Characterization of Glazes of the 11-14th Centurt Potsherds and Tiles from Comana Pontica in Tokat, Turkey

Date

2023-9-08

Author

Altundal, Canberk

Metadata

Show full item record

This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Item Usage Stats

356
views

560
downloads

Cite This

Comana Pontica, the ancient settlement in Tokat, Turkey, was one of the intersection points on the main ancient road connecting the inland cities of the Black Sea Region to the coastal cities. Archaeological findings from the layers of the Ottomans, the Seljuks-Danishmendids, the Byzantines, the Romans and the Greeks were revealed by the excavations from 2009 till present day. Totally 1479 glazed ceramic samples, 1842 unglazed cooking utensils and 2223 unglazed storage containers were unearthed from the Seljukid-Danishmendids’ layers until 2018. In this thesis, within these 5544 ceramic samples, totally 106 glaze spacimens with different colors on sides, together with 6 paste and 3 slip spacimens from the inner sides of 71 glazed potsherds and 2 glazed tiles, which were collected from the 11-14th-century Seljukid-Danishmendids’ layers, were analyzed by Fourier-Transform Infrared (FT-IR), X-Ray Powder Diffraction (XRD), Raman Spectroscopy, and X-Ray Fluorescence (XRF) Spectrometry. The aim was to find the composition of glaze colorants and production technologies of the glazes, to reveal the estimated firing temperatures of the ceramics and to characterize the samples with the help of the percentages of the oxide contents of the glazes. Moreover, these XRF oxide data were shown in bivariate and ternary plots, and then compared with the literature data in order to display whether the contents of materials used in the glazes of pottery and tile samples represent the characteristics of their period. As a result, with the help of the wavenumber positions of their Si-O asymmetric stretching peaks, firing temperature range of the Comana Pontica ceramics was estimated as changing between 600 and 925 °C, with a majority fired in the range of 800-900 °C. Brown glazes were determined to be applied in a reducing atmosphere and fired around 700-750 °C, followed by turquoise glazes in 725-800 °C, yellow and green glazes in 750-800 °C, orange glazes around 850 °C, and the black glaze around 900 °C. Red pastes were discovered to be fired around 850-900 °C, while white pastes behind the turquoise glazed potsherd and tiles, and the white slip layers around 900 °C, which can be suggested that the initial firing was done to the all potsherds except the turquoise one and at around 900 °C, then glazed and re-fired to 700-850 °C in an oxidating atmosphere for the light-colored glazes, as well as to colder temperatures in an reducing environment for the darker glazes. Furthermore, according to the vibrations around 3600-3700 cm-1 in an FTIR spectrum, the clay minerals in the glazes were identified as mainly in kaolinit (dickite, halloysite, kaolinite, nacrite) and montmorillonit-smectite (bentonite, montmorillonite, saponite, smectite, vermiculite) groups. On the other hand, exceptionally, paste of K104 was also including illite and sepiolite. In the rest of the slips and pastes, the peaks are almost not observed, probably due to the dehydroxylation, dehydration, and decarbonation occurred at higher firing temperature values such as 850-925 °C. Moreover, inclusions of slaked lime [Ca(OH)2], quicklime (CaO) and calcite (CaCO3 also provide information about the possible firing temperatures with signals at 3648 cm-1, 1647 cm-1 and 1430-1450 cm-1, recpectively, probably all taken from the crushed shells and bones. If CaCO3 peak is observed, it means the temperature was below 600 °C, CaO between 600-800 °C, Ca(OH)2 between 800-900 °C and disapperance of all the calcite peaks and -OH bands corresponds to the temperature above 900 °C. Additionally, due to the positive slope of the trend line in the chart of estimated average firing temperatures of the glazed ceramics differentiated by periods of production dates, the firing temperatures of the glazed ceramics of Comana Pontica were estimated to be increasing from the 12th till 14th centuries. Moreover, the colorant for the yellow glazes was considered mostly to be lead antimonate (Pb2Sb2O7) and green glazes as copper oxide, according to the XRF analysis results. By Raman spectrum, the presence magnetite and hematite were identified in the brown-glazed potsherd sample of K146. Moreover, XRF data showed that manganese is responsible for the color of the black tile, copper (3% for