LC-MS and FT-IR characterization of amber artifacts

Cent. Eur. J. Chem. • 10(6) • 2012 • 1882-1889 DOI: 10.2478/s11532-012-0103-5 Central European Journal of Chemistry LC-MS and FT-IR characterization of amber artefacts Research Article Georgiana I. Truică1,2, Eugenia D. Teodor2, Simona C. Litescu2, Gabriel L. Radu1* 1 Faculty of Applied Chemistry and Materials Science, University Politehnica Bucharest, Bucharest 011061, Romania Centre of Bioanalysis, National Institute for Biological Sciences, Bucharest 060031, Romania 2 Received 19 December 2011; Accepted 20 July 2012 Abstract: This work focuses on using analytical methods, such as Fourier transform Infrared spectroscopy (FTIR) and high performance liquid chromatography (HPLC) with mass spectrometry (MS) detection to assess archaeological and geological amber. The main goal of this study is to apply the previously developed and optimized analytical methods in verifying criteria to ascribe and characterize the origin of materials found in archaeological sites. The proposed LC-MS method was successfully applied for the quantification of succinic acid content both in geological and archaeological samples of amber and offers excellent linearity between 0.1 and 5µg mL-1. The developed FTIR method provided some criteria which is able to differentiate between Baltic and Romanian amber (Romanite) that furthermore validates on archaeological amber artefacts. Keywords: Baltic amber • Romanian amber • Archaeological amber • High performance liquid chromatography with mass spectrometry detection (LC-MS) • Fourier transform infrared spectroscopy- variable angle reflectance (FTIR-VAR) © Versita Sp. z o.o. 1. Introduction Amber is a precious stone resulting of the plant resin fossilization after millions of years and has the special color of honey-yellow, sometimes reddish-brown or greenish-black. Amber was used in ornamental and curative purpose because of its chemical and physical properties (it melts at 300oC, burns with a flame, and dissolves in organic solvents) and its investigation is of great importance for historians, both for gemological and archaeological purposes. Also amber shows different inclusions (gas, soil, liquids, insects, insect remains or fossil plants) that are well conserved; as reported data show this phenomenon took place 20-100 million years ago [1-5]. Amber is found all around the world but one of the biggest natural deposits of amber lies along the southern shores of the Baltic Sea and is known as Baltic amber or succinite. Other sources of amber are found in the Dominican Republic, Spain, Sicily, Lebanon, Burma, and Romania. Amber is an archaeological artefact with high interest for scientists because its analysis reveals information about life on earth at the moment of resin formation and ethnic and cultural aspects of the society that used it. During recent years, different modern techniques of analysis have been used to characterize amber composition, distinguish between types of amber from various historical periods, and determine their paleobotanic sources. Beck for the first time used infrared spectroscopy for the characterization of amber [6,7]. This was one of the most frequently and widespread method used by many authors to describe the fossil resin called amber, although its limits have become restrictive in the last years and other advanced methods were used in addition [8-13]. Using non-destructive techniques like Raman spectroscopy [14,15], X-ray diffraction or fluorescence and scanning electron microscopy [16-18], geological * E-mail: [email protected] 1882 G. I. Truică et al. origin of amber samples can be determined and maturation process of the resin can be observed [19-22]. In addition, studies by means of gas chromatography/ mass spectrometry coupled with pyrolysis or thermal desorption methods were published in the recent years [23-32]. This invasive techniques, however, involves extraction of volatile compounds from amber with organic solvents or decomposition using heat and acidic media. The literature identifies many places with possible appearances of amber in Romania, all along Eastern and Southern Carpathians [33], but in practice, it is very difficult to find anything else but „Colţi amber” (or locations up to 20 km away). As an indicator of the archaeological yield, the quantity and quality of amber artefacts from the territory of Romania are lower than those found in territories