1 Introduction
Spray drying is widely used in the industry for conversion of a suspension or solution into a dry powder product. In spray drying the suspension or solution feed is atomized and the droplets formed are contacted with
a hot gas.
The contact between the droplets and the heated gas causes the solvent in the droplets to evaporate, leaving a dry powder product. The spray drying process is described in detail by Masters [1].
In the ceramic industry, the powders produced by spray drying are usually used in pressing machines with subsequent sintering.
For the pressing process to be efficient the powder must have the right flow characteristics and the properties of individual particles, such as density and porosity, are also important. The final powder and particle properties are determined in the short process in the spray dryer where the droplets
are dried into particles.
To design particles with better properties it is important to have detailed information about the single droplet drying process taking place inside spray dryers. However, it is hardly possible to investigate the single
droplet drying process using conventional spray drying equipment because the droplets inside the dryer are travelling at high velocity in a chaotic flow field.
Therefore GEA Niro has developed the DRYNETICS™ concept which allows detailed investigations of the single droplet drying process. By using
DRYNETICS™ it is not only possible to investigate the morphology (i.e. external and internal particle structure) formation process, but it is also possible to obtain important processing parameters which may be
used for optimizing the design of new spray dryers and improving the operation of existing equipment.
In this paper the DRYNETICS™ concept is explained in detail. This includes a description of an experimental apparatus which has been developed to investigate the morphology formation of drying particles and to measure the specific drying rate. Additionally, it is explained how the data for the
drying rate are coupled with computer simulations of full scale spray dryers in order to optimise the dryer design and operation.
2 Equipment – the DRYING KINETICS ANALYZER™
To investigate the process of drying for a single droplet, GEA Niro has developed an experimental apparatus – the DRYING KINETICS
ANALYZER™ (DKA). The DKA is based on the principle of ultrasonic levitation as illustrated in Figure 1. An ultrasonic field is generated between the transmitter and reflector. Due to the forces of this ultrasonic field it is possible to hold a small droplet constant against gravity.
While the droplet is drying it is monitored by a CCD-camera and an infrared thermometer. The former is used to record a video file of the drying process while the latter is used to measure the droplet surface temperature during the drying process. The levitation unit is encapsulated in a small drying chamber so that the air temperature and humidity may be set to match the conditions in a spray dryer. Also, the conditioned drying gas is injected through small holes in the reflector (Figure 1) below the droplet to simulate the relative velocity between the gas and droplets in a spray dryer.
It is, however, important to note that the forces of the ultrasonic field and not the injected drying air are levitating the droplet.
Using the DRYING KINETICS ANALYZER™ the morphology formation may be investigated in different ways. Firstly, it is possible to review the video recorded during an experiment – an example is given in Figure 2.
Investigations of this kind may be conducted at, for example, different drying air temperatures to map the effect of this parameter.
Once the particle has been completely dried it is possible to recover it from
the ultrasonic field. Then the particle may be subjected to the analysis by optical or scanning electron microscopy as shown in Figure 3. It is also possible to conduct different kinds of analysis such as XPS (X-ray
photoelectron spectroscopy) or Raman spectroscopy to measure the distribution of different compounds – such as binders – inside
the dried particles. Beyond, the opportunity to study the morphology
formation process in detail, single droplet drying experiments offer advantages over experiments using pilot or full-scale spray drying equipment:
The experiments are quickly conducted; very little feed material is required. It is possible to conduct a full study of a feed by using less than 25 ml of feed.
The process parameters which can be used in the DKA are summarised in Table 1. Constant values for the drying gas temperature and humidity are used. The interval of these two parameters given in Table 1 should be
compared to the outlet conditions of a typical spray dryer.
A droplet drying in a spray dryer experiences the outlet temperature almost throughout the entire course of drying because the gas in the spray dryer is well mixed. Therefore, choosing the spray dryer outlet gas conditions in the DKA resembles a spray drying process best.
