Ultrasonics Sonochemistry 14 (2007) 750–756

www.elsevier.com/locate/ultsonch

Comparison of hydrodistillation and ultrasonic solvent extraction

for the isolation of volatile compounds from two unifloral honeys of
Robinia pseudoacacia L. and Castanea sativa L.
a,*
I. Jerković , J. Mastelić a, Z. Marijanović a, Ž. Klein a, M. Jelić b

a
Department of Organic Chemistry, Faculty of Chemistry and Technology, University of Split, N. Tesle 10/V, 21 000 Split, Croatia
b
Marko Marulić Polytechnics of Knin, Petra Krešimira IV 30, 22300 Knin, Croatia

Received 8 October 2006; received in revised form 13 December 2006; accepted 19 December 2006
Available online 23 January 2007

Abstract

A comparative study of ultrasound-assisted extraction (USE) with the mixture pentane:ether (1:2) and hydrodistillation (HD) with the
same trapping mixture is presented for the isolation of volatile compounds from two unifloral honeys of Robinia pseudoacacia L. and
Castanea sativa L. All HD isolates contained many thermally derived artefacts (especially phenylacetaldehyde with lower percentages
of furfural, cis- and trans-linalool oxides and others). USE method gave the most representative profile of all honey volatiles (without
artefacts). In addition, USE enabled extraction of low molecular weight semivolatile markers (especially benzoic, vanillic and phenyla-
cetic acids) that were not extracted by HD. In this regard, low percentage of benzoic acid (0.7–7.4%), vanillic acid (0.0–1.6%) and phen-
ylacetic acid (0.5–4.1%) was determined in Rp USE extracts, while Cs USE extracts contained phenylacetic acid (20.2–23.5%) as the
major constituent with low percentage of benzoic acid (2.5–5.5%).
 2007 Elsevier B.V. All rights reserved.

Keywords: Unifloral honey; Volatiles; Robinia pseudoacacia L. and Castanea sativa L.; Ultrasound extraction (USE); Hydrodistillation (HD); Gas
chromatography–mass spectrometry (GC–MS)

1. Introduction
Honey is nutritious food with great variability of flavour
and aroma. The unifloral origin of particular honey
remains difficult to determine [1]. Honey pollen analysis
has been combined with the analysis of its physicochemical
properties and, more recently, honey volatile fraction has
been used for identifying its botanical origin [2–4]. Vola-
tiles represent specific ‘‘fingerprint’’ of particular honey,
depending of its floral origin. However, the composition
of isolated honey volatiles greatly depends upon the isola-
tion method [5].
Hydrodistillation (HD) and simultaneous distillation-
extraction (SDE) by Likens–Nickerson [6] or its modifica-
*
Corresponding author. Tel.: +385 21 329 434; fax: +385 21 329 461.
E-mail address: igor@ktf-split.hr (I. Jerković).
tions [7,8] have been the most common methods for the
extraction of volatile compounds from different matrix.
Applying those methods for the isolation of honey volatiles
generates artefacts, mainly due to the effect of heat on
honey sugars and amino acids [3,5]. To avoid generation
of artefacts, some new methods have been developed
recently, that does not use high temperature such as solid
phase microextraction (SPME) that selectively isolates
headspace volatiles depending on the used fibre [4,5] or
simultaneous distillation-extraction under static vacuum
[9]. Ultrasound-assisted extraction (USE) was first intro-
duced by Alissandrakis et al. [10] for the isolation of vola-
tile and semi-volatile compounds from citrus honey as
rapid and promising method.
The volatiles of Robinia pseudoacacia L. (Rb) honey
were previously isolated by steam distillation and SDE
method [3], comprising thermal artifacts. The volatiles of

