Electrolytic assisted ion exchange Improves ion extraction efficiency and provides easter regeneration of the ion exchange material in the cell. In such a system, an electric field is applied across a water-spliffing Ion exchange membrane, as described in commonly assigned U. S. Patent No. 5,788,826 to Nyberg, which is incorporated herein by reference in its entirety. The water splitting membrane typically comprises a strong-acid cation exchange surface or layer (sulfonate groups ;-SO3M) and a strong-base anion exchange surface or layer (quaternary ammonium groups;@ NRsA). The membrane is positioned between electrodes so that its cation exchange surface faces the first electrode and tts anion exchange surface faces the Second electrode. During a deionizatlon process cycle, a solution stream is passed through the cell while a predetermined voltage level is applied to the electrodes to generate an electric field normal to the surfaces of tho water-splitting membrane. The electric field irreversibly dissociates water to split it Into its component ions H+ and OH- that migrate through the Ion exchange layers in the direction of the electrode having an opposite polarity (e. g., H6 migrates toward the negative electrode). The etectric fiefd aasiats the transport of the dissolved solid ions generated by the water-splitUng reaction in a direction perpendicular to the membrane to provide a short pathway through the

membrane. Thus, during deionization, the electric field is set at a single level that is sufficiently high to dissociate the water and effectively transport and remove a large majority of dissolved solids from the solution. The predetermined fixed field strength is generated by applying a single high DC voltage to the electrodes that maximizes ion extraction. Electrolytic ion exchange advantageously provides a uniform electric field in the cell that better utllfzes the entire surface area of the membrane and that increases ion exchange efficiency to be ab ! e to remove 90% or more of the dissolved solids.

Electrolytic assisted ion exchange systems also allow electrical regeneration of the membranes, which is advantageous over conventional chemical regeneration processes. Conventional cation exchange layers are commonly regenerated using acidic solutions, such as sulfuric acid ; and anion exchange layers are regenerated using basic solutions, such as sodium hydroxide. Regeneration is concluded with a rinsing step that removes entrapped regenerat solution. These chemical processes require large amounts of regenerating chemicals and/or water, and the cell has to be periodically shut down to allow the regeneration process. However, in electrolytic assisted ion exchange processes, the water-splitting membrane is regenerated by simply reversing the polarity of the voltage applied to electrodes to generate an inverted electric field that electrically regenerates the membranes by disgorging exchanged ions from the membrane. The reverse polarity voltage is also applied at a single voltage level that maximizes lon expulsion andRor rejection efficiency from the water splitting membrane during the regeneration cycle.

However, conventional electrolytic assisted ion exchange systems have several limitations. One limitation is that the TDS removal fraction can vary with influent solution quality and cell operating conditions. For example, a change in the flow rate or pressure of the influent solution can result in different fractions of dissolved ; sulids bl-ir iy removed. The total dissolved solids content of the effluent solution also varies due to the TDS content of the Influent solution changing with time, for example, the TDS of sewage changes drastically with heavy rains. As a result an electrolytic cell that provides a 90% TDS reduction will generate a treated effluent solution having a TDS of 15 ppm from an influent solution having a TDS of 150 ppm, but when the influent solution contains 1500 ppm TDS, the treated effluent solution will have a TDS of 150 ppm. Such variations in the output TDS content are undesirable. Conventional elsctolytic ion exchange cells also often exhibit a gradual increase in TDS content in

the effluent solution, as the working ion exchange capacity of the membrane is consumed during deionization, which further increases the variability In effluent solution TDS content. The slow increase of ion concentration during batch deionization processes can cause the premature end of the deionization cycle, well before the capacity of the membrane is truly exhausted. increasing capital and operating costs for the cell. Thus, the output TDS content can vary significantly with Influent TDS content and over time with conventional systems.

Furthermore, conventional electrolytic ion exchange systems also do not allow control of the ion concentration. While maximizing extraction of Ions from a solution stream is desirable to purify water and de ionized water in industrial applications, In some applications, it is desirable to maintain a redefined level of dissolved solids in the solution stream. For example, In drinking and cooking water applications, some dissolved solids are desirable so the water tastes better and does not taste synthetic. In other industrial water applications, it is also desirable to reduce the level of a particular ion in a wastewater stream, for example nitrate or arsenic ions, to meet an environmental standard. A particular ion level in a chemical solution can also be needed to provide precise control of the composition of the solution for industrial processes, for example, in cement manufacture and in electroplating solutions.

Thus, it is desirable to be able to treat influent solutions to provide uniform and consistent ion concentrations in the efnuent sotution, even if the quality or TDS of the influent sotuttoo vanes over the treatment process. There is also a need fier a water treatment system that can use electrical power for regeneration of ion exchange materials that uses electrical power rather than chemicals for regeneration to reduce or eliminate the incnnvenience and environmcntal hazmds associated with regenerant chemicals, and reduce rinse water volumes during cleaning cycles. It is also desirable to be able to treat influent solutions to maintain a predetermined or set level of dissolved solids in the treated solution.

