Characterization of archaeological beeswax by electron ionization and electrospray ionization mass spectrometry

Article abstract of DOI:10.1021/ac025637a

To better detect and identify beeswax in ancient organic residues from archaeological remains, we developed a new analytical methodology consisting of the analysis of (i) the trimethylsilylated organic extract by GC/MS and (ii) the crude extract by ESI-MS. Selective scanning modes, such as SIM or MRM, permit separate quantification of each chemical family (fatty acids, monoesters, monohydroxyesters, and diesters) and allow an improvement in sensitivity and selectivity, allowing the crude extract to be treated without further purification. GC/MS (SIM) was revealed to be a powerful method for the detection of components, with a detection limit down to a total lipid extract in the range of ～50 ng in a complex matrix, such as archaeological degraded material, whereas ESI-MS/MS is instead used for the detection of nonvolatile biomarkers. Identification by GC/MS (SIM) and ESI-MS/MS (MRM) of more than 50 biomarkers of beeswax in an Etruscan cup at the parts-per-million level provides the first evidence for the use of this material by the Etruscans as fuel or as a waterproof coating for ceramics.

Full text of DOI:10.1021/ac025637a

Anal. Chem. 2002, 74, 4868-4877
Characterization of Archaeological Beeswax by
Electron Ionization and Electrospray Ionization
Mass Spectrometry
Nicolas Garnier,† Ce´ cile Cren-Olive´,‡ Christian Rolando,‡ and Martine Regert*,†
Centre de Recherche et de Restauration des Muse´es de France, UMR CNRS 171, Laboratoire de Recherche
des Muse´es de France, 6, rue des Pyramides, 75041 Paris Cedex 01, France, and Universite´ des Sciences et
Technologies de Lille 1, UMR CNRS 8009, Chimie Organique et Macromole´culaire, UFR de Chimie, Baˆtiment C4,
59655 Villeneuve d’Ascq Cedex, France
products collected, their use, or their transformation by human
beings. Recent developments in the chemistry of archaeological
materials have shown that beeswax could be preserved for several
millennia, especially when trapped in the porous clay matrix of
ancient pottery.2,4-7
To better detect and identify beeswax in ancient organic
residues from archaeological remains, we developed a
new analytical methodology consisting of the analysis of
(i) the trimethylsilylated organic extract by GC/ MS and
(ii) the crude extract by ESI-MS. Selective scanning
modes, such as SIM or MRM, permit separate quantifica-
tion of each chemical family (fatty acids, monoesters,
monohydroxyesters, and diesters) and allow an improve-
ment in sensitivity and selectivity, allowing the crude
extract to be treated without further purification. GC/ MS
(SIM) was revealed to be a powerful method for the
detection of components, with a detection limit down to a
total lipid extract in the range of ∼5 0 ng in a complex
matrix, such as archaeological degraded material, whereas
ESI-MS/ MS is instead used for the detection of nonvolatile
biomarkers. Identification by GC/ MS (SIM) and ESI-MS/
MS (MRM) of more than 5 0 biomarkers of beeswax in an
Etruscan cup at the parts-per-million level provides the
first evidence for the use of this material by the Etruscans
as fuel or as a waterproof coating for ceramics.
From a chemical standpoint, beeswax is a very complex
material made of saturated and unsaturated n-alkanes, long-chain
wax esters, fatty acids, diesters, hydroxyesters, and other minor
constituents, such as flavonoids.8-14 Some of these biomarkers
were identified in ancient potsherds dating from the Neolithic to
the Medieval period by several authors who performed GC/ EI-
MS analysis on the total lipid extract of various archaeological
samples.2,4-7 The composition of beeswax is often altered through
time: the long-chain esters are partly hydrolyzed in palmitic acid
and long-chain alcohols, and the alkanes may have sublimed either
when preserved in a dry and warm environment as in Egypt or
under the action of men who may have heated the beeswax before
use.2,4,6 Despite these chemical alterations, beeswax is generally
characterized by the chromatographic pattern of n-alkanes and
long-chain esters usually associated with their degradation prod-
ucts.
Among the beehive products, beeswax was widely used until
the second half of the 19th century for a large variety of purposes
in the fields of technology, art, and medicine and in religious
rituals.1,2 Since the beginning of the Christian era, beeswax was
particularly used in lighting for candle-making before the develop-
ment of the oil industry and chemistry, which brought about the
replacement of beeswax by paraffin and stearin (a mixture of
saturated fatty acids obtained by saponification of animal fats
followed by their purification).3
However, such an analytical approach is not efficient for the
detection of the volatile molecular components that are known to
be present in beeswax, such as diesters and hydroxyesters. It must
also be noted that the analysis of archaeological beeswax by GC/
MS does not always allow a suitable identification of the charac-
teristic biomarkers, especially when it is mixed with other
degraded organic materials or when it is preserved in low amount.
(4) Heron, C.; Nemcek, N.; Bonfield, K. M.; Dixon, D.; Ottaway, B. S.
Naturwissenschaften 1 9 9 4 , 81, 266-269.
(5) Charters, S.; Evershed, R. P.; Blinkhorn, P. W.; Denham, V. Archaeometry
Several archaeological clues, especially rock shelter paintings
give evidence for the collection of wild beehive products since
prehistoric times, even before the development of beekeeping.
However, these paintings do not reveal any information about the
1 9 9 5 , 37, 113-127.
(6) Evershed, R. P.; Vaughan, S. J.; Dudd, S. N.; Soles, J. S. Antiquity 1 9 9 7 ,
71, 979-985.
(7) Regert, M.; Dudd, S. N.; Pe´trequin, P.; Evershed, R. P. Rev.′ Arche´om. 19 99 ,
23, 91-99.
* Corresponding author. Phone: 331 40 20 68 57. Fax: 331 47 03 32 46.
E-mail: martine.regert@culture.fr.
(8) Kolattukudy, P. E. Chemistry and Biochemistry of Natural Waxes; Elsevier
Science: Amsterdam, Oxford, New York, 1976.
† LRMF.
(9) Tulloch, A. P. J. Am. Oil Chem. Soc. 1 9 7 2 , 49, 609-610.
