Gravitational duality, Palatini variation and boundary terms: A synopsis
Extracted source text ยท arXiv 2210.10383
\documentclass[12pt]{article}
%%%%%%%%%%%%%%%%%%%%%%
%\usepackage{pdfsync}
\usepackage{amsmath,amsfonts,epsfig,mathrsfs,todonotes,yfonts}
%\usepackage{color}
\usepackage{eufrak}
\usepackage{tikz}
\usetikzlibrary{shapes,arrows}
\usepackage{xcolor}
\usepackage{cite}
\usepackage{stmaryrd}
\usepackage{mdframed}
%\usepackage{scalerel}
\usepackage[top=2.5cm, bottom=2.5cm, left=2.75cm, right=2.75cm]{geometry}
\usepackage{dsfont}
\numberwithin{equation}{section}
\usepackage{hyperref}
\hypersetup{
colorlinks=true,
linkcolor=black,
filecolor=blue,
citecolor=blue,
urlcolor=cyan,
}
\usepackage[most]{tcolorbox}
\def\baselinestretch{1.1}
%\parskip 3 pt
\newcommand{\re}[1] {(\ref{#1})}
\newcommand{\+}{{+\!\!\!+}}
\newcommand{\one}{{\mathds{1}}}
\newcommand{\pa}{\partial} % curly d
\newcommand{\pat}{\tilde \pa} % curly d tilde
\newcommand{\ta}{{\tilde A}}
\newcommand{\taB}{{\tilde\bA}}
\newcommand{\bA}{\mathbf{A}}
\newcommand{\bAB}{{\bar\bA}}
\newcommand{\tb}{{\tilde B}}
\newcommand{\bB}{\mathbf{B}}
\newcommand{\bBB}{{\bar\bB}}
\newcommand{\tc}{{\tilde C}}
\newcommand{\td}{{\tilde D}}
\newcommand{\bI}{\mathbf{i}}
\newcommand{\bJ}{\mathbf{j}}
\newcommand{\bIB}{{\bar\bI}}
\newcommand{\bJB}{{\bar\bJ}}
\newcommand{\hal}{{\hat\al}}
\newcommand{\hatq}{{\hat q}}
\newcommand{\aB}{{\bar a}}
\newcommand{\bbR}{{\mathbb{R}}}
\newcommand{\then}{\Rightarrow}
\newcommand{\lrow}{\leftrightarrow}
\newcommand{\al}{\alpha}
\newcommand{\be}{\beta}
\newcommand{\ga}{\gamma}
\newcommand{\Ga}{\Gamma}
\newcommand{\de}{\delta}
\newcommand{\ka}{\kappa}
\newcommand{\vp}{\varphi}
\newcommand{\La}{\Lambda}
\newcommand{\la}{\lambda}
\newcommand{\Om}{\Omega}
\newcommand{\tE}{\tilde{E}}
\newcommand{\tG}{\tilde{G}}
\newcommand{\bl}[1]{{\color{blue}{#1}}}
\newcommand{\ber}{\begin{eqnarray}}
\newcommand{\eer}[1]{\label{#1}\end{eqnarray}}
\newcommand{\eero}{\end{eqnarray}}
\newcommand{\balg}{\begin{align}}
\newcommand{\ealg}{\end{align}}
\newcommand{\beq}{\begin{equation}}
\newcommand{\eeq}{\end{equation}}
\newcommand{\bea}{\begin{eqnarray}}
\newcommand{\eea}{\end{eqnarray}}
\newcommand{\rd}[1]{{\color{red}{#1}}}
\newcommand{\nn}{\nonumber}
\newcommand{\na}{\nabla}
\newcommand{\nbar}{\bar\na}
\newcommand{\pbar}{\bar \pa}
\newcommand{\half}{{\textstyle{\frac12}}}
\newcommand{\ihalf}{{\textstyle{\frac i 2}}}
\newcommand{\quart}{{\textstyle{\frac14}}}
\newcommand{\Ka}{{K\"ahler}}
\newcommand{\bbX}[1]{\mathbb{X}^{#1}}
\newcommand{\smp}[1]{{\scaleto{({#1})}{7pt}}}
\newcommand{\bbXB}[1]{\bar{\mathbb{X}}^{#1}}
\newcommand{\bbD}[1]{\mathbb{D}_{#1}}
\newcommand{\bbDB}[1]{\bar{\mathbb{D}}_{#1}}
%\newcommand{\re}[1] {(\ref{#1})}
\def\HollowBox #1#2{{\dimen0=#1 \advance\dimen0 by -#2
\dimen1=#1 \advance\dimen1 by #2
\vrule height #1 depth #2 width #2
\vrule height 0pt depth #2 width #1
\llap{\vrule height #1 depth -\dimen0 width \dimen1}
\hskip -#2
\vrule height #1 depth #2 width #2}}
\def\BOX{\HollowBox{.100in}{.010in}}
%%%%%%%%%
\renewcommand{\baselinestretch}{1.2}
\font\bfcal=eusb10 at 12pt
\usepackage{fancyhdr}
%%%%%%%%%%%%%
