Non-Fouling, Flow-Through Capacitor, System and Method of Separation

Reference to Prior Applications This application is a continuation-in-part application of U.S. Patent Application Serial No. 08/439,310, filed May 11, 1995, which application is a divisional application of U.S. Patent Application Serial No. 08/194,609, filed February 10, 1994, now U.S. Patent 5,415,768, issued May 16, 1995, which is a continuation-in-part of U.S. Patent Application Serial No. 08/027,699, filed March 8, 1993, now U.S. Patent 5,360,540, issued November 1, 1994, which is a divisional of U.S. Patent Application Serial No. 07/819,828, filed January 13, 1992, now U.S. Patent 5,200,068, issued April 6, 1993, which is a continuation-in-part application of U.S. Patent Application Serial No. 07/792,902, filed November 15, 1991, now U.S. Patent 5,192,432, issued March 9, 1993, which is a continuation of U.S. Patent Application Serial No. 07/512,970, filed April 23, 1990, now abandoned. U.S. Patent Application Serial No. 07/760,752, a divisional application of U.S. Patent Application Serial No. 07/512,970, was filed on September 16, 1991 and is now U.S. Patent 5,196,115, issued March 23, 1993. All of these patents and co-pending application are hereby incorporated by reference.

Field of the Invention It has been discovered that a flow-through capacitor can be constructed with multiple, generally parallel open flow paths. The capacitor of the present invention utilizes this design and method to provide a fouling resistant flow¬ through capacitor with a compact and easy to manufacture construction.

Background of the Invention Flow-through capacitors have been described in the prior art (see for example U.S. Patents 5,360,540, 5,192,432, 5,196,115, 5,200,068 and 5,415,768 to Andel an; U.S. Patent 3,658,674 to Benak; and PCT International

Application No. US94/05364 to Andelman. The publication by Allen M. Johnson et al "The Electrosorb Process for Desalting Water", March 1970, The Office of Saline Water Research and Development Progress Report No. 516, U.S. Department of Interior PB 200 056 and U.S. Patent 5,425,858 to Joseph Farmer describe flow-through capacitors or flow¬ through deionization systems.

Flow-through capacitors of the prior art function by concentration of solutes in the feed stream into a concentrated waste stream. This has two serious disadvantages. The first disadvantage is that the method of waste recovery into a concentrated waste stream generates waste water. Waste water of any kind is often a significant process cost. Another disadvantage of the prior art is tendency to foul. Solutes present in many ambient waters, such as Calcium Sulfate or Calcium Carbonate or other solutes, are present at or near saturation. When concentrated past the saturation point, such solutes have a tendency to form scales and foul the apparatus. Fouling is a well known problem in waste water apparatus, such as reverse osmosis and electrodialysis, that share with the flow-through capacitor the tendency to concentrate solutes past the saturation point. Commonly present turbidity, colloids, and particles may also block flow channels and form deposits. Benak, U.S. Patent 3,658,674, describes a flow-through capacitor that rapidly plugs up when used with hard well water (column 4 lines 66-68). He further describes (column 5, lines 27 through 37) problems with the purification of Calcium Sulfate. Prior art U.S. Patents 5,360,540, 5,192,432, 5,196,115, 5,200,068 and 5,415,768 by Andelman, U.S. Patent 3,658,674, the publication by Allen M. Johnson et al "The Electrosorb Process for Desalting Water", March 1970, The Office of Saline Water Research and Development Progress Report No. 516, U.S. Department of Interior PB 200 056, and U.S. Patent 3,658,674 by Benak all describe flow-through

capacitors with a porous spacer layer. The pores in the spacer layer can be clogged up and fouled by turbidity already present in the feed stream, and well as by precipitates and crystals that form during use. U.S. Patent 5,425,858 by Joseph Farmer describes a flow-through capacitor whose spacer layers define an open channel with a long serpentine flow path. In the Farmer patent, the flow path is not also directly open to the outside, but is circumscribed by a gasket and forced to flows between holes in the successive multiple structural layers, and thence through an outlet. Serpentine channels provide bends where crystal and precipitates can settle, thereby blocking the flow path. Fluid flow-through holes in many successive structural layers further constrict the flow and offer many places for fouling to occur and multiple opportunity for solids to plug the flow path. A short flow path is preferable in order to flush the saturated waste out of the capacitor before the kinetic process of crystallization forms precipitates and causes fouling or crystals. The long flow path that a serpentine channel provides makes it difficult to flush the capacitor before crystallization occurs. Once solids do form inside the capacitor, it is that much more difficult to flush them through a long enclosed flow path. The Farmer patent suffers from other disadvantages. The plate frame design depends upon multiple sealing gaskets. This provides multiple opportunities for leaks, and therefore requires hardware such as heavy, structural, metal end plates and threaded rods to tightly compress the stack. The structural, metallic end plates are conductive, and therefore, in addition to the gasket spacers, require an extra non-conductive insulator layer between the end electrodes and the end plate. The electrodes in the Farmer device consist of a titanium metal sheet sandwiched between two high surface are layers. To effect leak tight seals, the titanium metal sheets also have to be thick and structural, rather than the thin metal foils described in

the Andelman patents. This adds to the cost and bulk of the system. Moreover, titanium is not the best choice, as titanium is a valve metal, and forms a non-conductive oxide coating under conditions which may occur during use of the flow-through capacitor. Finally, the Farmer device uses a conductive epoxy rather than a compression contact between the titanium sheets and the high surface area materials. This unnecessarily increases resistance of the electrical contact. The use of glue to form a contact would also shorten the lifetime of the capacitor device due to eventual deterioration of the bond.

