Natural Resources Canada, CanmetENERGY, 1615 Lionel Boulet Blvd., P.O. Box 4800, Varennes, QC, Canada J3X 1S6
a r t i c l e i n f o
Article history:
Received 24 October 2011
Received in revised form 3 January 2012
Accepted 5 January 2012
Available online 14 January 2012
Keywords:
Drying
Inert carriers
Pea starch
Kinetics
Diffusivity
Damage index
a b s t r a c t
This study aimed at investigating the drying of yellow pea starch dispersions on inert solid carriers and
determining the drying kinetics, moisture diffusivity and the product quality, quantied through damage
index and nal moisture content. Drying kinetics accomplished in a convective drying tunnel show that
the overall mass transfer is controlled by internal migration of moisture within the starch particles. For a
given inlet air temperature from 100 to 180 C, the apparent diffusion coefcient derived from the drying
curves increases exponentially with the instantaneous moisture content, with values ranging from
4 10
11
to 3 10
8
m
2
/s. Due to low diffusivity and the thin coat formed on the surface of solid carriers,
the resistance to internal diffusion is negligible as compared to the overall mass transfer resistance when
drying of starch dispersions takes place on inert solid carriers. Drying of yellow pea starch dispersion on
Teon particles as inert carriers was studied in laboratory and pilot fast spouted bed dryer for inlet air
temperatures from 140 to 240 C, and initial solid content of 38%, mass (d.b.). The starch damage index
for targeted product moisture content was below 2.5% in the inlet air temperature range from 120 to
210 C, when atomizing from the bottom of dryer.
2012 Published by Elsevier Ltd.
1. Introduction
Cereals (e.g., corn, wheat), potatoes and peas are the most pop-
ular sources of starch for many applications, including the produc-
tion of emerging biopolymers and animal diets. Several researchers
have also reported the possibility for the use of pea starch as an
encapsulating agent (Yilmaz et al., 2001; ngen et al., 2002; Korus
et al., 2003). Drying of pea starch dispersions is usually performed
in a spray dryer, a spin-ash dryer, or a pneumatic-ash dryer. For
required production rate, ranging typically from 5 to 150 kg of dry
starch per hour, the size of such dryers is so big that the related
capital and operating costs reduce greatly the prot margins. Smal-
ler equipment of similar performance is therefore needed. Drying
on inert solid carriers is one of the most efcient methods of drying
for this throughput range that offers low capital cost and compact
equipment, which facilitates its integration in the existing indus-
trial production site. It should be noted that the solid carriers can
be either active where the drying material is adsorbed by the car-
rier to form a composite product (contact-sorption drying), or inert
where the drying material is only deposited on the carrier surface
and removed from it by attrition once it is sufciently dry (Kudra
and Mujumdar, 2009). Drying on inert carriers has been laboratory
and pilot tested in various hydrodynamic congurations such as
turbulent uidized bed, swirling bed, mechanically spouted bed,
fast spouted bed (Spitzner-Neto et al., 1982; Barrett and Fane,
1989; Markowski, 1992; Ochoa-Martinez et al., 1993; Benali and
Amazouz, 2002; Kutsakova, 2004; Benali, 2004; Passos et al.,
2004; Wachiraphansakul and Devahastin, 2005; Benali and Ama-
zouz, 2006a; Barcelos and Freire, 2006; Rocha et al., 2009; Pereira
et al., 2010; Tatemoto and Miyazawa, 2011). A novel conguration
of the fast spouted bed has been developed (Benali and Amazouz,
2006b), which allows better control of the nal moisture content
and size distribution of a dry product. This novel conguration of
the dryer consists mainly of a conical drying chamber in conjunc-
tion with a reverse cone serving as a conical disengaging chamber.
