Journal of Food Engineering 59 (2003) 361–367
www.elsevier.com/locate/jfoodeng
Comparison of different methods for deacidification
of clarified passion fruit juice
Edwin Vera a, Jenny Ruales a, Manuel Dornier b,*, Jacqueline Sandeaux c,
Francßoise Persin c, Gerald Pourcelly c, Fabrice Vaillant b, Max Reynes b
a
c
Department of Food Science and Biotechnology, Escuela Polit
ecnica Nacional, P.O. Box 170 12759 Quito, Ecuador
b
CIRAD-FLHOR/ENSIA-SIARC, Avenue Agropolis, TA 50/PS4, 34398 Montpellier Cedex 5, France
Institut Europ
een des Membranes, Universit
e de Montpellier II, CC 047, Place Eug
ene Bataillon, 34095 Montpellier Cedex 5, France
Received 15 June 2002; accepted 1 December 2002
Abstract
The high acidity of passion fruit limits its addition in food preparation. In order to easy the uses of this juice to formulate high
aroma and flavour products, its citric acid content must be decreased. Various methods such as calcium salts precipitation, ionexchange resins and electrodialysis with homopolar and bipolar membranes were investigated to increase the pH of a clarified
passion fruit juice from 2.9 to 4.0. Each deacidification process was compared in terms of characteristics of deacidified juices such as
concentration of inorganic and organic ions, colour and flavour. The deacidification by precipitation using CaCO3 was not recommended because of CO2 release. The ion-exchange process gave a poor quality juice in terms of organoleptic characteristics.
Electrodialysis with homopolar membranes induced an increase in the sodium concentration. Precipitation with Ca(OH)2 and
electrodialysis with bipolar membrane were the most suitable processes in terms of sensorial properties of juices treated. The
physico-chemical analyses showed an increase in the calcium concentration with the use of Ca(OH)2 that could cause some precipitation problems in the final product. Inorganic anions were eliminated together with citrate by using electrodialysis and resins.
Nevertheless, electrodialysis with bipolar membranes presented great advantages: it was a continuous process without reagent
addition moreover allowing the production of a valuable solution of citric acid.
2003 Elsevier Science Ltd. All rights reserved.
Keywords: Passion fruit juice; Deacidification; Precipitation; Ion-exchange resin; Conventional electrodialysis; Bipolar electrodialysis
1. Introduction
Passiflora edulis v. flavicarpa or passion fruit is used
for production of concentrated juice (Serna & Chacon,
1988). Export of this juice at 50 Brix (500 g kg1 of total
soluble solids) is a very important source of income in
several South American countries such as Ecuador,
Brazil, Peru and Colombia. In Ecuador, the principal
exporter country (Loeillet, 1999), the passion fruit occupied in 2000 the second place in exports with 26 000
tons of concentrated juice (US$ 28 million).
The juice quality could be improved by a reduction of
its acidity that would facilitate its use in many food
products (Adhikary, Harkare, Govindan, & Nanjundaswamy, 1983; Bhatia, Dang, & Gaur, 1979; Couture &
*
Corresponding author. Fax: +33-4-67-61-44-33.
E-mail address: dornier@cirad.fr (M. Dornier).
Rouseff, 1992; Goloubev & Salem, 1989; Johnson &
Chandler, 1985, 1986; Lue & Chiang, 1989; Scott, 1995).
In previous studies, ion-exchange (IE) and electrodialysis (ED) processes were investigated for deacidification of clarified passion fruit juice (Vera, Dornier,
Ruales, Reynes, & Vaillant, 2003; Vera, Ruales, et al.,
2003). Ten weakly basic resins were tested and their
performances compared. The most suitable resins were
Amberlite IRA95 and Duolite A378, because they gave
the highest content of juice treated with the lowest
content of soda required for regeneration. Concerning
the electrodialysis process, different ED designs using
homopolar or bipolar membranes were investigated.
Among the three anion exchange membranes studied,
the AXE01 membrane was selected for its deacidification rate, current efficiency and energy consumption.
Moreover, its use was authorized in the food industry.
