Pres. Barack Obama has proposed that both the U.S. and Russia continue to deplete their nuclear weapons reserves, encouraging a one-third reduction to arsenal sizes. His June 19 proposal, which did not yet specify a time line for destruction of the weapons reduction, expands on the 2010 New START Treaty between Russia and the U.S., which calls for each country to possess no more than 1,550 warheads by 2018.

Efforts to cut stockpiles and thus decrease the risk of accidental nuclear detonations and launches have so far been hampered by risks posed by the verification process. Verification regimes require the exchange of classified information, which could encourage nuclear proliferation if it ended up in the wrong hands. So, a team of scientists from Princeton University has outlined a new verification system that would release no classified details about the weapons.

How to confirm destruction without risking proliferation has been an issue for several years, says Robert Goldston, a physicist at Princeton and co-principle investigator on the project. To prevent the release of classified information, the current method for checking compliance counts delivery systems for warheads, not the warheads themselves. Certain proposed systems make classified measurements and then feed the information directly to a computing system that records only the final verdict. But it is easier for a host country to fake the process or for others to snoop on the system when it relies on a computer, says Princeton physicist Alexander Glaser, the project’s other principal investigator. So the researchers came up with a “zero-knowledge” verification system in which the inspector learns nothing classified.

Image: Jen Christiansen and Susan Matthews

The inspector would visit the host country, which would have a device called a neutron beamer ready. Passing neutrons through a warhead creates a signature: nuclear material absorbs some neutrons, and fission deflects some neutrons, scattering them. The number of neutrons that reaches a detector reveals if the warhead is nuclear. Unfortunately, this signature number is classified because physicists could work backward from it to understand both what materials are being used and how fission is occurring, essentially deconstructing the bomb’s contents.

The process of concealing the classified signature number borrows from a simple principle: Imagine having to confirm two bags of blue marbles are equal without knowing the number of marbles in them. Have someone add a set number of red marbles to both bags without telling you how many. Then, with your eyes closed count the number of marbles in each bag. If the totals are equal, you will know that the bags started off with identical numbers of blue marbles. In the new nuclear verification process, before inspection the host country and the inspector would agree on a number, termed N-max, which would confirm a warhead is nuclear. N-max is the sum of the signature neutrons (blue marbles) and an unknown number of preloaded bubbles that indicate neutrons (red marbles) that will be added to each count to disguise the true number in the signature. The process is first tested on a confirmed nuclear weapon, so that all subsequent warheads that reach N-max can be determined to be nuclear as well.

Glaser admits that the approach is “almost embarrassingly simple.” But the system contains hidden cleverness. If the host country offers up a fake warhead, another number besides N-max could be reached, which may reveal the number of preloaded bubbles being added. This could allow the inspector to work backward to find the signature, potentially releasing the host country’s nuclear secrets—added incentive not to cheat.

Although the theory seems sound, says Steve Fetter, a nuclear-control expert at the University of Maryland, he is not convinced such secrecy is necessary. Two states with advanced nukes don’t have much that they need to steal from one another, he notes. Tom Cochran, a physicist with the Natural Resources Defense Council, has concerns on other grounds. He brought up a testing system he proposed in the 1990s that measures the gamma rays nuclear warheads emit, which are more plentiful than neutrons and are statistically more likely to be accurate, he says. His system, however, relies on a computer to shield the classified information.

Glaser disputes the notion that better verification tests are unnecessary. As evidence that current systems are not adequate, he points to the U.S. Department of Defense’s 2010 Nuclear Posture Review that noted a need for “expanded work on verification technologies.”

Fetter concurs that as the number of warheads diminishes, governments will want to verify that reduction and that the neutron system may be a good option, although it is still a few years away from reality. “If these kinds of ideas are on the table, it should facilitate moving forward in negotiations,” Goldston says.