Food Chemistry 82 (2003) 575582

www.elsevier.com/locate/foodchem

Ultrasound-assisted extraction of volatile compounds from citrus

owers and citrus honey
E. Alissandrakisa, D. Dafererab, P.A. Tarantilisb, M. Polissioub, P.C. Harizanisa,*
a
Laboratory of Sericulture and Apiculture, Agricultural University of Athens, 75 Iera Odos, 118 55 Athens, Greece
b
Laboratory of Chemistry, Agricultural University of Athens, 75 Iera Odos, 118 55 Athens, Greece

Received 5 December 2002; received in revised form 16 December 2002; accepted 16 December 2002

Abstract
The volatile fraction of honey is believed to facilitate satisfactory discrimination between honeys of dierent botanical origin. A
new methodology for extracting volatile compounds was developed, using n-pentane:diethylether organic solvent and a water bath
with ultrasound assistance. Analysis of the extracts of four Citrus species owers showed linalool to be the predominant compound
(11.3% in lemon, 51.6% in orange, 80.6% in sour orange and 75.2% in tangerine). The extracts from citrus honey were pre-
dominated by an array of linalool derivatives (more than 80% of the total extract). (E)-2,6-dimethyl-2,7-octadiene-1,6-diol was the
predominant compound (44.7%), while signicant proportions of 2,6-dimethyl-3,7-octadiene-2,6-diol (15.4%) and (Z)-2,6-dime-
thyl-2,7-octadiene-1,6-diol (7.2%) were also present.
# 2003 Elsevier Ltd. All rights reserved.
Keywords: Ultrasound extraction; Volatile compounds; Citrus honey; Linalool; Linalool derivatives

1. Introduction the geographical and botanical origin of honey were

reviewed by Anklam (1998). Although the major com-
Honey is a nutritious food, with economical impor- ponents of all honeys are sugars and water, there is a
tance for many countries worldwide. That is why great variety as far as the aroma and avour are con-
unambiguous ways of determining its botanical and cerned. Each honey possesses its own aroma and avour,
geographical origin have to be found. Traditionally, this depending mainly on the botanical sources that contribute
determination is achieved by pollen analysis, a tech- to its production. It is claimed that the chemical analysis
nique known as melissopalinology. Even though it gives of any particluar, unioral honey can give its nger-
satisfactory results, it cannot stand as a reliable method print, depending on the oral source (Tan, Wilkins,
on its own, mainly because it is tedious and very Holland, & McGhie, 1989). It is quite possible that the
dependent on the ability and judgement of the expert volatile fraction is potentially useful in the future as a
(Persano Oddo, Piazza, Sabatini, & Accorti, 1995). means for identifying the botanical and geographical
Moreover, there is great variability in the nectar con- origin of honey samples. It is therefore of great impor-
tribution of any particular ower compared with the tance to improve the extraction techniques used so far,
amount of its pollen found in honey (Tan, Wilkins, as well as to develop new ones in order to enable the
Molan, Holland, & Reid, 1989). More recently, pollen analysis of honey volatiles to become a routine proce-
analysis has been combined with the analysis of the dure. Bonaga and Giumannini (1986) suggested that
physicochemical as well as organoleptic properties of the next step in this type of research will be an attempt
honey. to correlate oral source with the presence of certain
In the early 1960s it was proposed that the origin of compounds originating either in the nectar or in some
honey could be determined by its chemical composition. biochemical modication carried out by the bee.
The analytical methods used for the determination of To date, the extraction of honey volatiles by means of
simultaneous distillationextraction (SDE) methodology,
* Corresponding author. Fax: +30-1-3466692. developed by Nickerson and Likens, (1966) or its modi-
E-mail address: [email protected] (P.C. Harizanis). cations (Bicchi, Belliardo, & Frattini, 1983; Bouseta &
0308-8146/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0308-8146(03)00013-X
576 E. Alissandrakis et al. / Food Chemistry 82 (2003) 575582

