The ubiquity of Sylvester forms in almost complete intersections (Q2515054)

The paper under review is a very explicit work to explain the Rees ideal and the depth of the Rees algebra of an almost complete intersection ideal of height \(n\) in a polynomial ring \(R=k[x_1,\cdots,x_n]\) over a field \(k\). It is well-known that a substantial subset of generators of the Rees ideal come from the syzygies, \(\mathcal{L}\). The focus of this work is on special situations where the Rees ideal can be generated by \(\mathcal{L}\) plus a set of Sylvester forms derived from \(\mathcal{L}\) by iteration. This idea is inspired by a previous work of Hong, Simis, Vasconcelos, [\textit{J. Hong} et al., J. Commut. Algebra 5, No. 2, 231--267 (2013; Zbl 1274.13015)]. In Theorem 2.5, the authors show that a generic specialization of the base ideal of a plance Cremona map provides a birational map from \(\mathbb{P}^1\) onto a plane curve and moreover the Rees algebra of the base ideal of the latter is \textit{almost Cohen-Macaulay} i.e., \(\mathrm{depth}\geq \dim-1\), if and only if the Rees algebra of the base ideal of former is Almost Cohen-Macaulay. As a special case, in Corollary 2.6, the authors show that the Rees algebra of the base ideal of a \textit{de Jonquieres} map given by polynomials of degree at least \(4\) has always depth \(3\) and they explain the minimal free resolution of its Rees ideal. In section \(3\), the authors focus on the Rees algebra of monomial ideals which are almost complete intersection of height \(n\). In Theorem 3.5, it is shown that the Rees ideal of the ideal \(I=(x^d,y^d,x^by^{d-b})\subseteq k[x,y]\) where \(\mathrm{gcd}(d,b)=1\) is generated by \(\mathcal{L}\) and the Sylvester forms. In Theorem 3.7 it is shown that if in addition \(2b\leq d\) then the Rees algebra \(\mathcal{R}(I)\) is almost Cohen-Macaulay. Although the situations seems very particular, the proofs are computationally very involved because the authors want to present a direct proof by focusing on the equations in the Rees ideal; otherwise a more general proof can be found in [\textit{M. E. Rossi} and \textit{I. Swanson}, Contemp. Math. 331, 313--328 (2003; Zbl 1089.13501)]. Finally in Theorem 3.14, the authors present the Rees ideal of the ideal \(I=(x^a,y^a,z^a,(xyz)^b)\subseteq k[x,y,z]\) where \(a>2b\) in terms of \(\mathcal{L}\) and the Sylvester forms and show that \(\mathcal{R}(I)\) is almost Cohen-Macaulay. They conjecture that for any \(n\) the Rees algebra \(\mathcal{R}(I)\) of the ideal \(I=(x_1^a,\cdots,x_n^a,(x_1\cdots x_n)^b)\subseteq k[x_1,\cdots,x_n]\) is always almost Cohen-Macaulay and it is Cohen-Macaulay if and only if \(a\leq 2b\).