Micro-filtration is a process of removing from a fluid flow target particles in the size range of 0.025 to 10 microns.

This is accomplished by passing the fluid through a microporous medium such as a membrane filter. A membrane filter acts as a microscopically fine sieve that captures particles on its surface that are greater in size than its pores. Although micron-sized particles can be removed by the use of non-membrane or depth filter materials, only a membrane filter can ensure proper quantitative particle retention. The retention boundary provided by the membrane filter also provides several important analytical advantages. Since the target particles removed from the fluid are confined to the surface of the membrane filter they can be analyzed in variety of ways not available with depth filters. For example, in addition to clarifying and sterilizing certain fluids, the contaminating organisms may be further cultured or examined microscopically.

Available techniques for the commercial production of high- quality membrane filters are presently limited. There are certain molding techniques that can, in principle, allow the fabrication of high-tolerance membrane filters (HTMF).

Ehrfeld et al. (US patent 4,872,888) describe a multi-stage procedure for creating a mold for the production of microporous membrane filters. The technique is particularly suited to the production of unitary structures where the supporting structure is integrally connected to the microporous membrane filter itself. However, the technique places limitations on the materials that may be used for the filters themselves; they must be castable in thin film form

and this is not always possible or desirable. For example, it is often preferred that the membrane filter material be preprocessed in large sheets before the pores are formed.

Secondly, the mold itself is necessarily rather delicate in nature and is unlikely to withstand the repeated use required of a mass-production system. In a related patent, Li (US patent 5,256,360) uses a similar technique to create a mold for membrane filter manufacture. In this technique, the mold is produced by means of a laser machining process and a liquid polymer resin is used in the casting process.

Despite the availability of these molding processes, the highest quality commercially available membrane microfilters are presently manufactured by a two step, track-etch process. These track-etched membrane filters begin as very thin polymer films that are exposed to a collimated beam of massive energetic nuclei ( for example, U235 fission fragments ). As the nuclei pass through the polymer film they cleave the polymer's backbone bonds and leave a trail of structural defects. In the second step, the polymer film placed in a warm caustic bath where the defect trails are preferentially etched against the bulk of the material. The result is a set of microscopic pores (see b in Figure 7) whose size can be controlled by the careful manipulation of the etch parameters and whose density is governed by the exposure time to the beam of fission fragments. The typical mean pore size tolerance for commercially available track-etched membranes is between +0 W to -20 % of the rated pore size for the membrane (Costar Life Science Filtration Catalog, 1992) While an inherent advantage of membrane microfilters is the precise control exerted over the pore sizes and densities, even the best examples suffer several drawbacks. Since the

fission fragment trajectories are random, there is a finite probability that two (or more) pores will merge (see a in Figure 1 ). This compromises the filtration properties and restricts manufacture to low pore densities in an effort to reduce this probability. The overlap of pores can be reduced by two pass techniques that essentially bore into the membrane from each side and then rely on the holes from each side of the membrane to link up as is described by Toulemonde et al (US patent 4,855,049). While this can alleviate the pore overlap difficulty, it creates a tortuous path for fluid flow that tends to compromise the hydro- mechanical qualities of the filter.

The etching procedure, though carefully controlled, must still lead to a statistical distribution of pore sizes about a mean pore size. As mentioned above, even this mean pore size is subject to a wide tolerance that can restrict the utility of the membrane filter. Moreover, the track-etch procedure requires that the film be very thin to allow the fission fragments to completely traverse the thickness, but that the film be thick enough to permit the required amount of etching without compromising its mechanical integrity, and that the film must be insulating. These requirements prevent certain types of materials from being formed into membrane filters that can offer unique advantages in filtration procedures.

If the membrane material is relatively thick and/or the etching process is not carefully controlled then the pore cross-sections will be tapered rather than square to the membrane surface. Fleischer et al. (US patent 3,713,921) describe the use of a surfactant added to the etchant to control this cross-section variation. This can reduce the cone angle of the pore but can only control it in a statistical manner. Track-etched membrane filters also

often fail to follow theoretical models of performance.

This makes designs difficult to predict and increases uncertainty in industrial applications.