Collin, 1995) have been the most popular methods used. sequently ltered through MgSO4 monohydrate in
Since the samples, in these methods, are heated, the order to hold back the water and solid matter. The
generation of artifactssuch as furan derivatives (Bic- extract was nally concentrated with a gentle stream of
chi et al., 1983) or Maillard reaction products (Mills, nitrogen to 0.1 ml, placed in a vial and sealed. It was
1978)is inevitable. Some new techniques have been kept in the freezer until the GCMS analysis.
recently developed, that do not use heat at all, such as
the headspace system (Bouseta, Collin, & Dufour, 1992; 2.3. Honey volatiles extraction
Radovic, Careri, Magnia, Musci, Gerboles, & Anklam,
2001) and simultaneous distillationextraction under Forty grams of honey were diluted with 22 ml of dis-
static vacuum (Maignial, Pibarot, Bonneti, Chaintreau, tilled water in a small beaker. In order to obtain high
& Marion, 1992). recoveries, we added 1.5 g of magnesium sulphate
Ultrasound-assisted extraction is used for the isolation of (MgSO4) hydrate, which was diluted by means of a
the volatile compounds from natural products at room magnetic stirrer. The solution (50 ml) was put into a
temperature with organic solvents. Moreover, some work 200-ml spherical ask and was extracted by means of
has been done for wine aroma compounds (Cocito, ultrasound for 10 min with 15 ml of n-pentane:diethyl-
Gaetano & Delni, 1995; Vila, Mira, Lucena, & Reca- ether (1:2). The top of the ask was covered and the
males, 1999), on which our methodology was largely based. whole procedure was repeated twice, using two asks
The rst aim of this work was to develop a metho- each time (a total of four extractions). After the end of
dology for isolating volatile compounds from honey each sonication, both samples were introduced in a
and the corresponding owers, in order to analyse them separation funnel and 20 ml of a saturated solution of
directly by gas chromatographymass spectrometry NaCl were added. The funnel was well-shaken and then
(GCMS). The second aim was to correlate the volatile left to rest at room temperature. When the two layers
compounds of citrus honey with those isolated from were well separated, the overlying emulsion was col-
citrus owers. lected. The ask was washed with another 15 ml of the
extraction solvent. The whole extract was centrifuged at
3000 rpm and the organic layer, containing the honey
2. Materials and methods volatiles, was collected and nally ltered to remove any
solid residues. The sample was concentrated with a gentle
2.1. Honey sample, plant material and solvents stream of nitrogen to 0.5 ml, placed in a vial and sealed.
It was kept in the freezer until the GCMS analysis.
The honey sample was collected from the area of
Argos, Greece. To be certain of the botanical origin of 2.4. GCMS
the honey, two bee colonies, containing unbuilt combs,
were placed in the middle of a large area of Citrus spe- The analysis of the extracts was performed using a
cies (orange trees were predominant). An eort was Hewlett-Packard 5890 II GC with a ame ionisation
made for no other honey to be present in any comb detector, equipped with a Hewlett-Packard 5972 MS
because bees tend to move honey from one comb to detector. In both cases, the column used was an HP-
another within the hive. The colonies remained in the 5MS (Crosslinked 5% PH ME Siloxane) capillary col-
orange groves only during the honey ow. The honey umn (30 m0.25 mm i.d., 0.25 mm lm thickness) and
was harvested by pressing the combs. No mechanical the gas carrier was Helium, at 1 ml/min rate. The injec-
treatment or heat was used. The owers of the four tor and detector temperatures were maintained at 220
Citrus species (orange, lemon, tangerine and sour and 290  C, respectively. Samples of 1 ml were injected
orange) used in this work were collected from the orch- manually and splitless. Electron impact mass spectra were
ard of the Agricultural University of Athens. recorded in the 40500 mass range. An electron ionisation
The solvents used were diethyl ether (Carlo Erba, pro system was used with ionisation energy of 70 eV.
analysi) and n-pentane (Merck, extra pure). For honey extracts, oven temperature was held at
40  C for 3 min, raised to 180  C at 2  C/min and then
2.2. Flower extraction to 250  C at 10  C/min (5 min hold). For ower
extracts, the oven temperature was held at 40  C for 3
In order to obtain the volatile compounds from the min, raised to 180  C at 4  C/min and then to 250  C at
owers of the four Citrus species, 5 g of fresh owers 10  C/min (5 min hold).
were placed in a 200-ml spherical ask, along with 30 ml
of n-pentane:diethylether (1:2). The ask was covered 2.5. Repeatability of the extraction procedure
and then placed in an ultrasound (US) water bath
apparatus for 10 min. The temperature of the US water The repeatability of the extraction procedure was tested
bath was carefully held at 25  C. The extract was sub- by comparing the results of ve extractions of the honey
 E. Alissandrakis et al. / Food Chemistry 82 (2003) 575582 577