\newcommand{\auth}{\large Ulf Lindstr\"om ${}^{a,b}$\footnote{email: ulf.lindstrom@physics.uu.se}
and {\"O}zg{\"u}r Sar{\i}o\u{g}lu ${}^a$\footnote{email: sarioglu@metu.edu.tr}}
\begin{document}
%\pagestyle{fancy}
%\fancyhf{}
%\rhead{DRAFT}
\begin{flushright}
%{\small AEI-2009-087}\\
{\small UUITP-43/22}\\
%{\small Imperial-TP-UL-02}\\
\vskip 1.5 cm
\end{flushright}
\begin{center}
{\Large{\bf Gravitational duality, Palatini variation and boundary terms: A synopsis}}
\vspace{.75cm}
\auth
\end{center}
\vspace{.5cm}
\vspace{.5cm}
\centerline{${}^a${\it \small Department of Physics, Faculty of Arts and Sciences,}}
\centerline{{\it \small Middle East Technical University, 06800, Ankara, Turkey}}
\vspace{.5cm}
\centerline{${}^b${\it \small Department of Physics and Astronomy, Theoretical Physics, Uppsala University}}
\centerline{{\it \small SE-751 20 Uppsala, Sweden }}
\vspace{1cm}
%\today
\centerline{{\bf Abstract}}
\bigskip
\noindent
We consider $f(R)$ gravity and Born-Infeld-Einstein (BIE) gravity in formulations where
the metric and connection are treated independently and integrate out the metric to find the
corresponding models solely in terms of the
connection, the archetypical treatment being that of Eddington-Schr\"odinger (ES) duality between cosmological Einstein and Eddington theories. For
dimensions $D\ne2$, we find that this requires $f(R)$ to have a specific form which makes the
model Weyl invariant, and that its Eddington reduction is then equivalent to that of BIE with certain parameters. For $D=2$
dimensions, where ES duality is not applicable, we find that both models are Weyl invariant and
equivalent to a first order formulation of the bosonic string. We also discuss the form of the boundary
terms needed for the variational principle to be well defined on manifolds with boundaries. This requires a modification of the Gibbons-Hawking-York boundary term for gravity. This modification also means that the dualities between metric and connection formulations include the boundary terms.
\vskip .5cm
\vspace{0.5cm}
\small
\renewcommand{\thefootnote}{\arabic{footnote}}
\setcounter{footnote}{0}
%\baselineskip=24pt
\pagebreak
\tableofcontents
\setcounter{page}{2}
\section{Introduction}
There has been a lot of interest in modified theories of gravity, often motivated by the singularities in black hole solutions of General Relativity (GR) or the abundance of dark energy and dark matter. A popular venue for modifying GR has been to treat the connection as an independent variable. This can lead to a dual theory, as in the Eddington-Schr\"odinger (ES) duality, or to a new model, as is the case already in the Palatini approach to GR if the matter coupling leads to a violation of the
metricity condition\footnote{Historical note: One of us pioneered applying the ``Palatini variational principle'' to certain matter coupled gravity theories \cite{Lindstrom:1975ry,Lindstrom:1976pq}, almost fifty years ago. Not many papers followed on these until the early 2000s, but since then the topic has grown impressively. Now there are many hundred papers, mainly due to applications in astrophysics.}.
