An attempt to derive the LVS from a 10d EFT perspective

Abstract

Non-perturbative scalar potentials are a central ingredient for stabilising moduli in flux compactifications. A key target in this context is the direct 10d derivation of the LVS AdS potential, i.e., the scalar potential of the Large Volume Scenario before the uplift. The familiar result, relying on 4d supergravity machinery and large volume expansion, consists of three terms of parametric form $\exp(-2aτ_s)/{\mathcal V}$, $W_0\exp(-aτ_s)/{\mathcal V}^2$, and $W_0^2/{\mathcal V}^3$. In plain 10d EFT language, the first term is an instanton-anti-instanton effect while the second term comes from a single instanton coupled to 3-form flux. When attempting to derive the suppression by the volume ${\mathcal V}$ directly from the 10d EFT, one encounters a puzzle: One expects a flux-dilution factor $1/\sqrt{\mathcal V}$ for the second term and a factor $1/{\mathcal V}^2$ from Weyl rescaling to the 4d Einstein frame for both the first and second term. Thus, the total volume suppression for these two terms is stronger than in the aforementioned 4d EFT expectations. We discuss the underlying logic in detail, also in the gaugino condensation case, but fail to find a satisfactory explanation for the mismatch. While log-corrections to the Kähler potential may be responsible, these would have to be strong enough to affect the outcome of the LVS scheme at the ${\mathcal O}(1)$ level. What is worse, the origin of the required correction remains unclear.