A non-vanishing conjecture for cotangent bundles on elliptic surfaces
Haesong Seo
TL;DR
This work resolves the non-vanishing conjecture for cotangent bundles on isotrivial elliptic surfaces, establishing that pseudoeffectivity of $\Omega_S^1$ forces the existence of nonzero symmetric differential sections $H^0(S, S^m\Omega_S^1)$ for some $m>0$. The authors adapt Höring–Peternell's framework by passing to a birational model with a finite cover removing orbifold divisors and then perform a detailed analysis of logarithmic and local obstructions arising from singular fibres of Kodaira types, including $II$, $III$, $IV$ and their stars. Key steps involve bounding the poles of logarithmic symmetric differentials along exceptional curves, embedding global sections into $(C\times E)$-models, and leveraging ramification bounds together with Riemann–Hurwitz, to deduce pseudoeffectivity of $\Omega_S^1$ if non-vanishing fails. Consequently, together with HP2020, the result completes the non-vanishing question for all surfaces with Kodaira dimension $\kappa(S)\le 1$, providing a full picture for cotangent bundles in this regime.
Abstract
In this paper, we prove the non-vanishing conjecture for cotangent bundles on isotrivial elliptic surfaces. Combined with the result by Höring and Peternell, it completely solves the question for surfaces with Kodaira dimension at most $1$.
