ABSTRACT
Convective drying of foods presents some limitations which make it difficult
to apply on specific fields. Among other things, the low drying rate and product
quality loss must be considered. Some of these limitations may be overcome
combining other technologies, which may be used as additional energy sources
during drying. Power ultrasounds are considered as an appropriate technology
since they may affect drying without significantly heating the material. This fact
may contribute to its application in the drying of heat sensitive materials or in drying
processes carried out at low temperatures, such as atmospheric freeze drying.
Power ultrasounds have been applied to affect mass transfer processes in
solid-liquid treatments, like meat and cheese brining, the osmotic dehydration of
fruits and several extraction processes. Nevertheless, applications on solid-gas
systems are much less frequent due to some technical difficulties, which prevent
this technology from being fully developed. Among other things, the high
impedance mismatch between the application systems and air, which makes the
acoustic wave transmission difficult, and the high acoustic energy absorption of the
air must be considered. These limitations may be overcome by adequately
designing the ultrasonic application system.
The mass transfer process that takes place during convective drying may be
influenced by a series of effects associated to power ultrasound application. On
one hand, the external resistance to mass transfer may be affected by pressure
variations, oscillating velocities and microstreaming at the solid-gas interfaces thus
reducing the boundary layer thickness and therefore improving water transfer from
solid surface to air medium. On the other hand, internal resistance may be reduced
by alternating expansion and compression cycles produced by ultrasound in the
material (a phenomenon known as the “sponge effect”) and also through some
effects on the interfaces of intercellular spaces, or even by cavitation which may
contribute to removing the strongest attached moisture to solid matrix.
The main aim of this work was to determine the effect of applying power
ultrasound on convective drying processes, establishing the influence of the main
process variables.
In order to reach this objective, a new ultrasonic system was designed to
attain a good transmission of acoustic energy to the gas medium. The design was
carried out by considering the walls of the drying chamber as the device for
radiating the acoustic energy, and as a consequence the drying chamber would be
the vibrating element transmitting the acoustic energy to the particles. The design
was made through modelling, considering a finite element model in 2 dimensions
using the ANSYS code. As a result, an ultrasonic system consisting of an
aluminium cylinder (external diameter 120 mm, thickness 10 mm and height 310
mm) driven by a piezoelectric composite transducer was developed. The driving
transducer consists of an extensional piezoelectric sandwich element together with
a mechanical amplifier. The ultrasonic system was characterized before being
installed on the convective drier, and for that purpose not only the electric
properties were measured using an impedance analyzer but also the response of
the system when applying an electric power of 90 W (frequency 21.8 kHz, voltage
60 V, intensity 1.55 A, phase 4ș and impedance 329 O). The acoustic field
generated inside the cylinder produced an average sound pressure level of about
154.3 dB when the electrical power applied to the transducer was 75 W. This figure
was very close to that estimated from modelling using the finite element method
(156.3 dB).
A laboratory scale convective drier was modified to install the different
elements of the ultrasonic system. The drier modifications were carried out in such
a way as to provide good working conditions for the vibrating cylinder and also to
maintain the automatic weighing system on the drier.
Drying kinetics of several foodstuffs were carried out: carrot, apricot,
persimmon and lemon peel. Sorption isotherms were obtained from literature
except for lemon peel, since no references were found for this product. Thus, the
sorption isotherms of lemon peel were determined at different temperatures (20,
30, 40 and 50 șC) using electric hygrometers.
The GAB model was rated as the best for describing the experimental
sorption data when considering the influence of temperature. The isosteric heats of
sorption were determined from the Clausius-Clapeyron equation using differential
and integral identification methods and also from Riedel equation. The isosteric
heats of sorption provided by both methods were very similar.
No significant effect of power ultrasound application (75W, 21.7 kHz) was
observed on fluidized bed drying kinetics of carrot cubes (side 8.5 mm) and eights
parts of apricots. Experiments were carried out at different temperatures (30, 35,
40, 45, 50, 55 and 60 șC) and air velocities between 10 and 14 m/s. Activation
energy figures were similar to others found in the literature for these products.