The aim of this paper is to compare both the above
physico-chemical techniques with the conventional
0260-8774/03/$ - see front matter 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0260-8774(02)00495-8
362
E. Vera et al. / Journal of Food Engineering 59 (2003) 361–367
chemical method based on precipitation of calcium
citrate, obtained by addition of calcium hydroxide or
calcium carbonate to the clarified juice.
In order to avoid microorganism growth and spoilage, the increase in pH value was limited to 4.0 for all
deacidified juices. Then, the physico-chemical and sensorial properties of juices treated were analysed and
compared to those of fresh juice.
2. Materials and methods
2.1. Passion fruit juice
The fruits (Passiflora edulis v. flavicarpa) were obtained at Quevedo (Ecuador) and treated as previously
described (Vera, Dornier, et al., 2003). After an enzymatic treatment, the juice was clarified to avoid fouling
of the resins and ion-exchange membranes by using
crossflow micro-filtration.
The characteristics of the clarified passion fruit juice
after pretreatment are given in Table 1. Citrate and
potassium ions were the most abundant anion (91%)
and cation (94%), respectively.
2.2. Deacidification by precipitation of calcium citrate
The reagents, CaCO3 and Ca(OH)2 , were products of
analytical degree from Sigma. The amounts required for
Table 1
Characteristics of the clarified passion fruit juice
Density (kg m3 )
Viscosity (mPa s)
pH
Colour
L
a
b
Total soluble solids (g kg1 )
Total sugars (g kg1 )
Titrable acidity
(g citric acid per kg)
(meq kg1 )
Citric acid (g kg1 )
Malic acid (g kg1 )
Tartaric acid (g kg1 )
Total minerals (g kg1 )
Chloride (mg kg1 )
Phosphate (mg kg1 )
Sulphate (mg kg1 )
Potassium (mg kg1 )
Calcium (mg kg1 )
Magnesium (mg kg1 )
Sodium (mg kg1 )
1050 0.13a
1.3
2.93 0.03a
30.8 0.56a
)1.4 0.07a
5.0 0.25a
132 1b
85.2 2b
43.3 1a
680 20a
36.8 1.3a
2.8 0.1a
0.09 0.08a
4.1 0.2a
90 15a
455 24a
194 21a
3206 144a
50 16a
126 11a
34 5a
Values expressed as x standard deviation.
a
Replications ¼ 8.
b
Replications ¼ 2.
precipitation of calcium citrate were calculated from
Eqs. (1) and (2). For 0.1 dm3 of clarified juice, 3.61 and
3.15 g of CaCO3 and Ca(OH)2 were necessary, respectively
2H3 Cit þ 3CaCO3 ! Ca3 ðCitÞ2 þ 3H2 O þ 3CO2
ð1Þ
2H3 Cit þ 3CaðOHÞ2 ! Ca3 ðCitÞ2 þ 6H2 O
ð2Þ
The clarified juice was slowly added to the reagent
(CaCO3 or Ca(OH)2 ) until the pH reached 4.5. At this
pH, the addition of juice was stopped because the calcium citrate is soluble at lower pH values. Then, the
mixture was maintained for 24 h at 4 C, and the precipitated calcium citrate was separated by filtration with
Whatman paper No. 40 (retention of 8 lm).
Finally, the deacidified juice was mixed with clarified
fresh juice in order to adjust the pH to 4.0.
2.3. Deacidification by ion-exchange resins
The operating conditions of deacidification by ionexchange resins were described in a previous paper
(Vera, Dornier, et al., 2003). A column 1.7 cm in diameter and 60 cm long, with a bed volume of 50 cm3 was
used. Among the most suitable resins tested, the macroporous weakly basic resin, Amberlite IRA95 from
Rohm and Haas, was chosen to compare ion exchange
with others processes. The resin matrix consisted of
styrene divinylbenzene. The tertiary amine functional
groups were equilibrated in the OH form before use.
2.4. Deacidification by electrodialysis
Electrodialysis experiments were performed using a
laboratory cell of two stack designs:
• ED3C, the stack was equipped with homopolar membranes comprising three compartments, C1 and C2
being two distinct compartments in this case (Fig. 1),
• EDBM2C, the stack was equipped with homopolar
and two bipolar membranes comprising two compartments (Fig. 2).