3 Investigation of morphology formation
The powder properties of a spray dried product such as mechanical strength, flowability, density, redispersion rate are very important.
These properties are determined by the final morphology of the individual particles. Both the feed formulation and the process parameters (e.g. drying gas temperature and humidity) determine the particle morphology. It is very difficult to predict the morphology outcome of a specific process
even if the formulation ingredients and process parameters are well known.
However, the importance of this subject is underlined by the extensive attention in scientific literature. Examples are the studies by Walton
[2–3] who characterised different morphologies by investigating a large number of different powder products dried at varying
air temperatures. Several authors propose theories to explain the formation of particles featuring indentations, donut shapes, voids, etc. [4–8].
As already mentioned in section 1 (“Introduction”) one of the two parts of the DRYNETICS™ concept is aimed to investigate the influence of different spray drying parameters being relevant for the particle morphology formation. This means that the investigations have to include parameters
such as drying air temperature and humidity but also feed properties such as composition of the feed, choice of binder and initial solids content.
4 Drying kinetics
From the previous section it is clear that every feed formulation has unique drying properties. However, this does not only apply to the morphology formation but also to the drying kinetics. The drying kinetics of a
suspension droplet is often defined as the rate of evaporation measured in unit mass of solvent per unit time or simply the residual droplet moisture content as a function of the drying time. A large number of parameters
influences the drying kinetics of a single droplet, for instance the drying gas
temperature and humidity as well as the relative velocity between the droplet and the drying gas. Equally important are the properties of the suspension being dried. The initial solids content obviously has a great impact.
But the addition of a binder could retard the migration of solvent from the droplet centre to the surface, which effects a decrease of the drying rate. The impact of the binder on the drying kinetics depends on the specific choice of the binder and the amount added.
It is very difficult to predict how a change in process parameters or formulation ingredients influences the drying kinetics.
However, as described in section 5 (“Computer simulations”) it is very important to obtain this knowledge to be used for designing and optimisation of spray drying equipment.
The importance of detailed knowledge of drying kinetics is well known and many studies have been devoted to this subject. Early studies include the investigations by Ranz and Marshall [9–10] who set up very accurate
mathematical models for the drying of pure liquid droplets. Numerous later studies are based on the results of Ranz and Marshall Most relevant to the ceramic industry are the works of Cheong [11], Malakhovskii [12],
Sano [13] and Jørgensen [14].
As part of the DRYNETICS™ concept it is possible to calculate the drying kinetics of a given feed using the measurements obtained
from the DKA. The following measurements are needed:
•• the droplet surface temperature measured by the infrared thermometer;
•• the droplet size found by analyzing the images taken with the CCD-camera;
•• finally, the vertical position of the droplet in the ultrasonic field.
The latter is also found by image analysis and it is an indication of the particle mass: as soon as the particle dries it becomes lighter and slightly rises in the ultrasonic field.
As already mentioned the drying kinetics is found from these measurements through advanced mathematical modelling. The corresponding principle is discussed in [15] and illustrated in Figure 4.
The mathematical modelling is also used to extrapolate the results for the drying kinetics.
Experiments regarding the drying kinetics are usually just performed at a few different conditions (drying gas temperature, humidity
and velocity). The mathematical model is used to calculate the drying kinetics for further conditions [15]. Hence, the drying kinetics can be obtained for all conditions relevant to spray drying. This is very important for conducting computer simulations dryers in industrial-scale, as explained in the next section.
5 Computer simulation
The spray drying process has become a frequent subject for Computational Fluid Dynamics (CFD) modelling [16–17]. In CFD a three-dimensional model of the spray dryer is drawn by a computer. Using this model, commercially available CFD software like FLUENT™ allows the calculation of temperatures, gas velocities, particle trajectories, etc. inside the spray dryer.
