Cleanroom Airflow Velocities

Unidirectional flow and be defined as an airflow moving in a single direction, in a robust and uniform manner, and at sufficient speed to sweep particles away from the critical processing or testing area with regularity. Hence the object of the unidirectional airflow is to push outward any contamination which might be deposited into the air-stream and to avoid the potential for contamination dropping out of the air, either though gravity or by striking a object, and falling onto a critical surface.

Part of the control of air rests with air direction and this is a factor of airflow velocity. Poor airflow uniformity leads to turbulent airflow and vortex formation. In terms of the velocity of the air, this is described in some regulatory documents: 0,.45 meters per second within arrange of 20%. Whether achieving good airflow (and thereby avoiding poor airflow) needs to conform to the range specified in regulatory guidance documents has been a long-standing issue, particularly given the non-scientific origins of the regulatory guidance values. This article considers regulatory guidance on airflow velocities and the way that these are verified, and whether satisfactory airflow can be achieved outside of these guidance values. The discussion extends to consideration of the verification of these parameters at working height, especially in light of if this the most appropriate location by which to measure air velocities.

Unidirectional Airflow and a Short History of Air Velocities

Unidirectional airflow is obtained through High Efficiency Particulate Air (HEPA) filters. HEPA filters function through a combination of three important aspects. With this, there are one or more outer filters that work like sieves to stop the larger particles of dirt, dust, and hair. Inside those filters, there is a concertina - a mat of very dense fibres - which traps smaller particles. The inner part of the HEPA filter catch particles as they pass through in the moving air stream. There are different grades of HEPA filters based on their ‘efficiency ratings (1).

The concept of laminar airflow (what is now described today as ‘unidirectional airflow’) was introduced with the first industrial cleanrooms and clean spaces of the 1960s. It was noted that when air is introduced into the cleanroom at a high velocity which causes the air to travel along a unidirectional path over a required distance. In doing so, contamination is swept away from the critical area unlike the more random distribution and transition of contaminants in turbulent flow cleanrooms (2).

The development of laminar flow technology was completed in 1961 by a team led by Willis Whitfield at the Sandia Corporation (later the Sandia National Laboratories) based at Albuquerque, New Mexico, USA, in partnership with the U.S. Atomic Energy Commission (3). This concept of laminar airflow led to the development of specialised airflow cabinets whereby greater levels of cleanliness could be achieved through air passing at a sufficient velocity. It was this work that showed that an airflow velocity of 90 feet per minute was adequate to achieve the necessary levels of particle cleanliness (avoiding settling particles of a diameter of 5.0 µm and greater) whilst maintaining unidirectional flow. This velocity also fitted with the capabilities of the fans in use at the time in relation to noise reduction. While this was effective, a fuller range – such as 25 to 250 feet per minute – was not explored (4). While the range of 90 feet per minute suited the Sandia Corporation conditions, this velocity part of the first U.S. cleanroom standard - Federal Standard 209 in 1963 (a forerunner to the ISO 14644 cleanroom standard). 90 feet per minute ±20 feet per minute (which, when metricised becomes 0.45 metres per second) also became adopted by regulatory authorities, when the FDA adopted the FS 209 document. The velocity also became part of EU GMP. At no stage was this velocity reconsidered to determine what the most applicable range was based on science, or even whether it was necessary to state a range at all.

The inclusion of 90 feet per minute in the first iteration of FS 209 came with the explanatory text that the velocity was “not mandatory”. With the first revision – FS 290A in 1966 – the text “individual circumstances may dictate other values” was added. The 1973 version changes the range from 90 feet per minute ±20 feet per minute to 90 feet per minute ±20% (72 -108 feet per minute) but continued with explanatory text indicating that other values could be considered. In 1987 the first U.S. FDA guidance for aseptic processing adopted the 90 feet per minute ±20% air velocity requirement, although in the same year the FS 209C standard dropped mention of any specific velocity completely, placing the emphasis upon air visualization (5).

Current Airflow Velocities Guidance

Regulatory standards for cleanroom unidirectional airflow velocities differ in terms of where measurements are to be taken from and in terms of how much weight should be placed upon specific velocities. In terms of position, the U.S. FDA guidance, the requirement is to measure airflow velocities below the filter face at a distance of 6 inches (6). Similarly, to meet ISO 14644 measurements of the airflow velocity should be at approximately 150 mm to 300 mm from the filter face (7). However, under EU (and WHO) GMP, the requirement is to measure airflows at working height, with working height to be defined by the user. The velocity is assessed using an anemometer, a device for measuring wind speed. There are two common designs – vane and hot wire. The typical testing frequency is six-monthly or following any maintenance work or filter changes (8).

Neither EU GMP or FDA provide any recommendations about the number of readings to take. According ISO 14644-3:2005 the number of measuring points should be sufficient to determine the supply airflow rate in cleanrooms and clean zones. This should be the square root of 10 times of area in square meters. However, not less than 4 readings should be taken. At least one point per filter should be measured.