Each ED unit had the same compartment spacing of
0.8 cm and the same electrode surface of 20 cm2 , corresponding to the effective area of each membrane.
The ion-exchange membranes were commercial ED
membranes, Selemion CMV (Asahi Glass) as cation
exchange membrane (CEM), Neosepta AXE01 (Tokuyama Co.) as anion exchange membrane (AEM), and
Neosepta BP-1 (Tokuyama Co.) as bipolar membrane
(BM).
The characteristics of solutions circulating in a batch
mode in all compartments are given in Table 2.
363
E. Vera et al. / Journal of Food Engineering 59 (2003) 361–367
Fig. 1. Configuration of the ED3C cell. C1 and C2 are separated
compartments, the two EC are connected (AEM and CEM: anion- and
cation-exchange membrane, respectively).
Fig. 2. Configuration of the EDBM2C cell. The two EC compartments
are connected (AEM and BM: anion-exchange and bipolar membrane,
respectively).
ED operations were carried out at constant current
density of 400 A m2 . Voltage, conductivity and pH
were monitored during all the experiments. ED operations were continued until pH 4.0 was reached in the
juice.
2.5. Analytical procedures
Amount of total soluble solids was determined by
refractometry with an Atago hand refractometer. Colour was measured with a Minolta CR-A70 colorimeter
in the L, a, b space. Colour variation was calculated
according to Eq. (3) using the non-treated sample as a
reference solution
DE ¼ ðDL2 þ Da2 þ Db2 Þ
1=2
ð3Þ
Acidity was assayed by titration with 0.1 N NaOH
and expressed as grams of citric acid or meq per kg of
juice. The cation and anion concentrations were measured by ICP and HPLC, respectively.
Sensory analyses were done by a panel of 15 semitrained members in a cabin under green light. Triangular
tests that are the most widely used of all the different
tests (Jellinek, 1985) were carried out to compare the
aroma of the fresh juice with the deacidified juices. To
avoid perturbations in the aroma perception, the pH of
deacidified juices was adjusted back to its initial value
with citric acid. Then, all the samples were prepared
with 18% w/w juice and 2% w/w sucrose, before testing.
Ranking tests in order of degree of preference were
directly done with the deacidified juices. The 19 panelists
received five random sorted samples of deacidified juice
treated by five different methods, and they were asked to
order the samples by their degree of preference. Then,
samples were ranked from 1 to 5, with 1 as the least liked
sample and 5 as the most liked sample. Final rank was
calculated as the addition of the rank given for each
sample by the 19 panelists.
3. Results and discussion
The juice acidity was mainly due to the presence of
citric acid (H3 Cit) which has three acid functions:
pK1 ¼ 3:1, pK2 ¼ 4:7, pK3 ¼ 6:4 (Merck Index, 1983).
The value of the initial pH (2:93 0:03) was close to the
pK1 of citric acid. Therefore, in the fresh juice, H3 Cit
and H2 Cit were the major species of citric acid, each
one being close to 50%.
The performances of each deacidification process
were compared in terms of characteristics of deacidified
juices such as concentration of inorganic and organic
ions, colour and flavour.
3.1. Deacidification by precipitation of calcium citrate
This method increased successfully the pH of the
juice. However, a major problem is the juice loss in the
precipitate which contained a moisture content of 70%
and 76% by using CaCO3 and Ca(OH)2 , respectively.
Table 2
Solutions and flow rates in the different compartments of the electrodialysis cells
Juice compartment
C1 compartment
C2 compartment
Electrode compartment
ED3C
Clarified juice
0.5 dm3
3.6 dm3 h1
0.1 N NaCl
2.5 dm3
3.6 dm3 h1
0.2 N NaOH
2.5 dm3
3.6 dm3 h1
0.1 N NaOH
1 dm3
10.2 dm3 h1
EDBM2C
Clarified juice
0.5 dm3
3.6 dm3 h1
0.5 N H3 Cit
2.5 dm3
3.6 dm3 h1
–
0.2 N Na2 SO4
1 dm3
10.2 dm3 h1
364
E. Vera et al. / Journal of Food Engineering 59 (2003) 361–367
That induced a juice loss of 19% and 27% with CaCO3
and Ca(OH)2 , respectively.
