An unreleased Department of Energy report concludes that a linear accelerator could produce one of the key ingredients in nuclear weapons more cheaply, quickly and safely than new production reactors could.

Such an accelerator would produce far less radioactive waste than a nuclear reactor and "should be more readily acceptable to the public," according to a draft summary of the report obtained by The Associated Press on Thursday.The U.S. supply of tritium, a radioactive gas used to produce nuclear warheads, is threatened by the crumbling condition of the Energy Department's existing weapons plants and reactors.

The report, prepared by scientists at the Brookhaven and Los Alamos national laboratories and at Westinghouse-Hanford Co. in Washington state, concluded that an accelerator could produce enough tritium to meet all U.S. weapons needs.

An accelerator is essentially a long tunnel in which subatomic particles are whipped at high speeds until they crash into a target, in this case lithium clad in aluminium. The resulting collision produces tritium.

According to the report, using such an accelerator to make tritium would not produce any "long-lived" nuclear wastes, and "costs associated with safety and environmental concerns should be minimal."

In addition, it said, an accelerator "does not have the licensing requirements associated with nuclear reactors and should be more readily acceptable to the public."

Last July, the department recommended building two new production reactors, one at the Savannah River complex near Aiken, S.C., and the other at the Idaho Falls National Laboratory in Idaho.

At present, the nation's only facilities for producing tritium are three reactors at Savannah River, but all three are shut down either for maintenance or because of safety problems.

Department officials have said that at least one of the reactors should be back on line this year, but the start-up date continues to slip as more problems are found.

The department's Energy Research Advisory Board looked "briefly" at using an accelerator to produce tritium, but ended up recommending construction of two new reactors, said Chris Sankey, a DOE spokeswoman.

"The board concluded that these technologies were not mature enough to provide new production capacity in the next 10 to 12 years," she said.

But the DOE report said that the nation's tritium needs could be met by using an accelerator "that could be in operation as early as 1996."

By comparison, the proposed new reactors are expected to take a minimum of 10 years to complete.

"The required accelerator technology has largely been demonstrated at the component level although some additional development work is desirable to effect cost reductions, improved performance and ease of operation," the report said.

The accelerator would take massive amounts of electricity to operate, roughly the entire output of a nuclear power plant. One potential site for it would be in the Pacific Northwest, which has the cheapest electric rates in the nation.

An accelerator could cost more to build, but because the cost of operating it would be about half that of a reactor, "total life cycle costs," including construction, would be about 25 percent less, said Pierre Grand, a scientist at Brookhaven.

The two planned new reactors would cost an estimated $6.8 billion.