WO2003082750A1 - Apparatus and a method for treating effluent - Google Patents

Abstract

Apparatus (10) for treating aqueous wastes, such as wastewater, landfill leachates and aqueous effluents, comprises an electrolytic cell (12). The aqueous (10) also includes electrical pulse generating means for generating a series of electrical pulses between the electrodes of the cell.

Description

APPARATUS AND METHOD FOR TREATING EFFLUENT

This invention relates to apparatus and a method for the treatment of aqueous waste, such as wastewater and aqueous organic effluents.

Wastewaters and aqueous organic effluents contain polluting substances such as organic chemicals and generally the waste products of industrial and commercial processes. A pollutant can be defined as any substance resulting from human activity that has a detrimental effect on the environment. Effluent can be defined as any liquid or flowable waste emanating from a process that is released to environment. Hazardous wastes are specific substances that are classified according to their source and characteristics.

Common organic pollutants generally fall into the following categories:

• Pesticides include insecticides, herbicides and fungicides. This group of compounds includes chlorinated hydrocarbons, organophosphates and carbamates.

• Organic Solvents such as benzene, toluene, ethylbenzene and xylene.

• Polycyclic Aromatic Hydrocarbons.

• Surfactants such as alkyl aryl sulphonate.

• Chlorinated Phenols such as Pentachlorophenol (PCP) and Polychlorinated Biphenyls (PCBs).

Many of these compounds are carcinogenic and hazardous to human health. They are generally toxic to the environment. Some are non-biodegradable and many only degrade over a very long period of time.

Organic pollutants may also be divided into those that are biodegradeable, high-priority pollutants, recalcitrant organic compounds, volatile organic compounds (VOCs) and malodorous compounds. These are described below:

• Biodegradable compounds include proteins, carbohydrates, fats and surfactants. Their presence may lead to depletion of oxygen in water thereby causing septic conditions to prevail. Some of these compounds, such as dyes and oils, spoil the environment aesthetically.

• High-priority pollutants are those known or suspected to be carcinogens, mutagens or toxins. Typical examples include benzene, ethylbenzene, toluene, chlorobenzene, chloroethene, dichloromethane, tetrachloroethene as well as many pesticides, herbicides and insecticides. • Recalcitrant organic compounds that are resistant to biodegradation.

• Volatile Organic Compounds (VOCs) include organic compounds with boiling points below that of water.

• Malodorous compounds include amines, mercaptans, and organic sulphides. Odour thresholds can be as low as 1 ppb by volume (such as mercaptans).

Most of the organic substances noted above have been identified in the leachate from landfill sites. As a significant number of these substances are either non-biodegradable or only degrade slowly over time, effective treatment by ordinary biological processes, such as aerobic and anaerobic reactors, is not possible.

Current treatment methods for pollutants include incineration and pyrolysis, air stripping, microbial treatment, precipitation and coagulation, chemical treatment, adsorption, membrane separation processes, distillation, and oxidation processes.

The latter category of "oxidation processes" includes:

• Homogeneous photolysis: UV photolysis of H2O2 and/or O₃ and other additives in solution to create hydroxyl (*OH) and other free radicals;

• Radiolysis: high-energy radiation (tf-rays) to irradiate wastewater. A variety of species, such as *OH, H^», e_aq (hydrated electrons) are created under these conditions.

• Dark oxidation processes: radicals generated by Fenton's Reagent, ozone at high pH and ozone/peroxide.

• Photocatalysis: semiconductor catalyst and a light source to induce photochemical reactions at the surface of the catalyst.

The following known electrochemical reactors have been designed and built for the treatment of various pollutants and chemical recycling:

• DEM Cell: EA Technology of Chester, UK, makes this electrochemical device. It is an anodic oxidation cell designed for the treatment of wastewater and toxic organic effluents including ammoniacal liquors, cyanides, organo-cyanides, solvents, chlorinated hydrocarbons, pesticides, mixtures of metals and organic compounds. This is a "plate and frame" design where dished electrodes are located between pairs of plastic frames. The electrodes have a catalytic coating and the dished design promotes high mass transfer rates.

• Axonic Cell: Produced by Axonics Ltd of Swansea, Wales, this is an electro-catalytic device designed to treat a variety of waste waters including effluent from metal finishers, pulp and paper makers, food and beverage producers, etc. A small electric charge is applied to a selective sacrificial electrode (anode) in order to produce metal hydroxides that act as flocculating agents. The floes are then removed by a further process such as Dissolved Air

Floatation ("DAF").

• Retec Cell: This cell design is marketed by Eltec Systems Corporation of Ohio, USA. It uses a three-dimensional electrode and contains 6-50 cathodes interspersed with dimensionally stable anodes, usually oxide-coated titanium mesh. The effluent solution flows perpendicularly through the electrodes by a "flow-through" arrangement. This design creates a large cathode volume and moderate mass-transfer coefficient with air sparging of the cell.

• The electrochemical cells mentioned above, in common, are supplied by Direct Current. None of them incorporate current switching devices across the cell at any frequency. These cells act in accordance with the normal established electrochemical principles.

Other known processes oxidise by means of a free radical mechanism. Many of them produce the high reactive hydroxyl radial (*OH) in water.

Examples of such oxidation process are given below:

• Perox-Pure System: This is a patented UV/H2O2 system made by Peroxidation Systems (USA). It is designed to destroy organic pollutants in water. An additional catalyst, such as iron, can be introduced if required. The system uses medium-pressure mercury vapour lamps in quartz sleeves and is designed to maximise the radiation output of the UV lamps.

These processes do not produce highly reactive radicals such as hydroxyl (*OH) radicals by electrochemical means.

According to one aspect of the invention there is provided apparatus for treating aqueous wastes, such as waste water, landfill leachates and aqueous effluents, the apparatus comprising electrolytic cell and electrical pulse generating means for generating a series of current pulses between the electrodes of the cell.

According to another aspect of the invention there is provided a method for treating aqueous wastes, such as waste water, land fill leachate and aqueous organic effluents, the method comprising introducing the aqueous waste into a electrolytic cell and creating a series of electrical pulses between the electrodes of the electrolytic cell. It is understood that the apparatus and method of the invention break down organic molecules in aqueous solution by way of the following electrochemical processes:

• Direct anodic oxidation (i.e. electron stripping at the anodes),

• The production of highly reactive radicals such as hydroxyl (OH) radicals resulting from the release of capacitance charge at the anodes, and

• Secondary oxidation reactions in solution by the creation of reactive intermediates such as hypochlorite.

• Reductive dehalogenation reactions (in the case of Carbon cathodes).

The apparatus and the method of the invention efficiently generates large numbers of highly reactive radicals, such as hydroxyl (OH) radicals and these create an aggressive oxidising environment in which dissolved organic pollutants are broken down and even mineralised to produce CO2. In this process, the bonds that hold organic molecules together are systematically broken-down by the above-mentioned forces. Complex aromatic molecules and long chain aliphatic molecules are broken down progressively.

Precise control over the electrochemical process maintains the optimum conditions under which oxidation takes place and organic compounds are broken down with a high degree of efficiency. When the apparatus is "switched-on" the oxidation process begins and continues until it is "switched-off.

The process breaks down recalcitrant organic compounds pollutants, as diverse as humic substances, endocrine disruptors, textile dye waste and sewage sludge liquors. The apparatus aggressively oxidises and destroys a wide range of organic compounds found in wastewater, landfill leachates and aqueous organic effluents. Such organic compounds are described above and include phenols, chlorinated phenols, PCBs, mercaptans, cyanides, organophosphates, amines and nitro compounds.

