Food Chemistry 154 (2014) 179–186

Contents lists available at ScienceDirect

Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem

Different compounds are extracted with different time courses

from fruits during microwave hydrodiffusion: Examples and possible
causes
Aurélie Cendres a,b,⇑, Mélanie Hoerlé a,b, Farid Chemat a,b, Catherine M.G.C. Renard a,b
a
UMR408 Sécurité et Qualité des Produits d’Origine Végétale, INRA, F-84000 Avignon, France
b
UMR408 Sécurité et Qualité des Produits d’Origine Végétale, Université d’Avignon et des Pays du Vaucluse, F-84000 Avignon, France

a r t i c l e i n f o a b s t r a c t

Article history: We set out to determine how nutrients diffuse during extraction, using fractional collection. The highest
Received 28 August 2013 concentrations of sugars (195.5, 64.8 and 60.8 g/L, respectively for grape, ‘Najbolia’ plum and apricot)
Received in revised form 3 January 2014 were found for the earliest stages of extraction, with a decrease in concentration (to 41.4 g/L, 48.2 g/L
Accepted 5 January 2014
and 1.7 g/L, respectively) at the end of extraction process. Total polyphenols showed the same trends
Available online 10 January 2014
for plum and apricot (from 4.1 g/L to 2.9 g/L for ‘Najbolia’ plum, from 2.2 to 0.2 g/L for apricot) but highest
concentrations of total polyphenols (for grape and cherry) were obtained at fraction 5 or 6 (out of 7).
Keywords:
Carotenoids from cherry tomato also had highest concentrations (at circa 25 mg/L) almost at the end
Fruit juice
Time courses
of extraction. For volatile molecules from sweet cherry, hexanal, 2-hexenal and linalool had their highest
Micronutrients concentrations at fractions 3–4 (out of 7).
Microwave hydrodiffusion Diffusion of nutrients depended on fruit destructuring, molecule solubility and localization of the com-
pounds. Fruit size seemed unimportant.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction varied markedly during the treatment. This finding prompted us to

Extraction, notably from fruits for production of juices or spe-
cific ingredients, is a highly studies subject. Several key aspects
of fruit extraction have been examined, including the nature of
the fruit and the location of the components to be extracted with
respect to tissue structures. Numerous pre-treatments of the tissue
prior to extraction have been proposed to increase yields or rates of
extraction: effects of cell-wall degrading enzymes (Ribeiro, Enri-
que, Oliveira, Macedo, & Fleuri, 2010), ultrasound (Vilkhu, Mawson,
Simons, & Bates, 2008), microwave (Gerard & Roberts, 2004),
pulsed electric fields (Turk, Vorobiev, & Baron, 2012), etc. However,
these studies at best compare initial fruit and global juice compo-
sition. The extraction is treated as a single step. To our knowledge,
no studies have described modifications of juice composition and/
or different time courses during the fruit extraction. There is a
‘‘black box’’ about the diffusion of different molecules during
extraction. While studying microwave hydrodiffusion (MWH) as
a means of juice production from fruits (Cendres, Chemat, Main-
gonnat, & Renard, 2011), we noticed that juice colour and viscosity
⇑ Corresponding author at: UMR408 Sécurité et Qualité des Produits d’Origine
Végétale, INRA, Domaine Saint-Paul, F-84914 Avignon Cédex, France. Tel.: +33 (0)4
32 72 25 28; fax: +33 (0)4 32 72 24 92.
E-mail addresses: aurelie.cendres@paca.inra.fr, aurelie.cendres@hotmail.fr
(A. Cendres).
investigate the variation in juice composition.
Microwave hydrodiffusion (MWH) can be used for the extrac-
tion of juice from fruits or for the production of various extracts
(Orio et al., 2012). Microwave heating and hydrodiffusion allow ra-
pid extraction of juice from fresh or frozen fruits. Using frozen fruit
gives better yields (on average + 30%) due to dual destructuring
(through both freeze–thaw and MW heating) (Cendres et al.,
2011, 2012). From these experiments, the process of extraction
from frozen fruit can be described independently of the plant ma-
trix. Ice occupies more space than liquid water and water transfers
between intra- and extracellular compartments occur during ice
crystal formation. Consequently, freezing and thawing cause loss
of structure and turgor, and disruption of membranes (Petzold &
Aguilera, 2009). Some water in the fruit is in liquid form (bound
water or localized thawing during transport to MW oven). This
water absorbs microwaves and forms hot spots. A temperature rise
then spreads out from these hot spots, causing the ice to melt. Re-
lease of the first drop during microwave hydrodiffusion is the re-
sult of the melting of ice crystals. More and more water
molecules become susceptible to microwave irradiation. The aver-
age temperature increases to approximately 100 °C, with a change
in the mechanism of extraction. In this second step, the extraction
of juice proceeds by the generation of steam as in fresh fruits. This
step corresponds to further destructuring of the plant matrix. It is
during this stage that the maximum flow rate of juice is reached.

