EP0433387A1

EP0433387A1 - Filtration apparatus and method

Info

Publication number: EP0433387A1
Application number: EP19890910890
Authority: EP
Inventors: Simon Jon Judd; George Stefan Solt
Original assignee: BTG International Ltd; National Research Development Corp UK
Current assignee: BTG International Ltd
Priority date: 1988-09-12
Filing date: 1989-09-12
Publication date: 1991-06-26
Also published as: WO1990002595A1; GB8920648D0; GB2224958B; JPH04500627A; GB2224958A

Abstract

Procédé permettant de filtrer en profondeur un fluide pour le débarrasser des particules qui s'y trouvent qui consiste à disposer dans un filtre à lit profond une certaine quantité d'éléments filtrants, dont certains au moins sont des conducteurs ioniques, à placer ledit filtre à lit profond dans un champ électrique qui agit individuellement sur lesdits éléments filtrants, à laisser ou faire couler ledit fluide chargé de particules dans le filtre à lit profond, et à influencer ensuite le mouvement des particules dans le filtre à lit profond au moyen du champ électrique pour en accroître la filtration.Method for depth filtering of a fluid to rid it of particles therein which consists in placing in a deep bed filter a certain quantity of filter elements, at least some of which are ionic conductors, in placing said filter deep bed in an electric field which acts individually on said filter elements, to let or run said particle-laden fluid in the deep bed filter, and then to influence the movement of particles in the deep bed filter by means of the electric field to increase filtration.

Description

FILTRATION

This invention relates to filtration and more particularly to the enhancing of filtration by electrical action.

The design of a filter to remove particles from a fluid must be a compromise between the obstruction the filter produces in the desired flow of the fluid and the efficiency of removal of particles from the fluid. It is known that in certain circumstances the action of a filter can be enhanced by the application of an electric field. An enhanced filter might for example offer less obstruction for a given removal of particles than the unenhanced version. Clearly the choice of parameter enhanced in any given filter, subject to design constraints, can be a matter of choice.

The way in which electrical action enhances filter action is believed to be understood for certain distinct classes of filter and electrical action.

One very important filter is the deep-bed filter in which the spaces are large compared with the size of the particles to be filtered and the filter medium is relatively loosely packed. The pore volume can be 20 to 30% or even more. A significant use of deep-bed filters is depth filtration for drinking water supplies. In this a relatively deep layer or bed of filter medium, such as sand, is used to adsorb particles on the surface of the medium with minimal pressure loss in the passage of the water through the relatively open filter. While such a filter is effective in removing many of the particles small particles tend to follow the hydraulic streamlines and pass through the filter medium.

It is an object of the invention to provide a method of electrically enhanced filtration of particles from a fluid and an electrically enhanced filter structure, to improve the removal of small particles. According to the invention there is provided a method of depth filtration of particle from a fluid carrying particles including providing in a deep filter a quantity of filter medium elements, at least some of the elements being ionically conductive, applying an electric field to said deep filter to cooperate with individual said elements,

permitting or causing said fluid with particles to flow into the deep filter,

influencing the movement of particles in the deep filter by the electric field to enhance the filtration of the particles from the fluid.

In such methods of depth filtration the filter medium elements may have an electrical conductivity varying from one part to another part of the element. The one and another part may be regions of a fibre and there may be several said parts in each fibre. Conveniently the fibres are manufactured fibres having within them one or more conductive parts of discontinuous longitudinal regions. One preferred fibre is the proprietary fibre ACTILEX B701 (RTM) a polyacrylic material, as recently available from Courtaulds Ltd.

The filter medium elements may be permitted or caused to have a specific state of charge, not the same as that of the particles to be filtered. Specifically the elements may be chemically treated, e.g. with NaOH, to have a charge opposite to that of the particles. The elements may be permitted or caused to imbibe water.

The deep filter type is well-known in the art and is characterised by a filter medium of a bed or layer of significant depth but open structure to have a high porosity. In a deep filter embodying the invention the porosity may be such that in the absence of the electric field particles could move through without encountering the filter medium.

