MAT E RI A LS CH A R A CTE R IZ A TI O N 6 1 (2 0 1 0 ) 5 9 –6 4

available at www.sciencedirect.com

www.elsevier.com/locate/matchar

PIXE analysis of medieval silver coins

H. Ben Abdelouahed a,⁎, F. Gharbi a , M. Roumié b , S. Baccouche a , K. Ben Romdhane c ,

B. Nsouli b , A. Trabelsi a
a
Centre National des Sciences et Technologies Nucléaires (CNSTN), Pôle technologique, 2020 Sidi Thabet, Tunis, Tunisia
b
IBA Laboratory, Lebanese Atomic Energy Commission, National Council for Scientific Research, 11-8281, Beirut, Lebanon
c
Faculté des lettres et des sciences humaines, Université de Tunis, Tunisia

AR TIC LE D ATA ABSTR ACT

Article history: We applied the proton-induced X-ray emission (PIXE) analytical technique to twenty-eight
Received 17 June 2009 medieval silver coins, selected from the Tunisian treasury. The purpose is to study the
Received in revised form fineness evolution from the beginning of the 7th to the 15th centuries AD. Each silver coin
16 October 2009 was cleaned with a diluted acid solution and then exposed to a 3 MeV proton beam from a
Accepted 16 October 2009 1.7 MV tandem accelerator. To allow the simultaneous detection of light and heavy
elements, a funny aluminum filter was positioned in front of the Si(Li) detector entrance
Keywords: which is placed at 135° to the beam direction. The elements Cu, Pb, and Au were observed in
PIXE the studied coins along with the major component silver. The concentration of Ag,
Silver coins presumably the main constituent of the coins, varies from 55% to 99%. This significant
GUPIX variation in the concentration of the major constituent reveals the economical difficulties
encountered by each dynasty. It could be also attributed to differences in the composition of
the silver mines used to strike the coins in different locations. That fineness evolution also
reflects the poor quality of the control practices during this medieval period. In order to
verify the ability of PIXE analytical method to distinguish between apparently similar coins,
we applied hierarchical cluster analysis to our results to classify them into different
subgroups of similar elemental composition.
© 2009 Elsevier Inc. All rights reserved.

1. Introduction method of PIXE, nuclear elastic backscattering spectrometry

(BS), channeling contrast microscopy (CCM), and secondary
Since 1953, several atomic and nuclear techniques have been electron microscopy to image the structural and elemental
developed and applied to the non-destructive elemental composition of materials and to get, from the images, reliable
analysis of archaeological artifacts [1–3]. During the 1960s quantitative information about materials.
and 70s, the development of solid-state detectors [4], proton Ancient coins are often struck with a well-controlled alloy
activation analysis (PAA) method [5] and ion beam analysis by a known mint with a date of issue and so, for most of them,
(IBA) techniques (mainly particle-induced X-ray emission references can be found in ancient documents dealing about
(PIXE), particle-induced γ-ray emission (PIGE), and Rutherford their typology, metrology and chronology. Nevertheless, cases
backscattering (RBS)) allowed growth of manufacturing and exist where the lack of written documentation brings obscu-
origin technologies for archaeological materials [6–10]. Now- rity to numismatists. For the investigation of such “obscure”
adays, expansion of new research areas in archeometry is ancient coins PIXE is very attractive: it combines interesting
noted [11–14] and new methods are established to improve analytical capabilities together with a non-destructive charac-
artifacts analysis. Jamieson et al. [15] proposed combinative ter. The PIXE method makes possible to analyze simulta-

⁎ Corresponding author. Centre National des Sciences et Technologies Nucléaires (CNSTN), Pôle Technologique, 2020 Sidi Thabet, Tunis,
Tunisia. Tel.: +216 21091877; fax: +216 71537555.
E-mail address: [email protected] (H. Ben Abdelouahed).
1044-5803/$ – see front matter © 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.matchar.2009.10.008
60 MAT ER IA LS CH A R A CTE R IZ A TI O N 6 1 (2 0 1 0) 5 9–6 4

