FIELD OF THE INVENTION

The present invention relates to an apparatus and process for the purification of a liquid by the removal of one or more impurities therefrom by adsorption onto a particulate adsorbent. The invention is particularly, though not exclusively, applicable to the removal of residual dye from dying effluent streams.

BACKGROUND OF THE INVENTION

In conventional dying technology, reactive dyes are widely used in the treatment of cotton, wool and regenerated cellulose. In order to encourage the dye to leave the dye solution and enter the fibres, it is normal to include a substantial concentration of a salt (such as sodium sulphate or sodium chloride) in the dying solution. The presence of the salt effectively reduces the solubility of the dye in the solution and promotes adsorption of the dye onto the cloth. Typically, 80% of the dye in the dying solution is transferred onto the cloth. Of this transferred dye, typically 70% is reacted with the cloth, whilst the remaining 30% of dye is unreacted.

Currently, the used dying solution comprising residual dye and salt is dumped. Such dumping is becoming environmentally unacceptable, and a process for removing

residual dye from salt-containing dying solution is disclosed in our patent application PCT/GB93/00135. This discloses contacting the used dye/salt solution with active carbon in a holding vessel such that the dye is adsorbed onto the carbon and then circulating the mixture of liquid and adsorbent carbon particles through a cross-flow membrane filter so as to remove a salt solution free of carbon particles and dye which may be recycled or disposed of.

In a further conventional processing stage, the unreacted dye remaining on the cloth is removed by a series of washing steps to ensure the removal of all dye nor firmly bound to the cloth. The disposal of wash water containing concentrations of dye also poses a substantial problem. Patent specification PCT/GB91/00947 Crosfield, discloses the use of a solid particulate adsorbent to remove dye from the wash water. The preferred adsorbent material is a hydrotalcite material.

Furthermore, there are many other industrial applications where solid particulate adsorbents may be used to purify liquids. By "purify" we mean the removal of at least one species from the liquid, though other species may remain in the liquid.

Generally speaking, it is desirable that the particulate solid adsorbent shall have a large surface area in order to promote adsorption onto the particulate solid. This generally leads to particles of a small

particle size being preferred, since they have a high surface to vo ^" -ie ratio. However., the use of small particles can _t jse problems, since the particles may not be easily immobilised in a filter bed, and may not easily be removed from the liquid.

Cross-flow filtration has proved to have a number of advantages in separating particulate solid adsorbents from liquids. Patent specification WO 91/04"? ! discloses removing nitrates from water by using an _^ on-exchange resin and a cross-flow filter. It is found, however, that there is a tendency for cross-flow filters to lose their filtration efficiency due to the build up of a layer of particles on the filter membrane, particularly when filtering liquids of relatively high content of solid particles.

It is an object of the present invention to address these problems.

SUMMARY OF THE INVENTION

The present invention provides an apparatus for purification of a liquid stream by adsorption onto a solid particulate adsorbent, the apparatus comprising:

(a) a vessel for containing liquid to be purified in contact with said particulate adsorbent such that impurity therein becomes adsorbed onto the particulate adsorbent;

(b) cross-flow filtration means connected to the vessel such that a recirculating liquid stream having particulate adsorbent suspended therein may be passed from the vessel to an inlet of the filtration means and returned from an outlet of the filtration means to the vessel, the filtration means comprising membrane means for permitting removal of a stream of purified filtrate free of particulate adsorbent from the recirculating liquid stream; and

(c) pump means for circulating said recirculating liquid stream; the arrangement being such that the contact time of the liquid with the particulate adsorbent is sufficient to ensure substantially complete adsorption of impurity onto the adsorbent; and such that the flow of liquid through the cross-flow filtration means is turbulent. The present invention further provides a method of purifying a liquid stream by adsorption of impurity therein onto a solid particulate adsorbent, the method comprisir.rr the steps:

(a) placing a liquid to be purified in a vessel in contact with a particulate adsorbent for a period sufficient to ensure substantially complete adsorption of impurity from the liquid onto the adsorbent;

(b) passing the liquid, with the particulate adsorbent suspended therein, through a cross--low filter including a membrane at a rate sufficient to produce turbulent flow over the membrane;

(c) removing a stream of purified filtrate, free of particulate adsorbent, from the liquid through the membrane; and

(d) returning the remaining liquid and the particulate adsorbent suspended therein, to said vessel.

