WO2010143950A1

WO2010143950A1 - Method for preventing fouling in a reverse electrodialyses stack

Info

Publication number: WO2010143950A1
Application number: PCT/NL2010/050350
Authority: WO
Inventors: Jan Willem Post; Hubertus Victor Marie Hamelers; Simon Grasman; Joost Veerman; Petrius Antonius Leijstra; Machiel Saakes
Original assignee: Stichting Wetsus Centre Of Excellence For Sustainable Water Technology
Priority date: 2009-06-09
Filing date: 2010-06-08
Publication date: 2010-12-16
Also published as: NL2002989C2

Abstract

The invention relates to a method and a device for preventing fouling in a reverse electrodialysis process. The method comprises the steps of : providing a reversed electrodialysis stack, comprising a number of membranes; supplying the stack with at least two feed streams; and - applying a preventive periodic osmotic shock to the fouling layers of the membranes.

Description

METHOD FOR PREVENTING FOULING IN A REVERSE ELECTRODIALYSES STACK

The present invention relates to a method for preventing fouling in a reverse electrodialysis (RED) stack. Such stack is used in a membrane-based RED process for the generation of energy by mixing river water and sea water, for example.

Existing RED operations generates sustainable energy in the form of electric power using ion-exchanging membranes that are stacked in an alternating pattern in an anode and a cathode. The stack comprises repeating units or cell pairs comprising a cation exchange membrane, an anion exchange membrane, and compartments filled with a spacer and solutions with different concentrations. Due to gradients between the solutions, transfer of ions occurs through the membranes. Electro neutrality in the system is maintained via oxidation at an anode surface and reduction at a cathode surface. Fouling of the ion exchange membranes limits the performance of these processes as it decreases energy production, and increases maintenance costs and use of chemicals, and results in shorter life times for the membranes. The fouling results from particles clogging the spacers, inorganic fouling or scaling, presence of organic matter and the adsorption through the membranes, presence of biodegradable substances that may be used as feed for microorganisms in the stack and as a result there of clogging of the stack and/or bio-film formation, and the presence of multivalent ions resulting in toxification of the membranes, for example.

The object of the present invention is to improve the efficiency of RED operations. This object is achieved with the method for preventing fouling in a reverse electrodialysis process according to the invention, the method comprising the steps of: providing a reverse electrodialysis stack, comprising a number of membranes; supplying the stack with at least two feed streams; and applying a preventive periodic osmotic shock to the fouling layers of the membranes. Several different and often related mechanisms affect the performance of a reverse electrodialysis (RED) process.

Due to the biofilm accumulation within the feed spacers, a build-up of the pressure drop may occur. For a RED proces, the clogging of the spacers may be seen as increased pressure drops from inlet to outlet of the stack for both inlet streams or feed streams.

Microorganisms composing biofilms may directly (via enzymes) or indirectly (via localized pH or redox potential changes) degrade the membrane polymer. If the membranes either foul or decompose during operation, it would mean both a decrease in selectivity and an increase in electrical resistance of the membranes.

Besides due to possible membrane degradation, the electrical resistance of a stack could also be increased due to an additional insulating effect of the biofilms adhered to the spacers .

As the biofilm comprised of micron-sized bacteria cells embedded in a matrix of EPS, the solutes now have to diffuse through the tortuous path of the deposit. The EPS matrix suppresses turbulent mixing at the membrane surface, resulting in enhanced concentration polarization.

Applying a preventive periodic osmotic shock to the fouling layers of the membranes prevents fouling layers building up and/or growing on membranes and/or other parts of the system. In a RED operation using river water and sea water as the two feed streams applying an osmostic shock would mean that microorganisms from a freshwater habitat are exposed to marine conditions, and vice versa. These microorganisms would get a hyper-osmotic shock, or a hypo- osmotic shock, respectively. When water is sucked out from the cytoplasm by osmosis, the cell membranes shrink. This shrinkage would cause a detachment of the microorganisms from the membrane surface.