According to the invention there is also provided a method of depth filtration including providing in quantity a filter medium of elements having an electrical conductivity varying from one part to another of said elements, forming a quantity of said elements into a deep filter, applying an electric field to cooperate with individual said elements in the deep filter, permitting or causing fluid with particles for filtration therefrom to flow into the deep filter, influencing the movement of particles in the deep filter by the cooperation of sai d appl i ed el ectri c fi el d and sai d i ndi vi dual el ements to enhance the filtration of said particles from the fluid.

According to another aspect of the invention there is provided a deep filter including filter medium arranged in filter layer, means to apply an electric field to the layer and means to move through the layer a liquid with particles for filtration, the filter medium being ion conductive.

Conveniently the filter medium is composed of fibres loosely packed to provide a high porosity ranging to in excess of 96% and up to 96.75% pore volume. The ion conductivity of an individual fibre may not be uniform from part to part of the fibre. Non-uniform conductivity can be provided by the fibres not being all of one type.

When the filter medium is of loosely packed fibres the fibres may be mainly in two of the three conventional orthogonal directions. It may then be advantageous for the means to movethe liquid to direct the liquid generally in the third of said three directions.

The filter medium may be polyacrylic fibres chemically modified to be set with elcctropositive groups. Embodiments of the invention will now be described with reference to the accompanying drawings and tables in which

Figure 1 is a schematic diagram, partly in cross-section, of a filter structure embodying the invention,

Figures 2 and 4 are graphs useful in understanding the behaviour of the filter, and

Figure 4 is a graph useful in comparing the behaviour of a filter embodying the invention with other filters.

Table 1 is a table of parameter ranges for the filter,

Tables 2 and 3 are tables useful in understanding the behaviour of the filter, and

Table 4 is a table useful in comparing the behaviour of a filter embodying the invention with other filters.

Referring first to Figure 1 this shows in schematic form an experimental form of a filter embodying the invention.

A deep-bed filter indicated generally at 1 is formed in tube

10 of acrylic material (specifically PERSPEX, RTM). The filter medium 20 is of fibres and is described in more detail below.

The filter medium 20 is held in the tube 10 by two similar retainers formed of plastic meshes 11 (conveniently of NETLON, R.T.M.) attached to 0-rings 12 and supports 13. The tube 10 is closed by similar closures 14 at each end. Each closure has a liquid flow tube 15 extending through it. Each closure also supports a disc electrode 30 and an insulated electrical connection 31 for the electrode. Conveniently the electrodes are of platinised titanium and the connection is of an insulated copper wire extending through the electrode. An adjustable source of electricity 32, with voltmeter 33 and ammeter 34, is connected to the connections 31. In this embodiment the source is of continuous, i.e. d.c, current, In this way an electrical field of desired strength can be applied in the filter. In one arrangement the tube 10 was about 35 millimetres in diameter and a filter length of about 10 to 25 and typically 15 millimetres was used.

The fibres and filter medium will now be described in more detail. The fibres were supplied by Courtaulds, under the trade name

"Actilex B701", and had a mean diameter of 25.81 micrometres ±

1.46 micrometres as determined by optical microscopic measurement of the cross section. The fibre density was calculated from the decitex measurement in g/10⁶cm supplied by the manufacturer. The available pore volume was calculated from specific gravity measurements of the wet and dry material.

The fibre was pre-treated by stirring a 0.4% slurry in 10^-2M NaOH for 5 hours before use to rid it of residual hydrochloric acid from the manufacturing process, and also to convert the ion-exchanging surface to the hydroxide form. The fibre zeta(ζ)-potentia1 was calculated from streaming potential measurements of fibre plugs using apparatus similar to that of Joy et al (1965). Measurements were corrected for surface conduction by the method of Fairbrother and Mastin (1924). The axial conductivity of the fibre was measured by the alternating current impedance method on continuous filaments of the material.

A buffer system of 2x10^-4M trishydroxymethylmethylamine :

10^-4M HCl was used. The measured conductivity of this buffer was 8 to 9x10^-6Ω^-1cm^-1. The water was supplied from an Elga (RTM) R03 deionisation unit which produced water of conductivity 0.5x10^-6Ω^-1cm^-1.

