Optimization of Maltodextrin and Tricalcium Phosphate for Producing Vacuum Dried Mango Powder

Drying of mango pulp into powder is a challenging operation, mainly due to the sticky issue of mango powder in the dryer and caking during handling and storage. To overcome the above problem, maltodextrin, MD (drying aid) and Tricalcium Phosphate, TCP (anti caking agent) were added to the mango pulp at three levels and vacuum dried. The dried powder was analysed for properties such as hygroscopicity, degree of caking, flowability and sticky point temperature. Factorial design of experiment was used to investigate the effect of MD and TCP on the mango powder properties. The amount of MD and TCP required to reduce powder stickiness and caking were optimized based on the powder properties. At the optimum amount of MD (0.527 kg per kg dry mango solid) and TCP (0.0167 kg per kg dry mango solid), the values of mango powder properties were: hygroscopicity = 6.4 %; degree of caking = 7.8%; flowability = 18.6 s; sticky point temperature = 47.4°C.

Keywords:

• Mango powder
• Stickiness
• Vacuum drying
• Food powder properties

INTRODUCTION

Production of fruit powders has many research challenges to improve process technologies and powder properties. One of the challenges of producing fruit powders is to lessen the stickiness of the powders for safe handling and storage. Stickiness of fruit powders are mainly due to the presence of low molecular weight sugars, such as fructose, glucose, sucrose and some organic acids present in the fruit pulp. These sugars and organic acids have low glass transition temperature and high molecular mobility at high temperatures[1,2]. They are also very hygroscopic in their amorphous state and loose their free flowing nature at high moisture content. While drying at temperatures about 150–180°C in the spray dryers, dried powder tend to stick at the bottom of the dryer walls and finally produce a paste like material instead of powder[3–5]. In order to reduce the stickiness of fruit powders, drying aids such as isolated protein, maltodextrin with different dextrose equivalent (DE) are added to produce non sticky, free flowing powders[1,6–12].

To improve the handling properties of the powders, food grade anti caking agents (e.g. tricalcium Phosphate, silicon dioxide, silicates, phosphates, salts of stearic acid and modified carbohydrates) are also often added along with drying aids. Anti caking agents are added to improve the flowability and to inhibit the tendency to cake[7]. James[13] used tricalcium phosphate, calcium silicate and calcium oxide at 0.15%, 0.10% and 0.10% of wet mixture respectively, along with 25 to 40% sucrose to produce powder from guava and pineapple juices using vacuum drying. He found that the addition of the above mentioned anti caking agents produced the desired free flowing property to the powder during storage. However, addition of drying aids and anti caking agents to each fruit pulps varies with its chemical composition and drying behaviour[14]. Addition of such materials must be optimized for each fruit pulps or juices to produce non-sticky fruit powders. Therefore, the main objectives of this present study were 1) to investigate the effect of maltodextrin (MD) and tricalcium phosphate (TCP) addition to mango pulp on the properties of mango powders such as sticky point temperature, hygroscopicity, degree of caking and flowability and 2) to optimize the amount of MD and TCP required to improve the powder properties.

MATERIALS AND METHODS

Raw Material

The mango pulp of Totapuri variety was used for production of mango powder. Due to the non-availability of fresh mango throughout the year, canned mango pulp (Tulip Products, Madhyamgram, WB, India) was used for all the experiments. The pulp was analyzed for total solids and moisture content as per the method prescribed by AOAC[15].

Production of Fruit Powder

Fruit pulps of about 250 g along with the required amount of additives such as MD and TCP was mixed using a laboratory model planetary mixer having its mixing arm operated at 75 m.min⁻¹ peripheral speed for 15 to 20 minutes. The mix was heated to 70°C, spread to a thickness of 3 mm on a Teflon coated tray and kept inside the vacuum chamber and dried under vacuum and temperature level of 710–750 mm Hg vacuum and 70 ± 2°C[16,17] respectively. A water ring type vacuum pump that was fitted to the dryer, could remove water vapour from the pulp during drying. While the water vapour was continuously removed from the dryer, the rate of drying was high in comparison to that obtained by a vacuum oven. The rate of drying of mango pulp mix using this vacuum drier was measured by conducting drying experiment with optimized amount of maltodextrin and tricalcium phosphate[18].

The dried pulp was removed from the dryer and kept in atmospheric air (80–83% RH) to attain the atmospheric temperature (26–32°C). Even though the dry product was obtained at very low moisture content (1–3% wb), it was lathery (but non-sticky) and found difficult to grind in to powder in the atmospheric condition.