(10) Tulloch, A. P. J. Am. Oil Chem. Soc. 1 9 7 3 , 50, 367-371.
(11) Tulloch, A. P.; Hoffman, L. L. J. Am. Oil Chem. Soc. 1 9 7 2 , 49, 696-699.
(12) Greenaway, W.; Scaysbrook, T.; Whatley, F. R. S. Proc. R. Soc. London 1987,
B 232, 249-272.
‡ USTL.
(1) Lucas, A. Ancient Egyptian Materials and Industries. Chatham: London, 1962.
(2) Regert, M.; Colinart, S.; Degrand, L.; Decavallas, O. Archaeometry 2 0 0 1 ,
43, 549-569.
(3) Gay-Lussac, J. L.; Chevreul, M. E. “Pour l’emploi des acides ste´ariques et
(13) Aichholz, R.; Lorbeer, E. J. Chromatogr. A 1 9 9 9 , 855, 601-615.
(14) Aichholz, R.; Lorbeer, E. J. Chromatogr. A 2 0 0 0 , 883, 75-88.
margariques dans l’e´clairage”; French Patent no. 4323; January 5, 1825.
4868 Analytical Chemistry, Vol. 74, No. 19, October 1, 2002
10.1021/ac025637a CCC: $22.00 © 2002 American Chemical Society
Published on Web 08/24/2002

Moreover, only very tiny samples are generally available for
analysis, because ancient ceramic remains are usually very
precious.
the organic phase was washed with saturated aqueous solutions
of citric acid (3 × 1 mL) and NaHCO3 (3 × 1 mL). The organic
extract was dripped through an anhydrous MgSO4 column to
remove water and was evaporated to dryness. The residue was
submitted to flash-column chromatography (SiO2, 230-400 mesh,
Sigma, Milwaukee, WI; 1 g, CH2Cl2) to obtain purified saturated
palmitates (C38-C46).
The monoesters derived from palmitoleic acid (C16:1), stearic
acid (C18:0), and oleic acid (C18:1) were synthesized and purified
by the same method.
To increase the number of biomarkers detected in ancient
samples containing beeswax and to improve the selectivity and
the sensitivity of the methods allowing the detection of this natural
resource in archaeological remains, we developed a methodology
permitting the study of both volatile and nonvolatile molecular
components. The first ones were characterized and quantified by
GC/ MS using both full scan and selected ion monitoring (SIM)
modes in order to obtain a better sensitivity. The higher-molecular-
weight compounds were identified by ESI-MS and ESI-MS/ MS.
To complete the semiquantification of volatile compounds per-
formed by GC/ MS using the SIM mode, total lipid extracts were
analyzed by ESI-MS/ MS using the multiple reaction monitoring
(MRM) operation.
Fractionation of Modern Beeswax Constituents. A solution
of raw beeswax (1 mg/ mL) from honey bee species Apis mellifera
provided by apiarists working in the Doubs (France) was prepared
by extraction with CHCl3/ MeOH (2:1, v/ v) by ultrasonication (10
min). An aliquot of this solution (100 µL) was evaporated to
dryness and dissolved in dichloromethane (25 µL). The total lipid
fraction was then fractionated by anion-exchange chromatography
with an aminopropyl cartridge (Varian, Palo Alto, CA) equilibrated
with n-hexane. The neutral components were eluted by 12 mL of
a i-PrOH/ dichloromethane mixture (1:2, v/ v), and the acid
constituents are recovered by applying 12 mL of a solution of 5%
HCO2H in dichloromethane.
To obtain valuable results on archaeological samples, we first
proceeded to the analysis of contemporary reference beeswax
previously fractionated into its main chemical constituents and
also the analysis of pure standards either commercially available
or directly synthesized in our laboratory.
Using both gas chromatography coupled to mass spectrometry
with electron ionization and mass spectrometry with electrospray
ionization was shown to be particularly efficient to identify a large
series of biomarkers of beeswax in both contemporary and
archaeological samples. In particular, the selective scanning modes
in mass spectrometry (SIM and MRM) allow a significant increase
in sensitivity, thus making possible the study of small ceramic
sherds containing beeswax (less than 400 mg of sherd). The ESI-
MS analysis is complementary to the GC/ MS data and allows the
characterization of high-molecular-weight compounds that were
never detected in any archaeological samples previously. This
analytical approach, based on a significant number of biomarkers
identified, allowed the unambiguous characterization of beeswax
in Etruscan samples. From a historical point of view, these results
provide interesting new data, completing the information published
by Latin poets and agronomists about the use of honey for the
aromatization or the chaptalization of the wine or game seasoning,
but lacking precision about beeswax exploitation.15
The neutral fraction (evaporated to dryness and dissolved in
50 µL of n-hexane) was fractionated by flash chromatography on
a silicagel column (SiO2, 230-400 mesh, Sigma, Milwaukee, WI)
washed with 10 mL of n-hexane. n-Alkanes were first recovered
by eluting with n-hexane (5 mL); diesters, by applying cyclohex-
ane/ dichloromethane (9:1, v/ v, 9 mL); palmitic esters, by using
n-hexane/ dichloromethane (1:1, v/ v, 9 mL); and hydroxyesters
and alcohols were recovered by eluting with dichloromethane (10
mL).
Archaeological Samples. These samples consisted of 30
ceramic potsherds that lacked visible surface organic residues;
consequently, this research was based on the study of the organic
matter trapped in the porous inner layer of the vessel. The
treatment applied to these samples was adapted from Evershed
and collaborators.5 Potsherds (0.4 to 1 g) were first cleaned by
scraping with a scalpel blade to remove all surface contaminants
and were then crushed using a glass mortar and pestle. After
addition of an internal standard (n-C34, 20 µL of a 0.1 mg/ mL
solution in dichloromethane), the powdered pieces were extracted
twice by sonication for 15 min (CHCl3/ MeOH, 2/ 1, v:v, 10 mL).
The organic extract was centrifuged at 4000 rpm for 20 min, and
the clear solution was evaporated to dryness (nitrogen stream,
40 °C). The lipid extract was diluted in dichloromethane (500 µL)
and an aliquot (100 µL) was treated further by derivatization.