SUMMARY An apparatus to treat an influent solution comprising ions to obtain a selectable ion concentration in a resultant effluent solution. The apparatus comprises an electrochemical cell comprising a housing comprising first and second electrodes. A water-splitting ion exchange membrane Is between the first and sennnd eteetrodes, tha membrane comprising a cation exchange surface facing the first electrode and an anion exchange surface facing the second electrode. The housing also has an influent solution inlet and an effluent solution outlet with a solution channel that allows influent solution to flow past both the anion and cation exchange surfaces of the water-splitting ion exchange membrane to form the muent solution. A variable voltage supply is capable of maintaining the first and second electrodes at a plurality of voltage levels during an ion exchange stage. The ion exchange stage can be, for example, a deionization or ion removal step, or a membrane regeneration or ion rejection step In one version, the apparatus includes an ion meter to measure an ion concentration of the influent solution, at least partially treated influent solution. or effluent solution, and generate an ion concentration signal. A controller receives the ion concentration signal from the ion meter, and smrtds a control signai to the vanable voltage supply to adjust the voltage level applied to the first and second electrodes in response to the ion concentration signal to achieve a redefined ion concentration range in the effluent solution. in another aspect, a method of treating an Influent solution comprising ions to control the concentration of ions in an effluent solution. is provided. The method comprises flowing the influent solution past both anion and cation exchange surfaces of a water-splitting ion exchange membrane to form the eflYuent solution ; maintaining a time averaged electric field across the cation and anion exchange surfaces of the water-splitting membrane ; and varying the strength of the time averaged electric field to control the ion concentration of the effluont solution.

DRAWINGS These features, aspects, and advantages of the present invention will become better understood with regard to the following description, appended claims, and accompanying drawings which illustrate exemplary features of the invention: Figure 1 is a schematic sectional side view of an embodiment of the apparatus of the present invention illustrating an electrolytic ion exchange cell, power supply and control system ; Figure 2 is as sectional view of sectional side view of another embodiment of the apparatus having an electrolytic ion exchange cell comprising a spiral membrane ; Figure 3 is a top view of the electrolytic ion exchange cell of Figure 2: Figure 4 is a graph showing a variable time averaged voltage for an Ion removal step (a) and an ion rejection (b); Figure s Is a circuit diagram of a variable voltage supply comprising a -phase-control-voltage-supply ;- Figures 6a-c are graphs showing exemplary Input AC voltage and DC output voltage waveforms produced by the phase control voltage supply of Figure 5; Figure 7a is block diagram of a switching voltage supply with pulse width modulation ; Figure 7b and c are graphs showing an input AC voltage signal and an output duty cycle modulated OC voltage generated by the voltage supply of Figure 7a : Figure 8 Is a circuit diagram of a switching DC voltage supply 90 ; Figure 9 is a schematic sectional view of an embodiment the apparatus that uses current and voltage measurements to control the voltage applied to the electrodes of a cell :

Figure 10a is an embodiment of the present invention comprising two cells connected m parallel ; Figure 10b is an embodiment comprising two cells connected in series ; Figure 10c Is an embodiment comprising two cells connected in series, and employing constant voltage and a variable voltage power supp@es; and Figure 11 is a graph that shows the comparative conductivity levels obtained for t-Muent solutions treated with a constant voltage level (constant voltage) or a variable voltage level (variable voltage).

DESCRIPTION An embodiment of the present apparatus 20 is napabie of treating an influent solution comprising ions to extract, replace, or add ions to generate an effluent solution having desired ion concentrations. Exemplary embodiments of the ion controlling apparatus 20 are provided to illustrate the invention and should not be used to limit the scope of the invention, and alternative cell arrangements and configurations as would be apparent to those of ordinary skill In the art are withio Lhe scope of the Invention. Generally, the apparatus 20 comprises an electrochemical cell 22 comprising a housing 26 that is an enclosed leak proof structure having at least one influent solution Inlet 30 and at least one offluent, solution outlet 35, as shown in Figure' 1. A suitable housing 25 typically comprises a plate and frame construction fabricated from metal or plastic. While one or more outlets 35 can be provided, the effluent solution from the housing 25 preferably comprises a single effluent solution stream that is formed before or after the outlet 35, for example in an exhaust manifold that combines the different solution streams. Uptionally, a pump 38, such as for exHrup) , a peristaltic pump or water pressure in combination with a flow control device can be used to flow influent solution into the inlet 30 of the housing 25.

The housing 25 has first and second electrodes 40, 45 therein. One or more of the electrodes 40, 45 can form a portion of the housing 25, for example an electrically conducting wall of the housing as shown in Figure 3 or can be a discrete structure such as a free standing plate inside the housing (as shown) In Figure 1.

Typically, the electrodes 40, 45 have conducting surfaces that tace one another. The anode and cathode electrodes 40, 45 are fabricated from electrically conductive materials, such as metals which are preferably resistant to corrosion in the low or high pH chemical environments created during pusitive and negative polarization of the electrodes during operation of the cell 22. Suitable electrodes can be fabricated from corrosion-resistant materials such as titanium or niobium, and can have an outer coating of a noble metal, such as platinum. The shape of the electrodes 40, 45 depends upon the design of the electrochemical cell 22 and the conductivity of the influent solution 70 flowing through the cell 22. A suitable electrode is a flat plate that provides : a uniform voltage across its surfaces, and that can have openings to allow influent solution to pass therethrough. In another arrangement, the electrodes 40, 45 are shaped as spirals, discs, or even conical shapes.

