- Email: [email protected]
Infrared spectroscopic characterization of residues on archaeological pottery through different spectra acquisition modes
Vibrational Spectroscopy 76 (2015) 48–54
Contents lists available at ScienceDirect
Vibrational Spectroscopy journal homepage: www.elsevier.com/locate/vibspec
Infrared spectroscopic characterization of residues on archaeological pottery through different spectra acquisition modes Mariateresa Lettieri * Institute of Archaeological Heritage – Monuments and Sites, CNR–IBAM, Prov.le Lecce-Monteroni, 73100 Lecce, Italy
A R T I C L E I N F O
A B S T R A C T
Article history: Received 9 September 2014 Received in revised form 16 December 2014 Accepted 17 December 2014 Available online 19 December 2014
Fourier transform infrared spectroscopy (FT-IR) is a versatile analytical method, very useful in many fields. Although a crucial step in producing good spectra is the use of the appropriate technique, the acquisition mode is sometimes not accurately selected and the results are partial or lacking. In recent years, FT-IR analysis has been proposed as a screening method for characterization of archaeological potteries and identification of the residues on these artifacts before turning to destructive, more expensive, and time-consuming techniques. In this study, a set of pottery shards, classified as fragments of amphorae, was subjected to FT-IR analyses. The results obtained from different sampling procedures and different spectra acquisition modes, were examined and compared. The as-received ceramic fragments were subjected to micro attenuated total reflectance (m-ATR) analyses. Investigations in diffuse reflectance (DRIFT) mode were performed on samples collected by abrading the surfaces of the shards with abrasive paper. Samples scraped from either the surfaces of the pottery fragments or the interior of the ceramic body, were analyzed in transmission mode as a powder in KBr pellets or after extraction with acetone. The sampling by abrasion of the surface with an abrasive disk, and consequently the analyses in DRIFT mode, were successful only in identifying the inorganic compounds coming from the pottery and/or the environment, while materials related to the content of the jar were not detected. Also the analyses in transmission mode provided information mainly about inorganic materials, which, even where in a limited amount, masked the signals of organic compounds. Just an extraction with a solvent made it possible a more detailed, but still partial, characterization of these organic substances. On the contrary, the content of the jar was easily detected using the m-ATR mode, even in areas where no residue was observed. ã 2014 Elsevier B.V. All rights reserved.
Keywords: Fourier transform infrared spectroscopy Coated surfaces Archaeometry Residue analysis Archaeological pottery
1. Introduction Fourier transform infrared spectroscopy (FT-IR) has become a very popular analytical technique, for many years now. This method is fast, relatively cheap, and easy to use, and because of these advantages, it is applied to the study of a wide range of materials in many fields. A crucial step in producing good FT-IR spectra is the use of the appropriate technique for presenting the sample to the spectrometer [1]. The selection of the suitable analysis method should be based on the type, form and amount of sample to be analyzed. The physical state of the sample, its preparation, and the analysis method have an effect on the resulting spectrum, modifying the
* Tel.: +39 0832 422219; fax: +39 0832 422225. E-mail address: [email protected] (M. Lettieri). http://dx.doi.org/10.1016/j.vibspec.2014.12.002 0924-2031/ ã 2014 Elsevier B.V. All rights reserved.
absorption pattern. In fact, the above mentioned factors influence the peak position, as well as the band shape and intensity. Nevertheless, the spectra acquisition mode is sometimes not accurately selected and, as a result, the information that can be gleaned from the spectra are partial or lacking. In recent years, FT-IR spectroscopy has been proposed as a quick and cheap method to screen archaeological samples before subjecting them to more expensive and time-consuming methods (e.g., gas chromatography–mass spectroscopy) [2–6]. Most of the studies concerning such research topics proved the effectiveness of FT-IR analyses for the detection of organic residues in archaeological materials, as well as for investigations of the fabrication conditions [7–11] and for provenance studies [12–14]. Since ancient time, ceramic artifacts have been largely employed for storing materials, as well as for preparing or cooking food. As a result, traces of the materials which were in
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
49
contact with the vessels can be found as surface deposits or as adsorbed residues into the porous structure [15]. The characterization of these residues is very helpful to provide information on cultural, technological, and commercial activities of ancient societies. Chromatographic techniques are frequently employed in identifying the organic residues [15–21]. However, these methodologies are destructive, while non-destructive or microdestructive diagnostic techniques would be preferred, and procedures for sample preparation, usually taking a long time, are required. Published papers dealing with characterization of archaeological potteries and identification of the residues in ceramic vessels, frequently discuss the application of FT-IR spectroscopy in transmittance mode [7,8,19,22–25]. Analyses performed in diffuse reflectance (DRIFT) [5,26,27] or in attenuated total reflectance (ATR) mode (sometimes coupled with a microscope, that is, the m-ATR mode) [6,9,28] are also used in these studies. However, despite the fact that the results depend on the sample, on its collection and preparation, as well as on the analysis methods, the different acquisition modes have been frequently used with no distinction. Consequently, the characterization of the materials under study has been not comprehensive and the FT-IR technique has been undervalued and underused. Starting from these issues, the present research was aimed at investigating residues on archaeological ceramic samples by means of FT-IR spectroscopy. A set of pottery shards, classified as fragments of amphorae, was subjected to analyses carried out in m-ATR mode, in DRIFT mode and in transmittance mode. The results obtained from different sampling procedures and different spectra acquisition modes were examined and compared.
kinds of analysis. Consequently, several characterization tests can be carried out on the same sample. In the next step, DRIFT analyses were performed on samples collected by abrading the surfaces of the shards with abrasive paper, as detailed in Section 2.4.2. In this manner, small amounts of sample can be taken from the surfaces with insignificant damage. Furthermore, the tools for sampling are portable and can be easily used without transferring in the laboratory the items under investigation. Subsequently, FT-IR analyses in transmission mode were carried out on samples scraped from the surfaces of the pottery fragments. This methodology made it possible to collect, and then analyze, materials from different parts of the archaeological fragment, with no limitation due to the original shape, dimensions or form of the sample. In fact, the presence of residues from the jar’s content was checked also inside the ceramic body. Moreover, FT-IR analyses in transmission mode were performed after extraction with acetone to identify the organic residues without interferences from the ceramic material. Finally, small flakes of the residues on the internal surfaces of the shards were analyzed in m-ATR mode. This method allows selective examination of the residue, particularly when other sampling procedures are difficult to perform or the archaeological artifacts cannot be moved into the laboratory. The analyses were carried out in the reported order, so that the destructive or micro-destructive samplings were performed last, because the removal of the residues from the surfaces could make ineffective the investigations performed directly on the shards (i.e., the m-ATR analyses).
2. Materials and methods
A FT-IR ThermoNicolet Nexus spectrometer equipped with a deuterated triglycine sulfate (DTGS) detector, was used to record the FT-IR spectra in transmission mode. A 13 mm KBr die (Model 129, Thermo Spectra-Tech) and a hydraulic press (Mod 660, Silfradent) were employed to shape the pellets for the analyses. Following the manufacturer’s suggestion, 6000 kg load was applied for approximately 1 min. A Smart Diffuse Reflectance accessory (ThermoNicolet) was inserted in the spectrometer to perform the DRIFT analyses. This module included a slide with a slot that can hold the sampling disk. A kit consisting of handle and plate (which keep the adhesivebacked silicon-carbide paper used to abrade the surface) was employed in the sampling. The ThermoNicolet Continuum IR microscope coupled with the spectrometer was used to acquire the FT-IR spectra. This was equipped with a mercury–cadmium–telluride (MCT) detector, which was cooled with liquid nitrogen. A 15 Reflachromat objective with a slide-on ATR attachment (Thermo Spectra-Tech), using a Si crystal (refractive index = 3.4; incident angle = 45 ; contact area = 50 50 mm), was employed to collect the m-ATR spectra. To ensure reproducibility and uniformity, the contact between the ATR crystal and the sample surface was automated and computer controlled. After each analysis, the crystal was cleaned with a soft cloth soaked in acetone. The employed instrumentations were not purged with dry and CO2-free air, therefore the contribute of absorption bands of both CO2 (appearing in the spectrum as a doublet around 2340 cm 1) and water vapor (resulting in sharp and very close peaks over 3700 cm 1) are observed in the reported spectra, although a background spectrum was acquired before each analysis and automatically subtracted (by software) from the sample spectrum. All the FT-IR data were processed with the OMNIC 8.1 software (Thermo Fisher Scientific Inc.).
