.527.11] 67.8 29.0 3.2 68.9 29.1 2.0 65.9 32.1 2.0 72/71 27/27 1/2
Adult's girdle, wallet bead 19 [A.1909.527.17] Not analysed 43.9 52.6 3.5 Not analysed 53 44 3
a
SEM-EDS values correspond to the mean compositions obtained for Surface/Bulk; except for the child's earring, where the EDS values correspond to the range of
composition measured and the adult girdle and earrings, where only surface analysis could be undertaken.

quantification of the major elements is less sensitive to the possible showed that each group is remarkably uniform. The necklace is
heterogeneity of the surface (Guerra and Calligaro, 2004); mPIXE made of an alloy containing, on average, 86 wt% Au, 12 wt% Ag and
provides a narrower range of depth analysis (10e15 mm for silver), 2 wt% Cu while the earrings are richer in gold with a composition of
but with comparable depth values for all the elements; in com- 95.4 wt% Au, 4.3 wt% Ag and 0.3 wt% Cu. The four bangles are
parison XRF provides variable depths of analysis with the greatest extremely homogeneous and made with an alloy close to the
values for the silver Ka lines. Finally, EDS analysis reflects the necklace, but with almost no copper (11.6e12.5 wt% Ag and 0.1 wt%
composition of the first 0.5 mm and is thus very sensitive to the Cu). Their great homogeneity in weight was noted by Petrie in his
heterogeneity of the surface. report; with the four bangles and the necklace following the same
All the Qurneh items were investigated using a combination of unit system6 (Petrie, 1909). Finally, the child's earrings are rather
these techniques and the polished sections of one of the rings from heterogeneous with silver contents varying from 13.5 wt% to
the adult's necklace and two silver-rich beads from the child's 15.3 wt% and copper contents from 1.7 wt% to 3.8 wt%. This variety
necklace allowed the composition of the surface and the core to be of the alloys could be explained by the low quality soldering of the
directly compared. As expected, differences were found (Table 2), rings giving rise to large melted regions, although it must also be
but the compositions determined from the polished core sections noted that visual inspection shows that these pieces have a dirty
of the metal were equivalent to those obtained by PIGE and XRF surface.
analysis. The compositions also agreed well with those obtained by The rest of the objects are made of electrum alloys with a silver
mPIXE (see Table 2). Thus while there may be some surface change content higher than 20 wt%. The adult's girdle is made of two types
from burial or from any deliberate surface treatment, this only af- of whitish coloured beads, the wallet beads being made of a single
fects a very thin layer; and in the absence of any surface preparation alloy with an average of 52.5 wt% Ag and 3.7 wt% Cu, while the
this only distorts measurements made using the shallow SEM-EDS spacer barrel beads are heterogeneous in composition, ranging
analysis. from 51 to 55 wt% Ag and 2.8e7.4 wt% Cu. The child's necklace is of

2.2. The analysis of the gold alloys

6
Petrie suggested in his report that several of the adult jewellery items were
2.2.1. Gold alloys from Qurneh following the 80-grain unit system: each bangle corresponded to 4 of the 80-grain
unit; the necklace was 20 of the 80-grain unit while each of the attachment
Fig. 2 presents the elemental composition of all the jewellery weighted half a unit on each side (Petrie, 1909). The adult earrings weighted 133.8
items from the Qurneh burial. The composition of the rings that and 138.8 grains (Petrie, 1909), corresponding together to almost 3 and a half 80-
constitute the adult's necklace and the tubes of the adult's earrings grain unit.
222 L.G. Troalen et al. / Journal of Archaeological Science 50 (2014) 219e226

