Ten-dimensional origin of nonperturbative large-volume-potential scaling

Use $c=\mathrm{\hslash }=1$. Let $X$ be a smooth compact Calabi–Yau threefold with a specified Type-IIB orientifold action and source data. Let ${\mathrm{\ell }}_s=\sqrt{{\alpha }^{\mathrm{\prime }}}$ and $L_s=2\pi {\mathrm{\ell }}_s$; $L_s$ is the string-length unit set to one in the cited ten-dimensional calculation. For a chosen string-frame Kähler metric $g_s$ with Kähler form $J_s$, define

${\mathcal{V}}_s=\frac{1}{6L_s^6}{\int }_X{J}_s\wedge J_s\wedge J_s,{\tau }_{s,s}=\frac{1}{2L_s^4}{\int }_{D_s}{J}_s\wedge J_s\mathrm{.}$

Let $\mathrm{\Phi }$ be the real dilaton profile, smooth away from the specified localized sources. If it is singular at a source, use a resolved source core or specify an integrable matching prescription that makes the displayed volume integrals well-defined. The corresponding ten-dimensional Einstein-frame internal metric is the Hermitian metric $g_E=e^{-\mathrm{\Phi }\mathrm{/}2}{g}_s$, with positive fundamental $(1,1)$-form $J_E:=e^{-\mathrm{\Phi }\mathrm{/}2}{J}_s$. When $\mathrm{\Phi }$ is constant this metric is Kähler; for a varying dilaton, $J_E$ need not be closed and is not generally a Kähler form. Define the Einstein-frame volume integral by

${\mathcal{V}}_E=\frac{1}{6L_s^6}{\int }_X{J}_E\wedge J_E\wedge J_E,{\tau }_{s,E}=\frac{1}{2L_s^4}{\int }_{D_s}{J}_E\wedge J_E\mathrm{.}$

In the constant-dilaton comparison regime $\mathrm{\Phi }={\mathrm{\Phi }}_{\mathrm{\infty }}$ and $g_{\mathrm{s}}=e^{{\mathrm{\Phi }}_{\mathrm{\infty }}}$, $J_E=g_{\mathrm{s}}^{-1\mathrm{/}2}{J}_s$, so

${\mathcal{V}}_E=g_{\mathrm{s}}^{-3\mathrm{/}2}{\mathcal{V}}_s,{\tau }_{s,E}=g_{\mathrm{s}}^{-1}{\tau }_{s,s}\mathrm{.}$

If $\mathrm{\Phi }$ varies over $X$, the pointwise conformal relation still defines $g_E$ and $J_E$, but the resulting volume ratios are no longer constant in general; the volume-scaling comparison below is restricted to the stated constant-dilaton regime. Fixed factors such as $2\pi$ and fixed powers of $g_{\mathrm{s}}$ affect coefficients and the numerical value assigned to a fixed divisor volume, but not powers of the total volume.

Here $D_s\subset X$ is a smooth complex divisor with no infinitesimal holomorphic deformations, $H^0(D_s,N_{D_s\mathrm{/}X})=0$, where $N_{D_s\mathrm{/}X}=TX{\mathrm{\mid }}_{D_s}\mathrm{/}T{D}_s$, and it is small in the Einstein-frame large-volume regime ${\mathcal{V}}_E\gg 1$ with ${\tau }_{s,E}\ll {\mathcal{V}}_E^{2\mathrm{/}3}$. A Euclidean D3 (E3) instanton is a D3-brane wrapped on $D_s$ and localized at a point of the noncompact four-dimensional spacetime. A D7 gauge sector is a four-dimensional gauge theory arising from a D7-brane wrapped on a divisor; no assumption that its gauge dynamics has the zero modes needed for a nonperturbative superpotential is made. The four-dimensional Einstein-frame potential $V_E$ is the scalar potential after the metric Weyl rescaling that fixes the Einstein–Hilbert coefficient to $M_P^2\mathrm{/}2$, with $M_P$ the reduced four-dimensional Planck mass.