In this letter we compare two ways of constructing a purely connection dependent gravity model starting from $f(R)$ theory or from Born-Infeld-Einstein (BIE) theory\footnote{Also known as Eddington-inspired Born-Infeld gravity.}. We find that certain Weyl invariant $f(R)$ theories lead to models that are equivalent to those from BIE theory for $D\ne2$ where they both give actions involving the square root of the determinant of the Ricci tensor. The constructions break down for $D=2$, where they lead to Weyl invariant models still involving both the metric and the connection. These models are then both equivalent to the first order action of the bosonic string.
For each model we find the appropriate generalisations of the Gibbons-Hawking-York (GHY)
boundary terms that ensure a consistent derivation of the field equations in manifolds with
boundaries under a variational principle that we specify. These terms are needed for complete duality.
The layout of the paper is that in secs. \!\ref{ES}-\ref{EIG} we remind the reader of ES duality,
$f(R)$ theory and BIE theory, respectively. There we display the purely connection and the dual
purely metric theories in the cases when they exist separately, as well as the mixed theories.
Sec. \!\ref{2D} contains the $2D$ string-limit of the theories. In sec. \!\ref{bdy} we discuss
variational principles and boundary terms. We collect some conclusions in sec. \!\ref{conc}.
\section{Eddington-Schr\"odinger }
\label{ES}
The original form of ES duality starts from the action
\cite{Schrodinger:2011gqa, Eddington},
\ber
S_{ES}=\half\int d^Dx \sqrt{|g|}\Big(g^{ab}R_{ab}(\Gamma)-\Lambda\Big)
\eer{act2}
with metric and (symmetric) connection as independent variables and $|g|$ the absolute value of the determinant of the metric $g_{ab}$. (The original discussion
has $D=4$.) Eliminating $\Gamma$ via its field equations determines the connection to be
the Levi-Civita connection $\Gamma^{(0)}$ via the Palatini variation of the Ricci tensor
\ber
&&\delta R_{ab}=\na_c\delta \Gamma^c_{~ab}-\na_{(a}\delta\Gamma^d_{~b)d}~.
\eer{pal}
This returns the GR action with a cosmological term when substituted into the action
\ber
S^g_{ES}=\half\int d^Dx \sqrt{|g|}\Big(g^{ab}R_{ab}(\Gamma^{(0)})-\Lambda\Big)~.
\eer{act22}
The boundary contribution generated in the process reads
\ber
\int \na_{[c}\Big(\sqrt{|g|}g^{ab}\delta \Gamma^c_{~a]b}\Big)~.
\eer{ESB}
Varying the metric using
\ber
\delta\sqrt{|g|} =-\half \sqrt{|g|}g_{ab}\delta g^{ab}
\eer{gvar}
gives the field equations
\ber
R_{ab}-\half g_{ab}\left(g^{cd}R_{cd}-\Lambda\right)=0~.
\eer{ConEq}
When $D\ne 2$ this can be solved to give
\ber
g^{ab}R_{ab}=\frac D {D-2}\Lambda~.
\eer{}
From \re{ConEq} it then follows that
\ber
g_{ab}\Lambda=(D-2)R_{ab}
\eer{}
and thus
\ber
\sqrt{|g|}=\Big(\frac {D-2}\Lambda\Big)^{\frac D 2}\sqrt{|\det R_{ab}|}~.
\eer{}
The dual theory, with the connection as independent variable is then, from \re{act2},
\ber
S^\Gamma_{ES}=\Big(\frac {D-2} \Lambda \Big)^\frac {D-2} 2\int d^Dx \sqrt{|\det R_{ab}(\Gamma)|}~.