poorer quality than the other pieces and the bead rings were scarab belonging to King Sobekemsaf, the finger ring bearing the
observed to be visually heterogeneous in shape and in colour with prenomen of King Nubkheperre Intef and the two spacer-bars from
silvery, reddish and yellowish alloys. The copper content of these a bracelet belonging to his wife Queen Sobekemsaf analysed at the
rings varies between 0.6 and 2.6 wt% while their silver content British Museum (Miniaci et al., 2013). The Qurneh girdle falls into
varies between 16.6 wt% and 32.1 wt%. Finally, the analysis of the the region described as whitish, close to the pale greenish-yellow
gold foil on the surface of the coffin showed an average composi- colour of the large group of aurian silver objects characterised by
tion of 17 wt% Ag and 1.5 wt% Cu on the ancient parts. This Gale and Stos-Gale (1981). Finally, only the rings from the child's
composition with a high level of silver is not unexpected and cor- necklace have similar composition to the majority of the objects
responds to the range of alloys characterised by Hatchfield and analysed at the Metropolitan Museum (Lilyquist, 2003).
Newman in their investigation of gold foils on Egyptian wooden It seems that at the end of the 2nd Intermediate Period a range
artefacts from the Middle to New Kingdom periods (Hatchfield and of coloured gold alloys close to the polychromy expected in the 18th
Newman, 1991). Dynasty was already being used (Schorsch, 2001), although for the
Qurneh burial, each object exhibits a single colour without the use
of surface treatment. It is reasonable to deduce from Fig. 2 that
2.2.2. Gold polychromy in the 2nd Intermediate Period different gold deposits with different Ag/Au ratios were exploited
The study published by Schorsch on 18th Dynasty artefacts from and that addition of different quantities of copper to the gold alloys
the time of Tutankhamun identified a range of techniques used to was current practice. The copper content is consistent with what
achieve polychromy: the use of different alloys containing variable would be expected for alluvial gold, below 2 wt% (Ogden, 2000;
amounts of silver and copper, and the application of surface Klemm and Klemm, 2013), with the exception of the girdle,
treatments to obtain a red-coloured aspect (Schorsch, 2001). The where the wallet beads contain up to 3.7 wt% copper, and the barrel
variety of alloys found in the jewellery group from Qurneh, ranging beads have a copper-content ranging from 2.8 to 7.4 wt%. Similarly
from very pure gold to silver-rich electrum, suggests the use of high amounts of copper were found in aurian silver objects from
polychromy in the 2nd Intermediate Period. Figs. 2 and 3 compare the Ashmolean Museum, ranging from 0.2 to 26 wt% (Gale and Stos-
the compositions obtained for the Qurneh items with compositions Gale, 1981). The addition of copper to debase gold is known to be
published for four jewellery items from the British Museum common practice from the 18th Dynasty onwards (Lucas and
attributed to the 2nd Intermediate Period (Miniaci et al., 2013) and Harris, 1962), but the amount of copper in the girdle indicates
the large set of jewellery items from the 18th Dynasty burial of the that this practice was also in use in the 2nd Intermediate Period,
foreign wives of Tuthmosis III analysed at the Metropolitan and according to the results published by Gale and Stos-Gale for
Museum (Lilyquist, 2003). In addition to these, the nine 18th Dy- aurian silver even in the Old and Middle Kingdom periods (Gale and
nasty aurian silver objects from the Ashmolean Museum and one Stos-Gale, 1981). Finally, the high gold content of several items from
from the 2nd Intermediate Period analysed by Gale and Stos-Gale the Qurneh burial demonstrates that the quality of gold alloys alone
(1981) are also considered. cannot be used as a criterion of authenticity.
The Qurneh items exhibit a range of colour that goes from yel-
low to whitish and it is notable that, with the exception of the
electrum objects, they are richer in gold content than the 18th 2.2.3. PGE inclusions and alluvial gold
Dynasty objects investigated at the Metropolitan Museum Platinum Group Element (PGE) inclusions were observed in all
(Lilyquist, 2003). Nevertheless, similar high gold content the jewellery from Qurneh, including the adult earrings made of
(80e89 wt% from XRF analysis) was characterised in the heart- high purity gold. The presence of these inclusions in Egyptian gold