For the four-dimensional reference description adopt reduced Planck units $M_P=1$. Thus $T_s={\tau }_{s,\mathrm{h}\mathrm{o}\mathrm{l}}+i{\rho }_s$, $a$, ${\mathcal{V}}_E$, and ${\tau }_{s,E}$ are dimensionless; $W_0$ and $A$ are the dimensionless Planck-unit representatives of quantities with mass dimension three, and $V_E$ is the dimensionless Planck-unit representative of a quantity with mass dimension four. The holomorphic Kähler coordinate $T_s$ is a complex scalar whose real part ${\tau }_{s,\mathrm{h}\mathrm{o}\mathrm{l}}$ enters terms such as $e^{-a{T}_s}$ and whose imaginary part ${\rho }_s$ is the normalized four-dimensional zero mode of the Ramond–Ramond four-form integrated over $D_s$, with its period fixed by the D3-charge convention used in $T_s$. At tree level the cited ten-dimensional Einstein-frame calculation identifies ${\tau }_{s,\mathrm{h}\mathrm{o}\mathrm{l}}={\tau }_{s,E}$ in its chosen $L_s$ units. Whether this identification receives a logarithmic or other correction is part of the open problem; ${\tau }_{s,\mathrm{h}\mathrm{o}\mathrm{l}}$ is not identified with the string-frame volume ${\tau }_{s,s}$ without the stated frame conversion. $W_0$ denotes the flux contribution to the four-dimensional superpotential, and $A{e}^{-a{T}_s}$ the one-instanton or gauge-condensation term, with $A$ its fluctuation prefactor and $a>0$ the dimensionless instanton-action or gauge-condensation coefficient. The instanton–flux interference term is the cross term between $W_0$ and $A{e}^{-a{T}_s}$; the instanton–anti-instanton term comes from a correlated pair with action factors $e^{-a{T}_s}$ and $e^{-a{\bar{T}}_s}$. A fluctuation determinant is the regulated product of nonzero eigenvalues of the quadratic fluctuation operators about that configuration. Collective coordinates are its bosonic and fermionic zero-mode parameters. Fermionic zero-mode lifting changes the effective action so the zero modes are no longer exact; zero-mode saturation supplies the Grassmann insertions needed for their integration to be nonzero.

The volume powers are compared asymptotically as ${\mathcal{V}}_E\to \mathrm{\infty }$ at fixed ${\tau }_{s,E}$, constant bulk string coupling, topology and specified discrete flux/source data. This is a comparison of terms in effective potentials, not an assertion that a family of stabilized vacua with these fixed data exists. In the conventional two-modulus four-dimensional large-volume-scenario (LVS) truncation, the instanton–flux interference and instanton–anti-instanton terms scale respectively as ${\mathcal{V}}_E^{-2}$ and ${\mathcal{V}}_E^{-1}$ under these conditions. For the E3-instanton realization, the 2026 local ten-dimensional effective-field-theory calculation estimates ${\mathcal{V}}_E^{-5\mathrm{/}2}$ and ${\mathcal{V}}_E^{-2}$, respectively, conditional on its local-propagation assumptions and on bulk-volume-independent local prefactors. Its determinant and correlated zero-mode interactions are not computed, so these are scaling estimates rather than a microscopic derivation. The paper treats D7 gaugino condensation separately: its local ten-dimensional power counting estimates the same two powers, while its alternative action-level matching recovers the four-dimensional powers only after inserting a volume-enhanced canonically normalized gaugino expectation value whose local origin is not derived. Thus the E3 and D7 comparisons remain distinct conditional arguments; no microscopic determinant, zero-mode or condensate normalization is assumed in advance.

For a type-IIB large-volume orientifold with a rigid small divisor supporting an E3 instanton or D7 gauge sector, compute from the localized ten-dimensional or microscopic theory the leading total-volume dependence of the instanton–flux interference and instanton–anti-instanton contributions to the four-dimensional Einstein-frame potential, including determinants, collective coordinates and zero-mode lifting. Determine whether and how these results match or correct the four-dimensional $N=1$ prediction when the tree-level holomorphic coordinate is identified with the ten-dimensional Einstein-frame divisor volume; any correction to that identification must be derived or otherwise independently justified.

Hebecker, Naturalness, String Landscape and Multiverse, §6.6

McAllister and Quevedo, Moduli Stabilization in String Theory, §4.2

Balasubramanian et al., Systematics of Moduli Stabilisation in Calabi-Yau Flux Compactifications, §§2–3

Hebecker et al., An attempt to derive the LVS from a 10d EFT perspective, §§3.2–3.3, 5 and 6

Status checked on 2026-10-26. The four-dimensional LVS potential has a specified three-term form under its stated Kähler-potential, superpotential, geometric, and large-volume assumptions. The 2026 ten-dimensional EFT analysis reports a different leading volume power for the localized instanton–flux and instanton–anti-instanton contributions under its local-propagation assumptions. This does not refute the conditional four-dimensional calculation.

The 2026 authors leave the interacting determinant and fermion zero-mode normalization uncomputed, and identify possible logarithmic corrections to the relation between the holomorphic coordinate and geometric divisor volume or long-range RR/Wilson-line propagation as unresolved routes to reconciliation. No microscopic derivation of the required correction is supplied. The matching problem therefore remains open; the stated source does not establish that the standard LVS branch is absent.