Derivatization. In some instances, derivatization of com-
pounds containing acidic hydrogen atoms was required prior to
high-temperature GC (HTGC) analysis. To avoid incomplete
derivatization due to the poor solubility of long-chain aliphatic lipid
components in organic solvents, all of the samples were deriva-
tized by using BSTFA (50 µL, Sigma-Aldrich, Milwaukee, WI) and
pyridine (5 µL) as solvent and nucleophilic catalyzer.13 The clear
mixture was held at 80 °C for 30 min, evaporated to dryness in a
flush of nitrogen (40 °C), and then dissolved by ultrasonic
treatment in dichloromethane (50 µL).
EXPERIMENTAL SECTION
Synthesis of Monoesters. These compounds were synthe-
sized by adapting protocols previously described.16,17 Aliquots (100
µL, 1 µmol) of five solutions of even-numbered n-alcohols (C22-
C30, Sigma-Aldrich, Milwaukee, WI) in dichloromethane (10 µmol/
mL) were mixed and evaporated to dryness under a stream of
nitrogen. A 50-µL portion of a 0.1 mmol/ mL solution of palmitic
acid (Sigma-Aldrich, Milwaukee, WI), 50 µL of a 100 µmol/ mL
solution of dicyclohexylcarbodiimide (DCC) in dichloromethane
(5 µmol), and 5 µL of a 100 µmol/ mL solution of 4-dimethylamino-
pyridine (DMAP) in dichloromethane were then added. The
mixture was purged with nitrogen and stored at room temperature
for 2 h. Approximately 1 mL of dichloromethane was added, and
(15) Apicius, C. De re coquinaria, VIII, 8, 1; VI, 9, 1. L′art culinaire, translated
from Latin by Andre´, J. les Belles Lettres: Paris, 1974.
(16) Camps, P.; Perez, F.; Soldevilla, N. Tetrahedron: Asymmetry 1 9 9 7 , 8, 1877-
1894.
(17) Lie Ken Jie, M. S. F.; Mustafa, J.; Khysar Pasha, M. Chem. Phys. Lipids 1999,
100, 165-170.
Gas Chromatography. A Hewlett-Packard (Palo Alto, CA)
6890 gas chromatograph fitted with an on-column injector used
Analytical Chemistry, Vol. 74, No. 19, October 1, 2002 4869

Table 1. Abbreviated EI-MS Mass Spectra of a TMS Characteristic Compound of Each Chemical Familya for the
Most Volatile Components of Beeswax
compound names
characteristic ions
tetracosanic acid, TMS ester
triacontanol, TMS ether
triacontanyl palmitate
117 (30), 129 (15), 201 (25), 243 (13), 257 (13), 355 (14), 440 (63), 4 4 3 (100)
55 (26), 57 (26), 73 (40), 75 (64), 83 (23), 97 (35), 111 (27), 421 (39), 495 (4), 5 1 3 (100)
55 (49), 57 (39), 67 (37), 69 (31), 81 (30), 83 (31), 97 (38), 111 (25), 2 5 6 (100), 257 (89),
420 (5), 465 (8), 676 (13)
tetracosanyl 15-hydroxypalmitate,
TMS ether
55 (23), 73 (68), 117 (58), 133 (21), 3 0 1 (100), 329 (93), 331 (59), 453 (7), 591 (2),
636 (14), 665 (12), 680 (0.5)
a
Fatty acid, alcohol, monoester, and hydroxymonoester.
in the track-oven mode and a FID detector (350 °C) was used for
analysis by HTGC. The gas chromatographic separation was
achieved using a CP-Sil 5 CB column (Varian, Palo Alto, CA, 15
m length, 0.32 mm internal diameter, 0.1-µm phase thickness)
preceded by a 1-m precolumn. Helium was used as carrier gas
with a programmed flow as follows: 2 mL/ min for 17 min; 1 mL/
min2 until 4 mL/ min; 4 mL/ min for 4 min; 4 to 6 mL/ min at a 1
mL/ min2 rate; 6 mL/ min for 16 min. The oven temperature was
held at 50 °C for 1 min, ramped from 50 to 350 °C at 10 °C/ min
and held at 350 °C for 10 min.
Gas Chromatography/ Mass Spectrometry. The GC/ MS
analysis was performed with a ThermoFinnigan (San Jose, CA)
GCQ device equipped with an ion trap analyzer linked to a Hewlett-
Packard (Palo Alto, CA) 5890 chromatograph. Injection was
achieved through a split/ splitless injector used in the splitless
mode and held at 325 °C. Modern beeswax and archaeological
samples were analyzed with the same column and temperature
as for HTGC analysis. Helium pressure was maintained at 16.0
psi and the GC/ MS interface, at 340 °C. Mass spectra were
recorded in the electron ionization mode at 70 eV, and the ion
source was held at 180 °C. In full scan mode, the mass range was
scanned from 50 to 800 in 0.6 s.
Electrospray-Mass Spectrometry. ESI-MS analyses were
performed on a Micromass (Manchester, U.K.) Quattro II triple-
stage quadrupole mass spectrometer fitted with an electrospray
ion source and controlled by Mass Linx software. Synthetic
monoesters and fractionated and raw modern beeswax (0.2 mg/
mL) were prepared in a 0.1% TFA methanolic solution. Samples
were infused (5 µL/ min) into the ESI source with a Harvard
(Holliston, MA) syringe pump. The electrospray interface was
operated with a capillary voltage fixed at 3.3 kV at a temperature
of 70 °C. The cone voltage was optimized by recording mass
spectra at voltages varying between 40 and 65 V. Nitrogen was
used as the nebullization and drying gas at flow rates of 15-20
and 300-400 L/ hr, respectively. For ESI-MS/ MS and MRM
experiments, the protonated molecular cations were selected in
the first quadrupole (Q1), collided with Ar (2.3 mTorr) in the RF-
only second quadrupole (Q2) using a collision energy of 5-40
eV and analyzed in the third quadrupole (Q3). A solution of
polypropylene glycol oligomers (mass range 30-2000 Da) was
used for mass calibration.