neously a great number of elements (from sodium to uranium) museum of Tunisia, and were chosen to cover a large medieval
in a wide concentration range. Ehrhardt et al. [16] used PIXE to minting period. We indicated in Table 1 the name and the period
determine the elemental compositions of artifacts and stated of the dynasty from which each group of coins is issued. The
that PIXE application contributes strongly to identify related PIXE spectra of all the analyzed coins were proceeded with the
manufacturing techniques and technical processes employed. well known GUPIX [24] software (Fig. 1).
However, when applied to corroded coins without sample pre- In this paper, we present the obtained results and we
paration, the accuracy of PIXE results can be affected [17–19]. It determine the fineness evolution during the medieval period,
is because PIXE method has the disadvantage of low penetra- ranging from 702 to 1574 AD, and deduce the difficult economi-
tion depths (the penetration range of 2.5–3.0 MeV protons is cal situation of the country encountered during successive
about 30 µm) [20]. If ancient silver coins are adequately cleaned dynasties. Also we introduce here our hierarchical cluster
and possible enriched layer is removed, the study of their analysis, using Stat-Graphics [25] program, to classify our
chemical composition with PIXE technique will bring a real studied coins into groups of similar elemental composition.
success to determine their fineness and to provide valuable in-
formation about the economy of the time of their minting [21].
It was demonstrated that, if six cited experimental condi- 2. Experimental Setup
tions [20] are carefully controlled, PIXE appears a very promising
method and the suitable one for determining ancient silver For fear of silver surface enrichment of silver–copper alloys
coins fineness. caused by suspect cleaning procedures, different kinds of
In this work, we applied the PIXE technique to study twenty- harmless cleaning protocols were adopted. Hence, our coins
eight Tunisian silver coins, called “Islamic dirhams” (a unit of were first immerged in a highly diluted acid solution for about
currency in several Arab nations) [22], belonging to six 5 h and then washed in distilled water. Dirt layers on the
successive dynasties from the beginning of the 8th century surface were thereby removed. The coins were then subjected
(Umayyad dynasty 702–748 AD) until the 16th century (Hafsid to PIXE analysis.
dynasty 1230–1574 AD). According to Islamic law, the Islamic According to Beck et al. [26], this cleaning procedure may
dirham has a specific weight of pure silver equivalent to 3.0 g, induce silver surface enrichment causing a limitation for the
nevertheless, there were many silver coins produced before
1468 AD which were issued in different weights [23]. The coins
were supplied from the treasury collection of the numismatic

Table 1 – Characteristics of the twenty-eight analyzed silver

coins.
Coin ref. Dynasty Period Mass Shape Dimension
(A.D) (g) (mm)

U1 Umayyad 702–748 2.87 Disc 13 < r < 14

U2 2.69
A1 Abbasid 775–833 2.96 Disc 13 < r < 14
A2 2.83
F1 Fatimid 909–975 2.25 Disc 7 < r < 10
F2 1.68
F3 1.39
F4 1.16
F5 1.67
F6 1.37
Z1 Zirid 975–1159 1.45 Disc 8<r<9
Z2 1.41
Z3 1.39
Z4 1.14
Z5 1.33
Z6 1.30
M1 Mowahid 1159–1230 1.50 Square 14 < l < 15
M2 1.53
M3 1.52
M4 1.53
M5 1.54
M6 1.52
H1 Hafsid 1230–1574 1.45 Square 14 < l < 15
H2 1.29
H3 1.52
H4 1.52
H5 1.52
H6 1.34 Fig. 1 – Six groups of medieval silver coins supplied from the
treasury collection of the numismatic museum of Tunisia.
 MAT E RI A LS CH A R A CTE R IZ A TI O N 6 1 (2 0 1 0 ) 5 9 –6 4 61

