Recent Works on Powder Mixing and Powder Coating Using an Optical Measuring Methodt

During the last few decades a great amount of effort has been addressed to the description, qualitative in most of the cases, of the basic mechanisms operating in the mixing of particulate solids. Given a particular process, the specific mechanism by which mixing proceeds depends fundamentally on the characteristics of the motion of the powder within the mixer, which in tum depend both on the physical properties of the powders and the characteristics of the mixing equipment. Another aspect of the powder mixing process is that concerning the quality of the resulting mixture and the time required for the particular mixer in consideration to yield the mixture of the desired quality. To evaluate the quality of the mixture it is necessary to know its composition at different points, task that has traditionally been done by stopping the mixer at a prescribed time to withdraw a number of samples from as many places within the mixture and analyzing the composition of each one of the samples. This procedure is, however, impractical if the performance evaluation of different types of mixer in order to select the most suitable for a given duty is to be accomplished. Both the flow pattern of powder within the mixer and the temporal variation of the mixture quality can be followed up by a recently developed on-line type optical measuring method. The method has already been applied to the evaluation of the mixing characteristics of a number of mixers. The results of these works are reviewed in the first part of this paper. The second part of the review is devoted to the recent works of the authors on powder


Introduction
During the last few decades a great amount of effort has been addressed to the description, qualitative in most of the cases, of the basic mechanisms operating in the mixing of particulate solids.Given a particular process, the specific mechanism by which mixing proceeds depends fundamentally on the characteristics of the motion of the powder within the mixer, which in tum depend both on the physical properties of the powders and the characteristics of the mixing equipment.
Another aspect of the powder mixing process is that concerning the quality of the resulting mixture and the time required for the particular mixer in consideration to yield the mixture of the desired quality.To evaluate the quality of the mixture it is necessary to know its composition at different points, task that has traditionally been done by stopping the mixer at a prescribed time to withdraw a number of samples from as many places within the mixture and analyzing the composition of each one of the samples.This procedure is, however, impractical if the performance evaluation of different types of mixer in order to select the most suitable for a given duty is to be accomplished.
Both the flow pattern of powder within the mixer and the temporal variation of the mixture quality can be followed up by a recently developed on-line type optical measuring method.The method has already been applied to the evaluation of the mixing characteristics of a number of mixers.The results of these works are reviewed in the first part of this paper.The second part of the review is devoted to the recent works of the authors on powder The composition of a mixture of different colored powders can be known by measuring the intensity of the light reflected by the mixture, onto which a light of certain wave length and constant intensity from a light source is projected.The block diagram of Fig. 1 shows the photometer and the computer system developed for these purposes.The details of the measuring system are to be found elsewhere [ 1] and, as an illustration of the capabilities of the method, the calibration curves relating the intensity of the reflected light (expressed in terms of output DC potential) to the concentration of the key (colored) component for a number of binary mixtures are shown in Fig. 2.
In mixing experiments, several probes containing two optical fibers each (see Fig. 1) are inserted into the mixing vessel and fixed at dif-      ferent positions.In this manner, with the data supplied by the sensors and the calibration curve of the powder system in consideration, the composition of the mixture at different points can be continuously measured.

Classification of mixing processes
Mixing processes can be classified in to several types according to the physical characteristics of the powders to be mixed.With the purpose of establishing this fundamental classification, a number of experiments were carried out [2] in a vibrated high speed stirred mixer (Kawata Ltd., type SFC-5) using various combinations of powders and different operating conditions.The physical properties of these powders are listed in Table 1.As shown in Fig. 3 probes used for measuring the concentration were inserted horizontally at different depths into the vessel and also at different distances from the wall.In the experiments, first the A component (white) is placed within the vessel, which is then vibrated at the prescribed frequency Nv (Hz), and the impeller is also set in motion at the desired speed NR (rpm).Once the A component has been agitated for a while under those fixed conditions, the B component (colored) is added at timet= 0 onto the bed in the form of a pulse input.
The concentration curves obtained and their corresponding mixing curves are shown in Fig. 4. The degree of mixing appearing in theY-axis of the mixing curves is defined as where aM is the standard deviation of the concentrations measured at the six sampling points and a 0 is the standard deviation of the completely unmixed state, i.e. at timet= 0.
In the remaining of this section the typical In particular, in some cases it is clearly seen the periodic circulation of the key component (black) up and down within the mixer and, in other cases, the delay of the key constituent in reaching the lower part of the vessel.This information is useful for the understanding of the basic mixing mechanism operating in this mixer.
3. 1 Coarse powders with similar physical properties