The influence of air velocity on the acoustic field generated by the ultrasonic
system was addressed due to the previous results from fluidized bed drying
kinetics. As the air velocity increased, there was a reduction in the average sound
pressure level, although, it seems to remain constant from an air velocity figure of
about 8 m/s onwards. Therefore, the air velocity increase provided lower acoustic
energy levels available for the particles in the drying chamber.
Drying kinetics of different products (carrot, persimmon and lemon peel) and
geometries (cubes, cylinders and disks) were carried out at air velocities of
between 0.5 and 14 m/s. Air velocity affected the non ultrasonic (SUS) drying
kinetics of the different products up to a value of about 5 m/s. This threshold was
established from the effective moisture diffusivities (De) identified with a diffusion
model considering external resistance to mass transfer as negligible (SRE model).
Power ultrasound application (75 W, 21.7 kHz) only increased the effective
diffusivity values identified with the SRE model on the drying kinetics of carrot and
persimmon for experiments carried out at air velocities lower than 6 m/s. As a
consequence, the influence of power ultrasound was negligible at high air flow
rates (low acoustic energy levels). However, power ultrasound application on
lemon peel drying significantly increased (p<0.05) effective diffusivities at all the air
velocities considered. A possible explanation of the fact that lemon peel behaves
differently compared to carrot and persimmon is its structure. Acoustic effects on
lemon peel were stronger, as it is considered to be a more porous product than
carrot or persimmon.
Porosity may be considered as one of the most important structural variables
for determining the acoustic effectiveness in foodstuffs. High porosity products may
be considered more prone to alternating compression and expansion cycles
produced by ultrasonic waves, improving water movements in its large intercellular
spaces. Small intercellular spaces are also found in low porosity products, that
means a high internal resistance to mass transfer. Thus, high acoustic energy
levels are required to affect mass transfer in low porosity products.
The influence of porosity may also be explained considering a greater
acoustic energy absorption in high porosity products. As a consequence, the
internal energy available in the particles would increase, leading to more intense
compressions and expansions (sponge effect), which could improve water removal
and therefore, reduce internal resistance. Furthermore, the acoustic effects on the
solid-gas interfaces of intercellular spaces could increase in high porosity products
due to a larger porous net. Indeed, this phenomenon also contributes to reduce
internal resistance to mass transfer.
Diffusion models that do not consider the external resistance to mass
transfer (SRE) presented a poor fit to the experimental data in experiments carried
out at low air velocities but a good fit at high air velocities. Therefore, diffusion
models considering external resistance (RE), which differ according to geometry,
were applied on experiments carried out at low air velocities. RE models were
solved using an implicit finite difference method using the programming language
available on Matlab. RE models provided percentages of explained variance higher
than 99 % and mean relative errors lower than 10 % in all the cases.
In persimmon drying experiments carried out at air velocities of lower than 6
m/s, power ultrasound application (75 W, 21.7 kHz) significantly increased (p<0.05)
the effective moisture diffusivity (De) and the mass transfer coefficient (k) values
identified with the RE model. Therefore, both external and internal resistance to
mass transfer were significantly affected by power ultrasound application when low
air velocities were used.
The influence of air temperature on power ultrasound assisted convective
drying was addressed from SUS (without ultrasound) and US (ultrasound, 75 W,
21.7 kHz) drying experiments of carrot cubes (side 8.5 mm) carried out at 1 m/s
and at several air temperatures: 30, 40, 50, 60 and 70 șC. De and k figures
identified from US experiments were only significantly (p<0.05) higher than those
identified on SUS experiments when air temperature was lower than 60 șC. The
influence of power ultrasound application decreased as the air temperature got
higher and it was almost negligible at 70 șC. Effective moisture diffusivities
identified from US experiments at the different temperatures showed a poor fit
when considering an Arrhenius equation, due to the cross effect of temperature
and ultrasound application. Figures identified at high temperatures (60 and 70 șC)
departed from the tendency shown by De figures identified at low temperatures (30,
40 and 50 șC).
The effect of mass load used in the experiments was also considered. In
order to clarify its influence, SUS and US (75 W, 21.7 kHz) drying experiments of
carrot cubes (side 8.5 mm) were carried out at several mass load densities: 12, 24,
36, 42, 48, 60, 72, 84, 96, 108 and 120 kg/m3, at 1 m/s and at 40 șC. There was a
significant (p<0.05) influence of mass load on drying kinetics. The drying rate
decreased as the mass load placed in the drying chamber increased. The results
provided by the RE model showed that mass load density did not affect De, but was
only observed to affect the mass transfer coefficient (k). The increase of the
number of particles placed on the drying chamber trays may disrupt the air flow
around the particles and create channeling. This fact increases external resistance,
and therefore contributing to reduce k.
