During archaeological excavation of complex statigraphies such as urban settings, the excavator faces multiple visual challenges where decision-making will strongly influence subsequent sampling and ultimately the archaeological interpretations (Discamps et al. Reference Discamps, Thomas, Dancette, Gravina, Plutniak, Royer and Angelin2023). Viewing archaeological excavation as an “unrepeatable experiment” (Barker Reference Barker1993; Carver Reference Carver2009) requires meticulous documentation of all features and contexts; however, as discussed by McPherron and colleagues (Reference McPherron, Dibble and Goldberg2005), Romagnoli and colleagues (Reference Romagnoli, Nishiaki, Rivals and Vaquero2018), and Discamps and colleagues (Reference Discamps, Bachellerie, Baillet and Sitzia2019), field layers are often uncritically adopted as more or less coherent units of chronological events, despite site formation processes often acting on a cm scale, as in urban contexts (e.g., Kristiansen and Tjellden Reference Kristiansen, Tjellden and Sindbæk2022). Both Discamps and colleagues (Reference Discamps, Thomas, Dancette, Gravina, Plutniak, Royer and Angelin2023:29) and Croix and colleagues (Reference Croix, Deckers, Feveile, Knudsen, Skytte Qvistgaard, Sindbæk and Wouters2019:35) argue that the level of objectivity achieved by single-context excavation—that is, the hypothesis that deposited material over time will “mirror the process of stratigraphic excavation” (Harris Reference Harris1989:113)—has limitations. They point out the fact that the interpretations are (1) restricted to the excavation window and (2) challenged by the fact that the different interfaces and the relation between matrix and assemblages do not always come together perfectly. Interpretations and sampling strategy are at risk of being influenced by preconceptions and heuristic constructs due to an idealistic idea of coherence between deposited sediments and artifacts. As a consequence, the sampling strategy and the prioritized level of documentation are based on the construction of coherence between noncoherent features produced by the need for a holistic interpretation. One way to address this problem would be to adapt a more “question-oriented approach,” as proposed by, for instance, Croix and colleagues (Reference Croix, Deckers, Feveile, Knudsen, Skytte Qvistgaard, Sindbæk and Wouters2019:35), or to reevaluate the field layers after excavation to accomplish a more spatially informed framework for interpretation, referred to as postexcavation stratigraphy (PES; Discamps et al. Reference Discamps, Thomas, Dancette, Gravina, Plutniak, Royer and Angelin2023). The current paradigm of recording separate layers and features within clear chronological units of depositional layers is hence questionable, and therefore we propose a third solution: the development of better tools for decision-making in the excavation process.
The principles of CT- and CBCT-scanning are well documented (e.g., Kak and Slaney Reference Kak and Slaney2001), and it is a well-known nondestructive documentation method prior to or as an alternative to actual excavation of archaeological material in sediment blocks (Adderley et al. Reference Adderley, Simpson and MacLeod2001; Lynnerup et al. Reference Lynnerup, Hjalgrim, Rindal Nielsen, Gregersen and Thuesen1997). However, the use of high-resolution X-ray CT has also looked promising in regard to geoarchaeological investigations when investigating micromorphological features (Adderley et al. Reference Adderley, Simpson and MacLeod2001).
