THIN- LAYER DRYING OF DICED CASSAVAROOTS

African Journal of Science and Technology (AJST)
Science and Engineering Series Vol. 2, No. 2, pp. 94-100
THIN- LAYER DRYING OF DICED CASSAVA ROOTS
Kajuna S.T.A.R.1 *, V.C.K. Silayo2, A. Mkenda3 and P.J.J. Makungu1
Department of Agricultural Engineering and Land Planning,
Sokoine University of Agriculture,P.O. Box 3179, Morogoro, Tanzania.
e-mail: kajuna@suanet.ac.tz, skajuna@yahoo.com
ABSTRACT: Fresh cassava (Manihot spp) roots were obtained from a farm and used in this study.
They were peeled and diced using a special dicing machine into cubes of side 0.5 cm. The cubes were
dried in thin layers (one to three layers) in a drier that was specifically designed and fabricated in
the Department of Agricultural Engineering and Land Planning, Morogoro, Tanzania for the purpose.
The drier had a motor driven fan, a heater and a tray chamber, with thermometers for determining
entry and exit temperature (dry and wet bulb) conditions of the air. The input variables were: depth
of thin layer (0.5, 1.0 and 1.5 cm), drying temperature of the air (55 and 65 oC) and drying time (from
0 min until the sample attained equilibrium, at intervals of 25 min). The response variable was the
moisture content of the cassava cubes. Weather conditions during the experiments were also monitored.
A parallel sun drying experiment was carried out to compare thin layer drying on the sun and thin
layer drying in the fabricated dryer. For the fresh cassava that was used in the experiments, a
duplicate sample was placed in an oven at 75 oC for 7 hours to determine the initial moisture content.
The results indicated that the average moisture content of fresh cassava roots was about 75.4
%(w.b.). Both temperature and depth of layers were found to affect the drying characteristics of
cassava cubes, with single layer and higher drying temperature giving faster approaches to
equilibrium moisture content. Comparing sun drying of one layer at an average temperature of 25 oC
with artificial drying at the above named temperatures, it was found that sun drying took 2 to 3 days
to reduce the moisture content to Equilibrium Moisture Content (EMC), while this was achieved
within 150 min and 125 min with artificial drying at 55 oC and 65 oC respectively. The generally
accepted thin layer drying equations were fitted to the drying data of cassava cubes, and the Page
model was found to agree with the drying data of one, two and three layers with high accuracy for
artifial drying, but not for sundrying. The exponential model only agreed accurately with drying of
one layer.
KEY WORDS: Cassava cubes, Thin layer drying, Temperature, Moisture content
INTRODUCTION
boiled and eaten as a main meal, or milled into flour and
prepared into stiff porridge (ugali). The latter may be either
Cassava (Manihot esculenta) is the third most important
a pure flour or a composite flour where it is mixed with
staple food in the tropics, after rice and maize (Odogola,
maize of sorghum (Bangu et al., 1998).
1988). It is believed to provide 37% of the calorie

requirement in Africa, and 12 % and 7% in Latin America
Among the problems limiting utilization and consumption
and Asia respectively (Odogola, 1988). In Africa, Tanzania
of cassava is the rapid spoilage of the roots after harvesting
is the third largest producer of cassava, next to the
(Nghiem, 1991). This problem can be minimized by cold
Democratic Republic of Congo and Nigeria (FAO, 1985;
storage or processing in to more stable products. Cassava
CIAT, 1993, quoted by Mlingi, 1995). However, in Tanzania,
roots stored at temperature and RH of 13oC and 85-90%
the crop is only important in half of the country’s twenty
respectively can last for a maximum of 2 months and 5-6
regions.
months when stored at temperature and RH of 0-2 oC and

85 - 90%, respectively (NRI, 1994).
In Tanzania, cassava is maily regarded as famine security
food. It is widely consumed as a snack when roasted,
9 4
AJST, Vol. 2, No. 2: December, 2001

---

Thin- Layer Drying of Diced Cassava Roots
In most cases, cold storage is not economically feasible,
Me
= equilibrium moisture content (% db)
hence processing remains the best alternative towards
M
= moisture content at time t (% db)
improving availability of the crop during off-season period.
t
= drying time (hr)

K and N = drying constants.
The most common processing methods are direct sun-

drying of peeled roots for days in to a storable product,
The above equation is a modification of the theoretical
known as makopa (Mlingi, 1985). Another method is
model, known as the exponential or the Newtonian model
fermentation of peeled roots by soaking in water for several
(Sun and Woods, 1994). The model is described as (Nellist,
days. A third method is fermentation of peeled root pieces
1976; Colson and Young, 1990; Pattey et al., 1988; Crisp
in covered heaps to enhance mould growth desired to
and Woods, 1994)
impart special characteristics to the end product after
d M
drying. The different dried products can then be stored or
k

