An Investigation into the Influence of Suspended Glass Particles on Bubble Diameter, Gas Holdup, and Interfacial Area in an Agitated Tank

An investigation into the influence of suspended glass particles on bubble diameter, gas holdup, and interfacial area in an agitated tank.

Effect of 200,Am glass particles on local 43 gas holdup and bubble diameter with increasing solids concentration at N = 250 RPM.

10
Literature comparisons. 57 x Al Local distribution of bubble diameter, 70 gas holdup, and interfacial area in an air-water dispersion (HA = T/2).

A4
Local distribution of bubble diameter, 79 gas holdup, and interfacial area in an air-water-solid (25~m) system at 0.6 wt percent for the upper impeller region (HA = T/2).

AS
Local distribution of bubble diameter, 81 gas holdup, and interfacial area in an air-water-solid (25,A.m) system at l.2 wt percent (HA = T/2).

A7
Local distribution of bubble diameter, 86 gas holdup, and interfacial area in an air-water-solid (70~m) system at l.2 wt percent (HA = T/4).

A8
Local distribution of bubble diameter, 88 gas holdup, and interfacial area in an air-water dispersion (HA = T/4).

A9
Local distribution of bubble diameter, 90 gas holdup, and interfacial area in an air-water-solid (200~m) system at 0.3 wt percent (HA = T/4).

All
Local distribution of bubble diameter, 92 gas holdup, and interfacial area in an air-water dispersion at a higher gas rate of 3CFM (HA = T/2).  Calderbank (1958).

2
The relation between the 17 observed two phase flow and the power consumption. 3 The gassed power curves for 17 different impeller speeds. From Warmoeskerken (1982 In many applications the solids are finely divided and generally fall into one of the following five categories: (1) Gas absorption into slurries, usually with some kind of chemical reaction; (2) precipitation of a solid resulting from absorption of a gas into a liquid; (3) slurry absorption of gases; (4) slurry adsorption of gases; and (5) slurry sorption of gases. A brief description of these processes and the influence that solids has on the physical characteristics are described below to demonstrate the diverse nature of slurry reactions.  (Shreve, 1956); absorption of sulfur dioxide into milk or lime as in the paper industry or water slurries of lime or limestone to remove so 2 from furnace gases (Mallette, 1955); chlorination of paper pulp (Shreve, 1956); aerobic fermentation (Blakeborough, 1967, Peppler, 1967; aeration of activated sludge in sewage treatment (Eckenfelder, 1963); and absorption of co 2 in thermal coal salvation with associated products and byproducts. Ramachandran and Sharma (1969), Uchida et al (1975, Uchida and Wen (1977), Sada et al (1977aSada et al ( , 1977bSada et al ( , 1977c, Tsao and Kempe (1960), Bennette and Kempe (1964) and Tsao et al (1972) all have looked at this type of system.
For the reactant solid, it has been found that the rate of absorption is considerably higher for small particles, that is smaller than the gas-liquid diffusional film thickness (Uchida et al, 1975). For large particles (diameter greater than the liquid film thickness) Sherwood and Farkas (1966), Satterfield (1970), and Zaidi et al (1979) indicates that the resistance in series concept works well for the system. For example, 2 Sherwood and Farkas analyzed the hydrogenation of methyl styrene and cyclohexane using 55mm and 30mm size palladium black as a catalyst respectively. For the hydrogenation processes, the gas dissolution, diffusion, reaction in series model would apply.
Solids can also behave as catalyst.
The addition of activated carbon to a gas-liquid system showed increases in kL (Alper et al 1980, Wichtendahl (1978 and Kars et al (1979) (Kohl and Riesenfeld, 1974) or in the Citrex process where so 2 is absorbed in a buffered citrate process and then 3 countercurrently contacted with HS to precipitate sulfur (Vassan, 1975).
In the above mentioned processes, the precipitated sulfur is of micron size. It is suggested that the solids coat the gas bubbles and act as a barrier to mass transfer of the gas phase.
may reduce kL considerably.