1350-4177/$ - see front matter  2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.ultsonch.2006.12.014
 I. Jerković et al. / Ultrasonics Sonochemistry 14 (2007) 750–756
Castanea sativa L. (Cs) honey were isolated by SPME [2],
by dynamic headspace GC–MS [4] and by SDE method
[3]. However, those honeys were not yet investigated apply-
ing USE for volatiles extraction and this is the first report
about these volatiles of Croatian origin.
It is known that the use of ultrasound-assisted extrac-
tion (USE) aids extraction by significantly reducing extrac-
tion times [11] in comparison with traditional methods
(exp. shake-flask extraction). The mechanical effect of
ultrasound provides a greater penetration of solvent into
matrix, via cavitation effects, and improves extraction.
The aim of this work was to compare hydrodistillation
and ultrasound-assisted extraction for the volatiles isola-
tion from two unifloral honeys. Two honeys were used to
investigate the impact of different sample matrix on the iso-
lated volatiles. Qualitative and quantitative composition of
the volatiles was determined by GC and GC–MS. Identi-
fied low and high molecular weight compounds provide
new data for markers of unifloral honey origin
determination.
2. Experimental
2.1. Honey samples and reagents
Ten unifloral honey samples (5 Rp and 5 Cs) were
selected from various honey producers in Croatia. Screen-
ing of honey unifloral origin was based on pollen analysis
and sensory test. All honey samples met Croatian require-
ments [12] for unifloral origin (for Rb the percentage of
pollen more than 20% and for Cs the percentage of pollen
more than 80%). All samples were stored at 4 C until
analysed.
The solvents used were diethyl ether and pentane pur-
chased from Kemika (Zagreb, Croatia). Diethyl ether was
distilled before usage. Anhydrous MgSO4 and menthol
were obtained from Fluka Chemie (Buchs, Switzerland).
2.2. Isolation procedures
2.2.1. Hydrodistillation (HD)
Hydrodistillation was performed in a Clevenger type
apparatus. Each honey sample (160 g) was added in 1 L
flask, dissolved in 200 mL distilled water and placed on
the apparatus. A mixture of pentane and diethyl ether
(2:1 v/v) was used for trapping the volatiles. Hydrodistilla-
tion was carried out for 2.5 h and the condenser was cooled
with water. For each batch of 10 honey samples, duplicate
hydrodistillation was performed.
2.2.2. Ultrasound-assisted extraction (USE)
Ultrasound-assisted extraction (USE) was performed in
an ultrasound cleaning bath (Transsonic Typ 310/H, Ger-
many) by the mode of indirect sonication, at the frequency
of 35 kHz at 25 ± 3 C. Each batch of ten honey samples
was extracted in duplicate as further described. Forty
grams of each honey sample were dissolved with 22 mL
751
of distilled water in a 100 mL flat-bottomed flask. Magne-
sium sulphate (1.5 g) was added and each sample was
extensively vortexed. A mixture (20 mL) of pentane:diethyl
ether (1:2) was used as the extraction solvent. Sonication
was held for 30 min. After sonication, the organic layer
was separated in a separation funnel and filtered over
anhydrous MgSO4. Aqueous layer was returned to the
flask and another batch of extraction solvent (20 mL)
was added and extracted by ultrasound for 30 min. Organic
layer was separated in the separation funnel, filtered over
anhydrous MgSO4 and aqueous was sonicated third time
for 30 min with another batch (20 mL) of the extraction
solvent. Joined organic extracts were concentrated up to
0.2 mL by distillation with Vigreaux column, and 1 lL
was used for GC–MS analysis.
2.3. Gas chromatography (GC)
Gas chromatography analysis was performed on a Hew-
lett-Packard Model 5890 Series II gas chromatograph
equipped with flame ionisation detector and capillary col-
umn HP-101 (Methyl silicone fluid, Hewlett-Packard,
Vienna, Austria), 25 m · 0.2 mm i.d., coating thickness
0.2 lm. Chromatographic conditions were as follows:
helium as carrier gas at 1.0 mL min1; injector and detector
temperatures were 250 C and 300 C. Oven temperature
was isothermal at 70 C for 2 min, then increased to
200 C, at a rate of 3 C min1 and held isothermal for
15 min. Volume injected was 1 lL and split ratio was 1:50.
2.4. Gas chromatography–mass spectrometry (GC–MS)
The samples were analyzed by gas chromatography–
mass spectrometry (Hewlett-Packard, model 5890, with a
mass selective detector, model 5971A) on two columns.
GC operating conditions [13,14] were: column HP-20M
(Carbowax 20M, Hewlett-Packard, Vienna, Austria),
50 m · 0.2 mm i.d., film thickness 0.2 lm; column temper-
ature programmed from 70 C isothermal for 4 min, then
increased to 180 C at a rate of 4 C min1; column HP-
101 (Methyl silicone fluid, Hewlett-Packard, Vienna, Aus-
tria), 25 m · 0.2 mm i.d., film thickness 0.2 lm; column
temperature programmed from 70 C isothermal for
2 min, then increased to 200 C at a rate of 3 C min1.
Carrier gas was helium at flow rate 1 mL min1. Injector
temperature was 250 C. Volume injected was 1 lL and
split ratio was 1:50. MS conditions were: ionization voltage
70 eV; ion source temperature 280 C; mass range: 30–300
mass units.
2.5. Quantization and identification
The individual peaks were identified by comparison of
their retention indices (relative to C8–C22 n-alkanes) to
those of authentic samples and the literature [15], as well
as by comparing their mass spectra with the Wiley 6.0
library (Wiley, New York) and NIST98 (National Institute
752 I. Jerković et al. / Ultrasonics Sonochemistry 14 (2007) 750–756
of Standards and Technology, Gaithersburg) mass spectral
database. The percentage composition of the samples was
computed from the GC peak areas using the normalization
method (without correction factors). For quantization,
internal standard (menthol) was added prior to the isola-
tion. Preliminary GC–MS analysis showed the absence of