Fig. 2. Ternary copper, gold, silver diagram in wt% of the Qurneh jewellery, with the colour of gold alloy adapted from McDonald and Sistare, 1978 (McDonald and Sistare, 1978).
 L.G. Troalen et al. / Journal of Archaeological Science 50 (2014) 219e226 223

Fig. 3. Ternary copper, gold, silver diagram in wt% of published Egyptian gold and electrum jewelleries dated from the 2nd Intermediate Period, 17th and 18th Dynasties (Gale and
Stos-Gale, 1981; Hatchfield and Newman, 1991; Lilyquist, 2003; Miniaci et al. 2013) with the colour of gold alloy adapted from McDonald and Sistare, 1978 (McDonald and Sistare,
1978).

jewellery was first reported by Petrie in 1895 in his report on and Tite, 1980), and furthermore two inclusions were analysed by
Naqada and Ballas (Petrie and Quibell, 1895). Their presence is Miniaci et al. in King Nubkheperre's finger ring and King Sobe-
characteristic of the use of alluvial gold deposits, and the classifi- kemsaf's heart scarab (Miniaci et al., 2013). In their study, Meeks
cation of PGE elements depends on their composition as defined by and Tite stressed the heterogeneity of these inclusions, sometimes
Harris and Cabri (Ogden, 1976; Harris and Cabri, 1991). Meeks and within the same object, but suggested the use of ruthenium con-
Tite published the composition of forty-nine inclusions in twelve centration to separate the gold sources in Egypt before and after the
Egyptian jewellery items dating from 3000 BC to 300 AD (Meeks 12th Dynasty (25e40 wt% to less than 25 wt%), reflecting the

Fig. 4. Ternary ruthenium, osmium, iridium diagram showing the average composition of the PGE inclusions characterised by SEM-EDS analysis in the Qurneh jewellery items,
expressed in wt% [Osmiridium: Os < 38%, Ru < 5%; Iridosmine: Os > 55%, Ru < 5%; Rutheniridosmine: Ru > 5%, Ru þ Os > 50%].(Ogden, 1976) These results should be regarded as
semi-quantitative due to the geometry of the inclusions and the lack of comparative analytical standards. The red lines indicate the ruthenium concentration defined by Meeks and
Tite (Meeks and Tite, 1980): 25e40 wt%.
224 L.G. Troalen et al. / Journal of Archaeological Science 50 (2014) 219e226

change from the exploitation of gold sources in the Eastern Desert 2013). Nevertheless, the presence of these inclusions in all objects
to those in the Nile Region north of the eighteenth parallel (Ogden, from the 2nd Intermediate Period analysed in this study and by
1976; Meeks and Tite, 1980). Miniaci et al. proves that either alluvial gold deposits were
Thirteen PGE inclusions were analysed in the adult's and child's exploited during this period or that re-melting of previously used
earrings and necklaces by EDS analysis. From these results, all the alluvial gold was a current practice. Finally, the PGE inclusions
inclusions were found to be Rutheniridosmine with an average analysed in the adult's earrings confirm the exploitation of gold-
composition of 22 wt% ruthenium, 35 wt% iridium and 43 wt% rich alluvial deposits. In principle a similarly high purity gold
osmium (Fig. 4). Most of the inclusions contain less than 25 wt% alloy could have been obtained using a cupellation refining process,
ruthenium, although several inclusions from the adult's necklace but there is no recorded evidence of parting (Ramage and Craddock,
exhibit up to 55 wt% ruthenium and in another case 69 wt% 2000) at that time in Egypt.
osmium. Even accepting that these compositions are quite variable,
we note that the level of ruthenium in most of the PGE inclusions is 3. Joining technique
higher than that found in the objects attributed to the 12th Dynasty
onwards investigated by Meeks and Tite. These results highlight the Several techniques can be used to join metal alloys: welding,
need for analysis of a larger number of PGE inclusions in Ancient soft-soldering, hard-soldering or brazing. Maryon defines the sol-
Egyptian artefacts, but would suggest the use of alluvial gold de- dering process as the use of ‘any metal or alloy whose melting point is
posits close to the Eastern Desert or Nubia. Furthermore the Qurneh lower than that of the metal or alloy to be soldered, which may be run
burial shows some links with Nubia (Eremin et al., 2000; Manley between the parts to be joined to fasten them together’ (Maryon,
et al., 2002) and gold mines from Nubia exhibit a lower level of 1949). The predominant use of soldering technique instead of
silver than the mines from the Eastern Desert (Klemm and Klemm, mechanical joints in gold jewellery was reported by Ogden in his

