Microwave/vacuum drying of model fruit gels

A.E. Drouzas, E. Tsami

*

, G.D. Saravacos

Department of Chemical Engineering, Zografou Campus, National Technical University, 157 80 Athens, Greece

Received 5 May 1998; accepted 10 October 1998

Abstract

Combined microwave (MW)/vacuum drying of fruit materials has a promising potential for high-quality dehydrated products. A

better knowledge of the drying kinetics of fruit products could improve the design and operation of ecient dehydration systems.

A laboratory MW/vacuum drier was used for drying kinetics experiments with model fruit gels, simulating orange juice con-

centrate. The system was operated in the vacuum range of 30±50 mbar and MW power of 640±710 W. The distribution of the

electromagnetic ®eld in the cavity of the oven was determined from the drying rate of samples, placed at 5 di€erent locations.

The drying rate was determined by periodic weighing of the sample. The rate constant (K) of the single-layer model of drying was

estimated by regression analysis of the experimental data. An empirical model is proposed for estimating the drying constant (K) as

a function of the absolute pressure and the MW power of the system. Ó 1999 Elsevier Science Ltd. All rights reserved.

1. Introduction

Dehydration of fruit materials, especially fruit juices,

is a dicult food processing operation, mainly because

of undesirable changes in quality of the dehydrated

product. High temperatures and long drying times, re-

quired to remove the water from the sugar containing

fruit material in conventional air-drying, may cause se-

rious damage to the ¯avor, color and nutrients of the

dehydrated product. Vacuum drying has been proposed

to overcome these problems, especially with orange and

other fruit juices. However, the vacuum-drying process

is expensive, due to high capital and operating costs.

Expensive vacuum equipment is required, and heat

transfer in vacuum may limit seriously the drying rate.

Freeze drying, which yields products of higher quality is

more expensive than vacuum drying and it is not eco-

nomical for fruit products. Heat transfer in vacuum and

freeze drying may be enhanced by infrared or micro-

wave radiation. Microwave (MW) or dielectric energy

has the advantage of higher penetration in the material

and its preferential absorption by the water molecules

(Drouzas & Schubert, 1996). The potential of MW en-

ergy to drying has found only limited applications

(Schi€mann, 1995). Improved dehydrated potato and

apples were obtained by MW drying (Huxsoll & Mor-

gan, 1968). Recent research has shown that pre-treat-

ment of food materials with MW energy increases

substantially the air-drying rate (Kostaropoulos &

Saravacos, 1995; Drouzas, Tsami & Saravacos, 1997).

The improved drying rate is ascribed to the development

of a porous structure of the food material, which facil-

itates the transport of moisture (Marousis, Karathanos

& Saravacos, 1991).

The MW and high frequency (HF) energy have two

advantages of high penetration in the solid material and

preferential absorption by the water molecules. How-

ever, they have the disadvantage of non-homogeneous

distribution in the processing cavity, creating problems

of non-uniform heating (Risman, Ohlsson & Wass,

1987; Ohlsson, 1990; Schubert, Gruneberg & Walz,

1991). The temperature distribution in the cavity is in-

¯uenced by the composition and the dielectric properties

of the food material, and its location in the oven. The

energy absorption by a sample is also a€ected by the

presence of other materials in the cavity (Kraisheh,

Nomenclature

K

drying constant (1/min)

K

0

drying constant corresponding to pressure

P

0

and MW power Q

0

(1/min)

m, n

empirical constants (±)

P, P

0

absolute pressure (mbar)

Q, Q

0

MW power output (W)

t

drying time (min)

X, X

0

moisture content (kg/kg dry matter)

Journal of Food Engineering 39 (1999) 117±122

*

Corresponding author. E-mail: [e-mail]

0260-8774/99/$ ± see front matter Ó 1999 Elsevier Science Ltd. All rights reserved.

PII: S 0 2 6 0 - 8 7 7 4 ( 9 8 ) 0 0 1 3 3 - 2

Cooper & Magee, 1997). The sugar content of agar gels

may have a signi®cant e€ect on the absorption of MW

energy (Padua, 1993).

