Electrodialytic Membrane Processes and their Practical Application

495

ELECTRODIALYTIC MEMBRANE PROCESSES AND THEIR PRACTICAL APPLICATION H. Strathmann, University of Twente, Faculty of Chemical Technology, Enschede, The Netherlands INTRODUCTION The desalination of brackish water by electrodialysis and the electrolytic production of chlorine and caustic soda are the two most popular processes using ion-exchange membranes. There are, however, many other processes such as diffusion dialysis, Donnan dialysis, electrodialytic water dissociation, etc. which are rapidly gaining commercial and technical relevance. Furthermore ion-exchange membranes are vital elements in many energy storage and conversion systems such as batteries and fuel cells. Although the large scale industrial utilisation of ion-exchange membranes began only 20 years ago, their principle has been known for about100 years [I]. Beginning with the work of Ostwald in 1890, who discovered the existence of a "membrane potential" at the boundary between a semipermeable membrane and the solution as a consequence of the difference in concentration. In 191 1 Donnan [2] developed a mathematical equation describing the concentration equilibrium. The first use of electrodialysis in mass separation dates back to 1903, when Morse and Pierce [3] introduced electrodes into two solutions separated by a dialysis membrane and found that electrolytes could be removed more rapidly from a feed solution with the application of an electrical potential. With the advent of ion-selective membranes it became feasible to transport ions against a concentration gradient. In 1940 Meyer and StrauB suggested a multicell arrangement, in which anion-selective and cation-selective membranes were installed in alternating series between two electrodes [4]. With such multi compartment electrodialysers, demineralisation or concentration of solutions could be achieved in many compartments with only one pair of electrodes. Thus the irreversible energy losses due to the decomposition potentials at the electrodes could be distributed over many demineralizing compartments and thus be minimised. With the development of highly selective ion-exchange membranes of low electric resistance in the late 40's by Juda and McRae [S] electrodialysis rapidly became an industrial process. The main use envisaged for ion-exchange membranes was in electrodialysis for the desalination of brackish water. In Japan electrodialysis was used also for concentrating sodium chloride from sea water to produce table salts [ 6 ] . A significant step towards the efficient application of electrodialysis was the introduction of a new operating mode referred to as electrodialysis reversal by Ionics. In this operation mode the flow streams and the polarity in an electrodialysis stack is reversed in certain time intervals [7] and membrane fouling and scaling can be reduced to a minimum.

496

With the development of a chemically extremely stable cation-exchange membrane based on sulfonated poly-tetra-fluor-ethyleneby Du Pont in the late 60 's the chlor-alkali ne membrane electrolysers were introduced [81. The practical use of bipolar membranes for the recovery of acids and bases from the corresponding salts by electrodialytic water dissociation in the early 80's by Liu et al. [9] opened a multitude of new applications in chemical industry and in waste water treatment. The combination of electrodialysis with conventional ion-exchange technology and the use of conducting spacers are both commercially and technically very interesting variations of the basic process [ 101. Diffusion dialysis through anion exchange membranes is used today on large scale to recover acids from pickling solution. Donnan dialysis is used for water softening or for recovering organic acids from their salts. PROPERTIES AND STRUCTURES OF ION-EXCHANGE MEMBRANES The most important parts in any electrodialitic process are the ion-exchange membranes. Their properties determine to a very large extent the technical feasibility and the economics of the processes. Therefore some fundamentals concerning the properties and structures of ion-exchange membranes shall be discussed. 1. Properties of ion-exchange membranes Ion-exchange membranes are ion-exchange resins in film form. They consist of highly swollen gels carrying futed positive or negative charges. There are two different types of ionexchange membranes: (1) cation-exchange membranes which contain negatively charged groups fixed to the polymer matrix, and (2) anion-exchange membranes which contain positively charged groups fixed to the polymer matrix.

negative fixed ion

Fig. 1:

~

negative co-ion

@positive counter-ion

Schematic diagram illustrating the structure of a cation-exchange membrane and the distribution of fixed and mobile ions in membrane matrix

In a cation-exchange membrane, the fixed anions are in electrical equilibrium with mobile

491

cations in the interstices of the polymer, as indicated in Figure 1, which shows schematically the matrix of a cation-exchange membrane with fixed anions and mobile cations, the latter are referred to as counter-ions. In contrast, the mobile anions, called co-ions, are more or less completely excluded from the poylmer mamx because of their electrical charge which is identical to that of the fixed ions. This type of exclusion is called Donnan-exclusion in honor of his pioneering work [2]. Due to the exclusion of the co-ions, a cation-exchange membrane permits transfer of cations only. Anion-exchange membranes carry positive charges fixed on the polymer matrix. Therefore, they exclude all cations and are permeable to anions only. Thus the selectivity of ion-exchange membranes results from the exclusion of co-ions from the membrane phase. For a cation-exchange membrane in a dilute solution of a strong electrolyte, the concentration of the cations is generally higher in the membrane than in the solution, because the cations are attracted by the negatively charged fixed ions of the cationexchange membrane. On the other hand, the concentration of mobile anions is higher in the solution than in the ion-exchange membrane. Thus concentration gradients are established between the membrane and the solution. These gradients act as driving forces for the mobile cations to move into the solution and the mobile anions to move into the membrane. Because electroneutrality is required the permeation of cations into the solution and of anions into the cation-exchange membrane leads to a counteracting space charge due to uncompensated ionsand an equilibrium is established between the attempt of diffusion on one side and the establishment of an electrical potential difference on the other. This electrical potential difference between an ion-exchange membrane and an adjacent salt solution is referred to as Donnan potential. It can be calculated but not be measured directly. The electrical potential gradient and the concentration profiles of mobile cations and anions in a cation-exchange membrane and a dilute adjacent salt solution is illustrated in Figure 2. The thermodynamical treatment by Donnan and Guggenheim [ l l ] is based on an equilibrium between the membrane phase, indicated by superscript M, and the outer phase, indicated by superscript 0 of the electrochemical potential Tli of all ions which are able to permeate through the

In both of the adjacent phases, the concentrations as well as the osmotic pressure and the electrical potential can be different. For the distribution of a specific ion, an established electrical potential difference (PM - (PO, the Donnan potential A q ~ ~ ~ be c adescribed n as a function of the different activities ai and the swelling pressure Ki [l8]:

where Zi is the valency of the ion species i, F the Faraday's constant, R the gas constant, T the absolute temperature, Vi the partial molar volume of the component i, and Xi the swelling

498 pressure. The numerical value of A ~ is negative D ~ for ~ the cation-exchange membrane and positive for the anion-exchange membrane. The Donnan potential cannot be determined by direct measurement, however, it can be used for the calculation of the distribution of the mobile ions between the solution and the membrane and hence for the determination of the membrane permselectivity

.

Membrane

Solution

)Qoo I

10 1

Q

8

0 Mobile Anions Q Mobile Cations - Fixed Charges

%

-Fixed

Charges Concentration Cation Concentration 0 Salt Concentration /

-Anion

concentration

Directional Coordinate

b

L irectiona oordinate

- t

Fig. 2:

Schematic diagram illustrating the electrical potential gradient and the concentration profiles of mobile cations and anions in a cation-exchange membrane and a dilute adjacent salt solution

The swelling pressure is proportional to the concentration of the fixed ions and inversely proportional to the concentration of the electrolyte. In ion-exchange membranes with high fixed ion concentrations and dilute solutions it can reach very high values well in excess of 100 bars [13]. The relationship between the electrolyte concentration in the solution and that in the membrane can be obtained by assuming chemical equilibrium between the two phases and electro neutrality in both the solution and the membrane. For negligible small pressure

499

differences between the membrane phase and the outer solution, and for membranes with a high fixed charge density compared to the salt concentration in the outside solution the coion concentration in the membrane for a monovalent salt, such as sodium chlorine, can be calculated to a first approximation by [ 141:

where Cco is the co-ion concentration, and COthe electrolyte concentration, Cfixed is the concentration of the fixed ions in the membrane and are the average activity and coefficients of the salt in the solution and in the membrane, respectively. This fundamental equation is based on the theories of Teorell and Meyer and Sievers [15,16]. However, the more complex structure of modem membranes cannot be adequately described exclusively by this theory [12]. The remaining differences in the observed and expected membrane behaviour are mainly due to a non-uniformity in the distributions of molecular components in the membrane. This results from structural irregularities on a molecular level and from the influence of the elecmc field. Additionally, the practical application of thermodynamics is rather limited by the difficulties in the experimental rneasm'ment of independent interaction, diffusion, resistance, and frictional coefficients. The Donnan exclusion equilibrium and thus the membrane selectivity depend on: (1) the concentration of the fixed ions, (2) the valence of the co-ions, (3) the valence of the counterions, (4) the concentration of the electrolyte solution, and (5)the affinity of the exchanger with respect to the counter-ions. Important parameters for the characterization of ion-exchange membranes are the density of the polymer network, hydrophobic and hydrophilic properties of the matrix polymer, the distribution of the charge density, and the morphology of the membrane itself. All these parameters do not only determine the mechanical properties, but they also have a considerable influence on the sorption of the electrolytes and the non-electrolytes and therefore on the swelling [131. The most desired properties for ion-exchange membranes are: High permselectivity - an ion-exchange membrane should be highly permeable to counter-ions, but should be impermeable to co-ions. Low elecmcal resistance - the permeability of an ion-exchange membrane for the counter-ions under the driving force of an electrical potential gradient should be as high as possible. Good mechanical and form stability - the membrane should be mechanically strong and should have a low degree of swelling or shrinking in transition from dilute to concentrated ionic solutions. High chemical stability - the membrane should be stable over a pH-range from 0 to 14 and in the presence of oxidizing agents.