2.1. Archaeological samples The investigated samples were collected from a set of pottery shards discovered in the archaeological site of Hierapolis (modern Pamukkale, Turkey) [29]. Archeologists classified these ceramic objects as fragments of amphorae manufactured and used in the early-Byzantine period (V–VI century A.D.) [30,31]. The unearthed shards were only unsoiled with a soft brush, while washing with either water or any other cleaning agent, was avoided. The examined fragments have different shapes and dimensions (not exceeding 5 5 cm2), with an average thickness of 0.5 cm. Their concave side (hereinafter called internal side) always was the interior surface of the jar. Shards with residues distinguishable to the naked-eye onto the internal side, were chosen. The observed residues were blackish and/or brownish in color and exhibited a quite good adhesion to the shards, but they never fully coated the pottery surface. An example of the pottery shards under investigation is shown in Fig. 1. 2.2. Sequence of the analyses A preliminary visual inspection of the pottery surfaces was performed through a binocular stereomicroscope (Zeiss, mod. Stemi SV11) at magnifications of up to 100. This examination was aimed at distinguishing the residues from the pottery material, in order to optimize the sampling. The as-received ceramic fragments were firstly subjected to the m-ATR analyses. The shape and the limited dimension of the shards examined in this study allowed a direct analysis. In this case, sampling or preparation procedures are not required, hence the specimen is not damaged at all and it can be used later for other
2.3. FT-IR instrumentation
50
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
[(Fig._1)TD$IG]
Fig. 1. A shard of amphora from Hierapolis: (a) internal side; (b) external side.
2.4. Sampling and analytical procedures 2.4.1. m-ATR mode The m-ATR analyses were performed on both the intact pottery shards and the flakes taken from the surfaces, placing the samples directly under the m-ATR objective. The spectra were collected with the Si crystal in the range of 4000–650 cm 1, with a resolution of 4 cm 1 and 200 scans for each measurement; the background spectrum was acquired in air. In the case of the shards, the spectra were acquired on the visible residues, as well as on apparently clean areas. Furthermore, the analyses were carried out on both the internal and external side of the shard, in order to avoid misinterpretation due to the presence of substances coming from the deposition soils. 2.4.2. DRIFT mode The samples for the tests in DRIFT mode were gathered by gentle abrasion with a small circle of abrasive paper (320 grit SiCarb, Thermo Spectra-Tech) mounted on a stick. The DRIFT analysis was carried out soon after the sampling. The spectra were acquired in the range of 4000–400 cm 1, with a resolution of 4 cm 1 and 200 scans per measurement; the background spectrum was collected on the same abrasive paper disk afterwards used for the sampling. For comparison, the sampling and the analyses were performed on both the internal and external sides of the pottery fragment. 2.4.3. Transmission mode The internal and the external surfaces of the shards were scraped with a scalpel to collect the samples as a powder to be used for the analyses in transmission mode. In the same way, the matter was taken also from the lateral side of the shard, at different depths, to investigate the presence of residues due to the jar’s content inside the ceramic body. These samples were finely ground with a pestle in an agate mortar, then mixed with KBr (suitable for
infrared analysis and provided by Mallinckrodt Baker Chemical Inc.) and compacted in a pellet 13 mm in diameter. The material gathered from the internal surface of the shard was divided in two parts. The first one was used to prepare KBr pellets. The second part, placed in a vial, was extracted with acetone (analytical grade, provided by Carlo Erba Reagents), using 1 ml of solvent per 5 mg of sample. Sonication was carried out for 22 min in an ultrasonic bath (FALC Instruments), followed by standing for 24 h in laboratory conditions. A drop of the obtained liquid fraction was placed on a KBr pellet, previously prepared, which was stored for 5 min at 40 C to enhance the solvent evaporation. All the KBr pellets were analyzed in transmission mode, immediately after the preparation. The spectra were acquired in the range of 4000–400 cm 1, with a resolution of 4 cm 1 and 200 scans per measurement; the background spectrum was collected on a pellet made of KBr only. 3. Results and discussion The observations through the stereomicroscope, preliminary to the samplings, allowed to identify residues having dissimilar features. Areas different in appearance were noticed on the surfaces, as described below: a Brown coatings; b Dark-brown residues filling holes and gaps of the surface; c Surfaces with no visible residue, where the ceramic body was
easily recognized; d Whitish incrustations.
In all the examined archaeological samples, the first three types of residues (1–3) were observed on the internal surface of the shards, as shown in Fig. 2a. Instead, the whitish incrustations (4) were mainly found on the external side (see Fig. 2b).
[(Fig._2)TD$IG]
Fig. 2. Images of the potsherd taken by means of the stereomicroscope: (a) internal side; (b) external side.
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
3.1. m-ATR mode The investigations carried out in m-ATR mode on the internal surfaces of the shards were very significant. As previously stated, spectra were acquired directly on visible residues, as well as on apparently clean areas, where the visual inspection, either to the naked-eye or through the stereomicroscope, did not reveal any residue. The most of the analyses of the visible residues gave rise to spectra where the bands at 2952, 2928, 2863, 1710, 1448, and 1172 cm 1 indicated the presence of vegetable resins [1,6,32], even if the peak around 1033 cm 1, due to the silicate compounds of the pottery, was the strongest signal (Fig. 3a-1). In these spectra, significant signals at 1584, 1385, and 1229 cm 1 were also detected. In other cases, actually less frequent, the spectra of the residues on the internal side of the shards exhibited bands around 1572, 1386 and 1228 cm 1 as main peaks (Fig. 3b). Other researchers attributed similar signals to crystals of tartrate salts [26,33–35]. In fact, tartrates, usually as calcium or potassium salts, can derive from tartaric acid occurring in large amounts in grapes and have been proposed in many studies as a marker of wine in archaeological potteries [36–38]. The results of the analyses performed in m-ATR mode on flakes of residues are reported in Fig. 3a-2. These spectra resulted very similar to those collected directly on visible residues (area (1) in Fig. 2a). These findings were supported by the results of the GC–MS analyses on the same samples as discussed in another study [39]. Abietic acid, dehydroabietic acid and 7-oxo-dehydroabietic acid were identified in the residues, indicating the presence of a pine resin and/or a pitch obtained from resinous materials of Pinaceae trees [25,38,40–42]. Tartrates was not detected, but it is to take into account that GC–MS technique can be not suitable to unambiguously prove the existence of tartaric compounds in archaeological samples [21]. The m-ATR spectrum acquired on clean areas is shown in Fig. 4. Although the main peak was due to silicate compounds of the ceramic body, organic matter ascribable to the content of the jar was recognized. Vegetable resins can be identified from the signals at 2960, 2925, 2866, 1712, and 1449 cm 1, while the peaks around 1586 and 1385 cm 1 could account for the presence of calcium tartrate. Traces of carbonates were also detected (1416 and 872 cm 1). On the outside of the shards greater amounts of carbonates (1415 and 872 cm 1) and no evidence of organic materials, were
51
found (Fig. 5). In addition, peaks at 3620 and 3690 cm 1, due to the OH stretching modes [43] of clay minerals belonging to the kaolinite-group [44], were detected. Analogous signals were not noticed in the spectra of the inner surface. Finally, it is to remark that the broad band around 3400 cm 1 and the peak centered at 1625 cm 1, both related to H—O—H vibrations of adsorbed water [44], were weaker in the spectra of the external side. Using the FT-IR analyses in m-ATR mode, the superficial substances were mainly recognized, while the absorption bands deriving from the ceramic body were not strong in intensity. As a consequence, the signals due to residues were not masked and the related substances were quite easily identified. Furthermore, residues from the content of the jar were detected even in areas where no residue was revealed by the visual observation. Shape and dimensions of the investigated object could impede a suitable spectral acquisition, but the ATR technique gave meaningful results also in analyzing selected small pieces of materials taken from the shard. This latter result makes the technique valuable when the sampling is carried out on unbroken artifacts. 3.2. DRIFT mode The results of the analyses performed in DRIFT mode are reported in Fig. 6. The bands due to silicate (1025 cm 1) and carbonate (1450 and 874 cm 1) compounds are the main peaks in the spectra of the material sampled from the internal surface of the shards. The peaks at 2960, 2920, and 2850 cm 1 suggest the presence of organic substances, even if no further information can be inferred from these data. However, these signals cannot be unambiguously attributed to organic materials, because peaks between 2850 and 3000 cm 1, similar in shape, are observed also in the standard reference DRIFT spectrum (HR Nicolet Sampler Library, Thermo Fisher Scientific Inc.) of calcium carbonate. Reasonably the external side of the potsherds did not have contact with the content, therefore, for comparison purpose, analyses in DRIFT mode were performed on specimens taken from these surfaces. In comparing the spectra (Fig. 6), peaks in the range of 2850– 2960 cm 1 were still noticed, even if some differences can be observed. Greater amounts of carbonates were found on the external surfaces, because the peaks at 1450 and 874 cm 1 were stronger in intensity. Furthermore, in both cases the overtone bands of the carbonate group were observed at 2516 and
[(Fig._3)TD$IG]
Fig. 3. m-ATR spectra of residues on the internal surfaces: (a) 1 = recurrent result in areas with visible residues, 2 = spectrum acquired on a flake of residue (shown in the inset box) taken from the internal side of a shard; (b) less frequent result in areas with visible residues.
