Scholarly article on topic 'Hydrodynamics of gas-agitated liquid-liquid extraction columns'

Hydrodynamics of gas-agitated liquid-liquid extraction columns Academic research paper on "Chemical engineering"

0
0
Share paper
Academic journal
Acta periodica technologica
OECD Field of science
Keywords
{""}

Academic research paper on topic "Hydrodynamics of gas-agitated liquid-liquid extraction columns"

Review

HYDRODYNAMICS OF GAS-AGITATED LIQUID-LIQUID EXTRACTION COLUMNS

Milan N. Sovilj

University of Novi Sad, Faculty of Technology, Bulevar cara Lazara 1, 21000 Novi Sad, Serbia

Although the non-agitated extraction columns (spray column, packed column, perforated plate column, sieve plate column, etc) can handle high flow rates and are simple and cheap, there have been relatively few applications of these columns because they suffer from serious backmixing of the continuous phase. It was shown that the backmixing is reduced when the spray column is operated with dense packing of drops. Another way of increasing the efficiency of a non-agitated extraction column is to introduce an inert gas (air, nitrogen, oxygen) as a mixing agent in the two-phase liquid-liquid (L-L) system. This method of energy introduction increases the turbulence within the new three-phase gas-liquid-liquid (G-L-L) system, which causes an improved dispersion of droplets, and, consequently, a higher dispersed phase holdup and therefore a great mass transfer area. The present study reports the hydrodynamics in the non-agitated extraction columns, as well as the axial dispersion for the two- and three-phase systems.

KEY WORDS: extraction columns; gas-liquid-liquid system; hydrodynamics

INTRODUCTION

Because of their simplicity, low cost, and versatility, non-mechanically agitated columns are still extensively used in extraction processes. They are also a convenient and inexpensive way to experimentally test theoretical models of mass transfer in simple extraction systems. In regard with the dispersed phase holdup, spray extraction column, as one of the simplest extraction columns, can operate in three modes of packing of the dispersed phase drops: dispersed, restrained, and dense.

Although they can handle high flow rates and are simple and cheep, there have been relatively few applications of these columns because they suffer from serious backmixing of the continuous phase. It was shown hat the backmixing was reduced when the spray column operated with dense packing of drops. One of the way to increase the efficiency of a spray column is to introduce an inert gas as a mixing agent in the two-phase L-L system. This method of energy introduction increases the turbulence within the three-phase G-L-L system, which causes the increase of average dispersed phase holdup and a

* Corresponding author: Milan N. Sovilj, University of Novi Sad, Faculty of Technology, Bulevar cara Lazara 1, 21000 Novi Sad, Serbia, e-mail: miso@uns.ac.rs

UDC: 66.069.82:532.5 BIBLID: 1450-7188 (2012) 43, 199-216 Review

larger mass transfer area. Mass-transfer or chemical reactions for G-L-L systems may be also encountered in gas absorption, gas-liquid reactions, and fermentation, often with a heterogeneous liquid catalyst, or liquid-liquid reactions with gas agitation. Some examples can be cited: absorption of SO2 into the aqueous emulsion of xylidine in water (1); purification of crude naphthalene with H2S04 accompanied by air sparging (2); air oxidation of hydrocarbon in aqueous emulsion, fermentation of hydrocarbons, in which the substrate is dispersed in an aqueous culture medium with air bubbling; and extractive fermentation of useful species, such as alcohols and steroids, which are produced in the aqueous phase by the metabolism of the relevant microorganisms and are extracted in situ into the coexisting organic phase of an extractant, shifting the reaction favorably. A few examples of more complicated systems containing solid particles are air oxidation of substituted benzyl alcohol catalyzed by palladium catalyst in the presence of an aqueous phase, which gives rise to the favorable formation of aldehyde (3), and competitive liquid-phase hydrogenation of cyclohexanone and cyclohexene catalyzed, by Ru catalyst in the presence of water (4). Packed towers operated under gas-liquid countercurrent conditions have found increased applications in distillation, absorption, and liquid-liquid extraction processes. They are also becoming increasingly important environmental protection technologies. The extraction of hydrogen peroxide by means of deionized water from anthraquinone working solution via anthraquinone process was carried out in a gas-agitated sieve plate extraction column (5). The effect of superficial velocity of air, dispersed phase and continuous phase on the overall plate extraction efficiency has been investigated in the mentioned paper. The correction for the prediction of the overall plate extraction efficiency was also presented.

Gas-liquid-liquid columns have many advantages over any other conventional gasliquid or liquid-liquid contactors,. They are of simple configuration without moving parts and require no seal, need little space and maintenance. In this columns it is easily and in the widely intreval adjusts the resistance time of the liquid phases. They allow comparatively large liquid-phase volumetric mass-transfer coefficients or interfacial area to be achieved with relatively low energy consumption. Namely, the efficiency of non-agitated extraction columns (spray, packed, perforated, and sieve plate column) can be cosiderab-ly increased by introducing an inert gas as a mixing agent in the two-phase L-L system. The transition and steady state behavior of the gas agitated two-phase L-L dispersions is well characterized for spray columns, where the gas is introduced into the continuous liquid phase at the base of the column and the second liquid phase is dispersed at the top of the column. Dispersions and emulsions can also arise as a consequence of liquid en-trainment by bubbles as they pass through a liquid-liquid interface. This mode of dispersion or emulsion formation is pertinent also to batch type processes, where neither liquid phase is dispersed initially, and the gas is again introduced from below. Such examples are pyrometallurgical processes for the production of nickel and copper, processes for electro-organic synthesis, and the dispersion or emulsification of oil slicks in breacking waves (6). The introducing gas as a mixing agent in the two-phase L-L dispersion caused the formation of large liquid-liquid interface area due to the presence of the smaller disperd phase drops. Coalescence times for drops at the liquid-liquid interface were found to be rapid and appear to be unafected by the rate of bubble passage. The energy introduced by the mixing agent increases the turbulence within the three-phase G-L-L system, which

UDC: 66.069.82:532.5 BIBLID: 1450-7188 (2012) 43, 199-216 Review

brings about an improved dispersion of the droplets and, consenquently, a higher dispersed phase holdups, and also higher backmixing in the continuous phase. Galkin et al. (7) concluded that the extraction efficiency was nearly three times greater for conventional columns when air was introduced into the sieve plate extraction column at the lower inlet, and claimed that the process was more efficient than by the use of stirring or pulsation of the column.