It is difficult to optimize the properties of ion-exchange membranes because the parameters determining the different properties often have opposing effects. For instance, a high degree of crosslinking improves the mechanical strength of the membrane but also increases its electrical resistance. A high concentration of fixed ionic charges in the membrane matrix leads to a low electric resistance but, in general, causes a high degree of swelling combined with poor mechanical stability. The properties of ion-exchange membranes are determined by two parameters, namely, the basic polymer matrix and the type and concentration of the fixed ionic moiety. The basic polymer matrix determines to a large extent the mechanical, chemical and thermal stability of the membrane. Very often the matrix of an ion-exchange membrane consists of hydrophobic polymers such as polystyrene, polyethylene or polysulfone. Although these basic polymers are insoluble in water and show a low degree of swelling, they may become water soluble by the introduction of the ionic moieties. Therefore, the polymer matrix of ion-exchange membranes is very often cross linked. The degree of cross linking then determines to a large extent the degree of swelling, and the chemical and thermal stability, but it also has a large effect on the electrical resistance and the permselectivity of the membrane. The type and the concentration of the fixed ionic charges determine the permselectivity and the electrical resistance of the membrane, but they also have a significant effect on the mechanical properties of the membrane. The degree of swelling, especially, is affected by the concentration of the fixed charges. The following moieties are used as fixed charges in cation-exchange membranes:

In anion-exchangemembranes fixed charges may be:

These different ionic groups have significant effects on the selectivity and electrical resistance of the ion-exchange membrane. The sulfonic acid group, e.g., - S q is completely dissociated over nearly the entire pH-range, while the carboxylic acid group -COO-, is virtually undissociated in the pH-range < 3. The quaternary ammonium group -R3N+ again is completely dissociated over the entire pH-range, while the primary ammonium group -NH3+ is only weakly dissociated. Accordingly, ion-exchange membranes are referred to as being weakly or strongly acidic or basic in character. Most commercially available ionexchange membranes have -SO3- or -COO- groups, and most anion-exchange membranes contain -R3N+ groups [ 121.

2. Structures of ion-exchange membranes The structures of ion-exchange membranes are closely related to those of ion-exchange

50 1

resins. As with resins, there are many possible types with different polymer matrixes and different functional groups to confer ion-exchange properties on the product. Most commercial ion-exchange membranes can be divided, according to their structure and preparation procedure, into two major categories, either homogeneous or heterogeneous membranes. Homogeneous ion-exchange membranes are produced either by polymerization of functional monomers, e.g., by means of polycondensation of phenolsulfonic acid with formaldehyde [ 171, or by functionalizing of a polymer film by sulfonation [18]. The polycondensation with formaldehyde according to the following reaction scheme was one of the first procedures for making cation-exchange membranes : OH

Phenol is treated with concentrated H2SO4 at elevated temperatures which leads to the phenolsulfonic acid in para form. This acid is reacted with a solution of formaldehyde in water. The solution is then cast into a film which forms the membrane after cooling to room temperature. Excess monomer is removed by washing the film in water. The polymerizing styrene and divinyl benzene and its subsequent sulfonation according to the following reaction scheme [ 191 is widely used for the preparation of commercial cationexchange membranes:

For the preparation of anion-exchange membranes positively charged groups are introduced into the polystyrene by chloromethylation and amination with mamine according to the following reaction scheme:

502

There exist numerous references in the literature for the preparation of ion-exchange membranes by polymerization [19, 20, 211. In recent years, membranes based on perfluorocarbon-polymershave proved to be very useful in the chlor-alkaline industry [22]. They were introduced fist by DuPont as Nafion@ [23] and are prepared according to the following reaction scheme in a several-stepprocedure:

This intermediate is reacted with hexafluoropropylene oxide to produce a sulfonyl fluoride adduct.

By heating with sodium carbonate the sulfonyl fluoride vinyl ether is formed which is then copolymerized with TFE.

*

-(CF2-CF,)n-

c-

0

F-CF2-

0

( Cq

- CF- 0 a-

I

II

- CF;- S - F

CF3

The resulting copolymer is extruded as a film and finally the ionogenic moiety is converted

503 to membranes which carry sulfone-groups as the charged moieties by reacting the -S02F groups with sodium hydroxide. The lifetimes membranes under the aggressive conditions of chlor-alkaline process are in the range of 3 years. Homogeneous ion-exchange membranes can also be prepared by the introduction of anionic or cationic moieties into a performed film. Starting with a film makes the membrane preparation rather easy. A typical example for this mode of preparing ionexchange membranes is the sulfochlorination and amination of polyethylene sheets according to the following reaction scheme [24]:

-

-

+ so, + CI

7 +

2NaOH

SQa

- HGI

- NaCI, - HO ,

S W

sQ- Na'

and

+ H,N-CI-$-N"' SQCI

L

'CH3

7 SQ-NH-CHyN H 3 d 'c%

Heterogeneous ion-exchange membranes consist of fine colloidal ion-exchange particles embedded in an inert binder such as polyethylene, phenol resins, or polyvinyl-chloride. Such membranes can be prepared simply by calandering ion-exchange particles into an inert plastic film [13].Also, ion-exchange particles can be dispersed in a solution containing a film-forming binder, and then the solvent is evaporated to give the ion-exchange membrane. Heterogeneous membranes with useful low elecmcal resistances contain more than 65% by weight of the cross-linked ion-exchange particles. Since these ion-exchange particles swell when immersed in water, it has been difficult to achieve adequate mechanical strength and freedom from distortion combined with low elecmcal resistance. In general, heterogeneous ion-exchange membranes have relatively high electrical resistances. Homogeneous ionexchange membranes have a more even distribution of fixed ions and often lower electrical resistances.

504

3. Special h p e r t y Ion-exchange Membranes In the literature, there are numerous methods reported for the preparation of ion-exchange membranes with special properties, e.g., to be used for the production of table salt, as battery separators, as ion-selective electrodes, or in diffusion and Donnan dialysis. Significant effort has also been concentrated on the development of anion-exchange membranes with low fouling tendencies. 3.1 Monovalent ion permselective membranes For the production of table salt by concentration of sea water monovalent cation selective membranes were prepared by forming a thin positively charged layer on the surface of a cation-exchange membrane. Monovalent anion permselective membranes have a thin highly cross-linked layer on the membrane surface have also been developed [25]. By such means the selectivity of sulfate compared to the one of chloride can be reduced from about 0.5 to about 0.01 and of magnesium compared to sodium from about 1.2 to about 0.1.

3 . 2 Anion-exchange membranes of high proton retention By means of traditional membranes, it is not possible to apply electrodialysis in the recovery of acid in order to reuse the acid because of high proton leakage through the anion-exchange membranes. In general, since protons permeate easily through an anion-exchange membrane, acids can not be concentrated to more than a certain level by electrodialysis with high efficiency. Recently developed membranes, however, exhibit low proton permeabilities and enable efficient acid concentration [25]. 3.3 Anti-fouling anion-exchangemembranes The anion-exchange membrane is more sensitive to fouling. The permeability of commercial anion-exchange membranes is limited in practical electrodialytical separations to components having a molecular weight of less than 100 Da [26]. A molecular weight of 350Da is to be considered as a maximum for any component to be transport through regular commercial membranes. Fouling of anion-exchange membranes often occurs when the anion is still small enough to penetrate into the membrane structure, but its mobility is so poor that the membrane is virtually blocked. To overcome this problem membranes were developed which are characterized by a high permeability for large organic anions. In general, the permselectivity of these membranes is lower than that of regular membranes. A membrane which is less sensitive to traces of detergents is commercially available today from Ionics. Another type of anti-fouling anion-exchange membrane is produced by Tokuyama Soda. The membrane is coated with a thin layer of cation-exchange groups causing electrostatic repulsion of organic molecules.