Fig. 5. (a) SEM-BSC micrograph of a single ring from a wallet bead from the Qurneh adult girdle, scale bar is 1 mm, (b) EDS line-scan illustrating the slight change in composition at
the hard-soldered joint.
 L.G. Troalen et al. / Journal of Archaeological Science 50 (2014) 219e226 225

review of Ancient Egyptian metalworking techniques, although it

was also suggested that colloidal hard-soldering or diffusion
bonding might have been used for more delicate soldering work,
but without specifying as from which period onwards (Ogden,
2000). The use of gold-based brazing alloys in Ancient Egypt is
also supported through archaeological evidence and it is reported
that some form of bellows were used in Egypt by about 2500 BC
(Wolters, 1975). However, little scientific analysis is available for
Ancient Egyptian gold jewellery and the use of hard-solder made by
the addition of copper and/or silver to a base gold alloy was so far
only considered in a few artefacts, including a granulated gold
cylinder-pendant from el-Hara geh (Ogden, 1992) and demon-
strated for some Middle Kingdom gold beads from the burial of
Wah (Schorsch, 1995). However, when considering the analysis
published on the jewellery items from the 18th Dynasty burial of
the foreign wives of Tuthmosis III, it is observed that an increase of
copper was also found in several of the items in the regions where
soldering was expected (Lilyquist, 2003 e Appendix 2).
With the exception of the child's necklace and the barrel beads Fig. 7. Comparison of the average composition of the Qurneh jewellery items with
from the adult's girdle, all the pieces from the Qurneh jewellery set their solder composition (values exhibiting the highest level of copper), by expressing
were hard-soldered. The soldering seams are more or less visible in the content of copper in wt% as a function of the ratios of silver/gold both in wt%.

