Ultrasonic Characterization of Slurries in an Autoclave Reactor at

Brian B. Anderson, Ken E. Creasy, Mel Koch, James F. Tatera, Alex L. Robinson, Leonard Bond, Lloyd W. Burgess, Gary N. Bokerman, Alan H. Ullman, G...
0 downloads 0 Views 205KB Size
Download PDF
Ind. Eng. Chem. Res. 1996, 35, 1807-1812

1807

Ultrasonic Characterization of Slurries in an Autoclave Reactor at Elevated Temperature Y. Soong,* I. K. Gamwo, A. G. Blackwell, F. W. Harke, R. R. Schehl, and M. F. Zarochak U.S. Department of Energy, Pittsburgh Energy Technology Center, P.O. Box 10940, Pittsburgh, Pennsylvania 15236

An ultrasonic technique was developed to measure the concentration of solids in an autoclave reactor. Preliminary measurements were conducted on slurries consisting of molten paraffin wax, glass beads, and nitrogen bubbles at 189 °C. The data show that the velocity and attenuation of the sound are well-defined functions of the solid and gas concentrations in the molten wax. The results suggest possibilities for directly measuring solids concentration during operation of a three-phase slurry reactor. Introduction The application of the three-phase slurry reactor system for coal liquefaction processing and chemical industries has recently received considerable attention. To design and efficiently operate a three-phase slurry reactor, the degree of dispersion of the solid (catalyst) in the reactor must be understood and controlled. The solids distribution within the reactor greatly affects its performance. Because it is crucial to understand the influence of various reactions and reactor configurations on the solids concentration profile, measurement of solids concentration must be made under reaction conditions. Several “on-line” techniques exist to monitor the concentration of solids in a three-phase slurry. Most of these methods are optically based techniques which use reflected, scattered, or absorbed light to measure the concentration of solids. These methods are sensitive to the opacity of the slurry phase (Lee and de Lasa, 1987). The direct sampling of solid particles from a slurry reactor suffers from either a need to perturb the reaction system or a difficulty of measurement because of high pressure and temperature (Kato et al., 1972). The static pressure method also suffers some difficulties when the solid and liquid densities are close (Hwang and Fan, 1986). Ultrasonic measurement is an established technique for many industrial applications, for example, medical imaging, materials testing, flow detections, and level measurements (Anson and Chivers, 1989; Lynnworth, 1989; McClements, 1991). The ultrasonic technique has advantages over many existing methods because it is a noninvasive and nondestructive measurement in systems which are concentrated, optically opaque, and electrically nonconducting (McClements, 1991). However, the main disadvantage of this technique is that there are no commercially available ultrasonic instruments which are designed to characterize the concentration of solids. Therefore, researchers have to develop their instruments and data interpretation. The utilization of ultrasonic techniques for slurry characterizations has received considerable attention recently (Abts and Dahi, 1986; Behrman and Larson, 1987; Bonnet and Tavlarides, 1987; Foote, 1985; Hilgert and Hofmann, 1986; Karplus and Raptis, 1978). Some ultrasonic * To whom all correspondence should be addressed. Telephone: (412) 892-4925. FAX: (412) 892-4152. E-Mail: [email protected].

