Mass bubble coalescence

The results of Massimilla et al., 0stergaard, and Adlington and Thompson are in substantial agreement on the fact that gas-liquid fluidized beds are characterized by higher rates of bubble coalescence and, as a consequence, lower gas-liquid interfacial areas than those observed in equivalent gas-liquid systems with no solid particles present. This supports the observations of gas absorption rate by Massimilla et al. It may be assumed that the absorption rate depends upon the interfacial area, the gas residence-time, and a mass-transfer coefficient. The last of these factors is probably higher in a gas-liquid fluidized bed because the bubble Reynolds number is higher, but the interfacial area is lower and the gas residence-time is also lower, as will be further discussed in Section V,E,3. 

Vaux (1978), Ulerich et al. (1980) and Vaux and Schruben (1983) proposed a mechanical model of bubble-induced attrition based on the kinetic energy of particles agitated by the bubble motion. Since the bubble velocity increases with bed height due to bubble coalescence, the collision force between particles increases with bed height as well. The authors conclude that the rate of bubble-induced attrition, Rbub, is then proportional to the product of excess gas velocity and bed mass or bed height, respectively,... 

Substituted phenols as well as phenol itself are typical constituents of (bio-)refractory waste waters and can increase a(0> 3 (Gurol and Nekouinaini, 1985). They studied the influence of these compounds in oxygen transfer measurements and attributed this effect to the hindrance of bubble coalescence in bubble swarms, which increases the interfacial area a. When evaluating the effect of these phenols on the ozone mass transfer rate, it is important to note that these substances react fast with ozone (direct reaction, kD= 1.3 103 L mol"1 s 1, pH = 6-8, T = 20 °C, Hoigne and Bader, 1983 b). 

An interesting class of exact self-similar solutions (H2) can be deduced for the case where the newly formed phase density is a function of temperature only. The method involves a transformation to Lagrangian coordinates, based upon the principle of conservation of mass within the new phase. A similarity variable akin to that employed by Zener (Z2) is then introduced which immobilizes the moving boundary in the transformed space. A particular case which has been studied in detail is that of a column of liquid, initially at the saturation temperature T , in contact with a flat, horizontal plate whose temperature is suddenly increased to a large value, Tw T . Suppose that the density of nucleation sites is so great that individual bubbles coalesce immediately upon formation into a continuous vapor film of uniform thickness, which increases with time. Eventually the liquid-vapor interface becomes severely distorted, in part due to Taylor instability but the vapor film growth, before such effects become important, can be treated as a one-dimensional problem. This problem is closely related to reactor safety problems associated with fast power transients. The assumptions made are ... 

Overall there will be a concentration of the hydrophobic particles in the upper regions of the froth layer and a concentration of the hydrophilic particles in the lower regions. Figure 10.7 provides an illustration of how froth structure changes from the bottom to the top of the froth layer. At the very top of the froth bubble, coalescence and rupture will be occurring as well. The amount of ore that can be separated for any given amount of liquid is proportional to the surface area of the foam. It has been estimated that a foam with a specific surface area of 0.2 m2g-1 can separate a thousand times more ore, by mass, than the mass of its own liquid [629]. 

In contrast, Fig. 72 shows the results of mass transfer in the system aqueous 1-n sodium sulphite solution/air. These measurements were carried out under steady-state conditions in vessels with hollow stirrers on the scale p = 1 5 [58/1, 92]. In this material system, the high salt concentration (70 g/1) fully suppresses bubble coalescence. In the case of the self-aspirating hollow stirrer (see Fig. 28), the stirrer power and gas throughput were coupled via the stirrer speed and were therefore dependent on each other. Consequently, v does not occur explicitly in the representation in Fig. 72, because it is a function of (P/V)". 

This is verified by the measuring data obtained in a column of 1.6 m 0, in which a slot injector was installed with a bottom clearance of 1 m and an angle of 25° towards the bottom. The liquid head H above the injector was varied in the range H = 1-7 m. It is shown that the influence of the bubble coalescence in the G/L free jet on the mass transfer - which occurs in a short distance from the nozzle orifice -is equilized only after H = 3 m see Fig. 75. This finding proves that the pi-set, eq. (13.19), is not complete but has to be widened by a pi-number which essentially contains liquid height H. It can be formulated by H = H (g/v2)1 3. 