3 . 4 Bipolar membranes Bipolar membranes have recently gained increasing attention as an efficient tool for the production of acids and bases from their corresponding salts by electrically enforced

505

accelerated water dissociation. Bipolar membranes can be prepared by simply laminating conventional cation- and anion-exchange membranes back to back [27]. Laminated bipolar membranes often exhibit unsatisfactory water splitting capability. But special surface treatment of commercial ion-exchange membranes and subsequent laminating may yield bipolar membranes with satisfactory properties. Single film bipolar membranes and multilayer bipolar membranes fulfil most of the practical needs [28,29]. ELECTRODIALYTIC PROCESSES AS UNIT OPERATION

In electrcdialytic processes the membrane is the most important component determining the overall performance in a given application. However, process design and chemical engineering aspects also have a significant effect on the efficiency and the economics of a process in a given application. The different electrodialytic processes are rather different in their basic concept and system design. 1. Electrodialysis The technically and economically most important electrodialytical process used for the separation of ionic components from an aqueous solution is conventional electrodialysis. The main application of electrodialysis is the desalination of brackish water. However, other uses, especially in the food, drug, and chemical process industry as well as in biotechnology and waste water treatment, have recently gained a broader interest. In its basic form electrodialysis can be utilized to perform several general types of separations, such as the separation and concentration of salts, acids, and bases from aqueous solutions, or the separation of monovalent ions from multiple charged components, or the separation of ionic compounds from uncharged molecules.

1. 1 Elecwodialysis process principles The principle of the process is illustrated in Figure 3, which shows a schematic diagram of a typical electrodialysis ceU arrangement consisting of a series of anion- and cation-exchange membranes arranged in an alternating pattern between an anode and a cathode to form individual cells. If an ionic solution such as an aqueous salt solution is pumped through these cells and an electrical potential established between anode and cathode, the positively charged cations migrate towards the cathode and the negatively charged anions towards the anode. The cations pass through the negatively charged cation-exchange membranes but are retained by the positively charged anion-exchange membranes. Likewise the negatively charged anions pass through the anion-exchange membranes, and are retained by the cationexchange membranes. The overall result is an increase in the ion concentration in alternate compartments, while the other compartments simultaneously become depleted.

506

Fig. 3.

Schematic diagram illustrating the electrodialysis process

The depleted solution is generally referred to as the diluate and the concentrated solution as the brine. The technical feasibility of electrodialysis as a mass separation process, i.e., its capability of separating certain ions from a given mixture with other molecules, is mainly determined by the ion-exchange membranes used in the system. The economics of the process are determined by the operating costs, which are dominated by the energy consumption and investment costs for a plant of a desired capacity, which again are a function of the membrane area and various design parameters such as cell dimensions, flow velocities, etc.. 1.2 Electrodialysis energy requirements The energy required in an electrodialysis process is an additive of two terms: (1) the electrical energy needed to transfer the ionic components from one solution, i.e. the feed through membranes into another solution, i.e. the brine, (2) the energy required to pump the solutions through the electrodialysis unit. Depending on various process parameters, particularly the feed solution concentration, either one of the two terms may be dominating and thus determining the overall energy costs. At high feed solution concentrations energy requirements for the ion transfer are dominating. At very low feed solution concentrations energy for pumping the solution through the stack may be more significant [30]. a) Minimum energy required for the separation of a molecular mixture In electrodialysis as in any other separation process there is a minimum energy required for the separation of various components from a mixture. For the removal of salt from a saline solution this energy is given by:

G

AGO = RT l n y

aw

(3)

507

Here AGO is the Gibb's free enthalpy required to remove one mole of water from a solution, R the gas constant and T the temperature in

OK;

c0 and a,S

are the water activities in pure

water and the solution. Expressing the water activity in the solution with a monovalent salt by the concentration of the dissolved ionic components the minimum energy required to remove water from a monovalent salt is given by [31]:

I

--1

E C O11

(4)

Here AGO refers to the Gibb's free enthalpy required energy for the production of 1 litre of diluate solution. C is the salt concentration in moles per litre, the superscripts 0,('), and (") refer to the feed solution, the diluate and the concentrate. The reversible Gibb's free energy can also be expressed by:

AG =

ni zF Acp (i = 1,2,3

... n)

(5)

Here F is the Faraday constant, z the electrochemical valance, n is the number of moles, and Acp is the potential drop due to the concentration difference in the diluate and concentrate. This potential drop between two solutions separated by a semipermeable membrane is generally referred to as concentration potential. Practical energy requirements for the ion transfer b) The total electrical potential drop across an electrodialysis cell consists only partly of the concentration potential, the other part is used to overcome the ohmic resistance of the cell. This ohmic resistance is caused by the friction of the various ions with the membranes and the water molecules while being transferred from one solution to another, resulting in an irreversible energy dissipation in form of heat generation. The potential drop to overcome the ohmic resistance can be and in most practicle relevant applications is significantly higher than the concentration potential, thus in electrodialysis the practical consumed energy is generally much higher than the required theoretical energy. Furthermore, a considerable amount of energy is also necessary to pump the feed solution, the diluate and the electrode rinse solution through the electrodialysis stack. The energy necessary to remove salts from a solution is directly proportional to the total current flowing through the stack and the voltage drop between the two electrodes in a stack. The energy consumption in a practical electrodialysis separation procedure can thus be expressed by:

E m = l*nRet (6) Here E m is the energy consumption, I the electric current flowing through the stack, Re the

508

resistance of a cell pair, n the number of cell pairs in a stack, and t is the time. The electric current needed to desalt a solution is directly proportional to the number of ions transferred through the ion-exchange membranes from the feed stream to the concentrated brine. It is expressed as [321:

I=

z F QAC

5

(7)

Here F is the Farady constant, z the electrochemical valence, Q the feed solution flow rate, and AC the concentration difference between the feed solution and the diluate, and 6 the current utilization. The current utilization is directly proportional to the number of cell pairs in a stack. A combination of equations (6) and (7) gives the energy consumption in electrodialysis as a function of the current applied in the process, the electrical resistance of the stack, i.e., the resistance of the membrane and the electrolyte solution in the cells, the current utilization, and the amount of ions removed from the feed solution: Eprod =

InRetzF QAC

5

Here EM is the energy requirement, I the electric current through the stack, n the number of cell pairs in a stack, Re the resistance of a cell pair, t the time, z the electrochemical valence of the components to be removed, F the Faraday constant, Q the volume flow rate of the feed solution, AC the ion concentration difference between the feed solution and the diluate, and 5 the current utilization. Electrical energy required in electrodialysis is not only directly proportional to the amount of salts, i.e. electrical charges that has to be removed from a certain feed volume to achieve the desired product quality but it is also a function of the electrical resistance of the cells. Electrical resistance of the cells, again, is a function of individual resistances of the membranes and of the solutions in the cells. Since, furthermore, the resistance of the solution is directly proportional to its ion concentration, the overall resistance of a cell will in most cases be determined by the resistance of the diluate solution. The concentration in the diluate cell, however, is decreasing during the desalting process and thus its resistance is increasing accordingly. Under the assumption, that the concentration in the diluate is much lower than that in the feed and brine the energy consumption can be expressed by 1241:

Here I is the electrical current passing through a cell stack, n the number of cell pairs in a

509

stack, Q the total volume of the diluate solution, C is the concentration and b a constant factor. The subscripts o and d refer to the feed and the diluate solution. A typical value for the resistance of an electrodialysis cell pair, i.e., the cation- and anion-exchange membrane plus the dilute and concentrated solution, e.g., in the desalination of brackish water, is within the range of 5 to 500 Q cm2. For other application the electrical resistance of a cell pair might be significantly higher or lower. Energy required to pump the solutions through the stack c) The operation of an electrodialysis system requires two or three pumps to circulate the diluate, the brine and eventually the electrode rinse solutions through the stack. The energy required for pumping these solutions are determined by the volumes to be circulated and the pressure drop. It can be expressed by:

Here Ep is the pumping energy, k a constant refemng to the efficiency of the pumps, Q volume flows, and Ap the pressure losses in the diluate; the subscripts D, B and E refer to brine and electrode rinse solution. The pressure losses in the various cells are determined by solution flow velocities and the cell design. The energy requirements for circulating the solution through the system may become significant or even dominant when solutions with rather low salt concentration are processed. Other energy consuming processes are the electrochemical reactions at the electrodes. In a stack with a multicell arrangement, however, the energy consumed at the electrodes is generally less than 1 % of the total energy used for the ion transfer and can therefore be neglected. d) Comparison of energy consumption in electrodialysis and competing processes In many applications electrodialysis is competing with other separation processes. While the theoretically required minimum energy is identical in all processes there are significant differences as far as the irreversible energy dissipation is concerned. For the desalination of a saline solution, e.g. different processes, such as reverse osmosis, ion exchange, distillation, are used in addition to electrodialysis. All processes require the same theoretical minimum energy. The irreversible dissipated energy is rather different in the different processes, as can be illustrated by comparing the basic principle of desalination by electrodialysis and reverse osmosis which is shown schematically in Figure 4.