techniques are based on the principle of scattered acoustic pulses. The process is similar to sonar, in which bursts of acoustic energy are emitted into a liquid and reflections from discontinuities (solid and bubbles) are detected and measured. Abts and Dahi (1986) have applied the ultrasonic technique to determine the concentration of oil droplets in an oil recovery system by detecting the forward scattering of ultrasonic energy from oil droplets in the oil recovery system. Foote (1985) has reported that a particulate in a fluid can be identified by sending an ultrasonic pulse across the fluid, measuring the amplitude of the portion of the pulse scattered from a particulate at a preselected angle. The relative size of the particulate is determined from the magnitude of the scattered signal. Behrman and Larson (1987) have applied on-line ultrasonic particle monitoring to brewing operations. These ultrasonic techniques are based on the scattering of ultrasonic pulses. Hilgert and Hofmann (1986) have also applied the ultrasonic Doppler technique to characterize the local bubble rise velocity in a bubble column reactor. Bonnet and Tavlarides (1987) have applied an ultrasonic transmission technique to determine the dispersedphase holdup of liquid-liquid dispersions by measuring the velocity of ultrasound in suspensions and emulsions. Recently, a method involving the measurement of ultrasound transmission has been reported in a slurryphase stirred-tank reactor which offers the possibility of using the ultrasonic technique for the measurement of solids concentration in a three-phase slurry reactor (Okamura et al., 1989). The ultrasonic transmission uses measurements of the velocity and attenuation of the sound wave which travels directly through the slurry sample. When an acoustic wave strikes the boundary between two different media (liquid and solid) and the acoustic impedances of the two media are different, some acoustic energy will be reflected and some will be transmitted. The reflected wave travels back through the incident medium (liquid) at the same velocity. The transmitted wave continues to move through the new medium (solid) at the sound velocity of the new medium. When the velocity of sound in a liquid (1324 m/s at 25 °C for kerosene) is significantly different from that in a solid (5968 m/s at 25 °C for fused silica), a time shift (a velocity change) in the sound wave can be detected when solid particles are present relative to that for the pure liquid. The amplitude of the sound wave is also reduced when a solid particle is present since the wave is partially scattered and absorbed.

S0888-5885(95)00771-8 This article not subject to U.S. Copyright. Published 1996 by the American Chemical Society

1808

Ind. Eng. Chem. Res., Vol. 35, No. 6, 1996

Figure 1. Effect of solids on the ultrasonic signal.

Therefore, a change in amplitude of the sound wave can also be detected when solid particles are present relative to that for the pure liquid. Figure 1 shows how the detected sound wave is varied in time and in amplitude when solids are suspended in the liquid. The arbitrary first distinct zero crossing time in liquid and in solidliquid are defined as ta and tb, respectively. The travel time between the transmitter and receiver in the liquid is defined as t0. Okamura et al. (1989) used a continuous stirred-tank reactor to correlate the concentration of solids to the relative time shift ((ta - tb)/t0). Thus, the concentration of solids in a slurry reactor can be measured by sending an ultrasonic pulse across the slurry and measuring the amplitude and time shift of that portion of the transmitted pulse received at the opposite side of the reactor. Then, comparing the value with those for known concentration, the concentration of solids is determined from the measured signal. Recent studies reported from our laboratories for the water-nitrogen-glass beads system at ambient conditions indicated that both the amplitude and the fractional change of transit time were affected by the variation of solids concentration (up to 35 wt %) under a constant nitrogen flow in the reactor (Soong et al., 1995). This paper gives the results of ultrasonic measurements of the concentration of solids in a three-phase slurry consisting of molten paraffin wax, nitrogen bubbles, and glass beads at 189 °C. Experimental Method Figure 2 shows a schematic representation of the 2.5-L autoclave reactor with which the ultrasonic investigation was conducted. The reactor vessel was a flat-bottom stainless steel cylinder, 10 cm in diameter and 30 cm in height, fitted with four equally spaced, vertical baffles, each 1/8 of the diameter of the vessel in width. The slurry was mechanically agitated by a motor-driven stirrer located 4 cm from the bottom of the vessel. A standard three-blade stirrer (4 cm in diameter) with a constant stirring speed of 600 rpm was used in this study and was slightly off the center of the reactor to avoid any interference with the ultrasonic signals, which were transmitted at 7 cm above the bottom of the reactor along the center of the reactor. Nitrogen bubbles were introduced through a sintered frit at the bottom of the reactor. The nitrogen flow was controlled electronically to a maximum of 400 mL/min through a mass-flow controller. Glass beads (80 ( 5 µm in diameter) were used as the solid in the slurry. The concentration of solids (solid weight/liquid weight) was varied from 1 to 25 wt % for each nitrogen flow in the reactor. Molten paraffin wax (Fisher Scientific, P-22) at a temperature of 189 °C was used as the liquid medium. Experiments were conducted at 0.1 MPa and 189 ( 0.5 °C in this study. Ultrasonic measurements