The lower cone has a supporting grid for inert particles. A conical
screen is attached to the dryer exit to avoid the entrainment of
the inert carrier when discharging the powdery product with the
exhaust gas but also to enhance removal of a dry coat from the in-
ert carriers. The experiments performed on both laboratory and pi-
lot units, revealed that heat transfer coefcients in this novel
design are sufciently high (from 21.5 to 35.8 W/m
2
K) to attain
rapidly the thermal equilibriumbetween the three phases involved
in the drying process: inert solid carriers, drying air, and aqueous
starch dispersion. The ideal drying cycle in a fast spouted bed is
based on four successive steps: (a) heating of the inert carrier,
(b) coating of the inert carrier with ne droplets of starch disper-
sions, (c) drying of the wet coat, and (d) attrition of the dry coat
due to multiple collisions within the dryer and entrainment of
0260-8774/$ - see front matter 2012 Published by Elsevier Ltd.
doi:10.1016/j.jfoodeng.2012.01.003
inlet
inlet air temperature (C)
T
air
outlet
outlet air temperature (C)
V inlet air velocity (m/s)
X moisture content on dry basis (kg/kg, d.b.)
X
c
critical moisture content (kg/kg, d.b.)
X
0
initial moisture content (kg/kg, d.b.)
X
k
w
_ _
5
Eq. (5) can easily be integrated to give the time of the constant
drying period (t
c
):
t
c
q
s
d
s
6jq
air
_ _
k
w
C
p;air
T T
i
_ _
X
0
X
c
6
The constant rate period frequently involves evaporation of the
majority of liquid water; however, it can take only a small fraction
of the total drying time depending on the type of the material. One
of the simplest approximations of the falling rate period can be de-
scribed by the rst order-type kinetic equation, because the instan-
taneous moisture content decay is proportional to the moisture
content:
dX
dt
kX 7
The integration of this equation gives the incremental time for
the falling rate period:
t
f
t
c
1
k
ln
X
Xc
_ _ _ _
8
Table 4 visualizes the dependence of a drying time on inlet air
temperature. As expected, increased inlet air temperature in-
creases the temperature difference between this temperature and
the material temperature and, hence, the mass ux increases and
the drying time decreases. The direct relationship between the
mass ux and this temperature difference occurs because the heat-
ing by conduction and radiation is negligible in the drying tunnel.
Drying kinetics obtained in the drying tunnel show that the overall
mass transfer is controlled by internal migration of moisture with-
in the starch particles. Considering the fact that the main moisture
transfer occurs through molecular diffusion, the drying kinetics can
be represented by the second Ficks law of diffusion. Given that the
starch dispersion sample contained in the Teon pan can be con-
sidered as a thin layer of thickness , the mass transfer mechanism
at non-steady state can be expressed as for a plane sheet:
Time (s)
0 200 400 600 800 1000
X
/
X
o
(
-
)
0.0
0.2
0.4
0.6
0.8
1.0
180
o
C
140
o
C
100
o
C
V
air
= 1.4 m/s
(a)
Time (s)
0 200 400 600 800 1000
X
/
X
o
(
-
)
0.0
0.2
0.4
0.6
0.8
1.0
V
air
= 1.4 m/s
V
air
= 2.8 m/s
T
air
= 140
o
C
(b)
Fig. 3. Drying kinetics curves based on drying tunnel experiments: (a) effect of inlet
air temperature (b) effect of air velocity.
Table 3
Drying kinetics parameters based on drying tunnel experiments.
Parameter Value
Mass transfer coefcient 0.051 m/s
Apparent density of pea starch 1430 kg/m
3
Drying surface 0.0053 m
2
Table 4
Drying time versus inlet air temperature based on drying tunnel experiments.
