Synchrotron tts-µXRD identification of secondary phases in ancient ceramics
- 151 Downloads
The study of the alteration process, and in particular of secondary phases precipitated in pottery after its production, can constrain the post-depositional environments or the type of use of the pottery. Detailed study of secondary phases can be done using different approaches, based on microscopic, microstructural and microchemical analyses, as well as mineralogical methods. In the present study are reported the results of the application of punctual mineralogical analysis by synchrotron through-the-substrate microdiffraction (tts-μXRD) to analyse secondary phases precipitated in ancient pottery (Bronze Age pottery form Frattesina in Italy and Mesolithic pottery form Al Khiday in Sudan). Although the chemical composition of these secondary phases in many cases almost matches with the stoichiometric composition of a known mineral, only some of them were actually true mineral phases (i.e. with crystallinity and therefore showing diffraction peaks), while others were amorphous. Some phases, the occurrence of which was previously attested only on the basis of their chemical composition (determined by scanning electron microscope or the electron microprobe analysis), were univocally identified as true vivianite (Fe3(PO4)2·8H2O) and crandallite (CaAl3(PO4)2(OH)5·H2O) minerals. Part of the phases detected in studied pottery, the chemical composition of which was consistent with specific mineral phases, were instead amorphous. The synchrotron through-the-substrate microdiffraction (tts-μXRD), therefore, represents a valid method for properly determine the mineralogical nature of secondary phases in pottery, as well as other small-sized and rare mineral phases occurring in polycrystalline materials of archaeological interest, such as pottery, mortars, painting and glassy materials.
KeywordsMineralogy Synchrotron through-the-substrate microdiffraction (tts-μXRD) Vivianite Crandallite Amorphous phases Secondary phases Pottery
energy dispersive X-ray spectrometry
electron microprobe analysis
scanning electron microscope
scanning electron microscope-back scattered electron
synchrotron radiation through-the-substrate microdiffraction
X-ray powder diffraction
Secondary phases in archaeological ceramics represent the products of post-depositional processes or contamination related to the pottery use. The study of these phases can therefore supply important information on the physical–chemical conditions of the burial environment and its changes over the time [1, 2, 3, 4, 5, 6, 7], as well as the type of use of the pottery .
In most of the cases, secondary phases precipitate in voids of the pottery paste and are therefore located in specific and often sub-millimetric or even micrometric sites. They can crystallize and hence constitute a true mineral or precipitate in form of an amorphous phase. Due to their small size and the fact that in most of the cases they occur in very little quantities, they cannot be effectively isolated from the ceramic body or mechanically concentred for being individually investigated. Only local analytical techniques, such as microanalysis at the scanning electron microscope or at the electron microprobe, can be used to retrieve successfully their chemical composition. However, determination of their crystalline nature is more difficult, since the structural methods commonly used (X-ray powder diffraction: XRPD) require large amount of material compared with that available for the secondary phases that sometimes appear in pottery. The dilution effect of the ceramic body, for instance in the case of XRPD analysis, determines that the secondary mineral phases are, in most of the cases, under the detection limit of the technique. Only local mineralogical analysis, undertaken using techniques such as micro-Raman or micro-diffraction, can solve this problem. In the case of micro-Raman, the possibility of using a small analytical spot (around 1 micron) determines this technique to be very useful for the mineralogical determination at a micro scale, therefore also focusing on specific site within the materials. But the limited available database of Raman spectra represents an important limitation in the spectra interpretation, especially of uncommon mineral phases. Micro-diffractometric technique in reflection mode can be used to characterise small sized crystals as well as small areas of archaeological materials, but the investigation areas of these methods are in many cases quite large (around 50 micron) [9, 10].
The use of synchrotron through-the-substrate microdiffraction (tts-μXRD) was here applied to the study of secondary mineral phases of different composition (phosphates, silicates) identified in a series of archaeological ceramics from different periods (Mesolithic, Bronze age) and sites (Sudan, Italy). This technique allows to make punctual analysis in thin section, using a spot size of few micrometres and it has proven to be a useful for the study of cultural heritage materials to mineralogically characterize phases, small in size, occurring in the coatings as well as in the ceramic body of pottery [11, 12, 13, 14, 15]. The Bronze Age pottery from Frattesina (north-eastern Italy), the secondary mineral phases of which were previously characterised under microscopic, microstructural and microchemical points of view , was deliberately selected among other possible case studies, to evaluate the application of the tts-μXRD for the mineralogical analysis of the secondary phases.