-

=

(
M
M

-


e )
d t

milled into flour to produce stiff porridge. However, due to
where k = drying constant.
weather uncertainities, sometimes drying can not proceed
Differentiating equation 2 gives (Nellist, 1976; Sun and
quickly enough. This may result in growth of undesirable
Woods, 1994)
moulds such as Aspergilus flavus that produces aflatoxins.
M

M

-

MR
=

e

=
Exp (

k

-

t
)
Quick drying of cassava is advantageous in the sence that
M o
-
M e
the risks of contamination and mould growth are minimized.
Equation 3 assumes that resistance to moisture movement
Moreover, drying reduces the levels of cyanogenic
and thus gradients within the material are negligible
glucosides which is a potential toxin naturally present in
(Colson and Young, 1990). At constant temperature,
cassava. The established 17 days of sun-drying of
pressure and hmidity, this equation is valid if drying is
longitudinally split roots only reduces cyanogenic
characterized by “falling-rate” regime (Nellist, 1976) which
glucosides to 27-37%, leaving more than 100 mg HCN
is a characteristic of drying of low moisture content
equivalent per kg dry weight of flour, that is ten times the
products such as grains. As quoted in Sun and Woods
safe level set by FAO/WHO (Mlingi, 1995).
(1994), this model has been successfuly used for barley

(Sharp, 1982; Bruce, 1985), wheat (Simmonds et al., 1953),
THEORY
paddy (Kachru et al., 1980), and shelled corn (Westerman

et al., 1973). The drying constants in thin-layer drying
Where heated air air is used as drying medium, the primary
equations vary with temperature (Yunfei and Morey, 1987;
factor influencing the rate of drying is temperature (Yunfei
Verma et al., 1985). In the work of Misra and Brooker (1980)
and Morey, 1987). In a study by Pathak et al., (1991), the
on yellow corn, the initial moisture content was also found
effect of drying temperature on thin-layer drying was high,
to affect the drying rate or constants.
followed by initial moisture content, air velocity and relative

humidity as the least factor that warranted its exclusion
OBJECTIVE
from consideration in the evaluation of thin-layer drying

rate of rapeseed. It has been reported, that drying is
The objective of this work was to determine the drying
independent of air velocity in the range of 0.15 to 0.81 m/s,
characteristics of diced cassava roots in an artificial dryer
but depends sharply on the drying temperature of the air
at temperatures higher than the ambient temperatures. The
from 21.1 to 76.7 oC (Simmonds et al., 1953, quoted by Sun
data obtained from thin-layer drying of the cubes would
and Woods, 1994).
be fitted to the generally accepted thin-layer drying models

(equations 1 & 3) to predict the drying characterics.
In drying of thin-layers of agricultural crops, the Page

equation has been used extensively. It was successfully
METHODOLOGY
used to describe drying characteristics of American

Ginseng (Yunfei and Morey, 1987), oilseeds such as
The Drying Equipment
sunflower (Yunfei et al., 1987). The equation is empirical,

and is given as:
A dryer (Figure 1) was designed and fabricated at the
Department of Agricultural Engineering and Land Planning
M M

-

MR
=

e

=
Exp (

-
K t N )
at Sokoine University of Agriculture, Morogoro, Tanzania.
M o M

-

e
where
MR
= Moisture ratio
Mo
= initial moisture content (% db)
AJST, Vol. 2, No. 2: December, 2001
95

---

S. T. A. R. KAJUNA
Figure 2. Schematic representation of the slicing machine.

The machine was calibrated for slicing 5mm thick slices.
The root was then sliced into 5mm thin rectangular slices,
and then re-sliced to produce cubes of 5 mm size .

Figure 1. Schematic representation of the thin-layer drying
equipement.
Drying Procedure


It consisted of a constant speed motor (0.5 hp) for driving
Before drying, the heater of the dryer was switched on and
a blower; a centrifugal blower (Type CNA-250/R, NR 8401-
the blower allowed to run for about 30 minutes to allow the
5141-001, Nordisk Ventilator, Denmark); heating chamber,
heated air to stabilize at the desired temperature. Two
consisting of two heater elements for heating air, and a
temperature settings, namely, 55 and 65oC were used for
circular column for leading air to and from the drying tray.
the drying work. Air velocity was measured by an air flow
The column was insulated with foam material and enclosed
meter (Type 5 manometer, Air flow Developments Ltd, High
in a 30 cm square column to minimize heat loss. Other
Wycombe, UK). Dry-bulb and wet-bulb temperatures of
components of the dryer were: the perforated drying tray
ambient air were measured by a sling psychrometer and
(10 cm diameter) for holding the sample to be dried; a pair
the data were interpretted into relative humidity using a
of thermometers, for recording temperature of the inlet and
psychrometric chart at atmospheric pressure.
outlet air; and a manometer for recording loss in head.