In the RED operations, the number of membranes can be one, however, for practical reasons it probably would be significantly higher. In RED operations a relatively large concentration difference or concentration gradient is present between both sides of a membrane. Further, there is a concentration difference or gradient difference over the membrane. These concentration or gradient differences result from the use of at least two different feed streams. In a practical device these feed streams include river water and sea water. However, other feed streams can be applied also. Relating to fouling problems, RED operations suffer from the fact that at least two sources are present for the feed streams each having their own specific characteristics involving ionic concentrations, composition and biological activity, for example.

In a preferred embodiment the periodic osmotic shock is applied by the periodic reversal of the at least two feed streams by switching needs.

By periodically switching feed streams the polarity of the system is reversed. This polarity change has as an effect that the driving force of the deposition of the charged particles also changes and thereby would hinder the initial deposition of micro-organisms, for example. Furthermore, this change would, at least partially, remove the deposited micro-organisms. This effect is in addition to the hyper-osmotic shock or hypo-osmotic shock as described above. Applying the shock prevents and/or removes fouling layers of the membranes. Possibly, also other components of the system are removed. This improves the overall performance of such operation. The periodical reversal of feed streams has especially the additional advantage that it can be applied to grow micro-organisms present in the biofilms formed in the system.

In a preferred embodiment according to the present invention the ratio of ion concentrations between at least two feed streams during at least the periodic osmotic shock is at least 30, preferably at least 50 and more preferably at least 100.

By applying feed streams with a ratio of at least 30, a significant periodic osmotic shock can be applied by reversing the feed streams for example. In RED operations using sea water with 25-35 g/1 NaCl and river water with 50- 250 mg/1 NaCl this ratio can be increased to at least 50 and preferably at least 10 to further increase the osmotic shock and thereby the resulting increased overall performance of this operation. Increasing the ratio would improve the killing rate of the micro-organisms in the biofilms formed in the system.

In a preferred embodiment according to the present invention the periodic osmotic shock is applied with a frequency in the range of 10^~2 to 10^~5 Hz as a first range, and preferably in the range of 2.5 " 10^~4 to 3 " 10^~4 Hz as a second range.

The first range corresponds to applying an osmotic shock in the range of every one and a half minute to once a day. The second range would mean applying an osmotic shock about every hour. Experimental results have shown that the first range for applying an osmotic shock to the operation results in prevention and removal of fouling layers in the system, while further experimental results have shown that applying an osmotic shock about every hour results in an optimal trade-off between cleaning on one hand and limited down time of the system on the other hand.

In a further preferred embodiment of the present invention the periodic osmotic shock is applied with a pulse width in the range of 1-1000 seconds, and preferably in the range of 1-10 seconds. By using a pulse width in the range of 1 to 1000 seconds, which corresponds to up to three hours a thorough cleaning of the system can be realized. Preferably, the pulse width is limited to 1-10 seconds. Experimental results have shown that also when using this relatively small pulse width a significant effect is realized, while on the other hand only a short interruption of the energy production of the RED operation is required. This improves the overall performance of the operation. In a further preferred embodiment according to the present invention the periodic osmotic shock is applied during a relaxation period without power production.

Applying an osmotic pulse during a short relaxation period without power production would mean that an anode can stay an anode and a cathode can stay a cathode.

This has a positive effect on the lifetime of these electrodes. In combination herewith, or as alternative hereto, the osmotic pulse can be applied during power production of the RED process. When applying the osmotic shock during power production also the electrical current is part of the reversal at the same time as the reversal of the feed streams. This results in the anode becoming an cathode and the anode becoming a cathode during this reversal. This current reversal has a positive effect on the deposition of micro-organisms and removal of deposited micro-organisms on the membranes.

In a further preferred embodiment according to a present invention preventing fouling in a reverse electro- dialysis process comprises the step of washing out foulants by reversal of inlet and outlet.