Highly-energetic photons from laser sources have long been used to work polymeric materials. Saunders (US patent 3,742,182) describes a technique for creating a scanning mask in a sheet material using laser light in the infrared region of the electro-magnetic spectrum. Specifically, the procedure makes use of a CO2 laser with a wavelength of 10.6 microns to produce holes in a sheet material with dimensions of roughly 125 microns. The drawbacks of this technique can be traced to the wavelength of laser energy being used. The relatively long wavelength, 10.6 microns, places a lower limit on the size of features that can be etched into the surface of the sheet material. This limit is a diffractive effect that limits the resolution of the feature to roughly the same dimension as the laser wavelength itself and thus makes it unsuited to the production of membrane micro- filters. The second drawback is also associated with the laser wavelength. The 10.6 micron energy is quite firmly in the heat-producing infrared region of the spectrum. This means that the material removal mechanism relies on the selective local heating of the sheet material. It is an inevitable consequence of this mechanism that there will be collateral damage to portions of the sheet that are not meant to be touched by the laser. As a consequence the mechanical properties of the sheet may be severely compromised and there may be an undesirable build-up of combustion products on the surface of the sheet after it is worked by the laser.

The difficulties associated with unwanted combustion products and the specific demands of electronic printed- circuit board (PCB) manufacture are addressed by a large

number of patents in the area of laser machining of polymers. Blum et al. (US patent 4,568,632) describe a technique for selectively etching polyimide films in PCB manufacture. The polyimides are valued in this area for their excellent electrical properties and high thermal stability. In this particular PCB application polyimides are preferred to polycarbonate because of the greater degree of control that can be exerted in the etching of the former.

In experiments it is shown that polycarbonate etches roughly ten times faster than polyimide under the same conditions and that it is more susceptible to temperature fluctuations.

This is a great concern for PCB manufacture since it is necessary to etch the polymer coating on the PCB without damaging the components or connections below. In Smyth, Jr et al. (US patent 5,066,357) ultraviolet lasers are used to make flexible circuit cards with laser-contoured vias and machined capacitors. The preferred material is again polyimide because of the control in the machining process that is offered by its reduced etch rates. Traskos et al.

(US patent 4,915,981) describe a technique for creating vias in fluoropolymers with ultraviolet lasers. The application of this method is the formation of substrate materials for PCB manufacture. The vias created in this way allow electrical communication between the various layers of multi-layer PCBs. The principle of this technique lies in the incorporation of a filler material in a fluoropolymer composite that has a higher ultraviolet absorption characteristic. The patent itself is limited to fluoropolymer composites and the specific application is in the production of multi-layer PCBs. The problem of contamination is the chief concern of Ludden et al. (U.S.

Patent 5,487,852) since even the use of ultraviolet laser radiation some small residues may be left behind on the worked material. In PCB manufacture this particulate contamination can be intolerable. Ludden et al. thus turn

their attention to aromatic and/or amorphous polyamides or polymers generally with aromatic rings and aliphatic chains repeating in their polymer backbones. In their experiments they find that these compounds produce the least amount of physical charring and are thus well-suited to PCB manufacture. In contrast, they note that aromatic polyesters (such as polycarbonates) will produce some charring over the ultraviolet ablation wavelengths from 193 nm to 303 nm. As a consequence, they would exclude them from their process as unsuitable. Another example of laser ablation of polymers comes from the area of biomedical elastomers. Knight (US patent 5,336,554) uses ultraviolet lasers to create a stretchable, tear-resistant elastomeric film element with a desired degree of air-permeability.

However, the rage of perforation size (50 microns) and the minimum perforation spacing (1 mm) render such a material entirely unsuited to micro-filtration. Moreover, the use of elastomeric materials means that such components would be incapable of withstanding even moderate pressures without the gross distortion of the pores and therefore provide unpredictable filtration properties. Mathiesen et al (US patent 5,486,546) use ultraviolet lasers to form microstructures in bioresorbable elements for implantation.

These materials are carefully engineered for these applications and contain mixtures of aliphatic polyesters and aliphatic polycarbonates together with copolymers.

These are used to selectively influence the healing process separating and guiding the tissues formed around the bioresorbable element so that regeneration is achieved.

This is accomplished by the selective patterning and machining of the elastomer to produce the guiding structures.

WO 87/03021 mentions the use of laser ablation in connection with filtration applications. The ablation is used to

produce elevations and depressions in fibers. These effectively increase the surface area of the fibers and the effective porosity of any depth filter that is made up of such patterned fibers. However, a depth filter is not at all the same as a membrane filter in either construction or separation properties.