each element either because of its level of wear, or because of the

craft worker's skill (or both), the latter being of such variability that
it suggests production by different goldsmiths. The most visible showed no particular signs of wear, only the clasp clearly shows
joints were found in the earrings from both bodies. In the adult's signs of hard-soldering. However, a clear joint made with an alloy
earrings, the joints are thick but perfectly crafted (S.I. 2), while for of higher copper content could be seen when one of the small rings,
the child's earrings the joints indicate a lack of temperature control removed during conservation treatment, was polished and etched
as the parts that were joined show signs of melting (S.I. 3). Petrie with an aqua regia mixture (Fig. 6, Tate et al., 2009). The compo-
noted in his report that the child's earrings ‘had been over-heated sition of the different solders was estimated by analysing the cross-
while on a mandril in the furnace for soldering; the solder had stuck section with SEM-EDS or in other pieces by mapping the region
them together, and they parted and began to drop away, being half around the joints using mPIXE or XRF systems. The solders stand out
melted' (Petrie, 1909). The joints in the bangles are very thin and due to their copper contents, which corresponds to an addition of
[near] invisible to the naked eye, but could be found by X-radiog- copper in order to lower the melting point and allow the parts to be
raphy with a 400 kV tube, showing that they had been formed from joined (Maryon, 1949). For all the Qurneh objects, the solders and
a solid tube, the joints cut at an angle and soldered (S.I. 4). the parts to be soldered contain very similar Ag/Au ratios, which
The least visible joints occur in the adult's necklace and girdle suggests that a similar soldering technique was used for all the
but for different reasons. The very high level of abrasion of the objects, with the use of the basic gold/electrum alloys mixed with
girdle due to its use has removed almost all of the joint seams so additional copper (Fig. 7).
that hard-soldering could only be detected by mapping Cu(Ka),
Ag(La) and Au(La) with SEM-EDS in the assumed joint region of one 4. Conclusions
of the wallet beads (Fig. 5a, b). For the adult's necklace, which
The study of the gold jewellery from Qurneh has provided an
initial overview of the workshop practices at the end of the 2nd
Intermediate Period in Egypt, revealing technologically high gold-
smithing skills across a large variety of gold items for two
mummified individuals.
New and old items were deposited together in the burial of the
young woman, while for the child, only objects made by recycling
without melting were used. The burial practices seem different for
both individuals, who were certainly of high-ranking, but for both
the presence of gold objects seems to be of fundamental impor-
tance. These observations support the hypothesis that some
jewellery items correspond to a re-use, in particular the adult's
electrum girdle and the child's necklace, neither of which follow
the unit system observed in the other adult jewellery items. This
suggests the continual use of ‘family’ jewellery such as the girdle,
handed down perhaps as heirlooms. It could be that such practices
resulted, at least in some regions, from the absence of the practice
of goldsmithing or else from a lack of raw materials leading to
recycling of older pieces; alternatively (or additionally), the older
objects may well have been treasured as manifestations of material
links with the ancestors.
Fig. 6. SEM-BSC micrograph of a single ring from the Qurneh adult necklace, polished
The composition of the alloys corresponds to what is expected
and etched with aqua regia (nitric acid and hydrochloric acid, 1:3 v/v), exhibiting the for secondary gold coming from different deposits and containing
presence of a hard-solder, scale bar is 200 mm (from Tate et al., 2009). naturally different silver contents: the goldsmiths combined a large
226 L.G. Troalen et al. / Journal of Archaeological Science 50 (2014) 219e226