Wash out of foulants, including inactivated biomass, improves the hydrodynamic resistance and electrical resistance. Furthermore, leaving the remnants of the fouling layer in place these could serve as substrate for new biofilms. Especially near the inlet of the membrane system biofouling occurs and the wash out is relevant.

In a further preferred embodiment according to the present invention the method further comprises the step of applying electrical pulses to the stack.

The detachment of biofilms and other foulants can also be achieved by periodically applying electrical back-pulses. Such pulse can be generated by charging and discharging capacitors in the electrical circuit in a relatively high frequency with optionally high amplitude that would be possible in a periodical flow switch as described above.

This has a beneficial effect on especially the initial deposition of the (charged) micro-organisms. In a further preferred embodiment according to the present invention the method comprises the step of using hypochlorite and/or acid produced at the electrodes.

As the chemistry of feed solutions in a RED operation is dominated by the high content of dissolved sodium chloride it is possible to produce hypochlorite at the anode that could be used as biocide. This biocide can be used either as preventive dosing step or as a cleaning agent. Similar considerations apply for the use of anolyte for an acid cleaning operation and katholyte for a basic cleaning operation.

In a further preferred embodiment according to the present invention, the method comprises the step of periodically heating the stack.

Periodical heating, of especially the solutions in the system, by Ohmic heating due to high applied voltages in the electrodes and/or introducing pre-heated solutions for example, would contribute to the inactivation of biomass.

The additional steps of washing out, applying electrical pulses, using hypochlorite and/or acid, and periodically heating the stack, can be applied in combination with applying a preventive periodic osmotic shock at the same time, in series, in different combinations, and also as separate measures independent form the application of preventive periodic osmotic shocks.

The present invention also relates to a device for performing a RED process, comprising: - a number of anode and cathode compartments provided with a number of electrodes; a number of membranes placed alternately between the electrodes forming electrolyte compartments; and - switching means to provide a periodic osmotic shock according to any of the clauses 1-13. Such devices provided the same effects and advantages as those related to the method.

Further advantages, features and details of the invention are elucidated on a basis of preferred embodiments thereof, wherein reference is made to the accompanying drawings : - figure 1 a schematic overview of the device according to the invention;

- figure 2 an illustration of a conceptual representation of a stack of figure 1 ; - figure 3A, B, C and D an illustration of experimental results for pressure drops, open-circuit voltage and stack resistance; figure 4A and B an illustration of experimental results for pressure drops correlated to biomass; - figure 5A, B, C and D an illustration of experimental results for pressure drops, open-circuit voltage and stack resistance;

- figure 6A-F an illustration of experimental results of TOC concentrations; and - figure 7A and B an illustration of experimental results of pressure drops related to electric back- pulses .

In a system 2 (figure 1) for a reversed electrodialysis process, a number of cation exchange membranes and anion exchange membranes 6 are placed between the anode 10 and the cathode 12. Between the anion exchange membranes 6 and cation exchange membranes 4 electrolyte compartments 8 are formed, wherein alternately sea water (s) 14 and river water (r) 16 flows. Due to the concentration differences of electrolytes in the sea water 14 and river water 16, the electrolytes in the sea water 14 will be inclined to move to the river water 16 to level the concentrations. For simplicity in figure 1 only sodium and chlorine ions are presented as positive and negative ions. As the anion exchange membranes 6 only allow anions to pass and the cation exchange membranes 4 only allow cations to pass, transport of anions and cations will proceed in opposite directions. The anions (Cl^") will move in the direction of anode 10, and the cations (Na⁺) will move in the direction of the cathode 12. In order to maintain electric neutrality in the compartments 18 where the anode 10 is placed, an oxidation reaction takes place. In the compartment 20 wherein the cathode 12 is placed, a reduction reaction takes place. Hereby a flow of electrons is generated in the electric circuit 22,23,24 wherein the anode 10 and cathode 12 are connected. In this electric circuit electric work is performed by an electric apparatus 24, here symbolically presented by means of a bulb. In the compartments 8 with flows of e.g. sea water 14 and river water 16 two layers of gaskets 26,28 are arranged. In the illustrated embodiment layer 26 has horizontal ribbons or guides and layer 28 has vertical ribbons or guides. A dialytic cell 30 is formed by a membrane couple of an anion exchange membrane 6 and a cation exchange membrane 4 and a mass of a solution having a low and a high electrolyte concentration. The number (N) of dialytic cells 30 (here N=I) may be increased to increase the potential difference between the anode and the cathode.

