While filling a sample cylinder with compressed hydrogen gas, a quick-disconnect coupler fitting came loose within a stainless steel laboratory hood, allowing a small purge of the hydrogen gas to escape directly into the hood through ~1/4-inch Tygon tubing. The stainless steel quick-disconnect fitting struck the stainless steel bottom of the laboratory hood and the hydrogen gas caught fire. It is not known what caused the hydrogen gas to catch fire. The most likely sources of a spark was from metal-to-metal contact of the quick-disconnect fitting with the laboratory hood floor, or the discharge of static electrical charge generated by flow of hydrogen gas through Tygon tubing. The resultant narrow jet of fire, directed toward the left side of the laboratory hood, extinguished itself within a few seconds of when the operator closed an upstream valve that was supplying the compressed hydrogen gas to the sample cylinder.

The operator was not injured and the only apparent damage was from minor melting of a short length of Tygon tubing. In addition to several 8-inch to 24-inch lengths of Tygon tubing and a plastic beaker, there was a 500 milliliter (ml) glass bottle containing approximately 300 ml of hydrazine-hydrate stored in the laboratory hood. This flammable chemical is periodically used as an additive for maintaining water chemistry. No additional fire extinguishing action was necessary inside the hood, but as a precautionary measure, the operator used a low-pressure demineralized water (LDW) tap within the hood to rinse/cool surfaces within the hood. Since there was no apparent remaining fire extinguishing action required, fire department assistance was not requested.

The operator immediately performed a walk down of the laboratory hood exhaust ventilation system to ensure no fire had occurred in the HVAC duct work, downstream of the laboratory hood. Some smoke was evident downstream of the laboratory HEPA filter, on the third floor mezzanine room where the laboratory hood exhaust fan is located; otherwise, all systems visible to the operator were found to be normal. Follow-up radiological surveys showed there was no spread of radiological contamination due to the operator having exited the room without surveying himself.

The direct and root causes for this event were a design problem. The quick-disconnect fitting on the chemical addition system became unintentionally disconnected. The release of hydrogen gas into the hood enclosure, combined with a spark generated from either the metal fitting striking the metal floor of the hood enclosure or discharge of static electricity from flow of the gas through non-metallic Tygon tubing created a fire hazard. To eliminate the hazard completely, two sets of quick-disconnect fittings (of the type that became unintentionally disconnected) and use of Tygon tubing were eliminated and replaced with a hard-piped system. The small quantity of hydrogen gas being purged through the sample cylinder was also redirected such that it was released near the top of the hood and immediately adjacent to the hood exhaust duct work. A locking device was then provided for the remaining two sample cylinder quick-disconnect fittings to prevent them from separating without first removing the locking device.