Drying 2004 Proceedings of the 14th International Drying Symposium (IDS 2004)

So Paulo, Brazil, 22-25 August 2004, vol. A, pp. 526-532

EXPERIMENTAL ANALYSIS OF FLUIDIZED BED FREEZE DRYING

P. Tomova, W. Behns, H. Haida, M. Ihlow and L. Mrl

Otto-von-Guericke-University of Magdeburg, Institute of Process Equipment and
Environmental Technology, Post office box 4120, D 39016 Magdeburg,
[email protected]
Keywords: freeze drying, fluidized bed, adsorption isotherm, modelling
ABSTRACT
The present work deals with freeze drying in a fluidized bed operating at atmospheric
pressure. This fluidized bed freeze drying process with integrated adsorption dryer
combines the advantages of the fluidized bed (high heat and mass transfer rates) with
the possibility of adsorption drying at low temperatures. The new process was experimentally studied. Experiments have been carried out with variation of the gas inlet temperature, the gas flow velocity, the bed mass and the particle diameter using a model
substance -Al2O3 and one biological product. The experimental investigations were
carried out in a fluidized bed plant with 100 mm nominal diameter which was built in
an industrial freezer container. The adsorption isotherms of the model material were
measured using the desiccator method at 1C and 5 C. The Brunauer-Emmett-Teller
(BET) and Guggenheim-Anderson-de Boer (GAB) models were applied to describe the
adsorption isotherm. This supplies information for a later mathematical modelling of the
fluidized bed freeze drying process.
INTRODUCTION
Freeze drying is the best method for drying temperature-sensitive products with regards to their composition. This process is widespread in food and pharmaceutical industry, when the natural properties of
the products do not allow other drying technologies. The most conspicuous properties of freeze dried
products are the voluminous state and high specific surface allowing a fast dissolution in solvents. Another important difference to conventional drying processes is the avoidance of denaturation which can
occur in case of conventional drying by thermal stress or by concentration increase in the liquid phase.
(Hayashi, 1989; Oetjen, 1999) The great benefits of the standard vacuum freeze drying process, like high
product quality, are faced by high investments in equipment and high operation costs, caused by long
drying times. To compare, spray drying is four times and vacuum drying is two times cheaper than the

526

standard vacuum freeze drying process (Wolff et al., 1990 Part 1). Therefore the new atmospheric freeze
drying alternative has been experimentally and theoretically investigated in order to determine energetic
and economic advantages.
MATERIALS AND METHODS
Two products have been used in the study of the new atmospheric fluidized bed freeze drying: the
spherical model substance Al2O3 with a mean diameter of
dP=1.93 mm and one biological product. The experiments
were prepared adjusting the
desired initial load of the
model substance and freezing
it at 18 C. The biological
product was peeled, reduced to
small pieces of cubic shape
and frozen also at 18 C. Figure 1 shows a flowsheet of the
experimental equipment used
in the study of this process.
The fluidized bed freeze dryer
is built in an industrial freezer
container facilitating drying
experiments at low temperaFigure 1 - Atmospheric freeze-drying experimental set up
tures. The fluidized bed chamber had a nominal diameter of 0.1 m. The drying medium i. e. the fluidizing medium was air. It is sucked
from the cold environment by means of a pressure fan and is introduced into the fluidizing chamber
through the flow distributor bottom. It leaves the apparatus at the top and flows through a pipework and a
dust filter into the adsorber. The dehumidification is performed there. Finally the dry fluidizing medium
with the desired temperature flows back into the fluidizing chamber. In case of a direct flow operation the
plant can be used for the adsorbent regeneration. The used adsorbent materials is molecular sieve with a
mean diameter of dP=2 mm. The adsorbent particles maintain the driving force for drying the water vapour pressure difference between the air and the product. The air velocity, air inlet and outlet temperature
and humidity were measured. The experiments were carried out as a batch process with periodic sampling. Therefore the process had to be interrupted regularly in order to determine the load values by
means of drying in an oven and weighing. In this manner, the particle load can be recorded in the course
of time.
RESULTS
The feasibility of atmospheric freeze drying was demonstrated by drying 0.450 kg model material Al2O3 with an initial moisture content of 0.56 g/g. The bed mass in this experiment was fluidized at fluid
flow velocity of 5.3 m/s. The air temperature below the flow distributor was controlled to be at 6 C.
The experimentally determined drying curve is shown in Figure 2. The course of the air humidity with
time is shown in Figure 3. The drying period in this case was approximately 4 hours. During the first 1
hour, a distinct decrease in the product moisture content and an increase in the air relative humidity at the
outlet have been observed. As in normal drying processes, the following 3 hours are characterized by decreasing the drying velocity which is an attribute of the second drying period. This second drying period
ends when the hygroscopic load is reached. The inlet humidity of fluidizing air was almost 0 % during the

