Electroextraction: A novel separation technique

Electroextraction: A novel separation technique

~hprjcal [email protected] Science, Vol. 47, No. 12, pp. 3015-3022. 1992. Printed in Great Britain. 0 ELECTROEXTRACTION: A NOVEL TECHNIQUE J. STICHLMAIR...

769KB Sizes 0 Downloads 21 Views


[email protected]

Science, Vol. 47, No.

12, pp. 3015-3022.


Printed in Great Britain.



fiir Therm&he

mx-250!3/92 s5.00 + 0.00 1992 Pergamon Press Ltd




Verfahrenstechnik der Universitit Essen, Fachbereich 13, Postfach 103 764, D-4300 Essen 1, Germany

(Received 16 July 1990, acceptedfor

publication in revisedform

9 December 1991)

Abstract-There exist several processes for mass transfer of charged molecules under the action of an electric field. Well-known examples are electrodialysisand electrophoresis.The paper presents a novel electric mass transfer process where the mass transfer takes place across the interface of two liquid phases. The principles of mass transfer under the action of an electric field and the criteria for choosing the second liquid phase are developed. Feasible fields of application and the design of a batch-flow and a continousflow apparatus are given.








mental technologies in chemical engineering has generated new separation tasks, which cannot often be addressed by conventional methods. A great majority of relevant substances are charged, either as ions in a solution or as particles in a dispersion. In theory, charged substances can be separated electrokinetically, using the physical effect of migration of charged particles under the action of an electric field. This material transport by migration has been used in the past to carry out several different separations. 2. HISTORICALDEVELOPMENT Reuss [ 1J was the first to describe the migration of clay particles in an electric field. Hittorf [2] observed concentration changes in a small tube on passage of an electric current. Figure 1 shows the Hittorf experiment which was carried out with silver electrodes immersed in a silver nitrate solution. He observed that, due to the ion migration, the concentration profile changed from a constant value. The solution concentration near the anode increased while the solution near the cathode became more dilute. The concentration in the middle remained constant. Since the mass of silver that dissolves at the anode is equal to the silver deposited at the cathode, the average silver nitrate concentration in the solution remains constant and the shaded areas in Fig. 1 are equal. Hittorfdiscovered that this experiment worked only in small tubes and at low current densities. In the case of larger currents, the Joule heating generated produces density differences in the solution, and the consequent convection currents cause the solution to re-mix, thus nullifying the possible concentration differences. The solution thus shows a constant concentration. Arrhenius [3] then postulated the existence of charged particles in a solution and Kohlrausch [4] determined the velocities of some ions in an electric field. The first investigations that used this electro-

phoretic migration to carry out a separation were done by Picton and Lindner [S, 61 using relatively simple colloidal systems about one hundred years ago. More extensive work at the beginning of the century by several other investigators established the potential of electrophoresis for the separation and characterization of charged materials. In the late thirties, an innovative “U-tube” apparatus, shown in Fig. 2, was described by Tiselius [77 and this opened a new era for preparative and analytical electrophoresis. In a typical use of the Tiselius apparatus, a mixture of proteins to be separated and a solution are put into the U-tube and covered by layers of electrolytic solutions with varying densities. The density gradient is such that the protein mixture remains at the bottom of the U-tube. Ionic species in the feed mixture will move under the action of an electric field towards the cathode or the anode depending on their charge. Species with the highest mobility (i.e. charge) will move faster towards the electrodes into the electrolytic solution, and thus be separated from the feed mixture. The cell is stabilized against convective disturbances by a density gradient and is often used to measure electrophoretic mobilities and protein interactions. On the other hand, the effectiveness of this system as a preparative tool is restricted to small power inputs (0.115 W/cm3) since its stability against convective re-mixing disappears at higher inputs in spite of cooling provisions.