510

+

Salt and water

Salt and water

Water

Water

Anions

AP

A P

AE

Water

Sah

Reverse osmosis

Fig. 4:

Electrodidysis

Schematic diagram illustrating the operating principles of reverse osmosis and electrodialysis

The basic difference between reverse osmosis and electrodialysis is that in reverse osmosis the water passes the membrane under a driving force of a hydrostatic pressure difference and in electrodialysis the salt is passing the membrane under the driving force of an electrical potential difference. The irreversible energy loss in reverse osmosis is caused by a friction loss of the individual water molecule on their pathways through the membrane matrix. This means, that the irreversible energy loss in reverse osmosis is independent of the feed water salt concentration. In electrodialysis the irreversible energy loss is caused by the friction of the individual ion on their pathway through the membrane from the diluate to the brine solution, thus in electrodialysis the irreversible energy loss is directly proportional to the concentration difference between the feed solution and the product water. For feed solutions with low salt concentrations the energy requirements are therefore generally lower in electrodialysis than in reverse osmosis, and at high feed solution salt concentration the situation is reversed. This is shown schematically in Figure 5 where the irreversible energy consumption is plotted versus the feed solution concentration assuming for both cases identical product water concentrations.

feed solution salt concentration

Fig. 5:

Schematic diagram showing the irreversible energy loss in electrodialysis, and reverse osmosis as a function of the feed solution salt concentration

511

A comparison of mass separation processes concerning their energy consumption has to take into account that in electrodialysis the energy is required in form of electricity, a relative expensive form, and in distillation, e.g., a relative inexpensive form of energy, i.e., heat can be used. In ion exchange very little energy is required directly. However, the chemical used for the regeneration of the resin required a significant amount of energy for their production. Electrodialysis process and equipment design Electrodialysis as a unit operation is determined by several process and equipment design parameters, such as feed flow velocities, cell and spacer construction, stack design etc. These parameters effect the costs of the process directly and also indirectly by means of the limiting current density and the current utilization [ 3 3 ] . 1. 3

Limiting current density and current utilization a) The limiting current density is the maximum current which may pass through a given membrane area without obtaining effects resulting in higher electrical resistance, lower current utilization, or other operational problems such as membrane fouling and scaling. The limiting current density is determined by the ion concentration in the dilute flow stream and by concentration polarisation effects as indicated in Figure 6, which shows the concentration profiles of cations in the boundary layer at the surface of a cation-exchange membrane during an electrodialysis process. Similar anion concentration profiles are obtained at the surface of an anion-exchange membrane.

t

1

Cathode

Anode

laminar boundary laier

Fig. 6:

Schematic diagram of the concentration profiles of cations in the laminary boundary layer at both surfaces of a cation-exchange membrane during electrodialysis. (C is the cation concentration, the subscripts b and m refer to the bulk solution and c and d refer to concentrate and diluate).

The transport of charged particles to the anode or cathode through a set of ion-exchange membranes leads to a concentration decrease of c o u n t w n s in the laminar boundary layer at the membrane surface facing the diluate cell and an increase at the surface facing the brine cell. The effect of concentration polarization due to a concentration increase in the

512

brine is less severe. The decrease in the concentration of counter ions directly affects the limiting current density and increases the electrical resistance of the solution in the boundary layer. The limiting current density is the current density at which the ion concentration at the surfaces of the cation- orJand anion-exchange membranes in the cells with the depleted solution will approach zero. The limiting current density can be calculated by a mass balance considering all mass transport through the membrane and the boundary layers. It can be described to a first approximation by [34]:

Here ilim is the limiting current density, C$ the bulk solution concentration in the cell with the depleted solution, D and z are the diffusion coefficient and the electrochemical valence of the ions in the solution, F the Faraday constant, Yb the boundary layer thickness, and TM and T the ion transport numbers in the membrane and the solution, respectively, and the subscripts + and - refer to cations and anions, respectively. The constant k is the mass transfer coefficient, taking into account the influence of the hydrodynamics of the feed solution flow, i.e. the flow channel geometry, the spacer design, the flow velocities, etc. According to equation (11) the limiting current density is proportional to the ion concentration in the diluate and the mass transfer coefficient, which is determined mainly by the cell geometry and the feed solution flow velocity. If in electrodialysis the limiting current density is exceeded, the process efficiency will be drastically diminished because of the increasing electrical resistance of the solution and because of water splitting which leads to both increasing energy consumption as well as pH changes in the solutions at the surface of the membrane causing additional operational problems. The mass transfer coefficient can be related to the Sherwood number, which again is a function of the Schmidt and Reynolds number [35].Introducing the proper relations in equation (1 1) leads to an expression which describes the limiting current density as a function to the feed flow velocity in the electrodialysis stack

d Here c b is the concentration of the solution in the bulk of the diluate cell, u is the linear flow velocity of the solution through the cells parallel to the membrane surface and a and b are constants, the value of which are determined by a series of parameters such as cell and spacer geometry. solution viscosity, ion-transfer numbers in the membrane and the solution, etc. The constants a and b are a function of the electrodialysis stack design and must be

513

determined experimentally. The limiting current density can be determined by several means [36], e.g. by measuring the electrical resistance of a cell pair or the pH-value in the diluate cell as a function of the current density. When the pH-value is plotted versus l/i a sharp decrease in the pH-value is noted when the limiting current density is exceeded. Likewise, when the total resistance of a cell pair is plotted versus l/i a minimum is obtained at the limiting current density. This is shown schematically in Figure 7 a and b. That usually a pH-drop in the diluate cell is observed is due to the fact that water splitting usually occurs first at the anion-exchange membrane probably because of the catalytic effect of the tertiary amine groups at the surface of the membrane [37] as will be discussed later when describing the water splitting mechanism in bipolar membranes.

al

G

0

al 3 -

2

Ia

r

.-

v)

2

I

1 '

4

1li

Limiting current density a)

Fig.7:

8 C

4

1 /i Limitin'g current density b)

Schematic diagram illustrating the determination of the limiting current density by plotting: a) the pH-value of the diluate cell versus l/i and b) by plotting the resistance of a cell pair versus l/i

The limiting current density determines the minimum membrane area required to achieve a certain desalting effect. Another very important parameter for the overall performance of the electrodialysis process is the current utilization. The current utilization determines the ponion of the total current that passes through an electrodialysis stack that is actually used to transfer ions from a feed solution. The current utilization which is always less than 100 % is affectea by three factors [38]: (1) The membrane selectivity, (2) osmotic and ion-bound water transport, and (3) current passing through the stack manifold.

Here 5 is the current utilization, q is an efficiency term and n is the number of cell pairs in a stack; the subscripts w, m and s are referring to efficiency losses due to water transfer, conductivity of stack components and membrane selectivity. The water transfer due to osmosis and ion hydration can be significant at higher brine salt concentrations. The membrane selectivity also depends on the salt concentration due to a Donnan equilibrium between the salt solution and the membrane as discussed earlier.

514

For a diluted feed or brine solution an ion-exchange membrane in general is more or less strictly semipermeable, i.e. the membrane is permeable to counter ions only. When, however, the ion concentration in the feed solution is of the same order as that of the fixed charges in the membrane co-ions may also enter the membrane and its selectivity will then be decreased, with the consequence that the current efficiency decreased, too. b)

The electrodialysis stack design A typical electrodialysis stack design is shown in Figure 8. An electrodialysis stack is essentially a device to hold an array of membranes between electrodes in such a way that the streams being processed are kept separated. Cation-exchange membrane

Spacer

Anion-exchange membrane Concentrate Diluate

Feed Dilu'ate cell Condentrate Cell

Fig. 8:

Exploded view of an electrodialysis stack

The gaskets not only separate the membranes but also contain manifolds to distribute the process fluids in the different compartments. The supply ducts for the diluate and the brine are formed by matching holes in the gaskets, the membranes, and the electrode cells. The distance between the membrane sheets, i.e. the cell thickness, should be as small as possible to minimize the electrical resistance. In industrial size electrodialysis stacks membrane distances are typically between 0.5 to 2 mm. A spacer is introduced between the individual membrane sheets both to support the membrane and to help control the feed solution flow distribution. The most serious design problem for an electrodialysis stack is that of assuring uniform flow distribution in the various compartments. In a practical electrodialysis system, 200 to 1000 cation- and anion-exchange membranes are installed in parallel to form an electrodialysis stack with 100 to 500 cell pairs. As any membrane separation process electrodialysis is effected by concentration polarization and membrane fouling. The magnitude of concentration polarization is largely determined by the electrical current density, by the cell and particularly spacer design, and by the flow velocities of the diluate and brine solutions [39]. Concentration polarization effects electrodialysis lead to a depletion in the laminar boundary layer at the membrane

515

surfaces in the cell containing the diluate flow stream and to an increase of ions in the laminar boundary layer at the membrane surfaces in the cell containing the brine solution. Concentration polarization is effecting the separation efficiency by decreasing the limiting current density 1401. More difficult to control are membrane fouling effects due to adsorption of polyelectrolytes, such as humic acids, surfactants, proteins etc.. The components often penetrate the membrane because of their size only partially and thus resulting in severely reduced ion permeability of the membrane 1411. In designing an electrodialysis stack several general criteria concerning mechanical, hydrodynamic, and electrical properties have to be considered. Since some of the criteria are counter effective, the final stack construction is generally a compromise between several conflicting parameters [38,391. A proper electrodialysis stack design should provide a maximum effective membrane area per unit stack volume. The dismbution of the solutions should ensure equal and uniform flow distribution through each compartment. Any leakage between the diluate, concentrate, and the electrode cells should be prevented. The spacer screen should provide a maximum of mixing of the solutions at the membrane surface and cause a minimum in pressure loss. Most stack designs used in today's large-scale electrodialysis plants are one of two basic types: tortuous path or sheet flow. These designations refer to the type of solution flow path in the compartments of the stack. In the tortuous-path stack, the membrane spacer and gasket have a long serpentine cut-out which defines a long narrow channel for the fluid path. The objective is to provide a long residence time for the solution in each cell in spite of the high linear velocity that is required to limit polarization effects. A tortuous-path and sheet flow spacer gaskets are shown schematically in Figure 9a) and b).