The aim of this paper is to give a critical review of the hydrodynamics characteristics of non-mechanically agitated extraction columns, which use an inert gas (air, oxygen, nitrogen, etc.) as a mixing agent in the two-phase liquid-liquid system. The energy thus introduced increases the turbulence within the now, three-phase G-L-L system, which brings about an improved dispersion of the droplets and, consequently, a higher holdup and larger mass transfer area.

HYDRODYNAMIC CHARACTERISTICS OF GAS-LIQUID-LIQUID SYSTEMS

The hydrodynamics of a system represents one of the main difficulties in the scale-up of liquid-liquid extractors. As for the design, difficulties arise mainly because of the dispersion in radial and axial directions; however, in most cases, the radial dispersion has a small influence. The main hydrodynamic characteristics in the non-mechanically agitated extraction column are the slip velocity, dispersed phase holdup, gas phase holdup, drop size distribution, and axial dispersion in the continuous phase. In the following text we will discuss the effects of these characteristics on the operation of non-mechanically gasagitated liquid-liquid extraction columns.

Dispersed phase holdup and slip velocity

In the generalization of dispersed phase holdup data, for a specified value of the slip velocity (us) for the countercurrent flow in an L-L system in the spray extraction column, use can be made of the familiar Thornton-Pratt relationship (8):

Us = "Z \ + — [1]

(1 -sd) sd

where: uc, ud - superficial velocity of the continuous and dispersed phase, respectively, sd - dispersed phase holdup. For a packed column voidage (e), the slip velocity has the next form:

Us = Z C ) + — [2]

e (1-Sd ) esd Equation (1) was corrected by the additon of a new part, as follows (9):

iS + S = U0 (1 -Sd) [3]

(1-Sd) sd wherein the viscosity ratio of the phases (m) is defined by:

m = 1.22 (^ / Vc )0 2 [4]

where: u0 - velocity of a single drop, /ud, M - dynamic viscosity of the dispersed and continuous phase, respectively. If the static dispersed phase holdup is defined as: £st = ud / us, when uc = 0, eq. [1] can be rewritten in the form:

u, u u, us = — = T \ + — [5]

— (1-— d ) —d

On the basis of the amount of experimental data simple empirical equations for the estimation of the slip velocity and dispersed phase holdup in the two-phase L-L columns were derived by several authors (11-15). The equation presented by Kumar et al. (11) predicts the slip velocity for the dispersed phase holdup (0.01 to 0.75) and Reynolds number (7 to 2450). This equation gives the average absolute value of the relative errors of 14.5% and 13.5% for the dispersed holdup and slip velocity, respectively. Kumar and Hartland (15) derived an empirical expression for the prediction of the dispersed phase holdup and slip velocity in droplet dispersions settled under gravity. This equation is valid in a wide interval of the dispersed phase holdup (0.01 to 0.76) and Reynolds number (0.61 to 3169), with the average absolute value of the relative error of 14.3% and 12.8% for the dispersed holdup and slip velocity, respectively. Sovilj (13) proposed an empirical relationship for the prediction of the dispersed phase holdup and slip velocity in the liquid-liquid spray extraction columns. A good agreement between the experimental and predicted values of the slip velocity by this equation were obtained for the dispersed phase holdup in the range (0.0097 to 0.362) and Reynolds number (58 to 1067). The average deviations for the slip velocity and dispersed phase holdup were 9.6% and 14.0%, respectively. On the basis of a large bank of published experimental data for eigth different types of extraction columns (rotation disc, asymetric rotation disc, Kuhni, Wirz-II, pulsed perforated-plate, Karr reciprocatind-plate, packed, and spray columns), Kumar and Hartland (15) presented a unified correlation for the prediction of dispersed phase holdup in the two-phase L-L dispersion. The average error of predicted data on the entire data sets using this equation was 18.1%, which is better than that achieved by most authors in attempting to correlate their own experimental results. The highest error was 22.7% for the rotating disk and asymetric rotating disk columns, and a lowest 14.1% for the spray extraction columns. The errors for the pulsed perforated-plate and packed columns were 19.0% and 18.3%, respectively.

The Experimental procedure applied in the paper (14) was as follows: at the beggi-ning of each run, the cylindrical part of the spray extraction column was filled to about half its volume with water (continuous phase) throuh a water distributor at the top of the column, and the level of continuous phase (Hc) was recorded. At that moment, the dispersed phase (toluene) was introduced at the bottom of the column with a chosen flow rate, and two-phase dispersion occurred. The position of the interface (Hb) corresponded to the height of the two-phase dispersion above the toluene inlet. When the volume of the two-phase dispersion (water-toluene) became constant, air (gas phase) was introduced at the bottom of the column via a gas distributor, and the interface level in the column increased. At that moment, the continuous phase was introduced in the column at the chosen flow rate. The new position of the interface (Ht) corresponds to the height of the three-phase G-L-L dispersion above the toluene and air inlet (14).

Review

On the basis of the experimental procedure for the estimation of dispersed phase holdup, explained in the previous text, the dispersed phase holdup (sdt) in the three-phase dispersion (air-water-toluene) was calculated from the following relationship (14):

Hb - Hc H

The mean value of the dispersed phase holdup in the three-phase dispersion was determined with an uncertainty of + 3%.

In the G-L-L system, the continuous phase holdup (sct) is defined by:

s0t = 1 - sdt - sg [7]

where: sdt - dispersed phase holdup in the G-L-L system, sg - gas phase holdup. On the other hand, the basic eq. [1] for the G-L-L system in the spray extraction column can be expressed as (16):

(1-^dt )

+ — = uCH (l

-sdt-Sg,

where: uch - characteristic velocity identified as the mean relative velocity of the droplets extrapolated to the essentially zero flow rate, defined by Thornton (17).