It has been found when using typical dye bath or wash bath volumes that w the rate of' recirculation of the liquid is controlled to give adequate residence times of the liquid (and contact time between the liquid and the particulate adsorbent) then the filtration efficiency in the cross-flow filtration means is poor. This arises due to the build up of a surface layer of particulate solids on the membrane which reduces flux through the membrane.

On the other hand, it has been found that filtration efficiency can be improved by increasing the velocity of the liquid through the cross-flow filtration means. In fact, the flow over the membrane should be turbulent to minimise build up of solids thereon, as may occur with laminar flow, and to promote uniform distribution of the solids throughout the liquid within the filtration means. In order to achieve turbulence, the Reynolds number R ₀

generally requires to be at least 2,000. However, higher Reynolds numbers are particularly preferred, for example at least 5,000 and especially at least 8,000. It is found that the flux through the filter reaches a maximum value as the Reynolds number increases, and it is desirable to attain this maximum flux. The R ₀ at which this is achieved may be determined experimentally and may depend on the nature and particle size of the particulate adsorbent, and its concentration; and also the nature of the filtration means, particularly its surface characteristics. Since the Reynolds number depends on the flow velocity, in order to achieve good filtration the liquid must be recirculated at such a high rate that, in conventional arrangements, the residence time of the liquid is reduced, so that the adsorption of impurity onto the adsorbent is impaired.

Thus, in conventional arrangements, good adsorption may result in poor filtration; whilst good filtration may result in poor adsorption.

In one embodiment of the present invention, good adsorption and good filtration are achieved by providing a by-pass duct connected between the inlet and the outlet of the filtration means, such that liquid, comprising liquid from the vessel and liquid from the by-pass duct, may be passed through the cross-flow filtration means at a velocity high enough to ensure turbulent flow. The relative rates of flow through the by-pass duct and

through the vessel may be controlled by choosing relative duct sizes which partition the flows appropriately. Alternatively, restrictions or valves may be employed. However, in a preferred embodiment a further pump means is provided in a by-pass circuit including the duct and the rates of flow are controlled by varying the rate of working of the two pump means.

In another embodiment, the residence time within the vessel is controlled so as to sharpen the residence time distribution by eliminating abnormally short or abnormally long residence times. This may be achieved by providing baffle means within the vessel to constrain flow through the vessel to a labyrinthine path, or otherwise forcing liquid to follow a tortuous path.

Ger-rally, the membrane filtration means comprises a bundle or bundles of hollow permeable tubes having porous walls (e.g. porous expanded polytetrafluoroethylene) . The liquid having dispersed solid particles therein is generally passed through the lumen of the tubes and the filtrate collected from the outside.

The invention is applicable to the removal of dye from seIt-containing dying solutions or from dye-co. _aining wash water using appropriate solid particulate adsorbents. Equally, the invention has wide application where a liquid is to be purified by means of a solid particulate adsorbent.

BRIEF DESCRIPTION OF THE DRAWINGS

Embodiments of the present invention will now be described by way of example only in conjunction with the drawings wherein:

Figure 1 shows schematically a first embodiment of the invention employing a by-pass;

Figure 2 shows schematically a second embodiment of the invention employing a baffled vessel; and

Figure 3 is an experimental graph showing the variation of filtrate flux with respect to flow velocity (and Reynolds Number) ; and

Figure 4 shows schematically apparatus utilised in determining pore sizes by bubble point methods.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

Figure 1 shows a tank T for receiving dye-containing liquor A from a dye bath. A solid particulate adsorbent is dosed into the tank. The outlet of the tank is connected via line 10 to the inlet of a cross-flow filter F comprising a bundle of hollow tubes 18 of internal diameter 4mm formed of a porous material, such as polytetrafluoroethylene (PTFE) (available from W.L. Gore & Associates (UK) Ltd.), polypropylene or polysulfone. The pore sizes are usually in the range 0.1 to 2 microns as measured by isopropanol bubble point measurements similar to those described in International Standard ISO 4003 and as described in some detail in attached Appendix I. The outlet of the cross-flow filter F is connected back to the

tank T via line 12.

