The significance of petroleum bitumen in ancient Egyptian mummies

Mummification was practised in ancient Egypt for more than 3000 years, emerging from initial observations of buried bodies preserved by natural desiccation. The use of organic balms (and other funerary practices) was a later introduction necessitated by more humid burial environments, especially tombs. The dark colour of many mummies led to the assumption that petroleum bitumen (or natural asphalt) was ubiquitous in mummification; however, this has been questioned for more than 100 years. We test this by investigating 91 materials comprising balms, tissues and textiles from 39 mummies dating from ca 3200 BC to AD 395. Targeted petroleum bitumen biomarker (steranes and hopanes) analyses by gas chromatography-mass spectrometry selected ion monitoring (GC-MS SIM, m/z 217 and 191) showed no detectable bitumen use before the New Kingdom (ca 1550–1070 BC). However, bitumen was used in 50% of New Kingdom to Late Period mummies, rising to 87% of Ptolemaic/Roman Period mummies. Quantitative determinations using 14C analyses reveal that even at peak use balms were never more than 45% w/w bitumen. Critically, the dark colour of balms can be simulated by heating/ageing mixtures of fats, resins and beeswax known to be used in balms. The application of black/dark brown balms to bodies was deliberate after the New Kingdom reflecting changing funerary beliefs and shifts in religious ideology. This article is part of the themed issue ‘Quantitative mass spectrometry’.

Mummification was practised in ancient Egypt for more than 3000 years, emerging from initial observations of buried bodies preserved by natural desiccation. The use of organic balms (and other funerary practices) was a later introduction necessitated by more humid burial environments, especially tombs. The dark colour of many mummies led to the assumption that petroleum bitumen (or natural asphalt) was ubiquitous in mummification; however, this has been questioned for more than 100 years. We test this by investigating 91 materials comprising balms, tissues and textiles from 39 mummies dating from ca 3200 BC to AD 395. Targeted petroleum bitumen biomarker (steranes and hopanes) analyses by gas chromatography-mass spectrometry selected ion monitoring (GC-MS SIM, m/z 217 and 191) showed no detectable bitumen use before the New Kingdom (ca 1550-1070 BC). However, bitumen was used in 50% of New Kingdom to Late Period mummies, rising to 87% of Ptolemaic/Roman Period mummies. Quantitative determinations using 14 C analyses reveal that even at peak use balms were never more than 45% w/w bitumen. Critically, the dark colour of balms can be simulated by heating/ageing mixtures of fats, resins and beeswax known to be used in balms. The application of black/dark brown balms to bodies

Material and methods (a) Samples and sample treatments
Samples of tissue, 'resin' and bandaging were obtained from mummies from a range of museums (electronic supplementary material, table S1). Samples were taken from various locations on the (b) Gas chromatography-mass spectrometry selected ion monitoring All the hydrocarbon factions were submitted to GC-MS using a Finnigan Trace instrument (Finnigan MAT GmbH, Bremen, Germany) equipped with an on-column injector. The mass spectrometer was set to scan in the range of m/z 50-700 in a total time of 0.6 s. For SIM, the mass spectrometer was set to monitor m/z 191 (hopanes and other triterpanes) and 217 (steranes) [20,23]. The GC column was a CPSIL-5 (60 m × 0.32 mm × 0.1 µm) and the operating conditions were 50-130°C at 20°C min −1 , to 300°C (held 30 min) at 40°C min −1 . He was used as carrier gas, the electron emission current was 300 µA, the ion source temperature was 170°C and the GC-MS interface was maintained at 350°C. The electron ionization potential was 70 eV.