Although combined MW±vacuum drying has found

some application in the dehydration of fruit juices, more

research and development is needed before the process is

used in large commercial scale. In particular, the e€ect

of vacuum and MW power on the drying kinetics should

be known quantitatively, so that the drying system can

be optimized from the cost and quality standpoints.

2. Experimental procedure

2.1. Apparatus

In the present study two MW ovens SHARP IEC 705

with di€erent power outputs, 800 and 700 W, both at

2450 MHz, were used. For each of the ovens there was a

glass vacuum desiccator, in which the samples being

dried were placed, as well as a vacuum pump for the

application of the vacuum (Drouzas & Schubert, 1996;

Kiranoudis, Tsami & Maroulis, 1997). The vacuum

system included a pressure regulator and cooling unit

for condensing and cooling the water vapor at low

temperature (ÿ25°C). A conventional air drier was also

used.

High-moisture samples of pectin gel were used

(X

0

ˆ 4±7 kg water/kg dry basis), in order to detect any

di€erences in the drying rate at the initial stage of drying.

2.2. Materials and methods

The material used was a pectin gel (model concen-

trated orange juice) with the following composition in %

(w/w):

2.3. Preparation of the pectin gel

(a) About 1/3 of the sugars was mixed with the pectin

and the mixture was added to water at 40±50°C with

constant stirring to ensure complete homogeneity of

the solution.

(b) The remaining 2/3 of the sugars was added to the

pH bu€er solution (pH ˆ 3) and boiled until it be-

came very viscous.

(c) The sugar solution (b) at 70°C was added with

constant stirring to the gelatinous sugar pectin solu-

tion (a).

The hot gel was poured quickly into petri dishes with

a diameter of 5 cm, where gelling occurred instanta-

neously. The mean thickness of the gel in the petri dish

was 5 mm.

The

0

Brix of the samples of the gel was measured

with an Abbe refractometer, and the initial moisture

content was determined, using the vacuum oven

method (AOAC, 1984) at 70°C, 25 mbar for 24 h. The

rest of the samples were preserved in a refrigerator

(ÿ4°C).

2.4. Drying experiments

One petri dish, containing the gel, was weighed using

an electronic balance and it was placed in a ®xed posi-

tion inside the vacuum desiccator, which was sealed air-

tight, and the desired pressure and radiation was ap-

plied.

The sample was radiated for 10 s, using the full power

of the oven, which corresponds to constant radiation,

then the power was paused for 30 s, resumed again for

10 s, paused again for 30 s, and ®nally resumed again for

10 s. In this way boiling and bubbling of the gel was

avoided.

The sample was weighed using a Mettler AE 160

electronic balance, it was placed once again into the

desiccator and the same procedure was repeated (eight

times) until the drying was completed. The total dura-

tion of the radiation was 4 min. When samples were

weighed, the radiation was paused and the vacuum was

released inside the desiccator. The amount of the dry

solids of the dried sample was determined using the

vacuum oven method (AOAC, 1984). The whole pro-

cedure was repeated three times.

For comparison of the color changes during drying,

the MW-drying apparatus was operated at atmospheric

pressure, using the same sample material and the same

MW power.

Samples of the gel were also dried in a laboratory

tunnel air drier at a temperature of 60°C, relative hu-

midity 15% and air velocity 4.5 m/s. At frequent inter-

vals (10 min) the sample was weighed in an electronic

Mettler scale for the determination of the drying curve

and comparison of the air drying with the vacuum±MW

drying.

The procedure was repeated three times.

The MW energy distribution in the oven was de-

termined indirectly at atmospheric pressure, by placing

samples of the model gel in 5 ®xed locations in the

cavity and estimating the drying rate (Fig. 1). The

same ®xed positions were used in all experimental

work.

