On the freeness of equisingular deformations of plane curve singularities (Q5960426)

From MaRDI portal
scientific article; zbMATH DE number 1724922
Language Label Description Also known as
English
On the freeness of equisingular deformations of plane curve singularities
scientific article; zbMATH DE number 1724922

    Statements

    On the freeness of equisingular deformations of plane curve singularities (English)
    0 references
    0 references
    7 April 2002
    0 references
    A germ \((V,0)\subset(\mathbb{C}^p,0)\) of complex analytic hypersurface is called free if the \({\mathcal O}_{\mathbb{C}^p,0}\)-module \(\text{Derlog} (V)\) is free (this terminology was proposed by K. Saito). Here, \(\text{Derlog} (V)=\{\alpha\in \Theta_p:\alpha (I(V))\subseteq I(V)\}\), where \(\Theta_p\) denotes the module of germs of vector fields on a neighborhood of the origin of \(\mathbb{C}^p\) (i.e., in local coordinates \(x_1,\dots,x_p\), \(\Theta_p\) is spanned by \(\partial/ \partial x_1,\dots,\partial/\partial x_p)\). Similarly, if \(F=0\) is an equation defining \(V\) near 0, we let \(\text{Derlog} (F):=\{\alpha\in \Theta_p: \alpha (F)=0\}\). Until recently, most nontrivial examples of free hypersurfaces were objects related to certain ``universal problems'', such as discriminants of versal deformations, certain singularities, bifurcation sets associated to suitable versal unfoldings, etc. Not any reduced hypersurface is free. Freeness is valid for \(p=1\) (i.e., reduced plane curves) but if \(p>1\), even hypersurfaces with isolated singularities may be not free. In this paper the author considers germs of hypersurfaces \((X,0)\subset(\mathbb{C}^p,0)\), where \(X\) is the total space of a one-parameter deformation of a plane curve singularity, say \(X:F=f(x,y)+ zg(x,y)=0\) (in suitable local coordinates \(x,y,z)\), where \(f\) and \(g\) are weighted homogeneous polynomials with \(wt(f)\leq wt(g)\) (such a deformation is necessarily equisingular), and proves a criterion to insure that the germ \((X,0)\) will be free. It involves a certain mapping \(\psi:\text{Derlog} (X)\to (J(f):g)\). Namely, if \(\xi\in \text{Derlog} (X)\), then its restriction to the ``plane'' \(z=0\) defines an element of the transporter \((J(f):g)\); in this way we get the mentioned mapping \(\Psi\). We say that \(g\) is fully extendable (for \(f)\) if \(\Psi\) is surjective. Then, in the situation above, the criterion says: if \(g\) is fully extendable for \(f\) and the deformation \(X:F=f(x,y)+ zg(x, y)=0\) is not analytically trivial, then \((X,0)\) is a free divisor in \(\mathbb{C}^3\) if and only if \((J(f):g)\) is a complete intersection ideal in \(\mathbb{C}\{x,y\}\), generated by weighted homogeneous elements \(h_1\), \(h_2\), and \(wt(g)+wt(h_1)+ wt(h_2)=2wt(f)- wt(x)-wt(y)\). As an application, if a quasi homogeneous polynomial \(f(x,y)\) defines a non simple curve singularity, \(H\) is the Hessian of \(f\) and \(X:f(x,y)+ zH(x,y)=0\), then \((X,0)\) is a free divisor in \((\mathbb{C}^3, 0)\). The author proves a partial converse to this result about Hessian deformations. He also gives other examples of applications of the main criterion. One of these shows that freeness is not a topological property.
    0 references
    0 references
    0 references
    0 references
    0 references
    free divisor
    0 references
    logarithmic derivation
    0 references
    equisingular deformation
    0 references
    nonisolated singularity
    0 references
    0 references