Chemical Analysis of Pottery to Characterize a Workshop Production
Although in many parts of the world the chief evidence of pottery production (workshops and kilns) is extremely infrequent, the reconstruction of pottery production should start with archeological excavations of pottery workshops and kilns (see Ceramics and Pottery). Then the pottery and other appropriate production debris would be subjected to a range of analytical, microscopy, and radiography techniques in order to establish what raw materials were used, how they were prepared, and how the objects were made.
This was the case of the local ceramic production during the Islamic period in the city of Zaragoza (in the northeast of Spain). In 1989, an Islamic potters’ suburb was located and excavated on the eastern side of the ancient city; the kilns were working from the end of the tenth century to the beginning of the twelfth century AD. Thousands of ceramic objects, both glazed and nonglazed, and also potters’ tools were found associated with ten kilns and some waster spots. The study of this archaeological material would allow a better knowledge of the ceramic technology during this period. Therefore a large program on technological features of this production was carried out, considering the different ceramic variants (forms and decorations).
The research started with the study of the material found inside one of the kilns, the most representative of the eleventh century production. Several fragments of ceramic objects, potter tools, and wastes were selected and submitted to chemical analysis. These results were completed with mineralogical and petrographic analyses and textural studies by image processing. From an overall selection of 100 samples, only 56 were chosen for chemical analysis after a preliminary observation of their thin sections (see Pottery Analysis: Petrology and Thin-Section Analysis). Determination of ten major and minor elements (Si, Al, Fe, Mg, Ca, Na, K, Ti, P, Mn) and some trace elements (Ba, Ce, Dy, Eu, La, Li, Nb, Nd, Sc, Sm, Sr, V, Y, Yb, Zr) were carried out by ICP-AES. This analysis was undertaken on representative 500 mg subsamples, drilled from freshly fractured surfaces of the sherds. Fusion was used to prepared sample solutions. Chemical results were submitted to a computer-assisted multivariate statistical treatment by cluster analysis in order to group samples with similar composition.
Chemical analyses showed the samples distributed in two main groups: one made from noncalcareous clays (CaO < 1wt.%) and another prepared from
Calcareous pastes (CaO > 14wt.%). The fragments with noncalcareous bodies belonged to cooking pots, and this relation between chemical composition (low CaO content) and utilization (objects exposed to fire) has been also proved in other ceramic productions as a technological character. Another observed technological division was that all objects covered and decorated with glaze (lead glazes in this case) were manufactured with calcareous clays. This type of clay would have the advantage of producing a buff color in an oxidizing atmosphere, and the ceramic body tends to have higher thermal expansion coefficient with best adherence between body and glaze.
The main calcareous group was formed by three subgroups (see Figure 3), each subgroup was related to different ceramic forms or decorations (tin glazes TG, cuerda seca CS, transparent colored glazes CG, red slips RS, or nondecorated ND). The analytical results indicated that a selection of raw material was used for each production variant, with significant technological differences detected in each group - different proportions of clay matrix and inclusions. Cluster 1 included small glazed objects, in cluster 2 were tin-glazed and decorated-glazed ceramics, while unglazed fragments belonged to cluster 3. Chemical composition of cluster 3 would be the most similar to the original raw clay.
In summary, chemical composition in major and minor elements was applied in order to know the nature of clay raw materials used in these Islamic workshops. The variation in some elements (aluminum, silicon, potassium, or rare-earth elements) also allows knowledge about raw material preparation. The technological features were also supported by petrographic and texture analyses.
The chemical analysis of this pottery production also led to the establishment of the compositional reference of these ceramics to be used in other provenance studies, in major and minor elements and also in trace elements, which can be compared with ceramic fragments found in other archaeological sites of the same historical period. In this way, four average chemical compositions with their respective standard deviations were defined for the four reference groups.
Study of Trade Routes and Cultural Contacts
The transformation of clay raw material into the final fired object is a complex process that could influence the final composition of the pottery. In order to compensate for all factors included in the ceramic process, the normal procedure of provenance studies is to compare the chemical composition of finished pottery bodies not with the ancient raw material source (usually very difficult to identify), but with other fired objects of certain or assumed provenance. Reference groups of known provenance are established from workshop products and kiln wasters, as with the later case described above, or by comparison with fragments of accepted provenance, certified by experts or other archaeological information.
This was the procedure selected to study the possible provenance of Roman glazed ceramics found in the western Mediterranean. Lead-glazed pottery was a tradition that originated in the eastern Roman world, but spread to all the other regions during the first century BC and several possible production centers had already been studied and suggested for the western Roman area, including Italy, Gaul, Britain, Germany, and Spain. Therefore a study was undertaken to shed new light on Roman glazed ceramic production and importation to the western Mediterranean (see Vitreous Materials Analysis).
Unfortunately, Roman glazed ceramic fragments are scarce at archaeological sites, so a very limited number of samples was selected for chemical analysis; 68 objects were chosen from different sites, chronologies, glaze colors, and potential origins. The selection was thought to be representative of the glazed ceramic found in the Ebro Valley (northeastern Spain). Laboratory subsamples to be analyzed were extracted with a diamond drill from the ceramic body core to avoid contamination with material from the lead glaze. Solutions were prepared from the powdered samples (50 mg) by acid attack in open beakers. Determination of eight major and minor elements (Na, Mg, Al, K, Ca, Ti, Mn, and Fe) was carried out by ICP-AES. A more exhaustive examination of the results needed a statistical treatment of the data, so HCA was used to classify fragments according to their chemical composition (see Figure 4).
In this case, every group is defined by an average chemical composition and these compositions should be compared with other reference groups of known provenance. At that moment, only two references had been published on body analyses of Roman lead-glazed ceramics with provenance established by experts and archaeological evidences; this is why only major and minor elements were chosen to be analyzed and not trace elements. Although inherently less powerful than the trace element analysis for provenance studies, this approach through major and minor element determination gives more insight into the nature of the clays used.
A statistical comparison (t test or Student’s test) of the different chemical compositions allowed to attribute some of the groups to production centers in Italy (second and third centuries AD) and Gaul (Lyon area, first century AD); however, other centers in the eastern Mediterranean and central Gaul were excluded.