F-theory Flux Vacua and Attractor Equations
Extracted source text ยท arXiv 1910.10725
\documentclass[11pt]{article}
\pdfoutput=1
\usepackage{amsmath,amsthm,amsfonts,amssymb,srcltx,empheq}
%\usepackage[dvipdfmx]{graphicx}
\usepackage{graphicx}
\usepackage[all]{xy}
\usepackage{color}
\usepackage{longtable}
\usepackage{hhline}
\usepackage{caption}
\captionsetup[figure]{format=plain, labelformat=simple, labelsep=period, font=small}
\captionsetup[table]{format=plain, labelformat=simple, labelsep=period, font=small}
\newcommand{\e}{{\rm e}}
\def\IN{\mathbb {N}}
\def\IZ{\mathbb {Z}}
\def\IQ{\mathbb {Q}}
\def\IR{\mathbb {R}}
\def\IC{\mathbb {C}}
\def\IP{\mathbb {P}}
\def\IG{\mathbb {G}}
\def\ICP{\mathbb {CP}}
\def\IRP{\mathbb {RP}}
\def\rank{\operatorname{rank}}
\def\Sym{\operatorname{Sym}}
\theoremstyle{plain}
\newtheorem{prob}{Problem}[section]
\newtheorem{conj}[prob]{Conjecture}
\newtheorem{ques}[prob]{Question}
\newtheorem{thm}[prob]{Theorem}
\newtheorem{prop}[prob]{Proposition}
\newtheorem{cor}[prob]{Corollary}
\newtheorem{lemm}[prob]{Lemma}
\newtheorem{defi}[prob]{Definition}
\theoremstyle{remark}
\newtheorem{remark}[prob]{\bf Remark}
\newtheorem{exam}[prob]{\bf Example}
\renewcommand{\thefootnote}{\fnsymbol{footnote}}
\topmargin=-45pt
\oddsidemargin=0cm
\evensidemargin=0cm
\textheight=23.0cm
\textwidth=16.5cm
\parskip=0.5em
\makeatletter
\def\Left#1#2\Right{\begingroup%
\def\ts@r{\nulldelimiterspace=0pt \mathsurround=0pt}%
\let\@hat=#1%
\def\sht@im{#2}%
\def\@t{{\mathchoice{\def\@fen{\displaystyle}\k@fel}%
{\def\@fen{\textstyle}\k@fel}%
{\def\@fen{\scriptstyle}\k@fel}%
{\def\@fen{\scriptscriptstyle}\k@fel}}}%
\def\g@rin{\ts@r\left\@hat\vphantom{\sht@im}\right.}%
\def\k@fel{\setbox0=\hbox{$\@fen\g@rin$}\hbox{%
$\@fen \kern.3875\wd0 \copy0 \kern-.3875\wd0%
\llap{\copy0}\kern.3875\wd0$}}%
\def\pt@h{\mathopen\@t}\pt@h\sht@im%
\Right}%
\def\Right#1{\let\@hat=#1%
\def\st@m{\mathclose\@t}%
\st@m\endgroup}
\makeatother
\makeatletter
\renewcommand{\theequation}{%
\thesection.\arabic{equation}}
\@addtoreset{equation}{section}
\makeatother
\renewcommand{\baselinestretch}{1.3}
\makeatletter
\def\eqnarray{%
\stepcounter{equation}%
\let\@currentlabel=\theequation
\global\@eqnswtrue
\global\@eqcnt\z@
\tabskip\@centering
\let\\=\@eqncr
$$\halign to \displaywidth\bgroup\@eqnsel\hskip\@centering
$\displaystyle\tabskip\z@{##}$&\global\@eqcnt\@ne
\hfil$\displaystyle{{}##{}}$\hfil
&\global\@eqcnt\tw@$\displaystyle\tabskip\z@{##}$\hfil
\tabskip\@centering&\llap{##}\tabskip\z@\cr}
\makeatother
\begin{document}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{titlepage}
\hfill{KEK-TH-2160}
\begin{center}
\vspace*{1cm}
{\Large \bf
F-theory Flux Vacua and Attractor Equations
}
\vskip 1.5cm
{\large Yoshinori Honma${}^{a}$\footnote[2]{yhonma@law.meijigakuin.ac.jp} and
Hajime Otsuka${}^b$\footnote[3]{hotsuka@post.kek.jp}}
\vskip 1.0cm
{\it
${}^a$%
Department of Current Legal Studies, \\
Meiji Gakuin University, \\
Yokohama, Kanagawa 244-8539, Japan \\
\vspace*{0.5cm}
${}^b$%
Theory Center, Institute of Particle and Nuclear Studies, \\
High Energy Accelerator Research Organization, KEK, \\
1-1 Oho, Tsukuba, Ibaraki 305-0801, Japan\\}
\end{center}
\vskip2.5cm
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{abstract}
We examine the vacuum structure of 4D effective theories of moduli fields
in spacetime compactifications with quantized background fluxes. Imposing the
no-scale structure for the volume deformations, we numerically investigate the
distributions of flux vacua of the effective potential in complex structure
moduli and axio-dilaton directions for two explicit examples in Type IIB string
theory and F-theory compactifications. It turns out that distributions of
non-supersymmetric flux vacua exhibit a non-increasing functional behavior of
several on-shell quantities with respect to the string coupling.
We point out that this phenomena can be deeply connected with a previously-reported
possible correspondence between the flux vacua in moduli stabilization problem and the
attractor mechanism in supergravity, and our explicit demonstration implies that
such a correspondence generically exist even in the framework of F-theory. In
particular, we confirm that the solutions of the effective potential we explicitly
evaluated in Type IIB and F-theory flux compactifications indeed satisfy the
generalized form of the attractor equations simultaneously.
\end{abstract}
\end{titlepage}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\renewcommand{\thefootnote}{\arabic{footnote}} \setcounter{footnote}{0}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Introduction}\label{sec:introduction}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Stabilization of extra massless scalar moduli fields of 4D effective theories arising
from spacetime compactifications is of particular interest in the field of string theory
and its broad applications to cosmology and phenomenology. In particular, remarkable
progress for cosmological observations in recent years has motivated us to realize de
Sitter spacetime with stabilized moduli in the framework of string theory and contrive a
concrete setup in which basic quantum gravitational issues can be thoroughly investigated.
In the study of this moduli stabilization problem, one looks at the scalar potential of
4D ${\cal{N}}=1$ effective theories arising from spacetime compactifications. In the
language of 4D ${\cal{N}}=1$ supersymmetry, there are two kinds of contributions to the
scalar potential of moduli fields, namely the K$\ddot{{\textrm{a}}}$hler potential and
the superpotential. String compactifications enables us to derive these quantities quantum
mechanically from the geometry of internal compact spaces, often taken to be a Calabi-Yau
manifold to use sufficient amount of supersymmetry while keeping the problem nontrivial.
Inclusion of background fluxes is a crucial ingredient to generate interactions between
moduli fields in superpotential, leading to their stabilization in a controllable way.
To grasp the background and clarify our motivation for the present paper in more detail,
let us begin with the 4D effective action arising from Type IIB Calabi-Yau orientifolds
or the F-theory on Calabi-Yau fourfolds \cite{Vafa:1996xn}, both of which can be described
by the 4D ${\cal{N}}=1$ supergravity language. Suppose the K$\ddot{{\textrm{a}}}$hler moduli
fields labeled by $\alpha$ are absent in the superpotential $W$, and also the
K$\ddot{{\textrm{a}}}$hler potential $K$ satisfies the condition $K^{\alpha \bar{\alpha}}
\partial_{\alpha} K \partial_{\bar{\alpha}} \overline{K}=3$, the $F$-term scalar potential
of moduli fields becomes a no-scale type
\begin{align}
V = e^K\left( K^{I \bar{J}} D_I W D_{ \bar{J}} \overline{W} \right),
\label{nosv}
\end{align}
in the reduced Planck mass unit $M_{\rm Pl}=1$. Here $I,J$ run over the complex structure
moduli of a Calabi-Yau threefold and the axio-dilaton in a Type IIB setup, which can be
also regarded as a unified complex structure moduli of a Calabi-Yau fourfold from the
F-theory perspective. Here $D_{I} \equiv \partial_I + (\partial_I K)$ denote the
K$\ddot{{\textrm{a}}}$hler covariant derivative and $K^{I \bar{J}}$ is the inverse of the
K$\ddot{{\textrm{a}}}$hler metric $K_{I \bar{J}} = \partial_I \partial_{\bar{J}} K$.
Throughout this paper, we focus on this no-scale type potential.
It is well known that the imaginary self-dual three-form fluxes in Type IIB string theory
or the self-dual four-form fluxes in F-theory can generate suitable flux superpotential
$W$ to stabilize moduli fields at $D_I W=0$, leading to stable Minkowski
minima \cite{Giddings:2001yu,Gukov:1999ya,Dasgupta:1999ss}. While this prescription completely
fix the complex structure moduli and axio-dilaton, the K$\ddot{{\textrm{a}}}$hler moduli of
the compactification remain unfixed. In order to fix all the moduli and construct realistic
cosmologies describing the accelerated expansion of the universe, quantum $\alpha'$-corrections
and non-perturbative effects should be further included. This issue has been resolved in the
explicit constructions of the Kachru-Kallosh-Linde-Trivedi (KKLT) model \cite{Kachru:2003aw}
and the LARGE Volume Scenario (LVS) \cite{Balasubramanian:2005zx,Conlon:2005ki}. There the
small value of de Sitter cosmological constant for the late-time cosmology has been realized
by an inclusion of the anti-D3-brane inducing an uplifting potential.
On the other hand, as has been pointed out for instance in \cite{Saltman:2004sn}, effective
theories equipped with the no-scale type potential can also accommodate supersymmetry-breaking
minima with $V \neq 0$ in complex structure moduli and axio-dilaton directions. There the
spontaneous supersymmetry breaking effect due to a small vacuum energy can be regarded as a
variant of the anti-D-branes to produce tiny cosmological constant. In the same spirit,
several supersymmetry-breaking scenarios has been discussed, for instance,
in \cite{Gallego:2017dvd,Blaback:2015zra}. While such an uplifting mechanism may have broad
applications to de Sitter model building, comprehensive study has not been fully elucidated,
possibly due to the complexity of analysis of the non-supersymmetric vacua. One of our goals
in the present paper is to fill this gap and understand the characteristics of the flux vacua
in general compactifications.
Specifically, we investigate the vacuum structure of two explicit examples in Type IIB string
theory and F-theory compactifications while assuming the no-scale structure (\ref{nosv}).\footnote{Note that, as we will discuss later, the no-scale analysis which ignores the K\"{a}hler moduli stabilization does not support the accelerated expansion in itself, and further modifications of the models are required if one wants to find out a landscape of explicit de Sitter minima following our approach.}
By solving the system numerically and finding the non-supersymmetric flux vacua in both setups,
we find that several on-shell quantities exhibit a non-increasing functional
behavior with respect to the string coupling. In the process of clarifying an underlying dynamics of
this phenomena, we find that this characteristic behavior of flux vacua can be regarded as a
direct consequence of a previously-studied possible correspondence between moduli stabilization
problem and attractor mechanism in supergravity emphasized in \cite{Kallosh:2005ax} (see
also \cite{Kallosh:2005bj,Kallosh:2006bt,Alishahiha:2006jd,Bellucci:2007ds,Larsen:2009fw}), and our
demonstration provides a nontrivial supporting evidence for this kind of correspondence.
Especially, one can check that our numerical solutions as well as analytic Minkowski solutions
indeed satisfy the appropriately generalized attractor equation simultaneously, even in the
framework of F-theory.