Fig. 9: Schematic diagram of a) a tortuous-path electrodialysis spacer gasket, and b) a sheet-flow electrodialysis spacer gasket

516

In stack designs employing the sheet-flow principle, a peripheral gasket provides the outer seal and the solution flow is approximately in a straight path from the entrance to the exit ports which are located on opposite sides in the gasket. This is illustrated in Figure 7 a) which shows the schematic diagram of a sheet-flow spacer of an electrodialysis stack. Solution flow velocities in sheet-flow stacks lie typically between 4 and 10 cm/sw, whereas in tortuous-path stacks solution flow velocities of 10 to 30 cm/s are required [42, 431. Because the higher flow velocities and longer flow paths, higher pressure drops in the order of 2 to 3 bars result in tortuous-path stacks than in sheet-flow systems where pressure drops of 1 to 2 bars occur. Electrdalysis process design and economics c) In addition to the actual stack and the power supply unit, an electrodialysis plant consists of several components essential for proper operation, such as pumps, process monitoring and control devices, feed solution pretreatment systems, etc.. There are two operating modes for the electrodialytic process described in the literature [44]. The first is referred to as the unidirectionally operated electrodialysis plant and the second is a reversed polarity operated electrodialysis plant [7]. A flow diagram of a typical unidirectional operated electrodialysis plant is shown in Figure 10. Feed Inlet

I

Concentrate Inlet

I Electrode Waste

Electrode Waste Product

Concentrate Recycle

Fig. 10:

-

....

cnncnntrate Blowdown -". ,"",

Flow diagram of a typical unidirectional electrdalysis desalination plant

After proper pretreatment, the feed solution is pumped through the actual electrodialysis unit, which generally consists of one or more stacks in series or parallel. A deionized solution and a concentrated brine is obtained. The concentrated and depleted process streams leaving the last stack are collected in storage tanks, when the desired degree of concentration or depletion is achieved, or they are recycled if further concentration or depletion is desired. Sometimes

517

acid is added to the concentrated stream to prevent scaling of carbonates and hydroxides. To prevent the formation of free chlorine by anodic oxidation the electrode cells are sometimes rinsed with a separate solution which does not contain any chloride ions. In many cases, however, the feed or brine solution is also used in the electrode cells. In the electrodialysis reversal operating mode the polarity of the current is changed at specific time intervals ranging from a few minutes to several hours. In this operating mode the hydraulic flow streams are reversed simultaneously, i.e. the diluate cell will become the brine cell and vice versa. The advantage of the reverse polarity operating mode is that precipitation in the brine cells are to a large extent prevented. Or if there is some precipitation, it will be redesolved when the brine cell becomes the diluate cell in the reverse operating mode. The flow scheme of a typical electrodialysis reversal plant is shown in Figure 11. Feed.Inlet , Feed

Electrode Waste

Electrode Waste

Fig. 11:

Flow diagram of a typical electrodialysis reversal plant

The process design and economics are closely related in electrodialysis. The total process costs are the sum of fixed charges associated with amortization of the plant capital costs and operating costs, such as energy and labour costs. Membrane replacement costs are sometimes regarded as a separate item because of their relatively short life of 5 to 7 years. Capital costs include depreciable items such as the electrodialysis stacks, pumps, elecmcal equipment, membranes etc., and non depreciable items such as land and working capital. The capital costs of an electrodialysis plant will strongly depend on the total membrane area required for a certain plant capacity. The required membrane area, however, is proportional to the number of ionic species removed from a given feed solution. It can be calculated by the

518

following relation [331: A =

zFQACn is

Here A is the membrane area, z the chemical valence, Q the volume of the produced potable water, A C the difference in the salinity of feed and product water, n the number of cells in a stack, i the current density which should be about 80 % at the limiting current density, and 5 the current utilization. The limiting current density is a function of the diluate concentration which is changing during the desalting process from the concentration of the original feed to the product solution concentration. The calculation of the minimum membrane area required for a given desalting capacity is based on an average limiting current density, which is a function of the average diluate concentration given by:

-

Here Flim is an average limiting current density, Cd is the average diluate concentration, a is a constant factor, which depends on the cell and spacer geometry and feed flow velocity. The subscripts o and d refer to the feed solution and the diluate. Introducing equation (15) into (14) leads to:

Here Amin is the minimum membrane area required for a certain plant capacity and feed and product solution concentrations, a is a constant for a given plant design and operating mode, z is the chemical valence, F the Faraday constant, Q the product solution volume, and COand Cd the feed and the product solution concentrations. For a certain plant capacity, the required membrane area is directly proportional to the feed water concentration. This is illustrated in Figure 12. For brackish water of ca. 3000 ppm TdS and an average current density of 12 mA/cm2, the required membrane area for a plant capacity of 1 m3 product per day is ca. 0.4 m2 of cation- and anion-exchange membrane. Other items such as pumps, electric power supplies, etc. depend on plant size. For desalination of brackish water with a salinity of ca. 3000 ppm the total capital costs for a plant with a capacity of 1000 m3/d will be in the range of US $200,000.- to US $300,000.-. The costs of the actual membrane is less than 30 % of the total capital costs. Assuming a useful life of 5 years for the membranes and 10 years for the rest of the equipment, a feed water salinity of 3000 ppm and a 24-hours operating day, the total amortization of the

519

investment is ca. US $0.10 to US $0.15 per m3 water with a salinity of less than 500 ppm.

1

10

100

Feed solution Concentration gll

Fig. 12:

Schematic diagram of the required membrane area in electrodialysis desalination as a function of the feed water concentration at constant current density, plant capacity and product water concentsation.

The operating costs are mainly determined by the required energy which is, as pointed out before, determined by the electrical energy required for the actual desalting process and the energy necessary for pumping the solution through the stack. The energy for the actual desalting process, i.e. the ion transfer from the feed solution to the brine is directly proportional to the number of ionic species to be removed, as indicated in equation (8) and (9), respectively. The energy requirements for the production of potable water as a function of the feed water concentration is shown in Figure 13. The case considered is a NaCl feed solution with the product having a salt concentration of less than 500 ppm.

x

P 0,

C

UJ

Fig. 13:

1 I 1

1

I

10

100

Feed solution concentration (g/l)

Energy requirements for the production of potable water with a solid content of 500 ppm as a function of the feed solution concentration (AU per cell pair = 0.8 V)

520

The pumping energy is independent of the feed solution salinity. Assuming a pressure drop in the unit of ca. 400 KPa (4 bar), a pump efficiency of 70 %, and 50 % product water recovery, the total pumping energy will be ca. 0.4 kWh per m3 product water. This indicates that at low feed water salt concenmtion the cost for pumping the solution through the unit might become quite significant. It should be noted, that according to equations (8) and (16),the energy costs increase with increasing current density while the required membrane area decreases with increasing current density. Thus the total desalination costs, which are the summation of capital, energy and operating costs, will reach a minimum at a certain current density as illustrated in Figure 14, where the total costs are shown as a function of the applied current density for a given feed solution. Total costs Energy cost

s

0

Capital costs Operating costs Current density

Fig. 14:

Schematic diagram of the electrodialysis process costs as a function of the applied current density

Quite interesting furthermore is a comparison of the cost of desalination by various processes as a function of the feed water salinity, as shown in Figure 15. O.Or

ion-exchange //

E!ectrodialysis

h

m

E

b

s

1.0

3

0

0.1

Na Ci Concentration (911)

Fig. 15:

Water desalination costs as a function of the feed solution concentration for ion- exchange, electrodialysis,reverse osmosis, and distillation

52 I

The graph in Figure 15 indicates that at very low feed solution salt concentration ion exchange is the most economical process. But its costs are sharply increasing with the feed solution salinity and at about 500 ppm TDS electrodialysis becomes the more economical process. While at around 5000 ppm reverse osmosis is the less costly process. At very high feed solution salt concentrations, in excess of 100 000 ppm multistage flash evaporation becomes the most economical process. The costs of potable water produced from brackish water sources are in the range of US $0.2 to US $0.5 per m3. 1.4

