Pressure effects on the hydrodynamic behavior of gas-liquid-solid

---

2322

Ind. Eng. Chem. Res. 1992, 31, 2322-2327

Pressure Effects on the Hydrodynamic Behavior of Gas-Liquid-Solid Fluidized Beds Peijun Jiang, David Arters, and Liang-Shih Fan* Department of Chemical Engineering, The Ohio State University, Columbus, Ohio 43210

Basic hydrodynamic behavior of a gas-liquid-solid fluidized bed at elevated pressures up to 1MPa was studied by measurement and visualization techniques in a specially constructed rectangular column. Conditions of gas and liquid velocities, particle size, distributor type, and pressure were varied in the apparatus, which was designed to allow operation in both two- and three-dimensional modes. Viualization in two-dimensional operation and measurement in three-dimensional operation were combined to provide a unique tie point between the two modes. The study indicated that bubble size decreases, bubble size distribution narrows, and gas holdup increases with increasing pressure. The type of distributor was found to have a marked impact on bed behavior, including bulk circulation, bubble coalescence, and gas holdup. Possible mechanisms for the observed pressure effects on the hydrodynamic behavior are also discussed, combined with available experimental observations. Analysis indicates that initial bubble size and maximum stable bubble size should decrease with increasing pressure due to changes in interfacial physical properties as well as in gas density.

Introduction Interest in the behavior of three-phase fluidized beds operated under high pressure has increased in recent years as the number of practical applications has increased (e.g., H-Oil and LC-Fining resid hydrotreating units). Knowledge of the fluidization behavior at high pressure is critical in predicting the bed performance as a chemical reactor. A complete understanding of pressurized three-phase fluidized beds requires knowledge of the effect of pressure on fundamental fluidization hydrodynamic properties. However, only a few publications have explored the nature of pressure effects on the basic hydrodynamic behavior. It is well-known that increased pressure will result in an increase in gas holdup in a particular fluidization region. Of fundamental importance is the extent to which these pressure effects occur and the reasons for them. Studies of the effects of pressure on the hydrodynamics of three-phase fluidized beds have not been comprehensive, and to date have not included direct visualization. A number of researchers (e.g., Deckwer et al., 1980; Tarmy et al., 1984) have found the measured gas holdup at pressure to deviate from the value predicted by any of the many correlations obtained in ail-water systems operated at ambient conditions. Tarmy et al. (1984) investigated the hydrodynamic characteristics of three-phase fluidized beds operated under pressures up to 0.621 MPa. Their results showed a 2-fold increase in gas holdup. Kurten (1982) reported that the gas holdup is in the range of 0.37-0.66 for gas superficial velocities of 6.9-13.9 cm/s at liquefaction conditions (480 "C and 31 MPa). Chiba et al. (1989) indicated that the gas holdup increases with pressure up to a critical pressure, beyond which there is no significant pressure effect on the gas holdup. Gas holdups were reported by Blum and Toman (1977)as high as 0.50at gas velocities of about 15 cm/s for a three-phase methanator operated at 350 "C and 6.8 MPa. Note that their unit was operated under a foaming-like flow regime. Contrary to these data were measurements of gas holdup carried out by Deckwer et al. (1980) in an A1203/ paraffin/N, slurry column with porous plate distributor (75 pm) operated under high pressure. They concluded that the gas holdup in the dispersed bubble regime is not influenced by the system pressure over the range of 0.4-1.1 MPa when the temperature is above 250 "C and solids

* To whom correspondence should be addressed.