sample, using linalool as the internal standard. In Many techniques have been developed for the isola-
Table 1, the relative standard deviations (RSD) for each tion of volatile compounds, most of them requiring
of the 14 compounds checked are presented. special equipment, while others are not easy to carry
out. This is not the case as far as this technique is con-
cerned. Finally, the duration of the whole procedure is
3. Results and discussion less than 2 h, making it quite rapid.
Even though the procedure can be considered as
3.1. Sample collection and honey extraction satisfactory, the data presented in Table 1 show that it
needs improvement concerning its repeatability, because
In order for the honey sample to be as pure as possi- some of the RSD values are above 20%.
ble, two bee colonies, containing unbuilt combs, were
placed in the middle of a large area of Citrus species, to 3.2. Flower organic extracts
ensure that the honey was unioral. In all published
work on honey volatiles, the botanical origin was sus- The amount of the extracted compounds is expressed as
tained using pollen analysis of the samples. Even though a percentage of the obtained peak area, compared with the
all workers have claimed that the analysis of the volatile total area of all the peaks of the chromatograph.
fraction of honey is a better way to prove its origin, yet The owers of four Citrus species were extracted by
they have used pollen analysis to check if the samples were means of ultrasound. The GC analysis proved linalool
unioral. To our knowledge, the safest way to ensure that (3,7-dimethyl-1,6-octadiene-3-ol) to be the predominant
a honey sample is unioral is to place bee colonies con- compound (see Table 2) in all, except for lemon (51.6%
taining unbuilt combs in the middle of a large area of in orange, 80.6% in sour orange and 75.2% in tanger-
the plants from which the honey is collected. ine). The ower extract of lemon exhibited ve major
One of the major drawbacks of the extraction techni- compounds: b-pinene (6,6-dimethyl-2-methylene-bicy-
ques used so far (such as the SDE or the purge and trap clo[3.1.1]heptane, 11.8%), limonene [1-methyl-4-(1-
system) is the formation of thermally created artifacts. methylethenyl)-cyclohexene, 16.1%], eucalyptol (1,3,3-
Furan derivatives are well known artifacts found in trimethyl-2-oxabicyclo[2.2.2]octane, 35.4%), linalool
thermally treated samples and so is hotrienol, as proved (11.3%) and a-terpineol (4-trimethyl-3-cyclohexene-1-
in our work. The technique developed in this work uses methanol, 9.1%), with eucalyptol being the pre-
an ultrasound water bath as a means of extracting dominant compound.
honey volatile and semi-volatile compounds. This tech- Signicant proportions of sabinene {4-methylene-1-
nique does not require heat, thus no thermally gener- (1-methylethyl)-bicyclo [3.1.0] hexane} were found in
ated artifacts are formed. orange (25.4%) and tangerine (10.8%) ower extracts.
Another advantage of this technique is that it enables Two compounds that exhibit similar structure to that
the extraction of compounds of molecular weight up to of linalool were found in orange and sour orange, that is
220 that could contribute to the determination of the ocimene [(E)-3,7-dimethyl-1,3,6-octatriene, < 1%] and
origin of honey. This is the most important advantage linalool acetate (10.6%), respectively. Finally, a well
of the US-assisted extraction, compared to the head- known linalool derivative, (Z)-2,6-dimethyl-2,7-octa-
space system. diene-1,6-diol, was found in low proportions in sour
orange ower extract.
Table 1
Relative standard deviations (RSD) of 14 compounds obtained from 3.3. Honey organic extracts
ve extractions of the honey sample

