Manganese welding fume, failed extraction and an untested respirator programme

Evaluation

• A personal result of 0.31 mg/m³ against a 0.2 mg/m³ limit means exposure is not adequately controlled. This is a health issue, not a marginal one: manganese fume can cause irreversible neurological damage with symptoms resembling Parkinsonism.
• The extraction is not working as designed: capture velocity at the hood face is half the commissioning specification and the system has never been thoroughly examined and tested. An unverified system cannot be assumed to control anything.
• FFP2 is a low-assurance choice for welding fume; FFP3 is normally appropriate. Storage in an open box contaminates and deforms masks. Three beards out of twelve means at least some wearers cannot achieve a face seal. Untested RPE should be treated as no RPE.
• Welding with the head in the plume puts the breathing zone exactly where extraction should not be.
• Noise at 87 dB(A) is above the 85 dB(A) action value, and peaks of 137 dB(C) are at a level where hearing protection is required.

Prioritised actions

1. Interim protection now: FFP3 with fit testing, a clean-shaven policy or powered respirators for those who cannot shave, working practices that keep the head out of the plume, and restriction of the worst-affected position until extraction is restored.
2. Re-commission the extraction: competent ventilation engineer, hood repositioning or redesign, ductwork and fan check, then verification by measurement.
3. Reduce fume at source: lower-manganese consumables, on-torch extraction, fume guns, enclosure or automated welding where feasible.
4. Noise: hearing protection zones, mandatory protection at the plasma cutter, and audiometry.
5. Health surveillance and repeat exposure monitoring after the engineering changes.

Justification

RPE is the last line of defence and depends on a seal that is easily broken; it is appropriate as an interim measure while engineering controls are restored, not as the answer. Engineering controls at the source and along the path protect every welder, every shift, including those who lift their helmets.

Extraction verification

• Use the commissioning data as the benchmark, then thorough examination and test at least every 14 months (or the shorter national interval) by a competent person, with a written report and defect schedule.
• Weekly user checks: hood position, airflow indicator, damage, unusual noise, recorded in a logbook kept at the hood.
• Any modification, relocation or repair triggers re-commissioning.
• Reason: extraction degrades silently. A slipping fan belt, a blocked duct or a displaced hood can halve capture without anyone noticing; only measurement detects it.

Exposure monitoring

• Personal sampling on representative welders across both shifts, worst-case positions and tasks, including plasma cutting, using documented methods and reporting the detection limit.
• Repeat after the engineering changes and then periodically; results are used to test the control, not just to file.
• Record who was sampled, where, what they were doing and which controls were running, otherwise the result cannot be interpreted or defended.

Health surveillance

• Baseline for all welders: occupational history, respiratory and neurological symptom questionnaire, and lung function.
• Periodic review with referral where symptoms suggest manganese effect (tremor, gait, mood or cognitive change) or respiratory disease.
• Be explicit about the limitation: blood and urine manganese correlate poorly with exposure and are not a reliable screening test; imaging changes are a research tool. Surveillance detects effect; it does not prevent it.

Noise

• Baseline and annual audiometry, a hearing protection zone around plasma cutting and mandatory protection for peak exposures.

Limitation

Health surveillance cannot compensate for exposure above the limit. It is an early-warning system, not a control measure.

What the gap means

• Three years without thorough examination means the real performance of the extraction is unknown for that period. The 0.31 mg/m³ result may represent the whole period or only recent degradation; either way, exposure above the limit cannot be excluded for the workforce.
• Records should be reviewed to see whether anyone has reported relevant symptoms, and all welders should be offered health surveillance promptly.
• Some jurisdictions require notification of occupational disease or of exposure incidents; check the national requirement and act on it.

To senior management

• Say plainly that the control failed, that exposure exceeded the limit and that the verification regime that should have caught it was not in place. Do not frame it as a paperwork gap.
• Set out the consequences: health risk to named individuals, potential enforcement, compensation and insurance implications, and the reputational cost. Balance that against the modest cost of remediation.
• Ask for a dated action plan with a named owner, capital for extraction improvements, and a standing board report until closure.

To the workforce

• What was found, what it means for them, what is being done and by when, and how to raise concerns - in a form they understand.
• Offer health surveillance and explain its purpose and confidentiality.
• Consult safety representatives. If management does not explain, rumour will.

Reasoning

Concealment destroys trust and increases legal risk. The organisation needs a defensible record showing it identified the problem and acted promptly once it knew, rather than one showing it sat on the finding.