In each case, the airflow velocity range is recommended to be in the range 0.45 metres per second, ±20% (that is 0.36 to 0.45 ms-1). EU GMP Annex 1, for example states that “Laminar air flow systems should provide a homogeneous air speed in a range of 0.36 to 0.54 m/s (guidance value) at the working position in open clean room applications” (9). However, more flexibility is provided with the FDA 2004 guidance which states: “at a velocity sufficient to sweep particles away from the filling / closing operation and maintain unidirectional airflow during operation.”

The 2004 guidance further states, via a footnote to the main text: “A velocity from 90 feet per minute is generally established, with a range of ±20% around the set point. Higher velocities may be appropriate in operations generating high levels of particulates.”

The air velocity range quoted in the regulatory documents is stated to be a “guidance value”, inferring that higher or slower air velocities could be used provided there is sufficient justification. In theory, the risk from lower air velocities is from insufficient laminarity and an inability of the air velocity to effectively sweep away any particles in the air-stream. The risk arising from faster airflow velocities is from turbulence, and a tendency for the air to potentially eddy. However, do the standards really seek to imply that, say, going to 0.45 ±30% would present a contamination risk? If they do, then this is not based on sound science.

While the air velocities remain guidance, experience suggests that some regulators are more open to considering velocities outside of these ranges than others. Furthermore, it is often the case that lower airflows can provide the same level of particle control and unidirectional pattern; and sometimes faster airflows are required, either as a result of equipment balancing or due to remove particles from certain operations (such as where powder is handled). In such cases satisfactory air patterns can be demonstrated through airflow visualisation. This is why, in this author’s opinion, the airflow velocities should be removed from future updates, and placed with individual user assessments based on air pattern visualisation and particle counting. Consideration of this is discussed next.

Why Consider Alternative Airflow Velocities?

As indicated above, an airflow velocity of 0.45 meters/second (90 feet per minute), with a range of plus or minus 20 percent around the setpoint, has been established, albeit open to different degrees of interpretation, since the 1960s and it has formed regulatory guidance since the mid-1980s. There has been regulatory drift towards seeing these airflow velocities are mandatory. This is a mistake, since lower velocities, requiring lower energy use, may achieve the same effect; and higher velocities may be appropriate in operations generating high levels of particulates.

What should be stipulated instead is where air velocity becomes related to performance expectations where the air in critical areas is supplied, via point of use as HEPA- filters, in a unidirectional manner and at a velocity sufficient to sweep particles away from the critical area during operations, irrespective of the velocity setting. This means setting air velocity parameters for each processing line or item of equipment and ensuring these are justified and appropriate to maintain air quality under dynamic conditions within a defined space.

Literature also supports this position. Work by Whyte, which looked at airflow velocities covering the range of 0.1 m/s to 0.6 m/s showed that airflow velocities below 0.3 m/s were insufficient to provide stable unidirectional airflow and for achieving the required low levels of particle and bacterial concentrations. Increasing the airflow velocity up to 0.6 m/s gave low airborne counter, although this was on the basis of a ‘law of diminished returns’ in that the amount of additional energy expended did not alter the particle levels significantly. The assessment suggested that an airflow velocity of 0.3 m/s was optimal (10).

Where alternate airflow velocities are proposed, these can be assessed through the recording of particle counts and by airflow visualisation studies. The revised text to EU GMP Annex 1 (not in force at the time of writing but profiled in a recent edition of the Journal of GxP Compliance) (11) puts greater emphasis upon airflow movement than with air velocity:

“Grade A: The local zone for high risk operations, e.g. filling zone, stopper bowls, open ampoules and vials, making aseptic connections. Normally, such conditions are provided by a localised air flow protection, such as laminar air flow work stations or isolators.”

The reference to ‘laminar airflow’ is confusing and outdated. Moreover, the inference that isolators require the same air velocities as to other Grade A devices is out of step with most studies (such as Peters et al (12) and Midcalf et al (13)). The draft goes on to read:

“Unidirectional air flow systems should provide a homogeneous air speed in a range of 0.36 - 0.54 m/s (guidance value), the point at which the air speed measurement is taken should be clearly justified in the protocol. During initial qualification and requalification air speeds may be measured either close to the terminal air filter face or at the working height, Where ever the measurement is taken it is important to note that the key objective is to ensure that air visualization studies should correlate with the airspeed measurement to demonstrate air movement that supports protection of the product and open components with unidirectional air at the working height, where high risk operations and product and components are exposed. The maintenance of unidirectional airflow should be demonstrated and validated across the whole of the grade A area. Entry into the grade A area by operators should be minimized by facility, process and procedural design.”

Despite this the draft guidance does not go as far to remove the ‘guidance’ airflow velocity values of 0.45 meters per second ±20%.  However, it is suggested in this paper that airflow visualisation studies can provide the means to consider alternate airflow velocities. This approach recognises actual performance, in the operational state with equipment running and person el carrying out the necessary activities, ahead of velocity.

