Bitumen from the Dead Sea in Early Iron Age Nubia

Bitumen has been identified for the first time in Egyptian occupied Nubia, from within the town of Amara West, occupied from around 1300 to 1050 BC. The bitumen can be sourced to the Dead Sea using biomarkers, evidencing a trade in this material from the eastern Mediterranean to Nubia in the New Kingdom or its immediate aftermath. Two different end uses for bitumen were determined at the site. Ground bitumen was identified in several paint palettes, and in one case can be shown to have been mixed with plant gum, which indicates the use of bitumen as a ground pigment. Bitumen was also identified as a component of a friable black solid excavated from a tomb, and a black substance applied to the surface of a painted and plastered coffin fragment. Both contained plant resin, indicating that this substance was probably applied as a ritual funerary liquid, a practice identified from this time period in Egypt. The use of this ritual, at a far remove from the royal Egyptian burial sites at Thebes, indicates the importance of this ritual as a component of the funeral, and the value attributed to the material components of the black liquid.


Results
Five samples of black paint were taken from palettes from the early walled town, and five from palettes from the Western Suburb. Two samples of applied black material were taken from coffin fragments from tomb G244, and five further samples were taken from the black friable lumps found in tomb G321. One reference sample of archaeological Dead Sea bitumen from the British Museum Reference Collection was analysed alongside the samples. This material was procured from Tel Aviv University in the 1970s. Results were compared to data in the literature. The presence in each sample of bitumen, lipids, and gums, determined using separate methods, is shown in Table 1. Positive   four-ring hydrocarbons derived from the degradation of steroids and sterols found in most higher plants and algae (but rare or absent from bacteria) via diagenesis and thermal maturation 16 .

Results -Bitumen.
Of the 10 samples of paint from palettes (PS numbers), 9 were found to contain bitumen, although those from the suburb only in trace amounts. All samples from the friable solid in tomb G321 (AS numbers) were found to contain bitumen. Samples PS295 and PS297 from the painted coffin fragments in G244 also contained bitumen.
Ratio data for a range of bitumen biomarkers are shown in Table 2. The results given in Table 2 are for the samples which provided good enough chromatograms from which to take data. Other samples showed hopanes and steranes but at trace levels such that peaks could not be reliably integrated. The ratios were calculated using the areas under the peaks obtained by manual integration. Selected ratios are plotted in Fig. 5 with comparative reference data.
Results -Lipids. The paint samples from the palettes (PS numbers) contained no resins, oils, fats, or waxes. Analysis results from G321 material (AS numbers) varied, due to the fact that this substance was not homogenous (Supplementary Table S1). Slightly different results were obtained for AS1932 and AS1994 by the two GC methods, most likely due to sample heterogeneity. AS1932 (Fig. 6), AS1933, and AS1994 contained several organic products. The presence of fatty acids, with stearic acid predominating, traces of triacylglycerols, and no diacids, suggests an animal fat 17,18 . Wax esters with carbon chain length 42, 44, 46, and 48, and long chain fatty acids, evidence a natural wax component 19,20 . The mass spectra of the wax esters have a peak at m/z 257 indicating an acid moiety with 16 carbons, thus are a series of even carbon number long-chain palmitate wax esters, probably indicating the presence of beeswax 21 , although no alcohols or alkanes were detected, possibly due to heating in ancient times 19 . In the mass spectra of the wax esters of carbon chain length 44, 46 and 48, m/z 257 is the base peak and there is also a peak at m/z 285 ( Supplementary Fig. S1). For the ester with 42 carbons observed in the chromatogram for AS1933 m/z 257 is present but the base peak is m/z 285. The m/z 285 ion indicates an acid moiety in the wax ester with 18 carbons, suggesting the presence of another waxy material, possibly plant based 22 . Peaks for oleanonic acid and moronic acid, with traces of masticadienonic and isomasticadienonic acids indicate that the resin from Pistacia sp. was a component of the mixture [23][24][25] . The samples taken from AS1941, AS1948, and AS1949 gave very poor chromatograms for lipids. Sample PS295 from the coffin gave a poor chromatogram with a peak for moronic acid, indicating that this material also contained pistacia resin. Given the heterogeneity of the samples from G321, it is possible that a larger sample from this coffin may have included a wider range of ingredients.
Results -Gums. Plant gums are sugary substances composed of monosaccharides. Identifying the range of monosaccharides present can sometimes enable the identification of the plant from which the gum was taken; published analyses of plant gums report the presence of the monosaccharides arabinose, fucose, xylose, mannose, rhamnose, galactose and glucose in varying quantities [26][27][28][29] .
Of the three samples from palettes analysed for monosaccharides, one (PS121) contained fucose, mannose, galactose, and other unidentified sugars ( Supplementary Fig. S2), which indicates that plant gum was used as a binder with the black pigment. The other two were either not mixed with a gum or the material was too degraded to be detected using this method. The presence of fucose suggests that the gum is tragacanth, obtained from the roots of Astragalus sp., but the presence of mannose points to a fruit gum, although mannose has been reported in tragacanth gum by one study 26,29,30 . Astragalus sp. grows in Turkey, Syria, Iraq and Iran, and would have been imported into Egypt 31  www.nature.com/scientificreports www.nature.com/scientificreports/ mummy mask) and two Third Intermediate Period objects (21, mummy mask; 22, falcon) from the Museum of Fine Arts, Boston 32 . The authors concluded that the binders may have included tragacanth but were probably a mixture of gums 32 .