literature,2,5,13,18-24 and Table 1 indicates the main fragments of
fatty acids, n-alcohols, monoesters, and hydroxymonoesters. One
must note that the analyses were performed with an ion trap
analyzer (see Experimental Section). Consequently, the mass
spectra detailed here may differ from those obtained on a
quadrupole or a magnetic sector, either as a result of a discrimina-
tion against ions with m/ z values below 100, as discussed
elsewhere,18 or because of secondary ion-molecule reactions
occurring with residual water present in the analyzer.19
Free Fatty Acids.20 The first separation on an aminopropyl SPE
cartridge (Varian, Palo Alto, CA) allows a fast separation of neutral
and acidic components. This procedure led to the isolation of a
homologous series of saturated fatty acids (C16:0, C18:0, and
C22:0-C34:0) with even carbon numbers of which tetracosanoic acid
(C24:0) is the main constituent, which is in agreement with other
investigations.14,21
The neutral components were then fractionated by flash
chromatograohy on a silica column into three different fractions:
(i) hydrocarbons, (ii) monoesters, and (iii) coeluted n-alcohols
and hydroxymonoesters.
Hydrocarbons. Alkanes and alkenes provide a typical pattern
for the recognition of beeswax, extensively described by Stra´nsky´
and Streibl,21 consisting of n-alkanes (C23-C33) and n-alkenes
(C27-C35) identified by their retention times by HTGC analysis
when compared to those of commercial standards. The alkane/
alkene ratio tends to decrease when the carbon length increases.
Isolation of hydrocarbons from polar compounds, such as fatty
acids and alcohols, avoids coelution phenomena and allows
identification of monunsaturated alkenes with 27, 29, 31, 33, and
35 carbon atoms, present at a very low level.
Monoesters.2,4,22 The major monoesters of beeswax contain an
even number of carbon atoms and are made of even-numbered
n-alcohols (C24 to C34) esterified by palmitic acid.
n-Alcohols. A homologous series of n-alcohols with even-
numbered chain length (C24-C34) was isolated by flash chroma-
tography. These compounds are degradation markers of beeswax
resulting from the partial hydrolysis of the palmitic esters. They
(18) Regert, M.; Rolando C. Anal. Chem. 2 0 0 2 , 74, 965-975.
(19) Munson, B. Int. J. Mass Spectrom. 2 0 0 0 , 200, 243-251.
(20) Brondz, I. Anal. Chim. Acta, 2 0 0 1 , 465, 1-37.
(21) Stra´nsky´, K.; Streibl, M. Collect. Czech. Chem. Commun. 1 9 7 1 , 36, 2267-
2297.
(22) Evershed, R. P. In: Modern Analytical Methods in Art and Archaeology; Wiley-
Interscience: New York, 2000.
(23) McLafferty, F. W.; Turecek F. Interpretation of Mass Spectra, 4th ed.,
University Science Books: Mill Valley, 1993.
(24) Reiter B.; Lechner M.; Lorbeer E. J. High Resolut. Chromatogr. 1 9 9 9 , 22,
514-520.
RESULTS AND DISCUSSION
Characterization of the Molecular Constituents of Modern
Beeswax by GC/ MS. The mass spectra of most volatile com-
pounds of beeswax have previously been described in the
4870 Analytical Chemistry, Vol. 74, No. 19, October 1, 2002

are present in the modern material at low-parts-per-million levels
but become major components in archaeological beeswax exposed
to hydrolytic degradations.2,5,6
Hydroxymonoesters. HTGC/ MS analysis allows the detection
of a homologous series of monohydroxyesters resulting from the
esterification of a long-chain alcohol (C24-C34) by 15-hydroxyp-
almitic acid, with triacontanyl 15-hydroxypalmitate dominating the
chromatogram. These compounds provide a small molecular peak
M+• (Table 1). At proximity of the molecular ion, a weak fragment
ion [M - 15]+ resulting from the loss of a methyl radical from
the silyloxy group and an ion [M - 44]+• resulting from the loss
of CO2 are observed. The spectrum is dominated by ions at m/ z
301 (base peak) and an intense fragment at m/ z 329. This latter
ion is obtained by the same pseudo-McLafferty rearrangement
observed for the saturated ester on ion [M - 15]+ obtained from
the molecular ion after the loss of a methyl group. The former
ion at m/ z 301 may result from a CO loss from the ion-dipole
complex between the acylium ion and the alkoxy radical, as found
in methyl lactate metastable decomposition.25
Figure 1. Profile of the monoesters of crude modern beeswax by
ESI-MS (0.2 mg/mL, TFA 0.1% in MeOH, cone 55 V).
Identification of the High-Molecular-Weight Constituents
of Beeswax by ESI-MS. The most volatile biomarkers of
beeswax, especially the n-alkanes, the fatty acids, and the long-
chain alcohols, are easily detected, separated, and identified by
GC and GC/ MS. The long-chain monoesters and hydroxy-
monoesters may also be studied by this method, but their elution
necessitates the use of limit conditions in GC separation, based
on the use of on-column injectors,26 short-length columns, high
temperatures (350 °C at the end of the run), and important flow
rates of helium. Their separation by GC/ MS is usually more
difficult as a result of the limit temperature of the interface; the
flow rate of helium, which must be maintained at a moderate flow
in order to keep an acceptable vacuum in the mass spectrometer;
and the presence of the mass spectrometer following the column,
which alters the separation, especially for the high-molecular-
weight constituents because of the pumping system. Higher-
molecular-weight components, in particular diesters, are not even
detected by this method, although they are known to be present
in beeswax. Analysis of diesters by HTGC and HTGC/ MS requires
appropriate instrumentation in order to achieve sufficient sensitiv-
ity at high elution temperatures and to avoid chromatographic
loss.14
Figure 2. Profile of the diesters of crude modern beeswax by ESI-
MS (0.2 mg/mL, TFA 0.1% in MeOH, cone 55 V).
range 500-900 corresponds to molecular ions of the monoesters
(Figure 1), and the high-mass range above m/ z 850 is character-
istic of the diester species (Figure 2).
