On the vanishing topology of isolated Cohen-Macaulay codimension 2 singularities (Q1983183): Difference between revisions

The authors study the Milnor fibre of a class of determinantal singularities with special attention to the case of simple isolated Cohen-Macaulay codimension-2 (ICMC2) singularities. Those were classified by [the first author and \textit{A. Neumer}, Commun. Algebra 38, No. 2, 454--495 (2010; Zbl 1193.32015)]. The main tool is a partial resolution called Tjurina transformation. In this paper this modification is first studied in more generality than needed for the present application. For a generic determinantal variety consisting of \((t+1) \times t\) matrices of rank less than \(t\) the transform is the closure of the graph of the rational map sending a matrix to its kernel, the latter considered as element of the Grassmannian of lines in \(t\)-space. For a special determinantal singularity the transform is defined by pullback from the generic set-up. The first application is a description of \(T^1\), the space of infinitesimal deformations, for the case \(t=2\) in terms of deformations of the Tjurina transform. The second is that all simple ICMC2 singularities of dimension \(d\leq2\) are rational. They have at most A-D-E singularities on their Tjurina transform. The main result concerns the homology of the Milnor fibre \(F\) of an ICMC2 threefold singularity \(X\) with only isolated singularities (necessarily complete intersections) on the Tjurina transform \(Y\): \(H_1(F)=0\), \(H_2(F)=\mathbb Z\) and \(H_3(F)=\mathbb Z^r\) with \(r\) the sum \(\mu_{\text{tot}}\) of the Milnor numbers of the singularities on the Tjurina transform if \(t=2\) and \(r=\mu_{\text{tot}}-1\) otherwise. In particular, the reduced Euler characteristic of the Milnor fibre is \(b_2-b_3\). It can be the case that \(b_3=0\). For the simple singularities, \(\tau=\dim T^1\) is compared to the topological invariants: \(\tau = h^1(Y,\mathcal{T}_Y^0)+b_3\). The authors discuss an example of a series of singularities where \(b_3=0\) and \( \tau\) grows linearly with the maximal number of determinantal singularities in which the given one can deform.