Archaeological Chemistry - ACS Publications - American Chemical

---

Chapter 7

Electron Microprobe and Neutron Activation Analysis of Gold Artifacts from a 1000 A.D. Peruvian Gravesite 1

2

3

Downloaded by PENNSYLVANIA STATE UNIV on June 28, 2012 | http://pubs.acs.org Publication Date: May 5, 1996 | doi: 10.1021/bk-1996-0625.ch007

Adon A. Gordus , Carl E. Henderson , and Izumi Shimada 1

2

Department of Chemistry and Department of Geology, University of Michigan, Ann Arbor, MI 48109 3Department of Anthropology, Southern Illinois University, Carbondale, IL 62901 The gold-silver-copper ternary alloy contents of gold objects from an unlooted Middle Sicán (ca. 1000 AD) burial site on the north coast of Peru were determined by instrumental neutron activation analysis (INAA) using 10 μg metal rubbings from the interior of the objects. Electron microprobe analysis (EMPA) was also performed on a few of the objects allowed to be exported from Peru in order to measure the interior and surface alloy contents and to determine if the INAA samples are representative of the interior alloy. The near-perfect agreement between the probe data for the interior metal and INAA sample data confirms that the INAA sampling method results in metal characteristic of the interior alloy. The thin outer edges, which are only about five μm thick, are depleted markedly in copper (18 wt% interior copper, 7% edge copper, for example) and suggest that, for the high-quality gold objects (typical contents: 45 wt% Au, 37% Ag, 18% Cu), surface depletion of copper (and enrichment of gold and silver) may not have been deliberately induced but could have been the result of the repeated hammering-annealing required to produce the thin gold sheets used for the construction of the objects. Lower quality tumbaga gold sheets (typical interior contents: 25 wt% Au, 45% Ag, 30% Cu) also show marked surface depletion of copper.

The Burial Site The northern coast of Peru was one of the primary centers of prehispanic metallurgy in the New World and, from the first millennium BC until the fourteenth century, The successive cultures inhabiting this region developed copper- and gold-based metallurgical expertise that was unmatched in innovative quality, sophistication, and overall production (7, 2). The Sicân Archaeological Project, under the direction of I. Shimada, is an interdisciplinary effort with some of the research directed toward an understanding of the technology and organization of metallurgical production and the role of the metal objects of the Middle Sicân culture that flourished on the northern coast of Peru during the 10th to 12th centuries AD. 0097H5156/96/0625-0083$12.00/0 © 19% American Chemical Society

In Archaeological Chemistry; Orna, M.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

Downloaded by PENNSYLVANIA STATE UNIV on June 28, 2012 | http://pubs.acs.org Publication Date: May 5, 1996 | doi: 10.1021/bk-1996-0625.ch007

84

ARCHAEOLOGICAL CHEMISTRY

The role of arsenical copper was first examined and excavations at various production centers in the Batân Grande region of Peru provided information on the technologies and uses of the arsenical copper products (3, 4). However, it is in the deep (ca. 11 meter) underground tombs of the Middle Sicân elite that unprecedented accumulations of precious metal objects were found (5) and between the 1930s to 1970s many of these tombs had been looted. Although many of the gold objects from these looted tombs found their way into private and public collections throughout the world, information about the specific gravesites, inventories of the sites, and other important contextual data are not available. Only a controlled archaeological excavation of an unlooted gravesite would provide the needed ancillary information. In 1991-1992 Shimada excavated an intact shaft-tomb of a Middle Sicân nobleman at the base of an adobe-brick pyramid of Huaca Loro in the capital city of Sicân in the Batân Grande region of Peru. This excavation represents the first scientific documentation of such an elite tomb (6). The site contained about 1.2 tons of diverse grave goods, over 75% consisting of arsenical bronze and precious metal objects and scrap. Systematic examination and conservation of the grave contents was done by Shimada and coworkers in the Museo de la Nation in lima where the objects will be on permanent display. The placement of the objects in the tomb and the positions of the main and sacrificial burials are illustrated elsewhere (7). Photographs of some of the objects are given in reference 6 and color photographs are given in reference 8. The Gold Objects About a hundred major gold objects (most constructed of separate pieces of gold) were contained in the tomb and included a gold mask with distinctive upturned eyes (which is characteristic of the Sicân style), a gold head ornament with an animal head, ceremonial tumi knives, crowns, various ornaments with attached bangles, dart throwers, ornate ear spools, long flat feathers, rattles, a pair of meter-long gold gloves, and piles of gold scraps and foils, including almost 2000 (ca. 1.4 cm square) pieces that had originally been sewn to a cloth mande (now disintegrated) that had been laid beneath the principal burial. Although all of the high-quality gold objects are very thin (usually only 0.1-0.5 mm thick) they are structurally sound due to the induced hardness which resulted from the repeated hammering that was required to produce the thin metal sheets used for fabrication of the final objects. The majority of the high-quality objects had interior gold contents in the range of about 40-75 wt% gold (7), the remainder being almost entirely copper and silver. Most of the gold scraps and foils are extremely thin (0.05 mm) and, hence, more flexible and fragile. As noted below, some of the scrap gold foil-fragments we analyzed have much lower gold contents in the range of 13-35 wt% gold and tend to be much more subject to corrosion. Surface Depletion. The composition and method of fabrication of the gold objects is of interest Based on the study of isolated pieces, it has been suggested that methods of deliberately removing copper and silver from the surface of ternary Au-Ag-Cu alloy objects were developed and used by some of the pre-Hispanic metalsmiths (977). (This process has sometimes been called depletion gilding, but modern goldsmiths limit the use of the word gilding to surface-additive, not subtractive processes.) Deliberate, induced surface depletion of silver and copper would allow the creation of an object with a surface of very high quality gold only a few um thick, even though the bulk alloy starting material was of low gold content. Some of the analytical data presented in this study address directly the question of deliberate (as compared with unintentional) surface depletion of the Sicân gold objects.