There is therefore a significant and unfulfilled need for a new and improved flow-through capacitor apparatus, method and system. Such a capacitor would be resistant to fouling. It would furthermore be desirable that such a capacitor reduce or even eliminate the waste water entirely. Finally, it would be desirable for such a capacitor to be easy to manufacture, and not have any unnecessary parts that increase cost, or limit the usefulness of the capacitor. Summary of the Invention

The invention relates to a foul-resistant capacitor, system and method, and in particular, a capacitor, system and method for the separation of solutes or fluids which tend to clog or foul the capacitor. The invention comprises a flow-through capacitor having at least one anode and at least one cathode adapted to be connected to a power supply, the capacitor arranged and constructed for use in the separation, electrical purification, concentration, recovery or electrochemical treatment or breakdown of solutes or fluids, particularly solutes and fluids which are saturated or substantially saturated and which tend to foul the capacitor.

The capacitor includes one or more spaced apart pairs of anode and cathode electrodes incorporating a high surface area electrically conductive material and characterized by an open, short solute or fluid flow path, which flow paths

are in direct communication with t-he outside of the capacitor.

The fouling-resistant, flow-through capacitor of the invention is able to treat saturated solutions. The flow- through capacitor takes into account that subsequent formation of fouling precipitates and crystals from a super¬ saturated solution is a kinetic process. Therefore, the flow-through capacitor design is configured such that the fluid flow path through the capacitor is short, generally straight and open. To allow optimal and unobstructed wash out of solids in the capacitor, it is also desirable for this flow path to be open directly to the outside of the capacitor and not have to pass directly through multiple holes or constrictions. Therefore, the open channels created by the spaced apart electrodes communicate directly with the outside surface of the capacitor. These channels are not circumscribed by a gasket, but preferably have one dimension completely free of obstruction to flow. Where possible, it is also preferable to make the flow path wide. A short, straight, open flow path, communicating directly with the outside surface of the capacitor has many advantages. A waste or feed solution, including a saturated or a super-saturated solution, can be flushed out of the capacitor before the process of crystallization takes place. Should the waste or feed solution be left in the capacitor long enough for crystals to form inside the capacitor, such fouling crystals can be flushed out through the straight and open flow path. Turbidity or particles present in the feed can also be flushed out through the capacitor without fouling the system. Because the flow path is directly open to the outside, it is also possible to clean mechanically between the electrodes without disassembling the capacitor.

The generally straight, short, open, wide flow path of the solute can be accomplished by replacing the porous spacer of the capacitors with multiple thin strips, shims, washers or open netting (such as sold by Nalle Plastics Inc. of Austin, Texas under the tradename Naltex) , preferably

including bidirectional filtration netting. Spacers may be any inert, non-conductive material, such as, but not limited to: fluorocarbon polymers, like Teflon®; ceramic beads; washer shapes; individual shims; or plastic netting, preferably biplanar filtration netting. The spacers may include microprotrusions screen printed onto the electrodes as disclosed in U.S. Patent 5,384,685, "Screen printing of microprotrusions for use as a space separator in an electrical storage device", to Tong et al, hereby incorporated by reference. In short, any material in any shape which is thin in cross section and electrically non¬ conductive can be used to space apart the anode and cathode electrodes of the capacitor.

Spacers need not be separate layers, but may also be built into internal or external supports. For example, an internal supporting rod may contain risers or notches which are used to space apart the electrodes. The same thing may be accomplished with an external support or scaffold. To keep internal resistance of the capacitor as low as possible, the space between the electrodes should be as narrow as possible. Too narrow a flow path, however, increases the possibility of accidental shorting between the layers and enhances fouling. The optimal spacer thickness in one embodiment is less than about 50 mils, for example, between about 5 and 20 mils.

Because the present invention does not require tightly sealed, gasketed spacers to form a flow path, the spacer material need not be elastic or rubbery in nature. The spacer material serves only to space the electrodes apart and not to form a seal in the capacitor. Therefore, only enough compression is required to hold the layers together, or to form an electrical contract between the conductive high surface area layers and any optional conductive backing layers to the electrodes. Either no end caps are required or less mechanically strong, non-conductive materials can be used, such as plastic materials. Therefore, the heavy structural metal end plates and connecting threaded rods can

be eliminated. As a consequence, the insulator layer of the former or prior art devices disposed between the end plate and the end electrodes may be eliminated. There are many possible geometric configurations that combine a generally straight and open flow path. Preferably, the flow path should alεo be short, and directly communicate with the outside surface. Spacers may also and simultaneously function to provide internal support for a compression contract between high surface area layer conductive backing layers. Where possible, a wide flow path is also desirable. Generally, the present invention in the preferred embodiment employs multiple straight, parallel flow paths through the spaced apart electrode layers.

The width of the flow path should be short and, like the thickness, less than about 50 mils, and more typically, about 5 to 20-30 mils. The length of the flow path should alεo be short and should usually be the lesser dimension of the X-Y-Z dimensions of the capacitor, and for example, less than about 12 inches, for example, about six inches or less. The electrodes may be made from any high surface area material. Where it is desirable to enhance the conductivity of the high surface area conductive material, a further conductive backing layer may be employed in direct contact with the electrodes. Where the intrinsic electrical conductivity of the electrodes is sufficient, this conductive backing layer may be eliminated. It is desirable to keep internal resistance capacitor as low as possible. Internal resistance sets a limit on charging time of the capacitor, which in turn directly limits the ultimate mass and fluid flow rates that it is possible to purify solutes and solutions. An internal resistance of less than about four ohms, e.g. one ohm, is preferred.

