Content uploaded by Zofia Anna Stos-Gale
Author content
All content in this area was uploaded by Zofia Anna Stos-Gale on Sep 07, 2021
Content may be subject to copyright.
Non-Stable Isotopes in
Archaeology
ZOFIA ANNA STOS-GALE
University of Oxford, UK; University of Gothenburg,
Sweden
Introduction to lead isotopes
Lead occurs in major amounts not only in
lead–zinc–silver ore deposits, but also in lesser
amounts in copper and iron ore deposits; it occurs
in at least parts per million levels in all rocks and
non-metallic minerals. Terrestrial lead consists
of four stable isotopes: 208 Pb, 207Pb, 206Pb, and
204Pb. e isotope 204Pb is non-radiogenic in
origin; it was incorporated into the Earth when
it was formed from the proto-solar system about
4.57 thousand million (109or billion) years
ago. e other three lead isotopes (LIs) derive
from the radioactive decay of longlived naturally
radioactive isotopes of uranium (U) and thorium
():
238U→206 Pb +energy; half-life =4,468Ma
(million years)
235U→207 Pb +energy; half-life =704Ma
232→208 Pb +energy; half-life =14,010Ma
At rst U, , and Pb were uniformly mixed
together in the hot, liquid Earth. When the Earth
startedtocool,therelativeconcentrationsofthese
elements dierentiated from place to place. Dur-
ing those years, the U and were continuously
producing more of 208Pb, 207 Pb, and 206Pb, but in
somewhat dierent amounts. A more important
stage of dierentiation of the amounts of the
isotopes came with the process of ore formation
starting about 3 billion years ago when, due to the
high temperatures and dierences in chemistry,
U and were separated from lead. erefore,
in most minerals the U/Pb and /Pb ratios are
so low that their lead isotopic compositions have
not changed appreciably with time since it was
e Encyclopedia of Archaeological Sciences. Edited by Sandra L. López Varela.
© 2018 John Wiley & Sons, Inc. Published 2018 by John Wiley & Sons, Inc.
DOI: 10.1002/9781119188230.saseas0413
laid down in a mineral deposit. ese “frozen” in
time lead isotope compositions can be used for
dating of the ore- and rock-forming processes.
If there are no additional events inuenc-
ing the lead isotope compositions in the ore
mineral or rock, then this system is referred
to as single-stage radiogenic growth of lead in
a closed chemical system. It can be character-
ized by growth curves with associated values
of 238U/204 Pb, as measured today for the parent
U–Pb system, and 232/204 Pb ratios for the
parent –Pb system. is model is called the
Holmes-Houtermans model; its more developed
versions were proposed by Stacey and Kramers,
and Cummings and Richards (see Faure 1986).
Using these mathematical models, it is possible
to calculate, from the measured LI compositions
of ore minerals, so-called “model ages” of the ore
deposits. e growth curve of the LI ratios related
to the age of the ore deposits is clearly represented
on the plots of ratios of 206 Pb/204Pb (xaxes) and
207Pb/204 Pb (yaxes) (Figure 1a).
eisotopecompositionsoftheleadintheores
depend on the age of their formation, the original
amounts of uranium and thorium, and the fur-
ther geological history of the deposit (mixing with
ore-forming uids, redeposition, later metamor-
phism). erefore, in principle, minerals of dier-
ent geographical origin, formed at dierent geo-
logical periods, have a high probability of display-
ing dierent LI compositions.
The measurements of lead isotope
compositions
Since the 1970s a standard method used for
LI analyses has been thermal ionization mass
spectrometry (TIMS). e LI ratios are mea-
sured using exactly specied chemistry of lead
separation and deposition, and methodology for
runningthemassspectrometer;therawdata
are corrected by comparisons with lead isotopic
standard US NBS (NIST) SRM 981. Strict adher-
ence to this methodology allows us to measure LI
compositions with very great accuracy (usually
10
10
12
12
14
14
16
16 18
4.57
4.0
4.0
3.7
3.5 3.0
3.0
2.0
A2.5 B
2.0 1.5
0.0 b.y.
1.0
1.0 0.
CDEFGHJKL
μ = 7
μ = 9
0
.0 b.y.