Temperature (C) 100 140 180
t
c
(s) 310 245 165
t
f
(s) 149 113 94
t
tot
(s) 459 358 259
M. Benali / Journal of Food Engineering 110 (2012) 337344 341
@X
@t
@
@z
D
e
@X
@z
_ _
9
The analytical solution of the partial differential Eq. (9) depends
on the conditions of the surface moisture on the material surface,
viz.:
If the surface moisture distribution is uniform and the zone
within the range 0 z 0 is initially at a uniform moisture
content X
0
, the analytical solution can be written as suggested by
Crank (1975) for a plane sheet with z 2:
X X
X
0
X
8
p
2
1
n1
1
2n 1
2
exp p
2
D
e
2n 1
2
4
2
_ _
t
_ _
10
If the surface moisture distribution is uneven and the thin layer
(i.e., starch coat) is initially at a uniform moisture content X
0
, the
analytical solution can be written as for a plane sheet with z
X X
X
0
X
8
p
2
1
n1
1
2n 1
2
exp p
2
D
e
2n 1
2
2
_ _
t
_ _
11
Based on experimental results of drying kinetics, the theoreti-
cally predicted values of the apparent diffusion coefcient (D
e
) at
various moisture contents, obtained from Eqs. (10) and (11) are de-
picted in Fig. 4 for 100, 140 and 180 C. Due to large scale of the
values of D
e
and to steer clear of any scale mix-up, a secondary
g. is added to make obvious that the variation of D
e
is not linear
at low moisture contents 6 0:90kg=kg; d:b:. For a given inlet air
temperature, Fig. 4 shows an exponential increase of the apparent
diffusion coefcient with increasing moisture content. The values
of D
e
vary from 4 10
11
to 3 10
8
m
2
/s at air temperature var-
ied from 100 to 180 C. At the low moisture content (0.840.88 kg/
kg, d.b.), the effect of temperature on the apparent diffusion coef-
cient is not signicant: the average deviation is about 4% when
temperature varies from 100 to 180 C, which is in the order of
magnitude of the experimental errors. A remarkable aspect is ob-
served at 100 C since the corresponding curve shows a right shift
which was not detected in the case of 140 and 180 C. Such results
can be explained by the fact that at 100 C, a partial shrinkage com-
bined with a partial adhesion on the walls of the pan were ob-
served while a full detachment of the dried product was
observed at 140 and 180 C which means that internal water
migration decelerates due to structural modications to the dried
layer at 100 C.
The extension of these ndings on drying kinetics along with
extrapolation of the diffusion results to the drying on inert solid
carriers suggest that the overall mass transfer within the rst coat
on the solid carrier is not signicantly controlled by the internal
diffusion. Even if a fresh coat envelops the surface of the already
dry coat when the particle passes again through the spray zone,
the resistance to a diffusion process is still negligible. This can be
explained by the fact that water removal occurs predominantly
from the recently deposited coat. Such ndings are similar to those
obtained by Kutsakova (2004) when drying liquids and pasty
materials in a modied spouted bed of inert carriers.
3.2. Powder discharge control
As indicated in Section 2, a conical screen is located at the dryer
outlet to avoid entrainment of inert particles. Therefore, the holes
of this screen should be maintained open-up along the drying
operation to allow the continuous powder discharge. To do so, a
differential pressure sensor is located at the top of the dryer to
monitor continuously the pressure drop across the conical grid.
Fig. 5 shows the pressure drop prole over this discharge screen
while increasing the feed rate of pea starch dispersion obtained
on both laboratory and pilot units. At the feed rate of pea starch
dispersion lower than 70 kg/h, the pressure drop is almost constant
and equal to 1.23 0.03 kPa, which means that in the feed range
from 0 to 70 kg/h, the openings in the screen are not blocked. How-
ever, the pressure drop growths substantially beyond 70 kg/h to a
maximum of 3.05 0.17 kPa, which indicates that either the open-
ing area in a currently installed screen is not sufcient to handle
high processing capacity or some of the holes are plugged with
the wet powder. Therefore, such a prole suggests that beyond
70 kg/h, an intermittent feeding of pea starch dispersion will favor
the continuous discharge of the powder. In a continuous process
control mode, the pressure transducer can be installed to monitor
the upper pressure drop limit determined from the following
experimental correlation based on the four-parameter Chapman
model (SigmaPlot
, 2000):
DP
screen
1:177 1:852 1 e
0:06 _ m
feed
167:5
12
The exponent in Eq. (12) may seem high. The value 167.5 is
actually due to particular feature of multiscale mathematical
modeling that the four-parameter Chapman model is based on.
The number of iterations performed was 28 to reach this correla-
tion that passes the ShapiroWilk test for normality and exponen-
tiality (i.e., principal goodness of t test for normal and uniform
data sets).