Experiment

A setup was developed in which the biofouling was accelerated by applying high concentrations of biodegradable substances (1.0 mg acetate-C/L) . With this setup, we were able to study the effects of biofouling in reverse electrodialysis and the effectiveness of preventive measures under controlled conditions.

Three reverse electrodialysis stacks 32 (figure 2) were run in parallel. These stacks were continuously fed with ^λsea water' 34 and 'river water' 36. For making-up of these feed solutions pre-filtered (> 1 μm) tap water was used. The tap water was supplied at a temperature of 10-12 ⁰C, without any disinfectant dosage or residual. Each feed stream contained three dosing points in series:

Addition of brine to increase the sodium chloride concentration of the feed solutions to appropriate levels. For river water the sodium chloride concentration was controlled at 1 g/L, and for sea water at 30 g/L.

Addition of a concentrated solution of biodegradable compounds. The main nutrient dosed to promote microbial growth was carbon, supplemented by phosphate and nitrate in the ratio of 100:10:20, as commonly found in biomass compositions. For both river water and sea water the sodium acetate concentration was controlled at 1.0 mg C/L. Sodium di-hydrogen phosphate was used as phosphate source and sodium nitrate as nitrate source. In order to prevent any microbial growth prior to the stacks, the pH of the nutrient solutions was leveled to pH 11 by adding sodium hydroxide and the dosing point was relatively close to the inlet of the stacks. - Addition of inoculates to introduce the microorganisms that may form a biofilm in these salinities. For river water the inoculate was taken from the Van Harinxma Canal (city of Leeuwarden, The Netherlands) , and for sea water from the Wadden Sea (harbour of Holwerd, The Netherlands) .

The pressure of the tap water was controlled by a pressure reducing valve at a level of approximately 0.5 bar. The feed flows through the stacks were controlled at 57.5 mL/min with upstream flow controllers (type FCA8842, Brooks Instruments) . The pressure drop over the stacks was measured with differential pressure transmitters (type deltabar S PMD70, Endress+Hauser) . The three square stacks 32 used in this study, consist of four membrane pairs 38 with a current-passing area of 104 cm². Between the cation exchange membranes (cem)40 and anion- exchange membrane membranes (aem)42, spacers 44 (Nitex 06- 700/53, Sefar) were used with a thickness of 0.5 mm.

Measurements are performed using Haber-Lugging capillaries placed in measurement compartments . The reference electrodes are placed within reservoirs that are connected to measurement compartments via Haber-Luggin capillaries. The tips of these capillaries were placed adjacent to the outside membranes (distance of 3 mm) . This adaptation was made to prevent the measurement to be disturbed by:

(i) gaseous products from the working electrodes; (ii) depreciation of the outer membranes by chlorine formed at the anode; and

(iii) bended current lines around the tips. Furthermore, the pressure within the measurement compartments was controlled at 1 bar in order to get a homogeneous tightening of the electrodialytic pile. The working electrodes are separated from the measurement compartments by chlorine-resistant auxiliary membranes (Ralex cation exchange membranes; Mega a.s.) . The working electrodes were connected to a galvanostat (IviumStat Electrochemical Interface & Impedance Analyser; Ivium Technologies) . With the galvanostat a square-wave signal was generated with cycles of a 180 s period of zero current conditions (open circuit) followed by a 3,420 s period with an adjusted current level of 100 mA (i.e. a current density of 10 A/m²) . The corresponding stack voltage was measured with an interval of 0.2 s. After each experiment, the spacers were exchanged for new spacers and the biofilm on the membranes was wiped off with a wetted tissue.