Particular applications of the invention include:

• Reduction of Chemical Oxygen Demand ("COD") and Biochemical Oxygen Demand ("BOD") in a wide range of wastewaters, landfill leachates, and aqueous organic effluents. In particular, by means of the invention the recalcitrant component of COD in aqueous organic effluents can be broken down (the so-called "Hard-COD"). This is described as the component of COD in an effluent that cannot be treated by ordinary biological treatment processes such as aerobic/anaerobic digestion. • Pre-treatment of aqueous organic effluent for biological digesters (bio-digesters). Intermediate products can be obtained as a result of partial oxidation. These products may have a lower toxicity than the starting compounds. In such cases, the wastewater, leachate or effluent may be suitable for biological treatment processes such as aerobic/anaerobic digesters and membrane bioreactors ("MBRs"). Energy consumption, and therefore process costs, can be reduced by the combination of these technologies.

• Removal of ammoniacal nitrogen from concentrated sludge liquors, landfill leachates and industrial effluents containing ammoniacal nitrogen. In this case, the end product of the process may be either nitrate or nitrogen gas.

• Breakdown of dyes and coloured effluent. The process rapidly destroys dyes of various types including acid, dispersed and reactive dyes and colouration is effectively removed from effluent streams.

• Destruction of malodorous compounds such as mercaptans, amines and organic sulphides.

• Disinfection of biologically active effluents. The process acts as an efficient biocide. The aggressive oxidation process disrupts cell walls of living organisms such as bacteria and protozoa and may also destroy viruses by breaking down of their protein shells. Oocysts of Cryptosporidium and Giardia, and other waterborne parasites are also subjected to the effects of oxidation and may be destroyed as they pass through the apparatus. Genetic material such as DNA, RNA and even protein prions may be broken down and rendered harmless by the process.

Possible electrode materials are described below:

• The anodes contained within the electrolytic cell can suitably be made from any of the following materials:

- Titanium coated with various metal oxides

- Titanium coated with Iridium doped Ruthenium oxide

- Carbon fibre fabric

- Composite materials having a free Carbon surface

- Allotropes of Carbon such as graphite

- Other materials having a carbon surface

- Any material that is not subject to electrolytic corrosion

• The Cathodes contained within the electrochemical cell can suitably be made from any of the following materials:

- Stainless steel, all types

- Titanium coated with various metal oxides - Titanium coated with Iridium doped Ruthenium oxide

- Carbon fibre fabric

- Composite materials having a free Carbon surface

- Allotropes of Carbon such as graphite

- Other materials having a carbon surface

- Any material that is not subject to electrolytic corrosion

• Electrodes made from Titanium coated with Iridium doped Ruthenium oxide are robust and resistant to corrosion when used for the electrochemical oxidation of dissolved organic pollutants.

• Electrodes made from certain Carbon fibre materials and composite materials having a free Carbon surface are robust and resistant to corrosion when used for the oxidation of dissolved organic pollutants.

• The electrodes described above can operate within the pH range 8 to 10. Typically most effluents fall within this range.

• The electrodes may be flat sheets but other shapes can be effective.

• Metal electrodes, such as oxide coated Titanium or stainless steel, can be in the form of plain flat sheets, or expanded mesh, or perforated material.

• Composite electrodes with a free Carbon surface are preferably in the form of flat rigid plates.

• The spacing between electrodes should be kept to the minimum that can be practically achieved based upon the electrode material used. This is to minimise the electrical resistance of the liquids passing between the electrodes. Increasing the separation distance increases the power requirement of the electrochemical cell. The electrodes are preferably less than 4 mm apart and conveniently more than 1 mm apart. Preferably the electrodes are about 2 mm apart.

Details of the flow of aqueous waste through the apparatus:

• The direction of flow of the wastewater or effluent through the electrochemical cell is determined by the nature of the electrodes. In the case of non-porous plate electrodes the aqueous waste simply passes over the surface of electrodes. In the case of perforated or expanded electrodes the aqueous waste can flow through the electrodes in a process described as "cross-flow".

• The optimum rate of flow of the aqueous waste through the apparatus is determined by a number of factors such as concentration of organic compounds within the aqueous waste and retention time of the liquid within the electrolytic cell. The flow rate may be varied to increase or decrease the retention time of the liquid within the electrolytic cell in order to achieve the desired level of oxidation and, therefore, processing requirement. The apparatus preferably includes means for passing aqueous waste, or allowing aqueous waste to pass, through the cell at a flow rate of less than 1 m5^"', preferably less than 10 mm5^''.

• Turbulent flow of the wastewater or effluent through the apparatus can be advantageous as this can enhance mixing of the dissolved organic compounds and may increase the contact time with the electrodes within the electrolytic cell. Turbulent flow can also aid the removal of precipitates, floes and scum from the electrolytic cell.

• The residence time of the aqueous waste within the electrolytic cell is one of several key factors that determine the rate of oxidation of organic pollutants. Large organic molecules are broken down progressively and the level of oxidation achieved is largely determined by the retention time of the liquid within the electrochemical cell.

• The Apparatus is preferably designed in such a way that the wastewater, leachate or effluent flows freely through the electrochemical cell and that any precipitate, floe or scum that may be formed in the treatment process can be removed.

Details of possible pre-treatments for wastewater, leachate or effluent:

• The wastewater, leachate, or effluent entering the apparatus is preferably pre-filtered to remove particles and suspended solids that may cause blockages within the electrochemical cell or electrical short circuits between the electrodes.

• The electrochemical oxidation process is generally more effective within the range pH 7 to pH 9, and preferably is at least PH 8.

• The electrochemical oxidation process is generally more effective in the presence of chloride. Concentrations should not exceed 2,000 mg/1.

• The efficiency of the electrochemical oxidation process can be reduced by the build-up of Calcium on the electrodes when treating wastewater or effluent containing high levels of Calcium, for example, liquid containing milk. In the case of some liquids it may be desirable to remove calcium compounds or reduce "hardness" by a softening process such as ion exchange prior to electrochemical treatment.

Details of the preferred electrical features and characteristics of the Apparatus:

• A variable DC electrical supply is preferably used which may output between 5 Volts and 25 Volts. The actual voltage applied to the apparatus will be selected dependent upon the nature of the wastewater leachate or effluent to be treated. For Titanium electrodes the applied voltage may suitably be from 6 Volts up to 12 Volts. In the case of Carbon electrodes the voltage may range from 10 Volts up to 20 Volts.

The current required is dependent upon the material of the electrodes and their surface area.

In the case of Titanium electrodes, typical current densities may be from 5 mA/cm² to 20 mA/cm² (50A/m² to 200 A/m²). For electrodes with free Carbon surfaces, the current requirement is generally higher and dependent upon the nature of the Carbon fibre material used.

Preferably the or each anode is connected to the positive line and a diode is included in the connection to inhibit reverse currents.

Preferably each cathode is connected to the negative line and a high-speed current switching device, such as a MOSFET, is included. There is preferably one switching device for each cathode.

Each MOSFET may be designed to turn the electrical power "ON" and "OFF" at a rate of up to 500,000 or 600,000 times per second even as high as 1 MHz. This process is described as the "Switching Frequency". The Switching Frequency is preferably greater than 20 kHz. A preferred Switching Frequency is in the range 100,000 (100 kHz) to 500,000 (500 kHz) per second and preferably is greater than 300 kHz.

The current when switched "ON" increases to a level related to the surface area and nature of the electrode. The current area relationship (Current Density) is typically between 5 mA/cm²

(50 A/m²) and 20 mA/cm² (200 A/m²) for Titanium electrodes and marginally higher for carbon based electrodes. In one embodiment, the source impedance and electrode capacitance of the cell are matched to thereby promote use at low current density.

To achieve the optimum process conditions under which electrochemical oxidation takes place for any particular wastewater, leachate or effluent, the optimum Current Density should be achieved at the electrodes as well as the optimum Switching Frequency.

Other factors that affect the efficiency of the Apparatus include the chloride concentration, acidity/alkalinity (pH), conductivity, composition and concentration of the organic pollutants, and molecular size.

