GB2322087A - Removal of pollutant - Google Patents

Abstract

A method for the removal from a fluid of inorganic pollutants, especially organo-tin biocides, used as antifouling agents on boat hulls, and which are removed during routine maintenance of the hull. The present invention attempts to overcome the problem of simply discharging the polluted water produced.

Description

Removal of Pollutants This invention is concerned with the removal of organic and organo-metallic pollutants from an aqueous environment. In particular, it is concerned with the removal of antifouling biocides from docks and drydocks, especially the removal of organo-tin biocides from the water generated as a consequence of the high pressure hosing of vessels in drydocking.

Organo-tin pollutants can be introduced into the environment by several activities; organo-tin is disposed of in used paint drums, it also enters the environment in the form of "overspray" of paint, whereby the organo-tin-containing paint is not applied correctly so that paint droplets fall onto dock-floors.

These problems have largely been addressed by the use of larger reusable containers for paint, which avoid the disposal of the previous containers. More advanced painting techniques have overcome the problem of overspray.

One important source of organo-tin pollution is the contamination arising from total reblasting of old and aging ships' hulls. The reblasting process involves high pressure hosing with fresh water and blasting grit.

High pressure hosing with fresh water is a normal prelude to further repair and maintenance procedures, and is necessary even when the vessel is free of macroscopic fouling, since the surface of the hull is almost invariably both slimy and salty. The high pressure hosing is necessary to remove both slime and salt, and the spent residues of previous antifouling, in preparation for recoating.

In worse cases where blasting of the hull is required each square metre of hull requires 40 kg of blasting grit. The waste produced in this process contains approximately 0.25% organo-tin.

We have discovered that the disposal of contaminated grit is a matter of minor concern with a positive environmental outcome. It can, for example, be incinerated in cement works with a beneficial outcome.

However, it is a substantial problem to dispose of the waste water produced. Allisson, [MOD] gives an estimate of 4 ppm of tri-butyl tin oxide (TBTO) in waste water, but this is about two million times the quoted Environmental Quality Standard (EQS). High pressure hosing requires 20-40 litres of fresh or potable water per square metre of hull, so that cleaning a 2000 square metre ship will generate 40 tonnes of waste water. Cleaning Ultra-Large Crude Carriers (ULCC) of 40,000 square metres will generate 800 tonnes of waste water. Depending on the integrity of the dock gates, this may be contaminated with an even larger volume of sea water. The organo-tin content, at several parts per million, is 10,000 to one million times greater than the Environmental Quality Standard, and most of the organo-tin must be removed before the water can be returned to the river or estuary under current UK legislation. Removal by tanker to landfill is both expensive and wasteful, given the small volume of pollutant present.

According to the present invention there is provided a means of treating a fluid which is contaminated with a pollutant, the method comprising mixing of said fluid with a second fluid immiscible with the first fluid, wherein the pollutant is more soluble in the second fluid than the first fluid.

The first fluid can be an aqueous solution.

The second fluid can be a hydrophobic fluid.

Preferably, the second fluid is not miscible with water.

Preferably also the surface area of the interface between the two fluids is maximised, e.g. by creating a dispersion of droplets of one fluid within the other.

Preferably also the pollutant is an inorganic species.

The second fluid can be mixed in small amounts with the first fluid. The pollutant in the first fluid then enters the second fluid phase, and the first and second fluids can then be separated and the second fluid and pollutant discarded.

Most preferably, the pollutant is a Tri-Alkyl Tin compound.

Preferably, the second fluid is chosen from the list of engine oil, slops oil, "waste" oil, silicone oil, isopropanol, pentanol, toluene or singly or multiply alkyl substituted toluene, where the alkyl substituents are chosen from methyl, ethyl, propyl or butyl.

Tri-butyl tin oxide or chloride have been the most commonly used antifoulants apart from copper for the last thirty years. The exact species present in water depends on the chloride concentration and the pH in the waste stream. The invention is especially (but not exclusively) concerned with that chemical, as it is that chemical which is usually released from a selfpolishing copolymeric antifouling material as described in Patent GB 1,457,590.

Preferably, the contaminated fluid is mixed with the immiscible fluid by passing through a rotorand screen (essentially a Silverson disperser) to form a dispersion of uniform particle size.

The dispersion is sustained optionally as long as it is necessary to effect exchange of the contaminant into the immiscible fluid, under varying conditions of temperature and salinity.

