Three quarters of the U.S. Polaris A1 model submarine-launched nuclear warheads probably would not have worked in the mid-1960s because of a mechanical defect, Department of Energy officials disclosed yesterday.

In November 1966, scientists who discovered the extent of the problem described it as "truly catastrophic," officials said. Publicly, however, there was nothing but praise at the time for the Polaris system.

The evident failure of one major part of the U.S. strategic triad was "closely held at the time," officials said.

It took several months in late 1966 and early 1967 to design a remedy for the difficulty and almost another year to refit the missile warheads already deployed on submarines, officials said after reviewing records of the Atomic Energy Commission, predecessor of DOE.

Accounts of past problems with the U.S. nuclear weapons stockpile -- such as the probable failure of the Polaris A1 -- have surfaced recently as part of the debate over a comprehensive nuclear test ban.

Scientists, many of them associated with the government's nuclear weapons laboratories at Los Alamos, N.M., and Livermore, Calif., have cited weapons problems to support arguments both for and against continued nuclear testing.

In the process, however, it has become clear that -- more often than the public knows -- stockpiled nuclear warheads do run into unexpected problems and do require and receive continued surveillance.

As one key laboratory official said yesterday, "nuclear weapons are pretty cotton-picking ornery," being made up of "strongly non-compatible materials" that are both radioactive and chemically unstable.

The 1960s Polaris problem, which did not appear until five years after the A1 warheads were deployed, is an illustration.

As pieced together from government sources and scientists with the Los Alamos and Livermore weapons laboratories, the Polaris story began in the late 1950s when the warhead was about to go into production.

Tests in 1958 and 1959 showed what a source described as a "nuclear safety problem," one that would not harm the warhead's operation but required additional safety mechanisms.

Livermore scientists, who had designed the troubled warhead, had an alternative warhead available. But because the nuclear test moratorium with the Soviet Union was then in effect, they could not test it. Therefore, they put the original warhead in production with an additional safety device.

The first Polaris A-1s and the submarines carrying them went into service in 1960.

In 1965, during a routine examination of the Polaris sytem, a problem of aging in the materials used in the safety device was discovered.

In November 1966, broader testing of a bigger Polaris A1 warhead sample found, according to one official, that "three were bad to one good."

A scientist said yesterday the materials in the safety device in effect "bonded together" so that the safety would not realese, thus preventing the nuclear warhead from exploding.

A DOE official said, "When three quarters failed in a lab test it was believed that three quarters would fail in operation."

A Livermore scientist of the time argued yesterday that the problem was not that serious. Rather, there were deviations from design specifications, he said, "and that would lead to degradation of the entire system."

He did concede, however, that "everybody gave finding a solution a very high priority and it got taken care of promptly."

The solution was to incorporate a nuclear component from the newer Polaris A3 warhead, which was then about to go into production.

A Polaris A1 warhead as modified by the new component was tested underground in early 1967, since the test moratorium had ended in 1961. One month after the successful test, production was begun on the components so that the whole A1 missile fleet could be changed over. The job, sources said, was finished in 1968.

Defense Secretary Harold Brown was, in 1958, director of the Liver-more laboratory where the Polaris was designed. When the Polaris problem turned up in 1966, Brown was secretary of the Air Force.

A spokesman said yesterday Brown did not want to comment on the incident. Attempts to reach Robert S. McNamara, then secretary of defense and now president of thhe World Bank were unsuccessful. President Lyndon B. Johnson is said to have been "aware of the situation."

Ironically, in August 1967 as the Polaris was being fixed the Navy made a big splash with the launching of the 41st and final Polaris submarine, the "Will Rogers." Newspapers in 1967 referred often to the Polaris as the most complex and successful of the U.S. nuclear weapons programs.

The first public indication of the Polaris problem came last Aug. 15 in a letter to President Carter from the scientists who were associated with the Los Alamos weapons laboratory.

They cited it in attempting to make a case for a comprehensive test ban -- one that would not permit any underground nuclear tests.

In their letter the scientists said, "one Polaris warhead problem could readily have ben solved" without "the change of nuclear system... "

This statement irritated Livermore scientists in two ways. Their weapon was being cited as having a problem, and most of them believe continued weapons testing is needed.