Anda di halaman 1dari 9

Journal of Food Engineering 88 (2008) 429437

Contents lists available at ScienceDirect

Journal of Food Engineering


journal homepage: www.elsevier.com/locate/jfoodeng

Review

Convective drying characteristics of Amelie mango


(Mangifera Indica L. cv. Amelie) with correction for shrinkage
A.O. Dissa a,*, H. Desmorieux b, J. Bathiebo a, J. Koulidiati a
a
Laboratoire de Physique et de Chimie de lEnvironnement (LPCE), Unit de Formation et de Recherche en Sciences Exactes et Applique (UFR/SEA),
Universit de Ouagadougou, BP7021, Burkina Faso
b
Laboratoire dAutomatisme et de Gnie des Procds (LAGEP), UCBL1-CNRS UMR 5007-CPE Lyon, Bt.308G, 43 bd du 11 Nov. 1918 Villeurbanne,
Universit Claude Bernard Lyon 1, Lyon, France

a r t i c l e

i n f o

Article history:
Received 11 December 2006
Received in revised form 8 March 2008
Accepted 10 March 2008
Available online 16 March 2008
Keywords:
Mango
Amelie
Shrinkage
Drying kinetics
Diffusivity

a b s t r a c t
Drying characteristics of Amelie mango were investigated through slices shrinkage behaviour at 50, 60
and 70 C and drying kinetics at temperatures ranging from 50 to 70 C and for slices of 2.5, 5 and
10 mm thick. Results showed that mango slices shrinkage was less inuenced by temperature and best
tted by the linear model of volume additivity. The density of solid matrix and the linear shrinkage coefcient were evaluated respectively at 1468 kg/m3 and 1.468. The tting of experimental data with two
models (Page and Henderson & Pabis) showed that drying curves were best described by the Page model.
Both water diffusivity values and drying rates curves were strongly inuenced by shrinkage. Drying rates
were underestimated and diffusivities overestimated when drying data are not corrected for shrinkage.
An Arrhenius dependence type of diffusivity was established from drying data corrected for shrinkage
and the activation energy deduced was 25.355 kJ/mol.
2008 Elsevier Ltd. All rights reserved.

Contents
1.
2.

3.

4.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Materials and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.
The mangoes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.
Maturity measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1.
Total soluble solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.2.
Acidity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.
Bulk shrinkage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.
Drying kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.
Shrinkage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.
Drying curves characterisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1.
Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.2.
Effective water diffusivity determination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3.
Calculation of equilibrium water content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.4.
Drying rate calculation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.5.
Drying rate curve correction for shrinkage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3.
Fitting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.
Maturity measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2.
Shrinkage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.
Drying kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.1.
Drying rates curves corrected for shrinkage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.2.
Influence of air temperature and slice thickness on drying kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

* Corresponding author. Tel.: +226 76 58 29 88.


E-mail addresses: dissa@lagep.univ-lyon1.fr, alfa_dissa@univ-ouaga.bf (A.O. Dissa).
0260-8774/$ - see front matter 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2008.03.008

430
430
430
430
430
430
430
431
431
431
432
432
432
432
433
433
433
433
433
434
434
434
435

430

5.

A.O. Dissa et al. / Journal of Food Engineering 88 (2008) 429437

4.3.3.
Kinetics models evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3.4.
Diffusivity and activation energy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

436
436
436
437
437

1. Introduction

2.2. Maturity measurement

Mango plantations in West Africa are much diversied and the


main fruit varieties composing orchards are Amelie, Brooks, Lippens, Keitt, Mangot, Kent and Smith. However, a substantial quantity of harvests is lost each season because of inadequate means of
preservation. Facing this problem, drying can constitute an efcient
solution of preservation. For mango fruit, the mode of drying used
in industry and traditionally is convective drying and Amelie is
the most commonly variety used for drying and exportation in this
region. The investigation of this variety drying characteristics is
then necessary for a better understanding of its drying process
and for having dried fruits which respect the standards of external
markets. Thus, during these last years, certain research related to
the drying of mango. Nieto et al. (2001) studied the inuence of
pre-treatments by blanching and/or osmotic dehydration with glucose syrups on air drying kinetics of mango during the falling rate
period. Goyal et al. (2006) studied the thin-layer drying kinetics
of raw mango slices at 55, 60 and 65 C air temperature with control, blanching and blanching in 1% potassium metabisulphide solution as pre-treatments. When studying drying of mango pulp in
vacuum, Jaya and Das (2003) developed an exponential correlation
which allowed determining mango diffusivity as a function of initial thickness and drying medium temperature. Tour and Kibangu-Nkembo (2004) studied the free convection sun-drying of
cassava, banana and mango and established an expression linking
slices initial moisture content to the maximal temperature difference between drying air and each product. Most of these works
did not deal with the inuence of the shrinkage on the drying kinetics of mango whereas that shrinkage could strongly affect the water
diffusion during drying. Therefore, the objective of the current work
was to investigate the shrinkage behaviours of mango slices, to observe the effect of drying temperature and slices thickness on drying characteristics of mango, to evaluate the inuence of shrinkage
on drying rates and water diffusion and to establish a suitable water
diffusion law with correction for shrinkage. This study thus contributed to mango dryer setting and dried mango preservation process.