The dispersion can be created by mixing a fraction of the contaminated fluid with the immiscible fluid, before mixing the dispersion thus produced with the remainder of the contaminated liquid to create a homogenous dispersion.

The dispersion can be separated into component fluids in a flow smoothing section and a cyclone separator.

Pre-accumulation from dockyard and ships' structures of the effluent to be treated generally comprises collection and gathering by systems or networks of gulleys, troughs, drains and low points, either existing, modified or specially created for the purpose. Effluent is then accumulated in a sump etc.

Solids materials exceeding eg 20 microns in diameter are preferably removed from the effluent before it enters treatment by the solvent dosing and mass transfer reaction stage of the process. The particulate profile of solids present in the accumulated effluent may vary widely in particle shape, size, density etc so that extraction can be effected by a series of stages, each stage being a process appropriate to the physical shape, dimensions and concentration of particles which are to be extracted in that stage. The series of extraction mechanisms may comprise, where functionally appropriate, gravitational settlement (at 1 g), mechanical screening, matrix filtration, or intensified gravitational treatments (centrifugation). The first two (gravitational settlement and mechanical screening) may be incorporated into the gathering and accumulation measures in the dockyard structures and systems.

Untreated solids extracted during pretreatment of the dockyard effluent are preferably disposed of appropriately under local regulations controlling such wastes contaminated by the particular effluent contaminant. Where the volume quantities of the extracted contaminated solids and/or operational circumstances are such that disposal under prevailing regulatory controls is commercially disadvantageous the solids may be accumulated and drench washed by irrigation with estuary or river water. The effluent run off from the washing process is in turn treated by the solvent dosing and mass transfer reaction system until levels of contaminant in the solids are sufficiently low to allow its return to the environment; irrigation of the solids may alternatively be continued until contamination of the solids is reduced to a level permitting more commercially attractive methods of disposal under local regulatory control.

An embodiment of the invention will now be described, by way of example with reference to the accompanying drawings, in which: Fig. 1 shows a schematic representation of the process; Fig. 2 shows a schematic representation of the preferred core process; and Fig. 3 shows a separator used in an outfall and solvent scrub in the process.

The method typically comprises four stages, described as follows: 1. Injection and mixing with a hydrophobic oil phase; 2. Extraction and equilibration; 3. Coalescence and demixing; 4. Return of the purified phase to the environment; and 5. Recirculation and/or removal of the oil phase.

1. Injection and Mixing of the Oil Phase Tributyl tin compounds are salt-forming compounds capable of forming ionic salts with hydrochloric acid and acetic acid. They also have hydrophobic properties.

The partition coefficient of tri-n-butyl tin chloride in a number of materials ranges from about 2000 to 20,000 in materials such as iso-Octanol, liquid paraffin, and is particularly high in silicone fluids and silicone elastomers, ranging from 10,000 to 20,000.

The partition coefficient is also high in stabilised North Sea crude oil, in slops-oil, and waste fabricating oil; all having a coefficient of approximately 5,000. The reasons for preferring such materials is that when it is not intended to recycle them they are sufficiently cheap to go directly to incineration. If the material used is to be recycled then more expensive materials like methyl siloxane fluids may be used.

In one preferred embodiment of the invention the rheology of the oil phase remains stable over a range of temperatures, which makes silicone fluids and engine oils, or mixtures with a significant proportion of waste engine oils convenient. Other reasons for preferring crude oils and waste engine oils is that the former have a significant amount of sulphur present, some of which is present in thiol (-SH) groups, with which tri-butyl tin oxide reacts strongly, while waste lubricating oils have a large amount of carbon particulates to which the tri-butyl tin oxide is also strongly absorbed. These materials, normally considered as waste, are therefore ideal for our purpose.

A number of methods of mixing are available, such as injection through an atomizer, for example. However, to maintain the most economic ratio of water to "oil" and to maintain good control over particle size distribution the preferred mixing method is the use of a Silverson type mixer in a side-loop of the main water duct.

The organic material is generally introduced to the polluted aqueous effluent by being atomised into small droplets or size 5-6ym approx. within the aqueous effluent creating an interface of a very large surface area between the two materials. Incorporation of the organic material, with appropriate atomisation, into the total volume flow of effluent is possible but the entire flow of organic material is preferably initially introduced into an extracted fraction of the total effluent flow, wherein it is dispersed to create a dispersion concentrate, which is re-introduced into and mixed with the remaining main effluent flow. Thus, producing a reduced dispersion concentration of organic fractional proportion, droplet size and interfacial surface area appropriate for reaction and transfer of the pollutant from dilute solution in the aqueous effluent to concentrated solution in the organic material.

