Information

Air & Gas Filtration

Air and gas filter, commonly including a highly hydrophobic membrane, are utilized in a multitude of different applications; for example tank, autoclave or lyophilizer venting, fermentor in and off-gas filtration, service gas filtration and blow-fill-seal venting and compressed gas supply filtration. All individual applications have very specific requirements. For this reason, air filter requirements can often be described as a best possible recipe. Too much or too less of an ingredient can make a big difference in the performance of such filters. If one mixes all the individual specifications of these air filters one could potentially come up with an optimized filter, which meets the stringent requirements given by all the different applications. To achieve a high effective filtration area, some filter cartridges contain a high pleat density, i.e. there is a multitude of membrane folds with the pleated cylinder. This causes the gap between the folds becoming very thin. So thin that capillary action can take place and water residue or moisture will be held within the pleats. If this happens the filter blocks and the air flow reduces drastically. This results into a longer blow-down time, the time required to reach the initial 100 % air flow through a dry filter cartridge.
Besides hydrophobicity and appropriate filter design to avoid water blockage, the design and hydrophobicity also ensure that the air flow through the vent filter is at an optimal level. High air flow through a vent filter is essential to achieve the required air flow need with low investments costs, but moreover air flow determines the running costs, especially energy, within a facility. The higher the differential pressure required to achieve the needed air flow rate the higher the energy consumption to achieve the pressure. Therefore air filters require to be optimized to flow rate and the flow rates established by the filter manufacturers require to be reliable to size the filter system properly. If the system is not sized right, it will cause a high running cost level and might result in equipment damage due to the fact that, for example a tank is not vented sufficiently.  In instances it has been experienced, that vent filter system are calculated and sized to small to be cost competitive in the investment phase of the project, to be far more expensive over the long term running costs.
An important factor is that the vent filter achieves the same air flow rate in forward and reverse direction, as vent filters are used in dual flow mode. For example a tank vent filter requires to vent the tank when it is filled and the air flows from the inner core to the outside and when it is emptied, which reverses the flow direction. In both instances the flow has to be the same to be able to size the filter accordingly.
Sterilizing grade, hydrophobic membrane filters are used to completely remove particulate and microbial contaminants. More stringent requirements also define the need of complete removal of phage or viruses. As these separation mechanisms for air filters differ to liquid filters, the retentivity of air filters is by far better than in liquid filtration. Most commonly the separation mechanism are not sieve retention but adsorptive properties. For this reason 0.2 micron rated air filters can have a retention capability of 0.02 micron. Generally the separation capability of an air filter is not measured by the retentivity but by the most particle penetrating point.

Retention Mechanisms

Sieving or Direct Interception: the particle is separated due to its size, which is bigger than the pores of the filter membrane itself. For hydrophobic membrane filters the average particle size commonly is > 0.2 µm

Diffusional Interception: particles of small sizes (< 0.3 µm) have a Brownian Motion, they move about randomly, due to collisions with gas molecules and the low viscosity of the air. Eventually they will collide with the filter material and be captured. The particle sticks to the membrane matrix or filter fiber due to van der Vaals forces. Once the particle sticks to the surface multiple G forces are required to dislodge it (see chart below).

Inertial Impaction: happens when particles with a higher mass or bigger size are not able follow the air streamline, which suddenly diverts around the filter matrix, and hit the filter matrix. This separation mechanism is indeed most effective with bigger particle sizes (> 1 µm), higher density of the particle and high velocities.

Interception: when the location of the streamline and the size of the particle result in a contact between the particle and the filter material, which results in the capture of the particle. The particle radius here is bigger than or equal to the distance of the streamline to the filter matrix. The particles stay in their streamline, that means these are usually particle of smaller sizes and lower mass.