3. CONTINUOUS FREEFLOW ELECTROPHORESIS Efforts in the development of preparative electrophoresis led to a truly continuous device that worked without a need for stabilization in paper, powders and gels. This free-flow technique was introduced in 1949 by Hanning and Grassmann [8] and is illustrated in Fig. 3. An electrolyte solution flows perpendicularly to the lines of force of the electric field. The mixture to be separated is fed at a single point into the electrolyte



3016 +


Cathode portion _- N Middle portion __ M Anode portion Concentration



Fig. 1. The Hittorf experiment[2]_




Fig. 2. TiseIius apparatus: (a) electrode compartments, (6) U-tube for separation [7].

such a chamber, the liquid is vigorously heated by the effect of the current flow between the electrodes. Local temperature differences can arise that induce uncontrolled thermal convection. These convective currents are much stronger than the electrophoretic transport and can destroy the separation. To ensure that the separation is not significantly hindered or disrupted, total laminar-flow conditions have to be approached. This in turn necessitates a very thin chamber ( -Z 1 mm thick) with intensive cooling and to carry out the separation at very low velocities or flows. Several efforts have been made in the past to stabilize the solution in an electrophoretic chamber against convection, e.g. the use of a density gradient produced by an arrangement in layers of different saccharose solutions of varying concentrations. But, in general, it is not possible to sensibly handle such a solution in a continuous apparatus on any significant scale. One potential solution to the convection problem would be to carry out electrophoresis in space under weightless conditions. Experiments of this type were carried out in space missions from Apollo 5 in 1971 to the Spacelab in 1985 and in nearly all Space Shuttle missions. All these restrictions to the implementation of this highly selective separation technique have meant that it can only be used successfully on a very small scale and for analytical purposes only. It has not been exploited on commercial production scales. Electrodialysis is a large-scale separation technique that uses Coulombic forces that act on charged particles in an applied electric field and is somewhat similar to electrophoresis. The disruptive effects of convection are avoided in electrodialysis by the presence of a membrane across which the separation occurs. The selectivity of such a process can be altered by the use of anion and cation exchange membranes. Many of the typical problems associated with membrane technology also apply to electrodialysis, namely fouling, large electrical resistances, and permeation limitations for large molecules (molecular weight no larger than 400 kg/kmol). This technique is only applicable to some specialized separations and requires significant capital investment. 4.

$_ + a


Fig. 3. Principle of free-flow electrophoresis. The fourcomponent mixture (abed) is separated. stream. Due to their electrophoretic mobilities, different species will move down the separation chamber along different paths and can be collected continuously at different points at the end of the chamber. In




A new separation technique, namely electroextraction has been developed at the Institute of Chemical Engineering at the University of Essen [9]. Electroextraction is similar to electrophoresis and electrodialysis on the one hand, and to liquid-liquid extraction on the other. The idea is based on the use of two or more liquid phases within the separation device to nullify the harmful effects of convection. One of the phases contains the mixture to be separated and the other acts as a solvent to remove the components separated. The state of the art for liquid-liquid extraction is fairly advanced and several excellent reviews are available in the literature [l&16]. In a two-phase system like the one shown in Fig. 4, there is one electrode located in each phase. An elec-



Fig. 4. Multiple-phase system for electroextraction [lo]. Charged species are transported across the interface and convective currents are constrained by the interfaces.

tric field is imposed on the system and charged particles tend to move towards the electrodes. In their motion towards an electrode, charged particles can cross a phase boundary, thus effecting a separation. It is possible to separate positive cations from negative anions into the separate phases since the thermal currents that generate convective mixing are constrained to a single phase and do not “cross” the interface. The separation process is no longer hindered by convective mixing and can even be helped by the convective transport of material to the region close to the interface. In a three-phase system, where the electrodes are located in the upper phase and in the lower phase, it is possible to remove anions and dations from the middle phase. If a positively charged species were transported into one phase and a negative species transported into the other phase without charge compensation, an electrical field would be induced that would neutralize the applied external field. There would thus result no net potential gradient in the solution. To avoid this, electroneutrality has to be ensured in every volume element. The charge of the substances transported has to be compensated in this case by reactions at the electrodes. Due to the standard electrode potentials of the substances in aqueous solutions, the electrode reactions at the anode are described in many cases by 20H - +&02 and at the cathode

t + 2e- + Hz0



H + e- d*Hzt.


