Organic Remains in Early Christian Egyptian Metal Vessels: Investigation with Fourier Transform Infrared Spectroscopy and Gas Chromatography–Mass Spectrometry

Organic remains preserved on eight copper alloy artifacts of the Byzantine Collection of the Benaki Museum with an Egyptian provenance were investigated, implementing a multi-analytical approach combining microscopy-FTIR and GC/MS. The transmission spectra of powder samples provided important information on the vessels regarding inorganic and organic components. In the latter case, subsequent extractions with a range of solvents allowed discrimination of components with different polarities and provided data on the suitability of the solvents for the acquisition of more informative spectra. GC/MS was implemented for the detailed characterization of the compounds present in the samples because of the complex nature of the residues preserved. A wide range of fatty acid oxidation products was identified, including a series of α, ω-dicarboxylic acids typical of such remains. In addition, vicinal dihydroxy-docosanoic and dihydroxy-eicosanoic acid, oxidation products of erucic and gondoic acid, respectively, were detected. Both are found in abundance in oils from plants belonging to the Brassicaceae family and imply their multiple uses in medieval Egypt.


Introduction
The application of analytical methodologies to archaeological material has revealed a range of organic molecules surviving over millennia. Organic remains have been identified in many archaeological contexts, and their study in the past 30 years has offered insights into plant and animal use and exploitation, trading of natural organic materials and their various uses [1,2]. Subsequently, archaeological research has focused on understanding the degradation pathways of the products utilized, during both use and deposition, and the identification of characteristic compounds that would allow the assignment of their origins [3][4][5]. Factors such as temperature, humidity, oxygen availability, and pH affect the preservation of different biomolecules [5][6][7].
The continuous development of new analytical methodologies has facilitated the characterization of a wide range of natural organic materials, such as animal fats and plant oils, beeswax, tars, and resins [1,8]. The identification of vegetable oils, known through literary sources to have been used in lighting, cooking, religious rituals, medications, and cosmetics, is particularly challenging, as they are very susceptible to degradation [9][10][11][12][13]. Partial hydrolysis of glycerol triesters occurs, gradually leading to mixtures of mono-and diacylglycerols, and eventually, free fatty acids and glycerol. Additionally, plant oils, rich in unsaturated moieties, undergo oxidative degradation mainly targeting double

Fourier Transform Infrared Spectroscopy
Fourier transform infrared spectra were recorded with a Perkin Elmer Spectrum GX 1 FTIR system. In the case of microscopy-FTIR, this was coupled with a PE AutoImage system. The recording conditions were as follows: for transmission spectra of KBr disks, 4000-400 cm −1 , 32 or 64 scans and 4 cm −1 resolution, DTGS detector at room temperature; for microscopy-FTIR spectra, 4000-700 cm −1 , 100 scans, 100 × 100 or 50 × 50 μm scanning areas, 4 cm −1 resolution using a cryo-cooled (liquid N2) MCT detector

FTIR Analysis of Powder Samples in KBr Discs
According to a typical procedure, mid-infrared spectra of powder samples pressed in 13-mm KBr discs using a hydraulic press were recorded in transmission mode.

Solvent Extraction for FTIR Analysis
A solvent extraction scheme involving hexane, dichloromethane, acetone, and methanol was followed; solvents were selected according to their polarity and volatility. Approximately 5-10 mg of the powder sample was sonicated with 1 mL of solvent in a glass vial for 30 min. From each vial, one droplet of the supernatant fluid was accordingly deposited with a syringe on a neat KBr disc and left to dry at room temperature; complete solvent evaporation was monitored by FTIR. The remaining material was then analyzed in transmission mode. All spectra of solvent-extracted samples are shown not normalized so that relative intensities reflect the approximate relative quantities of extracted material; besides baseline correction, no other treatment was applied to spectra.
For microscopy-FTIR analysis, extraction in xylene and chloroform was conducted; droplets of solvent-extracted samples were accordingly deposited using a syringe on a 13 mm circular gold mirror plate and were left to dry at room temperature; complete solvent evaporation was monitored by FTIR. Consequently, the remaining material was analyzed