The surface area of each electrode is limited by the capabilities of the high-speed switching device. Large electrodes require individual switching whereas several small electrodes can be operated in groups by one switch.

The minimum number of electrodes required by the Apparatus is one anode plus one cathode. Additional electrodes can be added in pairs to form a "Pack of Electrodes" or "Module". A typical Pack would consist of a series of anodes and cathodes, normally with a cathode at each end.

The size of a Pack of Electrodes is limited by various factors including available current, various mechanical factors, container size, and Pack weight.

In a Pack of Electrodes, the cathodes are preferably switched in groups and preferably in such a way that only one cathode adjacent to an anode would be switched "ON" at any given time. In a simple Pack of Electrodes, there would be two groups of cathodes. At any one time, only one group of these cathodes would preferably be switched "ON".

When a high-speed switching device, such as a MOSFET, in series with the or a cathode is switched "ON", current flows to the cathode. The current flow resembles that of charging a capacitor. When the device is switched "OFF", highly reactive radicals, such as hydroxyls

(•OH), are generated over the surface of the anode. First the current rises, then falls in a spike; it can over shoot and oscillate but this effect depends upon the circuit arrangements.

Oscillation of the current may be controlled by the use of split ferrite cores. In an alternative embodiment, the voltage spikes may by created by release of capacitive storage, or by means of a circuit reactive component such as an inductor.

For the process to operate most effectively, the high-speed switching device must be switched "ON" extremely rapidly and precisely. And, in turn, the device must switched

"OFF" extremely rapidly. State-of-the-art high frequency MOSFETs can provide this function.

When a MOSFET is switched "OFF" rapidly the voltage on the cathode also rises and then falls rapidly in a forward voltage spike. This spike can preferably exceed 100 Volts. The spike can be greater than 40 Volts, and preferably 60 Volts, in order to create a very significant effect in terms of electrochemical oxidation. This sign of the voltage spike is linked to cathode area and the characteristics of the MOSFET and its associated components.

The width at half voltage spike height may be less than about 500ns.

In a Pack of Electrodes, particularly where each electrode has a large surface area, each cathode preferably has its own high-speed switching device such as a MOSFET.

When a number of cathodes are switched "ON" and "OFF" at the same time in a group, then preferably there are two such groups. Depending on pulse time and frequency, however, it is also possible to switch more than two cathode groups, thereby further multiplexing the power source.

In particular when using a single power source to feed a Pack of Electrodes, it is preferred that only one group of cathodes is switched "ON" at a time. The first group of cathodes should be switched "OFF" before the second group of cathodes are switched "ON". It is further preferred that there should be a short delay between turning "OFF" the first group cathodes and turning "ON" the second. This delay should preferably be sufficient for the energy to dissipated. The length of this delay period may be determined by the time to collapse of the voltage spike.

A suitabled catalyst may be provided.

The cell may take any suitable form but in one embodiment the cell consists of an anode between two cathodes. This has been found to work particularly well. Also, as the anode may be made from precious metal, this arrangement, which takes advantage of both sides of the anode, is particularly cost effective.

According to a third aspect of the invention there is provided an electrolytic cell consisting of an anode between two cathodes.

Embodiments of the invention will now be described by way of example and with reference to the accompanying drawings, in which:

Figure 1 is a schematic drawing of an apparatus in a first embodiment;

Figure 2 is a schematic drawing of an apparatus in another embodiment;

Figure 3 is a connection diagram;

Figures 4 and 5 are diagrams of readings at the cathodes in the second embodiment;

Figures 6 and 28 are graphs illustrating the results of various experiments conducting using the apparatus of Figure 1 ;

Figure 29 is a schematic drawing of an apparatus in a further embodiment; and

Figure 30 is a graph of a control sine wave and the current on the two cathodes in the embodiment of Figure 25.

The apparatus of the first embodiment is shown diagrammatically in Figure 1. The apparatus 10 of the embodiment consists of an electrolytic cell 12 including, in this case, two cathodes 14, 16 positioned either side of an anode 18. The electrodes 14, 16, 18 are mounted centrally across the width of a glass trough 20 containing electrolyte. The anode 18 is connected to the positive terminal of a variable or voltage-controlled direct current power supply 22 through a diode 24 (BY328, peak reverse voltage of 1500 V). Alternatively, the anode could be connected to a positive terminal of a variable voltage supply from a lead-acid battery through the diode 24.

Each cathode 14, 16 is connected to the drain terminal of an IRFS450, a metal oxide semi- conductor field effect transistor (MOSFET) 26 (the MOSFET having a drain-source break-down voltage of 500 V, rise time of 65 ns, on resistance approximately 40 mS and fall time of 80 ns). The source electrodes of MOSFETs 26 are connected to the negative terminal of the power supply 22, or the battery 21, through a BY328 diode 25. A 4,400 μF electrolytic capacitor 27 is connected across the power terminals of the power supply 22.

The MOSFET gates are controlled by Telcom TC4420 MOSFET drivers (not shown) operating at up to 18 V with respect to earth potential. The negative terminals of the power supply controlling the MOSFET drivers and the main power supply are star-connected to earth potential.

Current-limiting resistors 28 of 2.2 ohms are used in the gate circuits of each MOSFET 26, giving a measured peak gate current of approximately 1 A. This enables each MOSFET 26 to switch with minimum rise and fall times.

The MOSFET drivers are controlled by 0 to 18 V complementary metal oxide semiconductor (CMOS) compatible square waves 30 from a function generator 32 and a specially constructed signal splitter 34. The signal splitter 34 splits the signal into two pulses each being identical but being out of phase by half a wavelength. In this way, it is possible, firstly, to cause each MOSFET to conduct for periods ranging from 0.4 μs either with a fixed pulse time independent of frequency or with a constant mark-space ratio, and secondly, to ensure that the delay time between the turn-on of one cell to turn-off of the other cell was, to within 20 ns, half the period of the function generator 32.

In a second embodiment (not shown) only a single cathode and anode are included in the electrolytic cell. In this case, the anode is connected to the positive terminal of a power supply through a BY328 diode and an IRFS450A MOSFET. The cathode is connected to the negative terminal of the power supply by a BY328 diode.

Many experiments have been conducted using this basic apparatus of the embodiments described above.

The apparatus 10 of Figure 2 is similar to that of Figure 1 and only the differences between the embodiments will be described. The resistors 28 are omitted, as is the diode 25. MOSFET drivers 29 are included between the MOSFETs 26 and an integrated switching box 31 which takes the place of the function generator 32 and signal splitter 34. Figure 2 also shows the system for flow of aqueous waste through the cell 12. The aqueous waste circuit includes a pump 100 and flow water 102. A filter 104 is provided in the index. After flow through the cell 12 the liquid can pass to the outlet or back though a recirculation line to enter the cell 12 again.

Experiment 1: Organic Effluent from Drum Cleaning Factory

This experiment used the apparatus 10 of Figure 2.

Background information:

• In this series of experiments, the industrial effluent comes from a company that specialises in the cleaning and reconditioning of drums of various types, Integrated Bulk Containers ("IBC's") and other bulk liquid containers. About 700 drums arrive on the company's site every day for processing. These drums come from a variety of sources and their contents include a variety of organic products, such as minerals oils, edible oils, fish oils, cocoa butters, solvents and other chemicals.

• The on-site process for cleaning and reconditioning drums and containers involves draining of any remaining contents, multistage washing including the use of Sodium Hydroxide and heat, shot blasting to remove any paint, and repainting.

• The initial draining of the drums is known as "Raw Drainings". These liquids pass into a tank where they are held for off-site disposal. All the subsequent washings from this and the IBC cleaning process are directed to large balancing tanks that feed a Dissolved Air Flotation ("DAF") treatment plant. The balancing tanks can hold volumes approximating to one day's effluent production.