The maturity of mangoes was evaluated only for the last series
of mangoes brought on April 2007 to France. However, these mangoes had the same ngering-maturity as those of the two rst series. The maturity index was evaluated by the ratio of total soluble
solids (in Brix) and acidity (in mmol/g) as:

2. Materials and methods


2.1. The mangoes
The Amelie variety used in this study was purchased from a local fruit market of Bobo Dioulasso, town in the west of Burkina
Faso. These mangoes belong to three series of fruit purchased from
a market of Burkina Faso and transported to France on the plane
traveller compartment (rst series on September 26th 2005, 2nd
on September 3rd 2006 and the third on April 13th 2007). During
their transport from Burkina to France, the fruits were preserved in
an ambient temperature of 20 C. These fruits were enough unripe
and did not ripen a lot during the 8 h of travelling. At the arrival,
the measures were made between the rst and seventh day after
the fruits transport. For measures, good quality fruits were selected and washed into water to which a small quantity of Natrium
hypochlorite was added as disinfectant, rinsed with drinking water
and peeled. The pulp was separated from the stone and sliced
through according to the desired thickness.

Imaturity

total soluble solids  Brix




Acidity mmol
g

2.2.1. Total soluble solids


The total soluble solids of mango samples was measured with a
hand refractometer Master-a (ATAGO P-1, 033 Brix, Japan) initially calibrated with a 20 Brix saccharose solution. To verify the
measure repeatability, three samples were taken once only among
the fruit in the same zone as those used for kinetics, shrinkage and
isotherms measurements. For the measures, a few drops of mango
juice obtained by pressing a sample were placed in a tissue and put
down on the refractometer. The total soluble solids was thus measured and obtained in Brix.
2.2.2. Acidity
A 0.01 mol/L NaOH (sodium hydroxide) solution added with a
0.05% phenolphthalein solution allowed measuring the sample
acidity. Before each measure, the sample was taken from the fruit
and ground in a mixer (MOULINEX, 140W, France) to have pulp of
mango. Three successive acido-basic titrants were obtained for
that pulp. From the sodium hydroxide volume used, the number
of moles of total acid in the mass of mango pulp is deduced from
the matter quantity preservation equation. The fruit acidity is then
dened as the ratio of millimoles of hydronium ion per mass of
fresh pulp used in grams.
2.3. Bulk shrinkage
Parallelepipedic samples of Amelie mango were used to establish the mango shrinkage curves. The initial mean dimensions of
the samples used during the experiments were presented in Table
1. To measure samples shrinkage, the product was dried at 50, 60
or 70 C in an oven-dryer (WTF BINDER). During drying, the weight
of the sample was regularly measured with a balance (SARTORIUS,
0.001 g precision, France) every 10 min during the rst 110 min.
Then, because of the drying rate decrease, measures were taken
every 15 min until 300 min drying and then every 30 min until
the end of drying, determined by constant weight. During drying,
the dimensions of slices were measured using a micrometer Mitu-

Table 1
Initial dimensions of samples used for shrinkage measurement
Temperature (C)

Maturity

Length (m)

Width (cm)

Thickness (cm)

50
60
70

Ripe
Ripe
Ripe

5.5
5.6
6

3.9
4.40
3.2

1.2
1.26
1.3

431

A.O. Dissa et al. / Journal of Food Engineering 88 (2008) 429437

Nomenclature
D
d
D0
DR
Ea
K
L
l
MR
m
Qv
qv
q0v
R
RH
R2
Sb
T
t
V
Va

diffusion coefcient (m2 s1)


thickness of the sample (m)
Arrhenius pre exponential factor (m2 s1)
drying rate (kg kg1 s1)
activation energy (J mol1)
drying constant (s1)
length of the slab (m)
half thickness of the sample (m)
moisture ratio
weight (kg)
rate of mass loss (kg s1)
mass ux per unit of initial surface area (kg s1 m2)
mass ux per unit of current surface area (kg s1 m2)
perfect gas constant, R = 8.3145 J mol1 K1
relative humidity (%)
coefcient of determination
shrinkage coefcient
temperature (C, K)
drying time (s)
volume (m3)
air velocity (m s1)

toyo (Japan) of 2  105 m precision. These dimensions were measured at several places of the piece of mango and their mean values
were considered. It was noticed that the slices initial form was
approximately conserved. The results of shrinkage measurements
obtained are the means of three trials at every temperature. For
each measure, the corresponding weights were noted. An ovendryer and a balance (SARTORIUS, France) of 0.001 g precision were
used to determine samplesmoisture content. The dry mass was
evaluated by drying samples at 105 C in the oven-dryer during
24 h (Nguyen et al., 2004). The shrinkage coefcient was then deduced from experimental data according to the following formula:
Sb

V
V0

where V is the sample volume at a time t of drying and V0 the initial


volume of the sample.
Assuming that mango slices form is one of a parallelepiped, their
volume was calculated as length  width  thickness. Due to the
similarity between the three relative dimensions variation d/d0,
W/W0, L/L0 for X/X0 > 0.2 as showed on Fig. 2a, the shrinkage of mangoes is considered isotropic. Changes in Sb with respect to the moisture content (reported to its initial value) constitute the shrinkage
curve. The experimental shrinkage curves obtained are then simulated with the fundamental linear model of volume additivity.

