A ‘STANDARD’ APPROACH TO VALIDATION BY PARTICLE COUNTING

It is not the intention to deride any particular standard or protocol, as they are in themselves vital to our industry. What is intended is to highlight possible differences in the validation work undertaken to achieve these standards. The diversity of the sources of the standards currently widely used is great, reflecting the diversity of the usage of cleanrooms by companies based within the UK.

There are inevitably conflicts and anomalies between standards that are thought to be comparable and of an equal status. Cleanroom users and owners beware; your test report may not tell you exactly what you think it tells you, or what you would like it to tell you!

All too often cleanroom owners do not know what validation standards to work to, and are of the opinion that there is no real difference between the standards. Essentially, there is very little difference in the environmental cleanliness class tables appearing in each standard, but in reality are some sampling protocol and application differences, a few of which are outlined below.

If cleanroom owners want to get more out of validation work, a better understanding of the standards and protocols will work in their favour. Test specifications should be tailored to your requirements, and should still be able to demonstrate conformance to a particular standard at the same time. Knowledgeable and reputable validation companies can help them to achieve this if advice is sought.

The three mostly widely used standards in the UK are: British Standard (BS) 5295: 1989 ‘Environmental Cleanliness in Enclosed Spaces’, American Federal Standard 209E ‘Airborne Particulate Cleanliness Classes in Cleanrooms and Clean Zones’ and the Medicines Control Agency Guide to Good Manufacturing Practice (GMP) for medicinal products 1997 – Annex 1 ‘Manufacture of Sterile Medicinal Products’. The ISO/TC 209 standard has not been considered, as voting on Part 1 ‘Classification of Airborne Particulates’ of the draft only ended on 12/05/97.

Cleanroom Mode

GMP has been recently revised and the grading classification system has been changed. Previously the grading system made no reference to mode of operation, only that ‘it is accepted that it may not always be possible to demonstrate conformity with particulate standards at the point of fill when filling is in process, due to generation of particles or droplets from the product itself’. The grading classification system has now been applied to facilities in the ‘at rest’ mode with an ‘in operation’ figure also provided for each grade. The guidance given is such that in order to meet the ‘in operation’ conditions, areas should be designed to reach certain specified air-cleanliness levels in the ‘at rest’ occupancy states.

The Note (b) beneath the grading table in Annex 1 states that ‘The guidance given for the maximum permitted number of particles in the "at rest" condition corresponds approximately to the US Federal Standard 209E and the ISO classifications as follows: grades A and B correspond with class 100, M3.5, ISO 5; grade C with class 10,000, M5.5, ISO 7 and grade D with class 100,000, M6.5, ISO 8’.

This terminology is essentially incorrect with regards to the Federal Standard 209E classification system. The Federal Standard 209E classification system is solely based upon the ‘operational’ mode, not the ‘at rest’, or ‘as built’ modes. In Federal Standard 209E terms, a grade B GMP installation is effectively a class 10,000 (M5.5), a grade C installation is effectively a class 100,000, and grade D installation would not attain a Federal Standard 209E classification.

The following table reproduced from Section 4.4 of the National Environment Balancing Bureau (NEBB) ‘Procedural Standards for the Certified Testing of Cleanrooms’ 2nd Edition, 1996, gives recommended acceptance criteria for cleanrooms in order to meet the design class under ‘operational’ mode.

General Acceptance Levels for Other Modes of Particle Count Testing

Percent of Operation Mode Classification

BS 5295 allows testing during any mode using the classifications given in Part 1 Table 2, as long as the mode is specified upon the test report sheet (i.e. ‘as built’, ‘unmanned’ or ‘manned’).

Number of Sample Locations

The American Institute of Environmental Sciences & Technology (IES-RP-CC006.2) ‘Testing Cleanrooms’ states the following under section 6.3.2 – Procedure for the airborne particle count: b) 1)

This principle is widely agreed with and understood by all. The same recommended practice advises that the number of test locations should be based on the equations given in the Federal Standard.

Federal Standard 209E sample location equations can cause problems when applied to different scenarios though. If we take a Class M3.5 (100) cleanroom that is unidirectional and a nonunidirectional Class M3.5 (100) cleanroom of the same dimensions, the equations appear fundamentally flawed.

