No label defined (Q4454552)

Recall that an integral domain \(D\) is a \(G\)-domain if there is a nonzero element \(u\) of \(D\) with \(D[u^{-1}]=K\), the quotient field of \(D\), or equivalently, if the intersection of the nonzero prime ideals of \(D\) is nonzero. A pair \((A,B)\) of integral domains \(A\subseteq B\) is called a \(G\)-domain pair if for each domain \(D\) with \(A\subseteq D\subseteq B\), \(D\) is a \(G\)-domain. Of course if \(D\) is a \(G\)-domain with quotient field \(K\), then \((D,K)\) is a \(G\)-domain pair. And it easily follows from \textit{I. Kaplansky}'s treatment of \(G\)-domains [theorems 21 and 22 in his book: ``Commutative rings'', 2nd ed. (1974; Zbl 0296.13001)] that \((A,B)\) is a \(G\)-domain pair if and only if \(A\) is a \(G\)-domain and \(B\) is algebraic over \(A\). However, the author gives an alternate proof. Main result: For a field \(F\), the following are equivalent: (1) every subring of \(F\) with quotient field \(F\) is a \(G\)-domain, (2) every subring of \(F\) is a \(G\)-domain, (3) \((E,F)\) is a \(G\)-domain pair where \(E\) is the prime field of \(F\), (4) \(F\) has no proper subrings with quotient field \(F\), and (5) \(F\) has characteristic \(p>0\) for some prime \(p\) and \(F\) is algebraic over \(F_{p}\).