Coating of Seed Particles in Tumbling Fluidized Bed by Atomizing the Suspensions of Clayey Particles t

Coating seed particles of 53 to 210 f.J.m with clayey particles was carried out by means of a tumbling fluidized bed in order to enlarge and reuse the clayey particles wasted in producing a carrier for driftless dust formulation. The influence of operat ing factors on the coating efficiency f/ and on the growth rate of seed particles was investigated. f/ varied depending upon thermal operating factors but was not in fluenced much by mechanical ones. It was found that f/ was related roughly to Rw defined as the ratio of the feed rate of water fed as slurry droplets to the maximum theoretical evaporation rate of water in a coating chamber. f/ increased with R w. The relationship between f/ and Rw varied with the additive ratio of binder PVA to clayey particles, Cb, and f/ became high with increasing Cb. The growth rate of seed particles was explained well by an equation derived on the basis of the mass balance of clayey particles in a coating chamber.


Introduction
Recently, the particle size of the dust among the various agricultural chemicals tends to be enlarged to reduce the public nuisances caused by the drift of fine particles generated on the spraying.
This agricultural chemicals called DL dust (particles size; 10 ~ 46 11m) contains few fine particles smaller than 10 11m. Therefore the fine particles smaller than 1 011m are generated as byproducts in large quantities, when they produce carrier of DL dust, which leads to a possibility of causing problems in the disposition.
As one of the means of solving the problem, the authors have investigated the coating of the seed particles (about 100 11m) with the above-mentioned by-product particles using a tumbling fluidized bed coaterO to develop a new type carrier for the agricultural chemicals, F micro-granules (particle size 63 ~ 210 11m). The tumbling fluidized bed coater has been devised as an apparatus that makes possible to coat particles less than 300 or 500 11m, which is a critical size for conventional fluidized bed coater. With regard to this apparatus, however, few studies have been reported so far and the instructions for optimum operating conditions to get desirable results have not been made clear yet at all. This work elucidates the effect of the operating factors of the tumbling fluidized bed coater on the coating efficiency, indicates the instructions for the optimum operating conditions. The growth rate of seed particles is also discussed in this work.
Roseki was a product from Kawatana factory of Goto Mining in Nagasaki pref. consisting mainly of quartz and agalmotalite. Kunimine clay was a product by Kunimine Kogyo Co., Ltd., which consisted mainly of quartz and feldspar and contained some clay minerals such as chlorite, sericite and talc.

