Generic noncommutative surfaces. (Q1826888)

The idea of interpreting certain noncommutative \(\mathbb{N}\)-graded algebras over a field \(k\) as coordinate rings of noncommutative projective schemes has emerged about a decade ago and has opened up a new subject area, now known as noncommutative projective geometry. Numerous ring-theoretic problems have been solved using this approach: for example, graded domains of dimension 2, which correspond to noncommutative curves, have been classified, and so have noncommutative analogues of the projective plane \(\mathbb{P}^2\). The survey article by \textit{J.~T.~Stafford} and \textit{M.~Van den Bergh} [Bull.~Am.~Math.~Soc., New Ser. 38, No. 2, 171-216 (2001; Zbl 1042.16016)] documents the considerable progress of the classification theory of graded algebras of dimension~3. In the paper under review, the author constructs examples of algebras of dimension \(\geq 3\) which provide answers to a number of open questions in the literature. Given a generic Zhang twist \(S=\bigoplus_{i\geq 0}S_i\) of a polynomial ring in \(t+1\) variables, \(t\geq 2\), over an algebraically closed field \(k\), let \(R\) denote any subalgebra of \(S\) which is generated by a generic subspace of \(S_1\) with codimension one. Then \(R\) has the following properties. (1) \(R\) is Noetherian, connected, finitely \(\mathbb{N}\)-graded, finitely generated in degree 1, and a certain technical homological condition, denoted \(\chi_1\), is satisfied (thus assuring the validity of the noncommutative version of Serre's Theorem due to \textit{M.~Artin} and \textit{J.~J.~Zhang} [Adv. Math. 109, No. 2, 228-287 (1994; Zbl 0833.14002)]), but the conditions \(\chi_i\) fail for all \(i\geq 2\), which answers a question posed by \textit{J.~T.~Stafford} and \textit{J.~J.~Zhang} [Math. Proc. Camb. Philos. Soc. 116, No.~3, 415-433 (1994; Zbl 0821.16026)]. (2) \(R\) is not strongly Noetherian, that is, \(R\otimes_kB\) is not necessarily a Noetherian ring for every commutative Noetherian \(k\)-algebra \(B\), thus putting an end to some hopes that all finitely generated \(\mathbb{N}\)-graded Noetherian \(k\)-algebras might be strongly Noetherian and that the point modules would therefore be naturally parametrized by a commutative projective scheme over \(k\), thus forming a nice geometric object. (3) \(R\) is the first example of a connected, finitely \(\mathbb{N}\)-graded maximal order for which \(\chi_i\) fails for \(i\geq 2\). This very interesting article will have considerable impact on what one can expect typical noncommutative surfaces to look like.