FLOODING IN PULSED SIEVE PLATE EXTRACTION COLUMNS WITH MASS TRANSFER EFFECTS

FLOODING IN PULSED SIEVE PLATE EXTRACTION COLUMNS WITH MASS TRANSFER EFFECTS

^*2

¹ Departamento de Engenharia Mecânica - Escola Politécnica da Universidade de São Paulo

Cx. Postal 61548 - CEP 05508-900 - São Paulo, SP - Brazil

² Departamento de Engenharia Química - Escola Politécnica da Universidade de São Paulo

Cx. Postal 61548 - CEP 05508-900 - São Paulo, SP - Brazil

(Received: January 30, 1997; Accepted: October 20, 1997)

Abstract - Measurements of flooding in a pulsed sieve plate extraction column were carried out using toluene - water and toluene - acetone - water systems. The influence of plate spacing and initial acetone mass fraction in toluene (in the case of the three-component system) on flooding curves was studied and empirical correlations for the maximum of these curves are presented. In the ternary system, acetone goes from toluene drops to water: the descending phase. An experimental pulsed column, with an internal diameter of 39.6 x 10^-3 m and an operational height of 2,670 m, and with stainless steel perforated plates, thickness of 1.5 x 10^-3 m, with a hole diameter of 3 x 10^-3 m and a free fractional area of 23%, was used. It was verified that the plate spacing and initial acetone mass fraction in the toluene have a great influence on throughput and frequency at the maximum; this influence depends on the phase ratio.

Keywords: Extraction, separation column, mass transfer.

INTRODUCTION

The pulsed sieve-plate extraction column (PSE) is the liquid-liquid extraction equipment most frequently used in industry due to its considerable flexibility, high throughputs (Haverland et al., 1987) and high separation efficiency (Lorenz et al., 1990). Nevertheless, according to the same authors, some factors of poorly understood, such as drops coalescence, plate wettability and effect of impurities, limit the use of analytical approaches. Therefore, as shown by many authors (Brandt et al., 1978; Berger and Walter, 1985; Haverland et al., 1987; Pilhofer, 1988 and Lorenz et al., 1990), the choice and design of an extractor depend upon extensive pilot plant tests with each system studied.

However, as pointed out by Lorenz et al. (1990), the cost of these tests can be drastically reduced if data on fluid dynamic and mass transfer are known for the following systems proposed by the European Federation of Chemical Engineering, EFCE (Misek, 1985 in Lorenz et al., 1990): n-butanol - succinic acid - water, n-butyl acetate - acetone - water and toluene - acetone - water. As these systems have very different interfacial tensions, they permit behavior prediction of extraction columns for other systems.

To study the flooding in a PSE column with mass transfer for different plate geometries, Berger et al. (1978) used the first and third systems above and the ethilhexanol - acetic acid - water system. Using the butanol - water, butyl acetate - water and toluene - water binary systems, Berger and Walter (1985) studied the flooding of the same PSE column used in the preceding paper and presented an empirical correlation for the maximum of the flooding curve as a function of the frequency of pulsation.

In these studies, the systems with high interfacial tension: toluene - water, in the absence of mass transfer, and toluene - acetone - water, in the presence of mass transfer effects, presented the most significant changes in flooding curves with the variation of plate geometry, phase ratio and pulsation intensity.

In the present work the same systems were thereby used for the purpose of extending these studies by analyzing the influence of plate spacing and initial acetone mass fraction in toluene (in the case of the three-component system). A correlation for the maximum of the flooding curves is also presented.

The experimental work was done for three phase ratios V_c / V_d , (the ratio between the continuous and dispersed phase flow rates), 5 , 1 and 0.2 and for four plate spacings, 25 x 10^-3, 50 x 10^-3, 75 x 10^-3 and 100 x10^-3 m. In all tests the dispersed and ascending phase was the toluene phase. In the cases where mass transfer was present, acetone went from toluene drops to water, i.e., from the dispersed to the continuous phase, in the d ® c direction. The toluene stream entered into the column with 0.05 ± 0.002 or 0.10 ± 0.002 acetone mass fraction.