The electrodes may be made out of any monolithic high surface area conductive materials, in at least one anode/cathode pair. Where the high surface area material is conductive, but not optimally conductive, an electrical conductive backing may be employed. High surface area

conductive materials suitable for use in the present invention include, but are not limited to: activated carbon; activated carbon treated with a halogen; carbon foams; carbon aerogel and aerogel composite materials; nanotubes; conductive polymers, especially in porous or network form; polymerized fullerenes; or any high surface area conductive material may be used. Conductive ceramics may also be used, either by themselves or impregnated onto high surface area substrates, including various forms of carbon such as fiber, foam, powder or aerogel. In general, absorbing any electrically actuated small or large molecule onto the conductive high surface area material that improves the capacitance will improve the function of the capacitor. Another preferred high surface area conductive material is conductive transition metal oxides, nitrides, or borides prepared using sol/gel technique. Powdered high surface area materials may be sintered into monolithic electrodes or bound together with binder materials.

Intrinsically conductive electrodes where no backing layer is required would include high surface area preparations of graphitic carbon, high surface area expanded metals, metal fibers, or metal meshes. For example, titanium fibers coated with high surface area platinum series black are known and are marketed as electrode materials. Other examples include platinum coated niobium and foamed metals. High surface area carbon materials may be mixed with metal or graphitic fibers or meshes and formed into monolithic units.

Another type of electrode includes a non-conductive high surface area material in intimate contact with a conductive backing. An example of this would be an oxide layer on etched tantalum or aluminum, oxidized conductive ceramics, or thin films deposited on any high surface area conductive material. A capacitor with this material has the advantage that the solution is protected from direct contact with a conductive material. Therefore, the capacitor can be operated under much higher voltages, up to 50 volts or

more, as opposed to a naked high surface area conductive material, which is limited by the breakdown voltage of the solution.

The invention will be described for the purposes of illustration only in connection with the embodiments; however, a person skilled in the art may make various modifications, changes improvements and additions to the embodiments or illustrated embodiments, all without departing from the spirit and scope of the invention. Brief Description of the Drawings

Figs. 1 and 2 are schematic, exploded views of a washer-type capacitor of the invention;

Figs. 3A-G are representations of various spacers which may be used in the capacitor of the invention; Fig. 4A shows an assembled capacitor cartridge of Fig. 1 or Fig. 2;

Fig. 4B shows the assembled capacitor cartridge of Fig. 4A within a cartridge holder with a partial cutaway view;

Fig. 5 illustrates in schematic view a short, fat, spiral wound capacitor of the invention in partial assembled form;

Fig. 6 illustrates the assembled, spiral wound capacitor of Fig. 5 within a cartridge;

Fig. 7A illustrates in partial assembled form a crescent-pleated capacitor;

Fig. 7B illustrates the assembled capacitor of Fig. 7A;

Fig. 8 shows a perspective, illustrative, partially cutaway view of a multiple manifold, flow-through capacitor of the invention; Fig. 9 illustrates in an exploded view a disassembled, flat-type capacitor of the invention;

Fig. 10 shows an assembled, flat, flow-through capacitor of the invention in a box-like cartridge;

Fig. IIA is a schematic, perspective, partial view of a hexagonal rod capacitor of the invention;

Fig. IIB is a schematic, perspective, partial view of another embodiment of a hexagonal capacitor of the invention;

Fig. 12 is a schematic, perspective, partially exploded view of a hexagonal rod capacitor of Figs. IIA or B in a cartridge;

Figε. 13A-F are schematic illustrations of various monolithic electrodes for use in the capacitor of the invention; Fig. 14 is a schematic illustration of a system employing the capacitor of the invention;

Fig. 15 is a schematic illustration of a multiple capacitor system of the invention;

Fig. 16 is a graphical representation of experimental data of electrical conductivity of a NaCl solution in microSiemens versus cumulative volume of fluid in ml;

Fig. 17 is a graphical representation of experimental data of electrical conductivity of a saturated CaS0 ₄ solution versus the number of charge and discharge cycles of the capacitor.

Different flow-through capacitor geometries can be employed that all share in common a short, straight and open, fluid flow path. All of such preferred designs in the drawings share in common multiple, parallel flow paths within the same capacitor, which for εimplicity, is referred to in the singular in the below discussion. All of the capacitors also have in common a flow path that is in direct communication with the outside. Figs. 1 and 2 show capacitors with washer-style electrodes about a central tube or rod. The flow path is between the washer electrodes and out through the central tube. Fig. 6 shows a short, fat spiral wound arrangement, where the flow path is between the electrodes in the longitudinal direction, parallel the central axis. The spiral wound capacitor may contain a structural central rod or tube. The electrodes may also be fashioned as interlocking or overlapping pleats in various flat or cylindrical geometries. Figs. 7A and 7B show a

crescent pleat design where the electrodes are placed as overlapping pleats about a central tube, where the flow path is between the electrodes in the horizontal direction, and thence longitudinally out through the central tube, which contains holes along its length. Alternatively, it may be advantageous to replace the central tube with a fluted rod, such that the flow path is through ribs or channels along the length of the rod. Fig. 9 shows a flat capacitor made up of rectangular electrodes, with the flow path across the short dimension of the electrodes, between the electrodes. Alternatively, these multiple electrodes could be replaced by two or more interlocking pleated electrodes.