• The effluent from the DAF plant is discharged to sewer and the sludge formed during the DAF process is added to the "Raw Drainings" tank.

• The quantity and quality of the effluent discharged to sewer is controlled by the local water services company using a regulatory instrument known as a "Consent to Discharge Trade Effluent to Sewer", as set out in the Water Industry Act 1989.

• The company consistently fails to meet the conditions of the "Consent to Discharge Trade Effluent to Sewer" as required to by the "Water Services Company" ("WSC"). The WSC have detected various "non-consented" substances in the trade effluent discharged from the company's site. Particular substances highlighted by the WSC include Volatile Organic Compounds

("VOCs"), Mercury, (3) Tributyl Tin, and Zinc. Discharge of substances not included in consent conditions is an offence under the "Water Industry Act 1989".

Samples of the company's effluent were taken for the experiments in accordance with BS 6068 and ISO 5667. The data record also follows the guidance set out in BS 6068. As there are a number of different factors, which could influence the performance of the Apparatus, a factorial design was developed for these trials.

There are two reasons for using a 'Factorial Experimental Design ("FED") rather than a classical design, where each factor is investigated in turn. These are (1) the FED detects and estimates any interactions, which the one-at-a-time experiments cannot do, and (2) if the effects of the factors are additive, then the FED requires fewer measurements than the classical approach in order to give the same precision.

The FED fulfils the requirements of BS 6068 Sections, 6.1, which deal with the design of the sampling programme.

On a daily basis effluent from the on-site Dissolved Air Flotation (DAF) was introduced into the Apparatus until a point was reached where the re-circulation pump could be activated. At this point an initial pre-treatment sample was taken from the outlet chamber.

Then electrical power was applied to the Apparatus and rate of flow and frequency were set in accordance with the FED. Subsequent samples and adjustments were also made in accordance with the FED.

The initial "Raw Effluent" (Pre-treatment) sample and subsequent samples were analysed for Total COD, Volatile Organic Compounds, FOG (Fats, Oils Grease), Chloride, Metals ( Zinc, Nickel, Chromium, Copper, Iron), conductivity, temperature and pH., using the methods set out below.

The following methods of analysis were used:

• COD was measured using Hydrocheck method HC6153

• Chloride was measured using a Hanna Instruments Chloride Ion Selective electrode and meter.

• Zinc was measured using a Quantofix 2- 100 mg/L field test kit

• pH and temperature were measured using a WTW 330 field pH monitor. The pH probe was buffered to pH 7 and 10 on a daily basis

Matched pairs of samples were delivered to the laboratory. The laboratory provided bottles, packaging and preservation materials in accordance with BS 6068 Section 6.3: 1986 ISO 5667-3:1985 plus any additional requirements stipulated by the United Kingdom Standing Committee of Analysts Guidelines.

Readings from the cathodes 14, 16 are shown in Figures 25 and 26. The input pulse width for the first cathode 14 is the same as for the second cathode 16. The "fall before rise" time is the same for the first cathode 14 as for the second cathode 16.

The results of the series of trials are set out below:

The statistical examination of the trial results, set out below, shows a 99% fit. This is considered to be excellent. This shows that the change in COD can be explained by the factors examined.

Normal Probability Plot of Residuals of COD Using Studentised Residuals in Model AOPFINALM5 (Sample size = 30)

Graph 1 Residuals of COD

27-JA -2001 14:16 Page 1 COD DROP PERCENT

TIMES = 2035, EFFLUENT = b, CURRENT = 10.086, CHLORIDE = 19544

FLOW COD COD DROP PER CENT

Graph 4

Linear effects explain 87% variation in COD reduction. Interactions and linear effects explain 98% in COD reduction. Linear, interaction and quadratics explain 99% in COD reduction, showing only a small contribution from curvature effects is present.

The initial fit on unbalanced trial was 64%, and on the second (D-optimal trial) was 98% indicating the trial had correctly identified the factors that were influencing the plant performance.

Furthermore there were no outliers that confirmed validity of the test method. The trial has shown that the interaction between time and flow rate shows that low flow with high time maximises COD reduction. High frequency and high current also enhance reduction of COD.

Graphs 3 and 4 show that the system was close but not at its optimum operating conditions. It should be noted that all these trials were carried out on the effluent from one site. The final full-scale trial (Graph 5 below) to test the plant under expected operating condition shows that the plant is capable of treating the effluent to within consent conditions.

COD vs. Time

Elapsed Hours

Graph 5

A full-scale plant is therefore expected to achieve consent compliance after approximately eight hours. This series of experiments also shows that it is possible to treat the effluent using AOP to a standard that would enable reuse within the site processes.

Using the mid-point of the electrical power consumption data, the electrical cost is calculated to be £0.28 per Kg COD reduced.

The overall outcome of these trials proved that the system can successfully treat a waste that has proved difficult using conventional industrial wastewater treatment technology. Independent consultants considered the system is technically βt for purpose.

Experiment 2

This experiment used the apparatus 10 of Figure 1.

By operation of a simple electrochemical cell in an aqueous electrolyte with a control system it is possible to generate pulse anodic potentials up to approximately 250 V relative to earth potential at frequencies from 100 to 500 KHz with widths at half peak height of less than 100 ns.

Both anode and cathode appeared to be almost identically charged when corrected for baseline voltage difference. Therefore, for a short time, each electrode acted simultaneously as an anode at a potential sufficient to oxidise water to form oxygen or hydroxyl radicals.

The temperature dependence of the average ionic current was much less than observed under direct current conditions, suggesting that the dominant conduction mechanism for this process is very different than for conventional electrolysis.

The cell destroyed dissolved ammonia with an average apparent Faradaic efficiency of -200% compared with that theoretically calculated and was about three times the efficiency of a separate direct current experiment at the same current density in the same apparatus.

The majority of the experimentation involved the use of anodes made from expanded and flattened titanium with 8mm by 6mm rhombohedral openings and an open area of approximately 40%) of the plan area. The anode was coated with iridium-doped ruthenium dioxide (with an Ir-Ru weight ratio of 30:70). Such an anode was a freely available commercial product and is known to evolve chlorine in the presence of chloride. The two cathodes 14, 16 were made from 0.4 mm diameter woven wire titanium or 316 stainless steel mesh with approximately 7 wires per cm. Such cathodes also had a 40% open area. The width of each electrode was 65 mm and the height was 80 mm. The anode 12 and cathodes 14, 16 were spaced 2mm apart.

The glass trough 20 in which the electrodes 12, 14, 16 were mounted centrally, had a length of 150 mm, width of 67 mm and height of 77 mm. The glass trough 20 was filled with electrolyte upto a depth of 60 mm, so providing a liquid volume of 600 cm³. The plan area of the electrode up to the level of the electrolyte was 39 cm³.

A peristaltic pump (not shown) was used to pump the electrolyte through the electrodes at an average flow rate of 0.1 mm/s taking into account porosity. During experimentation, three plate anodes were also used. These specifically being (i) a commercially available, iridium-doped ruthenium dioxide coated titanium anode, (ii) a titanium anode coated with antimony-doped tin, and (iii) an exfoliated graphite sheet, carbon is also known to be oxygen evolving. All the anodes had an area of 10 cm².

The results of the various experiments conducted using the apparatus of the present embodiments is set out below.

Initial experiments were done with a porous anode and single porous cathode configuration with cathode switching and liquid flowing perpendicular to the electrodes.

The electrolyte was demineralised water to which had been added 5 g/1 sodium chloride. Figure 2 shows graphs of the anode current and the anode voltage, anode switching at 60 kHz, pulse width of approximately 2.5 μs. The RMS current was 0.11 A. The voltage across the MOSFET per cell circuit was -5.45 V.