A water-splitting ion exchange membrane 100 is between the first and second electrodes 4D, 45. the membrane 100 comprising an anion exchange surface 4G facing the first electrode 40, and a cation exchange surface 48 facing the second electrode 45 as shown In Figure 1, or vice versa as shown in Figure 2. The water- splitting membrane 100 can comprise abutting anion and cation layers 49, 51, respectively, contained In an open frame 47 positioned between the electrodes 40, 45. or can be arranged in a spiral configuration as shown in Figures 2 and 3. Suitable anion exchange layers 49 of water-splitting membrane 100 comprise one or more basic functional groups capable of exchanging anions such as-NR3A,-NR2HA,-PRsA,- SR2A, or C5H5NHA (pyridine), where R is an alkyl, aryt, or other organic group and A is an anion (e. g., hydroxide, bicarbonate, chloride, or sulfate ion). The choice of anion exchange functional group also depends on the application. In water deionization,- NR3A is preferred for its ability to impart good membrane swelling, and thus provide low electrical resistances and high masa tranaport rates, over a wide range of pH. Weak base groups are preferred when particularly efficient regeneration Is required. For example, -NR2HA will react with OH in a very favorable reaction to form-NR2, HzO, and expel A. Suitable cation exchange layers 51 can comprise one or more acidic functional groups capable of exchanging cations such as-COOM,-SO3M,-PO3M2, and -C6H4OM, where M is a cation (e. g., hydrogen, sodium, calcium, or copper ion). Cation exchange materials also include those comprising neutral groups or ligands that bind cations through coordinate rather than electrostatic or ionic bonds (for example pyridine. phosphine and sulfide groups), and groups comprising complexing or cheating groups (e. g"those derlved from aminophosphoric acid, aminocarboxyiicacid, and hydroxamic acid). The choice of cation exchange functional group depends upon the applica ion of the cell 22. In water deionization for which the non-selective removal of ions is required.-SOgM groups are preferred for their ability to impart good membrane swelling, high mass transport rates, and low electrical resistances over a wide range of pH. For the selective removal of copper ion from a liquid containing other ions, for example sodium ion, ion exchange groups such as-COOM or a cheating group such as aminocarboxylic acid are preferred. These weak acid groups offer the additional benefit of particularly efficient regeneration due to the strongly favorable reaction of- (CO nM with t-T to form-COOH and expel M"', where M is a metal ion.

The water-splitting ion exchange membranes 100 can also comprise more than two anion and cation exchange layers 49, 51. For example, the water-splitting

membrane 100 can have two cation exchange layers and two anion exchange layers.

Each two cation or anion exchange layer differs In ion exchange capacities or ion exchange functions) groups. For example, a first inner cation exchange layer of water- splitting membrane 100 may comprise substantially-SO3 groups and an outer cation layer can comprise-COOH groups ; while the inner anion exchange layer comprises- NRs groups and the outer anion layer comprlses-NR2H groups. The anion and cation exchange layers can also be porous to hold solution, for example, an open cell foarn or other structure, to provide faster solution transport through the water-splitting ion exchange membrane.

A solution channel 62 in the housing 26 allows influent solution 70 from the inlet 30 to flow past both the anion and cation exchange surfaces 46, 48 of the water-splitting ion exchange membrane 100 to form the effluent solution 80'at the outlet 35. The flow path of solution channel 52 can be defined by the housing 25 and the structures in the housing 25. For example, the channel 52 can be formed between the surfaces of the water-splitting membranes 100, the electrodes 40, 45, and the sidewalls 54 of the housing 25, as shown In Figure 1. The channel 52 can also be a tubular walled structure in the housing (not shown). The solution channel 52 extends from the inlet 30 to outlet 35 where it outputs treated effluent solution 80.

An alternative embodiment of the electrochemical cell 22 comprises a first electrode 40 that forms an outer electrode structure and a second electrode 45 that forms a central electrode structure, as shown in Figure 2. The water-splitting membrane 100 comprises a plurality of anion and cation exchange layers 49, 51 that are positioned between the two electrodes 40, 45. The membrane layers 49, 51 are arranged in the housing 25 so that the cation exchange surfaces 48 of the membranes face the first electrode 40, and the anion exchange surfaces 46 of the membranes face the second electrode 45. The membranes layers 49, 51 form a spiral wrapped configuration with a spacer 43 which separates the cation surfaces 48 from the anion surfaces 46 and provides a solution channel 62, as shown in Figure 3. Any number of membranes 100 may be wrapped In parallel in this way to provide the desired pressure drop through the cell 22. For example, three membranes 100 can be spiral wrapped in a parallel arrangement. Spacers 43 can also be positioned within the solution channel 52. The spacers 43 can be, for example, layers of plastic netting material suspended form the sidewalls 54 ot the cell. I he spacers 43 can be used to separate the water-

ssplitting membranes 100, provide more uniform flow. or create turbulence in the solution channel 52 to provide hlgher ion transport rates. If two or more water-splitting membranes 100 are in direct contact, excess current may flow through this low resistance path, overheating the membranes 100 and bypassing the solution (thereby reducing cell performance). The spacers 43 can have openings sized 10 um or higher In dtarreLer.

The method of operation of the electrochemical cell 22 will be Illustrated for the treatment of an influent snlution comprising sodium chloride and for treatment of an influent solution comprising copper sulfate, in the cell 22 shown in Figures 2 and 3.

Although operation of a cell 22 is illustrated with respect to these two selected examples, it should ice understood that the cell 22 and apparatus 20 can be used for. many other applications, and should not be limited to the examples provided herein. In the tirst example, sodium chloride is removed and concentrated from the influent solution. The electrochemical cell 22, as shown in Figure 2, comprises water-splitting membranes 100 having sulfonate cation exchange layers 51 and quaternary ammonium anion oxchange layers 49 In the first step of the process, ion removal, a first electrode 40 facing the cation exchange layer 51 Is positively charged (the anode), and the second electrode 45 is negatively charged (the cathode) At the interface between the cation and anion exchange layers 61, 49 of the water-splitting ion exchange membranes 100, water in the solution stream is split into its component ions H+ and OH-, and OH- migrates through the cation exchange surfaces 48 toward the first (positive) electrode 40 while H+ migrates through the anion exchange surfaces 46. toward the second (negative) electrode 45. In this step, Na* diffuses and migrates through the cation exchange surfaces 48 in the direction of the negative electrode. The reactions occurring In the cation exchange layers 51 are : (1) P-SO3H + Na+ = P-SO3Na + H+ (2) H+ + OH- = H2O where"P"refers to the polymer or solid support to which the ion exchange group is attached. The OH-produced by the water-splitting reaction and minting through the cation exchange surfaces 48 reacts with H+ to form water. The consumption of H+ maintains the electric and concentration gradients which promote the rate of removal of Na* frum solution.