the turquoise potsherd glaze) and 6.7% SnO2 +1.6% CuO are the reasons for the turquoise coloring together with the high content of calcium oxide. By XRF, it could be seen that most of the samples contain high levels of lead up to 50%. The two tile samples are determined to be in the group of tin-opacified lead-alkali glaze, whereas the glazed potsherds are in the high lead and lead-alkali types. The results of XRF analysis of the glazes of both the potsherds and tiles showed that lead-oxide was preferred as the fluxing agent with natron. If the tin-opacified lead alkali type glaze is desired to be turquoise, then lead also should be in the medium, as could be understood from XRF data for the tiles. CuO is attributed to be responsible for the turquoise and green colors of the glazes as well as the yellow glaze is the result of Fe2O3 together with Pb2Sb2O7 and Sb2O3. This information is also supported by XRD spectrum of the turquoise glaze of the potsherd sample of K223, where Plagioclase [Na(AlSi3O8),Ca(Al2Si2O8)], Portlandite [Ca(OH)2], Quartz (SiO2) and Verdigris [Cu(CH3COO)2.H2O] might be possibly included in the glaze. For the turquoise color, the paste was preferred to be rich in calcite. For the color tones of green, it could be stated that increase in concentration of CuO could result in the yellow color of the glaze to turn into green. Yellow color of the glazes could be proposed to be reached with very low concentration (around 0.5%) of Fe2O3 in the oxidizing atmosphere and by addition of Sb2O3, PbO and TiO2. If the Fe2O3 concentration is increased to twice and 0.5% CuO and 0.06% As2O3 added in an oxidative environment, then its color turns into green. Glaze color of orange could be arisen from about 4% Fe2O3, 0.5% TiO2 and 0.13% CdO. CuO addition between 1.5-3.4% is thought to be the reason for the turquoise glaze and on a calcite paste. Purple color on the glaze of K117 potsherd could be due to the mixture of colors of red due to the reduced Cu2O and bluish green due to the oxidized CuO. Brown color could probably be occurred by FeO in the reductive atmosphere at lower temperatures together with around 0.2% TiO2, and also related with the high CO2 content. Likewise, black color could be gathered by MnO and FeO, also by contributions of Na2O and MgO. In tiles, SnO2 was used for the opacification. Cobalt is known to be responsible for the dark turquoise color of the glaze, but no cobalt is detected in any sample. Gas bubbles observed on the turquoise glaze, which could lead to consider no bisque-firing was applied prior to the glazing. Reddish color of the paste, occurred due to the clay and iron oxide, is an indicator for the atmosphere used in an oxidizing way in the kilns and its homogeneity is the result of the bisque-firing applied under high temperatures, whose temperature was estimated as around 900-925 °C from the FTIR Spectroscopy wavenumber data. Moreover, the glazes were determined to be produced using quartz, from not only calcareous but also non-calcareous clays. PbO is thought to be added most probably to lower the firing temperature, which was supported by FTIR spectroscopy that it was as low as 600-700 °C for some pottery samples. Firing temperatures estimated by FTIR spectra were mostly consistent with those determined by the positions of samples on the isothermal curves in the bivariate chart of PbO vs. (Na2O+K2O) acquired by XRF spectrometry. Binary and ternary diagrams also contributed to the possibility of the turquoise glazed sample K223 to be brought from Egypt or Iraq. From the ternary diagrams, the yellow glaze of K149 from the 13th century was found to be closer to Islamic glazes of Iran from the 10-11th centuries. On the other hand, since all the clay minerals were detected to be similar for the potsherds with respect to the 3600-3700 cm-1 region in FTIR spectra, it might be possible to produce all the pottery locally from the basin of Iris River, but influenced by different traditions or made by different potters. And by XRD spectroscopy, olivine, epidote and siderite were examined in both tiles, additionally with augite and fayalite in K002 black tile, which also supports the idea that glazing traditions in the Seljukids-Danishmendids periods were a mixture of those of the Byzantines and the Ottomans, since olivine, epidote, augite and fayalite were typical mostly to the Byzantines as well as siderite to the Ottomans. XRD spectra of the potsherd glazes were amorphous, which supports the information that they were transparent and formed by fast