Fourier Transform Infrared Spectroscopy
Fourier transform infrared spectra were recorded with a Perkin Elmer Spectrum GX 1 FTIR system. In the case of microscopy-FTIR, this was coupled with a PE AutoImage system. The recording conditions were as follows: for transmission spectra of KBr disks, 4000-400 cm −1 , 32 or 64 scans and 4 cm −1 resolution, DTGS detector at room temperature; for microscopy-FTIR spectra, 4000-700 cm −1 , 100 scans, 100 × 100 or 50 × 50 µm scanning areas, 4 cm −1 resolution using a cryo-cooled (liquid N 2 ) MCT detector

FTIR Analysis of Powder Samples in KBr Discs
According to a typical procedure, mid-infrared spectra of powder samples pressed in 13-mm KBr discs using a hydraulic press were recorded in transmission mode.

Solvent Extraction for FTIR Analysis
A solvent extraction scheme involving hexane, dichloromethane, acetone, and methanol was followed; solvents were selected according to their polarity and volatility. Approximately 5-10 mg of the powder sample was sonicated with 1 mL of solvent in a glass vial for 30 min. From each vial, one droplet of the supernatant fluid was accordingly deposited with a syringe on a neat KBr disc and left to dry at room temperature; complete solvent evaporation was monitored by FTIR. The remaining material was then analyzed in transmission mode. All spectra of solvent-extracted samples are shown not normalized so that relative intensities reflect the approximate relative quantities of extracted material; besides baseline correction, no other treatment was applied to spectra.
For microscopy-FTIR analysis, extraction in xylene and chloroform was conducted; droplets of solvent-extracted samples were accordingly deposited using a syringe on a 13 mm circular gold mirror plate and were left to dry at room temperature; complete solvent evaporation was monitored by FTIR. Consequently, the remaining material was Heritage 2021, 4 3614 analyzed with a Perkin Elmer AutoImage microscopy-FTIR system in reflection mode, where different classes of compounds were analyzed on various spots deposited on the mirror disc. The working principle of this phenomenon lies in the spontaneous aggregation of chemically similar compounds in certain spots on the disc surface during solvent evaporation in a like-goes-with-like mode [48][49][50].  Table S1, Supplementary Material) were labeled according to the object number and were extracted following standard extraction protocols (Charters et al. 1995;Stern et al. 2000). A solution of CH 2 Cl 2 :CH 3 OH/2:1 (2-4 mL) was added to the samples, that were then sonicated (2 × 15 min) and centrifuged (2000 rpm, 5 min). The excess solvent was transferred into a clean vial and evaporated under a gentle stream of nitrogen. A portion of the extract was saponified with NaOH/CH 3 OH (0.5 M in a 70 • C, 60-90 min water bath). The resulting samples were acidified with 6 M HCl aqueous solution and extracted with 3 × 3 mL of hexane. The supernatants were combined and transferred to glass vials and evaporated under a gently blown stream of nitrogen gas. All samples were subsequently converted to their trimethylsilyl derivatives using 80 µL of N,O-bis(trimethylsilyl)-trifluoroacetamide (BSTFA, Sigma-Aldrich, St. Louis, MO, USA, derivatization grade >99.0%) [51][52][53].

GC-MS Analysis
A Hewlett Packard series 6890 gas chromatograph equipped with a J&W GC fused silica capillary column, model DB-1HT (15 m, i.d.: 0.320 mm, film thickness 0.10 µm, stationary phase composition: dimethylpolysiloxane) with helium as carrier gas and coupled to an Agilent Technologies series 7683 injector were employed. Samples were injected into the chromatograph in splitless mode; the oven temperature program was held isothermal at 50 • C for 2 min following injection; and ramped at a rate of 10 • C/s to reach 340 • C, which was kept for 14 min. The Mass Spectrometer (EI, 70 eV) was set to scan at 50-700 m/z. The TIC was recorded in all cases. For quantitation purposes 0.155 mg/mL of hexadecane/CH 2 Cl 2 solution as internal standard was used.