Load Unit

Basket
Control Screen
Trays

Fan

Distilled Water
Heater

Heater

dry basis moisture content (kgw kgdm )


width of the sample (m)

Greek letters
q
density (kg m3)
c
samples shape factor
Subscripts
a
air
cr
critical
d.b.
dry basis
dm
dry mass
eff
effective
eq
equilibrium
exp
experimental
m
model
0
initial
s
solid
v
vapor
w
water
w.b.
wet basis

2.4. Drying kinetics


Drying Kinetics of Amelie mango slices were determined using
a laboratory dryer (climatic chamber Votsch Industrietechnik,
Germany) in the Process Engineering and Automatic Laboratory
(LAGEP UMR 5007 CNRS) of Claude Bernard Lyon 1 University.
The schematic diagram of the dryer is shown in Fig. 1. The samples were dried in a perforated square basket, which had a ow
cross-section of 10 cm  10 cm. They were weighted at regular
intervals with a Master Pro SARTORIUS precision balance coupled
to the chamber. Weight changes with respect to time were noted
at different temperatures (50, 60 and 70 C) and for different slice
thicknesses (2.5, 5 and 10 mm) at 70 C. The dry mass was determined by drying samples at 105 C in an oven-dryer during 24 h
according to Nguyen et al. (2004). Deduced drying curves were
simulated by the empirical models of Page and Henderson &
Pabis.

3. Theory
3.1. Shrinkage
Several previous works expressed the shrinkage coefcient Sb of
food products during drying. Several models were so developed by
Suzuki et al. (1976) and Lozano et al. (1980). In the current study,
the inuence of temperature on Amelie mango shrinkage was
investigated. Experimental measures of ripe mango shrinkage were
established and simulated using the fundamental linear model of
volumes additivity. This model is based on the hypothesis of the
additivity of volumes of water and solids, assuming that the volume of air in pores is negligible and that cellular tissues are incompressible. This last hypothesis has been used for fruit and
vegetables by several authors: Desmorieux (1992), Iglesias et al.
(1993), Leonardo and Dermeval (2004). If V is the total volume of
the sample and ms the mass of solid matrix, assuming a negligible
volume of air, the volume versus the moisture content on dry basic
X of the product is expressed as:

Balance Screen

V ms
Fig. 1. Schematic illustration of the experimental dryer set-up.

1

X
W

1
X

qs qw


3

432

A.O. Dissa et al. / Journal of Food Engineering 88 (2008) 429437

where qs is the density of solid matrix and qw the density of water.


From Eqs. (2) and (3), the shrinkage coefcient can be stated as:
Sb

qw qs X
qw qs X 0

Hence
Sb

 
1
uX 0
X

1 uX 0 1 uX 0 X 0

With u qqws the linear shrinkage coefcient and X0 the initial moisture content. Eq. (5) assumes that the shrinkage coefcient varies
linearly with the moisture content of the product as follows:
 
X
0
Sb b a0
6
X0
uX 0
7
where a0
1 uX 0
1
0
8
b
1 uX 0
The ratio of the slope a0 and the origin ordinate b0 of that linear
shrinkage curve (represented by straight-line of Sb versus XX0 ) allows
identifying the true values of the density of solid matrix qs and the
linear shrinkage coefcient u by the following formulae:
qs

qw a0
X 0 b0

a0
0
b X0

10

These expressions give an estimate of the density of solid matrix and the linear shrinkage coefcient which are difcult to
determine through direct physical measurements. The density of
water used is qw = 1000 kg m3.
3.2. Drying curves characterisation

3.2.2. Effective water diffusivity determination


The water diffusivity of mangoes is evaluated by using the simplied mathematical Ficks second law. Assuming one-dimensional
moisture transfer, homogenous and parallelepipedic shape for the
mangoes samples, uniform initial moisture distribution and a nonshrinking slab, the analytical solution of Ficks equation is (Crank,
1975):


1
2
X  X eq
8 X
1
2 p Dt
2
exp
2n

1
14
X cr  X eq p n0 2n 12
4 l2
where Xcr is the critical water content, l the half thickness of the
slice, D the water diffusivity, n the Fouriers series number and c
the shape factor equal to p82 for a parallelepiped.
In convective drying of solids, this solution is valid only for
the falling rate period when the drying process is controlled by
internal moisture diffusion for slice moisture content below the
critical value. Therefore, the diffusivity must be identied from
Eq. (14) by setting the initial moisture content to the critical
value Xcr and by setting the drying time to zero when the mean
moisture content of the sample reaches that critical moisture
content.
To take shrinkage into account in Eq. (14), the sample thickness must be considered as a function of time and l(t) is so
obtained from shrinkage measurements. Eq. (14) becomes (Crank,
1975):
"
#
1
2
X  X eq
8 X
1
Dt
2p
15

exp 2n 1
X cr  X eq p2 n0 2n 12
4 lt2
For a long diffusion time corresponding in our case to MRc < 0.35,
Eq. (15) becomes(Crank, 1975):
"
#
8
p2 Dt
16
MRcr 2 exp 
p
4 lt2
With MRcr

3.2.1. Simulation
Drying curves were simulated using two empirical models of reduced moisture content: Pages and Henderson & Pabis models.
These two empirical models coming from the fundamental diffusion model are generally suitable for fruits. The Page model has
been successfully used for the drying characteristics description
of many products such as carrots, okra and gs (Doymaz,
2004a,b, 2005), eggplant (Ertekin and Yaldiz, 2004) and fresh and
treated Papaya (Anoar et al., 2003). This model is stated as follow
(Page, 1949):
y

MR expkt

11

With
MR

X  X eq
X 0  X eq

12

where MR is moisture ratio, k and y the parameters of Page model,


X the average moisture content of the product, X0 the initial moisture content, Xeq the equilibrium moisture content and t the drying
time.
The Henderson & Pabis model is also an exponential model. This
model is obtained by simplication of the rst term of the series
solution of Ficks second law. It has been used by several authors
such as Simal et al. (2005) for kiwi fruits and by Doymaz (2004b)
for mulberries. This model effectively predicts the drying rate at
the beginning of the drying process but appears sometimes to be
less efcient for the last stages of the process. It is expressed as
(Henderson and Pabis, 1961):
MR a expbt

13

where a and b are the parameters of the Henderson & Pabis model.

