Sei sulla pagina 1di 8

Appl. Phys. A 83, 175182 (2006) DOI: 10.

1007/s00339-006-3503-6

Applied Physics A
Materials Science & Processing

w. kockelmann1,u s. siano2 l. bartoli2 d. visser1,3 p. hallebeek4 r. traum5 r. linke6 m. schreiner6 a. kirfel7

Applications of TOF neutron diffraction in archaeometry


1 2

Rutherford Appleton Laboratory, ISIS Facility, Chilton, OX11 0QX, United Kingdom Istituto di Fisica Applicata CNR, 50019 Sesto Fiorentino, Italy 3 The Netherlands Organisation for Scientic Research (NWO), 2593 CE Den Haag, The Netherlands 4 Netherlands Institute for Cultural Heritage (ICN), 1070 KA Amsterdam, The Netherlands 5 Mnzkabinett, Kunsthistorisches Museum Wien, 1010 Vienna, Austria 6 Institut fr Wissenschaften und Technologien in der Kunst, Akademie der Bildenden Knste, 1010 Vienna, Austria 7 Mineralogisch-Petrologisches Institut, Universitt Bonn, 53115 Bonn, Germany

Received: 5 December 2005/Accepted: 12 December 2005 Published online: 1 March 2006 Springer-Verlag 2006
ABSTRACT Neutron radiation meets the demand for a versatile diagnostic probe for collecting information from the interior of large, undisturbed museum objects or archaeological ndings. Neutrons penetrate through coatings and corrosion layers deep into centimetre-thick materials, a property that makes them ideal for non-destructive examination of objects for which sampling is impractical or unacceptable. A particular attraction of neutron techniques for archaeologists and conservation scientists is the prospect of locating hidden materials and structures inside objects. Time-of-ight (TOF) neutron diffraction allows for the examination of mineral and metal phase contents, crystal structures, grain orientations, and microstructures as well as micro- and macro strains. A promising application is texture analysis which may provide clues to the deformation history of the material, and hence to specic working processes. Here we report on instructive examples of TOF neutron diffraction, including phase analyses of medieval Dutch tin-lead spoons, texture analyses of bronze specimens as well as of 16th-century silver coins. PACS 61.12.-q;

81.05.Bx; 81.70.Jd

Introduction

Among the large variety of chemical, physical and microstructural analytical techniques that are employed to characterize cultural heritage materials, neutron methods play a special role. Neutrons are only weakly absorbed in matter in comparison to other elementary particle probes like photons, protons, or electrons, all of which have limited penetration power and are therefore typically absorbed within the rst millimeter of the objects surface. On the other hand, due to the weak interaction, neutrons are a ne probe for nondestructive examination of the interior of a thick and massive artefact that is not to be damaged by sampling. Neutron analysis is both unique and complementary to X-ray techniques. Compared to synchrotron sources, however, uxes at neutron sources are orders of magnitudes smaller, neutron counting times are typically in the order of minutes to hours, and beam
u Fax: 0044-(0)1235 445720, E-mail: w.kockelmann@rl.ac.uk

diameters are large. Compared to micro focus applications of synchrotron radiation, neutrons achieve a rather modest spatial resolution in the order of millimetres. Hence, neutron techniques have a disadvantage if a small sample is to be studied or if high spatial resolution is required on the surface of an artefact. On the other hand, since the neutron beam generally surveys a much larger sample volume than does an X-ray beam, the results of the neutron analysis have a good chance to truly represent the composition and material properties of the whole object. The high penetration power of neutrons can be understood from their basic particle properties and their interaction mechanisms with matter. The neutron is electrically neutral and foremost interacts with the nuclei of the atoms (leaving aside the magnetic interaction and the use of neutrons as a magnetic probe). A nucleus can be considered as a practically point-like scattering centre that offers a very small cross section for interaction. Hence, if neutrons are sent into a material, most of them will pass through it without any interaction. Some of them will interact in two ways: they can be scattered or they can be captured. These two basic reactions govern the applications of neutrons in material and in archaeological sciences. Table 1 gives an overview of neutron based methods which are, to a large extent, complementary among themselves. In this paper we will concentrate on neutron scattering, and in particular on time-of-ight (TOF) neutron diffraction. TOF neutron diffraction is usually applied at neutron spallation sources that are based on particle accelerators. The spallation source ISIS at the Rutherford Appleton Laboratory produces intense neutron radiation for material research. In particular, diffraction techniques can be straightforwardly applied to problems in archaeological sciences. Diffraction is a direct method for examining structural aspects of artefacts made out of a wide class of materials such as pottery, pigments, stone, marble and metal, providing qualitative and quantitative information on several levels: (i) on the phase composition, (ii) on the crystal and magnetic structures of each constituent phase, (iii) on microstrains as well as macro (residual) strains, and (iv) on texture, i.e., whether or not grains have random or preferred orientations. The microscopic structure is generally related to the material properties and to the specic fabrication processes. For example, the microstructure of a metal can be altered in a characteristic way