The effective slip velocity of the dispersed liquid was also analyzed by using the longitudinal dispersion coefficients of the measured dispersed liquid (18,19). However, the studies in these papers were limited only to the air-kerosene-water system. Wang et al. (20) expressed the slip velocity in the G-L-L system air-anthraquinone-aqueous working solution in the form:

1 - Sg - Sd

The average dispersed phase holdup in the spray extraction column increased with increasing the dispersed and gasesous phase superficial velocities, at the constant value of the continuous phase superficial velocity (5), fig. 1.

uG (x 10"3 m/s)

Figure 1. Effects of gas superficial velocity on the holdup of dispersed phase [Source: Cheng et al., Ind. Eng. Chem. Res. 47 (2008) 741-7418]

Review

The average dispersed phase holdup exibits a relatively small increase with the increase in the gas phase superficial velocity. Moreover, Billet and Braun (21) concluded that an initial sinking of dispersed phase holdup takes places at gas phase superficial velocities below 0.2 cm/s. They clamed that within that gas flow rate range, the energy input is sufficient to produce an intensive turbulence and, consequently, to form a large number of droplets (21). On the other hand, the increase in the dispersed phase holdup observed by them was below 20% for the gas phase superficial velocity within the range from 0.2 to 0.6 cm/s at a constant dispersed phase superficial velocity of 0.4 cm/s.

The average drop size in most of the investigated two-phase L-L systems can be expressed as a Sauter drop diameter, d32, in the form:

where: nt - number of the drop diameters in the limited range, dt - values of the drop diameters in the given range.

Nishikawa et al. (22) measured the effects of the volume fraction of dispersed phase, viscosity of liquids, impeller speed and impeller-to-vessel diameter on the average drop size of a dispersion in a mixing vessel for the two-phase: water (continuous phase)-honey bees' wax (dispersed phase). Hatate et al. (23) also measured the mean droplet size for several systems using two columns and correlated it as a function of the superficial gas velocity, interfacial tension, and column diameter. The same authors (24) measured the average gas holdups, the longitudinal distribution of the volume fraction of a dispersed liquid (droplet), and the longitudinal dispersion coefficients of a dispersed liquid, using two bubble columns (0.066- and 0.122 m i.d.) with a perforated plate as a gas sparger. Their columns were operated batchwise or continuously with respect to the liquids. They found smaller average gas holdups for the air-kerosene (dispersed liquid)-water (continuous liquid) system than for the corresponding ones without kerosene, over a range of superficial gas velocity: ug = 0.007-0.09 m3/(m s), analyzed the longitudinal distribution of the volume fraction of the dispersed liquid, using a dispersion model allowing for the slip velocity. Diaz et al. (25) examined the dependence of the mean droplet size on the superficial gas and liquid velocities, and measured the dispersion coefficients of both liquids, for the air-kerosene-water system. Priestly and Ellis (26) also found that the efficiency of non-mechanically agitated extraction column with different packings can be considerably increased by the introduction of an inert gas as a mixing agent in two-phase L-L systems. On the other side, Kato et al. (27) extended the studies of Hatate et al. (23,24) from a single-stage to the multistage bubble columns of the same diameter. Their study was also limited to the air-kerosene-water system with a few additional ones for the measurement of average gas holdups. Using the organic liquids (kerosene, dibutyl phtha-late, or groundnut oil) dispersed in water, Bandyopadhyay et al. (28) measured the average gas holdup of air in a bubble column (0.2 m in diameter) with a multiple nozzle sparger plate, operated batchwise with respect to the liquids. They found that the fractio-

Drop size and gas phase holdup

nal holdup depends on the gas velocity, liquid properties, phase inversion in the liquid mixture, as well as on the spreading coefficient of the organic liquid. In the presence of a liquid with a negative spreading coefficient, the holdup is a minimum at the phase inversion point but the reverse is true for a liquid with a positive coefficient of spreading. The model assumes that the particles rise or fall with the slip velocity caused by the density difference between the dispersed and continuous phases and explains well the behavior of the solid particles in the suspension bubble columns (16,17). The longitudinal distribution of the fractional gas holdup was measured in the bubble columns with two immiscible liquids (29). The columns were operated batchwise with respect to both liquids, over a wide range of relevant physical properties and average volume fraction of the dispersed liquid. The average gas holdups could be correlated by an empirical expression presented in the literature for a single liquid phase, when it was applied to the individual liquid phases, allowing for their volume fraction. Doungdeethaveeratana and Sohn (30) investigated a novel solvent extraction process without moving parts, in which the emulsion is generated the by bottom gas injection rather than by mechanical stirring. They found that this process had a number of advantages over the mixer-settler unit or the spray extraction column, which provideed a sufficiently large intefacial area for mass transfer. Yan (31) studied the process of extracting hydrogen peroxide from an anthraquinone working solution with the bottom air injection in a spray column. This result showed that the extraction efficiency was 2-3 times higher than that of conventional liquid-liquid extraction without air introduction. Lu et al. (32) presented the results of the extraction of hydrogen peroxide with deionized water from the anthraquinone solution via anthraquinone process, which was carried out in a gas-agitated sieve plate extraction column. Experimental procedure for the estimation of dispersed phase holdup in the three phase G-L-L system was described in the paper (13). The author presented the hydrodynamic characteristics of the air-water-toluene three-phase G-L-L system in a countercurrent spray extraction column. If the position of the interface was maintained constant in the cylindrical section of the column by the adjustable overflow tube, the average dispersed phase holdup on the three-phase G-L-L dispersion was calculated by the following relation (13):

H - Hb

Sg [11]

The uncertainty of the average gas holdup measurements in the phase system air-water-toluene was estimated to be + 5%. The average gas phase holdup increased with increasing superficial velocity of the gas phase in the three-phase G-L-L systrem, whereas the gas phase holdup determined at a constant ratio of the contiuous and gas phases decreased with incresing the superficial velocity of the dispersed phase (12). The hydrodynamics of a spray extraction column operated with the liquid-liquid and gas-liquidliquid systems was intensively investigated (35-37).