Due to its influence on drop size, the mass transfer direction strongly affects the hydrodynamic behavior of the column. In studies (Logsdail et al., 1957; Komasawa and Ingham, 1978a and 1978b; Bender et al., 1979; Shen et al., 1985 and Tsouris and Tavlarides, 1993) conducted on several types of extraction columns, an increase in drop size, with a consequent increase in the column throughput (due to the minor drops dragging by the continuous phase) was verified in the case of the d ® c direction, and a decrease was verified in the opposite direction. This behavior was also verified by Berger (1981) in Pilhofer and Schroeter (1986) and Kleczek et al. (1989) in PSE columns operating with the toluene - acetone - water system.

EXPERIMENTAL WORK

An experimental pulsed glass column, with an internal diameter of 39.6 x 10^-3 m and an operational height of 2,670 m, and with stainless steel perforated plates, thickness of 1.5 x 10^-3 m, with a hole diameter of 3 x 10^-3 m and a free fractional area of 23%, was used. This plate geometry presents the best performance for the toluene - acetone - water system according to Berger et al. (1978). Furthermore, it is a plate geometry usually found in the literature.

Four sampling probes were installed equidistant along the column for simultaneously collecting samples of the two phases, to determine the concentration profiles. The samples were analyzed in a chromatograph.

As described by Berger and Walter (1985), flooding points were determined by an indirect measurement of the hold-up via the differential pressure between the bottom and the top of the column. The procedure for obtaining each flooding point consisted of fixing the intensity of pulsation and varying the load (V_c+V_d) until the column flooded.

Experiments were performed at a constant amplitude of pulsation, as shown by Berger et al. (1978) and Berger and Walter (1985). An amplitude of pulsation of 8 x 10^-3 m was used, since in tests with the two-component system toluene - water, Aufderheide (1985) verified that, for amplitudes ranging from 6 x 10^-3 to 12 x 10^-3 m, the flooding remained almost constant with a maximum at 8 x 10^-3 m.

The experiments were conducted in the temperature range of 20 ± 2 ° C. Further details about the experimental apparatus and data can be found in Tribess (1995).

RESULTS AND DISCUSSION

The first series of experiments were done with the plate spacing of 50 x 10^-3 m due to the fact that this is the one most frequently found in the literature. The next experiments were planned as follows.

- The phase ratio of 0.2 was not used further due to the small influence of mass transfer on the flooding curves (Figure 1).

- The plate spacing of 25 x 10^-3 m was used only when the phase ratio equals 1 because, for the plate spacing of 50 x 10^-3 m and when the phase ratio equals 5, the acetone separation from toluene was already almost complete (Figure 6).

Influence of Mass Transfer

Figures 1 to ⁴ show the influence of mass transfer, i.e., initial acetone mass fraction in toluene, on the flooding curves: generally the column flooding increases in the presence of mass transfer and the curves are displaced to the right. These conclusions and the few exceptions are better observed in Figure 5 which represents the change in the curves’ maximums with phase ratio and plate spacing.

The increase of these curves’ maximums confirms the theory that mass transfer in the d ® c direction increases column throughput.

Regarding the displacement of the flooding curves to the right when mass transfer increases, this should be a consequence of the increase in coalescence as the acetone concentration in the phases increases (Groothuis and Zuiderweg, 1960); a greater tendency towards coalescence should require a higher energy of pulsation, i.e., greater intensity of pulsation, to form new small drops on the plates.

The largest variation in the maximum column loading and in the displacement of flooding curves was verified when the phase ratio equals 1, an initial acetone mass fraction in toluene of 10% and a plate spacing of 50 x 10^-3 m, as seen in Figure 1.

Influence of Plate Spacing

By analyzing Figure 5, one can also observe that the maximum column loading increases with plate spacing when the phase ratio equals 5, in both types of experiments: in the presence and absence of mass transfer. This fact was also shown by Cohen and Beyer (1953), Sege and Woodfield (1954) and Rouyer et al. (1974) and is due to the mean drop diameter increase when there are fewer plates.

