Medical Device Link.

FILTERS AND FILTRATION MATERIALS

Gabriel Spera

The market for medical and diagnostic filters is dominated by a few well-known suppliers with global operations. Traditionally, each manufacturer concentrated on supplying a particular filtration medium, and device manufacturers more or less knew where to go to find a specific type. That picture is changing. In recent years, an increase in demand has made it easier for smaller suppliers to enter the market, and larger manufacturers have begun to diversify their lines to appeal to a broader base of purchasers. As a result, the difficult task of specifying an appropriate filter or filtration medium could become even more complex.

As Eric Wigner of Gelman Sciences (Brackmills, Hants, UK) explains, "Each company has its own unique style. If you look at major competitors, you see each has its own type of membrane that is its bread and butter." As applications diversify, however, suppliers like Gelman are finding new opportunities in areas outside their traditional specialty. "Other companies have come to realize there are advantages to having different types of materials," he says. "Some are starting to promote new membranes they wouldn't have offered five years ago."

Matt Dunleavy of Millipore (St. Quentin, France) agrees. "It really stems from what people are good at. Historically, Pall was nylon, Gelman was PES, Millipore was PVDF." But recently, these clear demarcations have begun to blur. Millipore, for example, introduced a PES membrane, and Pall came out with a PVDF membrane. "It's a function of market need," he says. "If the market demand goes beyond what one company can supply, other companies will come out with competitive versions."

Because there is no standard recipe for each standard membrane, explains Wigner, different manufacturers might adjust temperature, humidity, and other processing factors to optimize a particular membrane characteristic. "We can define and control processes and put in specific test parameters to ensure that we get the end product," he says, and the actual manufacturing process leaves room for innovation. Developing a particular manufacturing style, according to Wigner, requires a long period of adjustment to "get your parameters set, validated, and reproducible." As a result, many media suppliers are reluctant to provide specific details of their manufacturing techniques. Gelman's processes "took 35 years to optimize," says Wigner. "Our manufacturing process is proprietary," says Jim Rudolph of W.L. Gore (Putzbrunn, Germany), "All I can say is that we expand PTFE."

Determining a Filter's Morphology

Still, the basic methods are well known. Wigner offers a simple analogy to illustrate the process: take a vat of gelatin mixed with tennis balls and spread it out; after it sets, remove the tennis balls. By varying the size, shape, and number of balls, the manufacturer can achieve a specific morphology. Dunleavy offers a more technical description: "In most cases you start with a polymer mixed with some type of solvent," which is called a lacquer. "You run the lacquer down a belt, direct jets of air over it, and the solvent evaporates, leaving the membrane behind. There's also wet casting, which begins with a similar type of lacquer; you expose it to a liquid, which causes the solvent to dissolve into the liquid," a process known as phase separation. "You can also take compounded polymer and melt it, extrude it into a sheet; as the material cools, it forms a porous structure. We use just about all of these."

As one might expect, the choice of a wet or dry casting method is partly determined by the membrane's hydrophobic (water-repellent) or hydrophilic (water-filtering) properties. According to Holger Ebbighausen of Akzo Nobel (Wuppertal, Germany), "We offer a variety of membranes, all produced using different processes. One of the biggest is dialysis. These are almost entirely made using a wet spinning process." Hydrophobic membranes, such as polyolefin membranes for oxygen filtration, are produced through the thermal inversion process. "The solvent dissolves the polymer at high temperature and you end up with a microporous structure."

Methods for producing paper and cellulosic media also leave room for individual style. According to Simon Vincent of Whatman International (Maidstone, Kent, UK), specialists in glass-fibre and nitrocellulose media, paper filters "are depth filters, produced by laying down fibres one on top of another. We can effectively control porosity by controlling the nature of the fibre. Also, in diagnostic operations, we can alter the thickness, absorption, and wicking rate. We're not necessarily offering a generic product."

Controlling Porosity

While some filtration media rely on a random microstructure, others seek to create a surface with a precise array of pores. As Daniela Rizzardi of Saati (Appiano Gentile, CO, Italy) explains, one of the main benefits of a woven medium is its consistent openings. "When you have a fabric, all the mesh has the exact same opening. You can exactly control the particles that go through and the ones that don't, because the mesh doesn't move. The fabrics have special finishes that fix the mesh so that it is always the same. Given a certain size, we can easily tell the number of pores and measure the mesh count per centimetre." Saati routinely creates woven synthetic mesh with micron-level openings for use as primary filter media or supports. The manufacturing process is based on the same techniques that are used to weave cloth. "It's not a simple weaving process," Rizzardi points out. "There are looms, but they are very sophisticated, and the fabrics are precision woven. It's a high-technology process."

