The paper resolves a long-standing open question in the mathematics of elasticity theory. When scientists model how solid materials deform, they use energy functions that describe how much energy is stored in a deformed material. Two important mathematical properties these functions can have are called rank-one convexity and quasiconvexity. Rank-one convexity is easier to verify and is implied by quasiconvexity, so mathematicians have long wondered whether the two conditions are actually equivalent, or whether rank-one convexity can hold without quasiconvexity. In two dimensions, a famous result from the 1990s showed they can differ, but the three-dimensional case, which is what physically matters for real elastic solids, remained open for decades.
The authors construct an explicit energy function in three dimensions that is rank-one convex but not quasiconvex, and that also respects a fundamental physical requirement called frame-indifference, meaning the stored energy does not change if you rigidly rotate the material before measuring it. The function is built as a carefully modified polynomial, starting from a known two-dimensional-style example, then adapting it to the geometric constraints of three-dimensional nonlinear elasticity through a sequence of corrections involving prestress terms and enlargements of certain polynomial components. The failure of quasiconvexity is subtle and occurs through a mechanism first introduced by Sverak, where specially constructed test functions reveal that the energy function can be made artificially low in a way that a truly quasiconvex function would not permit.
The work is notable not only for its mathematical content but also for its methodology. The authors used the AI model Claude Opus 5.5 as a research assistant during the construction, and then formally verified the proofs using Lean 4, a computer proof assistant that checks every logical step with machine precision. This combination of AI-assisted discovery and formal verification represents an emerging approach to tackling hard problems in pure mathematics, providing a level of certainty beyond traditional human-checked proofs.