Method of Grinding

Since the dry product was found difficult to grind, it was kept in an environment maintained at 5–6% relative humidity for a period of 30–45 minutes and then ground in a hammer mill[14]. Figure 1 shows the set up used for creating the low relative humidity (5–6%) environment. Atmospheric air is sucked through the saturated Sodium Hydroxide solution (4) kept in the glass vessel (5) by a diaphragm type vacuum pump (3). The delivery side of the pump is connected to a bed of silica gel (2), which was kept in a cylindrical chamber (1). The dried sample was placed on the perforated aluminium sheet (7). The chamber (1) was 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 powder was collected immediately in aluminium foil laminated flexible pouches. Pouches were heat sealed at atmospheric pressure using a heat- sealing machine.

Figure 1. Schematic diagram of conditioning apparatus for dried mango flakes.

Experimental Design

Based on the preliminary experimental results, the levels chosen for Maltodextrin, X_1, and Tricalcium phosphate X₂, were X₁: 0.35, 0.45 and 0.55; X₂: 0.0125, 0.015 and 0.0175 kg per kg of dry mango solid. The experimental design has been shown in Table 1. This design includes nine experiments and repeated for three times. Factorial design of experiment[20] was used, where the values of the independent variables ‘X’ were coded as the variables, ‘x’ in the range of +1 and –1. The relationship between ‘X’ and ‘x’ is given by the following equations.

(1)

(2)

(3)

(4)

Where, X_max is the maximum value of X and its coded value is ‘+ 1’. X_min is the minimum value of X and its coded value is ‘- 1’. For each combination of the independent variables in the experimental design, the dependent parameters ‘y’ of the powder were: Hygroscopicity (HG), Degree of caking (DC), Flowability (FL) and Sticky point temperature (T_s). To find out the effect of independent variables on the powder quality (dependent variables) and optimize the independent variables, Linear (Eq. 5) and Non-linear (Eq. 6) regression equations were fitted between ‘y’ and ‘x’.

(5)

(6)

where b_0, b₁, b₂, b₃, b₁₁ …… are regression coefficients and x_1, and x₂ are the coded values of the independent variables X, viz., maltodextrin (X₁) and tricalcium phosphate (X₂) respectively.

Table 1 Experimental design of MD and TCP variables

	Real values of independent variables
Coded values of independent variables	Maltodextrin, X1 (kg per kg dry mango solid)	Tricalcium phosphate, X2, (kg perkg dry mango solid)
−1	0.35	0.0125
0	0.45	0.0150
+1	0.55	0.0175
Serial no.	Experimental plan with coded values of the variables
1	+1	+1
2	+1	−1
3	−1	+1
4	−1	−1
5	−1	0
6	+1	0
7	0	−1
8	0	+1
9	0	0

Analysis of Powder

Moisture content

Moisture content of the powder was measured by using infrared Metler moisture analyzer. Moisture content of the sample was measured also by using vacuum oven (at 70°C for 24 h) as per the method given by AOAC[15]. It was found that the variation between both the methods was negligible. In order to assist quick measurement infrared moisture analyzer was used for further experiments.

Mean diameter of the powder

Mean diameter of the powder was measured by following the procedure for milk powder particle size analysis[18,19]. Sieves of sizes 0.6, 0.5, 0.25, 0.21, 0.15 mm were used. The set of sieves were shaken for 5 minutes. The weight of sample retained on each of the sieves was measured, and calculations using differential sieve analysis were done to find out the mean diameter of the particle.

Bulk density of the powder

Bulk density of the powder was measured using the procedure given for milk powder bulk density analysis[18,19]. According to the procedure about 100 g of dry powder was taken in a measuring cylinder. And the cylinder was tamped gently about 100 times to remove the air gap between the powder particles theoretically. After that the height of the powder sample inside the cylinder was measured. Using the inner diameter of the cylinder and the powder height, volume of the powder was calculated. From the weight and the volume of the powder bulk density was calculated.

Hygroscopicity

Hygroscopicity is expressed as the final moisture content attained after exposing the powder in humid air having 79.5% relative humidity[19,20]. Air at this relative humidity was allowed to pass through the powder sample kept for hygroscopicity analysis in a Gooch filter until a constant weight of the powder sample was obtained. It was decided by measuring the weight of powder sample at a particular interval. The relative humidity of 79.5% was developed using saturated salt solution of potassium nitrate at 20±2°C, maintained using air conditioning machine in the measurement room. The hygroscopicity of the powder was calculated using the relation given in Eq. 7. Generally, a powder having the hygroscopicity value less than 10% is considered as good ‘non-hygroscopic’ powder[19,20]. A detailed procedure of hygroscopicity measurement was used as explained by Jaya and Das[14].