The organic volume fraction in the dispersion concentrate is significantly higher than that required for reaction and transfer of the effluent pollutant.

The organic dosing and pre-treatment system is designed and configured to allow selection, by the system operator, from a range of candidate organic materials varying in specific gravity, rheology, surface tension and surface chemistry. The system operator can then select an appropriate organic material against locally dominant commercial, logistic, environmental and or other criteria.

l(a) Effluent Flow Proportioning A fraction of the total effluent flow volume is derived from the main flow pipe through the organic dosing and pre-treatment system by modulation of the flow rate allowed through the diversion valve. The size of this fraction is determined by comparison of the flow rate measured by the pre-treatment flow monitor to the flow rate measured by the total effluent flow monitor. The ratio of flow rates is continuously monitored by the control system and controlled by its proportional modulation of the diversion valve.

l(b) Organic Injection The chosen organic material is injected into the fractional effluent flow passing through the dosing and pre-treatment system. Organic injection is introduced upstream of the rotary atomisation system.

l(c) Organic Proportioning The chosen organic material is introduced at a rate such that its fractional proportion of the total effluent flow is appropriate for reaction and transfer of the pollutant from dilute solution in the aqueous effluent to concentrated solution in the organic material. Organic injection is by rotary internal gear pump.

Over the range of total effluent flow rates for any size category of apparatus, the magnitude and variation of prevailing pressures and flow heads of effluent upstream of the rotary atomisation system is preferably constrained to modest magnitudes and within narrow limits of variation, by fundamental design of the apparatus and by the action of the control system. The consequently modest pressure ratio developed across the internal gear injection pump encourages good proportionality between its volume flow rate and its rate of rotation when pumping any suitable organic material.

The fractional proportion of organic material to effluent can be adjusted in accordance with the salinity of the dock effluent water which will depend upon marine dock water treatment applications and by the choice of organic material. The volume flow rate of organic material required is determined by comparison of the total effluent flow rate measured by the total effluent flow monitor to the required fractional proportion set by the operator. Both the organic flow rate required and the injection pump rotation rate are continuously monitored by the control system, which also maintains the necessary fractional proportion of organic material to effluent by its proportional modulation of the organic injection pump rotation rate.

l(d) Organic Atomisation The organic material is atomised into the aqueous effluent within the rotary atomisation system, creating a dispersion droplet size distribution and interfacial surface area appropriate for reaction and transfer of the pollutant from dilute solution in the aqueous effluent to concentrated solution in the organic material.

Dispersion is preferably by a process of controlled viscous shear. The effluent and organic materials are constrained to pass between vaned impellers rotating within stationary perforated screens and thus through domains of high shear activity across which the gradients of shear are extremely steep.

The droplet size distribution parameters of the dispersion are both direct and indirect functions of shear gradient and therefore of the physical clearance dimension between the outer diameters of the impellers and the adjacent internal diameters of the stationary screens, and of the differential velocity between impellers and screens. The outer impeller and inner screen structures are nominally conical with equal included angles. The impeller-to-screen clearance dimension is varied by axial displacement of screen positions with respect to the impellers. The differential velocities between impellers and screens are a direct function of impeller angular velocities and are thus controlled by modulation and impeller speeds.

The droplet size distribution parameters of the dispersion are a further function of the duration of exposure of the effluent organic mixture to the effects of high shear. A secondary function of the impeller screen disperser is its development of a positive forward pressure head, which allows high shear exposure duration to be controlled within the rotary atomisation system by recirculation of part of the system flow drawn from the system outlet and returned to the system inlet.

The control system manages the recirculation flow rate and thus the high shear exposure duration by modulation of a restriction in the recirculation channel. It is also responsible for monitoring total effluent flow rate, effluent temperature, required fractional proportion of organic material, organic rheology, organic injection flow rate, and pre-treatment fractional flow rate. These parameters are set by the control system against operator input data including organic specific gravity and surface tension, and effluent salinity. The structures of the rotary atomisation system are positionally and dynamically manipulated as required by the control system in accordance with experimentally determined relational algorithms to create the appropriate dispersion of organic material within the aqueous effluent.