52
[(Fig._4)TD$IG]
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
3.3. Transmission mode
Fig. 4. m-ATR spectrum collected in an area of the internal surface without visible residues.
1796 cm 1 [3]. Other researchers attributed similar results to carbonate compounds in a powdered state [5] from layers covering the outside of the shards, realistically due to deposition from circulating fluids taking place in the burial conditions [45,46]. In the spectra collected on samples from the external side, peaks around 3620 and 3690 cm 1, due to clay minerals, were identified. This result is indicative of the presence of unfired raw material from the environment; in fact, these peaks are usually not observed in fired archaeological ceramics, where the high firing temperatures cause the dehydroxylation of the clays [12,47,48]. As already found in ATR spectra, the signals related to H—O—H vibrations of adsorbed water, that is, the band centered at 3400 cm 1 and the peak at 1630 cm 1, were stronger in the spectra of the internal side. The analyses carried out using the DRIFT technique allowed to detect just the most abundant compounds, that is, carbonates and silicates. Substances from both the pottery and the environment were detected, but evidence of the jar’s content was not clearly found. However, these results could be the consequence of the sampling method. The abrasion of the surface with an abrasive disk could be not able to collect the appropriate amount of residues. It is worth reminding that the substances contained in the jar could have been absorbed into the porous structure of the ceramic fabric, while residues remained on the surfaces only in a limited amount.
[(Fig._5)TD$IG]
Fig. 5. m-ATR spectrum collected on the external surface.
In the spectra acquired in transmission mode (Fig. 7a), the strong absorption band at 1032 cm 1 allowed to recognize silicate minerals as the main constituents of samples scraped off from the internal surfaces of the shards. The presence of quartz and iron oxide minerals was inferred from the characteristic doublet at 795 and 773 cm 1 and the signal around 520 cm 1, respectively. The peaks at 1420, 872 and 710 cm 1 indicated carbonate compounds in the samples. The presence of organic substances was identified by the peaks at 2954, 2925, 2850, 1730, 1457, and 1381 cm 1. The band at 1381 cm 1 could also be due to nitrate salts coming from burial soils. This hypothesis was not confirmed because the same peak was not observed in the spectra acquired on samples from the external surface. Therefore, the absorption band at 1381 cm 1 arose from a compound in contact only with the internal side of the amphora. On the outside of the shards greater amounts of carbonates and no evidence of organic materials, were found. In addition, as already observed in the spectra acquired in both ATR and DRIFT mode, peaks at 3616 and 3692 cm 1 testified to the presence of unfired clay minerals from the environment. Regarding the nature of the organic residues, position and shape of the related peaks suggested the presence of a vegetable resin [1]. The C—H stretching vibration modes of the methyl and the methylene group accounted for the peaks between 2850 and 2960 cm 1 [25]; the bands at 1457 and 1381 cm 1 were due to CH2 and CH3 bending; finally, the signal around 1730 cm 1 (ascribable to ketone groups) suggested the presence of resin derivatives [2], a likely result of aging processes that took place naturally or as a consequence of intended actions (such as the thermal treatment used in preparing pitch). Actually, rather than from substances in contact with the pottery during its use or preparation, organic matter can also be deposited throughout the burial period. The absence of significant signals due to organic compounds in the spectra of the samples collected on the external side excluded this event. Similar signals, that is, the same organic materials were identified inside the ceramic body, although only near the internal side of the shards. Quartz and iron oxides were constituent materials, because these minerals were recognized with no distinction between the inside and the outside of the ceramic body. Very low amounts of carbonates were identified in the ceramic material; their presence only on the outside surfaces indicated formation and deposition during the burial period.
[(Fig._6)TD$IG]
Fig. 6. Spectra collected in DRIFT mode.
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
[(Fig._7)TD$IG]
53
Fig. 7. Spectra of powder scraped off from either the shards’ surfaces or different depths into the ceramic body, collected in transmission mode: (a) the spectra in the full range of acquisition (400–4000 cm 1); b) the comparison in the range of 3000–3800 cm 1.
As can be seen from Fig. 7b, where the range between 3000 and 3800 cm 1 is expanded and compared, the band at 3425 cm 1 gradually decreased from the inside to the outside of the shards. Actually, a similar trend was observed for the peak at 1630 cm 1. These signals, due to water, very probably originated from hydration of minerals in the ceramic body. Since the exposure to water/humidity in the burial environment can be supposed identical for the whole shard, the stronger intensity could arise from a more prolonged contact of the internal surfaces with aqueous substances during the use of the vessels. The FT-IR spectra of liquid obtained by extraction with acetone from residues on the interior surfaces of the shards is reported in Fig. 8. Signals usually ascribed to terpenic compounds of vegetable resins [19,40] were observed. The —CH2 and —CH3 groups of the hydrocarbon skeleton in these resins gave rise to strong stretching vibrations at 2956, 2927 and 2854 cm 1 [3,23]. The bands at 1452 and 1381 cm 1 due to the —CH bending modes [24], were also recognized. The peaks in the range of 1265–1120 cm 1 addressed to different C—O vibration modes, while the stretching bands of carbonilic groups were observed at 1699 and 1722 cm 1. This latter (i.e., the peak at 1722 cm 1) suggested the presence of ketone groups, which were typically found in pitch [2,3,23]. The absence of peaks in the range of 1735–1745 cm 1, indicative of methyl esters of resin acids, should exclude the presence of wood tar [2].
[(Fig._8)TD$IG]
Fig. 8. Spectrum of liquid obtained by extraction with acetone from the residues, collected in transmission mode.