X  X eq
X cr  X eq

17

The water diffusivity is so evaluated using slopes method. Its


values are typically determined by plotting experimental drying
data in terms of ln(MRcr) versus t/[l(t)]2 data. Theses values identied for different temperatures: 40, 50, 60 and 70 for slices of 5 mm
initial thickness are deduced from the slopes of ln(MRcr) vs. t/[l(t)]2
straight-line. Activation energy was thus identied from diffusivity
according to Arrhenius dependence as:


Ea
18
D D0 exp
RT 273:15
where R is the perfect gas constant (R = 8.3145 J mol1 K1), D0 the
Arrhenius factor (m2 s1), T the temperature (C) and Ea the activation energy for internal mass transfer (J mol1).
In order to analyse the inuence of shrinkage on water diffusivity of Amelie mango, values of diffusivities for a non-shrinking
sample and for a shrinking sample were evaluated and compared.
3.2.3. Calculation of equilibrium water content
The equilibrium moisture content (Xeq) between the mango
sample and the drying air was expressed by Eq. (19) deduced
from the Modied Henderson model according to Dissa (2007).
This equation resulted from the tting of experimental data of
Amelie mango desorption isotherm at 40 C (R2 = 0.9982, RMSE =
6.4598  104). The modied Henderson model was used to take
drying temperature variation into account. The use of the parameters of this model for the whole drying air temperature range was
motivated by the results of Myara and Sablanis works (2001)
which showed that mango isotherm does not vary much with temperature in 1560 C range:

433

A.O. Dissa et al. / Journal of Food Engineering 88 (2008) 429437

1=0:3316
 ln1  RH
0:0193T 44:36

19

3.2.4. Drying rate calculation


The drying rate at different instants of drying was estimated by
derivation of the moisture content X (db) with respect to time
using a derivation program adapted to the software MATLAB version 6.1. The obtained drying rate curves were smoothed using a
Microsoft Excel macro-program. This macro initially tted series
by an appropriate polynomial function and replaced each value
of the series by a medium value in such a way that standard deviations from initial values are minimised.
3.2.5. Drying rate curve correction for shrinkage
While evaluating diffusivity, the critical moisture content
identication is possible only if a constant drying rate period
exists. This period is generally absent during drying of most
fruits and vegetables. However, it can sometimes be exhibited
a constant drying ux stage (mass per unit of area and time)
from a falling rate period (mass per unit of time) provided we
are able to determine and account for the sample exchange surface shrinkage. The drying rate is usually represented by the
mass loss of the product Qv (kg s1) per unit of dry mass. The
mass ux qv (kg s1 m2) can be obtained by dividing this mass
loss by the entire exchange surface area as (May and Perr,
2002):
qv

Qv
ms dX

S
S dt

20

where X is the moisture content on dry basis, ms the dry mass


and S the exchange surface area calculated from shrinkage
data.
Assuming that the shrinkage is isotropic, the bulk shrinkage is a
function of the samples current surface as:
Sb

S d
S 0 d0

21

where S0 is the sample initial surface area, d0 and d initial and current thicknesses of the sample.
With

d
Sb 1=3
d0

imental values, the degree of compatibility of each tting model


was evaluated by the values:
of the coefcient of determination (R2) and reduced chi square
v2 for the kinetics,
of the coefcient of determination (R2) and standard error given
by the root mean square errors (RMSE) for shrinkage.
The best t to the studied physical parameters were given for R2
values closer to 1 and standard error and v2 values closer to 0.
4. Results and discussion
4.1. Maturity measurement
For the last series of mangoes, the maturity of the fruit was
evaluated by Brix and acidity measurements. These two chemical
magnitudes for the whole fruits used were respectively 14.0
1.7 Brix and 8.9 3.1 mmol/100 gpulp. The maturity index represented by the Brix-acid ratio was estimated at 157.30
54.84 Brix g/mmolacid for the whole mangoes used. The rate of
dry mass was also evaluated at 13.66 g/100 gpulp.

1.0

Current dimension / initial dimension

X eq

0.8
0.7
0.6
0.5
L/L0
W/W0

0.4

d/d0

0.3
0.0

0.1

0.2

0.3

0.4

22

0.5
X/X0

0.6

0.7

0.8

0.9

1.0

Fig. 2a. Length, width and thickness shrinkage proles of Amelie mango slices at
50 C.

Hence
2=3

S S0  Sb

23

The mass ux and the mass loss of the product corrected for
shrinkage can then be expressed by the following equations:

0.9

24a

dX
dt

0.8

24b

Thus, in certain cases, the corrected drying curve could allow


exhibiting a constant drying ux period and identifying the critical
moisture content. From Eqs. (24a) and (24b), the drying rate corrected for shrinkage deduced from the mass loss corrected for
shrinkage is:
S2=3
b

25

0.7
0.6

V/V0

ms 2=3 dX
S
dt
S0 b
Q 0v S0 q0v

q0v 

DR

0.9

0.5
0.4
0.3
50C
60C
70C

0.2
0.1

3.3. Fitting
The different ttings were obtained through a regression program in the MATLAB software version 6.1. In order to nd the
models of drying kinetics and shrinkage that adjust the best exper-

Additivity linear model

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

X/X0
Fig. 2b. Inuence of temperature on the bulk shrinkage of Amelie mango slices:
experimental values and simulation by additivity linear model.

