Improving filters for process applications

FIltration is an essential, if unglamorous, part of pharmaceutical production. Stuart Nathan* reports on recent developments in filtration membranes

Process filtration is seen as a rather low-tech part of the gleamingly high-tech world of pharmaceutical production. Of course, it is one of the most vital operations, as it recovers the precious active ingredients from reaction mixtures, or removes impurities, boosting the yield of valuable materials. But glamorous it isn't — it invariably involves damp sludge. This might lead people to assume that development work in this field is somewhat low-key. This, however, is very far from the truth.

Membranes used for filtration tend to fall into two distinct groups. The simpler type, known as screen membranes, have cylindrical, capillary-type pores which are more or less at right angles to the surface of the membrane. These are generally used for analyses and other types of laboratory work. The other type has a more complex structure, with a rough surface and long, interconnecting networks of pores which follow tortuous paths through the thickness of the membrane. Known as depth membranes, these are the filters which are used as part of a process.

Screen membranes are usually made from polycarbonate or polyesters, using a manufacturing process which starts with a poreless sheet. This is bombarded with high-energy charged particles, generally from a radioactive source, which pass through the sheet leaving trails of weakness called damage tracks. These are etched out into pores by soaking the sheet in a strong alkaline solution — the diameter of the pores depends on the soaking time.

Depth filters, in contrast, are generally cast. For example, to make a filter from a cellulose-based polymer, the material is first dissolved in a mixture of organic solvents and various additives which control the properties of the finished material, and then cast as a film onto a moving belt. The solvents are evaporated under carefully controlled conditions, leaving voids in the film which form the characteristic labyrinthine pore structure.

Depth membranes can also be made by a process known as precipitaton casting. In this, the membrane film is cast from its organic solution, then immersed in a water-containing bath. The difference in solubility in the organic and aqueous solvents forces the polymer to crash out of solution, creating a porous structure with relatively few pores on the surface, but many interior channels. Such membranes are generally used for purification and desalination.

Manufacturing techniques can be the key to many properties. At the materials science department at MIT in the US, Anne Mayes has developed a method for making membranes with molecular-level 'bristles' which reduce fouling. The bristles are hydrophilic, so they repel organic materials, particularly proteins and other biological materials.

Mayes' membranes are made from two different polymers. The main body of the membrane is polyvinylidene fluoride (PVDF), a hydrophobic polymer commonly used as a depth filter. The 'bristles' are made from a 'comb' polymer, which consists of a hydrophobic backbone lined with short lengths of hydrophilic polymer chains. The type of polymers used can vary, imparting different properties to the final polymer.

The membranes are made by casting, similar to the typical process; however, the starting material is a solution of both the PVDF and the comb polymers. When cast, the two polymers segregate, with the combs migrating to the surface forming a dense layer of 'bristles'.

The membranes have a larger number of surface pores than membranes formed without the comb additive, Mayes says. This improves the flow of fluid through the membrane. Moreover, the bristles actively repel oily materials. Conventional membranes can be modified to resist fouling through treatment with various additives, but the comb polymers achieve comparable — or even better — results without the need for any further expensive, time-consuming and not always reliable treatment.

Moreover, the membranes appear to be 'self-repairing' to a degree. If they are damaged, for example by an aggressive cleaning solution, heating them to 90.8°C in water causes the combs to resegregate on the surface and relines damaged surfaces.

The addition of the comb polymer could open the way for highly specialised — but still reasonably low-cost — membranes. For example, the bristle polymers could be pH-sensitive for use in processes which depend on tight control of acidity or alkalinity. Bristles incorporating chelating agents could be used to capture metal ions in filtrant streams. Even more exotic structures could also be possible: Mayes is working with another MIT researcher, Linda Griffith, on comb polymers which incorporate cell-signalling peptides, which could be used in artificial tissues — the bristles would encourage cell attachment and growth.

Membrane materials are not limited to these, however. Boulder, CO, US-based firm Ellipsis specialises in metallic membranes. Like many components used in the process industries, these were originally developed for a completely different application from those where they are currently used. In this case, the genesis was a project funded by the US department of energy to remove radioactive particles from water which had been used in the laundries which handled the protective clothing used by workers involved in cleaning up nuclear installations. Ellipsis handled the design and fabrication of the various different types of filter, while a research centre at the University of Colorado was responsible for investigating and characterising their properties.

