washington

Superbrain: universities will gain access to the world's most powerful supercomputers, such as this one recently installed at Sandia. Credit: SANDIA NATIONAL LABORATORIES

US universities should reconsider their willingness to work closely with the nuclear weapons laboratories under the so-called ‘strategic alliances’ programme, according to the Natural Resources Defense Council (NRDC). The influential Washington-based lobby group says that such efforts could lead indirectly to the development of new nuclear weapons.

The programme, which started last summer, seeks to engage the universities' best brains on issues of interest to the weapons laboratories, without allowing the laboratories' secrets to leak out to the open environment of the universities.

But a report by the pressure group, which has been prominent in recent years in analysing and criticizing the US nuclear weapons programme, argues that the five major centres established under the programme are engaged in work that will help the United States to develop new and more powerful nuclear weapons in the future.

In particular, a draft of the report, obtained by Nature, says that work on shock physics at one of the centres — the California Institute of Technology (Caltech) — will pose a risk of proliferation, making it easier for aspiring nuclear powers to develop advanced nuclear weapons without having to test them.

Dual use: simulations of explosions are relevant to both ‘shock physics’ and weapons design. Credit: LAWRENCE LIVERMORE NATIONAL LABORATORY

In a move that reflects concern inside the weapons laboratories about the programme, foreign students and staff at the university centres have already been barred from having direct access to the laboratories' supercomputers (see below).

Critics say the unequal treatment of foreign students and staff reflects the ambiguous position of the whole strategic alliances programme, and that the programme implicates the universities in the development of weapons of mass destruction.

But the programme's defenders argue that research at the centres is not nuclear weapons research as such, and that helping the weapons laboratories with their new Science-based Stockpile Stewardship programme for maintaining the weapons without testing will enable the United States to comply with the Comprehensive Test Ban Treaty (see Nature 387, 541; 1997).

The five centres were set up last summer after a fierce competition between teams at forty research universities. The winners — Caltech, Stanford University and the universities of Chicago, Utah and Illinois — will each receive $25 million in research support over five years, with the strong prospect of a five-year extension.

They should also get access to the world's most powerful massively parallel supercomputers, now being developed and installed at the three weapons laboratories — Lawrence Livermore in California and Sandia and Los Alamos in New Mexico — under the Department of Energy's (DOE's) Advanced Strategic Computing Initiative (ASCI).

Although some universities originally drafted proposals which they considered of direct relevance to the weapons programme, the DOE has dissuaded them from this approach, asking them to simulate complex dynamic systems that are distinct from nuclear weapons. The energy department hopes the universities' involvement will help scientists at the weapons laboratories to solve the central technical problem of stockpile stewardship — the full computer simulation of a thermonuclear weapon's operation.

The University of Illinois centre, for example, will study solid-fuel rockets, while Chicago simulates astrophysical thermonuclear flashes. The Utah team will investigate accidental fires and explosions as these impinge on nuclear weapons safety, and Stanford will look at turbulence in gas-turbine engines.

But the Caltech centre will investigate the dynamic response of various materials to the detonation of high explosive — a problem of clear relevance to nuclear weapons design. According to the NRDC draft report, Caltech's proposal included the study of the behaviour of beryllium, uranium and actinide metals under shock from high explosives.

“If the Caltech program is permitted to continue for its five to ten year course, much of the work behind generating a bomb code will have been accomplished,” says the NRDC.

Caltech officials now say that these materials will not be included in their study. “We will not be working on nuclear weapons materials, or on anything that is a heavy metal,” says Daniel Meiron, head of the Caltech centre. Steve Koonin, provost and vice-president of research at Caltech, adds that he would “not allow research to be done at Caltech” for the nuclear weapons programme.

The NRDC argues, however, that the shock physics to be carried out at Caltech is exactly what proliferators would require if they wanted to advance from a primitive gun-type nuclear device to an implosion-based device, or even a hydrogen bomb boosted by a thermonuclear secondary, without testing.

In contrast to the vehement criticism of research related to President Ronald Reagan's Star Wars initiative fifteen years ago, campus reaction to the arrival of money from the nuclear weapons programme has so far been subdued. David Pershing, the head of the Utah centre, says that despite strong anti-nuclear feeling in the state (which is downwind of the Nevada site where US weapons tests took place), critics have accepted that his team is working on safety, not on building nuclear weapons.

At Caltech, however, the 100-strong Southern California Federation of Scientists has attacked the partnership in local news-papers, and called on David Baltimore, the new president of Caltech, to take a public position on the issue. Baltimore declined to be interviewed for this story, referring questions to Koonin.

The NRDC wants a government review of the proliferation implications of the partnerships with the universities, and a debate within the universities on whether they should be involved. Its draft report says that the universities' acceptance of the programme “represents an obvious — but financially fortuitous — failure to comprehend the full scope of the current nuclear weapons programme”.

The report argues that the $4.5-billion-a-year stockpile stewardship programme is less concerned with complying with the test ban treaty than with circumventing it, by developing computer models so powerful that the United States will be able to improve its nuclear weapons without testing.

The DOE has consistently denied such charges. But it does accept that the stockpile stewardship programme will help the United States to maintain its nuclear weapons design capability indefinitely, in case new weapons are required in the future, and that part of the function of the partnerships is to help the laboratories recruit scientists for that purpose.

A spokesman for Livermore referred questions about the partnerships to the DOE. But Tom Adams, an ASCI project leader at Los Alamos, says that staff there are enthusiastic about partnerships. “There's a lot of excitement,” he says. “People are most anxious to interact with the universities.”

Gilbert Weigand, deputy assistant secretary for strategic computing at the DOE, declines to comment on the NRDC's criticisms until its final report is published. But he stresses that the partnerships “are completely unclassified projects”, and adds: “They are things the universities wanted to do, in their own best interests.”