1(e) Organic Rheology Monitoring The droplet size distribution parameters of the dispersion are functions of the ratio of effluent and the rheology of the organic material. The rheology of the candidate aqueous effluents may be regarded as essentially constant, thus dispersion parameters become dependent upon the rheology of the organic material, which varies primarily as a function of the choice of the material. The apparatus may conceivably be required to process effluents in a range of climates from tropical to sub-arctic ambients with attendant variations in organic rheology, and is designed to accommodate these rheology variations.

The discharge port of the organic material gear injection pump is nominally restricted to ensure a minimum forward pressure rise across the pump for all operating conditions. The magnitude of this pressure rise is a function of the rheology of the organic material being pumped. The torque developed by the injection pump whilst pumping the organic material is a function of the pressure rise across the pump; the torque in turn influences the electrical current drawn by the injection pump drive motor. The control system continuously monitors the organic injection pump drive motor torque, thereby it is able, in accordance with calculated and experimentally determined relational algorithms, to continuously monitor the rheology (low to medium shear viscosity) of the organic material.

As a result of effluent temperature change the control system can thus shut down the system and report the attempted use of rheologically inappropriate organic material or the migration of a permitted material out of rheologically allowable limits.

l(f) Preserving the Integrity of the Dispersion Concentrate Following atomisation of the organic material, it is essential that the integrity of the dispersion concentrate is preserved intact, and that any natural tendency towards floatation, segregation, agglomeration or coalescence (strong in some preferred effluent organic combinations) is suppressed. The density fraction of organic droplets in the concentrate phase will be such as to represent a regime of reduced or hindered settlement velocity (cf free settlement velocity) thereby reducing natural tendencies to floatation segregation, agglomeration or coalescence.

The integrity of the dispersion concentrate can be further aided by the discharge network from the rotary atomisation system which is configured and dimensioned such as to ensure high linear flow velocities, thus ensuring fully turbulent high Reynolds number values.

In this way the transit time before incorporation into the main effluent flow is minimised and opportunity for decay of dispersion quality reduced.

l(g) Incorporating the Concentrate into the Main Effluent Flow The dispersion concentrate must be incorporated, preferably without damage, into the remaining bulk effluent flow, wherein the organic volume fraction concentration will be reduced to its working value.

This is attained in a venturi device situated in the main effluent flow pipe downstream of the organic dosing and pre-treatment system.

The dispersion concentrate is dynamically entrained into the main effluent flow by conventional venturi kinetics within the device. Violent mixing and powerful suppression of any natural tendency to floatation, segregation, agglomeration or coalescence, both within the device and in the outlet network immediately downstream of the device is promoted by the very high venturi nozzle and combination cone velocities. The fully developed effluent organic dispersion is discharged from the organic dosing and pre-treatment system to the reactor for transfer of the pollutant from the aqueous effluent and its concentration into the organic material.

2. Equilibration and Extraction The fully developed dispersion prepared in the organic dosing and pre-treatment system must be preserved intact for a time sufficient for optimal transfer of pollutant from solution in the aqueous effluent into the organic material.

The ratio of diameter and linear flow length of the reactor network is such that the Reynolds number of the network dictates that turbulent flow is maintained and there is appropriate residence time for the dispersion to equilibrate for all effluent flow velocities at a range of volume flow rates.

At the reduced working organic volume fraction concentration of the fully dilute dispersion a droplet regime of free settlement velocity (cf hindered settlement velocity) is created, which is significantly more fragile than the dispersion concentrate in the organic dosing and pre-treatment system.

The dispersion is much more stable when only a fraction of the aqueous solution is mixed with the organic phase. When this dispersion concentrate is mixed with the remainder of the aqueous solution the tendency of the two immiscible layers to separate is much greater.

Special orientational, structural and dynamic measures are incorporated throughout the reactor system to suppress all tendencies towards floatation, segregation, agglomeration or coalescence of the developed dispersion, as a result of gravitational dynamic activity within the reactor.

2(a) Residence Time. Dimensions and Configuration The outer structure of the reactor is comprised of fabricated pipework. The particular apparatus is configured to operate over a range of effluent volume flow rates. The length of the reactor is such that at the highest volume flow rate, with the least favourable (permitted) effluent-organic combination and under the least favourable ambient scenario, the residence time of the effluent organic dispersion in the reactor is adequate for the fullest degree of transfer and concentration of pollutant from solution in the aqueous effluent into the organic material. Commercial and practical size constraints upon the complete apparatus unit do not permit the dynamically ideal reactor configuration as a vertical continuously straight run length of pipe. Practical unit packaging and the reactor pipework require its configuration into a number of vertical run straight sections interconnected in series by 1800 return bends jointly making up the required minimum linear reaction length.