The gases produced by electrolysis have to be collected and continuously removed from the apparatus. These reactions also have an effect on the pH value of

the solutions, which is of interest for applications to the separation of amphoteric substances. In some cases, the solutes may take part in the electrode reactions. Depending on the nature of the process, this effect could be desirable or harmful. Nevertheless, the primary aim of electroextraction is the separation from one phase into another and not the electrochemical reactions of the substances to be separated. The effectiveness of electroextraction has been demonstrated and tested in comprehensive trials for several systems. Figure 5 illustrates an easy experiment that was carried out with a coloured dye (fuchsine acid) as the transported material. The U-tubes shown schematically in Fig. 5 are loaded with a mixture of butanol and fuchsine acid. Tubes b and c are additionally loaded with water to form two liquid phases. Transport of the solute between the phases in tube b occurs by natural diffusion only. Tube a and c are under the effect of an electric field, so that transport of the solute within the tubes is promoted by the action of the applied field. Tube a contains a homogeneous mixture of dye and butanol and Fig. 5 illustrates how, after 30 min, no separation of the dye is achieved near the electrodes because of convective disruptions. Tube b shows that diffusive separation across the interface is extremely slow and no mass transfer can be observed. Tube c shows that the presence of the interface nullifies the effects of convective mixing and that a separation across the interface is performed. The fuchsine acid is transported towards the anode and in the absence of




Begin t = 0 min

End t=30min




Fig. 5. Fundamental mechanism of electroextraction. (Schematicdemonstrationfor the transport of fuchsineacid across the interface of a butanol-water system, O/x = 50 V/cm.) (a) Fuchsineacid in a mixture of butanolwater, no liquid-phase boundary. (b) Two-phase system butanol-water. The light phase on the left side is loaded with fuchsineacid. No remarkablemass trantier takes place. (c) Two-phase systembutanol-water. The light phase on the left side is loaded with fuchsine acid at the beginning of the experiment. Due to the electric field, the fuchsine acid migratesinto the water phase.



convective disruptions, the butanol phase finds itself rather free of the dye after 30 min. Comparisons between conventional liquid-liquid extraction and electroextraction are in order to establish the potential advantages of electroextraction in large applications. Experiments on the separation of the dye from water with butanol as the solvent were carried out using fairly small dye concentrations in the water feed (0.25 g/l) by conventional extraction and by electroextraction. The electroextraction experiments were carried out in a batch apparatus. The heavy phase consisted of 0.11 of water saturated with butanol and with the fuchsine-acid concentration indicated. The top phase was 0.1 1 of n-butanol saturated with water. The liquid-liquid extraction experiments were carried out in cross-flow. fashion by the same bottom phase (0.1 1) 12 times with fresh organic phase (0.1 1 each time). Each time the phases were put in contact, equilibrium was achi&ed, and the phases separated. Experiments for electroextraction were carried out in one step, but with the action of an electric field. The results of these experiments are shown in Fig. 6. The ordinate is the fuchsine-acid concentration in the bottom phase relative to the primary concentration. On the abscissa, the time of the experiment is given in minutes. The parameter is the strength of the electric field. The lower figure shows the result of the conventional cross-flow liquid-liquid extraction. The ordinate is again the fuchsine-acid concentration ratio and the abscissa the number of extraction steps needed to achieve a certain concentration ratio. Figure 6 shows how, for the case of liquid-liquid extraction, 1.2 1 of solvent are required to separate 52% of the dye, while with electroextraction it is possible to separate over 90% with only 0.1 1 of solvent. The time required depends on the strength of the field and the electrical work is a function of the distance between the electrodes and the conductivity of the phases. 5,


X. net = X.mig + X.aXf + ~~-.GcB”v.


If the material balance of eq. (3) is done only across the phase boundary, eq. (3) can be rewritten in a onedimensional form since only the flux component perpendicular to the boundary (in x-direction) leads to a concentration change in either phase.