At the above temperatures, cassava cubes were dried in
Preparation of the Sample.
the dryer in one layer (5mm), two layers (10mm) and three

layers (15mm), each weighing 100g, 200g, and 300g
Freshly harvested cassava roots (Manihot spp) were
respectively, and forming a total of six runs. The runs were
obtained from a neighbouring farm and sent to the
replicated three times. The drying process was allowed to
laboratory for the study. They were peeled, and part of the
continue until no more weight loss was recorded between
meet removed from one side to create a flat surface. The
two successive readings. The control was a parallel
root was laid by the flat surface onto the board of a slicing
sundrying experiment of the same amount of material at
machine, schematized in Figure 2, which was also designed
ambient conditions. However, this was only done parallel
and manufactured for the purpose at the Sokoine
to the 55 oC. Concurrent with each drying run, three samples
University of Agriculture.
of cubes, about 100 g each were parallel dried in a ventilated
oven ( CARBOLITE model NR-30F, Fison way, Thetford,
Norfolk, England) at a temperature of 75 oC for six hours to
determine the initial moisture content.

The drying response variable measured was weight loss
at time intervals of 25 minutes.This involved quick
withdrawal of material from the dryer set-up and quick
weighing using a laboratory balance (Sartarius Universal,
96
AJST, Vol. 2, No. 2: December, 2001

---

Thin- Layer Drying of Diced Cassava Roots
Model U6100-A, GMBH, Gottingen, Germany). The sample
was quickly put back to continue with drying. Based on
the initial moisture content from oven drying, the weight
loss was used to calculate the moisture content using the
equation of Silayo (1995), given as
M i m
i w

-

l
M t
=

mi
-
wl
Where
M = m.c. at time t, (%w.b)
t
M = initial m.c, (%w.b)
i
m = initial weight, (g)
i
w = weight loss at time t, (g)
l
Figure 3. One layer drying of cassava cubes, showing change in
moisture ratio with time at temperatures of 55, 65 oC and sundrying.
The moisture content was converted to moisture ratio (MR)
using the non-exponential part of equations (1) and (3).

RESULTS AND DISCUSSION

The drying air velocity was found to be 0.67 m/s (0.02 m3/
s), which was within the range that does not affect the
drying rate (Simmonds et al., 1953 quoted by Sun and
Woods, 1994). The relative humidity of air ranged between
68 and 85 %.

The moisture content of freshly harvested tubers was
found to be 75.4 % wet basis. Literature reports a lower
value of 60% (Nghiem, 1991) and 62.5 % wet basis (Mlingi,
1995). The discrepancy could be attributed to the difference
in species used, difference in locality and seasonality of
the harvesting period (ie rainy or sunny seasons). In this
case, the work was done immediately after the heavy rain
season. The maturity of the root could be another reason
for the higher moisture content in this study because
experiene shows that as the roots of some varieties grow
old in the soil, they tend to develop water-soaked tissues,
especially around the pith. In this case, the roots tested
Figure 4. Two layer drying of cassava cubes at 55, 65 oC and
were about two years old in the soil. Another reason which
sundrying.
could have caused the difference in moisture content is
the method used in the determination of the moisture
content of which Nghiem (1991) is silent.

Equilibrim Moisture Content (EMC)

At the air temperature of 65oC, the EMC of cubes was
reached after 125, 150 and 175 minutes for one, two and
three layers respectively. This is indicated by moisture
ratio (MR) of zero as presented in Figures 3, 4 and 5.
AJST, Vol. 2, No. 2: December, 2001
97

---

S. T. A. R. KAJUNA
Table 1.Constants and regression coefficints for the expnential
model
Temperature of the air (oC)
55
65
1 layer 2 layers 3 layers
1 layer
2 layers 3 layers
A 0.991
0.6573
0.6346
0.9490
0.7766
0.6447
K -0.041 -0.0264 -0.026
-0.045
-0.0361 -0.0327
r2 0.993
0.975
0.973
0.999
0.990
0.974
A = 0.8544,
k = -0.0021,
r2 = 0.856