A.O. Dissa et al. / Journal of Food Engineering 88 (2008) 429437

4.2. Shrinkage
Shrinkage curves of parallelepipedic mango according to length,
width and thickness are showed on Fig. 2a. The similarity between
the shrinkage proles of the three sides allowed assuming Amelie
mango shrinkage to be isotropic. Slices bulk shrinkage coefcient
variations with reduced moisture content at 50, 60 and 70 C are
then presented in Fig. 2b. Three samples were used for each temperature and the average shrinkage values were presented. The
shrinkage curves obtained were overall linear and almost the same
at 50, 60 and 70 C. Therefore, it could be deduced that ripe Amelie
mangoes shrinkage is not much inuenced by drying temperatures. The experimental shrinkage curves were not strictly linear
but they could be tted by the fundamental linear model of
volumes additivity. The determination coefcients and standard
errors of experimental values tting showed in Table 2 were
respectively higher than 0.96 and 0.0004 for the three temperatures. Thus, the linear model of volumes additivity may be assumed to represent the bulk shrinkage mechanism of Amelie
mangoes at 50, 60 and 70 C with acceptable errors. This linear
model of volume additivity is based on physical assumptions and
supposes that the evaporated water volume during drying is
replaced by the same volume of shrinkage. The dried mango obtained is therefore a compact product with a low porosity. From
the slope of shrinkage straight-line, the intrinsic density of solid
matrix and the linear coefcient of shrinkage were respectively
evaluated at 1468 kg/m3 and 1.468. These results are in accordance
with the order of magnitudes given by Lozano et al. (1983), who
estimated the density of solid matrix for fruits and vegetables to
vary between 1300 and 1550 kg m3. For water diffusivity determination with correction for shrinkage, an empirical model of slice

thickness change was established in order to take shrinkage into


account in Crank (1975) solution. According to Fig. 2c the thickness
was assumed to decrease linearly with reduced moisture content.
The parameters a and b of the linear correlation established from
thickness measures are presented in Table 2. These parameters
were evaluated with standard errors lower than 0.0008.
4.3. Drying kinetics
4.3.1. Drying rates curves corrected for shrinkage
Assuming that samples maintain their parallelepipedic shape as
the drying progresses (assumption supported by the experimental
data for most of the drying process, i.e. X/X0 > 0.2) and considering
a linear and isotropic shrinkage, the mango slice bulk shrinkage is
expressed according to Eq. (4) and Section 4.2 as:
Sb

1 1:4687X
1 1:4687X 0

0.8
0.7
0.6
0.5

RH=15% d=5mm Va=0.8m/s


0.4
0.3

Bulk shrinkage

Thickness linear shrinkage

0.2

a = 0.9547
b = 0.0491

a = 0.4436
b = 0.5516

0.1

R2

50 C
60 C
70 C

RMSE (Standard error)

0.9846
0.9692
0.9920

4

4.5523  10
4.9229  104
1.7530  104

R2

50C Experimental data


60C Experimental data
70C Experimental data
50C Page model
60C Page model
70C Page model

0.9

Table 2
Mango slices shrinkage: experimental data tting at 50, 60 and 70 C with linear
additivity model

Ta (C)

26

According to Eqs. (24a) and (26), the mass ux per unit of current
exchange surface area can be stated as:

MR=(X-Xeq)/(Xo-Xeq)

434

RMSE (Standard error)

0.5

1.5

0.9181
0.9718
0.9782

2.5

7.8573  10
1.8407  104
2.3734  10 4

x 10

Time (s)

4

Fig. 3a. Inuence of temperature on the drying kinetics of Amelie mango slices and
simulation by Page model.

0.0025
0.00225

0.9

0.002

Drying rate (kgs-1m-2)

0.8

d/d0

0.7
0.6
0.5

70C

0.00175
0.0015
0.00125
60C
50C

0.001
0.00075
0.0005

50C
60C
70C

0.4
0.3

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

per unit of current surface area (with shrinkage)


per unit of initial surface area (without shrinkage)

0.00025
0

X/X0
Fig. 2c. Inuence of temperature on the thickness linear shrinkage of Amelie
mango slices.

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.1

X/X0
Fig. 3b. Inuence of temperature on the drying rate per exchange surface area of
Amelie mango slices of 5 mm thick.

435

A.O. Dissa et al. / Journal of Food Engineering 88 (2008) 429437

q0v 


2=3
ms 1 1:4687X
dX
dt
S0 1 1:4687X 0

27

Drying rates curves per unit of current surface area presented


on Figs. 3a and 4b were obtained from Eq. (27) by using experimental water contents. The critical water contents of Table 4 were
deduced from these curves. Evaluated critical water contents have
allowed calculation and comparison of diffusivities with taking
shrinkage into account and without taking shrinkage into account
in the processing of drying kinetics during the falling drying rate
period.
4.3.2. Inuence of air temperature and slice thickness on drying
kinetics
Moisture Ratio and drying rates curves (per unit of current surface area and per unit of initial surface area) for three temperatures
(50, 60 and 70 C) and for three thicknesses (2.5, 5, 10 mm) are presented respectively on Figs. 3a, 3b, 4a and 4b. The experiments
were carried out using a drying air of 15% relative humidity and
0.8 m s1 velocity. These characteristics of drying air were sufcient to reach equilibrium water contents corresponding to water
activities lower than 0.6, the typical preservation water activity
of dried mango according to several authors (Dissa, 2007; Pott
et al., 2005; Desmarais and Marcotte, 2002). It can be observed that
most of the drying takes place during the falling drying rate period.
The Figs. 3b and 4b show that when Amelie mangoes drying rates
curves are not corrected for shrinkage, almost no constant rate drying period exists. Moreover, drying rates obtained after correction
for shrinkage are always higher than those non-corrected. It can be
concluded that when shrinkage is not considered in drying data
processing, Amelie mango drying rates per unit of exchange surface area (mass ux) are very under-estimated. The corrected dry-