0308-8146/$ - see front matter Ó 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodchem.2014.01.004
180 A. Cendres et al. / Food Chemistry 154 (2014) 179–186
An added advantage of this process is that the freshly extracted
juice has very low contaminant levels, probably owing to (i) lim-
ited contacts with the outer surface of the fruit and (ii) elevated
temperatures. This extraction method preserves the colour and
natural taste of fruits.
Thus, to investigate the different step of extraction, and try to
understand how the extraction of the different molecules proceeds,
we decided to use MWH. This method had two advantages for our
investigations: (i) the extraction can be described independently of
the matrix, and (ii) we could collect samples for analysis at differ-
ent times along the extraction. The extraction of juice during MWH
is visibly inhomogeneous and probably involves different fruit con-
stituents during the extraction, depending on the specific charac-
teristics of the molecules such as localization and solubility.
Qualitative and quantitative data are required to study and under-
stand the difficulties met in the extraction of the different mole-
cules or families of molecules. To study this and assess impact
and factors of variability, we carried out fractional collection of
juices in a MWH experiment at different power density settings.
We linked the time courses of different components to their phys-
ico-chemical properties and their distribution in fruit.
We studied major compounds of fruits with sugars and acids,
using different fruits (plum, grape, sweet cherry, apricot and tomato)
in order to gather data on different size and nature of fruits. These
compounds are highly water-soluble and distributed throughout
the fruit. Concerning microconstituents, polyphenols and specifi-
cally anthocyans were the class of molecules for which we noticed
the existence of differences along the extraction (Cendres et al.,
2011). Therefore polyphenols, with lower, variable solubilities, were
studied in different fruits that presented different ratios of concen-
trations notably of highly visible anthocyans in skin and flesh (grape
and plum cv. ‘President’ et ‘Najbolia’ with red skin and yellow flesh,
sweet cherry (Burlat) with red skin and red flesh). For carotenoids,
which are hydrophobic, tomato was chosen due to the high concen-
trations of lycopene. Finally, volatile compounds were analysed for
cherry as its aromas are known and have been noticed to be very
different from flesh or stone.
2. Materials and methods
2.1. Chemicals and standards
Acetonitrile, methanol, (chromatographic quality) and hexane
were from Fisher Scientific (Illkirch, France), toluene-a-thiol was
from Merck™ (Darmstadt, Germany). Formic acid of purity
98–100% was from Sigma–Aldrich (Stenheim, Germany). Dichloro-
methane (for HPLC) was from Carlo Erba (France).
Enzymatic kits to measure D-glucose/fructose ref.: E 0139106,
D-glucose ref.: E 0716251, L-malic acid ref.: E 0139 068 and citric
acid ref.: E 0139076 were from R-Biopharm (Darmstadt, Germany),
and tartaric acid ref. 207.18.233 was from VWR (Fontenay-sous-
Bois, France).
All phenolic standards (p-coumaric acid, rutin, quercitrin, iso-
quercitrin, hyperoside, cyanidin-3-glucoside, cyanidin-3-rutino-
side and peonidin) were from Extrasynthèse (Lyon, France),
except for chlorogenic acid, (+)-catechin and ()-epicatechin,
which were from Sigma–Aldrich (Stenheim, Germany). Lycopene
was from Extrasynthèse (Lyon, France), b-carotene and apocarote-
nal were from Sigma–Aldrich (Stenheim, Germany). Benzaldehyde,
2-hexenal and linalol were from Merck™ (Darmstadt, Germany).
2.2. Plant material
Fruits free of visible defects were chosen randomly, washed,
dried using paper, weighed and frozen.
Plums (Prunus domestica L., cv. ‘President’ and ‘Najbolia’), grapes
(Vitis vinifera L., cv. Muscat), and apricots (Prunus armeniaca L.,
A3844) were obtained as described in Cendres et al. (2011). Cherry
tomato samples (Lycopersicon esculentum L., cv. sweet 100 F1) were
obtained from a local wholesaler from Avignon in August 2009.
Sweet cherries (Prunus avium L.var. Burlat) were obtained from a
private garden and harvested at maturity.
All fruits were portioned in batches of 500 g (except for apricot
250 g), frozen and stored at 18 °C.
2.3. Microwave juice extraction and procedure
Material and procedure is described in detail in Cendres et al.
(2011). Microwave extraction of fruit juice (Fig. 1) was performed
in a Milestone ‘‘DryDist’’ microwave laboratory oven (Milestone,
Bergamo, Lombardy, Italy). In a typical procedure performed at
atmospheric pressure, 500 g of frozen fruits (250 g for apricots)
were MW heated at power density 1 W g1. A mixture of hot
‘‘crude juice’’ (in situ water) and steam exited the microwave cavity
(through an opening underneath, due to gravity and pressure cre-
ated by steam). The juice ran out and the steam was condensed in a
condenser. Fractions of the fruit juice were collected (40 mL per
fraction, 20 mL for apricot) in 50 mL plastic tubes (Falcon, Becton
Dickinson, Franklin Lakes, New Jersey, USA). The extraction was
continued until no more fruit juice was obtained or overheating
was detected. At the end of the process, all the fractions were
individually frozen, stored at 18 °C and analyzed for sugars, acids,
polyphenols, carotenoids and/or volatile compounds depending on
the fruit species.
For cherries, juice (called steamer juice) was also produced by
using steam-cooker (Seb, France).
2.4. Colorimetric and enzymatic methods
The concentration of total phenolics was measured by the color-
imetric method of Folin–Ciocalteu (Singleton & Rossi, 1965). The
absorbance versus prepared blank was read at 730 nm in a Varian
UV–Visible spectrometer (Varian, Palo Alto, USA). Total phenolics
contents of juice were calculated against a calibration curve with
chlorogenic acid.
Total anthocyans were determined by the pH differential meth-
od of Wrolstad (1982), with the formula:
Concðg=LÞ ¼ Mw e ðA510 nm  A700 nm ÞpH1  ðA510 nm  A700 nm ÞpH7
The absorbances were read at 510 and 700 nm in a Varian UV–
Visible spectrophotometer (Varian, Palo Alto, USA). The concentra-
tions were expressed in cyanidin-3-rutinoside (plums cv. Najbolia
and President) or cyanidin-O-glucoside (apricot, grape, cherry)