Technically relevant applications of electrodialysis

Electrodialysis was developed first for the desalination of saline solutions, particularly brackish water. The production of potable water is still currently the most important industrial application of electrodialysis. But other applications, such as the treatment of industrial effluents [45], the production of boiler feed water, demineralization of whey [46], de-acidification of fruit juices [47], etc. are gaining increasing importance with large-scale industrial installations. An application of electrodialysis which is limited regionally to Japan has gained considerable commercial importance. This is the production of table salt from sea water. Diffusion dialysis and the use of bipolar membranes have significantly expanded the application of electrodialysis in recent years [48]. a) Desalination of brackish water In terms of the number of installations the most important large-scale application of electrodialysis is the production of potable water from brackish water. Here, electrodialysis is competing directly with reverse osmosis and multistage flash evaporation. For water with relatively low salt concentration (less than 5000 ppm) electrodialysis is generally the most economic process, as indicated earlier. One significant feature of electrodialysis is that the salts can be concentrated to comparatively high values (in excess of 18 to 20 wt.%) without affecting the economics of the process severely . Most modern electrodialysis units operate with reverse polarity, i.e. the anode and cathode, and with that the diluate and concentrate cell systems, are exchanged periodically, preventing a scaling due to concentration polarization effects. In brackish water desalination, more than 2000 plants with a total capacity of more than 1,000,000 m3 of product water per day are installed, requiring a membrane area in excess of 1.5 million square meters [49]. b) Production of table salt The production of table salt from sea water by the use of electrodialysis to concentrate sodium chloride up to 200 g/Lprior to evaporation is a technique developed and used nearly exclusively in Japan. More than 350,000 tons of table salt are annually produced by this technique requiring more than 500,000 square meters of installed ion-exchange membranes. Key to the success of this technology has been the low cost, high conductive membrane with

522 a preferred permeability of monovalent ions [6]. However, it should be noted that in Japan this procedure of salt production is highly subsidized. c) Waste water treatment The main application of electrodialysis in waste water treatment systems is in processing rinse waters from the electroplating industry. Here, complete recycling of the water and the metal ions can be achieved by electrodialysisin some applications [50]. Compared to reverse osmosis, electrodialysis has the advantage of being able to utilize more thermally and chemically stable membranes, so that processes can be run at elevated temperatures and in solutions of very low or high pH-values. Furthermore, the concentrations which can be achieved in the brine can be significantly higher. The disadvantage of electrodialysis is that only ionic components can be removed and additives usually present in a galvanic bath cannot be recovered. The recovery of nickel salts from electroplating rinse waters is an application which has been pursued by several companies [42]. Here electrodialysis has the function of a "kidney" by removing the nickel salts which have been dragged out of the plating tank into a still-rinse [50].The concentrated nickel salts can often be directly fed back into the plating tank, while the diluate is recycled into the still-rinse. Dump leach waters containing heavy metal ions have successfully been treated by electrodialysis. The removal of nitrate from drinking water by electrodialysis has been studied extensively and seems to compete well in this application with other treatment procedures, such as ion-exchange or reverse osmosis [47]. An application which has been studied in a pilot plant stage is the regeneration of chemical copper plating baths [45]. In the production of printed circuits, a chemical process is often used for copper plating. The components which are to be plated are immersed into a bath containing, besides the copper ions, a strong complexing agent, for example, ethylenediamintetraacetic acid (EDTA), and a reducing agent such as formaldehyde. During the plating process, formaldehyde is oxidized to formate. After prolonged use, the bath becomes enriched with Na2 SO4 and sodium formate and consequently loses its useful properties. By applying electrodialysis in a continuous mode, the Na2 SO4 and formate can be removed from the solution, without affecting the concentrations of formaldehyde and the EDTA complex and the useful life of the plating solution is significantly extended. Several other potential applications of electrodialysis in wastewater treatment systems which have been studied on a laboratory scale are reported in the literature. In most of these applications the average plant capacity, however, is considerably lower than that in brackish water desalination or table salt production. Concentration of reverse osmosis brines d) A further application of electrodialysis is the concentration of reverse osmosis brines. Because of limiting membrane selectivity and the osmotic pressure of concentrated salt solutions, the concentration of brine in reverse osmosis desalination plants can not exceed

523

certain values. Often the disposal of large volumes of brine is difficult, and a further concentration is desirable. This further concentration may be achieved at reasonable costs by electrodialysis. [421 Electrodialysis in the chemical, food, and drug industry e) The use of electrodialysis in food, drug, and chemical industries has been studied quite extensively in recent years. Several applications have considerable economic significance and are already well established today. One is the demineralization of cheese whey [46]. Normal cheese whey contains between 5.5 and 6.5 % of dissolved solids in water. The primary constituents in whey are lactose, protein, minerals, fat and lactic acid. Whey provides an excellent source of protein, lactose, vitamins, and minerals, but in its normal form it is not considered a proper food material because of its high salt content. With the ionized salts substantially removed, whey provides an excellent source for the production of babyfood. The partial demineralization of whey can be carried out quite efficiently by electrodialysis. The removal of tartaric acid from wine is another possible application of electrodialysis. In the production of bottled champagne, it is necessary to avoid the formation of crystalline tartar in the wine and tartaric acid must therefore be reduced to a value which does not exceed the solubility limit. This can be done efficiently by electrodialysis. Several other applications of electrodialysis in the pharmaceutical industry have been studied on a laboratory scale [51]. Most of these applications are concerned with desalting solutions containing active agents which have to be separated, purified, or isolated from certain substrates [52]. Here, electrodialysis is often in competition with other separation procedures such as dialysis, solvent extraction, etc. In many cases, electrodialysis is the superior process as far as economics and the quality of the product is concerned. Especially the separation of amino acids and other organic acids by electrodialysis seems to be of interest to the pharmaceutical and chemical industry [53].However, the deionization of cheese whey with an installed capacity of more than 35,000 square meters of membrane area for the production of more than 150,000 tons of desalted lactose per year is economically by far the most important application of electrodialysis in the food industry today. f) Production of ultrapure water More recently electrodialysis is being used in combination with mixed-bed ion-exchange

resins for the production of ultra pure water. In this application electrodialysis is used as a pre-treatment step and is in direct competition to reverse osmosis which has the advantage to remove also neutral components in addition to the salts. For certain feed water sources, however, electrodialysis is preferred for economic reasons. Overall the electrodialysis industry has experienced a steady growth since it made its appearance as an industrial scale separation process about 15 years ago and new areas of application in the food and chemical process industry are gaining interest rapidly.

524

OTHER ELECTRICALLY DRIVEN MEMBRANE PROCESSES Although electrodialysis is today by far the most important industrial membrane separation process utilising ion-exchange membranes and an electrical potential gradient as driving force there are several other processes gaining industrial significance rapidly, such as regular electrolysis used for the production of chlorine and caustic soda 1541, the electrodialysis with bipolar membranes used for the production of acids and bases from the corresponding salts [55], or the combination of conventional electrodialysis with regular ion-exchange techniques to produce ultra pure water. Most of these processes have been developed only recently and their large scale industrial utilization is still in the beginning. The chlorine-alkaline electrolysis The electrolytic production of chlorine and caustic soda using a cation-exchange membrane as a separation medium is already a technically and commercially well established process [54, 563.The principle of the process is illustrated in the schematic drawing of Figure 16, which shows an electrolysis cell arrangement consisting of two chambers separated by an cation-exchange membrane. 1.

t

Na CI

Fig. 16:

Schematic diagram illustrating the chlorine/alkaline production process

One compartment contains an anode and the sodium chloride feed solution. The other compartment contains the cathode and at the beginning of the process water. When an electrical potential difference between the two electrodes is applied, the positively charged sodium ions will migrate towards the cathode producing hydrogen and hydroxyl ions in an electrochemical reaction at the cathode. The negatively charged chloride ions move towards the anode and will be oxidized to form chlorine. Thus sodium chloride is converted into chlorine, caustic soda, and hydrogen. A migration of the hydroxyl ions is prevented by the cation-exchange membrane. Thus the current utilization in the electrolytic chlorine and caustic soda production is close to 100 %. The compartment containing the produced sodium

525

hydroxide is usually operated in a feed and bleed mode and its sodium hydroxide concentration is kept as high as possible. In industrial production processes sodium hydroxide concentrations in excess lOwt% are obtained. Since the sodium chloride concentration is also kept rather high the electrical resistance of the solutions is comparatively low, and the cell system can be operated at relatively high current densities up to a few thousand Mm2. The main problem in the electrolytic production of chlorine and caustic soda is the stability of the cation-exchange membrane which faces a strong caustic environment on one side and solution containing free chlorine on the other side. Today, membranes based on fluorinated hydrocarbone polymers have a useful life time of several years in operation at elevated temperatures [56]. 2. Electrodialytic water dissociation in bipolar membranes Bipolar membranes have recently gained increasing attention as efficient tools for the production of acids and bases from the corresponding salts by electrically forced water dissociation. The process, which has been known for many years, is economically very attractive and has a multitude of possible applications [%I. So far, however, the large-scale use of the process has been rather limited because of the shortcomings of today's bipolar membranes, which have to meet certain requirements as far as their water splitting capability, their electrical properties and chemical stability is concerned. But recent progress in the development of efficient bipolar membranes have increased the technical and industrial importance of this process. Principle of water dissociation in bipolar membranes a) The water dissociation in a bipolar membrane is illustrated in Figure 17 which shows a bipolar membrane consisting of an anion- and a cation-exchange layer arranged in parallel between two electrodes. anion-exchange

cation-exchange membrane

Y

\

/

bipolar membrane

Fig. 17:

Schematic diagram illustrating the principle of the electrodialytic water dissociation in bipolar membranes

526

If an electrical potential difference is established between the electrodes all charged components will be removed from an aqueous interphase between the two ion-exchange layers. If only water is left in the solution between the membranes further transport of electrical charges can only be accomplished by protons and hydroxyl ions which are available in very low concentrations in completely de-ionized water. Protons and hydroxyl ions removed from the interphase are replenished because of the water dissociation equilibrium. A bipolar membrane thus consists of a cation- and anion-exchange layer laminated together. The theoretical energy required for the process is that for establishing the desired concentration of H+ and OH--ions in the outer phases of the membrane from their concentration in the membrane which is approximately 10-7mow. The free energy of this process is:

AG=zRTln

i i aH+ %HaH+ %H-

Here AG is the free energy, z the electrochemical valance, R the gas constant, T the absolute temperature, and a the activity. The superscripts o and i refer to the outside and the interphase between cation- and anion-exchange membrane, respectively. 0

0

For the generation of a one molar ideal solution of H+- and OH--ions, i.e., %+= 1, aOH-= 1, and z = 1, equation (17)reduces to:

i

i

AG = R.T-l n(a H+.a OH-)=R T h K w Here Kw is the dissociation constant of water. At 25 OC the negative logarithm of the water dissociation constant, - log Kw is 13.99. The free energy for the dissociation of one mole water and thus the production of one molar acid and base at 25 OC is: 79887 Joule or 0.0222 kwh. The generation of protons and hydroxyl ions via an electrolysis process, however, requires considerably more energy. This is evident from the very nature of the process which entails co-production of H2 and 0 2 or chlorine. This step requires some additional energy input. The theoretical energy in electrolysis varies slightly, depending on the particular salt being processed, the concentration of acid and base generated, and the temperature of operation. For production of one normal acids and bases at 25OC the theoretical free energy varies between 0.056 and 0.58 kwh./mol. To minimize the irreversible energy losses in a bipolar membrane its electrical resistance should be as low as possible. Furthermore in practical applications, bipolar membranes are exposed to an aggressive chemical environment. The cationic side of the bipolar membrane is facing a strong acid and the anionic side of the membrane is in contact with a strong base. The preparation of cation-exchange membranes with excellent

521

stability even in strong acid is relatively easy. Anion-exchange membranes with the required alkaline stability and electrical properties especially at elevated temperature are far more difficult to make [57]. But today bipolar membranes with long term stability at pH-values in excess of 13 are commercially available. These membranes can be operated at current densities in excess of lo00 A m-2 with high current utilization [55]. Applications of the electrodialytic water dissociation b) The main application of electrodialytic water dissociation in combination with regular ionexchange membranes is the production of acids and bases from the corresponding salts. In this application bipolar and regular cation- and anion-exchange membranes are installed in alternating series between two electrodes to form a stack of individual cells similar to those used in conventional electrodialysis. In this case, however, a repeating unit consist of three individual cells. A typical arrangement of an electrodialysis stack with bipolar membrane as used for the production of an acid and a base is shown in Figure 18.

Acid

Base

&Pt

A

0

C

A

-

3

Anod

0 Cathode

X-

Salt

Fig. 18:

t

Salt

Schematic drawing illustrating an electrodialysis cell arrangement with bipolar membranes used for the production of acids and bases from the corresponding salts

The electrodialytic water dissociation has been evaluated on a laboratory scale and a multitude of applications have been identified mainly in generating acids and bases from the corresponding salts. The process shows significant advantages in terms of energy requirements over conventional acids and bases production procedures. The purity of the products, however, is often unsatisfactory especially when high acid and base concentrations are required. The process can be integrated in chemical or biochemical production processes when an adjustment of pH-values is required. Bipolar membranes have also been integrated in acids and bases scrubbers used for the removal of waste gases from air such as SO2 [%I.

528

3 Diffusion dialysis Another process utilizing ion-exchange membranes in an electrodialysis stack cell arrangement is referred to as diffusion dialysis. This process can be used, e.g. to separate acid from mixtures with salts [58]. The principle of the process is illustrated in Figure 19. In this case a diffusion dialysis cell system contains anion-exchange membranes only. A feed solution is pumped through alternating cells while water is pumped in counter current flow through the other cells. Protons and the anions can penetrate the anion-exchange membranes due to a concentration gradient while the cations are rejected. The net result is the removal of acids from a mixture with salts. Like-wise bases can be removed from salt solutions using cation-exchange membranes. Acid

(Salt+acid)

I

Acid

1 x

4 Fig. 19:

Schematic diagram illustrating the principle of diffusion dialysis

The process is used on a large scale to recover mineral acids from salt solutions obtained in pickling and etching processes. In this application only anion exchange membranes are installed in a stack as indicated in Figure 19. By feeding in alternating cells a mixture salt and acid and pure water in counter current flow more than 95% of the acids can be removed from the feed solution. Donnan dialysis D O M dialysis ~ is used to exchange ions between two solutions. The stack arrangement is identical to that used in diffusion dialysis. The principle of the process is illustrated in Figure 20, which shows a CuSO4 solution and H2SO4 separated by a cation-exchange membrane. Since the H+-ion concentration in the acid solution ' is significantly higher ( p ~ = 1) than the H+ ion concentration in copper sulfate solution " (PH = 7) there will be a driving force for the transport of H+-ions from solution I into solution II. Since the membrane is permeable to cations only, there will be a build-up of an elecmcal potential which will counter-balance the concentration difference driving force of the H+- ions. This electrical potential difference will cause a flux of Cu++-ions against their concentration

4.

529

gradient from solution " into solution '. As long as the H+-ion concentration difference between the two phases separated by the cation-exchange membrane is maintained, there will be the transport of Cu++-ions until their concentration difference is of the same order of magnitude as the H+-ion concentration difference. Solution

(PH = 7)

Solution'

(pH = 1)

"\cation exchange membrane

Fig. 20:

Schematic drawing illustrating the principle of Donnan dialysis by showing the transport Cu ++-ions through a cation-exchangemembrane utilizing an electrical potential build up by the flux of H+-ions

The process can be carried out accordingly with anions through anion-exchangemembranes. An example of anion Donnan dialysis is the sweetening of citrus juices. In this process hydroxyl ions furnished by a caustic solution replace the citrate ions in the juice. Electrodialytic regeneration of a cation- or anion-exchangeresins The process is illustrated in Figure 21 which shows a cation loaded resin placed between two electrodes. 5.

Salt free solution H+

I Cathode

Me

+

Anode

Cation exchange resin Me+ Feed solution

Fig. 21:

Schematic diagram illustrating the electrodialytic regeneration of a cationexchange resins

When an electric current is applied protons generated at the anode will move to the cathode into the resin and replace the metal ion on the resin which will then move from the resin to the cathode where they are collected, precipitated or concentrated. Electrodialytic regeneration of ion-exchange resins is less labour intensive than chemical regeneration and

530 adds no additional salts to the effluent. The process has been evaluated on a laboratory scale but low current efficiency due to the high mobility of the protons have hampered its practical application [59]. Recently introduced modifications in the process design, however, have eliminated part of this problem [60]. A anion-exchange resin is regenerated similarly by hydroxyl ions generated at the cathode. Using a mixed bed ion-exchangeresin completely de-ionized water is obtained. This process has recently been commercialised [61]. 6. Hybrid Processes Most separation processes are efficient only under certain conditions of feed solution concentration and required product quality. Therefore various separation processes are often combined, each operating its optimum range of application. A typical example for the combination of different processes is the combination of conventional ion-exchange process with electrodialysis. There are however, other combinations such as electrodialysis and reverse osmosis which are of growing technical and commercial interest. CONCLUSIONS The electrodialytic processes have experienced a steady growth since they made their appearance as industrial scale separation processes about 20 years ago. Currently the desalination of brackish water, the chlorine-alkaline elecwlysis and the production of table salt are still the dominant applications, but new areas of application in the food and chemical process industry and the use of hybrid processes are gaining interest rapidly. REFERENCES Ostwald, W. 1890. Elelctrische Eigenschaften halbdurchllssiger Scheidewiinde. Z. Physik. Chemie 6:71-82. Donnan, F.G., 1911. The theory of membrane equilibrium in presence of a nondialyzable electrolyte. Z. Electrochem. 12572. Morse, H.N., Pierce, J.A.. 1903. Z. physik. Chem., G:589. Meyer, K.H., Strauss. 1940. Helv. Chem. Acta 795. Juda, W., and W.A. McRae. 1950. Coherent ion-exchange gels and membranes. J. Am. Chem. Soc.22:1044. Nishiwaki, T. 1972. Concentration of electrolytes prior to evaporation with an electromembraneprocess. In: Industrial Processing with Membranes, Edts.: R.E. Lacey, and S. h b . Wiley & Sons, New York. Katz, W.E. 1979. The electrodialysisreversal (EDR) process. Desalination a:31-40. Connolly, D.J., and W.F. Gresham. 1966. Fluorocarbon vinyl ether polymers. U.S. Patent 3,282,875. Liu, K.J., F.P. Chlanda, and K.J. Nagasubramanian. 1977. Use of bipolar membranes

a:

531 for generation of acid and base: an engineering and economic analysis. J. Membr. Sci.