In order to characterize the accumulated biofouling, defined sections of membranes and spacers were stamped from different locations in the stack. These sections were placed in 30 mL of autoclaved sea water or river water in a capped tube. For the biomass analyses the biomass needed to be detached from the membranes and spacers. Therefore, the samples were placed in capped tubes filled with autoclaved sea water or river water. For the removal step, an approved procedure of ultrasonic treatment and sonification was used (Vrouwenvelder et al, Quantitative biofouling diagnosis in full scale nanofiltration and reverse osmosis installations, Water Research (42) (2008), p. 4856-4868). Total organic carbon (TOC) was used as a sum parameter to determine the total biomass amount (μg/cm² ) . The activity level of the present biomass was determined in terms of adenosine triphosphate (ATP) (ng/cm²).

Effect on parasitic hydrodynamic losses.

For membrane processes, biofouling is a term used to describe all instances of fouling where biologically active microorganisms are involved. This could vary from biofilm formation on the feed spacers, biofilm formation on the membrane surface or inside the membrane matrix, to biological degradation of the membranes. The effects of these fouling mechanisms on the performance of reverse electrodialysis are different. In the experiment with accelerated biofilm formation, the three stacks showed the same development of pressure drop in time (figure 3B) Performances of three parallel stacks: the pressure drops in mbar of feed flows over the stack in time in days (figure 3A and B for side 1 and 2 respectively) , the open-circuit voltage in V (figure 3C) , and the stack resistance in Ω (figure 3D) . The stacks (indicated with symbols Δ, O, and D) were operated similarly. The pressure drop over the river water compartments increased after 8 days. The salt water side has an operational period that is 5 days longer, although the growth potentials, expressed in the amount of biodegradable components, is the same for the fresh and salt water side. This difference is probably caused by a lower concentration of microorganisms in the salt water which leads to a lower rate of primary colonization of the stacks internal surface and thereby a slower build-up of biomass. In general, the pressure drop build-up graphs are similar to that observed in previous studies with membrane fouling simulators.

During sampling and disassembling the stacks after the experiments, it became clear that a biofilm was formed on membranes and spacers. The pressure drop increase can clearly be correlated to the build-up of biomass (Figure 4A: Pressure drop in mbar over the river water compartment correlated to amount of biomass (total organic carbon, TOC in μg/cm² ) , and Figure 4B: the activity of the present biomass (adenosine triphosphate, ATP in ng/cm² , in both cases with open symbols (0) representing samples taken at the inlet and closed symbols (♦) representing samples taken at the outlet) . The build-up of biomass appeared to be slightly higher at the inlet than at the outlet. At the end of the experiment, after the maximum pressure drop was reached and the feed flow started to decrease, the activity level of the biomass near the outlet appeared to be low. A possible explanation would be that the biomass near the outlet have faced nutrient shortage due to a combination of the starting hydrodynamic failure and a consumption of the preceding biomass.

Effect on thermodynamic efficiency. For RED, membrane degradation may be observed as a decreased open-circuit voltage (and a higher stack resistance) . The open-circuit voltage in the experiment with accelerated biofilm formation (figure 3C: voltage in V and time in days) seemed to be slightly affected after 8 days, i.e., when the rapid increase of the pressure drop occurred at the fresh water side. Likely, the stack was subject to minor internal leakage from the fresh water manifolds to the sea water compartments due to pressure differences over the membrane, or to salinity-gradients due to salinity-gradients within the fouled fresh water compartments, but not to failure of the membranes. This was confirmed after the experiment was finished as the open-circuit voltage was restored after the spacers were exchanged and membranes were wiped. After 10 days, i.e., when the pressure drop reached the maximum of 500 mbar and the fresh water flow could not longer be maintained at the same level, the open-circuit voltage decreased (not shown in figure 3) .