In the zone of the mass spectrum corresponding to the
monoester constituents, various homologous series consisting of
molecules differing by 28 amu (two carbon units) can be
identified: (i) the monounsaturated esters from m/ z 591 to 703,
which dominate the spectrum, indicative of their easy ionization;
(ii) the saturated monoesters from m/ z 593 to 705; (iii) the
monosaturated monohydroxylated esters from m/ z 609 to 693;
and (iv) the saturated monohydroxylated esters from m/ z 611 to
695 (Figure 1).
Among the diester species (Figure 2), this spectrum allows
distinguishing the monounsaturated (from m/ z 847 to 987), the
saturated (from m/ z 845 to 985), and the monohydroxylated
derivatives (from m/ z 863 to 1003) from each other.
Comparing the HTGC/ MS and the ESI-MS profiles of modern
beeswax reveals the complementarity of the ESI-MS analysis,
especially for polar trace constituents. Saturated esters are not
easily protonated, contrary to polarizable monounsaturated and
hydroxylated esters, allowing a better detection of these polar
constituents.
Detailed Structure of the Fatty Acid Moieties Constitutive of the
Mono- and Diester Species by ESI-MS/ MS. The structures of the
different mono- and diesters determined by ESI-MS/ MS are
detailed in Table 2. Compared to mass spectrometry analyses of
beeswax with electron impact ionization, the use of ESI-MS/ MS
appears to be a powerful structural tool to determine the formulas
To elucidate more properly the structure of the higher-
molecular-weight constituents of beeswax, it appeared necessary
to analyze them directly from the liquid phase.27
General ESI-MS Profile of Raw Contemporary Beeswax. The ESI-
MS profile was obtained by analyzing beeswax diluted in an
acidified methanolic solution (TFA 0.1%). The mass spectrum
presents three main mass zones on which major ions were
detected. The low-mass range of the spectrum, below 500, is
characterized by a complex mixture of components that were not
studied in detail. We noted the important fragments formed by
decomposition of the protonated esters in the ESI source, such
as the ions at m/ z 255 and 257 corresponding to the palmitoleic
and palmitic protonated acids, respectively. The intermediate mass
(25) Tajima, S.; Nagai, Y.; Sekiguchi, O.; Fujishige, M.; Uchida, N. J. Am. Soc.
Mass Spectrom. 1 9 9 5 , 6, 202-206.
(26) Ogden, J.; Pardo, C.; Tchapla, A. J. Chromatogr. Sci. 1 9 9 8 , 36, 287-292.
(27) Hartvigsen, K.; Ravandi, A.; Bukhave, K.; Hølmer, G.; Kuksis A. J. Mass
Spectrom. 2 0 0 1 , 36, 1116-1124
Analytical Chemistry, Vol. 74, No. 19, October 1, 2002 4871

Table 2. Formulas of Mono-, Di-, and Hydroxyesters
Identified by HTGC/MS or ESI-MS/MS Analysis of Crude
Modern Beeswax
Figure 3. Tandem mass spectrum of the ion at m/z 609 corre-
sponding to a mixture of two protonated hydroxymonoesters (cone
45 V, CID 15 eV).
rated by 2 amu. The first one corresponds to the monounsaturated
monoesters, and the second, to the saturated palmitates. Upon
CAD, MH+ ions of monoesters species provide an abundant
protonated fatty acid ion that reflects the identities of the fatty
acids and n-alcohols. Typically, the CAD mass spectrum of
protonated tetracosanyl palmitoleate synthezised in the laboratory
(MH+ at m/ z 591) reflects this neutral alkene loss observed at
m/ z 255 corresponding to [C16H30O2 + H]+. The formation of the
acylium cation [C15H29CO]+ at m/ z 237 is also observed along
with the more surprising appearance of the radical cation corre-
sponding to the acid side chain at m/ z 382. All of the protonated
saturated and monounsaturated monoesters from beeswax show
strong ions at m/ z 255 and m/ z 257, reflecting the neutral loss of
an alkene from the alcohol. Three different series of monoesters
corresponding to esters of palmitic acid, palmitoleic acid, and oleic
acid with long-chain alcohol (CnH2n+1OH, with n between 24 and
32) can be identified.
Hydroxymonoesters are characterized by the typical loss of a
neutral H2O fragment from MH+ at low collision energies (5-10
eV). The loss of a neutral hydroxy fatty acid, of an alkene from
the alcohol, or of an alcohol at higher energies (15-25 eV) from
the mixture of protonated hydroxypalmitoyltetracosane bearing
the alcohol function on the ultimate methylene group of the acid
or of the alcohol moiety (m/ z 609) results in the [M - (RCO2H)
+ H]+, RCO2H2+, and RCO+ ions at m/ z 339, 257 and 237,
respectively (Figure 3).
Figure 2 shows the detailed ESI-MS profile of protonated
diesters consisting of two homologous series of six biomarkers,
each one separated from the adjacent ones by 16 amu, which
corresponds to a hydroxyl function. ESI-MS/ MS allows elucidation
of the molecular structure of these entities resulting from the
esterification of diols or hydroxyacids by fatty acids, hydroxyacids,
or diacids. As for monoesters, the loss of neutral fatty acid reflects
the nature of the diols or hydroxyacid substituents. All of the
diesters present the same fragmentations as illustrated in Figure
4: the CAD tandem spectrum of protonated diester (16:0/ 22-diol/
16:0) contains ions at m/ z 591, 255, and 257, reflecting a neutral
of long-chain compounds with very similar structure. GC/ MS
analyses allow the detection and the identification of saturated
palmitate esters, whereas ESI-MS/ MS analyses reveal that the
monoester composition of beeswax is more complex than ex-
pected and is also made of long-chain esters containing stearic
and oleic acid moieties. Similarly, the structure of the diesters,
which are difficult to detect by GC/ MS because of their high
molecular weight and their low volatility, even when using the
SIM mode, could be determined by their study in the liquid phase.