176

Applied Physics A Materials Science & Processing

Neutron Activation Analysis for isotope and element analysis is based on the capture of neutrons. Characteristic gamma radiation is emitted during (prompt s) or after (delayed s) the neutron capture. INAA, Instrumental Neutron Activation Analysis [1]: high sensitivity to many trace elements; usually requires sampling; delayed s. NAAR, Neutron Activation Autoradiography [2]: non-destructive method for investigation of (mainly) paintings with cold neutrons; delayed s. PGAA, Prompt Gamma Activation Analysis [3]: applied non-destructively on intact objects based on thermal and cold neutron capture; high sensitivity for some light elements (H, S ,P ,K); prompt s. NRCA, Neutron Resonant Capture Analysis [4]: applied non-destructively on intact objects based on epithermal neutron capture; high sensitivity for some heavy elements (As, Ag, Sb, Sn); prompt s. Neutron Radiography/Tomography [5, 6]: real space imaging is based on the capture and scattering of thermal and cold neutrons to provide an inside view of objects with a spatial resolution down to 100 micrometers; exploits the attenuation of a neutron beam passing through an object; contrast for different elements, high sensitivities for some light elements (e.g. hydrogen); contrast variation by variation of neutrons energies [7]. Further imaging prospects are provided by neutron refraction [8] and Bragg-Edge transmission for residual strain mapping [9]. Neutron Diffraction [1012], based on the elastic scattering of neutrons by periodic, long-range (crystalline) or short range (glass) arrangements of atoms. Can be performed in angle-dispersive or energy-dispersive (time-of-ight) mode using thermal neutrons. Many structural aspects can be studied: phase and structure analysis, texture analysis [1318], microstructure analysis; residual stress analysis [9]. Porosity of a material, and size and surface characteristics of aggregates can be studied by small angle neutron scattering [19].
TABLE 1

Neutron-based techniques for archaeological research

through mechanical deformation or heat treatment. During the ring of pottery, the raw materials undergo chemical changes or mineral phase transformations. Diffraction techniques are ideal to record these kind of structural changes. A particular promising application is neutron texture analysis for recording grain orientation distributions [13]. Since the crystallographic texture of a material is critically dependent on the past mechanical and/or thermal treatment, texture analysis can provide important clues to the creation and deformation history of a sample, be it a mechanically deformed metal or a geological sample deformed by tectonic processes. Consequently, texture analysis may help to uncover historic production steps or the geological deformation history. Moreover, if the historical making techniques are known, texture maps may help to distinguish genuine from fake objects. Neutrons texture analyses have, for instance, been undertaken on chalcolithic copper axes [14], ancient bronze objects [15, 16], and historic silver coins [17, 18]. Here we present snap-shots of further neutron diffraction studies performed on the ROTAX diffractometer at ISIS, including tin-lead spoons from Amsterdam, bronze reference samples, and 16th-century Ag/Cu coins. The selected examples are intended to highlight a few characteristics of TOF neutron diffraction. Full analysis details and results of the studies are or will be given elsewhere.
2 Experimental and methodical aspects of TOF neutron diffraction

Instruments are equipped with several detector banks, viewing the object from different angles. Each detector yields a full diffraction pattern (counts versus crystallographic d -spacings). Generally, all patterns are used for data analysis, resulting in a robust evaluation of phase contents and structure parameters. Large detector coverage is most useful for texture analysis. The detectors at backscattering angles (diffraction angle > 90 deg) have a special relevance in TOF neutron diffraction because diffraction patterns of bulky samples and of objects with irregular shape can be easily collected. Actually, sharp diffraction peaks are mostly measured at backscattering angles, to a large extent independent of the sample thickness. This special feature of TOF diffraction is of advantage for multiphase analyses and for studying fabrication related peak broadening effects. ROTAX is one of the TOF diffractometers at ISIS. It is equipped with three detector banks positioned at forward and backscattering angles, thus yielding three diffraction patterns for each measurement. It has been demonstrated before on ROTAX that TOF-ND data collected from a large ceramic object and from a powder sample, bathing in the beam, basically yield the same phase fractions [10]. Data collection times are in the order of minutes (typically metal objects) up to hours (typically ceramics) depending on the size and complexity of the artefacts. The diffraction patterns are analysed with public-domain program suites such as GSAS [21] and MAUD [22]. TOF neutron data from multi-bank instruments are also ideally suited for a combined quantitative phase, structure, microstructure and texture analysis. The data treatment follows well-established procedures such as quantitative Rietveld analysis for phase identication, lattice constant determination and quantitative assessment of the materials composition as well as texture analysis. TOF neutron diffraction is particularly exible and versatile for studying the crystallographic texture of a material. Here texture stands for the lattice preferred orientation (or preferred crystallographic orientation) that refers to the orientation of the crystal lattice [13]. The texture of a polycrystalline material may result from plastic deformations (mechanical or geological), specic mechanical working processes or thermal treatments (e.g., triggering recrystalliza-