A by-pass line 14 is provide ^~; for circulating liquid at high velocity through the cross-flow filter whilst by-passing the tank T. A pump Q is provided for feeding liquid from the tank T to the cross-flow filter, whilst a pump P is provided principally for circulating liquid through the filter F and by-pass line 14. Valves Vr , Vj , "p and VQ are provided for controlling the flows. Pressure sensors P ₀ , Pj and Pp measure pressures at various points in the flow streams.

The filtrate free of adsorbent particles passes from the cross-flow filter along line 16, and may be directed to drain or recycled for further use in the process.

The apparatus may be operated as follows. After a dying or washing operation, some or all the contents A of the dye bath are dumped into the tank T and a suitable quantity of particulate solid adsorbent added. After a predetermined time, valve V-p is opened and pump Q started, so as to feed a slurry of liquid and solid particles towards the cross-flow filter. Pump P is also initiated and valves Vj and V ₀ are opened. Suitable time will have elapsed to allow substantial adsorption of dye onto the solid adsorbent. The slurry then passes through the lumen of tube-3 18 in the cross-flow filter F. The rate of working of recycling pump P relative to pump Q is adjusted such that a substantial proportion of the liquid slurry is recycled via by-pass 14 through the filter F such as to establish turbulent or super-turbulent conditions within

the tubes 18 and to maximise filtrate flux through the tube membranes. A clarified filtrate stream is removed along line 16 when valve Vp is opened.

The concentrated liquid slurry is returned via line 12 into the tank T for re-use.

The amount of liquid recirculated through the by-pass 14 is adjusted relative to the total amount of liquid leaving the tank T, such that substantially complete adsorption of dye onto the adsorbent is achieved, whilst at the same time turbulent flow within the filter tubes is also maintained leading to good flux of filtrate.

As liquid slurry is withdrawn from tank T, it may be replaced by fresh dye liquor A. In this way, the process may be operated continually. Periodically, spent solid particles may be dumped from tank T. Alternatively, a bleed of slurry may be removed from line 12, and continuous_y replaced by new solid particles introduced into tank T.

Figure 2 shows a second embodiment in which analogous parts are indicated by the same reference numerals. This embodiment is similar to the first embodiment, except that the by-pass line is omitted and the tank T is provided with a series of upwardly projecting baffles 20 and downwardly projecting baffles 22. The labyrinthine progress of the dye liquor A through the tank T as defined by baffles 20, 22 provides a more sharply defined residence time of liquor within the tank T in contact with

added particulate adsorbent. The residence time is also lengthened, so as to provide adequate contact time between the dye liquor and the adsorbent particles and to ensure substantially complete adsorption of dye onto the particles. The substantially purified liquor is then pumped by pump P via line 10 through the cross-flow filter F and a filtrate stream 16 removed as before. The rate of flow is adjusted such as to provide turbulent conditions and hence high flux within the filter F.

Example 1 (velocity and residence time)

The following demonstrates the correlation between residence time in the tank T and cross-flow velocity through the filter F, referring to Fig - 1. T tank T has a volume of 60 L and the filter F comprises 5hty one porous polytetrafluoroethylene tubes of internal diameter 4mm (total cross-sectional flow area = 0.00108 m ³ ) and nominal pore size of 0.45 microns. In all cases the flow of fresh liquor A is equal to the filtrate passing through line 16.

The cross-flow rate is the rate of flow through the filter F. The tank flow rate is the flow rate through tank T; and the bypass rate is the flow rate through the bypass 14 which recirculates liquor through the filter.

The table below shows different calculated tank residence times and cross-flow velocities, depending on

the bypass rate. Case 2 has a suitably long tank residence time for adsorption to occur, and a fast enough cross-flow velocity through the filter to maximise the filtrate flow out of the filter (see Figure 3) . In cases 1 and 3 either the tank residence time is too short for effective adsorption, or the cross-flow velocity in the filter is too low for good filtrate flow.

The chosen residence time will of course depend on the nature of the adsorbent, the nature of the impurity to be removed and its concentration, and the degree of purification to be achieved.

