Cleanroom environmental monitoring

Environmental monitoring in a cleanroom answers two questions that are usually run together and are not the same question at all. How much particulate is in the air, and is any of it alive.

What an optical counter actually reports

An airborne particle counter draws a metered flow — commonly 28.3 litres per minute — through a chamber where a laser crosses the stream. A particle passing through scatters light; a detector off-axis collects part of that scattered light and produces a voltage pulse. Each pulse is one particle, so the count rate divided by the sampled volume gives concentration. The pulse height is sorted into size bins.

The important qualification is what “size” means here. Scattered intensity depends on the particle’s diameter, but also on its refractive index, its absorption and its shape. The instrument is calibrated with polystyrene latex spheres of known diameter and refractive index, so every result is an equivalent optical diameter: the size of the PSL sphere that would have scattered this much light. A dark, absorbing particle scatters less and is reported smaller than it is; a fibre gives a pulse that depends on its orientation as it crosses the beam. Two counters from different makers can therefore disagree on the same air and both be within specification.

Where the limit comes from

ISO 14644-1 sets class limits as concentration per cubic metre at defined size thresholds, and EU GMP Annex 1 maps its grades onto them. At the cleanest grades the constraint stops being the instrument and becomes counting statistics. Particle arrivals are Poisson-distributed, so at a limit of a few thousand per cubic metre for ≥0.5 µm — and far fewer at ≥5 µm — a short sample collects so few counts that the confidence interval swallows the measurement. A single stray count can move a result across a limit. ISO 14644-1:2015 dropped the ≥5 µm threshold at ISO 5 for precisely this reason, while Annex 1 retained a large-particle limit in grade A as a trend indicator rather than a pass/fail statistic. Neither position is careless; they are answers to the same statistical problem with different aims.

The gap the count does not close

None of this speaks to sterility. A particle counter registers a skin flake, a fibre and a viable Bacillus spore identically, and a room can sit comfortably inside its class while carrying an organism that will spoil a batch.

Viable monitoring is therefore a separate measurement, and a weaker one. Impaction samplers accelerate air through slits or orifices onto an agar surface, relying on inertia: above a cut-off size set by jet velocity and geometry, particles cannot follow the deflected streamline and strike the medium. That same velocity dehydrates and mechanically stresses the organism, so recovery falls — most for fastidious species, and progressively over a long sample as the agar dries. What grows is then only what will grow on that medium, at that temperature, in that incubation time; viable-but-non-culturable cells and slow-growing environmental organisms are simply absent from the result.

The honest description of a monitoring programme is that it measures a proxy for a proxy, with a lag of days for the microbiological half. Its value lies in trend rather than in any single reading — which is why Annex 1 (2022) frames monitoring as evidence within a contamination control strategy, not as a test the room passes.