The manufacturing procedure for membrane filters described in this application addresses the problems considered above by exerting precise control over pore production in membranes. By making use of the photoablative properties of lasers operating in the ultraviolet and visible light regions, high-tolerance membrane filters ( HTMF ) can be constructed with exact pore size dimensions ( typically to within 1 micron ), exact pore positioning on membrane surface (typically to within 0.1 micron), ability to make use of a wide range of membrane materials including polycarbonates and polycarbonates that have been coated to render them electrically active, piezoelectrically active or biologically active, three-dimensional pore sculpting, and ability to use thicker or thinner membrane materials than with the track-etch process.

The fabrication of microscopic pores using ultraviolet lasers differs appreciably from techniques that employ much longer wavelengths. There are two principle advantages.

First, the ultraviolet radiation, being much shorter in optical wavelength, allows finer resolution for the fabrication of microscopic structures. Second, the material is removed from the workpiece by the rapid electronic excitation of molecules to bond-breaking energies. This cold working leaves little or none of the heat damage typically associated with lower wavelength lasers that rely on vibrational modes for material removal. An example of an ultraviolet laser is the excimer laser, excimer standing for

excited dimer. The excited dimers are formed in a high pressure plasma of a rare gas and a halogen. A transient high-voltage discharge ionizes the rare gas and the subsequent recombination of electrons and rare-gas ions yields electronically-excited rare-gas atoms that react strongly with the halogen atoms to produce short-lived excited dimers. These relax into their separate atomic components through the release of an ultraviolet photon.

These UV photons are then gathered and their numbers amplified through the usual stimulated emission to produce a coherent pulse of UV laser light.

To a lesser degree, similar advantages can be obtained using lasers producing wavelengths in the visible light band.The reduced quality of the filters obtained may be compensated by more economical manufacture using relatively higher powered and lower cost lasers. Thus Neodymium-doped, Yttrium Aluminum Garnet (Nd:YAG) solid state lasers are readily available producing relatively high output power levels, but typically produce laser energy at a wavelength of 1064 nm. In this form the laser energy cannot be utilized for strictly ablative working of polymers such as polycarbonate since the wavelength is firmly in the infrared region. However, it is possible to frequency double these laser outputs using well-known techniques to produce shorter-wavelength harmonics at the 532 nm wavelength. In this visible green range the laser photons can be used to produce pores in polycarbonate materials with a reasonable amount of precision but with some small amount of collateral damage due to local heating effects. However, these may be produced at a lower cost overall than the ultraviolet laser worked polymers.

Thus according to the invention, there is provided a method of preparing membrane microfilters, comprising generating by

means of a laser at least one pulsed beam of ultraviolet or visible light radiation having pulses of an intensity sufficient to ablate synthetic resin and having a beam width less than 10 microns, directing said at least one beam onto a synthetic resin film to form by ablation at least one through pore therein, and relatively moving the film and said beam to form an array of pores in said film.

Short Description of The Drawings Figure 1 is a photomicrograph of a portion of a track-etched membrane filter; Figure 2 illustrates diagrammatically how optical shaping of a laser beam may be used to adjust pore wall tapers; Figure 3 illustrates how multiple beams may be obtained from a single laser by means of a beam splitter; Figure 4 illustrates a multiple lens system; Figure 5 illustrates use of a pattern mask with magnification; Figure 6 is a photomicrograph showing a polycarbonate membrane micro-filter, 8 micron pore size, 50 micron pitch, produced by use of an ultraviolet laser; Figure 7 is a photomicrograph of a polycarbonate membrane micro-filter, 8 micron pore size, produced by use of a Nd:YAG laser, frequency doubled; and Figure 8 is a graph comparing pore size distributions in several polycarbonate membrane filters with a nominal 8 micron pore size; curves a, b and c relate to UV laser produced membrane micro-filters having 70, 30 and 50 micron pore pitches respectively, and curve d relates to a commercial track-etched membrane microfilter.

Description of the Preferred Embodiments Techniques for HTMF manufacture using ultraviolet or visible light lasers fall in two categories, focused beam and masked pattern.

A single, focused laser beam can be used to drill a single precise pore in a variety of membrane materials. This technique can make use of the optical manipulation of beam widths and envelopes to manufacture, in a single laser pulse, pores with varied cross-sections. All such techniques share the need to raster-scan or translate the target material beneath the laser between pulses in order to produce complete membrane filters with an array of pores.