Micromorphology has proved to be an ideal method for interpreting the palimpsest of microstratigraphical depositions, particularly at urban or settlement sites (Gebhardt and Langohr Reference Gebhardt and Langohr1999; Matthews et al. Reference Matthews, French, Lawrence, Cutler and Jones1997; Simpson et al. Reference Simpson, Milek and Guðmundsson1999). Here, the archaeologist typically faces two challenges: on the one hand, strongly homogenized deposits termed “Dark Earths” (Devos et al. Reference Devos, Vrydaghs, Degraeve and Fechner2009) and, on the other, very thinly stratified layers (Wouters Reference Wouters2020). Both types of deposit have proved difficult to investigate owing to the seemingly coherent matrix in the case of Dark Earth and the presence of micro laminations in the other (Devos et al. Reference Devos, Vrydaghs, Degraeve and Modrie2011, Reference Devos, Nicosia and Wouters2020; Macphail and Courty Reference Macphail and Courty1985; Nicosia et al. Reference Nicosia, Langohr, Mees, Arnoldus‐Huyzendveld, Bruttini and Cantini2012). Numerous impregnation technologies have been attempted to minimize the loss of fine material due to shrinkage and cracking during the water removal and resin impregnation stage, particularly in both organic- and clay-rich sediments (FitzPatrick Reference FitzPatrick1984); the long processing time—as long as three months (Adderley et al. Reference Adderley, Simpson and MacLeod2001)—prevents it from supplying ongoing excavations with crucial information. While portable X-ray fluorescence spectrometry (pXRF) and laser-induced breakdown spectroscopy have shown themselves to be useful in interpreting sediment layers from impregnated blocks (Holcomb and Karkanas Reference Holcomb and Karkanas2019), the ideal method would be an easily applicable screening tool, if possible one based on a swift visual assessment.
The aim of this study has been to develop and evaluate the usefulness of CBCT-scanning of sediment samples and quick, low-tech micromorphological screening as visual tools to interpret complex stratigraphies during the narrow window of the ongoing archaeological excavation. More specifically, the study explores whether micromorphological screening and CBCT-scanning can be adopted as standard tools for better archaeological prioritization and interpretation by adding information on features invisible to the naked eye.
Based on the abovementioned challenges, we have chosen four criteria for success when evaluating each method: (1) sampling treatment, from in situ sampling to end result (within five working days); (2) documenting features invisible to the naked eye; (3) acknowledging microstratigraphy in seemingly homogeneous layers; and (4) helping to prioritize where to sample for further micromorphological thin section analysis.
Methods and Materials
Sediment Samples and Sites
Sediment samples were retrieved from three Danish sites: Ginnerup at Djursland, a Neolithic cult site (Klassen et al. Reference Klassen, Rasmussen, Kveiborg, Richards, Orlando, Svenning and Ritchie2023); the Iron Age cult site of Sorte Muld at Bornholm (Adamsen Reference Adamsen2008); and Ladegård near Aarhus, an early Iron Age settlement with scattered housing and a large number of pits (not published). The sediment ranges from sandy loam to sandy.
In 2022, 24 sediment blocks (12 × 8 × 4 cm) were sampled at the southern profile of the Ginnerup site, Djursland, using transparent plastic boxes. In the spring of 2023, an additional nine samples were taken at the Sorte Muld site, Bornholm, using plastic cable conduits (40 × 5 × 4 cm), and, finally, a sediment block (20 × 5 × 4 cm) was sampled from the Ladegård site. The long cable conduit boxes allowed a longer and coherent sequence to be screened, and the fact that they were made of plastic enabled CT-scanning prior to impregnation.
The samples were taken from the layer of sediment following dark, anthropogenic material with ceramics and bones (Figure 1; see Supplementary Text 1 in Tjelldén Reference Tjelldén2025). To produce a heterogeneity in the samples, each sample contained the organic-rich layer as well as the sandy, nonanthropogenic layer. Each sample was tightly wrapped in thin film and bubble plastic during transportation. They were kept in a cold storage room at the Institute of Geoscience, Aarhus University.
The samples at the Ginnerup site, Denmark, were taken so that both a dark, anthropogenic layer with ceramics and bones and also a sandy, nonanthropogenic layer were present. Photo: Anna K. E. Tjelldén.
CT- and CBCT-Scanning of Sediment Samples
The aim of including CT- and CBCT-scanning as a tool was to acknowledge sedimentary microfeatures invisible to the naked eye, such as the presence of microstratigraphy. The samples from Ginnerup were CT-scanned at the Department of Nuclear Medicine and positron emission tomography (PET) of Aarhus University Hospital, using a Siemens Biograph Vision 600 PET/CT (120 kV, 300 Eff mAs, 150 mm field of view, 0.6 mm slice thickness, filter kernel Br62). The samples from Sorte Muld and Ladegård were CBCT-scanned at the Department of Dentistry and Oral Health, Aarhus University, on a NewTom 5G (Cefla medical NA, Verona, Italy) with an isotropic resolution of 0.3 mm and 14-bit grayscale. Image assessment was performed by an experienced observer, in a subjective manner (i.e., simple image observation considering the three planes of the acquired image volumes), targeting the identification of structures.