Actually, the lack of these signals would not be a conclusive test because methyl esters could have been hydrolyzed in free acids in the burial conditions [2]. Finally, the very strong absorption at 3450 and 1640 cm 1 can be due to water molecules soaked up during the tests or absorbed in the clay structure [5]. As previously stated, the GC–MS analyses, performed in another study [39], confirmed the presence of pine resin and/or pitch. he FT-IR analyses performed in transmission mode were successful in characterizing the constituent materials of the pottery. Also the residues from both the content of the jar and the burial environment were distinguished. Moreover, this technique allowed to have information about the distribution of the different compounds along the thickness of the shard. The presence of inorganic substances, even in small amounts, made difficult the exact identification of organic materials in the sample. A partial characterization of these compounds was possible only after extraction with a solvent. However, this procedure (i.e., the extraction) and the used solvent strongly affect the results. In fact, the analyses in transmission mode did not reveal the presence of tartrate salts which were easily identified by the m-ATR analyses. 4. Conclusions In this study a set of pottery shards classified as fragments of amphorae was analyzed by FT-IR spectroscopy. The results obtained from different sampling procedures and different analytical methodologies were examined and compared. Among the procedures used in this research, the sampling by abrasion of the surface with an abrasive disk, and consequently the spectroscopic analyses performed in DRIFT mode, were successful only in identifying the inorganic compounds coming from the pottery and the environment. The materials related to the content of the jar were not detected. These substances were trapped for the most part into hollows, gaps and pores of the pottery. Therefore, the sampling method used in this case allowed to collect the samples without damage, but precluded from taking a proper amount of material. Also the analyses in transmission mode on samples scraped as a powder provided information mainly about inorganic materials (i.e., constituent materials of the pottery, substances from the burial environment). The sampling inside the ceramic body was useful to investigate the distribution of the different compounds along the thickness of the shard. However, the presence of inorganic materials, even where in a limited amount, masked the signals of other organic compounds, usually found in traces. Just an
54
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
extraction with a solvent made it possible a more detailed, but still partial, characterization of these organic substances. The FT-IR analyses performed in m-ATR mode gave several advantages. Superficial substances were mainly recognized, the signals due to the ceramic body were weak in intensity and they did not hide the presence of other materials. In fact, the content of the jar was easily detected even in areas where no residue was observed. The m-ATR technique is a minimally invasive methodology, hence, it allows to use the same sample for further analytical investigations, or preserve it without damage. The m-ATR technique resulted really effective also in analyzing selected small pieces of materials taken from the archaeological samples. This sampling method can be applied in case of unbroken artifacts, when the archaeological objects cannot be moved into the laboratory and whenever other sampling procedures are difficult to perform. The findings of this study supported the FT-IR spectroscopy as a simple, fast and economical technique. The analytical results depend on the applied sampling method, but a limited or no preparation is usually suitable to obtain reliable and sound data. Finally, the versatility of the FT-IR technique is confirmed. In fact, the methods employed here allowed to investigate samples in different forms. Acknowledgements The author is grateful to Prof. Francesco D’Andria (Director of the Italian Archaeological Mission at Hierapolis) and Dr. Daniela Cottica (Department of Humanities, Ca' Foscari University of Venice) for the opportunity to analyze the artifacts from Hierapolis and for providing the archaeological samples. Thanks go to Dr. Florinda Notarstefano (Department of Cultural Heritage, University of Salento, Lecce) for her helpful suggestions in the interpretation of the results. References [1] M.R. Derrick, D.C. Stulik, J.M. Landry, Infrared Spectroscopy in Conservation Science, The Getty Conservation Institute, Los Angeles, 1999. [2] J. Font, N. Salvadó, S. Butí, J. Enrich, Anal. Chim. Acta 598 (2007) 119–127. [3] L.M. Shillito, M.J. Almond, K. Wicks, L.J.R. Marshall, W. Matthews, Spectrochim. Acta A 72 (2009) 120–125. [4] T.F.M. Oudemans, J.J. Boon, R.E. Botto, Archaeometry 49 (2007) 571–594. [5] G. Tarquini, S. Nunziante Cesaro, L. Campanella, Talanta 118 (2014) 195–200. [6] F. Mizzoni, S. Nunziante Cesaro, Spectrochim. Acta A 68 (2007) 377–381. [7] R. Ravisankar, S. Kiruba, C. Shamira, A. Naseerutheen, P.D. Balaji, M. Seran, Microchem. J. 99 (2011) 370–375. [8] G. Velraj, R. Ramya, R. Hemamalini, J. Mol. Struct. 1028 (2012) 16–21. [9] M.J. Ayora-Cañada, A. Domínguez-Arranz, A. Dominguez-Vidal, J. Raman Spectrosc. 43 (2012) 317–322. [10] S.A. Centeno, V.I. Williams, N.C. Little, R.J. Speakman, Vib. Spectrosc. 58 (2012) 119–124. [11] G.A. Mazzocchin, F. Agnoli, I. Colpo, Anal. Chim. Acta 478 (2003) 147–161. [12] R. Ravisankar, A. Naseerutheen, G. Raja Annamalai, A. Chandrasekaran, A. Rajalakshmi, K.V. Kanagasabapathy, M.V.R. Prasad, K.K. Satpathy, Spectrochim. Acta A 121 (2014) 457–463.
[13] G. Barone, S. Ioppolo, D. Majolino, P. Migliardo, G. Spagnolo, J. Cult. Herit. 3 (2002) 171–176. [14] D. Barilaro, G. Barone, V. Crupi, D. Majolino, P. Mazzoleni, M. Triscari, V. Venuti, Vib. Spectrosc. 42 (2006) 381–386. [15] R.P. Evershed, Archaeometry 50 (2008) 895–924. [16] R.P. Evershed, C. Heron, L.J. Goad, Analyst 115 (1990) 1339–1342. [17] M. Regert, Mass Spectrom. Rev. 30 (2011) 177–220. [18] S. Mitkidou, E. Dimitrakoudi, D. Urem-Kotsou, D. Papadopoulou, K. Kotsakis, J. A. Stratis, I. Stephanidou-Stephanatou, Microchim. Acta 160 (2008) 493–498. [19] M.P. Colombini, G. Giachi, F. Modugno, E. Ribechini, Microchem. J. 79 (2005) 83–90. [20] D. Arobba, F. Bulgarelli, F. Camin, R. Caramiello, R. Larcher, L. Martinelli, J. Archaeol. Sci. 45 (2014) 226–233. [21] K. Romanus, J. Baeten, J. Poblome, S. Accardo, P. Degryse, P. Jacobs, D. De Vos, M. Waelkens, J. Archaeol. Sci. 36 (2009) 900–909. [22] F. Dorrego, F. Carrera, M.P. Luxàn, Mater. Struct. 37 (2004) 369–374. [23] F.C. Izzo, E. Zendri, A. Bernardi, E. Balliana, M. Sgobbi, J. Archaeol. Sci. 40 (2013) 595–600. [24] E. Ribechini, F. Modugno, C. Baraldi, P. Baraldi, M.P. Colombini, Talanta 74 (2008) 555–561. [25] S. Zareva, I. Kuleff, Spectrochim. Acta A 76 (2010) 283–286. [26] P.E. McGovern, D.L. Glusker, L.J. Exner, M.M. Voigt, Nature 381 (1996) 480–481. [27] C. Cardell, I. Guerra, J. Romero-Pastor, G. Cultrone, A. Rodriguez-Navarro, Anal. Chem. 81 (2009) 604–611. [28] J. Baeten, K. Romanus, P. Degryse, W. De Clercq, H. Poelman, K. Verbeke, A. Luypaerts, M. Walton, P. Jacobs, D. De Vos, M. Waelkens, Microchem. J. 95 (2010) 227–234. [29] F. D’Andria, M.P. Caggia, I. Hierapolis di Frigia, Le Attivita delle Campagne di Scavo e Restauro 2000–2003, Ege Yayinlari, Istanbul, 2007. [30] D. Cottica, Hierapolis di Frigia I. Le attività delle campagne di scavo e restauro 2000–2003, in: F. D’Andria, M.P. Caggia (Eds.), MAIER–Missione Archeologica Italiana a Hierapolis, Ege Yayinlari, Istanbul, 2007, pp. 257–277. [31] M. Lettieri, F. Notarstefano, Il Dialogo dei Saperi. Metodologie integrate per i Beni Culturali, in: F. D’Andria, D. Malfitana, N. Masini, G. Scardozzi (Eds.), Edizioni scientifiche italiane, Napoli, 2010, pp. 337–352. [32] S. Prati, G. Sciutto, R. Mazzeo, C. Torri, D. Fabbri, Anal. Bioanal. Chem. 399 (2011) 3081–3091. [33] P.E. McGovern, D.L. Glusker, R.A. Moreau, A. Nunez, C.W. Beck, E. Simpson, E.D. Butrym, L.J. Exner, E.C. Stout, Nature 402 (1999) 863–864. [34] P.E. McGovern, J. Egypt. Archaeol. 83 (1997) 69–108. [35] X. Sahaya Shajan, C. Mahadevan, Cryst. Res. Technol. 40 (2005) 598–602. [36] M.R. Guasch-Jané, C. Andrés-Lacueva, O. Jáuregui, R.M. Lamuela-Raventós, J. Archaeol. Sci. 33 (2006) 98–101. [37] R.H. Michel, P.E. McGovern, V.R. Badler, Anal. Chem. 65 (1993) 408A–413A. [38] A. Pecci, G. Giorgi, L. Salvini, M.A. Cau Ontiveros, J. Archaeol. Sci. 40 (2013) 109–115. [39] F. Notarstefano, Ceramica e alimentazione: l'analisi chimica dei residui organici nelle ceramiche applicata ai contesti archeologici, Edipuglia, Bari, 2012. [40] M.P. Colombini, G. Giachi, F. Modugno, P. Pallecchi, E. Ribechini, Archaeometry 45 (2003) 659–674. [41] M.P. Colombini, F. Modugno, E. Ribechini, J. Mass Spectrom. 40 (2005) 675–687. [42] E. Ribechini, S. Orsini, F. Silvano, M.P. Colombini, Anal. Chim. Acta 638 (2009) 79–87. [43] R.L. Frost, Clays Clay Miner. 43 (1995) 191–195. [44] J. Madejová, Vib. Spectrosc. 31 (2003) 1–10. [45] M.A. Cau Ontiveros, P.M. Day, G. Montana, Modern Trends in Scientific Studies on Ancient Ceramics, in: V. Kilikoglou, A. Hein, Y. Maniatis (Eds.), Archaeopress, Oxford, 2002, pp. 9–18. [46] E. Andaloro, C.M. Belfiore, A.M. De Francesco, J.K. Jacobsen, G.P. Mittica, Appl. Clay Sci. 53 (2011) 445–453. [47] S. Shoval, Opt. Mater. 24 (2003) 117–122. [48] F. Berna, A. Behar, R. Shahack-Gross, J. Berg, E. Boaretto, A. Gilboa, I. Sharon, S. Shalev, S. Shilstein, N. Yahalom-Mack, J.R. Zorn, S. Weiner, J. Archaeol. Sci. 34 (2007) 358–373.