The filter medium was formed in place in the tube 10. The filter pad was formed on the lower retainer by sedimentation of the pre-treated fibres from a 4% slurry in the buffer solution. Fibre settlement was assisted by mechanical agitation using a modified Griffin and George mechanical shaker, which also helped remove trapped air bubbles, such that the fibres usually settled within 30 minutes. The supernatant was drained and the upper retainer placed in the cell to compress the fibre pad to the required filter length. Prior to filtration the fibre pad was equilibrated by passage of 250ml buffer solution through the cell. The electrical supply was connected and the field set at the required strength and the test material suspension flowed upwards through the cell (as shown in the drawing) at the desired flow rate from a constant head device.

The performance of the filter was measured using a test material of an aqueous suspension of homodispersed polystyrene latex particles.

The polystyrene latex particles were prepared by emulsion polymerisation according to the method of Kotera et al (1970). The initiator used was potassium persulphate and, in the case of the sub-micron particles, sodium lauryl sulphate was used as a surfactant. After fractionation of the suspensions, by sedimentation or centrifugation, particles of mean diameter 0.432, 1.13 and 2.09 micrometres respectively were obtained. These were then purified using nuclear grade ion-exchange resins (Rohm and Haas) according to the method of Van den Hull and Vanderhoff (1972). Particle diameters were determined using a model "T" Coulter counter (Coulter, Luton) or, in the case of the smallest particles, electron microscopy. The particle electrophoretic mobilities were determined using a Rank Mk II microelectrophoresis apparatus (Rank Bros., Bottisham, Cambridge) fitted with a flat cell.

The filtrate particle concentration was monitored nephelometrically using a Camlab Ratio Turbi dimeter (Camlab, Cambridge) fitted with a flow-through cell. The meter was pre-calibrated to give direct proportionality between the turbidity reading and the particle concentration. Filter efficiency is designated η, with suffix D,E,I,G respectively for diffusional, electrophoretic, interceptional, gravitational capture. The range of experimental parameters used in the study is given in Table 1.

Figure 2 shows the permeability values obtained for a fibre bed compressed to five different porosities, with comparison predicted values. The experimental data are in good agreement with those predicted by Happel, whose model yields a flow parameter:

A_F= 1/[ - ln α - 1 + α²/(1 + α²)] (01)

This expression for Ap was used in all subsequent calculations of the single fibre efficiency, α is the volumetric packing density.

In order to calculate η from consideration of the microscopic particle deposition onto the collector surface it is necessary to model the fluid flow through the filter medium. The models of Happel (1965), Kuwabara (1959) and Spielman and Goren (1968) are all based on creeping flow conditions for an array of cylindrical collectors randomly orientated transverse to the fluid flow. Under these conditions, when fluid inertia is negligible, D'Arcy's law is obeyed:

U = KΔP/μL (02) where U is the approach velocity of the fluid, K is the permeability, ΔP is the pressure drop across the filter and μ is the fluid viscosity.

The permeability is related to the packing density, according to the respective models, by:

K_Happel = a_F ²( - ln α - 1 + α²/(1 +α²))/8α (03) K_Kuwabara = a_F ²( - ln α + 2 - α²/2 - 1.5)/8α (04) α_Spiel/Gor = 2 + 4K₁(β)/K_zero(β) (05) where β = K^-1/2a_F and K_zero and K₁ are Bessel functions of the zero and first order respectively, a_F is the fibre radius. The zeta(ζ)-potential values calculated from the particle mobility and fibre bed streaming potential data are given in Table 2. The standard deviations for mobility measurements were less than 8% for the smallest particles and less than 5% for thetwo larger particles. Although the mobility measurements taken for suspensions sampled "in-situ" at the entrance of the filter bed were much less homogeneous, they were generally in reasonable agreement with these "ex-situ" values. Deviations of greater than 10% occurred only as a result of excessive electrode polarisation, thereby increasing in the ionic strength sufficiently to produce an increase in the particle ζ-potential by counter-ion adsorption. This effect was only significant at field strength to linear flow rate ratios of greater than 900V cm^-2s^-1.