In addition to shortening the flow path, the capacitor should be operated so that polarity is reversed in every charge cycle, so that no net buildup will occur on the electrodes. Since the electrodes are electrically active, reversing polarity makes the former anode into a cathode and the former cathode into an anode. Precipitates and deposits which may have favorably formed on one of either the anode or cathode will therefore be driven off, thereby further reducing fouling of the capacitor. To further minimize fouling, the flow-through capacitor should be backwaεhed by reversing the flow, for example, every charge or discharge cycle. This also helps avoid buildup of solids because there is no net flow in any one direction. It may also be desirable to continuously or occasionally operate the capacitor at slight overvoltage. This causes a small of electrolysis which keeps the electrodes clean of microbes and foulants. The combination of short, straight, open flow path, polarity reversal and optional backwash enables the flow¬ through capacitor of the invention to purify even saturated solutions, and to concentrate such solutions all the way until crystallization occurs. This phenomenon generates a further advantage. The fouling resistance offered by the straight, open flow path of the capacitor of the invention allows the flow-through capacitor to be operated to generate

97/13568 PCMJS96/16157 a εolid instead of a liquid waste. To achieve this result, the flow-through capacitor is connected to a tank of saturated waste. During the regeneration cycle, the capacitor is filled with saturated solution from the waste tank. The capacitor is discharged into this saturated water, which is then fluεhed back into the waste tank. Precipitation occurs because, upon desorbing its ions, the capacitor super-saturates this saturated waste water. The straight and open flow path of the present design allows the precipitates to be flushed directly out of the capacitor for recovery or removal. The precipitate settles into the bottom of the waste tank, which can be collected separately by decanting or filtration. This process can be repeated indefinitely. The waste solution may alternatively, instead of forming a precipitate, become super-saturated. In this case, the solution can be triggered to precipitate by heating, cooling, vibration, seeding with small crystal, adjusting pH, or other precipitation methods.

A further disadvantage of prior art that is overcome by the present capacitor design is the dilution of product solution with dead volume of the capacitor. The capacitor described in U.S. Patent 5,425,858 has a dead volume of 250 mis. This dead volume is deleterious, because waste solution left over after discharge needs to be flushed out with fresh feed solution, therefore generating additional waste water. If this waste solution is not adequately flushed out, then, upon charging the capacitor during a purification cycle, the capacitor repurifies in an inefficient manner the concentrated waste still present in the capacitor. This inefficiency gets worse with increasing feed solution concentration, due to the fact that the capacitor in this situation gets saturated faster, and needs to be regenerated more often. The more often the capacitor is regenerated, the greater the opportunity for contamination of product solution with the dead volume.

The flow-through capacitor of the present invention minimizes the dead volume problem, such as by connecting the

capacitor to a source of pressurized gas, such as air, N ₂ or other gases chosen not to react with or contaminate the product. The short and straight flow channels are optimal for allowing the solution inside to be displace with a gas. In contrast, liquid contained in εerpentine channels of a prior art capacitor would be more difficult to displace due to the tendency to form protected pockets. This displacement of dead volume with a gas or fluid is also difficult with the porous spacer design of the other patents, due to channeling. Any preference or unevenness in manufacture would cause channeling of gas or liquid fluids, and uneven displacement of the waste liquid with the gas. None of the prior art utilizes a gas to displace the dead volume. Two or more foul-resistant, short flow path capacitors can be operated continuously in a system, one charging while the other discharges. To recover energy, a discharging capacitor can be used to charge other capacitors. Single charged capacitors can only be used to charge another capacitor until their voltages equalize. Among two capacitors of equal size, this wastes half of the energy in the discharging capacitor. However, two or more of these half charged capacitors can be connected in series. The series voltage of capacitors connected in series is additive. Connection of capacitors in series allows continued use of the depleted capacitors energy to charge other capacitors. U. S. Patent 5,475,858 describes the use of single capacitors for energy recovery to recharge other capacitors; however connection in series makes it possible to recoup energy from the remaining half of the unutilized energy.

Multiple capacitors in a system allow for alternating and continuous, simultaneous charge and discharge in order to provide uninterrupted product flow. Uninterrupted flow can also be achieved with a single flow-through capacitor by including a hold-up tank downstream that is used to average

the flow between charge or discharge cycleε. Thiε allows for a simpler single capacitor system design.

A further improvement of the flow-through capacitor involves doping the carbon electrodes with a metal, εuch as platinum, titanium, or other metals with catalytic properties. This allows the capacitor to more easily electrochemically deεtroy chlorinated hydrocarbonε, chloroform and other hazardouε organic molecules. The metals overcome the activation energy of reactions such as removing a halogen from a hydrocarbon, or breaking down an organic molecule.

A preferred way to control the flow-through capacitor is to use the capacitor as its own sensor. In this syεtem, both current and time are recorded, with the current measured through a shunt resistor. A computer integrates current versus time to calculate total charge that has been transferred into the capacitor. After a preset value of total charge has passed into the capacitor, a controller automatically activates the regeneration cycle, which involves firεt disconnecting the power supply, waiting a short time to safeguard the electronics, shorting the capacitor through a load, and actuating the appropriate valves and pumps that select and isolate the waste stream from the product stream. Description of the Embodiments

Fig. 1 showε a stacked washer flow-through capacitor, whose high surface area electrodes contain a backing layer. The electrodes consist in combination of electrically conductive high surface area material 1 and conductive backing 2. The end electrodes may be either single or double sided, whereas the intermediate electrodes are preferably double sided. The electrical contact between the high εurface area layer 1 and the conductive backing layer 2 iε preferably a compression contact, which is afforded by the screw on end caps 7 tightened around central rod or tube 5 around threads 6. The electrodes are present in even numbers to form at least one anode/cathode pair. The anode

and cathodes so formed are εeparated by spacerε 3. Integral leads 4 extend from conductive backing (2).

These leads may be joined together to connect separately, in parallel allignment to themselveε, the alternate anode and cathode layers, or they may be gathered together to accompliεh the εame purpoεe and to form an electrical lead.