menthol among the Rp and Cs honey volatiles. The compo-
nent percentages (Tables 1 and 2) and standard deviations
were calculated as the mean value of duplicate GC and
GC–MS analysis for each batch of 10 honey samples.
3. Results and discussion
Tables 1 and 2 demonstrate minimum, maximum and
mean percentages, as well as standard deviations of Rp
and Cs honey volatile compounds isolated by hydrodistilla-
tion (HD) and ultrasonic solvent extraction (USE). Two
columns with different stationary phase polarity were used
for GC–MS analyses (polar HP-20M column and non
polar HP-101 column) for better separation and identifica-
tion of isolated compounds and it is generally know that
GC gives more representative results for compounds quan-
tization (peak area). The compounds are listed according
to their elution order on HP-20M column. Relative
amounts of isolated volatiles were calculated with reference
to the internal standard (menthol) and for USE extracts
varied from 3.9 to 10.1 mg kg1 (Rb honey) and from
21.1 to 38.6 mg kg1 (Cs honey). HD isolates contained
lower amount of volatiles for both honeys. From the tables
it can be seen that there are large qualitative and quantita-
tive differences among the isolated volatiles, depending on
the extraction procedure. The hydrodistillation conditions
were more drastic, with many thermal artefacts being gen-
erated, especially due to the impact of heat on sugars and
amino acids. Among them, phenylacetaldehyde was the
most abundant. Moreover, sensitive compounds were eas-
ily oxidized or decomposed and new compounds had arisen
that did not belong to the original aroma of honey, such as
furfural and linalool oxides that were present only in HD
isolates. In addition, low volatile and water soluble com-
pounds cannot be quantitatively determined using HD.
The USE method gave the most representative profile of
all volatiles, as no heat was applied. In addition, the use
of USE with organic solvent enabled extraction of semivol-
atiles with high boiling points that were not isolated by
HD. In this regard, USE was particularly useful for the iso-
lation of benzoic, vanillic and phenylacetic acids (although
it should be considered that their recovery was low due to
use of non polar solvents). They are important markers of
unifloral honeys that were not extracted by HD, SDE [1] or
SPME [2] and their contents can be used for distinguishing
different unifloral honeys [16].
3.1. Volatiles obtained by hydrodistillation (HD)
Seventeen volatile compounds were identified by GC–
MS in Rp honey HD isolates (Table 1), representing
95.3–97.0% of total peak area. The obtained volatiles con-
tained phenylacetaldehyde as the major component (66.2–
68.4%). In fact its formation was so favoured, that its peak
comprised 68% of all chromatogram area. Strecker degra-
dation of amino acids, especially during heat treatment,
can produce aromatic aldehydes and phenylacetaldehyde
is Strecker aldehyde of phenylalanine. Other important
HD volatiles were furfural (2.1–4.6%), hotrienol (1.4–
1.8%) and higher saturated alkanes, such as C21 and C24.
Hotrienol was present in Rp USE extracts, but its forma-
tion is favoured by high temperature employed in HD,
Table 1. Linalool oxides (cis- and trans-) were also present
(1.5–3.0% and 0.8–1.7%). In comparison with the results of
Rp honey volatiles [3] isolated by Likens–Nickerson steam
distillation-extraction (SDE) there are major differences
regarding dominance of 2-phenylethanol, 2,3-pentanedi-
one, ethyl-phenyl acetate, 2-phenylacetaldehyde and
methyl salicylate in SDE isolates. These differences arose
from different extraction technique and SDE gave a clearer
volatile profile due to less drastic isolation conditions in
comparison with HD. However, SDE was performed at
high temperature, also comprising thermal artefacts, such
as furans, pyrazines, aldehydes, ketones and others.
Forty-five volatile compounds were identified by GC–
MS in Cs honey HD isolates (Table 2), representing
98.1–99.2% of total peak area. These volatiles contained
mainly phenylacetaldehyde (10.8–15.6%) and higher linear
saturated hydrocarbons (particularly C21, C22 and C24).
Hydrocarbons were already found in cooked food, where
they originate from phospholipids by Maillard reactions
[17]. The percentage of thermally derived phenylacetalde-
hyde was significantly lower then in Rp honey, probably
due to different matrix composition in comparison with
Cs honey. However, phenylacetaldehyde was not identified
in Cs USE extract that additionally confirmed its thermal
formation during HD as Strecker aldehyde of phenylala-
nine. Since 4-aminoacetophenone (5.6–7.4%) was also con-
tained in Cs USE extract (even with higher percentages) it
can not be considered as HD artefact. It was also identified
in Cs SDE extracts [3]. HD isolate also contained thermally
derived furfural, 2-isopropylfuran, 2-acetylfuran, 5-methyl-
furfural and 2,2 0 -bifuran. Identified cis- and trans-linalool
oxides are thermal artefacts since they were not present
in USE extracts and it is known that monoterpene polyols
can be transformed into cyclic monoterpenols by the action
of high temperature and acid conditions [18,19]. Unstable
monoterpene polyols were not detected in HD isolate, how-
ever 3,7-dimethyl-1,5-octadiene-3,7-diol was present in
USE extracts (0.9–1.2%). Its degradation during HD pro-
duced hotrienol. Other oxidation artefacts were found only
among Cs HD volatiles, such as aldehydes: 3-phenyl-2-pro-
penal (trans-cinnamaldehyde) probably originated from
oxidation of 3-phenyl-2-propen-1-ol (cinnamyl alcohol)
found in USE extract and benzaldehyde (1.0–4.1%) proba-
bly formed from benzyl alcohol. Hotrienol comprised 0.3–
0.6% of HD isolate and is known thermally generated
product [5,20]. Nevertheless, some quantity seems to exist
 I. Jerković et al. / Ultrasonics Sonochemistry 14 (2007) 750–756 753