\eer{ES1}
\section{First order $f(R)$ and duality}
The second order $f(R)$ theories in $D$ dimensions are defined by an action
\ber
S^g_f= \int d^Dx \sqrt{|g|} f(R)~,
\eer{act1}
where
\ber
R(\Gamma^{(0)}) = g^{ab} R_{ab}(\Gamma^{(0)})~.
\eer{}
Here we will consider the action $S_f$, which is \re{act1} but with $R(\Gamma^{(0)})$
replaced by $R(\Gamma)$ where $\Gamma$ is a general (symmetric) connection.
Varying the metric and the connection independently as in the Palatini formulation of GR will
in general result in field equations inequivalent to those from \re{act1}, \cite{Olmo:2011uz}. Here
we are interested in the possibility of dualising the model to one written entirely in terms of
the connection. This procedure is related to the ES duality described
in sec. \!\ref{ES}.
The variation of the general action reads
\ber\nn
\delta S_f&=&\delta\!\!\int\!d^Dx \sqrt{|g|}f(R)\\[1mm]
&=&
\!\!\int\!d^Dx \Big(\sqrt{|g|}~\!\delta g^{ab}\left[-\half g_{ab}f(R)+f'(R)R_{ab}\right]+2g^{ab}f'(R)\sqrt{|g|}\na_{[c}\delta\Gamma^c_{~a]b}\Big)~,
\eer{vact2}
where use is made of the relations \re{pal} and \re{gvar}. Extremising the action will then result in a
boundary contribution
\ber
2\int\!d^Dx \na_{[c}\Big(g^{ab}f'(R)\sqrt{|g|}\delta\Gamma^c_{~a]b}\Big)
\eer{FRB}
and the field equations
\ber\nn
&&f'(R)R_{ab}-\half g_{ab}f(R)=0 \,, \\[1mm]
&&\na_c\big(\sqrt{|g|}g^{d b}f'(R)\big)-\delta^d_c\na_a\big(\sqrt{|g|} g^{ab}f'(R)\big)=0~.
\eer{feq}
Taking the determinant and the trace of the first of these results in
\ber
\sqrt{|g|}=\Big|\frac{2f'} f\Big|^{\frac D 2}\sqrt{|\det R_{ab} |}
\eer{detg}
and
\ber
D=\frac{2f'} f R \,,
\eer{ODE}
respectively. The equation \re{ODE} gives a first order ordinary differential equation for $f$ which we solve
\ber
\frac {df} f=\frac D 2 \frac {dR} R~~~\Rightarrow~~~f(R)=cR^{\frac D 2}
\eer{fRspec}
with $c$ a constant. Combining this with \re{detg} eliminates the metric completely and
gives the action
\ber
S_f=\int\!d^Dx\sqrt{|g|}f(R)~~ \to~~~ S_f^\Gamma=cD^{\frac D 2}\int\!d^Dx \sqrt{|\det R_{ab}(\Gamma )|} \,.
\eer{dual}
The action on the left hand side is invariant under Weyl transformations of the metric
\ber
g_{ab}\to \Omega(X)g_{ab}
\eer{}
in parallel to the construction of Weyl invariant (spinning and super) p-branes
\cite{Lindstrom:1987cv,Lindstrom:1987ps,Lindstrom:1987fr,Lindstrom:1988zm,Lindstrom:1988az}.
This becomes manifest if we rewrite the action $S_f^\Gamma$ in \re{dual} as
\ber
c\int\!d^Dx\sqrt{|\det (g_{ab}(g^{cd}R_{cd}))|}~.
\eer{}
Tracing the second equation in \re{feq} over $(c,d)$ implies
\ber
\na_c\big(\sqrt{|g|}g^{c b}f'(R)\big)=0~,
\eer{cfe}
in all dimensions $D\ne2$, so that, for the solution \re{fRspec} we have
\ber
0=\na_c\big(\sqrt{|g|}g^{d b}f'(R)\big)=\na_c\big(\sqrt{|g|}g^{d b}R^{\frac {D-2}2}\big)
=\na_c\big(\sqrt{|g|}g^{d b}\big)R^{\frac {D-2}2}+\sqrt{|g|}g^{d b}\na_cR^{\frac {D-2}2}.
