On Lie isomorphisms of rings (Q6940764)

In this paper, the authors significantly advance Herstein's longstanding problem of characterizing Lie isomorphisms between commutators \([A, A]\) and \([B, B]\) of associative rings, building on Herstein's conjectures and subsequent works by Martindale, Brešar, and others. By shifting focus from primeness and torsion-freeness assumptions to the presence of sufficiently many orthogonal full idempotents, they establish robust results that illuminate the interplay between associative and Lie structures, even in non-unital or torsion-present settings. By assuming at least three orthogonal full idempotents, they prove that such isomorphisms extended to standard mappings -- a homomorphism plus negative anti-homomorphism with orthogonal images -- in unital rings, applicable to matrix rings (\(M_n(R)\)) for (\(n \geq 3\)).\N\NFor non-unital rings, they establish a refined extension: the isomorphism lifted to a unique homomorphism from the universal annihilator extension of \(A \oplus A^{\mathrm{op}}\) to \(B\), incorporating central extensions akin to Schur-Garland. Additionally, they extend these insights to derivations, showing that they lift to associative derivations under the same hypotheses, with or without annihilator triviality. A counterexample with two idempotents highlights the necessity of three. The results discussed in this investigation add to a better understanding of these algebraic structures and their relationships between them, paving the way for future research and possible applications.