A relatively new addition to the filtration field is the track-etched membrane, which is created by literally drilling holes into membranes with an ion beam. Whatman supplements its line of paper and nitrocellulose materials with track-etched polycarbonate membranes. One of the advantages, according to Vincent, is that "you get a very precise porosity, and you can control the porosity and the filtration efficiency. Track-etched membranes tend to be used in applications where you want to retain material on the membrane surface--for example, in diagnostic tests, where you might want to tag materials so you can read how much you have." The main drawback, according to Wigner, is the number of holes: "It's simply impossible to get a lot of pores per square centimetre." Porosity directly relates to filtration speed: "How many end-users want a product that takes a long time to do its job?" he asks.

Polymer Recipes

Filter performance depends just as much on polymer composition as processing techniques, and here at least, there are a few rules of thumb. PTFE, for example, might be the logical choice for applications emphasizing hydrophobicity and durability, while PES is better suited for aqueous filtration, nylon is prized for its chemical resistance, and nitrocellulose still dominates the diagnostics market. "Because of their different surface tensions and microstructures," says Rudolph, "membranes made from different polymers have significantly different properties, including chemical inertness, water-entry pressures, airflows, and surface-release properties."

Polymer recipes can, of course, be varied, and very often a supplier can apply its expertise with a particular material to other media as well. For example, PVDF, on its own, does not normally exhibit exceptionally low protein binding, but Millipore's polymer engineers can change the basic chemistry to refine its low protein-binding properties. They apply this same technology to their PES membranes as well. Gore's expanded- PTFE membranes begin with the same basic polymer used in many commercial products, which they modify for medical applications. "It still has some of the same properties--nothing wants to stick to it--and it offers excellent durability," Rudolph says.

The same is true of paper media, as Vincent points out. "We have the ability to modify the surface of cellulose or glass fibre by actually attaching chemicals." On its own, glass fibre is a relatively weak material, and requires a binder to hold the integrity of the filter and minimize shedding. "Typically, the binder is performing a function of strengthening the paper, but we'll add chemicals so that it performs an additional function like reducing protein binding or nonspecific binding, or we'll add chemicals to make it hydrophobic or oleophobic."

Protein binding and fouling are two important measures of a membrane's performance. As Wigner explains, "Fouling is simply the process of blocking a membrane," caused when a thin layer of gel forms across the pores. "Protein adsorption is important in part because it can cause that layer, which inhibits flow. But it also means, in some instances, that important components of the fluids become attached to the membrane." In other words, a portion of the drug injected into an IV filter may not make it through. As a result, physicians might need to administer greater amounts of drugs to ensure that a patient receives the proper dosage. "Our Supor membrane has what we consider to be the most effective low drug-binding characteristic," Wigner says. With a bolus of 10 cm², for example, "You'd get comparatively 96% recovery. Only about 4% gets adsorbed into the membrane and housing." Protein binding varies among different materials: for nylon or PVDF, he notes, protein binding is higher. "Of course, not all drugs are protein. Drug adsorption is not synonymous with protein binding." The importance of protein binding is determined in part by the size and composition of the sample, and becomes more acute on the ultrafiltration level. According to Jim Camilleri of Gelman, "It depends on the particulate load--the greater the concentration, the greater the chance of fouling." When dealing with a small amount of targeted chemical suspended in a relatively large amount of fluid, any loss becomes proportionally significant. As a result, for most microfiltration purposes, protein adsorption only becomes significant at dilute protein concentrations.

Of course, certain filtration methods rely on some sort of molecular binding. Dunleavy notes that leukocyte filters are designed to attach or grab white blood cells out of whole blood not so much by size filtration but by attachment to the material. Nitrocellulose, according to Vincent, is commonly used in diagnostic devices such as pregnancy tests and infectious-disease tests precisely because proteins will stick to it readily. This property is especially conducive for immunochromatography.