Effect on ohmic and non-ohmic losses. For RED, the additional insulating effect of fouled spacers may be observed as an increased ohmic resistance. The ohmic resistance seemed to be slightly affected after 8 days (Figure 3) which again could be explained by mechanical or hydrodynamical failure of the dialytic pile. After 10 days, the ohmic resistance changed due to the decreasing flow of fresh water (not presented in figure 3D) .

For RED, the additional concentration polarization may be observed as an increased non-ohmic resistance. The non- ohmic resistance, however, seemed to be unaffected and even slightly lower at the end of the experiment (figure 3) .

Osmotic shock of bio-films by periodical reversal of seawater and river water.

From cleaning attempts after previous experiment, it can be concluded that the prevention of biofouling by control strategies is preferable over cleaning. A possible control strategy according to the invention is a periodical switch of feed streams and consequently of polarity.

Regarding RED, the cleaning mechanism of a reversal would work in two different ways. Firstly, like in electrodialysis, the electric current is the driving force for the deposition of charged particles. When a current reversal is applied, the driving force for the deposition changes and thereby would hinder the initial deposition of microorganisms and would (partially) remove the deposited microorganisms. From figure 5 with performances of three parallel stacks: figure 5A and B: the pressure drops in mbar of feed flows over the stack for side 1 and 2 respectively, figure 5C the open-circuit voltage in V, and figure 5D the stack resistance in Ω and time in days) illustrates results when the stacks were operated differently. (Δ without flow- switch, O with a daily flow-switch , and D with an hourly flow-switch) . It can be seen that the operational period of the stacks increased with increasing flow-switch frequency. Under given circumstances with the accelerated biofilm formation, the operational period of the stack without a flow-switch was limited to 5 days by the pressure drop increase over the river water compartments (8 days for sea water compartments) . For the stack with a reversal each twenty-four hours, the operational period was 14 days. This proofs that the periodical reversal has a beneficial effect and that the extension of the operational period is not simply a matter of a different distribution of biomass. In that case, one would expect the operational period to be in the range of 7-10 days only, i.e., between the average growth potentials of both feed streams (average of 5 and 8 days) and twice the growh potential of river water (2 times 5 days) . Moreover, the cleaning effect of a periodical flow- switch can be seen from the hitches in the pressure-drop curves . For the stack with a reversal each hour, the operational period was 17 days. Except for one data-point, the build-up of pressure drop seemed to be smoother when compared to that of the stack with the twenty-four hourly switch. This indicates that at a higher frequency the reversal works becomes more a preventive step and less reliant on the cleaning effect. However, an increase of the frequency would probably not feasible in practice as each flow-switch is associated with a period of sub-optimal power production . Apparently, the open-circuit voltages of the stacks with the periodical reversal were less stable. This could be explained by the appearance of unmixed zones in the stack where the convective flow-switch is not effective in changing the salt concentrations. A locally lower salinity- gradient decreases the open-circuit voltage of the stack.

This may well explain the fact that the stack with the daily switch is less stable than the stack with the hourly switch. The more rapid build-up of biomass in this stack would increase the heterogeneity of the flow distribution. Another point of concern is the fact that the membrane performance would become less stable due to the fast changes of concentrations (an equilibration period of twenty-four hours is minimal) and compositions of feed streams. At the end of the experiment, the determination of the biomass was solely based on TOC (figure 6A for sample-side, 6B for compartment, 6C for inlet and outlet for side 1 and 2, 6D for effect of no-flow switch, daily flow switch and hourly flow switch, 6E for flow switch and no-flow switch for inlet and outlet, daily flow switch, hourly flow switch on sides 1 and 2, and 6F for no-flow switch with TOC concentrations in μg/10 cm² on membrane and spacer after experiment with results shown in Figure 5 (for the stack Δ without flow-switch, O with a daily flow-switch , and D with an hourly flow-switch) . Since especially the un-switched stack and the daily-switched stack suffered from almost complete blockage of the stack at the end of the experiment. Consequently, the biofilm was exposed to nutrient depletion, which had resulted in a dead biofilm. It was expected that the final amount of TOC would not show a relation with the flow-switch frequency, since all three the stacks suffered from the same hydraulic resistance in the end. However, results clearly show that the final amount of TOC increases with increasing flow-switch frequencies. Apparently, the major effect of the flow-switch is not on the biofilm development, but on the stacks capability to function well with larger amounts of biofilm on the membranes.