In Archaeological Chemistry; Orna, M.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

7.

GORDUSETAL.

Gold Artifacts from a 1000 AD. Peruvian Gravesite

Methods of Analysis Two different methods of metal analysis were used. The first was based on instrumental neutron activation analysis (INAA) and involved taking very tiny (ca. 10 μg) metal rubbings of the interior alloy of the gold objects. The edge to be sampled was usually first stroked a fewtimesusing fine-grain emery paper in order to remove any surface corrosion products. This removed surface metal and oxides down to a depth of at least 20 μπι. Then, a small (ca. 6 mm long, 2.5 mm diameter) piece of roughened high-purity silica tubing was rubbed against this cleaned area (ca. 2-4 mm ), transferring about 10 μg of metal to the silica. A second sample was then taken from the same position and each sample packaged in a separate small polyethylene snap-cap vial for later irradiation and analysis. This method allowed the sampling of almost 400 separate gold pieces in the Museo de la Nation in Lima, transporting the rubbings to Ann Arbor, and irradiating the samples for 2.0 hours at a neutron flux of 3 Χ 10 neutrons χ cm* χ sec" in The University of Michigan nuclear reactor. The second method involved electron microprobe analysis (EMPA) in Ann Arbor of fragments from a few of the gold objects that were allowed to be exported from Peru. The INAA method has the advantage that sampling can be done in a museum and, if necessary, even while an object is mounted in a display case. The amount of metal required for each sample is so small (0.01 mg) that it is almost impossible to locate visually the place on the object where the metal rubbing was taken. Sampling can be done quickly; for example, 760 metal rubbings from 380 gold pieces were obtained during a three-day visit to Lima. A few hundred samples can be irradiated together with metal standards in the nuclear reactor. The activation analyses can be performed instrumentally using automatic sample changers followed by computer analysis of the resultant radioactivity spectra so that up to 100 samples can be analyzed in a day. The INAA method has the disadvantage that only the overall gross metal content of each object is measured. The INAA sampling and analysis method is described in detail in reference 7. Since there can always be a question that the INAA samples are truly representative of the interior composition, it is important to measure directly the interior composition by EMPA. The near-perfect agreement in the data presented in this paper on the comparison between the INAA and the EMPA (interior) measured contents substantiates the conclusion that the INAA samples are representative of the interior composition. The EMPA method has the advantage that very small areas of an artifact can be selectively analyzed to determine variations in the alloy content of a single object This is especially important when determining the differences, if any, between the surface-edge and interior metal alloy. It has the disadvantage that the analysis procedure is more complicated, that it requires mounting a fragment of the metal object that weighs at least a few mg and more commonly 10-20 mg, and that it requires the loan of objects or the release of metal fragments. Only the few objects described in this paper, which were allowed to be exportedfromPeru, were analyzed by EMPA. Small (ca. 2-7 mg) samples were mounted on edge, polished, coated with a very thin layer of electrically conducting carbon, and examined using a Cameca MBX microbeam electron microprobe analyzer. The electron beam was operated with a 15 kV accelerating voltage and a 20 mA beam current. Pure element standards were used to calibrate Au, Ag, and Cu. Compositional data were obtained for various areas of the thin surface edge and of the interior metal of each object. Areas ranging from a 1.0 um spot to a 15 X 15 μτη area were selected. The data for the 1 μπι spot allowed, if desired, determination of the composition of selected, individual metal phases,

Downloaded by PENNSYLVANIA STATE UNIV on June 28, 2012 | http://pubs.acs.org Publication Date: May 5, 1996 | doi: 10.1021/bk-1996-0625.ch007

2

13

2

1

In Archaeological Chemistry; Orna, M.; ACS Symposium Series; American Chemical Society: Washington, DC, 1996.

85

Downloaded by PENNSYLVANIA STATE UNIV on June 28, 2012 | http://pubs.acs.org Publication Date: May 5, 1996 | doi: 10.1021/bk-1996-0625.ch007

ARCHAEOLOGICAL CHEMISTRY

Gold

0 c,.,

ν*

S l l v e r

Ο

Ο

Ο

Ο

'