equivalent with e = 28,800 L/mol/cm and molecular weight
595 Da for plums, and e = 26,900 L/mol/cm and molecular weight
449 Da, respectively for apricot, grape and cherry.
Malic acid, tartaric acid, sucrose, glucose and fructose were ana-
lyzed with colorimetric enzymatic tests using kits by an automatic
analyzer (Hitachi, Maidenhead, United Kingdom). These results are
expressed as sum of sugars, i.e. the sum of fructose, glucose and su-
crose, and sum of acids, i.e. the sum of malic acid, citric acid and
tartaric acid (the last exclusively for grape).
2.5. Determination of specific polyphenols by HPLC
Polyphenol analysis of plum and cherry was carried out on
freeze-dried fruit and juice material. Thioacidolysis was performed
as described by Guyot, Marnet, Sanoner, and Drilleau (2001). The
HPLC apparatus was a Shimadzu LC-20AD equipped with an SPD-
M20A DAD detector (Shimadzu, Kyoto, Japan). The column was a
 A. Cendres et al. / Food Chemistry 154 (2014) 179–186
Fig. 1. Typical extraction of fruit juice by microwave hydrodiffusion. Fractions of 40 mL were collected and analyzed; depending on fruit species, sugars, acids, polyphenols,
carotenoids and/or aromas were measured. (This result corresponds to extraction of juice from frozen (cv. President) plum at power density 1W g1.) N: temperature during
extraction; h: experimental points of microwave extraction; –: model curve of microwave extraction (fit of the extraction data); . . .: derivative curve describing the
extraction flow.
Phenomenex (Torrance, USA) Synergy fusion C18 5 lm
(4  200 mm). A gradient of solvent A (water:formic acid 98:2)
and solvent B (acetonitrile:water:formic acid 80:18:2) was used:
initial composition 100% of A; linear gradient to 20% of B from 0
to 30 min; linear gradient to 60% of B from 30 to 45 min; linear gra-
dient to 100% of B at 50 min, followed by a plateau for 5 min then
return to initial conditions and 20 min washing and reconditioning
before injection.
Procyanidins were quantified at 280 nm against an epicatechin
standard. Chlorogenic acid and p-coumaroylquinic acid were
quantified at 320 nm against a chlorogenic and p-coumaric acid
standard, respectively. Neochlorogenic acid was identified by
retention time and peak spectrum according to Kim, Jeong, and
Lee (2003) and expressed as chlorogenic acid equivalents. Flavo-
nols were identified against authentic standards: rutin, quercitrin,
isoquercetin and hyperoside, and quantified at 350 nm. Catechin
and epicatechin were identified and quantified at 280 nm against
authentic standards. Anthocyans were quantified at 520 nm
against authentic standards for cyanidin-3-glucoside and cyani-
din-3-rutinoside and for peonidin-3-rutinoside against its agly-
cone, peonidin.
2.6. Determination of tomato carotenoids in juice
Carotenoids were extracted using the micromethod described
by Serino, Gomez, Costagliola, and Gautier (2009) with some mod-
ifications. Carotenoids were assayed in tomato and tomato juice
(Microwave extraction at 1 W g1). Carotenoids were extracted as
follows: 500 mg of juice was weighed out in 2 mL Eppendorf tubes
containing 100 lg of zirconium balls (Biospec products, Bartles-
ville, USA), 10 lL of apocarotenal (0.18 g L1 dissolved in dichloro-
methane) and 100 lL of saturated NaCl solution. Three solvents
were added successively (50 lL of hexane, 200 lL of dichlorometh-
ane, 800 lL of ethyl acetate), each addition being followed by cen-
trifuging at 10,000g at 4 °C. After each centrifugation the organic
phases were collected and filtered on a 0.45 lm PTFE filter. The
three extracts were pooled. An aliquot of the organic fraction
(upper phase) was filtered and injected in an HPLC instrument.
The column was a Phenomenex (Torrance, USA) C8 3 lm
(4.6  100 mm) (oven temperature: 60 °C). A gradient of solvent
A (ammonium acetate 0.1 mol/L/methanol (30:70%, v/v) and sol-
vent B (methanol) was used: initial composition 80% of A; linear
gradient to 75% of B from 0 to 14 min; linear gradient to 95% of B
from 14 to 24 min; linear gradient to 100% of B at 29 min, then re-
181
turn to initial conditions and 10 min washing and reconditioning
before injection.
Compounds were identified by comparing retention time and
spectra with those of standards, and quantified at maximal molec-
ular absorption: 290 nm for phytoene, 350 nm for phytofluene,
450 nm for b-carotene and lutein, and 501 nm for cis- and trans-
lycopene, using apocarotenal as internal standard, against a cali-
bration curve for lycopene and b-carotene. Other compounds were
quantified in b-carotene equivalent.
2.7. Determination of volatile compounds in cherry juices
Volatile compounds of cherry juice were analyzed by head
space/SPME/GC/MS.
Vials of volume 20 mL with 5 g of juice were incubated at 40 °C
(with stirring) for 10 min. The headspace volatile compounds were
adsorbed by polymer fibres of different polarity (PDMS: apolar,
DVB: polar intermediary, Carboxen: polar (Supelco, Sigma Aldrich,
Stenheim, Germany)). The duration of adsorption was 30 min at
room temperature, followed by desorption in GPC injector at
250 °C for 5 min, followed by elution in CPG/MS/QP2010 (Shima-
dzu, Kyoto, Japan). Injection was in split mode (ratio 1/10) at
250 °C. The vector gas was helium with a velocity of 39 cm s1.
The column was a CP SIL 8 CB (a polar column – Varian),
30 m  0.25 mm  0.5 lm. Oven temperature program was 35 °C
for 1 min ramped at 5 °C min1–230 °C, and held for 5 min. Mass
spectra were obtained by electron impact at 70 eV, with scanning
from 29 to 250 amu at 2 scans s1. Identification was carried out
by comparison with Wiley8 and NIST8 bank.
2.8. Statistical analysis
Two technological repetitions were carried out for each fruit
juice. For each analysis, samples were prepared and analyzed in
duplicate, giving 4 values (2 analytical  2 technological) per point.
Results are presented as mean values, and the reproducibility of
the results is expressed as pooled standard deviation. Pooled stan-
dard deviations were calculated for each series of replicates using
the sum of individual variances weighted by the individual degrees
of freedom (Box, Hunter, & Hunter, 1978). The variability in
composition between varieties was expressed by calculating the
average of the mean values for each variety and the standard devi-
ation of the mean.
182 A. Cendres et al. / Food Chemistry 154 (2014) 179–186

