domnick hunter - a division of Parker-Hannifin Corporation. World-Class Filtration, Purification and Separation Solutions to Industry

• View all products
• Customer Support
• Compressed Air Technical Centre
• Global Exhibitions and Events
• Careers
• Awards & Accreditations
• Legal

1. Introduction

The Water Intrusion (WI) or Water Flow Test (WF) was introduced a number of years ago and has been adopted by a small number of filter users looking for a method that eliminates the requirement for the introduction of a ‘contaminant’ such as IPA (Isopropylalcohol). The test has been designed for hydrophobic sterile gas cartridges. Existing liquid based tests for these cartridges involve the use of an alcohol to reduce the surface tension of the liquid and hence aid the wetting of the membrane. The diffusional flow of gas through the wetted pores is then measured. As this method wets the pores the filter must then also be dried prior to use. The use of IPA, especially if attempting to test multi round systems can also pose a health and safety issue due to flammability. The fact that only DI / WFI grade water is used during the WI test and the membrane is not ‘wet out’ means that in principle it is a very attractive test method. However, in practice the repeatability of the test is difficult to achieve due to a significant number of influences. This has been demonstrated through feedback from end users and work conducted in-house at domnick hunter. The purpose of this document is to highlight all the issues encountered to date that have influenced the values recorded during a WI test. This will give potential users of the test the best possible information to make an informed choice. If the decision is to use the test then the information contained within this document will provide the user with valuable support towards the development of a robust and repeatable test method.

2. The Water Intrusion Test:

i) Principles In Practice

• The upstream side of the filter is flooded with WFI or DI water.
• The water is then pressurised (typically to 2.5 barg (36 psig))
• The volume increase of an upstream compressible volume caused by the decrease in water volume is then measured either by:
  • Pressure drop (following isolation of the pressure source)
  • Gas flow required to maintain the test pressure

The WI value is essentially the increase in compressible volume expressed in Nml/10min i.e. the equivalent gas volume at s.t.p. The other form of test expression is water flow. This is a measure of the actual water volume drop.

ii) Theoretical Principles
The main question that has been debated over the years is; "what causes the drop in water volume on the upstream side of the filter?" The decrease in water volume (and associated increase in compressible volume) can be divided into two phases. The first is during the stabilisation phase the second during the test phase.

a) Stabilisation Phase
The largest pressure drop occurs during initial pressurisation. It is widely accepted that this increase in upstream compressible volume is due to:

• Pleat compression
• Cartridge deformation
• Expulsion of air within the membrane support layers.

b) Testing Phase
Once the test system has stabilised why does the water volume continue to decrease, especially when the membrane is hydrophobic? Various explanations have appeared in technical literature over the past recent years, primarily based around water intrusion into the pore structure of the membrane and evaporation losses:

Intrusion into Pore Structure
Water gradually intrudes into the membrane over the period of the test, with the presence of larger pores in the filter being detected by an increased flow due to bulk water flow through these pores. Tests conducted at domnick hunter show that no continuous intrusion occurs in the majority of the membrane structure but confirms that continuous intrusion into a minority of the larger pores can take place resulting in a small amount of water on the downstream side of the membrane. If continuous intrusion did occur through all the pores it has been calculated that the membrane structure would completely ‘wet out’ in approximately three hours. This does not happen as demonstrated by the fact that air can pass through the cartridge immediately after a WI test.

Evaporation Phase
Water level reduction on the upstream side of the cartridge is also the result of evaporation at the high-pressure water / downstream air interface. A cross section of membrane is shown below. This shows the slight intrusion of water into the membrane as well as the path for evaporation.

3. Correlation to Bacterial Challenge

For the test to be accepted in the pharmaceutical industry; the main application area for this test, a correlation to a liquid bacterial challenge ( ASTM 868/ Advemed) needs to be established. To guarantee filter performance a filter must be capable of being non-destructively integrity tested. This was recognised by the FDA, “Guideline on Sterile Drug Products produced by Aseptic Processing” (June 1987) which states: “After a filtration process is properly validated for a given product process and filter, it is important to assure that identical filter replacements (membrane or cartridge) used in production runs will perform in the same manner. One way of achieving this is to correlate filter performance data with filter integrity testing data. Normally, integrity testing of the filter is performed after the filter unit is assembled and sterilised prior to use. More importantly, however, such testing should be conducted after the filter is used in order to detect any filter leaks or perforations that may have occurred during the filtration”. To achieve this objective the correlation between bacterial challenge retention and a non-destructive integrity test must be proven. The procedure documented in ASTM F838-83, for determinating Bacterial Retention of Membrane Filters utilised for Liquid Filtration is used to test the manufactured product. The filter must be challenged with a minimum of 107 viable Brevundimonas diminuta (ATCC 19146) per cm² of effective filtration area.* Any organisms which pass the test filter are collected and cultured on the surface of analytical discs. In this way colonies may be counted and bacterial species identified. The bacterial challenge is quantified by expressing the filters efficiency to remove the challenge organism from the challenge suspension as a Log Reduction Value (LRV). To be classed as a sterilising grade filter the test filter must produce a sterile filtrate and LRV >7.