Figure 2. Schematic diagram of an autoclave reactor.

were taken by using a computer-based TestPro system, manufactured by Infometrics Inc., Silver Spring, MD. The ultrasonic unit was composed of a pulser/receiver board and an A/D and D/A board, both of which were controlled by a COMPAQ 386 computer. The software signals the pulser/receiver board to deliver pulsing instructions to the transducer. The transducer produces the ultrasonic burst by converting an electric pulse of a certain amplitude, duration, and frequency into an ultrasonic pulse using a piezoelectric crystal, which was mounted directly in the transducer. The transducer then emits an ultrasonic pulse which travels through the slurry in the reactor, and the receiver at the opposite side of the reactor receives the pulse. This information was then analyzed by the TestPro software. Data were obtained with longitudinal waves at a frequency of 2.5 MHz using lithium niobate transducers. Both the transducer and receiver were mounted inside a metal coupling rod which was screwed into a fitting inside the reactor wall (see Figure 2, enlarged area). The cooling air was provided to the metal coupling rod through the adapter to prevent the transducer/receiver from being exposed to high temperature. Results and Discussion Figure 3a shows a typical ultrasonic wave propagating in paraffin wax with a constant stirring in the setup described in this study. Here, ta is the arbitrary first distinct zero crossing time in liquid (ta ≈ 140.26 µs). The travel time between the transmitter and receiver in the liquid is defined as t0 (t0 ≈ 140.05 µs). We have noticed that the temperature of the paraffin wax greatly affects the ultrasonic signal. The effect of paraffin wax temperature on ultrasonic measurement under the conditions of a constant stirring speed and no gas flow in the autoclave reactor is presented in Figure 4. It shows that the measured transit time becomes significantly influenced by the paraffin wax temperature. The increase in transit time as the temperature increases is probably due to the change in

Ind. Eng. Chem. Res., Vol. 35, No. 6, 1996 1809

Figure 3. (a) Typical received signal in the autoclave reactor with paraffin wax only. (b) Effect of Solids on the ultrasonic signal.

one medium to another and the amount of reflection of sound at the boundary between the two media. If the impedances of two media are widely separated, e.g., nitrogen and paraffin wax, then most of the energy is reflected back in the first medium (paraffin wax) with little transmission into the second medium (nitrogen). It can be assumed that the ultrasonic pulse cannot penetrate through much of the nitrogen-paraffin wax interface at the current experimental frequency due to the acoustic impedance mismatch of this combination. Therefore, the amount of attenuation of the ultrasound beam by nitrogen bubbles is proportional to the amount of gas in the form of bubbles present in the path of the ultrasound. We also estimated the bubble sizes in separate experiments. The majority of the bubble sizes was found to be about 5 mm in the path between the transducer and the receiver within the range of flow studied. The number of bubbles increased as the nitrogen flow rate increased. The decrease of A/A0 as the nitrogen flow increased appears to be related to the bubbles. Chang et al. (1984) measured void fractions up to 20% in bubbly air-water two-phase flow using an ultrasonic transmission technique. Their results showed that the transmitted ultrasonic signal could be approximated by the exponential relationship:

A/A0 ) exp[-f(db) ]

Figure 4. Transit time as a function of temperature in paraffin wax.

where  is the void fraction and f(db) is a function dependent on the Sauter mean diameter. This correlation shows that the A/A0 ratio has an exponential relationship both with the void fraction and with a function dependent on the bubble diameter. The effect of air bubble diameter on the A/A0 ratio was found to be significant, with A/A0 decreasing with increasing bubble size. Bensler et al. (1987) also measured the interfacial area in bubbly flows in air-water systems by means of an ultrasonic technique. Their observations showed that the transmitted ultrasonic signal could also be expressed by an exponential relationship:

A/A0 ) exp[Γx/8S(kdb/2)]

Figure 5. Amplitude ratio and transit time as a function of nitrogen flow in paraffin wax.

densities and other physical properties in the paraffin wax which result in the variation of the ultrasonic signal with temperatures. To overcome this complication, the temperature in the autoclave was controlled at 189 ( 0.5 °C in this study. Figure 5 shows the change in the amplitude ratio of the transmitted ultrasonic signals A/A0 in the reactor as a function of nitrogen flow in paraffin wax at 189 °C. A and A0 are the amplitudes of the transmitted signals with and without the presence of nitrogen, respectively. Figure 5 also indicates that the amplitude ratio is approximately an inverse exponential function of the nitrogen flow. When a sound wave strikes the boundary between two different media and the acoustic impedances of the two media are different, part of the acoustic energy will be reflected and part will be transmitted. Acoustically, the impedances of the two media will determine the transmission of the wave from

(1)

(2)

where Γ is the volumetric interfacial area, x is the travel distance in the path, S is the scattering coefficient, k is the wavenumber of the ultrasonic wave which surrounds the bubble, and db is the Sauter mean bubble diameter. Equation 2 shows that the A/A0 ratio has an exponential relationship with the interfacial area and the scattering cross section, which is a function dependent on both the bubble radius and the wavenumber of the ultrasonic wave surrounding the bubble. Our observations of A/A0 in the nitrogen-paraffin wax system are in qualitative agreement with those reported by Chang et al. (1984) and Bensler et al. (1987). The decreasing A/A0 ratio as the nitrogen flow increased in Figure 5 may be attributed to a combination of the void fraction, the bubble size, the number of bubbles, and the scattering cross section. Figure 5 also illustrates the effect of the nitrogen flow on the transit time, tb. It was approximately 140.25 µs at all nitrogen flows. Apparently the transit time was unaffected by the nitrogen flow under the current experimental conditions, because what we measured was the signal not transmitted through the nitrogen. Thus, the measured transit time should not be affected by the nitrogen flow. The results indicate that only the amplitude and not the transit time of the ultrasonic signal are affected by the nitrogen flow rate in the reactor under the current

1810

Ind. Eng. Chem. Res., Vol. 35, No. 6, 1996

Figure 6. Amplitude ratio and transit time as a function of solids concentration (wt %) in paraffin wax.

experimental conditions. Chang et al. (1984) also reported that the amplitude of the transmitted sound pulses depends significantly on the number of bubbles; however, the transit time does not change with the void fraction. Our results obtained from the nitrogenmolten wax system are similar to those of Chang et al. (1984) for the air-water system. Figure 3b shows a typical ultrasonic signal as affected by 2 wt % of solids. Clearly, the presence of the solids caused a substantial decrease in the transit time. The first distinct zero crossing time, ta, which is 140.26 µs in Figure 3a, is shifted to 140.20 µs in Figure 3b, after the addition of 2 wt % of solids. When an ultrasound wave strikes the boundary between the paraffin wax and solids, the acoustic impedances of the two media are different and the ultrasound wave will be partially reflected and transmitted. The reflected wave travels back through the wax at the same velocity (943 m/s for paraffin wax at 189 °C). The transmitted wave continues to travel through the solids at the sound velocity of the new medium (5571 m/s for fused silica at 189 °C estimated from Lynnworth (1989)). This explains the shift in transit time. Furthermore, the concentration of the solids in the three-phase slurry greatly affects both the amplitude ratio and the transit time, tb, of an ultrasonic signal. Figure 6 shows the effect of solids concentration on the amplitude ratio of the transmitted ultrasonic signal A/A0 (A and A0 are the amplitudes of the transmitted signals with and without the presence of solids, respectively) in the autoclave at a constant temperature of 189 °C and with a constant stirring speed and without nitrogen flow. The A/A0 ratio decreased as the concentration of solids increased from 2 to 25 wt %. In general, the amplitude of the transmitted ultrasonic signals decreased as the solids concentration increased. When an ultrasonic pulse is sent across the slurry reactor, the amplitude of the pulse is reduced when it strikes a solid in the slurry because the pulse is partially scattered and absorbed. Scattering is often the predominant form of attenuation in heterogeneous materials. It occurs when some of the ultrasonic wave incident upon a discontinuity in a material, for example, a solid particle, is scattered in directions which are different from that of the incident wave. The unscattered pulse is transmitted through the solid at the sound velocity of the solid to the receiver. The amplitude of the transmitted portion of the pulse is measured by the receiver located on the opposite side of the autoclave. The more particles present in the path, the less transmitted pulse will be detected. The amplitude ratio of the pulse is inversely proportional to the quantity of solid particles present in the path. Therefore, the

