About 20 percent of the employees showed physiological responses that indicated some "deception" when questioned about unauthorized contacts, Edward J. Curran told The Washington Post.
But Curran, a former FBI official, said all of those people passed the exam when given the opportunity to answer the question again after being allowed to explain a minor transgression or admit past conduct that might cause slight feelings of guilt.
"We have nobody who hasn't gotten through the test, which is a pretty...good record," Curran told the newspaper. "These are not bank robbers or embezzlers. These are patriotic American citizens who already have [security] clearances - you expect them to pass."
Energy Secretary Bill Richardson approved the use of polygraphs for employees in highly sensitive positions within the nation's nuclear weapons program even before allegations of Chinese espionage emerged at Los Alamos National Laboratory in March 1999.
Following that case, Congress expanded Richardson's polygraph initiative, requiring a lie detector program to begin in January to cover 13,000 nuclear weapons scientists, technicians and security personnel.
The Post, however, quoted George Maschke, a leading polygraph critic and Army Reserve captain, who said he remains troubled by the program.
Maschke said because polygraphs are unreliable and inaccurate, the danger to employees lies in the interpretation of test results and employees' explanations.
He said if officials have reason to believe someone is a spy or may have broken security rules, investigators can find what they want in the test results.
©2000 CBS Worldwide Inc. All Rights Reserved. This material may not be published, broadcast, rewritten, or redistributed. The Associated Press and Reuters contributed to this report