Peak no. Compound (prominent MS peaks) RSD (%) Fig. 1 shows the GC prole of the Citrus honey
extracts, while in Table 3 the compounds found are lis-
5 Xylene 14.0
11 Phenylacetaldehyde 21.2
ted.
15 Hotrienol 22.7 The analysis of the honey extracts showed that more
16 Phenylethyl alcohol 12.8 than 80% of the total amount consisted of compounds
17 Lilac aldehyde 19.4 known as linalool derivatives (see Table 4 for mass
18 Lilac aldehyde 23.6 spectra data). This nding was not unexpected, as the
19 Lilac aldehyde 20.4
20 2,6-Dimethyl-3,7-octadiene-2,6-diol 7.71
analysis of the owers extracts showed linalool to be
23 1,3-Bis(1,1-dimethyl) benzene 14.8 the predominant compound. So, we can say that the
25 Lilac alcohol 13.3 precursors of the honey aroma compounds are expected
26 Lilac alcohol 16.9 to be found in the ower extract of the corresponding
27 Unknown (43, 55, 59, 60, 69, 97, 118) 7.22 botanical origin.
29 (Z)-2,6-dimethyl-2,7-octadiene-1,6-diol 9.62
30 (E)-2,6-dimethyl-2,7-octadiene-1,6-diol 7.28
Compounds belonging to this group have been
claimed to characterise New Zealands nodding thistle
578 E. Alissandrakis et al. / Food Chemistry 82 (2003) 575582

Table 2
Components of the essential oil from four Citrus sp. owers

Componenta tR (min)b % Of the total area

Orange Lemon Tangerine Sour orange

Sabinene {4-methylene-1-(1-methylethyl)-bicyclo [3.1.0] hexane} 12.98 26.7 1.3 10.8 <1

b-Pinene (6,6-dimethyl-2-methylene-bicyclo[3.1.1]heptane) 13.09 11.8
Limonene (1-methyl-4-(1-methylethenyl)-cyclohexene) 15.28 3.9 16.1 <1 1.6
Eucalyptol (1,3,3-trimethyl-2-oxabicyclo[2.2.2]octane) 15.39 35.7
Benzeneacetaldehyde 15.91 <1
Ocimene [(E)-3,7-dimethyl-1,3,6-octatriene] 16.13 2
4-Methyl-1-(1-methylethyl)-3-cyclohexen-1-ol 16.90 4.8 <1
Linalool (2,6-dimethyl-2,7-octadiene-6-ol) 18.28 51.6 11.3 75.2 80.6
Phenylethyl alcohol 18.76 1.8
a-Terpineol (4-trimethyl-3-cyclohexene-1-methanol) 21.82 3.8 9.1 2.7
Linalool acetate 24.24 10.6
Indole 25.66 2.9 2.2 2.1 2.7
(Z)-2,6-dimethyl-2,7-octadiene-1,6-diol 28.05 1.5
8-Heptadecane 38.00 1.1
Farnesol (3,7,11-trimethyl-2,6,10-dodecatrien-1-ol) 39.24 1.2 2.1 <1 <1
Caeine 41.90 5
Bis(2-ethylhexyl) adipate 49.76 1.5
a
The identication was based on the NBS75K mass spectra library.
b
Mean retention time.

Fig. 1. GC proles of citrus honey extractives. GC conditions: HP-5MS column, He as carrier gas (1 ml/min), 40  C (3-min hold) raised at 2  C/min
to 180  C and from there raised at 10  C/min to 250  C (5-min hold).

honey (Carduus nutans; Wilkins, Lu, & Tan, 1993). dimethyl-2,7-octadiene-1,6-diol (peak30, 44.7%) and it
Nevertheless, in that type of honey, the predominant was also present in New Zealands nodding thistle
compound was (E)-2,6-dimethyl-6-hydroxy-2,7-octadie- honey as a major constituent.
noic acid, which was not found in our honey. However, Hotrienol (3,7-dimethyl-1,5,7-octatrien-3-ol, peak15,
the extraction techniques used were dierent. The com- 4.7%) and 2,6-dimethyl-3,7-octadiene-1,6-diol (peak20,
pound that predominated citrus honey was (E)-2,6- 15.4%) have been reported as constituents of several
 E. Alissandrakis et al. / Food Chemistry 82 (2003) 575582 579

