Jordan derivations of some classes of matrix rings. (Q2470789)

For \(n\geq 2\) let \(R\) be a 2-torsion free ring with identity, and \(S\) a subring of \(M_n(R)\) containing all upper triangular matrices, so \(S=\sum I_{ij}e_{ij}\) where all \(I_{ij}\) are ideals of \(R\) and \(I_{ij}=R\) when \(i\leq j\). The author characterizes any Jordan derivation \(D\colon S\to S\) as the sum of a unique derivation \(d\colon S\to S\) and what is called a singular Jordan derivation \(\delta\colon S\to S\). The derivation \(d\) is the sum of an inner derivation of \(S\) and a derivation induced by entrywise application of a derivation of \(R\). The map \(\delta\) is additive and is given by \(\delta(I_{ij}e_{ij})=0\) except for \(\delta(xe_{n1})=f(x)e_{1n}\) for \(f\colon I_{n1}\to R\) satisfying certain necessary conditions; further, \(\delta=0\) at least when \(I_{n1}=0\), \(I_{n1}=R\), or \(R\) is a prime ring.