Discussion
Source of bitumen. Bitumen is composed of geologically old organic matter; its molecular make-up depends on the original living organisms that decayed to create it, which varies between formations 33 . Biomarkers are the "molecular fossils" from these organisms that are present in petroleum products, and can be used to identify types of source rock of petrochemicals and to match compounds from the same source 14 . A range of biomarkers should be considered because it is sometimes unclear to what extent each biomarker can predict the depositional environment 14,34 , and there are further issues with archaeological samples such as alterations to the chemical structure of the samples due to their archaeological depositional environment (rather than the www.nature.com/scientificreports www.nature.com/scientificreports/ geological one which formed the bitumen), and contamination from other substances. In addition, the pattern of hopanes (m/z 191) and steranes (m/z 217) can be studied for similarities with samples of known origin 35,36 .
Each biomarker is difficult to interpret independently, but the cumulative evidence seems to point to a marine carbonate source rock for most of the Amara West bitumen (Supplementary Table S2). The chromatograms of the friable material from G321, PS295 from the coffin, and palettes PS119, PS152, and PS415 (herein referred to as Group A) show a very similar pattern. Gammacerane (GCR) is prominent for Group A, and oleanane absent or very low. Diasteranes (m/z 259) are very low or absent for all of Group A. The patterns of the chromatograms for Group A are similar to those obtained for the Dead Sea from the reference sample and in the literature (Figs. 3 & 4). Dead Sea bitumen is characterised by a low pristane/phytane ratio of around 0.5, very low or absent diasteranes, low oleanane, high gammacerane, high C35/C34 homohopanes, and a complete series of tricyclic terpanes (C19-C30), maximizing at C23 35,[37][38][39] . Ts/Tm varies slightly for Dead Sea samples reported in the literature, from 0.04 in geological samples to 0.08 in archaeological samples 13,35 . Most of the samples analysed from Amara West fit into this range, with the exception of PS121. www.nature.com/scientificreports www.nature.com/scientificreports/ The black pigment from Amara West palette PS121 has different biomarker values from the other Amara West samples ( Table 2). The oleanane index is higher (0.14 compared to average 0.013 from the AS samples from G321), and the gammacerane index lower (0.25 compared to average 0.67 from the AS samples from G321; Table 2). These results are unlikely to be explained by biodegradation, as it would be expected that both the oleanane and the gammacerane levels would be elevated 40 . It appears likely that this sample had a non-Dead Sea origin, especially considering the presence of oleanane, which is very low or absent for the Dead Sea and other Amara West samples. Another source of bitumen in Egypt is Gebel Zeit 37,41 , but as can be seen from Table 2 and the radar plot in Fig. 5, the data from Gebel Zeit does not match any of the samples in this study. The origin of the bitumen in PS121 cannot currently be identified, and may be from a mixture of sources, which would make definitive identification difficult.
Significance of bitumen as a pigment. The ground black material from the palettes contained bitumen in 9 out of 10 cases, and one of these was shown to have been mixed with a plant gum binder to make a paint. This is the first molecular identification of bitumen being used as a ground pigment in a pharaonic context. There is one other published example of bitumen used as a pigment in Egypt prior to the Roman Period, on a 19th www.nature.com/scientificreports www.nature.com/scientificreports/ Dynasty model boat from Gurob, but this finding has not been confirmed by molecular analysis 42 . Most black paints and inks from ancient Egypt have been identified as carbon, obtained from burning organic matter [43][44][45] , but given the difficulties of distinguishing elemental carbon from bitumen, which is only really possible using molecular analysis (GC-MS), it is possible that the use of bitumen as a pigment in ancient Egypt has been under identified. The use of bitumen could have been a lot more extensive than is indicated by the evidence to date, and this should be taken into account in future pigment studies. While some of the samples come from occupation and rubbish deposits found within storage magazines (E13.14), and thus might reflect the period when apparatus of the pharaonic state was most prevalent at Amara West (inscriptions, storage facilities, seal impressions), other instances were identified in the latest phases of occupation, suggesting a continuing use of bitumen as a pigment. Nonetheless, the Dead Sea origin of the bitumen indicates that this trade would have been coming from the north, through Egypt.