Results and Discussion
The nature of preserved residues found in a larger number of vessels has been addressed in previous studies. In these, analysis of inorganic remains [45] and preliminary screening of the residues [46] through transmission FTIR of powder samples (Table 1, column 2) were conducted. In light of this, further inquiries arose regarding the use, primary or secondary, of the residues in the vessels and lamps (for example, reused during later periods). Other hypotheses were based on the relationship of the contained materials to past repairs or post-excavation treatments. Moreover, the possible interactions of the most reactive substances among them with the existing metal ions because of the formation of corrosion products and/or other sources from the burial or atmospheric environment were also possible.
Furthermore, carboxylate bands at 1587 and 1541 cm −1 (antisymmetric COO − stretch [61]), more prominent in samples #11633A, 11550, and 11551, were detected, assignable to salts and complexes of fatty acids with metals [62,63]. In light of detecting free fatty acids through SE-FTIR and GC-MS analysis (see below), the soaps may have been formed through the interaction between the free fatty acids and metal ions, arguably copper or calcium [62]. Between the two, copper appeared more plausible as it was the main metal in the body of all vessels, while very low amounts of calcium carbonate (the main geogenic calcium source [54]) were detected in the transmission spectra. The formation of metal soaps is chemically favored in the archaeological environment since they are relatively insoluble in water and, being more stable, they contribute to better preserving the lipid character in time [63].
The detection of metal soaps raised the question of possible contamination of the contained material through cleaning actions at an unspecified time. However, sodium fatty acid salts could be ruled out since elemental analysis did not detect sodium. Further supporting this hypothesis, no sodium carboxylate peaks (expected at 1560 and 1423 cm −1 for their antisymmetric and symmetric stretching vibrations, respectively [64][65][66][67]) were detected by FTIR.
Instead, calcium and copper soaps were most possibly detected as deterioration products of fatty acids in the presence of certain metal ions such as lead, zinc, copper, and calcium [66,[68][69][70]; besides, the relatively narrow line shapes of most bands suggested well-crystallized soaps [71]. This fact added an originality asset to the organic remains profile of the objects in this study.
Most of the above transmission spectra results were better evaluated in the light of solvent-extraction infrared spectroscopy (SE-FTIR) and GC-MS results.

Solvent-Extracted FTIR Results
An analytical scheme focusing on the organic fraction was implemented to overcome interpretation uncertainties in KBr infrared spectra of powder samples from peak overlaps between organic and inorganic components. Previous works have reported solvent extractions for acquiring infrared spectra of isolated components [42,72,73]. Here, a coordinated solvent extraction scheme based on polarity, involving methanol, acetone, and dichloromethane, while in microscopy-FTIR, chloroform and toluene, were considered. Infrared spectra were recorded from the dried remain on KBr discs after dripping solvent extracts of samples; this technique will be referred to as solvent-extracted FTIR (SE-FTIR, see Materials and Methods), with results for all samples shown in the Supplementary  Information file, Figures S1-S4.
Finally, fatty acid metal salts (FAMS) through a shoulder at 1560-1550 cm −1 were also detected in samples #11551, 11550, 11633A, 11622A, 11573, and 11596. Compared to transmission spectra of powdered samples (see above), soaps were detected in significantly lower amounts in SE-FTIR spectra because of their limited solubility in the used solvents. Acetone extracts: As seen through their SE-FTIR spectra, acetone extracts were particularly rich in detected components, comparatively shown in Figure S2. Esters were almost exclusively present in most samples (#11573, 11596, 11598, 11622, 11633) based on their carbonyl absorption at 1745-1730 cm −1 and confirmed by their methylene and methyl stretching (~2920, and 2955 cm −1 , respectively) and bending vibrations (~1470 and 1460/1380 cm −1 , respectively) [74,80]. In addition, evidence of fatty anhydrides or lactones was also visible (shoulders at 1780-1765 cm −1 [81]) as products of intense oxidation [28,82]. Spectra for samples #11550 and 11551 are shown in Figure 3c or Figure 4c, compared to corresponding spectra from other solvents.
Dichloromethane extracts: Extraction with dichloromethane resulted in rich SE-FTIR spectra ( Figure S3). A general characteristic of most spectra was the prominence of methyl groups' absorptions (i.e., stretching at 2958 cm −1 and bending at 1379 cm −1 ), suggesting a preference for short or mid-sized alkyl chains through these extractions. The above, assisted by the ~1714 cm −1 acidic carbonyl maxima and 1415 cm  In the light of GC-MS analysis, where a significant fraction of dicarboxylic acids was identified (see below), evidence for dicarboxylic fatty acids (di-FAs), typical degradation products of unsaturated oils, was provided in the infrared spectra of methanol extracts #11544 and 11550. This was based on the absence of the CH 3 vibrations, the low-shifted carbonyl maxima (shoulder at~1700 cm −1 ), their relatively high-shifted antisymmetric CH 2 stretch (shoulder at 2960-2950 cm −1 ), and their in-plane CH 2 and C-O-H bending (diagnostically appearing at~1425 and~1412 cm −1 , respectively) [64].
Finally, fatty acid metal salts (FAMS) through a shoulder at 1560-1550 cm −1 were also detected in samples #11551, 11550, 11633A, 11622A, 11573, and 11596. Compared to transmission spectra of powdered samples (see above), soaps were detected in significantly lower amounts in SE-FTIR spectra because of their limited solubility in the used solvents.
The above infrared results of the various solvent extracts valuably added to deconstructing the overall profile; for samples #11550 and 11551, this is exemplified in Figure  3d or Figure 4d.