1. 1 Proper operating conditions
When two powders of similar physical properties and good flowability are mixed, the powder flows homogeneously within the vessel, which means that the circulation period is the same everywhere (Fig. 4-1a), resulting in a monotonous increase of the degree of mixing (Fig. 4-1b).Also, the final degree of mixing attained is very good (almost 1 ).

1. 2 Bad selection of operating conditions
Fig. 4-2a, 2b shows the results obtained for the same powder system as in 1) but at different operating conditions (lower vibration frequency and lower impeller rotation speed).In this case, the mixing is poor, the circulation of powder within the vessel is very irregular and a complete mixed state cannot be attained even after a long mixing time.

2 Cohesive fine powders
Fine cohesive powders tend to form agglomerates, which have to be broken up in order to achieve a good mixture quality at the microscopic level.Therefore, in this case (Fig. 4-3a,  3b), the mixing time is much longer than in the case of large-sized particles.

3 Powders differing in particle size
In this case, the results depend on the way in which the components are charged into the mixer.

1 Addition of fines onto the coarser particles bed
As seen in Fig. 4-4a, the fines percolate 100 through the bed of the coarser particles, as a result of which a concentration gradient opposed to that existing at time t = 0 is progressively established.During this downwards flow of fines a point is reached at which the fines concentration is approximately the same everywhere.At this point, the degree of mixing attains its maximum value (Fig. 4-4b ).As the downwards flow of fines continues, the mixture loses progressively this temporary homogeneity and the degree of mixing decays until the steady state is reached.This is thus a typical case of mixing followed by demixing or segregation.

2 Addition of coarser particles onto the fine particles bed
In this case (Fig. 4-Sa, Sb), the larger particles, which occupy the upper part of the vessel at time t = 0, descend very slowly towards the lower part of the mixer.This downwards flow of large particles is not enough, however, to equalize the concentration within the bed and a mixture of poor quality is attained.As opposed to the former case, the degree of mixing increases monotonically.

4 Powders differing in density
This case (Fig. 4-6a, 6b) is somewhat similar to that of the mixing of different-sized particles.In fact, there is a downwards flow of the heavier component and at a given time the mixing curve passes through a maximum, after which segregation takes place.However, the segregation effect is not as strong as in the case of different particle size.

5 Coating of coarse particles by fine adhesive particles
Figure 4-7a, 7b shows the case of mixing of a coarse material with the same powder coated by a fine sized material.At the beginning, mixing proceeds in the same way as that occurring in Fig. 4-1 a, 1 b, since the physical properties of both components are (except the color) the same.After the concentration has been leveled off everywhere within the mixer, it is seen a further increase in the 'concentration' of the key component.Actually what is increasing is the blackness of the mixture.This occurs because the fine particles of dye (which are likely to exist initially in the form of agglomerates) attached to the black component, are being transferred to the white particles of their vicinity.In this manner, the dye is progressively being dispersed over the surface of all of the coarse material.This process will be later treated in more detail.