Assuming an entrance plane of 75 m²:

Consequently, the nonunidirectional cleanroom would yield substantially more sample locations than the unidirectional cleanroom. The equation for unidirectional cleanrooms takes no account of the class of the area, solely the floor area. A Class M1.5, M2.5 or M3.5 unidirectional cleanroom of the same size would give the same number of sample locations.

BS 5295 decreases the maximum floor area per location as cleanliness classes increase, but also increase the number of samples required per location. Our notional 75 m² cleanroom at Class E would only require 8 locations, but with 5 samples per location!

GMP does not however, lay down detailed methods for determining the particulate cleanliness of air.

Incidentally, using only 33 sample locations in a unidirectional cleanroom of 75 m² seems remarkably small, when balanced to the correct velocity, laminar airflows should exhibit parallelism characteristics that deviate from the vertical by no more than 14° (NEBB Procedural Standards for Certified Testing of Cleanrooms – 2nd Edition, Section 5.5). Many pinhole leaks in the filters and filter ceiling grid could be missed, unless regular filter installation leak testing of the filters and the ceiling grid is also undertaken.

Sample Volumes

BS 5295: 1989 stipulates that a full 1cfm sample must be taken at each location, whilst Federal Standard 209E allows the reduction in sample size. A minimum of 0.1 cfm must be sampled, with the proviso that; at the upper limit of the class classification of concern, the sample would be sufficient size to count at least 20 particles.

The following table reproduced from Section 7 of the NEBB ‘Procedural Standards for the Certified Testing of Cleanrooms’ 2nd Edition, 1996, details minimum sample volumes for particle sizes and classes.

*Volume of sampled air shall not be less than 0.00283 m³ of 0.1 ft³.

Federal Standard 209E requires that at least 5 samples be taken from at least 2 different locations.

BS 5295: 1989 stipulates both the number of samples required at each location and the maximum floor area per location. These are both dependent upon the actual cleanliness class. A minimum of 4 locations must be selected.

Referring back to our notional nonunidirectional class M3.5 (100) cleanroom of 75 m²: classification could be undertaken in as little time as 5 minutes when classifying to Federal Standard 209E (33 locations x 6 seconds sampling time at a flow rate of 1cfm). When running a BS 5295 classification to Class E this may take 50 minutes (8 locations (with 5 samples per location) x 60 seconds at a flow rate of 1 cfm).

Clients should decide what they require the particle counts to demonstrate. Federal standard is perhaps better at demonstrating spatial conformity, and BS 5295 may be better at demonstrating temporal conformity (but only for class E and better cleanrooms). However, all standards give minimum numbers of samples per location and minimum numbers of locations. There is nothing to stop a client specifying a test regime well in excess of these minima. Multiple samples and additional locations are advantageous around critical processes.

Particle Sizes

A classification to BS 5295: 1989 should be across the range of particle sizes highlighted in Table 2 of the standard for that classification. A Federal Standard 209E classification is normally performed at one or two particular particle sizes of concern. Often a cleanroom may meet the particular class limit at one particle size, and fail this class limit at other particle sizes. The results can be reported in this manner: passes class M3.5 at 0.3mm and larger.

In BS 5295: 1989, the smallest particle size quoted in the classification table is 0.3 mm and larger. This is suitable for most cleanroom applications; however, latest developments in the semiconductor industry have pushed the particle size of concern down to 0.1 mm and larger and Federal Standard 209E must be used for this purpose. BS 5295: 1989 is considered somewhat dated within the semiconductor and microelectronics industries for this reason. However, more consideration is given to larger particles for lower grade cleanrooms in BS 5295, with 5.0, 10.0 and 25.0 mm and larger particle sizes quoted in Table 2.

Extrapolation to other particle sizes (other than those quoted) in the Federal Standard 209E table is permissible, as long as these particle sizes are within the range of particle sizes quoted for the classification of interest.

Data Treatment

If there are less than 10 sample locations, Federal Standard 209E requires statistical treatment of the raw sampling data to ensure compliance with the standard by applying a 95% upper confidence limit to the overall mean results for each sample location mean. This is effectively a one-tailed student’s t-test to compensate for sampling errors due to the small number of data points and lack of precision.

All the sample location means and the upper confidence limit must fall within a particular class limit to be validated to that class. This is useful if occasional excursions above a class limit are observed at a particular location.

In the British Standard, only the sample location means must comply with the class limits.

Federal Standard 209E also offers a documented way in which to approach the validation of large class 10 and cleaner cleanrooms using a probability-based sequential sampling plan, which can reduce sampling time considerably.