1 Tumbling fluidized bed coater
A schematic diagram of a tumbling fluidized bed coater (Okada Seiko Co., Ltd., type SP-25) is shown in Fig. 1.
The air from a blower through a heater and damper is introduced into the coating chamber (250 mm¢) through the gap (10 mm) between a rotating disk and the chamber wall. By the disc rotation, the tumbling seed particles in the chamber are forced to exist in the vicinity of the chamber wall, where they are fluidized by the hot air. The coating particles dispersed in water are sprayed to such violently moving seed particles from a nozzle (Fuso Seiki, Co., Ltd., type Rumina ST-5, external mixing type pneumatic nozzle, liquid outlet diameter; 1.0 mm¢) set above in the center of the disc. The spray of the slurry is directed to the point 9 em away from the center on the disc face. The slurry droplets which adhered to the sur-  Fig. 1 Schematic diagram of coating apparatus 62 face of the seed particle are dried by the hot air, so that the particles in the droplets coat the surface of the seed particle. A part of the coating particles which failed to adhere to the seed particles as a result of peeling etc. are discharged from the coating chamber and collected in a cyclone on the way to the air exhaust of this system.
The experimental procedures were as follows. The disc rotation speed and the flow rate of inlet air were adjusted after charging the coating chamber with a required amount of seed particles. When the temperature of the air reached a fixed value, the slurry was fed to the nozzle at a constam rate by a gear pump and began to spray uncl cr the constant air pressure (0.294 MPa). During the experiment, the air temperature after passing through the fluidized bed was measured with a thermocouple set near the nozzle. After a fixed coating time, the coated particles were taken out from the outlet, as soon as the feed of the slurry was stopped. The representative sample of 0.04 ~ 0.06 kg out of the product taken out from the chamber was sifted through the standard sieve (37 pm) after stirring in hot water (343 ~ 353 K, 400 cm 3 ) for 4 hours. Having had confirmed that the particles of Roseki and the Kunimine clay passed the standard sieve (37 pm) entirely, the undersize was regarded as the coating particles. The mass of the undersize per total mass of the seed particles supplied to the coating chamber was adopted as the total mass of the coating particles adhering to the seed particles.
In this article, the coating efficiency 11 is defined as the ratio of the total mass of the coating particle adhering to the seed particles to the total mass of the coating particles fed into the coating chamber during the coating operation.
As for the operating factors, there are the mechanical operating factors such as the disc rotation speed N [s-1 ] and the mass of seed particles Ms [kg]  Additionally, PV A (Nippon Gosei Kagaku Kogyo Co., Ltd., Gohsenol GL-OSS) was used as a binder to improve the coating efficiency. The influence of the additive ratio of binder Cb on the coating efficiency was also examined.
The concentration of PV A in the water solution is 0.5 ~ 1.2 wt% under the condition of cb = 0.02, s = 0.2 ~ 0.37' and 0.2 ~ 1.7 wt% in the case of s = 0.3, Cb = 0.005 ~ 0.04.
According to Yamada et al., Roseki slurry with PV A indicates the flow characteristic of thixotropy, newtonian or dilatancy depending upon the additive ratio of PVA 2 ) and the viscosity increases with increasing slurry concen-tration3).
Roseki slurry in this work showed dilatancy when s = 0.3, cb = 0.02, and the apparent viscosity was 2.8 x I0-1 ~ 4 x I0-3 Pa·s at the shear rate of 81 ~ 40 1 s -1 and the tern perature of 307 K. In addition to the above-mentioned coating experiments, the coating experiments (using Kunimine clay as coating particles and calcite as seed particles) were carried out and compared with the case of coating ground Toyoura sands with Roseki, in order to examine whether the relation between the operating factors and coating efficiencies depends on the physical properties of the seed and coating particles or not. The coating time was 0.5 hr in all the experiments except the one conducted to ex-amine the growth rate of the seed particles.
The nozzle height is another mechanical operating factor in addition to the disc rotation speed and the mass of the seed particles. Harada et al. 4 ) have made clear already that the coating efficiency of a conventional fluidized bed coater is decreased as the nozzle height increases. Therefore, it becomes a problem how near the distance between the nozzle and the disc face can be fixed for the troubleless operation.
In preliminary experiments under the condition of Ms = 0.6 kg and N = 1.5 s-1 when the nozzle height varied from 22, 16, 13 down to 1 0 em, the seed particles on the disc face were eliminated by the air flow from the nozzle and the coating particles adhered to the disc face directly in case of the nozzle height less than 13 em. On the other hand, no difference in the coating efficiency between the nozzle heights 16 and 22 em was noticed. Therefore, all the following experiments were conducted with the nozzle height of 16 em or 22 em.

2 Measurement of the porosity of the coating layer
To consider the growth rate of the seed particles, the value of the porosity of the coating layer on the surface of particles must be known. The porosity was measured with two methods, a method of measuring the packed volume of the seed particle before and after the coating individually and the mercury porosimetry. 5 ) Assuming that the particles have the same shape before and after the coating and that the porosity Es in case of packing non-coated particles equals the one of interstices of particles in the case of packing coated particles, the porosity of the coating layer £ is given by,

VO-£s)-ms/Ps
(1) where ms [kg] is the mass of the seed particles packed in a vessel, me [kg] is the mass of the particles coating the seed particles having the mass of ms, V [m 3 ] is the packed volume of the particles having the mass of ms + me, P e, Ps [kg/m 3 ] is the density of the coating particles and the seed particles respectively. V was measured by tapping a vessel of 28 mm diameter under the condition of the stroke 1 em and counts of tapping 200.
On the other hand, POROSIMETO SERIES 1500 (CALRO ERBA STRUMENTAZIONE) was used as the instrument of the mercury porosimetry.
3. 3 Measurement of volume shape factor of the seed particles before and after coating To consider the growth rate of the seed particles, the value of volume shape factor is required in Eq. (11) that will be described later in this article.
On the basis of Eq.
(2), the volume shape factor rf>v was obtained by counting the number nP of the particles having the size D and the mass MP enlarged with a projector.