3.3. Quality aspects in drying on inert solid carriers
Pea starch in aqueous dispersion is a difcult-to-handle product
because of its tendency to settle very quickly if it is not
X (kg/kg, d.b.)
0.5 1.0 1.5 2.0 2.5 3.0
A
p
p
a
r
e
n
t
d
i
f
f
u
s
i
o
n
c
o
e
f
f
i
c
i
e
n
t
(
m
2
/
s
)
10x10
-9
20x10
-9
30x10
-9
40x10
-9
100
o
C
140
o
C
180
o
C
1
.
3
x
1
0
-
1
0
2
.
2
5
x
1
0
-
1
0
0.60 0.95
Fig. 4. Apparent diffusion versus moisture content: Effect of inlet air temperature
based on drying tunnel experiments.
Mass flow rate (kg/h)
0 50 100 150 200
P
r
e
s
s
u
r
e
d
r
o
p
o
v
e
r
t
h
e
u
p
p
e
r
g
r
i
d
(
k
P
a
)
0
1
2
3
4
Exponential phase
Plateau phase
Fig. 5. Pressure drop over the conical screen as a function of feed rate of pea starch
dispersion within conical fast spouted bed dryer.
342 M. Benali / Journal of Food Engineering 110 (2012) 337344
continuously owing or constantly mixed, which can induce pipe
clogging. Therefore, the position of a nozzle plays a key role for
such materials. Since the nozzle is in contact with the hot air
stream from the dryer inlet the solid deposit of pea starch can be
formed in the feeding pipe. A water jacket was designed to cool
down the nozzle and prevent the clogging of the nozzle. As shown
in Fig. 6, feeding the pea starch dispersion from the dryer bottom
offers a better product quality since the damage index (DI) ranges
from 1.86 to 1.88. The values of DI are in the range 1.982.22 when
atomizing the starch dispersion fromthe top of the dryer. The com-
parison with a commercial ash pneumatic dryer, where DI is in
the range 3.55.8, demonstrates the competitiveness of the pro-
posed drying method in terms of product quality in the tempera-
ture range from 120 to 210 C.
The feed rate-temperature effect on the nal moisture content
(FMC) of dry starch is shown in Fig. 7. As expected, the FMC
increases with the feed rate for a given inlet air temperature, and
decreases with increasing inlet air temperature, for a given feed
rate of the pea starch dispersion. As the damage index increases
by about 25% when increasing the inlet air temperature from
180 C to 210 C, the optimal inlet air temperature should be
accordingly in the range of 160180 C. The measurements of
FMC were repeated three times with the same measurement
procedure, the same operator, and the same instruments, used un-
der the same operating conditions. The differences between the
three replicates were in the order of 38%.
Our own experiments and literature data (Leontieva et al., 2002;
Konovalov et al., 2003; Kutsakova, 2004) indicate that Teon cubes
perform also well, and may be preferable with respect to the pur-
chase costs. To verify this general opinion for the pea starch drying,
a series of experiments were performed with Teon cubes of 4 mm
in edge (equivalent diameter of 4.96 mm) and Teon beads of
5 mm in diameter under the same operating conditions of inlet
air temperature, initial moisture content and feed rate of pea
starch dispersion. The effect of feed rate and Teon particles shape
on FMC is plotted in Fig. 8. At an inlet air temperature of 180 C, the
spherical particles outperform the cubic ones. Indeed, even tough
the heat transfer should be more efcient when the transfer sur-
face area A
cubes
=A
beads
1:22 of inert particles increases, there
is, however, an upper limit, above which the internal circulation
of inert particles is unstable and the spouting of the bed ceases.
These ndings suggest that the drying rates are not uniform when
using cubic shape for the inert carrier mainly because the wet coat
located at the corners of each Teon cube dries faster than the rest
of the wet coat. In other words, the residence time of the wet coat
in the eight corners of the Teon cube is almost instantaneous
whereas it is longer for the wet coat located on the six lateral sur-
faces of the cube, leading to a non-uniform drying cycle.