(7)

where b(g) is the increase in weight of powder, a_h is the amount of powder taken for the measurement and W_i (% wb) is the free water present in the powder before allowing in to the humid air environment.

Degree of caking

After the determination of hygroscopicity, the wet sample obtained at the end of hygroscopicity measurement was placed along with the Gooch filter in a drying oven set at 102°C±2°C for one hour to measure degree of caking. Normally, the degree of caking between 5 to 20% is called as ‘slightly caking’ powder[19,20]. After cooling the dried sample, it was weighed and transferred into a sieve of 500 μm size. The sieve was then shacked for 5 min in a shaking apparatus. The weight of the powder remaining in the sieve was measured. Degree of caking, DC (%) was calculated by using the Eq. (8)

(8)

where, d (g) is amount of the powder used sieving and c (g) is amount of the powder left on the sieve after sieving. For both hygroscopicity and degree of caking about 5 g of powder of having the mean diameter of 0.264 mm was used. It was spread uniformly in a Gooch filter of diameter 6.5 cm. The initial moisture content of the powder sample used was 3 – 4 % (wb).

Flowability

The schematic diagram of the apparatus used for the flowability measurement is shown in Figure 2. It consists of a stainless steel drum (1) of diameter 120 mm and length 90 mm. 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 70 mm length on the surface of the drum. Powder sample of known moisture content 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 rotary drum[19,20].

Figure 2. Schematic diagram of flowability measurement apparatus.

Sticky point temperature

A sticky-point apparatus has been developed and used since then by several other researchers[14,21–25]. A slightly modified form of the apparatus developed by the present authors was used for the measurement of sticky point temperature, T_s (°C). The measurement technique used in this study was based on the procedure given by Jaya and Das[14].

RESULTS AND DISCUSSION

Mean Diameter and Bulk Density of the Powder

The average mean diameter of the mango powder was calculated as 0.264 mm. The average bulk density of the powder obtained from the optimized ingredients was 537.2 kg.m⁻³.

Drying Rate of Mango Pulp Mix

Drying behaviour of mango pulp mix over time is shown in Figure 3 during vacuum drying at 70°C. During the initial stage of drying, the water evaporation rate was very high under vacuum. Within 900 s, more than 50% of the moisture was removed and the drying rate decreases continuously with improving drying time. The mango pulp thickness used in this study was about 3 mm[26].

Figure 3. Drying curve for a mango pulp mix under vacuum at 70°C.

Effect of MD and TCP on Mango Powder Properties

Table 2 shows the measured properties of mango powder at each combination of independent variables. Linear regression equations using the coded values (x₁ and x₂) of the independent variables (X₁ and X₂) have been developed to find out the effect of Maltodextrin (x₁), and Tricalcium phosphate (x₂) on the different mango powder properties (viz., hygroscopicity (HG), degree of caking (DC), flowability (FL) and sticky point temperature (T_s)) obtained. The equations developed are:

(9)

(10)

(11)

(12)

Table 2 Measured properties of mango powder at each combination of independent variables

∗X1	∗X2	∗HG (%)	∗DC (%)	∗FL (s)	∗Ts (°C)
0.55	0.0175	5.3	5.9	15	49.5
0.35	0.0125	10.3	17.1	35	38
0.45	0.0150	8.7	12.9	25.5	43.5
0.55	0.0125	6.7	13.02	28	45.5
0.35	0.0175	8.33	8.2	20	41.5
0.45	0.0175	7.35	7.6	18	44
0.45	0.0125	9.3	14.6	32	41
0.35	0.0150	6.1	9.2	23.5	47
0.55	0.0150	9.1	13.8	26	39
∗X1 – Maltodextrin (kg per kg dry mango solid).
∗X2 – Tricalcium phosphate (kg per kg dry mango solid).
∗HG – Hygroscopicity, %.
∗DC – Degree of caking, %.
∗FL – Flowability, s.
∗Ts – Sticky point temperature, °C.

The relationship between the real (X) and the coded (x) values are:

(13)

(14)

By observing the sign and magnitude of coefficients of x₁ and x₂ in Eqs. 9 to 12 we may arrive at the following conclusions.

Hygroscopicity: The hygroscopicity increases by decreasing the amount of maltodextrin and tricalcium phosphate (Eq. 9). The effect of maltodextrin (1.61) is the highest and is followed by tricalcium phosphate (0.89). We can find that the effect of MD is 1.8 (i.e., 1.61/0.89) times more than the effect of TCP.