Approximate area of Fig. 1b
207Pb/ 204Pb
206Pb/ 204Pb
A: Geneva Lake
B: Cobalt
C: Sudbury Errington Mine
D: SW-Finland
E: Broken Hill
F: Mount Isa
G: Sullivan
H: Balmat
I: Captains Flat
J: Cobar C.S.A. Mine
K: Bathurst
L: Halls Peak
(a)
(b) (c)
18.4
15.60
15.65
15.70
15.75
207Pb/204Pb
38.0
38.2
38.4
38.6
38.8
39.0
208Pb/204Pb
18.6 18.8 19.0
206Pb/204Pb
18.0 18.2 18.4 18.6 18.8 19.0
206Pb/204Pb
Calceranica. V. Imperina,
Vetriolo, etc.
Montefondoli, Parnera,
Maso Erdemolo, etc.
LBA copper slags Trentino-Bolzano
Cu ores S-W Spain
Cypriot ores
Lavrion
Over 100 oxhide ingots
Figure 1 (a) Graphic representation of the Holmes-Houtermans model ages of ore deposits (A, B, C, etc.) calculated from their lead isotope ratios. is
model demonstrates that not only ores of dierent ages, but also ores of the same age can have dierent lead isotope compositions due to the dierent
initial amounts of uranium and thorium in the ore-forming uids (μvalues) represented by the outer curves. e straight lines indicate ores of the same
age. (b) and (c) Plots of two sets of independent lead isotope ratios representing data for copper ore deposits in the Mediterranean, which were exploited in
the Bronze Age, and the second millennium copper ingots of the “oxhide” shape that were found in Cyprus, Greece, Italy, Egypt, Turkey, and Bulgaria.
All these ingots have lead isotope ratios and trace elemental compositions fully consistent with the Cypriot copper ores.
NON- STABLE ISOTOPES I N ARCHAEOLOGY 3
better than 0.1% for annual runs of the SRM
981 standard) and very good reproducibility of
measurements taken in dierent laboratories.
Further developments in instruments measur-
ing LI compositions came with the introduction
of instruments that use laser ablation of the
sample and ionization of lead in a plasma source
with multiple Faraday cups as ion collectors:
multicollector inductively coupled plasma mass
spectrometers (MC-ICP-MS). ese allow anal-
yses of much smaller samples with very high
overall accuracy (commonly 0.01%).
e TIMS and MC-ICP-MS are the only meth-
ods of measuring LI compositions that are suit-
able for archaeological provenance research. e
other instruments that can measure the LI ratios
(e.g., quadrupole mass spectrometer (Q-ICP MS)
or time-of-ight (ToF) mass spectrometers) have
much poorer accuracy due to substantially larger
and less reproducible mass fractionation.
Application of lead isotope compositions
in archaeology
In 1967 Robert Brill and John Wampler published
comparisons of the LI compositions of ancient
lead artifacts with those of ore deposits. Quite
soon, LI provenance studies broadened to include
analyses of glass, glazed pottery, silver coins, and
artist’s pigments. In 1982 the rst paper was pub-
lishedsuggestinguseofLIsforndingsourcesof
Bronze Age copper artifacts (Gale and Stos-Gale
1982), and that considerably raised the interest
of archaeologists in this method for provenanc-
ing Bronze Age metals. In the following years
archaeometallurgical projects conducted by sev-
eral research groups developed the methodology
by analyzing ores and metals to test the possibility
of changes of lead isotope compositions during
the smelting and rening processes, or during
the addition of tin to make bronze. e results
of these experiments proved that metallurgical
processes do not change the original LI compo-
sitionsofmetalsandpigments.Sincetheearly
1980s several large research projects into sources
of copper, lead, and silver in the Bronze Age
were undertaken. ese projects were based on
simultaneous collection of mineral samples from
ore deposits, and samples of ancient artifacts
from archaeological excavations and museums,
combined with archaeometallurgical surveys
of ancient mines and slag heaps, and provided
LI, geochemical and chronological information
about the earliest copper and lead/silver ore
extraction in the Old World (e.g., Stos-Gale and
Gale 2009; Artioli et al. 2014; Pernicka, Lutz, and
Stöllner 2016).
Since there is a necessity to take small samples
from artifacts for lead isotope analyses (1–20 mg),
the scope of the research can be limited by the
reluctance of museum curators and archaeol-
ogists to allow sampling of precious artifacts.
For MC-ICP-MS it might be possible to use
laserablationfromthesurfaceoftheartifactto
minimize the damage, but there is no possibility
of the use of a mass spectrometer in a museum
or on archaeological site, since no machine of the
required accuracy has yet been developed in a
suciently portable format.