Contents lists available at ScienceDirect
Vibrational Spectroscopy journal homepage: www.elsevier.com/locate/vibspec
Infrared spectroscopic characterization of residues on archaeological pottery through different spectra acquisition modes Mariateresa Lettieri * Institute of Archaeological Heritage – Monuments and Sites, CNR–IBAM, Prov.le Lecce-Monteroni, 73100 Lecce, Italy
A R T I C L E I N F O
A B S T R A C T
Article history: Received 9 September 2014 Received in revised form 16 December 2014 Accepted 17 December 2014 Available online 19 December 2014
Fourier transform infrared spectroscopy (FT-IR) is a versatile analytical method, very useful in many fields. Although a crucial step in producing good spectra is the use of the appropriate technique, the acquisition mode is sometimes not accurately selected and the results are partial or lacking. In recent years, FT-IR analysis has been proposed as a screening method for characterization of archaeological potteries and identification of the residues on these artifacts before turning to destructive, more expensive, and time-consuming techniques. In this study, a set of pottery shards, classified as fragments of amphorae, was subjected to FT-IR analyses. The results obtained from different sampling procedures and different spectra acquisition modes, were examined and compared. The as-received ceramic fragments were subjected to micro attenuated total reflectance (m-ATR) analyses. Investigations in diffuse reflectance (DRIFT) mode were performed on samples collected by abrading the surfaces of the shards with abrasive paper. Samples scraped from either the surfaces of the pottery fragments or the interior of the ceramic body, were analyzed in transmission mode as a powder in KBr pellets or after extraction with acetone. The sampling by abrasion of the surface with an abrasive disk, and consequently the analyses in DRIFT mode, were successful only in identifying the inorganic compounds coming from the pottery and/or the environment, while materials related to the content of the jar were not detected. Also the analyses in transmission mode provided information mainly about inorganic materials, which, even where in a limited amount, masked the signals of organic compounds. Just an extraction with a solvent made it possible a more detailed, but still partial, characterization of these organic substances. On the contrary, the content of the jar was easily detected using the m-ATR mode, even in areas where no residue was observed. ã 2014 Elsevier B.V. All rights reserved.
Keywords: Fourier transform infrared spectroscopy Coated surfaces Archaeometry Residue analysis Archaeological pottery
1. Introduction Fourier transform infrared spectroscopy (FT-IR) has become a very popular analytical technique, for many years now. This method is fast, relatively cheap, and easy to use, and because of these advantages, it is applied to the study of a wide range of materials in many fields. A crucial step in producing good FT-IR spectra is the use of the appropriate technique for presenting the sample to the spectrometer [1]. The selection of the suitable analysis method should be based on the type, form and amount of sample to be analyzed. The physical state of the sample, its preparation, and the analysis method have an effect on the resulting spectrum, modifying the
* Tel.: +39 0832 422219; fax: +39 0832 422225. E-mail address: [email protected] (M. Lettieri). http://dx.doi.org/10.1016/j.vibspec.2014.12.002 0924-2031/ ã 2014 Elsevier B.V. All rights reserved.
absorption pattern. In fact, the above mentioned factors influence the peak position, as well as the band shape and intensity. Nevertheless, the spectra acquisition mode is sometimes not accurately selected and, as a result, the information that can be gleaned from the spectra are partial or lacking. In recent years, FT-IR spectroscopy has been proposed as a quick and cheap method to screen archaeological samples before subjecting them to more expensive and time-consuming methods (e.g., gas chromatography–mass spectroscopy) [2–6]. Most of the studies concerning such research topics proved the effectiveness of FT-IR analyses for the detection of organic residues in archaeological materials, as well as for investigations of the fabrication conditions [7–11] and for provenance studies [12–14]. Since ancient time, ceramic artifacts have been largely employed for storing materials, as well as for preparing or cooking food. As a result, traces of the materials which were in
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
49
contact with the vessels can be found as surface deposits or as adsorbed residues into the porous structure [15]. The characterization of these residues is very helpful to provide information on cultural, technological, and commercial activities of ancient societies. Chromatographic techniques are frequently employed in identifying the organic residues [15–21]. However, these methodologies are destructive, while non-destructive or microdestructive diagnostic techniques would be preferred, and procedures for sample preparation, usually taking a long time, are required. Published papers dealing with characterization of archaeological potteries and identification of the residues in ceramic vessels, frequently discuss the application of FT-IR spectroscopy in transmittance mode [7,8,19,22–25]. Analyses performed in diffuse reflectance (DRIFT) [5,26,27] or in attenuated total reflectance (ATR) mode (sometimes coupled with a microscope, that is, the m-ATR mode) [6,9,28] are also used in these studies. However, despite the fact that the results depend on the sample, on its collection and preparation, as well as on the analysis methods, the different acquisition modes have been frequently used with no distinction. Consequently, the characterization of the materials under study has been not comprehensive and the FT-IR technique has been undervalued and underused. Starting from these issues, the present research was aimed at investigating residues on archaeological ceramic samples by means of FT-IR spectroscopy. A set of pottery shards, classified as fragments of amphorae, was subjected to analyses carried out in m-ATR mode, in DRIFT mode and in transmittance mode. The results obtained from different sampling procedures and different spectra acquisition modes were examined and compared.
kinds of analysis. Consequently, several characterization tests can be carried out on the same sample. In the next step, DRIFT analyses were performed on samples collected by abrading the surfaces of the shards with abrasive paper, as detailed in Section 2.4.2. In this manner, small amounts of sample can be taken from the surfaces with insignificant damage. Furthermore, the tools for sampling are portable and can be easily used without transferring in the laboratory the items under investigation. Subsequently, FT-IR analyses in transmission mode were carried out on samples scraped from the surfaces of the pottery fragments. This methodology made it possible to collect, and then analyze, materials from different parts of the archaeological fragment, with no limitation due to the original shape, dimensions or form of the sample. In fact, the presence of residues from the jar’s content was checked also inside the ceramic body. Moreover, FT-IR analyses in transmission mode were performed after extraction with acetone to identify the organic residues without interferences from the ceramic material. Finally, small flakes of the residues on the internal surfaces of the shards were analyzed in m-ATR mode. This method allows selective examination of the residue, particularly when other sampling procedures are difficult to perform or the archaeological artifacts cannot be moved into the laboratory. The analyses were carried out in the reported order, so that the destructive or micro-destructive samplings were performed last, because the removal of the residues from the surfaces could make ineffective the investigations performed directly on the shards (i.e., the m-ATR analyses).
2. Materials and methods
A FT-IR ThermoNicolet Nexus spectrometer equipped with a deuterated triglycine sulfate (DTGS) detector, was used to record the FT-IR spectra in transmission mode. A 13 mm KBr die (Model 129, Thermo Spectra-Tech) and a hydraulic press (Mod 660, Silfradent) were employed to shape the pellets for the analyses. Following the manufacturer’s suggestion, 6000 kg load was applied for approximately 1 min. A Smart Diffuse Reflectance accessory (ThermoNicolet) was inserted in the spectrometer to perform the DRIFT analyses. This module included a slide with a slot that can hold the sampling disk. A kit consisting of handle and plate (which keep the adhesivebacked silicon-carbide paper used to abrade the surface) was employed in the sampling. The ThermoNicolet Continuum IR microscope coupled with the spectrometer was used to acquire the FT-IR spectra. This was equipped with a mercury–cadmium–telluride (MCT) detector, which was cooled with liquid nitrogen. A 15 Reflachromat objective with a slide-on ATR attachment (Thermo Spectra-Tech), using a Si crystal (refractive index = 3.4; incident angle = 45 ; contact area = 50 50 mm), was employed to collect the m-ATR spectra. To ensure reproducibility and uniformity, the contact between the ATR crystal and the sample surface was automated and computer controlled. After each analysis, the crystal was cleaned with a soft cloth soaked in acetone. The employed instrumentations were not purged with dry and CO2-free air, therefore the contribute of absorption bands of both CO2 (appearing in the spectrum as a doublet around 2340 cm 1) and water vapor (resulting in sharp and very close peaks over 3700 cm 1) are observed in the reported spectra, although a background spectrum was acquired before each analysis and automatically subtracted (by software) from the sample spectrum. All the FT-IR data were processed with the OMNIC 8.1 software (Thermo Fisher Scientific Inc.).