6 Electrostatic effect in powder mixing
It is observed (Fig. 4-9a, 9b) a very irregular pattern at the beginning, with mixing stages followed by segregation periods due to the difference in particle density.However, after a certain time the mixing curve becomes monotonous and a fairly good final degree of mixing is attained.The surface electrostatic potential was also measured continuously during this mixing run and it was found that the potential first increased and later decreased continuously attaining a highly negative steady value after about 15 minutes of mixing.The electrostatic behavior of the powder bed was thought to be the main cause of the irregular pattern observed in the mixing curve.
The cases discussed in this section are the typical mixing processes which can be found in the practical situations.In the remaining of the present review, two of these cases, namely, the mixing of cohesionless powders with similar physical properties and the coating of coarse powders by fine adhesive particles, are discussed more extensively.

Mixing of non-segregating cohesionless powders
Using the optical measuring method described above, the mixing performance of two stopper rod f1ber bundle multichannel photometer new types of powder mixers was studied.For the experiments, glass powder was mixed with the same powder coated by a black dye.Therefore, this is the case corresponding to Fig. 4-7a, 7b, discussed before.However, as long as macromixing is concerned, the coating stage occurring here can be neglected.Furthermore, since the powders used do not segregate, the final degree of mixing is very close to one and the rate of mixing (or equally well the mixing time) is the only parameter to take into account for the evaluation of the optimum operating conditions.The mixing time is taken to be that at which the concentrations measured by the probes become equal, that is, the point at which the concentration curves (Fig. 4-7a) converge.

1 Rocking mixer®
This type of mixer (Aichi Electric Co., RM-30) consists in a rotating drum with simultaneous rocking motion in the longitudinal direction (Fig. 5).The mixing mechanisms operating in this type of mixer are [3] diffusion and convection.Due to the rocking motion of the vessel, the powder circulates quite freely from the one end to the other of the mixer.Besides, the motion of the powder also has a diffusive component, consequence of the rotating movement of the mixer.Figure 6 shows the influence of the operating conditions on the mixing time.It is noticed that the rate of mixing, which is very low when only rotation is imparted to the mixer, is greatly enhanced by the rocking component of the vessel's motion.The effect of rocking is perhaps

2 Twist-hole mixer®
The mixing chamber of this type of mixer (Matsui MFG Co., Ltd., type TW-20) is cylindrical and the agitator is a twisted perforated plate (Fig. 7).
The flow pattern of powder within the mixer consists of [ 4] an upwards stream due to the rotation of the agitator, a fraction of which passes through the holes of the agitator and returns back to the lower part of the vessel.The relative amounts of upwards and downwards flows depend on the agitation speed, the inclination of the vessel (15° in the experiments) and on the total area of holes in the agitator.
In this study, the optimum values of charge ratio (volume of powder/volume of mixer), agitation speed and total area of holes in the agitation speed and total area of holes in the agitator were obtained.These optimum values resulted to be the following: charge ratio, 31 ~ 38%; agitation speed, 85% of the critical speed; and total area of holes, 25 ~ 30% of the agita-102 tor surface area.