= JLi,

FieM strength U/s (V/cm)





Time t (min) -

s .is +



8 0

0.2 -Co - 0.25 gil fuchsin acid in heavy phase Volume retio 1:l per step










, 1

steps n -

Fig. 6. Separation of fuchsine acid (0.25 g/l) from water with n-butanol as solvent. The dye is transportedfrom the heavy into the light phase. (Electroextraction at the top and conventional cross-flow liquid-liquid extraction at the bottom.)


To acquire a deeper understanding of the concept of continuous electroextraction, it is necessary to examine the total mass flux equation that includes transport due to the migration of charged particles in an electric field, transport by conventional diffusion due to concentration gradients, and transport through convection as shown in eq. (3).


et al.

+ Jtm

+ Jf&nv.


If the assumption is made that convection cannot take place across the phase boundary, eq. (4) further sim-

plifies to eq. (5). JL,



+ J;dirr.


The expressions for Jtmi, and Jcdiffcan be derived using chemical-potential considerations. The fundamental equation for the motion in the x-direction of species i due to differences in chemical potential is w.=-_kti I


’ dx *

In general, the chemical potential depends on the activity of i and on the electrical state of the system, i.e. Pi




The dependence on the activity is given by /A{= ~2.i + RTln





Equation (19) now simplifies to

with the activity itself expressed as (9)

a, = yici



with yi = 1 for Ci + 0. Equations (8) and (9) lead to /.J$= P(.9i+ RTln(Ciyi).


On the other hand, the dependence of the chemical potential on the electrical state of the system can be expressed by


The use of the definition of the diffusion coefficient, which is given in eqs (21) and (22), leads to eq. (23), i.e. D ..r,=Dn( with D,,

To transform eq. (11) so that it uses the term d&dx needed in eq. (6), one can write the following expression by simple differentiation:

(12) The partial derivatives ( api/aCi),, y, and ( &.+/ayi)~_c, can be obtained by differentiation of eq. (10).

(z > aPi

= & *a,.,Yt

af.4 (-> @J



LP2.i + RTln (Y&l

(13) I







= 5



Substitutingeqs(13),(14)and(ll)intoeq.(12)leadsto RTdy. dPi -~~+->+ziF~_ dx



Yi dx

Inserting eq. 15 into eq. 6 we get


w; =

RT dci


’ ci




’ yi dx

(16) To obtain the flux, it is necessary to multiply each side of.eq. (16) by the concentration ci, i.e. (17)

J; = wfci Jf=





k.c.2 I I







Equation (18) can be cast in the form (with the insertion of dcJdc, in the second term on the RHS) J+




k_RT I

&i + dX



dyi dci dx dc,

(19) EES 47-12-c

= kiRT dc.




D,, % dx



A comparison between eq. (23) and eq. (5) makes it clear that the first term represents the diffusive transport and the second the migration in the field. Since the balance boundary was located at the phase boundary, it is sensible to use overall transport coefficients in the first term of eq. (23) and refer them to concentration differences across the interface. J;=





DT1ciziF RT

[email protected], dx


with C;* = NC;.

= y.




The study of eq. (24) illustrates that mass transfer in the direction of the potential gradient depends on mobility, concentration, valency, and temperature. Mass transfer in either direction depends on the concentration gradient generated within the phases by the migration flux. The concentration gradient in a phase induced by the migration flux limits the separation in a batch device even at very high field strengths. It is obvious then that a continuous countercurrent device is required to optimize the process and get away from gradient limitations. Figure 7 shows a schematic drawing of a continuous electroextraction apparatus with countercurrent flow of the phases. The two phases flow co-currently within each stage, but the flow in the apparatus as a whole is countercurrent. The heavy phase is fed at the top of the device and leaves from the bottom. The lighter phase is fed at the bottom and leaves the top in the same way as in conventional liquid-liquid extraction. The height of the phase boundary in each stage is fixed by an overflow weir. A provision for cooling has to be present in order to maintain a suitable operating temperature. Figure 8 shows a flowsheet of an experimental continuous electroextraction apparatus installed in OUT laboratory. Cooling at each stage is achieved by the flow of a coolant within the stage decks themselves. This apparatus has a total of 10 stages and a total volume of 1.8 1. The redox electrodes are made of platinum.