Slurry Absorption of Gases
Also the particle barrier This type of operation is commonly found in the air pollution control industry in which a solute gas is removed from a gas mixture by absorption into a slurry.
The controlling step is the rate of absorption of the gas into the liquid phase. The solids are usually inert. An investigation by Lee et al (1982)  Other studies of this type of operation are by Joosten et al (1977) and Schmitz et al (1982) who indicated little influence on gas-liquid interfacial area with moderate solids concentration.

Slurry Adsorption of Gases
This operation is commonly used in purifying 4 gaseous streams.
Resistance to mass transport through the ·liquid phase near the gas bubble to near the particle is negligible for agitated slurries. This was confirmed by experimental studies (Kolbel and Siemes (1957); Siemes and Weiss (1959). So the predominate mass transfer mechanism is the adsorption of the gas by the solids. Mehta and Calvert (1967) found that the adsorption capacity could be as high as that for the dry particles.
In this case, it is better to adsorb the gas into slurries because it is easier to handle for continuous operation and regenerating adsorbing sites. Misic and Smith (1971) studied a similar type of system where adsorption capacities of aqueous slurries of carbon particles were established for benzene.
The investigation of gas-liquid-solid agitated systems is complicated by the lack of understanding of gas-liquid agitated systems. Although there has been an effort to unify the results for both gas-liquid and gas-liquid-solid contactors, great confusion remains.
To date, little is available in the literature as to the influence particle 5 properties have in relation to gas holdup, bubble diameter or gas-liquid mass transfer rates.
In order to compare funct i ona 1 it i es of two phase with three phase systems, a fundamental understanding of the past research and methodology performed is essential. So the following will discuss some of the theory and work which has been done in the area.

Bubble Diameter in Dispersions
There are various methods for measuring the bubble sizes. Calderbank (1958) and Lee and Meyrick (1970)  Here bubble size depends on a balance of forces due to surface tension and to turbulence. Yoshida and Miura (1963) Ganguli (1975) observed this for insoluble surfactants in liquids. Levich (1962) also predicts this type of behavior in his theoretical work.

Gas Holdup in Dispersions
The functional gas holdup in a gas-liquid system is defined as the volume of gas divided by the total of gas volume and liquid volume. Normally, gas holdup is measured by observing the change in height above the tank bottom between the gas-liquid dispersion and the ungassed liquid.
The smaller fines can be adsorbed at the gas-liquid interface and thus strengthen these by playing a role analagous to soluble surfactants. This phenomenon is strong for low solid concentrations.

Interfacial Area in Dispersions
Interfacial area is an important mathematical parameter which appears in most mass transfer models. Studies in the past to determine this parameter have been generally applied to gas-liquid dispersions using one of these methods: (1) chemical measurement of surface area, (2) light transmission, or (3) photography. Each of these methods can give good results in gas-liquid dispersions but appear somewhat impractical in a gas-slurry reactor.
The chemical technique involves absorption followed by a fast chemical reaction. This technique has been widely used (Danckwerts (1970), Sharma et al (1970), Ganguli et al ( , 1980, Robinson et al (1970Robinson et al ( , 1974, Mehta et al (1971), Westerpterp et al (1963), Yoshida et al (1960).  (Calderbank (1958), Vermuelen et a 1 ( 1 9 5 5 ) , Lee and Mey r i ck ( 1 9 7 0 ) ) • When add i n g so 1 ids , this method is difficult when measuring interfacial area in the tank. However, if the bubbles can be removed from the vessel through a glass capillary and analyzed by a light transmission source like a laser, it may be a practical means of studying local distribution of interfacial area.
Indeed, this method was investigated further. Kawecki et al (1967) and Reith and Beek (1970) have had success in removing bubbles in an air-water system. The method appears to be quite practical for removing bubbles in gas-liquid systems and also gas-liquid-solid. Also, it .