After turn on, the anode current rose exponentially to about +1.2 A. At the instant the MOSFET ceased to conduct, the current rose for a short time to about +1.6 A then fell rapidly to a negative value of about 1.4 A. It then rose to about +0.8 A before decaying to zero in under-damped sinusoidal oscillation.

At turn on, the anode voltage rose transiently to near the gate voltage then fell to about 4.5 V before stabilising at about 5.45 V. At turn off, the voltage fell for a short time to near zero then rose to about 4 V for about 120 ns. This period coincided almost exactly with the short rise in current just before its rapid fall. The voltage then rose sharply to about 40 V in a narrow spike with width at half peak height of about 100 ns, the peak occurring almost exactly at the point of inflection of the rapid anode current fall. Afterwards, the voltage fell below zero before rising to a second narrow spike of about 17 V before approaching about 6 V in under-damped oscillation. The second voltage spike appeared to be associated with the second current peak.

The cathode current, Figure 3, followed the same general form as the anode current except for the superimposition of high frequency oscillations, the significantly less negative value of the negative current and a much sharper secondary peak.

The cathode voltage showed an essentially similar form to the anode voltage except for the lower baseline value associated with the voltage drop across the cell. After turn off, it fell for a short time below 0 V then rose to about 0 V for about 120 ns. Again, this period corresponded almost exactly to the slight current rise just before its rapid fall.

Both sets of data show that the rapid fall in current took place -200 ns after the MOSFET ceased to conduct.

Substituting the BY328 diode 24 in the anode circuit with a diode having lower peak reverse voltage than the anode voltage spike reduced the anode voltage spike to about the peak reverse voltage of the new diode yet did not alter the cathode voltage spike. This proved that the energy producing the cathode spike was transmitted from the electrolyte to the cathode rather than through the external electrical circuit. Similarly, connecting a zener diode with breakdown voltage lower than the cathode voltage spike between the drain of a MOSFET 26 and ground caused the cathode voltage spike to fall to about the breakdown voltage of the zener diode yet did not alter the anode voltage spike. This proved that the energy causing the anode voltage spike was transmitted from the electrolyte rather than through the external electrical circuit.

Connecting a low dielectric loss capacitor of approximately 80 pF/cm2 of cathode plan area between the cathode and earth reduced the peak value of the voltage spike on the cathode to about half its value in the absence of the capacitor and increased its width with little loss of energy. This demonstrated that simple modification of the electronic circuit permits the use of MOSFETs with substantially lower peak reverse voltage and on resistance thereby reducing resistive power loss in the control circuitry.

The maximum value of the anode voltage spike was substantially reduced if the BY328 diode 24 was placed between the positive terminal of the 4,400 μF capacitor and the positive terminal of the power source 22 probably because of the known low impedance of electrolytic capacitors at high frequencies.

The shape of the voltage spike was essentially independent of frequency. Figures 4 and 5 show anode and cathode voltages at 10 Hz and 500 kHz. Also shown is the difference between the two signals. At the onset of electrolysis, before gassing caused short-term signal fluctuations, the maximum value of the voltage peak for each electrode was identical to within about 50 mV when a correction was made for baseline difference. The anode peak occurred slightly earlier than the cathode peak; hence the rise and fall of the difference between the two signals near the peak voltage. Switching probes proved this was not an artefact due to differences in signal velocity in the cable connecting the probe to the oscilloscope.

An oscilloscope probe placed in the electrolyte 7.3 cm from the cathode showed a very similar signal. Not only were all the voltage fluctuations, including the main spike, observed with approximately the same magnitude as the cathode voltage fluctuations, the time displacement increased to about 3 ns. Figure 6 shows the oscilloscope trace of the falling edge of the main peak. These data were AC coupled and averaged over 128 traces. By calculating the centroids of the data, a more accurate estimate was about 3.1 ns.

With anode switching, the main voltage spike was absent. Figure 7 shows anode current and voltage. Figure 8 shows cathode current and voltage. The positive voltage spike was completely absent from both anode and cathode signals, although a minor negative spike was seen for the cathode voltage.

A calibration experiment with a non-inductive carbon resistor of value equal to the direct current solution resistance of the cell proved that induction in the external electronic circuit was much less than required to explain these phenomena. Therefore, with cathode switching there appeared to be a pseudo-inductive effect within the cell that gave a reactive potential when the anode ceased to conduct.

Lenz's Law states that any induced quantity has a polarity or direction that opposes the cause that induces it. It is demonstrated by the reactive voltage change, ΔV, generated when the electrical current through an inductor is changed. ΔV is superimposed on the previously existing electrical potential and is proportional to -ΔV/dl/dt, where I is the current. The coefficient of proportionality is the inductance. Therefore, if the voltage spike is an induced phenomenon, a plot of -dl/dt vs. time should show the same general shape as the high frequency components of the voltage-time data and -ΔV/(dI/dt) is the pseudo-inductance.

Figures 9 to 12 are graphs of -dl/dt vs time calculated from the current data in Figures 2, 3, 7 and 8 by smoothing with a Gaussian kernel algorithm and differentiating sub-interval polynomial fitting functions. Also shown are the real voltage-time data. The data in Figures 9 and 10 show that -dl/dt replicates the shapes of the main voltage spike reasonably well. However, Figures 11 and 12 predict a voltage spike for the anode switching data that is not observed in the real data. Figures 13 and 14 shows 10⁶.V/(-dI/dt) calculated from current data in Figures 2 and 3. Also shown are the voltage data from Figures 2 and 3. The results indicate that within the period of the voltage spike, the pseudo-inductance was about 0.5-3 μH for both sets of data. The variation of pseudo-inductance with time is to be expected because inductance can show strong dependence on frequency and the centre of the voltage spike represents the highest frequency. The different forms of variation of pseudo-inductance with respect to time for the anode and the cathode data are less understandable. Outside this time range, the data were highly scattered. Figure 14 shows data for a similar experiment at slightly higher current and cell voltage. The trend of the pseudo-inductance values data was very similar to those in Figure 13.

The proof that the effect was due to a Lenz's Law phenomenon was obtained by systematically increasing the gate resistor to reduce the fall times of the MOSFET and the current. Figure 15 shows pseudo-inductance, current fall time and voltage spike height above baseline, all at peak voltage spike level, as a function of gate resistor value. The data show that within about +/-5% the pseudo-inductance was independent of fall time. This level of inductance should be compared with the typical value of the drain inductance for power MOSFETs of ~5 nH, i.e. two orders of magnitude lower than observed in the present work.

Further insight into the conduction mechanism was obtained by examining the temperature and frequency-dependence of conductivity. Figure 16 shows the ratio of the average current at 25°C to that at 6°C for direct current and cathode switching from 10 kHz to 500 kHz with a mark-space ratio of 0.2. The data show that the direct current conductivity ratio was -1.45, about the same as reported elsewhere both. However, as frequency increased, the ratio fell rapidly then plateaued at about 1.15 before falling rapidly at >400 kHz to about unity.

The absence at high frequencies of any temperature-dependence suggests the conduction is by a mechanism significantly different than direct current conduction. It is believed that this based upon the slow diffusional drift of hydrated ions. One possibility is the increase in ionic conductivity at excitation frequencies >3MHz, believed to be due to ions breaking free from solute atmospheres. However, the apparent propagation velocity of the pulses is many orders of magnitude faster than can be explained by ionic diffusion. Furthermore, the maximum rate of change of current observed in the present work of about 108 A/cm2 corresponds to a frequency of about 100 MHz.

