On March 16, 2016, a 29-year-old postdoctoral student at the University of Hawaii, Manoa, was working alone in the lab. Thea Ekins-Coward was bleeding off hydrogen, carbon dioxide, and oxygen from pressurized tanks to feed to experimental cells. The project had been ongoing since 2008 without a mishap, and the lab had passed its most recent safety inspection in January 2016—but in March something went wrong. An explosion destroyed the lab and seriously injured Ekins-Coward, who lost an arm and suffered burns to her face and temporary hearing loss in the incident.
 |
In July 2016, The University of California Center for Laboratory Safety completed an investigation of the incident and issued a 73-page technical analysis, and another 38-page set of recommendations for improving lab safety at the University.
So what went wrong, in a process that had run safely for 8 years? Here’s what the University of California Center for Laboratory Safety found.
Proximate Causes and Systemic Failures
The immediate cause of the accident was a discharge of static electricity—probably from the lab worker’s skin, when she touched a 13-gallon pressurized tank containing hydrogen/oxygen gas. The tank was not grounded, so the charge instead went to ground through the tank’s pressure gauge, instantly igniting the highly flammable mixture in the tank. But, leading up to that explosive moment was a series of systemic failures that included:
Failure to manage change. In February, the procedure for the experiment changed. The PI and Ekins-Coward had decided to scale up their process, by premixing the 3 gases in a 13-gallon gas storage tank. They did not perform a risk analysis for the scaling-up procedure. The tank was filled with varying mixtures of hydrogen, oxygen, and carbon dioxide—but all of the mixtures were explosive. The PI and Ekins-Coward reviewed the experiment weekly, but apparently assumed that the process was safe because they stayed well below the tank’s maximum rated pressure.
They could not have been more wrong. The tank itself was not designed for the use they put it to, and neither was the pressure valve. The operating instructions for the tank specifically state “Do not use this tank to store or disperse liquids or flammable gases,” warning that “Failure to comply with these guidelines could result in personal injury and/or property damage due to air tank failure.”
But this and other problems were never identified because of a:
Failure to perform risk assessment. If the PI and Ekins-Coward had performed a risk assessment, they might have identified safety issues that included not only the fact that the tank itself was unsuited to the use they made of it, but also that:
- The tank was never grounded. The 13-gallon tank was filled from tanks in two different areas before being carried to the experimental location. At no point was the tank—which was filled with a flammable gas mixture—grounded.
- The digital pressure gauge installed on the tank was not intrinsically safe. In other words, it was not safe for use in a flammable environment.
- The ball valve used to add gas to the tank and to withdraw samples of gas from the tank was stainless steel. Stainless steel valves can ignite explosive gases, either as a result of friction from the high-velocity flow of flammable gases through the valve, or through a buildup of static electricity resulting from friction with the Teflon® tank seals.
The issues with static electricity and the tank showed themselves very soon after the tank was placed in service. Ekins-Coward reported to the PI in February that sometimes when she touched the tank, it shocked her; other lab workers noticed and reported the same thing.
But no one inquired further about the shocking tank, because the lab was also guilty of:
Failure to investigate close calls and near misses. A tank of flammable gas that is constantly shocking people should be considered a “close call,” but no one in the lab investigated the problem. Similarly, the lab failed to investigate an incident that occurred just the day before the explosion, when Ekins-Coward heard a “cracking” sound inside the tank. When she opened the vessel, she noticed a burning smell. There had been a small deflagration, but that day, the mixture did not explode. Ekins-Coward reported the incident to the PI, who told her not to use the vessel again. But no investigation was initiated, and the prohibition was not enforced: the pressure vessel was in use in the exact same way the very next day, when it did explode.
The lab had at least three opportunities to identify and prevent the explosion hazard that their procedural change created, and they missed all three.
Need more information on anticipating hazards and preventing incidents? Safety.BLR.com® will put you ahead of the game.