1The archaeometric study of ancient ceramic materials is often based on the use of different analytical techniques, typical of hard science (e.g. Maritan 2019). Methods that are normally selected depend on both the type of ceramic itself (coarse—versus fine-grained body), and the archaeological issues to be solved (Gliozzo 2020). Results from various types of analysis are often integrated within the same research to reconstruct the specific stages of the ceramic “life”, from the raw materials’ selection, to the production recipes, to its firing, use and alteration during burial (e.g. Ballirano et al. 2014; Bajeot et al. 2020; Botticelli et al. 2020; Maritan et al. 2005; Maritan et al. 2017; Medeghini, Nigro 2017; Medeghini et al. 2019; Russo et al. 2018). Among the numerous works describing the analytical techniques for ancient ceramic analysis, we suggest the detailed and comprehensive Oxford handbook of archaeological ceramics analysis edited by Hunt (2017), and a very recent topical collection of review papers on “Ceramics: Research questions and answers”.1 In this section we will focus on the description of the main analytical techniques (optical microscopy, scanning electron microscopy, X-ray diffraction, chemical analysis and statistical treatment of data) to investigate different aspects of ancient ceramic materials, such as the provenance of the raw material used and production technology.

Optical Microscopy

2Often examination of ceramics begins with a hand-lens (20 × magnification) or a small binocular microscope (up to 100 × magnification). This enables the basic components and appearance of the sherd to be described. For more specific details, a petrographic microscope is employed that examines a thin slice of a sherd on a slide (fig. 9). This provides important information on the mineralogy of the inclusions in the paste, be they natural or added, and clay characteristics (see section ii.3; Ownby, Brand 2019). Such data can be related to geological formations for provenance purposes. Petrography also clarifies technological features such as paste recipe, general firing temperatures, and surface features.

Figure 9. Petrographic images of a Marl C sherd showing paste and scum surface (plane-polarized light, image width is 5 mm), and a glaze ware sherd showing paste and glazed surface (plane-polarised light, image width is 5 mm)

Scanning Electron Microscopy (SEM)

3For questions of firing temperatures, the chemical composition of glazes or particular inclusions, and other features requiring very high magnification, scanning electron microscopy (SEM) is used. This enables both great magnification and also chemical data to characterise features that often relate to particular workshops creating vessels from specific recipes at certain times. Understanding the technology of production at this level increases knowledge on the organisation of ceramic manufacture and distribution. SEM was used to clarify the white surfaces of Marl C pottery from Egypt highlighting the technological reasons these surfaces are present and the deep knowledge potters had of their raw materials (Ownby, Griffiths 2009). Samples for SEM are often prepared by imbedding them in a resin block and finely polishing them, nonetheless whole specimens can also be examined.

4SEM secondary electron imaging (SEM-SEI) is normally performed on freshly-broken pieces of ceramic body to study their microstructure, especially the degree of vitrification to constrain their firing temperature, according to the method proposed by Tite and Maniatis (1975).

5The development of accessible and fast-processing software (information technology) favoured the possibility of processing digital images to quantitatively describe the abundance, grain-size distribution, and shape of textural features (inclusions and voids) of the ceramic materials (fig. 10). This quantitative approach can be performed on photomicrographs, scanning electron microscope images in back-scattered mode, or even on multi-spectral images such as the chemical maps acquired by SEM-EDS, using open access software (Maritan et al. 2020). Possible differences in the quantity of inclusions and/or their grain-size distribution, can be related to different production recipes in diachronic or ceramic function/provenance terms (Dal Sasso et al. 2014; Baklouti et al. 2016).