After having tried a variety of materials and manufacturing techniques, the project partners finally developed a membrane which met the DoE's specifications — only for the department to change track completely and decide to send contaminated clothing elsewhere for cleaning and treatment. This left Ellipsis with a product which was obviously in search for a market — a very robust, durable filter which could be subjected to harsh environments and could be cleaned very quickly and very thoroughly.

It was quickly obvious that pharmaceuticals and biotechnology would be among the most important markets for the membranes, says Ellipsis president Robert Herrmann. The pore size of the filters is relatively large — around 0.5mm or around 0.2mm — which makes them suitable for uses such as buffer solution sterilisation in biotechnology, or recovery of intermediates and end products in pharmaceuticals. The smaller pore size is also suitable for removing microorganisms such as Pseudomonas diminuta from aqueous streams.

Unlike conventional depth filters, the metal filters are not susceptible to fouling — particles are trapped on the surface, rather than in the depths of the material. This means that cleaning by back-flushing is quick and efficient.

Ellipsis has had some problems in bringing the membranes — which it has named 'Metall-O-Pore' — to market. The investment climate in Colorado is not geared towards this kind of product, Herrmann explains. However, the company is currently working on a strategy of licensing the membranes for contract manufacturing of filtration equipment, and is in negotiations with two such companies. 'Networking and timing turned out to be the critical ingredients,' Herrmann says.

Filtration does not have to involve solid membranes at all, however. Liquid membranes — a solvent and carrier layer separating two aqueous solutions with different pHs — have been investigated for extractions of biosynthetic substances since the 1980s. Carboxylic acids, ranging from acetic acid to fumaric and phenoxyacetic acids, can all be extracted through liquid membranes, as can amino acids and penicillin-related antibiotics. In effect, this represents filtration of individual molecules. In fact, according to Romanian Oniscu Corneliu and colleagues, recent research has suggested that use of liquid membranes can speed up some enzyme-mediated reactions, such as the conversion of penicillin G to 6-aminopenicillanic acid.

Liquid membranes can be made by emulsification of the solvent and carrier, or by supporting the solvent in hydrophobic porous polymer. In a paper currently in press, Corneliu and colleagues discuss a method for separating penicillin V from phenoxyacetic acid through a liquid membrane of Amberlite LA-2 using 1,2-dichloroethane as a carrier. They found that if the pH difference between the feed phase (the aqueous solution containing the penicillin) and the stripping phase (the solution which receives the extracted drug) is small, the concentration of Amberlite in the liquid membrane in low, and the two aqueous phases are mixed vigorously, the extraction becomes highly efficient and selective.

It seems that the expansion of technology in the filtration area is targeting the life sciences areas above most others. While the drinking water market is still taking the bulk of revenues for filters, specialised uses are driving development into ever more diverse areas.

Lilly uses Delta Neu filter on Merseyside

Delta Neu has replaced a troublesome fabric filter with a wet scrubbing system at Eli Lilly's Merseyside, UKsite. The bag media was becoming clogged very quickly because the dust and vapour passing through it were sticky, humid and at a temperature of 90°C. Delta Neu replaced the filter with an Aqualine R Wet Scrubber unit, designed specifically for treating wet, hygroscopic, sticky or explosive dusts. To minimise the cost, existing extraction hoods and dirty-side ductwork were reused, but clean-side ductwork was replaced. A high-velocity air discharge section with dust sampling points was connected to the outlet. The design ensured that conveying velocities of 18;20m/s were maintained, so preventing dust drop out and condensation. A control panel and alarm sensors were supplied for all inputs and outputs. The filter operates at peak air volume without the need for media reconditioning. Dust and vapour are conveyed through a water spray and fan. A cyclonic separator removes the liquid effluent from the air stream. Because the collection medium is water, it minimises the risk of fire and explosion. The filter achieves 95% efficiency or greater. Contact:Delta Neu, Woking, Surrey, UK;
tel +44 1483 737888; fax +44 1483 755177.