Methodology of interpretation of lead
isotope data
eidenticationoforiginofarchaeological
materials requires comparisons of their lead
isotope ratios with the ores, and therefore an
extensive lead isotope database for ore deposits is
the most important requirement. Calculating the
geological age (model age) from the lead isotope
ratios of the ancient artifact and comparing with
the known ages of ore deposits is not sucient,
due to the variations in lead isotope compositions
of ores not related to their age (e.g., the U and
contents and other geological processes). So far
there is only one open LI database on the internet;
OXALID makes available data for Mediterranean
oredepositsandBronzeAgemetals(University
of Oxford 2017). e other datasets are still only
available in published papers. Currently there
are about 8,000 published lead isotope results for
ores relevant to Old World archaeology.
In principle, the ancient artifact can be
regarded as fully consistent with an origin
from a given ore deposit if the three independent
lead isotope ratios of the artifact are each iden-
tical to the lead isotope ratios of the ores from
this deposit. So, the initial stage of identifying
the possible sources of archaeological artifacts
4NON- STABLE ISOTOPES I N ARCHAEOLOGY
relies on nding, for each artifact, the identical
data for ores. is can be done by calculating
the Euclidean distances between the three inde-
pendent lead isotope ratios of the artifact and all
available ratios for ore samples. However, this is
only a preliminary step to select a group of ore
deposits that have lead isotope ratios within the
appropriate range. Following this initial selection,
it is necessary to check the geochemistry and
historyofexploitationofthepossibleoredeposits
and eliminate those which cannot have been used
as a source of given material on geochemical or
historical grounds.
Plotting the lead isotope ratios of the artifacts
and ores on two lead isotope diagrams to include
all three independent lead isotope ratios provides
the most accurate visual representation of the
consistencyofleadisotoperatiosoftheoresand
artifacts. e lead isotope ratios used for such
plots in geochronology always relate to 204Pb,
as seen in Figure 1b and c: X =206Pb/204 Pb,
Y1=207Pb/204 Pb and Y2=208Pb/204 Pb. However,
because the isotope 204 Pb in naturally occurring
lead is always present in much smaller quantities
than the other three radiogenic isotopes, the mea-
surements of the ratios to 204Pb are less accurate.
erefore, in principle, the plots of radiogenic
ratios can show more accurately the comparative
positions of data points. For this reason, in the
research publications of archaeological applica-
tions of lead isotope measurements, most oen
the two plots represent the data along the axes:
X=207Pb/206 Pb, Y1=208Pb/206 Pb and Y2=
206Pb/204 Pb. Some researchers also used other
ratios,butessentiallythechoiceofleadisotope
ratios for comparative plots always leads to the
same conclusions if all three independent ratios
are considered.
Problems and additional help
with interpretation of lead isotope data
Not all ore deposits have dierent ranges of lead
isotope ratios, and therefore there are common
instances of overlap of the data for deposits
from dierent geographical regions. In such
cases,therearetwomainpointsthatmustbe
considered: the geochemistry of the ores in
each of the deposits, and the probability of their
exploitation in the period under consideration.
erefore, it is necessary to compare the ele-
mental compositions of ancient artifacts with
the geochemistry of the ores and the periods
of exploitation of the selected ore deposits. e
elemental compositions of metals reect the type
of ore mineral used for smelting, and typical
impurities are oen characteristic of certain
deposits.Furthermore,thetraceelementsigna-
tures can discriminate metal groups that not only
reect various ore types, but are important for
comparison with previously analyzed artifacts
from various regions and certain chronologies.
However, it is important to stress that the iden-
tical lead isotope characteristics of a group of
metals do not guarantee that their elemental
characteristics will also form a group, and vice
versa.
ere are two other serious problems that
can aect the provenancing of ancient metals by
their lead isotope ratios: one is the possibility
of addition of lead to the metal (copper, tin,
bronze, or silver), and the other is remelting of
metal pieces from dierent locations and making
a new artifact from the metal obtained in this
way. e addition of lead should be considered
if the lead content is above 1–2%. en the
comparisonsmustbemadewiththesourcesof
lead or copper/lead ores. e problem of mixing
can be assessed by examining the distribution
of the data points on the lead isotope plots:
each metal obtained from mixing two other
sources will plot somewhere along a straight line
between the data points representing these other
sources.