Hikita et al. (38) derived an empirical correlation for the prediction of the average gas phase holdup in a three-phase G-L-L system, as follows:

eg = 0.672 Ca0578 Mo -

in \ 0 06V „ \

where: Ca=(uc/L)/c and Mo= // g /pL c3 - characteristic parameters, c - interficial tension, g - acceleration due to gravity, pL - density of the liquid phase, pg - gas phase density, /o, /L - dynamic viscosity of the gas and liquid phase, respectively. The predictions may be seen to be in reasonable agreement with their experimental data. Asai and Yoshi-zawa (29) showed the relation for the calculation of the average gas holdup in the three phase air-water-kerosene system which was based on the relationship for the G-L system, when the mean volume fraction of the dispersed phase was (dt = 0.50:

sg =(l -(dt )Sgc + Sgd (dt [13]

where: sgc, sgd - average gas phase holdup in continuous and dispersed phase of the G-L-L dispersion. As a check of the presented static pressure measurements, the values of (%) for the system air-water-kerosene were determined by measuring the difference in the total liquid height between the sparged and unsparged conditions. Although this technique of the measurements of the liquid level under the sparged conditions was not so easy because of the violent fluctuation, the average values of several measurements agreed with the values obtained from the static pressure measurements within an error of about 10%. Asai and Yoshizawa (29) presented experimental values of the gas-phase holdup in the two-phase air-water system and three-phase air-water-kerosene system. Both systems gave the same observed values and were in good agreement with the predictions, eq. [13]. At the same time, figure 4. in the paper (29) shows the data for gas-phase holdup in the three-phase system air-water-kerosene obtained from empirical relationship of Bandiot-phayay et al. (28), whose form is:

/■ \0.011 ' PM cl

a/;C 0,496 i 0.25 = 065 U g (

Vg Vm J

0.20 > (dt < 1.0 [14]

where: ug - superficial velocity of the gaseous phase, pm, /4 - density and dinamic viscosity of the liquid mixture, respectively. Asai and Yoshizawa (29) concluded that this expression predict worse values for the air-kerosene-water system than those for the air-water system, but still it gives rise to larger values than their data for both systems. This is not in line with the findings of Hatate et al. (23), Kato et al. (24) and Bandyopadhyay et al. (28), who found lower average gas holdups for the air-kerosene-water system. However, these experiments were performed with a perforated plate or a multiple nozzle, in the region of lower superficial gas velocity, where the flow mechanism was known to vary with the configuration of the gas sparger. Therefore, this different observation may be possibly attributed to the different effect of the liquid physical properties on the fractional gas holdups in the different flow regimes. In fact, the data of Bandyopadhyay et al. (28) reveal a reduction of the difference in the gas phase holdups between both systems with an increase of the gas flow rate. Bandyopadhyay et al. (28) claimed that eq. [14] correlates their data for the air-kerosene-water system worse than those for the air-water systems, but still it gives rise to larger values than for the experimental data of Asai and Yoshizawa (29) for both systems. These authors presented graphically the relation between the average gas phase in the G-L (air-water) and G-L-L (air-kerosene-water) system, fig. 2.

Review

Key System

1 Air-Kerosene-Water

2 Air-Water ^

uG (m/s)

Figure 2. Average gas holdups for the air-kerosene-water and air-water systems: DT = 0.064 m; t = 9.7 + 1.9oC [Source: Asai and Yoshizawa, Ind. Eng. Chem. Res., 30 (1991) 745-751]

Wang et al. (20) presented an empirical expression for the calculation of the gas-phase holdup in a gas-agitated sieve plate extraction column, as follows:

eg = 0.215

/ \ 3.104

(u 4 gA

- 0.0071-

where: pg, pc - density of the gaseous and continous phase, respectively, /ug, ¡j.c - dynamic viscosity of gaseous and continuous phase, respectively.

The effects of the gaseous superficial velocity on the gas holdup in different liquid-liquid and gas-liquid-liquid systems have been widely investigated, which is presented, fig. 3.

ud (cm/s); uc = 0.1115 cm/s

1 0.0318

2 0 1486

3 0.2550 [1-5] Sovilj and Knezevic (13)

4 0.3840

5 0.4520

6 Xiong and Zhang (51)

7 Therning and Rasmuson (52)

8 Bandyopathyay et al (28)

9 Hikita et al.(38)

10 Asai and Yoshizawa (29)

u , cm/s

Figure 3. Average gas holdups for the different gas-liquid-liquid and gas-liquid systems

Drop size and interface area

In chemical engineering, the rate of mass transfer between two different phases often directly determines the production rate of the process (e.g., the gas absorption rate in gasliquid systems). The mass transfer rate is directly proportional to both the mass transfer coefficient and the specific interfacial area between the different phases. Both parameters depend mainly on the (local) hydrodynamic situation inside the system. For the design purposes as well as for the improvement of the existing production facilities, it is very important to have a better insight into the phenomena that affect these parameters.

The knowledge of the dispersed phase drop size is of primary importance in the design of liquid-liquid non-mechanicall agitated extraction columns. It affects the dispersed phase holdup, the residence time of the dispersed phase, and the free throughputs. Furthermore, together with the dispersed phase holdup, it determines the interfacial area disposable for mass transfer and affects both the continuous and dispersed phase mass transfer coefficients. It is therefore important to be able to predict the drop diameter as a function of the column geometry, physical properties of the liquid-liquid system, and direction of mass transfer (39).

Seibert and Fair (40) proposed a new equation for the prediction of the Sauter drop diameter in the packed and spray extraction columns, as follows:

d32 = 1.15 r

LApg _

where j is a correction factor calculated from the experimental drop diameter data assumed from the literature, Ap - difference of density. Its values are r = 1.0 for no mass transfer or transfer from the continuous phase to the dispersed phase and r = 1.0 - 1.8 for mass transfer from the dispersed phase to the continuous phase. Kumar and Hartland (39) presented a relationship for the limiting value of the drop size in the absence of agitation or at low levels of agitation in the liquid-liquid extraction columns in the following form:

d32 = CCc/ Apg)1/2 [17]

where the constant, C1, is a function of the column geometry, mass transfer, and the characteristics of the liquid-liquid system employed. Vedaiyan et al. (41) proposed an empirical correlation for the calculation of the Sauter drop diametar, given by:

d 32 = 1.59

( u 2 y

0 [18]

where: u0 - superficial velocity of the dispersed phase at the nozlle, d0 - diameter of the nozlle of the distributor of dispersed phase. The gas-liquid interfacial area, which is determined by the gas holdup and the Sauter mean bubble diameter, determines the production rate in many industrial processes. The effect of additives on this interfacial area is often not undrestood, especially in multiphase systems (gas-liquid-solid, gas-liquid-liquid). The addition of a third phase can cause the gas-liquid system to become completely opaque, which means that conventional techniques to study the interfacial area cannot be used (41). The influence of different additives (1-octanol; dodecane, and toluene) on

the interfacial area was studied in a stirred vessel and in a bubble column under coalescing and noncoalescing conditions (42). It was found that the addition of toluene to a noncoalescing electrolyte system decreased the interfacial area to a large extent by turning it into a coalescing system, due to the interaction between gas bubbles and liquid organic droplets. Furthermore, around the toluene solubility concentration, both the gas holdup (measured using an electric conductivity technique) and the interfacial area increased to the values similar to those observed in noncoalescing systems. The cause of this remarkable phenomenon lies probably in the presence of a small toluene layer around the gas bubbles, which can be formed beyond the solubility point (43). This layer is absent at the concentrations below the solubility limit and a large surface tension gradient exists between these two situations, which can be responsible for the sharp change in the coalescence behavior.

