Monday, May 5, 2008

Yucca corrosion data found to be suspicious

Findings fuel further criticism

WASHINGTON -- Government scientists raised questions in recent weeks about Department of Energy experiments on how long it will take canisters containing highly radioactive nuclear waste to corrode after being placed within Yucca Mountain.

The discovery led DOE to replace the data that were part of its license application to build a repository at the Nevada site, a project official said Wednesday.

At the same time, DOE has launched a review of the challenged research, which involved the nickel-based Alloy 22 that will be the outer cover of waste-containing packages. In the experiments, Alloy 22 samples were subjected to a solution of corrosive chemicals and then weighed to determine how much they had degraded.

Technicians reviewing the results reported "documented, repeated and potentially significant excursions" from the American Society for Testing and Materials standard for handling corrosion test specimens, according to a March 5 document that surfaced this week.

The activity is taking place weeks before the department has said it plans to apply for a construction license. It fueled further criticism from Nevada critics of Yucca Mountain who charge DOE is rushing unduly to file for a license.

Russ Dyer, the chief scientist for the Yucca Mountain Project, said the suspicious corrosion data "was roped off" and is not part of the Yucca application.

Dyer said DOE initiated a corrective action to determine "what exactly happened in this experiment and the results that came out of it and the processes we used."

In the meantime, DOE is using corrosion rates resulting from a separate set of experiments that sought to determine how corrosion might develop in canister welds and other crevices of the waste package.

"What is the potential impact on total system performance, and the answer is none," Dyer said. To gain a license, DOE must show that the canisters together with other features of the repository can prevent radioactive material from leaking for periods close to a million years.

Steve Frishman, technical policy coordinator for the Nevada Agency for Nuclear Projects, maintained DOE was "papering over" a problem. He said the state may challenge DOE on the corrosion data during license hearings.

The discovery came as scientists from Sandia National Laboratories were reviewing corrosion data. They said they uncovered a "vulnerability" in the data that were collected over five years.

The Sandia findings were contained in a March 5 Power Point presentation that became available on a Yucca Mountain public document database.

After emerging from the corrosive bath, the Alloy 22 "coupons" were cleaned of corrosion before being weighed. Sandia reported the cleaning process "may have been incomplete."

As a result, salts and other residue may have skewed the weight of the samples, raising questions about how much corrosion had taken place. A heavier piece might suggest the metal could last longer.

Sandia said there is "less than a 50 percent chance" the corrosion data were invalid. "But given the critical nature of this parameter (it) must be confirmed."

"The corrosion rate is the core of the total system performance analysis," Frishman said. "When they are talking about containers that don't fail for hundreds of thousands of year, it is possible they are off by orders of magnitude."

The corrosion experiments were conducted at the Lawrence Livermore National Laboratory in California. In 2006, DOE issued a stop work order on a separate set of corrosion experiments after Nuclear Regulatory Commission inspectors reported the work was based on humidity gauges that were not calibrated.

No comments: