Journal of Food Engineering 63 (2004) 125–134

www.elsevier.com/locate/jfoodeng

Effect of maltodextrin, glycerol monostearate and tricalcium

phosphate on vacuum dried mango powder properties
S. Jaya, H. Das *

Department of Agricultural and Food Engineering, Indian Institute of Technology, Kharagpur 721302, India
Received 3 October 2001; received in revised form 24 May 2002; accepted 13 April 2003

Abstract
Vacuum drying of mango pulp was carried out at various levels of maltodextrin (MD) ranging between 0.25 and 0.65 kg per kg of
mango solid. Glycerol monostearate (GMS) and tricalcium phosphate (TCP) were added at the levels varying between 0.01 and 0.02
kg per kg of mango solid. MD was used to eliminate the stickiness of the mango powder and to get less hygroscopic powder. GMS
was used as foam stabilizer and TCP as anticaking agents respectively. Hygroscopicity, degree of caking, dispersibility, flowability,
sticky point temperature of the dry powder at 5% (db) moisture content and overall color difference between the reconstituted
powder and the pulp were found out. Based on these properties of mango powder, an optimum feed mix composition of 0.43–0.57
kg MD per kg of mango solids was obtained. The optimum requirement for the TCP and GMS were found to be 0.015 kg per kg of
mango solid.
Ó 2003 Elsevier Ltd. All rights reserved.

1. Introduction (e.g. tricalcium phosphate, silicon dioxide, silicates,

phosphates, salts of stearic acid and modified carbohy-
Drying of sugar-rich mango pulp into powder is diffi- drates) are added to get nonsticky, free flowing powder
cult, mainly due to the low molecular weight sugars, such (Bhandari, Datta, Crooks, Howes, & Rigby, 1997; Peleg
as fructose (0.0265 kg per kg of pulp
1 ), glucose (0.0068 & Hollenbech, 1984; Roos, 1995; Roos & Karel, 1991c;
kg per kg of pulp
1 ), sucrose (0.095 kg per kg of pulp
1 ) Roos, Karel, & Kokini, 1996; Slade & Levine, 1994;
and acids (0.00346 kg per kg of pulp
1 as citric acid) Schwarz & Penn, 1948).
present in the pulp (Jagtiani, Chan, & Sakai, 1988). These Bhandari, Senoussi, Dumoulin, and Lebert (1993)
materials have low glass transition (sucrose: 62 °C, fruc- produced black current, raspberry and apricot juice
tose: )5 °C, glucose: 32 °C) temperature (Roos, 1995; powders by spray drying a mixture of the juice and
Roos & Karel, 1991a, 1991b, 1991c). Due to their low maltodextrin having dextrose equivalent (DE) 36 at
molecular weights (sucrose: 342, glucose: 180, fructose: 60 °C. They fixed the proportion of juice solid and
180) the molecular mobility of the materials is high when maltodextrin in the ratio 65:35, 55:45 and 60:40 respec-
the temperature is just above the glass transition tem- tively. Mizrahi, Berk, and Cogan (1967) used isolated
perature (Tg ). They are very hygroscopic in their amor- soya protein for preparing spray dried banana powder.
phous state and loose free flowing nature at high moisture Anticaking agents are added to improve the flow-
content. While drying at temperatures normally prevail- ability and to inhibit the tendency to cake (Peleg &
ing in spray dryers, they tend to stick to the walls of the Hollenbech, 1984). James (1971) used tricalcium phos-
dryer and finally give a paste like structure instead of phate, calcium silicate and calcium oxide at 0.15%,
powder (Bhandari, Datta, & Howes, 1997; Masters, 1985). 0.10% and 0.10% of wet mixture respectively, along with
To overcome these problems, addition of drying aids, 25–40% sucrose to produce powder from guava and
such as isolated protein, maltodextrin with different pineapple juices by using vacuum drying. He found that
dextrose equivalent (DE), food grade anticaking agents the addition of the above-mentioned anticaking agents
gave the desired free flowing property to the powder
during storage.
*
Corresponding author. During foam mat drying, the foam stabilizers, such as
E-mail address: hd@agfe.iitgp.ernet.in (H. Das). glycerol monostearate (GMS), carboxyl methyl cellulose,
0260-8774/$ - see front matter Ó 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0260-8774(03)00135-3
126 S. Jaya, H. Das / Journal of Food Engineering 63 (2004) 125–134

pectin and modified soya protein allowed the puffed it is subjected to increase its temperature. The sticky
structure to be retained (Bates, 1964; Morgan, Ginnette, point temperature measurement assesses the influence of
Randall, & Graham, 1959). Bates (1964) while studying temperature and moisture content on powder cohesion
the foam mat drying of some tropical fruit juices using and is applicable to nearly dry powders, which form
the above-mentioned foam inducers observed that 0.5% immobile liquid bridges of water present in the particles
of GMS (total solid basis) was able to produce foam (Papadakis & Babu, 1992).
density of 330–360 kg m
3 which was impossible to get Due to the absence of any specified range of values
by using modified soya protein. GMS was found to be for fruit juice powders, the above-mentioned properties
superior over other edible foam inducers for tomato juice of commercially available instant coffee and tomato
concentrates and aqueous food system by the authors. It soup powders were measured. Instant coffee powder was
was used in the present investigation for retaining the chosen for its high hygroscopicity. The tomato soup
puffed structure created in the initial stage of vacuum powder, being made of high molecular weight sub-
drying of mango pulp. stances like starch, maltodextrin, corn and wheat flour,
The method of drying used for this study was vacuum had its hygroscopicity lower than the instant coffee
drying. In this method since, the drying takes place in powder. It was considered that mango powder would
the absence of oxygen, the oxidative degradation e.g. have the desirable quality if its properties lie within the
browning is low in the final product. The rate of drying range of properties of the above-mentioned food pow-
is fast due to the creation of a frothy or puffed structure ders.
in the mango pulp. This expanded structure creates the The following were the objectives fixed for the present
desired property of ‘‘instant’’ reconstitution and pro- study.
vides large surface area to volume ratio for good heat
and mass transfer. The temperature range used for the ii(i) To produce vacuum dried mango powder having
vacuum drying is usually kept within 65–75 °C (Anon, various combinations of maltodextrin, glycerol
1952; Copley, Kaufman, & Rasmussen, 1956). monostearate and tricalcium phosphate.
The quality of food powders may be defined in terms i(ii) To measure the above-mentioned quality parame-
of (i) hygroscopicity, (ii) degree of caking, (iii) disper- ters in the mango powder as well as the commer-
sibility and (iv) flowability (Haugaard, Krag, Pisecky, & cially available tomato soup powder and instant
Westergaard, 1978; Pisecky, 1985), (v) sticky point coffee powder.
temperature (Lazar, Brown, Smith, Wang, & Lindquist, (iii) To find out the effect of maltodextrin, glycerol
1956) and (vi) resemblance of color of fresh juice monostearate and tricalcium phosphate on the
(Ranganna, 1987). quality of mango powder and to optimize the
(i) Hygroscopicity is the ability of food powder to amounts required for production of vacuum dried
absorb moisture from high relative humidity environ- mango powder having the desirable qualities.
ment. In the case of fruit powders, glucose and fructose
are responsible for strong interaction with the water
molecule due to the polar terminals present in these 2. Materials and methods
molecules (Haugaard et al., 1978; Slade & Levine, 1991).
(ii) Degree of caking is expressed as percentage of 2.1. Raw material
powder, which will remain on a sieve of defined mesh
size, after re-drying and sieving the cake formed in the The mango pulp of Totapuri variety was used for
hygroscopicity test (Pisecky, 1985). production of mango powder. Due to the nonavail-
(iii) Dispersibility is the ability of powder to get ability of fresh mango throughout the year, this pulp
wetted without formation of dry lumps in water. The was sterilized in tin cans (Tulip Products, Madhyam-
International Dairy Federation method of dispersibility gram, West Bengal) and used for the experiments. The
measurement specifies a stirring procedure following pulp was analyzed for total sugar, total solids, sucrose,
which the solution is poured through a 210 lm size sieve fructose, glucose, moisture content, titrable acidity, bulk
(Pisecky, 1985). density and color value (L, a and b values) as per the
(iv) Flowability is an important but complex property method prescribed by Ranganna (1987) and Sadasivam
controlled by particle size distribution, angle of repose and Manickam (1996). The experimental values are
and moisture content of powder. The flowability is presented in Table 1.
usually measured as the time needed for a certain vol-
ume of powder to pass through slots of a specified size in 2.2. Experimental design
a rotating drum (Pisecky, 1985).
(v) Sticky point temperature is the temperature at Based on the literature and the preliminary experi-
which a powder sample starts to stick to each other and ments conducted, the levels chosen for maltodextrin, X1 ,
also to the wall of the container in which it is kept when glycerol monostearate, X2 and tricalcium phosphate, X3
 S. Jaya, H. Das / Journal of Food Engineering 63 (2004) 125–134 127

Table 1 ðXmax
XM Þ
Average composition and physical properties of mango pulp XD ¼  ð2Þ
2
Constituents Amount present
Moisture content (kg per kg pulp) 0.80–0.85 ðX
XM Þ
Total solids (kg per kg pulp) 0.15–0.20 x¼ ð3Þ
XD
Sucrose (kg per kg pulp) 0.06–0.0636
Fructose (kg per kg pulp) 0.025–0.0255
Glucose (kg per kg pulp) 0.046–0.0465 X ¼ x  XD þ XM ð4Þ
Total sugar (kg per kg pulp) 0.131–0.1356
Acidity as anhydrous citric acid (kg per kg pulp) 0.00306–0.00346 where Xmax is the maximum value of X and its coded
Bulk density (kg/m3 ) 1015–1060 value is Ô+2’. Xmin is the minimum value of X and its
Color value
coded value is Ô)2’.
L 28.76 The experimental design has been shown in Table 2,
a )6.56 the design includes six repeated experiments at the
b 14.71 central points of the coded variables. This was necessary
for finding out the Ôerror sum of square’ and the Ôlack of
fit’ of the regression equations developed between the
were X1 : 0.25, 0.35, 0.45, 0.55 and 0.65; X2 : 0.01, 0.0125,
dependent ðyÞ and the independent ðxÞ variables (Myers,
0.015, 0.0175 and 0.02; X3 : 0.01, 0.0125, 0.015, 0.0175
1971). It may be observed from Table 2 that 21 experi-
and 0.02 kg per kg of dry mango solid.
ments were carried out.
The values of the independent variables ÔX ’ were co-
For each combinations of the independent variables
ded as the variables Ôx’ in the range of +2 and )2. The
in the experimental design the dependent parameters Ôy’
relationship between ÔX ’ and Ôx’ is given by the following
of the powder, were: hygroscopicity (HG), degree of
equations.
caking (DC), dispersibility (DS), flowability (FL), sticky
ðXmax þ Xmin Þ point temperature ðTs Þ and overall color difference
XM ¼ ð1Þ
2 ðDEc Þ.

Table 2
Experimental design for feed mix with coded variables
Coded values of independent Real values of independent variables
variables
Maltodextrin, X1 Glycerol monostearate, X2 Tricalcium phosphate, X3
(kg per kg of dry mango solid) (kg per kg of dry mango solid) (kg per kg of dry mango solid)
)2 0.25 0.01 0.01
)1 0.35 0.0125 0.0125
0 0.45 0.015 0.015
+1 0.55 0.0175 0.0175
+2 0.65 0.02 0.02

S. no. Experimental plan with coded form of independent variables

1 +1 +1 +1
2 +1 +1 )1
3 +1 )1 +1
4 +1 )1 )1
5 )1 +1 +1
6 )1 +1 )1
7 )1 )1 +1
8 )1 )1 )1
9* +2 0 0
10* )2 0 0
11 0 +2 0
12 0 )2 0
13 0 0 +2
14 0 0 )2
15 0 0 0
16 0 0 0
17 0 0 0
18 0 0 0
19 0 0 0
20 0 0 0
21 0 0 0
128 S. Jaya, H. Das / Journal of Food Engineering 63 (2004) 125–134

Table 3 glass wool. The chamber was connected with vacuum

Properties of instant coffee and tomato soup powder gauge (3) and vacuum release valve (4). A 3.7 kW,
Property Instant coffee Tomato soup 0.0236 m3 s
1 capacity water ring vacuum pump (5) was
powder powder used to create the desired vacuum of 710–730 mmHg
Moisture content % (wb) 4.18–5.25 3.4–4 (30–50 mmHg absolute pressure) in the system. Sealing
Hygroscopicity (%) 9.09–10.32 4.20–6.64 water was supplied to the vacuum pump from an over
Degree of caking (%) 31.27–34.48 8.96–10.23 head tank. The temperature of the water available was
Dispersibility (%) 99.95–99.98 68.19–70.21
Flowability (s) 15–18 36–42
at 28–32 °C. The pump was connected to the vacuum
Sticky point temperature 46  1 >70 chamber through a pipe of 0.0191 m diameter (13) and
at 5% (db) moisture (°C) an one way valve (14). A pump (6) was used for sup-
plying hot water from a hot water tank to the vacuum
chamber tank (2). Electrode boiler (8) was used to
To find out the effect of independent variables on the produce steam and the steam was used for heating the
quality of powder, which are the dependent variables, water kept in the hot water tank.
linear (Eq. 5) and nonlinear (Eq. 6) regression equations All flanges and joints were made air tight by using
were fitted between Ôy’ and Ôx’. neoprene rubber gaskets. Light and sight glasses were
y ¼ b0 þ b1  x 1 þ b2  x 2 þ b 3  x 3 ð5Þ provided in the vacuum chamber to observe the interior
of the vacuum chamber. Mango pulp was kept on a
y ¼ b0 þ b1  x1 þ b2  x2 þ b3  x3 þ b11  x21 þ b22  x22 Teflon coated tray for its drying inside the vacuum
þ b33  x23 þ b12  x1  x2 þ b13  x1  x3 þ b23  x2  x3 chamber. The temperatures of vacuum chamber plate,
Teflon coated tray and the pulp were sensed by thermo
ð6Þ
couples and were indicated by a digital temperature
where b0 ; b1 ; b2 ; b3 ; b11 ; . . . are regression coefficients and indicator through a selector switch. Controlling of the
x1 , x2 , and x3 are the coded values of the independent temperature of the vacuum chamber plate was done
variables X , viz., maltodextrin ðX1 Þ, glycerol monostea- manually by regulating the flow of the water from hot
rate ðX2 Þ and tricalcium phosphate ðX3 Þ respectively. water tank (16) to the vacuum chamber tank (2).
The desired range of the quality parameters of the
mango powder was fixed as those of instant coffee and 2.4. Method of grinding
tomato soup powders available in market. Table 3 gives
the above-mentioned range of quality parameters for the The dried product was kept in an environment
two powders. maintained at 5–6% relative humidity for a period of 30–
45 min and then ground in a hammer mill.
2.3. Drying of mango pulp Fig. 2 shows the setup used for creating the low rel-
ative humidity (5–6%) environment. Atmospheric air is
Fig. 1 shows the components of the dryer. The drying sucked through the saturated sodium hydroxide solution
chamber (1) was cylindrical (0.35 m diameter  0.45 m (4) kept in the glass vessel (5) by a diaphragm type
length) and it was made of stainless steel. The dryer was vacuum pump (3). The delivery side of the pump is
provided with inbuilt water tank (2) with drain valve connected to a bed of silica gel (2), which was kept in a
(15). Hot water was used as the heating medium. To cylindrical chamber (1). The dried sample was placed on
avoid the falling down of condensed vapours on the the perforated aluminum sheet (7). The chamber (1) was
material being dried inside the vacuum chamber, hot
water was supplied through pipes coiled over the surface
of the chamber. The entire chamber was insulated with

Fig. 2. Schematic diagram of conditioning apparatus for dried mango

Fig. 1. Schematic diagram of vacuum drying setup. flakes.
 S. Jaya, H. Das / Journal of Food Engineering 63 (2004) 125–134 129

covered with a lid (6), which had a small hole (4 mm

diameter) at its center through which the air escaped to
atmosphere. The apparatus was kept inside a room
maintained at 20–22 °C by using a room air conditioner.
The relative humidity of the air inside the room was 40–
50%.
A hammer mill was used for grinding the dried
product. The mill was fitted with a swinging hammers
and a stationary sieve having holes of 0.5 mm. The
weight of the powder obtained from the mill in a single
experiment varied between 15 and 60 g.

2.5. Measurement of physico-chemical properties

Fig. 3. Schematic diagram of hygroscopicity measurement apparatus.
Hygroscopicity, degree of caking, dispersibility and
flowability of mango powders were measured using the
methods explained by analytical Methods for Dry Milk 79.5  2% at 20 °C) was kept in a glass wash bottle (2)
Products (Haugaard et al., 1978; Pisecky, 1985). Sticky having two passages for air inlet and outlet. Powder
point temperature was measured as per method pro- sample weighing 0.5 g, which had passed through 500 l
posed by Lazar et al. (1956) and the overall color dif- sieve was spread uniformly in the Gooch filter (4), fitted
ference was measured as per the method given by to the conical glass flask (5). The Gooch filter was
Ranganna (1987). Three replicate measurements were covered by a box (6) made of Perspex sheet. A dia-
done and the average values of the properties are pre- phragm type vacuum pump (1) was used to suck the air
sented in Table 4. through the salt solution and the sample at a flow rate of
Hygroscopicity: It is expressed as the final moisture 300–400 ml per minute. The increase in weight of the
content attained after exposing the powder in humid air sample at every 15 min was noted. This measurement
having 79.5% relative humidity (Haugaard et al., 1978; was continued till the difference between two successive
Pisecky, 1985). weightings did not exceed by 0.5%. The entire operation
The apparatus used for the Hygroscopicity measure- was carried out in a room maintained at 20 °C.
ment has been shown in Fig. 3. A saturated solution of The hygroscopicity, HG (%) was calculated by using
potassium nitrite salt (equilibrium relative humidity ¼ Eq. (7)

Table 4
Quality parameters of mango powder
S. no. X1 X2 X3 HG DC DS FL DEc Ts
1 0.65 0.015 0.015 6.07 13 70.44 18 9.50 54
2 0.25 0.015 0.015 14.76 48.4 99.43 22 4.95 37
3 0.35 0.0125 0.0125 8.74 19.1 99.8 35 5.25 40
4 0.35 0.0125 0.0175 5.33 18.5 100 25.5 5.59 40.5
5 0.35 0.0175 0.0125 10.04 35 92.51 33.5 5.21 41.5
6 0.35 0.0175 0.0175 7.4 13.4 96.9 21 5.08 42
7 0.55 0.0125 0.0125 5.094 25.92 72 30 8.55 48
8 0.55 0.0125 0.0175 6.57 25.14 79.53 22 8.21 48.5
9 0.55 0.0175 0.0125 6.7 25.02 74.62 26 8.68 49
10 0.55 0.0175 0.0175 5.26 24.9 81.08 25.5 8.83 49.5
11 0.45 0.015 0.01 10.29 33.4 99.01 33 6.68 41
12 0.45 0.015 0.02 7.035 22 83.85 15 6.75 45
13 0.45 0.01 0.015 3.29 36.11 98.97 22 6.76 42
14 0.45 0.02 0.015 5.83 35.17 90.6 20.5 6.15 45
15 0.45 0.015 0.015 10.33 9.4 81.05 23 6.07 46.5
16 0.45 0.015 0.015 10.3 9.3 81.09 22.5 6.05 46
17 0.45 0.015 0.015 10.32 9.5 82 23.5 6.03 47
18 0.45 0.015 0.015 10.29 9.6 80.5 24 6.04 46.75
19 0.45 0.015 0.015 10.28 9.2 81.5 23 6.02 46.25
20 0.45 0.015 0.015 10.3 9.3 81.03 23.5 6.01 46
21 0.45 0.015 0.015 10.3 9.5 81.02 22.5 6.08 47

X1 , X2 and X3 (kg per kg dry mango solid) are respectively the amount of maltodextrin, glycerol monostearate and tricalcium phosphate added to the
mango pulp.
130 S. Jaya, H. Das / Journal of Food Engineering 63 (2004) 125–134

b=a þ Wi
HG ¼ ð7Þ
1 þ b=a
where b (g) is the increase in weight of powder, a (g) is
the amount of powder taken for the measurement and
Wi (%, wb) is the free water present in the powder before
measurement.
Degree of caking: After the determination of hygro-
scopicity, the Gooch filter along with the wet sample
was placed in a drying oven set at 102  2 °C for 1 h
(Haugaard et al., 1978; Pisecky, 1985). After cooling the
dried sample, it was weighed and transferred into a sieve
of 500 l size. The sieve was then shacked for 5 min in a
shaking apparatus. The weight of the powder remaining
in the sieve was measured.
The degree of caking, DC (%) was calculated by using
Eq. (8)
DC ¼ c=d  100 ð8Þ
Fig. 4. Schematic diagram of flowability measurement apparatus.
where d (g) is amount of the powder used sieving and c
(g) is amount of the powder left on the sieve after siev-
Sticky point temperature: A sticky point apparatus
ing.
developed by Lazar et al. (1956) has been in use since
Dispersibility: Thirteen grams of mango powder was
then by several others (Buhler & Liedy, 1989; Downton,
stirred with 100 g of water at 50 °C in a beaker
Flores-Luna, & King, 1982; Wallack & King, 1988).
(Haugaard et al., 1978; Pisecky, 1985). Stirring was
A slightly modified form of the apparatus (Fig. 5)
carried out to make 25 complete back and froth move-
developed by the present authors was used for the
ments across the whole diameter of the beaker for 15 s.
The reconstituted mango powder was then poured
through the sieve of 210 l size. The dry matter of the
filtered juice was estimated by measuring its moisture
content using Metler moisture analyzer (Mettler LJ16).
The dispersibility, (DS) was calculated by using Eq.
(9)
ðW þ aÞSp
DS ¼ ð9Þ
a  Sj
where a (g) is the amount of powder used, W is the
weight of water taken for reconstitution, Sp (%) is the
total solid present in the mango powder and Sj is the %
dry matter present in the reconstituted mango juice after
it has been passed through the sieve.
Flowability: Schematic diagram of the apparatus used
for the flowability measurement has been shown in Fig.
4. It consists of a stainless steel drum (1) of diameter 12
cm and length 9 cm. One end of the drum was fitted with
a transparent lid (3), while its other end was rigidly fixed
to the shaft of motor (2). The drum had two slots, each
having 4 mm width and 7 cm length on the surface of the
drum. Powder sample weighing 25 times of its bulk
density in g cm
3 was kept in the drum. The drum was
rotated at 30 rpm by a geared motor and the powder
was allowed to flow through the slits. Time required for
all the powder to come out of the slits provided on the
drum was noted. Flowability, FL (s) was expressed as
the time in seconds necessary for the powder to leave the Fig. 5. Schematic diagram of sticky point temperature measurement
rotary drum (Haugaard et al., 1978; Pisecky, 1985). apparatus.
 S. Jaya, H. Das / Journal of Food Engineering 63 (2004) 125–134 131

measurement of sticky point temperature, Ts (°C). It pulp used for powder production and the reconstituted
consists of a sample holder (1) made of glass for holding powder was calculated by using Eq. (10).
the powder sample (2) of known moisture content. A qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
copper–constantan type thermocouple was used to sense 2
DEc ¼ ðL0
Lp Þ þ ða0
ap Þ þ ðb0
bp Þ
2 2
ð10Þ
the powder temperature, which was indicated by a
digital temperature indicator. The thermocouple wire where L0 , a0 and b0 , are the L, a, and b values of the
was taken out of the sample holder by the method mango pulp and Lp , ap and bp are corresponding values
shown in Fig. 5. of the reconstituted powder.
Sealing liquid holder (3) fitted with a glass tube (4) Moisture content: Moisture content of the powder
was inserted inside the sample holder (1). Mercury was was measured by using Metler moisture analyzer (Met-
filled into this sealing liquid holder to a height of about tler LJ16), by placing the sample in the sample pan and
1.5 cm. The sample could be isolated from atmosphere evaporating the moisture at 100 °C (as given in Mettler
via this mercury seal. A sealing tube (5) was inserted into Manual for fruits) by infrared heating arrangement. The
the liquid holder (3) to the required depth decided by the value of moisture content was checked by the standard
position of the Teflon ring stopper (6) such that the AOAC method. It was observed that there was no dif-
bottom of the tube (5) could be dipped into the mercury ference between the two methods.
seal. The stopper (6) was fitted to the outside of the The above mentioned properties, in three replicate
sealing tube and its position on the sealing tube was measurements, were obtained for tomato soup and in-
fixed by use of an adhesive such that the lower surface of stant coffee powders, except the overall colour differ-
the stopper (6) touched the upper edge of the sealing ence. The average of three replications has been given in
liquid holder (3). Table 3.
A stainless steel stirring rod (7) fitted with a rubber
cork (8) was then inserted and tightly fitted to the sealing
tube (5) such that the rod (7) and the sealing tube (5)
could rotate together while rotating the sealing tube (5). 3. Results and discussions
The sealing tube (5) and the stirring rod (7) assembly
were then inserted into the sealing liquid holder (3). The From Table 1 it may be observed that the typical
length of the stirring rod (7) was such that its lower end values of fructose (0.0265 kg per kg of pulp
1 ), glucose
came out of the bottom of the glass tube (4). (0.0388 kg per kg of pulp
1 ), sucrose (0.075 kg per kg of
The propeller (9) was then fitted to the threaded pulp
1 ) and acids such as citric acid (0.00346 kg per kg
bottom of the stirring rod (7). The assembly consisting of pulp
1 ) content of mango pulp reported by Jagtiani
of the sealing liquid holder (3), sealing tube (5), stirring et al. (1988) are close to the measured values of the
rod (7) and propeller (9) was inserted into the sample Totapuri variety of mango pulp used in the experiments.
holder (1). Since the diameter of the propeller (9) was
slightly lower than the inside diameter of the sample 3.1. Properties of instant coffee and tomato soup powders
holder (1) the propeller could rotate freely inside the
sample holder through rotation of the sealing tube (5). The measured values of the quality parameters such
The whole assembly consisting of sample holder (1), as hygroscopicity, (%), degree of caking, (%), dispersi-
powder sample (2), sealing liquid holder (3), sealing tube bility, (%), flowability, (s), and sticky point temperature,
(5), stirring rod (7) and propeller (9) was then immersed (°C) of the instant coffee and tomato soup powders are
into a constant temperature water bath. The tempera- given in Table 3.
ture of the bath was slowly raised while the powder was From Table 3 the desired range of the quality
intermittently stirred by hand by rotating the sealing parameters of mango powder was fixed based on the
tube (5) and the temperature of the powder noted by the average of range of each property obtained for the in-
digital temperature indicator. At a certain temperature stant coffee and tomato soup powders. From that
of powder, the force required to turn the propeller average the minimum was fixed as lower range and the
rotating inside the powder increased dramatically. The maximum as upper range of mango powder (e.g. HG:
temperature at which this occurred was usually distinct ð4:206 þ 6:64Þ=2 ¼ 5:13 is the lower range, ð9:091 þ
and is noted as the sticky point temperature of the 10:322Þ=2 ¼ 9:38 is upper range). These values are given
powder. in Table 5. The range of overall color difference (DEc ) of
Color of powder: An amount of powder having same mango powder was fixed from the maximum and min-
level of total solids as in the mango pulp used for powder imum color difference obtained from the experiments
production was reconstituted with water at 30 °C. The L, (Table 4). The Ôdesirable’ values were used to find out
a, and b values of the reconstituted powder were mea- the optimum range of composition of feed mix con-
sured by using Hunter Lab colorimeter (Hunter lab taining maltodextrin, glycerol monostearate and trical-
D25). The total color difference (DEc ) between the mango cium phosphate for production of mango powder.
132 S. Jaya, H. Das / Journal of Food Engineering 63 (2004) 125–134

Table 5 By observing the sign and magnitude of coefficients of

Desired range of properties for mango powder x1 , x2 and x3 in Eqs. (11)–(16) we may arrive at the
S. no. Property of mango powder Desired range following conclusions.
1 Hygroscopicity (HG%) 5.13–9.38 Eq. (11): The hygroscopicity will increase by
2 Degree of caking (DC%) 8.96–33.97 decreasing the amount of maltodextrin and tricalcium
3 Dispersibility (DS%) 67.05–99.98 phosphate and will increase by increasing the amount of
4 Flowability (FL, s) 17.53–35.27 glycerol monostearate. The effect of maltodextrin (1.58)
5 Sticky point temperature (at 5% dry 45.78–53
basis moisture content, Ts )
is the highest and is followed by tricalcium phosphate
6 Overall color difference (DEc ) 4.95–5.08 (0.783) and glycerol monostearate (0.547).
Eq. (12): There will be a decrease in degree of caking
with increase in amount of all the three ingredients. The
tricalcium phosphate has got more effect on percentage
3.2. Experimental values of mango powder quality
degree of caking with the value of 3.49 followed by
maltodextrin with 2.864 and then by glycerol monoste-
For the various combinations of maltodextrin, gly-
arate with 0.138. The effect of glycerol monostearate will
cerol monostearate and tricalcium phosphate mentioned
be about 21–26 times lower than maltodextrin and tri-
in Table 2, the measured values of quality parameters
calcium phosphate added to the pulp.
(viz., hygroscopicity, (HG, %) degree of caking, (DC, %)
Eq. (13): It shows that decreasing all the additives will
dispersibility, (DS, %) flowability, (FL, %) sticky point
increase the dispersibility. Maltodextrin, glycerol mon-
temperature (Ts , °C) and overall color difference) of
ostearate and tricalcium phosphate have the coefficient
mango powder are given in Table 4.
values of 8.74, 1.435 and 0.734 respectively. Maltodex-
trin, therefore, will have the highest influence followed
3.3. Effect of feed mix formulations on mango powder by glycerol monostearate and tricalcium phosphate.
quality Eq. (14): The flow time will decrease by increasing the
amount of maltodextrin, glycerol monostearate and
Linear regression equations using the coded values tricalcium phosphate. The effect of tricalcium phosphate
(x1 , x2 , and x3 ) of the independent variables (X1 , X2 and (4.16) is the highest, followed by maltodextrin (1.22) and
X3 ) has been developed to find out the effect of malto- glycerol monostearate (0.6).
dextrin (x1 ), glycerol monostearate, (x2 ) and tricalcium From Eq. (15) we find that there will be decrease in
phosphate (x3 ) on the different powder qualities (viz., overall color difference due to the increase in amount of
hygroscopicity (HG), degree of caking (DC), dispersi- glycerol monostearate and decrease in tricalcium phos-
bility (DS), flowability (FL), over all color difference phate and maltodextrin. The effect of maltodextrin (1.4)
(DEc ) and sticky point temperature (Ts )). The equations is the highest, followed by glycerol monostearate (0.06)
developed are: and tricalcium phosphate (0.01).
HG ¼ 8:311
1:58x1 þ 0:547x2
0:783x3 ; Eq. (16) indicates that there the sticky point tempera-
ture will increase by increasing the amount of malto-
r2 ¼ 0:94 ð11Þ dextrin, glycerol monostearate and tricalcium phosphate.
The effect of maltodextrin (4.06) is the highest, followed
DC ¼ 20:518
2:864x1
0:138x2
3:49x3 ;
by glycerol monostearate (0.688) and tricalcium phos-
r2 ¼ 0:78 ð12Þ phate (0.625).
Since amorphous sugars are generally hygroscopic,
DS ¼ 86:045
8:74x1
1:435x2
0:734x3 ;
they easily cause stickiness by reducing the value of glass
r2 ¼ 0:99 ð13Þ transition tempertaure (Tg ). Roos and Karel (1991d,
2 1991e) and Slade and Levine (1991) reported that sticky
FL ¼ 23:96
1:22x1
0:6x2
4:16x3 ; r ¼ 0:85 ð14Þ
point temperature of food powders decreased with de-
2
DEc ¼ 6:09 þ 1:4x1
0:06x2 þ 0:01x3 ; r ¼ 0:97 ð15Þ crease in molecular weight and Tg value. Levine and
Ts ¼ 45:167 þ 4:062x1 þ 0:688x2 þ 0:625x3 ; Slade (1986) and Roos and Karel (1991c, 1991d) re-
ported that materials such as fruit juice powders with
r2 ¼ 0:99 ð16Þ high amount of monosaccharides (e.g. fructose) exhibit
low Tg values and thereby low sticky point temperature.
The relationship between the real (X ) and the coded (x) Stickiness of such products can be reduced and stability
values are: increased by adding compounds having higher Tg value.
X1 ¼ 0:1x1 þ 0:45; ð17Þ Since the Tg value increases with increasing molecular
weight, the high molecular weight amorphous material
X2 ¼ 0:0025x2 þ 0:015; ð18Þ
like maltodextrin is generally used to increase the glass
X3 ¼ 0:0025x3 þ 0:015; ð19Þ transition temperature, decrease stickiness and improve
 S. Jaya, H. Das / Journal of Food Engineering 63 (2004) 125–134 133