Glucose

14.2

Fructose

15.8

Sucrose

27.6

Citric acid

1.2

Pectin

2.8

Water

38.4

100 g

118

A.E. Drouzas et al. / Journal of Food Engineering 39 (1999) 117±122

2.5. Isotherms and color of product

Samples of dried gel were also used for the determi-

nation of the moisture sorption isotherm at 25°C, using

a Rotronic-hygroscop BT apparatus attached to a water

circulator. The dried in MW-oven samples were placed

above water in desiccators and they were humidi®ed to

di€erent levels of relative humidity which were then

determined with the use of the apparatus mentioned

above. Equilibration of the samples was ensured by

leaving them in the apparatus for long enough time to

reach constant weight. Each sample was weighed after

each measurement and, using the vacuum oven method,

the dry mass of each sample was recorded and the

equilibrium moisture content was calculated.

The color of pectin gels dried under vacuum and

without the use of a vacuum was determined with the

use of the Hunter Lab program and apparatus. Three

replications were made, and the mean of three mea-

surements of each replication is reported.

2.6. Modeling of MW vacuum drying

The drying curves were prepared by plotting the

moisture content X (kg moisture/kg dry matter) vs. time

t (min). Assuming the thin-layer theory of drying, the

drying rate can be expressed by the equation
…X ÿ X

e

†=…X

0

ÿ X

e

† ˆ exp…ÿKt†;

…1†

where K is the drying constant and X

0

, X, X

e

are the

moisture contents at the beginning, after time (t) and at

equilibrium. For vacuum drying it can be assumed that

X

e

ˆ 0.

The drying constant (K) can be expressed by the

following empirical model

K ˆ K

0

…P=P

0

†

n

…Q=Q

0

†

m

;

…2†

where (P,P

0

) and (Q,Q

0

) are, respectively, the operating

and reference pressure and MW power output. The

empirical constants K

0

, m and n can be estimated by

nonlinear regression of the experimental drying data

(Kiranoudis et al., 1997).

3. Results and discussion

3.1. MW-energy distribution

A signi®cant variation of the drying rate was ob-

served for samples placed in di€erent locations in the

MW-oven cavity (drying curves of Fig. 2). The di€er-

ences in drying rate were caused by the uneven (multi-

mode) distribution of the electromagnetic energy in the

cavity (Risman et al., 1987; Ohlsson, 1990).

Locations (1) and (4) showed the highest and lowest

drying rates, corresponding to ``hot'' and ``cold'' spots

in the oven. At three locations (2,3,5) the absorbed MW-

energy seemed to converge as the drying of the samples

progressed (after about 6 min). These locations were

used for placement of the samples in the subsequent

MW-drying experiments.

In drying applications, the complex modeling of MW

distribution in the oven can be simpli®ed (Ohlsson,

1990; Kraisheh et al., 1997). The wavelength of MW at

2450 MHz is 12 cm in the air, which is close to the di-

mensions of the laboratory MW ovens. The penetration

length of MW power at low moisture contents estimated

from Lambert's absorption law, is about 30 cm, which is

higher than the normal thickness of the food pieces

being dehydrated. Thus, attenuation e€ects of foods at

low moisture content can be neglected.

The absorbed MW energy has been found to increase

linearly with the diameter of the food material (Kraisheh

Fig. 2. Drying rates of samples of pectin gel at di€erent locations (1, 2,

3, 4, 5) in the MW-oven cavity (atmospheric pressure).

Fig. 1. Location of samples in the MW cavity for estimation of the

MW-energy distribution.

A.E. Drouzas et al. / Journal of Food Engineering 39 (1999) 117±122

119

et al., 1997). However, the absorbed energy decreased

nonlinearly with the loading of the oven. Thus, the en-

ergy distribution in an MW oven can be made more

uniform by changing the load pattern in the cavity.

3.2. MW±vacuum drying

Combination of MW heating and vacuum drying

resulted in acceleration of the drying rate of model fruit

gels. The experimental pectin gel of 38.4 moisture con-

tent dried to less than 3% moisture within 4 min (Fig. 3).