2.1. Archaeological samples The investigated samples were collected from a set of pottery shards discovered in the archaeological site of Hierapolis (modern Pamukkale, Turkey) [29]. Archeologists classified these ceramic objects as fragments of amphorae manufactured and used in the early-Byzantine period (V–VI century A.D.) [30,31]. The unearthed shards were only unsoiled with a soft brush, while washing with either water or any other cleaning agent, was avoided. The examined fragments have different shapes and dimensions (not exceeding 5 5 cm2), with an average thickness of 0.5 cm. Their concave side (hereinafter called internal side) always was the interior surface of the jar. Shards with residues distinguishable to the naked-eye onto the internal side, were chosen. The observed residues were blackish and/or brownish in color and exhibited a quite good adhesion to the shards, but they never fully coated the pottery surface. An example of the pottery shards under investigation is shown in Fig. 1. 2.2. Sequence of the analyses A preliminary visual inspection of the pottery surfaces was performed through a binocular stereomicroscope (Zeiss, mod. Stemi SV11) at magnifications of up to 100. This examination was aimed at distinguishing the residues from the pottery material, in order to optimize the sampling. The as-received ceramic fragments were firstly subjected to the m-ATR analyses. The shape and the limited dimension of the shards examined in this study allowed a direct analysis. In this case, sampling or preparation procedures are not required, hence the specimen is not damaged at all and it can be used later for other
2.3. FT-IR instrumentation
50
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
[(Fig._1)TD$IG]
Fig. 1. A shard of amphora from Hierapolis: (a) internal side; (b) external side.
2.4. Sampling and analytical procedures 2.4.1. m-ATR mode The m-ATR analyses were performed on both the intact pottery shards and the flakes taken from the surfaces, placing the samples directly under the m-ATR objective. The spectra were collected with the Si crystal in the range of 4000–650 cm 1, with a resolution of 4 cm 1 and 200 scans for each measurement; the background spectrum was acquired in air. In the case of the shards, the spectra were acquired on the visible residues, as well as on apparently clean areas. Furthermore, the analyses were carried out on both the internal and external side of the shard, in order to avoid misinterpretation due to the presence of substances coming from the deposition soils. 2.4.2. DRIFT mode The samples for the tests in DRIFT mode were gathered by gentle abrasion with a small circle of abrasive paper (320 grit SiCarb, Thermo Spectra-Tech) mounted on a stick. The DRIFT analysis was carried out soon after the sampling. The spectra were acquired in the range of 4000–400 cm 1, with a resolution of 4 cm 1 and 200 scans per measurement; the background spectrum was collected on the same abrasive paper disk afterwards used for the sampling. For comparison, the sampling and the analyses were performed on both the internal and external sides of the pottery fragment. 2.4.3. Transmission mode The internal and the external surfaces of the shards were scraped with a scalpel to collect the samples as a powder to be used for the analyses in transmission mode. In the same way, the matter was taken also from the lateral side of the shard, at different depths, to investigate the presence of residues due to the jar’s content inside the ceramic body. These samples were finely ground with a pestle in an agate mortar, then mixed with KBr (suitable for
infrared analysis and provided by Mallinckrodt Baker Chemical Inc.) and compacted in a pellet 13 mm in diameter. The material gathered from the internal surface of the shard was divided in two parts. The first one was used to prepare KBr pellets. The second part, placed in a vial, was extracted with acetone (analytical grade, provided by Carlo Erba Reagents), using 1 ml of solvent per 5 mg of sample. Sonication was carried out for 22 min in an ultrasonic bath (FALC Instruments), followed by standing for 24 h in laboratory conditions. A drop of the obtained liquid fraction was placed on a KBr pellet, previously prepared, which was stored for 5 min at 40 C to enhance the solvent evaporation. All the KBr pellets were analyzed in transmission mode, immediately after the preparation. The spectra were acquired in the range of 4000–400 cm 1, with a resolution of 4 cm 1 and 200 scans per measurement; the background spectrum was collected on a pellet made of KBr only. 3. Results and discussion The observations through the stereomicroscope, preliminary to the samplings, allowed to identify residues having dissimilar features. Areas different in appearance were noticed on the surfaces, as described below: a Brown coatings; b Dark-brown residues filling holes and gaps of the surface; c Surfaces with no visible residue, where the ceramic body was
easily recognized; d Whitish incrustations.
In all the examined archaeological samples, the first three types of residues (1–3) were observed on the internal surface of the shards, as shown in Fig. 2a. Instead, the whitish incrustations (4) were mainly found on the external side (see Fig. 2b).
[(Fig._2)TD$IG]
Fig. 2. Images of the potsherd taken by means of the stereomicroscope: (a) internal side; (b) external side.
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
3.1. m-ATR mode The investigations carried out in m-ATR mode on the internal surfaces of the shards were very significant. As previously stated, spectra were acquired directly on visible residues, as well as on apparently clean areas, where the visual inspection, either to the naked-eye or through the stereomicroscope, did not reveal any residue. The most of the analyses of the visible residues gave rise to spectra where the bands at 2952, 2928, 2863, 1710, 1448, and 1172 cm 1 indicated the presence of vegetable resins [1,6,32], even if the peak around 1033 cm 1, due to the silicate compounds of the pottery, was the strongest signal (Fig. 3a-1). In these spectra, significant signals at 1584, 1385, and 1229 cm 1 were also detected. In other cases, actually less frequent, the spectra of the residues on the internal side of the shards exhibited bands around 1572, 1386 and 1228 cm 1 as main peaks (Fig. 3b). Other researchers attributed similar signals to crystals of tartrate salts [26,33–35]. In fact, tartrates, usually as calcium or potassium salts, can derive from tartaric acid occurring in large amounts in grapes and have been proposed in many studies as a marker of wine in archaeological potteries [36–38]. The results of the analyses performed in m-ATR mode on flakes of residues are reported in Fig. 3a-2. These spectra resulted very similar to those collected directly on visible residues (area (1) in Fig. 2a). These findings were supported by the results of the GC–MS analyses on the same samples as discussed in another study [39]. Abietic acid, dehydroabietic acid and 7-oxo-dehydroabietic acid were identified in the residues, indicating the presence of a pine resin and/or a pitch obtained from resinous materials of Pinaceae trees [25,38,40–42]. Tartrates was not detected, but it is to take into account that GC–MS technique can be not suitable to unambiguously prove the existence of tartaric compounds in archaeological samples [21]. The m-ATR spectrum acquired on clean areas is shown in Fig. 4. Although the main peak was due to silicate compounds of the ceramic body, organic matter ascribable to the content of the jar was recognized. Vegetable resins can be identified from the signals at 2960, 2925, 2866, 1712, and 1449 cm 1, while the peaks around 1586 and 1385 cm 1 could account for the presence of calcium tartrate. Traces of carbonates were also detected (1416 and 872 cm 1). On the outside of the shards greater amounts of carbonates (1415 and 872 cm 1) and no evidence of organic materials, were
51
found (Fig. 5). In addition, peaks at 3620 and 3690 cm 1, due to the OH stretching modes [43] of clay minerals belonging to the kaolinite-group [44], were detected. Analogous signals were not noticed in the spectra of the inner surface. Finally, it is to remark that the broad band around 3400 cm 1 and the peak centered at 1625 cm 1, both related to H—O—H vibrations of adsorbed water [44], were weaker in the spectra of the external side. Using the FT-IR analyses in m-ATR mode, the superficial substances were mainly recognized, while the absorption bands deriving from the ceramic body were not strong in intensity. As a consequence, the signals due to residues were not masked and the related substances were quite easily identified. Furthermore, residues from the content of the jar were detected even in areas where no residue was revealed by the visual observation. Shape and dimensions of the investigated object could impede a suitable spectral acquisition, but the ATR technique gave meaningful results also in analyzing selected small pieces of materials taken from the shard. This latter result makes the technique valuable when the sampling is carried out on unbroken artifacts. 3.2. DRIFT mode The results of the analyses performed in DRIFT mode are reported in Fig. 6. The bands due to silicate (1025 cm 1) and carbonate (1450 and 874 cm 1) compounds are the main peaks in the spectra of the material sampled from the internal surface of the shards. The peaks at 2960, 2920, and 2850 cm 1 suggest the presence of organic substances, even if no further information can be inferred from these data. However, these signals cannot be unambiguously attributed to organic materials, because peaks between 2850 and 3000 cm 1, similar in shape, are observed also in the standard reference DRIFT spectrum (HR Nicolet Sampler Library, Thermo Fisher Scientific Inc.) of calcium carbonate. Reasonably the external side of the potsherds did not have contact with the content, therefore, for comparison purpose, analyses in DRIFT mode were performed on specimens taken from these surfaces. In comparing the spectra (Fig. 6), peaks in the range of 2850– 2960 cm 1 were still noticed, even if some differences can be observed. Greater amounts of carbonates were found on the external surfaces, because the peaks at 1450 and 874 cm 1 were stronger in intensity. Furthermore, in both cases the overtone bands of the carbonate group were observed at 2516 and
[(Fig._3)TD$IG]
Fig. 3. m-ATR spectra of residues on the internal surfaces: (a) 1 = recurrent result in areas with visible residues, 2 = spectrum acquired on a flake of residue (shown in the inset box) taken from the internal side of a shard; (b) less frequent result in areas with visible residues.
