Two cleanroom fire safety protocols accepted by NFPA 318
10/01/2001
Meg Villeneuve
QUINCY, MA—The National Fire Protection Association (NFPA) committee 318 recently adopted two protocols for testing non-fire propagating materials used to manufacture cleanroom tools and equipment—Factory Mutual Research 4910 Protocol and Undewriter's Laboratories (UL) 2360.
With the adoption of the two protocols, the association hopes to see a decrease in the number of cleanroom fires.
 Flammability test apparatus for cleanroom plastics being used within Factory Mutual Research's lab. |
"NFPA 318 has accepted the concept of non-fire propagating materials, or materials listed for use without internal fire detection and suppression," says Steve Zenofsky of FM Global, the parent company of Factory Mutual Research. "NFPA 318 also has indicated that the test methods of the Factory Mutual Research 4910 Protocol and UL 2360 provide the means to obtain such materials," says Zenofsky.
Since the adoption of its protocol, UL has experienced an increase in testing requests. To date, the company has worked with seven different companies, of which three are classified to UL 2360. "[The companies] manufacture plastic materials, which are used to fabricate tools used in cleanrooms," says Bob Backstrom, engineering group leader at Underwriters Laboratories.
The NFPA 318 committee was formed in 1987. The association's goal was to establish guidelines and standards for cleanrooms, including but not limited to wafer fabs. "However, as time went on it was necessary to limit the scope to wafer fabs because of the complexity of cleanrooms," says David Quadrini, chairman of the NFPA 318 committee.
While, Factory Mutual Research can't predict how many fires will be avoided, the company did say it expects to see a decrease in damages, which should lead to a decrease in downtime when fire does occur.
"The industry now has a common protocol that hinges on the use of the parallel panel test, regardless of the small scale test used to perform the initial qualification of the material," Zenofsky adds.
The parallel panel test is a large-scale test where two vertical parallel panels separated by 1 ft. are exposed to an ignition source that is located at the bottom between the panels. Each panel is 8 ft. long, 2 ft. wide and 0.25-1 inches thick. The ignition source consists of a 2 ft. long, 1 ft. wide and 1 ft. high propane sand burner. The propane gas flow is adjusted to provide a heat release rate of 60 kW. During the test measurements are made for: release rates of material vapors (mass loss rate), chemical and convective heats, and smoke; heat fluxes to the panels; and flame height.
According to FM Global, there is no such thing as a small disaster within a cleanroom. Any type of fire, small or large, can shut a line down for months and cause cleanroom operators to replace and reconfigure tools and equipment. The extent of damage caused by fire is usually dependent upon what technologies are being processed. "The down time from a fire depends on the processing—a fire that occurs in a cleanroom processing smaller line widths may take up to six months to recover from. It also depends on what tools are involved," says Heron Peterkin, of Factory Mutual Research.
"Over the last twenty years," Peterkin adds, "there have been over 400 reported fires/chemical spills within cleanrooms worldwide acknowledged incidents by FM Global, Industrial Risk Insurers and Munich Reinsurance Company."
Factory Mutual Research began testing non-propagating plastics in 1997. The company says it thought the new protocol would be ideal in semiconductor fabs, however, it found that the semiconductor industry is not quick to adopt the new materials or processes. "We learned that we had to prove ourselves in the industry," says Peterkin. "But that was four years ago."