You are on page 1of 6

Journal of Food Engineering 79 (2007) 828833

www.elsevier.com/locate/jfoodeng

Drying characteristics of unshelled kernels of candle nuts


Elieser Tarigan a,b, Gumpon Prateepchaikul b, Ram Yamsaengsung c,
Anchalee Sirichote d, Perapong Tekasakul b,*
a

The Joint Graduate School of Energy and Environment (JGSEE), King Mongkuts University of Technology Thonburi, Bangkok, 10140, Thailand
b
Department of Mechanical Engineering, Faculty of Engineering, Prince of Songkla University, Hat Yai, Songkhla 90110, Thailand
c
Department of Chemical Engineering, Faculty of Engineering, Prince of Songkla University, Hat Yai, Songkhla 90110, Thailand
d
Department of Food Technology, Faculty of Agro-Industry, Prince of Songkla University, Hat Yai, Songkhla 90110, Thailand
Received 24 August 2005; accepted 24 February 2006
Available online 21 April 2006

Abstract
Candle nuts are cultivated in approximately 170 000 ha. in Indonesia, and yield around 140 000 tons of fruit per year. Prior to removing the shells, drying is an important process in the post-harvesting practice of candle nuts to obtain kernels as the nal products. This
paper gives a brief description of the eorts to obtain the basic drying characteristics of candle nuts, which is a pre-requisite for dryer
design. The measurements were taken for unshelled kernels of candle nuts, for both fresh and stored samples, at temperatures of 40, 50,
and 60 C and drying air velocity of 0.67 m/s and 1 m/s using a forced convection dryer. The total drying time was substantially reduced
with an increase in temperature, and the eect of drying air velocity was relatively small. Pages model adequately describes the drying
behavior for the range of temperatures measured. The dependence of the drying constant on air temperature can be described by the
Arrhenius model.
 2006 Elsevier Ltd. All rights reserved.
Keywords: Candle nuts; Unshelled kernels; Moisture content; Drying rate

1. Introduction
Candle nut (Aleurites moluccana) is one of the important
multipurpose trees of the world. It is indigenous to IndoMalaysia, and today is widespread throughout the tropics
(Elevitch & Manner, 2004). Candle nuts, locally called
kemiri in Indonesia, have been utilized for various purposes: the lipid-rich seeds provide material for illumination, cooking, pharmaceuticals, and the trunks provide
timber. Candle nuts are cultivated in approximately
170 000 ha. in Indonesia, and yielded around 140 000 tons
of fruit in the year 2003. The production has been for both
local consumption and export. Additionally, candle nuts
trees have been planted as a part of the rehabilitation of
forest areas in some regions of Indonesia (Koji, 2002).
*

Corresponding author.Tel.: +66 74 287035; fax: +66 74 212893.


E-mail address: perapong.t@psu.ac.th (P. Tekasakul).

0260-8774/$ - see front matter  2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jfoodeng.2006.02.048

The physical features of candle nuts have been explained


in a previous study by Tarigan, Prateepchaikul, Yamsaengsung, Sirichote, and Tekasakul (2006). A candle nut consists of a kernel covered with a shell and an outer bulk
skin. The outer bulk skin of a mature nut is pliable and
can be easily removed. In Indonesia, harvesting of candle
nuts includes collecting the nuts which have dropped from
trees and removing the outer bulk skin, thus leaving the
unshelled kernels. Candle nuts can be stored either in the
form of shelled or unshelled kernels. Drying is an important process prior to removing the shells to obtain kernels
as the nal products. The removal of the shells can only be
done for the dry product.
Knowledge of the drying kinetics of biological materials
is essential to the design, optimization, and control of the
drying process (Mohamed et al., 2005). Several studies
have been conducted to investigate the drying behavior
of dierent kinds of nuts, such as peanuts (Chen, 2000),