Non-electrophoretic filtration performance is considered first. Under the conditions employed (see Table 1) the η_D/η_I ratios were approximately 2:1, 1:1 and 1:2 for the particle radii of 0.216, 0.565 and 1.05 micrometres respectively. Equations from Natanson (1957a), Langmuir (1942) and Stechina et al (1969) respectively for diffusion (D), interception (I) and gravitation (G) were used in calculating the theoretical collector efficiency. The classical value for interception (Langmuir 1942) was modified to include the effects of van der Waal's attraction and hydrodynamic retardation according to the predictions of Natanson (1957b) or, where appropriate, the computations of Spielman and Fitzpatrick (1973). According to Natanson, the modified collector efficiency for capture by interception is given by

ηI_m = η_I(3 N_Ad/4)^1/3 (06) where N_Ad is the dimensionless adhesion number Qa_F ²/9μa_pA_FU and Q is the Hamaker constant, estimated from the work of Hough and White (1980) to be 0.91x10^-20 J. Although this explicit expression is adequate at high adhesion numbers, when N_Ad < 1 it is more accurate to use the Spielman and Fitzpatrick solution. The results (Table 2) indicate an increase in n_exp/ η_DIG with particle size. This is in qualitative agreement with the computations of Adamczyk and van de Ven, who predict an increase in capture over the classical value in the presence of double layer attraction at high Peclet numbers. The relevant data is summarised in Table 2. According to Adamczyk and van de Ven, capture is expected to increase with the double layer number N_D1 and the modified Peclet number Pe_m, and decrease with increasing a_P, the particle radius to double layer thickness ratio. As Pe_m is proportional to a_p4 this enhanced capture effect becomes significant for large particles in solutions of low ionic strength. Note η_exp is the experimental results.

Turning now to electrophoretically-assisted filtration it was found that the collector efficiency for capture by electrophoretic transport, η_E, was independant of filter history in the range of conditions employed. This was checked by increasing flow rate by increments of 10ml min^-1 every 4 to 6 minutes during the course of a run, allowing the outlet turbidity to reach equilibrium for each flow rate, and then reverting to the original flow rate. In 90% of the runs where this method was applied η_init = η_end ±5%. Further experiments at constant field and flow rate showed η_E to increase by less than 10% over 500ml of suspension passed and less than 5% over 200ml. This allowed several isopotential measurements to be taken at a number of flow rates in the course of a single run.

Plots were made to correlate η_E with N_E (the electric field) according to the following equations:

η_E = mN_E (07) η_E = mN_E/(l+N_E) (08) ηE = mN_E/(l-mN_E) (09) where m is the electrophoretic capture coefficient and where η_E = η_exp - η_DIG; i.e. non-electrophoretic capture is assumed to be given by the calculated theoretical collector efficiency, ignoring the effect of the double layer attractive force. As the experimentally measured electrophoretic capture efficiency values were generally large in comparison to those obtained in the absence of an electric field, the error induced by this assumption is small.

The statistical analyses of the plots are given in Table 3, and the data from equation 09, chosen because of its more consistent intercept values, shown in Figure 3. Figures 3a, 3b, 3c relate respectively to particles of 0.216, 0.565, 1.05 micrometre diameter. The results at high N_E values are scattered to too great an extent to establish which, if any, of these equations fits the data exactly. However, as η_E cannot exceed unity equation (07), analagous to the expression of Stechina and co-workers for capture by sedimentation at R<<1 , cannot fully describe capture if the electrophoretic or settling velocity exceeds that of the fluid. At low N_E values ηE approaches mN_E where the coefficient m = 0.32 - 0.44 for positive (co-flow) fields and 0.22 - 0.29 for negative (counter-flow) fields. Capture appears to show some particle size and/or particle charge dependancy.

