Non-destructive fabric analysis of prehistoric pottery using high-resolution X-ray microtomography: a pilot study on the late Mesolithic to Neolithic site Hamburg-Boberg

https://doi.org/10.1016/j.jas.2012.02.029Get rights and content

Abstract

The characterisation of prehistoric pottery fragments presents a quite complex task. In provenance studies, petrographic and chemical analyses of the ceramic materials are employed to investigate potential production areas in respect to the geolocial background. Moreover, also the production technology of the firing process, as well as the forming techniques used by the prehistoric potters are of great interest. Their investigation is most often accompanied by a destructive preparation of the samples. In this paper, we want to present high-resolution X-ray microtomography (μ-CT), a non-destructive and non-invasive method, as a supplementary research tool in the study of prehistoric pottery.

Ceramic fragments from the Endmesolithic–Neolithic site Hamburg-Boberg 15 (northern Germany) were analysed by X-ray microtomography. μ-CT inspection combines quantification and shape analysis of fabric components by means of computer aided image processing. As the μ-CT method is sensitive to material densities, qualitative and quantitative analyses of different temper materials are possible. Furthermore, the μ-CT method permits the characterisation of the connectivity within the porous system, as well as the analysis of the orientation of the pore structures, which are indicative for vessel forming techniques. Although limited by the resolution of the reconstructed images, distribution analysis of heavy minerals in the clay matrix can offer distinctive features to discriminate various clay sources.

Moreover, X-ray microtomography can be used to infer the nature of organic temper even with all plant remains completely burnt out during the firing process. The visualisation of the high-resolution true volume renderings and their detailed morphometric characterisation enables new avenues in the study of ceramic technology.

Highlights

► Endmesolithic–Neolithic ceramic fragments were analysed by X-ray microtomography. ► μ-CT combines quantification and shape analysis of fabric by image processing. ► μ-CT infers the use of organic material even with plant remains completely burnt out. ► The characterization of true 3D data enables new avenues in the study of pottery.

Introduction

Pottery-making can be considered as a complex activity combining technical and social constraints. The study of archaeological ceramic materials bears a cornucopia of insights into behaviour and environment of the people involved into both production and use of pottery. To facilitate subsequent interpretations, e.g., concerning organisational aspects of pottery production such as craft specialisation, or implications of provenance studies for regional trade and interaction, a thorough reconstruction of the production technology as well as scientific investigation of the desired use of the ceramic products (e.g., storage vs. cooking pots) is essential. In this respect, geochemical, petrographic and spectroscopic techniques have proven to contribute substantially to our understanding of the technological aspects such as raw materials, manufacturing techniques and vessel function.

Both provenance studies and investigations of production technology involve the determination of compositional characteristics in terms of chemical or petrographic attributes of the bulk ceramic or its constituent compounds. Analyses of major and trace element bulk chemistry can be obtained e.g. by X-ray fluorescence (XRF; e.g., Cultrone et al., 2011; Maggetti et al., 2011; Maritan, 2004), neutron activation analysis (see Tite, 2008, and references therein) or inductively coupled plasma spectrometry (Tite, 2008; ICP-OES and ICP-MS). More recently, also laser ablation ICP-MS has been utilised in archaeometric research (LA-ICP-MS; e.g., Speakman and Neff, 2005). Chemical characterisation using Scanning Electron Microscope/Energy Dispersive Spectrometry (SEM/EDS; e.g., Spataro, 2011) permits compositional area mappings on a smaller scale or point-focused analyses of temper compounds. Petrographic analyses of thin sections (Braun, 2012; Cultrone et al., 2011; Knappett et al., 2011) yield phase characterisation of temper materials and clay matrix, which can be indicative for both geological background of the raw materials and firing conditions of the ceramics. X-ray diffraction (XRD; e.g., Schmidt et al., 2012; Maggetti et al., 2011), and, more recently also used in archaeometric studies, electron backscatter diffraction (EBSD; see Peruzzo et al., 2011) are utilised to determine the crystal structure and therefore the chemistry of minerals.