Fluid flow iε between the spaced apart electrodes and through the holes 9 and then out through the central tube 5. Instead of a tube with holes, a ribbed rod may be substituted with fluid flow alongside the longitudinal ribε. Washer meanε 8 are provided to allow the cartridge to form a leak proof εeal inεide cartridge holders (see Fig. 4).

Fig. 2 showε a waεher-εtyle flow-through capacitor with high εurface area electrodes that are sufficiently conductive that no conductive backing is required. Integral leads 4 are attached to high surface area conductive material 1 , which forms alternating anode cathode pairs separated by εpacers 3. Figs. 3A-G show various styles of spacers that may be used in the flow-through capacitor of the present invention. The spacer may be in the form of an open net (Fig. 3A), such aε manufactured filtration netting (e.g. Naltex) . This netting may be symmetrical or non-symmetrical. Biplanar netting is preferred, because this sort of weave does not obstruct the flow path along the surface of the netting. The spacer material may also consist of protrusions, such as the screen printed microprotrusions of U. S. Patent 5,384,685 by Robert Tong et al. One preferred embodiment would be washer-εtyle circular εpacers, such as the star- shaped spacer (Fig. 3C) . Another preferred embodiment (Fig. 3D) would include individual shi ε, small rods, or threads, laid out between the anode/cathode layers to space these apart. Also depicted are εpacerε that are integrally formed from a central or external εupport, εuch aε spaced notches in a central tube, or an external scaffold arrangement with spaced risers whereupon the electrodes are placed (Fig. 3E) .

Under some circumstanceε, it may be desirable to deviate from a short straight flow path and provide instead for flow paths which are multiple, parallel, and generally open to the outside, but are alεo constricted; that is, long and serpentine. In this case, spacers as depicted in Fig. 3F and Fig. 3G may be used, as long as they are tightly sealed against the electrode surface. Constricting the flow to a spiral or serpentine flow path increaseε the linear velocity, which may have the advantage of providing greater turbulence and εcouring action to remove foulants. This multiple flow serpentine flow path eliminates the likelihood of a severe pressure drop and provides a more effective technique over the prior art. While these spacers are shown to be used with the washer-style capacitor of Figs. 1 and 2, it is recognized that they may be adapted to be used with other capacitors.

It should also be noted that the multiple parallel flow paths allow the addition of extra length to the individual flow paths without a problem with pressure drop due to excesεive elongation. Flow paths that are short and εtraight in the εtricteεt sense are an ideal that is preferred in most cases. An exception would be, for example, a gel-like foulant such as iron hydroxide formation, which would require a turbulent or very fast flow to break up. The most common example requiring the εhorteεt, straightest and widest flow path possible would be rapidly crystallizing solutions. In either case, the flow¬ through capacitor of the invention enables optimum function of the procesε due to the utilization of the εhorter, multiple flow path deεign as desired.

Figs. 4A-B show the assembled flow-through capacitor of Figs. 1 or 2. The εpacerε create open εpace 10 between the high εurface area conductive material 1. The washer εhaped electrodeε are assembled around central tube 5 and held in place by end caps 7. Leads 4 interconnect the alternate electrodes to form anode/cathode pairs. Anodes are connected to anodes and cathodes connected to cathodes, in parallel

arrangement. Fig. 4B depicts the asεembled flow-through capacitor of Fig. 4A in a cartridge holder 17. The cartridge holder iε fitted with a εcrew-on lid 15 with threads 16. The lid contains two graphite rods 12. When lid 15 is screwed down upon cartridge holder 17, graphite rods come into electrical contract with the two concentric electrical contacts 18. These two contacts connect electrically to the bundled leadε, which in turn interconnect the anodes and cathodes to themselves. The graphite rods 12 are attached to a spring 11 which provides a controlled tension and downward force that makes an electrical contact with concentric conductor 18. Metal caps 13 form a compression electrical contract with graphite rods 12 to connect electrically to wire leads 14, for connection to a DC power supply. The lid 15 is provided with an inlet

19, and the cartridge holder 17 is provided with an outlet

20. Washer meanε 8 εeal against both the lid and the cartridge holder to form leak tight sealε. The flow path is through the inlet, then between the spaced apart electrode layers, then out through the outlet.

Fig. 5 depicts a spiral wound capacitor of the present invention utilizing conductive high surface area material 1, optional conductive backing 2, and spacer material 3 in a netting or open mesh form. Electric leads 4 extend from the electrodes formed from material 1 or the optional conductive backing 2. The capacitor may optionally be wound around a structural central rod 5. This capacitor is preferably made short and fat, with the width wider than the length of the capacitor as measured down the central axis. Fig. 6 depicts the short, fat capacitor of Fig. 5 in a cartridge holder or piece of pipe 17. The pipe 17 is fitted with lids 15 containing inlet 19 and outlet 20. Integral leads 4 are bundled together in parallel. Wire leads 14 extend in a leak tight sealed manner through lid 15. A contact 23 is made between wire leads 14 and bundled leads 4. In caseε where integral leadε 4 are graphite foil, a compression contact is preferable, optimally using a gold or