The set-up of a TOF neutron diffractometer at a neutron spallation source as well as its potential use for archaeological research, have been discussed earlier [10, 12, 20]. We, therefore, just summarize some main experimental features: A TOF diffractometer works with a white neutron beam containing a broad band of neutron wavelengths. The beam size is typically between 5 5 up to 20 20 mm2 in cross section, thus illuminating a rather large volume. The only preparation for the analysis required is to produce a safe object holder. Specimens are, if necessary, wrapped in aluminium foil to record aluminium Bragg peaks as calibration markers. Objects can, if required for protection from soiling or scratching, also be wrapped in a thin layer of paper tissue.

KOCKELMANN et al.

Applications of TOF neutron diffraction in archaeometry

177

tion) during manufacturing. Many working processes (casting, rolling, hammering, heating) cause distinct grain orientation distributions. Diffraction provides an elegant approach to determine them, and to displayed them as pole gures (pole: the normal direction to a lattice plane). If all crystallite orientations are equally realized corresponding to a random distribution of poles in the pole gure, then the material is said to be texture-free. In the following, we summarize a few details of how to read texture maps: Each single pole gure represents the orientation distribution of a specic lattice plane, e.g., (200), with respect to the sample body. Each particular point in a pole gure denotes a specic direction in the sample. For the at samples, bronzes and coins in the gures below, the centre of the pole gure corresponds to the normal direction, i.e., the direction perpendicular to the slab/coin surface. Preferred alignments of lattice planes shows up as maxima. Different colour/grey shadings refer to different pole densities given as multiples of the random distribution (mrd). Pole gures are normalized, with mrd = 1 marking the average distribution. Consequently, for a completely random distribution, the density would be 1.0 mrd everywhere, i.e., pole gures would be of single-colour or grey shade. For reasons of clarity, pole densities are often plotted only for mrd-values greater than one. This is done in all illustrations of this paper. The maximum mrd-values indicate the texture strength. Several pole gures are required to fully describe the texture of a phase. Distributions of higher order planes, e.g., (100), (200) and so on, are obviously identical. Moreover, the orientation distributions of the lattice planes are correlated, as given by the crystal symmetry. For a fcc metal, a preferred alignment of (220) poles (see maximum in 220 pole gures below) along the normal direction imposes preferred alignments of equivalent (202) and (022) planes at angles of 60 degrees (pole density ring in pole gures below) and 90 degrees (less pronounced and hardly visible on the circumference of the pole gure), respectively. The pole strength and symmetry of a pole gure reects the deformation strength and symmetry of a working or deformation process. For example, rotation-symmetric pole densities indicate a single working direction, e.g., the hammering direction. Neutron texture analysis is an elegant method to measure pole gures, i.e., to count how many crystallites are oriented in a certain direction with respect to the sample body and its geometry. Due to the high penetration capabilities of neutrons, particularly large sample volumes of coarse grain materials, can be investigated. Pole gures are measured by recording diffraction patterns from many different angles (1) by rotating the sample, or (2) or by moving the sample or building a detector around the sample. The rst option (1) is the conventional texture set-up where a sample is mounted on a goniometer in order to realize a multitude of orientations. The second approach (2) is realized with instruments like GEM [23] and HIPPO [24], characterised by a large detector coverage. For both experimental options, Bragg peak intensities are recorded as a function of the sample orientation angle, and subsequently transferred into pole gures. In