product stability. In confirmation to the above obser- The range of optimum values of x1 , x2 and x3 at which
vations, we find from Eq. (16) that the effect of malto- the desired values (Table 5) of hygroscopicity
dextrin on sticky point temperature is 5.9–6.5 times (5:13 6 HG 6 9:38), dispersibility (67:05 6 DS 6 99:98),
more than that of the amount of glycerol monostearate degree of caking (8:96 6 DC 6 33:97), sticky point tem-
and tricalcium phosphate added to the pulp. perature (45:78 6 Ts 6 53), flowability (17:53 6 FL 6
By observing the coefficient values of the individual 35:27) and overall color difference (4:95 6 DEc 6 5:08)
variables in all Eqs. (11)–(16) we find that maltodextrin can be obtained were found out by solving the above
and tricalcium phosphate are affecting the powder equations using Micro Soft Excel Solver. Table 6 gives
quality more than that of the glycerol monostearate. It the optimum range of coded (x1 , x2 and x3 ) values of
was also observed from the experiments carried out that maltodextrin, glycerol monostearate and tricalcium
the glycerol monostearate was affecting the foam for- phosphate.
mation and stability only during drying instead of The range of values of x1 , x2 and x3 that lie within the
affecting the powder quality. maximum and minimum values of x1 , x2 and x3 men-
tioned in Table 6 are:
3.4. Optimization of feed mix Maltodextrin :
0:21 < x1 < 1:0 ð26Þ
Glycerol monostearate :
0:12 < x2 < 0:28 ð27Þ
In Table 5 we have mentioned that the desired range of
mango powder properties. In order to find out the range Tricalcium phosphate :
0:12 < x3 < 0:25 ð28Þ
of maltodextrin (X1 ), glycerol monostearate, (X2 ) and The above-mentioned range of maltodextrin (x1 ), gly-
tricalcium phosphate (X3 ) required to attain the range of cerol monostearate (x2 ) and tricalcium phosphate (x3 )
powder properties mentioned in Table 5, second order are given in coded form. The real range (viz., X1 , X2 and
regression equations were fitted with the coded values of X3 ) of the ingredients were obtained by using Eqs. (17)–
X1 , X2 , and X3 by using the experimental values of the (19) and they are given below.
properties given in Table 4. The regression equations are
given in Eqs. (20)–(25). The second order regression Maltodextrin (kg per kg dry mango solid):
coefficients, which do not have the significant contribu- 0:429 < X1 < 0:55 ð29Þ
tions to the response were not included in these equations.
Glycerol monostearate (kg per kg dry mango solid):
HG ¼ 9:92
1:58x1 þ 0:55x2
0:783x3 þ 0:761x1 x3 0:0147 < X2 < 0:0157 ð30Þ

1:57x22
0:54x23 ; r2 ¼ 0:93 ð20Þ
Tricalcium phosphate (kg per kg dry mango solid):
DS ¼ 81:01
8:75x1
1:43x2 þ 0:841x21 þ 1:82x1 x2 0:0147 < X3 < 0:0156 ð31Þ
þ 3:3x22 þ 2:5x23 ; r2 ¼ 0:94 ð21Þ
From Eq. (29) we observe that the amount of maltod-
DC ¼ 8:81
2:86x1
3:5x3 þ 4:96x21 þ 3:9x1 x2 þ 6:19x22 extrin required for getting the desired properties of

3:79x2 x3 þ 4:2x23 ; r2 ¼ 0:86 ð22Þ mango powder should lie within 0.43–0.55 kg per kg dry
mango solid and the requirement of tricalcium phos-
Ts ¼ 46:54 þ 4:1x1 þ 0:69x2 þ 0:63x3
0:23x21
0:73x22 phate and glycerol monostearate were 0.015 kg each per

0:85x23 ; r2 ¼ 0:99 ð23Þ kg of dry mango solids (Eqs. 30 and 31).
Using the optimum ranges of the independent vari-
FL ¼ 23:11 þ 7:029x1 þ 2:32x21 þ 0:375x1 x2
0:5x1 x3 ables, x1 , x2 and x3 (Eqs. (26)–(28)) the values of
þ 1:17x22 þ 0:814x23 ; r2 ¼ 0:85 ð24Þ hygroscopicity, HG (%), dispersibility, DS (%), degree
of caking, DC (%), sticky point temperature, Ts (°C),
DEc ¼ 6:03
0:96x1
0:755x2 x3
1:114x23 ; r2 ¼ 0:97 flowability, FL (s) and overall color difference, DEc of
ð25Þ mango powder was obtained from Eqs. (20)–(25). The

Table 6
Maximum and minimum coded values of maltodextrin, glycerol monostearate and tricalcium phosphate
Properties Maltodextrin, x1 Glycerol monostearate, x2 Tricalcium phosphate, x3
Maximum Minimum Maximum Minimum Maximum Minimum
Hygroscopicity, % 1.682
0.587 1.21
1.03 0.34
0.12
Dispersibility, % 1.682
1.682 0.55
0.272 0.324
0.125
Degree of caking, % 1.43
0.245 1.56
0.15 1.5339
0.724
Sticky point temperature, C 1.67
0.21 0.282
0.12 0.257
0.124
Flowability, s 1.682
1.674 1.55
0.284 1.25
0.24
Overall color difference 1
2 0.55
0.12 0.25
0.5
134 S. Jaya, H. Das / Journal of Food Engineering 63 (2004) 125–134