When the phase ratio equals 1, this also occurs in the absence of mass transfer; however in the presence of mass transfer there is a maximum for the plate spacing of 50 x 10^-3 m.

All these facts can be observed by analyzing the concentration profiles corresponding to the phase ratios of 5 and 1, presented in Figures 6 and 7.

Figure 1: Flooding curves for plate spacing of 50 x 10^-3 m.

Figure 2: Flooding curves for plate spacing of 75 x 10^-3 m.

Figure 3: Flooding curves for plate spacing of 100 x 10^-3 m.

Figure 4: Flooding curves for plate spacing of 25 x 10^-3 m.

The higher the average acetone concentration in the aqueous phase for the entire column, the larger the average drop size will be; this is due to coalescence (Groothuis and Zuiderweg, 1960), which tends to increase the maximum column loading.

At a phase ratio of 1 (Figure 7) the higher average concentration in the aqueous phase throughout the column occurred for the plate spacing of 50 x 10^-3 m, due to a very great concentration in the inferior column part.

In addition, the greater the acetone concentration in the aqueous phase at the bottom of the column, the more severe the maximum coalescence condition will be, which requires a higher energy of pulsation, with displacement of flooding curves to the high intensity region. This behavior also tends to increase the maximum column loading.

So the higher column loading and its greater percent increase at a phase ratio of 1 occurred at a plate spacing of 50 x 10^-3 m (Figure 5).

Figure 5: Maximum column loading (V_c+V_d) in the flooding curves.

Figure 6:
Concentration profiles at a phase ratio of 5.

Figure 7:
Concentration profiles at a phase ratio of 1.

EMPIRICAL CORRELATIONS OF THE MAXIMUM OF THE FLOODING CURVE

According to Berger and Walter (1985) and Palma (1991), even for a given system, there is no general empirical correlation capable of representing flooding data in all ranges of PSE column parameters, even without mass transfer. So it is necessary to restrict correlations only for relevant experimental data.

Since Berger et al. (1978) have verified that the optimum relation between separation efficiency and throughput occurred at a pulsation frequency around 10 - 20 strokes/min higher than the frequency at the maximum of the flooding curve, the determination of this maximum has a practical purpose.

Berger and Walter (1985) proposed empirical correlations to determine the maximum point of the flooding curve - throughput, (V_c+V_d)_m , and frequency of pulsation, f_m - using their data obtained for the n-butanol - water, n-butyl acetate - water and toluene - water systems, i. e., data from experiments with an absence of mass transfer:

.

(1)

+ (2)

In these equations, the interfacial tension, s , is given in terms of 10^-3 N/m; (V_c+V_d)_m and f_m in m³/m² h and min^-1, respectively; the phase ratios, V_c / V_d and V_d / V_c , and the free fractional plate area, e , are dimensionless.

These correlations were used in the present research as the basis to obtain the following correlation for the condition of the maximum of the flooding curves for the system of toluene -water, in the absence of mass transfer, and the system of toluene - acetone - water, in the presence of mass transfer.

Throughput at the Maximum

Using the experimental data obtained in this work, in the presence and absence of mass transfer, and the data from Berger and Walter (1985), in the absence of mass transfer, the following empirical correlation was obtained by applying multilinear regression techniques to the experimental data:

.

. (3)

or,

.

where the column diameter, D, the plate spacing, H_p , and the hole plate diameter, d_p , are given in mm and the phase ratio, V_c / V_d , and the initial acetone mass fraction in toluene, c_t , are dimensionless.

Frequency of Pulsation at the Maximum

Initially, a modification was made in the second polynomial of the correlation of Berger and Walter (1985), Equation (2), in order to relate parameters in a product form. This will make it easier to obtain new correlations using that one as a starting point.