Figure 7. Transit time and time ratio versus solids concentration (wt %) with different nitrogen flow in paraffin wax.

measured amplitude of the sound wave transmitted through a slurry mixture is expected to be a strong function of the concentration of solids. Figure 6 also presents the effect of the solids concentration on the transit time, tb. The transit time decreased as the solids concentration increased in the reactor. It was approximately 140.25 µs with no solids but decreased to approximately 139.25 µs as the solids concentration increased to 25 wt %. These phenomena can be understood by assuming that the presence of solid particles in the path of sound wave reduces the transit time because sound waves travel at a much faster velocity in the solid (5571 m/s for fused silica at 189 °C) than in the paraffin wax (943 m/s at 189 °C). The transit time of the transmitted ultrasonic pulse should depend on the quantity of solid particles present in the path. The higher the concentration of solid particles present in the path, the shorter the transit time that would be observed. Figure 7 shows the effect of solids concentration on the transit time as a function of nitrogen flow. The solids concentration has a significant effect on the transit time at a constant flow of nitrogen of 100 mL/ min: it was about 140.20 µs at 2 wt %, decreased to approximately 139.79 µs as the solids concentration increased to 15 wt %, and decreased further to 139.27 µs as the solids concentration increased to 25 wt %. As discussed earlier, the higher the concentration of solid particles present in the path, the shorter the transit time that would be observed. We speculate that the decrease of transit time as the solids concentration increased is due to the number of solid particles present in the path. It was interesting to note that the presence of nitrogen bubbles affects the transit time. For most cases, the presence of more nitrogen bubbles causes a slight decrease in transit time. It should be noted that the principal effect of the nitrogen bubbles is to improve the suspension of the beads, thereby increasing the attenuation and decreasing the transit time over the values obtained with the stirrer alone. Separate tests conducted with the constant stirring speed and nitrogen bubbles without beads confirm that the presence of nitrogen bubbles alone causes no effects on the transit time (Figure 5). Figure 7 further illustrates the fractional change in transit time ∆t/t0 (∆t ) ta - tb) as a function of solids concentration under different nitrogen flows. The fractional change in transit time (time ratio) increased as the solids concentration increased. The results illustrated in Figure 7 can be described by a simple model in which it is assumed that the presence of solid particles in the path of the sound wave reduces the transit time because the sound velocity in the solid is faster than that in the liquid. If the solid particles

Ind. Eng. Chem. Res., Vol. 35, No. 6, 1996 1811

are assumed to be randomly distributed, the time ratio ∆t/t0 is related to the concentration fraction ω by the equation:

∆t/t0 ) Rω1/3(1 - vl/vs)

(3)