For capture by gravitational or electrostatic forces m = 1 according to classical theory, confirmed empirically for positive gravity in aqueous media by Fitzpatrick and Spielman (1973). In the same study it was found that capture by negative gravitation was some 500 times greater than that predicted by theory CSpielman and Fitzpatrick (1973)]. The current study suggests that when an electrophoretic force opposes fluid flow η_E reaches a mi ni mum between N_E = 0 and -0.02. At val ues of η_E exceeding η_DIG, η_E increases linearly with N_E at a rate approximately 60% that of positive fields.

There is insufficient data to allow correlation between the coefficient m and the size and charge of the particle. However the results indicate that, under the experimental conditions imposed, capture by electrokinetic transport is a factor of 2-3 lower than that predicted by analogy with gravitation or aerosol deposition under a coulombic force. There are several mechanisms that could contribute to the divergance, which have so far been examined only cursorily if at all.

Tentative experimental studies using citrate to poison the fibre surface have shown that a reduction in the surface charge has little effect on η_E as long as the fibre remains oppositely charged to the particles. The discrepancy between deposition by sedimentation and electrophoresis cannot therefore be explained by electroosmotic motion of the fluid at the fibre surface shear plane. Under the same conditions, reducing the particle surface charge. using cetyltrimetyhl ammonium bromide (CTAB) caused a very significant reduction in the electrophoretic collector efficiency.

Throughout this treatment it has been assumed that fibre polarisation does not affect particle capture. Alternating current studies have shown the ionic conductivity of the fibre to be an order of magnitude greater than that of the electrolyte used in the filtration experiments. This implies that the fibre should have a significantly higher permittivity than the fluid surrounding it, and thus capture should be enhanced according to Zebel (1965). However, it is questionable that coulombic forces could alter the particle trajectory in colloidal suspensions. In liquids surface charges are compensated by the diffuse part of the double layer, such that the associated electrical forces act over only very short distances.

Specific gravity measurements revealed that the fibre has a fractional pore volume of 0.05. Optical microscopic studies showed that the fibre has a large degree of surface roughness, and therefore only approximates to a perfect cylindrical collector. Furthermore, its conductivity, although significantly higher than that of the suspension electrolyte, is at least two orders of magnitude lower than would be expected of a homogeneous ion-exchange resin of equivalent ion-exchange capacity. This suggests that the fibre is chemically heterogeneous with an uneven distribution of quaternary ammonium functional groups throughout the fibre material. The effect of such a material on the applied field cannot be predicted.

The above describes particle capture by a fibrous filter in the presence and absence of an external electric field. The field was applied parallel to the the fluid flow in both the co-flow (positive field) and counter-flow (negative field) directions. The fibres of the filter, by way of example polyacrylic, are preferably chemically modified to be set with electropositive groups. ACTILEX fibres are believed to be those described in European Published Patent Application 0 194766 Al.

In the absence of an applied field the measured values of the single fibre efficiency are in reasonable agreement with those predicted by theory, if the influence of double-layer attractive forces is taken into account.

In positive fields the electrophoretic capture efficiency η_E follows a trend qualitatively analogous to classical sedimentation or electrostatic aerosol deposition, η = mup/U where u_p is the particle velocity induced by the applied force. However, the analogy is not quantitative, the values for electrophoretic deposition being a factor of 2 to 3 lower than the theoretical values, where m = 1 in each case. This was attributed to the observed fibre heterogeneity.

In negative fields, η_E minimises at low negative field strengths and then follows the same relationship as positive fields but with a reduced proportionality factor m. No extensive empirical study of particle capture by negative sedimentation has been carried out in the past, but the results for negative fields are in qualitative agreement with the observations of Fitzpatrick and Spielman (1973). The relation of field and flow direction may be important. Even better filtration may occur when the field is across the flow. It is believed that the values given for factor m include the effect of ohmic resistance at the filter supports but this affects all measurements similarly. An estimate of this effect is a multiplier of 2-3 for m. The electrophoretic fibre efficiency was found to be independent of filter history under the conditions employed in the study and dependent to a small extent on the size and/or surface charge of the particle.