2(b) Dispersion Integrity - General Considerations The onset of floatation, segregation, agglomeration or coalescence of the fully developed dispersion may be provoked by any one of several mechanisms within the reactor. Provocation at even small highly localised sites may seed a divergent growth of precipitation, expanding rapidly throughout the bulk of the dispersion, destroying its stability and capacity for reactive transfer of effluent pollutant.

The problems of separated dispersion are slightly different in the straight sections and curved sections of the reactor.

Flow in Straight Sections All organic effluent dispersions will tend to separate by gravity. The opportunity for gravitational separation will be minimised by vertical orientation of the reactor straight pipe runs. Also, the reactor can be generally configured to ensure a high minimum linear flow velocity such that a high fully turbulent Reynolds number prevails at all locations for all effluent volume flow rates. Micro eddies in the turbulent flow regimes will be created, which eliminate the tendency towards gravitational separation.

The general macro flow velocities throughout the reactor will be many times greater than the highest natural floatation velocities of any of the candidate organic effluent dispersion combinations. However, despite the general prevalence throughout the reactor of a highly turbulent regime at the macro level, a laminar regime boundary sub-layer will inevitably be present adjacent to the stationary wetted surfaces of the reactor structure. This will present opportunity for the onset localised of separation a potential for the seeding of progressive catastrophic collapse of the effluent organic dispersion.

To help prevent this occurring, a series of perforated baffles are located within the reactor straight pipework transversely to the direction of flow. The configuration and disposition of the baffles is such that the flow velocity adjacent to the pipe walls is accelerated into the turbulent regime, destroying the low velocity regime in the laminar boundary sub-layer and thereby also destroying opportunity for localised separation, and subsequent collapse of the dispersion.

The baffles also provide a secondary macro fixing function in the straight reactor pipe sections.

Changes in Flow Direction The configuration of the reactor structure will require the incorporation of 1800 degrees return and other bends, enforcing a change in direction with linear flow, thereby imparting an angular velocity component to the effluent organic dispersion. The imparted angular velocity induces a centrifugal force in the flowing dispersion, proportional to the algebraic product of the bend mean radius and the square of the angular velocity. This centrifugal force strongly promotes separation and the collapse of the dispersion at a very significant rate.

Centrifugal separation at changes of flow direction is fully suppressed by inducing precession of the flowing dispersion throughout the change in direction. The precession is induced by helically angled vanes placed in the flow stream, attached to the reactor structure immediately upstream of changes in direction, and at intervals throughout the change in direction. The helical vanes induce a secondary angular velocity component in the flowing dispersion, normal to and superimposed upon that induced by the change in direction. This causes the flowing dispersion to precess violently, creating an internal viscous shear in the flowing dispersion, which suppresses any natural tendency towards floatation, segregation, agglomeration or coalescence, promoted by the primary centrifugal field in the change of directional flow.

The primary angular velocity component at a bend radius is a direct function of mean linear flow velocity and bend radius. The secondary angular velocity component is a direct function of mean linear flow velocity and vane helix angle. Precession is maximised by the configuration and helix angle of the vanes with respect to the bend radius, such that the rates of angular rotation on primary and secondary axes are equal. The ratio of primary and secondary angular rotation rates thus remains constant and maximum for all effluent flow rates. Any residual rotational component prevailing within the flow is suppressed by flow straightening vanes at the exit from a bend, thus ensuring no centrifugal fields are induced in the baffled reactor pipe sections.

3. Coalescence and Demixing The rate of the dispersion is reduced to the point where the difference in specific gravity allows partial demixing, alternatively the flow can also be passed into a cyclone separator. Other embodiments of the invention use parallel plate coalescers, hydrophobic sieves or air lift separators to effect separation.

A final polishing phase involving passage through a fine coalescing sieve of hydrophobic epitropic fibre, or extraction through a packed column of wood flour, sawdust, straw, all of which may have been rendered hydrophobic, may be required.

Fig. 1 shows a schematic view of one embodiment of apparatus for performing the invention. The core process without any pretreatment, postreatment or ancillary structures is shown in Fig. 2. The apparatus is designed to treat process water from the cleaning and anti fouling of ships previously treated with, undergoing treatment with or being retreated with antifouling paint containing TBT; and other effluents, and infiltration water arising from the dock which contain TBT owing to prior contamination of the dock with TBT. This apparatus was designed to treat up to ten tonnes of effluent per hour.