Table 1
Volatiles from Pseudoacacia robinia L.
HD USE
Min. Max. Avg. d Min. Max. Avg. d
1 2-Methyldecane – – – – 1.6 3.1 2.23 0.78
2 Tridecane – – – – 0.6 2.9 1.50 1.23
3 Nonanal – – – – 0.6 0.7 0.60 0.10
4 Acetic acid – – – – 0.3 1.3 0.80 0.50
5 Decanal – – – – 0.4 0.5 0.43 0.06
6 cis-Linalool oxide 1.5 3.0 2.23 0.75 – – – –
7 2-Furancarboxaldehyde (Furfural) 2.1 4.6 3.27 1.26 – – – –
8 trans-Linalool oxide 0.8 1.7 1.20 0.46 – – – –
9 1H-Pyrrole 0.0 0.6 0.33 0.31 0.7 1.6 1.13 0.45
10 Benzaldehyde 0.0 0.8 0.40 0.40 – – – –
11 Pentadecane – – – – 0.2 0.6 0.40 0.20
12 5-Methyl-2-furancarboxaldehyde 0.0 0.4 0.23 0.21 0.0 0.3 0.17 0.15
13 Hotrienol 1.4 1.8 1.60 0.20 0.3 0.8 0.57 0.25
14 Phenylacetaldehyde 66.2 68.4 66.53 1.72 – – – –
15 Hexadecane 0.2 0.3 0.27 0.06 0.0 0.3 0.20 0.17
16 Heptadecane – – – – 0.5 2.4 1.37 0.96
17 Hexanoic acid – – – – 0.4 0.4 0.40 0.00
18 Octadecane – – – – 0.7 0.8 0.73 0.06
19 Benzyl alcohol – – – – 0.6 0.7 0.67 0.06
20 2-Phenylethanol 0.0 0.5 0.27 0.25 0.6 1.3 0.90 0.36
21 1-Nonadecene – – – – 0.0 0.5 0.23 0.25
22 Nonadecane 0.0 0.9 0.47 0.45 0.0 0.3 0.17 0.15
23 Phenol 0.0 0.4 0.23 0.21 0.2 1.0 0.63 0.41
24 2-Hydroxy-3,5,5-trimethyl-2-cyclohexen-1,4-dione – – – – 0.0 1.3 0.60 0.66
25 1-Octadecene – – – – 0.0 0.6 0.27 0.31
26 Octanoic acid – – – – 1.8 2.4 2.07 0.31
27 Heneicosane 1.8 2.7 2.20 0.46 0.0 1.8 0.80 0.92
28 Nonanoic acid – – – – 1.9 4.2 2.87 1.19
29 3,5-Dimethoxy-4-hydroxy-benzaldehyde – – – – 0.0 1.2 0.53 0.61
30 4-Vinyl-2-methoxyphenol 0.0 0.4 0.23 0.21 3.4 3.6 3.53 0.12
31 5-Hydroxy-2-decenoic acid lactone 0.0 1.6 0.70 0.82 0.0 1.1 0.53 0.55
32 2,6-Dimethoxyphenol – – – – 0.0 0.7 0.37 0.35
33 Docosane – – – – 1.1 3.2 2.27 1.07
34 Decanoic acid (Capric acid ) – – – – 2.8 3.7 3.17 0.47
35 2,3-Dihydrobenzofuran (Coumaran) – – – – 1.6 2.2 1.93 0.31
36 Benzoic acid – – – – 0.7 7.4 3.87 3.37
37 3,7-Dimethyloct-1-en-3,6,7-triol – – – – 2.2 4.1 3.30 0.98
38 5-Hydroxymethyl-2-furancarboxaldehyde – – – – 0.0 0.5 0.23 0.25
39 1,4-Benzenediol (Hydroquinol ) – – – – 0.0 0.6 0.27 0.31
40 4-Hydroxybenzyl alcohol – – – – 2.4 3.3 2.83 0.45
41 Tricosane 0.0 4.9 2.37 2.45 1.2 4.5 2.90 1.65
42.Dodecanoic acid (Lauric acid ) – – – – 0.7 1.1 0.90 0.20
43 Phenylacetic acid – – – – 0.5 4.1 2.53 1.85
44 Dibuthylphtalate – – – – 3.0 3.6 3.30 0.30
45 1-Hexadecanol (Cetal ) – – – – 4.7 11.4 8.03 3.35
46 Hexadecanoic acid (Palmitic acid ) – – – – 11.4 12.9 11.43 1.45
47 Octadecanoic acid (Stearic acid ) – – – – 0.0 0.7 0.37 0.35
48 (Z)-9-Octedecen-1-ol – – – – 0.0 6.5 3.50 3.28
49 (Z)-9-Octadecenoic acid (Oleic acid ) – – – – 0.0 2.1 1.03 1.05
50 Tetracosane 12.1 12.9 12.00 0.95 2.1 6.9 12.33 0.49
51 3-Methoxy-4-hydroxy-benzoic acid (Vanillic acid ) – – – – 0.0 1.6 0.80 0.80
Total yield (mg/kg) 2.0 2.8 2.40 0.40 5.6 10.1 8.27 2.36
Min., minimal percentage; r, standard deviation; I1, retention indices on HP-20M column; –, not detected on this column.
Max., maximal percentage; Avg., average percentage; I2, retention indices on HP-101 column.