Figure 5. Three layer drying of cassava cubes at 55, 65 oC and
In the study of Verma et al., (1985) the constant “A” was
sundrying.
used as a shape factor. But in this study, “A” was used to
show the degree of fitness of the exponential model on top
The trend was almost the same at 55 oC (Figures 3 to 5)
of the regression coefficient.
where EMC of cubes was attained after 150 minutes for

one layer and 200 minutes for both two and three layers.
For one layer, deviation of the constant “A” from 1.0, which
This was expected because experience shows that at high
presents the ideal case of the exponential model is about
drying temperature, the drying rate is high and vice versa.
0.88% and 5.1% at air temperatures of 55 and 65oC,

respectively. For two layers the deviation increased to
For the sun-drying trials, the EMC for the single layer was
34.3% and 22.3% at 55 and 65 oC respectively, and 36.5 and
attained after 31 hours of continuous drying (day and
35.5% at the respective temperatures of 55 and 65 oC for 3
night). However, this was a lot of time compared to heated
layers. These show that the exponential model is best
air drying at 55 oC and 65 oC. Sun-drying of two and three
represented with the one/single layer drying data as in the
layers, took 2-3 days and by then the bottom cubes had
case of grains. However, this is interesting since the model
started to be discoloured (deteriorate) even before the EMC
has been useful to single layer drying of low moisture
was reached. For the ease of comparison only the Moisture
content crops that exhibited “falling-rate” regime (Nellist,
Ratio (MR) obtained on the first day of sun drying were
1976) while in this case, , initial moisture content of cassava
compared with MR obtained from heated air drying. These
used was 75.4 % wet basis. This was probably because in
results show that heated air drying is superior to sundrying,
this work there was a short rapid drying period (Fig. 3-5),
even at moderate temperatures, for example 55oC.
initially. For the two and three layers, the regression

coefficients were found to be high (r2>0.9) which is an
Thin-Layer Dying Models
indication that the data obeyed an exponential law, but

with decreasing approach to equilibrium moisture with
Moisture ratio data obtained at air temperatures of 55 and
increasing number of layers as shown by the values of k,
65oC for one layer, two layers and three layers of cassava
parallel with decreasing values of “A” (Table 1).
cubes were fitted into both the Page (equation 1) and the

exponential (equation 2) thin-layer drying models. From
However, for sun-drying of one layer, the deviation of “A”
the results of this study, the exponential model was
from 1 was about 14.6% and a value of r2 lower than the
modified to take the form (Verma et al., 1985)
rest (0.856) was obtained. This was due to the fluctuating
nature of solar radiation and the fact that the experiment
MR

=


A exp (

-
K t
)
was done continuously including day and night times.

The constants for equation 5 and regression coefficients
The page model was linearized to form the equation (Pathak
for one, two and three layers are presented in Table 1.
et al., 1991; Yunfei and Morey, 1987; Misra and Brooker,
1980; Yunfei et al., 1987)
ln
(-
ln

MR

=

)

ln K
+
N ln t

98
AJST, Vol. 2, No. 2: December, 2001

---

Thin- Layer Drying of Diced Cassava Roots
The constants for equation 6 and regression coefficients
Bruce D.M. 1985. Exposed Layer Barley Drying: Three
for one, two and three layers are presented in Table 2.
Models Fitted To New Data Up To 150 oC. J. Agric.

Engg. Res. 32:337-347.
Table 2. Constants and regression coefficients for the Page model.
CIAT. 1993. Cassava: The latest Facts About An Ancient
Crop. Cali, Colombia.
Temperature of the air (oC)
Colson K.H. and J.H.Young. 1990. Two-Component Thin-
55
65
Layer Drying Model for Unshelled Peanuts. TRANS.
ASAE, 33(1): 241-146.
1 layer 2 layers 3 layers
1 layer
2 layers 3 layers
Crisp J. and J.L. Woods. 1994. The Drying Properties of
A 0.061
0.6573
0.1237
0.1217
0.0649
0.1047
Rapeseed. J. Agric. Engg. Res. 57, 89-97.
FAO. 1985. Year Production Book, Rome, Italy.
K 0.909
-0.0264 0.7071
0.7087
0.9172
0.7826
Kachru R.P., T.P. Ojha and G.T. Kurp. 1980. Drying
Characteristics of Indian Paddy Varieties. J. Agric.
Engg. Res. 7:16-23.
r2 0.983
0.975
0.990
0.991
0.996
0.999
Misra M.K., and D.B. Brooker. 1980. Thin-Layer Drying
For sun-drying of one layer at 25 oC:
and Re-wetting Equations for Shelled Yellow Corn.
K = 0.0063,
N =
0.8557,
r2 = 0.925
TRANS. ASAE page 1254-1260.