fact, pole gures obtained from diffraction represent maps of the Bragg intensities, and as such, maps of the crystallite orientations. The orientation distribution function (ODF) is reconstructed, even from incomplete experimental pole gures, using the WIMV [25] algorithm by means of a program like MAUD [22]. We have used the ROTAX diffractometer to measure pole gures of small objects such as coins. For 100 sample orientations this is achieved in a matter of a few hours. On a diffractometer like GEM at ISIS with its large number of detectors around the sample position, a high orientation coverage is already achieved with just one sample orientation, i.e., an object needs only to be positioned inside the sample chamber and no sample rotations are required at all. The texture analysis of a bulky and uneven object can be achieved in a single shot in a matter of a few minutes. In the experimental context, it should be noted that objects irradiated with neutrons become temporarily radio-activated. The rate at which the activation decays depends on the element composition of the material, on the time of exposure, and the neutron ux. For ceramics, for instance, activation decay times are usually in the order of minutes. For some metals like copper, the activity monitored after completion of a 10 hours diffraction experiment typically drops below the detection limit of 0.1 Sv/hour after 2-3 days. Some elements such as iron are not activated in the neutron beam, others like cobalt may emit radiation for months, depending on the neutron load.
3 3.1 Results and discussion Characterisation of tin-lead spoons from Amsterdam

In a recent study, we have carried out TOF neutron diffraction analyses of 25 tin-lead spoons from Amsterdam from the period 13501775. Figure 1a shows one of the 14th-century objects. Within the town boundaries, the department of Archaeology of Amsterdam excavated over time more than 800 complete tin spoons and numerous incomplete parts. Many of these objects have been well preserved due to the peaty soil found in the Amsterdam area. The spoons are roughly dated by hammer and rose-marks, form and decoration. Iron core reinforced spoons were produced during 13501500 AD. For many of the excavated spoons the patina was removed over an area between 2 5 mm2 for investigation by X-ray uorescence in order to determine their metal content. The patina itself was investigated by X-ray diffraction revealing metal sulphides (CuS2 , FeS2 ) as black corrosion layers, and chalcopyrite (CuFeS2 ) and pyrrhotite (FeS) as golden corrosion layers. A large variation in Pb content has been found by XRF, the early spoons (13501450) may contain up to 40% 50% Pb. Due to the recognized health hazards of Pb, the production of tin spoons was regulated and overseen by the Guild after 1530. From then on, only a low Pb content ( 5% 10%) was allowed from then on. However, the X-ray data seem to suggest that this control was more or less lost after 1600. Figure 1b shows a snap-shot of a tomograph on spoon NDK-340 providing insight into the placement and possible manifacture of the iron reinforcement core. The tomography also shows the distribution of a corrosion lump at the forked

178

Applied Physics A Materials Science & Processing


FIGURE 1 (a) Photo of a 14th-century tin-lead spoon (NDK-340) from Amsterdam. (b) Neutron tomography from the NEUTROGRAPH station at the ILL, Grenoble, reveals the iron reinforcement core and the distribution of a corrosion extrusion. (ce) TOF neutron diffraction patterns from the scoop, the corrosion lump, and the handle. The neutron beam size was 20 20 mm2 for scoop and handle, and 10 10 mm2 for the corrosion extrusion, respectively. The arrows in (d) mark the Bragg peaks of vivianite, Fe3 (PO4 )2 8H2 O

end of the iron bar. Radiography however does not give direct evidence of the crystal and microstructures of the phases. Figure 1ce show diffraction patterns collected on the scoop, the corrosion extrusion, and the handle. SnPb is an eutectic system, hence we observe separately a tetragonal (Sn) phase and a cubic (Pb) phase in the diffraction patterns; the Bragg peak positions are found to coincide with lattice spacings of the pure metals, Sn and Pb. The scoop exhibits 37.3 wt. % (Sn) and 62.7 wt. % (Pb), ratio Pb/Sn = 1.68. The handle exhibits, in addition to the Sn and Pb phases (26 wt. %, 42 wt. %, ratio Pb/Sn = 1.62), strong iron Bragg peaks (31.5 wt. %) and 0.5 wt. % of a SnO corrosion phase (tetragonal crystal symmetry). It is worth mentioning that the Bragg peaks of the bcc Fe are composed of nuclear and magnetic diffraction from the Fe nuclei and magnetic Fe iron moments, respectively. The extrusion at the end of the iron bar exhibits (in wt. %): 28 (Sn), 59 (Pb), 12 Fe as metal phases. Yet the diffraction data from the corrosion extrusion reveal further Bragg peaks (Fig. 1d) from about 1 wt. % of vivianite, an iron phosphate hydrate (hence, the sum of phase fractions is 100%). Figure 2 shows the Pb/Sn ratios plotted versus the production dates of some of the spoons, as determined by the archaeological classication. The minimum in the Pb/Sn ratio marks roughly the regulation of the lead content by the Guild. The encircled values belong to spoon artefacts that are characterized by an additional SnPb phase which is related to