By comparison, similar samples of pectin gel required

more than 8 h to reach a moisture of about 10% in an air

drier at atmospheric pressure and 60°C (Fig. 4). The

high sugar content of the gel caused shrinkage and

collapse of the gel structure during air drying, resulting

in low transport rate (di€usion) of water and prolonged

drying time.

The MW energy and vacuum drying created a very

porous structure (pung) of the gel samples, facilitating

the transport of the water vapor. Evaporation of water

within the sample is accelerated by the preferential ab-

sorption of microwave energy by the water molecules.

Pretreatment of fruit and vegetable materials with

MW energy has been found to increase the drying rate

during the early stages of air drying (Drouzas et al.,

1997). However, the drying rate of sugar-containing

food materials was reduced in the last stages of drying,

evidently due to the collapse of the porous structure,

created earlier in the drying process. It is evident that

vacuum drying maintains the porous structure through-

out the drying process, reducing sharply the required

drying time.

3.3. E€ect of pressure and MW power

The optimum operating pressure in vacuum-drying

processes depends on the process economics and the

quality of the dried products. High vacuum yields nor-

mally better quality but the equipment and operating

costs may be too high for most food products. Most

vacuum-drying operations use the pressure range of 30±

50 mbar, in which water evaporates from the liquid

phase and product pung takes place. Freeze drying

requires pressures lower than 5.33 mbar (evaporation

Fig. 3. MW±vacuum drying curve of model pectin gel (P ˆ 40 mbar,

Q ˆ 710 W).

Fig. 4. Atmosphere air-drying curve of model pectin gel at 60°C.

Fig. 5. E€ect of pressure (P) and MW power (Q) on the drying rate

constant (K).

Fig. 6. Moisture adsorption isotherm of MW±vacuum dried pectin gel

(25°C).

120

A.E. Drouzas et al. / Journal of Food Engineering 39 (1999) 117±122

from the frozen state) and long drying times, increasing

considerably the cost of the process.

The drying rate constant (K), estimated from Eq. (1),

was found to increase signi®cantly as the pressure (P)

was reduced from 50 to 30 mbar (Fig. 5). A signi®cant

increase of the drying rate constant was observed when

the MW power output (Q) was increased from 640 to

710 W.

Regression analysis of the experimental data of dry-

ing rate (K) vs. pressure (P) and MW power output (Q),

using Eq. (2), yielded the following empirical constants:

m ˆ 0.698, n ˆ ÿ0.318 and K

0

ˆ 0.857 1/min, for ref-

erence P

0

ˆ 40 mbar and Q

0

ˆ 710 W.

3.4. Moisture sorption isotherms

Fig. 6 shows a typical moisture adsorption isotherm

at 25°C of a sample of vacuum-dried pectin gel. The

experimental adsorption data ®tted well the empirical

GAB model (Tsami, Marinos-Kouris & Maroulis,

1990). The estimated empirical constants of the model

were X

m

ˆ 7.89, C

0

ˆ 1.25, DH

c

ˆ 1.02, k

0

ˆ 1.10,

DH

k

ˆ ÿ0.48. The shape of the sorption isotherm is

characteristic of the high-sugar fruit materials, which

show a sharp increase of sorption capacity at higher

(above 0.65) water activities (Tsami et al., 1990). The

sorption isotherm indicates that dehydrated high-sugar

pectins or fruit juice powders should be handled and

stored as hygroscopic materials in order to preserve their

quality and functionality.

3.5. Color of dried pectin gel

The color of MW±vacuum dried model pectin gel was

better (lighter) than the color of the MW±air dried

product at atmospheric pressure. Table 1 shows the re-

sults of color measurements of the dried products, using

the Hunter lab system. The three color parameters (L, a,

b) are quantitative indicators of the lightness (L), the

redness (a), and the yellowness (b) of the product.

The color of the MW±vacuum dried material was

much lighter (higher L value) than the MW±air dried

product. At the same time MW±air drying increased the

redness (a value) and decreased the yellowness (b value).

The undesirable browning of the MW±air dried product

is a characteristic color change of high-sugar food ma-

terials, heated at high temperatures.