I-'
Vl 16 the radial distribution of the bubble disappear and the gas rises di re ct ly through the impeller to the 1 iqu id surf ace.
In practice for large uni ts, normally the agitated vessels are op er a ted near the f loading reg ion. In our research, studies were carried out near the flooding region and the power of agitation in the aerated medium was calculated based on this work by Warmoeskerken (see Figure 2,3).
This summary has presented essential information from the literature to understand the behavior of gas-liquid agitated reactors. This information includes bubble diameter, gas holdup, interfacial area, and the gassed power.
To extend this knowledge to a gas-liquid-solid system is difficult and complex. Without preliminary experimentation, it is almost impossible to determine the interfacial area accurately and centers on one of the main objectives of our study. Other required apparatus is the following: (1) a He-Ne laser, (2) neutral density filters, (3) a light sensitive photodiode, (4) a storage oscilloscope, (5) electrical power supply and associated circuitry, (6) an air rotameter with air filter and pressure gauge, (7) a glass sample probe to traverse the tank, separator and collector, (9) a vacuum pump, and associated manual control valves, block valves, supports, filters and plastic tubing.   When a gas bubble passes in front of the photodiode through the glass tube, a signal from the photocell is amplified and sent through the associated circuitry (see Figure 7)  Common flow patterns exist with increasing impeller speed (see Figure 8). It is generally accepted to feed the gas beneath the impeller because it encourages the gas to pass outwards through the high shear region, thereby improving the change of gas dispersion.
In ( Samples of a gas-liquid or a gas-slurry system are taken using the combination suction method/laser technique.  The effect of increasing solid concentration of 2~ m glass particles on local gas holdup and bubble diameter is shown in Tables 1-6. In Table 1        To evaluate the overall changes in holdup and bubble diameter, more data is required at the lower levels.   It may be speculative to say at this point possibly the large particles such as 7Y m size occupy the space in the 1 iqu id phase more easily than they do in the 1 iqu id film surrounding the gas bubbles. Therefore, there is little particle interaction with the gas bubbles and relatively no change in local gas holdup or bubble size.   Calderbank (1958). The absolute error of the actual bubble diameter versus the predicted bubble diameter is -9.3%.
Further correlation of the experimental data by a linear regression will show improved functionalities of 0 and PG. The improved coefficients were due to only including data taken on locations 7-18 which will be the coalescing region. Shinnar (1961) Rande and Ulbrecht (1978) found that the gas holdup decreased as the polyacrylamide concentration increased possibly because of a growth of elasticity of the gas-liquid bubble interface and therefore inhibited further divisions of the bubbles. For this to be true, no further coalescence between bubbles wi 11 occur and you would expect that the data would need to be correlated in two groups (one with solids and one without solids). However, results by Nagaraj ( 198 4) and this au th or indicate that the presence of so lids leads to more coalescence and can be correlated as one group. This will be discussed later.
Results similar to those found here for the influence of suspended solids were reported by other investigators such as Lee et al (1982), Ching (1983), and Kato et al (1973).
In the investigation by Lee et al (1982), glass particles of sizes 41-10~ m and wetted ~ m polyacrylonitrile particles at volume fractions 0-0.5 were evaluated for its effect on gas-liquid mass transfer. In one series of experiments using 5~m glass ballotini, holdup decreased by more than 10% with increasing solids content. Furthermore, Lee studied Orlon particles ')4 m range) and results showed substantial decreases in gas holdup and interfacial area (>12% 0, >50% a). the results are interpreted in terms of the particles obstructing the diffusion path and damping the turbulence.
Ching (1983)  However, they were vague as to why a decrease and then increase in interfacial area.
In a stirred tank at a specific turbulent intensity, bubble breakup and coalescence are in equilibrium with one another and wi 11 usually determine the mean bubble size.