A comparison of ultrasonic spectroscopy with a digital camera technique was performed in a flat (20x3x150 cm) bubble column using the ultrasonic technique in combination with the electrical conductivity method and a digital camera technique with digital image analysis, simultaneously (43). The camera was placed 10 cm in front of the column, and the ultrasonic transducers were mounted into the wall of the column (the measurement path length was 20 cm). Measurement of the exact size distribution using the ultrasonic technique was difficult, mainly due to the small attenuation and ultrasonic velocity differences. These differences were small due to the low gas holdups that were applied («1%), which was necessary for the digital camera technique to work optimally. The interfacial area could, however, be determined accurately, and together with the measurement of the gas holdup using the electrical conductivity technique, the Sauter mean bubble diameters were calculated. The value of the interfacial area (a) of the bubble size distribution can be calculated from the Sauter mean diameter (d32) and the gas phase holdup (%), according to the following relation:

a =-a- [19]

The liquid-liquid interfacial area and liquid-phase mass-transfer coefficients in the emulsion bubble columns were measured by Fernandes and Sharma (43), who took advantage of the alkaline hydrolysis reaction of several esters for their determination. For the analysis they assumed complete mixing of both the continuous and dispersed liquids. Yoshida et al. (44) measured the mean diameter of kerosene dispersed in the water phase of bubble columns, operated batchwise with respect to both liquids. They studied the variation of the oxygen absorption into water with the addition of kerosene, liquid paraffin, toluene, and oleic acid. They claimed that in the previous studies, the effects of physical properties on the various characteristics of the bubble columns were not clarified. Assai and Yoshizawa (29) presented the longitudinal distribution of volume fraction of the dispersed liquid over a wide range of relevant physical properties and average volume fraction of the dispersed liquid. The observed longitudinal distribution of the volume fraction of the dispersed liquid was analyzed by means of the dispersion model, allowing for the slip velocity caused by the density difference between both liquid phases. The observed Peclet numbers based on the slip velocity were empirically correlated as a function of the relevant system parameters.

UDC: 66.069.82:532.5 BIBLID: 1450-7188 (2012) 43, 199-216 Review

The hydrodynamic characteristics of air-anthraquinone working solution-water three-phase system used for the production of hydrogen peroxide were determined in a gasagitated sieve plate extraction column (5). The effects of the superficial velocities of the gaseous phase, organic dispersed phase and continuous phase on the organic dispersed phase holdup were investigated. The organic dispersed phase holdup increased with the increase of the superficial velocity of the gaseous phase and organic dispersed phase. The effect of the superficial velocity of the continuous phase on the organic dispersed holdup could be neglected. Based on the equation of the relative velocity between the organic dispersed phase and continuous phase, a method used to calculate the organic dispersed holdup was proposed (20). The organic dispersed holdup of gas-liquid-liquid systems (including liquid-liquid systems) in this study and in data from the literature (32) were calculated using the method proposed in that study (20). The calculation data were well consistent with the corresponding experimental data (20) and the literature data (33,34), and the relative error were 4.9-15.5%.

Axial mixing

A theoretical and experimental study has been carried out on the dynamics of two-phase countercurrent flow with interfacial transfer in a packed bed absorption column (45). An eight-parameter model has been formulated consisting of axially dispersed plug flow for the gas phase and a piston-diffusion exchange model for the liquid phase. In addition, three limiting cases of this model have been analyzed. Solutions of the models have been obtained in the Laplace domain with four possible transfer functions for each model as a result. Only two of these transfer functions have been found useful for an experimental study of the absorption of a poorly soluble gas. Experimental measurements of these two transfer functions, in the form of frequency characteristics, have been carried out in a 0.105 m diameter column packed to a height of 2.1 m by glass spheres 0.01 m in diameter. The absorption system studied was water-air-oxygen. Evaluation of the parameters of the formulated models was carried out in the frequency domain. The results showed that the models with a stagnant liquid zone are considerably better than the axi-ally dispersed models. For a more reliable assessment of the various models, however, a combination of several independent measurements is recommended.

Axial mixing arises in the packed columns from the fact that the „packed" fluid do not all move through a packed bed at a constant and uniform velocity, either because of either velocity gradients in the fluid, or eddy motion in the packed voids. Axial mixing tends to reduce the concentration driving force for mass transfer that would exists for piston flow (45). To achive a given separation, more transfer units are required for the axial-mixing case owning to the reduced drivning force. Longitudinal dispersion coefficients of the continuous phase were experimentally obtained in spray type liquid-liquid extraction columns (45). The method used was unsteady-state measurements of a KCl solution as the tracer. It was concluded that the increase in the continuous phase velocity greatly increased the axial mixing coefficient in binary mixtures (Ecb), and the increase in the dispersed phase velocity decreased the axial mixing coefficient (45). Small dispersion coefficients were found for small tower lenghts and these coefficients increased as the tower lenght increased. Also, at a long lenghts, where the end of the effects became negligible, Ecb was

UDC: 66.069.82:532.5 BIBLID: 1450-7188 (2012) 43, 199-216 Review

indipendent of the lenght. A decrese in the tower diameter from 35.8 mm to 27 mm caused a decrease in Ecb of approximately 20% for a range of the continuous phase velocities. A comparison of the Peclet and Reynolds numbers for spray towers with those for packed beds gave comparable values. Using the data of the work (46) and the values from the literature (47), the following realatioship was obtained using the method of least squares:

Ecb = 3,43x10~4 u(0A2 ( uc < 4.5 mm/s) [20] where Ecb - continuous phase axial dispersion cofficient of the two-phase system, uc was used in mm/s. The correlation coefficient was 0.94, with an average deviation of + 8%.