placed on the track of the Amber Route (like Hungary and Serbia). Romanian museums preserve some collections of amber artefacts found through archaeological excavations from Romanian territory and the problem which appears is whether these artefacts have Baltic or Romanian origin. Therefore, in the present study, attention has been focused on source identification/confirmation of archaeological amber found on two different locations, Cioclovina (Romania), and Denmark, using two methods, a non-destructive (FTIR-VAR) and destructive one (LC-MS), developed, optimized and applied in a previous study on geological amber samples [10]. analysis and further subjected to solvent extraction for the LC-MS analysis. The last technique cannot always be performed due to impossibility of sampling when working with historical heritage, which explains the small number of archaeological samples available for destructive analysis. 2.2. FTIR-VAR analysis 2. Experimental procedure The infrared spectra were acquired on a Bruker TENSOR 27 instrument using the FTIR-VAR technique. The following instrumental settings were used: optimal beam incidence angle of 45o, range from 4000 to 600 cm-1, 96 scans, aperture of 4 mm, spectral resolution ±4 cm-1, OPUS software version 6.0. The samples were used, as a whole piece fixed on a gold mirror, and all the spectra were registered versus a background of clean gold foil. The ascription of the amber origin was attempted by comparison between registered FTIR spectra for archaeological amber samples and spectra of controlled origin amber, both Baltic and Romanite considered as reference. Amber has an amorphous structure, and, as a consequence, the reproducibility of determinations is affected by the diversity of the sample, the obtained spectra depending on the analysed part of the sample (the spectra of the same sample could be slightly different due to the technique’s specific scattering). For each sample two or three spectra were registered for different zones of the sample, and differences appear only in the spectral region related to methyl/methylene. In conclusion the characteristic spectral region for amber was preserved for each duplicate spectrum acquired. 2.1. Samples 2.3. LC-MS analysis A controlled geological origin set of seven Baltic amber (succinite, B1-B7) and four Romanian amber (Romanite, R1-R4) samples were selected as reference material. Baltic amber originated from KaliningradRussia (B4, B5), Bitterfield-Germany (B1, B6, and B7), Palanga-Lithuania (B2) and Earth Museum of Warsaw- Poland (B3). Romanite samples originated from Buzău County (R1, R3, R4) and Sibiciul de Jos village (R2). The archaeological material consisted of 13 samples excavated from Cioclovina hoard, Hunedoara County (provided by the National History Museum from Bucharest) and 6 samples from Denmark (kindly provided by Yvonne Shashoua, Department of Conservation, National Museum of Denmark). Both categories are dating from Prehistory (Bronze Age). These samples were small pieces (small fragments of 5 to 10 mg detached from different archaeological pieces) and were used without any pre-treatment for FTIR-VAR High performance liquid chromatography coupled with mass spectrometry method was developed to assess the succinic acid amount (well-known component in the amber structure). This aliphatic dicarboxylic acid lack a good ultraviolet (UV) chromophore, as a consequence mass spectrometry (MS) is used to achieve sensitive detection. The extraction of succinic acid was adapted to very small amounts of samples. It was done in mild conditions, so the extraction of succinic acid was partial. The samples were grounded to powder by hand into a porcelain mortar without stirring and extracted in acetonitrile acidified with formic acid (pH =3.00), 1:1 ratio (mg mL-1) at room temperature, for a period of 7 days. The extracted fractions were filtered and analysed by LC-MS. Experiments were performed on a Shimadzu system equipped with two pumps LC-20AD, 1883 LC-MS and FT-IR characterization of amber artefacts Figure 1. FTIR-VAR spectra of Romanite and Baltic geological amber samples, range 2000-600 cm-1. photodiode array detector SPD-M20A (operating at wavelengths between 200 and 600 nm) and coupled to a MS detector Shimadzu LCMS-2010EV equipped with an electrospray interface (ESI). UV and MS data were acquired and processed with a LC-MS Solutions operating system. The separation