in non-thermally treated ripe honey [20], and we also iden-
tified it in some Cs USE extracts. Benzopyrroles 3-methyl-
1H-indole (0.4–0.7%) and 1H-indol (0.4–1.0%) with
undesirable odour were only identified in HD isolate. They
were derived from the metabolism of tryptophan [17], since
it is known that pyrroles occur among volatiles of many
heated foods, where they can arise from Maillard reactions
or from the pyrolysis of amino acids [17]. The obtained
Cs honey HD isolate was different from previously
published SDE extract [2] containing 2-ethyl-1-hexanol,
754 I. Jerković et al. / Ultrasonics Sonochemistry 14 (2007) 750–756

Table 2
Volatiles from Castanea sativa L.
HD USE
Min. Max. Avg. d Min. Max. Avg. d
1 Undecane – – – – 0.0 1.6 0.67 0.83
2 Limonene – – – – 0.9 4.5 2.23 1.97
3 p-Menth-1-en-7-al (Phellandral) 0.0 1.5 0.50 0.87 – – – –
4 1,2,4-Trimethylene cyclohexane 0.0 1.2 0.40 0.69 – – – –
5 Tridecane – – – – 0.0 0.3 0.10 0.17
6 Acetic acid – – – – 1.4 5.0 3.17 1.80
7 1-Methyl-4-(1-methylethenyl)-benzene (o-Isopropenyltoluene) 0.0 1.6 0.53 0.92 – – – –
8 cis-Linalool oxide 0.3 0.6 0.40 0.17 – – – –
9 Furfural 1.8 3.7 2.73 0.95 0.2 0.4 0.30 0.10
10 trans-Linalool oxide 0.0 0.4 0.07 0.23 – – – –
11 1-(2-Furanyl)-ethanone(2-Acetylfuran) 0.8 1.2 0.97 0.21 – – – –
12 Benzaldehyde 0.7 4.1 1.93 1.88 – – – –
13 Propanoic acid – – – – 0.0 0.3 0.17 0.15
14 6-Methyl-hept-5-en-2-ol* – – – – 0.0 0.3 0.17 0.15
15 Pentadecane – – – – 0.9 2.0 1.3 0.61
16 3,9-epoxy-p-mentha-1,8-diene 0.0 0.5 0.17 0.29 – – – –
17 5-Methylfurfural 0.0 0.4 0.13 0.23 – – – –
18 2,3-Dihydrobenzofuran-2-one (2-Coumaranone) 0.0 0.9 0.43 0.45 – – – –
19 2,2 0 -Bifuran 0.0 0.6 0.30 0.30 – – – –
20 2,3-Butanediol – – – – 0.0 0.4 0.13 0.23
21 Terpinene-4-ol 0.0 0.6 0.30 0.30 – – – –
22 Hotrienol 0.3 0.6 0.43 0.15 0.0 0.1 0.03 0.06
23 Butanoic acid – – – – 0.4 0.6 0.33 0.30
24 2-Isopropylbenzaldehyde 0.0 1.5 0.70 0.75 – – – –
25 Phenylacetaldehyde 10.8 15.6 12.80 2.50 – – – –
26 1-Phenylethanone (Acetoin) 0.0 0.3 0.20 0.17 0.3 0.6 0.43 0.15
27 2-Furanmethanol – – – – 0.0 0.3 0.13 0.15
28 Hexadecane 0.0 1.3 0.63 0.65 – – – –
29 o-Mentha-1,5,8-triene 0.0 0.8 0.30 0.44 – – – –
30 Ketoisophorone 0.0 0.9 0.30 0.52 0.0 0.8 0.27 0.46
31 2-(4-Methylphenyl)-propanal 0.0 1.0 0.47 0.50 – – – –
32 p-Mentha-1,5,8-triene 0.0 0.3 0.10 0.17 – – – –
33 cis-Carveol* 0.0 0.9 0.37 0.47 – – – –
34 Heptadecane 0.0 0.5 0.27 0.25 1.0 2.7 1.87 0.85
35 Phenylvinylketone 0.7 1.0 0.83 0.15 – – – –
36 4-(1-Methylethyl)-benzaldehyde 0.0 0.5 0.20 0.26 – – – –
37 a-Methylbenzyl alcohol* 0.0 0.5 0.20 0.26 0.5 0.7 0.60 0.10
38 Hexanoic acid (Caproic acid) 0.0 0.6 0.30 0.30 0.6 1.2 0.87 0.31
39 Octadecane – – – – 0.0 0.3 0.13 0.15
40 Benzyl alcohol 0.0 0.4 0.20 0.20 1.4 2.1 1.83 0.38
41 2-Phenylethanol 0.0 0.4 0.20 0.20 1.1 1.7 1.30 0.35
42 1-H-indol 0.4 1.0 0.63 0.32 – – – –
43 3,7-Dimethyl-1,5-octadiene-3,7-diol – – – – 0.9 1.2 1.03 0.15
44.1-Nonadecene – – – – 0.0 3.0 1.27 1.55
45 Nonadecane 0.4 1.2 0.77 0.40 1.4 5.0 2.87 1.89
46 Phenol – – – – 0.5 1.3 1.13 0.57
47 3-Phenyl-2-propenal (trans-Cinnamaldehyde) 0.8 1.4 1.07 0.31 – – – –
48 Octanoic acid (Caprylic acid) 1.7 2.0 1.83 0.15 0.9 1.2 1.07 0.15
49 2-Pentadecanone 0.0 0.5 0.20 0.26 – – – –
50 Heneicosane 8.2 14.6 10.93 3.30 1.0 2.8 2.10 0.96
51 Nonanoic acid 0.0 5.4 2.47 2.73 2.0 2.1 2.03 0.06
52 1-(2-Aminophenyl)-ethanone (2-Aminoacetophenone) – – – – 2.1 2.5 2.27 0.21
53 4-Vinyl-2-methoxyphenol (4-Vinylguiacol) 1.0 1.7 1.27 0.38 2.6 5.1 3.83 1.25
54 1-(4-Aminophenyl)-ethanone (4-Aminoacetophenone) 5.6 7.4 6.33 0.95 3.8 10.4 7.83 3.54
55 5-Hydroxy-2-decenoic acid lactone* – – – – 0.0 0.6 0.30 0.30
56 2,3-Dihydro-3,5-dihydroxy-6-Methyl-4H-pyran-4-one – – – – 4.4 8.2 5.97 1.99
57 1,4-Benzenediol (Hydroquinol) – – – – 0.0 0.3 0.13 0.15
58 Docosane 1.3 7.7 4.03 3.30 1.5 1.7 1.60 0.10
59 3-Phenyl-2-propen-1-ol (Cinnamyl alcohol) – – – – 1.8 9.1 4.63 3.91
60 Decanoic acid (Capric acid) 0.9 1.7 1.30 0.40 – – – –
61 2-(p-Methoxyphenyl)-ethanol – – – – 0.0 0.6 0.30 0.30
 I. Jerković et al. / Ultrasonics Sonochemistry 14 (2007) 750–756 755