FIGURE 2

Pb/Sn ratios versus archaeological datings for several medieval Amsterdam SnPb spoons. The minimum in the curve approximately corresponds to the regulation year of the Pb content by the Guild. The encircled points belong to English spoons which contain an additional SbPb phase with the approximate composition of the eutectic. The apparently linear trend of the Pb/Sn ratio with the archaeological dating of the three spoons before 1500 is rather accidental

the eutectic Sn0.6 Pb0.4 composition. This extra SnPb phase indicates a different manufacturing cycle which probably involved rapid quenching of the spoons. This case study on medieval spoons illustrates that neutron techniqes can very well add extra information that is difcult or impossible to obtain otherwise. While the XRF analyses required removal of the patina and were restricted to a few

KOCKELMANN et al.

Applications of TOF neutron diffraction in archaeometry

179

analysis spots. Neutron diffraction allows for performing nondestructive quantitative analyses of the metal-alloy phases in all parts of the spoons. The results will be elsewhere presented in full detail.
3.2 Texture analysis of bronze reference samples

Considering texture analysis, a relevant amount of technical renement and interpretation work is still required on reference materials. In this context, a systematic measurement campaign on standards was aimed at accompanying neutron diffraction characterisations of archaeological bronze artefacts. To this purpose, the synthetic samples are used to derive reference behaviours associated with different alloy compositions and working treatments. Here we report on neutron analyses on some of the standards in order to illustrate the implications of non-destructive texture analysis in archaeometallurgy. The reference samples (planar slabs) were cast in a cast iron ingot mould, then milled to a standard thickness of 7 mm in order to remove any irregularities and other boundary effects. Subsets from A4 (Cu 96%, Sn 4%) and A5 (Cu 94%, Sn 6%) were produced and cut into pieces of 4 20 30 mm to investigate the effect on texture produced by hardening and annealing cycles. The samples A41 , A42 , A43 , A44 , taken from the raw casting were hammered up to a thickness reduction of 7.9%, 19.3%, 30.7%, and 45.7%, respectively. A5 was left as-cast. Figure 3 shows the pole gures of the bronze specimens A5, A41, A42, and A44, measured on ROTAX. These texture maps can be regarded as signatures of the past treatment, and the absence of any preferred orientation can represent an important piece of information. The texture maps of specimens A41 A44 bear signs of both the working strength and working direction. For specimen A5 irregular, random grain distributions are found; the remaining maxima are to be interpreted to originate from some random poles associated with a few large crystallites. The hammered specimens A4x are characterised by the alignment of the (110) poles normal to the specimen surface. In other words, there is a statistical predominance of (110) planes that are parallel to the specimen surface. Rotation symmetry associated with hammering and multiplicity of the reported plane sets allow for the interpretation of the displayed pole gures. They represent a bre texture given by a superposition of Brass {110} 112 and Goss {110} 001 components driven by the dominant slip system of the fcc lattice, {111} 110 . The sequence of Fig. 3 provides a quantitative evaluation of the dependence of the preferred orientation on the degree of hardening. The largest pole density was 2.16 mrd, which in general represents a moderate degree of texture compared to strong texture cases, where a maximum pole density can be in the order of 10 mrd. A signicant variation by a factor of about two was observed in the texture parameter upon stepping from 7.9% to 45.7% degree of hardening. This result can be of practical importance in associating texture strength to the hardening range of artefact zones investigated.
3.3 Texture analysis of Ag/Cu coins

FIGURE 3 Pole gures of bronze standards. From top to bottom: irregular pole densities of the as-cast sample A5 ; regular axisymmetric pole densities of samples A41 , A42 , and A44 (hammered up to a thickness reduction of 7.9%, 19.3% and 45.7%, respectively). The centre of each pole gure corresponds to the normal direction to the slab-shaped samples. The maxima in the (220) A4x pole gures indicate alignment of the (110) poles parallel to the normal direction, i.e. preferred alignment of (110) planes parallel to the surface of the slabs