Here R is a constant for particles of a given composition and vl and vs are the speed of sound in the liquid and solid, respectively. A derivation of this equation is given elsewhere (Soong et al., 1995). The above equation, which predicts that the time ratio should be dependent upon solids concentration, is approximately valid for the experiments presented here for solids well suspended in the liquid phase. However, it is clear that the predicted experimental dependence on solids concentration does not adequately describe the data. The observed dependence is closer to a linear rather than a cube root dependence on ω. This discrepancy is perhaps not surprising because eq 3 is only expected to be true for ω , 1, and this condition is violated over the range of solids concentration investigated here. It should be noted that the simple model (assuming that the ultrasound wave can transmit through the nitrogen bubble) predicts that the number of nitrogen bubbles in the liquid should increase the transit time, because the speed of sound in nitrogen is less than that in the liquid. This increase will not be detected using the current setup due to the acoustic impedance mismatch and will not affect our measurements because the signal that we measured was not transmitted through the nitrogen. In concentrated systems, multiple scattering is important; i.e., a significant proportion of the energy scattered from one particle is incident upon neighboring particles (McClements, 1991). The velocity of the sound and attenuation depends on particle size as well as concentration under this condition (Anson and Chivers, 1989). Two important parameterssthe ultrasound wavenumber, k, and the radius of the particle, asare often utilized to describe the ultrasound propagation and attenuation. For this study, the ka range is about 1.3. As indicated previously, the bubble size was about 5 mm. A slight change in the transit time (velocity) observed in the presence of more nitrogen bubbles is probably due to the improved suspension. The theoretical ultrasonic propagation of polystyrene spheres in water and glass spheres in epoxy systems has been studied by Anson and Chivers (1989). Although their systems are different from those in this study, their theoretical approachessan effective medium approach and a multiple scattering approachscan be applied to further theoretical study. The results from our study are in qualitative agreement with those previously reported by Okamura et al. (1989). Our results with molten wax are similar to the previous results from the nitrogen-water-glass beads system (Soong et al., 1995). However, our previous results for the nitrogen-waterglass beads system give larger values of the time ratio at a given solids concentration fraction than our current nitrogen-paraffin wax-glass beads data. A possible explanation of this discrepancy could be that the geometry of our previous apparatus and nitrogenwater-glass bead system differs considerably from that employed here. In general, both the amplitude and the fractional change of transit time (or time ratio) were affected by the variation of solids concentrations. It appears that the time ratio depends mainly on the solids concentration. The variation of nitrogen flow has little effect on the observed fractional change of transit time

(or time ratio) within the flow conditions studied. Based on these results, it can be concluded that the ultrasonic technique with further development has potential applications for solids concentration measurements in three-phase slurry reactors under high temperature. The subsequent development will involve the effect of operating variables on the wave propagation. At low solids concentration, the linear relationship between ω and the transit time and the time ratio is not well conserved. This deviation from the linear dependency which affects the accuracy of the ultrasonic technique is under investigation. Conclusion The results presented in this study led to the conclusion that both the amplitude and the transit time of an ultrasonic signal are influenced by the variation of solids concentration in molten paraffin wax. The variation of nitrogen flow has little influence on the observed transit time within the conditions studied. The ultrasonic technique has potential applications for solids concentration measurements in three-phase slurry reactors at elevated operating temperature. Acknowledgment The authors appreciate the assistance from R. R. Anderson and C. E. Taylor. I.K.G. acknowledges the support of the U.S. Department of Energy through ORISE. Reference in this report to any specific commercial product, process, or service is to facilitate understanding and does not necessarily imply its endorsement or favoring by the U.S. Department of Energy. Nomenclature a ) bubble or particle radius, m A ) amplitude of the transmitted signal A0 ) amplitude of the transmitted signal in the absence of nitrogen or solid k ) wavenumber of the ultrasonic wave which surrounds the bubble or particle f(db) ) function dependent on the Sauter mean diameter ta ) arbitrary first distinct zero crossing time in liquid, s tb ) arbitrary first distinct zero crossing time in the presence of solid in slurry, s t0 ) arbitrary travel time of the sound wave between the transmitter and the receiver in the absence of solid in slurry, s vl ) speed of sound in liquid, m/s vs ) speed of sound in solid, m/s ∆t ) (ta - tb), s ∆t/t0 ) time ratio x ) travel distance in the path, m S( ) ) scattering coefficient Greek Letters R ) constant for particles at a given composition  ) void fraction ω ) solids concentration fraction Γ ) volumetric interfacial area Fl ) Fluid density, kg/m3 Fs ) particle density, kg/m3

Literature Cited Abts, L. R.; Dahi, P. H. Ultrasonic Determination of Component Concentrations in Multi-Component Fluids. U.S. Patent 4,580, 444, 1986.

