The Virginia Beach (Va.) Fire Department recently conducted an investigation into a fire and catastrophic failure of a portable oxygen cylinder.

A fire occurred Feb. 5 inside of a medical “jump” bag aboard a department ladder truck. An oxygen cylinder contained in the bag soon ruptured violently.

A number of ignition sources were examined, and attempts were made to establish an ignition sequence. Attention initially focused on a 3-volt penlight found melted to the underside of the aluminum oxygen cylinder. It was theorized that the penlight's push-to-operate clasp had been depressed through the soft-side bag, and that a crack in the lamp cover or a spark from the clasp may have been the source of ignition.

Investigation also found the T-handle adjuster on the regulator was bent. The presence of a steel adjustable wrench in the bag suggested the non-sparking plastic wrenches that had been issued were unable to provide firefighters with sufficient torque to tighten down on a leaking seal between the cylinder and regulator, and a steel wrench was being used instead to torque the T-handle tighter, thus bending the “t.”

Additional investigation identified the root of the problem. There are two types of Compressed Gas Association 870 seals commonly found in use with medical oxygen cylinders. One type is a rubber-style sealing gasket that can be reused and retightened. The other is a plastic crush gasket that seals well only once. The department had been sent a new-style “dust cover” to protect the valve, accompanied by new plastic seals. Firefighters had begun using the new seals — including reusing and retightening them — just as they had the rubber-style seals for years.

A forensic engineering report noted that “experience indicates that the differences between the multiple-use sealing washers and the single-use crush gaskets are generally not known or appreciated by most oxygen users.” The Virginia Beach Fire Department was no exception.

The ignition source was found not to be the penlight closure as first thought. The CGA 870 crush gasket was to blame here, as well. The report states, “The probable ignition mechanism … was associated with the heat developed by oxygen leaking across the cga 870 gasket in a weeping manner. This ignition mechanism is referred to as flow friction and has been implicated in several previous oxygen-related fires.”

The conditions necessary for ignition by this mechanism were exacerbated by the use of the adjustable steel wrench to open and close the cylinder valve, which probably led to an overload of the valve seat and subsequent leakage.

The complete incident report is available by contacting Bttn. Chief Stephen Miles, Virginia Beach Fire Department, 2408 Courthouse Drive, Municipal Center-Building 21, Virginia Beach, Va. 23456-9065; or e-mail smiles@vbgov.com.