Furthermore, when so 1 ids are present, the turbulent in tensity in the system is affected, and hence the size and behavior of the gas bubbles. Nagaraj (1984) shows that the presence of suspended glass particles actually dampens the high level of turbulent intensity to a certain extent and leads to more coalescence between bubbles.
Our results tend to agree. There apparently is an increase in the number of coalesces occurring.
Nagaraj mentioned that coalescence may occur in one of the following ways: (1) causing rupture of the gas-liquid film (2) solids adhering to the gas-liquid bubble interface may cause adhesive forces between bubbles (similar to van der Waal forces).
(3) According to Kirkpatrick and Lockett (1974), the bubbles will not coalesce if the approach velocity is greater than a certain critical velocity.
Solids may (a) increase this critical velocity (most probable) or (b) decrease bubble approach velocity by increasing the drag.
In correlating the results, we have treated the three phase system as a two phase system and per formed a 1 i near 50 regression.
Our results indicated the following: This empirical result was in good agreement with already well known correlations by Shinnar (1961) and Calderbank (1958).
According to Shinnar, if a dispersion remains in a quasi-stationary flow field for a sufficient duration, a dynamical equilibrium between coalescence and breakup is established.
In the breakup region, the maximum diameter is estimated to be a function of the agitator speed: It is the belief of this author and others that show the majority of the collisions end up in coalescence. Howarth's (1964) results show that almost every collision 51 resulted in coalescence. Nagaraj says that better than 50% of the collisions coalesce if the approach velocity is less than the critical velocity.
Shinnar predicted the coalescence of droplets. He found that the forces of adhesion and those of inertia are different functions of droplet diameter. Hence, for very small droplets, the turbulent energy input in the impeller region may be insufficient to overcome the adhesion energy however and thus results in coalescence.
The energy of adhesion, E , and the energy required to a separate two droplets of unit diameter and separated by a minimum distance, ho, is related by The inertia 1 forces of two droplets relative to each other a r e proportional to p u 2 (d) d 3 . This must be larger than the energy of adhesion in order for coalescence not to occur. -~ Coalescence region: DB o<.. N ~ cY-As discussed before, we suspect that our system behaves similar to a coalesing one and our results point this out. Near the impeller region (break up dominates) gas bubbles will be subjected to a region of high shear and will result in breakup. Correlating of our results in this region proved unsatisfactorily. Okamoto et al (1981)  He also showed that some coalescence does take place within the impeller discharge stream. The coalescence efficiency is dependent upon the size of the bubble and the fluctuating velocity of the energy dissipation eddies.
Generally speaking, the larger the bubble and the greater the velocity of approach of the bubbles, the greater the probability of coalescence. For a given bubble size, however, there is a maximum velocity of approach for which 54 bubbles will deform to such an extent that the bubbles will bounce off of each other rather than coalesce.
The role of the solids appears to be to prevent this rebounding effect.
If the solid sizes are within the size range of the energy dissipation eddies, these solids tend to get caught up in the wake of the moving bubbles. When the bubbles attempt to rebound from a collision, these particles resist this rebounding effect (inertia of the bubbles attempt to reverse direction) and promote coalescence.
Larger particles (7rm, 200m) tend to move more independently from the fluid and bubble wake flow patterns and appear to play little to no role in the coalescence process.
As we move away from the impeller region, the energy dissipation eddies are larger with a slower velocity. This allows the bubbles to approach each other at a velocity at which the bubbles do not deform to the extend which prevents coalescence.
The fluid between these bubbles is now allowed to drain from between the bubbles and coalescence is more -4 fleet a difference in the gassed power exponent, PG · • We suspect that one may expect an exponent between -.25 and -.4 for the overall bubble size.
For a summary of the literature comparisons, see Table   10. Possibly this is because the large particles tend to move more independently from the fluid or bubble wake nor do the particles adsorb onto the gas-1 iquid interface and therefore appear not to be playing any role in coalescence.