Table 2 (continued)
HD USE
Min. Max. Avg. d Min. Max. Avg. d
62 Geranicacid* 0.0 1.1 0.53 0.55 0.0 2.2 1.10 1.10
63 3-(2-Hydroxypheyl)-2-propenoic acid – – – – 0.0 4.0 1.73 2.05
64 2,3-Dihydrobenzofuran (Coumaran) – – – – 0.0 4.9 1.63 2.83
65 Benzoic acid – – – – 2.5 5.5 4.33 1.61
66 3-Methyl-1H-indole 0.4 0.7 0.53 0.15 – – – –
67 5-Hydrohymethyl-2-furancarboxaldehyde – – – – 0.0 13.3 6.10 6.72
68 Tetracosane 23.0 27.8 24.93 2.53 0.0 1.8 0.80 0.92
69 Phenylacetic acid – – – – 20.2 23.5 21.57 1.72
70 Isodihydrocarveol* – – – – 0.0 1.2 0.57 0.60
71 (Z)-9-octedecen-1-ol – – – – 0.0 2.5 1.0 1.32
72 Hexadecanoic acid (Palmitic acid) 0.2 11.0 4.90 5.53 – – – –
Yield (mg/kg) 8.4 12.1 10.17 1.86 21.1 38.6 29.90 8.75
Min., minimal percentage; I1, retention indices on HP-20M column.
Max., maximal percentage; I2, retention indices on HP-101 column.
r, standard deviation; –, not detected on this column.
Avg., average percentage; *, tentatively identified.