ing rates curves allowed identifying critical moisture contents


presented on Table 4 for many drying conditions. It is easily
remarkable on this table that these critical values depend on the
drying operating conditions.
The inuence of temperature on the moisture ratio and drying
rate curves per unit of exchange surface area of 5 mm thick Amelie
mango slices at 50, 60 and 70 C is illustrated in Figs. 3a and 3b.
The time required to reduce the moisture ratio to any given level
in Fig. 3a depended on the drying temperature, being highest at
50 C and lowest at 70 C. This tendency exists for both corrected
drying curve and non-corrected. Thus, for a 5 mm thick slice, equilibrium is reached in 250 min at 50 C, 166 min at 60 C and 83 min
at 70 C. Fruits and vegetables drying time reduction with air temperature increasing was reported by many authors: Doymaz
(2005) and Gogus and Maskan (1999) for okra, Doymaz (2004b)
for carrot, Ertekin and Yaldiz (2004) for eggplant, Stamatios and
Vassilios (2004) for g and Rosell et al. (1997) for green bean. This
drying time reduction should be linked to the energies necessary
for water extraction from the product and its vaporisation. These
energies are as high as the drying air temperature is high. However,
increasing temperature to decrease drying time should take into
account the conditions of retention of certain nutrients of mangoes
such as vitamin C, the risk of volatile components lost at higher
drying temperatures and the product crusting.
The inuence of slice thickness on the Moisture Ratio and drying rate curves per unit of exchange surface area of Amelie mango
for three slices of 2.5, 5 and 10 mm thick at 70 C is presented in
Figs. 4a and 4b. We can note that at any given level in Fig. 4a the
time required to reduce the Moisture Ratio was dependent on
the slice thickness. Moreover, corrected drying rate and non-corrected drying rates are as high as the thickness is small. When
the slice thickness decreases, the drying rate per unit of exchange

0.003

0.8
0.7

0.0025

Drying rate (kgs-1m-2)))))

0.9

MR=(X-Xeq)/(Xo-Xeq)

0.00275

2.5mm Experimental data


5mm Experimental data
10mm Experimental data
2.5mm Page model
5mm Page model
10mm Page model

0.6
RH=15% Ta=70 C Va=0.8m/s

0.5
0.4
0.3

2.5mm

0.00225
0.002

5mm

0.00175
0.0015
0.00125
0.001

10mm

0.00075
0.0005

0.2

per unit of current surface area (with shrinkage)


per unit of initial surface area (without shrinkage)

0.00025

0.1

0
0
0

0.5

1.5

2.5

Time(s)

3
x 10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.1

X/X0

Fig. 4a. Inuence of slices thickness on the drying kinetic of Amelie mango at 70 C
and simulation by Page model.

Fig. 4b. Inuence of slices thickness on the drying rate per exchange surface area of
Amelie mango at 70 C.

Table 3
Fitting statistics of drying kinetics
Dimensions
L (cm)

Henderson and Pabis


W (cm)

d (cm)

Air characteristics: Va = 0.8 m s


4.2
1
5
4
1
5
3.8
0.5
5
1.6
0.6
2.5
4
2.7
10

Ta (cm)

Page
2

1.0150
1.0605
1.0824
1.0232
1.0335

2.8761  104
5.1666  104
4.2015  104
0.0012
1.3123  104

0.9971
0.9923
0.9896
0.9967
0.9986

1.4957  104
6.1110  104
0.0010
2.8154  104
1.4277  104

R2

v2

1.7562  104
2.9391  104
6.2462  105
7.0988  104
6.0183  105

1.0612
1.0648
1.2313
1.0736
1.0817

0.9968
0.9940
0.9983
0.9985
0.9996

1.6445  104
4.7429  104
1.7277  104
1.3217  104
4.3428  105

1

, RH = 15%
50
60
70
70
70

A.O. Dissa et al. / Journal of Food Engineering 88 (2008) 429437

surface area increases and the time required for attaining equilibrium decreases. At 70 C, this time is 666, 250 and 83 min, respectively for slices of 10, 5 and 2.5 mm thick. That means that, when
the thickness is doubled, more than the double of the time required
to attain the equilibrium point is necessary.