Diaz et al. (25) showed that high axial dispersion coefficients were deduced in both the liquid phases for the three-phase air-water-kerosene system (Ect), in which water was the continuous phase and kerosene the dispersed phase. They also found that the values of the axial dispersion Peclet number of the water phase were 0.1 to 1.2, decreasing with the increase in the flow rates of air or kerosene, or when the flow water rate was reduced. Diaz et al. (25) concluded that the Peclet number for the kerosene phase decreased to the values between 0.4 and 0.1 when the air flow rate increased. Kato et al. (27) investigated the axial dispersion in the multistage bubble columns for the air-water-kerosene system. They concluded that the dispersion-phase coefficient in the three-phase G-L-L system (Edt) increased with increasing the gas phase superficial velocity and coluumn diameter, and was independent of the the total liquid velocity in the range from 0.05 to 1.0 cm/s. Kato et al. (27) derived also an empirical equation in which Edt depends on the gas phase superficial velocity, column diameter and gravitattion acceleration.

Asai and Yoshizawa (29) measured the longitudinal dispersion coefficients of the continuous (Ec) and dispersed phase (Ed) in bubble columns (with air as a gaseous phase) operated batchwise with respect to two immiscible liquids (2-ethylhexanol, water or kerosene). They concluded that the longitudinal coefficients Ec and Ed of the continuous and dispersed phase were independent of the clear liquid height. It was shown that Ec and Ed increased with the increase in the volume fraction of disperesed liquid (p) in the system air-2-ethylhexanol-water, which had a highly viscous dispersed liquid. For the air-kerose-ne-50% aqueous sucrose solution system with a highly viscous continuous liquid, Ec increased and Ed decreased with an increase in p. All observed longitudinal dispersion coefficients of both liquids were correlated by empirical correlations (29). Figure 4 shows the effect of the gaseous superficial velocity (ug) on the longitudinal coefficient of the continuous and dispersed phase of (Ec, Ed) in the system air-etylhexanol-water (29). These authors concluded that the increase in the visosity of the dispersed liquid apears to rather improve the dispersioin of both liquids, fig. 4.

The continuous phase axial dispersion coefficients of the three-phase G-L-L system in a gas-agitated spray extraction column, described above, were examined by Sovilj (48). The system used was water as a continuous, toluene as a dispersed, and air as a gaseous phase. The experimental values of the continuous phase axial mixing coefficients were obtained by unsteady-state measuring of the concentrations of a tracer solution (solution of potassium chromate in water) in the continuous phase. The increase in the gas phase superficial velocity increased the continuous phase axial mixing coefficient. A nonlinear

UDC: 66.069.82:532.5 BIBLID: 1450-7188 (2012) 43, 199-216 Review

dependence between the continuous phase axial mixing coefficient and continuous phase superficial velocity was observed.

♦ Ec ED

0.50 1 2 0.25 3 4

UG (m/s)

Figure 4. Effects of the gas phase superficial velocity on the longitudinal dispersion coefficients Ec and Ed or the air-2-ethylhexanol-water system:

DT = 0.064 m; t = 14.0 ± 2.9oC [Source: Asai and Yoshizawa, Ind. Eng. Chem. Res., 31 (1992) 587-592]

No correlation was found between the continuous phase axial dispersion coefficient and the dispersed phase superficial velocity. The increase in the dispersed phase holdup generated a growth of the continuous phase axial dispersion coefficient. The continuous phase axial dispersion coefficients in the spray extraction column were higher for the three-phase air-water-toluene system (48) than those obtained for the two-phase water-toluene system (49) under the same operating conditions. Regression analysis showed that the mean increase in the continuous phase axial dispersion coefficient in the three-phase system Et) was approximately 90%. In the paper (48), an equation for the prediction of the continuous phase axial mixing coefficient was developed, as is given below:

f 2 \0.77 f \

d32 n 1 o„l U 2 d0 Pc I uc d0 P.

= 0.124

-0.24 gg0.23 [20]

where: Ect- continuous phase axial dispersion coefficient of the three-phase system, d0 -orifice diameter, sdt - dispersed phase holdup in the G-L-L dispersion. The average deviation for eq. [16] was 17.7%. Seventy-one percent of the predicted continuous phase axial dispersion coefficients lied within the + 20% limits and 84% within the + 30% limits. These results are in accordance with the results for the two-phase system (48) and with the conclusion of Horvath et al. (50) that an average deviation within 30% was sufficient for the use with back mixing models.

UDC: 66.069.82:532.5 BIBLID: 1450-7188 (2012) 43, 199-216 Review

CONCLUSION

This review article deals with the hydrodynamic characteristics of the non-mechani-cally agitated extraction columns. An inert gas (air, nitrogen, oxygen) as a turbulence agent was introduced in the two-phase liquid-liquid system. In the new, three-phase gasliquid-liquid system, the gaseous phase causes intensive turbulence, which caused improving of the average dispersed phase holdup and a larger mass transfer area. Mass-transfer or chemical reactions for three-phase systems may be also encountered in the gas absorption, gas-liquid reactions, and fermentation, often with a heterogeneous liquid catalyst, or liquid-liquid reactions with gas agitation. Different empirical equations which describe a function of the dispersed phase holdup, gas phase holdup, and axial distribution coefficient were analyzed and compared.

REFERENCES

1. Kohl, A. L. and Riesenfeld, F. G.: Gas Purification, 4th ed., Gulf, Houston (1985) 223-234.

2. Doraiswamy, L. K. and Sharma, M. M.: Heterogeneous Reactions, Vol. 2, Wiley, New York (1984) 187-193.

3. King, W. M.: U. S. Pat. 4306083, Dow Chemical Co. 1981; Chem. Abstr. 96 (1982) 68598P.

4. Koopman, P. G. J, Buurmans, H. M. A, Kieboom, A. P. G. and Van Bekkum, H.: Solvent-Reactant-Support Interactions in Liquid Phase Hydrogenation. Reel. Tray Chim. Pays-Bas. 100 (1981) 156-161.