analysis of the silver–copper alloys; hence in order to check any kept the beam current lower than 3 nA in order to avoid high
surface alteration, RBS was used simultaneously with PIXE. counting rates at the detector which could increase the
The probed depth, in such material, of 3 MeV proton using RBS background noise and then alter the detection sensitivity.
mode is more or less 20 μm. The RBS analysis didn't show any The detector signal was shaped and amplified and finally,
significant gradient concentration of either Cu or Ag, at least in through a pulse height analysis, the energy spectrum was
the 20 μm probed depth of the coin. Consequently, we can stored and displayed in a multichannel analyzer. We present
assume that a signal of a bulky material was seen. Also in order in Fig. 3 the typical spectrum of one of our silver coins.
to check the same phenomena, two distinguished parts (where GUPIX [24] was used for PIXE spectrum fitting and
part of the surface was removed mechanically while the other calculation of elemental composition of the coins.
part was just cleaned) of one of the coins showed closed PIXE
values, where the Ag content in the first part was slightly
higher than the other one but not more than 5%. 3. Data Analysis
Various kinds of backgrounds in PIXE should take into
account to enable determination of absolute concentrations of For thick target, as it is the case here, GUPIX takes into
major components and to point out the presence of existing account: the energy loss of the 3 MeV incident protons, the
trace elements. The PIXE detection limit has been theoretically variation of X-ray production cross-sections with the decreas-
derived and a contour representation of the detection limit for ing proton energy, the absorption of X-rays from different
trace-element analyses has been calculated [27], and it is depths in the target, and the elemental effect (enhancement of
shown that the best detection limit is obtained with about the X-ray yields). GUPIX uses all the inputted specification of
3 MeV proton beam, depending of course on the sample the Si(Li) X-ray detector to generate a theoretical curve for its
matrix. Hence, we used 3 MeV proton beam delivered from the efficiency. GUPIX allows also for the escape peak, sum peaks
1.7 MV tandem accelerator of the Lebanese Atomic Energy and low energy tailing of X-rays.
Commission [28]. At first step, to calculate the elemental concentration from
We present in Fig. 2 the target chamber in which a vacuum unknown coins, each spectrum was qualitatively character-
of about 10− 7 Torr is maintained and in which we can see a ized to introduce to GUPIX the various X-ray lines of the
wheel target holder (where we put coins) monitored by a PC elements present in the spectrum and the total charge Qp
through a stepping motor. The chamber has several view ports carried by Np protons (Qp = 1.60210 10− 13 Np µC) on the analyzed
for multipurpose studies, two of them being designed to carry coin. Those given inputs are used then by the program to
out PIXE and RBS analysis. generate a theoretical spectrum with all the visible elements.
From the PIXE view port of the target chamber and at 135° to The calculation of the theoretical intensity of the emitted
the beam direction, a retractable collimated ORTEC Si(Li) X-ray line It(Z) is based on the hypothesis of a homogenous
detector (FWHM 170 eV at 5.9 keV), with 30 mm2 active area and flat sample. According to our experiment setup, the
and 12.7 µm thick Be window, is placed to detect X-ray emission proton beam is bombarding the coin at an angle φ equal to 45°
from targets. To make possible the detection of light and heavy with the normal to coin surface. The X-ray detector is
elements simultaneously, Roumié et al. [29] proposed to place positioned at an angle θ equal to 135° to the proton beam
an aluminum funny filter in front of the detector window. direction, subtending a small solid angle. The theoretical
The beam charge deposition on the target is measured via a intensity of the emitted X-ray line, of element Z present with a
charge integrator connected to a digital counter-timer. We mass concentration CZ, from a volume of thickness dδ at the

Fig. 2 – Target chamber: coins are placed in the wheel target Fig. 3 – PIXE spectrum of Umayyad coin obtained by using
holder monitored by PC through a stepping motor. At 135° to 3 MeV protons and an Al funny filter in front of the detector.
the beam direction, a retractable collimated Si(Li) detector is The X-ray lines of the elements present in the spectrum are
placed to detect X-ray emission from targets. those of Ag, Pb, and Au.
62 MAT ER IA LS CH A R A CTE R IZ A TI O N 6 1 (2 0 1 0) 5 9–6 4

depth δ behind the coin surface, due to a number Np of Table 2 – Elemental composition of medieval silver coins
impinging protons, is well described by the following integral obtained by PIXE analysis; n.d.: not detected.
over the coin thickness: Coin ref Ag (%) Cu (%) Pb (%) Au (%) Zn (%)