The use of CaCO3 induced a release of CO2 making
foam which hindered the mixture of CaCO3 and juice.
This explains the lower content of precipitate obtained
with the latter reagent, 23% compared to 32% obtained
with Ca(OH)2 .
Moreover, precipitation of calcium citrate requires an
accurate control of pH which must be maintained at
values higher than 4.5 during the experiment, the adjustment of pH to 4.0 being made by addition of fresh
juice. Consequently, stability and re-precipitation
problems could occur in the final product.
3.2. Deacidification by ion-exchange resins
Deacidification by ion-exchange resins was previously
studied (Vera, Dornier, et al., 2003). All the resins tested
gave similar results in terms of physico-chemical properties and sensorial analysis of deacidified juice. To
select the best resin, the production of deacidified juice
has been related to the soda consumption used for regeneration. Taking into account these considerations,
the most efficient resins were the Amberlite IRA95 and
Duolite A378. The results obtained with the Amberlite
IRA95 resin were used in the present paper to compare
the performances of ion exchange with other processes.
The resin used was in OH form and the ionexchange equilibrium was described by the following
equation:
OH þ H2 Cit () OH þ H2 Cit
ð4Þ
Consequently, the pH at the column outlet was initially higher than 6 and decreased with the volume of
juice treated. The experiment was stopped when pH 4.0
was reached.
3.3. Deacidification by electrodialysis
The principle of the two electrodialysis configurations
used is the extraction of citrate anions from the juice and
their replacement by hydroxyl ions provided either by
the NaOH solution in the C2 compartment (ED3C,
Fig. 1) or by the bipolar membrane sandwiched between
the juice and electrode compartments (EDBM2C,
Fig. 2). In the latter configuration, citric acid was
formed in the C1 compartment by citrate ions extracted
from the juice and protons provided by the second
bipolar membrane separating the C1 and electrode
compartments.
The ED3C configuration led to NaOH consumption
in the C2 compartment, while the EDBM2C configuration avoided the reagent consumption.
In the two cases, the cations in the juice were not
eliminated because the juice circulated between two
anion exchange membranes. On the other hand, all
organic and inorganic anions were partially extracted.
The performances of the two ED operations are
compared in Table 3. The EDBM2C configuration leads
to a higher energy consumption required by the water
splitting in the bipolar membranes. Note that these
values of energy consumption only aim to make comparison for two laboratory scale cells. But we can suppose that this classification will be confirmed in
industrial trials.
3.4. Comparison of the properties of juices deacidified by
various methods
Physico-chemical and sensorial analyses were performed on the juices deacidified by the various methods
investigated and compared to those of fresh juice.
3.4.1. Physico-chemical properties of deacidified juices
Table 4 shows that titrable acidity and total soluble
solids varied up to 30% according to the deacidification
method applied. Resin treatment induced the greatest
change in total soluble solids and colour.
Fig. 3 shows the variation in anion concentration. In
all the deacidified juices, a decrease in the citrate and
malate concentrations was obtained. The precipitation
of calcium citrate is the most selective technique where
citrate ions were preferentially eliminated, the chloride,
sulphate and phosphate concentration remaining unchanged. Nevertheless, the use of CaCO3 induced a
lower elimination of citrate and malate ions, 40% and
12.5%, respectively, than with Ca(OH)2 where it attained 65% and 32%, respectively. This result is in
agreement with the lower content of precipitate obtained
with CaCO3 than with Ca(OH)2 , as previously noticed.
The gap between the extraction ratios of organic
anions was lower with the other methods, about 65%
and 53% for citrate and malate, respectively. Moreover,
the inorganic anions were partially extracted together
with organic ions when using resin and electrodialysis.