Micromorphological Sediment Screening
The primary focus of developing the MSS method was therefore to provide a low-tech supplementary analysis to the classical micromorphological thin section analysis. The main scope was to gain information based on an impregnation and a subsequent screening of the cut-through surface instead of cutting a thin section.
For the micromorphological screening result to be useful during an ongoing excavation, the success criterion was to reach a sample preparation time and screening result within a week. Each step of the sample’s treatment was therefore evaluated to find the fastest procedure possible (Figure 2).
Overview of the experimental setup to minimize sample preparation time prior to micromorphological screening.
Choosing the three different low-viscosity impregnation materials to be tested (epoxy, methyl methacrylate [MMA], and tetraethyl orthosilicate [TEOS]) was done by conferring with laboratory thin section experts at the Technological Institute at Taastrup, Denmark, experts in thin-section production at the company PELCON at Ballerup, Denmark, and geologists at the Department of Geoscience, Aarhus University. Also, chemists from each manufacturer were interviewed to find the impregnation material with the lowest viscosity in combination with the fastest hardening. Hence, a TEOS variant with a relatively high gel percentage was chosen to minimize the hardening time.
The following low-viscosity impregnation materials were tested:
1. Epoxy: Conpox (Resin BY 158/Hardener HY 2996) from Condor Kemi A/S with and without the powder Epodye from Struers
2. MMA: Wecryl 821with catalysator from Condor Kemi A/S
3. TEOS: SILRES BS OH100 from Wacker Industries.
The impregnation experiments were carried out at the Department of Geoscience at Aarhus University by lab technician Rikke B. Jensen and conservator Anna K. E. Tjelldén. For the impregnation procedure, please see Tjelldén (Reference Tjelldén2025).
Results
CT- and CBT-Scanning Results
The CT-scanning results proved useful for documenting artifacts and ecofacts such as bones and fragments of ceramics in the cultural layer of the archaeological samples. However, it was challenging to document the difference between the organic-rich and the sandy nonanthropogenic layers in the Ginnerup samples, as well as between the ash layer and the organic-rich layer in the Sorte Muld samples. These features were better acknowledged macroscopically, as exemplified in Figure 3.
Two comparisons between the sediment surface of samples from anthropogenic layers and the corresponding CT-scan. A1/A2 show the sample REF3 from Sorte Muld, Dk, res. 0.3 mm, and B1/B2 show the sample G2 from Ginnerup, Dk, res. 0.6 mm. The macroscopic heterogeneity seen in the photos A1 and B1 is not significantly obvious in the CT-scans. Photos: Anna K. E. Tjelldén.
The difference between the upper ash layer and the organic-rich anthropogenic layer in the Sorte Muld sample (A1/A2 in Figure 3) is barely visible in the CT-scan. Also, the CT-scan of the Ginnerup sample did not show a significant difference between the dark cultural sediment and the sandy layers. In Supplementary Figure 1 in Tjelldén (Reference Tjelldén2025) a comparison between sample surfaces and the CT-scans can be viewed next to each other (the CT-scans are from the same level in the sample as the surface shown in the photos).
MSS Results
The fixation experiments prior to a micromorphological screening proved successful, in the sense that the procedure of sampling/drying/impregnating/hardening/cutting/evaluating a cut-through surface could be done within a week if a weekend of drying was included in the preparation time. When impregnated and cut, the nonfluorescent epoxy samples did not provide any useful information, as the epoxy covered all features in a whitish layer. However, it was interesting how significantly the sample heterogeneity stood out when the cut-through surface of an epoxy/Epodye-impregnated sample was viewed in UV reflective light (Figure 4).