CZ Nav τZ U1 98.7 ± 0.1 0.45 ± 0.01 0.12 ± 0.01 0.67 ± 0.01 n.d.
It ðZÞ = Np ωZ εZ fZ ∫σZ ðδ; Ep Þe−μZ δ=tgφ dδ ð1Þ
AZ τ U2 96.1 ± 0.2 2.49 ± 0.04 0.97 ± 0.02 0.41 ± 0.03 n.d.
A1 90.6 ± 0.2 8.43 ± 0.06 0.88 ± 0.02 0.05 ± 0.01 n.d.
which takes into account the decreasing energy Ep of the A2 90.3 ± 0.2 8.47 ± 0.07 1.08 ± 0.03 0.04 ± 0.01 n.d.
proton beam entering the thick sample and the attenuation of F1 75.8 ± 0.2 22.4 ± 0.1 1.46 ± 0.03 0.22 ± 0.02 < 0.08
the X-rays emitted by the element in the direction of the F2 74.3 ± 0.2 22.4 ± 0.1 2.64 ± 0.05 0.33 ± 0.02 0.20 ± 0.03
detector. In the relation (1): σZ(δ,Ep) corresponds to the F3 69.3 ± 0.2 28.3 ± 0.1 1.96 ± 0.05 0.19 ± 0.02 0.12 ± 0.04
F4 68.8 ± 0.2 28.6 ± 0.1 1.83 ± 0.04 0.33 ± 0.03 0.33 ± 0.05
ionisation cross section, which is studied at our previous
F5 77.2 ± 0.2 19.0 ± 0.1 3.27 ± 0.05 0.30 ± 0.02 < 0.08
work [30], varies with depth δ due to the energy loss dEp of the F6 64.2 ± 0.2 33.2 ± 0.1 2.15 ± 0.05 0.25 ± 0.02 < 0.08
proton beam; ωZ is the fluorescent yield and τZ /τ is the fraction Z1 57.6 ± 0.2 40.2 ± 0.1 1.76 ± 0.04 0.22 ± 0.02 0.08 ± 0.05
of the total X-ray emission of the analyzed line; AZ is the Z2 70.6 ± 0.2 26.7 ± 0.1 2.19 ± 0.05 0.31 ± 0.03 < 0.08
atomic mass of the considered element, Nav is the Avogadro Z3 63.0 ± 0.2 35.0 ± 0.1 1.50 ± 0.04 0.23 ± 0.02 0.15 ± 0.05
number; μZ is the absorption coefficient of emitted X-rays, ɛZ Z4 54.9 ± 0.2 42.9 ± 0.1 1.79 ± 0.04 0.15 ± 0.02 0.09 ± 0.05
Z5 59.8 ± 0.2 38.3 ± 0.1 1.49 ± 0.04 0.19 ± 0.02 < 0.08
is the detector efficiency, and fZ corresponds to attenuation
Z6 73.0 ± 0.2 25.0 ± 0.1 1.67 ± 0.05 0.21 ± 0.02 < 0.08
fraction coming from the aluminum funny filter placed in
M1 91.0 ± 0.2 8.60 ± 0.08 0.23 ± 0.02 0.10 ± 0.01 n.d.
front of the detector. M2 99.8 ± 0.2 0.03 ± 0.01 0.05 ± 0.01 < 0.03 n.d.
At a second step of the spectral data analysis, we calibrated M3 99.7 ± 0.2 0.13 ± 0.01 0.02 ± 0.01 0.08 ± 0.01 n.d.
GUPIX by analyzing two coins: the inner parts of 1 Euro (75% M4 99.8 ± 0.2 0.10 ± 0.01 0.02 ± 0.01 < 0.03 n.d.
Cu, 25%Ni) and 2 Euro (75% Cu, 20%Zn, 5% Ni). The M5 97.4 ± 0.2 1.76 ± 0.03 0.80 ± 0.02 < 0.03 n.d.
composition of those recent coins is well defined and could M6 96.3 ± 0.1 3.50 ± 0.03 0.06 ± 0.01 0.03 ± 0.01 n.d.
H1 90.8 ± 0.2 8.80 ± 0.08 0.25 ± 0.02 0.09 ± 0.02 n.d.
be used as reference standards. The calibration step is done to
H2 99.8 ± 0.2 0.06 ± 0.01 0.05 ± 0.01 < 0.03 n.d.
obtain the calibration factor, known as H value [31], which
H3 99.8 ± 0.2 0.03 ± 0.01 0.08 ± 0.01 < 0.03 n.d.
relates the calculated intensity (from the GUPIX database) It(Z) H4 99.9 ± 0.2 0.04 ± 0.01 < 0.02 < 0.03 n.d.
of the emitted X-ray line from an element Z with the measured H5 99.9 ± 0.2 0.03 ± 0.01 0.03 ± 0.01 < 0.03 n.d.
one Im(Z) [32]: H6 99.4 ± 0.2 0.03 ± 0.01 0.13 ± 0.01 0.08 ± 0.01 n.d.