The table below demonstrates that a 250 mm (10") HIGH FLOW TETPOR II 0.2 micron filter with a water intrusion of <16.5 ml/10 min will produce sterile filtrate under these test conditions. To build in an added level of security to the user, domnick hunter has reduced the allowable water intrusion. A 250 mm (10") HIGH FLOW TETPOR II 0.2 micron filter therefore has a maximum allowable water intrusion of 13.5 ml/10 min at a test pressure of 2.50 bar (36.3 psi) at 20°C (68°F). The maximum water intrusion for other HIGH FLOW TETPOR II cartridges is based upon the ratio of their surface area compared to a 250 mm (10") cartridge.

4. Key Considerations when Water Intrusion Testing

A number of critical areas can be identified as worthy of consideration when conducting a water intrusion test;

i) Water condition

• Temperature
• Quality

ii) Cartridge condition

• Contamination
• Dryness

iii) Procedure for filling the upstream volume with water

• At atmospheric pressure
• Under pressure

iv) Performing the test

v) Post test filter use

i) Water Condition

Temperature
As the water loss on the upstream side of the filter is due primarily to evaporation of water through the membrane it is essential that the temperature of the water is equal to that of the ambient air. If the temperature of the water is greater than that of the air the evaporation losses will be greater and hence a larger WI value will be recorded. Therefore the water used for test purposes should be stored in the test area for a minimum of 24 hrs for the temperature to equilibrate. The actual temperature of the water will also influence the measured water intrusion value. Correlation of water intrusion test values to a liquid bacterial challenge have been conducted in line with industry standards at a test liquid temperature of 19-21°C(66-70°F). If this temperature range cannot be achieved in practice, an appropriate correction factor should be used. See the graph detailed below. The test should ideally not be performed at temperatures exceeding 25°C(77°F) or below 15°C(59°F).

Correction factors for WI at various temperatures

Quality
The quality of water is very important if consistent results are to be achieved. It is recommended that DI or WFI quality water is used. Using untreated tap water can result in spurious values being recorded due to possible contaminants.

ii) Cartridge Condition

Contamination
Cartridge contamination may influence results if these contaminants affect the hydrophobicity of the membrane. Possible contaminants include:

• Chemical Additives used in steam boilers. This becomes evident when too much additive is used due to boiler shut down or low steam usage. It would normally also be associated with a discolouration of the cartridge; typically a green or red tinge.
• Rust from corroded pipework
• Contaminants from fermenter broth on off gas filters. Fermentation broth often includes antifoams and these can be deposited onto the membrane throughout the fermentation in the form of aerosols. The antifoams can potentially reduce the hydrophobicity of the membrane again leading to incorrect Water Intrusion values.

Dryness
The cartridge needs to be fully dried to ensure the WI value is correct. Any residual water within the pore structure of the membrane will result in a higher WI value being recorded. The few companies that have adopted the WI test have often incorporated a drying cycle (Typically 60°C for 1 hr) into their testing procedure to ensure consistent results. If a drying cycle is used it is essential to validate this to guarantee consistent results.

Temperature
As a result of a drying cycle the cartridge may be at an elevated temperature compared to ambient. It is essential that the cartridge is allowed to cool down to ambient conditions before the water is introduced into the housing. Experience has shown that as the cold water contacts the warm membrane a thin film of water is heated which can result in localised water penetration resulting in an abnormally high result. This can be most important when testing a vent filter on a WFI tank that utilises a heater on the housing to prevent condensation during use.

iii) Filling the upstream volume with water

Without Pressure
One method is to simply pour water into the top of an open housing and to flood the upstream side of the filter. The problem with this method is that on pressurisation the volume will drop due to pleat compression, air expulsion etc. In some cases this can result in the water level dropping below the cartridge resulting in bulk air flowing through the membrane. This is especially apparent on smaller cartridges (<10"). To overcome this problem it is recommended that a small additional volume is attached to the top of the housing. This is essentially a DEMICAP housing with 1" tri-clover connections (Order no XAKPP08). The empty DEMICAP is attached to the housing as shown below. It is then filled with water and connected to the integrity test equipment. As the water is forced into the pleats the level of the water in the housing is maintained as it is replenished from the additional volume.