J. STICHLMAIR et al. Heavy phase

Fig. 7. Continuous electrocxtractionapparatus.Distance betweenthe electrodes(= 1 cm); liquid surface area on one stage (= 0.0145 mZ).


I .or

(1) Electrosxtractioncolumn (2) Light-phass feed tank (3) Heavy-phasefeed tank (4) Light-phase receiver tank (5) Heavy-phase redver tank (6) Power supply

0.06 Fig. 8. Flowsheet of a laboratory continuouselectroextraction apparatus.(Cooling provisionfor the column has been omitted for clarity.)

Several experimental tests have indicated that the design of the continuous apparatus works very effectively. Figure 9 shows some results of ekctroextraction

for the recovery of citric acid from water with IIbutanol as solvent. The results indicate that a nearly complete removal is achieved with a volume ratio of unity between solvent and feed. Results from [12] show that the removal of citric acid from an aqueous solution by high-performance extraction (complex formation) with ternary amines in different solvents is much less effective (with the same volume ratio of solvent) and entails a high expenditure on chemicals. 6. LIQUID-LIQUID SYSTEMSFOR ELECTROEXTRACI’ION

The phases required for electroextraction should, in general, have some special characteristics. Some of the properties required for conventional liquid-liquid extraction are also here desirable, such as sufficient density difference to allow phase separation, adequate interfacial tension to present a well-defined interface, and low viscosities to enhance mass transfer and to allow efficient gas removal. The special characteristics

cbO= 0 mol/l citric acid in light phase

50 ’

0 100 ’

’ 150 ’

c, (molll) -4 0.1 0 0.05 A 0.025 * 0.01

’ 200 ’

’ 250 L

Field strength U/s (V/cm) Fig. 9. Recovery of citric acid from water by continuous elcctroextractionwith n-butanol.

of electroextraction make additional phase properties indispensable, such as good electrical conductivity. This property is easily achieved when the phases have a high water content or if other chemicals are added. Binary mixtures of water and an organic solvent can be used in cases when adequate solubility of water in the solvent makes the organic phase conductive. Examples of organic solvents that can be used are nbutanol and 4-chloro-l-butanol. Another type of system that can be used is a ternary system where water distributes in both phases, such as in the case of glycerine-acetone-water. It is also possible to form two-phase systems where water is the main constituent of both phases. Highly polymerized substances such as polyethylene glycol and/or dextran can be added to water and two phases will ensue. This type of system is commonly used in extractions of biotechnology products [13]. Still a fourth kind of system is the one formed by water and a non-ionic surfactant (non-ionic alkylphenyl polyether alcohols) that distributes between the water and a co-acervate phase. This auxiliary phase contains a larger amount of surfactant, e.g.



Electroextraction can be utilized in many separations, particularly those from dilute systems. The design of the apparatus is determined by the two-phase system selected, the electrode reactions, and by the substances to be separated. The required energy input depends strongly, among others, on the composition and conductivity of the solutions. Laminar-flow conditions persist in the apparatus described so that scale-up is possible and the maximum size of a stage is limited by the cooling capacity of the apparatus.

Two-phase region I


81 .O 0.1 ,

0.2 I

0.3 I

0.4 I

0.5 I

0.6 I

Mass fraction x (Triton X-l 14) (kg/kg) Fig. 10. Binode of the system Triton X114-water.