Besides age and provenance of raw materials, the most distinct characteristics of ancient pottery derive from actions by the manufacturers of the artefacts. The different modes of forming a vessel leave characteristic signatures in the ware (e.g., Lindahl and Pikirayi, 2010; Berg, 2008), and also firing conditions have, in addition to control the mineralogy of the ceramics, an impact on the development of porous structures within the sherds (for a review, see Tite et al., 2001). Aside from thin section microscopy and visual examination from surface markings, further information on forming techniques by inspection of void and temper distribution and orientation can be provided using radiography. Although applied to pottery in earlier studies (Van Beek, 1969), the potential of X-radiography for ceramics was fully appreciated with the work of Rye (1977). Since the field of diagnostic radiology was revolutionised by the introduction of X-ray computed tomography (CT), X-ray CT has had broad applicability to material sciences. Also in archaeological investigations, CT was used to non-destructively infer the interior of valuable, so far intact artefacts (e.g., analysis of Roman cremation vessels by Anderson and Fell, 1995). Promoted by increased capability of computers and imaging software, and particularly by the invention of a new type of micro-focus X-ray source, both feasibility and resolution of the CT technique improved drastically. To date, X-ray sources with small (less than 10 μm) diameter focal spots operate in bench-top systems that achieve object resolutions down to the sub-μm range (X-ray microtomography, or μ-CT). The X-ray microtomography method, like medical tomography, employs X-rays to inspect the three dimensional distribution of matter inside the object of investigation. The volume reconstruction of the material density arrangement in the samples allows detailed inspection of the fabric of the pottery sherds on micrometre scale in 2D and as well as in 3D.

The appearance of high-resolution X-ray microtomography fuelled research in many disciplines, e.g. earth sciences (quantification of rock textures and reservoir characterisation: Ketcham, 2005; Jerram et al., 2009; Remeysen and Swennen, 2008), palaeontology (dinosaurs: Sereno et al., 2007; dental tissue: Rossi et al., 2004; trace fossils: Kiel et al., 2010), metallurgical engineering (grain boundary fractures: Garcia et al., 2009). High-resolution microtomography (synchroton radiation) has also been used in archaeometric context, e.g., to investigate the carbonized remains of a Citrus-like fruit in a funerary offering deposit dated back to the beginning of the 6th century BC (Coubray et al., 2010), and, e.g. to study the impregnation of archaeological waterlogged wood with consolidation treatments (Bugani et al., 2009).

As high-resolution μ-CT detects – and localises – attenuation differences within the sampling volume, it is sensitive to chemical composition in terms of mean proton number and electron density of the material. Thus, it does not provide single element analysis. However, if the phase composition is known from other techniques, it is straightforward to recognise the phases in the volume reconstruction. Therefore, the method is distinctive for different temper materials as well as variations in the clay matrix. Moreover, because of the high density contrast between solids and voids, the μ-CT method is suited extraordinarily well to survey occurrence and orientation of the porous structure within the sherds. Thus, the application of μ-CT on ceramic fragments can be applied to study details related to production technology, and, by quantitative analysis of fabric compounds, utilised to perform compositional investigations.

The present study deals with an archaeometric approach to ceramics from the final Mesolithic – early Neolithic site Hamburg-Boberg 15, which is situated on one of several small, occupied sand dunes located in the marshlands of the Elbe Urstromtal in the southeast of Hamburg (Fig. 1). The site is considered an important location for the documentation of the Neolithisation of the Jutland peninsula, which comprises the mainland part of Denmark and parts of northern Germany. The Boberg area harbours pottery artefacts covering the final Mesolithic Ertebølle culture as well as the early to middle Neolithic Funnelbeaker culture. Moreover, there are numerous finds of sherds that exhibit a foreign morphology. Excavated already in the 1950s, the material is known because of its typologically foreign ceramics. During the last decades, the same typologically foreign sherds were related to different archaeological cultures, e.g. Linearbandkeramik, Stichbandkeramik, Rössen, as well as Baalberge, Gatersleben and newly Michelsberg (summarized in Klassen, 2004; Ramminger, 2012). Although Boberg is one of the most important sites for the debate on the Neolithic transition in northern Germany and southern Scandinavia, the material has never been presented completely, and the interpretations of the function of the site have differed strongly (e.g. Schindler, 1953, 1961, 1962; Schwabedissen, 1994; Hartz et al., 2000). Regarding northern Germany, it has been suggested that changes in the subsistence strategies rely on a network of connections established between the Mesolithic people of northern Germany and the Neolithic people of more southern regions (e.g. Schindler, 1961; Laux, 1986; Schwabedissen, 1994; Hartz et al., 2000). These studies have often drawn upon ceramic typology and stylistic comparison with other regional cultures, while issues of pottery technology have rarely been addressed, and the question of import or imitation of the typologically foreign ceramics at Hamburg-Boberg has never been discussed. Therefore, we think that a more sound understanding of the site can be achieved by addressing also key aspects of the manufacturing process and the desired use of the ceramics. Thus, the combination of analytical methods and typological analysis, and not an approach solely based on the latter, may strengthen the discriminative power of hypotheses concerning the complex archaeological record of the site. As this is the first step in an ongoing project to assess the potentials and limitations of a technological approach to ceramic style and exchange for the site Hamburg-Boberg, a discussion concerning the cultural or technical roots of ceramics is outside the scope of this work.