benzaldehyde, furfuryl propionate, acetophenone, 2-phenyl- oxyphenol and phenol. They are mainly produced by bio-
ethanol, 2-aminoacetophenone, aliphatic and aromatic chemical degradation of phenolic acids in honey.
methyl esters and phenylacetamide. Forty-five volatile compounds were identified in Cs
honey USE extracts (Table 2), representing 77.1–94.9% of
total peak area. This extract contained phenylacetic acid
3.2. Volatiles obtained by ultrasound-assisted extraction
(20.2–23.5%) as the major constituent, and its high percent-
(USE)
age is important as new Cs honey marker. Namely, phen-
ylacetic acid was not detected in previous papers on Cs
Rp honey extract isolated by USE enabled another
honey [2,3], due to applied SDE and SPME methods.
important view on its volatiles qualitative and quantitative
The percentage of benzoic acid was 2.5–5.5% and its low
composition. Forty-six volatile compounds were identified
percentage should also be noted for unifloral identification.
by GC–MS (Table 1), representing 88.7–91.3% of total peak
Specific volatile markers of this honey were 4-aminoace-
area. The percentage composition of compounds was more
tophenone (3.8–10.4%) and 2-aminoacetophenone (2.1–
uniformly, without striking dominance of one compound,
2.5%), found only in Cs honey, as was already reported
as in HD. The major components were hexadecanoic acid
[2,3]. Identified 3,7-dimethyl-1,5-octadiene-3,7-diol (0.9–
(11.4–12.9%) and hexadecanol (4.7–11.4%). Other linear
1.2%) is monoterpene polyol and other monoterpene poly-
saturated acids were also present such as acetic, hexanoic,
ols (especially linalool derivatives) were abundant in citrus
octanoic, nonanoic, decanoic and dodecanoic acid, as well
honey [10,20]. Identified phenylpropane derivatives benzyl
as other unsaturated acids such as octadecanoic acid and
alcohol (1.4–2.1%) and 2-phenylethanol (1.1–1.7%) are
(Z)-9-octadecenoic acid. Benzoic acid (0.7–7.4%) and vanil-
ubiquitous among honey volatiles and we also identified
lic acid (0.0–1.6%) were identified as aromatic carboxylic
3-phenyl-2-propen-1-ol (cinnamyl alcohol) with percent-
acids, originated mainly from cinnamic acid degradation.
ages 1.8–9.1%. The presence of a-methylbenzyl alcohol
In addition, phenylacetic acid (also derived from shikimate
(0.5–0.7%) is another characteristic of Cs honey [4]. Vola-
pathway) was also present (0.5–4.1%). Low percentage of
tile phenols were present with the following percentages:
these aromatic acids determined by USE could be impor-
4-vinyl-2-methoxyphenol (2.6–3.8%) and phenol (0.5–
tant for comparison with other unifloral honeys of different
1.3%). Linear acids were also identified such as acetic acid
botanical origin. Linear saturated hydrocarbons represent
(1.4–5.0%), propanoic acid (0.0–0.3%), butanoic acid (0.4–
another abundant chemical class of Rp USE honey volatiles
0.6%), hexanoic acid (0.6–1.2%), octanoic acid (0.9–1.2%)
including C13, C15, C16, C17, C18, C19, C21, but C22, C23 and
and nonanoic acid (2.0–2.1%). Hydrocarbons were less
C24 were the major n-alkanes. In general, the pattern of
abundant then in HD isolate, with the major ones: penta-
hydrocarbons in honey was very similar to beewax [21],
decane, nonadecane, heneicosane and docosane.
and in some cases plant wax may be transported by bees
from visited plants. It is assumed that they can also origi-
nate by condensation–dexarboxylation–reduction–elimina- 4. Conclusion
tion mechanism [22] or others. The aliphatic components