inert metal contact. The capacitor is sealed inside a shrink wrap plastic tube 22. In this embodiment, the central tube or conduit has one or more holes at a εelected position in the tube, generally at the middle of the tube, and usually spaced apart about the periphery of the tube. Instead of sealing the ^• capacitor layers themselves, the central tube is connected at both ends to outlet 20 (Fig. 6), or sealed at one end and connected at the other end to outlet 20 (Fig. 6). A gap is required in the surrounding spiral wound high surface area material layer (or layers) and optional conductive backing layer (or layers) by cutting slots or holes 21 (Fig. 6) therein, and aligning the cut holes or εlotε with the holeε in the central tube to form a radial flow path. In this embodiment, fluid flow passes in the inlet 19, then moves both downwardly and upwardly between the electrode layers, through the space created by the spacer mesh or net spiral, until the fluid reaches the central gap or hole in the central tube, and the fluid is withdrawn from one or both ends of the central tube or conduit, then on through the outlet 20. Alternatively, a spiral wound capacitor can be tightly fitted inside a pipe or cartridge 17 with O-ring gaskets, in the same fashion as depicted with the bundled rod capacitor of Fig. 12. Here, the flow path is in through the inlet 19, down through the capacitor between the electrode layers, through the space created by the spacer mesh or net, and thence on through outlet 20. Lids 15 may also be in the form of end capε. Lids 15 and pipe 17 may be any material including metal, plastic, or ceramic, including PVC, Teflon, or stainless steel.

Fig. 7 depicts a crescent pleat design, where the individual layers do not complete a circuit around the center axis. The layers are made up of high surface area conductive material 1, optional conductive backing 2, and netting or open mesh material 3. Integral leads 4 lead directly off of material 1, or from conductive backing 2, where thiε optional conductive backing layer is employed.

The layers are formed around central tube 5. Ends are sealed with any sealing means 24, including resins, such as urethane, epoxy, thermo-molded plasticε, etc. The flow path iε between the spaced-apart, conductive high surface area layers along the space created by spacers 3, through holes 9 and out through central tube 5. This flow path is not exactly straight, as it follows the curve of the crescent pleats along the shorteεt direction.

Fig. 8 depicts multiple capacitors of any style, such as for example the capacitorε of Fig. 7, manifolded together with manifold plate 25 inside a single cartridge holder 17. Integral leads 4 are connected anode to anode and cathode to cathode by wires 14 with compression contacts 23. As shown in Fig. 6, wires 14 extend in a leak tight manner through lid 15 fitted with inlet 19. Flow path is through inlet 19, between spaced apart electrodes in the individual capacitors, then combined after flowing through manifold plate 25, thence out through outlet 20.

Fig. 9 depictε a box-εtyle flow-through capacitor. High εurface area electrodes 1 are rectangular shaped, and are spaced apart by spacers 3. Spacers in Fig. 9 are depicted as shimε, but may alεo be thin rodε, threads, nets or open mesh, protruεionε, or an outεide εcaffold. Integral leadε extend from electrodes 1, and are bundled together in parallel, anode to anode and cathode to cathode, against electrode lead collectors 28. Electrode lead collectors are εcrew tightened against inside lid 15 and washers 29 with fasteners 30 to form an electrical connection with wire leads 14. The capacitor is placed snug inside box 26, which may be polygonal or circular in shape. Leak proof lids 15 with inlets 19 and outlets 20 are fitted onto box 26.

Fig. 10 depicts the assembled flow-through capacitor of Fig. 9. The flow path is in through the inlet, between spaced apart high surface area layers 1. Optionally, the high surface area layers may sandwich a conductive backing.

Fig. 11 depicts a rod-type electrode design. The conductive, high surface area material 1 is in the shape of circular or polygonal, e.g. hexagonal, rods. This conductive material may have a central conductive rod or tube 32, to form a conductive backing for the high surface area material. High surface area electrodes 1 are spaced apart by εpacer εcaffold 33 connecting central conductive rodε 32. Alternatively, the high εurface area electrodes 1 may be spaced apart by protrusions or shims 34. Wire leads 14 connect alternating electrodes anode to anode and cathode to cathode in parallel connection. This forms anode/cathode pairs exactly similar to flat electrode designs. Wire leads 14 are attached to central conductive rods 32 via metal cap connector 31. The metal cap connectors 31 are preferably an inert metal, such aε gold or platinum, and form a compression fitting where the central conductive rods 32 are graphite. Central conductive rods may be any inert conductive material. Where the central conductive rods 32 are metallic, metal cap connectors 31 can be omitted and the rod may simply be drawn out or directly attached to wires 14.

Fig. 12 is an illustration of an assembled rod-style capacitor with rods εealed together with εhrink wrap plastic tube 22 and held against cartridge holder 17 with leak proof O-ring 21. Wire leads extend in leak tight fashion through lid 15, which seals against cartridge holder 17. Flow iε through inlet 19, between spaced apart high surface area conductive polygonal electrodeε, and out through outlet 20. Figs. 13A-F depict various monolithic electrode designs that incorporate an inner conductive backing. This is useful for all the above flow-through capacitors because a compression fitting is no longer required to make a contact between the high surface area layers and the conductive backing layers. The electrodes of Fig. 13 contain an inner conductive backing layer 2, which may be a metal foil, graphite foil, a fibrous material, or an interpenetrating