Within the anion exchange layer 49, Ci"* diffuses and migrates through the anion exchange layer 49 in the direction of the positive electrode. The reactions occurring In the anion exchange layers 49 are : (3) P-NR3OH + Cl- = P-NR3Cl + OH- (4) H+ + OH- = H2O Analogous to the situation In the cation exthange layer 51, the M+ produced by the water-splitting reaction reacts with OH-to promote the rate of removal of Cl"from solution.

While the ion removal step can be performed without voltage (or current), ion removal rates are substantially slower due to the absence of the migration and chemical neutralization effects described above. Without voltage the irreversible water- splitting reaction does not occur, and thus neutralization reactions (2) and (4) cannot take place, Ion removal retce are then governed exctusivety by the diffusion rates of Na+, H+, Cl~, and OH-through the cation and anion exchange layers 51, 49.

After thA ion exchange groups within the water-splitting membrane's cation and anion exchange layers 51,49 are fully occupied by Na+ and ions, respectively, a regeneration process step Is required to return the water-splitting ion exchango membranes 100 to their original chemical form prior to reactions (1) to (4). In this step, a solution stream 70 Is introduced into the cell 22, as for example, the same sodium chloride solution treated in the first stop, and the polarity of the electrodes is reversed so that the first electrode 40 is now the negative electrode and the second electrode 45 is the positive electrode. In this step, H+ produced by the water-spliting reaction moves through the cation exchange layar 51 in the direction of the negative electrode causing reaction (5) : (5) P-SO3Na + Hot = P-503H + Na' Similarly, OH- produced in the water-splitting reaction moves through the anion exchange layer 49 in the direction of the positive electrode, causing reaction (6) to occur : (6) P-NR3Cl + OH- = P-NR3OH + Cr The Na and Cl~ replaced by H+ and OH- in the cation and anion exchange layers 51, 4S combine to form NaCI in tha effluent solution stream 80. By controlling the solution flow rate during this regeneration step, a concentration of NaCi substantially higher than was present in the Incoming solution stream 70 during the ion removal step can be

produced. ! f ion removal in reactions (1) to (4) was intended to deionize a solution stream, then the NaCi concentrate formed during reactions (5) and (6) is discarded. If the purpose of NaCl removal in this apparatus was to isolate this salt, then the NaCl concentrate is saved for further processing.

The operation of this exemplary cell 22 will be described In the context of the removal and subsequent concentration of CuSO4 from a solution. Any suitable ion exchange materials can be used In the cation and anion exchange layers 51, 49 of the water-splitting membranes 100. For example, the cation exchange layers 51 may comprise P-COOH groups, and the anion exchange layers 49 may comprise P-NR2H groups. During ion removal, OH- produced in the water-splitting reaction again moves through the cation exchange layers 51 toward the positive electrode, and reactions (7) and (8) occur within the cation exchange layers 61 : (7) P-COOH + OH- = P-COO- + H2O (8) 2 P-COO"+ CU+2 = (P-COO) 2CU Cu+2 cannot replace H+ in P-COOH direct@y due to unfavorable thermodynamics, and thus reaction (7) is required to remove H+ from P-COOH, forming P-COO- which readily reacts with Cu+2.

In the anion exchange layers, 49, H+ produced in the water-splitting reaction moves through the anion exchange layers 49 toward the negative electrode, causing roactions (9) and (10) to occur; (9) P-NR2 + H+ = P-NR2H+ (10) 2 P-NR2H+ + SO4-2 = (PNR2H)2SO4 Reaction (9) forms an ionic group from the neutral P-NRz, enabling the subsequent reaction with 804-2 in reaction (10).

When ion removal is complete, the cation and anion exchange layers 61, 49 are returned to their condition prior to ion removal by flowing the same or a different solution through the call 22 and reversing the polarity of the electrodes 40, 45. HO produced in the water-splltting reaction now moves through the cation exchange layers 51, resulting in reaction (11) : (11) (P-COO) zCu + 2 H+ = 7 P-COOH + CU+2 and OH'-produced in the water-splitting reaction moves through the anion exchange layers 49, resulting in reaction (12):

(12) (P-NRaH) 2SO4 + 2 OH- = 2 P-NR2 + 2 H20 + SO4-2 Both reactions (11) and (12) are thermodynarnicallyfavorable, allowing the concentration ot CU804 In te sulution stream, In a subsequent step, the copper and/or sulfate may be recovered from the solution stream or discarded.

The method of replacing lons in an ion exchange material Illustrated in the two preceding examples may be described in general terms as follows. The water- splitting membranes 100 comprise ion exchange layers A and B, one a cation exchange layer 51 and the other an anion exchange layer 49, which layers contain ions I1A and I1B, respectively-An ion-contalning solution electrically connects the electrodes 40, 45 and the water-splitting membranes 100, and upon application of a sufficient voltage to the two electrodes 40, 45, water at the interfacial region between ion exchange layers A and B is"splif into its component ions H' and OH-. This phenomenon, refend to as the"watcr-oplitting reaction,"invntvps the spontaneous dissociatlon of water into component ions H+ and OH- (dissociation occurs with or without the electric fleld), followed by the migration of these ions Into ton exchange layers A and B under the influence of the electric field. Hs migrates toward the negative electrode, and OH'toward the positive electrode As the H- and OH- migrate into layers A and B, more water diffuses to the Interfaclai region tu continue the water-splitting reaction cycle. As H and OH'move through ion exchange layers A and B, they cause the replacement of ions I1A and by ions ISA and 12B from the solution, respectively.