52
[(Fig._4)TD$IG]
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
3.3. Transmission mode
Fig. 4. m-ATR spectrum collected in an area of the internal surface without visible residues.
1796 cm 1 [3]. Other researchers attributed similar results to carbonate compounds in a powdered state [5] from layers covering the outside of the shards, realistically due to deposition from circulating fluids taking place in the burial conditions [45,46]. In the spectra collected on samples from the external side, peaks around 3620 and 3690 cm 1, due to clay minerals, were identified. This result is indicative of the presence of unfired raw material from the environment; in fact, these peaks are usually not observed in fired archaeological ceramics, where the high firing temperatures cause the dehydroxylation of the clays [12,47,48]. As already found in ATR spectra, the signals related to H—O—H vibrations of adsorbed water, that is, the band centered at 3400 cm 1 and the peak at 1630 cm 1, were stronger in the spectra of the internal side. The analyses carried out using the DRIFT technique allowed to detect just the most abundant compounds, that is, carbonates and silicates. Substances from both the pottery and the environment were detected, but evidence of the jar’s content was not clearly found. However, these results could be the consequence of the sampling method. The abrasion of the surface with an abrasive disk could be not able to collect the appropriate amount of residues. It is worth reminding that the substances contained in the jar could have been absorbed into the porous structure of the ceramic fabric, while residues remained on the surfaces only in a limited amount.
[(Fig._5)TD$IG]
Fig. 5. m-ATR spectrum collected on the external surface.
In the spectra acquired in transmission mode (Fig. 7a), the strong absorption band at 1032 cm 1 allowed to recognize silicate minerals as the main constituents of samples scraped off from the internal surfaces of the shards. The presence of quartz and iron oxide minerals was inferred from the characteristic doublet at 795 and 773 cm 1 and the signal around 520 cm 1, respectively. The peaks at 1420, 872 and 710 cm 1 indicated carbonate compounds in the samples. The presence of organic substances was identified by the peaks at 2954, 2925, 2850, 1730, 1457, and 1381 cm 1. The band at 1381 cm 1 could also be due to nitrate salts coming from burial soils. This hypothesis was not confirmed because the same peak was not observed in the spectra acquired on samples from the external surface. Therefore, the absorption band at 1381 cm 1 arose from a compound in contact only with the internal side of the amphora. On the outside of the shards greater amounts of carbonates and no evidence of organic materials, were found. In addition, as already observed in the spectra acquired in both ATR and DRIFT mode, peaks at 3616 and 3692 cm 1 testified to the presence of unfired clay minerals from the environment. Regarding the nature of the organic residues, position and shape of the related peaks suggested the presence of a vegetable resin [1]. The C—H stretching vibration modes of the methyl and the methylene group accounted for the peaks between 2850 and 2960 cm 1 [25]; the bands at 1457 and 1381 cm 1 were due to CH2 and CH3 bending; finally, the signal around 1730 cm 1 (ascribable to ketone groups) suggested the presence of resin derivatives [2], a likely result of aging processes that took place naturally or as a consequence of intended actions (such as the thermal treatment used in preparing pitch). Actually, rather than from substances in contact with the pottery during its use or preparation, organic matter can also be deposited throughout the burial period. The absence of significant signals due to organic compounds in the spectra of the samples collected on the external side excluded this event. Similar signals, that is, the same organic materials were identified inside the ceramic body, although only near the internal side of the shards. Quartz and iron oxides were constituent materials, because these minerals were recognized with no distinction between the inside and the outside of the ceramic body. Very low amounts of carbonates were identified in the ceramic material; their presence only on the outside surfaces indicated formation and deposition during the burial period.
[(Fig._6)TD$IG]
Fig. 6. Spectra collected in DRIFT mode.
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
[(Fig._7)TD$IG]
53
Fig. 7. Spectra of powder scraped off from either the shards’ surfaces or different depths into the ceramic body, collected in transmission mode: (a) the spectra in the full range of acquisition (400–4000 cm 1); b) the comparison in the range of 3000–3800 cm 1.
As can be seen from Fig. 7b, where the range between 3000 and 3800 cm 1 is expanded and compared, the band at 3425 cm 1 gradually decreased from the inside to the outside of the shards. Actually, a similar trend was observed for the peak at 1630 cm 1. These signals, due to water, very probably originated from hydration of minerals in the ceramic body. Since the exposure to water/humidity in the burial environment can be supposed identical for the whole shard, the stronger intensity could arise from a more prolonged contact of the internal surfaces with aqueous substances during the use of the vessels. The FT-IR spectra of liquid obtained by extraction with acetone from residues on the interior surfaces of the shards is reported in Fig. 8. Signals usually ascribed to terpenic compounds of vegetable resins [19,40] were observed. The —CH2 and —CH3 groups of the hydrocarbon skeleton in these resins gave rise to strong stretching vibrations at 2956, 2927 and 2854 cm 1 [3,23]. The bands at 1452 and 1381 cm 1 due to the —CH bending modes [24], were also recognized. The peaks in the range of 1265–1120 cm 1 addressed to different C—O vibration modes, while the stretching bands of carbonilic groups were observed at 1699 and 1722 cm 1. This latter (i.e., the peak at 1722 cm 1) suggested the presence of ketone groups, which were typically found in pitch [2,3,23]. The absence of peaks in the range of 1735–1745 cm 1, indicative of methyl esters of resin acids, should exclude the presence of wood tar [2].
[(Fig._8)TD$IG]
Fig. 8. Spectrum of liquid obtained by extraction with acetone from the residues, collected in transmission mode.
Actually, the lack of these signals would not be a conclusive test because methyl esters could have been hydrolyzed in free acids in the burial conditions [2]. Finally, the very strong absorption at 3450 and 1640 cm 1 can be due to water molecules soaked up during the tests or absorbed in the clay structure [5]. As previously stated, the GC–MS analyses, performed in another study [39], confirmed the presence of pine resin and/or pitch. he FT-IR analyses performed in transmission mode were successful in characterizing the constituent materials of the pottery. Also the residues from both the content of the jar and the burial environment were distinguished. Moreover, this technique allowed to have information about the distribution of the different compounds along the thickness of the shard. The presence of inorganic substances, even in small amounts, made difficult the exact identification of organic materials in the sample. A partial characterization of these compounds was possible only after extraction with a solvent. However, this procedure (i.e., the extraction) and the used solvent strongly affect the results. In fact, the analyses in transmission mode did not reveal the presence of tartrate salts which were easily identified by the m-ATR analyses. 4. Conclusions In this study a set of pottery shards classified as fragments of amphorae was analyzed by FT-IR spectroscopy. The results obtained from different sampling procedures and different analytical methodologies were examined and compared. Among the procedures used in this research, the sampling by abrasion of the surface with an abrasive disk, and consequently the spectroscopic analyses performed in DRIFT mode, were successful only in identifying the inorganic compounds coming from the pottery and the environment. The materials related to the content of the jar were not detected. These substances were trapped for the most part into hollows, gaps and pores of the pottery. Therefore, the sampling method used in this case allowed to collect the samples without damage, but precluded from taking a proper amount of material. Also the analyses in transmission mode on samples scraped as a powder provided information mainly about inorganic materials (i.e., constituent materials of the pottery, substances from the burial environment). The sampling inside the ceramic body was useful to investigate the distribution of the different compounds along the thickness of the shard. However, the presence of inorganic materials, even where in a limited amount, masked the signals of other organic compounds, usually found in traces. Just an