One must note that such a method, based on complementary
results obtained by the analyses in the gas and liquid phases,
avoids any saponification preceding the analysis, which was used
by several authors and is very prejudicial to the knowledge of
the complete structure of these dimers.
When infused into the ion source in an acidified methanolic
solution, monoester species yield two homologous series sepa-
4872 Analytical Chemistry, Vol. 74, No. 19, October 1, 2002

Figure 5. Tandem mass spectrum of protonated diester hexa-
cosanediol-1,15-dipalmitoyl (cone 60 V, CID 15 eV).
Scheme 1. Relationship between the Different
Homologous Ester Series Present in Beeswax.
Figure 4. Tandem mass spectrum of the ion at m/z 847 reflecting
a mixture of three protonated diesters (cone 45 V, CID 20 eV).
loss of palmitic acid (16:0) or of tetracosanyl palmitate, respec-
tively, from MH+ at m/ z 847. The presence of the weak ions at
m/ z 563 and 283 issued from the loss of a neutral stearic acid
(18:0) and a docosanyl stearate proves the existence of two
isomers with the same molecular weight, resulting from the
esterification of a diol by two molecules of palmitic acid for the
first one, and a molecule of palmitic acid and one of stearic acid
for the other one (Figure 4).
Hydroxylated diesters show as their hydroxymonoesters
homologous an intense loss of neutral water at low CID energy
and a loss of a neutral palmitic acid or a neutral 15-hydroxypalmitic
acid from the MH+ ion (m/ z 875), providing the [MH - (H2O)]+,
[MH - (C16:0)]+, and [MH - (HO-C16:0)]+ fragments at m/ z 857,
619, and 601 respectively (Figure 5).
In conclusion, ESI-MS/ MS analysis permits straightforward
characterization of the structure of high-molecular-weight biom-
arkers, such as diester, avoiding any sample preparation, especially
hydrolysis. The results obtained are summarized in Scheme 1.
From a biosynthetic point of view, the different esters, hydroxy
esters, and diesters may be considered as all derived from the
palmitate esters, as presented in Scheme 1. The hydroxylation of
the palmitate esters is conducive to the formation of hydroxy
esters (H1 and H2), which are then esterified by palmitic acid
(diesters D1 and D2). A second step of hydroxylation is responsible
for the presence of diesters D3.
SIM and MRM: Two Complementary Methods for the
Semiquantification of Biomarkers by GC/ MS and ESI-MS,
Respectively. Each molecular family presents characteristic
fragmentation of their biomarkers as M+• in HTGC/ MS analysis
or as MH+ in ESI-MS/ MS sequence, as described above. The
specific fragmentation that reflects various functional groups can
be exploited to enhance the selectivity and the sensitivity of the
biomarker detection. In the case of HTGC/ MS analysis on the
ion-trap analyzer, the selected ion monitoring mode (SIM) was
chosen, and the multiple reaction monitoring sequence (MRM)
was exploited for the ESI-MS/ MS analysis with a triple quadrupole
analyzer. The scanning parameters for the SIM experiment in
HTGC/ MS analysis are presented in Table 3. Specific fragmenta-
tions, such as loss of a TMSOH group and the addition of an H2O
molecule for the alcohols, formation of the two intense ions at
Analytical Chemistry, Vol. 74, No. 19, October 1, 2002 4873

Table 3. Scanning Parameters for the SIM Sequence in
HTGC/MS Analysis
Table 5. General Composition of the Trimethylsilylated
Total Lipid Extract (TLE) of Modern Beeswax from
Apis Mellifera Species, of an Etruscan Cup, and of
Scoria from the Castellina Settlement, Determined by
HTGCa
compound
SIM
n-alcohols
401, 429, 457, 485, 513, 541
saturated monoesters
monohydroxyesters
free fatty acids
256, 257
301, 329
117
modern
A. mellifera, %
Etruscan
cup, %
archaeological
scoria, %
component
hydrocarbons
n-alkenes
n-alkanes
fatty acids
unsaturated
saturated
24.6
6.8
17.8
3.6
0.4
3.2
36.8
10.3
26.5
11.1
n.d.
11.1
trb
n.d.
n.d.
n.d.a
5.8
0.2
5.6
Table 4. Scanning Parameters for the MRM Sequence
in ESI-MS/MS Analysis of Beeswax Esters
selected energy collision monitored ion
n-alcohols
4.3
1.7
compound
ion (Q1)
(Q2), eV
(Q3)
monoesters
hydroxymonoesters
diesters
â-sitosterol
cholesterol
total identified
components
39.2
9.5
5.6
n.d.
n.d.
86.8
13.1
10.2
tr
1.5
4.6
67.0
12.5
tr
0.7
0.1
saturated monoesters
unsaturated monoesters
monohydroxyesters
diesters (D1, D2)
MH+
MH+
MH+
MH+
MH+
20
20
5
20
5
257
255
[MH - H2O]+
+
[MH - C16:0]
hydroxydiesters (D3)
[MH - H2O]+
77.3
87.8
a
Not detected. b Trace.
m/ z 256 and at m/ z 257 for the monoesters or at m/ z 301 and
m/ z 329 for the monohydroxyesters in the HTGC/ MS procedure,
or loss of a water molecule or of an alcohol from the mono- or
diesters in the ESI-MS/ MS analysis are exploited with the
intention of improving the selectivity of the detection (Table 4).
Comparison between results obtained on the monoesters and
hydroxymonoesters by full scan and SIM scanning modes reflects
an improvement of the ratio signal/ noise by a factor of 10-50.
The high selectivity of these two scanning procedures allows direct
analysis of the crude lipid extract of modern or ancient material
without further purification.
The amount of each molecular component was calculated by
reference with the internal standard added to the solution before
its analysis by HTGC/ MS (see Experimental Section). This
procedure allowed quantification of the different main homologous
series (hydrocabons, fatty acids, n-alcohols, monoesters, hydroxy-
monoesters, and diesters) and analysis of the proportions of
saturated and unsaturated components for the most volatile series,
such as hydrocarbons and fatty acids (Table 5).