variety of polychrome gold alloys, ranging from high-carat gold to Eremin, K., Goring, E., Manley, W.P., Cartwright, C., 2000. A 17th dynasty Egyptian
queen in Edinburgh? KMT 11, 32e40.
silver-rich electrum with the addition of different quantities of
Gale, N.H., Stos-Gale, Z.A., 1981. Ancient Egyptian silver. J. Egypt. Archaeol. 67,
copper. However, the composition of the adult's earrings (items 103e115. http://dx.doi.org/10.2307/3856605.
which show very little evidence of wear) is particularly notable, and Guerra, M.F., Calligaro, T., 2004. Gold traces to trace gold. J. Archaeol. Sci. 31,
could correspond to a particular deposit, bearing in mind that the 1199e1208.
Grimwade, M., 1999. The surface enrichment of carat gold alloys e depletion
burial shows definite links with Nubia. All the jewellery items were gilding. Gold Technol. 26, 16e23.
found to contain several PGE inclusions, indicating the systematic Harris, D.C., Cabri, L.J., 1991. Nomenclature of platinum group element alloys: re-
use of alluvial gold instead of a more intensive use of primary gold. view and revision. Can. Mineral. 29, 231e237.
Hatchfield, P., Newman, R., 1991. Ancient Egyptian gilding methods. In: Bigelow, D.,
This suggests that alluvial deposits were exploited during the 2nd Cornu, E., Landrey, G., van Home, C. (Eds.), Gilded Wood Conservation and
Intermediate Period or else that re-melting and re-use of earlier History. Sound View Press, Madison CT, pp. 291e299.
alluvial gold was a current practice. Klemm, R., Klemm, D., 2013. Gold and gold mining in Ancient Egypt and Nubia. In:
Herrmann, B., Wagner, G.A. (Eds.), Geoarchaeology of the Ancient Gold Mining
The study of the joints revealed the systematic use of hard- Sites in the Egyptian and Sudanese Eastern Deserts, Series in Natural Science in
soldering techniques with the use of basic gold/electrum alloys Archaeology. Springer-Verlag, Berlin-Heidelberg.
mixed with additional copper to lower the melting point. This Kraut, J.C., Stern, W.B., 2000. The density of gold-silver-copper alloys and its
calculation from the chemical composition. Gold Bull. 33, 52e55.
confirms a certain unity for the workshop practices in Egypt. All the Lilyquist, C., 2003. The Tomb of the Three Foreign Wives of Tuthmosis III. The
items are soldered except the female's girdle spacers and the beads Metropolitan Museum of Art, New York.
that constitute the child's necklace, which by their form could at Lucas, A., Harris, J.R., 1962. Ancient Egyptian Materials and Industries. Edward
Arnold Publishers Ltd.
least partially be considered as spacers deriving from other items.
Manley, W.P., Eremin, K., Shortland, A., Wilkinson, C., 2002. The facial reconstruc-
This joining technique was nevertheless practiced at very different tion of an Ancient Egyptian Queen. J. Audiov. Media Med. 25, 155e159. http://
skill levels, from invisible small joints for each ring in the adult's dx.doi.org/10.1080/0140511021000051144.
necklace, to an excess of solder with large melted areas in the Maryon, H., 1949. Metal working in the Ancient World. Am. J. Archaeol. 53, 93e125.
Maxwell, J.A., Campbell, J.L., Teesdale, W.J., 1989. The Guelph PIXE software. Nucl.
fastening of the adult's necklace and the child's earrings, suggesting Instrum. Methods.Phys. Res. B 43, 218e230.
the work of a less-skilled craftsman but not a divide between the McDonald, A.S., Sistare, G.H., 1978. Part I - coloured gold alloys. Gold Bull. 11,
two burials. Furthermore, the work of different goldsmiths is 128e131.
Meeks, N.D., Tite, M.S., 1980. The analysis of platinum-group element inclusions in
indicated in the same item, specifically where two types of chisels gold antiquities. J. Archaeol. Sci. 7, 267e275.
and characteristic differences in craftsmanship were visible in the Miniaci, G., 2011. Rishi Coffins and the Funerary Culture of Second Intermediate
wallet beads of the girdle (S.I. 5). Period Egypt. In: Egyptology, vol. 17. Golden House Publications, London.
Miniaci, G., La Niece, S., Guerra, M.F., Hacke, M., 2013. Analytical Study of the First
Finally, with regard to the composition of the alloys, it was found Royal Egyptian Heart-scarab, Attributed to a Seventeenth Dynasty King, Sobe-
that while there may be some surface changes, these are only a very kemsaf. In: British Museum Technical Research Bulletin 7. Archetype, London,
