Woodland Pulp contests the $794,456 OSHA penalty over the fatal January H2S release
Woodland Pulp LLC has contested the citations and $794,456 in penalties that the US Occupational Safety and Health Administration proposed after the fatal January release of hydrogen sulfide at its Baileyville, Maine mill. The contest, reported in early September, moves the case to the independent US Occupational Safety and Health Review Commission for adjudication.
According to OSHA’s 30 July 2026 news release, the agency issued four willful and eight serious citations. On 27 January 2026, workers performing a cold-weather shutdown were exposed to high concentrations of uncontrolled hydrogen sulfide; two employees died and ten others were exposed. OSHA’s investigation also covered a separate 7 March 2026 chemical fire that began after roughly 4,400 gallons of 50% hydrogen peroxide were released and mixed with wood pulp that had accumulated on the ground. The willful citations — OSHA’s highest severity classification — covered exposing workers to lethal H2S levels without proper alarms or breathing apparatus, and allowing overflow pulp to build up in storage areas.
A contest is a routine legal step, not a finding; the citations are allegations until the Review Commission rules, and nothing here presumes the outcome. The US Chemical Safety Board is separately investigating the same incident and issued an investigation update on 14 July. What follows is drawn from the public enforcement record and offered as lessons for study and revalidation teams, not as a verdict on the facility.
Detection layers and the modes a study forgets
The recurring lesson in H2S fatalities is the distance between what a process hazard analysis credits and what is physically present when the release happens. H2S is the reference case for why that distance is dangerous. It is immediately dangerous to life or health at 100 ppm; at the concentrations that matter, olfactory fatigue removes the human nose as a sensor within minutes, so “someone will smell it and leave” is not a safeguard. When a study assigns a protection layer of “gas detection plus escape,” the credit is only real if fixed or personal detection is actually installed, maintained, calibrated and alarmed in the specific area — and in the specific operating mode — where the release is credible.
That operating-mode point is the sharper one. Steady-state operation is the easy part of a HAZOP; the hazardous surprises cluster in the transients. A cold-weather shutdown is exactly such a non-routine mode: process conditions differ from normal, equipment may be drained or isolated in unusual line-ups, and staffing and supervision assumptions can quietly diverge from what the study imagined. A PHA that examines only normal running can leave a shutdown scenario un-nodes and its detection assumptions untested.
Transferable questions for facilitators
Two questions carry across from this case regardless of how the citations are ultimately resolved. First, does the study explicitly address startup, shutdown, and other transient and non-routine modes — with their own deviations, causes and safeguards — rather than treating them as variations on steady-state operation? Second, for every safeguard recorded as gas detection or personal monitoring, is that detection genuinely in place and functional in the area and mode where the hazardous scenario unfolds, and does the mechanical-integrity and management-of-change record show it staying that way?
The second incident in the same enforcement action — a strong-oxidiser release meeting accumulated combustible material — is a reminder that housekeeping and material-accumulation scenarios belong in the hazard review too, not only in the general-duty safety programme.