Table 3
Components of Citrus honey

Peak no. Componenta tR (min) % Of the total areab Prominent MS peaksc

1 Heptane 3.60 1.3 43, 57, 71, 100

2 Methyl-cyclohexane 4.09 1.1 41, 42, 55, 69, 70, 83, 98
3 Toluene 5.34 1.2 50, 51, 65, 91, 92
4 Octane 6.51 <1 43, 57, 71, 85, 114
5 m- (Or p-) xylene 9.72 3.2 50, 62, 65, 77, 91, 105, 106
6 o-Xylene 10.97 <1 51, 52, 63, 65, 77, 91, 106
7 Cyclohexanone 11.11 <1 42, 55, 69, 70, 98
8 Nonane 11.53 1.1 43, 57, 71, 85, 99, 128
9 Decane 18.20 1 43, 57, 71, 85, 99, 142
10 Limonene 20.04 1.3 53, 67, 68, 79, 93, 107, 136
11 Benzeneacetaldehyde 21.22 1.7 51, 65, 91, 92, 120
12 1-Methyl-2-pyrrolidinone 21.23 2.5 56, 71, 98, 99
13 n-Octanol 23.55 <1 41, 44, 55, 56, 69, 84
14 Undecane 25.53 <1 43, 57, 71, 85, 99, 156
15 3,7-Dimethyl-1,5,7-octatrien-3-ol (hotrienol) 25.89 4.7 41, 43, 55, 67, 71, 79, 119
16 Phenylethyl alcohol 26.40 1.3 51, 65, 91, 92, 103, 122
17 Lilac aldehyded 28.60 1.5 41, 43, 55, 69, 81, 93, 111, 153
18 Lilac aldehyded 29.22 3 41, 43, 55, 67, 69, 71, 81, 93, 111, 153
19 Lilac aldehyded 30.63 1.3 41, 43, 55, 67, 71, 83, 93, 111, 153
20 2,6-Dimethyl-3,7-octadiene-2,6-diol 32.56 15.4 41, 43, 55, 67, 71, 82, 83, 137
21 Dodecane 32.89 <1 43, 57, 71, 85, 170
22 2,3-Dihydro benzofuran 34.46 <1 51, 65, 91, 119, 120
23 1,3-Bis(1,1-dimethyl) benzene 36.66 <1 41, 57, 65, 91, 111, 175, 190
24 Benzeneacetic acid 37.32 2.3 51, 65, 91, 92, 136
25 Lilac alcohold 39.63 1.4 43, 55, 68, 71, 75, 81, 93, 111
26 Lilac alcohold 40.59 1.1 43, 55, 68, 71, 75, 81, 93, 111
27 Unknown 41.60 3.2 43, 55, 59, 60, 69, 97, 118
28 (E)-2,6-dimethyl-6-hydroxy-2,7-octadienald 42.97 <1 41, 43, 55, 67, 71, 87, 98, 135
29 (Z)-2,6-dimethyl-2,7-octadiene-1,6-diold 43.07 7.2 41, 43, 55, 67, 71, 79, 82, 93, 119, 137
30 (E)-2,6-dimethyl-2,7-octadiene-1,6-diold 44.80 44.7 41, 43, 55, 67, 71, 79, 82, 93, 119, 137, 152
31 1-Isocyanato-2-methyl benzene 51.38 <1 51, 78, 104, 133
32 Degraded carotenoid 62.36 <1 43, 77, 79, 108, 150
33 Dibutyl phthalate 76.86 <1 57, 73, 149, 150, 205, 223
34 Tricosane 82.55 <1 43, 57, 71, 99, 113, 127, 141
35 Bis-(2-ethylhexyl) adipate 84.42 <1 57, 70, 83, 112, 129, 147, 241
a
The identication was based on the NBS75K mass spectra library or/and on published data.
b
Only for components over 1%.
c
Italics indicate the basic peak, underline indicates the molecular ion.
d
The identication was based on the NBS75K mass spectra library and on published MS or MS and NMR data of Wilkins et al. (1993).