4. Conclusions

The drying rate constant of the thin-layer model of

drying of a model fruit gel was found to increase with

increasing MW-power output and decreasing absolute

pressure in vacuum drying. Due to the uneven distri-

bution of the MW energy in the MW oven the location

of the material in the cavity should be speci®ed. The

color of the MW±vacuum dried fruit gel was signi®-

cantly lighter than the color of the MW±air dried

product at atmospheric pressure.

References

AOAC (1984). Ocial methods of analysis of the association of ocial

analytical chemists (14th ed.) Association of Ocial Analytical

Chemists, Washington, DC.

Drouzas, A. H., & Schubert, H. (1996). Microwave application in

vacuum drying of fruits. Journal of Food Engineering, 28, 203±209.

Drouzas, A. E., Tsami, E., & Saravacos, G. D. (1997). E€ect of

microwave pretreatment on the moisture di€usivity of fruits and

vegetables. In R. Jowitt, Engineering and Food at ICEF 7, Part 1.

Academic Press, Sheeld (pp. A57±A60).

Huxsoll, C. C., & Morgan, A. J. (1968). Microwave dehydration of

potatoes and apples. Food Technology, 22, 47±51.

Kostaropoulos, A. E., & Saravacos, G. D. (1995). Microwave pre-

treatment for sun-dried raisins. Journal of Food Science, 60, 344±

347.

Kiranoudis, C. T., Tsami, E., & Maroulis, Z. B. (1997). Microwave

vacuum drying kinetics of some fruits. Drying Technology, 15(10),

2421±2440.

Kraisheh, M. A. M., Cooper, T. J. R., & Magee, T. R. A. (1997).

Microwave and air drying. I. Fundamental considerations and

assumptions for the simpli®ed thermal calculations of volumetric

power absorption. Journal of Food Engineering, 33, 207±219.

Marinos-Kouris, D., & Maroulis, Z. B. (1995). Transport properties in

the drying of solids (pp. 113±140). In A.S. Mujumdar, Handbook of

industrial drying (2nd ed.). Marcel Dekker, New York.

Marousis, S. N., Karathanos, V. T., & Saravacos, G. D. (1991). E€ect

of physical structure of starch materials on water di€usivity. J.

Food Proc. and Pres., 15, 183±195.

Ohlsson, T. (1990). Temperature distribution in microwave heating ±

In¯uence of oven and food related factors (pp. 231±241). In W.E.

L. Spiess, H. Schubert, Engineering and food, vol. 2. Elsevier,

London.

Table 1

Color parameters of dried pectin gel

L

a

b

MW±Vac

MW±Air

MW±Vac

MW±Air

MW±Vac

MW±Air

51.56

23.97

3.51

7.08

14.94

10.60

49.50

21.60

3.84

6.08

15.17

8.40

50.60

21.40

3.20

5.92

14.43

7.20

A.E. Drouzas et al. / Journal of Food Engineering 39 (1999) 117±122

121

Padua, G. W. (1993). Microwave heating of agar gels containing

sucrose. Journal of Food Science, 58, 1427±1428.

Risman, P. O., Ohlsson, T., & Wass, B. (1987). Principles and models

of power density distribution in microwave oven loads. Journal of

Microwave Power, E

2

, 193±198.

Schi€mann, R. F. (1995). Microwave and dielectric drying (pp. 345±

371). In Handbook of industrial drying (2nd ed.). Marcel Dekker,

New York.

Schubert, H., Gruneberg, M., & Walz, E. (1991). Erwarmung von

Lebensmittel durch Mikrovellen: Grundlagen, Messtechnik, Be-

sondorheiten. ZFL, 42(4), 14±21.

Tsami, E., Marinos-Kouris, D., & Maroulis Z. B. (1990). Water

sorption isotherms of raisins, currants, ®gs, prunes and apricots. J.

Food Science, 55(6), 1594±1597, 1625.

122

A.E. Drouzas et al. / Journal of Food Engineering 39 (1999) 117±122