network mesh material. In foil form, this backing material has many holes 35 therethrough to allow communication and interconnection with the high surface area material that forms a sandwich on both sideε in a flat electrode. Alternatively, a rod-style conductor can be used, with the high surface area material 1 formed directly around a central rod or wire conductor 32. This material iε bonded together or calcined as a εingle, monolithic piece, containing the conductive backing internally. For example, activated carbon or aerogel powder may be mixed with a phenolic binder and hot preεεed to form the εhapeε in Fig. 13, prior to calcining in the absence of air. The interconnections formed through the holes in the conductive backing hold the high surface area material together and prevent it from pulling away from the backing due to εhrinkage during calcining. Alternatively, carbon filmε or layers may be deposited onto conductive backings, and activated in place. Integral leads 4 are formed from the internal conductive backing layer or rod 32. Fig. 14 represents a diagram of a flow-through capacitor system. The flow-through capacitor 36 connected to manifold valves 37 and DC power supply 40 are controlled by computer or programmable logic 39. The feed solution 49 contains εolution that is pumped or fed through capacitor 36 during the charge cycle to form purified product. Hold-up tank 46 pools the product flow produced during the charge cycles to provide an averaged, continuous product flow 53. Counter 55 measures how much current versus time flows into the capacitor, preforms an integral of current versus time to calculate charge, and automatically starts the discharge cycle after a preprogrammed amount of charge has passed into the capacitor. This charge is a cutoff point that correspondε with a deεired level of saturation of the capacitor. Alternatively, the capacitor may also be controlled with senεors that monitor conductivity, pH, or concentration of feed, product, and waste solution. At thiε point, the

power supply is first disconnected, and, after a short time interval for safety purposes, the capacitor is discharged through a load. During discharge of the capacitor, waste εolution εtored in tank 43 iε εhunted through three way valveε 38 and manifold valveε 37 and pumped through the capacitor in a loop back into εettling tank 41. The ions abεorbed during the charge cycle of the capacitor are desorbed and discharged into this waste solution during discharge of the capacitor. The waste stream 50 gradually becomeε εaturated or supersaturated. Settling tank 44 is fitted with a heat exchanger 44 which functions to change the temperature of super-saturated wastes in order to provoke crystallization from solution.

Cryεtals 42 that flow into or form in the tank 41 settle to the bottom, where they may be recovered through outlet 73. The saturated waste is decanted into tank 43, for use in the next discharge cycle. Saturated waste may be bled off directly through outlet 52. This may be replaced with make up solution through make up loop 51, which cross connects with feed stream 49. After discharge of the capacitor, saturated waεte iε returned to the tanks 41 and 43. Manifold valves 37 then actuate to close off the capacitor to both the waste and the feed streams, and open it up to a compresεed gas stream 47. Thiε gas stream displaceε the waste solution from the capacitor, where it is driven into gas separation tank 56. Saturated waste liquid from tank 56 is recombined with the saturated waste εtream through three way valve 57. Gas is exported from tank through outlet 48. It may also be deεirable to likewiεe displace the feed solution with compressed gas after a charge cycle. Compresεed gas stream 47 may be replaced with steam, which reactivates activated carbon electrodes and sterilizes the capacitor from microorganisms. Shunting loop 58 allows the direction of the feed solution through the capacitor to be reversed on alternate charge cycles in order to backwash the capacitor. This further prevents net build up of turbidity and foulant.

All liquid solutions should first be fed into the bottom of the capacitor in order to diεplace all of the air, after which the flow may be maintained in thiε direction or reverεed via εhunting loop 58 to provide a backwaεh cycle. Finally, reconditioning tank 45 containε reconditioning εolutionε that are uεed to remove εuch abεorbed materialε, such as metals which electroplate onto the electrodes, organic foulants, or microbiological contaminants. It may also be important to adjust the pH of the feed solution, especially when dealing with eaεily plated metalε. The object iε to adjuεt the pH or add other ingredients to the feed solution in order to keep the plating voltage as high as poεsible. This allows the capacitor to operate more effectively via its capacitance mode and electrostatic absorption rather than by electrowinning and electroplating of metals onto the electrodeε. Electroplated metalε require acidε, bases, oxidizers, solvents, or other additional chemicals to recondition the electrodes and recover the plated metals, via reconditioning tank 45. As desired, a pump 75 may be included along feed streams, such as feed stream 49; waste εtreamε, such as waste stream 50; or any other streamε.

Fig. 15 shows a multiple capacitor system, with continuously charging and discharging capacitors to achieve continueε product flow, aε oppoεed to the single capacitor and flow averaging tank of Fig. 14. The flow path of capacitors 36 may either be connected in parallel, or may alternatively be connected in serieε through εhunt 54. Serieε connection of the flow path allows successive staged purification of a concentrated solution to very high purity product solution 53. Other components function and are as numbered in Fig. 14.

In another embodiment, it may be desirable to provide baffles in one or more or all of the flow passages to create a flow turbulence to the fluid through the flow passages to scour the surfaces to discharge foulants or to remove foulants from the flow paεεageε. The value of the uεe of

baffles for flow turbulence εhould be balanced with any pressure drop which results from the use of such baffles.

In order to create turbulence, baffles can be formed between the electrode layers, either constitutively with the electrode material or more easily with the spacer material. Increasing the flow rate of the capacitor may also be desired. The baffles create turbulence, which scours foulants from the electrode surface.

Other methods to prevent and treat fouling include washing the capacitor with purified product solution. This can be accomplished simply by reversing a pump 75 placed along the feed or product stream, and pumping purified solution back from hold-up tank 46 depicted in Fig. 14. Alternatively, an extra shunting loop may be added. Organic foulantε tend to accumulate on the electrode surface. Microorganismε tend to grow and multiply on the electrodes, especially activated carbon. Temporarily running the capacitor at a high enough voltage to cause the electrolysis will actively clean the electrodes. Also, continuously running the capacitor at a slight over-voltage to cause continuously a little electrolysis will also keep the electrodes clean.

The capacitor may be part of a train that includes other water purification technologieε, including electrowinning, ion exchange, RO, ED, microfiltration and ultrafiltration, either upstream or downstream in the capacitor. For example, it may be desirable to add microfilters aε a prefilter to remove fouling causing turbidity, or as a postfilter to remove carbon fines. A granular activated carbon prefilter may be desirable to further protect the capacitor from organic foulants. Greensand or oxidation means may be employed upstream of the capacitor to remove iron, which forms a slimy hydroxide that would tend to foul the capacitor. To produce ultrapure water, it may alεo be deεirable to include a deionization ion exchange bed downεtream of the capacitor aε a final polishing step.