Solvent-Extraction-μFTIR
Microscopy-FTIR spectra of selected solvent-extracted samples deposited as films on a gold mirror (abbreviated as SE-μFTIR) were recorded in reflection-absorption infrared spectroscopy (RAIRS) mode according to the procedure described in Materials and Methods. This aimed to investigate organic components based on their spatial micro-separation during solvent evaporation on the gold surface [43][44][45], and therefore, it supported the standard SE-FTIR investigation (see above) by offering more detailed insight into FTIRdetectable components.
SE-μFTIR investigations for the sample from oil lamp #11551, after acetone, chloroform, and xylene extractions, gave the most interesting results. The acetone extract of sample #11551 showed the spatial micro-separation and detection of an almost pure mediumchain fatty acid (Figure 4e), based on its antisymmetric stretch at 2928 cm −1 .
In the xylene extract, four spots containing fatty acids and various esters (arguably, acylglycerols) in variable amounts on the basis of their carbonyl maxima at 1746 (TAG), Dichloromethane extracts: Extraction with dichloromethane resulted in rich SE-FTIR spectra ( Figure S3). A general characteristic of most spectra was the prominence of methyl groups' absorptions (i.e., stretching at 2958 cm −1 and bending at 1379 cm −1 ), suggesting a preference for short or mid-sized alkyl chains through these extractions. The above, assisted by the~1714 cm −1 acidic carbonyl maxima and 1415 cm −1 due to C-O-H in-plane bending for #11544, 11550, 11551, 11633, supported the detection of mid-and possibly short-chain FAs in these samples. In other samples, however, such as #11596, only traces of the above were found. Additionally, intense ester carbonyl maxima at 1738-1743 cm −1 were detected in most samples from acylglycerols. From the maxima at 1261 (CH 2 wagging) and 1182, and 1100 cm −1 (C-O-C antisymmetric stretching of ester links), and 800 cm −1 (C-O-C bending of ester links) [28,74,75,84], and the low OH stretching intensities (~3450 cm −1 ), it can be inferred that diacylglycerols are the predominant esters detected in this solvent.
The above infrared results of the various solvent extracts valuably added to deconstructing the overall profile; for samples #11550 and 11551, this is exemplified in Figure 3d or Figure 4d.

Solvent-Extraction-µFTIR
Microscopy-FTIR spectra of selected solvent-extracted samples deposited as films on a gold mirror (abbreviated as SE-µFTIR) were recorded in reflection-absorption infrared spectroscopy (RAIRS) mode according to the procedure described in Materials and Methods. This aimed to investigate organic components based on their spatial micro-separation during solvent evaporation on the gold surface [43][44][45], and therefore, it supported the standard SE-FTIR investigation (see above) by offering more detailed insight into FTIRdetectable components. SE-µFTIR investigations for the sample from oil lamp #11551, after acetone, chloroform, and xylene extractions, gave the most interesting results. The acetone extract of sample #11551 showed the spatial micro-separation and detection of an almost pure medium-chain fatty acid (Figure 4), based on its antisymmetric stretch at 2928 cm −1 .