Im ðZÞ = H It ðZÞ: ð2Þ Results are given by GUPIX program.

The last step is the iteration procedure that GUPIX runs to

fit the experimental spectrum with the generated one, decrease in silver composition of silver coins, which we call
providing the elemental composition on the computer. The debasement, is then noted only in the period from Fatimid's
result of the elemental analysis of the twenty-eight silver dynasty to the end of Zirid's dynasty. There are several
coins obtained by GUPIX program is shown in Table 2. possible reasons for debasement to occur [33–35] but it is
assigned in most cases to economical difficulties leading to
the replacement of the noble metal by a less expensive one.
According to Mujani [23], one of the reasons which caused
4. Results and Discussion debasement during that period was the government's in-
creasing financial needs. The revenues from the agriculture,
The elements Cu, Pb, Au and Zn were observed in the coins industrial and economical sectors were not sufficient to cover
along with the major component silver. The percentage of all its expenses. This caused to manipulate the currency in
silver varies between 54 and 99%. The fineness evolution of order to achieve fiscal advantage in the short term. Thus it was
the Tunisian silver coins during all the medieval period is
summarized in Fig. 4. According to the diagram (Fig. 4), we can
say that until 833, that is from Umayyad reign (U) to Abbasid
dynasty (A), the fineness of the coins varies from about 90.3%
silver to 98.7%. From the beginning of the Fatimid dynasty (F)
the devaluation of the silver coin is noted. But for the period of
Zirid's dynasty (Z) the silver content is seriously decreasing to
an average value of 63.1%. According to Balog [33], in the
Fatimid Dynasty, the fineness of the dirham fell gradually
from 86% at close to the tenth century to 24% at the close the
eleventh century. The same reference notices that the weight
of the dirham was slightly reduced during that period and full
dirham became increasingly rare. Hence, our analyses results Fig. 4 – Fineness evolution diagram of the medieval silver
come to confirm the fact. coins in Tunisia. The diagram is deduced from results
The same diagram shows that the dynasties coming after corresponding to the concentration of Ag element in each
Zirid's one, Mowahid (M) and Hafsid (H) present good level of group of coins which are Umayyad (U), Abbasid (A), Fatimid (F),
fineness reaching an average silver content of 97.8%. The Zirid (Z), Mowahid (M) and Hafsid (H).
 MAT E RI A LS CH A R A CTE R IZ A TI O N 6 1 (2 0 1 0 ) 5 9 –6 4 63

necessary to lower the weight of the dirhams as well as to

debase them.
Our results show (at Fig. 5) that silver is replaced by copper
in coins at the period of Fatimid's and Zirid's dynasties. We
noted a strong negative correlation between Ag and Cu for all
the analyzed medieval coins. Such a correlation is an
indication that copper was deliberately added to silver.
To classify our PIXE analytical results of the twenty-eight
silver coins into groups of similar elemental composition, we
applied to results of Table 2 a hierarchical cluster analysis
[28,29]. For that analysis we used Stat-Graphics program [25] Fig. 6 – Classification of medieval silver coins according to
which is an all-purpose statistics package on statistical their elemental composition by using Stat-Graphics program.
process control and design of experiments; it disposes over
than 250 statistical analysis ranging from basic statistical
analysis to the latest and most sophisticated tools. As shown essential to indicate that PIXE is a surface analysis (the
in Fig. 6, the hierarchical cluster analysis result indicates that penetration range of 2.3–3.0 MeV protons is about 30 µm) and
only two major groups can be distinguished: so the positioning of the coin has to be carefully controlled to
avoid a random analysis of the altered part of the surface of
■ Group 1, consisting of only Umayyad coins. the coin.
■ Group 2, consisting of two subgroups of coins: subgroup of Our study of fineness evolution revealed a silver coins
Abbasid, Mowahid and Hafsid coins, and subgroup of debasement at a limited period of the medieval epoch
Fatimid and Zirid coins. (between 909 and 1159 AD). To solve the real reason behind
the demonstrated debasement, our PIXE analytical results are
According to Lowick's reference [36], Umayyad coins were under numismatic studies.
minted at al-Basrah and not at Afrikya, the minting place of Application of hierarchical cluster analysis to our results,
the other five groups. Consequently, it seems that our study allowed us to classify them into different subgroups of similar
comes to support such a fact. elemental composition and to verify the ability of PIXE
Statistical analysis of large data sets (cluster analysis) is a analytical method to distinguish between apparently similar
powerful tool that is increasingly being used in materials groups of coins.
characterization. The fact that the coins can be classified into
two major clusters, with the second cluster having two major
sub-clusters is a fine observation, and the fact that these Acknowledgments
groups correlate perfectly by mint location shows how this
tool might be used to analyze coins of unknown provenance. The authors acknowledge the scientific organizers and
General Directors, Mr. Mahmoud Nasreddine before and Mr.
Abdelmajid Mahjoub presently, of the Arab Atomic Energy
5. Conclusion Agency for the assistance of this Tunisian–Lebanon project.