holdup is in the range of 5-16 w t %. This is in good agreement with Kolbel et al. (1961), who studied the pressure effect in an air-water system with porous plate distributor (50-60 pm) at pressures up to 1.6 MPa. Kolbel et al. also indicated that the bubble size is independent of pressure. Likewise, no marked pressure effect on the gas holdup was found in the results of Clark (1990) using a sintered metal plate (60pm) distributor for low hydrogen velocities, obtained in their glass bead/H2 (and N2)/ methanol system. Note that porous plate or sintered plate distributors were used in each of these three latter studies. It is likely that the use of porous plate distributors is the reason for their findings of the independence of the gas holdup from pressure at low gas velocities. This is also confiied by Idogawa et al. (19861, who found the bubble size to be almost independent of operating pressure in the air-water system with porous plate distributor. Idogawa et al. (1986,1987b)investigated the effects of the properties of gas and liquid phases and operating pressure on hydrodynamics in gas-liquid bubble column systems. They found an increase in gas holdup with increasing pressure up to 15 MPa and a decrease in mean bubble size. They noted that the effects of pressure on gas holdup are more profound at higher gas velocity, in good agreement with Tarmy's (1984) results. A more profound effect of pressure on gas holdup at higher gas velocity was also reported by De Bruijn et al. (1988),who investigated the effects on the gas holdup of pressure over the range 5.57-11.03 MPa in a bubble column at a constant temperature of 300 "C. The mechanisms of pressure effect on the fluidization behavior are uncertain. Tarmy et al. (1984) attributed the pressure effect to the small bubble size formed due to the increased contribution of gas momentum to the bubble formation process. This implies that an increase in system pressure operates in much the same way as an increase in gas density. However, separate roles of gas density and system pressure were found by Saberian-Broudjenniet al. (1987) in a three-phase fluidized bed and Clark (1990) in a slurry bubble column. Saberian-Broudjenniet al. (1987) used He, N2,and C02,which have significantly different molecular weights, as the gas phase in a three-phase fluidized bed with a packed bed distributor of 4-mm nickel beads 8 cm in height. Their results demonstrated that the minimum velocity for fluidization is influenced to only a very small extent by the properties of the gas and the effect of gas properties on gas holdup is insignificant. In comparing the gas holdup of H2 and N2 at low gas velocities,

OSSS-5SS~/92/2631-2322$03.00/0 0 1992 American Chemical Society

Ind. Eng. Chem. Res., Vol. 31, No. 10, 1992 2323 Clark (1990) found that the holdup of N2,even with a 14-fold increase in gas density, was lower than that of H2 On the basis of these results Clark suggested that gas holdup is influenced by some pressure-dependent properties other than gas density. Results reported by Kuo et al. (1985)in gas-liquid bubble columns of product wax also showed that the gas type (N,and H,)does not have an effect on gas holdup when the bed is operated in the dispersed bubble regime. It is interesting to point out that any substance which stabilizes the bubbleliquid interface suppresses the coalescence. This stabilization is generally attributed to surface tension gradients around the bubbles which resist the interfacial motion and thus retard the rate of drainage of liquid film between two bubbles. Slowinski et al. (1957) and Massoudi and King (1974) studied the pressure effeds on the surface tension between the gas and liquid. Their results showed that the surface tension decreases approximately linearly with increasing gas pressure. It has been found that bubble coalescence is reduced with reduced surface tension (Calderbank, 1958; Kim and Lee, 1987; Fan and Tsuchiya, 1990). It was also found that the bubble coalescence rate is only significantly affected when the pressure exceeds a critical value, which depends on the gas/liquid interaction, e.g., 2.7 MPa for the nitrogen-water system and 2.0 MPa for carbon dioxidewater (Sagert and Quinn, 1976,1977, and 1978). The extent of reduction of the surface tension with pressure due to the type of gas varies: gases with a high extent of reduction yield low critical pressures. Thus another possible reason for the smaller bubble size under higher pressure is due to the pressure effects on the interfacial properties, which in turn affect liquid film thinning and rupture processes and bubble instability. Wilkinson and Dierendonck (1990) suggested that the increase in gas holdup is due to the decrease in maximum stable wavelength, which results in a smaller maximum stable bubble size in bubble columns. The objectives of the present study are to examine the bubble behavior in three-phase fluidized beds operated under high system pressures, to elucidate the system pressure effects on the gas holdup, and to study the system pressure effects on the performance of the gas/liquid distributor. To accomplish this, measurements in a three-dimensional system are tied to direct visualization of bubble behavior in the corresponding two-dimensional system through videotaping of events in a specially constructed three-phase fluidized bed apparatus.

