The heat equation describes how temperature (or any diffusing quantity) spreads through a material over time. A key ingredient in this equation is the "diffusivity," a function that describes how easily heat flows from place to place and can vary across the material. A natural and practically important question is: if you cannot directly measure the diffusivity everywhere, can you figure it out by observing how the temperature is distributed at some later time? This paper answers that question in a strong and precise way.
The authors prove that, under reasonable mathematical conditions on the class of materials you are considering, a single snapshot of the temperature field at one chosen moment in time is enough to uniquely determine the diffusivity everywhere in the material. Crucially, they show you can design the experiment ahead of time, before knowing the specific material, by choosing a simple uniform starting temperature and a particular exponential heating pattern applied at the boundary. This single experimental setup works simultaneously for an entire class of materials, which is a much stronger and more practical statement than needing a tailored experiment for each unknown material.
Beyond just proving that the diffusivity can be recovered uniquely, the paper establishes "Lipschitz stability," meaning that small errors in your temperature measurement at the final time lead to only proportionally small errors in your estimate of the diffusivity. The authors also show that the type of measurement norm they use is essentially the sharpest possible within a standard mathematical framework. Taken together, these results give solid theoretical foundations for the feasibility of recovering material properties from a single well-chosen observation, resolving an open problem that had been posed in the field.