This paper is organized as follows. We first take a brief look at an effective theory arising
from Type IIB toroidal orientifold compactification and study its vacuum structure in Section
\ref{sec:typeIIB}. In Section \ref{sec:Fth}, we move on to analyze the effective potential of
moduli fields arising from F-theory compactified on a Calabi-Yau fourfold. Finally in
Section \ref{sec:Att}, we reconsider our results from another viewpoint and confirm that the
moduli stabilization problem discussed in Sections \ref{sec:typeIIB} and \ref{sec:Fth} can be
rephrased in terms of the attractor mechanism in supergravity. Section \ref{sec:conclusion}
is devoted to conclusions and discussions. In Appendix \ref{app:attF}, we describes the F-theory
generalization of the attractor equation.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Type IIB Toroidal Orientifold}\label{sec:typeIIB}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
First we consider a well-known Type IIB toroidal orientifold model and investigate its vacuum
structure. Explicit demonstrations in this simple setup would make the arguments in subsequent
sections more intelligible.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Setup}\label{subsec:example1}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Here we briefly look at the effective potential of moduli fields in Type IIB string theory
compactified on a $T^6/\mathbb{Z}_2$ toroidal orientifold. We refer the reader
to \cite{Kachru:2002he,Frey:2002hf} for the details about the construction.
Let us represent the periodic six real coordinates on $T^6$ by $x^{i}$ and $y^i$ with
$i,j=1,2,3$, and take the holomorphic one-forms of the $T^2$ submanifolds as
$dz^i=dx^i +\tau^{ij}dy^j$. By choosing the orientation as $\int_{T^6} dx^1\wedge dx^2\wedge
dx^3 \wedge dy^1 \wedge dy^2 \wedge dy^3=1$ and representing the three-form cohomology basis
in $H^3(T^6, \mathbb{Z})$ as
\begin{align}
\begin{split}
\alpha_0 &= dx^1 \wedge dx^2 \wedge dx^3, \ \ \ \ \ \ \ \ \ \ \ \alpha_{ij} = \frac{1}{2}
\epsilon_{ikl} dx^k \wedge dx^l \wedge dy^j, \\ \beta^{ij} &= -\frac{1}{2} \epsilon_{jkl}
dy^k \wedge dy^l \wedge dx^i, \ \ \ \ \beta_0 = dy^1 \wedge dy^2 \wedge dy^3,
\end{split}
\end{align}
the holomorphic three-form $\Omega \equiv dz^1 \wedge dz^2 \wedge dz^3$ on $T^6/\mathbb{Z}_2$
can be described as
\begin{align}
\Omega =\alpha_0 +\alpha_{ij}\tau^{ij} -\frac{1}{2}\beta^{ij}(\epsilon_{ikl}
\epsilon_{jmn}\tau^{km}\tau^{ln})+\beta^0 ({\rm det}\tau).
\end{align}
Here the matrices $\tau^{ij}$ describe the complex structure moduli of the geometry and the
three-form cohomology basis is normalized as $\int_{T^6} \alpha_I \wedge \beta^J =\delta_I^J$.
The three-form fluxes $F_3$ and $H_3$ in RR and NSNS sectors in Type IIB string theory can be
expanded in the same basis as
\begin{align}
\begin{split}
&\frac{1}{(2\pi)^2 \alpha'}F_3 = a^0 \alpha_0 +a^{ij}\alpha_{ij} +b_{ij}\beta^{ij} +b_0 \beta^0,
\\
&\frac{1}{(2\pi)^2 \alpha'}H_3 = c^0 \alpha_0 +c^{ij}\alpha_{ij} +d_{ij}\beta^{ij} +d_0 \beta^0,
\end{split}
\end{align}
where the coefficients $a^0,a^{ij},b_0,b_{ij},c^0,c^{ij},d_0,d_{ij} \in 2\mathbb{Z}$
describe the quanta of background fluxes. Note that here we focus on even numbers of flux
quanta to prohibit the appearance of delicate exotic O3-planes \cite{Frey:2002hf}.
The explicit form of the K$\ddot{{\textrm{a}}}$hler potential for moduli fields is given by
\begin{align}
K=-\ln{\left[ -i(S-\overline{S}) \right]} -\ln{\left[ -i\int \Omega \wedge \bar{\Omega}
\right]} -2\ln{\cal{V}},
\label{torK}
\end{align}
where $S$ denotes the axio-dilaton. The hermitian norm of $\Omega$ corresponds to the
Weil-Petersson metric of the complex structure moduli space and ${\cal V}$ is a volume of
the internal manifold $T^6/\mathbb{Z}_2$ in the Einstein-frame, measured in units of
$2\pi \sqrt{\alpha^\prime}$. Plugging (\ref{torK}) into the scalar potential (\ref{nosv}),
one finds that $V$ no longer depends on the K$\ddot{{\textrm{a}}}$hler moduli fields
except for a prefactor $\frac{1}{{\cal V}^2}>0$ treated as a constant throughout this paper.
On the other hand, the superpotential induced by background fluxes in Type IIB
compactifications\footnote{Here we use a convention where the factor $(4\pi)^{-1/2}$ appear
in the superpotential to match with the dimensional reduction of Type IIB supergravity
action.} takes the following form \cite{Gukov:1999ya}
\begin{align}
W = \frac{1}{\sqrt{4\pi}(2\pi)^2 \alpha'}\int \Omega \wedge G_3,
\end{align}
where $G_3 \equiv F_3-SH_3$ denotes the combined three-form fluxes. On $T^6/\mathbb{Z}_2$
geometry, this becomes
\begin{align}
\begin{split}
W=\frac{1}{\sqrt{4\pi}} \Bigg( &(a^0 -Sc^0){\rm det}\tau -\frac{1}{2}(a^{ij} -Sc^{ij})
\epsilon_{ikm}\epsilon_{jpq}\tau^{kp}\tau^{mq} \\
& -(b_{ij} -Sd_{ij})\tau^{ij} -(b_0 -Sd_0) \Bigg) .
\end{split}
\end{align}
Recalling that there exist $2^6$ O3-planes on $T^6/{\mathbb Z}_2$ orientifold, the tadpole
cancellation condition
\begin{align}
32 -2n_{{\rm D}3}= \frac{1}{(2\pi)^4 (\alpha')^2}\int H_3 \wedge F_3 = c^0b_0 +c^{ii}b_{ii}
-d_{ii}a^{ii} -d_0a^0,
\label{tadcan}
\end{align}
must be also satisfied to ensure the global conservation of fluxes inside the compact manifold.
Here $n_{{\rm D}3}$ denotes the number of mobile D3-branes, which is set to be zero in the
following discussions.
To simplify the analysis, we further focus on an isotropic case for the system, namely we only
consider diagonal components of $\tau^{ij}$ and fluxes satisfying the following condition
\begin{align}
\begin{split}
\tau^{11} &=\tau^{22} =\tau^{33} \equiv \tau, \\
a^{11} &=a^{22}=a^{33} \equiv a^1, \\
b_{11} &=b_{22}=b_{33} \equiv b_1, \\
c^{11} &=c^{22}=c^{33} \equiv c^1, \\
d_{11} &=d_{22}=d_{33} \equiv d_1,
\end{split}
\end{align}
as well as the fluxes $\{ a^0, b_0, c^0, d_0 \}$. Then the K$\ddot{{\textrm{a}}}$hler potential
(\ref{torK}) of the $T^6/{\mathbb Z}_2$ orientifold model reduces to be
\begin{align}
K=-\ln{\left[ (S-\overline{S})(\tau-\overline{\tau})^3 \right]}-2\ln{\cal{V}},
\label{isok}
\end{align}
and the flux-induced superpotential becomes
\begin{align}
W = (4\pi)^{-1/2}\left( (a^0 -S c^0) \tau^3 -3 (a^1 -S c^1) \tau^2 -3(b_1 -S d_1)\tau
-(b_0 - S d_0)\right).
\label{isow}
\end{align}
Combining the above setup with the definition of the no-scale type potential (\ref{nosv}),
one can study the vacuum structure of the model explicitly. Here we consider a condition
\begin{align}
a^0=c^1=b_1=d_0=0,
\end{align}
and non-zero otherwise, which corresponds to the imaginary self-dual condition allowing (2,1)
and (0,3) piece for $F_3$ and the imaginary anti-self-dual condition allowing (3,0) and (1,2)
piece for $H_3$. This asymmetric choice of background fluxes was chosen to completely fix the
complex structure moduli and axio-dilaton, coming from the fact that the $T^6/{\mathbb Z}_2$
toroidal orientifold model originally preserves 4D ${\cal{N}}=4$ supersymmetry.\footnote{See for
instance \cite{DAuria:2002qje} about 4D ${\cal{N}}=4$ gauged supergravity description of the
model.} In this setup, the model has a Minkowski minimum\footnote{In order to be supersymmetric,
the minimum also need to satisfy $W=0$ and allowed fluxes are further constrained.} as a solution
to the $F$-term conditions $D_I W=0$, where $V$ becomes zero and the values of moduli fields are
fixed as
\begin{align}
{\textrm{Re}}\tau ={\textrm{Re}}S=0, \ \ \ \ {\textrm{Im}}\tau =\left(\frac{b_0 {d_1}'}{a^1 c^0}
\right)^{1/4}, \ {\textrm{Im}} S =\left( \frac{(a^1)^3 b_0}{c^0 {d_1}'^3}\right)^{1/4},
\label{torvacuum}
\end{align}
where ${d_1}' \equiv -d_1$. On the other hand, when we consider a condition
\begin{align}
c^0=a^1=d_1=b_0=0,
\end{align}
and non-zero otherwise, this choice corresponds to the imaginary self-dual condition for $H_3$ and
the imaginary anti-self-dual condition for $F_3$. Then the model has a Minkowski minimum at which
the values of moduli fields are fixed as
\begin{align}
{\textrm{Re}}\tau &={\textrm{Re}}S=0, \ \ \ \ {\textrm{Im}}\tau =\left(\frac{{b_1}' {d_0}'}{a^0
{c^1}'}\right)^{1/4}, \ {\textrm{Im}}S=\left( \frac{a^0 ({b_1}')^3}{({c^1}')^3 {d_0}'}\right)^{1/4},
\label{torvac2}
\end{align}
where ${b_1}'\equiv -b_1, {c^1}'\equiv -c^1$ and ${d_0}' \equiv -d_0$. Later we will utilize these
simple analytic solutions (\ref{torvacuum}) and (\ref{torvac2}) when we discuss about a
correspondence between the moduli stabilization and a seemingly different topic in supergravity.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Distribution of non-supersymmetric flux vacua}
\label{subsec:resultIIB}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Having introduced the setup of the $T^6/\mathbb{Z}_2$ toroidal orientifold model, now we turn to
analyze the non-supersymmetric flux vacua with fixed moduli $\Phi_I \equiv \{ \tau^{ij}, S \}$,
while taking into account the tadpole cancellation condition for quantized background fluxes
appropriately.
As emphasized in \cite{Saltman:2004sn}, the no-scale type potential (\ref{nosv}) can also
accommodate the supersymmetry-breaking minima characterized by
\begin{align}
(\partial_I V)|_{\Phi_I^0} =0, \qquad
V|_{\Phi_I^0} =\frac{\epsilon}{{\cal V}^2}>0,
\label{nonsv}
\end{align}
with the mass squared of the moduli fields being strictly positive. Note that the nonzero
vacuum energy at $\Phi_I^0$ is proportional to a constant Calabi-Yau volume arising from
the prefactor $e^K$ in the scalar potential, as mentioned in the previous section. Numerical
constant $\epsilon$ determined by fluxes can take sufficiently small value owing to the
richness of the choice of background fluxes, and this becomes a candidate of the source of
an uplifting potential. Throughout this paper, we only pick up the solutions satisfying
$\epsilon <1$ as true flux vacua, otherwise the effective theory description of the system
turns out to be unreliable.
In our numerical analysis, we utilize the ``FindMinimum'' function in Mathematica to find
supersymmetry-breaking minima of the effective potential.\footnote{More precisely, we
performed the analysis with working precision of 50 digits and accuracy goal of 20 digits.
We also checked that our results are all insensitive to such parameter choices and pick up
the accurate solutions only.} To simplify the analysis, let us again focus on the isotropic
case for $\tau^{ij}$ and background fluxes in which the the K$\ddot{{\textrm{a}}}$hler
potential and the superpotential are given by (\ref{isok}) and (\ref{isow}), respectively.
