On April 14, 1998, a Pacific Northwest National Laboratory (PNNL) researcher performing work in the Building 324 facility approached facility management and asked if facility management could turn off the tritium sampler in the main exhaust stack. The researcher was demonstrating the feasibility of treating components from dismantled nuclear weapons in a device called a plasma arc furnace and was concerned that the sampler would compromise classified information. B and W Hanford Company (BWHC) operated the facility, and PNNL conducted research as a tenant in the facility. The treatment of 200 components in the furnace would result in the release of up to about 20 curies of tritium through the facility stack. The exact quantity of tritium was calculated from the manufacturing data for the weapons components and was known to be less than 20 curies. The Notice of Construction (NOC) approved by the Washington State Department of Health (WDOH) had been modified to allow releasing 20 curies of tritium through the stack in support of this research. However, there were irregularities in the way the NOC modification was processed. The researcher was concerned that data performed on the sampler could be used to back-calculate the tritium content of the components, revealing classified information about the design of nuclear weapons. He had discussed this with the PNNZ security organization, and they had told him that data from the sampler would be classified. He was also concerned that if he could not proceed with operation of the plasma arc furnace, the furnace would be damaged. The researcher told BWHC management that the last time the furnace was shut down and restarted it had cost $0.5 million and caused a six month delay in the project`s schedule. He had already begun heating up the furnace before recognizing the security problem and was concerned that stopping the heatup could damage the furnace. The NOC that allowed the research did not have an explicit requirement to operate the sampler during a release. The sampler was installed several years previously for other research. After reviewing the NOC and other safety basis documents, and after consulting environmental compliance specialists, facility management agreed to turn off the sampler.