1812

Ind. Eng. Chem. Res., Vol. 35, No. 6, 1996

Anson, L. W.; Chivers, R. C. Ultrasonic Propagation in SuspensionssA Comparison of a Multiple Scattering and an Effective Medium Approach. J. Acoust. Soc. Am. 1989, 85, 535-540. Behrman, C. E., Larson, J. W. On-line Ultrasonic Particle Monitoring of Brewing Operations. MBAA Tech. Quart. 1987, 24, 7276. Bensler, H. P.; Delhaye, J. M.; Favreau, C. Measurement of Interfacial Area in Bubbly Flows by Means of an Ultrasonic Technique. Proc. ANS Natl. Heat Transfer Conf. 1987, 240246. Bonnet, J. C.; Tavlarides, L. L. Ultrasonic Technique for DispersedPhase Holdup Measurements. Ind. Eng. Chem. Res. 1987, 26, 811-815. Chang, J. S.; Ichikawa, Y.; Irons, G. A.; Morala, E. C.; Wan, P. T. Void Fraction Measurement by an Ultrasonic Transmission Technique in Bubbly Gas-Liquid Two-Phase Flow. In Measuring Techniques in Gas-Liquid Two-Phase Flows; Delhaye, J. M., Cognet, G., Eds.; Springer-Verlag: New York, 1984; pp 319335. Foote, K. G. Ultrasonic Particulate Sensing. U.S. Patent 4,527, 420, 1985. Hilgert, W.; Hofmann, H. Characterization of Gas Phase Flow in Bubble Columns at Low Superficial Gas Velocities with the Aid of Ultrasonic Doppler Technique. Ger. Chem. Eng. 1986, 9, 180190. Hwang, S. J.; Fan, L. S. Some Design Considerations of a Draft Tube Gas-Liquid-Solid Spouted Bed. Chem. Eng. J. 1986, 33, 49-56.

Karplus, H. B.; Raptis, A. C. Flow Measurement of Dense Slurries Using the Sonic Doppler Principle. Ultrason. Symp. Proc. IEEE 1978, 291-295. Kato, Y.; Nishiwaki, A.; Fukuda, T.; Tanaka, S. The Behavior of Suspended Solid Particles and Liquid in Bubble Column. J. Chem. Eng. Jpn. 1972, 5, 112-118. Lee, S. L. P.; de Lasa, H. I. Phase Holdups in Three-Phase Fluidized Beds. AIChE J. 1987, 33, 1359-1370. Lynnworth, L. C. Ultrasonic Measurements for Process Control; Academic Press, Inc.: San Diego, CA, 1989. McClements, D. J. Ultrasonic Characterization of Emulsions and Suspensions. Adv. Colloid Interface Sci. 1991, 37, 33-72. Okamura, S.; Uchida, S.; Katsumata, T.; Iida, K. Measurement of Solids Holdup in a Three-Phase Fluidized Bed by an Ultrasonic Technique. Chem. Eng. Sci. 1989, 44, 196-197. Soong, Y.; Blackwell, A. G.; Schehl, R. R.; Zarochak, M. F.; Rayne, J. A. Ultrasonic Characterization of Three-Phase Slurries. Chem. Eng. Commun. 1995, 138, 213-224.

Received for review November 27, 1995 Revised manuscript received March 22, 1996 Accepted March 23, 1996X IE950771P

X Abstract published in Advance ACS Abstracts, May 15, 1996.