MP
( 2) where D is the arithmetic mean of sieve aperture, Pp is the average particle density calculated from Eq. (3) using the porosity € of the coating layer. (3)

1 Droplets size
The results of the photographical size analysis of more than 1000 droplets collected by the immersion liquid method 6 ) are shown in Fig. 2.
It's seen that the droplets having the mass median diameter of 17 ~ 28 pm are formed under the following conditions; the additive E' Ratio of mass feed rate of suspension to that of air  Table 1, the ratio of mass feed rate of slurry to that of air from the nozzle was within the range of 0.33 to 0.69.

2 Influence of thermal operating factors on the coating efficiency ( 1) Evaporation of water in fluidized bed
Some examples of the relationship between the thermal operating factors and the coating efficiencies are shown in Fig. 3.
As seen from the figure, the coating efficiency 71 increases as the feed rate of water WE increases when other conditions are kept constant. Conversely, 71 decreases with increasing the difference between the adiabatic saturation humidity and the humidity Hs-H, the temperature ti, or the volume flow rate of the inlet air Q. These experimental results lead to the following consideration.
In Fig. 3 (a), the evaporation rate of water in the fluidized bed must have a maximum value because the operating factors except WE are kept constant. If only WE increases under the condition, a certain amount of water exceeding the evaporation rate is accumulated in the fluidized bed.
Similarly in Figs. 3 (b), (c) and (d) the decrease of Hs-H, ti or Q makes the accumulation rate of water increase, if the other factors are constant. Above-mentioned considerations suggest that the coating efficiency 71 might be determined by the accumulation rate of water in the fluidized bed and 71 becomes higher as this rate increases. The evaporation of water in the tumbling fluidized bed will be discussed in the following.
If the tumbling fluidized bed coater is regarded as an adiabatic humidifier in terms of the inlet air, the variation of air temperature and humidity through the apparatus ought to be illustrated as Fig. 4.
The air of temperature t r and humidity H from the btower is heated by the heater until the temperature reaches ti. This heated air "'   gives the latent heat of vaporization to the slurry droplets in the fluidized bed. As a result, the temperature of the air decreases along the adiabatic cooling line through the point (ti, H) and the humidity increases. If the contact time between the air and the seed particles is assured long enough, the maximum theoretical evaporation rate of water in the coating chamber Wr [kg H 2 0/h] is ex- KONA No.4 (1986)  is the temperature at the humidity of H +WE/ We on the adiabatic cooling line of the inlet air. Figure 5 shows the comparison of the measured values of the air temperature in the fluidized bed obtained from all the coating experiments in this work with the theoretical values from a humidity chart on the basis of the above-mentioned procedure. The difference between the measured values and the theoretical ones is at most within ±2 K as seen from the figure so that it is considered that adiabatic humidification takes place actually in the tumbling fluidized bed coater. (2) Effects of moisture on the coating efficiency As is already mentioned, adiabatic humidification takes place practically in the tumbling fluidized bed coater and the coating efficiency tends to become higher as the accumulation rate of water in the fluidized bed increases.
Since the accumulation rate of water in the fluidized bed is determined by the synthetic effect of the thermal operating factors, we use Rw defined by Eq. (6) as an index to evaluate the synthetic effect of the thermal operating factors and examine whether the coating efficiency can be definitely related to Rw. WE Rw=- (6) Wr Figure 6 shows the relation between the coating efficiency 77 and Rw under the condition of Cb = 0.02, N = 1.5 s-1 , and Ms = 0.6 ~ 0.69 kg.
The figure indicates that as far as the thermal operating factors vary in the range shown in Table 1, 77 is strongly correlated with the index Rw.
The relationship between 77 and Rw at slurry concentration of s = 0.5, however, shifted to high efficiency in the figure. The reason has not been clear and is the subject to be investigated in the future. Therefore, it is not to affirm that 77 should be determined by Rw de- finitely without any restriction. Nevertheless, the relationship between 77 and Rw is considered to be valuable since more synthetic instructions to find the optimum operating conditions can be obtained from the relations than from those between the individual operating factor and the coating efficiency shown in Fig. 3. The broken line in the figure indicates the relationship between 77 and Rw in case of N = 1.5s-1 shown in Fig. 6. Though a slight difference is noticed between the experimental results in case of N = 2.1 s-1 , 3.3 s-1 and N = 1.5 s-1 in the figure, the coating efficiency is hardly influenced by the disc rotation speed as far as it varies from 1.5 s-1 to 3.3 s-1 • The relation between the mass of the seed particles Ms and coating efficiency 77 is shown in Fig. 8. The figure makes clear that 77 scarcely varies with Ms. Therefore, the mass of the coating particles per unit mass of the seed particles Me /Ms naturally decreases in inverse proportion to the mass of the seed particles as seen in The reason why the coating efficiency 11 be-68 comes high with increasing Rw and a parameter cb will be discussed in the following.