Experimental Section The experimental apparatus for the three-phase fluidized bed system is shown schematically in Figure 1. A rectangular carbon steel column of 0.25- X 0.05-m cross section and 1.0-m height with a transparent section (0.25 X 1m) was used. To prevent the internal surface (made of Lexan)from abrasion by particles, two plate glass inserts were affixed inside the column symmetricallyagainst both transparent walls. The bed can be operated in either a purely two-dimensional mode (1 X 0.25 X 0.008 m) by inserting two pieces of thick glass (20 mm thick) against the transparent sides of the column or a pseudo threedimensional mode (1X 0.25 X 0.05 m) using thin inserts. The two-dimensionalsystem setup was used in this study for visualization of the bed phenomenon and the threedimensional system for collection of quantitative measurements such as holdups. There inevitably exist wall effects in a two-dimensional system which are exerted on the bubble behavior. Certain hydrodynamic behavior, however, such as pressure effects on flow patterns, bubble size, and gas holdup are maintained qualitatively the same

.

Pressure Controller

--

-Gas

to Vent

Pressurized Liquid

Drain Valve

Figure 1. Schematic diagram of two-dimensional pressurized three-phase fluidized bed.

in the two-dimensional setup as in the three-dimensional system. Ten pressure taps are flush mounted along the side wall of the column. A bubble-cap type distributor, a perforated plate with 1-mm hole diameter and 8-mm pitch, and a 75-pm mean diameter porous plate were all used as distributors. The bubble-cap distributor consists of four cylindrical bubble caps which are evenly spaced. Each cap is 30 mm in diameter with 12 rectangular flow holes (9 X 4 mm) evenly spaced on the vertical surface of the cylindrical cap. Tap water and air were employed, respectively, as liquid and gas phase. The particles were spherical glass beads of 2500 kg/m3 density and average diameters of 0.32, 0.46, 1, and 6 mm. System pressure in the bed was maintained at the desired value by a back-pressure control valve and a computerized controller located downstream of the bed. The gas and liquid flow rates were measured with a set of calibrated rotameters. Pressure taps connected to water manometers were used to measure the pressure gradient along the column. The gas holdup in sections between any two adjacent pressure taps is calculated from the pressure gradient along the bed on the assumption of a uniform solids holdup along the bed height. The average gas holdup is estimated using the weighted average of the gas holdup in the individual sections in the pseudo three-dimensional mode. A video camera with 1/1000-s shutter speed was employed to monitor the bubble behavior in the purely two-dimensional mode. The bed height was determined quantitatively as the intersection of the two pressure curves obtained for the bed and freeboard regions.

Results and Discussion Flow Pattern Observation. Flow visualization studies, such as undertaken in this study, can be instrumental in developing physical concepts or yielding qualitative information about the hydrodynamic behavior of three-phase fluidization under pressure. In this study, the flow patterns were first observed over a wide range of gas velocity and system pressure utilizing the video camera with the column in two-and three-dimensional modes. Typical flow patterns are graphically illustrated in Figure 2. At low gas velocity, the distributor-generated turbulence is a dominant factor in the liquid-solid emulsion phase and the bubbles ascend individually along linear paths (see Figure 2a). At higher gas velocity the most evident features observed are that the uniform bubble swarm begins to

2324 Ind. Eng. Chem. Res., Vol. 31, No. 10,1992

(a)

(b)

(C)

(d)

_____f

Increase Gas Velocity

Figure 2. Graphical illustration of orpieal flow pattern m two-dimemional three-phase fluidized beds. (Reconstructed from video pictures.)

