The stochastic heat equation is a fundamental mathematical object that describes how heat spreads through a randomly fluctuating environment. In two dimensions, this equation is particularly tricky because the randomness is too wild for standard approaches to work directly. A special limiting object called the Critical Stochastic Heat Flow (SHF) was recently constructed to capture the behavior of such systems at a precise threshold of disorder. Understanding what kinds of models converge to this object is an important open problem in probability theory.
This paper studies a family of models called directed random polymers, which describe the statistical mechanics of a long flexible chain (like a polymer) moving through a disordered medium. Specifically, the models live in two spatial dimensions and evolve in continuous time, with the randomness coming from correlated Brownian noise. The central quantity of interest is the partition function, a measure of how the polymer explores its environment. The authors show rigorously that as these discrete polymer models are scaled appropriately, their partition functions converge to the Critical SHF, establishing these polymers as natural discrete approximations of this exotic limiting object.
The proof strategy combines several modern tools from mathematical physics. The key step is showing that all statistical moments of the partition functions converge to their predicted limits. This is achieved using a technique called the resolvent method, which reformulates the moment calculations in terms of quantum mechanical operators and allows the authors to handle the complicated correlations in the system. Once moment convergence is established, the authors verify additional technical conditions needed to guarantee full convergence to the SHF, using a recently developed axiomatic characterization of the SHF that pins down what properties uniquely define it. The result strengthens the universality of the Critical SHF as a canonical object in two-dimensional disorder problems.