The influence of the fibre conductivity on the filter performance is more difficult to assess. It is certain that a fibre of different conductivity to the medium surrounding it will induce a deviation in the electric field in the vicinity of the fibre. The extent to which this can affect the particle transport and deposition has not previously been investigated either theoretically or empirically in the case of liquid depth filtration, even though there is a long history of research into this aspect of aerosol filtration.

Accordingly three other fibres were characterised with respect to their surface charge and bulk conductance, and their respective efficiencies compared to that of the anionic Actilex B701 fibre. The fibres were Courtelie (RTM) (a polyacrylic fibre of zero bulk conductivity), 12k High Modulus carbon fibre of very high electrical conductivity, and a cationic ion-exchange fibre of comparatively high ionic conductivity, Actilex A801. Well characterised negatively charged polystyrene latex particles were used throughout.

The apparatus and conditions described above were used except that the surfactant used to adjust the surface charge of the fibres was "Percol 1597", a polypyrollidine polyelectrolyte. CTAB was found to adsorb reversibly on the fibre surface and was therefore not used in these measurements.

Approximate zeta-potential values, calculated from ΔP:E_S plots are given in the first half of Table 4 for all fibres, some of which are base materials modified by the adsorption of the polyelectrolyte. The fibre conductivities are shown in the same table. The latter refer to the ratio of the actual fibre plug conductivity to that predicted in the absence of any fibre conductance [Fairbrother and Mastin (1926)]. The carbon fibre has an immeasurably large K_meas/K_pred value - a 10mm plug at a packing density of approximately 15% has a resistance of only 2 ohms. This means that the fibre has a negligible zeta potential, because the plug offers no resistance to a potential difference set up across it. The surface charge of the carbon fibre can therefore not be determined by this method. However, it seems probable that a fibre of substantial bulk conductivity will not undergo charge reversal on saturating with any surfactant because of its large density of charge carrying groups. This is demonstrated by the cationic Actilex A801 fibre, which remains weakly negative on saturation with the polyamide. In the case of Courtelie such charge carrying groups are probably less highly charged and more thinly distributed.

The filtration data is summarised in Table 4, and the electrophoretic capture data shown graphically in Figure 4. In the absence of an external field it can be seen that both the modified Courtelie fibre and the anionic ion-exchange fibre Actilex B701 approximate to a perfect sink. This may have been predicted from the streaming potential data, indicating positively charged surfaces for the respective fibres. All other fibres are negatively charged and this is reflected in their low measured capture efficiencies, both with and without an applied electric field.

The most interesting aspects of the results are:

(a) The ubiquity with which the equation η_E = mu_e/U = mN_E appears to apply, (where u_e is the electrophoretic velocity of the particle and U the approach velocity of the fluid),

(b) The significant impovement in the efficiency on reversing the surface charge using the polyelectrolyte,

(c) The ineffectiveness of carbon fibre as an electrophoretic filter fibre, even though the "field absent" result is not unreasonably low (see table),

(d) The marked similarity between the electrophoretic performance of Courtelie and the cationic fibre Actilex A801. It appears that a very high fibre electrical conductivity actually impai rs electrophoretic filtration. On the other hand, Actilex A801 is significantly conducting but still responds almost identically to an electric field as modified Courtelie, which is an insulator. It is evident that the excessively low resistance of the carbon fibre breaks down the field across the fibre plug, even at the high bed porosities used in the filtration experiments (96.75% void space)and preferably 95% to 96% in a range down to 90%.

Notably, the transient response of an electrophoretic depth filter composed of the ion-exchange materials differs noticably to that of Courtelie (figure 5). The filtrate concentration remains almost unchanged over a period of 30 minutes for both the ion-exchangers. A gradual increase in the filtrate concentration with time occurs with Courtelie under the same conditions.

The anionic fibre Actilex B701, from the slopes in Figure 4, clearly has a collector efficiency significantly greater than the other fibres although Courtelie and Actilex A801 when surface modified show behaviour which might be useful.

While the reasons for the unexpected filter efficiency of the Actilex B701 fibre are not known for certain the efficiency has been demonstrated by the above examples.

