The emergence of torsion in the continuum limit of distributed edge-dislocations (Q2516262)

The paper under review is motivated by the study of defects in solids, such as crystalline dislocations and point defects. Defects are quantified by the Burgers vector which is essentially the discrepancy between closed loops in real space and closed loops in the discrete crystallographic space. Defects in amorphous materials can be similarly understood. Here, the continuum is modelled by a topological manifold that is smooth except at the loci of the defects. The smooth part of the manifold is equipped with a locally-flat Riemannian metric, the defects are singularities of the topological manifold and manifest themselves through parallel transport with respect to the associated Levi-Civita connection. The Burgers vector is identified with the translational component of the monodromy. In material science one often studies materials with distributed defects. Using smooth manifolds to model such materials was pioneered by Nye, Kondo and Bilby \& Smith, see the reference list of the paper under review. In these early works from the 1950's the singularities were smoothed out, the resulting manifold comes with a flat metric, but a non-vanishing torsion that represents the Burgers vector density. The main drive of the paper under review is to make mathematical sense of torsion emerging in the continuum limit of discretely distributed dislocations. To do this the authors construct a sequence of manifolds with isolated dislocations, such that the dislocations become increasingly dense, but their total magnitude remains constant. They show how the sequence converges as a sequence of metric spaces to a flat simply connected Riemannian manifold. Moreover, they show that this sequence converges as a sequence of manifolds equipped with a connection. The main result of the paper is that they show that torsion can arise as a rigorous limit process from a torsion free Riemannian manifold.