I his replacement of ions in ion exchange layers A and B may be either a direct replacement, for example His for Na* on a P-SO3- ion exchange group, or it may be an Indirect replacement, for example as occurs when OH'reacts in a first step with P- COOH to form P-COO', followed by reactinn to bind Cu+2.

The apparatus 20 further comprises a variable voltage supply so to maintain the first and second electrodes 40, 45 at a plurality of voltage levels during an ion exchange stage. The voltage bias levels are selectable voltage levels that have different magnitudes during the ton exchange stage. For example, the voltage levels can be time averaged voltage levels that are each a fixed constant voltage level or fixed pulsed voltage level, during a pre-specified time period that is a portion of the time period of an ion exchange stagc, ouch as (i) a solution treatment step, fnr example, a deionization or ion removal step in which ions are removed from the influent solution 70 to form the effluent solution 80, or (ii) a membrane regeneration step or ion rejection

step in which the membrane 100 is regenerated by expelling or replacing lons in the membrane, (ii) but not both. The ion removal step is a pass of the influent solution across thé anion and cation surface 46, 48 of the water-sptitting membrane 100 in the electrochemical cell 22 to change the ion concentration level in the influent solution 70 by removing or replacing ions from the influent solution 70. The ion rejection step Is a regeneration cycle In which thP water-splittin membrane 100 is regenerated by expelling or rejecting ions in the membrane 100. In each step. the time averaged voltage level is determined by averaging the magnitude of the voltage applied to the electrodes 40, 45 over a selected time period. The time period is a length of time in which the voltage has a peak absolute magnitude that remains substantial the same.

Figure 4 shows an exemplary embodiment of variable time averaged voltage levels for an ion removal step (graph (a) ) and an ton rejection step (graph (b) ).

In this flgure, the magnitude of the pulsed voltage is plotted over time to show different time cycles in each of which, the time-averaged voltage is maintained at one level.

Graph (a) demonstrates a variable time averaged voltage applied to the electrode that has a first magnitude (V1) that remains substantially the same over a first time period Cri), a second first magnitude (V2) that remains substantially the same over a second time period (T2), a third magnitude (V1) that remains substantially the same over a third time period (T3), etc. Curve (b) shows a time averaged voltage having a reversed polar@ty with negative values of voltages (V4) (V5) (V6) that are used during different time, periods (T4) (Ta) (Ta) 0 an ion rejection step. In each step. the time average voitage generates a time averaged electric field In the cell of different field strengths., The varying electric field has a field strength proportional to the different applied voltage tevets. it shouid be noted that in the ion rejection step, the voltage can also remain substantially constant over the entire cycle, since the objective is only to expel ions from the membrane 100.

In one version, the variable voltage supply 60 provides a time modulated or pulsed direct current (DC) voltage having a singlo polarity that remains either positive or negative, during an Ion removal step, or during an ion rejection step. In contrast, a non-DC voltage such as an alternating current (AC) supply voltage, has a time- averaged AC voltage that would be approximately zero. Employing one polarity over the course of either an ion removal (deionization) or ion rejection (regeneration) step in the operation of the electrolytic ion exchange cell 22 allows ions in the Influent solution

70 being treated to travel in a single direction toward or away from one of the electrodes 40, 48, thereby providing a net mass transport of ions either into or out of the water-spotting membranes 100. The magnitude nfthe average DC vottage is obtained by mathematically integrating the voltage over a time period and then dividing the integral by the time period. The polarity of the integration tells whether one is in ion removal or reJection mode, and the magnitude of this calculation is proportiona ! to the electrical energy made available for ion removal or rejection. A reasonable estimate of this integrated value is obtained from the time-averaged measurement observed Witt'i a slow response OC voltmeter, hr example, a voltmeter with a characteristic response time in the tens of seconds, such as 20-30 milliseconds-The voltage supply 50 can also includes a polarity switch 44 to switch polarity to the electrodes 40, 45, for example, to either momentarily add lons to solution to maintain a desired concentration in the solution in the cell 22 during the ion removal step, or to regenerate the membrane 100 in the cell 22 to prepare for a. subsequent ion deionization step.

The variable voltage supply 50 can provide a variable DC voltage or current by full-wave or half-wave rectification of an AC voltage. Such a voltage supply can be, for example, (i) a phase control voltage supply 50a, or (ii) a switching voltage supply with putze width modulation 50b. These circuits and descriptions are only two examples of how to provide a variable time-modulated DC voltage, and other variable DC voltage supply circuits apparent to those of ordinary skill in the art can also be employed. An embodiment of a phase control voltaga Ruppty 50a using phase control thyristors (SCR) 81 is Illustrated in Figure 6. The phase control voltage supply bOa can provide linear or unregulated power.) he voltage supply 50a operates on conventional single phase At power 82 (120 volts, 60 Hz) as shown in Figure 6a, inputted directly . y the AC into DC. Besides rectification, the SCRs 81 arc also able to switch from non-conducting to conducting very rapidly. When an AC power source 82 is switched by an SCR 81, the current is turned off every time the current drops to zero using the zero crossing trigger 91 that operates in combination with the trigger transistor 89 Figure 6a shows an input AC cycle signal 72 of an SCR 81 that is triggered into conduction at a trigger point (t) that is about 2 milliseconds after the AC cycle crosses zero voltage. At that point the SCR 81 turns on and remains on until the alternating wave crosses a zero voltage level, The back-to-back SCRs 81 function as full-wave rectifiers whose voltage output would be represented by the dotted line 76 in Figure 6b. Figure 6c shows the resulting waveform comprising a variante

magnitude pulsed DC voltage 78, which is then provided to the electrodes 40, 45 of the cell 22. In the phase control voltage supply 50a, the SCRs 81 are turned on by an optotriac trigger sCR 83 and their timing determined by an RC circuit 84 comprising a 555 and 4066 control chips. The trigger occurs at a set time delay after the zero crossing 85 which Is determined by varying the resistances 86 in RC timing circuit 84.