Gas Chromatography-Mass Spectrometry
In all eight samples, a wide range of fatty acids, comprising saturated and monounsaturated monocarboxylic acids, as well as oxo-, hydroxyl-, dihydroxy-, and diacids were identified. (Figure 5. Exemplar chromatogram of sample #11551.) Dicarboxylic acids (or diacids) were of both odd and even carbon numbers, ranging from 4 to 15 carbon atoms and in the majority of samples represented the most abundant class of compounds (i.e., samples #11550, 11551, and 11598), with nonanedioic (azelaic) acid predominating. Dihydroxy acids, bearing 18, 20, and 22 carbon atoms (namely, TMS derivatives of 9,10-dihydroxyoctadecanoic acid, 9, 12-dihydroxy-octadecanoic acid, 11,12-dihydroxy-eicosanoic acid, and 13,14-dihydroxy-docosanoic acid, in many cases as pairs of threo-erythro isomers) were also detected in most of the samples, more pronounced in #11550, 11551, and 11633. Fatty acids were primarily of even carbon number (C8 to C26), while odd-numbered members, bearing 9, 15, and 17 carbon atoms, were present in lower abundance and might indicate the addition of animal fat and/or post-depositional contamination (Evershed et al. 1997). The mono-unsaturated C18:1 and C22:1 were found to be present in low abundance and the latter solely in samples #11550 and 11551.
Monoacylglycerols and glycerol were detected, indicating degraded fats/oils. A few samples also appeared to contain traces of the plant sterol β-sitosterol, implying residue of plant origin. In sample #11622A detached from ingot-containing connection areas, low amounts of natural resin-related diterpenic molecules, such as dehydroabietic and 7-oxo-dehydroabietic acid, components of Pinaceae resins in their oxidized state, were detected [89]. As these were found in the connection area of the pedestaled bowl #11622, resin as a flux for soldering purposes was suggested; alternatively, the addition of resins for their aromatic properties was considered, even though they were not detected in any other sample. A synopsis of the compounds identified in all the samples is provided in Figure 6; the chromatograms of all samples and the entire range of identified compounds are presented in detail in Supplementary Data ( Figure S2 and Table S1, respectively).
A distinctive feature in the chromatograms was the series of dihydroxy fatty acids (diOH FA) and the relatively polar α,ω-dicarboxylic acids detected in all the samples, typically considered as oxidation products of unsaturated fatty acids [8,16,17,90]. Aging tests in Brassicaceae seed oil and investigations of illuminant residues in replica pottery vessels through the burning of various oils [91] explored degradation markers for identifying commodities used as fuel in lamps. The distribution of α,ω-dicarboxylic acids, was suggested to directly reflect the position of the double bond in the original product, because vicinal dihydroxy carboxylic acids are formed through dihydroxylation of the double bonds and hence indicate the original position of the double bond in the precursor fatty acids. In particular, it was shown that both 11,12-dihydroxy-eicosanoic (C20diOH) and 13,14dihydroxy-docosanoic acids (C22diOH) were formed during degradation of 11-eicosenoic (gondoic) and 13-docosenoic acids (erucic), respectively. Similarly, it was suggested that the high proportion of α,ω-undecanoic (C11diFA) and α,ω-tridecanoic (C13diFA) acids, as well as the presence of shorter chain homologous compounds, were also related to oxidative mechanisms affecting the higher non-polar unsaturated fatty acids gondoic and erucic [8,16]. The chromatographic profile obtained, comprising degradation products of unsaturated fatty acids, implied the presence of vegetable oils. A typical fatty acid profile, including 15-tetracosenoic (nervonic), 13-docosenoic, 11-eicosenoic and 9-octadecenoic (oleic) acids, as well as their degradation markers is found in members of the Brassicaceae family [8,91,92]. In addition, 9,12-dihydroxyoctadecanoic acid is produced through hydration of 12-hydroxy-octadecenoic acid (ricinoleic acid), found in high abundance in castor oil. Rapeseed and radish oils, products of the same plant family, and castor oil are widely available in the Eastern Mediterranean and have been historically reported as significant sources of natural material with various uses in antiquity [3,6,8,16,91,93,94]. Finally, in the light of FTIR results (see above), the aforementioned detected acidic compounds also included their metal soap counterparts transferred in the worked-up samples, since the BSTFA reagent (see Experimental Procedures) was capable of derivatizing fatty acids and their salts [95].