Table 4
Mass spectra of linalool derivatives

Compound (peak No) Prominent MS peaks

Linalool 71 (100), 43 (70), 41 (69), 93 (64), 55 (57), 80 (27), 67 (20), 121 (17), 136 (6)
3,7-Dimethyl-1,5,7-octatrien-3-ol (hotrienol) (15) 71 (100), 43 (74), 82 (59), 67 (31), 41 (24), 55 (18), 79 (7), 119 (2)
Lilac aldehydes (17, 18, 19) 55 (100), 43 (94), 41 (60), 111 (52), 71 (49), 93 (49), 81 (32), 69 (32), 153 (26)
55 (100), 43 (82), 41 (49), 93 (44), 71 (44), 111 (35), 67 (32), 69 (28), 81 (23), 153 (20)
55 (100), 43 (74), 71 (47), 41 (41), 93 (36), 67 (30), 111 (29), 153 (24), 83 (24)
2,6-Dimethyl-3,7-octadiene-1,6-diol (20) 82 (100), 43 (88), 71 (84), 67 (52), 41 (22), 55 918), 83 (9), 137 (1)
Lilac alcohols (25, 26) 43 (100), 111 (78), 55 (73), 75 (57), 93 (54), 81 (36), 68 (35), 71 (16)
43 (100), 111 (85), 55 (72), 75 (65), 93 (63), 81 (37), 68 (36), 71 (19)
(E)-2,6-dimethyl-6-hydroxy-2,7-octadienal (28) 71 (100), 43 (78), 55 (39), 41 (36), 87 (25), 67 (15), 98 (12), 135 (4)
(Z)-2,6-dimethyl-2,7-octadiene-1,6-diol (29) 43 (100), 71 (73), 67 (62), 55 (43), 41 (41), 82 (21), 79 (16), 119 (16), 93 (12), 137 (9)
(E)-2,6-dimethyl-2,7-octadiene-1,6-diol (30) 43 (100), 71 (70), 67 (52), 55 (38), 41 (35), 79 (23), 93 (19), 82 (18), 119 (10), 137 (8), 152 (1)
580 E. Alissandrakis et al. / Food Chemistry 82 (2003) 575582

honeys. They have also been found in the essential oil of toch, Morales, Duque, & Schreier, 1993). It has been
a large number of plants. These two compounds were shown (Rowland et al., 1995) that this dehydration can
found to predominate in the extract of leatherwood take place when the sample passes through the hot
honey (Eucryphia lucida; Rowland, Blackman, DArcy, injection port in GC analysis. In order to further sustain
& Rintoul, 1995). They were also found in Eucalyptus the nding that hotrienol is heat generated, the sample
melliodora honey (DArcy, Rintoul, Rowland, & Black- was extracted by means of steam distillationextraction,
man, 1997). using a LikensNickerson apparatus. The sample was
The thermal dehydration of 2,6-dimethyl-3,7-octa- deliberately heated up to 70  C during the procedure.
diene-1,6-diol leads to the formation of hotrienol (Win- The predominance of hotrienol (peak 15) in Fig. 2 states
that this compound is indeed heat-generated.
The avour of hotrienol has been described as sweet
and owery (Nakatani, Sato, & Yamanishi, 1969), while
2,6-dimethyl-3,7-octadiene-1,6-diol has been reported to
be odourless (Wintoch et al., 1993).
Another linalool derivative found was (Z)-2,6-dime-
thyl-2,7-octdiene-1,6-diol (peak 29, 7,2%). This com-
pound was also found in traces in the orange ower
extract. As reported by Wilkins et al. (1993), it was a
major compound of New Zealands nodding thistle
honey aroma. It was also found in the unmethylated
extracts of leatherwood honey (Rowland et al., 1995).
Peaks 17, 18 and 19 were identied as lilac aldehydes
(5.8% in total), while peaks 25 and 26 were lilac alcohols
(2.5% in total). The comparison of their retention time
and their mass spectra with data published by Wilkins
et al. (1993) allowed their identication. According to
the Wilkins et al. (1993) work, both lilac alcohols and
lilac aldehydes are produced from linalool acetate, fol-
lowing the path shown in Fig. 3. Lilac alcohols and lilac
aldehydes have been reported as nodding thistle honey
constituents (Wilkins et al., 1993), while lilac aldehydes
were also found in haze honey (Rhus succedanea) (Shi-
moda, Wu, & Osajima, 1996).
Fig. 2. GC prole of the sample that was heated during the extraction, Peak 28 was identied as 2,6-dimethyl-6-hydroxy-2,7-
showing the predominance of hotrienol. octadienal. It was found in low proportions.