Continuous variation In the output voltage can be accomplished by using a potentiometer. This circuit 5Qa uses fixed resistors 86 to set six (or some other number) of discrete time delays to trigger the SCRs 81 at particular polnts in the alternating voltage phase thus setting five discrete output voltages at the output leads 87, for example. 0 (off), 30, 100, 200, 300 volts.

In embodiment of a phase control voltage supply 60a Illustrated in Figure 5, there are a number of standard components, such as resistors, capacitors, diodes and transistors. Suitable values and part numbers that may bR used for these components include: C1=C2=250µF (250Xf), C3=0.01µF (50V), C4=0. 1 µF(25V), D1 and D2 are T6A1002 Q1 and Q2 are 2N65D7 (SCRs), R2=100#, R4=1k#, R7=20k#, R8=22k#, R9=28k#, and R11=14k#. The above-listed values and components are only exemplary and different values or components can be used- An alternative variable voltage supply 50 comprising a switching voltage supply with pulse width modulation 50b is illustrated in the btock diagram of Figure 7a.

Given the Input signal In Figure 7c, Figue 7b illustrates the output of the voltage supply 50b. There is only one voltage level because the DC voltage or current power is controlled by varying the duty cycle, which is the ratio of on-time to off-time of the vottagc. A vcrcion of the switching voltage supply SOb that is a switching DC vottage supply 90, is illustrated in Figure 8. In the voltage supply 90. the AC power is rectified by unregulated DC rectifier and filter 93 before being transformed into a higher frequency and voltage. The modulator and inverter 94 uses a pulse width modulator 92 to regulate the voltage applied to the high frequency transformer 99. The feedback loop compares the sampling feedback voltage 95 with a reference voltage 96 and control the duty cycle of the primary transistor 98 to control the output voltage 97.

The apparatus 20 can also include an ion sonsor 69 to measure an ion concentration of the solution and generate a solution ion concentration signal. The ion sensor 69 can measure, for example, concentration, species type, or a ratio of

concentrations of different ions. In one version, the ion sensor 59 is a conductivity sensor, which is useful to determine and control total dissolved solids (TDS) concentration in the treated efftuent sofution 80. Alternatively, the ion sensor 59 can be a sensor specific to a particular ionic species, for example nitrate, arsenic or lead. Tne ion specific sensors can be for example, 1SES (ion selective electrodes).

The location of the ion sensor 59 in the solution stream is selected depending on the nature ot the zon concentration signal that Is desired. The ! on sensor can be used to (i) measure an ion concentration of the influent solution, at least partially treated solution, or effluent solution, and (ii) generate an ion concentration signal, and In one version, the ion sensor 59 is placed in the at least partially treated solution after the influent solution 70 has passed across at least some of the cation and anion surfaces 48,46 of the water-splitting membrane 100. For example, the ion sensor 59 can be placed near In the effluent solutinn stream beyond the effluent outlet 35. as shown in Figure 2. In this version, the ! on concentration signal represents the concentration of an ion, or ions, in the treated solution, and ! s used to adjust or regulate the ion extraction efficiency of the cell 22 to obtain a desired ion concentration in the cell 22.

A controller 53 receives the signal from the ion sensor 59 via a feedback circuit loop 57 and sends a control signal to the variable voltage supply 50 to control the voltage output to the electrodes 40, 4. The controller 53 comprises etectronic circuitry and program code to receive, evaluate, and send signals. For example, the controller 53 can comprise (i) a programmable integrated circuit chip or a central processing unit (CPU), (II) random access memory and stored memory, (iii) peripheral input and output devices such as keyboards and displays, and (IV) hardware Interface boards comprising analog, digital input and output boards, and communkion boards. The controller can also comprise program code instructions stored in the memory that is capable of controlling and monitoring the electrochemical cell 22, ion sensor 59. and voltage aupply 50. The program oodo may be written in any conventional computer programming language. Suitable program code Is entered into single or multiple files using a conventional text editor and stored or embodied in the memory. If the entered code text Is In a high level language, the code is compiled, and the resultant compiler code Is then linked with an object code of pre-compiled library routines. To execute the

linked, compiled object code, the user invokes the object code, causing the CPU to read and execute the code to perform the tasks identified in the program.

The controller 53 receives an ion concentration signal from the ion sensor 59 and sends a control signal to the variable voltage supply 50 to adjust the voltage level provided by the voltage supply 50 to the electrodes 40, 45. The signal sent by the controller 53 to the voltage supply SO instructs the voltage supply 50 to adjust a time averaged voltage level applied to the electrodes 40, 45 in response to the signal from the ion sensor 59. For example, the controller 53 can send a voltage control signal to the voltage supply 50 to adjust the time averaged voltage level provided by the voltage supply 60 to achieve a redefined ion concentration range in the effluent solution 80, as described below. The controller 63 can also send a control signal to the variable voltage supply 50 to select the degree of power modulation or duty cycle of the time averaged voltage level.