5. Cheng, Y., Wang, L., Lü, Sh., Wang, Y. and Mi, Zh.: Gas-Liquid-Liquid Three-Phase Reactive Extraction from the Hydrogen Peroxide Preparation by Anthraquinone Process. Ind. Eng. Chem, Res. 47 (2008) 7414-7418.

6. Shaw, J. M. and Konduru R.: The Behaviour of Large Gas Bubbles at a Liquid-Liquid Interface. Part 2: Liquid Entrainment. Can. J. Chem. Eng. 70 (1992) 381-384.

7. Galkin, N. P., Gorianov, N. E., Tikhomirov, V. B. and Fedorov, V. D.: The Mecha-nizm of Distribution of Liquids in a Plate Extractor. J. Nucl. Energy (A and B) 14 (1961) 132-133.

8. Treybal, R.: Liquid Extraction, Mc-Graw Hill, New York (1963).

9. Varfolomeev, B. G., Pebalk, V. L., Chigogidze, K. Sh, Lan, N. N. and Fernando, R. S.: Spray Extraction Columns: Drop Size and Dispersed Phase Holdup. Theor. Found. Chem. Eng. 34, 6 (2000) 556 -561.

10. Pilhofer, Th.: Extraction Liquide-Liquide. Hydrodynamik von Tropfenschwärmen in Flüssig-Flüssig Sprüchkolonnen. Chem.-Ing.-Tech. 46, MS 133/74 (1974).

11. Kumar, A., Vohra, D. K. and Hartland, S.: Sedimentation of Droplet Dispersion in Counter-Current Spray Columns. Can. J. Chem. Eng. 58 (1980) 154-159.

12. Kumar, A. and Hartland, S.: Prediction of Drop Size, Dispersed Phase Holdup, Slip Velocity, and Limiting Troughputs in Packed Extraction Columns. Chem. Eng. Res. Des. 72 (1994) 89-104.

UDC: 66.069.82:532.5 BIBLID: 1450-7188 (2012) 43, 199-216 Review

13. Sovilj, M.: Slip Velocity and Hydrodynamic Parameters in Liquid-Liquid Spray Columns, Collect. Czech. Chem. Commun. 54 (1989) 990-995.

14. Sovilj, M. and Knezevic, G.: Gas-Agitated Liquid-Liquid Extraction in a Spray Column. Collect. Czech. Chem. Communications. 59 (1994) 2235-2243.

15. Kumar, A., Hartland, S.: A Unified Correlation for the Prediction of Dispersed Phase Holdup in Liquid-Liquid Extraction Columns. Ind. Eng. Chem. Res. 34 (1995) 39253940.

16. Ziebland, G. and Hackl, A.: Begaste Flüssig-Flüssig Sprühkolonnen. VDI-Berichte, 315 (1978) 509-526.

17. Thornton, J. D.: Limiting Holdup and Flooding Rates in Spray-Extraction Columns, Chem. Ind. 46 (1954) 1581-1588.

18. Suganuma, T. and Yamanishi, T.: Behavior of Solid Particles in Bubble Columns, Kagaku Kogaku 30 (1966) 1136-1140.

19. Cova, D. R.: Catalist Suspension in Gas-Agitated Tubular Reactors. Ind. Eng. Chem. Process Des. Dev. 5 (1966) 20-25.

20. Wang, L., Cheng, Y., Lü, Sh., Mi, Zh.: Calculation of Organic Dispersed Holdup in Gas-Agitated Extraction Columns. J. Tianjan Uviversity 40, 3 (2007) 260-264.

21. Billet, R., Braun, Chr.: Gas Agitated Liquid-Liquid Extraction in Columns Filled with Regularly Arranged Bialecki Rings. International Solvent Extraction Conference, (ISEC'86), München, Book of abstracts, 3 (1986) 123-126, DECHEMA, Frankfurt/M.

22. Nishikawa, M., Mori, F. and Fujieda, Sh.: Average Drop Size in a Liquid-Liquid Phase Mixing Vessel. J. Chem. Eng. Japan 20, 1 (1987) 82-88.

23. Hatate, Y., Okuma, S., Kato, Y.: Longitudinal Dispersion Coefficient and Holdup Distribution of Droplets in Bubble Columns. Kagaku Kogaku Ronbunsyu 1 (1975) 577-582.

24. Hatate, Y., Mori, S., Okuma S. and Kato, Y.: Drop Size in Gas-Liquid-Liquid System Bubble Columns. Kagaku Kogaku Ronbunsyu 2 (1976) 133-137.

25. Diaz, M., Aguayo, A. T. and Alvarez, R.: Hydrodynamics of Liquid-Liquid Coun-terurrent Extraction Column with Upflow Gas Agitation. Chem.-Ing.-Tech. 58 (1986) 74-75.

26. Priestly, R., Ellis, S. R. M.: Gas Agitated Liquid Extraction Columns. Chem. Ind. 7 (1978) 757-760.

27. Kato, Y., Kago, T. and Morooka, S.: Longitudinal Concentration Distribution of Droplets in Multi-Stage Bubble Columns for Gas-Liquid-Liquid Systems. J. Chem. Eng. Jpn. 17 (1984) 429-435.

28. Bandyopadhyay, N., Ray, P. and Dutta, B. K.: Gas Holdup in Bubble Column with Immiscible Liquid Mixtures. Can. J. Chem. Eng. 66 (1988) 995-999.

29. Asai, S. and Yoshizawa, H.: Longitudinal Holdup Distribution of Gas and Dispersed Liquid in Bubble Columns with Two Immiscible Liquids. Ind. Eng. Chem. Res. 30 (1991) 745-751.

30. Doungeethaveeratana, D. and Sohn, H. Y.: The Kinetics of Extraction in a Novel Solvent Extraction Process with Bottom Gas Injection Without Moving Parts. J. Hydro-metallurgy 49 (1998) 229-254.

31. Yan, T.: Ph.D. Thesis, Tianjin University, Tianjin, P. R. China, 2003.

UDC: 66.069.82:532.5 BIBLID: 1450-7188 (2012) 43, 199-216 Review

32. Lü, Sh., Mi, Zh., Wang, Y. and Wang, L.: Experimental Investigation and Simulation of a Gas-Agitated Sieve Plate Column. Chem. Eng. Technol. 27, 8 (2004) 903-908.