Ten 16th-century silver Taler from the Mnzkabinett of the Kunsthistorisches Museum Vienna were examined

by quantitative phase and texture analyses in order to distinguish between minted and cast copies on the basis of the microstructures of the silver/copper phases. Usually genuine coins are minted and should therefore display characteristic crystallographic textures whereas forgeries are often cast. The genuine coins were issued by Ferdinand I (15191564) and Archduke Ferdinand (15641595), respectively. Some of them are expected to conform to one of the Mnzordnungen (MO), e.g., Hall 1524, Hall 1566, MO Hall 1577, prescribing an Ag nenesse of approx. 90 wt. %. In the 16th century up to 1566, the silver coins were struck with a hammer and a die before a new revolutionary minting technique by rolling was introduced that allowed mass production of coins. The examined coins all have a silvery appearance, diameters of approximately 40 mm and thicknesses of about 3 mm. Two ancient Greek Tetradrachms and several replica coins were also measured for comparison. The replicas consisted of Ag, Cu or Ag/Cu and were produced by imitating different minting techniques. Figure 4 compares neutron diffraction patterns of several Taler coins. Two crystalline fcc phases, and , constitute the AgCu eutectic system with coexisting (Ag) and (Cu) phases. The former is an Ag rich solid solution, which con-

180

Applied Physics A Materials Science & Processing

tains some Cu as solute component for non-equilibrium cooling. Conversely, (Cu) represents the solvent of the phase. As a consequence of the different lattice parameters associated with the two fcc phases, the neutron pattern exhibits separate (Ag) and (Cu) peaks. Peak intensities depend on the relative weight fractions and preferred orientations. Inv.196.732 is a pure Ag coin, two coins (Inv.205.437 and 205.492) contain an extra cuprite, Cu2 O, corrosion phase. The Rietveld analysis provides the phase fractions as well as the structural details that would be difcult to obtain by other methods. Figure 5a shows an example of a Rietveld tted pattern. The quantitative phase analysis of a texturised material is not trivial and can reliably only be performed by collecting and averaging data over many different sample orientations. Even better, the phase analysis can be carried out in combination

FIGURE 4 Neutron diffraction patterns of 16th-century silver coins. The phase contents as determined from neutron diffraction are (in wt%) (from bottom to top): Inv.205.437 (11.9 Ag, 87.5 Cu, 0.6 Cu2 O), Inv.196.732 (100 Ag), Inv. 196.733 (3.0 Ag, 97.0 Cu), Inv. 161.090 (93.8 Ag, 6.2 Cu), Inv. 205.491 (89.5 Ag, 10.1 Cu). The hkl-labels in the upper-left corner of each pole gure correspond to the Bragg peaks in the diffraction pattern. Maximum multiples of a random distribution (mrd) are given in the lower-left corner of each pole gure

FIGURE 5 Rietveld tted diffraction pattern (a) and pole gures (b) of a 16th-century silver Taler (c) (Kunsthistorisches Museum Vienna, Inv.161.090). The (111), (200) and (220) labels for Ag and Cu phases refer to the Bragg indices in the diffraction pattern

KOCKELMANN et al.