thin layer and hence are observed only by the shallow SEM-EDS ISBN 9781909492073, pp. 53e60.
Mongiatti, A., Meeks, N., Simpson, St J., 2010. A Gold Four-horse Model Chariot from
analysis without surface preparation. The compositions deter- the Oxus Treasure: a Fine Illustration of Achaemenid Goldwork. In: British
mined from the polished core sections of the metal were equivalent Museum Technical Research Bulletin 4. Archetype, London, ISBN
to those obtained by both mPIXE/PIGE and XRF analysis. 9781904982555, pp. 27e38.
Ogden, J., 1976. The so-called “platinum” inclusions in Egyptian goldwork. J. Egypt.
Archaeol. 62, 138e144. http://dx.doi.org/10.2307/3856354.
Acknowledgments Ogden, J., 1992. Interpreting the Past: Ancient Jewellery. University of California
Press/British Museum, Berkeley and Los Angeles, pp. 51e52.
The authors are grateful to Dr. William P. Manley, formerly Se- Ogden, J., 2000. Metals. In: Nicholson, P., Shaw, I. (Eds.), Ancient Egyptian Materials
and Technology. Cambridge University Press, Cambridge, pp. 148e175.
nior Curator at NMS, and to Ms Margaret Maitland Curator of Petrie, W.M.F., Quibell, J.E., 1895. Naqada and Ballas. B. Quaritch, London.
Ancient Mediterranean collections at NMS for supporting this Petrie, W.M.F., 1909. Qurneh. British School of Archaeology in Egypt and EgyptIan
research, as well as Ms Lesley-Ann Liddiard, Assistant Curator for Research Account. Fifteenth Year. B. Quaritch, London.
Petrie, W.M.F., 1932. Seventy Years in Archaeology. H. Holt & Co., London.
providing access to the objects. At C2RMF the authors gratefully Ramage, A., Craddock, P., 2000. King Croesus's Gold. Excavations at Sardis and the
acknowledge Mr Thierry Borel, Mr Dominique Bagault, Mr Emma- History of Gold Refining. In: Archaeological Exploration of Sardis, Monograph
nuel Ple
, Dr. Stefan Ro€ hrs, Mr. Laurent Pichon and Dr. Thomas 11. British Museum Press with Harvard University Press, London, Cambridge
MA.
Calligaro and at Historic Scotland, Dr. Maureen Young, for addi- Rapson, W.S., 1996. Tarnish resistance, corrosion and stress: corrosion cracking of
tional p-XRF measurements. Finally, the authors express their gold alloys. Gold Bull. 29, 61e69.
gratitude to Prof. Stephen Quirke, Professor of Egyptian Archae- Roehrig, C.H. (Ed.), 2007. Hatshepsut: from Queen to Pharaoh. The Metropolitan
Museum of Art - Yale University Press, New Haven and London. http://www.
ology at University College London, for comments, suggestions and metmuseum.org/research/metpublications/Hatshepsut_From_Queen_to_
encouragements during the progress of this paper. Financial sup- Pharaoh.
port was granted through 6th F.P. Eu-ARTECH (FP 6, European Schorsch, D., 1995. The gold and silver necklaces of Wah: a technical study of an
unusual metallurgical joining method. In: Brown, C., Macalister, F., Wright, M.
Union, contract number RII3-CT-2004-506171) and CNRS funded
(Eds.), Conservation in Ancient Egyptian Collections. Archetype Publications,
project PICS 5995 EBAJ-Au. London, pp. 127e135.
Schorsch, D., 2001. Precious-metal polychromy in Egypt in the time of Tutank-
Appendix A. Supplementary data hamun. J. Egypt. Archaeol. 87, 55e71. http://dx.doi.org/10.2307/3822371.
Scott, D.A., 1983. The deterioration of gold alloys and some aspects of their con-
servation. Stud. Conserv. 28, 194e203.
Supplementary data related to this article can be found at http:// Tate, J., Eremin, K., Troalen, L.G., Guerra, M.F., Goring, E., Manley, W.P., 2009. The
17th dynasty gold necklace from Qurneh, Egypt. Arche
oSciences 33, 121e128.
dx.doi.org/10.1016/j.jas.2014.07.010.
ISSN 9782753511811.
Troalen, L.G., Guerra, M.F., Tate, J., Manley, W.P., 2009. Technological study of gold
References jewellery pieces dating from the Middle Kingdom to the New Kingdom in
Egypt. Arche
oSciences 33, 111e119. ISSN 9782753511811.
Drouin, D., Real Couture, A., Joly, D., Tastet, X., Aimez, V., Gauvin, R., 2007. CASINO Trouslard, P., Tirira, J.: Pyrrole Software V 3.93, CEA (INSTN).
V2.42-A fast and easy-to-use modelling tool for scanning electron microscopy Wolters, J., 1975. Zur Geschichte der Lo € ttechnik (The origin of gold brazing). In:
and microanalysis users. Scanning 29, 92e101. Degussa Gesch€ aftsbereich Verbindungstechnik Metall (Germany).