The paper tackles a deep question in the mathematics of elasticity: when is a material's stored energy function "well-behaved" in a precise mathematical sense? For elastic materials, the energy stored under deformation must satisfy certain convexity-like conditions to ensure the physics makes sense, meaning that minimizers of energy actually exist and solutions to equilibrium equations are stable. Three nested conditions are relevant here, from strongest to weakest: polyconvexity, quasiconvexity, and rank-one convexity. For the quadratic energy functions that describe linear elastic materials, rank-one convexity and quasiconvexity turn out to be equivalent, but polyconvexity is a strictly stronger requirement. A long-standing open problem is to identify which quadratic forms are quasiconvex but not polyconvex, as these "extremal" cases sit on the boundary of what is mathematically tractable.
The paper organizes this search according to the symmetry class of the material, since real materials like crystals have internal symmetry that constrains the form of their elasticity tensor. The symmetry classes range from the most symmetric (isotropic) to the least (triclinic), with intermediate classes like cubic, hexagonal, trigonal, tetragonal, and orthotropic. The authors prove a striking result for orthotropic materials, which include many common engineering materials like wood or rolled metals: within this symmetry class, quasiconvexity and polyconvexity are actually equivalent. This means there are no exotic extremal forms hiding in that class, which simplifies the mathematical landscape considerably.
Moving to the less symmetric trigonal class, which describes crystals like quartz and calcite, the equivalence breaks down. The authors explicitly construct a one-parameter family of quadratic forms that are quasiconvex but not polyconvex, making them genuine extremal examples. The paper also develops general tools for verifying rank-one convexity, polyconvexity, and extremality, improving on earlier work. Together, the results draw a clearer map of where the gap between polyconvexity and quasiconvexity first appears as one moves through the hierarchy of material symmetries, with the answer being precisely at the trigonal level.