Two-way analysis of variance (ANOVA) by Fisher’s test (F) was

(a)
used to compare the means, and performed using the Excelstat 70
package of Microsoft Excel. Differences were considered significant 60

Sugars (g.L-1 )
at P < 0.05. 50
40
30
3. Results and discussion 20
10
3.1. Fruit description 0
1 2 3 4 5 6 7
Collected fraction
The largest fruit were plums cv. ‘President’ (on average 74 g)
followed by ‘Najbolia’ and apricot (52 and 51 g respectively); cher- 100 (b)

Anthocyans (mg.L-1 )
ry tomato, grape and sweet cherry were much smaller (<10 g). Cor- 80
respondingly, the proportions of flesh tissue were highest in plums,
60
followed by apricots and grapes (for details, see Supplementary
data Table SI). Most sugars and acids were found in the flesh of 40
fruits, as written by Glew et al. (2005). 20
Polyphenol concentrations were much higher in skin than in
0
flesh. For example total polyphenol concentrations were of 3.50, 1 2 3 4 5 6 7
3.98, 6.60 and 5.99 g kg1 of FW in the skin of apricot, grape, plums Collected fraction
cv. ‘President’ and cv. ‘Najbolia’, respectively, while concentrations
in the flesh were of 1.10, 0.52, 3.30 and 3.99 g kg1 FW. Anthocyans
35 (c)
30

Carotenoïds (mg.L-1)
were exclusively present in skins for apricot, grape and both
25
plums. However, considering total amounts, 25%, 37%, 50% and
20
76% of polyphenols were found in the skin fraction of plum cv.
15
‘President’’’ and cv. ‘Najbolia’, apricot and grape, respectively (for
details, see Supplementary data Table SII). 10
Detailed analysis carried out for plum cv. ‘President’ (Table 1) 5
indicated that the concentration in skin was higher for all classes, 0
with anthocyans and flavonol present only in the skin. Procyani- 1 2 3 4 5 6 7 8
dins were the main polyphenols of plum. This was in agreement Collected fraction
with Usenik, Kastelec, Veberic, and Stampar (2008). However, gi-
Fig. 2. Impact of microwave power (0.5 (N), 1 (h) and 1.5 ( ) W g1 of fruit) on
ven the proportions of skin (14%) and flesh (84%) in whole plum, time courses of extraction of different molecules: (a) sugars (sum of glucose,
most procyanidins and phenolic acids were in the flesh. Phenolic sucrose and fructose) from frozen plums cv. ‘President’, (b) total anthocyans from
compositions of the different tissues were clearly differentiated frozen plums cv. ‘President’ and (c) carotenoïds from frozen cherry tomato. Each
in all fruit used, allowing to relate composition of juice to tissular fraction corresponds to 40 mL of juice.
origin.

3.2. Impact of MW power
The influence of microwave energy on extraction is strictly
thermal. The microwave energy quantum is given by the usual
equation W = hm. Within the frequency domain of microwaves
and hyperfrequencies (300 MHz–300 GHz), the corresponding
energies are in the range 1.24  106–1.24  103 eV. These ener-
gies are much lower than the usual ionisation energies of biological
compounds (13.6 eV), covalent bond energies, e.g. OH (5 eV),
hydrogen bonds (2 eV), Van der Waals intermolecular interactions
(lower than 2 eV) and even lower than the energy associated with
Brownian motion at 37 °C (2.7  103 eV). Hence direct molecular
activation by microwaves can be excluded. Any step-by-step accu-
mulation of the energy, giving rise to a high-activated state can be
totally excluded due to fast relaxation. With increasing power,
time to reach a given temperature is reduced.
Fig. 2a shows time courses of plum juice extraction for total
sugars (sum of fructose, glucose and sucrose) at three microwave
power densities; similar patterns were obtained for total acids
(see Supplementary data Fig. S1a) (sum of malic and citric acids)
and total polyphenols (see Supplementary data Fig. S1b). The same
configuration was followed at the different powers: the highest
concentrations were observed in the first juice fraction for all the
compounds, where the concentrations were similar whatever the
power applied. Concentrations stayed constant or tended to
decrease slowly until the very last fraction, for which a marked
decrease was observed. By contrast, for anthocyans (Fig. 2b) and
carotenoids (Fig. 2c), higher concentrations were found in the
middle or end of extraction, especially at low power (Fractions 5
and 6, respectively). At high power density, diffusion of anthocyans
and carotenoids was facilitated, probably due to increased destruc-
turing. Carotenoids are hydrophobic and were extracted when par-
ticles were carried along at high flow rates and with fruit
disintegration, i.e. at the highest power densities. Two phases were
observed in MW hydrodiffusion time courses of carotenoids
(Fig. 2c). The first one, where carotenoid concentration ranged
from 7 to 15 mg kg1 depending on MW power, could correspond
to easily extracted compounds, i.e. carotenoids present in inner
layers of fruit flesh.
The final fractions collected at low power were generally low in
sugars, acids, polyphenols and anthocyans, explaining identical fi-
nal concentrations in juices obtained at different powers (Cendres
et al., 2012). Extraction at high power densities decreases yield of
juice (Cendres et al., 2011); it therefore did not lead to exhaustion
of the fruit, as temperature increase became the limiting factor.
Power did not seem to have any significant impact on extraction
of acids, sugars, or total polyphenols, but seemed to affect diffusion
of anthocyans and carotenoids, compounds more difficult to
extract.
3.3. Concentration variation for different compounds during MWH
3.3.1. Sugars and acids
Sugars and acids are soluble compounds mainly present in the
flesh of the fruit, and are located in vacuoles in the flesh. The
 A. Cendres et al. / Food Chemistry 154 (2014) 179–186 183