Moreover, in a chemical family, the ability to ionize is roughly
almost constant, which permits one to calculate a relative semi-
quantification of each detected biomarker in each family as well
as in GC/ MS (SIM) and ESI-MS/ MS (MRM). Consequently, we
proceeded to a relative quantification in each homologous series:
all molecular components are expressed as a percentage of the
main one in each chemical family (Figure 6).
A comparison of the SIM and MRM showed that volatile
components are more easily analyzed and quantified by SIM,
whereas nonvolatile components with high molecular weight are
rightly studied by MRM (Figure 6). Comparing the two analytical
procedures for the saturated monoesters shows that quantification
by these modes is similar but, for high dilution, MRM is more
precise as a result of the low volatility of these compounds in the
gas phase.
the analysis of archaeological beeswax preserved in ancient
ceramic vessels. Analysis of these ancient samples necessitates a
very sensitive method in order to minimize the required sample,
which has to be as low as possible (1-2 g for the domestic
ceramics, 200-400 mg for the funerary or luxurious ceramics).
Identification of Low- and High-Molecular-Weight Bio-
markers of Beeswax in Archaeological Samples. The archaeo-
logical samples studied in this investigation were collected during
summer 1999 by one of us (N.G.) on the Etruscan settlement
named La Castellina (late 8th-7th century B.C.) located on the
Mediterranean cost of Italy, 70 km northwest of Rome. A series
of 30 ceramic potsherds were chosen according to their size and
morphology.
All of the sherds lacked visible organic residue on their surface,
as has been described for other ceramic vessels.2 However, these
sherds were chemically investigated after having been ground into
powder and extracted in order to analyze the organic matter
potentially trapped in the porous clay matrix of the pottery.
Other samples, which were not associated with any pottery
but were found as small lumps in the sediment, were also
analyzed. These scoria, which look as burned dark stones, light
and very porous, were discovered in the stratigraphic unit
corresponding to the 8th-6th century B.C. of a supposed palatial
construction of the village.28
Among the 30 potsherds, two-thirds present a total lipid extract
(TLE) > 20 µg/ g, limit value for further investigation of the
organic matter by GC/ MS,29 generally found to vary between 40
and 100 µg/ g in the ceramics from La Castellina. The chromato-
graphic patterns obtained revealed the use of a wide range of
natural materials which are still under investigation. In this article,
we develop in detail the case of two samples, the organic content
of a domestic cup and the scoria that were found to contain traces
of beeswax. The pottery that yielded ancient beeswax was a
Although integration of a noisy signal in HTGC is impossible
for trace biomarkers, the combination of HTGC/ MS with SIM
scanning mode and of ESI-MS/ MS with MRM operation permits
a consequent improvement of the selectivity and the sensitivity
of these analyses. This new analytical protocol allows detection
of traces of beeswax in a sample and is especially well-adapted to
(28) Gran-Aymerich, J.; Prayon, F. Me´langes Ecoles Fr. Rome Athe`nes (Antiquite´)
2 0 0 0 , 112, 487-496.
(29) Charters, S.,; Evershed, R. P.; Goad, L. J.; Leyden, A.; Blinkhorn, P. W.;
Denham, V. J. Archaeol. Sci. 1 9 9 3 , 35, 211-223.
4874 Analytical Chemistry, Vol. 74, No. 19, October 1, 2002

Figure 6. Semiquantifications by EI-MS (SIM mode) and ESI-MS/MS using the MRM sequence of the total lipid extracts of (a) a modern crude
beeswax and (b) a bucchero Etruscan cup. Labels refer to carbon number length.
domestic cup made in bucchero, an old type of ceramic, typically
from Etruria, that imitates the metal Greek vessels (Figure 7).
HTGC Analyses of Archaeological Samples. The total lipid extract
from the domestic cup was of 72 µg/ g of sherd. HTGC allowed
detection of the major biomarkers (hydrocarbons and palmitic
esters) and degradation markers (fatty acids and traces of
n-alcohols) of beeswax in both samples (Figure 7).
full scan mode allows detection of major fatty acids (palmitic and
tetracosanoic acids), degradation markers issued from the partial
hydrolysis of beeswax.2 The presence of â-sitosterol, a biomarker
from plant origin, suggests that beeswax was mixed with a plant
material.30
Contrary to the total lipid extract of the cup, the scoria are
hydrocarbon-free, and we note on the HTGC-profile that mono-
The low amount of palmitic monoesters in the bucchero cup,
its degraded pattern, and the n-alkanes profile do not permit a
direct conclusion that this cup contained beeswax. HTGC/ MS in
(30) Logan, G.; Collins, M.; Eglinton, G. In Taphonomy: releasing the data locked
in the fossil record; Allison, P. A., Briggs, D. E. G., Eds.; Plenum Press: New
York, 1991; pp 1-24.
Analytical Chemistry, Vol. 74, No. 19, October 1, 2002 4875

Figure 7. HTGC profile of the lipid extract of (a) an Etruscan bucchero cup and (b) scoria from the Castellina settlement (late 8th-7th century
B.C.). 16:0 and 24:0 are palmitic and tetracosanoic acids, C25-C31 are n-alkanes, and the saturated monoesters and hydroxymonoesters (primed
numbers) are labeled by their carbon number length.
esters and monohydroxyesters are well-preserved. The presence
of palmitic (16:0) and stearic (18:0) acids is certainly due to an
external addition, and the association of these two saturated fatty
acids with cholesterol suggests that beeswax was mixed with fats
of animal origin.31 The preservation of esters and the hydrocarbon
depletion suggest that the material was not subjected to any
chemical alteration during preservation in the sediment but was
heated at high-temperature inducing the fast alkane vaporization.2
The association of chemical evidence and archaeological observa-
tions (porous, heated aspect of the material) suggests that
beeswax was mixed with animal fat and could have been used as
a fuel.
SIM and MRM Analyses of the Archaeological Beeswax. If HTGC
permits detection and quantification of only the major biomarkers,
investigation of the SIM and MRM procedures allows consider-
ation of the preservation of more than 60 beeswax biomarkers,
and semiquantification of these biomarkers was made by these
scanning operations (Table 5 and Figure 6). This method was
essentially investigated on the bucchero cup because (i) its TLE
is very low and the characterization of the preserved organic
matter requires more sensitive and selective methods, and (ii)
this ceramic sample is more representative of the studied material
exhumed during excavations, and analysis of the organic matter
preserved in the inner porous sides can bring to light the function
and the use of the pottery.