Each hydrocarbon fraction was run twice, first without internal standards. Only those bitumen hydrocarbon fractions displaying detectable peaks (s/n > 5 : 1) at appropriate retention times in the initial sterane and hopane analyses were submitted to quantitative analysis. The integral of all the peak areas in the m/z 217 and 191 mass chromatograms provided the basis of the quantification based on co-injected standards. Quantification of the sterane and hopane biomarkers was achieved through electronic integration of the peak areas and comparison with the area of co-injected standards; 5α-cholestane was selected for the steranes and hop-21-ene for the hopanes as these have similar fragmentation patterns and ion yields to the group of biomarkers being quantified. A calibration curve was determined prior to analysis to aid the choice of a suitable standard concentration. The limit of detection of these standards was found to be 0.05 ng of hopane and 0.005 ng of sterane injected.
The concentration of biomarkers in the extracted mummy balm was then calculated using these standards and using the weight of the aliquot of the total lipid extract fractionated into the saturated hydrocarbon fraction and the volume the saturated hydrocarbon fraction was dissolved in for analysis. The integral of all the peak areas in the m/z 191 and 217 mass chromatograms provided the basis of the quantification. The concentrations of hopanes and steranes were calculated as follows: The difference in the way the sterane and hopane concentrations are determined is due to the way the internal standards are used. The area of the added hop-21-ene is readily determined as it does not co-elute with any component of the bitumen. However, the area of the co-injected 5α-cholestane must take account of the cholestane endogenous to the archaeological or reference bitumen. This is overcome in the standard addition method by using the difference in peak area with and without the co-injected 5α-cholestane standard. In all calculations, the reasonable assumption is made of similar response factors between the analytes and co-injected standards. The errors associated with determination of the biomarkers using the co-injection method are principally associated with calculation of the concentration of the standard, the measurement of the aliquot of both the standard and sample for co-injection and the initial weighing of the sample, insoluble residue and the aliquot used for fractionation. The various points at which the sample was weighed have an associated error of ±0.0005 g, which for a sample size of 50 mg accounts for an error of approximately 1%. As each sample was weighed three times (initial sample, insoluble residue and aliquot for fractionation), this gives an overall error of 2% (≈ √ (1 2 + 1 2 + 1 2 )). The error in the concentration of the standard is 1% for the weighing of the standard and 10% for the volume, giving an overall error of 10%. Finally, the error of the co-injection is also calculated to be 10%. These errors combined give a total error for the determination of the biomarker concentration in bitumen of approximately 15%. Given the latter biomarker concentrations are reported in electronic supplementary material, table S1 as semi-quantitative concentration ranges rather than precise concentrations: + = 0.01-0.1 µg g −1 and 0.1-1 µg g −1 ; ++ = 0.1-1 µg g −1 and 1-10 µg g −1 ; +++ = 1-10 µg g −1 and 10-100 µg g −1 ; ++++ = 10-100 µg g −1 and 100 to greater than 1000 µg g −1 , for steranes and hopanes, respectively.

(c) Radiocarbon analyses by accelerator mass spectrometry
A subset of samples of 'resin' and bandaging (purified cellulose [28]) were analysed at the Oxford Radiocarbon Accelerator Unit (ORAU, UK) using a continuous-flow CHN analyser (Europa-ANCA) fitted with a CO 2 collection facility to provide CO 2 as the target material for gas source AMS [29]. From AMS analysis, a value for the 14 C content can be derived, expressed as % modern 14 C. Isotopic fractionation effects are accounted for by normalizing the measurements to the common δ 13 C value of −25‰ and adding or subtracting 8.2 14 C years for each 1‰ difference. The radiocarbon age can be expressed as a radiocarbon age (in years BP) using the following expression: radiocarbon years (BP) = −τ × ln %mod 100 , (2.4) where τ is the Libby mean-life (8033 years) and %mod is the percentage of 14 C remaining relative to modern levels (i.e. AD 1950). This corrected age is then be calibrated against the [30] calibration curve using the OxCal v. 3.9 program [31] to provide a calendar date range. The radiocarbon ages of balm and the textile samples (e.g. figure 1) were used to determine bitumen content on the basis of the differences in their dates, δ radiocarbon years. The presence of 1% radiocarbon 'dead' carbon shifts the true age by 80 years (see equations (2.5) and (2.6); [32]). Contamination by radiocarbon-dead carbon will affect the measured radioactivity of the balm according to the expression where A m is the measured activity, A x the activity of the contaminant (i.e. petroleum bitumen), A s the activity of the true sample and f the fraction of the contamination. Conversion of measured activity to time is achieved using the following standard radioactivity equation (substituting A x and A s ): where 8033 years is Libby's mean-life and A 0 is the modern activity. Where contamination is by infinitely old carbon, A x = 0, so, using the above equations, it can be shown that 1% of infinite age (dead) carbon adds ca 80 years to the apparent age of the sample. Rearrangement of the radiocarbon equations gives the percentage of 'dead' carbon present from the difference in radiocarbon ages as the following equation: where radiocarbon years = radiocarbon age (resin/tissue)-radiocarbon age (textile). Bitumen from the Dead Sea contains 78% carbon [12], which can be used to convert % dead carbon to % bitumen. The difference, radiocarbon years, was calculated using the convolution of the two functions (equation (2.9)) which also gives the associated error [31]: The convolution effectively 'blends' one function (p 1 or the radiocarbon age of the bandage) with another function (p 2 or the radiocarbon age of the 'resin') giving the distribution of the difference in radiocarbon age.