Within a flux range $-12\leq a^0,a^1,b_0,b_1,c^0,c^1,d_0,d_1\leq 12$ while imposing the
tadpole cancellation condition (\ref{tadcan}) with $n_{\rm D3}=0$, we found 962
supersymmetry-breaking minima with no tachyonic instability.
\begin{figure}[ht]
\begin{minipage}{0.5\hsize}
\begin{center}
\includegraphics[width=80mm]{T61.pdf}
\end{center}
\end{minipage}
\begin{minipage}{0.5\hsize}
\begin{center}
\includegraphics[width=80mm]{T62.pdf}
\end{center}
\end{minipage}
\caption{Numerical results for the distribution of flux vacua with nonzero vacuum energy.
Each dot corresponds to a solution with different set of fluxes. Note that, however,
generically these dots can be degenerated in this 2D plot, due to the existence of another
direction of the stabilized moduli $\tau$ which is not depicted here for the simplicity.
The right figure represents the log-log plot of the same set of solutions.}
\label{fig:T6}
\end{figure}
The result is depicted in
Figure \ref{fig:T6}, where we plot the height of the normalized potential
$\epsilon ={\cal V}^2V$ with respect to the string coupling $g_s \equiv ({\rm Im}S)^{-1}$
at the each of the minima. The other on-shell quantities are also shown in Figure \ref{fig:T3}.
Note that generically an ambiguity originating from the
Type IIB SL(2,$\mathbb{Z}$) duality transformation generates physically-equivalent solutions
and a careful analysis is required. Here and in what follows, we deduct this
ambiguity\footnote{For the present example, there exists another type of redundancy arising
from SL(2,$\mathbb{Z}$) $\subset$ SL(6,$\mathbb{Z}$) transformation acting on $T^6$ and we
have also deducted this ambiguity by requiring $|\textrm{Re} \tau| \leq 1/2$ and $|\tau|
\geq 1$.} by selecting only the solutions satisfying $|\textrm{Re} S| \leq 1/2$ and $|S|
\geq 1$ while keeping the string coupling small but finite: $0 < g_s <1$.
\clearpage
To provide typical examples of our numerical results, here we pick up three independent
flux vacua determined under the following choice of fluxes:
\begin{table}[ht]
\centering
\begin{tabular}{|c|c|} \hline
Vacuum & Set of fluxes $(a^0, a^{1},b_0, b_1, c^0, c^1, d_0, d_1)$ \\ \hline \hline
A & $(-2, 4, 0, -4, 0, 0, 4, -2)$ \\ \hline
B & $(4, -4, 2, 2, 0, 0, -2, 2)$ \\ \hline
C & $(-2, 2, 4, 4, 0, 0, 10, -2)$ \\ \hline
\end{tabular}
\end{table}
\noindent The vacuum A, B and C are represented in Figure \ref{fig:T6} by the red, green and blue
points, respectively. Explicit values of stabilized moduli and various on-shell quantities at the
each of these flux vacua are summarized in Tables~\ref{tab:T61} and~\ref{tab:T62}.
\begin{table}[ht]
\centering
\begin{tabular}{|c|c|c|c|c|c|} \hline
Vacuum & $\tau$ & $S$ & $g_s$ & $\epsilon$ & $W$ \\ \hline \hline
A & $0.355+1.64 i$ & $0.271 + 3.98 i$ & 0.251 & $1.52\times 10^{-2}$ & $22.6 +5.08i$ \\ \hline
B & $0.0468 +1.19 i$ & $0.493 + 2.84 i$ & 0.352 & $6.99\times 10^{-2}$ & $-11.6 -3.90i$ \\ \hline
C & $-0.393+2.48 i$ & $-0.440 + 2.29 i$ & 0.438 & $8.23\times 10^{-2}$ & $14.3 +12.6i$ \\ \hline
\end{tabular}
\caption{Explicit values of stabilized moduli fields and on-shell quantities in $M_{\rm Pl}=1$
unit.}
\label{tab:T61}
\end{table}
\begin{table}[ht]
\centering
\begin{tabular}{|c|c|} \hline
Vacuum & Eigenvalues of mass matrix $\partial_I \partial_J V \times {\cal V}^{2}$
\\ \hline \hline
A & (3.41, 0.603, 0.222, $1.91\times 10^{-2}$) \\ \hline
B & (6.66, 1.30, 0.324, $6.85\times 10^{-2}$)\\ \hline
C & (2.18, 0.690, $7.29\times 10^{-2}$, $9.76\times 10^{-4}$)\\ \hline
\end{tabular}
\caption{Mass eigenvalues of normalized scalar potential $\epsilon$ in $M_{\rm Pl}=1$ unit.
Note that we have diagonalized the mass matrix ${\cal V}^{2} \partial_I \partial_J V$ by taking
the linear combinations of the basis and the eigenvalues are displayed in the descending order.}
\label{tab:T62}
\end{table}
\begin{figure}[htb]
\begin{minipage}{0.5\hsize}
\begin{center}
\includegraphics[width=80mm]{T63.pdf}
\end{center}
\end{minipage}
\begin{minipage}{0.5\hsize}
\begin{center}
\includegraphics[width=80mm]{T64.pdf}
\end{center}
\end{minipage}
\caption{Numerical results for the absolute values of flux superpotential $|W|$
and a quantity $|{\cal Z}|=e^{K/2}|W|$ at the each of the solutions depicted in
Figure \ref{fig:T6}.}
\label{fig:T3}
\end{figure}
From Figures \ref{fig:T6} and \ref{fig:T3}, one finds that the flux vacua of the model seems to
be determined such that several on-shell quantities are governed by a nontrivial dynamics, rather
than just providing a completely random distribution. Especially, the distribution of a normalized
superpotential $|{\cal Z}|=e^{K/2}|W|$ indicates a monotonic, non-increasing functional dependence
with respect to the string coupling. As we will see in the next section, another concrete example
based on F-theory flux compactification exhibits this kind of dependence of on-shell quantities
more clearly. Therefore let us save the arguments for a physical background of this suggestive
behavior and its consequences for later discussions in the final section, after looking at the
F-theory example and its vacuum structure.
Here we provide some comments on a further consistency check condition to be satisfied,
which can be straightforwardly applied to the F-theory compactification described in the
next section. To ensure the validity of the effective theory description of the system in
the framework of spacetime compactifications, the vacuum energy at the each of the minima
needs to be smaller than the mass scales of the Kaluza-Klein and the stringy modes as
\begin{align}
\langle V \rangle \ll m_s^4, m_{\rm KK}^4.
\end{align}
In the Type IIB language, these scales are given by
$m_s=(\alpha^\prime)^{-1/2} \simeq \frac{g_s^{1/4} \sqrt{\pi}}{ {\cal V}^{1/2}}$ and
$m_{\rm KK} \simeq \frac{\sqrt{\pi}}{ {\cal V}^{2/3}}$ in the reduced Planck mass unit.
Therefore the above conditions can be rephrased as a requirement that the dimensionless
constant $\epsilon$ must satisfy the condition
\begin{align}
\epsilon \ll {\rm min}({\cal V}^{-2/3}\pi^2 , g_s \pi^2).
\label{eq:epsilon}
\end{align}
For instance if we take a sufficiently large volume ${\cal V}$, e.g. larger than $\simeq
{\cal{O}}(10^2)$, one can estimate that all the numerical solutions summarized in
Figure \ref{fig:T6} satisfy (\ref{eq:epsilon}). Of course in order to conduct a detailed
analysis, ${\cal V}$ needs to be completely fixed by considering an extension of the
K$\ddot{{\textrm{a}}}$hler moduli sector beyond our no-scale setup\footnote{No-scale
potential in itself generically provides a runaway potential for the
K$\ddot{{\textrm{a}}}$hler moduli and does not support the accelerated expansion
\cite{Frey:2002qc,Dine:1985he}, which requires the inclusion of quantum corrections to
the K$\ddot{{\textrm{a}}}$hler moduli sector. This means that after doing this kind of
extension in our Type IIB and F-theory setups, the exact relationship between our analysis
and the de Sitter swampland conjecture \cite{Obied:2018sgi,Garg:2018reu,Ooguri:2018wrx} would come into
view directly.} in (\ref{torK}).
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{F-theory Model}
\label{sec:Fth}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Let us move on to discuss the 4D ${\cal N}=1$ effective theory based on F-theory flux
compactification on a Calabi-Yau fourfold $X_4$. After introducing the setup of a model
constructed in \cite{Honma:2017uzn}, we numerically analyze the distribution of
supersymmetry-breaking minima of its scalar potential in the same manner as performed in the
previous section.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Setup}
\label{subsec:setupF}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
From the perspective of the F-theory framework, the K$\ddot{{\textrm{a}}}$hler potential for
complex structure moduli space of a Calabi-Yau fourfold $X_4$ is defined by
\begin{align}
K=-\ln{\int_{X_4}} \Omega \wedge \overline{\Omega},
\label{Kahp}
\end{align}
where $\Omega$ denotes a holomorphic $(4,0)$-form on $X_4$. The F-theory compactification
also generically admits a superpotential of the form \cite{Gukov:1999ya}
\begin{align}
W = (4\pi)^{-1/2}\int_{X_4} G_4 \wedge \Omega,
\end{align}
in the presence of four-form fluxes $G_4$, which is inherited from a duality between F-theory
and M-theory on the same manifold \cite{Becker:1996gj,Sethi:1996es,Haack:2001jz,Denef:2008wq}.
As is the case in the Type IIB orientifold model, background fluxes are required to satisfy
the tadpole cancellation condition given by
\begin{align}
\frac{\chi}{24}=n_{\rm{D3}}+\frac{1}{2}\int_{X_4} G_4 \wedge G_4,
\end{align}
in order to be conserved within a compact manifold $X_4$ globally. Here $\chi$ is the Euler
characteristic of $X_4$ and $n_{\rm{D3}}$ denotes the total number of the mobile D3-branes.
For a Calabi-Yau fourfold $X_4$ with $h^{3,1}(X_4)$ complex structure moduli, the period
integrals of holomorphic $(4,0)$-form defined by
\begin{align}
\Pi_i = \int_{\gamma^i} \Omega,
\label{4peri}
\end{align}
encode the moduli dependence of the system, and the K$\ddot{{\textrm{a}}}$hler potential for
complex structure moduli (\ref{Kahp}) can be rewritten as
\begin{align}
K=-\ln{\left[ \sum_{i,j} \Pi_i \eta^{ij} \overline{\Pi}_j \right]},
\label{Kper}
\end{align}
where $\gamma^i$ with $i=1, \ldots , h^4_H (X_4)$ denote a basis of primary horizontal
subspace of $H_4 (X_4)$. Here we introduced an intersection matrix $\eta^{ij}$ and a dual
basis $\hat{\gamma}^i$ in $H^4_H (X_4)$ defined by
\begin{align}
\eta^{ij} = \int_{X_4} \hat{\gamma}^i \wedge \hat{\gamma}^j,
\ \ \ \ \ \ \ \int_{\gamma^i} \hat{\gamma}^j = \delta^{ij}.
\label{intm}
\end{align}
Similarly, when we turn on background $G_4$ fluxes whose integer quantum numbers are given
by
\begin{align}
n_i = \int_{\gamma^i} G_4,
\label{Gflux}
\end{align}
they generate a superpotential of the following form
\begin{align}
W = (4\pi)^{-1/2} \sum_{i,j} n_i \Pi_j \eta^{ij}.