E. Tarigan et al. / Journal of Food Engineering 79 (2007) 828833

pistachio nuts (Yanniotis & Zarmboutis, 1996), Turkish


zilgen & O
zdemir, 2001), and Macadamia nuts
hazelnuts (O
(Palipane & Driscoll, 1994; Wang, Kanoh, Kaneko, Lu, &
Duong, 2002). However, very little data on the drying
behavior of candle nuts is available for the engineering
design of drying. In a previous work, Tarigan et al.
(2006) reported the experimental results on sorption isotherms of candle nuts for both unshelled and shelled kernels at temperatures of 30, 40, 50, and 60 C. The
reported equilibrium moisture content of unshelled kernel
samples is higher than shelled kernel samples at any given
temperature and relative humidity. In both unshelled and
shelled kernel samples, hysteresis was pronounced and
was reduced as the temperature increased, and the isosteric
heat of desorption was higher than the isosteric heat of
adsorption. The study did not involve the candle nuts drying characteristics, such as drying time and eect of drying
air temperature and velocity.
The main objective of this work was to study drying
characteristics of candle nuts, so that post-harvest operating conditions could be established. The measurements
were taken for unshelled kernels. Drying must be a part
of the unshelled kernel of post-harvesting processes, and
this is a form in which the product can be stored. In this
study, temperatures of 40, 50, and 60 C were chosen. This
temperature range represents the preliminary stage of study
of candle nuts drying in a solar dryer working with an auxiliary heating system.
1.1. Traditional and alternate drying methods
After the harvest, the unshelled fruits of candle nuts are
commonly rst dried by sun drying for 45 days (for sunny
days), then the shells are cracked to obtain the kernels.
Cracking of the shells is usually done by farmers using a
small ladle-like tool made of rattan or wood. A candle
nut is placed on the tool and struck against a large stone
to crack the shell. The product is sold under the name candle nuts or kemiri is the shelled kernels. After the nuts are
extracted, the remaining shells can be directly used as fuel.
During the sun drying process, the dry level is commonly assessed by feeling with the hand. This method does
not provide signicant information about the moisture
content of the fruit. The cracking process is very sensitive
to the dry level. If they are not dry enough when cracked,
the kernels will be broken. On the other hand, some deterioration in the obtained kernels is commonly found in
the over-dried products. These is associated with poor
quality of the products.
Sun drying is not considered fully appropriate for drying
candle nuts, especially during high product market
demand, and during poor weather conditions. Solar drying
with a back-up system seems to be a good alternative for
candle nuts. To design any new dryer system, basic drying
parameters of the products to be dried, such as the drying
rate under dierent conditions must be known. Such basic
information is not available for many kinds of products

829

such as candle nuts. This paper gives a detailed account


of eorts made to determine this information.
2. Materials and method
The experiments were conducted for both fresh and
stored unshelled candle nuts. Fresh product refers to the
product that is recently separated from its outer bulk skin.
Stored product sample refers to the unshelled kernel sample that was separated from its outer bulk skin, pre- sun
dried by sun drying for about 2 days, and stored at room
temperature about for 1 month before the experiments.
The measurements were taken using a forced convection
dryer. The measured parameters were drying time, temperature, drying air velocity, and mass of dried products. Measurements were taken at temperatures of 40, 50, and 60 C
and drying air velocity of 0.67 m/s and 1 m/s. The dryer
was installed in an environment with a relative humidity
of about 9095%, and the ambient temperature of about
2740 C.
The dryer consists of two fans, heater block, drying
chamber, and control and measurements unit. The schematic diagram of the dryer is shown in Fig. 1. The drying
chamber of 80 40 40 cm3 was made from galvanized
sheet metal of 1.5 mm thickness. The drying chamber has
a door at the front for insertion and removal of the sample.
The drying chamber, heater block, and air tunneling were
covered with 30 mm rock wool and aluminum foil to minimize heat during the experiments. Air supplied by the fans
was heated to the required temperature in the air heater
block with a 3000 W electrical heater. The air velocity
and temperature were controlled with an assembled electrical controller and measurements unit. The air velocity was
directly measured in the drying chamber using a hot-wire
anemometer as a sensor with a measurement range of 0
5 m/s. The temperature and relative humidity in the drying
chamber were measured by a themocouple and RenseHT733M06 relative humidity sensor, respectively.
In the drying chamber, the dried samples were placed in
a cubic basket made from large-mesh metal in which the

Fig. 1. Schematic diagram of dryer.

830

E. Tarigan et al. / Journal of Food Engineering 79 (2007) 828833

products can be stacked so they are evenly dried. The


experiments were considered as thin layer drying. About
300 g of candle nuts was tested for each experiment. The
dried samples weights were recorded on a scale/balance
having an accuracy of 0.01 g. Weighing was made at least
once every hour during the rst 5 h of drying test, and at
least once every 3 h afterwards. Removing, weighing, and
replacing the sample took about 2 min. The moisture contents of the samples were measured by air-oven at a xed
temperature of 100 C for 72 h until the mass did not
change (Hall, 1980). At least three replicates of each experiment were made.