Results and discussion
The investigation proceeded to two stages. Initially, extracts of mummy balms, textiles and tissues were screened for the presence of diagnostic hopane (and other triterpane) and sterane petroleum biomarkers using GC-MS with SIM. The results of this screening phase were used to map the use of petroleum bitumen through time and also to identify a subset of samples for subsequent radiocarbon analysis to quantify the petroleum bitumen concentration in mummy balms.

(a) Screening mummy balms for petroleum biomarkers
All the balms were initially screened by GC-MS to determine the major fat/oil, di-/triterpenoid resin and beeswax components of the balms [24]. Since the biomarkers for petroleum bitumen are known to be present in trace concentrations, extracts were fractionated to yield saturated hydrocarbon fractions (figure 2a) that were analysed by GC-MS SIM (figure 2b,c; m/z 191 and 217) for the targeted analysis of bitumen sterane and hopane biomarkers at high sensitivity (results are summarized in electronic supplementary material, table S1). The analytical procedure was validated by analysing reference bitumens including Dead Sea, Gebel Zeit and Abu Durba, which gave analogous SIM chromatograms [24] to those published in Harrell & Lewan [20]. Steranes concentrations were always lower than hopanes (and other triterpanes). The difficulties in detecting steranes and hopanes in mummy balms indicate that the bitumen biomarkers are present at low concentrations. Quantification of biomarkers was performed where they were deemed to be present in sufficiently high concentrations (electronic supplementary material, table S1). The results show that in the majority of mummy balms the concentration of biomarkers range between approximately 10 µg g −1 and 500 µg g −1 ; the highest concentration detected was from the tissues of a female Greek mummy (MTB 7700/4963) where the concentration of hopanes was approximately 1500 µg g −1 . In all cases, the concentrations of steranes and hopanes are considerably lower (by several orders of magnitude) than the concentrations of lipids from fats/oils, beeswax and resins found in the balm, which are found in typically mg g −1 concentrations. Where detected the majority of the bitumens identified can be attributed to the previously recognized Dead Sea source [18][19][20][21]24]. Significantly, none of the mummies dating before ca 1000 BC contained detectable bitumen biomarkers. An example of one of these early mummies was the male adult Khumnnakht (Middle Kingdom, ca 1985-1795 BC) previously shown to comprise mainly fat/oil [22,33]; GC/MS with SIM indicated no detectable steranes or hopanes (figure 3c). By contrast, many mummy balms from later periods of Egyptian history exhibited evidence for the presence of steranes and hopanes (figure 3a,b; electronic supplementary material, table S1). Their occurrence was most common in mummies from the Ptolemaic to Roman Periods (332 BC on), rather than from those of the  In the TIC, the numbers on the peaks correspond to the carbon numbers of the major n-alkanes, which derive predominantly from beeswax. In the GC-MS SIM m/z 191 and 217 mass chromatograms, the horizontal bars correspond to the retention time windows within which the major hopanes (and other triterpane biomarkers, i.e. oleanane, small peak eluting just before the C 30 hopane and gammacerane, small peak eluting between the C 31 and C 32 hopanes) and sterane biomarkers elute. The numbers denote the carbon numbers of the components eluting within those ranges. The multiple peaks within the carbon number group under each horizontal bar correspond to isomeric mixtures produced during the petroleum formation process. Further explanation of the biomarker compositions are given by Connan [23]. The m/z 191 ion for hopanes is formed by EI cleavage of the C-ring with charge retention on the A + B ring containing fragment, while the m/z 217 ion for steranes is formed by D-ring cleavage with charge retention on the A + B + C ring fragment (see Fig. 2 in reference [20]). The rising baseline in the m/z 191 mass chromatogram arises from the presence of the ion in the column bleed accentuated by the low concentration of hopanes.    appears to be a general increase in mummification of individuals, which peaked in the Graeco-Roman era when mummification became even more common across social classes and age groups [5].