The use of a high-quality single lens for laser beam shaping and focusing allows for high precision control of pore manufacture. After the laser beam is produced by an W or visible light laser, for example a frequency doubled Nd:YAG laser, it is homogenized to reduce intensity variations across its profile and collimated. It is then directed onto a single lens element that brings the laser energy to a focus at or near the target material. The use of lenses with varying focal lengths and numerical apertures allows control over the pore wall taper, as illustrated in Fig. 2.

Lenses with long depths of focus will produce straight, cylindrical pore walls (see pore c in Fig. 2). Lenses with short depths of focus will produce waisted beams which yield significant tapers to these walls (see pore d in Figure 2).

The taper an be controlled by lens position or focal length and can be used to manufacture so-called anisotropic membrane filters. After a single laser pulse or a train of pulses has drilled a pore the material is translated to a new position and the procedure is repeated.

If the pore is to be very large or if it is to have a shape that departs from that of the laser beam waist, then material can be removed in a series of precise pulses as the target material is translated beneath the laser. The energy and duration of the laser pulse can be used to selectively cut away, without completely penetrating the membrane, only a.small depth of material. In this way the target material

is micromachined into the proper shape and cross-section.

Under computer numerical control this is repeated all over the surface of the membrane material to complete a desired pore pattern. This technique lends itself to the production of complex or specialised filter designs.

In addition, the envelope of the pulses produced by the laser may be manipulated to improve performance. If the highest optical power is concentrated at the leading ede of a pulse, this may be used to overcome interface resistance at the surface of the material, while reduced power during the remainder of the pulse may be sufficient to sustain ablation without risking ionization of removed debris to form a plasma forming a barrier to the transmission of further energy.

If the energy of the laser pulse is sufficiently high then the laser beam can be divided into several smaller beams and used to produce several pores simultaneously. Many techniques are known for producing multiple focused beams from a single laser source. A simple beamsplitting procedure together with several appropriate lenses can drill multiple pores using a single laser pulse as shown in Fig.

3, in which a laser beam 2 is incident on a beam splitter 4 which deflects part of its energy onto a mirror 6 such that parallel beams 8, 10 are focused by lenses 12 onto a film 14 to be ablated.

The laser beam 2 could instead be directed as shown in Fig.

4 onto a multiple "insect-eye" type array of lenses 12 to produce an array of focused beams for pore drilling, or a holographic lens could be used to define a multiple set of focused beams to produce pores of desired sizes and wall tapers. As before, translation of the target material 14 with respect to the laser beam would allow the entire

membrane filter to be constructed in a series of laser pulses.

Another technique for the production of membrane filters with uV or visible light lasers uses microlithographic procedures to create a suitable mask from dielectric reflectors. A large and complex membrane filter pattern can be produced by illuminating the mask with the UV or Nd:YAG laser light. Where the dielectric reflectors impinge on the laser beam, the membrane material is left untouched.

Moreover, the use of subsequent imaging optics allows the pattern to be easily and quickly enlarged or reduced depending on the requirements of the membrane filter itself.

Such an arrangement is shown in Figure 4, in which a laser beam 2 is shown impinging on a dielectric reflector mask 20 which passes multiple beams 22. The spacing of these beams may be adjusted by a lens system 24 to provide a desired pore size and spacing. Such techniques may be combined with the multi-lens systems described above to increase the throughput of the manufacturing system provided that the laser energy is high enough to overcome losses and provide the necessary degree of ablation to form pores in the membrane.

The energy produced by available lasers is a consideration in the type of laser selected. While UV excimer lasers undoubtedly produce membrane filters of higher quality with fewer residues and greater precision than visible light lasers, the availability at relatively low cost of relatively high powered Nd:YAG lasers capable of producing, with frequency multiplication, high energy pulses in the visible spectrum, makes their use attractive where a lower quality product is acceptable. To upconvert Nd:YAG laser energy, two techniques are commonly employed. One is three way sum-difference mixing using multiple incident

wavelengths, but more commonly the infrared laser energy is used to drive a non-linear crystalline material with absorptive and radiative properties which generate harmonics of the incident radiation, e.g. 1064nm > 532nm. For example, potassium dihydrogen phosphate provides a high efficiency (83%) second order frequency conversion for Nd:YAG lasers operating at a fundamental wavelength of 1064nm. The non-linear crystalline material may be located either inside or outside of the lasing cavity, although the former provides greater efficiency and is preferred.

Example 1 A series of polycarbonate membrane filters were produced by ultraviolet laser ablation using a raster-scanned mask with optical reduction, as described with reference to Figure 5.