Substituting Equation (2), the resulting correlation is:

. (4)

From Equation (4), a new correlation was obtained for the frequency of pulsation at the maximum by adopting a procedure similar to the one adopted to obtain the correlation for (V_c+V_d)_m , Equation (3), resulting in:

.

(5)

or,

.

where the amplitude of pulsation, a , the column diameter, D, the plate spacing, H_p, and the hole plate diameter, d_p, are given in mm and the phase ratio, V_d / V_c , and the initial acetone fraction in toluene, c_t , are also dimensionless.

Figures 8 and 9 show a comparison of the experimental results obtained in this work, for the systems toluene - water and toluene - acetone - water, and those obtained by Berger and Walter (1985), for the system toluene - water, with the correlations above, Equations (3) and (5).

For c_t = 0, i.e., the acetone mass fraction in the feed stream of toluene equals zero, Equations (3) and (5) are reduced to a correlation for the two-component system, toluene - water, in the absence of mass transfer.

Figure 8:
Comparison of experimental results obtained in this work and by Berger and Walter (1985) and the correlation proposed, Equation (3), for the throughput at the maximum of the flooding curve for the toluene - water and toluene - acetone - water systems.

Figure 9:
Comparison of experimental results obtained in this work and by Berger and Walter (1985) and the correlation proposed, Equation (5), for the frequency of pulsation at the maximum of the flooding curve for the toluene - water and toluene - acetone - water systems.

CONCLUSIONS

The results show that for the toluene - water system, maximum throughputs increase with plate spacing, while for the toluene - acetone - water system, this does not always occur, and the influence varies with phase ratio and plate spacing.

These effects increase with initial acetone mass fraction in toluene.

Taking into account the experimental data and the correlations for two-component systems existent in the literature, the correlations (3) and (5) presented here - throughput and frequency of pulsation at the maximum - are more general, due to the inclusion of the effect of column diameter, plate spacing and initial solute mass fraction in the solvent for the two-component system, toluene - water, and for the toluene - acetone - water system, in the presence of mass transfer.

NOMENCLATURE

a

Amplitude of pulsation

c

Initial acetone mass fraction in toluene

D

Column diameter

d_p

Hole diameter of the sieve plates

f

Frequency of pulsation

H_p

Plate spacing

V_c

Continuous phase flow rate

V_d

Dispersed phase flow rate

V_c / V_d

Ratio between the continuous and dispersed phase flow rates

V_d / V_c

Ratio between the dispersed and continuous phase flow rates

V_c + V_d

Column loading

Greek letters

e

Free fractional area of the sieve plates

s

Interfacial tension

Subscripts

c

continuous phase

d

dispersed phase

m

maximum of the flooding curve

p

plate

t

toluene

ACKNOWLEDGEMENTS

The authors are grateful to the Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) for the support which made this work possible.

REFERENCES

Aufderheide, E., Hydrodynamische Untersuchungen an pulsierten Siebbodenextraktionskolonnen, Ph.D. diss., Technischen Universität Clausthal (1985).

Bender, E., Berger, R., Leuckel, W. and Wolf, D., Untersuchungen zur betriebscharakteristik pulsierter Füllkörperkolonnen für die Flüssig/Flüssig-Extraktion, Chem. Ing. Tech., 51, n^o 3, 192-99 (1979).

Berger, R., Leuckel, W. and Wolf, D., Investigations into the Operating Characteristics of Pulsed Sieve Plate Columns, Chem. Ind., 760-64 (1978).

Berger, R., Lecture, Jahrenstrefen der Verfahrens-Ingenieure, Düsseldorf, 1981. In Pilhofer, T. and Schröter, J., Design and Performance of Countercurrent Extraction Columns, Ger. Chem. Eng. J., 1-7 (1986).

Berger, R. and Walter, K., Flooding in Pulsed Sieve Plate Extractors, Chem. Eng. Sci., vol. 40, 2175-84 (1985).