highest concentrations of sugars and acids were found in the early
stage of extraction with a variably marked decrease at the end of
extraction. Fruit size did not seem to have any significant impact
on sugar and acid extraction, similar time-courses being observed
for plums cv. ‘President’ (Fig. 2a), plum cv. ‘Najbolia’ (Fig. 3a and b)
and grape. However fruit structure had a major impact, as apricot,
with a similar size as plum cv. ‘Najbolia’, formed a foam in the reac-
tor and released much less juice than the other fruit species tested,
as already described in Cendres et al. (2011).
3.3.2. Total polyphenols
Total polyphenols (Fig. 3c) were mainly released early in MW
extraction of plum (whatever the power) and apricot. Polyphenols
in plums and apricot were mainly localized in fruit flesh, 75% and
50%, respectively. For grape juice, polyphenol concentrations
peaked in the middle of the extraction process. This typical profile
may be explained by the distribution of polyphenols in grape. Un-
like the other two fruits, the phenolic compounds in grapes were
mainly localized in skin (76%) (Supplementary data, Table SII).
200
(a)
Sugar (g.L-1)
With sweet cherry, the polyphenol concentration was more
constant during the process, although a peak was reached at
Fraction 6. In sweet cherries, anthocyans comprise 90% of total
polyphenols (Jakobek, Seruga, Seruga, Novak, & Medvidovic-Kosa-
novic, 2009). Although anthocyans were present in skin and flesh,
these compounds are more concentrated in skin (Lenucci, Cadinu,
Taurino, Piro, & Dalessandro, 2006).
Thus localization had an important impact on hydrodiffusion of
polyphenols. When polyphenols were mainly localized in fruit
flesh, the diffusion was immediate, whereas when polyphenols
were mainly localized in fruit skin, the highest concentrations
were reached in intermediate fractions. This could be related to
difficulty in rupturing the cells in the epidermis, which are smaller
and have more resistant cell walls than in the flesh. This would
mean that the outermost fractions of the fruit juice were not
extracted first in this method.
3.3.3. Total anthocyans
Two profiles were observed in Fig. 3d. Firstly, for apricot, grape
and plum, in which anthocyans were present only in the epidermic
region, release of anthocyans was gradual with maximal concen-
tration at the middle and decrease at the end of extraction. A long
time was required to release anthocyans, owing to their epidermic
localization. With sweet cherry, skin and flesh both have antho-

150 cyan (Lenucci et al., 2006), and so concentration was more constant
100
during the extraction, although a peak was reached at Fraction 6,
suggesting release of the anthocyans from the cherry skin, where
50 their concentrations are higher than in the flesh. Again, the
0
thickness of the wall and cell size could account for more difficult
0 1 2 3 4 5 6 7 rupture of skin cells.
Collected Fraction

20 (b)
3.3.4. Specific polyphenols
Acid (g.L-1)

15 Specific polyphenols were identified and quantified by HPLC in

10
plums and cherry; only results from the plum cv. ‘President’ are
presented in Fig. 4a and Table 1. They comprise in addition to
5 anthocyans, highly soluble phenolic acids (550 mg kg1), located
0
mainly in the flesh (61% in the flesh), flavonols (74 mg kg1, spe-
0 1 2 3 4 5 6 7 cific for skin, and flavan-3-ol both as monomers (catechin
Collected Fraction 7 mg kg1) or oligomers (procyanidin 881 mg kg1).
(c) Extraction profiles obtained for specific polyphenols classes
Total polyphenols (g.L-1)

5
(Fig. 4a) varied and could be different from those observed for to-
4
tal polyphenols (Fig. 3c). Maximum concentrations were ob-
3 served in the first fraction for flavonols (with decrease along the
2 time course) and in the third fraction for phenolic acids, catechin
and anthocyans. Procyanidins were released more gradually to a
1
maximum concentration in the fourth fraction. Extraction of cat-
0
0 1 2 3 4 5 6 7 8
echin, anthocyans and phenolic acids before procyanidins could
Collected Fraction be explained by the greater solubility of the monomeric polyphe-
nols. Also, the difficulty in extracting procyanidins could be
700 (d) caused by retention by cell walls, as in conventional pressing
Anthocyans (mg.L-1 )

600
500
(Le Bourvellec, Guyot, & Renard, 2009). Individual polyphenols
400
(individual phenolic acids, anthocyans or flavonols) are not illus-
300 trated: for anthocyans and flavonols, the relative proportions of
200 the different molecules did not vary. For phenolic acids, better
100 extractions were observed for neochlorogenic and p-coumaroyl-
0 quinic acids, mostly present in the flesh, than for chlorogenic acid
0 1 2 3 4 5 6 7 8 (data not shown). With HPLC analysis, the sum of individual
Collected Fraction polyphenols presented a triangular profile, whereas using the
Folin–Ciocalteu methods, the total polyphenol concentrations
Fig. 3. Variation of the time course of microwave extraction with the original fruit: were highest at the beginning of extraction (except for grape).
time course of different constituents (sugars, acids, total polyphenols and anthoc- This difference could be related to degradation of polyphenols;
yans) from frozen grape (N), apricot (h), plum cv. ‘Najbolia’ () and sweet cherry
(j) (only anthocyans and total polyphenols for sweet cherry) during microwave
HPLC analysis considered only native polyphenols, whereas all
extraction. All treatments were carried out at 1 W g1 and 40 mL fractions were polyphenols and reductive compounds react in the Folin-Ciocal-
collected except for apricot (20 mL). teu assay (Brat et al., 2007).
184 A. Cendres et al. / Food Chemistry 154 (2014) 179–186

