The existence of identity of orthodox semigroup rings (Q1206776)

Let \(S\) be an orthodox semigroup, and let \(E(S)\) be the band of idempotents of \(S\). For any \(x\in S\), put \(V(x) = \{y \in S\mid xyx = x\), \(yxy = y\}\). \(S\) is said to be \(N\)-perfect iff \(V(x)V(y) = V(yx)\) for all \(x,y \in D\). Assume that \(R\) is a ring with identity. The author gives a criterion for the existence of an identity in a contracted semigroup ring \(R_ 0[S]\) for \(S\) orthodox \(N\)-perfect. Assume that \(E(S)\) is a semilattice \(Y\) of rectangular bands \(E_ \alpha\) \((\alpha \in Y)\). Denote by \(M(Y)\) the set of all maximal elements of \(Y\) (under the usual semilattice order). If \(R_ 0[S]\) possesses an identity element then the following is true: (i) \(| E_ \alpha| = 1\) for every \(\alpha \in M(Y)\). So suppose that (i) holds; if this is the case then (Theorem 11) the following three conditions are equivalent: (ii) \(R_ 0[S]\) is a ring with identity, (iii) \(R_ 0[E(S)]\) is a ring with identity, (iv) \(R_ 0[Y]\) is a ring with identity. Thus the problem of the existence of an identity in a semigroup ring \(R_ 0[S]\) for \(S\) orthodox \(N\)-perfect is completely reduced to the same problem for a semilattice semigroup ring. The last problem was solved by Wenger: if \(Y\) is a semilattice then \(R_ 0[Y]\) is a ring with identity iff the set \(M(Y)\) is finite nonempty such that \(Y = Y\cdot M(Y)\). The author also poses an interesting problem which may be formulated as follows: let \(S\) be a regular semigroup such that the ring \(R_ 0[S]\) is a ring with identity \(1_{RS}\); is it true that \(\text{supp}(1_{RS}) \subset E(S)\)?