The evidence suggests that the high voltage spikes were caused by under-damped oscillation of ionic current at the anode following turn-off of the cathode. One problem was that when the maximum current exceeded about 7.5 mA/cm2, the system oscillated in an uncontrolled fashion. To counter this, split ferrite cores were added. These had impedance Z at 25 MHz of about 100 TJ corresponding to an inductance at this frequency calculated from the formula L=Z/2Mf (f=frequency) of about 0.6 μH. One core added to the anode conductor cable near the anode stopped oscillation, but two with the second either on the anode line or on the cathode-MOSFET drain line ensured the voltage spike height was maximised for a given rms current. This corresponded to a total inductance of about -1.7 -3.5 μH including the pseudo-inductance of the cell. Experiment proved that this reduced the maximum rate of change of current by about two orders of magnitude.

Figures 2 and 3 show that the first event after turn-off is -30% rise in anode and cathode current. Careful analysis of the data showed that the potential of the cathode fell by a little under 2 V compared with the previous steady state value. This is reasonable evidence of fast rearrangement of the cathode diffuse electric double layer and could account for the transient rise in cathode current. Further evidence that the effect was initiated by the rearrangement of the cathode diffuse electric- double layer was obtained by examining the characteristics of a larger experimental rig with 1575 cm² electrode area. The rms cell current for a 5 g/1 sodium chloride/tap water electrolyte at a frequency of 277 kHz, cell voltage 7.2 V and pulse width about 0.5 μs was only about 100 mA or about 6.3x10-5 A/cm2, a factor of about 40 lower than for the smaller cell operating under the same conditions, almost exactly the ratio of the electrode areas and capacitances. The implication is that even for operation at the higher current density of the smaller cell, most of the energy per pulse was stored capacitatively in the diffuse electric double layer.

However, despite the much lower current density of the larger cell, substantial voltage spikes were still observed, approximately 30 V high and about 400 ns wide at half maximum height. To get approximately the same voltage spike height and width as the smaller cell, it was necessary to have a pulse width of -1 μs giving -1 A rms current. Thus, the current density was a quarter of that of the smaller cell. One interpretation is that it is not necessary fully to establish the cathode diffuse electric double layer in order to obtain adequate voltage spike height. If so, and the effectiveness of the system results solely from the voltage spikes, suitable matching of source impedance and electrode capacitance might enable the cell to be used at very low current densities compared with other electrochemical devices, substantially reducing running costs and the effect of coulombic side reactions. Ionic drift is too slow a process to account for the apparently simultaneous increase in anode and cathode current in Figures 2 and 3 especially since there was no evidence at this time of a significant change in anode potential compared with the previous steady state value. Therefore, another process has to be responsible for communicating across the cell to the anode what has happened at the cathode. Figure 17 shows an oscilloscope trace of anode voltage and current for operation of this circuit with dual cathodes, average current 0.59 A, pulse length 0.4 μs, base frequency 600 kHz (giving a frequency of 1200 kHz in the anode measurements). The peak value of the voltage spikes on start-up was about 250 V, but onset of gassing at the electrodes caused this to fall to about 150 V. This apparatus was used for most of the subsequent porous electrode experiments.

With regard to experiments conducted for non-porous electrodes, the same control electronics was used to test the operation of the three plate anodes. Using an electrolyte of 5g/l sodium chloride in tap water, the three plate anodes produced essentially the same voltage effects as the expanded titanium anode. Indeed, the heights of the voltage spikes for a given current density were about the same thus proving the effect was not caused by the presence of porosity.

Experiment 3: Treatment of Industrial Effluents containing Ammonia.

Experiments were also conducted using the present apparatus to treat ammoniacal effluents. To assess the performance of the porous electrode cell for ammonia, the electrolyte consisted of 600 ml tap water with 3 g sodium chloride and 2.5 ml saturated ammonia solution (density 0.88 g/cm3), giving a starting concentration of ammonia of about 900 mg 1 by weight (as nitrogen) and pH approximately 1 1. The temperature was 6-8°C. Ammonia levels were measured using a gas sensing combination electrode. Before each measurement, the electrode was calibrated precisely against a solution of ammonium chloride of known concentration. The results are reported as the concentration of nitrogen in mg/1. The limit of accuracy of the ammonia measurement was about +/- 5%. Nitrate and nitrite levels were assessed using commercially available photometric test kits.

Ammonia in water exists in two forms; NH₃ and NHT ions. NH₃ is volatile and the equilibrium shifts as pH and temperature change. Standard data enable the shift in equilibrium to be predicted accurately. At 8°C and pH 11, 94.1% is molecular so to reduce ammonia loss, the cell was covered with a plastic sheet.

The theoretical value of e/NH3 can be calculated from the assumed reaction. Two possible reactions have been proposed for chloride-containing electrolytes. An indirect reaction:

2NH₄ ⁺ + 3HOC1 ' N₂ + 3H₂O + 5 H⁺ -(9)

This requires 11 electrons to discharge the protons and to recreate the hypochlorite ions.

A direct reaction: 2NH₃ + 6OH- ' N₂ + 6H₂O + 6e^" -( 10) This requires twelve electrons, six direct and six to produce the chlorine required to react with the hydroxyls to make hypochlorite. Therefore, a reasonable estimate of the theoretical e/NH₃ is 11 for an ammonium ion reaction or 12 for an ammonia gas reaction.

Figure 18 shows ammonia concentration as a function of time for 384.6 kHz excitation, rms current 0.33 A, 10.17 V cell voltage, peak spike voltage 82 V, 5 g/1 NaCl. The ammonia concentration fell approximately linearly with time. Also shown is a graph showing the cumulative number of electronic charges required to destroy each ammonia molecule (e/NH₃). The data show that this tended to a limiting value at long times of about 4.3, or an apparent faradaic efficiency of about 230% assuming an ammonia gas reaction.

The solution was tested for the presence of nitrate and nitrite ions; nitrite was never more than the detection limit, nitrate levels reached about 20 mg/1 as N then levelled off. It was concluded that either the process did not oxidise all the nitrogen in the ammonia to nitrate or nitrite or that it also decomposed nitrite and nitrite ions.

Figure 19 shows results of an experiment with lower ion concentration (0.5 g/1 NaCl). The limiting value of e/NH₃ was about 5.5.

Figure 20 shows results of an experiment similar to Figure 18, but at lower current and voltage spike height. e/NH₃ at about 3.7 was lower and the apparent faradic efficiency of 324%) was substantially higher.

Figure 21 shows results of an experiment similar to Figure 20 but at 500 kHz. e/NH₃ tended to about 5.

Figure 22 shows results of an experiment similar to Figure 18 except that the cell voltage was increased substantially thereby giving about 50% higher current and voltage spikes of 95 and 105 V (a result of asymmetry of anode-cathode spacing). The limit of e/NH₃ was apparently substantially lower, at about 3.4.

An experiment was also carried out to assess ammonia loss rate in the absence of electrolysis. Over a 24-hour period, the loss was -200 mg/1, an average of about 8 mg/l/hr. Therefore, it is believed that the data in Figures 18-22 may have been overstated, but only by a maximum of -10%. Experiment 4

An experiment was also carried out to measure the rate of ammonia loss using direct current of 0.33A, approximately the same as for Figure 18, other conditions being the same. After 24 hours, ammonia reduced from 850 mg/1 to 440 mg/1. This corresponds to e/NH3 about 16.5 or an apparent Faradic efficiency of about 71%. Nitrate levels continuously increased indicating that for direct current operation the dominant process appeared to be oxidation by hypochlorite ions. An estimate of ammonia loss due to partitioning to the hydrogen can be made if it is assumed that the partial pressure of ammonia in hydrogen is the same as in air. The concentration of ammonia was about 0.039 M giving a total quantity of about 0.023 mols. Standard data (Perry et. al., 1998) indicate that the partial pressure of ammonia over a non-pH corrected 0.05 M solution in water at 10°C is 0.032 Bar. Therefore, an upper bound estimate of ammonia loss is the hydrogen produced (0.067 mols in 12 hours at 0.3 A) multiplied by this partial pressure which is 0.0021 mols or about 9% of the total. Therefore, partitioning to the hydrogen did not exceed 10%. Furthermore, all the data showed that e/NH₃ approached a horizontal asymptote; if substantial ammonia partitioning had taken place, e/NH₃ would have tended to increase as [NH₃] decreased.