was done on a 50×2.1 mm Kromasil 100-3.5 C18 column. The mobile phase used was: solvent A, deionised water with formic acid, pH=3.00; solvent B, acetonitrile with formic acid, pH=3.00, isocratic mode, A:B (25:75). The flow rate was 0.1 mL min-1 and the analyses were performed at 20oC. Samples of 10 µL of amber extract were directly injected into the column, at the autosampler temperature of 20oC. ESI source and negative ionization mode was used with detector voltage 1.5 kV. Tunning was performed on suitable mass range before the experiment. The MS acquisition with ESI ionization was performed under the following condition: nebulizing gas (N2) flow rate, 1.5 L min-1; curved desolvation line temperature (CDL), 250oC, heat block temperature, 200oC, acquisition mode, SIM (selected ion monitoring), m/z 117. The succinic acid standard was dissolved in formic acid 0.1% in acetonitrile and directly injected to determine the specific retention time. The stock solution of succinic acid was diluted so to obtain six different standard solutions that covered the range of concentration between 0.1 and 5 µg mL-1. 1884 3. Results and discussion 3.1. Interpretation of FTIR-VAR spectra Fourier transform infrared (FTIR) spectroscopy has been applied to amber mainly to establish whether it is a fossil resin or a recent one and to obtain information about origin. It is an effective technique for the examination of amber because the polymerization process preserves all the original functional groups of the compounds found in amber structure with the exception of C=C double bond which suffers a saturation process by exposure to air. Concerning the archaeological samples, the FTIRVAR is a reliable tool for non-destructive investigation of amber. The obtained spectra are more complex than FTIR-transmittance spectra, the signals are slightly difficult to ascribe, but corroborated with other techniques (LC-MS). It is a trustworthy method to diagnose the origin of amber from archaeological sites. FTIR-VAR spectra were analysed and assigned on the three wavenumbers domains of significance for amber, namely those between 3600-2000 cm-1, 1820-1350 cm-1 and the 1250-1045 cm-1 regions, according to our previous experience [10], which correlate with the presence of hydroxyl groups, carboxyl groups, carbonyl groups and with C=C unsaturation. Assignments of the observed functional groups to FTIR bands are summarized in Table 1. A characteristic G. I. Truică et al. Table 1. Wavenumber positions (cm-1) of FTIR-VAR bands and their suggested vibrational assignments for Baltic and Romanian amber in the 4000-600 cm-1 range. Band, cm-1 Group Remarks 3900-3400 -alcohols, phenols O-H streching 3000-2800 -alkyl C-H streching 1735 -carbonyl C=O 1642 -double bond unsaturation C=C stretch 1595 - alcohols intramolecular -H bonds -alkyl C-H bending 1375 -methyl C-H bending 1155 -esters C-O streching 1050-1000 -esters C-O streching -alkyl C-H out of plane bending of H atoms 1460-1420 880-860 FTIR spectrum of Romanian and Baltic geological amber is presented in Fig. 1. The FTIR spectrum of geological amber samples first presents a large number of peaks between 3900-3400 cm-1, more or less intense, due to the OH stretching bands of alcohols and/or carboxylic acids [34]. As can be noticed in Fig. 2 from overlapping archaeological sample R5 with Romanite geological sample R1 and Baltic geological sample B1, in Romanite case no matter the source- either archaeological or geological – the band intensity is higher than Baltic especially in region 1045 cm-1. The bands corresponding to the alkyl stretching show, between 3000 and 2800 cm-1, a characteristic pattern with a principal band at 2926 cm-1 and two bands at 2961 and 2859 cm-1 of comparable intensities. The assignment generally admitted of these absorptions are the following: the CH3 asymmetric stretching vibration occurs at 2961 cm-1 and may easily be distinguished from the nearby CH2 absorption at about 2926 cm-1 while the symmetric stretching absorption band of the methylene group occur at 2859 cm-1. This wavenumber range indicates an important level of methyl groups in amber structure. Two other bands can be assigned to alkyl groups: the first between 1460 and 1420 cm-1 involve CH2 and CH3 bending, the second at 1375 cm-1 is only due to CH3 bending. The spectral region 1820-1690 cm-1 is a very complex one, the predominant absorptions bands that have been highlighted