To characterize in more detail the degraded beeswax, the SIM
procedure was applied to the trimethylsilylated TLE and unques-
tionably revealed homologous series of even-numbered n-alcohols
(C20 to C34), of even-numbered saturated monoesters (C40-C50 with
the C34-C38 palmitates), and of even-numbered hydroxy-
monoesters (C40-C50 mentioned for the first time in archaeological
samples). The alcohols, monoesters and hydroxymonoester
profiles are similar to those of modern material. However an
increase in the proportion of alcohols, markers of degradation by
hydrolysis of esters, and relative reduction in the monoesters and
diesters prove that beeswax was not subjected to any specific
alteration, such as heating. The preservation of more than 50
biomarkers in the archaeological samples, which were susceptible
to hydrolysis by microorganisms or bacteria present in the
sediment, reflects the good preservation condition of the beeswax
in the bucchero clay. Fatty acids present a more complex profile
consisting of an important pattern of constituents with an even
number of carbon atoms (24:0 to 34:0) corresponding to beeswax
biomarkers; palmitic acid (16:0) as beeswax hydrolytic degradation
marker; myristic (14:0), oleic (18:1), and stearic (18:0) acids
coming from an external source of plant oil; and odd-numbered
fatty acids (13:0, 15:0 and 17:0) as bacterial biomarkers.32 By this
new scanning procedure, biomarkers present at low-parts-per-
million levels (1150 pg/ g for C16:0, <250 pg/ g for major
(31) Eglinton, G.; Logan, G. A Philos. Trans. R. Soc. London, Ser. B 1 9 9 1 , 333,
315-328.
(32) Dudd, S. N.; Regert, M.; Evershed, R. P. Org. Geochem. 1 9 9 8 , 29, 1345-
1354.
4876 Analytical Chemistry, Vol. 74, No. 19, October 1, 2002

monoesters, and <70 pg/ g for the major hydroxymonoesters)
reflect the powerful improvement in the sensitivity.
importance for understanding the exploitation of beeswax through
time. Enhancement of the detection limits allows henceforth the
study of the chemical structure of the organic matter preserved
in all potsherds, even if no charred residue is visible on its surface.
This more specific detection of beeswax biomarkers and degrada-
tion markers permits direct analysis of the crude or the trimeth-
ylsilylated lipid extract by ESI-MS/ MS (MRM) and HTGC/ MS
(SIM) respectively, and avoids a long time- and solvent-consuming
protocol, as described previously for the biomarkers’ isolation from
modern beeswax.
From the perspective of archaeology, the unambiguous iden-
tification of beeswax in ceramic vessels from the Etruscan village
of La Castellina is proof of the uninterrupted use of this natural
resource through time. Indeed, this material has been used in
Bercy (France) since the Neolithic period, during the Egyptian
period as a cosmetics or painting lacquer in the Fayum portraits
or in Crete (Mochlos) as fuel, but use of beeswax in archaic Italy
was not reported. The present study brings to light new elements
that suggest the use of beeswax as a waterproofing agent for the
porous Etruscan bucchero ceramics, or as fuel when mixed with
other fats. The absence of oil lamps in meridional Etruria during
the 8th-7th centuries B.C. could be explained by the use of
beeswax candles for lighting.
Investigating the MRM sequence to determine the molecular
structure of the mono- and diesters of this archaeological beeswax,
we prove the preservation of two homologous series of monoesters
(saturated and unsaturated derivating from palmitic acid and from
a mixture of palmitoleic and oleic acids respectively). As for the
diesters, apolar biomarkers, the ionization of which requires very
high cone tension (>70 eV) or very acidic strength (>0.2% in TFA)
do not present a sufficient response coefficient allowing a
representative semiquantification, but are detected as traces,
though hydroxylated diesters (type D3 in Table 2 and Scheme 1)
on which the alcoholic function confers a greater ionizability are
more easily detected and can be semiquantified. Figure 6 sum-
marizes the different results obtained by SIM and MRM analyses
of modern and archaeological beeswax preserved in the bucchero
cup.
CONCLUSION
Although gas chromatographic analysis, using flame ionization
detection or coupled to mass spectrometry, is the most common
means for the study of the molecular composition of beeswax,
the use of ESI-MS technique has significant potential for extending
our knowledge of the structure of the high-molecular-weight
constituents of beeswax. Working in the liquid instead of the gas
phase allowed determination of the structure of several biomarkers
present in beeswax, such as monounsaturated and saturated esters
containing either palmitic, palmitoleic, stearic, or oleic moieties
as well as hydroxyesters and diesters.
However, such analyses do not allow the detection and
identification of low-molecular-weight constituents, such as fatty
acids, alcohols, and n-alkanes. Indeed, beeswax appears to be a
very complex lipidic material containing an important number of
molecular constituents on a wide range of molecular weights. In
consequence, further studies of this material should be based on
both the use of GC/ MS and ESI-MS analyses.
ACKNOWLEDGMENT
The authors thank Dr. Jean Gran-Aymerich for providing the
archaeological samples and for all the fruitful discussions on
Etruscan pottery. We also express our gratitude to Bernadette
Coddeville and Adeline Page for their assistance in ESI-MS
analyses. We gratefully aknowledge the support of the GDR CNRS
2114 Physico-Chimie des Mate´riaux du Patrimoine Culturel. The
mass spectrometers from the Centre Commun de Spectrome´trie
de Masse (USTL) were acquired thanks to grants from the Conseil
Re´gional Nord, Pas-de-Calais and the European Community
(FEDER Program).
Received for review March 20, 2002. Accepted July 16,
2002.
If such research opens new avenues for the study of contem-
porary beeswax, especially in the case of adulterations which can
alter the quality of beehive products, they are also of considerable
AC025637A
Analytical Chemistry, Vol. 74, No. 19, October 1, 2002 4877