The influence of power ultrasound was observed in the whole range of mass
load density tested increasing both De and k figures. The effect was not significant
(p<0.05) at mass load densities higher than 90 kg/m3. This fact may be explained
considering the decrease of the amount of energy per unit mass as the load
increases.
An important variable to take into account is the ultrasonic power level
applied. Drying experiments of carrot cubes (side 8.5 mm) and lemon peel slabs
(thickness 7 mm) were carried out at 1 m/s and 40 șC at different ultrasonic
powers: 0, 10, 20, 30, 40, 50, 60, 70, 80 and 90 W. A significant (p<0.05) influence
of this variable on the drying kinetics was observed. In the case of lemon peel, a
linear correlation was found between De or k and the ultrasonic power applied, and
this relationship was valid for all the range tested. Nevertheless in carrot drying, the
ultrasonic effects were negligible below ultrasonic power figures of around 20-30
W. Above this threshold, the linear relationship between De or k and the power
ultrasonic was also observed. Furthermore, the slopes found in the linear
relationships were 10 times lower in the case of carrot regarding those found on
lemon peel experiments. As a consequence, the ultrasonic effects were more
intense on lemon peel (porosity 0.4) than in carrot (porosity 0.04), which is
considered a low porosity product. Therefore, the influence of raw material on the
ultrasonic effects observed during convective drying is confirmed.
To test if the conclusions reached demonstrated a bias in the model, the
empirical model of Weibull was used in a complementary way to SRE and RE
diffusion models to describe the different kinds of drying experiments. The Weibull
model adequately described the drying kinetics of the different products, providing
percentages of explained variance similar to those found by the RE model and
much higher than the SRE model at low air velocities. The Weibull parameters (a
and ß) convey similar information about the influence of power ultrasound on
convective drying to that obtain through diffusion models.
Atmospheric freeze drying technology was addressed to consider the use of
power ultrasound in this process. Atmospheric freeze drying presents low drying
rates due to the use of air temperatures below freezing point. Power ultrasound
application may be considered interesting in these conditions, since it could
increase the drying rate without significantly heating the material. Thus, addressing
ultrasonic application in this process may be considered as a matter of relevant
research. Before addressing ultrasonic application, the influence of atmospheric
freeze drying on the quality parameters of a highly valuable product, like cod fish,
was considered. Thus, drying kinetics of granulated and cubic (side 5 mm)
samples of cod fish were carried out at different temperatures (-10, -5, 0, 15 and 30
șC) using a heat pump drier. Experiments were also conducted by combining
temperatures. Figures below freezing point (-10 or -5 șC) were used until moisture
contents of close to 0.4 kg water/kg product, while a high temperature (30 șC) was
applied in the last drying stage. Drying kinetics of cod fish cubes were fitted using
diffusion (SRE) and Weibull models. Different activation energy figures were
assessed from the effective moisture diffusivities identified for experiments carried
out below freezing point (-10 and -5 șC, 71.1 kJ/mol) and at high temperatures (15
and 30 șC, 30.7 kJ/mol). Higher quality parameters were found for samples dried
below freezing point, these samples presented higher brightness, lower shrinkage,
lower bulk density and higher rehydration ability than samples dried using hot air
(15 and 30 șC). Samples dried at 0șC, on the other hand, presented intermediate
quality parameters. The combination of temperatures (-10/30șC or -5/30 șC)
significantly increased the drying rate and provided similar quality parameters in
the samples to those dried under atmospheric freeze drying conditions.
Power ultrasound application would increase mass transfer rate in
atmospheric freeze drying processes without significantly heating the material, and
therefore without affecting product quality parameters. Based on ultrasonic
experiments an effective moisture diffusivity increase of about 55 % would be
expected by ultrasound application, similar to those obtain on experiments carried
out in this work. As a consequence, the drying time to reach a moisture content of
0.15 kg water/kg product would be reduced by 45000 s (12.5 hours) in experiments
carried out at -10 șC.