\eer{IdId}
Tracing on $(d,b)$ and using that
\ber
g_{db}\na_c(\sqrt{|g|}g^{db})=(D-2)\na_c(\sqrt{|g|})~,
\eer{}
we find
\ber\frac{\na_c\sqrt{|g|}}{\sqrt{|g|}}+\frac D 2 R^{-1}\na_cR=0~~~
\Rightarrow~~~\pa_c \ln{\sqrt{|g|}}-\Gamma_c+\frac D 2 {\na_c \ln R}=0~,
\eer{}
where $\Gamma_c$ is the contracted connection. Noting that $\pa_c \ln{\sqrt{|g|}}=\Gamma^{(0)}_c$, the contracted Levi-Civita connection, and defining $R^{-1}\na_cR=:\pa_c\varphi$, we have
\ber
\Gamma_c=\Gamma^{(0)}_c+\frac D 2\pa_c\varphi~.
\eer{Gam3}
Making use of this in \re{IdId} yields
\ber\nn
&&\na_c\big(\sqrt{|g|}g^{d b}\big)+\big(\frac {D-2}2\Big)\sqrt{|g|}g^{d b}R^{-1}\na_cR=0~,\\[1mm]\nn
&&\implies \pa_c(\ln \sqrt{|g|})g^{d b}-\Gamma^{(0)}_cg^{d b}+\na_cg^{d b}-\frac D 2\pa_c\varphi g^{d b}+
\big(\frac {D-2}{2}\Big)g^{d b}\pa_c\varphi=0\\[1mm]
&&\implies \na_cg^{d b}-\pa_c\varphi g^{d b}=0~,
\eer{setup}
where we have used \re{Gam3} in an intermediate step.
The last equation in \re{setup} may be solved in the way the Christoffel symbols are found from the metricity condition to give
\ber
\Gamma^a_{~cd}=:\tilde\Gamma^a_{~cd}=\Gamma^{(0)a}_{~~~~cd}+\half g^{ab}\left(2g_{b(c}\pa_{d)}\varphi - g_{cd }\pa_b\varphi\right)~,
\eer{GAM}
with $\Gamma^{(0)a}_{~~~~cd}$ the Levi-Civita connection\footnote{In $D=2$ the solution contains an independent vector as part of the first order version of the bosonic string in \cite{Lindstrom:1987sg}. There it also corresponds to an additional symmetry of the model, unique for $D=2$.}. This is consistent with \re{Gam3} and shows that the connection $\tilde\Gamma^a_{~cd}$ equals the Levi-Civita connection $\Gamma^{(0)a}_{~~~~cd}$ up to a Weyl transformation
\ber
g^{d b} \to e^\varphi g^{d b}~.
\eer{}
Since the model is Weyl invariant, we can thus always choose the connection to be Levi-Civita.
In the preceding derivation we only used the equation \re{ODE} from the metric
field equations to determine the form of $f(R)$. So if we assume that form to begin with, the
result \re{GAM} implies that the action $S_f^\Gamma$ in \re{dual} is dual to
\ber
S_f^{g}=\int\!d^Dx\sqrt{|g|}R^{\frac D 2}(\tilde\Gamma)~.
\eer{}
\section{Born-Infeld-Einstein Gravity}
\label{EIG}
In this section we relate the previous results to a gravitational model based on an analogy
to the Born-Infeld action. We shall be brief in our presentation, since we recently discovered
that several of the calculations already exist in the literature (for $D=4$)
\cite{Banados:2010ix,Nascimento:2019qor}.
