Bed structure - Tin Bronze - Roman Times - Switzerland

Marianne. Senn (EMPA, Dübendorf, Zurich, Switzerland) & Christian. Degrigny (HE-Arc CR, Neuchâtel, Neuchâtel, Switzerland)

Stratigraphic representation: none

Fig. 4: Stratigraphic representation of the object in cross-section using the MiCorr application. This representation can be compared to Fig. 8.

Analyses performed:
Metallography (etched with ferric chloride reagent), Vickers hardness testing, SEM/EDX, XRD, Raman spectroscopy.

The remaining metal is a porous tin bronze (Table 1) with tiny iron oxides inclusions (Table 2) and secondary re-deposited copper. Intergranular corrosion has developed locally inside the metal (Figs. 4 and 5). The etched alloy shows a structure principally consisting of large polygonal grains, some of which include strain lines (Fig. 6). Locally an isolated alpha + delta eutectoid can be seen. Much smaller twinned grains can be found on the left side of the sample (Fig. 5). The average hardness of the metal is HV1 100. However on the left side of the sample, where the grains are smaller and twinned, the hardness is much higher (more cold work followed by annealing).


Elements Cu Sn Fe
mass% 82.0 13.0 ?

Table 1: Chemical composition of the metal. Method of analysis: SEM/EDX, Lab of Electronic Microscopy and microanalysis, IMA (Néode), HEI Arc.



O Mn Fe Cu Sn Total
Inclusions 23 0.7 64 8.7 1.8 99

Table 2: Chemical composition (mass %) of the tiny inclusions on Fig. 4. Method of analysis: SEM/EDX, Laboratory of Analytical Chemistry, Empa.

The metal - corrosion products interface is irregular (Fig. 3) and the average thickness of the corrosion crust is 300μm. The latter can be divided into three layers (Fig. 4). In bright field, the inner layer is light-grey and dense. In polarised light it is dark-red/orange (Fig. 7). It is Sn-rich, contains O and Cl but is Cu depleted (Table 3 and Fig. 8). The outer part of this layer looks slightly lighter in bright field and light-red in polarised light. It has a composition similar to cuprite/Cu2O (Table 3). This was confirmed by Raman spectroscopy (Fig. 9). The second layer is thin, has (in bright field) a colour similar to the first layer (light-grey) and contains black pores. Under polarised light it looks brilliant black (Fig. 7) and is rich in Cu and Sn with O (Fig. 8). Using XRD, it was identified as a mixture of tenorite/CuO, cuprite/Cu2O and cassiterite/SnO2 (Krieg 2009). Raman spectroscopy could only detect tenorite in this layer (Fig. 10). In bright field the thick outer layer is marbled in dark-grey and black, in polarised light it appears blue-green (Fig. 7). XRD analysis gave a composition close to malachite/Cu2CO3(OH)2 (Krieg 2009). In the present study no C was detected (Fig. 9), but Raman spectroscopy confirmed its presence (Fig. 11). The spectrum obtained shows a good match with published spectra (Bouchard and Smith 2003). Peaks are similar to those expected for malachite but most of them have a Raman shift: 3322, 1503, 1108, 1073, 759, 731, 603, 540, 442, 357, 276, 225, 186, 163 cm-1. The underlined Raman shifts correspond to the peaks of greatest intensity.  Environmental elements such as Si are concentrated on the top surface (Fig. 8).



O Cu Sn Si Cl Mn Fe Pb Total
C3i, red outer edge of the inner corrosion layer (average of 2 similar analyses) 19 69 5.1 0.7 < < < < 94
C3i, dark-orange inner corrosion layer (average of 3 similar analyses) 25 52 25 0.9 0.6 < < 1.0 104

Table 3: Chemical composition (mass %) of the corrosion crust from Fig. 8. Method of analysis: SEM/EDX, Laboratory of Analytical Chemistry, Empa.

Corrected stratigraphic representation: none

One side of the tin bronze artefact has been repeatedly annealed after cold working whereas the rest of the metal was cold worked. The metal is covered with a terrestrial type corrosion layer. Cl is found in proximity to the metal in a Sn-rich but Cu-depleted inner area of the first cuprite layer. It is followed by an intermediate thin layer composed of a mixture of tenorite and tin oxides. The third voluminous layer is malachite. These results confirm Krieg’s conclusions (Krieg 2009). Since the bed fragments have been excavated from a burnt down house, it is difficult to judge if the tenorite is an intentional black patina or if it is the result of heating during the fire. The limit of the original surface is outlined by the clear demarcation of Sn content (internal marker) in the first two corrosion layers and can be located between layers 2 & 3. Intentional or no, the black tenorite/cassiterite layer most likely represents the surface that will retain most surface details. In this case, the corrosion type is difficult to attribute after Robbiola et al. 1998.

References on object and sample

Reference object

1. Delbarre-Bärtschi, S., Fischbacher, V., Krieg, M. (2009) Lits en bronze à Avenches: état de la question et pistes de recherche, Bulletin de l’Association Pro Aventico 51, 7-57.
2. Krieg, M. (2009) Conservation-restauration de fragments de cadre d'un lit romain en bronze. Travail de master Filière Conservation-restauration, Haute Ecole Arc de Conservation-restauration, La Chaux-de-Fonds.


Reference sample

3. Krieg, M. (2009) Conservation-restauration de fragments de cadre d'un lit romain en bronze. Travail de master Filière Conservation-restauration, Haute Ecole Arc Conservation-restauration, La Chaux-de-Fonds

References on analytic methods and interpretation

4. Bouchard, M., Smith, D.C. (2003) Catalogue of 45 reference Raman spectra of minerals concerning research in art history or archaeology, especially on corroded metals and coloured glass. Spectrochimica Acta, Part A 59, 2247-2266.
5. Robbiola, L., Blengino, J-M., Fiaud, C. (1998) Morphology and mechanisms of formation of natural patinas on archaeological Cu-Sn alloys, Corrosion Science, 40, 12, 2083-2111.