It appears that there may be some property pertaining to the Actilex B701 fibre that makes it very suitable for use in electrophoretic depth filters. It is possible that the Actilex A801 fibre may have some usefullness for such filters. The enhanced performance over fibres of similar surface charge characteristics, and therefore similar field-absent filter properties, is evidenced by a doubling of the electrophoretic collector efficiency and a less time-dependent filtrate concentration. It is possible that the higher efficiency of the Actilex fibres is a consequence of either their ionic conductivity or their permittivity. Although the ionic conductivity of these fibres is by no means large, they at least will not actively obstruct the passage of ions through the filter bed, in the way that an insulator must. It is quite possible that any fibre that retains an appreciable amount of water (Actilex B701 imbibes 50% of its own weight of water) will respond in the same way. A significantly large fibre dielectric constant, compared to that of the surrounding water, would also be expected to enhance the electrokinetic capture. However, interfacial dielectric relaxation, due to the electrical double layer, is actually far more significant than that of the bulk material [Einol and Carstenen (1971)]. As similar dielectric behaviour would be expected from any charged fibre in aqueous media CSchwarz (1962)] one would not expect on this basis the marked difference in the results from the Courtelie and the Actilex fibres.

The values of the electric field appear from the results but it should be noted that fields of above some 4000V/m^-1 may cause polarisation and gassing which could inhibit the filter action.

Earlier patents describe various filter structures and materials, for example US PS 4350590 (Robinson), US PS 4594138 (Thompson) and UK published Application 2198365. However none of these refer to anionic fibres in an applied electric field as set out herein.

(i) η_DIG = η_D + η_I + η_G

(ii) From uncorrected Smoluchowski equation ζ = m_e/εε₀ (iii) K = Debye-Huckel parameter

(iv) N_Dl = dimensionless double layer number according to Adamczyk and van der Ven, 4πεε_Oζ_Pζ_Fa_p/kT

I

<_o

* Surface modified using polyamide electroylte

+ Refers to a random fibrous mat of volumetric packing density

0.2 ± 0.02 REFERENCES

Happel, J., Brenner, H. 1965. Low Reynolds Number Hydrodynamics Kuwabara, S. 1959. J. Phys. Soc. Japan 14(4):527

Spielman, L.A., Goren, S.L. 1968. Environ. Sci. Technol. 2(4):279 Joy, A.S., Watson, D., Botten, R. 1965. Research Techniques and

Instrumentation 1(1):6

Fairbrother, F., Mastin H. 1924. J.Chem. Soc. 75:2318,1494

Kotera, A., Furusawa, K., Takeda, Y. 1970. Kolloid-Z. 239:677

Van den Hull, H.J., Vanderhodd, J.W. 1972 J. Electroan. Chem. 37:161 Natanson, G. 1957a. Dokl. Akad. Nauk SSSR 112:100

Natanson, G. 1957b. Dokl. Akad. Nauk SSSR 112(4) :696

Langmuir, I. 1942. Report on Smokes and Filters

Sect. 1 US Off. Sci. Res. Dev. 865:iv

Stechkina, I.B., Kirsch, A.A., Fuchs, N.A. 1969. Ann. Occup. Hyg. 12:1 Spielman, L.A., Fitzpatrick, J.A. 1973. J. Colloid Int. Sci. 42(3):607 Hough, D.B., White, L.R. 1980. Adv. Colloid Int. Sci. 14:3

Adamczyk, Z., van der Ven, T.G.M. 1981. J. Colloid Int. Sci. 84(2):497 Zebel, J. Colloid Int. Sci. 20, 522(1965)

Fitzpatrick, J.A., Spielman, L.A. 1973. J. Colloid Int. Sci. 43:350 Einolf, C.W., Carstensen, E.L. (1971), J.Phys. Chem. 75(8) 2636

Schwarz, G., (1962), J. Phys. Chem. 66 2636

Claims

1. A method of depth filtration of particles from a fluid carrying particles including

providing in a deep filter a quantity of filter medium elements, at least some of the elements being ionically conductive,

applying an electric field to said deep filter to cooperate with individual said elements,

permitting or causing said fluid with particles to flow into the deep filter,

2. A method according to Claim 1 including providing filter medium elements having an electrical conductivity varying from one part to another part of the element.