Experiment 5

To establish the effect of absence of chloride and reduction of pH, an experiment was done with 3 g/1 anhydrous magnesium sulphate and 2.5 ml saturated ammonia solution in 600 ml tap water. Sulphuric acid was used to reduce the pH to 7. The ammonia reduction kinetics were much slower and showed an exponentially-decreasing form. The experimental value of e/NH₃ for a reduction of 800 mg/1 was about 21. Increasing pH to 10.5 increased the kinetics considerably. The rate of decrease of ammonia concentration appeared linear and tended to an average value of e/NH3 of about 8. Thus, chloride gave faster kinetics and pH had to be sufficient to ensure a substantial proportion of the ammonia was in the molecular form.

Experiment 6

An experiment was also conducted to examine the effect of reducing pH in the presence of chloride. The electrolyte was the same as for Figure 18 except that pH was modified by hydrochloric acid to approximately 7 and 9. At pH 7, the rate of ammonia removal was similar to that of the pH 7 magnesium sulphate experiment. At pH 9 (15% free NH3) the kinetics tended to be linear with time and e/NH3 was approximately 7. The conclusion is that ammonia destruction kinetics were fastest for high pH and in the presence of chloride. The highest reliable measure of the apparent Faradaic efficiency was 324% with 200% being the norm. These results appear substantially better than reported elsewhere and because the fall was in linear with respect to time in those cases where pH was than about 8 appear to have kinetics independent of concentration.

Experiment 7: Destruction of Ammonia in Landfill Leachate.

Two representative landfill leachates were tested. Figure 23 shows a graph of the concentration of ammonia vs. time for a leachate that had been diluted with rainwater. The starting pH was -7.8 and the starting ammonia concentration was about 170 mg/1 as N. The electrochemical cell consisted of the non-porous coated titanium anode with two, non-porous, 3161 stainless steel sheet cathodes of area equal to the anode mounted either side with an inter-electrode spacing of 2mm. Fluid flowed at an average velocity of approximately of 1 cm/s along one channel then reversed to flow along the second channel. This flow velocity was insufficient to cause turbulent flow. The pulse frequency was scanned using a saw tooth wave at 40 Hz to sweep the main frequency from 200-400 kHz and vice-versa. The voltage spikes were 60-40 V depending on pulsing frequency.

The ammonia concentration fell approximately linearly and e NH₃ tended to -10.5 even down to 3.3 mg 1 ammonia as N, low enough for the effluent legally to be discharged to a water course. Furthermore, the calculated power cost of treatment assuming £0.05/kWhr was about £0.01/m³, very low compared with the typical £l/m³ power cost for air stripping and also much lower than the cost of treatment by microbes.

Figure 24 shows data for a much more concentrated effluent treated in the porous electrode cell. C.O.D. was measured by light absorption of Cr(III) using acidified dichromate test kits. Sodium chloride was added to make the total concentration about 10 g/1. Starting pH was 8.1, ammonia >3,000 mg/1 and C.O.D. about 9,000 mg/1. Frequency was scanned from 80-500 kHz. The voltage spikes were 60-40 V depending on pulsing frequency. Ammonia and COD fell with time approximately linearly. The limiting e/NH₃ was approximately 4. Nitrite levels were consistently about 7 mg/1 as N. Nitrate remained consistently about 1,400 mg/1 as N. These data showed that there was no relationship between nitrite and nitrate levels and the quantity of ammonia destroyed, again suggesting that the nitrogen in the ammonia was converted to a species other than nitrite or nitrate and that nitrite and nitrate may have remained unaffected by the process.

When a sample of this leachate was treated for an extended period then analysed by a gas chromatograph fitted with a mass spectrometer detector, it was discovered that a significant quantity, about 9,000 μg/1 of trichloromethane and -5,000 μg/1 of romodichloromethane had appeared when none had been detected in the raw feed. Haloforms are relatively stable compounds of moderate solubility in water and having a relatively low Henry's Law coefficient so would be expected preferentially to accumulate. They can only have been formed by reaction of halogen ions with C-l free radicals produced by the cleavage of larger molecules by an active agent such as hydroxyl radicals.

Experiment 8: Destruction of Gentian Violet (4-dimethylaminophenyl methane)

This intensely light absorbing dye is used in relatively large quantities for forensic science and general medicine. The typical concentration is 1 wt/vol %. Measuring the light absorbance at 650 nm followed the process of electrochemical destruction.

0.3 g sodium chloride and 0.6 g crystal violet were dissolved in 600 ml tap water. An experiment was done using direct current, rms current 0.23 A, cell voltage 4.7 V in the porous electrode cell. After 24 hours, the colour remained intense although an oily scum had formed on the surface.

A pulsed current experiment using the same fluid was done at 9.0 V, 384.6 kHz and 0.32 A. The voltage spikes were about 60 V. In the first twenty minutes, an oily, highly coloured foam separated. This was removed using a spatula. After two hours, the liquid became effectively transparent with a light grey colloidal precipitate. After this had settled, the absorbance at 650 nm was 0.088 and there was no evidence of the intense specific light absorption between 350 nm and 650 nm showed by the pure dye.

This experiment proved that the process rapidly destroyed the dye molecule making an insoluble precipitate. Most of the precipitate was bought to the surface by the hydrogen bubbles generated in the process. The power used was 5 kWhr/m3 which at £0.05/kWhr would cost ~£0.30/m3, much less than the ~£75/m³ cost of disposing the solution as toxic waste.

Calcium may be deposited on the electrodes over a period of time depending upon the nature of the effluent passing through the apparatus. Arrangements to clean the electrodes can be arranged in this event.

Further Embodiment

The apparatus 10 of the embodiment of Figure 26 consists of an electrolytic cell 12, the electrodes 114, 116 of which are connected to an AC power supply 118 through a circuit 120. The tank 122 of the electrolytic cell 12 is connected to an effluent through flow system 124. In the embodiment the AC power supply 118 is a single phase 230 volt 30 amp supply, in other words the domestic mains supply. It powers electronic controls and instrumentation 126 through a switch 128. The power supply 118 is connected to the electrodes 114, 116 through a solenoid operated contactor switch 130, fuses 134 and a timer 136 connected in series to a transformer 138. The transformer 138 is connected to a bridge rectifier 140 and then through smoothing inductors 142 to the electrodes 114, 116. A switch 144 is provided in the connection to each cathode 114 so that the switches 144 are all in parallel.

The tank 122 includes four partitioned cells 146. Each cell 146 includes an anode 116 between two cathodes 114. Each cell includes a catalyst in the form of a platinum mesh 148.

The tank 122 further includes fluid level controllers 150 and thermometers 152 which act in conjunction with an over temperature switch 154 to control a radiator 156 in the effluent supply line to heat the effluent. The effluent supply circuit 124 consists of a pump 158 to circulate effluent towards the radiators 156, which may be a 5 kW radiator. The radiator 156 is connected through a flow switch 160 and a flow meter 162 into the tank 122.

In use, the AC supply operated upon by the rectifier bridge 140 and inductors 142 creates a waveform which approximates to DC. The switch 144 on each cathode 114 switches the DC supply to the cathode 114 on and off in the form of a square wave so that first one cathode and then the other cathode receives current in sequence. The switches 144 are MOSFET switches operated by a sine wave 164 at upper and lower thresholds 166, 168 to give the waveforms 170, 172 shown in Figure 2 at the cathode 114. The frequency of alternation of current between the cathodes 114 can be raised as high as 600 kHz giving excellent results in breakdown of organic molecules.