A method of controlling ion or TDS concentration using the cell 22 shewn in Figure 2, will be described using a measurement of the ionic conductivity of the treated effluent solution 80 by an ion sensor 59 comprising an ion conductivity sensor located just exterior to the outict 35 of the cell 22. Assuming an influent solution 70. having a conductivity of 1000 µS/cm (#500 ppm TDS) enters the cell 22 at inlet 30, that a desired conductivity of the treated solution 130 is 200 µS/cm (#100 ppm TDS), and that the cell 22 is initially powered at 100 V. As influent solution 70 passes through the cell 22, cations are extracted by the ion exchange layers having surfaces 48 and anlons are extracted by the ion exchange layers having surfaces 46. The ion sensor 59 measures a lower ionic conductivity of the effluent solution 80 exiting the cell 22 via outlet 35 that represents a reduced TDS concentration and generates a proportional ion concentration signal. The measured ion concentration signal Is relayed to the controller 53, which in turn sends a control signal to the variable voltage supply 50 by feedback loop 57. If the measured conductivity of the effluent solution 80 is lower than the desired level of 200 µS/cm, then the controller 53 sends a control signal to the variable power supply 50 to reduce the applied voltage. If the measured conductivity is greater than the desired level, then the controller 53 sends a control signal to the variable power supply SO to increase the voltage. The controller 53 can contain a programmed microprocessor or program code to adjust the voltage to a proportional level based upon the difference between the target and measured ionic conductivity. If the

measured conductivity is nearer the target, the control signal instructs the variable . voltage supply 5D to apply smaller voltage changes; and if measured conductivity is further from the target, the control signal instructs the voltage supply 60 to apply larger voltage changes. For example, If the measured ion concentration using the apparatus of Figure 1 is a specific ion, for example nitrate, then the controller 53 can send a control signal to the variable power supply 60 containing instructions to adjust voltage to a predetermined level to decrease of increase the concentration of this specific ion.

In another version. the ion sensor 59 is located in the Influent solution 70 as far upstream as possible to obtain the earliest measurement of ion concentration in the influent solution 70, as for example, shown in Figure 1. It may also he located before the influent inlet 30 or Interior to the cell 22 in a location that allows the desired sensing of the ion content of the influent solution 70 before it is treated by the water- Splttting membrane 100. In this locatlon, the lon senRer 59 generates an ion concentration signal that Is indicative of the ion sensor 59. The earlier the ion sensor 59 measurement can be determined in this embodiment, the more precisely can the ion concontration of the treated solution be controlled. The ion sensor 59 measures an ion concentration of the influent solution 70 and generates an Influent ion concentration signal relating to the influent ion concentration. The controller 53 receives the Influent ion c. oncentration siqnal and sends a control signal to the variable voltage supply 50 to adjust the time-averaged voltage level in response to this signal. Feedback loop 57. controMer S3 and variable voltage supply 50 adjust the vnltage to cell 22 In response to the control signal to increase the voltage if the influent so ! ut) on 70 contains a larger concentration of ions or to decrease the voltage if the influent solution 70 contains a smaller concentration of ions.

When the influent ton composition or concentrations are known and predictable over time. the controller 53 can also be preset to issue control signals that instruct the voltage supply 60 to pass redefined selected time average voltages during an ion removal or rejection step of tale cela 22. In this example, the controller 53 is preprogrammed with knowledge of the application, namely, the composition of the influent solution 70 and/or the behavior of cell 22 over time or with the Input volume of the innuent solution 70. The ion sensor 59 and feedback toop are absent. Controller 53 instructs variable supply 60 to adjust the voltage to cell 22 as a function of time or

volume without an ion sensor 59 or closed loop control to sense a change in the ion concentration of the influent or effluent solutions 70, SO.

In yet another version, the electrical resistance of cell 22 is used to determine or predict the conductivity of the effluent solution 60. as shown in Figure 9, to control TDS in the affluent without contrniting specific ion concentrations. In this version, the apparatus 20 comprises a voltmeter 62 to measure the voltage bias in the cell 22 and generate a voltage signal and an ammeter 60 to measure the current in the cell 22 and generate a current signal. The controller 53 receives a voltage signal from the voltmeter 62 and the current signal from the ammeter 60 and calculates an ion concentration in the solution from the ratio of the voltago and current signal^. The ratio of currentlvoltage represents the electrical resistance of the call 22 and is an indicator of the conductivity of the solution in the cell 22, and thus a predictor of the effluent solution conductivity. Feedback toops 57 trnm the ammeter 60 and voltmeter 62 provide the signals to controller 53 to allow the controller 53 to adjust the applied voltage to the cell 22 to maintain the ion concentration at a cesired level. The measured cet ! resistance is a sum of water and membrane resistances, and thus controller 53 can also Interpret the signals from feedback loops 57 to account for changes in the membrane electrical resistance over time. In operation, the current/voltage ratio measurement is used to estimate ion concentration, and the controller 53 sends a control signal to the variable voltage supply 50 to adjust voltage in response to tho measured current/voltage ratio.

Instead of a single cell 22, the apparatus 20 can a ! so comprise a plurality nt c., alls 22 with solution Inlets 30 and outlets 35 fluidly connected in a parallel or series arrangement. For example, an arrangement comprising two calls 2Ra, b that have their Influent Inlets 30a, b connected In parallel is shown in Figure 10a. In this arrangement, the influent water stream is split into two streams each of which provided to an influent inlet 30a, b of a separate cell 22a,b. The effluent solution exits at each of the cell's outlets 35a, and is recombined prior to passing through or past an ion sensor 59. The voltage supply 50 powers both cells 22a, b in parallel. The signal from ion sensor 59 is carried to controller 53 by feedback loop 57 and the variable voltage supply 50 is adjusted accordingly in response to instructions in a control signal from the controller 63. This version is useful for operation at lower pressures or higher flow rates.

