Severity
Incident
Leak
Yes
Ignition
Yes
Ignition Source
(Spontaneous combustion)

While performing hydrogen gas release experimentation by thermally reacting a slurry of ammonia borane powder in silicone oil in a plug flow reactor, a discharge port on the test reactor became loose. A foaming white product was leaking from the fitting and discharging in the direction of the heat tape and insulation (back over the reactor). As a result, hot, reacting slurry flowed out of the port and was exposed to air. In the presence of oxygen, the slurry ignited, producing a green flame. A small green flame was noticed at the leak site and flaming product dripped onto the hood deck surface. The flame at the end of the reactor was ~10-12 inches tall at the highest point. The flame on the deck was ~4-6 inches in height.

The incident occurred behind the lowered sashes in the test hood. No one was injured. The hood was not damaged and the test equipment appears to be fully functional, except for a few thermocouple cables, heat tape, controller power cable and a section of polymer purge tubing that was in the path of the flame.

The building fire alarm was not pulled as the cognizant scientists determined the incident to be isolated, contained and that the fire could be extinguished with the equipment available. One researcher left the area to call for an emergency response while another researcher located the building fire extinguisher (<10ft away) and extinguished the fire with two short bursts. The sash was raised ~12 inches to aim the fire extinguisher discharge at the flame and was pulled back down after successfully suffocating the flame. The flames were extinguished within 15-20 seconds of ignition.

Incident Date
Sep 19, 2013
Setting
Equipment
  • Ventilation System
  • Glove Box/Fume Hood
  • Laboratory Equipment
  • Glassware
  • Motive Power Systems
  • Pump
Damage and Injuries
When Incident Discovered
Lessons Learned

The researcher's failure to pull the fire alarm was an oversight of required facility practice. The alarm should have been triggered in consideration of the potential for greater harm to personnel and facilities.

Hot, reacting ammonia borane produces hydrogen as well as other pyrophoric impurities. Reactions should be carried out in inert atmospheres or purged with inert gases. Furthermore, efforts should be made to prevent oxygen from coming in contact with the material while hot.

As a consequence of this incident, an updated procedure was put in place to check critical fittings before each subsequent test and to purge the apparatus with argon in the area surrounding the reactor/fittings.