Mlingi N.L.V.1995. Cassava Processinng and Dietary
The constant K shows a general increase with increasing
Cyanide Exposure inTanzania. PhD. Thesis, Faculty
number of layers, whereas N has shown a decreasing trend
of Medicine, University of Uppsala, Sweeden.
at one temperature (55 or 65oC). Moreover, both K and N
Nellist M.E. 1976. Exposed Layer Drying of Ryegrass Seeds.
values at 65 oC were generally higher than those at 55oC.
J. Agric. Engg. Res. 21, 49-66
For sun drying of one layer, these coefficients were different
Nghiem Q. 1991. Development of Milksap-Free Cassava
from the rest. However, the regression coefficients for this
Half-Product in Vietnam. In Scott G, Wiersema S. And
model were high (r2>0.9). From the empirical nature of the
P.I. Ferguson (eds). Proceedings of the International
Page equation, the high r2 values are an indication that the
workshop on Product Development for Root and
model fits all the drying trials satisfactorily. Also, the
Tuber Crops, Vol I, held in April 22-May 1, 1991
coefficients K and N can not be fixed and their values
Vasays State College of Agric, (VISCA) , Baybay,
depended very much on the prevailing experimental
Leyte, Phillipines.
conditions K and N.
N.R.I. 1994. Manual for Horticultural Export Quality

Assurance: Part II, page 8.3. Natural Resources
CONCLUSIONS
Institute, Overseas Development Administration,

London.
The exponential (thin-layer drying) model was found to be
Odogola O.W. 1988. The Potential of Cassava As a Cash
suitalbe to describe drying of one layer of 5 mm cassava
Crop For Small Holder Farmers Through The
cubes at a temperature range of 55 and 65 oC. On sun-
Development of Commercial and Industrial
drying of one layer, the model represent the data poorly,
Production and Their Markets. Paper presented at
but it is better than when used for two and three layers at
the workshop on the contribution of cassava to food
these temperatures. The Page model can be fairly used for
security in the member states of SADDC, Malawi,
all the conditions similar to the ones that prevailed in this
28th Nov-1st Dec, 1988.
work, but with variable values of the constants.
Pathak P.K., Y.C. Agrawal and B.P.N. Singh. 1991. Thin-

Layer Drying Model for Rapeseed. TRANS. ASAE,
REFERENCES
34(6):2505-2508.
Pattey E., P. Savoie and P.A. Dube. 1988. The Effect of a
Bangu N.T.A., S.T.A.R. Kajuna and G.S. Mittal. 1998.
Hay Tedder on The Field Drying Rate. Canadian
Storage And Loss Moduli For Various Stiff Porridges.
Agricultural Engineering. Vol. 30(1):43-50.
Proceedings of the Conference of The Tanzania
Sharp J.R. 1982. A Review of Low-Temperature Drying
Society Of Agricultural Engineers 17-19th November,
Simulation Models. J.Agric. Engg. Res. 27:169-190.
1998. TANESCO Kihonda, Morogoro, Tanzania. Vol.
Silayo V.C.K. 1995. Sundrying of Grains. Ph.D Thesis,
8 page 99-108.
University of Newcastle upon Tyne, UK.
AJST, Vol. 2, No. 2: December, 2001
99

---

S. T. A. R. KAJUNA
Sun D. and J.L Woods. 1994(a). Low-Temperature Moisture
Transfer Characteristics of Barley: Thin-Layer Models
and Equilibrium Isotherms. J.Agric. Engg. Res. 59:273-
283.
Sun D.W. and J.L. Woods. 1994(b). Low Temperature
Moisture Transfer Characteristics of Wheat in Thin
Layers. TRANS ASAE. Vol. 37(6):1919-1926.
Yunfei Li.R. and V. Morey. 1987. Thin-Layer Drying Rates
and Quality of Cultivated American Ginseng.
TRANS. ASAE. 30(3):842-847.
Yunfei Li.R., V. Morey and M. Afinrud.1987. Thin-Layer
Drying Rates of Oilseed Sunflower. TRANS. ASAE.
30(4):1172-1175.
Verma L.R., R.A. Bucklin, J.B. Endan and F.T. Wratten. 1985.
Effedts of Drying Air Parameters on Rice Drying
Models. TRANS. ASAE, 296-301.
Westerman P.W, G.M. White and J.J. Ross. 1973. Relative
Humidity Effect on The Hight Temperature Drying of
Shelled Corn. TRANS. ASAE, 16: 1136-1139.
100
AJST, Vol. 2, No. 2: December, 2001

---