Two epoxy/Epodye-impregnated samples from the Ginnerup site, Denmark, viewed in normal reflected light (left) and UV reflected light (right). A1/A2 is sample G9 and B1/B2 is sample G-REF. The fluorescent material Epodye makes porosities and air voids brighter, which enhances the visibility of density difference between the layers. Photos: Mikael Dissing.
Based on the results from each step, the best fixation result was observed when following the protocol below:
DAY 1
Drying: Drying in an oven at 45oC for three nights.
Compared with vacuum drying by boiling the water off, warming the sample to 45oC for three nights in an air-ventilated oven proved most efficient in regard both to time and to the number of samples you could dry concurrently.
DAY 4
Impregnation: Impregnating in a vacuum chamber.
We soon realized that it was necessary to impregnate using a vacuum impregnation chamber or else the sample would not be fully fixated, which both impeded the cutting of the sample and made screening impossible. As this project demanded large sediment samples, the standard chambers were too small, and we therefore designed a suitable chamber (L: 52 cm, inner diameter: 13.5 cm), which was manufactured by PELCON A/S. For a protocol for the use of vacuum-impregnating sediment samples, see Supplementary Text 2 in Tjelldén (Reference Tjelldén2025). Vacuum chambers are standard equipment at most laboratories, and this may also prove perfect for the purpose of impregnating samples.
Impregnation material: Conpox epoxy with/without Epodye (hardening for 24 hours) proved the best impregnation.
The MMA (Wecryl 821) did not harden for several days and continued to reek from its components for over one week. TEOS only fixated a few centimeters of the surface, as the hardening process requires oxidation and the samples were too thick to be sufficiently exposed to the air.
DAY 5
Cutting: A fine diamond blade with no teeth proved best for cutting the sample in the laboratory.
Screening was possible without subsequently polishing the surface.
Screening: Screening the samples impregnated with Conpox epoxy supplemented with 1% Epodye proved to give the best results.
Discussion
Several highly regarded protocols for thin section preparations for soil science have been published over the years (e.g., Benyarku and Stoops Reference Benyarku and Stoops2005; FitzPatrick Reference FitzPatrick1981; Guilloré Reference Guilloré1985; Kubiëna Reference Kubiëna1938; Murphy Reference Murphy1986). As a primary success criterion for this study was to produce a screening result within a short period of time, preferably within a week, it soon became apparent that the real challenge was to define how much information is enough to be used on a screening level to aid decision-making during an ongoing excavation. When questioning the field archaeologists, however, it became clear that the answer to “enough information” seemed more directed toward traditional micromorphological thin-section interpretations than to what fast answers actually could be used for: questions on formation processes, for example, such as “Did these layers accumulate over a long period of time or can some be regarded as single events?,” or on identifying features, such as “Is this flooring?” Based on these initial interviews we decided to evaluate the CT and MSS methods according to scanning abilities in relation to four success criteria, discussed below.
Criterion 1: Sample Treatment from in Situ Sampling to End Result
CT. The CT-scanning needed one day (i.e., half-hour sampling in situ, transport to university or hospital, 10 minutes for one scan, half-hour data treatment). Although the CT-scanning per se is a fast method, it does require easy access to a CT-facility either at a university or a hospital. This may present a challenge if the results are needed within a few days, as such facilities are often dedicated to “real patients.”
MSS. A micromorphological fixation and screening requires five days (half-hour sampling in situ, three days drying, 24 hours for impregnation and cutting, 10 minutes data treatment).
Criterion 2: The Identification or Documentation of Features Invisible to the Naked Eye
CT. As expected, the CT-scan was very useful when documenting small objects within the sediment samples. If this is the aim of the scanning, it is a useful, fast, and nondestructive method of documentation.
MSS. The use of a fluorescent powder added to the epoxy resin is traditionally done in the concrete industry to evaluate the amount of air voids and porosity within the samples using UV light. However, it also proved useful when screening the archaeological stratigraphy, since it enables differentiation between layers of varying porosity and density. This could prove indispensable for rapidly identifying features formed by compaction, such as trample zones in floor layers.