3 Influence of mechanical operating factors on the coating efficiency
The coating operation in the tumbling fluidized bed coater may be regarded as a process where the sprayed slurry droplets being dried by the hot air collide with and adhere to the seed particles being tumbled or fluidized and the coating particles in the droplets coat the surface of the seed particles by further drying.
Regarding the slurry droplets as water droplets to simplify the phenomena, the water droplets need longer time with increasing R w for disappearance by the evaporation in the range of Rw < I. On the other hand, when Rw > I, only a fraction of Wr in the fed water is evaporated and the water droplets equivalent to WE-Wr still remain. Moreover the higher R w is, the larger are the diameters of the remained droplets. Therefore the probability of the collision between the droplets and the seed particles which would dominate the coating efficiency becomes higher with increasing Rw, whether Rw is smaller than I or not. Consequently, the coating efficiency will become higher with increasing Rw.
According to the above-mentioned considerations, the increase of Cb is surmised to cause more droplets to collide with and to adhere to the seed particles and to give rise to the higher coating efficiency in consequence, because the drying rate of the droplets of PV A solution is small in comparison with that of the water alone 7 ).

S Variation of the coating efficiency with
the difference in physical properties of the seed particles or the coating particles Figure 10 shows the experimental results in the case of coating calcite having the size of 297 ~ 590 pm or ground Toyoura sands as the seed particle with Roseki and the case of coating calcite with Roseki or Kunimine clay. The experimental conditions were s = 0.5 and Cb = 0.02 in the both cases.
The coating efficiency 11 is correlated with Rw as well as in the case of Fig. 9.
If the seed particles differ in material such as calcite or Toyoura sands as well as in particle size, the relation 11 vs Rw is unvaried as far as the coating particles are the same. Conversely, the relation 11 vs Rw is different obviously depending upon the physical properties of coating particles if the same seed particles are used. For example, the coating efficiency of Kunimine clay is lower than that of Roseki at the same value of Rw. This seems to be due to the fact the tensile strength of the wet powder bed reduces as the particle size increases 8 ). It would be also attributed to the trend that the sprayed droplets of larger solid particles contain less fraction of liquid as confirmed by Yamada et a1. 2 ). In this way, the particles of Kunimine clay are less adhesive to the seed particles than Roseki and are more likely to come off from them even having adhered, since the median diameter of the former is twice as large as that of the latter.
To examine the coating efficiency more minutely and quantitatively, hereafter it is necessary to analyze the drying process of the slurry droplets in the tumbling fluidized bed and that in the coating layer of fine particles on the surface of the seed particles, particularly the case with addition of high molecular binder such as PV A.  Fig. 9, it is obvious that the coating efficiency is not affected by the coating time.