3. A method according to Claim 2 in which the one and another part are regions of a fibre.

4. A method according to Claim 3 in which the fibres are manufactured fibres having within them one or more conductive parts of discontinuous longitudinal regions.

5. A method according to Claim 3 in which the fibre is the proprietary polyacrylic material fibre ACTILEX B701 (RTM, Courtaulds Ltd.).

6. A method according to Claim 1 including permitting or causing the elements to imbibe water.

7. A method according to Claim 1 including permitting or causing the filter medium elements to have a specific state of charge, not the same as that of the particles to be filtered.

8. A method according to Claim 1 including chemically treating the elements to have a charge opposite to that of the particles.

9. A method according to any one of the preceeding Claims including providing a deep filter of a filter medium of a bed or layer of significant depth but open structure to have a high porosity, such that in the absence of the electric field particles move through without encountering the filter medium.

10. A method of depth filtration of a fluid carrying particles including

providing in quantity filter medium elements having an electrical conductivity varying from one part to another of said elements and differing from the fluid,

forming a quantity of said elements into a deep filter, applying an electric field to cooperate with individual said elements in the deep filter,

permitting or causing fluid carrying particles for filtration therefrom to flow into the deep filter,

influencing the movement of particles in the deep filter by the cooperation of said applied electric field and said individual elements to enhance the filtration of said particles from the fluid.

11. A deep filter including filter medium arranged in a filter layer, means to apply an electric field to the layer and means to move through the layer a liquid with particles for filtration, the filter medium being ion conductive.

12. A deep filter according to Claim 11 in which the filter medium is composed of fibres loosely packed to provide a high porosity, up to 96.75% pore volume.

13. A deep filter according to Claim 12 in which the ion conductivity of an individual fibre is not uniform from part to part of the fibre.

14. A deep filter according to Claim 11 in which the filter medium, in operation, imbibes water.

15. A deep filter according to Claim 12 in which the non-uniform conductivity is by the fibres not being all of one type.

16. A deep filter according to Claim 11 in which the filter medium is of loosely packed fibres, the fibres being mainly in two of the three conventional orthogonal directions.

17. A deep filter according to Claim 16 in which the means to move the liquid is effective to direct the liquid generally in the third of said three directions.

18. A deep filter according to Claim 16 in which the deep filter is a filter medium of a bed or layer of significant depth but open structure to have a high porosity and the porosity is such that in the absence of the electric field particles could move through without encountering the filter medium.

19. A deep filter according to Claim 11 in which the filter medium is an anionic ion-exchange fibre.

20. A deep filter according to Claim 19 in which the fibre ion-exchange surface is in the hydroxide form.

21. A deep filter according to Claim 11 in which the filter medium has been chemically set with electropositive groups.

22. A deep filter including filter medium arranged in a filter layer, means to apply an electric field to the layer and means to move through the layer a liquid with particles for filtration, the filter medium being polyacrylic fibres chemically modified to be set with electropositive groups.

23. A deep filter according to Claim 22 in which the polyacrylic fibres are cationic.

24. A deep filter according to Claim 22 in which the porosity is no more than 96.75%.

25. A deep filter according to Claim 22 in which the fibres are of ACTILEX B701 (a trade name of Courtaulds) treated with sodium hydroxide to modify the fibre surface to the hydroxide form.

26. A deep filter according to Claim 22 in which the polyacrylic fibres imbibe water.

27. A deep filter according to Claim 22 in which the fibres are ionically conductive.

28. A deep filter according to Claim 12 in which the ion conductivity of an individual fibre is not uniform from part to part of the fibre, the filter medium being of ionically conductive fibres capable of imbibing water.

29. A deep filter according to Claim 28 in which the fibre ionic conductivity is higher than that of the liquid.

30. A method of depth filtration substantially as herein described with reference to the accompanying drawing.

31. A deep filter substantially as herein described with reference to the accompanying drawing.