Given the ease with which carbon can be obtained, i.e. by burning anything organic, the use of bitumen as a pigment must have had a significance. We do not know the end-use of the black pigment in the palettes, but it is possible it was being reserved for a particular use.
Significance of bitumen in a funerary context. The friable material from tomb G321 was found to consist of resin from Pistacia sp., a lipid component (fat or oil), wax, and bitumen. The composition of the G321 material is very similar to mummification "balms", however the black material from Amara West was not found on the body or in body cavities, but instead in scattered fragments and in one area quite a thick, flat puddle close to but not on a (disturbed) body. The black material was applied as a liquid, which allowed textile impressions to form on the surface of the substance when it solidified. This suggests it may have been applied to the exterior of a wrapped body. Black ritually applied liquids are known from the exterior of wrapped bodies and funerary containers from Egypt, and are the subject of current research at the British Museum. Analyses of these externally applied black liquids has shown them to consist of lipids, beeswax, bitumen, and conifer resin or pistacia resin in various combinations (for example, EA6662, EA6660, EA48001, and EA24906 in the collections of the British Museum 46,47 ). The components of the black material from Amara West and the context in which it was found are consistent with the Egyptian black funerary liquids. The black material from the coffin fragments from G244 also contained bitumen and pistacia resin, and appears to have been applied to the surface of the coffin. Given the similarity in components and context to the material in G321, it seems likely that this was also a ritually applied black liquid. The use of similar ingredients in mummification balms and black varnishes on funerary statues suggests that this black liquid had multiple uses in funerary practice 13,33,[48][49][50][51][52][53][54] . A link to Osiris may be inferred from the colour of the substance, Osiris is known as "the black one" and shown with black skin, and from the similarity of the liquid to mummification balms, the deceased is identified as an aspect of Osiris 55 .
Significance of bitumen at Amara West. The biomarkers in the black materials from Amara West are consistent with those of Dead Sea examples, which is likely to be evidence for a trade in solid bitumen from the Dead Sea into Nubia over the 19th and 20th dynasties (c. 1300 to 1070 BC). Evidence for the trade in bitumen into the Nile Valley during the New Kingdom has so far been very limited, so this would be a major contribution to this dataset. Alternatively, the bitumen found in G321 may relate to the later use of the tomb, in the period after Egyptian occupation, as ceramics diagnostic of that date and distinctive Nubian wooden funerary bed fragments were found in the same context. If this is the case, it may reflect the adoption, and perhaps reinterpretation, of Egyptian funerary practises by individuals who identified as Nubian. Previous studies have found bitumen in mummification materials from the Third Intermediate Period to the Roman Period (c. 1086 BC to 300 AD), most of which was shown to have come from the Dead Sea 13,33,41,56-60 , and a trade route for Dead Sea bitumen into Egypt in the 4th to 3rd millennium BC has been identified from lumps of archaeological bitumen 38 . Molecular www.nature.com/scientificreports www.nature.com/scientificreports/ evidence for bitumen from the New Kingdom (pre-dating the Third Intermediate Period) is limited to the black coating on the coffin of Henutmehyt in the British Museum (EA48001) 46 , the balm of a mummified man from Thebes 13 , an identification of Dead Sea bitumen in a 19th Dynasty "mummy balm" 12 , and the presence of hopanes in the black coatings on an 18th Dynasty canopic chest and anthropoid coffin 49 . Amara West was founded by the pharaonic state but the presence at the town of individuals who identified as Nubian is suggested by the production and use of Nubian pottery, and building. E12.11 that reflects Nubian architectural traditions 9,61-63 . The evidence from the tombs at Amara West appears to demonstrate an integration of Egyptian funerary and technological traditions with those from Nubia, such as the tumulus superstructure over tomb G244, from which the coffin fragment comes 9 . In this context it is interesting to see evidence for the use of a ritual black liquid that is linked to the Egyptian embalming tradition.

