Logarithmic foliations (Q2053316): Difference between revisions

The authors study singular holomorphic foliations of arbitrary codimension on projective spaces defined by logarithmic forms. Here a logarithmic form on a complex manifold \(M\) is a meromorphic \(q\)-form \(\eta\) on \(M\) such that the pole divisors of \(\eta\) and \(d\eta\) are reduced. One of the main results of the paper (Theorem 2.1) concerns normal forms for closed logarithmic forms: Let \(\eta\) be a germ at \(0\in\mathbb{C}^n\) of a closed logarithmic \(p\)-form with poles along a hypersurface \(X = (f_1 \dots f_r = 0)\) with strictly ordinary singularities outside \(0\in\mathbb{C}^n\). Assume that \(n > p + 2\). Then: \begin{itemize} \item[(a)] If \(r < p\) then \(\eta\) is exact; \(\eta = d\Theta\), where \(\Theta\) is logarithmic non-closed and has the same pole divisor as \(\eta\). \item[(b)] If \(r \geq p\), then \[\eta = \sum \lambda_I \frac{df_{i_1}}{f_{i_1}}\wedge\dots\wedge\frac{df_{i_p}}{f_{i_p}} + d\Theta,\] where, either \(\Theta =0 \), or \(\Theta\) is logarithmic non-closed and has pole divisor contained in \(X\). \end{itemize} As a consequence of Theorem 2.1, in the general case one gets normal forms in the case of logarithmic \(p\)-forms on \(\mathbb{CP}^n\) (Corollary 1.3). A natural question one might ask concerning the structure of logarithmic foliations is the following (Problem 1.9): Does a foliation on \(\mathbb{CP}^n\), defined by a logarithmic \(p\)-form, \(2 \leq p < n\), is an intersection of \(p\) codimension-one logarithmic foliations? A partial answer to this question is given by Theorem 1.10 in the case of strictly ordinary singularities.