Solvent vapour is three to four times heavier than air and settles where painters kneel. Ventilation rate and gas monitoring are the method, not the paperwork.
Coating work is judged on the film. It is also, occasionally, judged in a coroner's report, and the two failure modes share a cause: somebody assumed the air in the space was the air they were breathing in the yard.
Vapour density decides where the danger sits
Solvent vapour from the products used in heavy-duty coatings is heavier than air. Xylene, the workhorse thinner in epoxy and polyurethane systems, has a vapour density around 3.7 relative to air. Toluene is about 3.1, and the light aliphatic fractions used in some primers are lower but still above 1. Vapour from an open bucket or a freshly coated floor does not disperse upward and away. It settles, flows along the bottom of the tank, and accumulates in pits, sumps, bilges and the lower third of any enclosed volume.
This is why air monitoring at the manhole reads acceptable while the base of the tank is not. It is also why the person who enters to retrieve a dropped tool is exposed to a concentration the person standing at the opening never experiences.
Ventilation is a calculated quantity
The ventilation requirement is not a matter of putting a fan at the opening. It follows from three numbers: the evaporation rate of the product in use, its lower explosive limit, and its vapour density. Those three are on the safety data sheet, and the calculation is standard practice in every coating specification that addresses enclosed spaces. What the calculation produces is usually uncomfortable, because it is measured in air changes per minute rather than per hour.
Two consequences follow. First, the extract must be taken from the low point, where the vapour is, not from the high point where the air is easiest to move. Second, supply and extract need to be arranged so the working zone is swept rather than merely diluted; a single fan blowing into a tank produces a jet and a dead zone, and the dead zone is usually where the painter is kneeling.
The numbers that stop work
Gas monitoring for entry and work in a confined space is commonly set at ten percent of the lower explosive limit, with work halted well before twenty-five percent. Oxygen is monitored in the 19.5 to 23.5 percent band. For solvent exposure itself, the eight-hour occupational limit for xylene sits around one hundred parts per million, and odour is no guide: the smell disappears long before the concentration becomes harmless, and at higher concentrations the sense of smell is dulled further.
A monitor is not a substitute for the assessment. Readings taken at the opening, at head height and at the low point, before entry and at intervals during work, describe the space. A single reading at the manhole describes the manhole.
The practices that keep people alive
- Never enter without a permit, an attendant at the opening in continuous communication, and a rescue plan that does not rely on the attendant entering
- Continuous forced ventilation for the whole working period, not only during application; freshly coated surfaces keep releasing solvent for hours
- Keep solvent, rags and open containers out of the space; wipe-down rags in a closed metal container
- Monitor at the low point as well as at the working level, and log the readings with times
- Respiratory protection selected for the actual concentration, with fit testing, not a nuisance dust mask
- Bond and earth equipment and containers where flammable atmospheres are possible
None of this is unusual, and all of it is written into the specifications that clients audit. The recurring failure is not ignorance of the rules but the small exception: ten minutes without ventilation to finish an edge, one entry without a monitor to check a defect. Solvent vapour does not negotiate over ten minutes.