Degree of caking: There is a decrease in degree of caking with increase in amount of both the ingredients (Eq. 10). The tricalcium phosphate has got more effect on percentage degree of caking with the value of 3.84 followed by maltodextrin with 1.83. It shows that TCP is having the effect 2.1 (i.e., 3.84/1.83) times more that the effect of MD added to the pulp.

Flowability: The flow time decreases by increasing the amount of maltodextrin, and tricalcium phosphate. The effect of tricalcium phosphate (7.0) is the highest, followed by maltodextrin (2.42). From Eq. (11), it can be found that TCP is having the effect 2.9 (i.e., 7/2.42) times more than the MD.

Sticky point temperature: Eq. (12) indicates that the sticky point temperature increases by increasing the amount of maltodextrin and tricalcium phosphate. The effect of maltodextrin (3.92) is the highest, followed by tricalcium phosphate (1.75). For the sticky point temperature of mango powder, MD is having the effect 2.24 times more than the effect of TCP.

Optimization of Independent Variables

Figure 4 shows the superimposed contour plots of four dependent variables viz., hygroscopicity, degree of caking, flowability and sticky point temperature with respect to coded values of maltodextrin and tricalcium phosphate. Contour lines of all the dependent variables are meeting at one point (marked by circle). The point at which all the four contour lines were joining together was considered as the optimum point to get the desired properties of mango powder, i.e., satisfactory levels of four dependent properties. The amount of maltodextrin and tricalcium phosphate at this point was considered as the optimum. Similar approach was also used by many researchers[27,28].

Figure 4. Superimposed contour plots of different dependent variable, showing optimum levels of independent variables.

The coded value of optimum point is 0.767 for maltodextrin and 0.672 for tricalcium phosphate. By converting this coded value in to real value using the Eqs. (13 & 14), the optimum amount of MD and TCP is: X₁ = 0.527 and X₂ = 0.0167, respectively. At the optimum amount of maltodextrin and tricalcium phosphate, the values of dependent variables were calculated using the 2^nd order regression Eqs. 15, 16, 17 and 18 developed for the coded values of independent variables and respective dependent variables. The calculated values are: Hygroscopicity (HG) = 6.4 %; Degree of Caking (DC) = 7.8%; Flowability (FL) = 18.6 s; Sticky point temperature (T_s) = 47.4°C. The mango powder properties obtained at the optimum amount of MD and TCP were close to the desired range of powder properties for tomato soup powder and instant coffee powder [14]. Second order regression equations developed between x and y are given as follows:

(15)

(16)

(17)

(18)

The significance of 2^nd order regression equation was found out by analysis of variance (ANOVA). Table 3 shows the result of analysis of variance for each properties of mango powder. The significance of each term in the 2^nd order regression equations was calculated and tabulated. It was also observed that the linear terms were significant for all the dependent variables at 10% significant level. If we consider the whole model (whole regression equation), except for degree of caking, all other three properties were significant at 10% level. Cross product terms were not significant for all dependent variables. Similar kind of result is observed for quadratic terms of all the properties except flowability. This analysis shows that the prediction model was only significant with the linear terms.

Table 3 Analysis of variance of independent variables as linear, quadratic and cross product on each of dependent variables

		Sum of squares
Source of variation	Degrees offreedom	Hygroscopicity	Degree of caking	Flowability	Sticky point temperature
Model	9	21.73∗	113.13 ^ns	335.06∗	112.56∗
Linear	2	20.14∗	108.49∗	329.09∗	110.03∗
Quadratic	2	1.50 ^ns	3.83 ^ns	4.97∗	2.47 ^ns
Cross product	1	0.08 ^ns	0.801 ^ns	1.0 ^ns	0.063 ^ns
Error	4	0.66	19.5	0.6	5.75
r^2		0.99	0.99	0.99	0.99
∗Significant at 90% level.
^nsNot significant.

CONCLUSIONS

Addition of maltodextrin and tricalcium phosphate significantly increased the sticky point temperature and decreased the hygroscopicity, flowability and degree of caking of the mango powder, which is desirable for efficient handling and storage of fruit powders. The optimal amount of MD and TCP were calculated as 0.527 and 0.0167 kg per kg dry mango solids respectively. Therefore, MD and TCP can successfully be used to improve the mango powder properties and drying conditions. However, the optimal amount of MD and TCP has to be further examined, if different drying methods or operating conditions are used.

Notes

5. Sudhagar, M. Spray drying of fruit juices. M. Tech Thesis. Department of Agriculture and Food Engineering, Indian Institute of Technology: Kharagpur, India, 2000