At approximately 9:14 p.m., on March 4, 1996, a DOT-E- 8059, 4,500-psi, fiberglass- wrapped composite aluminum breathing air cylinder ruptured without warning while stored in an equipment compartment on a reserve engine (Engine 1-7) of the Humboldt No. 1 Fire Protection District in Eureka, California. Engine 1-7 suffered approximately $5,000 worth of damage; and Engine 1-4, adjacent to Engine 1-7, sustained $600 in damages, predominantly minor dents and scratches in the paint. Engine 1-7, a 1,000-gpm pumper with equipment compartments, hosebed, and crew cab with jump seats in conventional arrangement, housed the cylinders in two left-side rear compartments, aft of the pump panel. Fortunately, no one was injured in the incident.

Air cylinders were found six and 10 feet away, and there were indications that they had been airborne that distance. The compartment doors were severely damaged; the rear door was completely separated from the engine body and lay on the floor, curled into a “U” shape. The overpressure caused the diamond-plate top of the compartment to expand up and outward, twisting it and pulling it off the bolts holding it in place. Also, the rear seam of the compartment floor spread open, leaving a one-half-inch gap between the compartment floor and the back wall. Directly behind the location of the ruptured bottle, on the compartment`s back wall, was a dent about six inches in diameter and one inch deep, resulting from a direct hit from the bottle.

The force of the rupture caused the cylinder–a half-hour-capacity fully-wrapped composite cylinder (manufactured by EFI Corporation) and part of an SCBA manufactured by International Safety Instruments (ISI)–to split open and its walls to violently shred the plywood backing board in the compartment. The ruptured cylinder was found on the apparatus room floor, in front of its holding bracket and almost directly beneath the damaged cylinder tube door on the other engine. Several of the cylinder valves showed serious damage after the incident and were immediately determined to be unfit for service.

At the time of the rupture, the on-duty firefighters were in the station day room. The rupture shook the station and created a few moments of concern. Captain Dave Gibbs investigated and ordered firefighters out of the apparatus room. After 30 minutes and following a telephone conference with Chief Robert Heald, Gibbs donned full turnouts, a helmet, and a flack jacket and entered the apparatus room. He opened the cylinder valves on the involved cylinders to bleed off the pressure. He left the area and waited for the cylinders to drain their contents before allowing his crew to enter the area. Meanwhile Gibbs radioed to dispatch that his station was out of service.

After another 30 minutes, Gibbs and his crew entered the apparatus room to survey the damage.


The sales manager and chief engineer of EFI Corporation, representatives of CAL-OSHA and the Department of Transportation (DOT), marketing and sales managers from ISI, safety experts, insurance adjusters and inspectors, breathing apparatus users from several disciplines including the U.S. Coast Guard and the U.S. Department of Energy, representatives of large and small fire departments, and the news media all visited the site beginning the day after the incident.

ISI notified EFI Corporation of the cylinder failure the same evening of the incident. Since the cause of the cylinder`s failure was unknown at the time, EFI Corporation, issued a quarantine for the 199 other cylinders manufactured in the same lot. The quarantine was rescinded when the investigation subsequently revealed that the cylinder rupture was the result of the cylinder`s coming in contact with a highly acidic cleaning fluid.

EFI Corporation contracted with Failure Analysis Associates, Inc. (FaAA) to investigate the incident. The failed cylinder and the cylinder storage rack holding the cylinder at the time it failed were transported to FaAA, which undertook an investigation that included information gathering, visual examination, optical and scanning electron microscopy, chemical analysis, stress corrosion cracking testing, and chemical exposure testing.

The investigation revealed that the Humboldt Fire Department SCBA cylinders had been used for a training exercise. The following sequence of events was reconstructed:

On February 27, the SCBA cylinders were used during a multiagency fire training exercise for the officers of Humboldt County Sheriff`s Office and several correctional facilities. Nine fire personnel from Humboldt No. 1 Fire Protection District served as technical advisors and safety support personnel.

On the evening of February 27, a deputy sheriff with the Humboldt County Sheriff`s Department transported the cylinders in his flatbed trailer to the Eureka Fire Department for refilling. The cylinders were in the trailer about 30 minutes. One fiberglass cylinder and four to five metal cylinders were on the floor of the trailer; other cylinders were on top of them.