The above results reveal denitely that the inlet air tempera-
ture and the feed rate of pea starch dispersion control the quality
of the dried pea starch, and the product nal moisture content
can be predicted by measuring continuously the relative air
humidity at the dryer outlet. Indeed, as illustrated in Fig. 9, the
Fig. 7. Effect of inlet air temperature on nal moisture content of dried pea starch
within conical fast spouted bed dryer.
Feed rate of yellow pea starch solution (kg/h)
0 20 40 60 80 100 120 140 160
F
i
n
a
l
m
o
i
s
t
u
r
e
c
o
n
t
e
n
t
o
f
d
r
i
e
d
p
r
o
d
u
c
t
(
%
,
m
a
s
s
,
d
.
b
.
)
0
2
4
6
8
10
12
14
16
18
Teflon spheres
Teflon cubes
(T
air
)
inlet
= 180
o
C
initial solid content = 38%, mass (d.b.)
m
Teflon
= 12 kg
Fig. 8. Effect of Teon particles shape on nal moisture content of dried pea starch
within conical fast spouted bed dryer.
Inlet air temperature (
o
C)
100 120 140 160 180 200 220
D
a
m
a
g
e
i
n
d
e
x
(
%
)
1
2
3
4
5
6
7
8
Pilot unit - Feeding from the bottom
Laboratory unit - Feeding from the top
Laboratory unit - Feeding from the bottom
Industrial flash dryer
Targeted value of the industrial quality requirements
Fig. 6. Damage index versus inlet air temperature within conical fast spouted bed
dryer.
IMC = 62%, kg/kg
Relative humidity of outlet air (%)
0 2 4 6 8 10 12 14 16
F
i
n
a
l
m
o
i
s
t
u
r
e
c
o
n
t
e
n
t
o
f
y
e
l
l
o
w
d
r
i
e
d
p
e
a
s
t
r
a
c
h
(
%
,
m
a
s
s
,
d
.
b
.
)
0
2
4
6
8
10
12
14
16
(T
air
)
inlet
= 120
o
C
(T
air
)
inlet
= 140
o
C
(T
air
)
inlet
= 160
o
C
(T
air
)
inlet
= 180
o
C
(T
air
)
inlet
= 210
o
C
b
is
e
c
t
in
g
lin
e
initial solid content = 38%, mass (d.b.)
m
Teflon
= 12 kg
Fig. 9. Final moisture content of dried pea starch versus relative air humidity: Effect
of inlet air temperature within conical fast spouted bed dryer.
M. Benali / Journal of Food Engineering 110 (2012) 337344 343
nal moisture content increases with the raise of relative humid-
ity, for a given inlet air temperature. All experimental results are
located below the bisecting line, i.e., the relative humidity is al-
ways higher than the nal moisture content of dried pea starch,
which together can be predicted by the following correlation de-
rived from these results:
FMC kg=kg; d:b: 1:896e
0:163 HR %
0:397 13
As the relative air humidity sensors are designed and accurate
for temperatures up to 180 C and as the exhaust air temperature
does not exceed 110 C, the key benet of this correlation is to eas-
ily set the operating conditions according to the targeted nal
moisture content of the dried product. In practice, for a given inlet
moisture content of pea starch dispersion, the user of such correla-
tion will measure the relative humidity of outlet air and then cal-
culate the corresponding FMC for a set of fT
air
inlet
; _ m
s
g
conditions. If the calculated value of FMC does not match the de-
sired value, then the end-user will adjust either inlet air tempera-
ture or the feed rate of starch dispersion.
4. Concluding remarks
Drying on inert carrier is more suitable to process yellow pea
starch than ash dryer regarding the quality of dried product,
which is affected by both inlet air temperature and feed rate of
pea starch. Therefore, these two variables have to be well con-
trolled to maintain the desired product quality. Increasing the dry-
ing inlet air temperature led to higher processing capacity, thus the
process yield, and to lower nal moisture content of the pea starch.