The action for BIE gravity \cite{Deser:1998rj}, with independent (symmetric)
connection reads
\ber
S_{BIE}=\frac 1 \epsilon \int d^Dx \Big(\sqrt{|\det(g_{ab} + \epsilon R_{ab}(\Gamma)|}
-\lambda \sqrt{|\det(g_{ab}|}\Big)~,
\eer{act3}
where the Ricci tensor is assumed symmetric and $\epsilon$ is a parameter of dimension (length)${}^2$. Abbreviating
\ber
h_{ab}:= g_{ab} + \epsilon R_{ab}~,~~~h^{ab}h_{bc}=\delta^a_c~,
\eer{metric2}
and extremising $S_{BIE}$ by varying the metric and the connection using \re{pal} and \re{gvar},
yields a boundary contribution
\ber
\int d^Dx\na_{[c}(\sqrt{|h|}h^{ab}\delta\Gamma^c_{~a]b})
\eer{EiBIB}
along with the field equations
\ber
\sqrt{|h|}h^{ab}-\lambda \sqrt{|g|} g^{ab}=0
\eer{f1}
and
\ber
\na_{c}(\sqrt{|h|}h^{d b})-\delta^d_c\na_{a}(\sqrt{|h|}h^{ab})=0~.
\eer{f2}
%Tracing the second of these over $i$ and $c$ and using the result in the first, gives
%\ber\nn
%&&\na_a(\sqrt hh^{ab})=0\\[1mm]
%&&\na_a(\sqrt gg^{ab})=0~.
%\eer{}
When $D\ne 2$ we can solve \re{f1} to give
\ber
\sqrt{\frac{|h|}{|g|} }=\lambda ^\frac{D}{D-2}
\eer{}
which substituted back yields
\ber
g_{ab}=\lambda^\frac{-2}{D-2}h_{ab}=\lambda^\frac{-2}{D-2}(g_{ab} + \epsilon R_{ab})~,
\eer{geqn}
so that
\ber
g_{ab}=\frac {\epsilon R_{ab}(\Gamma)}
{(\lambda^\frac2{D-2}-1)}~,
\eer{gtoR}
thus eliminating the metric in favour of the connection. Inserting this in \re{act3}, we find an action
with the connection as the only variable
\ber
S^\Gamma_{BIE}=\frac 1 \epsilon \int d^Dx\big(\sqrt{|h|}-\lambda \sqrt{|g|}\big)=\epsilon^{{\frac D 2 }-1}(\lambda^\frac2{D-2}-1)^{\frac {2-D}2} \int d^Dx\sqrt{|\det R_{ab}(\Gamma)|}~.
\eer{EDD}
For
\ber
\lambda =\Big[\frac {\epsilon\Lambda +D-2}{D-2}\Big]^{\frac{D-2}2}~,
\eer{Lrel}
this agrees with the ES result in \re{ES1}\footnote{For $D=4$ \re{Lrel} implies
$2(\lambda-1)= \epsilon\Lambda$.}.
The two models are thus equivalent as starting points for deriving the $R(\Gamma)$ theory,
although only one of them is Weyl invariant.
Finding a metric action dual to \re{EDD} by solving \re{f2} meets with several obstacles.
The solution to \re{f2} is formally
\ber
\Gamma=\Gamma^{(0)}(h_{ab})~,
\eer{Gamh}
i.e. the Levi-Civita connection for $h_{ab}$. But this is incomplete since $h_{ab}$ still contains
$R_{ab}(\Gamma)$. So it is rather the full metric and connection theory with field equations
\re{f1} {\em and} \re{f2} one has to consider.
Using \re{gtoR} we write
\ber
h_{ab} := g_{ab} + \epsilon R_{ab}=\lambda^\frac{2}{D-2}g_{ab} ~.
\eer{}
Since this is a constant Weyl transformation, plugging it into \re{Gamh} gives
\ber
\Gamma^a_{~bc}=\Gamma^{(0)}(h_{ab})=\Gamma^{(0)}(g_{ab})~,
\eer{Chris}
in perfect agreement with the right hand side of \re{f1}. The on-shell theory thus found
satisfies \re{gtoR} with connection \re{Chris} is
\ber
{(\lambda^\frac2{D-2}-1)}g_{ab}= {\epsilon R_{ab}\big(\Gamma^{(0)}(g_{ab})\big)}~,
\eer{gtoRR}
which is equivalent to Einstein's vacuum field equations with a cosmological constant
\ber
\half(2-D)\big(\lambda^\frac 2{D-2}-1\big)~.