To obtain useful results from the CFD simulations it is necessary to include calculations of the evaporation process occurring within the drying droplets. Using the FLUENT™ CFD software it is possible to include the evaporation process by applying a simple drying model which assumes that
the feed droplets are drying like pure liquid droplets. But this assumption leads to a too optimistic drying rate as any solid material in the droplets will retard the drying. To overcome this problem a user-defined
function has been developed as part of the DRYNETICS™ method. This user-defined function couples the results of the drying kinetics based on the DKA measurements with the commercial FLUENT™ CFD software.
Consequently, very accurate results from the CFD software are obtained because the unique drying properties of the feed in question are included in the simulations.
6 Improving the spray dryer design
The coupling between the DKA experiments and the CFD simulations is an important part of the DRYNETICS™ method because it improves the accuracy of the CFD simulations so that these may be used for evaluation
of process parameters and equipment design. This can be elucidated for example by a validation study where results of the simulations are compared to a full-scale experiment.
The dryer considered (Figure 5) was an existing small-scale dryer with a top
air disperser for inlet of the heated process air. The feed was atomised into the drying chamber by a pressure nozzle located in the centre of the air disperser. The exhaust air exited the drying chamber along with the
dried particles through an outlet in the bottom of the dryer. The mass flow of the drying air was 1900 kg/h while 90 kg/h of feed was spent. A non-ceramic compound – maltodextrin DE18 – with an initial solids content
of 29 mass-% was chosen. If the dryer is operating under inappropriate conditions this compound leaves distinct rubbery deposit.
It is explained below why this feed property is important. In accordance with the DRYNETICS™ method a drying kinetics model of the feed
was set up based on measurements on the DKA. This model was used in CFD simulations of the small-scale spray dryer choosing an air inlet temperature of 179.5 °C which corresponds to an outlet temperature
of 75.8 °C. The results are illustrated in Figure 6. To generate an easily interpreted image it simply shows the tracking of a few hundreds of particles. Overall several hundred thousand particles were tracked during
the simulation. In Fig. 6 the moisture content of the particles is shown. Figure 7 reveals a problematic area in the spray dryer where green (half-dried) particles hit the conical part of the dryer. The right hand side of Figure 7 points out the exact area. When half-dried particles hit the walls of
the chamber there is a great risk that they will stick to the wall as deposits rather than come out as a powder product. This can lead to severe equipment fouling, which means that the plant has to be shut down for cleaning.
The DRYNETICS™ analysis was conducted using an existing spray dryer as mentioned above. Therefore, it is appropriate to compare the results of the analysis with a spray drying experiment from this plant. The plant has to operate under the same conditions used for the DRYNETICS™ analysis.
The results of the spray drying experiment are shown in the left hand side of Figure 7.
This picture was taken inside the spray dryer after the experiment and it shows the lower conical part of the spray dryer. Although some dry powder is lying on the conical part, a distinct rubbery deposit of maltodextrin
can be seen. The area where this deposit has formed corresponds to the area predicted by computer simulation. The DRYNETICS™ method is normally used before the spray dryer is manufactured. This gives engineers
the opportunity to modify the process parameters or the equipment design in order to avoid areas with potential risks regarding equipment fouling. In the example above, the problem may be remedied by increasing the air inlet temperature. In other cases this is not possible. If so it might be an
option to redesign the air distributor to get a better air flow field inside the drying chamber.
In some cases it is helpful to consider a larger size chamber.
7 Conclusions
In this paper the two parts of the DRYNETICS™ concept are presented: One part concerns the morphology formation during the drying of single droplets. The other part is related to the coupling of drying kinetics
measurements with computational fluid dynamics in order to achieve high-precision results. It is planned to use the first part of the concept to improve the quality of existing products as well as the aid product development.
Hence, the particle morphology and the resulting powder properties can be
optimised. Additionally, further experiments have to be performed to obtain a deeper understanding of the morphology formation process.
The computer simulations are well suitable for the troubleshooting of existing plants, for designing new plants as well to develop
better components for spray dryers.