Co-Higgs bundles on \(\mathbb P^1\) (Q2439662)

Co-Higgs bundles have been of interest recently. They have been studied in connection with generalized holomorphic bundles in [\textit{N. Hitchin}, J. Geom. Phys. 61, No. 1, 352--362 (2011; Zbl 1210.53079)]. Later, the present author investigated them in more detail in his thesis [``Geometry of co-Higgs bundles'', D. Phil. thesis, Oxford (2011), \url{http://people.maths.ox.ac.uk/hitchin/hitchinstudents/rayan.pdf}]. In this paper, which is derived from that thesis, he studies co-Higgs bundles on the projective line. These are defined as vector bundles \(E\) on \(\mathbb{P}^1\) together with a morphism of vector bundles \(\phi: E \to E \otimes \mathcal{O}(2)\). The concept of (semi)stability of co-Higgs bundles can be defined in a manner similar to that of Higgs bundles. In the case of co-Higgs bundles on a curve, the author points out (see Section 3) that these are meaningful only when the curve has zero genus. Accordingly, this paper treats the genus zero case. By a well-known result of Grothendieck, any vector bundle \(E\) on the projective line can be written as a direct sum of line bundles, which are in turn given by a unique sequence of decreasing integers. The main result of the paper (Theorem 6.1) gives necessary and sufficient conditions on these integers for the existence of a semistable co-Higgs field on \(E\). Moreover, the author proves that the generic co-Higgs field in this case is in fact stable. He uses the main result to study moduli spaces for rank 2 co-Higgs bundles; and also to calculate the Betti numbers of moduli spaces when the rank is 3 and 4.