On some 2-dimensional Hermitian manifolds (Q1082621): Difference between revisions

Let (M,g,J) be a Hermitian manifold of complex dimension n and \(\Omega (X,Y)=g(X,JY)\) the Kähler form. Recall that g is called locally conformal Kähler (l.c.K.) if one can find an open covering \(M=\cup_{\alpha}U_{\alpha}\) such that \(g| U_{\alpha}=e^{\sigma_{\alpha}}g_{\alpha}\), where \(\sigma_{\alpha}\) is a differentiable function and \(g_{\alpha}\) is a Kähler metric on \(U_{\alpha}\). Let \(\Omega_{\alpha}\) be the Kähler form of \(g_{\alpha}\), \(\Omega_{\alpha}=e^{- \sigma_{\alpha}}\Omega\). The condition \(d\Omega_{\alpha}=0\) is equivalent to \(d\Omega =\omega \wedge \Omega\) with \(\omega | U_{\alpha}=d\sigma_{\alpha}\). From this identity one obtains \(\omega =(1/(n-1))i(\Omega)d\Omega\). Since \(\omega =d\sigma_{\alpha}\), one has \(d\omega =0.\) Suppose now that g is not necessarily l.c.K. and consider the 1-form \(\omega\) obtained from \(\Omega\) by the above expression. If \(\dim_ C M=2\) then one obtains easily that \(d\Omega =\omega \wedge \Omega\). \(\omega\) is called the Lee form of the Hermitian 2-manifold. \(d\omega =0\) is equivalent to say that M is l.c.K. The aim of the paper is to discuss 2-manifolds defined by other interesting properties of the Lee form \(\omega\). For this purpose the authors use a classification method due to \textit{A. Gray} and \textit{L. M. Hervella} [Ann. Mat. Pura Appl., IV. Ser. 123, 35-58 (1980; Zbl 0444.53032)]. For a unitary vector space (V,g,J) of real dimension 4 they consider the vector space \(\tau\) of the 2-covariant tensors having the same ''symmetries'' as \(\nabla \omega\). They write a (non-unique) decomposition of this space as sum \(\tau =\tau_ 1\oplus \tau_ 2\oplus \tau_ 3\oplus \tau_ 4\oplus \tau_ 5\) of 5 irreducible components of the natural action of U(2). In this way they obtain the following 5 classes of Hermitian manifolds: 1) \((\nabla_ x\omega)(Y)=\lambda g(X,Y)\). 2) \(d\omega =0\), \(\delta \omega =0\), \(L_ Ag=0\) (where \(A=-JB\) and B is such that \(g(B,X)=\omega (X)\), \(\forall X)\). 3) \(d\omega =0\), \(L_ B\Omega =0\). 4) \(L_ Bg=0\), \(L_ B\Omega =0\). 5) \(L_ Bg=0\), \(L_ Ag=0.\) The l.c.K. case is characterized by \(\nabla \omega \in \tau_ 1\oplus \tau_ 2\oplus \tau_ 3\). They prove, for example, that when M is compact and belongs to the classes (1), (2), (5) then one has necessarily \(\nabla \omega =0\) (generalized Hopf manifolds) and find examples, when M is not compact, of non-generalized Hopf manifolds belonging to classes (3) and (4).