To mollify Congress, the Department of Energy is preparing to give polygraph tests to thousands of nuclear scientists. But at the same time, a Senate panel has asked the CIA and FBI to explore alternatives to polygraphing because of the "potential unreliability" of the so-called lie-detector exams.

In the wake of allegations of Chinese nuclear espionage, Secretary of Energy Bill Richardson has ordered polygraphs for an estimated 5,000 scientists and other employees involved in nuclear weapons research, the first large-scale use of the tests outside of the CIA and National Security Agency.

Yet experts inside and outside the government agree that the tests are far from foolproof, which is why they are not admissible as evidence in court.

"Polygraphing has been described as a 'useful, if unreliable' investigative tool," the Senate intelligence committee said in its report on the fiscal 2000 intelligence spending bill, which the Senate passed Wednesday. The panel instructed CIA Director George J. Tenet and FBI Director Louis J. Freeh to assess "alternative technologies to the polygraph" and report back within 90 days after the measure becomes law.

The polygraph is based on sensitive monitoring of pulse rates while the subject answers questions. Its accuracy is highly dependent on the skill and experience of the examiner.

Among the other potential lie-detection systems that the CIA and FBI already have investigated are skin sensors, voice stress recorders, brain wave studies and even hypnotism. "We have expended a lot of research into alternatives," said an FBI spokesman. "But no real other technologies exist today."

Concern about the unreliability of the standard polygraph is also growing at the Department of Energy's nuclear weapons labs. In a recent letter in the Los Alamos National Laboratory's employee newsletter, one scientist noted that the expected rate of false positives--tests indicating someone is lying when that person is not--is about 2 percent. "In our situation, that's 100 innocent people out of 5,000 whose reputations and careers would be blemished," wrote the scientist, John D. Fowler Jr.

And what will happen if a prominent weapons designer or the leader of a research team fails a polygraph? That is a question many lab employees are now posing, according to one Energy Department official.

Sensitive to the growing opposition among scientists, Energy Department officials are trying to reassure employees that the scope of polygraph questions will be narrow. No one will be asked about "thoughts and beliefs" or "conduct and lifestyles"--meaning that political, religious and sexual questions are off-limits. The focus, officials say, will be on any involvement in espionage, sabotage or terrorism; unauthorized disclosure of classified information; or unauthorized foreign contacts, including failure to report an approach by someone seeking secret information.

Two of the main nuclear weapons labs, Los Alamos and Lawrence Livermore, are run by the University of California under contract with the Energy Department. The third big research center, Sandia Laboratories, is run by Lockheed Martin Corp. Since the labs are in the private sector, employees cannot be required to take polygraphs as a condition of employment. But workers can be switched out of sensitive assignments, deprived of access to secret information, and shunted to other research projects if they refuse to take the tests.

C. Paul Robinson, director of Sandia Laboratories, told his employees earlier this month that by law, they can choose not to take a polygraph. "However," he said, "they will not be permitted to work in areas where such a requirement is established by the customer," the federal government.

Some employees working on "special access programs"--the most secret projects--already have signed agreements to take the tests. But polygraphs are a new requirement for most researchers involved in nuclear weapons design and operations.

In another recent missive to the Los Alamos employee newsletter, a scientist named Mario E. Schillaci challenged the government to prove the validity of polygraphs. "If Energy Secretary Bill Richardson wants our enthusiastic concurrence in the planned tests, he must do more than simply assert that the FBI, CIA and NSA consider the tests a strong enough indicator of truth," he wrote. "The question is, Exactly how strong?"