It is sometimes concluded in the literature
that lead isotope provenance methods allow only
“rejection” of a source of the ancient metal, and
not a positive conclusion of its provenance. is
is far too weak an assessment of this provenance
tool. e lead isotope compositions, particularly
combined with geochemical and archaeometal-
lurgical data, give much more information than
a simple rejection of a source on the grounds of
dierence between lead isotope ratios of ores and
the ancient metal. While the lead isotope data
do not allow a denite positive assignment, they
tell us which assignments are possible, and that
this information should be taken further to look
at which of the ore deposits can be considered as
NON- STABLE ISOTOPES I N ARCHAEOLOGY 5
sources, particularly if evidence exists that these
were exploited at the times in question.
The impact of lead isotope provenance
studies in archaeology
Figures 1b and c show an example of a com-
parative plot of lead isotope ratios for some of
the copper ore deposits in the Mediterranean
that were exploited in the Bronze Age, and those
measured for over 100 ingots of copper of a
very specic shape (“oxhide ingots”) dated to the
Late Bronze Age and found on various sites in
southern Europe, Turkey, and Egypt (Stos-Gale
et al. 1997). is plot demonstrates clearly that
all analyzed oxhide ingots are consistent with the
origin of the ores from Cyprus. is conclusion
was at rst highly surprising for archaeologists,
since many regions where the ingots were found
(i.e., Sardinia, Egypt, Greece) exploited their own
copper ores in the Late Bronze Age, but chemical
and archaeological evidence fully supports this
conclusion. e lead isotope research aimed
at provenance of Bronze Age metals brought
some other unexpected results concerning the
exploitation of copper in Europe and import of
copper into Scandinavia (e.g., Ling et al. 2014;
Artioli et al. 2014; Cattin et al. 2009; Gale et al.
2003).
Strontium as tracer of origin
Strontium(Sr)isamemberofthealkaline
earthsandisknowntoreplacecalciuminmany
minerals. Strontium, therefore, is dispersed in
all calcium-bearing minerals, its salts are dis-
solved in fresh and seawaters, and it is present in
sedimentary and igneous rocks. Strontium com-
pounds in the soil and water enter plants, animals
and humans through direct uptake or through the
food chains. Strontium has four naturally occur-
ring isotopes: 88Sr, 87Sr, 86Sr, and 84Sr, all of which
are stable. eir relative abundances are 82.5%,
7%, 9.9% and 0.56% respectively. e isotopic
abundances of strontium are variable because
of the formation of radiogenic 87Sr through the
decay of terrestrial 87Rb with a half-life of 4.88
×1010. For this reason, the precise composition
of strontium in a rock or mineral that contains
rubidium depends on the age and Rb/Sr ratio
of that rock or mineral. Strontium isotopes can
be measured very accurately using TIMS and
MC-ICP-MS; the typical overall accuracy for
TIMS is ±0.01% in measuring the isotopic ratio
of 87Sr/86 Sr for about 0.5 μgofSrextractedfrom
samples of minerals. TIMS has been used for
longer than the other instruments and is the
most routinely used procedure for strontium
isotope analyses. Although using the TIMS is
time-consuming (each analysis could take 1–2 h),
it requires less strontium compared with other
solution-based analyses, and has the highest
reproducibility and precision.
Rb/Sr dating is used in geochronology, but the
characteristic Sr isotope compositions of minerals
frompossiblesourcesofrawmaterialscanbeused
for nding the origin of artifacts made of natural
sedimentary rocks, bone, and ivory. e earliest
research applying Sr isotopes combined with sul-
furtoprovenancinggypsumusedinMycenaean
and Minoan palaces was published in 1988 (Gale
et al. 1988).
Isotope ratio analysis of the radiogenic stron-
tium is a well-established procedure used in
geology, and therefore a considerable data-
base of Sr isotope ratios in rocks in dierent
locations—regional isoscapes—are available for
archaeological projects. 87Sr/86 Sr values change
little as they pass from weathered rocks through
soils to the food chain, and it has been proven that
the isotopic compositions of strontium deposited
in human teeth and bones reect the composition
of the natural environment that supplied the food
consumed by the individuals. erefore, analyses
of radiogenic strontium isotopes combined with
application of other scientic techniques (for
example, DNA, as well as oxygen and lead iso-
topes) in either hard and/or so tissue allow the
reconstruction of ancient human mobility at the
individual level.One of the most iconic projects
basedonthismethodologywasthestudyofthe
remains of the Egtved Girl, in a Bronze Age oak
con burial in Denmark (Frei et al. 2015).