Applications of TOF neutron diffraction in archaeometry

181

with a full texture analysis, as it was done here. Three coins (Inv.196.676, 205.437, and 205.492) exhibit high Cu contents of 23.1, 87.5, and 84.3 wt. % respectively. Three coins (171657, 171675, and 196733) contain about 3 wt. % Cu. Two other coins (Inv.205.491 and 205.518) show a Ag/Cu wt. % ratio of 90/10 complying with the values for MO Hall from 1577. Taler Inv.161.060 has 6 wt. % Cu in agreement with MO Joachimstal 1528. Inv.196.732 is a pure silver coin. The structure analysis involves the renement of the lattice parameters of the two main phases, (Ag) and (Cu). For most objects, the lattice constants are slightly different compared from those of the pure metals (Ag: a = 4.085 , Cu: a = 3.6145 ), indicating deviations from the equilibrium constitutions of AgCu. For Ag/Cu mixtures quenched at 500 C, it is reported [26] that the Ag-lattice can incorporate Cu up to 3 mol %, which reduces the lattice parameter to 4.073 . Likewise, the Cu-lattice can accommodate Ag up to 1 mol %, which increases the lattice parameter to 3.6188 . We observe such lattice contractions and expansions on this scale for the AgCu Taler coins. For Inv.196.676, the lattice constant of the Ag phase is slightly lower than that of pure silver indicating the incorporation of Cu atoms, along the lines of the quenching experiments [26]. The large lattice parameter of (Cu) 3.6344 , however, can not be explained by incorporation of Ag alone, it requires an additional alloying element, for example tin in the order of 23 wt. %. Figure 5b shows the texture maps of coin Inv.161.090. The grain orientations are represented by six pole gures, 3 for the Ag phase, and 3 the Cu phase, corresponding to the Bragg peaks marked in Fig. 5a, i.e. the textures of all phases are obtained simultaneously by diffraction analysis. It is worth noting, that the Ag and Cu pole gures in Fig. 5b show the same type of texture, even the same texture strength as expressed by the maximum mrd-values. This means that Ag and Cu have been subjected to the same working treatment. This is not necessarily the case, for example the silver coating and the copper core of coins may exhibit different types of textures. Here, both phases exhibit the typical fcc (110) bre texture as observed for the bronze samples (Fig. 3). Since the pole density symmetry reects the symmetry of the deformation process, this coin was clearly minted by a uniaxial compression, e.g., by striking with a hammer and a dye. Pole gures of the majority phase (usually Ag, Cu in case of Inv.205.437) of different coins are compared in Fig. 6. Inv.205.491 (and Inv.205.518) show the hallmarks of a rolling texture in agreement with the production of the Ferdinandtaler in the second half of the 16th century, when coins were minted by cold-rolling a Ag/Cu sheet. The pole gures are not rotation-symmetric; the rolling direction is perpendicular to the pole density stripes. Inv.161.090 exhibits the (110) bre texture reecting the deformation symmetry by hammer striking as used for minting in the rst half of the 16th century. Three coins with 3 wt. % Cu (Inv. 171.657, 171.675, and 196.733) exhibit the same texture in terms of the (110) bre texture and of the texture strength. The pole densities are, however, signicantly smaller than for Inv.161.090. Hence the minting technique was the same but the degree of reduction was clearly different. The crystallite distribution of the pure silver coin, Inv.196.732, is rather indifferent and irregular, although some features of a bre texture seem to be

FIGURE 6 (111), (200), and (220) pole gures of the majority (Ag or Cu) phases in 16th-century Taler coins. The type of the texture (characterised by streaks, rings, or irregular pole densities) and the strength of the texture (maximum mrd) point to different manufacturing procedures

present. Inv.196.676, Inv.205.437, and Inv. 205.492 clearly show irregular grain orientation distributions, typical of cast materials. In summary, the texture analysis reveals different minting techniques for the Talers coins. Texture maps that point to specic minting processes and Ag/Cu contents in agreement with the minting regulations strongly indicate genuine objects. Coins with irregular grain distributions are more likely to be fakes. It is interesting to note that neither the hammered replica coins nor the ancient Tetradrachms exhibit the bre textures observed for some of the Ferdi-

182

Applied Physics A Materials Science & Processing


support during the neutron diffraction experiments at ISIS by J.W. Dreyer, C.M. Goodway and A. Church is gratefully acknowledged.

nandtaler. Also these results will be presented in full detail elsewhere.