54
M. Lettieri / Vibrational Spectroscopy 76 (2015) 48–54
extraction with a solvent made it possible a more detailed, but still partial, characterization of these organic substances. The FT-IR analyses performed in m-ATR mode gave several advantages. Superficial substances were mainly recognized, the signals due to the ceramic body were weak in intensity and they did not hide the presence of other materials. In fact, the content of the jar was easily detected even in areas where no residue was observed. The m-ATR technique is a minimally invasive methodology, hence, it allows to use the same sample for further analytical investigations, or preserve it without damage. The m-ATR technique resulted really effective also in analyzing selected small pieces of materials taken from the archaeological samples. This sampling method can be applied in case of unbroken artifacts, when the archaeological objects cannot be moved into the laboratory and whenever other sampling procedures are difficult to perform. The findings of this study supported the FT-IR spectroscopy as a simple, fast and economical technique. The analytical results depend on the applied sampling method, but a limited or no preparation is usually suitable to obtain reliable and sound data. Finally, the versatility of the FT-IR technique is confirmed. In fact, the methods employed here allowed to investigate samples in different forms. Acknowledgements The author is grateful to Prof. Francesco D’Andria (Director of the Italian Archaeological Mission at Hierapolis) and Dr. Daniela Cottica (Department of Humanities, Ca' Foscari University of Venice) for the opportunity to analyze the artifacts from Hierapolis and for providing the archaeological samples. Thanks go to Dr. Florinda Notarstefano (Department of Cultural Heritage, University of Salento, Lecce) for her helpful suggestions in the interpretation of the results. References [1] M.R. Derrick, D.C. Stulik, J.M. Landry, Infrared Spectroscopy in Conservation Science, The Getty Conservation Institute, Los Angeles, 1999. [2] J. Font, N. Salvadó, S. Butí, J. Enrich, Anal. Chim. Acta 598 (2007) 119–127. [3] L.M. Shillito, M.J. Almond, K. Wicks, L.J.R. Marshall, W. Matthews, Spectrochim. Acta A 72 (2009) 120–125. [4] T.F.M. Oudemans, J.J. Boon, R.E. Botto, Archaeometry 49 (2007) 571–594. [5] G. Tarquini, S. Nunziante Cesaro, L. Campanella, Talanta 118 (2014) 195–200. [6] F. Mizzoni, S. Nunziante Cesaro, Spectrochim. Acta A 68 (2007) 377–381. [7] R. Ravisankar, S. Kiruba, C. Shamira, A. Naseerutheen, P.D. Balaji, M. Seran, Microchem. J. 99 (2011) 370–375. [8] G. Velraj, R. Ramya, R. Hemamalini, J. Mol. Struct. 1028 (2012) 16–21. [9] M.J. Ayora-Cañada, A. Domínguez-Arranz, A. Dominguez-Vidal, J. Raman Spectrosc. 43 (2012) 317–322. [10] S.A. Centeno, V.I. Williams, N.C. Little, R.J. Speakman, Vib. Spectrosc. 58 (2012) 119–124. [11] G.A. Mazzocchin, F. Agnoli, I. Colpo, Anal. Chim. Acta 478 (2003) 147–161. [12] R. Ravisankar, A. Naseerutheen, G. Raja Annamalai, A. Chandrasekaran, A. Rajalakshmi, K.V. Kanagasabapathy, M.V.R. Prasad, K.K. Satpathy, Spectrochim. Acta A 121 (2014) 457–463.
[13] G. Barone, S. Ioppolo, D. Majolino, P. Migliardo, G. Spagnolo, J. Cult. Herit. 3 (2002) 171–176. [14] D. Barilaro, G. Barone, V. Crupi, D. Majolino, P. Mazzoleni, M. Triscari, V. Venuti, Vib. Spectrosc. 42 (2006) 381–386. [15] R.P. Evershed, Archaeometry 50 (2008) 895–924. [16] R.P. Evershed, C. Heron, L.J. Goad, Analyst 115 (1990) 1339–1342. [17] M. Regert, Mass Spectrom. Rev. 30 (2011) 177–220. [18] S. Mitkidou, E. Dimitrakoudi, D. Urem-Kotsou, D. Papadopoulou, K. Kotsakis, J. A. Stratis, I. Stephanidou-Stephanatou, Microchim. Acta 160 (2008) 493–498. [19] M.P. Colombini, G. Giachi, F. Modugno, E. Ribechini, Microchem. J. 79 (2005) 83–90. [20] D. Arobba, F. Bulgarelli, F. Camin, R. Caramiello, R. Larcher, L. Martinelli, J. Archaeol. Sci. 45 (2014) 226–233. [21] K. Romanus, J. Baeten, J. Poblome, S. Accardo, P. Degryse, P. Jacobs, D. De Vos, M. Waelkens, J. Archaeol. Sci. 36 (2009) 900–909. [22] F. Dorrego, F. Carrera, M.P. Luxàn, Mater. Struct. 37 (2004) 369–374. [23] F.C. Izzo, E. Zendri, A. Bernardi, E. Balliana, M. Sgobbi, J. Archaeol. Sci. 40 (2013) 595–600. [24] E. Ribechini, F. Modugno, C. Baraldi, P. Baraldi, M.P. Colombini, Talanta 74 (2008) 555–561. [25] S. Zareva, I. Kuleff, Spectrochim. Acta A 76 (2010) 283–286. [26] P.E. McGovern, D.L. Glusker, L.J. Exner, M.M. Voigt, Nature 381 (1996) 480–481. [27] C. Cardell, I. Guerra, J. Romero-Pastor, G. Cultrone, A. Rodriguez-Navarro, Anal. Chem. 81 (2009) 604–611. [28] J. Baeten, K. Romanus, P. Degryse, W. De Clercq, H. Poelman, K. Verbeke, A. Luypaerts, M. Walton, P. Jacobs, D. De Vos, M. Waelkens, Microchem. J. 95 (2010) 227–234. [29] F. D’Andria, M.P. Caggia, I. Hierapolis di Frigia, Le Attivita delle Campagne di Scavo e Restauro 2000–2003, Ege Yayinlari, Istanbul, 2007. [30] D. Cottica, Hierapolis di Frigia I. Le attività delle campagne di scavo e restauro 2000–2003, in: F. D’Andria, M.P. Caggia (Eds.), MAIER–Missione Archeologica Italiana a Hierapolis, Ege Yayinlari, Istanbul, 2007, pp. 257–277. [31] M. Lettieri, F. Notarstefano, Il Dialogo dei Saperi. Metodologie integrate per i Beni Culturali, in: F. D’Andria, D. Malfitana, N. Masini, G. Scardozzi (Eds.), Edizioni scientifiche italiane, Napoli, 2010, pp. 337–352. [32] S. Prati, G. Sciutto, R. Mazzeo, C. Torri, D. Fabbri, Anal. Bioanal. Chem. 399 (2011) 3081–3091. [33] P.E. McGovern, D.L. Glusker, R.A. Moreau, A. Nunez, C.W. Beck, E. Simpson, E.D. Butrym, L.J. Exner, E.C. Stout, Nature 402 (1999) 863–864. [34] P.E. McGovern, J. Egypt. Archaeol. 83 (1997) 69–108. [35] X. Sahaya Shajan, C. Mahadevan, Cryst. Res. Technol. 40 (2005) 598–602. [36] M.R. Guasch-Jané, C. Andrés-Lacueva, O. Jáuregui, R.M. Lamuela-Raventós, J. Archaeol. Sci. 33 (2006) 98–101. [37] R.H. Michel, P.E. McGovern, V.R. Badler, Anal. Chem. 65 (1993) 408A–413A. [38] A. Pecci, G. Giorgi, L. Salvini, M.A. Cau Ontiveros, J. Archaeol. Sci. 40 (2013) 109–115. [39] F. Notarstefano, Ceramica e alimentazione: l'analisi chimica dei residui organici nelle ceramiche applicata ai contesti archeologici, Edipuglia, Bari, 2012. [40] M.P. Colombini, G. Giachi, F. Modugno, P. Pallecchi, E. Ribechini, Archaeometry 45 (2003) 659–674. [41] M.P. Colombini, F. Modugno, E. Ribechini, J. Mass Spectrom. 40 (2005) 675–687. [42] E. Ribechini, S. Orsini, F. Silvano, M.P. Colombini, Anal. Chim. Acta 638 (2009) 79–87. [43] R.L. Frost, Clays Clay Miner. 43 (1995) 191–195. [44] J. Madejová, Vib. Spectrosc. 31 (2003) 1–10. [45] M.A. Cau Ontiveros, P.M. Day, G. Montana, Modern Trends in Scientific Studies on Ancient Ceramics, in: V. Kilikoglou, A. Hein, Y. Maniatis (Eds.), Archaeopress, Oxford, 2002, pp. 9–18. [46] E. Andaloro, C.M. Belfiore, A.M. De Francesco, J.K. Jacobsen, G.P. Mittica, Appl. Clay Sci. 53 (2011) 445–453. [47] S. Shoval, Opt. Mater. 24 (2003) 117–122. [48] F. Berna, A. Behar, R. Shahack-Gross, J. Berg, E. Boaretto, A. Gilboa, I. Sharon, S. Shalev, S. Shilstein, N. Yahalom-Mack, J.R. Zorn, S. Weiner, J. Archaeol. Sci. 34 (2007) 358–373.