Chemical Leaks at West Virginia DuPont Plant Appear Unrelated

Three recent chemical leaks at a West Virginia DuPont plant, one of which killed a worker, don’t appear to be related, a federal investigator said.

U.S. Chemical Safety Board investigator Johnnie Banks promised a thorough investigation into the leaks, which prompted the plant’s temporary shutdown over the weekend.

The release of about 1,900 pounds of methyl chloride went unnoticed for five days before the company reported it.

A worker died after being exposed to phosgene — a chemical used as a choking agent during World War I that now is used in pesticides and plastics. That worker was exposed last Saturday, the same day the plant reported that less than 20 pounds of sulfuric acid had leaked from its spent acid recovery process.

The plant is on a 600-acre site in Belle and employs about 400 workers with an additional 250 contractors.

The leaks “were not in close proximity to one other where there could be a sense of something happening in one building transferring to another,” Banks said. “From all outward appearances, they weren’t related processes.

“That’s what makes this so compelling is that they’re in disparate parts of the plant and we’re trying to figure out if there is a common theme.”

The investigation will include a review of equipment age, maintenance and inspection, as well as the plant’s monitoring system.

The chemical board has said it was aware of six earlier leaks at the DuPont plant since December 2006. Banks said his team also plans to examine how thorough any investigations were after those leaks.

No one appeared to be in the immediate vicinity when 58-year-old Carl Fish, a 32-year DuPont employee, was exposed to phosgene Saturday and died a day later.

“He was conducting normal rounds,” Banks said. “There was nothing extraordinary going on that would give them a sense of dread or something terribly amiss.”

Fish was exposed by an 18-inch braided steel transfer hose that ruptured. Banks said the hose was frayed, but it wasn’t immediately determined whether the hose wore down quickly or over time.

Although company officials confirmed the five-day leak of methyl chloride, Banks said his team hasn’t determined how long it was leaking.

“We will make that part of our investigative process to determine how long the leak went and why it wasn’t detected, if there was some type of mechanical integrity issue with the monitoring system,” he said.

The federal Occupational Safety and Health Administration also is investigating the leaks.

The safety board is already stretched thin by 16 other open investigations, the largest number in its 11-year-history. Board member William E. Wright has said the latest case would likely delay other cases, including those at the Bayer CropScience facility in Institute, where a worker was killed in an August 2008 explosion.