Factorially closed subrings of commutative rings (Q2353499)

Let \(A\subseteq B\) be integral domains. Then \(A\) is said to be \textit{factorially closed} (\textit{fc}) in \(B\) if for any two nonzero elements \(b_1, b_2\in B\), \(b_1b_2\in A\) implies \(b_1, b_2\in A\). In the paper under review, the authors generalize this notion in the following way: for any integral domains \(A\subseteq B\), \(A\) is said to be \textit{quasifactorially closed} (\textit{qfc}) in \(B\) if for any nonzero \(b\in B\), if there exists some nonzero \(b'\in B\) such that \(bb'\in A\), then there exists a unit \(u\in B\) such that \(bu\in A\). The authors prove several results about fc and qfc subrings of commutative rings from algebraic and geometric viewpoints. For example: {\parindent=6mm \begin{itemize} \item[(1)] If an inclusion of complete normal domains \(A\subseteq B\) over \(k\) (which is always assumed to be an algebraically closed field of characteristic zero) is qfc, then \(A\) is algebraically closed in \(B\). Further, any irreducible element of \(A\) is irreducible in \(B\). \item [(2)] An fc subring of a polynomial ring over \(k\) in at most three variables is again a polynomial ring. Similarly, a complete qfc subring of a power series ring over \(k\) in at most three variables is is again a power series ring. \item [(3)] If an inclusion of affine normal domains \(A\subseteq B\) (with a suitable hypothesis) is fc, then a general fiber of the corresponding morphism of affine varietes does not have any nonconstant invertible regular functions. \end{itemize}} As an application of their results, the authors give a short proof of a result of L.~D.~Tráng in connection with the Jacobian problem. They also give some examples of ring extensions which shed more light on the fc and qfc properties. In the last section of the paper some open problems are given. For example: (1) Is any fc subring of a PID also a PID? (2) Let \(A\subseteq B\) be an qfc inclusion of normal complete domains over \(k\). Is \(\dim A\leq \dim B\)?