The arithmetic mean of product samples moisture content from the experimental replicates was determined and
plotted versus drying time. For each experimental point,
the standard deviation (S) was determined by
s
PN
i1 M i  M
S
;
1
N 1
where M i ; M and N are moisture content at a certain point,
mean value of moisture content from replicates, and number of replicates, respectively. In addition to S, the relative
% deviation (D) between the mean and individual value
was determined by

!
N 
Mi  M
100 X
D
.
2
N
M
i1
The description of the drying kinetics experimental data
was veried according to the following three models (Babalis, Papanicolaou, Kyriakis, & Belessiotis, 2006; Kaleemullah & Kailappan, 2006; Sacilik, Elicin, & Unal, 2006):
Pages model
3

Logarithmic model
M R a  expk  t c;

Two-term exponential model


M R a  expk  t b  expk 0  t;

where MR is dimensionless moisture ratio; t is drying time


(h); k and k0 are drying constants (1/h); a, b, and n are
dimensionless constants. The moisture ratio is given as
MR

M  M eq
;
M 0  M eq

^ R;i is the estiwhere MR,i is the measured value data, M


mated value through tting of the model, and m is the
number of experimental points.
The dependence of the drying constant on air temperature, calculated from the best t of the there models above,
was analyzed using the Arrhenius model, which is given as
(Sogi, Shivhare, Garg, & Bawa, 2003)
k A  expB=T ;

where T is the absolute temperature (K); A and B are


coecients.

2.1. Data analysis

M R expk  tn ;


!
m 
^ R;i 
M R;i  M
1 X
;
MRD
m i1
M R; i

where M, M0, Meq are the moisture content at any time, the
initial moisture content, and equilibrium moisture content
in % d.b., respectively. The values of Meq were obtained
from a previous study by Tarigan et al. (2006).
The goodness of t for tting of the three models was
evaluated with the correlation coecient (r2) and the mean
relative deviation (MRD), which is dened as

3. Results and discussion


From the experimental data, we observed a small variance for measured moisture contents (replicates), where
the average standard deviation (S) were equal to 0.16 and
0.23 for stored and fresh samples, respectively. The average
of relative % deviation (D) were 1.5% and 2% for stored
and fresh samples, respectively. For most cases, the higher
deviation was observed at lower moisture contents.
Measurement results indicated that the initial moisture
content, in dry basis (d.b.), of fresh samples was 16.6%,
and the initial moisture content of stored samples was
8.3%. This implies that after the harvest, pre-sun drying,
and one month storing at room temperature and relative
humidity, the candle nuts moisture content remain the
same as half of fresh condition. Another separate measurement for the samples that were stored without pre-sun drying indicated that it almost reached equilibrium moisture
content of 10% after one month of storing.
The most suitable of level of dryness for cracking process was 46%. The observation showed that the cracking
process can only be done if the dryness level of the samples
is lower than 6%, otherwise the kernel broke or cracked
during the shells cracking. Broken or cracked nuts indicates bad quality of the nal product. However, some deterioration in the obtained kernels was found in the samples
with the level of dryness lower than 3%.
The moisture contents of samples as a function of drying
time are presented in Figs. 2 and 3 for stored and fresh
samples, respectively, at drying air velocity of 1 m/s. The
gures also show the standard deviation bars for each
experimental data. The periods of drying that can be identied were mostly falling rate periods. This indicates that
diusion-controlled process in which the rate of moisture
removal is limited by diusion of moisture from inside to
the surface of the product.
The increase of drying air velocity for both stored and
fresh samples from 0.67 to 1 m/s has a relatively small
inuence to decrease drying time. Fig. 4 shows the eect
of drying air velocity between 0.67 m/s and 1 m/s to the
drying time at drying temperatures of 50 C for the stored

E. Tarigan et al. / Journal of Food Engineering 79 (2007) 828833

Fig. 2. The drying experimental data with the standard deviation bars for
the stored candle nuts at drying air velocity of 1 m/s.