0.5 (a)

Specific polyphenols (g.L-1 )

0.4
0.3
0.2
0.1
0
1 2 3 4 5 6
Collected fraction
100 (b)
Specific Carotenoïds

80
(mg.L-1 )

60

40

20

0
1 2 3 4 5 6 7 8
Collected fraction
10
(c)
Hexanal & 2-hexenal

8
Area (x 100000)

6
4
2
0
1 2 3 4 5 6 7

Collected fraction
& Furfural (x 1000000)

(d) 10
4.0 9

Area Linalol (x10000)

Area Benzaldehyde

3.5 8
3.0 7
2.5 6
2.0 5
1.5 4
1.0 3
0.5 2
1
0 0
1 2 3 4 5 6 7
Collected fraction
Fig. 4. Time courses for extraction of specific microconstituents from the different fruits during MWH treatment. All treatments were carried out at 1 W g1 and 40 mL fractions
were collected. Polyphenol classes (sum of concentrations obtained for the individual compounds measured by HPLC) are illustrated for plum cv. ‘President’ in (a): procyanidins
( ), phenolic acids ( ), catechins ( ), flavonols (j), and anthocyans (h). Specific carotenoids during tomato juice extraction are presented in (b). Carotenoids were divided into
groups of compounds: the most abundant in tomatoes native lycopene ( ) and b-carotene (h) are presented individually; the minor compounds (j)(phytoene, lutein and
phytofluene) were grouped so that their distribution be visible. Lycopene isomers ( ) were grouped to compare their distribution with that of all-trans lycopene. Volatile
compounds during cherry juice extraction are presented in (d) for hexanal ( ) and 2-hexenal ( ) and (e) for linalol (j), furfural ( ) and benzaldehyde (h).

3.3.5. Carotenoids (hydrophobicity scale: lycopene > b-carotene > xanthophylls) and
Commercial tomato ‘‘juices’’ are actually thin cold-break tomato were extracted when particles were carried along at high flow
purées resulting from sieving and heat treatment, rather than rates and with fruit disintegration. All classes of carotenoids fol-
pressing (Van Breemen & Pajkovic, 2008). With MW hydrodiffu- lowed the same time course (Fig. 4b). Isomers of lycopene were
sion, it is possible to extract a tomato juice, which we did to follow present from the beginning of extraction, and their concentration
the extraction of fat-soluble carotenoids in microwave hydrodiffu- increased with fruit temperature. Increase in isomers was due to
sion. Lycopene is the most abundant carotenoid in tomatoes (Lyc- thermal degradation (Nguyen, Francis, & Schwartz, 2001).
opersicon esculentum L.) with concentrations in the range 50– In plant cells, carotenoids are synthesized and stored within the
120 mg/kg depending on the variety (Gomez-Romero et al., chloroplasts, and so the extraction of carotenoids requires the
2007). In cherry tomato, concentration of b-carotene in the range rupture of the plant cell wall, three membranes (plasmalemma,
0.5–12 mg/kg (Lenucci et al., 2006) have been reported. The carot- plast and thylakoid) and may need the destruction of protein
enoid concentrations in cherry tomatoes used were within these assemblies.
rangs with 81 ± 2.6 mg/kg FW for lycopene and 4.2 ± 0.1 mg/kg
for b-carotene. 3.3.6. Volatile compounds
Yields of carotenoids were low overall, representing, for total We chose sweet cherry to analyze volatile compounds: cherry
juice extracted at 1 W/g, only one sixth of total carotenoids present aroma as used in products such as yogurt or cherry juice has an
in tomato fruit with concentrations of 13 ± 3 mg/L for lycopene and important component of ‘‘bitter almond-like’’ odour, arising from
1.4 ± 0.1 mg/L for b-carotene. Carotenoids are hydrophobic compounds such as benzaldehyde (Girard & Kopp, 1998) initially
 A. Cendres et al. / Food Chemistry 154 (2014) 179–186 185

Table 1
Specific polyphenol concentrations in mg kg1 in the flesh and skin plums cv. ‘President’ analyzed by HPLC Values for whole plum are calculated from content and proportion of
skin and flesh.

Fruit part DM Pyrocyanidins Flavan-3-ols Phenolic acids Flavonols Anthocyans

monomer
g/g g/g Conc Dp CAT nCQA Coum CQA Tot Rut Gly. Querc. Tot Cya-3- Cya-3- Tot
gly rut
Plums cv. President
Flesh 0.84 0.17 8.67 5.9 7 310 40 50 400
SD 53 1 2 0 1
Skin 0.14 0.19 1095 3.3 8 1040 70 480 1590 314 214 529 234 224 458
SD 21 1 6 0 4 12 41 12 8
% in skin 20% 15% 36% 25% 62% 39% 100% 100% 100% 100%
Whole plums 0.17 881 5.4 7 400 40 110 550 44 30 74 33 31 64

With fruit part: the proportions in wet weight relative to g of fruit; DM, dry matter; DP, degree of polymerization; Tot, total; CAT, catechin; nCQA, neochlorogenic acid; Coum,
p-coumaroylquinic acid; CQA, chlorogenic acid; rut, rutin; Gly. Quer., other glycosides of quercetin; Cya-3-glu, cyanidine-3-glucoside; Cya-3-rut, cyanidin-3-rutinoside; SD,
standard deviation. The percent value represent the content in skin compared to total fruit.