With Pressure
Filling under pressure ensures that the water is forced into the pleats during this process resulting in a much reduced initial volume drop when the test is initiated. This method eliminates the requirement for an additional upstream volume. When filling under pressure it is important that the cartridge is not pressurised above the test pressure. The housing should be filled through a drain port allowing the air to expel out through the vent. The fill rate should be low to ensure that the full upstream volume is filled and to reduce the risk of over pressurisation.

iv) Performing the integrity test

The integrity test sequence is typically 10 minutes stabilisation and 10 minutes test. The main factors that can affect the test at this stage are :

• Fluctuation in ambient temperature
• Problems with the integrity of the test system i.e. not leak tight.

Fluctuation in ambient temperature
As the values being measured are extremely small compared to a standard diffusional integrity test with an alcohol / water mix (typically 1/20th) the effects of temperature difference have to be considered. This includes not only ambient temperature fluctuations but also differences between the water and air temperature in the compressible volume. As discussed previously the test water should be at the same temperature as the test environment. If this is not ensured false failures / passes will be recorded. Similarly it is important to control the ambient air temperature. However, whilst this may be readily achievable within a laboratory environment, in production, appropriate controls may not be possible. For example, using the WI test to integrity test filters on a bulk fermentation facility (multi round housings) is not practical. For these reasons, implementation of the water intrusion test within a production environment requires careful risk assessment (contact domnick hunter for further advice).

Housing Integrity
As the water intrusion values measured are small, all leaks on the upstream / non sterile side of the housing have to be eliminated. This can involve the replacement of all isolation / vent valves within older facilities which can result in significant additional expense not visualised when the test introduction was agreed. When testing small filtration systems, the maximum allowable WI value will be extremely low (e.g. 3.5 ml/10min). This means that any small leak in the system can result in an incorrect failure being recorded. Whilst this is obviously a fail safe result, i.e. an integral cartridge will register as a failure, it leads to frustration for the operator conducting the test and complicates manufacturing operations

v) Post test filter use

After the test has been performed the water has to be drained from the upstream side of the filter. To facilitate draining, pressure should first be vented from the housing either manually or by the integrity test machine. Once the pressure has vented the water can be drained from the housing. The vent valve should remain open to aid draining and to prevent a partial vacuum forming. It is recommended that the drain valve is left open for approximately 30 minutes to ensure the maximum amount of liquid is removed from the filtration support layers. If air is processed immediately following the WI test residual water can be forced onto the surface of the membrane leading to high initial differential pressures. This is not generally problematic in pressurised systems but if the cartridge is being used as a vent the consequences can be a collapsed tank. A number of manufacturers cartridges have been tested and all would exhibit the same potential problem. It is therefore advised that a housing bandheater is used for at least 30 minutes to minimise the risk of high differential pressures being generated.

5. Typical Production Test Procedure

On-Line Water Intrusion Testing of a Vent filter on a WFI Tank
Whilst the implementation of the water intrusion test has been limited, the most common application of this test has been for integrity testing of vent filters on WFI holding tanks. This has therefore been used as a case study for illustration purposes. To perform the integrity test a fully automated integrity test instrument such as the PORECHECK 3 should be utilised. The most practical way to initially fill the housing with the test water is the use a pressurised delivery system.

Procedure
Important : Ensure the test water temperature is equal to the surroundings

Test start: Valves B and D closed ; A, C and E open

1. Switch off the bandheater and allow system to cool for 1 hour.

2. Connect the water delivery tube to the base of valve C.

3. Open valve D and begin to slowly fill the upstream side of Housing through valve C using a separate pressure vessel.

4. When water exits from valve A, close valve C and D. n.b.If filling under pressure is not possible an extra upstream volume should be used to prevent the water level dropping below demi type cartridges.

5. Connect the Integrity tester to valve A and start the WI test.

6. Following the test, disconnect the integrity tester and drain the water from the housing via valve C. Allow the water to drain for approximately 30 minutes to ensure the majority of the water has drained from the membrane surface.

7. Switch on the bandheater and allow to attain temperature prior to re-using the vent filter.

6. Conclusion

The Water Intrusion test is practical in an environment which is controlled with respect to temperature and cleanliness. It is essential to observe the recommendations in this document to achieve consistent repeatable results. Ideally the test protocol should be validated by repeatedly testing a single cartridge and verifying consistent results. The test maybe affected by contamination on the membrane surface and so should only be implemented on sterile gas cartridges which are adequately protected by both pre-filters and steam filters.

7. Technical Support Group activities

domnick hunter have a trained team of scientists and engineers available to answer questions regarding the technical capabilities of our products, to assist in the selection and design of appropriate filtration systems and to provide user training programs. The following services can be delivered both on site and in-house;

• filterability testing to optimise filter system design
• advice on the development of integrity testing, steam sterilisation and clean in place procedures
• development of validation procedures
• troubleshooting
• facility audits to ensure continued optimisation of filter use
• operator training including filtration theory, filter system design and management, validation, etc.

For more information on any of the above support services please contact your local domnick hunter representative. email: tsg@domnickhunter.com