are assigned to ester, ketone and, the last, carboxylic acid groups. In the carbonyl range, a predominant ester band around 1735 cm-1 is present and a medium-weak band at 1155 cm-1 is always observed that can be ascribed to the C–O simple bond stretching of esters (Fig. 1). Baltic amber shows this characteristic broad band at 1155 cm-1, followed by two absorption peaks, which reaches maximum intensity at 1015, and 879 cm-1, attributed to C-O stretch. There are vibration frequency shifts determined by the degree of ethers and esters formation, which appear frequently in amber type resins with younger ages. For example, a shift in the infrared bands ascribed to C-C bend from 1595 cm-1 specific to Romanian amber compared to the Baltic amber where it occurs around 1642 cm-1 was observed during oxidation. This suggested a transition from conjugated to isolated double bonds. Two bands at 1045 and 863 cm-1 are observed in about 80% of the acquired spectra for Romanian amber samples and could be assigned to different C–O bonds. This region has signals of higher intensity in archaeological samples. Some differences were observed in the spectral region 900-600 cm-1 that could lead to discrimination between the two species of amber. The second peak from Baltic amber spectra (Fig. 1.) at 667 cm-1 is absent indicating a final process of contraction/reticulation of the polymeric chain. The Baltic shoulder was shifted toward 1045 cm-1, in the Romanite spectra, providing evidence about older age of the amber samples. Although FTIR spectra of amber present similar features for all samples, some differences can be observed between geological and archaeological amber and are of great interest for the origin diagnosis of fossil resins (see Table 2). The FTIR-VAR spectra of archaeological amber samples (R5-Cioclovina and B7-Denmark) are presented in Figs. 2-3. In light of the FTIR-VAR bands of archaeological samples from the regions 900-1050 cm-1 and 1510-1640 cm-1, we could conclude that the majority of the sampled amber from Cioclovina, is of Romanian origin. A large part of the samples present characteristic bands of Romanite from 1588-1600 cm-1 and from 1045-1050 cm-1, correlated with C-O stretching. The archaeological samples from Denmark present the predominant ester band around 1750 cm-1 well correlated with a band at 1152 cm-1 always observed at Baltic amber samples, ascribed to the C–O simple bond stretching of esters. These differences can be explained by the geological conditions of evolution for the resin. 3.2. LC-MS analysis The concentration of extractable succinic acid was determined by LC-MS method described above using a 1885 LC-MS and FT-IR characterization of amber artefacts Figure 2. Table 2. FTIR-VAR spectra of Romanite (R1) and Baltic (B1) geological amber overlaid with Romanite archaeological amber (R5), range 4000-600 cm-1. Differences between amber samples with controlled origin (geological) and from archaeological sites. Wavenumber domain cm-1 Vibrational assignment Differences between geological and archaeological amber 3900-3400 OH stretching bands of alcohols and carboxylic acids Archaeological samples present a less intense signal compared with geological amber samples probably due to environmental conditions in which they were found. 3000-2800 CH2, CH3 stretching vibrations More intense signals in archaeological samples indicating a higher level of methylation which correlates with the polymeric structure of amber (probably due to a reduction in the number of functional bonds and an increase in aromaticity). 1820-1690 The signals from these region present similar shapes in both cases studied C=O stretching bands from esters, (geological and archaeological) but different intensities and shifts of the specific ketones, aldehydes and carboxylic acids wavenumbers indicate some degradation and/or transformation processes for archaeological samples. 1695-1510 C=C stretching This region is quite similar and with comparable intensity in all samples from Romanian territory (geological and archaeological amber); Baltic amber samples present less intense signals for geological origin compared with samples from Denmark (archaeological). 