Figure 10. Example of a scanning electron microscope, back scattered electron (Sem-Bse) image of a ceramic body and of the binary image obtained from the segmentation of its quartz inclusions

X-ray Powder Diffraction

6X-ray powder diffraction (XRPD) is an analytical technique widely used to determine the mineralogical composition of crystalline materials (Maritan et al. 2015). The interaction between the crystalline lattice of the mineral-phases composing the analysed material and the X-ray beam produces (when the Bragg law is satisfied) a diffraction pattern (fig. 11). These peaks can be directly linked, through comparative mineralogical databases (such as the PDF—Powder Diffraction File—produced by the International Centre for Diffraction Data), to specific mineral phases. In order to obtain the mineralogical composition of a ceramic material with such a technique, a representative fragment of the body of a potsherd, after the mechanical removal of possible coatings and altered/dirty external surfaces, needs to be finely grinded (possibly in an agate mortar).

Figure 11. Example of a XRPD pattern (diffactogram) of a ceramic material, with indication of the relative firing temperature, and a table reporting mineral associations as determined by XRPD data

7The mineralogical composition of a ceramic body depends on the mineralogical composition of the raw materials used (of both clay and possibly added temper) and their grain-size, as well as the firing conditions, and in particular, on the maximum temperature reached and firing atmosphere (oxidising versus reducing). Therefore, the mineral associations recognised in a ceramic can be used to characterise its firing conditions, based on comparison to the results of firing experiments on similar raw materials (Daszkiewicz, Maritan 2017). The original mineralogical composition of the ceramic body and its possible coating (slip or glaze) can undergo important changes (precipitation of secondary minerals, transformation of pristine phases into new products, or dissolution of mineral phases) during the post-depositional processes (Maritan 2020 and quoted literature), that can be determined by XRPD analysis as well.

8Other methods can be used to determine the mineralogical composition of ceramic materials, such as micro-diffraction, synchrotron light diffraction, the electron back-scattered diffraction, micro-Raman spectroscopy and infra-red spectroscopy (see Maritan 2019 and quoted literature). Each technique has advantages, such as the possibility of preforming the analysis in a non-destructive way, or requiring very little material, or making the analysis on very restricted areas of the ceramic such as painted decorations or certain specific portions of the body.

Chemical Analyses and Statistical Treatment of Data

9One of the most common analytical methods employed for studying ceramics is to acquire chemical data, typically the full set of elements and their quantity in a powdered sherd. These data are used to assess those ceramics made from similar raw materials and pastes. Often this relates to a common production source enabling the distribution of such pottery to be revealed and those economic factors involved to be understood. Thus, while many vessels can appear similar in form and paste, chemical data will identify those likely to have been produced in the same workshop and/or area. Such research has revealed the organised workshop production of marl pottery in Egypt and clarified fabric groupings (and their relation to each other) for ceramic specialists (Bourriau et al. 2006).

10Chemical analysis is normally performed on fine-grained ceramics, the provenance of which cannot be inferred by petrographic analysis due to the very fine inclusions.

11A number of different instruments can acquire such data. In the past, neutron activation analysis employing a nuclear reactor was a common method for precise and highly sensitive (parts per billion) data acquisition. As reactors are less readily available, other methods have become more popular. Inductively coupled plasma mass spectrometry or atomic emission spectrometry (ICP-MS, ICP-AS), proton induced X-rays (PIXE) and gamma rays (PIGME), and X-ray fluorescence spectrometry (XRF) have also been used (For a case study of the latter method see Ownby 2009). Portable XRF has also become more commonly employed for non-destructive chemical analysis, however, for ceramics the method is less accurate due to material heterogeneity.

12All of these methods produce tables of elements and their values that must be examined statistically. The data are usually normalised first (by base 10 logarithms or other equations) so the variances between common and rare elements do not overly influence the data. Then several statistical methods are used to explore the data and discover those samples with similar compositions (fig. 12). Principal components analysis, correspondence analysis, hierarchical cluster analysis, and discriminant analysis are the most frequently employed (Baxter 2003; Shennan 1997). Use of multiple statistical tests to verify groups of chemically similar pottery is the best practice.

Figure 12. Principal components graph of XRF data from Giza samples

13Overall, these scientific methods have greatly improved our understanding of pottery manufacture, distribution, and use. Such information is vital for clarifying the role of pottery in past societies and providing information on the economic, religious, and social systems of those groups.