During loading or transport, a 12-ounce spray bottle on the floor of the trailer, containing about four ounces of aluminum cleaning fluid, broke and spilled its contents.

After delivering the SCBA cylinders to the Eureka Fire Department, the deputy emptied and washed his trailer.

About two hours later, the deputy retrieved the refilled bottles from the Eureka Fire Department and returned them to the Humboldt Fire Department.

The potential exposure of a fiberglass SCBA cylinder to a corrosive fluid was a key piece of information in the failure investigation.

FaAA reported the following after its investigation:

The macroscopic and microscopic fracture surface is characteristic of stress corrosion cracking in fiberglass composites.

The results of the chemical analysis indicate that fiberglass samples of the failed cylinder contained 2-butoxyethanol, a chemical constituent of a highly acidic fluid that chemically attacks glass. The fluid is a commercially available aluminum cleaner used also for chrome and other metal surfaces that contains among other ingredients hydrofluoric acid, phosphoric acid, sulfuric acid, ethylene glycol, monobutyl ether, and nonylphenoxypoly (ethyleneoxy) ethanol.

The rapid failure of specimens in stress corrosion cracking experiments provides qualitative verification that the fluid caused stress corrosion cracking of the accident cylinder. The appearance of fracture surfaces on the accident cylinder was very similar to the stress corrosion cracking observed on the sample exposed to the acidic fluid.

Personal interviews indicated that during the training exercise on February 27, a fully wrapped composite cylinder was exposed to an aluminum cleaning fluid, which was likely the one in the container on the trailer. The cylinder ruptured about six days after exposure to the fluid.

The FaAA report noted: The cylinder failure “… is consistent with environmental stress-assisted cracking of the fiberglass composite (overwrap). Glass fibers failed due to the combined action of an acidic chemical environment and the stress caused by the internal pressure. This combination of chemical attack and stress acting over several days led to the failure of the cylinder.”


The National Institute for Occupational Safety and Health (NIOSH) has issued an advisory that includes the following precautions:

End-users of all composite fiberglass-wrapped cylinders could experience similar cylinder failure if they fail to take precautions to ensure that acidic cleaning fluids or other acidic chemicals do not come in contact with such cylinders. Besides contact with acidic cleaning agents in common use, NIOSH noted, fiberglass-wrapped cylinders could also be exposed to acidic chemicals in chemical plants, warehouses, and other haz-mat sites.

Fiberglass-wrapped composite cylinders should only be used, handled, cleaned, maintained, and transported by individuals who have been made aware of the need to avoid exposing such cylinders to acidic materials.

If a fiberglass-wrapped cylinder is suspected to have come in contact with an acidic chemical, the cylinder should be depressurized immediately and removed from service. The SCBA manufacturer should then be contacted for further instructions.

Additional information is available from Robert S. Frankle, managing engineer, Failure Analysis Associates, 149 Commonwealth Drive, Menlo Park, CA 94025. n


Report, Chief Robert Heald, Humboldt No. 1 Fire Protection District, Eureka, California.

“Investigation of the Failure of an SCBA Cylinder,” Robert S. Frankle, Harry F. Wachob, Ph.D.; Failure Analysis Associates, Inc., Menlo Park, California, July 2, 1996.

Letter, David C. Woodward, Vice President & General Manager, EFI Corporation, Fremont, California, July 19, 1996; EFI Corporation, “Incident Report Conclusion,” March 21, 1996.

“NIOSH Respirator Users Notice,” National Institute for Occupational Safety and Health, Morgantown, West Virginia, July 30, 1996.

Click here to enlarge image

Click here to enlarge image

Click here to enlarge image

Click here to enlarge image

(Top left, top right, bottom left) The failed cylinder caused significant damage to the engine on which it was stored and minor damage to the adjacent engine. The damage caused by the failed cylinder (on the floor between the two parked engines) included blowing off the door of the compartment in which the cylinder was stored, blowing open and deforming an adjacent storage compartment door, expelling the majority of the contents of the compartment, and dismounting and throwing to the floor 10 feet of large hard-suction hose. (Bottom right) Close-up of damaged cylinder. The cause of the cylinder`s failure was a previous contact with a highly acidic fluid that chemically attacks fiberglass. (Photos by David Gibbs.)

No posts to display