Table 3
Comparison of performances of the two electrodialysis configurations tested
Time to obtain pH 4.0 in
the juice (min)
ED3C
EDBM2C
496
549
Current efficiency for citric
acid (%)
Deacidification rate
(eq h1 m2 )
Energy consumption
(kWh dm3 of juice)
33
31
4.9
4.6
0.38
0.50
365
E. Vera et al. / Journal of Food Engineering 59 (2003) 361–367
Table 4
Physico-chemical analysis of the juices deacidified by different methods
Deacidification
method
CaCO3
Ca(OH)2
IRA95
ED3C
EDBM2C
pH
Titrable aciditya
(g citric acid per kg)
4.0
4.0
4.0
4.0
4.0
d
13.5
10.4e
10.8e
12.3f
12.0f
1
(meq kg )
d
210
160e
170e
190f
190f
Total soluble solidsb
(g kg1 )
d
130
112e
100f
111e
111e
Colourc
L
a
d
30.90
31.00d
31.66d
30.33d
31.26d
DE
b
d
)1.50
)1.46de
)1.49d
)1.26f
)1.30ef
d
5.14
4.94d
4.09e
4.62de
4.94d
0.2
0.2
1.5
0.5
0.4
Maximum error on the measure: a 2%, b 1%, c 5%.
d;e;f
Homogeneous groups determined by one-way ANOVA analysis with 95% of confidence.
Fig. 3. Anions concentrations variation.
In the latter method, the elimination of anions is related
to their mobility both in the solutions and membrane,
and consequently the extraction of inorganic anions is
easier than that of organic anions. Almost all the chloride ions were eliminated and about 90% and 70% of
phosphate and sulphate ions were extracted, respectively.
Fig. 4 shows the variation in cation concentration.
The potassium and magnesium concentrations were not
affected whatever the methods used, while the calcium
and sodium concentrations were significantly increased
by using the precipitation method and electrodialysis
with three compartments, respectively.
The calcium amount in the juice reached 48% and
25% of cations after using CaCO3 and Ca(OH)2 , respectively, compared to 1% in the untreated juice. This
increase was due to an incomplete precipitation of added
calcium ions.
The sodium concentration was increased from 1% to
9% by ED3C treatment because of NaOH and NaCl
dialysis arising from adjacent compartments. Indeed, it
is well known that a concentration gradient between
ion-exchange membranes induces a dialysis phenomenon.
There was no change in the cation concentration by
using resins and electrodialysis with bipolar membranes.
To conclude, the method based on precipitation of
calcium citrate leads to a selective decrease in the citrate
concentration and an increase in the calcium one, the
use of Ca(OH)2 giving better performances than CaCO3 .
Ion exchange with resins and electrodialysis induce an
extraction of inorganic anions together with organic
Fig. 4. Cations concentrations variation.
366
E. Vera et al. / Journal of Food Engineering 59 (2003) 361–367
Table 5
Results of triangular test between treated juices and fresh juice
a
Number of tests
Number of good answers
Minimum number of good answers required to find
a difference between the samples compared
a
CaCO3
Ca(OH)2
IRA95
ED3C
EDMB2C
12
3/12
8/12
15
7/15
9/15
11
4/11
7/11
14
8/14
9/14
12
6/12
8/12
Replications ¼ 1.
Fig. 5. Degree of preference with passion juices deacidified by different
methods (a 6 0:05).
ones, without change in the cation concentration, except
for electrodialysis with homopolar membranes where
the sodium concentration is significantly increased.
3.4.2. Sensory properties of deacidified juices
The triangular tests did not highlight significant differences between the deacidified juices and the fresh
juice (Table 5).
To classify the deacidified juices, a ranking test in
order of degree of preference was done by directly
tasting the deacidified juices. The results of rank sums
required for significance at the 5% level (a 6 0:05) were
depicted by Fig. 5. The rank position of samples within
the range delimited by dashed lines is not significant
(Jellinek, 1985). Nevertheless, one can conclude that the
juice treated by ion-exchange resin was highly significant
as the sample liked least, the precipitation method with
calcium hydroxide giving the most liked sample by a
narrow margin.
Besides, citric acid could be recovered as a by-product
by dissolving the precipitated calcium citrate in sulphuric acid, and converting the calcium citrate into
calcium sulphate and citric acid (Milsom & Meers,
1985). However, the application of the precipitation
method has two limitations: the legislation of some
countries and the fact that consumers prefer natural
products, without chemical addition. It is also necessary
to take into account that this method induces an increase in the calcium concentration and could cause
some precipitation problems in the final product.
The ion-exchange process does not seem to be a good
option, because of changes in the organoleptic characteristics of juice treated and the large amounts of effluent
produced during the regeneration phase of resins.