It was notable that a number of extremely black mummies lacked detectable bitumen biomarkers, notably the Roman era mummy with the folded arms (TUR Pravv 540), and the XXI dynasty mummy (BM 6660, Third Intermediate Period; figure 1). Both balms were very black and visually might easily be interpreted as being bitumen or containing a significant concentration of bitumen. Interestingly, the mummy with the folded arms (TUR Pravv 540) has been dated to the end of the Roman period and might have been expected to contain bitumen. A number of other mummies dating to the Ptolemaic and Roman periods similarly lacked detectable bitumen biomarkers, such as the male adult with the prosthetic hand (DUR 1999.31.1) and the 'resin' from the head of a female adult (RMO 41). These results clearly indicate that the use of bitumen was a complex activity, and that the widespread use of true bitumen was a later introduction. Clearly, the black colour of mummies is an unreliable indicator of the presence of bitumen.

(c) Determination bitumen concentration in mummy balms
A further important question concerning the importance of bitumen in mummy balms is the proportion of bitumen used to prepare organic balms relative to the other ingredients, such as fat/oil, beeswax and resin [16][17][18]22,24]. One method of determining the concentrations of bitumen that was used in the past is based on the concentrations of the biomarker components of balms relatively to the source bitumen [15,23]. However, the results obtained using this approach are affected by the natural variability in concentrations of biomarkers in the sources due to varied diagenetic histories, together with possible mixing of bitumen from different sources during the preparation of balms. In view of the latter, we adopted a radiocarbon approach reasoning that since bitumen is of geological age it would be radiocarbon 'dead'. Thus, the 14 C content would be negligible, and the presence of any bituminous material in the balm would dilute the 14 C present in the balm, thereby causing a shift in radiocarbon date towards older ages [41,42]. By comparing this date with the date from other materials from the same mummy, contemporaneous with the body and free of bitumen, it is possible to apportion the  concentration of bitumen present in the balm. Hence, a subset of the collection of mummy balms was chosen according to the following criteria: (i) they had well established dates based on archaeological/stylistic/contextual/typological criteria; (ii) they covered a wide range of dates, which did not fall in flat areas of the radiocarbon calibration curve; (iii) samples of bandaging and balm were available from the same mummy; and (iv) variable bitumen concentrations were suggested, based on bitumen biomarker concentrations ranging from mummies with no bitumen, barely detectable bitumen biomarkers to those with readily detectable biomarkers. The results of the radiocarbon analyses of the balms and bandaging 'pairs' are shown in figure 5 and table 1 giving the radiocarbon age, the calibrated age, the difference, radiocarbon years and the percentage of dead carbon (attributed to the presence of bitumen) that would cause the observed differences in the dates. The results from Khumnakht (MAN 21471) show a small negative difference between the dates from the bandaging and the 'resin' (table 1) corresponding to the presence of 0-3% of 'dead' carbon in the bandaging. As cellulose was purified from the bandaging, the possibility of contamination by bitumen was eliminated and the percentage of dead carbon must be zero. The 'resin' from the Glasgow male (MAN G6) also shows a small difference in the radiocarbon age between the bandaging and the 'resin' of 40 and 310 years (table 1). This difference in age corresponds between 0.5% and 4% of 'dead' carbon and, therefore, a maximum of only 0.6-5% of this balm comprised bitumen. Given the blackened nature of this mummy, which would normally be attributed to the presence of bitumen in the balm, the low fraction of 'dead' carbon suggests that the blackened nature of this mummy is due to other factors. Components identified in the latter balm included fatty acids originating from the application of a fat or oil, wax esters from beeswax and diterpenoids from coniferous resin [24]. Hence, the black colour of the mummy must result largely through their darkening during balm production and use or through ageing. Experiments aimed at replicating ancient Egyptian balms have shown that when resins are melted with oils very dark blackened coatings are indeed produced (S. Ikram 2012 and 2014, unpublished data).