6 Growth rate of the seed particles
Harada et al. 4 ) derived the growth rate equation of uniform sized spherical seed particles in the case of coating with a eductional solution (sucrose) in a conventional fluidized bed.
In the present work, the authors deal with the coating of the seed particles having size distribution with fine particles. Then the variation of the mean diameter of the seed particles with time in the coating with fine particles is discussed below.
From the material balance of the coating particles, the following equation is obtained So~nf(D,8 )<t>vD 3 dD-S 0~n fo (D,8) In Eq. (11 ), e equals 0 for the coating with eductional solutions, and ~vli"vo equals 1.0, unless the shape of the particles varies during the coating. In this case, Eq. (11) is reduced to the equation proposed by Harada et al.
Experimental values of the volume mean diameter at each coating time are compared with the calculated ones from Eq. (11) in the following. To make use of the equation, the volume shape factor and porosity of the coating layer must be known. Table 2 shows the measured values of the volume shape factor of the particles sifted in the ranges of 125 ~ 149 J.tm and 177 ~ 210 J.tm. It is clear from the table that though there exists a difference in the volume shape factors between the two groups, they increase with increasing e, which means the particles get spherical gradually with time. This tendency was also confirmed visually on a projector. 70 Table 3 shows the measured values of the porosity of coating layer e by the packing method. The value of the porosity of the noncoated seed particles es was 0.37. The measured values of e were 0.5 5 regardless of the coating ratio mc/ms.
The relation between the pore radius and the specific pore volume measured by the mercury porosimetry is shown in Fig. 12.
The pore volume consists of both the volume of interstices of coated particles and the pore volume of coating layer, as generally seen on the penetration of mercury into a layer of fine granules 5 • 9 ). The median of the pore radius of coating layer is about 0.4 J.tm. The specific pore volume of the coating layer gives the porosity of 0.54, which is nearly equal to the value obtained by the packing method.
The assumption at the time when Eq. (1) was derived is not valid in the strict sense because the particle shape varies with coating time as shown in Table 2. However, the values of the coating ratio demonstrated in Table 3 are not so large that the difference of the volume shape factor before and after the coating is small. Hence, it appears that the values of porosity obtained by both methods were almost the same. Table 4 shows the experimental values of   11) regarding as e = 0.54, the variation of the particle shape being taken into account and not respectively.
15 was calculated from Eq. (12) on the basis of the particle size distribution shown in Fig.  11, assuming the volume shape factor was independent of particle size.
In calculation of Eq. (ll ), the arithmetic mean values of volume shape factor of two groups shown in Table 2 w~re used for if>v, if>vo.
As seen from the table Deal calculated taking the variation of shape factor with time into account agrees better with the measured values than Deal' obtained assuming if>vo /fv = 1.

Conclusions
Coating of seed particles in the tumbling fluidized bed was carried out by atomizing the suspensions of clayey particles. Investigating the influence of operating factors on the coating efficiency and the growth of seed particles, the following conclusions have been obtained.
( 1) The coating efficiency hardly depends KONA No.4 (1986) on the mechanical operating factors such as the mass of the seed particles, the disk rotation speed and the coating time.
(2) Introducing the additive ratio of binder, Cb, as a parameter, the coating efficiency T/ is roughly related to and increased with Rw, which is an index to evaluate the synthetic effect of the thermal operating factors and the ratio of the feed rate of water as slurry droplets to the maximum theoretical evaporation rate of water in the tumbling fluidized bed coater.
T) becomes high with increasing cb under the same value of Rw. There exist critical values of Rw depending upon Cb beyond which it is no longer possible to continue the coating operation.
The instructions for the optimum operating condition of the tumbling fluidized bed coater can be obtained according to the relationship between T/ and Rw.
(3) The growth of the seed particles can be known as the variation of the volume mean diameter with coating time calculated from the growth rate equation proposed in this article on the basis of the material balance of the coating particles.