\label{fluq}
\end{align}
As an explicit example of F-theory compactification, here we consider the background studied
in \cite{Honma:2017uzn}
(see also \cite{Alim:2009bx,Grimm:2009ef,Jockers:2009ti,Grimm:2010gk,Honma:2015iza}),
whose period integrals and topological intersection matrix are given by
\begin{align}
\begin{split}
\Pi_1&=1, \ \Pi_2=z, \ \Pi_3=-z_1, \ \Pi_4=S, \\
\Pi_5&=5Sz, \ \Pi_6=\frac{5}{2}z^2, \ \Pi_7=2z_1^2, \ \Pi_8=-\frac{5}{2}Sz^2-\frac{5}{3}z^3, \\
\Pi_9&=-\frac{2}{3}z_1^3, \ \Pi_{10}=-\frac{5}{6}z^3, \ \Pi_{11}=\frac{5}{6}Sz^3+\frac{5}{12}z^4
-\frac{1}{6}z_1^4,
\end{split}
\label{per}
\end{align}
\begin{align}
\eta=\begin{pmatrix}
0& 0 & 0 & 0 & 1 \\
0 & 0 & 0 & I_3 & 0 \\
0 & 0 & \widetilde{\eta} & 0 & 0 \\
0 & I_3 & 0 & 0 & 0 \\
1 & 0 & 0 & 0 & 0
\end{pmatrix}, \ \ \ \ \ \
\widetilde{\eta}=\begin{pmatrix}
0& \frac{1}{5} & 0 \\[5pt]
\frac{1}{5} & \frac{2}{5} & 0 \\[5pt]
0 & 0 & -\frac{1}{4}
\end{pmatrix},
\label{imat}
\end{align}
with the Euler characteristic $\chi = 1860$. Here the complex structure moduli of the
fourfold $z,z-z_1,S$ are originated from a bulk quintic modulus, a brane modulus and the
axio-dilaton in Type IIB description, respectively. Note that here we have picked up the
leading interactions only. While this choice is sufficient for the current purpose, further
quantum corrections can be easily calculated by using mirror symmetry technique or
supersymmetric localization approach as in \cite{Honma:2013hma}.
Substituting the topological data (\ref{per}) and (\ref{imat}) into the generic formula
(\ref{Kper}), one can obtain the explicit form of the K$\ddot{{\textrm{a}}}$hler potential
as
\begin{align}
K=-\ln{\left[ -i(S-\overline{S}) \right]} -\ln{\widetilde{Y}}-2\ln{\cal{V}},
\label{FK}
\end{align}
where
\begin{align}
\widetilde{Y}=\frac{5i}{6}(z-\bar{z})^3+\frac{i}{S-\overline{S}}
\left( \frac{5}{12}(z-\bar{z})^4 -\frac{1}{6}(z_1-\bar{z}_1)^4 \right).
\label{FY}
\end{align}
Here we also added the contribution from the classical K$\ddot{{\textrm{a}}}$hler moduli
sector. The characteristic property of this kind of F-theory construction based on a particular
class of Calabi-Yau fourfold is that the next-to-leading-order correction with respect to the
string coupling $g_s = ({\rm Im}S)^{-1}$ has been incorporated, owing to the elliptically fibered
structure of the background geometry.\footnote{See \cite{Grimm:2009ef,Jockers:2009ti} for a detailed
analysis and general construction based on mirror symmetry technique with D-branes.} Similarly,
the flux-induced superpotential is given by
\begin{align}
\begin{split}
W = \frac{1}{\sqrt{4\pi}} & \Bigg( n_{11}+n_{10}S+n_8 z+n_6 S z+\frac{1}{2}\left( n_5+2 n_6 \right)
z^2 -\frac{5n_4}{6} z^3 -n_2 \left( \frac{5}{2}Sz^2+\frac{5}{3}z^3 \right) \\
& \ \ -n_9 z_1-\frac{n_7}{2}z_1^2 -\frac{2n_3}{3}z_1^3+n_1 \left( \frac{5}{6}Sz^3+\frac{5}{12}z^4
-\frac{1}{6}z_1^4 \right) \Bigg),
\label{W}
\end{split}
\end{align}
Note that the tadpole cancellation condition needs to be satisfied by background fluxes (\ref{Gflux})
as
\begin{align}
\frac{1860}{24} =& \ n_{\rm{D3}}+n_1 n_{11}+n_2 n_8+n_3 n_9+n_4 n_{10}+\left( \frac{n_5+n_6}{5}\right)
n_6-\frac{n_7^2}{8},
\label{tadcanF}
\end{align}
which requires that $n_7$ must be $2+4k$ with $k \in {\mathbb{Z}}$ and $n_5+n_6$ or $n_6$ must be
$5k'$ with $k' \in {\mathbb{Z}}$, to preserve the integrality of each of the flux quanta.
Here let us impose a condition
\begin{align}
n_2=n_3=n_4=n_8=n_9=n_{10}=0,
\end{align}
and non-zero otherwise, which corresponds to consider the self-dual $G_4$ fluxes only.
In this setup, the model has a Minkowski minimum with $V=0$ as a solution to the $F$-term condition
and the values of the moduli fields are fixed as \cite{Honma:2017uzn}
\begin{align}
\begin{split}
{\textrm{Re}}z &={\textrm{Re}}z_1={\textrm{Re}}S=0, \\
{\textrm{Im}}z&=\left(\frac{6n_{11}}{5n_1}\right)^{1/4}\frac{2\sqrt{n_6}}
{(8n_6(n_5+n_6)-5n_7^2)^{1/4}},\\
{\textrm{Im}}z_1&=\left(\frac{30n_{11}}{n_1}\right)^{1/4}\frac{\sqrt{n_7}}
{(8n_6(n_5+n_6)-5n_7^2)^{1/4}}, \\
{\textrm{Im}}S &=\left(\frac{6n_{11}}{5n_1}\right)^{1/4}\frac{n_5}{\sqrt{n_6}
(8n_6(n_5+n_6)-5n_7^2)^{1/4}}.
\label{vacuum}
\end{split}
\end{align}
Again, later we will utilize this analytic solution to discuss about a correspondence between
the moduli stabilization problem and a seemingly different topic in supergravity.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Distribution of non-supersymmetric flux vacua}
\label{subsec:resultF}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Based on the above setup, we are now ready to investigate the non-supersymmetric flux vacua
with fixed moduli $\Phi_I \equiv \{ z, z_1, S \}$ in the framework of F-theory compactification,
in a similar way as we demonstrated in Section \ref{subsec:resultIIB}.
In order to find general supersymmetry-breaking minima (\ref{nonsv}) for the no-scale potential
with (\ref{FK}) and (\ref{W}), we again utilize the ``FindMinimum'' function in Mathematica and
also check the absence of tachyons at thus obtained flux vacua by confirming the mass squared of
the moduli fields being strictly positive. In the following analysis, we restrict ourselves to
the case with $n_{\rm D3}=0$ for simplicity, and we only pick up the solutions satisfying
$\epsilon <1$ as true flux vacua in order to ensure the effective theory description of the system.
We also need to be careful about the validity of the leading term approximation of the underlying
fourfold periods (\ref{per}) from which the scalar potential is determined. During our numerical analysis, thereby we further impose the conditions ${\rm Im}\Phi_I > 1$ to ensure that the leading terms of the periods are sufficiently large compared to the possible quantum loop and instanton corrections, and thus all the obtained minima become
well-defined.\footnote{We have also checked that our numerical results are consistent with large
complex structure expansion of the periods, whose radius of convergence can be extracted from
discriminants of the underlying fourfold.}
As a result, we numerically found 10058 supersymmetry-breaking flux vacua under the following
range of background fluxes\footnote{By using the Monte-Carlo methods, we randomly generated $10^6$
data set of fluxes within $-10\leq n_{1,2,\cdots,11}\leq 10$ and confirmed that the possible minima
are almost clustered around $n_1=n_2=0$. While this indication makes our present artificial choice
of fluxes due to the limited performance of our computers quite conceivable, it would be instructive
to re-analyze the model in a broader range of fluxes.}
\begin{align}
n_1=n_2=0, \quad -8\leq n_{3,4,8,9,10,11}\leq 8, \quad -10\leq n_{5,6} \leq 10, \quad -6\leq n_7\leq 6,
\label{eq:flux_F}
\end{align}
satisfying the tadpole cancellation condition~(\ref{tadcanF}). Our numerical results are summarized in
Figures \ref{fig:F} and \ref{fig:F3}.
Note that we have not drawn the area over a range $0.6 < \epsilon <1$, where we confirmed that there
are no appropriate solutions in our present setup.
\begin{figure}[t]
\begin{minipage}{0.5\hsize}
\begin{center}
\includegraphics[width=80mm]{F1_new.pdf}
\end{center}
\end{minipage}
\begin{minipage}{0.5\hsize}
\begin{center}
\includegraphics[width=80mm]{F2_new.pdf}
\end{center}
\end{minipage}
\caption{Numerical results for the distribution of non-supersymmetric F-theory flux vacua.
Each dot represents a solution with different set of fluxes and the ambiguity from
SL(2, $\mathbb{Z}$) duality has been deducted. Note that, as in Figure \ref{fig:T6},
these dots can be degenerated in this 2D plot, due to the existence of other directions of the
stabilized moduli $z$ and $z_1$. The right figure is the log-log plot of the same results.}
\label{fig:F}
\end{figure}
As advocated in the previous section, one can see clearly that the scalar potential and other
on-shell quantities thus obtained exhibit a monotonic, non-increasing functional behavior over
a wide range of parameter $g_s$. One can also check that all the numerical solutions in
Figure \ref{fig:F} satisfy the condition (\ref{eq:epsilon}), if we take a sufficiently large
volume ${\cal V}$, e.g. larger than $\simeq {\cal{O}}(10^2)$ for the background geometry.
\clearpage
Here we exemplify our numerical results by picking up three independent flux vacua determined by
the following set of fluxes:
\begin{table}[htb]
\centering
\begin{tabular}{|c|c|} \hline
Vacuum & Set of fluxes $(n_1, n_2 n_3, n_4, n_5, n_6, n_7, n_8, n_9, n_{10}, n_{11})$ \\ \hline \hline
D & $(0,0, -2, -8, -8, -2, -6, -4, -7, -8, 0)$ \\ \hline
E & $(0,0, -3, -7, 4, 1, -2, -7, -7, -8, -2)$ \\ \hline
F & $(0,0, -4, 1, -10, -10, 2, 7, -8, 6, 0)$ \\ \hline
\end{tabular}
\end{table}
\noindent These three examples D, E and F are represented in Figure \ref{fig:F} by the red,
green and blue points, respectively. Explicit values of stabilized moduli and various on-shell
quantities at the each of the vacua are summarized in Tables~\ref{tab:F1} and~\ref{tab:F2}.\footnote{
More precisely, at the each of the minima, we found that the VEVs of moduli fields exhibit the following distributions:
\begin{align*}
1 < {\rm Im}z_1 < 1.34, \ \ \ \ 2< {\rm Im}z < 4.54, \ \ \ \ 1.19< {\rm Im}S < 74.6.
\end{align*}
Naively the smallness of ${\rm Im}z_1$ may suggest that all the numerical solutions we have obtained are likely to be in a region
outside the ``parametrically large field distance" forbidden by the swampland hypothesis \cite{Obied:2018sgi,Garg:2018reu,Ooguri:2018wrx}
and our results are consistent with the recent developments on this field. We would like to thank the referee for making us realize this point.}
Comparing these results as well as the plots of Figure \ref{fig:F} in F-theory setup with those
in Type IIB orientifold model in Section \ref{subsec:resultIIB}, it appears that within a finite
range of background fluxes of the same amount, the model based on F-theory flux compactification
equipped with a next-to-leading-order $g_s$ correction prefers a smaller value of the vacuum energy.
This preference would provide strong motivation for further studies toward the de Sitter model
building in the framework of F-theory dealing with the finite string coupling corrections.
\begin{table}[ht]
\centering
\begin{tabular}{|c|c|c|c|c|c|} \hline
Vacuum & $z_1$ & $z$ & $S$ & $g_s$ & $\epsilon$ \\ \hline \hline
D & $-0.703+1.00 i$ & $-0.0908 + 2.08 i$ & $-0.275 + 6.74i$ & 0.148 & $9.37\times 10^{-3}$ \\ \hline
E & $-0.181+1.01 i$ & $0.0584 + 2.15 i$ & $-0.461 + 6.51i$ & 0.154 & $2.16\times 10^{-6}$ \\ \hline
F & $0.243 + 1.01 i$ & $0.435 + 2.03i$ & $-0.354 + 1.25 i$ & 0.802 & $3.30\times 10^{-3}$ \\ \hline
\end{tabular}
\caption{Explicit values of stabilized moduli and on-sell quantities in $M_{\rm Pl}=1$ unit.