831

Fig. 4. The experimental data (with the standard deviation bars) on the
eect of drying air velocity on the drying time for stored and fresh
unshelled candle nuts at temperature of 50 C.

respectively, to reach a moisture content of 6%. For all


temperatures tested, the predominant direct eect of air
temperature on the drying rate becomes less important
after reaching approximately 3% of moisture content.
From the experiment, the maximum drying temperature
of 60 C for candle nuts is recommended. When the samples were dried at a temperature higher than 70 C some
deterioration was observed in the nal products.
The moisture ratio, MR, of candle nuts at drying air
temperatures of 40, 50, and 60 were tted in three models:
Pages model (Eq. (3)), Logaritmic model (Eq. (4)), and
Two-term exponential model (Eq. (5)). The estimated
values of the parameters of the models are listed in Tables
1 and 2 for stored and fresh samples, respectively. From

Fig. 3. The drying experimental data with the standard deviation bars for
the fresh candle nuts at drying air velocity of 1 m/s.

and fresh samples. The gure also shows the standard deviation bars for each experimental data.
The eect of drying temperature is signicant to the drying time for both stored and fresh samples, and total drying
time was substantially reduced with an increase in temperature. When the drying temperature was 40 C, more than
25 h was required for stored samples to reach moisture
content of 6%, and for fresh samples more than 33 h was
required. This agrees with sun drying of the nuts by the
farmers which usually takes approximately 34 days of
drying before the cracking process. When the drying temperature was 50 C, the drying time to reach a moisture
content of 6% was about 6.5 h for stored samples, and
for fresh samples about 21 h. The drying time was signicantly shortened when drying temperature was 60 C, and
it only took 4 h and 7.5 h for stored and fresh samples,

Table 1
Parameters of the three drying models for drying kinetics of stored
unshelled candle nuts
Drying temperature, C

Model

Model parameters/
goodness of t

40

Page

k
n
r2
MRD

0.1461
0.5956
0.9965
0.01955

0.1839
0.7165
0.9989
0.0092

0.2786
0.7172
0.9937
0.0301

Logarithmic

a
k
c
r2
MRD

0.6344
0.3020
0.0927
0.9808
0.0457

0.7452
0.2195
0.1608
0.9916
0.0322

0.7443
0.2538
0.3349
0.9932
0.0381

Two-term

a
k
b
k0
r2
MRD

0.7173
0.2764
0.0261
0.3835
0.9964
0.0196

1.6940
0.7708
0.0907
0.0918
0.9756
0.0571

0.6685
0.3312
0.3712
0.0252
0.9996
0.0370

50

r2 = correlation coecient; MRD = mean relative deviation.

60

832

E. Tarigan et al. / Journal of Food Engineering 79 (2007) 828833

Table 2
Parameters of the three drying models for drying kinetics of fresh
unshelled candle nuts
Model

Model parameters/
goodness of t

Drying temperature, C
40

50

60

Page

k
n
r2
MRD

0.0871
0.7603
0.9959
0.0262

0.1206
0.7437
0.9988
0.0113

0.2173
0.9270
0.9981
0.0264

Logarithmic

a
c
k
r2
MRD

0.7996
0.1481
0.0534
0.9916
0.0291

0.7056
0.2671
0.1108
0.9950
0.0213

0.9471
0.0437
0.2093
0.9973
0.0301

Two-term

a
b
k
k0
r2
MRD

0.8873
0.1121
0.0356
0.8926
0.9980
0.0265

0.8267
0.1668
0.0474
0.4683
0.9980
0.0134

15.010
43.990
0.2395
0.2404
0.9911
0.04731

r2 = correlation coecient; MRD = mean relative deviation.

the values of coecient of determination (r2) and mean


relative deviation (MRD), calculated from Eq. (7), it is
clear that the Pages model has the best t in predicting
the moisture ratio of both stored and fresh samples candle
nuts during thin layer drying. The values of correlation
coecient of tting to Pages model were always greater
than 0.99, and in most cases, the mean relative deviations
were less than 0.03. Comparisons of the experimental
and predicted moisture ratios obtained using Pages model
are shown in Figs. 5 and 6 for stored and fresh samples,
respectively. The standard deviation bars for each
experimental data for the moisture ratio are shown in the
gures.

Fig. 6. Comparison of experimental (with the standard deviation bars)


and predicted moisture ratios of fresh candle nuts by tting of Pages
model.