Table 7 James, R. (1971). Vacuum puff freeze drying of tropical fruit juices.
Estimated values of quality parameters of mango powder obtained Journal of Food Science, 36, 906–910.
from the optimized combinations of the ingredients Lazar, M. E., Brown, A. H., Smith, G. S., Wang, F. F., & Lindquist,
F. E. (1956). Experimental production of tomato powder by spray
S. no. Quality parameters Estimated values
drying. Food Technology, 13, 129–134.
1 Hygroscopicity (HG%) 8.33–10.27 Levine, H., & Slade, L. (1986). A polymer physico-chemical approach
2 Degree of caking (DC%) 10.24–11.5 to the study of commercial starch hydrolysis products (SHPs ).
3 Dispersibility (DS%) 73.63–83.19 Carbon polymer, 6, 213.
4 Flowability (FL, s) 22–24.56 Masters, K. (1985). Spray drying handbook (4th ed.). England:
5 Sticky point temperature (at 5% dry 45.5–50.65 Longman Scientific and Technical.
basis moisture content, Ts ) Mizrahi, S., Berk, Z., & Cogan, U. (1967). Isolated Soya Protein as a
6 Overall color difference (DEc ) 5.8–7.9 banana spray drying aids. Cereal Science Today, 12(8), 322–325.
Morgan, A. I., Jr., Ginnette, L. F., Randall, J. M., & Graham, R. P.
(1959). Technique for improving instants. Food Engineering, 31(9),
86–87.
estimated values of the above parameters are given in
Myers, P. H. (1971). Response surface methodology. Boston: Allyn and
Table 7. Bacon.
We observe from Table 7 that except the overall color Papadakis, S. E., & Babu, R. E. (1992). The sticky issues of drying.
difference, DEc all the values of the above-mentioned Drying Technology, 10, 817–837.
quality parameters either very close or within the range Peleg, M., & Hollenbech, A. M. (1984). Flow conditioners and anti
caking agents. Food Technology, 38(3), 93–102.
of Ôdesired’ values presented earlier in Table 4. The value
Pisecky, J. (1985). Standards, specifications and test methods for dry
of DEc was higher than the Ôdesired’ value possibly due milk products. In M. C. Diarmuid (Ed.), Concentration and drying
to the addition of maltodextrin in the pulp. We come to of foods (pp. 203–220). New York: Elsevier Science Publishing Co.
this conclusion by observing Eq. (15), which shows that Ranganna, S. (1987). Hand book of analysis and quality control for fruit
the coefficient of maltodextrin in the equation is positive and vegetable products. New Delhi: Tata McGraw-Hill Publica-
tions.
and higher than the coefficient of tricalcium phosphate
Roos, Y. H. (1995). Glass transition-related physicochemical changes
glycerol monostearate. in foods. Food Technology, 49(10), 97–102.
Roos, Y. H., & Karel, M. (1991a). Phase transitions of mixtures of
amorphous polysaccharides and sugars. Biotechnological Progress,
References 7, 49–53.
Roos, Y. H., & Karel, M. (1991b). Plasticizing effect of water on
Anon (1952). In Torrey, M. (Ed.), Dehydration of fruits and vegetables thermal behavior and crystallization of amorphous food models.
(pp. 66–80). Food Technology Review no. 13. Park Ridge, New Journal of Food Science, 56, 38–43.
Jersey: Noyes Data Corporation. Roos, Y. H., & Karel, M. (1991c). Phase transition of amorphous
Bates, R. P. (1964). Factors affecting foam production and stabiliza- sucrose and sucrose solution. Journal of Food Science, 56, 38–49.
tion of tropical fruit products. Food Technology, 18(1), 93–96. Roos, Y., & Karel, M. (1991d). Water and molecular weight effects on
Bhandari, B. R., Datta, N., Crooks, R., Howes, T., & Rigby, S. (1997). glass transitions in amorphous carbohydrates and carbohydrate
A semi-empirical approach to optimize the quantity of drying aid solutions. Journal of Food Science, 56(1), 1676–1681.
required to Spray dry sugar-rich foods. Drying Technology, 15, Roos, Y., & Karel, M. (1991e). Amorphous state and delayed ice
2509–2525. formation in sucrose solutions. International Journal of Food
Bhandari, B. R., Datta, N., & Howes, T. (1997). Problem associated Science and Technology., 56, 1676–1681.
with spray drying of sugar-rich foods. Drying Technology, 15(2), Roos, Y. H., Karel, M., & Kokini, J. L. (1996). Glass transition in low
671–684. moisture and frozen foods. Journal of Food Technology, 11, 95–107.
Bhandari, B. R., Senoussi, A., Dumoulin, E. D., & Lebert, A. (1993). Sadasivam, S., & Manickam, A. (1996). Biochemical Methods (pp. 1–
Spray drying of concentrated fruit juices. Drying Technology, 11(5), 21). New Delhi: New Age International.
33–41. Schwarz, H. W., & Penn, F. E. (1948). Production of orange juice
Buhler, W., & Liedy, W. (1989). Characterization of product qualities concentrate and powder. Industrial Engineering Chemistry, 40(5),
and its application in drying process development. Chemical 938–944.
Engineering Process, 26(1), 27–34. Slade, L., & Levine, H. (1991). Beyond water activity: Recent advances
Copley, M. J., Kaufman, V. F., & Rasmussen, C. L. (1956). Recent based on an alternative approach to the assessment of food quality
development in fruit and vegetable powder technology. Food and safety. Critical Reviews in Food Science and Nutrition., 30, 115–
Technology, 13, 589–594. 360.
Downton, G. E., Flores-Luna, J. L., & King, C. J. (1982). Mechanism Slade, L., & Levine, H. (1994). Water and the glass transition––
of stickiness in hygroscopic amorphous powders. Industrial Chem- Dependence of the glass transition on composition and chemical
istry Fundamentals, 21, 447–451. structure: Special implications for flour functionality in cookie
Haugaard, I. S., Krag, J., Pisecky, J., & Westergaard, V. (1978). baking. Journal of Food Engineering, 22, 143–188.
Analytical methods for dry milk powders. Denmark: Niro Atomizer. Wallack, D. A., & King, C. J. (1988). Sticking and agglomeration of
Jagtiani, J., Chan, H.T., & Sakai, S.W. (1988). Tropical Fruit hygroscopic, amorphous carbohydrate and food powders. Biotech-
Processing (pp. 45–97). San Diego: Academic Press. nology Progress, 4(1), 31–35.