cooling of the glaze also affected by the addition of Al2O3 and MgO. On the other hand, XRD signals of the tiles were easily identified with the help of their crystalline content in the glaze probably contributed by ZnO, TiO2, CaO, Fe2O3, SiO2, opacified by SnO2, and formed by the slowly-cooling process. Those oxides were also checked by the XRF analysis. Anatase, calcite, albite, quartz, graphite, epidote, carbon iron silicon, siderite, magnetoplumbite, ramsdellite were identified in the turquoise tile by XRD spectroscopy. Due to the observance of anatase and calcite signals in XRD, the firing temperature was estimated between 600-850 °C, which is also parallel with the temperature found as 600-700 °C by FTIR spectroscopy. Moreover, pyroxene, quartz, graphite, albite, tin oxide, carbon iron silicon, tridymite, magnetoplumbite and ramsdellite were discovered by XRD spectrum of the glaze of the black tile, in which pyroxene and tridymite are formed at temperatures higher than 870-900 °C, which was also relevant to the same temperature range calculated as 900-925 °C by FTIR spectroscopy. From the biplot of SnO2/PbO ratios, K002 black tile was found to be extremely similar to the Iznik (Nicaea) tiles of the Ottoman Empire from the 17th century and K001 turquoise tile was observed to be similar to the tiles of an Ottoman mosque (1429) in Edirne and those of the Timurids (2nd half of the 14th century – first quarter of the 16th century) where Iran might be the possible origin of influence. With respect to the biplot of ratios of Pb/Si vs. Sn/Si, the potsherd K149 with yellow glaze from 13th century was determined to be close to the tiles from the mosques of the 15th and 16th centuries in Edirne, while K161 with green glaze and K172 with orange glaze from the 12-13th centuries were found to be close to the 15th century mosque in Edirne, which could bring the consideration that pottery making tradition with addition of similar proportions of lead and quartz might have continued from the periods of Seljukids in the 12th century to the Ottomans in the 16th century. This consideration is consistent with the information that continued workshop of ceramic arts and techniques of the Seljuks initiated the tile production in the Ottoman Empire. Also the locations of the Comana potsherds and tiles in the biplots were consistent with the other previously analyzed glazes from Seljukid Era and the same time periods were condensed in close positions, which could make the glazes found in the layer of the Seljukids unique and its differentiation possible with respect to the oxide contents. Also this uniqueness could be seen from Figure 12 that the FTIR and Raman spectra of the potsherds display both high PbO content similar to Byzantine Era as well as specified by addition of more NaO that also cause a change in the shape with a shoulder in the wavenumber region on the stretching modes. Furthermore, with the help of the biplots, mineral natron was also identified to be added as the flux during firing the glazes, that is also consistent with the traditional method in Anatolia in that period. From the characterization biplot of PbO/SiO2 vs. CaO/SiO2, both tiles were closest to the 12th century Islamic Iran and the turquoise -glazed K223 was closest to the 11-12th century Islamic Syria which could find an estimation where it could be imported from if this is the case as recommended from Dr. Bilgi Er. The locations of the Comana potsherds in the biplots were consistent with the other previously analyzed Seljukid glazes and the same time periods were condensed in close positions, which could make the glazes found in the layer of the Seljukids unique and its differentiation possible with respect to the oxide contents. The firing temperatures of the glazes of some potsherds samples are also estimated from Ip calculation of the Raman spectra, whose results are in accordance with the ones gathered by FTIR analyses, but if one method was to be preferred most probably it would be Raman Ip values since they are more accurate with sharper signals compared to the larger area under the maximum point of the curves of FTIR spectra that seems to be more misleading. Analysis of ceramics and glazes are important in not only describing and dating the archaeological remains, but also providing information about the sociocultural structure and economic relationship.

Subject Keywords

FTIR, XRF, XRD, Raman, Comana Pontica

URI

https://hdl.handle.net/11511/105947

Collections

Graduate School of Natural and Applied Sciences, Thesis