In yet another arrangement, two or more cells 22 are fluidly connected in series, with the influent solution entering a first cell 22a at inlet 30a, exiting the first cell's outlet 35a. entering the inlet 30b of a second cell 226, exiting at the second cell's outlet 35b, and finally passes past the ion sensor 59, as shown in Figure 10b. The electrodes 40a, b, 45a, b of both cells 22a, b can be connected in parallel to a single voltage supply 50, or each cell can have a separate voltage supply 50. The controller 53 receives a signal from an ion sensor 59 relating to the ion content of the sequentially treated solution and controls the voltages applied to the electrodes 40a,b, 45a,b In each cell 22a, b to control ion removal in each cell 22a, b. In the version shown in Figure 10c, the first cell 22a is power by a constant voltage supply sOc that provides a single constant voltage to the cell 22a. and the second cell 22b is powered by a variable voltage supply 50. The ion sensor 69 Is located in the treated effluent solution 80 outside outlet 36b of the second cell 22b. In this embodiment, cell 22a provides maximum deiontzation as a pretreatment for the influent solution of the second cell 22b, where the ion content of the cell is controlled to achieve desired ion levels. Ion sensor 69, feedback loop 57, controller 53, and variable voltage supply 50 allow adjustment only of the voltage to ceil 22b and not cell 22a.

EXAMPLE 1: Constant Voltage This example demonstrates the loss in ion removal efficiency of a water splitting ion exchange membrane 100 over an inn removal cycle. A cell 22 comprising a spiral wound water-splitting membrane 100 in a cylindrical housing 25 was employed for a I L) content reduction experiment A constant voltage power supply with a current limit of 1 A was used so a feedback circuit was not employed. The cell 22 comprised 30 layers of membrane 100 between a centra ! electrode and an outer electrode. The cation exchange layer 51 of the membrane 100 faced outward and the anion exchange layer 49 Inward. Initially, the cell 22 was regenerated by an ion rejection step for 20 minutes at 160 mllmin and 300 V using a feed water stream of #50 µS/cm to prepare it for the doionlzation or ion removal step. During the regeneration step, the current to the electrodes 40, 45 was maintained at 0. 5 to 1-0 amps, and the influent solution 70 was flowed from the inside to the outside of the spiral membrane 100 in the cell 22.

Subsequently. in a deionization or ion removal step, the influent solution flow direction and polarities of the time averaged voltage applied to the electrodes 40, 45 of the celt 22 were reversed. Influent solution 70 comprising water with dissolved solids or ions and having an initial ionic conductivity of 504 µS/cm was pumped through the cell 22 at 1.5 liters/minute. The voltage applied to the electrodes 40, 45 was the maximum voltage available from the variable voltage power supply 50, namely, 300 V limited to 1 Amp maximum current. The result is shown in the graph of Figure 11, as the constant voltage curve in which full power was applied throughout the : ion exchange cycle, The early high conductivity is due to purging the last regenerant solution from the cell 22 for the first 30 seconds, followed by a fairly stable conductivity output for 4 minutes, then a steady increase in effluent water conductivity occurrcd to 250 µS/cm at 16 minutes. when the experiment was stopped. This example demonstrates that the ion content in t he effluent solution 80 gradually increased over time as the water . sptitting membrane 100 became satuoted wtth ions removed from the influent solution 70. Consequently, the ionic concentration of the effluent solution 80 did not remain constant during the treatment cycle, but Instead varied by a factor of 5 corresponding to an increase in conductivity from 160 to 250 µS/cm

EXAMPLE 2: Variable voltage In this example, the apparatus 20 comprise the aell 32 of Example 1 having an ion sensor 59 that was a conductivity sensor that was connected to a controller 53 comprising a microprocessor to interpret the ion concentration signal and control thf autput voltage of a variable voltage supply 50. The target ionic conductivity in this experiment was 125 µS/cm. Previous experiments determined the output voltages, which would be employed for various deviations of the ionic conductivity of the effluent solution So from the target conductivity. The result is presented in Figure 11 as the variable voltage curve. When the measured conductivity exceeded the 125 µS/cm target, the controller 53 increased the votiage to the cell 22, and the conductivity of the effluent 80 was reduced soon thereafter. The slow variation up and down for the effluent solution 80 is a result of the finite response time of the system. This is largely dependent on the flow rate and void volume of the cell 22. It is evident thot the effluont solution conductivity is controlled in this illustration to 125 25 µS/cm for a full'16 minutes, or 24 liters of treated solution. In contrast, the full (or single) power experiment in Example 1 provided treated effluent solution 80 in this tart range for only 3 minutes, or for only 4. 6 liters of volume.

The present apparatus 20 and methods allow treatment of influent solutions 70 to provide a selected or controlled ion concentration in the effluent solution 80. For example, they can provide a treated effluent solution 80 comprising water having controlled concentration of ions most suitable for a particular purpose, whether it be taste for drinking water, hardness for softened water, or a particular TDS content-for an industrial process. Tho present invention can also extend the working capacity of the delonization cycles of electrolytic ion exchange devices by avoiding the over- deionization of water, thereby preserving ion exchange capacity for further water treatment. By employing variable voltages rather than chemical systems, the complexity of the apparatus 20 Is reduced, as is the capital and operating cost.

Furthermore, the influent and effluent solutions avoid the risk of cross-contamination, which can occur when bleeding solutions to control ion concentrations. The apparatus 20 also provides good ion selectivity, resistance to mineral scale fouling, and concentrated regenerant effluent solutions typical of conventional ion exchange.

The present invention has been described in considerable detail with reference to exemplary versions thereof, however, other versions are possible, as would be apparent to one of ordinary skill in the art. For example, other arrangements of electrochemical cells can also be used depending on the influent ion concentration, volume, or treatment desired. Also, alternative voltage supplies can be designed to provide voltages suitable for other applications, for example, a single polarity voltage having an AC component. Further, relative terms, such as first, second, outer, inner are provided only to illustrate the Inventlon and are interchangeable with one another, for example, the first electrode can be the second electrode. Therefore the spirit and scope of the appended claims should not be limited to the description of the preferred versions contained herein.