The deacidification by electrodialysis has some advantages over the above methods, especially electrodialysis with bipolar membrane which is a continuous
process without added reagents, giving a high quality
juice in terms of physico-chemical and sensorial analyses. It was observed that there was (i) no change in the
cation concentration, (ii) slight colour variation and (iii)
good flavour. Only organic and inorganic anions were
partially eliminated. Nevertheless, electrodialysis is
more expensive than the other techniques but the citric
acid simultaneously produced during the deacidification
could improve the cost of this technique. Therefore, ED
on BM could be a promising alternative to the conventional calcium precipitation process for the deacidification of the passion fruit juice. Further studies are
necessary under conditions similar to the industrial ones
which are more favourable than those carried out in a
laboratory cell.
Acknowledgements
4. Conclusion
All the methods tested allowed the deacidification of
the clarified passion fruit juice in which pH was increased from 2.9 to 4.0. Nevertheless these methods have
some different advantages and disadvantages.
The deacidification by precipitation using CaCO3 is
not recommended, because problems induced by the
liberation of CO2 , foam making and poor pH control,
occurred during the precipitation phase. The use of
Ca(OH)2 is easier and gives a good quality product.
The authors thank the French Embassy in Quito and
IPICS (Sweden) for their financial support.
References
Adhikary, S. K., Harkare, W. P., Govindan, K. P., & Nanjundaswamy, A. M. (1983). Deacidification of fruit juices by electrodialysis. Indian Journal of Technology, 21, 120–123.
Bhatia, A. R., Dang, R. L., & Gaur, G. S. (1979). Deacidification of
apple juice by ion exchange resins. Indian Food Packer (January–
February), 15–19.
E. Vera et al. / Journal of Food Engineering 59 (2003) 361–367
Couture, R., & Rouseff, R. (1992). Debittering and deacidifying sour
orange (Citrus aurantium) juice using neutral and anion exchange
resins. Journal of Food Science, 57(2), 380–384.
Goloubev, V. N., & Salem, B. (1989). Traitement a lÕelectrodialyse du
jus dÕorange. Actualites des Industries Alimentaires et Agro-industrielles (France), 106(3), 175–177.
Jellinek, G. (1985). Sensory evaluation of food. Theory and practice.
Chichester, England: Ellis Horwood Ltd, pp. 184–288.
Johnson, R. L., & Chandler, B. V. (1985). Ion exchange and adsorbent
resins for removal of acids and bitter principles from citrus juices.
Journal of Science and Food Agriculture, 36(6), 480–484.
Johnson, R. L., & Chandler, B. V. (1986). Debittering and de-acidification of fruit juices. Food Technology in Australia, 38(7), 294–297.
Loeillet, D. (1999). Jus de fruit de la passion. Un modele dÕinstabilite.
Fruitrop, 56, 8–15.
Lue, S. J., & Chiang, B. H. (1989). Deacidification of passion fruit juice
by ultrafiltration and ion-exchange processes. International Journal
of Food Science and Technology, 24, 395–401.
367
Merck Index (1983). Merck & Co., Inc., Rahway, NJ, USA.
Milsom, P. E., & Meers, J. L. (1985). Citric acid. In H. W. Blanch, S.
Drew, & D. I. C. Wang (Eds.), Comprehensive biotechnology. The
practice of biotechnology: Current commodity products: Vol. 3 (pp.
665–680). Oxford: Pergamon Press.
Scott, K. (1995). Handbook of industrial membranes. Amsterdam:
Elsevier Advance Technology, pp. 760–770.
Serna, J., & Chacon, C. (1988). El cultivo de maracuya (3 ed). Bogota:
Federaci
on Nacional de Cafeteros de Colombia, Colombia
Editolasers.
Vera, E., Dornier, M., Ruales, J., Reynes, M., & Vaillant, F. (2003).
Comparison between different ion exchange resins for the deacidification of passion fruit juice. Journal of Food Engineering, 57(2),
199–207.
Vera, E., Ruales, J., Dornier, M., Sandeaux, J., Sandeaux, R., &
Pourcelly, G. (2003). Deacidification of clarified passion fruit juice
using different configurations of electrodialysis. Journal of Chemical Technology and Biotechnology, in press.