Looking at the waveform of the cathodes 114 in more detail, as can be seen from Figure 2, the left hand electrode 1 14 is switched on by the respective switch 144 for a period t and then is switched off for a period T in cycle. The current is switched to the right electrode 114 in the same pattern, but out of phase and it will also be seen that there are delay periods d in which neither of the cathodes 114 is switched on, the periods d being shorter than the period t (which in turn is shorter than the period T). It is when the current is switched from one cathode to the other that the voltage spike is observed of around 46 volts. At a frequency of 20 kHz the technique is a vast improvement over known techniques in breaking down large organic molecules and thereby reducing COD. The efficiency of the process is observed to rise in linear relationship with frequency so that extremely good results are achieved at 300 kHz and even better results at 600 kHz, which is approaching the limit of the switched employed.

A brief 46 volt pulse is seen over microseconds. The inductor 42, which was originally included to smooth the DC supply, exhibits a reverse EMF facilitating the voltage spikes.

An explanation is that water is being split into H+ and OH- with creation of OH radicals which have very powerful oxidising properties. Hydroxyl radical techniques are already known elsewhere, as described, but the present technique is a great improvement over them.

Claims

1. Apparatus for treating aqueous wastes, such as waste water, landfill leachates and aqueous effluents, the apparatus comprising electrolytic cell and electrical pulse generating means for generating a series of current pulses between the electrodes of the cell.

2. Apparatus as claimed in claim 1, wherein the apparatus is arranged such that the pulses create a series of voltage spikes across the cell.

3. Apparatus as claimed in claim 2, wherein the voltage spikes are greater than 10 volts with respect to the potential of the anode in the absence of current pulsing.

4. Apparatus as claimed in claim 2, wherein the voltage spikes are greater than 40 volts with respect to the potential of the anode in the absence of current pulsing.

5. Apparatus as claimed in claim 2, wherein the voltage spikes are greater than 60 volts with respect to the potential of the anode in the absence of current pulsing.

6. Apparatus as claimed in claim 2, wherein the voltage spikes are greater than 100 volts with respect to the potential of the anode in the absence of current pulsing.

7. Apparatus as claimed in any preceding claim, wherein the current pulses are at a frequency of greater than 20 kHz.

8. Apparatus as claimed in any preceding claim, wherein the current pulses are at a frequency of greater than 100 kHz.

9. Apparatus as claimed in any preceding claim, wherein the current pulses are at a frequency of up to 500 kHz.

10. Apparatus as claimed in any preceding claim, wherein the electrodes are less than 4 mm apart.

1 1. Apparatus as claimed in any preceding claim, wherein the electrodes are more than 1 mm apart.

12. Apparatus as claimed in any preceding claim, wherein the electrodes are about 2 mm apart.

13. Apparatus as claimed in any preceding claim, wherein the current density of the electrodes is between 50 and 250 Am^'2.

14. Apparatus as claimed in any preceding claim wherein at least the anode or anodes are made of titanium coated with a metal oxide.

15. Apparatus as claimed in any preceding claim, wherein at least the anode or anodes are made of titanium coated with iridum doped ruthenium oxide.

16. Apparatus as claimed in any preceding claim, wherein at least the anode or anodes present a free carbon surface.

17. Apparatus as claimed in any preceding claim, wherein at least the anode or anodes comprise carbon fibre fabric.

18. Apparatus as claimed in claim 14 or claim 15, wherein the voltage applied to the apparatus is between 6 volts and 12 volts.

19. Apparatus as claimed in claim 16 or claim 17, wherein the voltage applied to the apparatus is between 10 volts and 20 volts.

20. Apparatus as claimed in any preceding claim, wherein a diode is included in the connection to the or each anode to inhibit reverse currents therefrom.

21. Apparatus as claimed in any preceding claim, wherein a high speed current switching device is provided in the electrical connection to the or each cathode.

22. Apparatus as claimed in claim 21 wherein a plurality of cathodes are provided and a high speed current switching device is provided in the connection to each cathode.

23. Apparatus as claimed in claim 21 or claim 22, wherein the or each high speed current switching device comprises a MOSFET.

24. Apparatus as claimed in claim 23, wherein the electrical pulse generating means is connected to the or each MOSFET through a MOSFET driver circuit, and the or each MOSFET and its associated MOSFET driver circuit is mounted on the associated cathode or cathode support.

25. Apparatus as claimed in any preceding claim, wherein there are a plurality of anodes and cathodes.

26. Apparatus as claimed in claim 25, wherein the cathodes are arranged in at least two groups, and the electrical pulse generating means is arranged to generate pulses in the groups at different times.

27. Apparatus as claimed in claim 26, wherein there is a time delay between switching off one cathode group and switching on the next cathode group.

28. Apparatus as claimed in claim 26 or claim 27, wherein there are only two groups of cathodes.

29. Apparatus as claimed in claim 26, 27 or 28, wherein there are a plurality of cathodes in each group, and the cathodes of different groups are arranged alternate so that each anode has a cathode from two different groups mounted alongside it.

30. A method for treating aqueous wastes, such as waste water, landfill leachate and aqueous organic effluents, the method comprising introducing the aqueous waste into a electrolytic cell and creating a series of electrical pulses between the electrodes of the electrolytic cell.

31. A method as claimed in claim 30, wherein the pulses are arranged to create a series of voltage spikes across the cell.

32. A method as claimed in claim 31, wherein the voltage spikes are greater than 10 volts with respect to the potential of the anode and in the absence of pulsing.

33. A method as claimed in claim 31, wherein the voltage spikes are greater than 40 volts with respect to the potential of the anode and in the absence of pulsing.

34. A method as claimed in claim 31, wherein the voltage spikes are greater than 60 volts with respect to the potential of the anode in the absence of pulsing.

35. A method as claimed in claim 31, wherein the voltage spikes are greater than 100 volts with respect to the potential of the anode in the absence of pulsing.

36. A method as claimed in any of claims 30 to 35, wherein the electrical pulses are at a frequency of greater than 20 kHz.

37. A method as claimed in any of claims 30 to 36, wherein the current pulses are at a frequency of greater than 100 kHz.

38. A method as claimed in any of claims 30 to 37, wherein the electric pulses are at a frequency of up to 500 kHz.

39. A method as claimed in any of claims 30 to 38, wherein the applied current density of the electrodes is between 50 and 250 Am^"2.

40. A method as claimed in any of claims 30 to 39, wherein at least two groups of cathodes are provided and an electrical pulse is created in each group at a different time.

41. A method as claimed in claim 40, wherein there is a delay between applying a pulse to one cathode group and applying a pulse to the next cathode group.

42. A method as claimed in any of claims 30 to 41, wherein the method includes filtering the aqueous waste before introducing it into the electrolyte cell.

43. A method as claimed in any of claims 30 to 42, wherein the method includes maintaining the ph of the aqueous waste in the electrolyte cell at between PH7 and PH9.

44. A method as claimed in any of claims 30 to 43, wherein the method includes maintaining the ph of the aqueous waste in the electrolyte cell at at least PH8.

45. A method as claimed in any of claims 30 to 44, wherein the method includes adding a halogen to the aqueous waste.

46. A method as claimed in any of claims 30 to 45, wherein the method includes adding a chloride to the aqueous waste.

47. A method as claimed in claim 46, wherein the method includes adding a chloride to the aqueous waste to a concentration not exceeding 2000 ml of chloride per litre of aqueous waste.

48. A method as claimed in any of claims 30 to 47, wherein the aqueous waste is softened prior to the application of electrical pulses.

49. A method as claimed in any of claims 30 to 48, wherein the method includes passing the aqueous waste or allowing the aqueous waste to pass tluough the electrolytic cell at a flow rate of less than 1 m5^"'.

50. A method as claimed in claim 49, wherein the method includes passing the aqueous waste or allowing the aqueous waste to pass through the electrolytic cell at a flow rate of less than 10 m.5^"1.