Materials and methods
A set of secondary phases in various types of pottery, already petrographically, microstructurally and microchemically characterised at the scanning electron microscope (SEM) and by electron microprobe analysis (EMPA) [5, 6, 16, 17] were analysed by synchrotron through-the-substrate microdiffraction (tts-μXRD). In particular, the pottery here analysed comes from: (i) the Bronze Age site of Frattesina in north-eastern Italy (pottery samples FRCON34 and FRE48); and (ii) the Mesolithic occupation levels (~ 7000–5250 bc) of the archaeological sites of Al Khiday in central Sudan [18, 19, 20, 21, 22, 23, 24] (pottery samples 203 and 211). tts-μXRD analysis was performed at the Material Science and Powder Diffraction beamline  at ALBA synchrotron facility (Cerdanyola del Vallès, Barcelona, Spain). This end-station is equipped with Kirkpatrick-Baez mirrors providing a monochromatic focused beam of 15 × 15 μm (FWHM) size and a Rayonix SX165 CCD detector. Since the measurement is in transmission mode, the lateral resolution provided by this beam size is of the same order as the thin-section thickness (25–30 μm) . The energy used in the experiments was 29.2 keV (λ = 0.4246 Å). The sample-detector distance (180–250 mm) and the beam centre positions were calibrated from LaB6 diffraction data measured at exactly the same conditions as the sample. The tts-μXRD allows collection of data directly on the thin sections preserving the textural context and allowing local mineral phase identification and/or structure determination [15, 26, 27, 28]. Mineralogical composition was retrieved for the analysed areas using Sleve+ plugin within PDF4+ database (ICDD).
Results and discussion
The mineralogical analysis of the secondary phases observed in the studied samples shows that only some of them are crystalline, while others are amorphous.
Moreover, the detailed observation of the diffraction pattern of vivianite (Fig. 1c) show the occurrence of some peaks (low in intensity) such as at 2 Theta value of 6.71°, which does not find a correspondence with the reference pattern of vivianite (PDF file (00-030-0662). The identification of the phase responsible of this peak was not possible due to the small number of reflexions.
Bulk mineralogical analysis by XRPD of the Frattesina pottery detect only traces of phosphate, the broad peaks and very low intensity of which did not allow to the definition of the mineralogical phases. The small size of these secondary precipitates and their occasional occurrence is also attested by the low concentration of P2O5 measured by X-ray fluorescence bulk chemical analysis (FRE48: 1.02 wt%; FRCON34: 0.69 wt%), which fall in the average values of other ceramic classes found in the region, in which none evidence of secondary phosphates were attested. Phosphates precipitated in these potsherds within both long channels (some hundreds microns in length) and small pores (some tens microns in size), forming geode-like structure, partially or completely filling the cavities.
As for the origin of the calcium and iron phosphate, the change in the burial environment, particularly the more oxidising condition and the interaction with Ca-rich circulating ground water, promoted the transformation of vivianite into a hydrated calcium and iron phosphate. Mitridatite is in these samples not crystalline since it formed through an Ostwald step sequence after vivianite. In such situation, the transformation stopped at a higher Gibbs energy, with the amorphous particles characterised by a lower superficial energy than if the mitridatite were in crystalline form. Therefore, the transformation determined the formation of an amorphous phase with a chemical composition consistent with that of mitridatite, which could in fact eventually evolve to crystalline mitridatite. Both thermodynamic and kinetic factors influenced the transformation and inhibited the crystallization of mitridatite. The amorphous state of this secondary phase can be also observed from its optical behaviour (Fig. 2a, b) under a petrographic microscope, as the lack of interference colours is compatible with a cubic or, as it is the case, an amorphous phase. It is important to remark that the bulk mineralogical composition of the pottery obtained using conventional XRPD bearing these secondary phases does not comprise any phosphate , due to their low content and to the fact that only part of them (vivianite) is crystalline.
Vivianite and amorphous iron and calcium phosphate are post-depositional phases, precipitated in the pottery during the burial, according to the model proposed by Maritan et al. .
The precipitation of this phase took place mainly on the elongated channels of the outermost portion of the inner surface of the pottery, up to about 1 mm in depth, and filling pores of thickness of some tens microns. This pottery is characterised by a pervasive pore system, mainly constituted by channels, planar voids and elongated vughs, as well as shrinkage rims around the quartz inclusions, covering about 20% of the total area according to image analysis determination .