Criterion 3: Discerning Microstratigraphy
CT. Our hypothesis was that it would be possible to sample, scan, and document possible microstratifications within one day, thereby pointing out further areas for micromorphological analysis. However, CT and CBCT unfortunately did not prove successful when scanning heterogeneous samples. One can speculate that this is owing to the fact that both CT and CBCT images rely on the X-ray attenuation provided by the samples, and in this case the composition in regard to that parameter is quite homogeneous, therefore avoiding differentiation among substances. Besides that, in the present study we focused on a simple, yet subjective, image assessment. One could speculate as to further quantitative assessments that rely on the gray shades within the volume to perform a more “computer-oriented” assessment. As no previous methodologies were found that could provide a reliable approach to perform such a task, the simple, observational assessment was favored at this stage.
MSS. However successful the screening method was when investigating the epoxy/Epodye-impregnated samples macroscopically, it is necessary to further evaluate whether it can differentiate microlaminations from each other. This requires a microscope with UV light capability.
Criterion 4: Where to Sample for Further Micromorphological Thin-Section Analysis
CT. Since it was not even possible to macroscopically acknowledge the heterogeneity in the samples, it is doubtful whether microstratigraphy can be documented using this screening method.
MSS. The micromorphological screening would be a useful tool to point out areas for further micromorphological thin section analysis to address such questions as “Are we inside or outside a building?” or “What is the corresponding matrix for this truncation?”
Conclusion
The aim of this study was to develop and evaluate new screening methods that would enhance the interpretation of complex stratigraphies while the archaeological excavation is still ongoing. The agenda was to enable the field archaeologist to choose from a larger toolbox, both when interpreting the site formation and when prioritizing further sampling strategy.
Two methods were evaluated in this study: CT-scanning (and CBCT-scanning) and micromorphological sediment screening, and while the former is a well-known method for documenting artifacts within soil blocks and burial urns, the latter was developed within this project as a low-tech method supplementary to traditional thin section analysis.
Based on the scanning and fixation of 34 sediment samples from three different Danish archaeological sites, we conclude that CT-scanning does not offer any information on microstratigrafication, as it did not clearly differentiate between obvious macroscopic heterogeneous layers within the samples. As for micromorphological screening, the method showing the most potential was to vacuum-impregnate the sediment samples with a low-viscosity epoxy supplemented with the fluorescent powder Epodye. When viewed in UV-reflected light, the features and layers stood out significantly.
This low-tech screening of homogenous organic rich layers could benefit the decision-making process enormously during the narrow window of the archaeological excavation, as it would draw attention to those areas and sequences that provide the most important information.
Acknowledgments
We would like to thank Mikael Dissing (Technological Institute at Taastrup, Denmark) for the useful conversations and considerations during the whole impregnation process. The director of PELCON, Peter Laugesen, was very helpful in giving hands-on advice regarding the vacuum impregnation process, and we would like to thank Ole Munk and Aage Kristiansen for kindly CT-scanning our samples from the Ginnerup site. Special thanks go to the project leaders of each archaeological excavation for letting us sample from their profiles: Lutz Klassen at Ginnerup, Michael S. Thorsen at Sorte Muld, and Rasmus Iversen at Ladegård. For help with sampling at the Ginnerup site, we thank Hjalte W. Mølgaard, and, finally, we would like to thank Professor Cristiano Nicosia, University of Padova, for offering us insight into the thin section standards, as well as giving us critical advice during our “learning by doing” process when fixating soil sediment samples. Finally, we would like to thank the anonymous peer-reviewers who helped improve the manuscript.
Funding Statement
This work was supported by the Augustinus Foundation under Grant No. 21-1585. The archaeological excavations have been separately financed by other sources and are reported. References to these data and discussions of them can be found in the cited articles. Permits were not required for this research.
Data Availability Statement
Raw data and supplementary material can be viewed by accessing following link to data repository: https://figshare.com/projects/Data_repository_Micromorphological_screening/246980.
Competing Interests
There are no competing interests directly or indirectly related to this study and the article submitted for publication. None of the material is being considered for publication elsewhere, and we are unaware of any conflicts of interest that would bias the review process.
Open access