SEE ALSO: DNA: Mitochondrial; DNA: Next
Generation Sequencing; ICP-MS and Material
Analysis;Isotopes;MobilityandLeadIsotopes;
MobilityandOxygenIsotopes;Mobilityand
Strontium Isotopes
6NON- STABLE ISOTOPES I N ARCHAEOLOGY
REFERENCES
Artioli, Gilberto, Ivana Angelini, Paolo Nimis, and
Igor M. Villa. 2014. “A Lead Isotope Database of
Copper Ores from the Southeastern Alps: a Tool
for the Investigation of Prehistoric Copper Metal-
lurgy.” Journal of Archaeological Science 75: 27–39.
DOI:10.1016/j.jas.2016.09.005.
Brill, Robert H., and J. Marion Wampler. 1967. “Iso-
tope Studies of Ancient Lead.” American Journal of
Archaeolog y 71: 63–77.
Cattin, Florence, Barbara Guénette-Beck, Marie Besse,
and Vincent Serneels. 2009. “Lead Isotopes and
Archaeometallurgy.” Journal of Archaeological and
Anthropological Sciences 1 (3): 137−48. DOI:10.1007/
s12520-009-0013-4.
Faure, Gunter. 1986. Principles of Isotope Geology.New
Yo r k : Wi l e y .
Frei, Karin Margarita, Ulla Mannering, Kristian Kris-
tiansen,MortenE.Allento,AndrewS.Wilson,
Irene Skals, Silvana Tridico, et al. 2015. “Tracing
the Dynamic Life Story of a Bronze Age Female.”
Nature Scientic Reports 5: 10431. DOI:10.1038/
srep10431.1-7.
Gale, Noel H., and Zoa A. Stos-Gale. 1982. “Bronze
Age Copper Sources in the Mediterranean: A
New Approach.” Science 216: 11–19. DOI:10.1126/
science.216.4541.11.
Gale, Noel H., Hubert C. Einfalt, Henry W. Hubberten,
and Richard E. Jones. 1988. “e Sources of Myce-
naean Gypsum.” Journal of Archaeological Science 15:
57–72. DOI:10.1016/0305-4403(88)90019-2.
Gale, Noel H., Zoa Stos-Gale, Ana Raduncheva, Ivan
Ivanov, Peter Lilov, Ted Todorov, and Ivan Panay-
otov. 2003. “Early Metallurgy in Bulgaria.” In Min-
ing and Metal Production rough the Ages, edited
by Paul Craddock and Janet Lang, 122–73. London:
British Museum Press.
Ling,Johan,ZoaStos-Gale,LenaGrandin,KjellBill-
ström, Eva Hjärthner-Holdar, and Per-Olaf Persson.
2014. “MovingMetals II: Provenancing Scandinavian
Bronze Age Artefacts by Lead Isotope and Elemen-
tal Analyses.” Journal of Archaeological Science 41:
106–32. DOI:10.1016/j.jas.2013.07.018.
Pernicka, Ernst, Joachim Lutz, and omas Stollner.
2016. “Bronze Age Copper Produced at Mitterberg,
Austria, and its Distribution.” Archaeo logia Austria ca
100: 19–55. DOI:10.1553/archaeologia100s19.
Stos-Gale, Zoa A., and Noel H. Gale. 2009. “Metal
Provenancing using Isotopes and the Oxford
Archaeological Lead Isotope Database (OXALID).”
Archaeological and Anthropological Sciences 1(3):
195–213. DOI:10.1007/s12520-009-0011-6.
Stos-Gale, Zoa A., George Maliotis, Noel H. Gale, and
Nick Annetts. 1997. “Lead Isotope Characteristics of
the Cyprus Copper Ore Deposits Applied to Prove-
nance Studies of Copper Oxhide Ingots.” Archaeom-
etry 39 (1): 83–124. DOI:10.1111/j.1475-4754.1997.
tb00792.x.
University of Oxford. 2017. OXALID Database, http://
oxalid.arch.ox.ac.uk.
FURTHER READINGS
Gale, Noel H., and Zoa A. Stos-Gale. 2000. “Lead
Isotope Analyses Applied to Provenance Studies.”
In Modern Analytical Methods in Art and Archaeol-
ogy, edited by Enrico Ciliberto and Giuseppe Spoto,
Chemical Analyses Series, Vol. 155, Chapter 17,
503–84. Chichester: Wiley.
Pernicka, Ernst. 2014. “Provenance Determination of
ArchaeologicalMetal Objects.” In Archaeometallurgy
in Global Perspective, edited by Benjamin W. Roberts
and Christopher P. ornton, 239–68. New York:
Springer Science +Business Media.