527

whole experiment which confirms the adsorber efficiency. The air temperature at the inlet and outlet of
the fluidized bed apparatus is shown in Figure 4. The air inlet temperature has been adjusted at 1 C by
the controller and maintained constant during the course of the process. The air outlet temperature drops
during the first drying period because of the sublimation heat consumption and, then, this temperature
increases to a steady state value because of the reduction in the drying velocity. The obtained temperature
curves show that the period of time necessary for reaching the steady outlet temperature at the apparatus
exit is equal to the period of time necessary for reaching the stationary relative humidity and the stationary product moisture content. The temperature and humidity curves correspond to those under normal
drying conditions. The difference between the inlet and the outlet temperature is ascribed to the heat
transfer through the apparatus wall.

Figure 2 Experimentally determined

drying curve

Figure 3 Experimentally determined gas

inlet and outlet humidity

Figure 4 Experimentally determined

gas inlet and outlet temperature

In order to better understand the mass and heat transfer processes in atmospheric fluidized bed freeze
drying, experiments have been carried out with variation of the gas inlet temperature, the gas flow velocity, the bed mass and the particle diameter using a model substance -Al2O3. Table 1 shows the main parameters of all experiments carried out.
Table 1: Main parameters of atmospheric fluidized bed freeze drying experiments
Experiment code
air inlet temperature variation
V1
V2
V3
fluid flow velocity variation
V4
V5
V6
product mass variation
V7
V8
V9
product diameter variation
V11
V12

Tair [C]

wair [m/s] mP [kg]

1
4
6

3
3
3

0.450
0.450
0.450

6
6
6

3
4
5

0.450
0.450
0.450

6
6
6

3
3
3

0.225
0.450
0.600

1
1

4
4

0.500
0.500

The influence of the gas inlet temperature on the drying kinetics is shown in Figures 5 and 6. The
curves represent freeze drying at temperatures of 1 C, 4 C and 6 C in the case of a constant gas

528

flow velocity of 3 m/s for the three experiments, corresponding to a gas mass flow rate of 108 kg/h. The
initial bed mass was 0.450 kg with a load of 0.6 kg/kg for all experiments. As shown in Figure 5, the increase in the gas inlet temperature causes a decrease in drying time, which is expressed by a faster decrease in the particle moisture content values and a faster reaching of a stationary gas outlet temperature
(Figure 6). The reason is that at equal gas mass flow rates the higher gas inlet temperature makes that the
higher average gas temperature is achieved causing a greater driving force for the drying process.

Figure 5 Comparison of the measured drying

curves for the experiments V1, V2 and V3

Figure 6 Comparison of the measured gas outlet

temperature for the experiments V1, V2, V3

The influence of the gas flow velocity on the drying period, the temperature and the humidity curves is
displayed in Figures 7 and 8. The experiments were performed with a gas inlet temperature of 6 C. The
bed mass was 0.450 kg with an initial moisture content of 0.6 kg/kg. The increase in the gas flow velocity, while keeping other process parameters constant, leads to a decrease of the drying period and a faster
reaching of stationary gas outlet humidity and gas outlet temperature values. This can be explained by the
higher enthalpy flow rate introduced by the higher gas mass flow rate.

Figure 7 Comparison of the measured drying

curves for the experiments V4, V5 and V6

Figure 8 Comparison of the measured gas outlet

temperatures for the experiments V4, V5, V6

529

The influence of the bed mass on the drying process is illustrated in Figures 9 and 10. The experimentally obtained drying curves of three experiments with a gas flow velocity of 3 m/s and a gas temperature
of 1 C are compared in Figure 9. The bed mass was varied from 0.225 kg to 0.4 kg and 0.6 kg. The initial product moisture content was 0.6 kg/kg. In all cases the same enthalpy flow rate is introduced into the
system by the fluidizing gas. Because of the same initial product moisture content, the fluidized bed contains different frozen water masses. Therefore a longer drying period is needed in the case of higher bed
masses because a higher ice mass has to be sublimated. The load is related to the dry mass so that the particle load decreases faster in case of smaller bed masses.