All have 8 micron pores in 10 micron thick polycarbonate sheet material. The pores were produced on a hexagonal grid and three pore spacings were manufactured, 70 microns, 50 microns (shown in Fig. 6) and 30 microns. It should be noted that the 70 micron pore spacing results in a pore density that is effectively equivalent to the rated pore density of a comparable track-etched membrane filter (100,000 per square centimeter).

The laser utilised was a xenon chloride excimer laser producing pulses of ultraviolet radiation at a repetition rate of 200Hz and a power of 50 watts to provide an estimated energy of 15 micro joules per pulse per pore site.

12 pulses were required to break through the film, corresponding to a duration of 0.06 seconds per set of pores.

To estimate the performance of these filters in relation to a commercially-available track-etched membrane filter such as that shown in Fig. 1, a bubble point measurement (ASTM

method F 316-70 (1976)) was performed on each. Using these measurements it is possible to estimate the distribution range of pore sizes, as shown in Fig. 7 and reported in Table 1 below. In Figure 7, curves a, b and c relate to the 70, 50 and 30 micron filters of this example, and curve d relates to the filter of Figure 1. It is evident that the laser-ablated membrane micro-filters show a much tighter distribution range than a similar commercial track-etched polycarbonate membrane micro-filter.

Table I Pore size estimations from bubble point measurements for 8 micron nominal pore diameter membrane micro-filters membrane filter mean pore standard deviation of mean pore diameter diameter (microns) (microns) laser etched, 7.9 0.28 30 micron pitch laser etched, 8.5 0.32 50 micron pitch laser etched, 7.6 0.66 70 micron pitch track-etched 7.4 1.30 membrane Example 2 A further membrane microfilter was produced from the same polycarbonate film, with the same nominal 8 micron pore diameter and a pore spacing of 50 microns, but using a frequency doubled Nd:YAG laser using an intra cavity doubling medium to produce light in the visible spectrum at 532 nm wavelength, at a pulse repetition rate of 200Hz and 60 nanosecond duration. The estimated energy applied per

pore site was 20 microjoules, and a single pulse was sufficient to break through the film, equivalent to an applied energy of 40 joules per cm2. As will be noted from Fig. 7, the quality of the resulting filter is reduced as compared with those of the preceding example,with more visible residues and less uniform pore size, but the use of less expensive equipment and increased production rate may compensate for this where the very highest performance is not required. Although it appears that the energy required to form a pore in this example was considerably lower, the pore forming technique used was different (the pores were formed one at a time) and figures for energy applied per pore site in both examples are estimates intended for guidance only.

The advantages of laser-produced polycarbonate membrane filters are numerous. The laser allows a precise control over pore size and shape even allowing for complex geometries. The method of manufacture allows the pores to be positioned any where over the surface of the membrane filter eliminating unwanted pore overlaps and allowing for complex pore arrangements and composite patterns. The use of a laser system reduces restriction on the dimensions and thickness of the membrane used for manufacture opening the field to a wide variety of novel membrane materials utilizing polycarbonate as a support material, that may offer advantages in separation technology over conventionally produced membranes. The manufacture of laser drilled membranes reduces or eliminates the possibility of chemical (or other) contamination of filter surfaces that may interfere with desired function. The use of masked patterns to create filters is a rapid and simple technique for mass production in a single step process. The regular pore geometries and sizes allow these filters to be accurately described by existing filtration theories thus

yielding shorter and more effective design cycles. The laser machining of the filter surface allows the membrane's cross-section to adopt complex and potentially beneficial shapes for various types of filtration procedures.

Although the use of polycarbonate films has been exemplified and is preferred, the method of the invention may also be used with other synthetic resin film materials, preferably thermosetting, which can be ablated utilising UV or visible light laser beams. Successful tests have been made using an W examer laser to form pores in a 12.5 micron film of an aromatic polymide film marketed under the trade mark KAPTON.

The choice of laser wavelength is not critical, although shorter ultraviolet wavelengths will provide better results, and will largely depend on the availability of lasers capable of economically providing pulses of sufficient power that membranes may be prepared at an acceptable rate relative to their quality. The method of the invention may be combined with other fabrication techniques: for example, pores formed by the method of the invention could be enlarged by chemical etching particularly where a visible light laser is used, undesirable residues can be removed from the membrane surface by rubbing or scrubbing.