present in the stone of the fruit. Zhang et al. (2007) report that hex-
anal, 2-hexenal and benzaldehyde are the main volatile com-
pounds in sweet cherry. Benzaldehyde originates from enzymatic
hydrolysis of amygdalin in stone fruits or can be derived from
precursors such as phenylalanine and benzyl alcohol (Souty &
Reich, 1978). C6 aldehyde compounds, including 2-hexenal and
hexanal, are the main volatile compounds of sweet cherries in
the commercial state (Zhang et al., 2007). C6 aldehydes are associ-
ated with green/grassy notes and are formed by the lipoxygenase
pathway from unsaturated fatty acids. The initial concentration
of these compounds in the sweet cherries were 1.24 mg/kg for
2-hexenal, 1.64 mg/kg for benzaldehyde.
Extraction of hexanal and 2-hexenal showed similar profiles,
these two compounds having a ‘‘triangular diffusion profile’’ com-
parable to that of anthocyans (Fig. 4c). Hexanal was more easily ex-
tracted when the temperature reached 100 °C. C6 aldehydes,
hexanal and (E)-2-hexenal were present in fruit before juice
extraction. These polar compounds would be expected to be
strongly extracted by the polar water as demonstrated by Ramir-
ez-Rodrigues, Balaban, Marshall, and Rouseff (2011) in water infu-
sion. Probably, lipoxygenase became inactive due to temperature
rise, which would account for their concentration decrease in the
second part of the extraction time course.
Benzaldehyde, also a native compound of cherry, was present at
the beginning of extraction, whereas linalol was absent. Linalol ap-
peared in the second fraction when temperature reached 100 °C.
Linalol and benzaldehyde (Fig. 4d) concentration increased with
temperature of extraction, and these compounds reached their
highest concentrations at the end of extraction. Even though the
stones were macroscopically unchanged, microfissures could be
formed during heating and might explain the linear release of
benzaldehyde from stones. Comparison with juice produced by
steam heating supported this hypothesis, as we had a concentra-
tion seven times higher in MW juice (31 ± 7.5 mg/L) than in ‘‘stea-
mer juice’’ (4.3 mg/L).
Furfural, probably a product of sugar dehydration, linked to the
high temperatures reached in the fruit, is not detected in raw fruits
(Zhang et al., 2007). This compound appeared during the process.
In this juice production, furfural appeared in the last fraction (peak
area around 108), when temperature exceeded 100 °C (Fig. 4d).
Except for furfural, induced by this thermal reaction, release of
aroma compounds was complex and depended on enzymes, on sol-
ubility and on interactions among the constituents such as sugars
or organic acids.
In summary, we can say that during thermal treatments, the
impact compounds of the fresh cherry flavour (originating from
the endogenous pathways in fresh cherry) decreased as observed
in tomato (Buttery, Takeoka, Teranishi, & Ling, 1990); whereas
the sugar dehydration reaction activation promoted the genesis
of volatile compounds related to the cooked flavor (Servili, Selvag-
gini, Taticchi, Begliomini, & Montedoro, 2000). Furfural is the most
representative aroma generated due to heating.
4. Conclusions
Microwave hydrodiffusion generated juices with different com-
positions during the course of extraction, which could allow spe-
cific juices to be obtained. Following extraction step by step, we
found that times of diffusion were related to the localization and
physico-chemical properties of the various constituents. Localiza-
tion in epidermis or parenchyma had a marked influence. This
could be related to relative cell sizes and cell wall thickness or
resistance, explaining why compounds present in the epidermis
were extracted less or later.
A high concentration of sugar, acids, and total polyphenols
(except in grape) were markers of beginning of extraction with di-
rect hydrodiffusion of these compounds. A high concentration of
anthocyans, some specific polyphenols of plums and sweet cher-
ries, carotenoids and some volatiles (hexanal, 2-hexenal, linalol)
were specifically reached in the middle of extraction. Furfural
and benzaldehyde were characteristic of the end of extraction in
sweet cherries. Using high power densities, thus allowing more
fruit destructuring, seems useful when compounds are hydropho-
bic (carotenoids) or in the epidermis.
This research demonstrates that the extract composition varies
during the course of extraction, and qualitative and quantitative
data are required to study and understand the difficulties met in
the extraction of compounds.
The specific aspects of extraction are of interest for the creation
of new products and the manipulation of organoleptic and nutri-
tional characteristics of the juice. These findings could be of inter-
est for juice production to adapt time and temperature to reach
maximal extraction or to create a juice that contains greater
amounts of one or more nutrients, thus offering a product with
higher nutritional value, and improved acceptance by consumers.
Acknowledgements
The authors thank C. Ginies and M. Bogé for their excellent
technical help.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.foodchem.2014.
01.004.
186 A. Cendres et al. / Food Chemistry 154 (2014) 179–186