1
40C (RH30% Va0.8m/s)
50C (RH15% Va0.8m/s)
60C (RH15% Va0.8m/s)
70C (RH15% Va0.8m/s)
60C (RH20% Va0.48m/s)
60C (RH40% Va1.6m/s)

0.9

(X-Xeq)/(Xcr-Xeq)

0.8
0.7
0.6
0.5
0.4
0.3
0.2

R= 0.9907

11

-20.7

10

-20.8
-20.9

9
8

-21

-21.1

-21.2

3.1
1/T (1/K)

4.3.4. Diffusivity and activation energy


Experimental drying kinetics during falling drying rate period of
Amelie mango slices of 5 mm thick for various drying conditions
(Ta = 40, 50, 60 and 70 C, RH = 15%, 20%, 30% and 40% and
Va = 0.48, 0.8, 1.6 m s1) are showed on Fig. 5a. From these drying
curves, were deduced water diffusivities identied from the slopes
of ln[(X  Xeq)/(Xcr  Xeq)] vs. ratio time [l(t)]2 tted straight-line.
This linear data tting was done with a determination coefcient
R2 = 0.9935. Identied diffusivities were presented in Table 4,
likewise shrinkage and temperature inuences on their values
were illustrated. These diffusivities varied from 2.61  1010 to
1.2973  109 m2 s1 for drying data corrected for shrinkage and
from 1.04  109 to 4.05  109 m2 s1 for non-corrected drying

-10

12

-20.6

-21.3
2.9

x 10

R= 0.9917

-20.5

ln (D)

4.3.3. Kinetics models evaluation


The experimental results of drying kinetics were t using Page
and Henderson & Pabis models and the tting statistics are presented in Table 3. The experimental data were well tted by the
two models with similar values of tting parameters. However,
Pages model tting parameters R2 from 0.9940 to 0.9996 and v2
from 4.3428  105 to 4.7429  104 were the highest. As showed
in Figs. 3a and 4a, Pages model suitably describes the experimental drying curves of mango slices and may be assumed to represent
Amelie mango desorption kinetics prole.

13

-20.4

D(m/s)

436

3.2
x 10

-3

310

320

330

T(K)

340

350

Fig. 5b. Inuence of temperature on Amelie mango water diffusivity: convective


drying of slices of 5 mm thick at 40, 50, 60 and 70 C.

data. These diffusivities values were in the same range as those obtained by Ruiz-Lpez and Garca-Alvarado (2007) for a Mexican
mango variety, which varied from 1010 and 2  109 m2 s1.
According to Table 4, enough signicant differences existed between diffusivities obtained from corrected data and those obtained from non-corrected data. Relative differences between the
two types of diffusivity ranged from 45% to 76%. These results
showed that Amelie mangoes diffusivities were much overestimated when shrinkage was not taken into account in drying data
processing. Corrected diffusivities from Table 4 and Fig. 5b increased with temperature in 1010 m2 s1 at 40 C to 109 m2 s1
at 70 C. Also, ln(D) decreased almost linearly with the inverse of
temperature according to Fig. 5b. From the slope of the curve D
vs. 1/T (Fig. 5b), the activation energy value for a sample of
5 mm thick was evaluated at 25.355 kJ/mol. This value was close
to 27 kJ/mol, that obtained by Ruiz-Lpez and Garca-Alvarado
(2007) for the Mexican variety. From experimental data tting
according to Arrhenius law and for temperature ranging from 40
to 70 C, Amelie mango diffusivity was best expressed by the following equation:


25471
D m2 =s 1:003105 exp
28
RT 273:15

0.1

5. Conclusion

0
0

10

20

30

40
x103

Time (s)

Fig. 5a. Inuence of drying conditions on the drying curve of mango slices of 5 mm.

Table 4
Inuence of dryings conditions on the water effective diffusivity value
Ta
(C)

d
(mm)

X0
(db)

RH
(%)

Va
(m s1)

Xcr/
X0

D
(m2 s1)

Dv,eff
(m2 s1)

(Dv,eff  D)/
Dv,eff (%)

40
50
60
60
60
70
70
70

5
5
5
5
5
2.5
5
10

5.03
6.22
6.59
7.05
6.11
6.52
6.43
7.23

30
15
15
20
40
15
15
15

0.8
0.8
0.8
0.48
1.6
0.8
0.8
0.8

0.87
0.72
0.31
0.38
0.37
0.41
0.33
0.57

5.63  1010
7.49  1010
1.07  109
3.15  1010
2.61  1010
5.44  1010
1.30  109
1.17  109

1.04  109
1.37  109
3.15  109
1.12  109
1.09  109
1.45  109
4.05  109
3.18  109

45.86
45.29
66.11
71.88
76.06
62.48
67.97
63.21

D: diffusivities corrected for shrinkage; Deff: diffusivities non corrected for


shrinkage.

In this work, convective drying of the Amelie mangoes variety


was characterised. Drying kinetics at 50, 60 and 70 C and bulk
shrinkage at 50, 60 and 70 C were experimentally established
and simulated by models. Results showed that Amelie mangoes
shrinkage was less inuenced by drying air temperature and best
tted by the fundamental linear model of volumes additivity.
This model allowed evaluating the mango solid matrix density
and linear shrinkage coefcient at respectively 1468 kg/m3 and
1.468. Drying curvesstudy showed that Amelie mango drying
kinetics were best tted by Page model. There was a marked
inuence of shrinkage on drying rates and water diffusivities.
Drying rates were underestimated and diffusivities overestimated
when drying data are not corrected for shrinkage. A temperaturedependence law of diffusivity was established from shrinkagecorrected data and activation energy (Ea) was evaluated at
25.355 kJ/mol. This study allowed estimating drying parameters
of Amelie mango; the most used, produced, dried and exported
variety in sub-Saharan Africa. However, some intrinsic parame-