Bubble Points

Membrane efficiency is ultimately determined by material characteristics and morphology: the size, shape, number, and distribution of the pores. As Wigner points out, "More holes mean higher flow rates." Also, the symmetry of the pores--funnel-shaped or tubular, for example--can really affect flow rate, adds Dunleavy. "Classically, the pore size correlates to a bubble point. For a wetted membrane, it correlates to the pressure required to push the water out of the pores. Pore size literally means the largest size of pore. For example, the bubble point for a 0.2-µm membrane is specified as 60 psi, which could mean that the largest pore will break through at 60 psi. The pore size follows a distribution beneath that. The average pore size could be significantly smaller. You need to know what the biggest pore is. Other properties, such as how long it takes to get clogged, are dependent on distribution." According to engineers at Pall Filtron (Portsmouth, Hants, UK), surface roughness can also affect filtration efficiency, particularly for biotechnology applications such as cell processing. A rough surface may cause cells to become entangled and trapped, causing a higher degree of polarization (gel formation) at the membrane surface. This phenomenon can impair the recovery of both membrane and cell products.

Specification of paper filters is slightly different, according to Vincent. "When you measure the filtration property of a paper, it's unusual to measure it as a discrete pore size--it's measured in terms of how well it will capture a defined particle. It's very difficult to determine a pore size, because the medium consists of very large pores and very small pores. The structure consists of an almost random orientation." Morphology is generally less important for depth filters used to process air and gases. As Vincent explains, "Most of our medical applications are in respiratory filters. For example, a patient under anaesthesia, using bacterially retentive papers--here, the important criteria is pressure drop, resistance of air through the system--how difficult will it be for the patient to breathe? Papers are also used in venting applications: for example, IV vents, where you need to allow air to pass in or out, or on an IV line, where you need to remove air."

Manufacturers Should Adopt a Methodical Approach

With so many variables involved, the process of specifying a filter or filtration medium requires a highly methodical approach. Dunleavy describes some of the initial considerations involved: "We start out by asking the OEM, what are you trying to filter? What pore size do you need? Do you need to remove bacteria, to sterilize, to clarify? Selecting the right pore size is the first step. Then you have to select the right polymer membrane; it needs to be compatible with the fluid. You need to consider the type of material for the housing; not all membranes can be welded to all housings. The other materials in the filter all have to be compatible, and all have to be sterilized. If all materials meet requirements, then we work with the customer to establish specifications concerning flow rates, pressure, life expectancy, maximum flow, maximum pressure, how many millilitres before the filter clogs."

There are a number of mechanical considerations that must be addressed during the early design stages. In some cases, sterilization presents a problem. PTFE, for example, is commonly used for vents, but according to Dunleavy, it can't withstand gamma radiation. "Even some of the supposedly gamma-stable PTFE membranes, which are just supported PTFE, have been known to delaminate under sterilization." Millipore's Durapel membrane was developed in part to provide an alternative to PTFE in devices requiring gamma irradiation. Wigner notes that "sealing a variety of porous media into devices requires a special talent. But that's our business. Our core business is membrane manufacturing, true, but also packaging those membranes into devices. Not just sealing technology, but designing devices to optimize fluid- and airflow characteristics. Determining the path of least resistance, the dynamics of flow--basically, how to take advantage of the membrane." Gelman employs a variety of different sealing techniques, each of which depends on the membrane and how it is coated or laminated onto a support. Ultrasonic, RF, and thermal sealing methods can all present difficulties, and so can insert moulding. Manufacturers address the issues of welding and sterilization in various ways. Saati, for example, uses plasma treatments to facilitate bonding as well as improve flow characteristics. But, of course, this process is not applicable to all media types.

Sometimes, a particular targeted substance presents unusual difficulties. Viruses, for example, fall into the transitional range between ultrafiltration and microfiltration. "They're bigger than molecules, smaller than bacteria," Dunleavy says. "Filtering them would require extremely small pores. In the case of mycoplasma, we found that just stepping down to 0.1 µm is adequate to reduce these noncellular organisms." So far, viral removal is still pretty much restricted to process streams. In the case of medical applications, "it's virtually impossible to remove a virus from blood. You would remove the blood cells. You would need an adsorptive technique, or some other technique like viral inactivation."

Nonetheless, according to Rudolph, hydrophobic membranes can function as effective viral barriers without actually filtering any viruses; rather, the membranes block the liquids that contain viruses, and filter virus-laden aerosols from air. "In any vent or barrier design, it is necessary to choose a membrane that will not allow passage of a liquid at a given application pressure," he says, because liquids can carry infectious particles. "As long as the liquid and aerosols are contained, the membrane acts as a sterile barrier," protecting the device, equipment, patients, and health-care providers.

The materials and manufacturing processes involved in filter production are not likely to change much in the coming years. Rather, engineers will look for new applications of existing technologies. In the meantime, device designers should not expect the process of specifying a filter to get any simpler. But more available options should increase the chances of finding one that meets the requirements of a particular application.