}
\label{tab:F1}
\end{table}
\begin{table}[ht]
\centering
\begin{tabular}{|c|c|c|} \hline
Vacuum & $W$ & Eigenvalues of mass matrix $\partial_I \partial_J V \times {\cal V}^2$ \\ \hline \hline
D & $16.9-32.0i$ & (5.13, 2.13, 1.85, 1.26, 0.0446, $2.01\times 10^{-4}$)\\ \hline
E & $-9.16-33.8i$ & (3.99, 3.83, 2.89, 1.42, 0.0393, $1.54\times 10^{-6}$ )\\ \hline
F & $25.9+2.33i$ & (7.24, 4.56, 3.27, 0.509, 0.154, $7.06\times 10^{-3}$)\\ \hline
\end{tabular}
\caption{Explicit values of flux superpotential and mass eigenvalues in $M_{\rm Pl}=1$ unit. Note that
we have diagonalized the mass matrix and the eigenvalues are displayed in the descending order.}
\label{tab:F2}
\end{table}
\begin{figure}[htb]
\begin{minipage}{0.5\hsize}
\begin{center}
\includegraphics[width=80mm]{F3_new.pdf}
\end{center}
\end{minipage}
\begin{minipage}{0.5\hsize}
\begin{center}
\includegraphics[width=80mm]{F4_new.pdf}
\end{center}
\end{minipage}
\caption{Numerical results for the absolute values of the flux superpotential $|W|$ and
$|{\cal Z}|=e^{K/2}|W|$ on each of the flux vacua depicted in Figure \ref{fig:F}.
}
\label{fig:F3}
\end{figure}
Before ending this section, we would like to discuss about a possible extension of our F-theory
setup toward the stabilization of K$\ddot{{\textrm{a}}}$hler moduli and explicit construction of
de Sitter spacetime. As can be seen from Figure \ref{fig:F3}, the absolute value of the
superpotential is of, at lowest, ${\cal O}(10)$. This implies that LVS-type extension is more
desirable to construct an explicit de Sitter spacetime, rather than utilizing the KKLT-type
construction which generically requires an extremely small on-shell value of the flux superpotential.
Suppose we employ the LVS-type argument when we extend the K$\ddot{{\textrm{a}}}$hler moduli sector
in our F-theory setup, and assume that quantum corrections for K$\ddot{{\textrm{a}}}$hler moduli
fields breaking the no-scale structure give rise to a potential
\begin{align}
V \sim -e^{%K_{\rm CS}
-\ln \left( 2\widetilde{Y} {\rm Im}S \right)}|W_0|\frac{\ln {\cal V}}{{\cal V}^3}.
\end{align}
Then the dimensionless constant $\epsilon$ determined in the dynamics of complex structure moduli
sector is further constrained to satisfy a condition $\epsilon\sim e^{K_{\rm CS}}|W_0|{\cal V}^{-1}
\ln {\cal V}$ such that the total effective potential becomes close to zero as a candidate of the
tiny cosmological constant. Since the numerical value of $\epsilon$ obtained in our analysis is
sufficiently small to realize this naive expectation, it would be interesting to cast this idea in
a rigorous setup.\footnote{For the details about quantum corrections in F-theory, we refer the
reader to, for instance, \cite{Grimm:2013gma,Grimm:2013bha,Minasian:2015bxa,Weissenbacher:2019mef}.}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Flux Vacua and Attractor Equations}
\label{sec:Att}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
So far we have analyzed the vacuum structure of effective theories of moduli fields arising
from Type IIB and F-theory flux compactifications, aiming to construct appropriate 4D spacetime
describing our universe in a consistent and systematic way. There we have shown that the on-shell
quantities such as vacuum energy and flux superpotential exhibit a monotonic, non-increasing
functional behavior with respect to the string coupling. While this indication strongly motivates
us to devote special attention to the significance of F-theory framework for the de Sitter model
building, at the same time the following natural question arises: how exactly and why these
on-shell quantities on the flux vacua in moduli stabilization problem are determined to exhibit
this fascinating behavior?
Here we argue that our findings are deeply connected with a seemingly different topic in the 4D
supergravity referred to as the attractor mechanism \cite{Ferrara:1995ih} (see
also \cite{Ferrara:2008hwa,Moore:2004fg}), where the normalized flux superpotential ${\cal Z}=e^{K/2}W$
of a system can be identified as a central charge of the underlying algebra which is subject to a
monotonic flow along the radial direction of a background. %related to the entropy by $|{\cal Z}|=\sqrt{A/4\pi}$.
More precisely, the attractor mechanism means that in a class of extremal
black holes in 4D ${\cal N}=2$ theories, moduli fields are drawn to fixed values at the horizon,
regardless of the initial values at the asymptotic infinity. There the fixed values of the moduli
fields are determined by the so-called attractor equation. In fact, a possible existence of a
correspondence between flux vacua in moduli stabilization and the attractor mechanism in
supergravity has been pointed out in \cite{Kallosh:2005ax} (see
also \cite{Kallosh:2005bj,Kallosh:2006bt,Alishahiha:2006jd,Bellucci:2007ds,Larsen:2009fw}). From this perspective,
the systematic behavior of various quantities at the distributed flux vacua we observed may
reflect the dynamics of a corresponding attracting object, if this kind of correspondence truly exists.
Here we provide a supporting evidence for this conjecture by confirming that various solutions we
obtained indeed satisfy the suitably generalized attractor equations simultaneously.
First let us reconsider the $T^6/\mathbb{Z}_2$ toroidal orientifold model as a limit of F-theory on an uplifted fourfold $T^6 \times T^2/\mathbb{Z}_2$, in accordance with a prescription illustrated in \cite{Kallosh:2005ax,Kallosh:2005bj}.
Seen from this perspective, the
K$\ddot{{\textrm{a}}}$hler potential and flux superpotential in the isotropic case with moduli
fields $\Phi_I = \{ \tau, S \}$ in Section \ref{subsec:example1} can be derived from the generic
formulas (\ref{Kper}) and (\ref{fluq}) in F-theory compactifications with the following assignment
of the fourfold periods
\begin{align}
\begin{split}
\Pi_1 &=1, \ \Pi_2=-S, \ \Pi_3=\tau, \ \Pi_4=-S\tau, \\
\Pi_5 &=3\tau^2, \ \Pi_6=-3S\tau^2, \ \Pi_7=-\tau^3, \ \Pi_8=S\tau^3,
\end{split}
\label{peritoro}
\end{align}
and the intersection matrix
\begin{align}
\eta=\begin{pmatrix}
0 & 0 & 0 & 0 & 0 & 0 & 0 & 1 \\
0 & 0 & 0 & 0 & 0 & 0 & -1 & 0 \\
0 & 0 & 0 & 0 & 0 & 1 & 0 & 0 \\
0 & 0 & 0 & 0 & -1 & 0 & 0 & 0 \\
0 & 0 & 0 & -1 & 0 & 0 & 0 & 0 \\
0 & 0 & 1 & 0 & 0 & 0 & 0 & 0 \\
0 & -1 & 0 & 0 & 0 & 0 & 0 & 0 \\
1 & 0 & 0 & 0 & 0 & 0 & 0 & 0
\end{pmatrix},
\label{intmtoro}
\end{align}
where the background three-form fluxes in the $T^6/\mathbb{Z}_2$ model can be reproduced from (\ref{Gflux})
with
\begin{align}
\vec{n} = \int G_4=(-c^0, a^0, -c^1, a_1, -3d_1,-3b_1,d_0,-b_0)^T.
\end{align}
The above reformulation corresponds to geometrize the SL(2, $\mathbb{Z}$) duality of Type IIB string
theory into an auxiliary two-torus and combine the various three-forms in Section \ref{subsec:example1}
with additional cohomology classes dual to the A- and B-cycles of the torus.\footnote{See for instance
Section 3 in \cite{Denef:2008wq} for more details and precautions.}
By using above quantities and the central charge ${\cal{Z}}(\tau,\bar{\tau},S,\bar{S}) = e^{\frac{K}{2}}
W(\tau,S)$, one can check that the Minkowski solutions given by (\ref{torvacuum}) and (\ref{torvac2})
of the effective potential also satisfy the attractor equation of a generalized
form \cite{Kallosh:2005ax,Denef:2004ze}
\begin{align}
n_i = 2 {\textrm{Re}} \left[ \overline{{\cal{Z}}}\hat{\Pi}_i + \overline{D}^{\tau} \overline{D}^S
\overline{{\cal{Z}}} D_{\tau} D_S \hat{\Pi}_i \right],
\end{align}
where $\overline{D}^I \equiv K^{I\bar{J}}{\overline{D}}_{\bar{J}}$ and we have
K$\ddot{{\textrm{a}}}$hler-normalized the periods as $\hat{\Pi}_i \equiv e^{K/2}\Pi_i$. The appearance
of this equation can be also understood from the viewpoint of the Hodge structure of allowed four-forms
of the system \cite{Denef:2004ze}. In a similar fashion, we numerically checked that the
non-supersymmetric flux vacua with nonzero scalar potential described in Section \ref{subsec:resultIIB}
also satisfy the equation
\begin{align}
n_i = 2 {\textrm{Re}} \left[ \overline{{\cal{Z}}}\hat{\Pi}_i - \overline{D}^I \overline{{\cal{Z}}}
D_I \hat{\Pi}_i + \overline{D}^{\tau} \overline{D}^S
\overline{{\cal{Z}}} D_{\tau} D_S \hat{\Pi}_i \right].
\end{align}
More precisely, we have picked up several points in Figure \ref{fig:T6} as well as the points A, B and C
and confirmed that in all the examples the equation indeed holds up to the accuracy ${\cal{O}}(10^{-20})$,
as far as we investigated.\footnote{It is worth performing further consistency checks about the appearance
of this equation for non-supersymmetric flux vacua from another independent way, either by analytical or
more sophisticated numerical approach.}
Before moving on to the analysis about F-theory example, we would like to mention that a whole picture of
a possible correspondence between flux vacua and attractor mechanism is not completely elucidated even in
a simple situation in Type IIB compactification as we actualized above. The main problem is the lack of
knowledge about the identification of the exact metric of a corresponding black object, from which the
attractor behavior and the entropy can be extracted to confirm the coincidence with the analysis of the
flux vacua. Although we naively expect that a warped geometry in 10D spacetime with appropriate fluxes and
horizon topology would be relevant, a detailed study of this subject is beyond the scope of the present
paper. The simple $T^6/\mathbb{Z}_2$ toroidal orientifold with a suitable set of background fluxes we
thoroughly investigated here may be regarded as a moderate starting point to address this important problem.
Finally, let us discuss about more nontrivial check concerning the F-theory model discussed in
Section \ref{sec:Fth}, equipped with the moduli fields $\Phi_I = \{ z, z_1, S \}$. After a careful
calculation utilizing the fourfold periods (\ref{per}) and the central charge ${\cal{Z}}(z, \bar{z}, z_1,
\bar{z_1}, S, \bar{S}) = e^{\frac{K}{2}}W(z, z_1, S)$, one can check that the Minkowski solution
in (\ref{vacuum}) simultaneously satisfies the attractor equation of the following form
\begin{align}
n_i = 2 {\textrm{Re}} \left[ \overline{{\cal{Z}}}\hat{\Pi}_i + C^{IJ} D_I D_J \hat{\Pi}_i \right],
\label{Fatt}
\end{align}
where $C^{IJ}$ is determined by a condition\footnote{See Appendix \ref{app:attF} for the details about
our F-theory generalization of the attractor equation.}
\begin{align}
\overline{D}_{\bar{K}} \overline{D}_{\bar{L}} \overline{Z} = C^{IJ} \biggl[ R_{I\bar{K}J\bar{L}}
+K_{I\bar{K}}K_{J\bar{L}} +K_{I\bar{L}}K_{J\bar{K}}\biggl]
+ \overline{C}^{\bar{I}\bar{J}} e^K \overline{Y}_{\bar{I}\bar{J}\bar{K}\bar{L}}.