Crandallite was previously attested as a secondary phase in Amazonian pottery but it was identified only on compositional grounds, no mineralogical evidence supported its occurrence . In the present study, the possibility of performing the mineralogical analysis also on small-crystal aggregates by tts-μXRD, allowed to univocally identify this secondary phase from a mineralogical viewpoint. In spite of the fact that a very large set of samples (more than 400 potsherds) from the archaeological sites of Al Khiday have been petrographically, microstructurally and chemically analysed [16, 17, 29], only in one of them (sample 211) it was possible to identify crandallite. This potsherd contains a higher percentage of P2O5 (1.82 wt%) than almost all the others produced with the same recipe (around 0.30 wt%), with the exception of few potsherds coming from the same area (site 10-W-4), in which the P2O5 range around 1.5–2 wt%. In all these samples with a high phosphorous content no phosphates were observed by X-ray powder diffraction (XRPD) performed on the bulk. The microstructural analysis at the SEM indicated that only in sample 211 crandallite could have precipitated as an independent phase and tts-μXRD results confirm that the secondary phase is truly crandallite and not an amorphous precipitate. In all the other P-rich potsherds, phosphorous was probably adsorbed by the ceramic paste, without the precipitation an individualized secondary phase, according also to what previously observed by Freestone et al.  and Freestone . Interestingly all the samples with high phosphorous content in Al Khiday come from the same site (10-W-4) dated to the Late Mesolithic (6000 to 5250 bc).
The presence of crandallite in the potsherd can be interpreted as a proof that the pot was used for cooking for a very long time. More specifically, phosphorous and calcium were probably supplied by the cooked meat, and in particular by the bones, mineralogically composed by hydroxilapatite. The experimental work of Rodrigues and Costa  supports this interpretation, in their work on the effect of phosphorous incorporation in cooking pots, variscite, another hydrated aluminium phosphate (AlPO4·2H2O), crystallized after a long time of contact with an experimental solution containing 0.03 wt% P2O5 and 0.03 wt% CaO (more than 600 h of cooking). Rodrigues and Costa experiments were performed using a solution and not real food, and the formed secondary phosphates were so abundant that could be observed by XRPD on the bulk . In any case, their study clearly indicates that the accumulation and crystallization of secondary phosphates is time dependent. The crandallite found in sample 211 would have formed similarly: this phosphate would have crystallized after reaction between alumina and the phosphorous and calcium available in the cooking fluids. Alumina could come from the ceramic body, after the phyllosilicate decomposition (and in particular kaolinite, the breakdown of which determine at temperatures exceeding 550 °C the formation of a metastable and amorphous phase, metakaolin, a known source of alumina and silica). This site has been considered a seasonal occupation devoted to hunting activities, as inferred by analysis of faunal remains . Cooking the meat, most probably stewing it, may have provided some advantages: softening muscle fibres, eliminating or reducing bacteria and pathogens, increasing energy yields, longer preservation [33, 34]. The lack of crandallite in other samples form the same site may be explained by the use of the pottery for: (i) cooking other type of food than meat, such as vegetables and other plants; (ii) making few cooking, not enough to determinate the accumulation and precipitation of crandallite. A series of residues analysis (on about 100 potsherds) are at the moment under execution to define the type of lipids occurring in the pottery form Al Khiday. This on-going research is also coupled with a series of experimental cooking of different types of food and for varying duration (number of times that the pot is used), using locally produced pottery with analogous characteristic in terms of ceramic paste.
Considering the shape of the portion, the high content of silicon, as well as the amorphous state, this deposit could be interpreted as the product of precipitation of amorphous silica enriched in impurities. According to the experimental work by Wang et al. , magadiite, a hydrous sodium silicate phyllosilicate (NaSi7O13(OH)3·4(H2O)), precipitates from a hydrous silica solution in hydrothermal conditions (120 °C) after 1 day of heating. A similar process could be responsible of the silica-rich crust in samples 203. It may represent the accumulation and deposition of silica from the boiling of plants. The tissue of some plants, in fact, is characterised by a microscopic structure composed of amorphous silica containing up to about 10% of water, namely phytoliths. The accumulation of these skeletal portions of amorphous silica from the processing of some plant through cooking, determined the formation of this crust of amorphous silica, then contaminated by other chemical elements circulating in solution in the fluid that permeated in the post-depositional environment. The transformation of this amorphous into a crystalline mineral phase/s was probably prevented by the short duration of the heating process with respect to the minimum time necessary to promote nucleation and growing .
The tts-μXRD analysis on secondary phases in ancient pottery represents an important method to univocally define whether these phases are true minerals or amorphous precipitates. The small size of the focused spot (15 × 15 µm) of tts-μXRD allows to analyse small-size phases, often precipitated in submillimetric pores within the ceramic body, which cannot be mechanically separated. Other methods that allow local analysis, such as micro-Raman spectroscopy, could also be used to determine the mineralogical nature of these phases, although in some cases, when dealing with archaeological porous materials, micro-Raman may encounter fluorescence effects that add noise to the recorded experimental data. Besides that, the availability of XRPD databases containing thousands of mineral phases (PDF-type databases) almost guarantees to find adequate correspondences without the need to acquire experimental patterns from standard samples.