meander, especially under high-pressure conditions, indicating the occurrence of local emulsion-phasecirculation near the side wall (see Figure 2b). With increases in gas velocity, bubbles transfer a significant amount of kinetic energy to the emulsion phase. The gross and local circulations of the emulsion phase develop randomly in space and time. The bubble cluster appears in the bulk flow and follows the circulation. Coalescence and breakup may occur in the bubble cluster and the system transits into the c o a l d bubble regime. With further increases in the gas velocity, gross circulation currents sweep the entire column. The downward motion of the emulsion phase near the side wall affecta steady bubble formation at the distributor and suppresses the bubble rise in the near side wall region, but continuous sparging of gas with a high jetting velocity makes large enough bubbles to counteract, and damp, the gross circulation. Meanwhile, another circulation commences to form on the other side (see Figure 2c). Such a flow creates locally and temporally different flow structures throughout the column, and strong circulation patterns cany toward the bottom of the column small and medium-sized bubbles which otherwise would have disengaged. At still higher gas velocity, the dynamic interaction between the gross circulation and bubbles or gas jet becomes frequent and violent. A symmetrical dual-loop circulation upward in the center and downward in the side wall region develops (see Figure 2d). The flow regime transition with gas velocity follows the same trend regardless of system pressure. However, the transition velocity between regimes varies with the system pressure. In the bed of 0.46-mm glass heads, for example, the transition gas velocity from region a to region b is O.O35m/s under atmospheric pressure while the transition velocity is delayed to 0.045 m/s when the bed is operated at a pressure of 0.7 MPa. The results obtained by Tarmy et al. (1984)have been confirmed by the present work. Their drift flux analysis indicated that the transition velocity from the dispersed bubble regime to coalesced bubble regime increases with system pressure. The increased in pressure leads to a tendency toward uniform bubble size distributions, which chacterizea the dispersed bubble regime. Pressure Effects on the Bubble Behavior. The bubble behavior in the two-dimensional bed was visually observed over a wide range of operating conditions. Photographs of typical conditions are shown in Figure 3. Investigation of such photographs showed that the bubble size is much larger under atmospheric pressure than that at 1 MPa, especially at high gas velocity. At low system pressure, very large bubbles were frequently observed which pass through the bed very quickly. These large bubbles disappear with higher pressure, as shown in the

(S)

p=0.1 Mn

(a) P=I.oMPa

F%mre 3. Hydmdynamie behavior of bubble m the two-dimensional three-phase fluidized bed operating under different pressures (d, = 0.32 mm, H. = 48 cm,U, = 0.5 cm/s, Us = 5 m/s,perforated plate). -0.40

E

P=1.0 MPa P=O.1 MPa

-0.30

a o'ooO

5

10

15

20

25

30

Bubble diameter (mm) Figure 4. Comparison of bubble size distribution in beds operated under different prwures.

figure. Note that all pidures obtained at these conditions are similar in nature to those shown in this figure. It also can be seen that the bubble size distribution is much narrower at the higher pressure condition. A comparison of bubble size distributions between the beds operated under pressure of 0.1 MPa and 1 MPa is given in Figure 4, obtained from averages of freeze-frame images from the video camera. At high pressure, it was found that the bubbles do not readily coalesce. Coalescence of two hubbles occurs in three steps. First, bubbles collide, trapping a small amount of liquid between them. This liquid then drains until the liquid f h separating the bubbles reaches a critical thickness. At this point, film rupture occurs through an instability mechanism, resulting in coalescence (Sagert and Quinn, 1976). From the first step and also from a macroscopic point of view, it is seen that the bubble coalescence rate is intimately connected to the collision rate. It is clear that colliisions may result from the random motion of bubbles due to nonuniformity of the local flow field. Or, bubbles of different sizes have different rising velocities which may lead to collision. Finally, a trailing bubble accelerates on entering the wake of a leading bubble and overtakes the leading bubble, coalescing through wake capture (Fan and Tsuchiya, 1990). A t a given gas and liquid flow rate, higher pressure yields smaller, more uniform bubble size and higher gas holdup. Wake size and intensity decreases with decreasing bubble size, trapping fewer trailing bubbles; the difference in bubble velocities decreases with increased size uniformity; and nonuniformity of the local flow field decreases as bubble size and wake intensity decreases, leading to fewer random bubble motions. Thus each of the mechanisms for bubble collision will be reduced due to the smaller bubble size. The experimental data seem to corroborate that bubble size is

Ind. Eng. Chem. Res., Vol. 31, No. 10, 1992 2328

-

04

I

W

-

I

z=O.l-0.2 m - m z=O.6-0.7 m

03 -

0

I

b

I

P=0.7 MPa Perforated plate Ul=1.5 cm/s -

0 -0

I

P=O.1 MPa P=O.i' MPa

I

0

1

5 0.20

---*a

0

& v) 0.10

1

1

a

c

cd I

Perforated plate U1=0.5 cm/s d,=0.46 mm $=48 c q

I

2326 Ind. Eng. Chem. Res., Vol. 31, No. 10, 1992 I

I

0

1

I

P=O.1 MPa

0

c

/-

a ? a 0.20

.’ /

I

Perforated plate U1=5.04 cm/s 0 00

1

4

1

8

12

16

Gas velocity (cm/s) Figure 8. Effect of gas velocity on gas holdup in a three-phase fluidized bed operating under different pressures using coarse particles.