Through this application of PIXE analytical technique to the REFERENCES

study of fineness evolution of the Tunisian country during the
medieval centuries, we underlined the accuracy of PIXE
[1] Ambrosino G, Pindrus P. Non-destructive analysis of ancient
method to carry out fast and precise information about
metal objects. Rev Metall 1953;50:136–8.
major and minor elements in silver coins. It is however [2] Emeleus VM. Neutron activation, as applied to trace elements,
determination in pottery and coins. Archeometry 1958;1:6–15.
[3] Kraay CM. Gold and copper traces in early Greek coins.
Archaeometry 1959;15:53–78.
[4] Fleischer RL, Price PB, Walker RM. Solid-state detectors:
applications to nuclear sciences and geophysics. Annu Rev
Nucl Sci 1965;15:1–28.
[5] Barrandon JN, Debrun JL. Activation analysis with charged
particles of intermediate energy: performances and
applications. Accuracy in trace analysis: sampling, sample
handling, analysis, vol. 422. NBS special publication; 1976.
p. 1215–32.
[6] Condamin J, Picon M. Changes suffered by coins in the course
of time and the influence on the results of different methods
of analysis. Methods of chemical and metallurgical
investigation of ancient coinage, vol. 8. Royal Numismatic
society; 1972. p. 49–66.
[7] Friedman AM, Conway M, Kastner M, Milsted J, Metta D, Fields PR.
Fig. 5 – Correlation between silver and copper concentration Copper artifacts: correlation with source type of copper ores.
in all the studied silver coins. Science 1966;152:1504–6.
64 MAT ER IA LS CH A R A CTE R IZ A TI O N 6 1 (2 0 1 0) 5 9–6 4