Example 1

A stacked washer flow-through capacitor of Fig. 2 was assembled using sintered activated carbon for the high surface area layer (1) and 5 mil thick graphite foil from Polycarbon for the conductive backing (2). The sintered activated carbon washers were obtained from Kansai Coke and Chemicals, Ltd. These washers had an outside diameter of 1.5 inches and an inside diameter of .75 inches. They were .046 inches thick. The specific density waε .7 gramε/ml, and the ratio of activated carbon to binder was 7:3. The activated carbon used had a BET surface area of 2000 square meters per gram. The central tube was 1 inch long, 3/8 inch OD, and 1/4 inch ID PVC, with holes drilled in the side with end threadε (6). End caps were .25 inch thick PVC, 1.8 inch diameter, with internal threads that allow the end caps (7) to screw onto both ends of the central tube (5). Spacers consisted of six Teflon shims .75 inches long, .1/16 inch wide, and .01 inches thick. These were laid out radially on top of the electrodes, as depicted on Fig. 3D. The anode and cathode electrodes where formed of ten εintered carbon waεherε of ten grams total which formed four intermediate double-sided electrodes, with the sintered activated carbon conductive high surface area material (1) εandwiching the conductive backing (2), and two single sided end electrodes with the conductive backing on the outside, between the εintered activated carbon conductive high surface area material (1) and the end caps (7). Tabs (4) formed integral to the conductive backing layers (2) where gathered up to form anode and cathode leads, which were connected via a gold compression contact to wire leads which led to a DC power supply.

This capacitor was not placed inside a cartridge holder. One end of the central tube was sealed with urethane resin. The other end was connected to a length of plastic tubing, which in turn led to a pump. The flow¬ through capacitor was placed inside an open 200 ml container of feed solution, which was pumped through the capacitor

under negative presεure. Alternatively, the tubing end could be placed in the feed εolution, and εolution could be pumped through under poεitive preεεure.

Fig. 16 graphically illustrates Experiments 1 and 2 with non-εaturated NaCl feed to demonεtrate the utility of the present invention. Theεe experimentε uεed the flow¬ through capacitor deεcribed above in Example 1.

Experiment 1 - Continuouε Purification of NaCl Conditionε: Feed solution 1035 micro Siemens NaCl Flow rate .58 ml/minute

Voltage- 2 volts Results:

This experiment demonεtrateε more than a two order of magnitude purification εtarting from 1035 micro Siemenε feed to 9 micro Siemenε product. Thiε demonstrates that a very pure product stream can be produced starting with a concentrated feed solution. This further demonεtrateε that a capacitor deεigned with many εhort parallel flow paths, aε opposed to long εerpentine flow paths, produces very high purity product.

Experiment 2 - Continuouε Purification of NaCl at Fast Flow

Rate

Conditions: Feed Solution 1035 micro Siemenε NaCl Flow rate 4.2 ml/minute Voltage 2 voltε.

Product solution - to as low as 223 micro Siemens

This experiment demonstrateε a greater than 75% purification with the a flow rate that iε more than seven times faster than Experiment 1. Fig. 17 graphically illustrates Experiment 3 which testε the utility of the preεent invention with a εaturated feed stream. The capacitor used in Experiment 3 is as described in Example 1 above.

Experiment 3 - Simultaneous Batch Purification and Batch Concentration of Saturated CaSO _c

Saturated CaS0 ₄ is made by mixing diεtilled water with an excesε of CaS0 ₄ and letting the mixture sit on the excess solidε overnight.

Initial Conditionε:

Initial conductivity of saturated CaS0 ₄ solution is 1906 micro Siemens.

This saturated solution is divided into two equal 200 ml volumes.

Flow rate: 12.5 ml/minute, batch flow. Batch flow means that fluid iε pumped in a circular loop through the capacitor and recombined with the feed solution. Procedure A capacitor is connected to one of the two 200 ml saturated CaS0 ₄ solutions and run in batch mode at 2 volts for five minutes. Air iε then pumped through the capacitor until capacitor iε completely drained. Solution from the εecond 200 ml volume of εaturated CaS0 ₄ iε then pumped through the capacitor while simultaneously short circuiting the capacitor. This proceεε is repeated twenty times, always reusing the same solutions for the purification cycle and concentration cycle. Polarity was reversed between every charge cycle. A graph of the results is depicted on Fig. 17. After twenty charge cycleε, the batch purified solution is purified down to 802 micro Siemens. After twenty discharge cycles, the batch concentrated solution is concentrated up to 2470 micro Siemenε. Thiε concentrated εolution haε become εuper εaturated. Upon raiεing the temperature of thiε εuper εaturated solution in a hot water bath, crystals were observed to drop out of solution. Decanting thiε εolution provides a saturated solution which can be saved and reused as a regeneration εtream. The εeparated crystals may be collected as a solid waste.

An interesting phenomenon iε obεerved when purifying εaturated εolutionε. After about 10 charge/diεcharge

cycles, the capacitor is observed to purify the batch purification solution in a step wise manner. This iε shown on the graph of Fig. 17. Purification of the batch 200 ml solution begins to chiefly occur every other charge cycle. Precipitates that form from the εaturated εolution form more favorably on one of either the anode or the cathode and muεt firεt be deεorbed before any new net abεorption can occur in the purifying εolution. Thiε causes a step wise behavior that is observed in Fig. 17.