Gas Chromatography-Mass Spectrometry
In all eight samples, a wide range of fatty acids, comprising saturated and monounsaturated monocarboxylic acids, as well as oxo-, hydroxyl-, dihydroxy-, and diacids were identified. (Figure 5. Exemplar chromatogram of sample #11551.) Dicarboxylic acids (or diacids) were of both odd and even carbon numbers, ranging from 4 to 15 carbon atoms and in the majority of samples represented the most abundant class of compounds (i.e., samples #11550, 11551, and 11598), with nonanedioic (azelaic) acid predominating. Dihydroxy acids, bearing 18, 20, and 22 carbon atoms (namely, TMS derivatives of 9,10dihydroxy-octadecanoic acid, 9, 12-dihydroxy-octadecanoic acid, 11,12-dihydroxy-eicosanoic acid, and 13,14-dihydroxy-docosanoic acid, in many cases as pairs of threo-erythro isomers) were also detected in most of the samples, more pronounced in #11550, 11551, and 11633. Fatty acids were primarily of even carbon number (C8 to C26), while oddnumbered members, bearing 9, 15, and 17 carbon atoms, were present in lower abundance and might indicate the addition of animal fat and/or post-depositional contamination (Evershed et al. 1997). The mono-unsaturated C18:1 and C22:1 were found to be present in low abundance and the latter solely in samples #11550 and 11551.    Table S1.

Conclusions
The analysis of visible residues found on the interior surfaces of copper alloy vases from the Benaki Museum Byzantine Collection, dating from the 5th to the 8th centuries AD, provided valuable insights into the potential of residue analysis in non-ceramic material. Organic remains were well preserved, and it was possible to assign the origin of the oil-rich organic materials used/processed in the vessels. In particular, the materials identified in the oil lamps could be related to lighting purposes. On the other hand, the oil residue detected in the small bowls could be related to food processing, cosmetic purposes or other domestic activities.
In the detached samples, the oil material was found in various stages of degradation: esters (possibly of glycerol), fatty acids of long or shorter alkyl chains (indicating fatty substances of both animal and plant origin), and metal salts (soaps), products of possible interaction of the above chemical species with metal cations of the corrosion products or other inorganic remains. The identification of solvent-extractable fatty acid oxidation products (i.e., α,ω-dicarboxylic acids and dihydroxy carboxylic acids, including 9,12-dihydroxyoctadecanoic acid, 11,12-dihydroxy-eicosanoic acid, and 13,14-dihydroxy-docosanoic acid), biomarkers and degradation markers of castor oil and Brassicaceae seed oils, allowed the determination of the origin of the residues preserved on the surfaces of the eight metallic vases. Egypt's dry and arid climate offers optimal conditions for the preservation and survival of organic matter [93,95]. The use of cruciferous oil in Egypt has been previously reported in pottery vessels and mollusk shells [91,92,96] and is consistent with ancient sources. However, this is the first time that residues of use in metallic artifacts from Egypt are reported, expanding the range of artifacts [22] that can be studied, providing interpretable results.
After the KBr infrared profile containing numerous inorganic and organic components was broken down by analyzing selected solvent-based organic extracts, compound classes, such as mono-, and di-glyceryl esters, free fatty monoacids with various alkyl chain lengths, as well as metal carboxylates, were confirmed. GC-MS provided detailed chemical data and, together with solvent extraction spectra, added significant information to the preliminary FTIR results. This twofold approach facilitated the determination of the specific nature of the residue preserved, providing information on the current condition and the origin of the investigated remains. Another important aspect of this approach addressed the significance of the container remains, with implications for conservation in compliance with widely adopted ethics and for proposing an ultimately informative museological exhibition approach.  Table S1.