\label{Ccon}
\end{align}
Here $Y_{IJKL}$ represent the classical quadruple intersection numbers of the background Calabi-Yau fourfold whose explicit values can be easily extracted from (\ref{FY}), and $R_{I\bar{J}K\bar{L}}$
is the Riemann curvature tensor of the complex structure moduli space defined by
\begin{align}
R_{I\bar{J}K\bar{L}} = K^{M\bar{N}}(\partial_{\bar{L}}\partial_{\bar{J}}\partial_M K)
\partial_{I}\partial_{\bar{N}}\partial_K K -\partial_{\bar{L}}\partial_I\partial_{\bar{J}}\partial_K K.
\end{align}
Note that the covariant derivative $D_I$ appearing in (\ref{Fatt}) and (\ref{Ccon}) is not a simple
K$\ddot{{\textrm{a}}}$hler covariant derivative, but also includes the information about the curvature
of the moduli space as symbolically represented by $D = \partial + a \partial K + \Gamma.$ Here $a$ is
the K$\ddot{{\textrm{a}}}$hler weight\footnote{For instance, ${\cal{Z}}$ and $\overline{{\cal{Z}}}$ have
K$\ddot{{\textrm{a}}}$hler weights $1/2$ and $-1/2$, respectively.} of a quantity the covariant derivative
acts on, and $\Gamma$ denotes the Christoffel symbol of the second kind of the moduli space metric. For
the case of $T^6/\mathbb{Z}_2$ model discussed above, $\Gamma$ is exactly zero and therefore can be
identified with the ordinary K$\ddot{{\textrm{a}}}$hler covariant derivative.
Moreover, we numerically checked that, up to the accuracy ${\cal{O}}(10^{-20})$, the
non-supersymmetric flux vacua with nonzero scalar potential found in Section \ref{subsec:resultF} satisfy
the equation
\begin{align}
n_i = 2 {\textrm{Re}} \left[ \overline{{\cal{Z}}}\hat{\Pi}_i - \overline{D}^I \overline{{\cal{Z}}}
D_I \hat{\Pi}_i+ C^{IJ} D_I D_J \hat{\Pi}_i \right],
\end{align}
where the coefficient $C^{IJ}$ is constrained to satisfy (\ref{Ccon}). These results strongly support a
possible existence of a F-theory generalization of the attractor mechanism, going beyond the conventional
Type IIB setup. Especially, all the explicit results we have shown in the F-theory setup should be regarded
as a prediction that if there exists a black object exhibiting the attracting behavior in the framework
of F-theory dealing with finite string coupling, its Bekenstein-Hawking entropy and other on-shell quantities
would have the same characteristics we have demonstrated above.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Conclusions and Discussions}
\label{sec:conclusion}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
In this paper, we studied several aspects of moduli stabilization in the framework of spacetime
flux compactifications by explicitly constructing flux vacua and analyzing their characteristics.
Especially we thoroughly investigated the 4D effective theories arising from Type IIB string
theory on a toroidal orientifold $T^6/\mathbb{Z}_2$ and F-theory on a Calabi-Yau fourfold with
appropriately quantized background fluxes. As well as clarifying the Minkowski minima with vanishing
scalar potential analytically, we numerically solved the extremal conditions of the moduli fields
and confirmed that the richness of the choice of background fluxes leads to the existence of vast
numbers of stable non-supersymmetric minima of the potential with sufficiently tiny vacuum energy.
Moreover, it turned out that the allowed values of the potential minima and various on-shell
quantities show a non-increasing functional behavior with respect to the string coupling $g_s$.
In particular it appears that within a finite range of background fluxes of the same amount,
F-theory model equipped with a next-to-leading-order $g_s$ correction prefers a smaller value of
the vacuum energy, compared with the simple Type IIB toroidal orientifold model. We hope that
several extensions of our results would enrich the stringy construction of the de Sitter spacetime.
We also argued that an interesting conclusion follows from the suggestive behavior of on-shell
quantities of Type IIB string and F-theory flux vacua. In the process of clarifying the underlying
dynamics of the non-increasing functional behavior of the vacuum energy and the superpotential, we
come up with a conclusion that the previously-reported possible correspondence between flux vacua
in moduli stabilization problem and attractor mechanism in supergravity lies behind. To provide a
nontrivial evidence for this surmise, we have also checked that our analytic and numerical solutions
in various flux backgrounds indeed satisfy the suitably generalized attractor equations simultaneously.
Especially, our demonstrations in F-theory framework go beyond the conventional Type IIB setups
describing the attractor mechanism and may shed new light on this intriguing subject.
Finally, we comment on several possible future research directions in a random order.
\begin{itemize}
\item
Our studies and demonstrations about the vacuum structure of effective theories would be straightforwardly
applicable to broader class of examples based on other type of toroidal orientifolds such as
$T^6/(\mathbb{Z}_2 \times \mathbb{Z}_2)$ \cite{Blumenhagen:2003vr,Cascales:2003zp} and various
Calabi-Yau manifolds. Besides providing concrete examples of moduli stabilizations toward the explicit
de Sitter model building, this would be connected with a further development of the attractor mechanism
in various setups. Within this context, the recent works about asymptotic flux compactifications
in \cite{Grimm:2019ixq} and attractor black holes in \cite{Hulsey:2019xdb} can be also involved.
\item
Since our analysis has been restricted to the complex structure moduli as well as the dilaton in Type IIB language, it is indispensable to address the dynamics of the K\"{a}hler moduli and their stabilization at the quantum level if one wants to find a landscape of explicit de Sitter minima. As has been emphasized for instance in \cite{Dasgupta:2014pma}, it will be required to perform a careful analysis with respect to the consistency with the Einstein equations of motion in order to realize the de Sitter spacetime correctly, which should be a highly nontrivial task.
\item
We naively guess that the standard monotonic gradient flow of the central charge of ordinary attracting extremal BH with respect to the radial direction of a background in SUGRA might be extended at the finite string coupling regime such that the central charge or its derivatives related to the vacuum energy will become also a function of string coupling, whose dependence becomes monotonic if the radial evolution or energy scale in AdS/CFT context and the string coupling are also monotonically correlated.
Clarifying this point would be an interesting but highly difficult problem.
\item
The attractor mechanism states that the evolution of moduli fields is governed by the gradient flow in
the radial direction of the 4D background and the central charge ${\cal{Z}}$. On the other hand, it has
been recently pointed out that there exists a close relation between the swampland argument and the
gradient flow in gravity \cite{Kehagias:2019akr}. Therefore, naively one can expect that our findings
about the appearance of the attractor equations in both of the Type IIB string and F-theory flux
compactifications might be connected with the basics of the swampland hypothesis. It would be interesting
to figure out a more complete description about this expectation.
\item
About KKLT construction, recent analysis in \cite{Bena:2018fqc} about the conifold modulus parametrizing
the size of three-cycle on the bottom of a warped deformed conifold throat indicates that one must
require a large amount of three-form background fluxes to avoid the destabilization, which is generically
incompatible with the tadpole cancellation condition in Type IIB flux compactifications. It would be
interesting to clarify whether F-theory framework we studied can provide one way to resolve this
destabilization problem.
\end{itemize}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection*{Acknowledgements}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
We would like to thank K. Ohta, A. Otsuka, K. Sakai and T. Watari for useful discussions and
comments. H. O. was supported by a Grant-in-Aid for JSPS Research Fellow No. 19J00664. We
appreciate the Yukawa Institute for Theoretical Physics at Kyoto University, where this work
was presented during the YITP-W-19-10 on ``Strings and Fields 2019". We are also grateful to
the participants of the workshop ``KEK Theory Workshop 2018" held at KEK Theory Center for
illuminating discussions.
\appendix
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Hodge Structure and Attractor Equation}\label{app:attF}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
We have described a F-theory generalization of the attractor equation in (\ref{Fatt}) with a
condition (\ref{Ccon}). Here let us explain how we arrived at these expressions. In fact, a
slight modification of the analysis in \cite{Denef:2004ze} leads us to find out appropriate
form of the equation.
Along the line of \cite{Denef:2004ze} (see also \cite{Bellucci:2007ds}), here we consider an
expansion of a generic real four-form flux on a Calabi-Yau fourfold $X_4$ as\footnote{Although
the third and forth order covariant derivatives of $\hat{\Omega}$ are irrelevant to derive
(\ref{Fatt}), it would be instructive to clarify their relationship to other basis and obtain
a more complete view of the Hodge structure of the system.}
\begin{align}
G_4 &= A \hat{\Omega} + B^I D_I \hat{\Omega} +C^{IJ} D_I D_J \hat{\Omega}+\overline{A}
\overline{\hat{\Omega}} +{\overline{B}}^{\bar{I}} \overline{D}_{\bar{I}}\overline{\hat{\Omega}}
+{\overline{C}}^{\bar{I}\bar{J}} \overline{D}_{\bar{I}}\overline{D}_{\bar{J}}\overline{\hat{\Omega}},
\label{eq:G4expand}
\end{align}
where the indices $I, J$ label the complex structure moduli of a Calabi-Yau fourfold $X_4$,
and we have K$\ddot{{\textrm{a}}}$hler-normalized the holomorphic four-form as $\hat{\Omega}
(\Phi_I,\overline{\Phi}_{\bar{I}}) \equiv e^{K/2}\Omega(\Phi_I)$ such that the covariant derivative
would acts on $\hat{\Omega}$ as $D_I\hat{\Omega}=(\partial_I +K_I/2)\hat{\Omega}=e^{K/2}D_I\Omega$.
Let us denote the intersection products of the covariant derivatives of $\hat{\Omega}$ as
\begin{align}
\begin{split}
{\cal N}_{I...J|K...L} &\equiv D_I \cdots D_J \hat{\Pi}\; \eta^{-1} D_K\cdots D_L \overline{\hat{\Pi}}
= \int_{X_4} D_I\cdots D_J \hat{\Omega} \wedge D_K \cdots D_L \overline{\hat{\Omega}}, \\
{\cal N}^\prime_{I...J|K...L} &\equiv D_I \cdots D_J \hat{\Pi}\; \eta^{-1} D_K\cdots D_L \hat{\Pi}
= \int_{X_4} D_I\cdots D_J \hat{\Omega} \wedge D_K \cdots D_L \hat{\Omega},
\end{split}
\end{align}
where we have K$\ddot{{\textrm{a}}}$hler-normalized the fourfold periods of $X_4$ as $\hat{\Pi}
\equiv e^{K/2}\Pi$. Their explicit forms are known to be given by (see for
instance \cite{Denef:2004ze})
\begin{align}
\begin{split}
{\cal N}_{I|\bar{J}} &= -K_{I\bar{J}}, \\
{\cal N}_{IJ|\bar{K}\bar{L}} & = R_{I\bar{K}J\bar{L}} +K_{I\bar{K}}K_{J\bar{L}} +K_{I\bar{L}}K_{J\bar{K}}, \\
{\cal N}_{\bar{I}J|K\bar{L}} &= K_{J\bar{I}}K_{K\bar{L}}, \\
{\cal N}^\prime_{IJ|KL} &= e^K Y_{IJKL},
\end{split}
\end{align}
where $Y_{IJKL}$ represent classical quadruple intersection numbers of $X_4$, and $R_{I\bar{J}K\bar{L}}$
is the Riemann curvature tensor of the complex structure moduli space of $X_4$.
Utilizing the above formulas, it turns out that the explicit forms of the coefficients $A$ and $B^I$
in (\ref{eq:G4expand}) can be determined by
\begin{align}
\begin{split}
\overline{{\cal{Z}}} &= \int G_4 \wedge \overline{\hat{\Omega}} = A \int \hat{\Omega} \wedge
\overline{\hat{\Omega}} = A, \\
\overline{D}_{\bar{I}}\overline{{\cal{Z}}} &= \int G_4 \wedge \overline{D}_{\bar{I}} \overline{\hat{\Omega}}
= B^J \int D_J \hat{\Omega} \wedge \overline{D}_{\bar{I}} \overline{\hat{\Omega}} = B^J (-K_{J\bar{I}}),
\label{ABco}
\end{split}
\end{align}
whereas $C^{IJ}$ and its conjugate are constrained to satisfy
\begin{align}
\begin{split}
\overline{D}_{\bar{K}} \overline{D}_{\bar{L}} \overline{{\cal{Z}}} &= \int G_4 \wedge \overline{D}_{\bar{K}}
\overline{D}_{\bar{L}} \overline{\hat{\Omega}} \\
&=C^{IJ} \int D_I D_J \hat{\Omega} \wedge \overline{D}_{\bar{K}} \overline{D}_{\bar{L}} \overline{\hat{\Omega}}
+\overline{C}^{\bar{I}\bar{J}} \int\ \overline{D}_{\bar{I}} \overline{D}_{\bar{J}} \overline{\hat{\Omega}}
\wedge \overline{D}_{\bar{K}} \overline{D}_{\bar{L}} \overline{\hat{\Omega}} \\
&= C^{IJ} \biggl[ R_{I\bar{K}J\bar{L}}+K_{I\bar{K}}K_{J\bar{L}} +K_{I\bar{L}}K_{J\bar{K}}\biggl]
+\overline{C}^{\bar{I}\bar{J}} e^K \overline{Y}_{\bar{I}\bar{J}\bar{K}\bar{L}}.