For both stored and fresh samples, the drying constant,


k, which is a measure of drying rate, increased with temperature. The dependence of the drying constant on air temperature can be described by the Arrhenius model (Eq.
(8)). Relationships between the parameters of the model
and drying air temperature calculated based on Pages
model were found to be for:
Stored samples:
k 1:106  106  exp5149=T r2 0:9760;

Fresh samples:
k 1:125  104  exp3539=T r2 0:9800

10

4. Conclusion
The most suitable of level of dryness for the cracking
process for unshelled candle nuts was found to be about
45%. The eect of drying temperature is signicant on
the drying time. The maximum drying temperature for candle nuts is recommended to be 60 C. When drying temperature was 60 C, it took about 4 h for stored samples and
7.5 h fresh samples to reach a moisture content of 6%
(d.b.). The inuence of drying air velocity was relatively
small. Pages model adequately describes the drying behavior for the range of temperatures measured. The drying
constant increased with temperature and its dependence
follows the Arrhenius model.
Acknowledgements

Fig. 5. Comparison of experimental data (with the standard deviation


bars) and predicted moisture ratios of stored candle nuts by tting of
Pages model.

This research has been nancially supported by the Joint


Graduate School of Energy and Environment (JGSEE),
King Mongkuts University of Technology Thonburi,
Bangkok, Thailand.

E. Tarigan et al. / Journal of Food Engineering 79 (2007) 828833

References
Babalis, S. J., Papanicolaou, E., Kyriakis, N., & Belessiotis, V. G. (2006).
Evaluation of thin-layer drying models for describing drying kinetics
of gs (Ficus carica). Journal of Food Engineering, 75(2), 205214.
Chen, C. (2000). A rapid method to determine the sorption isotherms of
peanuts. Journal of Agricultural Engineering Research, 75, 401408.
Elevitch, C. R., & Manner, H. I. (2004). Aleuritus moluccana(kukui).
Species proles for Pacic Island agroforestry. Permanent Agricultural
Resources (PAR), Holualoa, Hawaii.
Hall, C. (1980). Drying and storage of agricultural crops. Westport,
Connecticut: The Avi Publishing Company, Inc (pp. 6971).
Kaleemullah, S., & Kailappan, R. (2006). Modelling of thin-layer
drying kinetics of red chillies. Journal of Food Engineering, 76(4),
531537.
Koji, T. (2002). Kemiri (Aleurites moluccana) and forest resource
management in Eastern Indonesia: an eco-historical perspective. Asian
and African Area Studies, 2, 523.
Mohamed, L. A., Kouhila, M., Lahsasnia, S., Jamalia, A., Idlimama, A.,
Rhazia, M., et al. (2005). Equilibrium moisture content and heat of
sorption of Gelidium sesquipedale. Journal of Stored Products
Research, 41, 199209.

833

zilgen, M., & O


zdemir, M. (2001). A review on grain and nut
O
deterioration and design of the dryers for safe storage with special
reference to Turkish hazelnuts. Critical Reviews in Food Science and
Nutrition, 41(2), 95132.
Palipane, K. B., & Driscoll, R. H. (1994). The thin-layer drying
characteristics of macadamia in-shell nuts and kernels. Journal of
Food Engineering, 23(2), 129144.
Sacilik, K., Elicin, A. K., & Unal, G. (2006). Drying kinetics of Uryani
plum in a convective hot-air dryer. Journal of Food Engineering, 76(3),
362368.
Sogi, D. S., Shivhare, U. S., Garg, S. K., & Bawa, A. S. (2003). Water
sorption isotherm and drying characteristics of tomato seeds. Biosystems Engineering, 84, 293301.
Tarigan, E., Prateepchaikul, G., Yamsaengsung, R., Sirichote, A., &
Tekasakul, P. (2006). Sorption isotherms of shelled and unshelled
kernels of candle nuts. Journal of Food Engineering, 75(4), 447452.
Wang, Z. M., Kanoh, H., Kaneko, K., Lu, G. Q., & Duong, D. (2002).
Structural and surface property changes of Macadamia nut-shell char
upon activation and high temperature treatment. Carbon, 40,
12311239.
Yanniotis, S., & Zarmboutis, I. (1996). Water sorption isotherms of
pistachio nuts. Lebensmittel-Wissenschaft and Technologie, 5, 372375.

You might also like