In this case study some mineral phases, the occurrence of which was previously defined only on chemical composition grounds either using scanning electron microscope  or electron microprobe [5, 6] analyses, have been now univocally identified as minerals whilst for others it has been shown that they precipitated as amorphous phases during the pots use or in the burial environment, although showing a chemical composition that sometimes is stoichiometrically consistent with that of specific mineral phase (mitridatite).
The tts-μXRD, finally, seems to be an interesting analytical method in archaeological study, since the precise mineralogical definition allows to grasp functional aspects of prehistoric pottery assemblages that may otherwise remain obscured, as well as defining the post-depositional processes and conditions which determined their formation. In the case of Al-Khiday Mesolithic specimens it provided a hint on some possible indications connected to the subsistence system of this group of hunter-gatherers on which it seems worth to investigate thoroughly.
The synchrotron through-the-substrate microdiffraction (tts-μXRD) represent therefore a powerful method to determine univocally the mineralogical nature of small-sized remnants of secondary phases in pottery, but also in all kind of archaeological materials.
LM, LC, AC, JR designed the research; LM and DU performed the sampling; LM and EG performed the microstructural analysis; LM, LC, AC, JR, OV performed the tts-μXRD analysis; DU did the archaeological interpretation; all the authors contributed in the writing. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
This research was possible thanks to the financial support of University of Padova (project: BIRD160990/16), Spanish Ministerio de Economía y Competitividad (project: CGL2013-42167-P), MINECO/FEDER (project MAT2015-67953-P) and the synchrotron facilities (experiment AV-2017022081).
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 1.Heimann RB, Maggetti M. Experiments on simulated burial of calcareous terra sigillata: mineralogical changes-preliminary results. In: Hughes MJ, editor. Scientific studies in ancient ceramics. London: British Museum Occasional Paper 19; 1981. p. 163–77.Google Scholar
- 3.Buxeda i Garrigós J, Cau Ontiveros MA, Madrid i Fernández M, Toniolo A. Roman amphorae from the Iulia Felix shipwreck: alteration and provenance. In: Hars H, Burke E, editors. Proceedings of the 33rd international symposium on archaeometry, geoarchaeological and bioarchaeological studies, Institute for Geo- and Bio-archaeology. Amsterdam: Vrije Universiteit; 2005. p. 149–51.Google Scholar
- 9.Berthold C, Zimmer KB, Scharf O, Koch-Brinkmann U, Bente K. Nondestructive, optical and X-ray analytics with high local resolution on ATTIC white-ground lekythoi. J Archaeol Sci Rep. 2017;16:513–20.Google Scholar
- 10.Russo G, Amicone S, Berthold C, Siddall R, Sconzo P. Early bronze age painted wares from Tell el-’Abd, Syria: a compositional and technological study. J Archaeol Sci Rep. 2018;21:359–66.Google Scholar
- 17.Gravagna E. Tracing provenance of Mesolithic and Neolithic pottery along the Nile river by trace elements and Sr isotope analysis. Unpublished Ph.D. thesis of University of Padova.Google Scholar
- 19.Salvatori S, Usai D, Faroug MA, Di Matteo A, Iacumin P, Lindseele V, Magzoub MK. Archaeology at Al Khiday: new insight on the prehistory and history of Central Sudan. In: Anderson JR, Welsby DA, editors. The fourth cataract and beyond. Leuven: Peeters; 2014. p. 243–7.Google Scholar
- 21.Zerboni A. The geoarchaeological contribution to the el-Salha Project: from site to landscape at Al-Khiday (central Sudan). In: Anderson JR, Welsby DA, editors. The fourth cataract and beyond. Leuven: Peeters; 2014. p. 259–70.Google Scholar
- 22.Usai D. Recent advances in understanding the prehistory of central Sudan. In: Anderson JR, Welsby DA, editors. The fourth cataract and beyond. Leuven: Peeters; 2014. p. 31–44.Google Scholar
- 31.Freestone IC. Post-depositional changes in archaeological ceramics and glasses in. In: Brothwell DR, Pollard AM, editors. Handbook of archaeological sciences. Chichester: John Wiley; 2001. p. 615–25.Google Scholar
- 34.Speth JD. When did humans learn to boil? PaleoAnthropology. 2015;2015:54–67.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.