size is about 4 mm for the air-water system with a 1mm diameter hole perforated plate at atmospheric conditions. Pressure effects on the initial bubble size are negligible over the range of the present study. However, as gas velocity increases, bubbles of smaller size and a higher formation frequency are formed in the higher pressure system, which results in a higher gas holdup compared to those in the system operated under atmospheric pressure. A similar observation also can be found in some of the results reported by Tarmy et al. (19841, who investigated the hydrodynamic behavior in a liquefaction reactor. It appears that at low gas velocities the incremental increase in gas holdup with system pressure is lower compared to that at higher gas velocities. The trend of the pressure effect on the gas holdup observed in the present studies is also consistent with the results obtained in bubble columns (De Bruijn et al., 1988; Wilkinson and Dierendonck, 1990). The explanation can be best obtained by combining results with the bubble behavior observed in the bed. In the dispersed bubble regime, uniform bubbles of ellipsoidal shape were observed throughout the column. The bubble size under this operating condition is about 4-43 mm. As discussed above, the transition from dispersed bubble regime to coalesced bubble regime occurs at a higher gas velocity at higher system pressure. This implies that bubbles start to coalesce at higher gas velocity in the higher pressure system and, consequently, the gas holdup is higher under the same superficial gas velocity. Moreover, a strong gross circulation, and therefore pronounced coalescence, were observed in the coalesced bubble regime for the atmospheric pressure system, which results in a decrease in gas holdup. The change of bubble rise velocity due to increased gas density is a negligible factor in the present studies. The insignificant pressure effect on the gas holdup in the low gas velocity region seems to suggest that pressure does not exert a strong effect on small, noncoalescing bubbles. The similar trend of pressure effect on gas holdup in the bed of 1-mm glass beads was also observed in the present study. Pressure effects on the gas holdup in a bed of 6-mm g h beads is shown in Figure 8. It appears that the bubble dynamic pattern for systems of coarse particles is different from that for the system involving smaller particles. As can be seen, no significant system pressure effects can be found over the range of operating gas velocity. Larger bubblea formed at the distributor are disintegrated by their axial mixing with the large particles and dispersed uniformly throughout the bed, as observed by video camera. The presence of coarse particles enhances the bubble breakup, especially at low bed expansions, due to the large inertia of coarse particles available to penetrate the roof of a bubble (Fan, 1989; Chen and Fan, 1989). Similar

results also can be found in the work of Ostergaard (1969) and Lee et al. (1974). Note that the type of distributor does not exhibit any significant effect on the gas holdup in the bed of 6-mm glass beads. The interaction between the coarse particles and bubbles dominates the bubble size and rising velocity in the bed of coarse particles. Pressure Effects on the Performance of Dietributors. When fluidized systems are operated in the dispersed regime, there is insignificant interaction between the bubbles due to the small bubble size and low turbulent intensity. The condition for the appearance of a pressure effect on the gas holdup seems to be concerned with the process of bubble formation. An investigation of various possible causes for the reduction of bubble size under the high-pressure condition demonstrated that the initial bubble size does play a part. The bubble formation procees becomes even more important when the system is operated at elevated pressure. Initial bubble size has been recognized as a strong function of geometry or design of distributor (Idogawa et al., 1987a). In order to study this effect, three distributors were used in the present study; a porous plate with a mean pore diameter of 75 pm, perforated plate with hole diameter of 1mm, and bubble caps. In the case of the perforated plate, it seems that the initial bubble size is almost independent of gas velocity, though smaller bubbles with a higher formation frequency were observed under the high-pressure condition. At low pressure, bubbbles coalesce immediately after they detach from the distributor and the coalescence rate decreases along the bed height. However, with increasing pressure the coalescence tendency decreases rapidly, especially in the upper zone. This implies a suppression effect by pressure on the coalescence rate, possibly due to the reduced turbulent intensity or changes in the interface properties and decrease in surface tension. It was also noted that the perforated plate more evenly distributes the gas and liquid under the high-pressure condition, thereby reducing recirculation. A very different phenomenon was observed in the bed with bubble-cap type distributor. Such a distributor creates large bubbles and very strong global circulation patterns which carry small and medium-sized bubbles, which otherwise would disengage, toward the bottom of the column. Note that, despite this entrainment, the downward circulation areas had very low gas holdups, resulting in the low values obtained for the bed as a whole. Coalescence does play some part, but only seems to occur near the top of the jet, and has little influence on the average bubble size. Besides, this was not attributable to any phenomenon occurring in the single bubble case, but was due to effects arising from the interaction between bubbles and the bubble-turbulent flow pattern. In this case the gross circulation patterns caused by the bubble caps overshadowed the pressure effects on the flow pattern. However, an increased gas holdup, as shown in Figure 9, was obtained in the high-pressure condition, indicating the suppression of coalescence even at the distributor but increased entrainment of the smaller bubbles in the gross (downward) circulation. In the porous plate system, uniform gas and liquid distribution and the formation of smaller bubbles were obtained due to the high distributor resistance. Very fine bubbles are maintained at almost the same size along the axis of the bed through the range of low gas velocity. No significant pressure effect on the bubble size and gas holdup can be observed at low gas velocity. A t high gas velocity, 3-4-mm bubbles in bubble clusters coalesce to form larger bubbles which rise in the center region with