The findings from the Ptolemaic mummies show considerably greater differences between the radiocarbon ages obtained from the resin and bandaging (table 1) Table 1. Results of AMS radiocarbon analyses performed on textiles and balms to estimate petroleum bitumen content of balms.

Conclusion
It has been demonstrated that for the first 2000 years in which mummification was practised prior to the New Kingdom petroleum bitumen (or natural asphalt) was not used in embalming as a general practice. The earliest evidence for the presence of bitumen in a mummy balm derives from a single individual dating to the end of the New Kingdom (1250-1050 BC; [18]). The use of bitumen in balms becomes more prevalent during the Third Intermediate Period, ca 750 BC and was extensively used during the Ptolemaic and Roman periods. Radiocarbon analyses have shown that even when present, balms were likely never wholly composed of bitumen. This might reflect its initial rarity, or the belief that some of the traditional materials had to be used if the mummification were to be efficacious. Although the use of bitumen became widespread in later periods, it was not ubiquitous, as confirmed by sterane and hopane biomarker analyses of mummy balms from these periods. The increase in its use is attributable to a variety of factors, some practical, e.g. antimicrobial properties, as with other components of balms [22,43], and others cultural. It probably provided a simpler and speedier means of mummification-many Graeco-Roman mummies lack the excerebration and in some cases evisceration that were commoner in earlier periods [6,36]allowing for the embalming of larger numbers of people, across social classes and age groups. Additionally, the various sources likely became increasingly accessible as trade routes opened up and the control of the whole area was in Roman hands.
Another explanation for the introduction of bitumen during the Late and Graeco-Roman Periods might be due to a shift in funerary beliefs that involved colouring the body black. The symbolism associated with the colour black is significant: black was associated with the colour of the rich, fertile silt deposited by the annual Nile flood, a symbol of regeneration, rebirth and resurrection, and a colour, together with green, attached to Osiris, god of the dead, lord of the afterlife [5,42,44], and master of resurrection. By darkening the deceased's body during the final phases of mummification so that it became black, he or she was literally transformed into Osiris (see discussion in [7]), living eternally.
Bitumen itself seems to have been regarded as a commodity associated with sacredness and divinity. The ancient Egyptian word for bitumen is usually translated as mnnn, which has parallels with mny [45]. Other texts mention 'mny on the flesh of the gods' [46] and mnît mixed with îhmt (an unknown substance) to prepare the ointment, mrhet, for application to the limbs of the god Amun [47], also a fecundity deity, especially in his guise as Amun-Min/Kamutef [48], who was also shown with black flesh. A recipe from the temple of Edfu lists mnn as an ingredient of aat-netrjret translated as 'divine stone', which was applied to images of the ithyphallic fertility god Min, who himself was often described as black like mn [49,50], and was also associated with aspects of Osiris. Thus, based on the results of this study, it would seem that both practical and theological associations with bitumen are responsible for the increase in its use, and of dark coloured balms generally, in the latest periods of Egyptian history, as it democratized death and the transformation of the deceased into Osiris [5,51].