\label{CconApp}
\end{split}
\end{align}
By plugging (\ref{ABco}) into (\ref{eq:G4expand}) and regarding (\ref{CconApp}) as an additional condition,
one can obtain the formulas about the attractor equations discussed in Section \ref{sec:Att}.
In a generic Type IIB Calabi-Yau compactification, the special K$\ddot{{\textrm{a}}}$hler geometry relation
of an underlying Calabi-Yau threefold yields a simplification of the form $D_iD_j \hat{\Omega} = i Y_{ijk}
K^{k\bar{l}}D_{\bar{l}}\overline{\hat{\Omega}}$ with $Y_{ijk}$ being classical triple intersection numbers
of the threefold with moduli fields $\Phi_i$. This makes the basis $D_i D_j \hat{\Omega}$ in an analogous
expansion of a real three-form flux in Type IIB compactifications redundant and as a result the attractor
equation in conventional Type IIB setups takes a quite simple form. By contrast, the F-theory compactification
based on a generic Calabi-Yau fourfold does not have such a simple structure of special
K$\ddot{{\textrm{a}}}$hler geometry and the basis $D_I D_J \hat{\Omega}$ becomes independent. This is the
main reason why a slight modification to the attractor equation is necessary in F-theory framework.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{thebibliography}{99}
\parskip=-2pt
%\cite{Vafa:1996xn}
\bibitem{Vafa:1996xn}
C.~Vafa,
%``Evidence for F theory,''
Nucl.\ Phys.\ B {\bf 469}, 403 (1996)
%doi:10.1016/0550-3213(96)00172-1
[hep-th/9602022].
%%CITATION = doi:10.1016/0550-3213(96)00172-1;%%
%1316 citations counted in INSPIRE as of 21 Oct 2019
%\cite{Giddings:2001yu}
\bibitem{Giddings:2001yu}
S.~B.~Giddings, S.~Kachru and J.~Polchinski,
%``Hierarchies from fluxes in string compactifications,''
Phys.\ Rev.\ D {\bf 66} (2002) 106006
%doi:10.1103/PhysRevD.66.106006
[hep-th/0105097].
%\cite{Gukov:1999ya}
\bibitem{Gukov:1999ya}
S.~Gukov, C.~Vafa and E.~Witten,
%``CFT's from Calabi-Yau four folds,''
Nucl.\ Phys.\ B {\bf 584}, 69 (2000)
Erratum: [Nucl.\ Phys.\ B {\bf 608}, 477 (2001)]
%doi:10.1016/S0550-3213(01)00289-9, 10.1016/S0550-3213(00)00373-4
[hep-th/9906070].
%%CITATION = doi:10.1016/S0550-3213(01)00289-9, 10.1016/S0550-3213(00)00373-4;%%
%951 citations counted in INSPIRE as of 27 Sep 2019
%\cite{Dasgupta:1999ss}
\bibitem{Dasgupta:1999ss}
K.~Dasgupta, G.~Rajesh and S.~Sethi,
%``M theory, orientifolds and G - flux,''
JHEP {\bf 9908}, 023 (1999)
%doi:10.1088/1126-6708/1999/08/023
[hep-th/9908088].
%%CITATION = doi:10.1088/1126-6708/1999/08/023;%%
%693 citations counted in INSPIRE as of 27 Sep 2019
%\cite{Kachru:2003aw}
\bibitem{Kachru:2003aw}
S.~Kachru, R.~Kallosh, A.~D.~Linde and S.~P.~Trivedi,
%``De Sitter vacua in string theory,''
Phys.\ Rev.\ D {\bf 68} (2003) 046005
%doi:10.1103/PhysRevD.68.046005
[hep-th/0301240].
%\cite{Balasubramanian:2005zx}
\bibitem{Balasubramanian:2005zx}
V.~Balasubramanian, P.~Berglund, J.~P.~Conlon and F.~Quevedo,
%``Systematics of moduli stabilisation in Calabi-Yau flux compactifications,''
JHEP {\bf 0503} (2005) 007
%doi:10.1088/1126-6708/2005/03/007
[hep-th/0502058].
%\cite{Conlon:2005ki}
\bibitem{Conlon:2005ki}
J.~P.~Conlon, F.~Quevedo and K.~Suruliz,
%``Large-volume flux compactifications: Moduli spectrum and D3/D7 soft supersymmetry breaking,''
JHEP {\bf 0508} (2005) 007
%doi:10.1088/1126-6708/2005/08/007
[hep-th/0505076].
%\cite{Saltman:2004sn}
\bibitem{Saltman:2004sn}
A.~Saltman and E.~Silverstein,
%``The Scaling of the no scale potential and de Sitter model building,''
JHEP {\bf 0411} (2004) 066
%doi:10.1088/1126-6708/2004/11/066
[hep-th/0402135].
%\cite{Gallego:2017dvd}
\bibitem{Gallego:2017dvd}
D.~Gallego, M.~C.~D.~Marsh, B.~Vercnocke and T.~Wrase,
%``A New Class of de Sitter Vacua in Type IIB Large Volume Compactifications,''
JHEP {\bf 1710} (2017) 193
%doi:10.1007/JHEP10(2017)193
[arXiv:1707.01095 [hep-th]].
%\cite{Blaback:2015zra}
\bibitem{Blaback:2015zra}
J.~Bl\r{a}b$\ddot{\textrm{a}}$ck, U.~H.~Danielsson, G.~Dibitetto and S.~C.~Vargas,
%``Universal dS vacua in STU-models,''
JHEP {\bf 1510} (2015) 069
%doi:10.1007/JHEP10(2015)069
[arXiv:1505.04283 [hep-th]].
%\cite{Kallosh:2005ax}
\bibitem{Kallosh:2005ax}
R.~Kallosh,
%``New attractors,''
JHEP {\bf 0512}, 022 (2005)
%doi:10.1088/1126-6708/2005/12/022
[hep-th/0510024].
%%CITATION = doi:10.1088/1126-6708/2005/12/022;%%
%128 citations counted in INSPIRE as of 06 Oct 2019
%\cite{Kallosh:2005bj}
\bibitem{Kallosh:2005bj}
R.~Kallosh,
%``Flux vacua as supersymmetric attractors,''
hep-th/0509112.
%\cite{Kallosh:2006bt}
\bibitem{Kallosh:2006bt}
R.~Kallosh, N.~Sivanandam and M.~Soroush,
%``The Non-BPS black hole attractor equation,''
JHEP {\bf 0603}, 060 (2006)
%doi:10.1088/1126-6708/2006/03/060
[hep-th/0602005].
%%CITATION = doi:10.1088/1126-6708/2006/03/060;%%
%136 citations counted in INSPIRE as of 06 Oct 2019
%\cite{Alishahiha:2006jd}
\bibitem{Alishahiha:2006jd}
M.~Alishahiha and H.~Ebrahim,
%``New attractor, entropy function and black hole partition function,''
JHEP {\bf 0611}, 017 (2006)
%doi:10.1088/1126-6708/2006/11/017
[hep-th/0605279].
%%CITATION = doi:10.1088/1126-6708/2006/11/017;%%
%55 citations counted in INSPIRE as of 06 Oct 2019
%\cite{Bellucci:2007ds}
\bibitem{Bellucci:2007ds}
S.~Bellucci, S.~Ferrara, R.~Kallosh and A.~Marrani,
%``Extremal Black Hole and Flux Vacua Attractors,''
Lect.\ Notes Phys.\ {\bf 755}, 115 (2008)
[arXiv:0711.4547 [hep-th]].
%%CITATION = ARXIV:0711.4547;%%
%85 citations counted in INSPIRE as of 06 Oct 2019
%\cite{Larsen:2009fw}
\bibitem{Larsen:2009fw}
F.~Larsen and R.~O'Connell,
%``Flux Attractors and Generating Functions,''
JHEP {\bf 0907}, 049 (2009)
%doi:10.1088/1126-6708/2009/07/049
[arXiv:0905.2130 [hep-th]].
%%CITATION = doi:10.1088/1126-6708/2009/07/049;%%
%5 citations counted in INSPIRE as of 06 Oct 2019
%\cite{Kachru:2002he}
\bibitem{Kachru:2002he}
S.~Kachru, M.~B.~Schulz and S.~Trivedi,
%``Moduli stabilization from fluxes in a simple IIB orientifold,''
JHEP {\bf 0310} (2003) 007
%doi:10.1088/1126-6708/2003/10/007
[hep-th/0201028].
%\cite{Frey:2002hf}
\bibitem{Frey:2002hf}
A.~R.~Frey and J.~Polchinski,
%``N=3 warped compactifications,''
Phys.\ Rev.\ D {\bf 65}, 126009 (2002)
%doi:10.1103/PhysRevD.65.126009
[hep-th/0201029].
%%CITATION = doi:10.1103/PhysRevD.65.126009;%%
%226 citations counted in INSPIRE as of 27 Sep 2019
%\cite{DAuria:2002qje}
\bibitem{DAuria:2002qje}
R.~D'Auria, S.~Ferrara and S.~Vaula,
%``N=4 gauged supergravity and a IIB orientifold with fluxes,''
New J.\ Phys.\ {\bf 4}, 71 (2002)
%doi:10.1088/1367-2630/4/1/371
[hep-th/0206241].
%%CITATION = doi:10.1088/1367-2630/4/1/371;%%
%75 citations counted in INSPIRE as of 08 Oct 2019
%\cite{Frey:2002qc}
\bibitem{Frey:2002qc}
A.~R.~Frey and A.~Mazumdar,
%``Three form induced potentials, dilaton stabilization, and running moduli,''
Phys.\ Rev.\ D {\bf 67}, 046006 (2003)
%doi:10.1103/PhysRevD.67.046006
[hep-th/0210254].
%%CITATION = doi:10.1103/PhysRevD.67.046006;%%
%37 citations counted in INSPIRE as of 09 Oct 2019
%\cite{Dine:1985he}
\bibitem{Dine:1985he}
M.~Dine and N.~Seiberg,
%``Is the Superstring Weakly Coupled?,''
Phys.\ Lett.\ {\bf 162B}, 299 (1985).
%doi:10.1016/0370-2693(85)90927-X
%%CITATION = doi:10.1016/0370-2693(85)90927-X;%%
%310 citations counted in INSPIRE as of 09 Oct 2019
%\cite{Obied:2018sgi}
\bibitem{Obied:2018sgi}
G.~Obied, H.~Ooguri, L.~Spodyneiko and C.~Vafa,
%``De Sitter Space and the Swampland,''
arXiv:1806.08362 [hep-th].
%%CITATION = ARXIV:1806.08362;%%
%281 citations counted in INSPIRE as of 09 Oct 2019
%\cite{Garg:2018reu}
\bibitem{Garg:2018reu}
S.~K.~Garg and C.~Krishnan,
%``Bounds on Slow Roll and the de Sitter Swampland,''
arXiv:1807.05193 [hep-th].
%\cite{Ooguri:2018wrx}
\bibitem{Ooguri:2018wrx}
H.~Ooguri, E.~Palti, G.~Shiu and C.~Vafa,
%``Distance and de Sitter Conjectures on the Swampland,''
Phys.\ Lett.\ B {\bf 788}, 180 (2019)
%doi:10.1016/j.physletb.2018.11.018
[arXiv:1810.05506 [hep-th]].