Figure 9 Comparison of the measured drying

curves for the experiments V4, V5 and V6

Figure 10 Comparison of the measured gas outlet

temperatures for the experiments V4, V5 and V6

Experiments with a diced biological product

have been performed in order to verify the applicability of the new process for temperature
sensitive products. The particle diameter was
varied from 5 to 8 mm for the experiment V 11
and from 8 to 12 mm for the experiment V 12.
The air inlet temperature was 1 C and the gas
flow velocity was 4 m/s for both experiments.
As shown in Figure 11, the drying period is
shorter in the case of the smaller fraction (V11),
which can be explained by the greater transfer
area and the higher drying rate.
Previous knowledge and the analogy between conventional drying processes and atmospheric fluidized bed freeze drying serves as
Figure 11 Comparison of the measured drying curves for the basis for the mathematical description of
the experiments V11 and V12
this new process. The water vapor adsorption
isotherms of the product serve as the basis for theoretical considerations (Wolff et al., 1990 Part 2). The
needed adsorption isotherm of the model material was measured using the desiccator method at 1 C and
5C. The adsorption isotherm was described by the Brunauer-Emmett-Teller (BET) model:
X eq = X m

530

(1 )

C BET
(1 + CBET )

(1)

and the Guggenheim-Anderson-de Boer (GAB) model. This is a modified BET-equation with an additional factor k and is valid in the relative humidity range from 0 % up to 90 % (Sablani et al., 2001; San
Martin et al., 2001):
(2)

CGAB
.
X eq = X m
(1 k) (1 k + CCAB k )
The GAB-model is often used in the literature in order to describe adsorption equilibria of biological
matter and showed better conformity with the experimental results than the BET-model Figure 13.

Figure 12 Adsorption isotherm of the model

material -Al2O3 at 1 C and 5 C

Figure 13 Experimental and predicted by BET

and GAB models isotherms of -Al2O3 at 1 C

CONCLUSIONS
Based on experimental investigations, the technical feasibility of the new atmospheric fluidized bed
freeze drying has been confirmed in this paper. The experimental results with the model substance Al2O3 have shown that the drying rate can be increased by increasing the gas flow velocity and gas temperature and by decreasing the bed mass as in conventional drying processes using heated air. The experiments with a diced biological material have demonstrated the applicability of this process for the drying of temperature sensitive products. The adsorption isotherms of the -Al2O3 material have been experimentally measured and mathematically described. These will be used to develop a mathematical
model for describing the freeze drying process.
ACKNOWLEDGMENT
The research has been supported by the InnoRegio-Netzwerk REPHYNA and by the scholarship program of the Deutsche Bundesstiftung Umwelt (German Federal Environmental Foundation).
NOTATION
C
k
m

sorption isotherm parameter

sorption isotherm parameter
mass

kg

531

T
t
w
X
Greek Symbols

temperature
time
velocity
particle moisture content, dry basis

C
s
m/s
g/g

relative air humidity,

Subscripts
BET

bed
eq

GAB

gas
m
P

BET- equation
fluidized bed
equilibrium
GAB- equation
gas
monolayer
product
LITERATURE

Hayashi H. (1989), Drying technology of toods their history and future, Drying Technology 7, 2 pp.
315-369
Oetjen G. W. (1999), Freeze drying, Weinheim, Wiley-VCH, pp. 200-230
Sablani, S.S., Myhara R. M., Mahgoub, Al-Attabi, Z. H. and Al-Mugheiry, M. M. (2001), Water sorption
isotherms of freeze dried fish sardines, Drying Technology, 19 (3&4), pp. 673-680
San Martin, M. B., Mate, J. I., Fernandez, T. and Virseda, P. (2001), Modelling adsorption equilibrium
moisture characteristics of rough rice, Drying Technology, 19 (3&4), pp. 681-690
Wolff, E. and Gibert, H. (1990), Atmospheric Freeze Drying Part 1: Design, experimental investigation
and energy saving advantages, Drying Technology, 8 (2), pp. 385-404
Wolff, E. and Gibert, H. (1990), Atmospheric Freeze Drying Part: 2 Modelling drying kinetics using
adsorption isotherms, Drying Technology, 8 (2), pp. 405-428

532