2:109. 10 Kedem, O., and Y. Maoz. 1976. Ion conducting spacer for improved electrodialysis. Desalination Ip( 1-3):465-470. 1 1 Donnan, F.G., and E.A. Guggenheim. 1932. Exact thermodynamics of membrane equilibrium. Z. Phys. Chem. A162:346-360. 12 Bergsma, F., Ch.A. Kruissink. 1961. Ion-exchange membranes. Fortschr. Hochpo1ym.Forsch. a:307-362. 13 Helfferich, F. 1962. Ion Exchange. New York: McGraw-Hill Book Co. 14 Gregor, H.P. 1951. Gibbs-Donnan equilibria in ion-exchange resin systems. J. Am. Chem. Soc. 22:642-650. 15 Teorell, T.Z. 1951. Z. Elektrochem. Z:460-469. 16 Meyer, K.H., J.F. Sievers. 1936. Permeability of membranes. Helv. Chim. Acta B: 649-677. 17 Adams, B.A., E.L. Holmes. 1936. Brit. Pat. 450 308. 18 Friedlander, H.Z. 1968. Membranes Encycl. Polym. Sci. Technol. &620-638. 19 Kusomoto, K., T. Sata, and Y. Mizutani. 1976. Modification of anion-exchange membranes with polystyrene sulfonic acid. Polym. J. &:225-226. 20 McRae, W.A., and S.S. Alexander. 1960. Sulfonating reagent and its use in preparing cation-exchange membranes. U.S. Patent 2,962,454. 21 Flett, D.S. 1983. Ion Exchange Membranes. Chichester, U.K.:E. Horwood Ltd. 22 Eisenberg, A. and H.L. Yeager. 1982. Perfluorinated Ionomer Membranes. ACS Symposium Series No. 180. Washington, DC: American Chemical Society. 23 Grot, W.G. 1973. Laminates of support material and fluorinated polymer containing pendant side chains containing sulfonyl groups. US.Patent 3,770,567. 24 Ishigaki, I., N. Kamiya, T. Sugo, and S. Machi. 1978. Synthesis of an ion-exchange membrane by radiation-induced grafting of acrilyc acid onto poly(tetrafluorethy1ene). PO~YIYI.J. J.Q(5):513-5 19. 25 Sata, T., R. Izuo, and Y. Mizutani. 1984. Study of membrane for selective permeation of specific ions. Soda to Enso 3(415):313-336. 26 Gudernatsch, W., Ch. Krumbholz, H. Strathmann. 1990. Development of an AnionExchange Membrane with Increased Permeability for Organic Acids of High Molecular Weight. Desalination B:249-260. 27 Leitz, F.B. 1972. Apparatus for electrodialysis of electrolytes employing bilaminar ionexchange membranes. U.S. patent 3,654, 125. 28 Chlanda, F.P., L.T.C. Lee, and K.J. Liu. 1976. Bipolar membranes and method of making same. U.S. Patent 4, 116,889. 29 Bauer, B., F.-J. Gerner, and H. Strathmann. 1988. Development of bipolar membranes. Desalination 48:279-292. 30 Schaffer, L.H., and M.S. Mintz. 1966. Electrodialysis. In: Principles of Desalination,

532

31 32

33 34

35 36 37 38

39 40 41

42 43 44

45 46

47 48 49

Ed.: K.S. Spiegler, pp. 3-20. New York: Academic Press. Spiegler, K.S. 1956 in: 'Ion-Exchange Technology", Edts.: F.C. Nachod, J. Schubert. pp.118-181, Academic Press, New York, N.Y. Strathmann, H., 1979. Trennung von Molekularen Mischungen mit Hilfe Synthetischer Membranen. Darmstadt: Steinkopff-Verlag. Mintz, M.S. 1963. Electrodialysis: principles of process design. Ind. Eng. Chem 2 5 (6):18-28. Lacey, R.E. 1972. Basis of electromembrane processes. In: Industrial Processing with Membranes, Edts.:. R.E. Lacey and S.Loeb. New York: John Wiley & Sons. Bird R.B., W.S. Stuart,E.N. Lightfoot, 1960. "Transport Phenomena". John Wiley & Sons, New York. Cowan, D.A:, J.H. Brown. 1959. Effect of turbulence on limiting current electrodialysis cells. hid. Eng. Chem. 51:1445. Simocs, R. 1984. Electric field effects on proton transfer between ionizable groups and water in ion-exchange membranes. Electrochemica Acta 2: 151-158. Huffmann, E.L., R.E. Lacey. 1972. Engineering and economic considerations in electromembrane processing. In: "Industrial Processing with Membranes". Edts.: R.E. Lacey, S. Loeb, pp. 39-55. John Wiley & Sons. Jonsson, G., C.E. Boezen. 1984. Polarization phenomena in membrane processes. In: "Synthetic Membrane Processes". Ed.: G. Belfort. Academic Press, New York. Spiegler, K.S. 197 1. Polarization at ion-exchange membrane-solution interfaces. Desalination 9367-385. Korngold, E., F. De KBriSsy, R. Rahav, and M.F. Taboch. 1970. Fouling of anion selective membranes in electrodialysis. Desalination &(2):195-220. Ionics Inc. 1988. Product Bulletin, Watertown, MA 02172. Asahi Chemical 1988. Product Bulletin, Tokyo, Japan. Strathrnann, H., 1990. Membranes and Membrane Separation Processes. "Ullmann's Encyclopedia of Industrial Chemistry" Vol. A 16, pp. 187-263.Verlag Chemie, Weinheim, FRG. Korngold, E., IS.Kock, and H. Strathmann. 1978. Electrodialysis in advanced waste water treatment. Desalination B(1-3):129-139. Ahlgren, R.M. 1972. Electro membrane processing of cheese whey. In: Industrial processing with Membranes, Edts.:. R.E. Lacey and S. Loeb, pp. 71-81. New York: John Wiley & Sons. Kneifel, K., G. Liihrs, H. Wagner. 1985. Nitrate removal by electrodialysis for brewing water. Desalination 68:203-209. Sata, T. 1986. Recent trends in ion-exchange research. Pure & Appl. Chemistry 58:1613-1626. Wangnick, K. 1985. Desalting plant inventory report. International Desalination Association (IDA) Report.

533

50 Deuschle, A., E. Kiibler. 1984. Elektrodialyse: Wertstoffriickgewinnungaus galvanischen Spiilwlssern. Galvanotechnik x:968:97 1. 51 Reed, Ph.B. 1984. Electrodialysis for the purification of protein solutions, Chem. Eng. Progress, Dec. 1984: 47-50. 52 Voss, H. 1986. Deacidification of citroacid solutions by electrodialysis. J. Membrane Sci. =:165-171. 53 Itoh, H., T. Yoshizumi, M. Saeki. 1986. Sieving effect in electrodialysis with an ionexchange membrane. J. Membrane Sci. z:155-163. 54 Coulter, M.V. 1980. Modem chlor-alkali technology. Ellis Honvood Ltd., Chichester, U.K. 55 Mani, K.N. 1991. Electrodialysis water splitting technology. Journal of Membrane Science 3(2):117. 56 Seko, M., S. Ogawa, K. Kimoto. 1988. Perfluorocarboxylic acid membrane and membrane chlor-alkali process developed by Asaki Chemical Industry. In: "Perfluorinated Ionomer Membranes". Edts.: A. Eisenberg, H.L. Yeager. ASCSymposium Series L8Q,American Chemical Society, Washington. 57 Bauer, B., F. Effenberger, H. Strathmann. 1990. Anion-Exchange Membranes with Improved Alkaline Stability. Desalination 2125-144. 58 Kobuchi, Y., H. Motomura, Y. Noma, F. Hanada. 1987. Application of ion-exchange membranes to recover acids by diffusion dialysis. J. Membrane Sci. u:173-179. 59 Strathmann, H., Kock, K. 1982. Effluent free electrodialytic regeneration of ionexchange resins, in: "Polymer Seperation Media", Edt.: A.R. Cooper, Plenum Press, New York, N.Y. 60 Johan, J., G. Eigenberger Chem. Eng. Techn. in Press 61 Millipore Corp. Bulletin. 1987. No. CJ 001. Bedford, Ma.