Fig. 3. Synthesis of lilac alcohols and lilac aldehydes from linalool acetate.
 E. Alissandrakis et al. / Food Chemistry 82 (2003) 575582 581

Fig. 4. GC proles of the sample analysed a few days (a) and 1 month (b) after the extraction.

Peak 35 was bis(2-ethylhexyl) adipate. As shown in The analysis of the extracts of the owers of four
Fig. 4, this compound was found in traces when the Citrus species showed that the precursors of the honey
honey extract was analysed a few days after the extraction aroma compounds are found in the owers of the plant
(Fig. 4a), while it predominated in the extracts when ana- of the corresponding botanical origin. Thus, we can say
lysed 1 month later (Fig. 4b). Thus, it is possible that it is that the analysis of the owers can give us information
an artifact produced in storage. Adipic acid is produced about what to expect in the corresponding honey.
by the oxidation of cyclohexane and cyclohexanone, Citrus honey is characterised by the predominance of
compounds that were detected in our GC analysis linalool derivatives in the honey extract. The analysis of
(peaks 2 and 7, respectively). Alternatively, adipic acid more citrus honey samples, as well as the quantication
is possibly a honeybee pheromone, as Tan, Holland, of these derivatives, could lead to the establishment of a
Wilkins, and Molan (1988) mention that diacids are threshold to distinguish citrus honeys from others of
related to the pheromonal system of the honeybee. dierent oral origin. Moreover, signicant proportions
Esterication of adipic acid results in the increased of linalool derivatives in other unioral honeys (such as
amount of bis(2-ethylhexyl) adipate in the second case. r) would indicate honey adulteration.
Consequently, sonicated extracts must be analysed
immediately after the extraction.
Acknowledgements

4. Conclusion The authors would like to thank the Greek Ministry

of Agriculture and the European Union for nancially
The use of water bath ultrasound as a means of supporting this research according to the Council of
extracting honey volatile and semi-volatile compounds Regulation (EC) 1221/97.
seems to be a promising technique. It does not require
heat, thus no artifacts (such as furan derivatives or References
hotrienol) are generated. Furthermore, the whole proce-
dure that was developed is quite rapid, easy to be carried Anklam, E. (1998). A review of the analytical methods to determine
out and does not necessitate special equipment. Moreover, the geographical and botanical origin of honey. Food Chemistry, 63,
this technique allows the extraction of compounds of 549562.
molecular weight up to 220 that could possibly contribute Bicchi, C., Beliardo, F., & Fratini, C. (1983). Identication of the
volatile components of some Piedmontese honeys. Journal of Api-
to the determination of the origin of honey. Optimisa- cultural Research, 22, 130136.
tion of the procedure is required for quantitative and Bonaga, G., & Giumanini, A. G. (1986). The volatile fraction of
qualitative evaluation of honey volatiles. chestnut honey. Journal of Apicultural Research, 25, 113120.
582 E. Alissandrakis et al. / Food Chemistry 82 (2003) 575582