4 Future outlook

REFERENCES
1 M.D. Glascock, H. Neff, Meas. Sci. Technol. 14, 1516 (2003) 2 A. Kalicki, E. Paczyk, L. Rowinska, B. Sartowska, L. Walis, K. Pytel, Radiat. Meas. 34, 557 (2001) 3 Z. R vay, T. Belgya, Principles of the PGAA method. In Handbook e of Prompt Gamma Activation Analysis with Neutron Beams, ed. by G.L. Moln r (Kluwer Academic Publishers, Dordrecht, The Nethera lands, 2004) pp. 130 4 H. Postma, P. Schillebeeckx, R.B. Halbertsma, Archaeometry 46, 635 (2004) 5 E. Deschler-Erb, E.H. Lehmann, L. Pernet, P. Vontobel, S. Hartmann, Archaeometry 46, 647 (2004) 6 T. Materna, S. Baechler, J. Jolie, B. Masschaele, M. Dierick, N. Kardjilov, Nucl. Instrum. Methods A 525, 69 (2004) 7 N. Kardjilov, S. Baechler, M. Bastrk, M. Dierick, J. Jolie, E. Lehmann, T. Materna, B. Schillinger, P. Vontobel, Nucl. Instrum. Methods A 501, 536 (2003) 8 W. Treimer, M. Strobl, A. Hilger, C. Seifert, U. Feye-Treimer, Appl. Phys. Lett. 83, 398 (2003) 9 J.R. Santisteban, L. Edwards, M.E. Fizpatrick, A. Steurer, P.J. Withers, Appl. Phys. A 74, S1433 (2002) 10 W. Kockelmann, A. Kirfel, E. Hhnel, J. Archaeol. Sci. 28, 213 (2001) 11 S. Siano, L. Bartoli, M. Zoppi, W. Kockelmann, M. Daymond, J.A. Dann, M.G. Garagnani, M. Miccio, Proc. Archaeometall. Eur. 2, 319 (2003) 12 W. Kockelmann, S. Siano, M. Schreiner, Studio e Conservazione di Manufatti Archeologico (Nardine Editore, Firenze, 2004) pp. 431451 13 H.-R. Wenk, P. Van Houtte, Rep. Prog. Phys. 67, 1367 (2004) 14 G. Artioli, M. Dugnani, I. Angelini, L. Lutterotti, A. Pedrotti, A. Fleckinger, Proc. Archaeometall. Eur. 2, 19 (2003) 15 E. Pantos, W. Kockelmann, L.C. Chapon, L. Lutterotti, S.L. Bennet, M.J. Tobin, J.F.W. Mosselmans, T. Pradell, N. Salvado, S. But, R. Garner, A.J.N.W. Prag, Nucl. Instrum. Methods B 239, 16 (2005) 16 S. Siano, L. Bartoli, W. Kockelmann, M. Zoppi, M. Miccio, Physica B 350, 123 (2004) 17 W. Kockelmann, A. Kirfel, E. Jansen, R. Linke, M. Schreiner, R. Traum, R. Denk, Proc. Numismatics & Technology, Kunsthistorisches Museum Wien (2003) p. 113 18 Y. Xie, L. Lutterotti, H.-R. Wenk, F. Kovacs, J. Mater. Sci. 39, 3329 (2004) 19 A. Botti, M.A. Ricci, G. DeRossi, W. Kockelmann, A. Sodo, J. Archaeol. Sci. 33, 307 (2006) 20 W. Kockelmann, A. Kirfel, Physica B 350, e581 (2004) 21 A.C. Larson, R.B. VonDreele, GSAS: Generalized Structure and Analysis Software (www.ccp14.ac.uk) 22 L. Lutterotti, MAUD, Material Analysis Using Diffraction (www.ing.unitn.it/luttero/maud/index.html) 23 P. Day, J.E. Enderby, W.G. Williams, L.C. Chapon, A.C. Hannon, P.G. Radaelli, A.K. Soper, Neutron News 15, 19 (2004) 24 H.-R. Wenk, L. Lutterotti, S. Vogel, Nucl. Instrum. Methods A 515, 575 (2003) 25 S. Matthies, G.W. Vinel, Phys. Stat. Solidi B 112, K111 (1982) 26 L. Junqin, Z. Yinghong, L. Weibo, J. Alloys Compd. 191, 187 (1993)

Some of the neutron activation and radiography techniques listed in Table 1 are rather mature and established in archaeological sciences, whereas the potential of neutron diffraction applications has not yet been fully exploited. Meanwhile, advances in diffraction instrumentation at continuous and pulsed neutron sources and new data analysis strategies over the past few years, make it possible to quantitatively characterise the internal microstructure of an archaeological artefact in a relatively short time. Neutron texture studies on engineering and geological samples used to last many hours or even days, and were carried out on cut and extracted engineering and geological samples. Present-day TOF neutron diffractometers provide unique and new possibilities for nondestructive texture studies of bulky and complex objects. The instrumentation progress is particularly related to the increase of detector coverage. Combined phase, structure and texture analyses can be performed on stationary objects within minutes. For instance, on ROTAX the texture analysis of a coin required about 4 5 hours for a total of about 100 sample orientations. Data collection on GEM could be performed in less than one hour. Such rapid texture measurements are of interest not only for achieving a higher throughput of analyses, but also for minimizing the neutron activation of objects. A recent neutron application with high potential is the mapping of strain and texture in archaeological objects [11] based on the so-called Bragg-edge transmission technique [9]. Neutron strain and stress scanners allow for mapping of residual strains, and for scanning of a small diffraction volume ( mm3 ) across the surface or into the depth of an object in order to record position dependent strain elds or phase changes. Radiography and tomography type 2D and 3D spatial imaging techniques are currently under development in order to map also other structural parameters such as phase contents and microstructures, aiming at further enhancing the attraction of TOF neutron diffraction in archaeological sciences.
ACKNOWLEDGEMENTS The authors are grateful to A. Hillenbach, H. Ballhausen and M. Engelhardt for collecting the spoon tomography at the NEUTROGRAPH facility at the Institute Laue Langevin, Grenoble, France. We wish to thank J. Veerkamp and W. Krook, Department of Archaeology, Amsterdam University, for helpful discussion. Technical

Potrebbero piacerti anche