1500-1200 C-H bending High intensity signals in archaeological samples confirming the presence of methyl/ methylene groups in the structure of amber in accordance with the observations from region 3000-2800 cm-1 . C-O stretching and C-H out-of-plane bending This fingerprint zone is different in most archaeological samples when compared to geological ones with bands of higher intensity and slight shifts of the characteristic wavenumbers although confirming the functional group region. 1200-600 calibration curve, A = 12306C + 422; R2= 0.9898 where A-peaks area and C-concentration, (µg mL-1). Excellent linearity within 0.1 to 5 μg mL-1 is observed and the method detection limit is 0.07 μg mL-1. The mass spectra of the peak eluting at approximately 2.57 min reveals the deprotonated molecular ion [M-H] - at m/z 117 (as seen 1886 in Fig. 4). TIC chromatogram of some representative samples selected from the 30 ones analyzed by this method is presented in Fig. 5. Identification of succinic acid in amber samples was accomplished both on ion monitoring and the specific retention time at 2.57±0.11 min for standards and samples. The results G. I. Truică et al. Figure 3. Overlaid FTIR-VAR (reflectance) spectra, region 4000-600 cm-1 for geological Baltic amber (B1), and archaeological Baltic amber from Denmark (B7). Figure 4. Ion chromatogram and signal intensity for succinic acid standard solution. of succinic acid extractable concentrations obtained from LC-MS analysis ranged from 1.1 μg mg-1 for Baltic samples to 10.25 μg mg-1 in case of Romanite samples. This fact could be explained by a partial extraction of succinic acid due to more or less compact structures of different amber samples. Because archaeological material should be preserved intact, only small amounts (maximum 1-2 mg) of sample were prevailed, therefore a total extraction of succinic acid from amber could not be performed. As consequence, we attempted to develop a simple procedure applicable to small amounts of amber (even so, this technique cannot always be applied because in most cases the sampling is not possible on artefacts; that explains the relative small numbers of analyzed samples). Therefore, our previous stipulation [10] that is possible to differentiate Romanian from Baltic amber using the extractable succinic acid determination by this technique (because the quantities 1887 LC-MS and FT-IR characterization of amber artefacts Figure 5. Ion chromatogram for A-Baltic (B3) and B-Romanite (R4) amber samples. of succinic acid extracted under given conditions are higher for Romanite samples), also validates for archaeological samples. All the archaeological samples from Denmark which are for sure of Baltic origin present very low content of extractable succinic acid (< 0.57) and all the samples from Cioclovina have a high content of extractable succinic acid (~ between 8 and 10 µg). Samples from Cioclovina were analyzed as well by Raman spectroscopy [35], which confirms the origin of these artefacts as mainly Romanian. Our results are not comparable with those published in the literature [36] concerning the different ratios of succinic acid from different types of amber because of the small quantities of samples we had to work with (we did not made a total extraction of succinic acid). 4. Conclusions The main objective of this study is to examine amber from diverse archaeological sites using analytical techniques, such as FTIR and LC-MS, which were previously developed and applied on geological samples of amber. Because FTIR spectroscopy was a non-destructive technique, samples were subsequently available for 1888 investigation by an additional technique, a destructive LC-MS analysis (in this case). The study made on a relative large number of amber samples (from different locations) leads us to conclude that archaeological material is different from geological material, so the origin diagnosis of archaeological material is more difficult and the comparison with geological standards requires an association between data obtained by FTIR analysis (reflectance) and other analytical techniques, such as LC-MS analysis if is possible, to achieve satisfactory results. Concerning the archaeological samples, the FTIR-VAR spectra pointed out some differences between the archaeological and geological material. However relaying on specific vibration bands from geological samples spectra, with minor shifts and intensity variation due to transformation processes that occur in the environmental condition where they were discovered, we could identify the origin of the archaeological artefacts. The LC-MS method results concluded that Baltic amber shows a lower