Most Particle Penetration Point: The individual separation mechanisms are most effective at specific parameters, but especially at specific particle dimensions. The diffusive interception mechanisms works most effectively at very small particle sizes, therefore the retentive effectiveness becomes smaller with increasing particle size. To the contrary, the retentive efficiency of the interception and impaction. It becomes higher with larger particle size, but loses its efficiency with smaller particle sizes. Therefore the retention graph of these mechanisms will cross over at a point of lowest retention, which is called the Most Penetrating Particle Size (MPPS) (see chart above). At this specific point the air filter has its highest penetration rate, nevertheless it still will not reach 100 % penetration. The position of the Most Penetrating Particle Size depends very much on the air velocity through the filter. When the contact time within the filter is low, the retentive capacity of the diffusion mechanism no longer works as well as at low velocities. Also the Brownian movement of the particles will be influenced by the velocity. Therefore the MPPS is shifting to a higher penetration of smaller particles. The direct interception and impaction are not influenced to such an extent by the velocity. Nevertheless to achieve an optimal retention, the design criteria velocity needs to be viewed appropriately and the sizing of the system needs to be accordingly. To achieve the highest effect of retention, the velocity has to be low to get the highest residence time within the filter matrix, the higher the velocity the lower the retention effectiveness at small particle sizes. The velocity very much influences the penetration of sub-micron particles through a depth filter. If the flow rate increases the penetration of smaller particles increases due to the fact that the separation mechanism of Diffusion (Brownian Motion) will not be as effective as with a lower velocity. At high flow rates the contact time within the filter is shortened, therefore the chance that a particle - filter contact time becomes lesser. The Interception, is not influenced by the velocity.  Also the humidity needs to be low, otherwise the particles will be surrounded by moisture and loose their ability to move around freely. Same is valid for the viscosity of the air which will be influenced by the humidity. Once captured Van der Waals forces, covalent and ionic bindings (depending on the particle material), keep the particles on the filter material and they will not be released again, even at pressure pulsations, due to the fact that the binding forces are up to thousand times higher than the mass of the particle.

Validation

Most filter manufacturers supply validation guides with their air filter cartridges. In these validation guides, ASTM F838-05 Bacteria Challenge test correlations to non-destructive tests, like the diffusion, Bubble Point, pressure hold or Water Intrusion test, are documented. Also test results of the various required USP test, e.g. Plastic Class VI, particulate matter or conductivity test, are documented. In general flow rates are quoted, which can be used to size systems, as well as all necessary technical data, like sizes, effective filtration areas, maximum allowable differential pressures at different temperatures etc. These guides are important documentation and help one to fulfill process parameters and support further, on-site validation work. Users of sterilizing grade filter cartridges, especially Q.C. departments should always request these documents, if they are not automatically supplied.
Validation and qualification of the air filter or better air filter system needs to achieve highest attention, when a new system is installed, but also when filter types are changed. Main aspects, which need to be checked when a new system is installed are usually performance criteria, as described by 21 CFR 211.220 or Article 8 and 10: 91/356/EEC. The filter system has to cope with the required flow rates at specified differential pressures. Commonly the filter user specifies the max. required flow rates to vent fermenter at the highest oxygen consumption of the organisms within the fermentation process, at tank venting systems rapid condensation vacuums need to be overcome, at Blow-Fill Seal machines minimum flow rates and pressures need to be achieved to mold the containers, etc. Such required flow rates depend on differential pressures and pressure losses within the air system. Therefore clean differential pressures need to be specified and also maximum allowable differential pressures, when the filter needs to be changed and to control the air system. Besides the required flow rates need also to be achieved at parameters like excess of humidity or after steam sterilization. If the filter cartridges are challenged with an excess of moisture, the flow rate needs to recover quickly or be not influenced by the condensate at all. It is essential in the qualification stage to check whether the filter system meets the requirements or not. Logically such qualification needs to be thoroughly documented for inspection requirements. Same aspects count when an existing filter is exchanged by another type, as described in the EC GMP Guide Section 5.23.

Performance Qualification 

As in all validation processes, these qualification steps are known as Installation Qualification (IQ), Operation Qualification (OQ) and Performance Qualification (PQ). Besides these qualification steps, new air system will go through Design Qualification (DQ) and Equipment Qualification (EQ).

Focusing on PQ, the main aspects which need to be tested are:
- Steam sterilization cycles at process conditions
- Blow-Down-Time after steaming at process conditions or water based integrity testing
- Flow rates at specified differential pressure conditions
- O-rings seat at multiple steaming cycles
- Integrity test performance and consistency
- Thermal and mechanical compatibility at process conditions