Experiment 2

The experimental setup of the first experiment is used to illustrate the effect of pulsed reversed current on the fouling processes (figure 7 with pressure drop in mbar and time in days for sides 1 (figure 7A and B with Δ without pulse and O with reversed electric pulse for side 1 and 2 respectively) .

Detachment of biofilms (and other foulants) can also be achieved by a periodically applied electrical back-pulse. This back-pulse can be generated within the reverse electrodialysis system itself, e.g., by charging and discharging capacitors in the electrical circuit in a much higher frequency and amplitude than possible with a periodical flow-switch as discussed previously. Therefore, a pulsed reversed current is able to prevent the initial deposition of (charged) micro-organisms, in our experiments, for instance, a back-pulse of 2 ms per 1 s with amplitude of 75 A/m2 (i.e., 7.5 times the operational current density). The present invention is by no means limited to the above described preferred embodiments. The rights sought are defined by de following claims, within the scope of which many modifications can be envisaged. So would it be possible to apply the method and the device according to the present invention to prevent scaling/ biofouling on the electrodes in membrane-based operations like reversed electrodialysis and electrodialysis for example .

Claims

1. Method for preventing fouling in a reversed electro dialysis process, comprising the steps of: - providing a reverse electrodialysis stack, comprising a number of membranes; supplying the stack with at least two feed streams; and applying a preventive periodic osmotic shock to the fouling layers of the membranes.

2. Method according to claim 1, wherein the fouling layers comprise bio-films.

3. Method according to claim 1 or 2, wherein the periodic osmotic shock is applied by periodic reversal of the at least two feed streams by switching means.

4. Method according to claim 1, 2 or 3, wherein at least during the periodic osmotic shock the ratio of ion- concentrations between the at least two feed streams is at least 30, preferably at least 50, and more preferably at least 100.

5. Method according to any of the claims 1-4, wherein the periodic osmotic shock is applied with a frequency in the range of 10^"2 to 10^"5 Hz.

6. Method according to claim 5, wherein the frequency is in the range of 2.5 ^• 10^"4 to 3 ^• 10^"4 Hz.

7. Method according to any of the claims 1-6, wherein the periodic osmotic shock is applied with a pulse width in the range of 1-1000 seconds, and preferably in the range of 1-10 seconds.

8. Method according to any of the claims 1-7, wherein the periodic osmotic shock is applied during power production of the reversed electrodialysis process.

9. Method according to any of claims 1-8, wherein the periodic osmotic shock is applied during a relaxation period without power production.

10. Method according to any of claims 1-9, further comprising the step of washing out foulants by reversal of inlet and outlet.

11. Method according to any of claims 1-10, further comprising the step of applying electrical pulses to the stack.

12. Method according to any of the claims 1-11, further comprising the step of using hypochlorite and/or acid produced at the electrodes.

13. Method according to any of the claims 1-12, further comprising the step of periodically heating the stack.

14. Device for performing a reverse electrodialysis process, comprising: a number of anode and cathode compartments provided with a number of electrodes; a number of membranes placed alternately between the electrodes forming electrolyte compartments; and switching means to provide a periodic osmotic shock according to any of the claims 1-13.