[8] Gordus AA. Rapid non-destructive activation analysis of silver [22] Masudul Alam Choudhury. Money in Islam: a study in Islamic
coins. In: Brill RH, editor. Science and Archaeology. MIT Press; political economy. Volume 3 of Routledge International
1971. p. 145–55. Studies in Money and Banking, Routledge; 1997.
[9] Meyers P, Van Zelst L, Sayre EV. Major and trace elements in [23] DR Wan Kamal Mujani, The monetary system from historical
Sassanian silver. J Radiation Chem 1973;39:99–111. perspectives, http://www.ialsnet.org/meetings/business/
[10] Guerra MF, Barrandon J-N, Le Roy Ladurie E, Morrisson C, KamalWan-Malaysia.pdf.
Colin B. The diffusion of the silver from Potosi in the XVI [24] Maxwell JA, Teedale WJ, Campbell JL. Nucl Instr Meth B
century European coinage. Archaeometry, vol. 90. Basel: 1995;95:107.
Birkhauser Verlag; 1991. p. 11–8. [25] STATGRAPHICS Plus5.0 Software, http://www.pcworld.com/
[11] Hajivaliei M, Garg ML, Handa DK, Govil KL, Kakavand T, downloads/file/fid,58081-order,1-page,1/description.html,
Vijayan V, Singh KP, Govil IM. PIXE analysis of ancient Indian 2006.
coins. Nucl Instr Meth B 1999;150:645–50. [26] Beck L, Bosonnet S, Réveillon S, Eliot D, Pilon F. Silver
[12] Linke R, Schreiner M, Demortier G. The application of photon, enrichment of silver–copper alloys: a limitation for the
electron and proton induced X-ray analysis for the analysis of ancient silver coins by surface techniques. Nucl
identification and characterization of medieval silver coins. Instr Meth B 2004;226:153–62.
Nucl Instr Meth B 2004;266:172–8. [27] Ishii K, Morita S. Theoretical estimation of PIXE detection
[13] Santra S, Mitra D, Sarkar M, Bhattacharya D, Denker A, limits. Nucl Instr Meth B 1988;34(2):209–16.
Opitz-Coutureau J, Rauschenberg J. Analysis of some coins by [28] Roumie M, Nsouli B, Zahraman K, Reslan A. Application of
energy dispersive X-ray fluorescence (EDXRF) and high PIXE using Al funny filter for cluster analysis of Byzantine
energy particle induced X-ray emission (PIXE) technique. Nucl amphorae from Beirut. Nucl Instr Meth B 2004;398:219–20.
Instr Meth B 2005;229:465–70. [29] Roumié M, Reynolds P, Attallah C, Elias Bakraji, Zahraman K,
[14] Mendoza-Lopez ML, Pérez-Bueno JJ, Rodriguez-Garcia ME. Bsouli B. Provenance study of excavated pottery from Beirut
Characterization of silver alloys used in modern Mexican using PIXE cluster analysis. Nucl Instr Meth B 2006;249:612–5.
coins. Mater Charact 2009;60(9):1041–8. [30] Ben Abdelouahed H, Incerti S, Mantero A. New Geeant4 cross
[15] Jamieson David N, Romano Linda T, Grime Geoff W, Watt Frank. section models for PIXE simulation. Nucl Instr Meth B
Nuclear microscopy: a new way of analyzing materials. Mater 2009;267:37–44.
Charact 1990;25:3–15. [31] Denker A, Bohne W, Campbell JL, Heide P, Hopman T, Maxwell JA,
[16] Ehrhardt Kathleen L, Nash Samuel K, Swann Charles P. Opitz-Coutureau J, Rauschengerg J, Rohrich J, Strub E.
Metal-forming among the seventeenth century Illinois, High-energy PIXE using very energetic protons: quantitative
1640–1682, Materials Characterization 45. Issues 2000; analysis and cross-sections. X-ray Spectrom 2005;34:376–80.
4–5:275–88. [32] Shariff A, Bulow K, Elfman M, Kristiansson P, Malmqvist K.
[17] Demortier G, Bodart F, Hackens T. Analyse de monnaies de Calibration of new camber using GUPIX software package of
bronze grecques au moyen de la méthode PIXE et par PIXE analysis. Nucl Instr Meth B 2002;189:131–7.
activation aux neutrons rapides, problèmes de méthodes. Rev [33] Paul Balog. History of the Dirham in Egypt from the Fatimid
Arch 1981;5:63–71. conquest until the collapse of the Mamluk empire, 358 H–922
[18] Beauchesne F, Barrandon J-N, Alves L, Gil FB, Guerra MF. Ion H (968 A D–1517 A D). Rev Numis 1961;6(N3):109–46.
beam analysis of copper and copper alloys coins. [34] Rolnick Arthur J, Velde François R, Webber Warren E. The
Archaeometry 1988;30:187–97. debasement puzzle: an essay on Medieval history. J Econ Hist
[19] Carter FG. Preparation of ancient coins for accurate X-ray 1996;56(4):789–808.
fluorescence analysis. Archaeometry 1964;7:106–13. [35] Guerra MF. Elemental analysis of coins and glasses. Appl
[20] Weber G, Guillaume J, Strivay D, Garnir HP, Marchal A, Radiation Isot 1995;46(6/7):583–8.
Martinot L. Is the external beam PIXE method suitable for [36] Lowick NM. Une monnaie alide d'al-Basrah datée de 145 H
determination ancient silver artifact fineness? Nucl Instr (762-3 après J.-C). Rev Numis 1979;6(N21):218–24.
Meth B 2000;161–163:724–9.
[21] Beck L, Alloin E, Berthier C, Réveillon S, Costa V. Silver surface
enrichment controlled by simultaneous RBS for reliable PIXE
analysis of ancient coins. Nucl Instr Meth B 2008;266:2320–4.