In this paper, we discuss the birational invariance of the class of balanced hyperbolic manifolds.
IntroductionIn his celebrated paper [Gro91], M. Gromov introduces an important notion called the Kähler hyperbolicity. It is pinched between the real hyperbolicity and the Kobayashi hyperbolicity [Kob98], and helps to settle the Kähler case of the Chern conjecture [Gro91]. After that, it leads to fruitful applications and improvements such as [CX01, CY18, Kol95, Eys97, Hit00, McM00] and so on.However, since the class of Kähler manifolds in general is not invariant under the birational transform, it would be desirable to have a birational variant of the Kähler hyperbolicity developed. It is an open problem posed by J. Kollár in [Kol95]. Kollár suggests to require Gromov's condition for a degenerated Kähler form, and [BDET22, BCDT24] introduce the weakly Kähler hyperbolicity by asking the cohomology class to be nef and big rather than Kähler. Weakly Kähler hyperbolic manifolds possess many key features as Kähler hyperbolic manifolds, and are invariant under the birational transform. Whereas in this paper, we are trying to investigate a more general situation, namely the balanced hyperbolicity.More precisely, let X be a compact complex manifold of dimension n. A Hermitian metric ω on X is called balanced if dω n−1 = 0. X is called a balanced manifold if it possesses a balanced metric. Obviously a Kähler metric must be balanced, but there do exist non-Kähler balanced metrics. Hence a balanced form is also regarded as a degenerated Kähler form in some aspects. A celebrated theorem in [AB95] asserts that the class of compact balanced manifolds is invariant under the birational transform, which directly inspires this paper. Now, let π : X → X be the universal cover, and fix a Riemannian metric g on X. Recall that a k-form α on X is called d-bounded, if there exists a (k − 1)-form β on X such that π * α = dβ and sup X β π * g < ∞. Note since X is compact, this notion is actually independent of the choice of g. We say that a Hermitian metric ω on X is balanced hyperbolic if ω is balanced and ω n−1 is d-bounded. It is notable that the balanced hyperbolicity was first introduced in [MP22, MP23], and we will continue this streamline to make a discussion in a wider range.Observe that if ω is balanced, [ω n−1 ] must be nef and big as an (n−1, n−1)-class. It allows us to talk about more degenerated cases. We say that a real smooth (1, 1)-form ω is semi-balanced hyperbolic if ω n−1 is d-closed, non-negative, strictly positive on a Zariski open set, and d-bounded. Fix an integer k | (n − 1), say kt = n − 1. We say that a real smooth (k, k)-form β is weakly