Zero-knowledge Isotopic Discrimination for Nuclear Warhead Verification

Abstract

Future nuclear arms control agreements seeking reductions to both deployed and stockpiled nuclear warheads will require new verification approaches to support their implementation. A key challenge will be to establish the authenticity of individual items presented as warheads before dismantlement. Warhead designs are highly sensitive, however, and the verification process is likely to be constrained by national security sensitivities that may not allow the revelation of any design information to inspecting parties. To address this conundrum, cryptographic techniques have been proposed for making meaningful measurements while addressing secrecy requirements.

This thesis directly contributes to the development of cryptographic methods for warhead verification by developing and demonstrating key aspects of a zero-knowledge isotopic comparison system comparing the isotopic composition of objects without learning what they are. It builds upon the work of Glaser, Barak, and Goldston who proposed the application of the cryptographic concept of zero-knowledge proofs to physical measurements, and the first experimental demonstration by Philippe et al focusing on geometry and opacity to 14 MeV-neutrons. In mathematics, zero-knowledge proofs have the unique and counterintuitive ability to show that a mathematical statement is true without revealing why it is true.

This thesis extends existing capabilities of zero-knowledge verification to include discrimination of an object's isotopic composition. This is fundamental since the isotopic composition of weapon-grade fissile materials is a key feature of its ability to sustain an explosive fission chain reaction. Through the design, simulation, and construction of a compact, steel-moderated collimator that can discriminate an inspected item's uranium or plutonium isotopic composition, this work allows the zero-knowledge framework to cover more inspection scenarios, and provides greater assurance of the validity of objects claimed to be nuclear weapons.

This thesis also introduces an optical method for counting bubbles in superheated emulsion detectors. To avoid the use of untrusted electronics in the measurement process, the zero-knowledge warhead verification technique uses analog detectors to record neutron fluence. As the effectiveness of a zero-knowledge comparison system increases with higher counting statistics, this 3D optical tomographic method provides reliable bubble counts at higher neutron exposure than other existing methods that tend to rapidly saturate.