In a more recent work, Slesser et al.128 showed that small amounts of fine solid particles in an agitated three-phase slurry reaction can change the magnitude of the volumetric mass-transfer coefficient considerably. Chandrasekaran and Sharma14 reported a similar conclusion in the case of the oxidation of sodium sulfide in the presence of activated carbon. They argued that the presence of solids prevents bubble coalescence and thus increases the gas-liquid interfacial area. 

Joosten et al.51 explained the data based on the increase in the apparent viscosity of the slurry by the addition of solids. The volumetric mass-transfer coefficients, as a function of the relative viscosity of slurry obtained by them, are shown in Fig. 9-16. These data show that as the density of the solids decreased the value kLaL decreased faster with the increase in the relative viscosity. These data also show that for particle sizes < 250 pm, the suspended solid particles do not significantly affect the gas-liquid volumetric mass-transfer coefficient when the apparent viscosity of the slurry is not higher than four times that in the liquid. At high solids concentration, bubble coalescence and subsequent reduction in gas holdup can be the major cause in the reduction of fcLaL. The data show that kLoL in a three-phase slurry depends on the difference in density between the solids and liquids. The greater inertia of the heavier particles may create a stronger disturbance at the gas-liquid interface and thus affect the value of kL. 

Massimilla et al.79 found that the presence of solid particles reduced the gas-liquid mass-transfer rate in a bubble-column. This was explained as being due to the higher rate of bubble coalescence and. consequently, a lower gas liquid interfacial area obtained in the presence of solids. They also found that the absorption rate increased with an increase in the nominal liquid velocity and a decrease in particle size. 

The dependence of heat and mass transfer coefficientes on the scale factors dp/L and dp/D can also be rationalized in terms of gas bypassing through the bed in the form of bubbles. Since bubbles coalesce and grow as the rise from the distributor, a longer bed, big L/D values, will operate with larger bubbles in its upper part. This will lead to smaller values of the Sherwood and Nusselt numbers since the interchange coefficient between bubble and emulsion phase varies inversely with bubble diameter. 

The presence of vertical membrane elements or modules also helps to prevent bubble coalescence, thus favoring heat and mass transfer in the reactor. Their spacing should be sufficiently small so that the maximum number of the membrane tubes or modules may be provided in the reaction zone and large enough that no blockage or bridging of the fluidized bed occurs. 

All these flow types appear more or less in a series one after the other during the evaporation of a liquid in a vertical tube, as Fig. 4.30 illustrates. The structure of a non-adiabatic vapour-liquid flow normally differs from that of an adiabatic two-phase flow, even when the local flow parameters, like the mass flux, quality, etc. agree with each other. The cause of this are the deviations from thermodynamic equilibrium created by the radial temperature differences, as well as the deviations from hydrodynamic equilibrium. Processes that lead to a change in the flow pattern, such as bubbles coalescing, the dragging of liquid drops in fast flowing vapour, the collapse of drops, and the like, all take time. Therefore, the quicker the evaporation takes place, the further the flow is away from hydrodynamic equilibrium. This means that certain flow patterns are more pronounced in heated than in unheated tubes, and in contrast to this some may possibly not appear at all. 

As regards mass transfer in the G/L system, it can be concluded that Ac, the driving force of the process, is higher in an oil/water dispersion than with water alone. In addition it can be expected that depending upon the solubility of the organic phase in water (distribution coefficient ) the bubble coalescence will he suppressed (increase of a). 

In Eq. (9.47), (Pg/V) is in W/m, D, in m, and in mm/s. The effect of D, is somewhat uncertain, since it is based only on two sizes, and for a conservative scaleup, this factor could be ignored. Of course, conditions dose to flooding may not be optimum for mass transfer, since bubble coalescence in the regions away from the impeller could greatly reduce the surface area. 

---