How DOCA tackled dust cross-sensitivity in cleanroom compressed air

Australian pharmaceutical manufacturers in Sydney and Melbourne, semiconductor fabs in Adelaide, and hospital systems across Brisbane depend on ISO 14644-classified cleanrooms to protect critical processes. Compressed air feeding these spaces must stay free from oil contamination, yet optical instruments designed for hydrocarbon detection frequently confuse airborne dust with oil, raising false alarms or missing genuine events.

The DOCA Project, backed by European Union research funding, developed an online optical sensor capable of detecting oil in compressed air in liquid, aerosol, and vapour phases. Cross-sensitivity to inert particulates became one of the project's most demanding technical hurdles, particularly for industries where false positives trigger expensive shutdowns or compromise sterility.

This article examines how the consortium investigated dust interference, the laboratory framework behind the work, and what the results mean for facilities operating under Australian standards including AS/NZS ISO 14644 and TGA manufacturing guidance.

Why dust causes false positives in optical oil sensors

Optical instruments identify hydrocarbons by analysing scattering, fluorescence, or absorption of light. Dust particles from skin cells, textile fibres, concrete, and bushfire ash common across eastern Australian cities produce their own scattering signatures that overlap with the response from microscopic oil droplets, especially when those droplets fall below one micrometre.

Cleanroom ventilation compounds the challenge. HEPA filtration traps most particulates above 0.3 micrometres, yet airflow disturbances during filter changes or process upsets release fine particles into supply lines. Facilities hit by seasonal dust storms or summer bushfire smoke events face inlet air carrying mineral fragments and combustion residue that single-channel instruments cannot reject.

Without a sound discrimination strategy, the sensor would generate nuisance trips inside pharmaceutical isolators or allow real oil ingress to slip past unnoticed in a semiconductor lithography bay.

Laboratory setup and dust loading methodology

Cross-sensitivity trials took place in dedicated aerosol chambers fitted with calibrated particle generators, condensation nuclei counters, and laser particle spectrometers serving as reference instruments. The team introduced known concentrations of ISO 12103 A1 Arizona test dust alongside background aerosols typical of cleanroom air, then injected trace oil quantities in controlled pulses.

Two variables were logged simultaneously: the magnitude and time profile of the DOCA sensor's optical response, and an independent particle count. Trials spanned multiple humidity bands because moisture causes hygroscopic particles to swell, shifting their optical behaviour. Engineers also tested particulates drawn from Australian sources, including red desert clay from South Australian mining regions and bushfire-derived charcoal aerosols, confirming resilience to dust profiles that diverge from European baselines.

How the DOCA sensor separates dust from oil

Rather than depending on a single measurement channel, the sensor pairs laser-induced fluorescence with scattered-light analysis. Algorithms compare fluorescent intensity against the scattering cross-section, because mineral particles produce predominantly scattering signatures while hydrocarbons generate strong fluorescence at defined wavelengths. This ratio behaves as an effective discriminator.

Key outcomes from the testing programme included:

• Particulate-only events triggered less than 0.3 percent fluorescent response relative to oil at equivalent mass loading
• Mixed aerosols containing dust and oil were classified correctly in 94 percent of pulse trials
• Sensor drift stayed inside specification after 200 hours under 5 milligrams per cubic metre dust loading
• Humidity compensation restored discrimination accuracy above 85 percent relative humidity

False-positive rates dropped by roughly an order of magnitude compared with single-channel instruments under identical conditions, while true-positive detection of oil above 0.1 milligrams per cubic metre held steady.

Field trials at pharmaceutical and electronics sites

After laboratory success, the prototype was installed at collaborating European facilities and at independent sites in Australia. A Sydney-area contract manufacturer running TGA-licensed aseptic filling lines operated the sensor alongside existing compressed air monitors for three months. Site engineers reported that the DOCA unit caught a compressor fault involving polymeric oil carryover from a separator element, a real event the legacy particulate counter had missed entirely.

In Adelaide, a precision electronics assembler used the prototype to monitor supply lines feeding wave-soldering and conformal coating stations. Cross-sensitivity to solder fumes surfaced as an additional consideration and was resolved through a software filter update released during final testing. The deployment confirmed that hydrocarbon contamination can be reliably distinguished from inert particulates in busy production environments, well beyond the constraints of a clean laboratory.

Alignment with Australian cleanroom and pharmaceutical rules

Australia's regulatory framework for cleanroom air quality draws on international standards. AS/NZS ISO 14644-1 and 14644-2 define classification limits, while the TGA's PE009 guidance specifies compressed air quality for sterile pharmaceutical manufacturing. Food Standards Australia New Zealand adds further contamination limits on air contacting edible products.

The DOCA sensor supports compliance documentation through continuous logged measurements with full audit trails. Operators can configure alarm thresholds to match facility-specific risk assessments, whether protecting an Adelaide semiconductor fab, a Melbourne teaching hospital's surgical air supply, or a Brisbane cosmetics production line.

Considerations when selecting a compressed air sensor

Engineers comparing compressed air monitoring options should weigh several criteria:

• Request cross-sensitivity data under realistic dust loading, not oil-only specifications
• Verify behaviour under local particulates, including bushfire smoke and construction dust
• Confirm the unit supports multiple detection channels with ratio-based discrimination
• Check integration capability with building management systems and audit logging
• Ask for evidence of field validation rather than laboratory benchmarks alone

Choosing the right instrument reduces nuisance alarms, prevents unnecessary production halts, and catches genuine contamination before products are affected.

For facilities interested in learning more about the DOCA sensor or arranging a demonstration, the project team welcomes direct inquiries through their contact page.