conclusion
This study provides evidence of the use of bitumen from the Dead Sea and another origin, in Nubia during the early Iron Age pharaonic occupation of the region and its immediate aftermath. This bitumen was found in two contexts, ground and mixed with gum to make a black paint, and as a component of an organic ritual liquid.
The results provide the first molecular identification of bitumen used as a ground pigment in a pharaonic context, albeit outside Egypt itself. It is possible that bitumen has hitherto been under recognised as a painting material in the ancient Nile Valley and should be considered by future pigment studies.
The black substance from G321 and the coffin fragments are examples of an Egyptian funerary ritual using long-distance imported ingredients, for at least two individuals on two separate occasions. The use of this ritual at a far remove from the royal Egyptian burial sites at Thebes, and in graves reflecting the entanglement of Egyptian and Nubian funerary traditions, indicates the importance of this ritual as a component of the funeral, and the value attributed to the material components of the black liquid. The liquid probably had important ritual associations with Osiris, who is associated with the dead and the colour black.
Given that evidence for bitumen use in Egypt in the New Kingdom has previously been limited to a few individual samples from objects with poor provenance 12,13,42,46,49 , this study provides proof for a much more extensive use than might have been suspected, with a secure archaeological context.

Methods
Three methods were used, each designed to analyse the samples for a different set of components: bitumen, lipids, and gums. The aim was to determine the extent of the use of bitumen at the site and whether the other components of the materials would provide clues for different end uses. Previous analysis of Egyptian paints has shown that plant gums were used as a paint binder 31,64-68 , whereas black funerary liquids were complex mixtures of organic products 13,46,50,52,53,56,69,70 . As stated in other publications 13,71 the analysis method will determine the components that can be detected, so different methods were applied.
Method A -bitumen analysis. Samples were dissolved in 1 ml dichloromethane (DCM), and heated at 40 °C for 2 hours, after which the solution was decanted, and dried under a stream of nitrogen. This was done 3 times, combining the extracts. 20 µl DCM and 1 ml hexane were added to the soluble fraction, the asphaltene fraction precipitated out, and this was left overnight to settle. The solution was then decanted and dried under a stream of nitrogen to obtain the maltene fraction.
Each maltene fraction was then fractionated using column chromatography. 100 µl hexane was added to the maltene fraction. Each was decanted into a glass pipette held upright and plugged with glass wool and half filled with dried silica (chromatography grade 60-120 µm, pre-extracted with DCM/methanol 97:3 (v:v), followed by hexane, then oven dried) to which hexane had been added to exclude moisture. The first fraction was extracted using 3 ml hexane washed through the pipette; the second using 3 ml DCM:hexane 1:3 (v:v); the third using 3 ml DCM:methanol 2:1 (v:v). The elutes were collected and dried in a stream of nitrogen. For analysis, 50 µl of hexane was added to the first fraction and this was decanted to a micro vial.
The Method B1. The GC-MS analysis was carried out with a SGE HT5 GC column (12 m × 0.22 mm; 1 µm film thickness) with splitless injection, coupled to an Agilent 5975 C MSD. The mass spectrometer was operating in the electron impact (EI) mode at 70 eV and scanning m/z 50 to 1000. The oven was set at 50 °C to 370 °C at 10 °C per minute, isothermal for 15 minutes.
Method B2. Two further samples were taken from AS1932 and AS1994, which were prepared in the same way, and run on a longer method to further separate the peaks. The oven was set at 50 °C to 320 °C at 5 °C per minute and then 10 °C to 370 °C, isothermal for 15 minutes. PS295 was only run on the longer method, because the sample was taken at a later date. Method C -Gum analysis. Samples for gum analysis were taken from PS119, PS121 and PS139. The method followed for the analysis of Amara West paints for plant gums was the standard operating procedure used at the British Museum for the preparation of polysaccharide samples for GC-MS analysis of neutral sugars and uronic acids, based on a published method 26 .
Samples and reference samples were hydrolysed by the addition of 500 µl of 0.5 M hydrochloric acid and heated at 80 °C for 20 hours. The solution was decanted and dried under nitrogen. Samples were derivatised by the addition of 300 µl Sigma-Sil A (1:3:9 ratio of trimethylchlorosilane (TMCS), hexamethyldisilazane (HMDS) and pyridine), and heated at 80 °C for 2 hours. Samples were dried under nitrogen and dissolved in 100 µl hexane in preparation for injection into the GC-MS instrument. A blank and three reference samples were prepared alongside the samples using the same method.
The instrument and column used were the same as for the bitumen analysis. The oven was set at 40 °C to 130 °C at 9 °C/min, then to 290 °C at 2 °C/min, with the final temperature held for 10 mins.
In all cases the data were analysed using Masshunter software and the NIST database.

Data availability
All relevant data are within the paper and its Supporting Information files. Archive documents referred to in the paper are held in the archives of the British Museum, Department of Scientific Research (Project Record no. 7671) and can be viewed in hard copy or electronically by appointment through science@thebritishmuseum.ac.uk.