Orthogonal complex structures on domains in \(\mathbb R^4\) (Q1006805): Difference between revisions

This very nice paper studies orthogonal complex structures in domains of \(\mathbb R^4\). Recall that a complex structure (which we always assume to be integrable) on Euclidean space is called orthogonal if it is compatible with the Euclidean metric. The existence of an orthogonal complex structure on a domain is equivalent to a conformally invariant first-order system of partial differential equations. The goal of the paper is to prove Liouville-type uniqueness theorems for this system. The first result says that if \(J\) is an orthogonal complex structure defined in \(\mathbb{R}^4\backslash \Lambda\) where \(\Lambda\) is a closed set with zero one-dimensional Hausdorff measure, then either \(J\) is constant or it can be maximally extended to the complement of just one point. In either case, \(J\) is the image of the standard complex structure on the \(4\)-sphere under a conformal transformation. The second result says that if \(\Lambda\) as above is a round circle or a straight line, and \(J\) is not conformal to a constant complex structure, then \(J\) is unique up to sign and cannot be further extended. These complex structures are all conformal (up to sign), and are induced from a real quadric in \(\mathbb{CP}^3\) via the twistor construction. The authors give two applications of their results: they give a new proof of \textit{M. Pontecorvo}'s classification of compact locally conformally flat Hermitian surfaces [Differ. Geom. Appl. 2, No. 3, 295--305 (1992; Zbl 0766.53052)], and they classify nondegenerate quadrics in \(\mathbb{CP}^3\) up to the action of the conformal group.