However, the increase of this variable causes damage to pea starch:
the results suggest that the effective inlet air temperature should
not exceed 180 C to maintain a good quality of pea starch. One
of the proposed reliable technical solutions is a continuous moni-
toring of the relative humidity of the outlet air that allows the pre-
diction of the desired nal moisture content. Clearly, the drying
kinetics performed on convective drying tunnel can be used to ana-
lyze and predict the drying mechanism since the liquid lm which
coats the inert carriers (Teon beads) is very thin.
Acknowledgment
The author expresses his sincere thanks to Dr. Tadeusz Kudra
from Natural Resources Canada/CanmetENERGY for his valuable
advices.
References
AACC., 2000. Approved methods of American Association of Cereal Chemists. AAAC
Inc., St. Paul, Minnesota, U.S.A.
Barcelos, M.S., Freire, J.T., 2006. Stability of spouting regimes in conical spouted
beds with inert particle mixtures. Industrial Engineering and Chemistry
Research 45 (2), 808817.
Barrett, N., Fane, A., 1989. Drying liquid materials in a spouted bed. In Drying89. In:
Mujumdar, A.S., Roques, M. (Eds.), Hemisphere Publishing Corporation, New
York, USA, pp. 415420
Benali, M., 2004. Thermal drying of foods: loss of nutritive content and spoilage
issues. In: Mujumdar, A.S. (Ed.) Drying of products of biological origin, India:
Science Publishers, pp. 137152.
Benali, M., Amazouz, M., 2002. Effect of drying aid agents on processing of sticky
materials. Developments in Chemical Engineering Mineral Processing 10 (3 and
4), 114.
Benali, M., Amazouz, M., 2006a. Drying of vegetable starch dispersions on inert
particles: Quality and energy aspects. Journal of Food Engineering 74, 484489.
Benali, M., Amazouz, M. 2006. Apparatus for producing powder from biomaterials.
US Patent, No. 6993856.
Crank, J., 1975. The mathematics of diffusion. Oxford University Press, New York,
pp. 4468 and 89103
Konovalov, V.I., Gatapova, N.Z., Kudra, T., 2003. Drying of liquid dispersions a
unied approach to kinetics and modeling. Drying Technology 21 (6), 1029
1047.
Korus, J., Tomasik, P., Lii, C.Y., 2003. Microcapsules from starch granules. Journal of
Microencapsulation 20 (1), 4756.
Kudra, T., Mujumdar, A.S., 2009. Advanced drying technologies. 2nd edition, CRC
Press, Taylor $ Francis Group., New York, USA, pp. 2954
Kutsakova, V.E., 2004. Drying of liquid and pasty products in a modied spouted
bed of inert particles. Drying Technology 22 (10), 23432350.
Leontieva, A.I., Bryanklin, K.V., Konovalov, V.I., Utrobin, N.P., 2002. Heat and mass
transfer during drying of a liquid lm from the surface area of a single inert
particle. Drying Technology 20 (4 and 5), 729747.
Markowski, A.S., 1992. Drying characteristics in a jet-spouted bed dryer. Canadian
Journal of Chemical Engineering 70 (5), 938944.
Ochoa-Martinez, L.A., Brennan, J.G., Niranjan, K., 1993. Spouted bed dryer for liquid
foods. Food Control 4 (1), 4145.
ngen, G., Yilmaz, G., Jongboom, R.O.J., Hennink, W.E., 2002. Encapsulation of a
amylase in a starch matrix. Carbohydrate Polymers 50 (1), 15.
Passos, M.L., Trindade, A.L.G., dAngelo, J.V.H., Cardoso, M., 2004. Drying of black
liquor in spouted bed of inert particles. Drying Technology 22 (5), 10411067.
Pereira, N.R., Godoi, F.C., Rocha, S.C.S., 2010. Drying of starch suspension in spouted
bed with inert particles: physical and thermal analysis. Drying Technology 28
(11), 12881296.
Rocha, S.C.S., Donida, M.W., Marques, A.M.M., 2009. Liquid-particle surface
properties on spouted bed coating and drying performance. Canadian Journal
of Chemical Engineering 87 (5), 695703.
SigmaPlot