%%CITATION = doi:10.1016/j.physletb.2018.11.018;%%
%163 citations counted in INSPIRE as of 09 Oct 2019
%\cite{Honma:2017uzn}
\bibitem{Honma:2017uzn}
Y.~Honma and H.~Otsuka,
%``On the Flux Vacua in F-theory Compactifications,''
Phys.\ Lett.\ B {\bf 774} (2017) 225
%doi:10.1016/j.physletb.2017.09.062
[arXiv:1706.09417 [hep-th]].
%\cite{Becker:1996gj}
\bibitem{Becker:1996gj}
K.~Becker and M.~Becker,
%``M theory on eight manifolds,''
Nucl.\ Phys.\ B {\bf 477}, 155 (1996)
%doi:10.1016/0550-3213(96)00367-7
[hep-th/9605053].
%%CITATION = doi:10.1016/0550-3213(96)00367-7;%%
%436 citations counted in INSPIRE as of 29 Sep 2019
%\cite{Sethi:1996es}
\bibitem{Sethi:1996es}
S.~Sethi, C.~Vafa and E.~Witten,
%``Constraints on low dimensional string compactifications,''
Nucl.\ Phys.\ B {\bf 480}, 213 (1996)
%doi:10.1016/S0550-3213(96)00483-X
[hep-th/9606122].
%%CITATION = doi:10.1016/S0550-3213(96)00483-X;%%
%328 citations counted in INSPIRE as of 29 Sep 2019
%\cite{Haack:2001jz}
\bibitem{Haack:2001jz}
M.~Haack and J.~Louis,
%``M theory compactified on Calabi-Yau fourfolds with background flux,''
Phys.\ Lett.\ B {\bf 507}, 296 (2001)
%doi:10.1016/S0370-2693(01)00464-6
[hep-th/0103068].
%%CITATION = doi:10.1016/S0370-2693(01)00464-6;%%
%132 citations counted in INSPIRE as of 29 Sep 2019
%\cite{Denef:2008wq}
\bibitem{Denef:2008wq}
F.~Denef,
%``Les Houches Lectures on Constructing String Vacua,''
Les Houches {\bf 87}, 483 (2008)
[arXiv:0803.1194 [hep-th]].
%%CITATION = ARXIV:0803.1194;%%
%365 citations counted in INSPIRE as of 29 Sep 2019
%\cite{Alim:2009bx}
\bibitem{Alim:2009bx}
M.~Alim, M.~Hecht, H.~Jockers, P.~Mayr, A.~Mertens and M.~Soroush,
%``Hints for Off-Shell Mirror Symmetry in type II/F-theory Compactifications,''
Nucl.\ Phys.\ B {\bf 841}, 303 (2010)
%doi:10.1016/j.nuclphysb.2010.06.017
[arXiv:0909.1842 [hep-th]].
%%CITATION = doi:10.1016/j.nuclphysb.2010.06.017;%%
%57 citations counted in INSPIRE as of 01 Oct 2019
%\cite{Grimm:2009ef}
\bibitem{Grimm:2009ef}
T.~W.~Grimm, T.~W.~Ha, A.~Klemm and D.~Klevers,
%``Computing Brane and Flux Superpotentials in F-theory Compactifications,''
JHEP {\bf 1004}, 015 (2010)
%doi:10.1007/JHEP04(2010)015
[arXiv:0909.2025 [hep-th]].
%%CITATION = doi:10.1007/JHEP04(2010)015;%%
%85 citations counted in INSPIRE as of 01 Oct 2019
%\cite{Jockers:2009ti}
\bibitem{Jockers:2009ti}
H.~Jockers, P.~Mayr and J.~Walcher,
%``On N=1 4d Effective Couplings for F-theory and Heterotic Vacua,''
Adv.\ Theor.\ Math.\ Phys.\ {\bf 14}, no. 5, 1433 (2010)
%doi:10.4310/ATMP.2010.v14.n5.a3
[arXiv:0912.3265 [hep-th]].
%%CITATION = doi:10.4310/ATMP.2010.v14.n5.a3;%%
%55 citations counted in INSPIRE as of 01 Oct 2019
%\cite{Grimm:2010gk}
\bibitem{Grimm:2010gk}
T.~W.~Grimm, A.~Klemm and D.~Klevers,
%``Five-Brane Superpotentials, Blow-Up Geometries and SU(3) Structure Manifolds,''
JHEP {\bf 1105}, 113 (2011)
%doi:10.1007/JHEP05(2011)113
[arXiv:1011.6375 [hep-th]].
%%CITATION = doi:10.1007/JHEP05(2011)113;%%
%19 citations counted in INSPIRE as of 01 Oct 2019
%\cite{Honma:2015iza}
\bibitem{Honma:2015iza}
Y.~Honma and M.~Manabe,
%``Open Mirror Symmetry for Higher Dimensional Calabi-Yau Hypersurfaces,''
JHEP {\bf 1603}, 160 (2016)
%doi:10.1007/JHEP03(2016)160
[arXiv:1507.08342 [hep-th]].
%%CITATION = doi:10.1007/JHEP03(2016)160;%%
%1 citations counted in INSPIRE as of 01 Oct 2019
%\cite{Honma:2013hma}
\bibitem{Honma:2013hma}
Y.~Honma and M.~Manabe,
%``Exact Kahler Potential for Calabi-Yau Fourfolds,''
JHEP {\bf 1305}, 102 (2013)
%doi:10.1007/JHEP05(2013)102
[arXiv:1302.3760 [hep-th]].
%%CITATION = doi:10.1007/JHEP05(2013)102;%%
%36 citations counted in INSPIRE as of 01 Oct 2019
%\cite{Grimm:2013gma}
\bibitem{Grimm:2013gma}
T.~W.~Grimm, R.~Savelli and M.~Weissenbacher,
%``On \alpha' corrections in N=1 F-theory compactifications,''
Phys.\ Lett.\ B {\bf 725} (2013) 431
%doi:10.1016/j.physletb.2013.07.024
[arXiv:1303.3317 [hep-th]].
%\cite{Grimm:2013bha}
\bibitem{Grimm:2013bha}
T.~W.~Grimm, J.~Keitel, R.~Savelli and M.~Weissenbacher,
%``From M-theory higher curvature terms to $\alpha'$ corrections in F-theory,''
Nucl.\ Phys.\ B {\bf 903} (2016) 325
%doi:10.1016/j.nuclphysb.2015.12.011
[arXiv:1312.1376 [hep-th]].
%\cite{Minasian:2015bxa}
\bibitem{Minasian:2015bxa}
R.~Minasian, T.~G.~Pugh and R.~Savelli,
%``F-theory at order $\alpha'^3$,''
JHEP {\bf 1510}, 050 (2015)
%doi:10.1007/JHEP10(2015)050
[arXiv:1506.06756 [hep-th]].
%%CITATION = doi:10.1007/JHEP10(2015)050;%%
%32 citations counted in INSPIRE as of 20 Oct 2019
%\cite{Weissenbacher:2019mef}
\bibitem{Weissenbacher:2019mef}
M.~Weissenbacher,
%``F-theory Vacua and $\alpha'$-Corrections,''
arXiv:1901.04758 [hep-th].
%%CITATION = ARXIV:1901.04758;%%
%5 citations counted in INSPIRE as of 29 Sep 2019
%\cite{Ferrara:1995ih}
\bibitem{Ferrara:1995ih}
S.~Ferrara, R.~Kallosh and A.~Strominger,
%``N=2 extremal black holes,''
Phys.\ Rev.\ D {\bf 52}, R5412 (1995)
%doi:10.1103/PhysRevD.52.R5412
[hep-th/9508072].
%%CITATION = doi:10.1103/PhysRevD.52.R5412;%%
%834 citations counted in INSPIRE as of 13 Oct 2019
%\cite{Ferrara:2008hwa}
\bibitem{Ferrara:2008hwa}
S.~Ferrara, K.~Hayakawa and A.~Marrani,
%``Lectures on Attractors and Black Holes,''
Fortsch.\ Phys.\ {\bf 56}, 993 (2008)
%doi:10.1142/9789814293242_0003, 10.1002/prop.200810569
[arXiv:0805.2498 [hep-th]].
%%CITATION = doi:10.1142/9789814293242_0003, 10.1002/prop.200810569;%%
%66 citations counted in INSPIRE as of 13 Oct 2019
%\cite{Moore:2004fg}
\bibitem{Moore:2004fg}
G.~W.~Moore,
%``Strings and Arithmetic,''
%doi:10.1007/978-3-540-30308-4-8
hep-th/0401049.
%%CITATION = doi:10.1007/978-3-540-30308-4_8;%%
%82 citations counted in INSPIRE as of 23 Oct 2019
%\cite{Denef:2004ze}
\bibitem{Denef:2004ze}
F.~Denef and M.~R.~Douglas,
%``Distributions of flux vacua,''
JHEP {\bf 0405} (2004) 072
%doi:10.1088/1126-6708/2004/05/072
[hep-th/0404116].
%\cite{Blumenhagen:2003vr}
\bibitem{Blumenhagen:2003vr}
R.~Blumenhagen, D.~Lust and T.~R.~Taylor,
%``Moduli stabilization in chiral type IIB orientifold models with fluxes,''
Nucl.\ Phys.\ B {\bf 663}, 319 (2003)
%doi:10.1016/S0550-3213(03)00392-4
[hep-th/0303016].
%%CITATION = doi:10.1016/S0550-3213(03)00392-4;%%
%191 citations counted in INSPIRE as of 07 Oct 2019
%\cite{Cascales:2003zp}
\bibitem{Cascales:2003zp}
J.~F.~G.~Cascales and A.~M.~Uranga,
%``Chiral 4d string vacua with D branes and NSNS and RR fluxes,''
JHEP {\bf 0305}, 011 (2003)
%doi:10.1088/1126-6708/2003/05/011
[hep-th/0303024].
%%CITATION = doi:10.1088/1126-6708/2003/05/011;%%
%162 citations counted in INSPIRE as of 07 Oct 2019
%\cite{Grimm:2019ixq}
\bibitem{Grimm:2019ixq}
T.~W.~Grimm, C.~Li and I.~Valenzuela,
%``Asymptotic Flux Compactifications and the Swampland,''
arXiv:1910.09549 [hep-th].
%%CITATION = ARXIV:1910.09549;%%
%\cite{Hulsey:2019xdb}
\bibitem{Hulsey:2019xdb}
G.~Hulsey, S.~Kachru, S.~Yang and M.~Zimet,
%``Distributions of extremal black holes in Calabi-Yau compactifications,''
arXiv:1901.10614 [hep-th].
%%CITATION = ARXIV:1901.10614;%%
%\cite{Dasgupta:2014pma}
\bibitem{Dasgupta:2014pma}
K.~Dasgupta, R.~Gwyn, E.~McDonough, M.~Mia and R.~Tatar,
%``de Sitter Vacua in Type IIB String Theory: Classical Solutions and Quantum Corrections,''
JHEP {\bf 1407} (2014) 054
%doi:10.1007/JHEP07(2014)054
[arXiv:1402.5112 [hep-th]].
%\cite{Kehagias:2019akr}
\bibitem{Kehagias:2019akr}
A.~Kehagias, D.~L$\ddot{\textrm{u}}$st and S.~L$\ddot{\textrm{u}}$st,
%``Swampland, Gradient Flow and Infinite Distance,''
arXiv:1910.00453 [hep-th].
%%CITATION = ARXIV:1910.00453;%%
%1 citations counted in INSPIRE as of 06 Oct 2019
%\cite{Bena:2018fqc}
\bibitem{Bena:2018fqc}
I.~Bena, E.~Dudas, M.~Gra$\tilde{n}$a and S.~L$\ddot{\textrm{u}}$st,
%``Uplifting Runaways,''
Fortsch.\ Phys.\ {\bf 67} (2019) no.1-2, 1800100
%doi:10.1002/prop.201800100
[arXiv:1809.06861 [hep-th]].
\end{thebibliography}
\end{document}