content in extractable succinic acid compared with Romanian amber samples. Under the extraction conditions used, we obtained the same extraction ratio between Baltic amber and Romanian amber for geological amber as well as for the archaeological material. The explanation could be that Baltic amber posses a more compact structure with strong polymeric links that make it difficult to extract succinic acid under the conditions outlined. These conclusions were drawn based on a relatively small number of archaeological samples and therefore are not definitive. Further studies will be conducted whenever it will be possible to get sample from archaeological material to obtainmore definitive conclusions. Acknowledgments The work has been funded by the Sectoral Operational Programme Human Resources Development 2007-2013 of the Romanian Ministry of Labour, Family and Social Protection through the Financial Agreement POSDRU/88/1.5/S/61178. The authors are grateful to Romanian research programm, PN09360106/2009-2011, and to Senior Research Scientist Yvonne Shashoua, Department of Conservation, National Museum of Denmark. G. I. Truică et al. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] E. Penalver, X. Delclos, C. Soriano, Cretaceous Res. 28, 791 (2007) M. Dierick, V. Cnudde, B. Masschaele, J. Vlassenbroeck, L. Van Hoorebeke, P. Jacobs, Nucl. Instrum. Meth. A 580 (1) 641 (2007) P. Crespo, L. Blasco, M. Poza, T.G. Villa, Int. Microbiol. 10, 117 (2007) D. Bellis, D. L. Wolberg, Palaeogeogr. Palaeocl. 97, 69 (1991) A.R. Schmidt, H. Dörfelt, Rev. Palaeobot. Palyno 144, 145 (2007) C.W. Beck, E. Wilbur, S. Meret, Nature 201, 256 (1964) C.W. Beck, Appl. Spectrosc. Rev. 22, 57 (1986) E.Trevisani, C.A.Papazzoni, E. Ragazzi, G. Roghi, Palaeogeogr. Palaeocl. 223, 260 (2005) I. Angelini, P. Bellintani, Archaeometry 47, 441 (2005) E.D. Teodor, S.C. Liţescu, A. Neacşu, G. Truică, C. Albu, Cent. Eur. J. Chem. 7(3), 560 (2009) Y. Shashoua, M-B.L.D. Berthelsen, O.F. Nielsen, J. Raman Spectrosc. 37, 1221 (2006) M. Guiliano, L. Asia, G. Onoratini, G. Mille, Spectrochim. Acta A 67, 1407 (2007) E. Penalver, E. Alvarez-Fernandez, P. Arias, X. Delclos, R. Ontanon, J. Archaeol. Sci. 34, 843 (2007) P. Vandenabeele, D.M. Grimaldi, G.M. Howell Edwards, L. Moens, Spectrochim. Acta A 59, 2221 (2003) H.G.M. Edwards, D. Farwell, S.E.J. Villar, Spectrochim. Acta A 68, 1089 (2007) A. Matuszewska, M. Czaja, Talanta 56, 1049 (2002). I. Pakutinskienea, J. Kiuberisa, P. Bezdickab, J. Senvaitienec, A. Kareivac, Can. J. Anal. Sci. Spect. 52(5), 287 (2007) J.T. Senftle, S.R, Larter, Org. Geochem. 13(4-6), 973 (1988) L. Lowe, J.L. Ruvalcaba Sil, Antropologia y tecnica 6, 41 (1993) K.A. Rodgers, S. Currie, Thermochim. Acta 326,143 (1999) [21] E. Ragazzi, G. Roghi, A. Giaretta, P. Gianolla, Thermochim. Acta 404, 43 (2003) [22] M. Feist, I. Lamprecht, F. Muller, Thermochim. Acta 458, 162 (2007) [23] G.C. Galletti, R. Mazzeo, Rapid Commun. Mass. Sp. 7, 646 (1993) [24] A.M. Shedrinsky, D.A Grimaldi, J.J. Boon, N.S. Baer, J. Anal. Appl. Pyrol. 25, 77 (1993) [25] E.C. Stout, C.W. Beck, K.B. Anderson, Phys. Chem. Miner 27, 665 (2000) [26] S. Pipatmanomai, C.A. Islas, I. Suelves, A.A. Herod, D.R. Dugwell, R. Kandiyoti, J. Anal. Appl. Pyrol. 58-59, 299 (2001) [27] A.M. Shedrinsky, T.P. Wampler, K.V. Chugunov, J. Anal. Appl. Pyrol. 71, 69 (2004) [28] S. Yamamoto, A. Otto, G. Krumbiegel, B.R.T. Simoneit, Rev. Palaeobot. Palyno 140, 27 (2006) [29] L. Tonidandel, E. Ragazzi, G. Roghi, P. Traldi, Rapid Commun. Mass. Sp. 22, 630 (2008) [30] R. Pereira, I.S. Carvalho, B.R.T. Simoneit, D.A. Azevedo, Org. Geochem. 40, 863 (2009) [31] M. Virgolici, C. Ponta, M. Manea, D. Negut, M. Cutrubinis, I. Moise, R. Suvaila, E. Teodor, C. Sarbu, A. Medvedovici, J. Chromatogr. A 1217, 1977 (2010) [32] C. Menor-Salvan, M. Nijarro, F. Velasco, I. Rosales, F. Tornos, B.R.T. Simoneit, Org. Geochem. 41, 1089 (2010) [33] H.U. Kasper, In: M. Ganzalewski, R. Slotta (Eds.), Der Rumänische Berstein, in Bernstein. Tränen der Götter. Katalog der Austellung des Deutschen Bergbau-Museums Bochum (Verlag Glückauf GmbH, Bochum, 1996) 357-362 (In German) [34] G. Socrates, Infrared and Raman Characteristic Group Frequencies, 3rd edition (John Wiley & Sons, Chichester, England, 2001) [35] E.S. Teodor, E.D. Teodor, M. Virgolici, M.M. Manea, G. Truica, S.C. Litescu, J. Archeaeol. Sci. 37(10), 2386 (2010) [36] M. A. Bogdasarov, Geol. Ore Deposit+ 49 (7), 630 (2007) 1889