References Nguyen, M., Francis, D., & Schwartz, S. (2001). Thermal isomerisation susceptibility
of carotenoids in different tomato varieties. Journal of the Science of Food and
Agriculture, 81, 910–917.
Box, G. E. P., Hunter, W. G., & Hunter, J. S. (1978). Statistics for experimenters, an
Orio, L., Cravotto, G., Binello, A., Pignata, G., Nicola, S., & Chemat, F. (2012).
introduction to design, data analysis, and model building (653 pages). New York:
Hydrodistillation and in situ microwave-generated hydrodistillation of fresh
Wiley J. et Sons.
and dried mint leaves: A comparison study. Journal of the Science of Food and
Brat, P., Mennen, L., George, S., Scalbert, A., Bellamy, A., Amiot-Carlin, M. J., et al.
Agriculture, 3085–3090.
(2007). Determination of the polyphenol content of fruits and vegetables.
Petzold, G., & Aguilera, J. M. (2009). Ice morphology: Fundamentals and
Establishment of a database and estimation of the polyphenol intake in the
technological applications in foods. Food Biophysics, 4, 378–396.
French diet. Acta Horticulturae, 61–69.
Ramirez-Rodrigues, M. M., Balaban, M. O., Marshall, M. R., & Rouseff, R. L. (2011).
Buttery, R. G., Takeoka, G., Teranishi, R., & Ling, L. C. (1990). Tomato aroma
Hot and cold water infusion aroma profiles of Hibiscus sabdariffa: Fresh
components – Identification of glycoside hydrolysis volatiles. Journal of
compared with dried. Journal of Food Science, 76, 212–217.
Agricultural and Food Chemistry, 38, 2050–2053.
Ribeiro, D. S., Enrique, S. M. B., Oliveira, L. S., Macedo, G. A., & Fleuri, L. F. (2010).
Cendres, A., Chemat, F., Maingonnat, J. F., & Renard, C. M. G. C. (2011). An innovative
Enzymes in juice processing: A review. International Journal of food Science and
process for extraction of fruit juice using microwave heating. LWT – Food Science
Technology, 54, 635–641.
and Technology, 44, 1035–1041.
Serino, S., Gomez, L., Costagliola, G., & Gautier, H. (2009). HPLC assay of tomato
Cendres, A., Chemat, F., Page, D., Le Bourvellec, C., Markowski, J., Zbrzezniak, M.,
carotenoids: Validation of a rapid microextraction technique. Journal of
et al. (2012). Comparison between microwave hydrodiffusion and pressing for
Agricultural and Food Chemistry, 57, 8753–8760.
plum juice extraction. LWT – Food Science and Technology, 49, 229–237.
Servili, M., Selvaggini, R., Taticchi, A., Begliomini, A. L., & Montedoro, G. (2000).
Gerard, K. A., & Roberts, J. S. (2004). Microwave heating of apple mash to improve
Relationships between the volatile compounds evaluated by solid phase
juice yield and quality. LWT – Food Science and Technology, 37, 551–557.
microextraction and the thermal treatment of tomato juice: Optimization of
Girard, B., & Kopp, T. G. (1998). Physicochemical characteristics of selected sweet
the blanching parameters. Food Chemistry, 71, 407–415.
cherry cultivars. Journal of Agricultural and Food Chemistry, 46, 471–476.
Singleton, V. L., & Rossi, J. A. (1965). Colorimetry of total phenolics with
Glew, R. H., Ayaz, F. A., Millson, M., Huang, H. S., Chuang, L. T., Sanz, C., et al. (2005).
phosphomolybdic–phosphotungstic acid reagents. American Journal of Enology
Changes in sugars, acids and fatty acids in naturally parthenocarpic date plum
Viticulture, 48, 144–158.
persimmon (Diospyros lotus L.) fruit during maturation and ripening. European
Souty, M., & Reich, M. (1978). Effets de traitements technologiques: congélation et
Food Research and Technology, 221, 113–118.
appertisations sur certains constituants de l’arôme des pêches. Annales de
Gomez-Romero, M., Arraez-Roman, D., Segura-Carretero, A., & Fernandez-Gutierrez,
Technologie Agricole, 27, 837–848.
A. (2007). Analytical determination of antioxidants in tomato: Typical
Turk, M., Vorobiev, E., & Baron, A. (2012). Improving apple juice expression by
components of the Mediterranean diet. Journal of Separation Science, 30, 452–461.
pulsed electric field on an industrial scale. LWT – Food Science and Technology,
Guyot, S., Marnet, N., Sanoner, P., & Drilleau, J. F. (2001). Direct thiolysis on crude
49, 245–250.
apple materials for high-performance liquid chromatography characterization
Usenik, V., Kastelec, D., Veberic, R., & Stampar, F. (2008). Quality changes during
and quantification of polyphenols in cider apple tissues and juices. Methods for
ripening of plums (Prunus domestica L.). Food Chemistry, 111(4), 830–836.
enzymology – Flavonoids and other polyphenol (Vol. 335, pp. 57–70). Amsterdam,
Van Breemen, R. B., & Pajkovic, N. (2008). Multitargeted therapy of cancer by
The Netherlands: Elsevier Science Publishers.
lycopene. Cancer Letters, 269, 339–351.
Jakobek, L., Seruga, M., Seruga, B., Novak, I., & Medvidovic-Kosanovic, M. (2009).
Vilkhu, K., Mawson, R., Simons, L., & Bates, D. (2008). Applications and opportunities
Phenolic compound composition and antioxidant activity of fruits of Rubus and
for ultrasound assisted extraction in the food industry – A review. Innovative
Prunus species from Croatia. International Journal of Food Science and Technology,
Food Science & Emerging Technologies, 161–169.
44, 860–868.
Wrolstad, R. E. (1982). Description of the pH differential method for determination
Kim, D. O., Jeong, S. W., & Lee, C. Y. (2003). Antioxidant capacity of phenolic
of total anthocyanins and indices for anthocyanin degradation as applied to
phytochemicals from various cultivars of plums. Food Chemistry, 81, 321–326.
fruit juices and wines. In R. E. Wrolstad (Ed.), Current protocols in food analytical
Le Bourvellec, C., Guyot, S., & Renard, C. M. G. C. (2009). Interactions between apple
chemistry (pp. 1417–1423). New York: Wiley.
(Malus x domestica Borkh.) polyphenols and cell walls modulate the
Zhang, X., Jiang, Y. M., Peng, F. T., He, N. B., Li, Y. J., & Zhao, D. C. (2007). Changes of
extractability of polysaccharides. Carbohydrate Polymers, 75, 251–261.
aroma components in Hongdeng sweet cherry during fruit development.
Lenucci, M. S., Cadinu, D., Taurino, M., Piro, G., & Dalessandro, G. (2006). Antioxidant
Agricultural Sciences in China, 6, 1376–1382.
composition in cherry and high-pigment tomato cultivars. Journal of Agricultural
and Food Chemistry, 54, 2606–2613.