Powder coating
When a fine cohesive powder is mixed with a coarser granular material, the structure of the resulting mixture consists of a layer of fines adhered on to the surface of the larger particles.In general, the fine powder, because of its inherent cohesiveness, is forming agglomerates.During the coating process the agglomerates of fines have to be broken up if a mixture of good quality at the microscopic level is to be achieved.Accordingly, the amount of mechanical energy inputted by the mixer plays a central role in this type of mixing process.This point has been clearly confirmed throughout our experimental work on powder coating [5, 6] using three types of mixers with different mechanical energy input.
These three mixers are the high speed stirred mixer and the rocking-type mixer described before, and the Angmill Mechanofusion Sys-tem® [7] (Hosokawa Micron Co., type AM-15F).The use of these three mixing equipments has permitted the elucidation of the basic mechanism involved in the process.A brief description of the mechanism is presented in the following.
At the beginning, i.e., in the completely unmixed state, the fine component is distributed in lumps within the bulk, thus existing large zones of the bed with no fine particle at all.Accordingly, the first step of the process consists in the reduction of the size of these zones.The agglomerates of fines attach to the surface of the larger particles existing in their vicinity.These few primary coated particles act as carriers of fines, distributing them randomly throughout the mixer.During their travel, the carriers suffer a vast amount of collisions with other large particles, most of which remain, at this early stage of the process, without fines on their surface.In these collisions, the transfer of fines (either as individual entities or in the form of agglomerates) between large particles takes place.The kinetics of fines transfer among carriers was studied experimentally [ 5] using the optical measuring method described above.It was found that the fines transfer kinetics resembled the kinetics of a second order autocatalytic chemical reaction, with a rate equation of the form where Xc is the mass fraction of powder that has been coated at time t, k is the coating rate constant, and m is the partial order of the coating 'reaction'.The reaction analogy was, however, limited by the fact that not only the rate constant k but also the 'partial order to reaction' m depended on the operating conditions.On the contrary, the overall order of reaction (two) is independent of the operating conditions and hence characterizes the coating process.
The fines transfer kinetics has also been simulated by a discrete model based on a population balance relating the distribution of fines within the carriers at a given time to that existing at a later time [ 8] .The results of the simulation are in agreement with the kinetics expressed by Eq. ( 2), although Xc is to be reinterpreted as the fraction of carriers which at time t are coated by the average number of fines per carrier existing in the mixture.
Once the fines transfer process has led to a spatially uniform distribution of fines within the carriers, the agglomerates of fines are gradually dispersed onto the surface of the larger particles, which results in an increase in the coated surface area.The dispersion of fines, which is actually occurring since the earliest stages of the process, takes also place by friction and collision between the particles.
In the rocking-type mixer, the mechanical forces acting upon the particles are relatively weak and do not suffice to break up the smaller or more cohesive agglomerates of fines and, as a result, the quality of the mixture attained is, from a microscopic point of view, relatively poor.
The high speed mixer, because of its higher mechanical energy input, yields mixtures of better quality.However, even this type of mixer was not able to disperse completely the agglomerates of fines in experiments with coarse particles of PMMA and fine particles of magnetite [ 6] .The degree to which the fines (black) are dispersed onto the large particles (white) can be qualitatively examined by measuring the blackness (or, conversely, the lightness) of the mixture.Figure 8 shows the temporal variation of the lightness of the mixtures prepared by the high speed mixer at different agitation speeds.It is readily noticed the influence of the mechanical force imparted by the mixer on the rate and degree of dispersion of fines.
The mixtures obtained through a five-hour mixing in the high speed mixer were furthered processed in the Angmill Mechanofusion System [7) (Fig. 9).The corresponding values of the lightness versus time are shown in Fig. 10.At the beginning, the lightness decreases, which means that the agglomerates of fines remaining after the five-hour mixing in the high speed mixer are further dispersed onto the carriers surface.This demonstrates once again the strong influence of the level of mechanical energy input on the quality of the mixture.Figure 10 also illustrates a peculiar feature which did not appear in the high speed mixer experiments.After a certain time, the lightness of the mixture passes through a minimum and then increases, reaching the steady state.This phenomenon is a consequence of the gradual penetration of the small spheres of magnetite into the body of the PMMA particles.The mechanofusion process results in a remarkable modification of the particle surface texture.Up to this point, the basic mechanism of the coating-mechanofusion processing of powders (illustrated in Fig. 11) has been reviewed.Further details concerning the modification of the physical properties of the powder (flowability, compressibility) as a result of the coating process are available in the original paper [ 6] .

Fig. 1
Fig. 1 Photometer and computer system

Fig. 6 Fig. 7
Fig. 6 Influence of rocking speed on mixing time at different rotation speeds

Fig. 8
Fig.8 Influence of the agitation speed on the temporal variation of the lightness of the mixture.(Highspeed stirred mixer)

Table 1
Physical properties of the powders used