Weapon inspection scheme would test for nukes but keep designs secret
Protocol could encourage dismantling of stockpiled nuclear warheads
By Andrew Grant
A proposal for authenticating nuclear weapons carries no risk of leaking classified information. The scheme would keep a weapon’s design hidden while allowing inspectors to say whether or not a warhead carried nuclear material. The advance removes a hurdle toward achieving a thorough plan for global nuclear disarmament.
“It’s ingenious,” says James Acton, a nuclear policy expert at the Carnegie Endowment for International Peace in Washington, D.C., who trained as a physicist. “It solves a major problem for warhead verification.”
In 2010 the United States and Russia agreed to the New Strategic Arms Reduction Treaty, which has provisions for inspectors to count nuclear warheads on long-range delivery systems including missile launchers and submarines. By 2018, each country should have no more than 1,550 deployed warheads. That’s an important step, says Princeton University physicist Alexander Glaser, but it doesn’t account for the unknown number of weapons in storage in both countries. “Why have all this inspection on 1,550,” he says, “if you have 5,000 sitting in your basement?”
Glaser hopes that a future treaty would reduce the total number of nuclear weapons, counting the stockpiled as well as the deployed warheads. But no one knows how to implement such an agreement.
Rather than simply count warheads, inspectors would have to examine individual weapons, ensuring that countries agree to dismantle real nukes rather than fakes. But no country would agree to inspections if classified information about the weapons’ design could leak out to rivals or terrorists. In other words, Glaser’s dream treaty would require an inspection protocol that could “show that something is authentic without learning anything about it,” he says.
Glaser learned from a colleague that computer scientists have already solved a similar conundrum with what’s called a zero-knowledge proof, which is used in electronic voting and online auctions to gather information while ensuring security.
In a zero-knowledge protocol for nuclear weapons, detailed by Glaser and colleagues in the June 26 Nature, inspectors would use detectors that measure how many neutrons pass through and are emitted by a potential warhead. Although the number of transmitted neutrons can reveal the amount of weapons-grade uranium and the design — both confidential details — a roundabout scheme would prevent inspectors from determining the neutron count.
Under the scheme, the country holding the weapon in question would provide a detector with cells that are preset with a particular neutron count unknown to the inspectors. The country would also provide a target result for the test, which the actual neutron count and the detector’s preset count would add up to. For example, if Russia knew one of its weapons would transmit 100 neutrons (a classified number) to a given detector cell, it could preset the cell to a neutron count of 900 and tell the inspectors that the weapon test result should be 1,000.
That step alone does not verify a weapon is authentic. Glaser’s scheme also calls for inspectors to use the same preset detectors to test a template warhead that both countries agree is a nuke. If the detectors come up with the same readings for both devices, then the inspectors can be confident that they have authenticated a nuclear weapon.
The big advance of this protocol, Acton says, is that it makes suspicions between countries and inspectors irrelevant. Previous protocol proposals would hide measurement results from inspectors using software, but those processes could be tampered with by either country or a malicious third party. Under Glaser’s plan, Acton says, “you end up taking an entirely unclassified measurement.”
Acton stresses that Glaser’s plan isn’t ready for implementation yet, and that even a foolproof scientific method would not ensure a political agreement between two feuding countries. Still, he says that nuclear diplomacy is impossible without progress from science. “This is a major development,” he says.