This study was mainly sparked by the prospect of evaluating the potential of an integrative assessment of a composition-related (type) as well as a manufacturing-related (preparation) description of clay matrix, solid temper compounds and the porous system of the pottery fragments. The porous system, consisting of voids created by organic temper that vanished during the firing process and of porosity gained by the forming process and the firing technique, is a key to decipher certain steps in the production chain. Especially the use of organic material bears potential to infer many ancillary conditions of pottery production. Mixing organic matter with clay material is a basic recipe for making building material and sometimes pottery. Many traditional societies use plant temper. In Middle European prehistory, it is especially known from the early phase of the Linearbandkeramik (Cladders, 2001) as well as within Epi-Rössen and Michelsberg groups, e.g. in France and Belgium (Constantin and Kuijper, 2002). Some recent studies have underlined the technological function of plant temper (Skibo et al., 1989; Sestier, 2005; Tsetlin, 2003). Sestier et al. (2005) pointed out, that, despite the fact that organic-tempered pottery is quite common in some cultures, only few studies have been carried out on the use of plant temper in Neolithic sherds (Sestier et al., 2005, p. 252). As a potential reason they mention the lack of an analytical technique appropriate to the characterisation of organic fragment “ghosts” in pottery. Plant temper of prehistoric ceramics is probably more common than previously thought, and therefore deserves more intense research. The inclusion of fibres enhances the tensile strength of the wet and dry paste, and improves water exchange through macroscopic porosity (Sestier, 2005). Earlier on some trials were made to show the structure and form of plant inclusions by filling the voids with a fluorescent polymer (Sestier et al., 2005), aiming at the botanical identification and a quantitative evaluation. While SEM studies (e.g., Tomber et al., 2011) that focus on analysis of organic inclusions explicitly require organic remains, the μ-CT method is able to investigate the very cavities that have been left by now vanished organic material (see, e.g., Kiel et al., 2010, who investigated the nature of trace fossils in oligocene whale bones by morphometric analyses of borehole cavities made by a bone-eating worm).

In this paper, we want to present high-resolution X-ray microtomography as a supplementary research tool in the study of prehistoric pottery. We want to evaluate the potential and the limitations of the μ-CT method, concerning the charaterisation of nature, preparation and abundance of temper agents. Furthermore, we want to assess the applicability of μ-CT observations on cavity morphology and orientation of the pore structure to infer the coiling technique that was applied in the forming process of the vessels.

Section snippets

Materials and method

In this pilot study, we report the application of a non-destructive and non-invasive method for textural analysis of ceramic artefacts. Five selected potsherd samples have been investigated using high-resolution X-ray microtomography to find distinctive markers for technological features which supplement the typological characteristics of the sherds.

The morphometric analysis of the ceramic microtexture involved the following steps:

  • (1)

    qualitative analysis of the temper material (grain size and

Results

The volume reconstruction of the material density arrangement in the samples allows detailed inspection of the fabric of the pottery sherds on micrometre scale in 2D and as well as in 3D. Each voxel represents the mean linear attenuation coefficient for X-rays of the corresponding volume of the sample and depends on density and atomic number of the scanned material. In the reconstructed images, areas of high attenuation (e.g. dense minerals in the clay matrix and rock fragment temper) are

Discussion

In order to find distinctive markers for technological features that can supplement findings on typological characteristics as well as results of petrographic and chemical analyses of the ceramic material, we employed high-resolution X-ray microtomography to distinguish the ceramics on the basis of their plastic and non-plastic components.

By means of true 3D morphometric analyses of the reconstructed μ-CT images it has been possible to characterise nature, abundance and differences in the

Conclusion and archaeometrical perspectives

In this study, we used high-resolution X-ray microtomography as a tool to investigate the microstructure of selected prehistoric potsherd samples from the Endmesolithic–Neolithic site Hamburg-Boberg 15 (northern Germany). The potential and the limitations of the μ-CT method, concerning the charaterisation of nature, preparation and abundance of temper agents, have been addressed. Furthermore, we reported the excellent applicability of μ-CT observations on cavity morphology and orientation of

Acknowledgements

We are grateful to Helms-Museum, Hamburg for the loan of artefacts for analysis. Dr. Markus Helfert and Dr. Ole Stilborg are thanked for intensive discussions of the material, Ester Gütschow and Markus Helfert for photographs of the sherds, and Dr. Niels Jöns for his skilled usage of GMT.

References (50)

  • K. Remeysen et al.

    Application of microfocus computed tomography in carbonate reservoir characterization: possibilities and limitations

    Marine and Petroleum Geology

    (2008)
  • M. Rossi et al.

    MicroCT scan in paleobiology: application to the study of dental tissues

    Nuclear Instruments and Methods in Physics Research B

    (2004)
  • P. Schmidt et al.

    Crystallographic and structural transformations of sedimentary chalcedony in flint upon heat treatment

    Journal of Archaeological Science

    (2012)
  • M. Spataro

    A comparison of chemical and petrographic analyses of Neolithic pottery from South-eastern Europe

    Journal of Archaeological Science

    (2011)
  • R. Tomber et al.

    Rice temper: technological solutions and source identification in the Indian Ocean

    Journal of Archaeological Science

    (2011)
  • S. Bugani et al.

    Study on the impregnation of archaeological waterlogged wood with consolidation treatments using synchrotron radiation microtomography

    Analytical and Bioanalytical Chemistry

    (2009)
  • M. Cladders

    Die Tonware der Ältesten Bandkeramik. Untersuchung zur zeitlichen und räumlichen Gliederung

    (2001)
  • C. Constantin et al.

    Utilisation de mousse comme dégraissant dans des céramiques néolithiques de France et de Belgique

    Bulletin de la Société Prehistorique Franςaise, Paris

    (2002)
  • G. Cultrone et al.

    Iberian ceramic production from Basti (Baza, Spain): first geochemical, mineralogical and textural characterization

    Archaeometry

    (2011)
  • L.A. Feldkamp et al.

    Practical conebeam algorithm

    Journal of the Optical Society of America A: Optics and Image Science and Vision

    (1984)
  • S. Hartz et al.

    Frühe Bauern an der Küste. Neue 14C-Daten und aktuelle Aspekte zum Neolithisierungsprozess im norddeutschen Ostseeküstengebiet

    Prähistorische Zeitschrift

    (2000)
  • B. Hulthén

    On Ceramic Technology During the Scandinavian Neolithic and Bronze Age

    (1977)
  • Kahl, W.-A., Holzheid, A., 2009. Quantifying porosities in sandstones: how to assess accuracy? SkyScan User Meeting...
  • R.A. Ketcham

    Computational methods for quantitative analysis of three-dimensional features in geological specimens

    Geosphere

    (2005)
  • S. Kiel et al.

    Fossil traces of the bone-eating worm Osedax in early Oligocene whale bones

    Proceedings of the National Academy of Sciences of the United States of America

    (2010)
  • Cited by (56)

    • Fine pottery shaping techniques in Predynastic Egypt: A pilot study on non-destructive analysis using an X-Ray CT scanning system

      2021, Journal of Archaeological Science: Reports
      Citation Excerpt :

      Furthermore, for fine pottery with few minerals and little porosity, a two-dimensional X-radiograph cannot reveal internal structure in detail. Recently, X-ray Computed Tomography (CT) scanning systems have been introduced in pottery analysis (e.g., Kahl and Ramminger, 2012) to observe and record an object’s surface and internal information. CT can basically reconstruct volume data of an object.

    View all citing articles on Scopus
    View full text