Brandt, H.W., Reissinger, K.H. and Schröter, J., Moderne Flüssig/Flüssig-Extraktoren - Übersicht und Auswahlkriterien, Chem. Ing. Tech., 50, n^o.5, 345-54 (1978).

Cohen, R.M. and Beyer, G.H., Performance of a Pulse Extraction Column, Chem. Eng. Progress, vol. 49, n^o. 6, 279-86 (1953).

Groothuis, H. and Zuiderweg, F.J., Influence of Mass Transfer on Coalescence Drops, Chem. Eng. Sci., 12, 288-89 (1960).

Haverland, H., Vogelpohl, A., Gourdon, C. and Casamatta, G., Simulation of Fluid Dynamics in a Pulsed Sieve Plate Column, Chem. Eng. Technol., 10, 151-57 (1987).

Kleczek, F., Cauwenberg, V. and Van Rompay, P., Effect of mass transfer on droplet size in Liquid-Liquid Dispersions, Chem. Eng. Technol., 12, 395-99 (1989).

Komasawa, I. and Ingham, J., Effect of System Properties on the Performance of Liquid-Liquid Extraction Columns - I: Packed Column, Chem. Eng. Sci., vol. 33, 341-47 (1978a).

Komasawa, I. and Ingham, J., Effect of System Properties on the Performance of Liquid-Liquid Extraction columns - II: Oldshue-Rushton Column, Chem. Eng. Sci., vol. 33, 479-85 (1978b).

Logsdail, D.H., Thornton, J.D. and Pratt, H.R.C., Liquid-Liquid Extraction Part XII: Flooding Rates and Performance Data for a Rotary Disc Contactor, Trans. Instn. Chem. Engnrs., 35, 301-15 (1957).

Lorenz, M., Haverland, H. and Vogelpohl, A., Fluid Dynamics of Pulsed Sieve Plate Extraction Columns, Chem. Eng. Technol., 13, 411-22 (1990).

Misek, T. (Ed.), Standard Test Systems for Liquid Extraction, EFCE Publication Ser. No. 46 (1985). In Lorenz, M., Haverland, H. and Vogelpohl, A., Fluid Dynamics of Pulsed Sieve Plate Extraction Columns, Chem. Eng. Technol., 13, 411-22 (1990).

Palma, M.S.A., Contribuição ao estudo de colunas pulsadas de pratos perfurados: Aplicação à remoção de fenol de soluções aquosas, Master’s thesis, Escola Politécnica da Universidade de São Paulo, São Paulo (1991).

Pilhofer, T., Scale-up of Mechanically Agitated Extraction Columns, Chem. Eng. Technol., 11, 259-63 (1988).

Rouyer, H., Lebouhellec, J., Henry, E. and Michel, P., Present Study of Extraction Pulsed Columns, Proceedings of the International Solvent Extraction Conference - ISEC, 2339-53 (1974).

Sege, G. and Woodfield, F.W., Pulse Column Variables, Chem. Eng. Progress, vol. 50, n^o.8, 396-402 (1954).

Shen, Z.J., Rama Rao, N.V. and Baird, M.H.I., Mass transfer in a Reciprocating Plate Extraction Column - Effects of Mass Transfer Direction and Plate Material, Can. J. Chem. Eng., vol. 63, 29 - 36 (1985).

Tribess,A., Influência do Transporte de Massa e do Espaçamento entre Pratos na Hidrodinâmica e na Eficiência de Separação em Colunas Pulsadas de Pratos Perfurados, Ph.D. diss., Escola Politécnica da Universidade de São Paulo, São Paulo (1995).

Tsouris, C. and Tavlarides, L.L., Mass Transfer Effects on Droplet Phenomena and Extraction

Column Hydrodynamics Revisited, Chem. Eng. Sci., vol. 48, n^o.8, 1503-15 (1993).

* To whom correspondence should be addressed.

To whom correspondence should be addressed.

Publication Dates

• Publication in this collection
  09 Oct 1998
• Date of issue
  Mar 1998

History

• Received
  30 Jan 1997
• Accepted
  20 Oct 1997