Bouseta, A., & Collin, S. (1995). Optimized Likens-Nickerson metho- analysis of aroma compounds to authenticity testing of honey. Food
dology for quantifying honey avors. Journal of Agricultural and Chemistry, 72, 511520.
Food Chemistry, 43, 18901897. Rowland, C. Y., Blackman, A. J., DArcy, B., & Rintoul, G. B. (1995).
Bouseta, A., Collin, S., & Dufour, J.-P. (1992). Characteristic aroma Comparison of organic extracts found in leatherwood (Eucryphia
proles of unioral honeys obtained with a dynamic headspace GC lucida) honey and leatherwood owers and leaves. Journal of Agri-
MS system. Journal of Apicultural Research, 31, 96109. cultural and Food Chemistry, 43, 753763.
Cocito, C., Gaetano, G., & Delni, C. (1995). Rapid extraction of Shimoda, M., Wu, Y., & Osajima, Y. (1996). Aroma compounds from
aroma compounds in must and wine by means of ultrasound. Food aqueous solution of haze (Rhus succedanea) honey determined by
Chemistry, 52, 311320. adsorptive column chromatography. Journal of Agricultural and
DArcy, B., Rintoul, G. B., Rowland, C. Y., & Blackman, A. J. Food Chemistry, 44, 39133918.
(1997). omposition of Australian honey extracts. 1. Nor- Tan, S. T., Holland, P. T., Wilkins, A. L., & Molan, P. C. (1988).
isoprenoids, monoterpens, and other natural volatiles from blue Extractives from New Zealand honeys. 1. White clover, manuka
gum (Eucalyptus leucoxylon) and yellow box (Eucalyptus mellio- and kanuka unioral honeys. Journal of Agricultural and Food
dora) honeys. Journal of Agricultural and Food Chemistry, 45, Chemistry, 36, 453460.
18341843. Tan, S. T., Wilkins, A. L., Holland, P. T., & McGhie, T. K. (1989).
Maignial, L., Pibarot, P., Bonetti, G., Chaintreau, A., & Marion, J. P. Extractives from New Zealand unioral honeys. 2. Degraded car-
(1992). Simultaneous distillation-extraction under static vacuum: otenoids and other substances from heather honey. Journal of Agri-
isolation of volatile compounds at room temperature. Journal of cultural and Food Chemistry, 37, 12171221.
Chromatography, 606, 8794. Tan, S. T., Wilkins, A. L., Molan, P. C., Holland, P. T., & Reid, M.
Mills, F. D. (1978). Ring contraction products from 2,3-dihydro-3,5- (1989). A chemical approach to the determination of oral sources
dihydroxy-6-methyl-4(H)-pyran-one: 4-hydroxy-5-hydroxymethyl- of New Zealand honeys. Journal of Apicultural Research, 28, 212
2-methyl-3(2H)-furanone and 2,5-dimethyl-4-hydroxy-3(2H)-fur- 222.
anone. Journal of Agricultural and Food Chemistry, 26, 894898. Wilkins, A. L., Lu, Y., & Tan, S. T. (1993). Extractives from New
Nakatani, Y., Sato, S., & Yamanishi, T. (1969). 3S-(+)-3,7-Dimethyl- Zealand honeys. 4. Linalool derivatives and other substances from
1,5,7-octatrien-3-ol in the essential oil of Black tea. Agricultural and nodding thistle (Carduus nutans) honey. Journal of Agricultural and
Biological Chemistry, 33, 967968. Food Chemistry, 41, 873878.
Nickerson, G. B., & Likens, S. T. (1966). Gas chromatographic evi- Wintoch, H., Morales, A., Duque, C., & Schreier, P. (1993). (R)-(-)-
dence for the occurrence of hop oil components in beer. Journal of (E)-2,6-Dimethyl-3,7-octadiene-2,6-diol-6-O-b-D-glucopyranoside:
Chromatography, 21, 15. Natural precusor of hotrienol from Lulo Fruit (Solanum vestissi-
Persano Oddo, L., Piazza, M. G., Sabatini, A. G., & Accorti, M. mum) peelings. Journal of Agricultural and Food Chemistry, 41,
(1995). Characterization of unioral honeys. Apidologie, 26, 453 13111314.
465. Vila, D. H., Mira, F. J. H., Lucena, R. B., & Recamales, M. A. F.
Radovic, B. S., Careri, M., Mangia, A., Musci, M., Gerboles, M., & (1999). Optimization of an extraction method of aroma compounds
Anklam, E. (2001). Contribution of dymanic headspace GCMS in white wine using ultrasound. Talanta, 50, 413421.