The drydock where the apparatus was installed was of a size sufficient to accommodate vessels of up to 115m length, 16.6m beam and 7m draught. Process and infiltration water were collected via two longitudinal gullies running along the dock's length on either side, which feed into a transverse trough running the entire width of the dock and of approximately lm depth. The transverse trough was fitted with vertical baffles at intervals, which act to facilitate the settling of heavier particulates (eg spent grit from blasting, rust and metal particles and swarf). Water overflowed from the surface of the trough at one side of the dock into a subterranean cavern or sump, approximately 10m square and 1.5m depth, where further settlement of particulates may occur. Water from this sump was discharged via a pump which draws water from such a depth in the trough as to avoid discharging any surface oil, and prior to the installation of the apparatus was discharged directly back into the harbour. This discharge, on installation of the apparatus, was diverted to pass through the apparatus for the removal of TBT.

In operation, water is pumped from the dock sump via a commercial filtration device (not shown) which removed suspended particulates greater than 20 microns diameter. The particulates collected in the filter are periodically and automatically backwashed (without interfering with the correct operation of the filter) and passed back into the sump. This continuous backpassing of the particulates to the sump, combined with the continuous passage of water through the sump, over time removes TBT from the particulates and reduces their TBT content, facilitating their eventual disposal more economically by reducing their content of toxic material.

The apparatus is provided with a prefill and start flush facility 1 for priming the system with clean water before intake of effluent. The priming tank 1 can be charged from a reflux tank collecting treated water from the outflow pipe of the system.

The effluent is then pumped through the primary flow control 2 into the fractionation and remixing section 3 directly to the reaction and transfer unit (5, Fig. 2), but a small proportion (5% - 7%) is drawn down via a disperser loop 3d to a dosing and concentrate dispersion unit 4. This part of the apparatus comprises two modified Silverson Screen devices.

In the disperser units 4, the sample drawn through pipe 3d is mixed with a quantity of solvent such that the solvent is dispersed into fine droplets of very consistent size of around 5-6mm. This solvent-water emulsion is passed through the second screen and then reintroduced into the main flow through 3r and a venturi joint and the remixture passes to the reactor and transfer tubes 5. Factors governing the nature and quantities of solvent required are discussed below.

The purpose of the tubes 5 is to maintain the integrity of the emulsion until transfer of TBT from the water to solvent phase has taken place. Coalescence of the emulsion begins as the liquid leaves the reactor tubes and enters the separation and extraction stage 6.

At present, this stage consists of an oil-water separator which is designed to remove oil from water down 6 the limits required by international regulations governing the discharge of oil to water, that is to a concentration of a few milligrams of oil per litre of water. In the prototype apparatus, the solvent is collected for disposal through 6d, and the water is discharged to the environment. However, for maximal economy in the use of solvent, a further stage of oil separation is desirable, because the solvent can contain such a quantity of TBT that the effluent, though meeting oil emission standards, occasionally approaches or exceeds the proposed standard for TBT (200 ng/l).

Fig. 3 shows a centrifugal separator for use in an outfall scrub. The separator is fed with contaminated fluid from the separation and extraction phase through inlet pipe 50 which enters an axial aperture in the separator body 30. The main body of the separator has a back plate 30b, sidewalls 30s and a front plate 30f which are welded together as shown in the drawings.

The body 30 of the separator is in turn welded to a shaft 35 which rotates on bearings with the body 30 of the separator 30. Back plate 30b and front plate 30f are in the form of discs with axial apertures. A series of baffle discs 36, 37 and 38 are welded to guide pins 40 extending from the back plate 30b.

Baffle discs 37 are welded at their outer radial ends to side wall 30s and baffle plate 38 is welded at its inner radial end to closer plate 45. The arrangement of baffle plates and the closer plate 45 forms a serpentine path indicated by arrows V, W, X, Y and Z along which fluid flowing from inlet pipe 50 can travel through the separator. Additionally, closer plate 45 defines a boundary between two annular channels 52, 51.

Channel 52 draws fluid from the portion of the separator at the inner radial ends of the baffles, and channel 51 draws fluid from the outer radial edges of the baffles by virtue of its connection to baffle 38 at its inner radial end. Channels 52 and 51 (which rotate with the body of the separator 30) empty into stationary collectors 55 and 56 respectively.