“i =i

c: C!’1

cc* 10 wt% at 4O”C, and the original water phase in equilibrium with the co-acervate phase contains only a few ppm of surfactant. Examples of such chemicals are Triton X114 (Rohm and Haas) or Marlophen 87 (Hiils AG). The binode of the system Triton X114-water is shown in Fig. 10. The possibility of the existence of chemical and especially electrochemical reactions between the solvent and the solutes has to be accounted for in the selection of a solvent as well.

di F J; Ki

ki Ni






The technique is generally applicable to the separation of materials that are charged in solution. Comprehensive experiments in our laboratories have shown that the separations of organic and inorganic acids (citric and sulphuric acid) are possible and that the conversion of inorganic and organic salts (salts of sulphonic acids) into their acids and bases can also be done effectively by electroextraction. This salt-splitting technique suggests that separations of substances that have a low degree of dissociation in solution can be separated by electroextraction if they are converted previously to highly dissociated species. For example, an organic acid can be converted first with a base into a highly dissociated salt. From this point, the separation of the acid residue is possible by electroextraction. Examples of this are the separation of lactic acid via sodium lactate or acetic acid via its sodium salt. If amphoteric molecules are to be separated by this technique, special precautions have to be made to maintain constant pH values in the phases. The main provision in such cases is to separate the electrodes from the actual separating device by membranes as in free-flow electrophoresis. Electroextraction is basically applicable to the separation of proteins, enzymes and other amphoteric substances, as long as they are charged in solution.

x Z


letters coefficient of component i, dimensionless standard chemical potential of component i, kJ/kmol chemical potential of component i, kJ/kmol electrical potential, V activity






transport velocity of componen’t i in x-direction, m/s length, m valency, dimensionless



activity of component i, kmol/m3 concentration of component i, kmol/m3 concentration of component i in phase’, kmol/m3 concentration of component i in phase”, kmol/m” equilibrium value of cf. kmol/m’ diffusion coefficient of component i, m2/s Faraday constant 94.49 x 10P6, A s/km01 flux of component i in x-direction, kmol/(m’ s) overall mass transfer coefficient of component i, m/s proportional constant of component i, m2 kmol/(kJ s) Nernst distribution coefficient of component i, dimensionless universal gas constant, kJ/(kmol K)



Reuss, F. F., 1809, Memoires de la Societe Imperiate des Naturalistes de Moscou 2,324-347. PI Hittorf, J. W., 1853, 1856, 1858,1859, Poggendmfs Ann. 89,177; 98, 1; 103, 1; 106,337-513. c31 Arrhenius, S., 1887, 2. physik. Chem. 1, 631. c41 Kohlrausch, F., 1876, Gatt. Nachr. 213. c51 Picton, H. and Lindner, S. E.. 1892, J. &em. Sot. 61, 148. WI Picton, H. and Lindner, S. E., 1897, J. them. Sot. 71, 568. c71 Tiselius, A., 1937. Trans. Famduy Sot. 33, 524-535. C81 Harming, K. and Grassmann, W., 1949, DBP 805399, 24 May. t9-lStichlmair, J., 1987, DBP 3742292, 14 December. Stichlmair, J. and Schmidt, J., 1989, Stofftrennung WI durch Zweiphasen-Elektrophorese. Vortrag auf dem Jahrestreffen der Verfahrensingenieure, Hannover.

3022 [ll]


Stichlmair,J. and Schmidt, J., 1989. Zweiphasen-Elektrophorese.Techn. Mitt. 82 6, 397. [f2] Wenerten, R, 1983, The extraction of citric acid. J. chent.Tech. Biotechnol. 338, 85. [13] Kula, M.-R., Kroner, K. H. and Hustedt, H., 1984, Bioengng 24, 1015. [14] Miiller, E.. 1972, FlGssigljlGssig-Extraktion: UIlmanns



Enzykloptiie der technischen Chemie, 4. Au& Bd. 2. Verlag Chemie, Weinheim. [lS] Treybal, R. E., 1963, LiquidExtraction. McGraw-Hill, New York. [16] Perry, J. H. and Chilton, H. C., 1973, Chemical Engineers Handbook, 5th Edition. McGraw-Hill (Kogakusha), New York, Tokyo.