Airflow Visualisation Studies

The purpose of flow visualization is to confirm the smoothness, flow patterns and other spatial and temporal characteristics of airflow in an installation. For this, the airflow is examined through airflow visualisation mapping whereby smoke is generated, and the behaviour of the smoke is studied and then captured by a video camera. Air-flow studies can demonstrate a significant amount of information. This can relate to contamination control in assessing whether air-flows are drawing potentially contaminated air towards a critical zone of whether certain objects in the air-stream cause contamination by forcing the air to change direction (14).

To measure the aerodynamic performance of the unidirectional airflow unit, smoke should be introduced at the filter face so that the distribution of the smoke downwards and away from the critical zone can be seen. Smoke should also be introduced at the working height, immediately above the area where product or product components are exposed. The assessment should note the impact of the machine upon the airflow. Does the smoke, for example, entrails inwards when the air impacts upon the filling machine guarding? What is the effect of disturbances caused by the motion of machine operations? The biggest risk will be when potential airborne particles accumulate in vortex regions.  When a unidirectional air flow strikes and object, an obstacle will create a 'wake region'.  Such regions should be studied for vortices and potential particle accumulation (15).

Any regions of stagnation should be detected.  When unidirectional air flow meets an object, wakes and vortex streets can be formed.  This causes turbulence.  This can lead to pockets of stagnation in front of machinery and work surfaces that are perpendicular to the main direction of the air flow.  Such pockets can be unpredictable in speed and direction and require mapping. Another potential risk from large surfaces is that wake regions can entrain ambient air into clean zones. A consideration of the impact of Grade B areas upon Grade A zones should be considered in such circumstances. Ideally, the pattern should show that the air is characterized by a smooth flow, free of any disturbances (such as small and temporary vortices or eddies) and unimpeded.

From the above, the importance of airflow patterns is demonstrated and it is arguably more important to ensure that an appropriate airflow pattern is in place than with seeking to achieve an airflow within a particular range. However, even where a case can be made to vary the airflow velocity it remains important to measure airflow velocities at commissioning, periods of requalification and at the start and end of each test session to ensure consistency and to verify that airflows remain within validated parameters. Where airflows are found outside of range it is important that an airflow velocity study be conducted in the ‘as found’ state to assess whether the air pattern have changed to the extent that they pose a contamination concern. This leaves one question to consider: where, in terms of position, should airflow velocities be assessed?

Airflow Velocity Measurements: Working Height or Filter Face

European regulatory guidance indicates that airflow velocities should be measured at working height. This leads to two considerations. First, what is working height? And second, is working height the correct location to select, especially as a determinant of airflow patterns as described above.

Assuming first that working height needs to be measured, this is something to be determined by the user. For aseptic filling, for example, an appropriate definition of working height would be a point just above the vial neck opening, to ensure that any particles that might enter the air-stream are directed away from the open neck position. The complexity that emerges from this is that there will be various ‘working heights’ should vials of different sizes be used on a given line.

However, is working height the right location to measure airflow velocities? With airflow velocity studies, the measurement of the velocity at the working position can be highly variable due to the equipment size and configuration within the unidirectional airflow device and it can be argued that proper airflow pattern at the working position is more important than achieving the specified airflow velocity at the working position. In a sense, the velocity measurement at the working position is more for information purposes in terms of helping to understand the observed airflow pattern. This means that the FDA guidance in relation to ensuring consistent measurements below the filter face is more accurate predictor of airflow patterns than the European position of measuring airflow velocity at working height.

A final point to make is that if there are areas of concern arising from airflow visualization studies then these are be partly verified through environmental monitoring (16). This is notwithstanding some of the inherent weaknesses associated with environmental monitoring. However, trend data can be particularly useful for assessing clean air device performance particularly if a problem is detected at a later stage such as an airflow velocity reading out of range or damage to HEPA filter media. Satisfactory environmental monitoring data, provided the monitoring locations are representative of contamination concerns, can be a useful risk mitigation factor.

Summary

In summary, the consensus of the regulations, especially as enshrined in European GMP, is for airflow velocities of 0.45 ms-1 (90 feet per minute) +/‐20%. However, these values are arbitrary and their origin lying in the early days of cleanrooms. Hence other velocities maybe more suitable for achieving contamination control. The way this can be achieved is through focusing on airflow visualization. If a company intends to do this, it will be important to ensure the rationale and justification are sound. The rationale will need to include:

The “operational” airflow visualization under the actual ranges.For the height at which the velocity is measured (either the working position or, more appropriately, below the filter face) is recorded and justified.
In addition, it is good practice that where airflow velocities need to be adjusted, that airflow visualization patterns are always repeated before any processing recommences. Here airflows are a critical factor affecting the distribution of particles within a clean space (17). Moreover, it is similar best practice for environmental monitoring locations to be selected in relation to airflow visualization patterns.