In operation of the separator, contaminated fluid enters through inlet pipe 50, and is subjected to centrifugal force which enhances separation between the baffle plates of the aqueous phase to the outer radial edge towards sidewall 30s; the oily phase is forced towards the radial inner edge against the pipe 50. Any oil droplets migrating from V to W will again be subjected to high G forces and will be forced radially inwardly towards the pipe 50. Therefore with each turn V, W, X, Y and Z through the baffle plates, the fluid passing through the separator 30 will be stripped of oil droplets which will naturally migrate towards the inner radial ends of the baffles and collect against the pipe 50. The cleaned aqueous phase will eventually pass over the last baffle 38 at Z and by virtue of the continued flow through the inlet pipe 50 will be forced through channel 51 and into collector 55 where it can be drawn off and assessed for further treatment or discarded as appropriate. The oily phase containing the contaminant will be forced through channel 52 and into collector 56 where it can be drawn off and collected for disposal in a concentrated form. A similar device is used for scrubbing the solvent drawn through 6d and the aqueous phase is returned to the reflux tank 8.

The remaining features shown in Fig. 1 comprise various arrangements for connecting the apparatus to the existing dock drainage system, for priming the apparatus on startup and for collecting and if necessary recycling the solvent.

The solvent represents a cost of operation, and contaminated solvent must somehow be disposed of, representing a further cost. These factors dictate the need for economy in the quantity of solvent used. The ultimate factor governing the quantity of solvent required is the partition coefficient of TBT from water into the solvent. The higher this is, the less solvent is required for a given level of performance. Some solvents with high partition coefficients are ruled out on grounds of cost, or other difficulties, eg flammability or other handling problems. The ideal solvent would have a suitable combination of low cost, ease of disposal or re-use, and the right rheological characteristics. The ideal solvent might therefore differ in different locations or applications, depending upon local conditions. (The disperser units are designed to be able to cope with a variety of different solvents for this reason).

One solution is to use waste oil as a solvent. Many dockyard facilities have a ready supply of waste oils which in normal circumstances is disposed of, at a cost, by incineration, re-refining or other means.

Therefore to use a suitable waste oil would offer the advantages of small cost, since the oil is available at no cost and its disposal costs are already accounted for. Another solution is to use diesel oil, which is readily available at modest cost. Further, since the diesel oil is recovered, separated from the water, it can be used subsequently for its original purpose as fuel. At engine combustion temperatures, TBT breaks down to its hydrocarbon moieties which combust, and inorganic tin, which is relatively harmless. The tin is likely to end up in an inorganic insoluble form in the engine lubricating oil, or in the engine sludge.

In practice, one tonne of diesel oil would contain no more than a few grams of TBT; there are therefore few, if any, environmental issues raised by this means of disposal.

The apparatus can operate at any volume from 1.5 to 10 tonnes per hour. There are no serious constraints on smaller or larger units being constructed. This range corresponds to the requirements of the vast majority of dockyards. Using diesel oil as solvent, at a rate of 2% of the bulk flow (ie 2 tonnes diesel oil to treat 100 tonnes of effluent), one core unit can produce a reduction in TBT concentration of up to 480 times.

Factors governing the performance of the apparatus include the quality of the influent (in particular, the presence of particulates and TBT-contaminated oil), and the efficiency of oil-water separation, dictating the requirement for post-treatment (scrubbing or final polishing of the effluent to remove the last traces of solvent carry-over). In practice, a single core unit can be relied upon to reduce TBT concentrations by about 100-fold. Therefore, two units connected in series can produce a 10,000-fold reduction.

In practice, dockyard effluents contain the highest concentrations of TBT for relatively brief times, during the actual operations of ship washing and/or painting, when concentrations up to 500 micrograms/litre have been recorded. A 10,000-fold reduction of this yields an effluent concentration of 5 ng/l, compared to the proposed emission standard of 200 ng/l. It is therefore concluded that a 2-unit installation would be adequate to cater for all but the most extreme circumstances. No dockyard yet studied has required a capacity of more than 25 tonnes/h, and then only for short periods. The majority of docks have sufficient retention capacity in sumps or tanks to treat their effluent over a period of time at the lower 10 tonne/h rate. Note, however, that the use of multiunit installations in the largest docks may have some advantages. For example, a 2, 3 or 4-unit installation could easily be arranged so that the units could be connected either in series or in parallel; this would offer the choice of treating extremely concentrated effluents at a modest rate to a very high standard, or the more likely situation of large quantities of more dilute effluent to a more modest standard, while still remaining within the emission limit. In this connection, note that the use of multiple units does not require the use of any additional solvent. This is importance because solvent usage, and eventual disposal, represents the major running cost of the process.