A.O. Dissa et al. / Journal of Food Engineering 88 (2008) 429437

ters of the fruit such as the degree of maturity could have effects
on Amelie mango drying and should be taken into account in
coming work.
Acknowledgements
The authors are grateful to the AUF (Agence Universitaire de la
Francophonie) and the French Cooperation service in Burkina Faso
for nancial support.
References
Anoar, A.E., Patricia, M.A., Fernanda, E.X.M., 2003. Drying kinetics of fresh and
osmotically pre-treated papaya. Journal of Food Engineering 59, 8591.
Crank, J., 1975. The Mathematics of Diffusion, second ed. Clarendon Press, Oxford,
London.
Desmarais, G., Marcotte, M., 2002. Cte dIvoire: Coup denvoi au schage des
mangues et des papayes. CRDA Saint-Hyacinthe, Canada.
Desmorieux, H., 1992. Le schage en zone subsaharienne: Une analyse technique
partir des ralits gographiques et humaines. Thse de lINPL, France.
Dissa, A.O., 2007. Schage convectif de la mangue: tude de linuence des
paramtres arauliques et intrinsques, conception et modlisation du
fonctionnement dun schoir solaire indirect. Thse de lUniversit de
Ouagadougou, Burkina Faso.
Doymaz, I., 2004a. Convective air drying characteristics of thin layer carrots. Journal
of Food Engineering 61, 359364.
Doymaz, I., 2004b. Drying kinetics of white mulberry. Journal of Food Engineering
61, 341346.
Doymaz, I., 2005. Drying charactristics of okra. Journal of Food Engineering 69,
275279.
Ertekin, C., Yaldiz, O., 2004. Drying of eggplant and selection of a suitable thin layer
drying model. Journal of Food Engineering 63, 349359.
Gogus, F., Maskan, M., 1999. Water adsorption and drying characteristics of Okra
(Hibisus Esculentus L.). Drying Technology 20, 83894.
Goyal, R.K., Kingsly, A.R.P., Manikantan, M.R., Ilyas, S.M., 2006. Thin-layer drying
kinetics of raw mango slices. Biosystems Engineering 95 (1), 4349.
Henderson, S.M., Pabis, S., 1961. Grain drying theory I: temperature effect on drying
coefcient. Journal of Agriculture Research Engineering 6, 169174.
Iglesias, O., Garcia, A., Roques, M., Bueno, J.L., 1993. Drying of water gels:
determination of the characteristic curve agaragar. Drying technology 3 (11),
571587.

437

Jaya, S., Das, H., 2003. A vaccum drying model of mango pulp. Drying Technology 21
(7), 12151234.
Leonardo, D.S.A., Dermeval, J.M.S., 2004. Dependence analysis of the shrinkage and
shape evolution of a gel of system with the force convection drying periods. In:
Proceedings of International Drying Symposium (IDS 2004), Sao Paulo, Brazil,
vol. A, pp. 152160.
Lozano, J.E., Rostein, E., Urbicain, M.J., 1980. Total porosity and open-pore porosity
in the drying of fruits. Journal of Food Science 45 (5), 14031407.
Lozano, J.E., Rotstein, E., Urbicain, M.J., 1983. Shrinkage, porosity and bulk density of
foodstuffs at changing moisture contents. Journal of Food Science 48, 1497
1502, 155.
May, B.K., Perr, P., 2002. The importance of considering exchange surface area
reduction to exhibit a constant drying ux period in foodstuffs. Journal of Food
Engineering 54, 271282.
Myara, M.R., Sablani, S., 2001. Unication of fruit water sorption isotherms using
articial neural networks. Drying Technology 19 (8), 15431554.
Nguyen, T.A., Verboven, P., Daudin, J.D., Bart, M.N., 2004. Measurement and
modelling of water sorption isotherms of Conference pear esh tissue in the
high humidity range. Postharvest Biology and Technology 33 (3), 229241.
Nieto, A., Castro, M.A., Alzamora, S.M., 2001. Kinetics of moisture transfer during air
drying of blanched and/or osmotically dehydrated mango. Journal of Food
Engineering 50, 175185.
Page, G.E., 1949. Factors inuencing the maximum rates of air drying shelled corn in
thin layers. M.S. thesis. Department of Mechanical Engineering, Purdue
University, Purdue, USA.
Pott, I., Neidhart, S., Mhlbauer, W., Carle, R., 2005. Quality improvement of nonsulphited mango slices by drying at high temperatures. Innovative Food Science
and Emerging Technologies 6, 412419.
Rosell, C., Simal, S., San Juan, N., Mulet, A., 1997. Nonisotropic mass transfer model
for green bean drying. Journal of Agriculture and Food Chemistry 45, 337342.
Ruiz-Lpez, I.I., Garca-Alvarado, M.A., 2007. Analytical solution for food-drying
kinetics considering shrinkage and variable diffusivity. Journal of Food
Engineering 79, 208216.
Simal, S., Femenia, A., Garau, M.C., Rosselo, C., 2005. Use of exponential, Pages and
diffusional models to simulate the drying kinetics of kiwi fruit. Journal of Food
Engineering 66 (3), 323328.
Stamatios, J.B., Vassilios, G.B., 2004. Inuence of the drying conditions on the drying
constants and moisture diffusivity during the thin-layer drying of gs. Journal
of Food Engineering 64, 445449.
Suzuki, K., Kubota, K., Hasegawa, T., Hosaka, H., 1976. Shrinkage in the hydration of
vegetables root. Journal of Food Science 41, 11891193.
Tour, S., Kibangu-Nkembo, S., 2004. Comparative study of natural solar drying of
cassava, banana and mango. Renewable Energy 29, 975990.