Conclusions
The analysis of visible residues found on the interior surfaces of copper alloy vases from the Benaki Museum Byzantine Collection, dating from the 5th to the 8th centuries AD, provided valuable insights into the potential of residue analysis in non-ceramic material. Organic remains were well preserved, and it was possible to assign the origin of the oil-rich organic materials used/processed in the vessels. In particular, the materials identified in the oil lamps could be related to lighting purposes. On the other hand, the oil residue detected in the small bowls could be related to food processing, cosmetic purposes or other domestic activities.
In the detached samples, the oil material was found in various stages of degradation: esters (possibly of glycerol), fatty acids of long or shorter alkyl chains (indicating fatty substances of both animal and plant origin), and metal salts (soaps), products of possible interaction of the above chemical species with metal cations of the corrosion products or other inorganic remains. The identification of solvent-extractable fatty acid oxidation products (i.e., α,ω-dicarboxylic acids and dihydroxy carboxylic acids, including 9,12-dihydroxyoctadecanoic acid, 11,12-dihydroxy-eicosanoic acid, and 13,14-dihydroxydocosanoic acid), biomarkers and degradation markers of castor oil and Brassicaceae seed oils, allowed the determination of the origin of the residues preserved on the surfaces of the eight metallic vases. Egypt's dry and arid climate offers optimal conditions for the preservation and survival of organic matter [93,95]. The use of cruciferous oil in Egypt has been previously reported in pottery vessels and mollusk shells [91,92,96] and is consistent with ancient sources. However, this is the first time that residues of use in metallic artifacts from Egypt are reported, expanding the range of artifacts [22] that can be studied, providing interpretable results.
After the KBr infrared profile containing numerous inorganic and organic components was broken down by analyzing selected solvent-based organic extracts, compound classes, such as mono-, and di-glyceryl esters, free fatty monoacids with various alkyl chain lengths, as well as metal carboxylates, were confirmed. GC-MS provided detailed chemical data and, together with solvent extraction spectra, added significant information to the preliminary FTIR results. This twofold approach facilitated the determination of the specific nature of the residue preserved, providing information on the current condition and the origin of the investigated remains. Another important aspect of this approach addressed the significance of the container remains, with implications for conservation in compliance with widely adopted ethics and for proposing an ultimately informative museological exhibition approach.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/heritage4040199/s1, Figure S1: Infrared spectra of samples extracted from methanol. Intensities of spectra reflect relative quantities of their components; (a) full spectrum, (b) C-H stretching region, and (c) carbonyl stretch and C-H deformation region. Figure S2. Infrared spectra of samples extracted from acetone. Intensities of spectra reflect relative quantities of their components; (a) full spectrum, (b) C-H stretching region, and (c) carbonyl stretch and C-H deformation region. Figure S3. Infrared spectra of samples extracted from dichloromethane. Intensities of spectra reflect relative quantities of their components; (a) full spectrum, and (b) carbonyl stretch and C-H deformation region. Figure S4. Microscopy FTIR spectra of sample #11551 extracted from xylene and deposited on gold mirror disc; spectra (i)-(iv) correspond to four different micro-separated spots on the disc; (a) full spectra; (b) the C-H stretching region, (c) carbonyl region, and (d) methylene rocking region. Figure S5. (a)-(h): Partial total ion gas chromatograms of all samples with main identified com-ponents; For a key of abbreviated analytes, see Table S1. IS: Hexadecane (internal standard, see Materials and Methods in main text); x: unknown. Table S1: Organic compounds with relative integration areas detected in the various samples through gas chromatography-mass spectrometry.
Author Contributions: K.K.: data curation; investigation; writing-review and editing; S.C.B.: conceptualization; data curation; methodology; project administration; supervision; writing original draft; writing-review and editing; M.R.: data curation; methodology; supervision; writing-review and editing; N.K.: resources; project administration; writing-review and editing; D.K.: conceptualization; project administration; writing-review and editing. All authors have read and agreed to the published version of the manuscript.
Funding: Part of this research was pursued in partial fulfillment of Koupadi's undergraduate thesis; no external funding was received.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Acknowledgments:
The authors wish to thank A. Stefanis, Assistant Professor in the Dept. of Conservation of Antiquities and Works of Art, for providing access to digital optical microscopy facilities; Anthea Phoca, Conservator of Antiquities and Works of Art, Benaki Museum, for her involvement in the materials investigations of the Byzantine containers; Anastasia Drandaki, Assistant Professor of Byzantine Archaeology and History of Art, National and Kapodistrian University of Athens, for providing access to the archaeological material. We would also like to thank M. Christea at the Laboratory of Food Chemistry, Biochemistry and Physical Chemistry, Harokopio University, for her valuable help.

Conflicts of Interest:
The authors declare no conflict of interest.