Ind. Eng. Chem. Res., Vol. 31, No. 10, 1992 2327 0.30 n

I W

Bubble-cap U1=0.9 cm/s

$0.20

20 A

0

P=O.1 MPa P=0.7 MPa

0 10

v)

(d

r k

1

0.00

4

I

a

12

Gas velocity (cm/s) Figure 9. Effect of gas velocity on gas holdup in a three-phase fluidized bed with bubble-cap type distributor operating under different pressures.

higher velocity. This was observed in all operating pressure conditions, but the tendency to coalescence was much stronger at low pressures, and the extent of pressure effects are small compared with results obtained in the perforated plate system. In other words, an influence of pressure was observed in the coalesced bubble regime in the bed with the porous plate distributor when the system pressure was changed from atmospheric condition to 0.7 MPa. The another notable feature is that the bed can be operated in a dispersed regime over a wide range of gas velocity at high pressure. This qualitative feature highlights again the probable effects of the initial bubble size.

Concluding Remarks Effects of operating pressure on the hydrodynamic behavior of gas-liquid-solid fluidized beds were studied by means of direct visualization in two-dimensional systems and measurement in three-dimensional systems. It was found that the transition between the dispersed bubble regime and coalesced bubble regime occurs at a higher gas velocity under the higher pressure condition. Visual observation revealed a smaller bubble size and a narrower bubble size distribution in the high-pressure system. It was also observed that the gas holdup increases as pressure increases in three-dimensional systems. The distributor characteristics can play a critical role in determining the gas holdup, bubble coalescence, and bulk circulation. The experimental results also indicated that pressure affects the hydrodynamic behavior through the initial bubble size and the interaction between the bubble and turbulence. Acknowledgment The support of The National Science Foundation under Grants CTS-8905463 and CTS-9200793 for this work is gratefully acknowledged.

Nomenclature db = bubble size (mm) d = particle diameter (mm) If = bed expansion height (cm) H,= static bed height (cm) P = operating pressure (MPa) U = superficialgas velocity (cm/s) L( = superficial liquid velocity (cm/s) z = bed height (m) Literature Cited Blum, D.; Toman, J. J. Three-phase fluidization in a liquid phase methanator. In Fluidization Theories and Application; AZChE