33. Hancu, D., Green J. And Beckman, E. J.: Hydrogen Peroxide in CO2/H2O Biophasic Systems: Green Synthesis and Epoxidation Reactions. Ind. Eng. Chem. Res. 41, 18 (2002) 4466-4474.

34. Wang, Q., Wang, L., Wang, Y.: Study on the Activation and Regeneration of Pd/ Al2O3 Catalyst in Hydrogen Peroxide Production by Anthraquinone Process, React. Kinet. Catal Lett. 81, 2 (2004) 297-304.

35. Sovilj, M.: Holdup and Backmixing in Liquid-Liquid Spray Column. The International Solvent Extraction Conference-ISEC'88, July 18-24, Moscow, USSR, Proceedings, 2 (1988) 274-276.

36. Sovilj, M.: Prediction of Slip Velocity and Holdup in Liquid-Liquid Spray Columns. 2nd International Conference on Separation Science and Technology - ICSST, Oktober 1-4, Vol. II, Hamilton, Canada, Proceedings (1989) 252-258.

37. Sovilj, M.: Hydrodynamics of a Gas-Agitated Spray Extraction Column. XVth Congress of Chemists and Technologists of Macedonia, October 2-4, 1997, Skopje, Macedonia, Proceedings (1997) 372-373.

38. Hikita, H., Asai, S., Segawa, K., Tanigawa, K. and Kitao, M.: Gas Holdup in Bubble Columns. Chem. Eng. J. 20 (1980) 59-67.

39. Kumar, A. and Hartland, S.: Gravity Settling in Liquid-Liquid Dispersions. Can. J. Chem. Eng. 63 (1985) 368-3676.

40. Seibert., A. F. and Fair, J. R.: Hydrodynamics and Mass Transfer in Spray and Packed Liquid-Liquid Columns. Ind. Eng. Chem. Res. 27 (1988) 470-481.

41. Vedaiyan, S., Degaleesan, T. E. and Laddha, G. S.: Mean Drop Size and Characteristic Velocity of Droplet Swarms in Spray Columns. India J. Technol. 12 (1974) 135139.

42. Cents, A. H. G., Jansen, D. J. W., Brilman, D.W. F. and Vesterteeg, G. F.: Influence of Small Amounts of Additives on Gas Holdup, Bubble Size, and Interface Area. Ind. Eng. Chem. Res. 44 (2005) 4863.

43. Fernandes, J. B. and Sharma, M. M.: Air-Agitated Liquid-Liquid Contactors. Chem. Eng. Sci. 23 (1968) 9-16.

44. Yoshida, F., Yamane, T. and Miyamoto, Y.: Oxygen Absorption into Oil-in-Water Emulsions. Ind. Eng. Chem. Process Des. Dev. 9 (1970) 570-577.

45. Moravec, P. and Stanek, P.: Counter-current Absorption of Oxygen by a Frequency Response Technique: Die Untersuchung der Gegenstromabsorption von Sauerstoff mit der Frequenzgangtechnik. Chem. Eng. Process.: Process Intensification, 24, 2 (1988) 93-103.

46. Hazlebeck, D. C. and Geankoplis, C. J.: Axial Dispersion in Spray-Type Extraction Tower. Ind. Eng. Chem. Fundam. 2 (1963) 310-315.

47. Geankoplis, C. J., Shap, J. B., Arnold, F. C. and Marroquin, G.: Axial Dispersion Coefficients of the Continuous Phase in Liquid-Liquid Towers. Ind. Eng. Chem. Fun-dam. 21 (1982) 306-311.

48. Sovilj, M.: Axial Dispersion in a Three-Phase Gas-Agitated Spray Extraction Column, Collect. Czech. Chem. Commun. 63 (1998) 283-292.

UDC: 66.069.82:532.5 BIBLID: 1450-7188 (2012) 43, 199-216 Review

49. Sovilj, M. and Zoric, M.: Axial Dispersion in a Spray Extraction Column for Water-Toluene System. Chem. Ind. 51, 5 (1997) 207-211.

50. Horvath, M., Steiner, L., Hartland, S.: Prediction of Drop Diameter, Holdup and Backmixing Coefficients in Liquid-Liquid Spray Columns. Can. J. Chem. Eng. 56 (1978) 9-18.

51. Xiong, J. and Zhang, L.: Effects of Gas-Agitated and Packing on Hydrodynamic and Mass Transfer of Extraction Column. Can. J. Chem. Eng. 82 (2004) 1076-1080.

52. Therning, P. and Rasmuson, A.: Liquid Dispersion and Gas Holdup in Packed Bubble Columns at Atmospheric Pressure. Chem. Eng. J. 81 (2001) 69-81.

ХИДРОДИНАМИКА ЕКСТРАКЦИОНИХ КОЛОНА ТЕЧНО-ТЕЧНО АГИТОВАНИХ ГАСОМ

Милан Н. Совил Технолошки факултет, 21000 Нови Сад, Булевар цара Лазара 1, Република Србща

У овом раду дат jc приказ и анализа хидродинамичких карактеристика екстрак-ционих колона течно-течно код ко_|их се користи инертан гас као агитатор. Уво^е-аем гаса у колонски уре^ са двофазним системом течно-течно формира се знатно ефикаснщи трофазни систем гас-течно-течно, пошто се ситаеаем капи дисперго-ване фазе повеЬава специфична површина одговорна за пренос масе у систему. Дат ]е и анализиран утица] средаег садржа] дисперговане и гасне фазе на пренос масе у трофазном систему гас-течно-течно. У исто време, приказане су и корелацще за сваку од хидродинамичких величина, као и аихова тачност у предви^аау ових величина у колонском уре^у. Коначно, приказан ]е и утица] повратног мешала на хидродинамику екстракционих колона течно-течно, као и гас-течно-течно. Анали-зиране су и одговара]уЬе емпирщске корелацще ще да]у везу измену коефицщента повратног мешала и хидродинамичких карактеристика екстракционих колона.

Кшучне речи: екстракционе колоне, систем гас-течно-течно, хидродинамика

Received: 4 April 2012 Accepted: 13 June 2012