with long chains emanate from the wax of bees, cannot be While HD isolates contained many thermally derived
removed completely from honey [23]. Volatile phenols artifacts (especially phenylacetaldehyde) dependent on dif-
were also identified: 4-vinyl-2-methoxyphenol, 2,6-dimeth- ferent Rp and Cs matrix, the USE method gave the most
756 I. Jerković et al. / Ultrasonics Sonochemistry 14 (2007) 750–756
representative profile of all honey volatiles. Therefore we
can propose USE as a method to get the volatiles finger-
print of different honeys. In addition, USE enabled extrac-
tion of low molecular weight semivolatiles (especially
benzoic, vanillic and phenylacetic acids) that can not be
extracted by HD (it should be considered that their recov-
ery was low due to the use of non polar solvents for USE).
Low percentage of aromatic acids (benzoic, vanillic and
phenylacetic) in Rb USE extracts could be important for
comparison with other unifloral honeys of different botan-
ical origin. The percentage composition of USE com-
pounds was more uniformly, without striking dominance
of one compound, as in HD extract. Cs USE extracts con-
tained phenylacetic acid as the major constituent, and its
high percentage is important as new Cs honey marker.
Low percentage of benzoic acid should be noted. Specific
volatile markers of this honey were 4-aminoacetophenone
and 2-aminoacetophenone. We also identified cinnamyl
alcohol and a-methylbenzyl alcohol (known characteristic
of this honey).
References
[1] C. Guyot, V. Scheirman, S. Collin, Food Chem. 64 (1999) 3.
[2] L. Piasenzotto, L. Gracco, L. Conte, J. Sci. Food Agric. 83 (2003)
1037.
[3] J.S. Bonvehi, F.V. Coll, J. Sci. Food Agric. 83 (2003) 275.
[4] B.S. Radovic, M. Careri, A. Mangia, M. Musci, M. Gerboles, E.
Anklam, Food Chem. 72 (2001) 511.
[5] E. Alissandrakis, P.A. Tarantilis, P.C. Harizanis, M. Polissiou, J. Sci.
Food Agric. 85 (2005) 91.
[6] G.B. Nickerson, S.T. Likens, J. Chromatogr. 21 (1966) 1.
[7] C. Bicchi, F. Beliardo, C. Fratini, J. Apicult. Res. 22 (1983) 130.
[8] A. Bouseta, S. Collin, J. Agric. Food Chem. 43 (1995) 1890.
[9] L. Maignial, P. Pibarot, G. Bonetti, A. Chaintreau, J.P. Marion, J.
Chromatogr. 606 (1992) 87.
[10] E. Alissandrakis, D. Daferera, P.A. Tarantilis, M. Polissiou, P.C.
Harizanis, Food Chem. 82 (2003) 575.
[11] I. Rezić, A.J.M. Horvat, S. Babić, M. Kaštelan-Macan, Ultrason.
Sonochem. 12 (2005) 477.
[12] Narodne Novine 20 (2000) 280.
[13] I. Jerković, J. Mastelić, M. Miloš, Int. J. Food Sci. Tech. 36 (2001)
649.
[14] J. Mastelić, I. Jerković, Food Chem. 80 (2003) 135.
[15] R.P. Adams, Identification of Essential Oil Components by Gas
Chromatography and Mass Spectroscopy, Allured Publishers, Carol
Stream, IL, USA, 1995.
[16] K. Speer, A. Montag, Dtsch. Lebensm.-Rdsch. 80 (1984) 103.
[17] D.S. Mottram, ‘‘Meat’’, in: Hank Maarse (Ed.), Volatile Compounds
in Foods and Beverages, Marcel Dekker Inc., New York, 1991, pp.
107–177.
[18] E. Alissandrakis, P.A. Tarantilis, P.C. Harzizanis, M. Polissiou, Food
Chem. 100 (2007) 396.
[19] Y. Vasserot, A. Arnaud, P. Galzy, Acta Biotechnol. 15 (1995) 77.
[20] C.Y. Rowland, A.J. Blackman, B. D’Arcy, G.B. Rintoul, J. Agric.
Food Chem. 43 (1995) 753.
[21] G. Bonaga, A.G. Giumanini, G. Gliozzi, J. Agric. Food Chem. 34
(1986) 319.
[22] P.E. Kolattukudi, R. Croteau, J.S. Buckner, Chemistry and Bio-
chemistry of Natural Waxes, Elsevier, Amsterdam, 1976 (chapter 8).
[23] S.T. Tan, P.T. Holland, A.L. Wilkins, P.C. Molan, J. Agric. Food
Chem. 36 (1988) 453.