TABLE

Case Cross-flow Tank flow Bypass Tank Cross-flow

Residence Vel m ³ /h m ³ /h m ³ /h s m/s

1 20 20 0 10.8 5.14 2 20 4 16 55 5.14 3 4 4 0 55 1

Example 2 (filtrate flow and cross-flow velocity)

An apparatus of the type shown in Figure 1 and described in Example 1 was employed to investigate the effect of cross-flow velocity through the filter tubes 18

on the rate of filtrate leaving the filter along line 16; except that flow along line 14 was zero, pump P was omitted, and a tank of volume 40L was employed together with a filter comprising nineteen porous polytetrafluoroethylene tubes.

First 40L of water at 60°C was circulated from the tank through the tubes and returned to the tank with valve Vp closed, to condition the system. Then 1.1kg of Celite Z850NW activated carbon was added to the water together with 0.4 kg of common salt (sodium chloride). The filtrate valve Vp was opened to give an average permeate driving pressure ( (P ₀ - Pj)/2 - Pp) of 0.5 bar.

The filtrate flow was recorded every 5 mins over a 30 minute period.

The procedure was repeated at various cross-flow velocities. Figure 3 shows the filtrate flux as a function of filter cross-flow velocity. Reynolds numbers R ₀ calculated on the basis of pure water in a tube of 4mm internal diameter are also given.

APPENDIX I

Determination of Bubble Test Pore Size

This method of determining the pore size of a permeable material involves the steps of:

- impregnation of a test piece with a test liquid;

- immersion of the test piece in the test liquid and introduction of a gas into the test piece at gradually increasing pressure;

- determination of the pressure at which bubbles are emitted from the surface of the test piece; and

- evaluation of the equivalent bubble test pore size by means of a mathematical formula.

The bubble test pore size is the maximum equivalent capillary diameter in the test piece which is calculated from the measured minimum pressure required to force the first bubble of gas through the test piece impregnated with a liquid.

The first bubble of gas will form at the pore having the greatest throat, the throat being the narrowest section of this pore.

For calculation purposes, it is assumed that this bubble forms at the end of a capillary tube of circular cross-section which is initially filled with the same liquid of known surface tension.

For a circular capillary, the diameter is related to the bubble pressure by the equation:

where d is the capillary diameter corresponding to the bubble test pore size in metres; t is the surface tension of the test liquid, in newtons per metre;

Λ _> P is the differential pressure in pascals, across the test piece under static conditions, i.e.

Δp = p _g - PI (2 )

Pg being the gas pressure, in pascals;

Pl being the pressure in the liquid at the level of bubble formation, in pascals;

where l is the density of the test liquid, in kilograms per cubic metre; h is the height of the surface of the test liquid, in metres, above the highest throat in the test piece.

The apparatus utilised to determine the bubble test pore size is illustrated somewhat schematically in Figure 4 of the drawings. The apparatus includes a source of dry and filtered air 30, a pressure regulator 32 affording control of the gas pressure, a device 34 to measure the effective gas pressure and an assembly 36 for observing the bubble appearance at the surface of the test piece 38 and for ensuring that the test piece is completely saturated with the liquid and immersed un ^r ' --r a c ^o nstant depth of the liquid throughout the test. .-_s sh. in Figure 4, the assembly 36 in this example includes tank 40 filled with liquid 42 and containing an arrangement in communication with the air supply and having a wall 44 formed of the material to be tested.

A suitable test liquid 42 for polytetrafluoroethylene is isopropanol (density 0.79g/cm ³ , surface tension at 20°C 0.0215N/m). The test is carried out at room temperature.

The test piece 38 should be clean, dry and free from extraneous material and any trace of grease or similar substances likely to hinder the perfect and uniform wetting action of the isopropanol.

The test piece 38 is impregnated completely with the test liquid 42. The test piece 38 is then inserted into the bubble test apparatus and fixed in place. The test piece 38 is then immersed under the smallest depth of isopropanol consistent with the convenient observation of the appearance of the bubbles. This depth h and the temperature of the liquid are measured.

From an effective gas pressure of zero, the pressure is regularly increased, while the surface of the test piece is under constant observation. The first bubble pressure is noted when a string of bubbles occurs from one distinc- point or perhaps several distinct points at the same time. The bubble test pore size is then calculated using the equation (1) above.