Resulting Standard Operating Procedures (SOP´s), which are part of the validation documentation need to specify maintenance cycles and procedures, as for example described in the CGMP for Medicinal Products, Annex 1, Section 30. Such maintenance parts should describe, all the specific locations and what type of air filter is installed at the specific location. Every validation should be done with at least two different types of sterilizing air filter to have an alternative, if it is required. It would be foolish, just to rely on one specific type due to risk of a possible lack of availability and purchasing reasons. Additionally to thorough written SOP´s, staff training needs to be performed on a regular basis. Especially, in terms of the maintenance of the filters, i.e. the integrity test, systematic O-ring and cartridge inspection and exchange periods of the filters. Also the installation of the air filter, steaming at max. allowable differential pressures and blow-down pressures need to be trained. This can either be done by in-house personnel or by outside source like the filter manufacturers support services. Sufficient training can prolong the service-life time of the filter and avoid detrimental situations for the production process. As often experienced, main reasons for flaws and damages of filter systems, were insufficient descriptions in an SOP or lack of training.
One main aspect of validation work and continuous process control, is integrity testing. Several regulatory guidelines describe the need and necessity for integrity testing, these are the FDA 2004 "Guidance for the Industry - Sterile Drug Products Produced by Aseptic Processing - Current Good Manufacturing Practice"; the FDA "Guide To Inspections Of High Purity Water Systems", "Guide To Inspections Of Lyophilization Of Parenterals", both July 1993; GMP Document No. 212.72 and ISO/TC 198. If not done, regulatory approval can be failed. As can be seen these guidelines and recommendations are valid for a variety of applications, whether vent for water and product tanks, lyophilizers or autoclaves. Common sense will result in appropriate measurements to be taken to evaluate the performance of the air filter by using an integrity test, on a routine basis, if not after every batch cycle. Especially after steaming of the filter an integrity test creates the highest assurance level.
As described in the FDA 2004 "Guidance for the Industry - Sterile Drug Products Produced by Aseptic Processing - Current Good Manufacturing Practice" "Compressed gas should be of appropriate purity (e.g., free from oil)..". The same guideline also quotes that critical vent filters on lyophilizers, autoclaves and tanks, where the air comes into contact with the product, should be bacterial retentive filters. These filters should be dry, either by heat or by the hydrophobicity of the filter material. These quotation certainly give the user a guideline or hints, what air quality is required by the regulatory bodies and what kind of filter is expected at critical applications.

Materials

PTFE is certainly the most hydrophobic material, the benefits of which were discussed in the previous paragraph. It has no water adsorption and the best thermal as well as mechanical strength, in comparison with the other materials. The membrane is produced by a stretching process under specific temperature conditions. The advantages are a highly porous, but also mechanically resistant filter membrane. Due to the stretching process, the membrane experiences a certain amount of mechanical and thermal stress which creates a stable membrane. Nevertheless, not every PTFE filter will be the same, as the designs of the filter cartridges are different. Blow-down time to regain the initial air flow rate of a dry filter cartridge can be prolonged due to high pleat densities or too fine pre filter fleece structures. At this point the water will be retained within the fleece structure or in between the pleats. The hydrophobic properties of the PTFE membrane do not play a substantial role anylonger. Additionally, the pore size distribution of PTFE membrane can be different as these membranes are commonly produced in different facilities and processes. Some filter manufacturers specify tighter pore size distributions than others. An open pore structure commonly allows a higher flow rate, but could also mean a higher risk of microbial penetration. Therefore pore structures and distribution are a well defined balance between the optimal flow rate and retentivity.

PVDF is produced within a quenching process, i.e. the membrane is cast. Commonly such membranes need to be pretreated. If not the membrane properties, especially pore size distribution might be altered during steam sterilization. As air filters are steam sterilized multiple times, the alteration of a casted membrane can be drastic. For example changes in the hydrophobicity or shrinkage of the membrane can reduce the flow rate severely. In one specific case, it was reported that a PVDF sterilizing grade filter experienced a flow rate reduction of up to 20 % of its initial flow rate, after repeated steam sterilization. This would mean that the system needs to be sized with a higher security margin to avoid damage or failure.  PVDF has also a low short term temperature resistance, certainly 10 degC higher than PP, but again for a thin membrane this physical property reflects an extremely negative factor of membrane shrinkage during steam sterilization which results in lower air flow rates after repeated steaming. Membrane shrinkage can stretch the membrane to such a degree, that flaws and micro-leaks occur, especially in the area of end-cap welding.

PP and PVDF both can be effected by long term degradation, i.e. disintegration of the filter membrane to a powder. This effect happens due to the loss of the stabilizers, in cases when the filter cartridges are repeatedly steam sterilized over a longer period, for example in the large scale fermentation applications. There the usual lifetime of air filter cartridges can be up to 2 years. Besides the membrane materials the support fleeces can also be effected by oxidation, especially when the temperature of the supplied air is higher than 80 °C, which is not unusual in large scale fermentation. Such higher air temperatures are found in plants which used to use glass fiber towers. The higher air temperatures were required to avoid condensate in the filter matrix of the towers. Such elevated temperatures, plus repeated steaming initiates an oxidative breaking up of the Polypropylene support fleeces. This effect can be seen after a period of around half a year. At that point the PP starts to become brittle and disintegrates. Similar experiences can be found with heat traced housings. The elevated temperatures of the heat trace will promote the oxidation of the polypropylene support fleece. In these instances heat stabile vent filter need to be used.