Symp. Ser. 1977, 73 (No. 161),115-120. Calderbank, P. H. Physical rate processes in industrial fermentation. Part 1. Trans. Znst. Chem. Eng. 1958,36,443-463. Chen, Y.-M.; Fan, L.-S. Bubble breakage mechanisms due to collision with a particle in liquid medium. Chem. Eng. Sci. 1989, 44, 117-132. Chiba, S.; Idogawa, K.; Maekawa, Y.;Moritomi, H.; Kato, N.; Chiba, T. Neutron radiographic observation of high pressure three-phase fluidization. In Fluidization VI; Grace, J. R., Shemilt, L. W., Bergougnou, M. A., Eds.; Engineering Foundation: New York, 1989;pp 523-530. Clark, K. N. The effect of high pressure and temperature on phase distributions in a bubble column. Chem. Eng. Sci. 1990, 45, 2301-2307. De Bruijn, T. J.; Chase, W. J. D.; Dawson, W. H. Gas holdup in a two-phase vertical tubular reactor at high pressure. Can. J. Chem. Eng. 1988,66,330-333. Deckwer, W.-D.; Louisi, Y.; Zaida, A.; Ralek, M. Hydrodynamic properties of the Fischer-Tropsch slurry process. Znd. Eng. Chem. Process Des. Deu. 1980,19,699-708. Fan, L A . Hydrodynamics of cocurrent upward fluidized bed systems. In Gas-liquid-solidfluidization engineering; Butterworths: Stoneham, MA, 1989;Chapter 2. Fan, LA.;Tsuchiya, K. Single bubble rise Characteristics. In Bubble wake dynamics in liquids and liquid-solid suspensions; Butterworths: Stoneham, MA, 1990; Chapter 2. Idogawa, K.; Ikeda, K.; Fukuda, T.; Morooka, S. Behavior of bubbles of air-water system in a column under high pressure. Znt. Chem. Eng. 1986,26,468-474. Idogawa, K.; Ikeda, K.; Fukuda, T. Formation and flow of gas bubbles in a pressurized bubble column with a single orifice or nozzle gas distributor. Chem. Eng. Commun. 1987a,59,202-212. Idogawa, K.; Ikeda, K.; Fukuda, T.; Morooka, S. Effects of gas and liquid properties on the behaviors of bubbles in a column under high pressure. Znt. Chem. Eng. 1987b,27,93-99. Kim, J. W.; Lee, W. K. Coalescence behavior of two bubble in stagnant liquids. J. Chem. Eng. Jpn. 1987,20,448-453. Kolbel, H.; Borchers, E.;Langemann, H. Griiasenverteilung der gasblasen in blasendulen. Chem. Zng. Technol. 1961,33,668-675. Kuo, J. C. W. “Two-stage process for conversion of synthesis gas to high quality transportation fuels”. Final report to the U.S.Department of Energy for Contract DE-AC22-83PC60019,Mobil Research and Development Co., Paulsboro, NJ, 1985. Kiirten, H. Verfahrenstechnik der Kohlehydrierung in Sumpfphasen-Reaktoren. Chem. Zng. Tech. 1982,54,409-416. Lee, J. C.; Sherrad, A. J.; Buckley, P. S. Optimum particle size in three phase fluidized bed reactors. In Fluidization and its Application; Angelino, H., Ed.; Cepadues-Editions: Toulouse, 1974; pp 407-416. Massoudi, R.; King, A. D. Effect of pressure on the surface tension of water: Adsorption of low molecular weight gases on water at 25 OC. J. Phys. Chem. 1974, 78, 2262-2266. Ostergaard, K. Studies of gas-liquid fluidization; Danish Technical Press: Copenhagen, 1969. Saberian-Broudjenni, M.; Wild, G.; Charpentier, J.-C.; Fortin, Y.; Euzen, J.-P.; Patoux, R. Contribution to the hydrodynamic study of gas-liquid-solid fluidized bed reactors. Znt. Chem. Eng. 1987, 27,423-440. Sagert, N. H.; Quinn, M. J. The coalescence of Ha and COz bubbles in water. Can. J. Chem. Eng. 1976,54,392-398. Sagert, N. H.; Quinn, M. J. Influence of high pressure gases on the stability of thin aqueous films.J. Colloid Interface Sci. 1977,61, 279-286. Sagert, N. H.; Quinn, M. J. Surface viscosities a t high pressure gas-liquid interfaces. J. Colloid Interface Sci. 1978,65,415-422. Slowinski, E. R.;Gates, E. E.; Waring, C. E. The effect of pressure on the surface tensions of liquids. J. Phys. Chem. 1957, 61, 808-810. Tarmy, B.; Chang, M.; Coulaloglou, C.; Ponzi, P. Hydrodynamic characteristics of three phase reactors. Chem. Eng. 1984, Oct, 18-23. Wilkinson, P. M.; Dierendonck, L. L. V. Pressure and density effects on bubble break-up and gas holdup in bubble columns. Chem. Eng. Sci. 1990,45,2309-2315. Receiued for review November 18,1991 Revised manuscript receiued July 15,1992 Accepted July 21, 1992