This arises because the factor governing the amount of TBT which the solvent can absorb is not, in practice, the solubility of TBT in the solvent, but the concentration difference between the TBT in water and the TBT in solvent, taking into account the partition coefficient of TBT between the two phases. Thus, if multiple units are arranged in such a way that the cleanest solvent is passed through the last unit in the series, collected and recycled and passed, countercurrent fashion, to the previous one, and so on, the same volume of solvent can serve several units while still achieving the required reduction in TBT content of the water.

It is implicit in the above that a further application of the invention is in the treatment of solid wastes contaminated by TBT, such as particulate matter arising from dockyard operations and contaminated silt. The process of grit-blasting of ships' hulls generates considerable particulate waste material, consisting of spent grit, paint flakes, rust and other material removed from the hull. Where ships have been treated with TBT, the whole mass of the particulate matter becomes contaminated with TBT. Environmental regulations in many regions dictate that if the quantity of TBT in such solid material exceeds a certain limit, the material must be disposed of as "special" waste, which is extremely expensive. Similar considerations apply to the disposal of dredged silt from dock mouths and quaysides traditionally associated with shipbuilding and repair, which are historically contaminated with TBT.

Such solid material could at low cost be continuously irrigated, preferably with aerated water and possibly with some pH adjustment, to facilitate the passage of adsorbed TBT from the particulate matter into the water. After settling, the liquid can be treated in accordance with the invention after a period of time resulting in the TBT content of the solid material falling below the critical level, allowing its disposal by more economical methods. In principle, a dredging vessel employed to dredge harbour channels and quaysides, where TBT contamination is a problem, could be equipped with a shipboard apparatus for continuous treatment of the dredge spoil prior to dumping, under regulation, at sea; or to other means of disposal presently precluded by virtue of the material's high TBT content.

Modifications and improvements can be incorporated without departing from the scope of the invention.

Claims (16)

1. A method of treating a contaminated fluid to remove a pollutant therefrom, the method comprising mixing said fluid with a second fluid immiscible with the first fluid, wherein the pollutant is more soluble in the second fluid than the first fluid.

2. A method as in Claim 1, wherein said contaminated fluid is aqueous.

3. A method as claimed in Claims 1 and 2, wherein the contaminated fluid is waste water from pressure hosing.

4. A method as claimed in Claims 1, 2 and 3 wherein the second fluid is emulsified in the contaminated solution.

5. A method as claimed in any of the preceding Claims wherein the second fluid is first emulsified in a carrier fluid which is subsequently introduced into the contaminated fluid.

6. A method as claimed in Claim 5, wherein the carrier fluid is a sample of the contaminated fluid.

7. A method as claimed in Claims 5 or 6, wherein the carrier fluid sample is approximately 5 - 7% of the bulk flow of contaminated fluid.

8. A method as claimed in any of Claims 4 to 7, wherein the emulsion comprises particle sizes of 5-6ym.

9. A method as claimed in any of Claims 4 to 8, wherein the emulsion is maintained for a time sufficient for transfer of the pollutant from the contaminated fluid to the second fluid.

10. An method as claimed in any preceding Claim, wherein the second fluid is chosen from the list of engine oil, slops oil, "waste" oil, diesel oil, silicone oil, iso-propanol, pentanol, toluene or singly or multiply alkyl substituted toluene, where the alkyl substituents are chosen from methyl, ethyl, propyl or butyl.

11. A method as claimed in any preceding Claim, wherein the contaminated fluid is mixed with the second fluid by passing through a rotor and/or screen to form a dispersion of uniform particle size.

12. A method as claimed in any preceding Claim, wherein the mixture is maintained while being passed along a conduit defining a straight path.

13. A method as claimed in any one of Claims 1 to 11, wherein the mixture is maintained while being passed along a conduit following a serpentine path.

14. A method as claimed in Claim 13, wherein the conduit defining the serpentine path has vanes at the bends to prevent demixing of the fluids while passing the bends.

15. A method as claimed in any preceding Claim, wherein after mixing the mixture is subsequently separated in a centrifuge.

16. A method as claimed in any preceding Claim, wherein the concentration of pollutant in the contaminated fluid is higher than in the second fluid.