\eer{}
This confirms the $D=4$ discussion of the matter coupled system in \cite{Banados:2010ix,Nascimento:2019qor} where it is shown that the vacuum solutions
are the Einstein solutions.
We saw in Sec.\ref{ES} that in $D=4$ the $\Gamma$-action \re{dual} is dual to GR with a cosmological constant
$-\half\Lambda = -\frac 1{16 c}$. For general $D$, it corresponds
to BIE with cosmological term
\[ \lambda =\Big[\frac {\epsilon\Lambda +D-2}{D-2}\Big]^{\frac{D-2}2}\]
We thus have an interesting chain of equivalences between ``metric'', ``affine'' and ``metric/affine'' models. See Fig. below for the interrelations.\\
\bigskip
\tikzstyle{decision} = [diamond, draw, fill=blue!20,
text width=4.5em, text badly centered, node distance=3cm, inner sep=0pt]
\tikzstyle{block} = [rectangle, draw, fill=blue!20,
text width=5em, text centered, rounded corners, minimum height=4em]
\tikzstyle{line} = [draw, -latex']
\tikzstyle{cloud}= [draw, ellipse,fill=red!20, node distance=4cm,
minimum height=3em]
\noindent
{\bf Fig.1. A schematic picture of the interrelation between the models discussed:
\bigskip
\noindent
\begin{tikzpicture}[node distance = 2cm, auto]
% Place nodes
\node [block] (init) {Eddington
$\int\!\!\sqrt{|\det R(\Gamma)|}$};
\node [cloud, above left of=init] (expert) { $S_{ES}(g,\Gamma,\Lambda)$};
\node [cloud, left of=expert] (expert2) {$S^g_{ES}(g,\Lambda)$};
\node [cloud, above right of=init] (system) {$S_{f(R)}(g,\Gamma)$};
\node [cloud, right of=system] (system2) { $S^g_{R^{D /2}}(g)$};
%\node [block, below of=init] (identify) {Superstrings};
\node [cloud, below of=init, node distance=3cm] (evaluate) {$S_{BIE}(g,\Gamma,\lambda)$};
\node [cloud, left of=evaluate, node distance=6cm] (update) {GR +$\Lambda$ eqns};
%\node [decision, below of=evaluate] (decide) {Killing tensors};
%\node [block,right of=evaluate, node distance=4cm] (stop) {Higher spin};
% Draw edges
\path [line] (expert)--node {$\delta\Gamma$} (expert2);
\path [line] (evaluate) -- node {$\delta g~~\delta\Gamma$} (update);
\path [line] (evaluate) -- node {$\delta g$} (init);
\path [line] (system) -- node {$\delta\Gamma$} (system2);
%\path [line] (update) |- (evaluate);
%\path [line] (decide) -- (stop);
\path [line] (expert) -- node {$\delta g$} (init);
\path [line] (system)--node {$\delta g$} (init) ;
\path [line,dashed] (init) -- node {Dual}(expert2);
\path [line,dashed] (expert2) -- (init);
\path [line,dashed] (init) -- (system2);
\path [line,dashed] (system2) -- node {Dual} (init);
\path [line] (expert2) -- node {$\delta g$} (update);
\end{tikzpicture}
}
\section{The $2D$ string limits}
\label{2D}
Clearly the results in the previous sections do not hold when $D=2$.
When $D=2$, the discussion of the model \re{act3} is as follows: Taking determinants in equation \re{f1}
leads to
\ber
\lambda^2=1~~~\Rightarrow~~~\lambda=\pm 1~,
\eer{}
so that equation \re{geqn} gets replaced by
\ber
\pm g^{ac}h_{cb}=\pm (\delta^a_b+\epsilon g^{ac}R_{cb})
\eer{bosact}
leading to
\ber
\sqrt{|h|}-\lambda \sqrt{|g|}=\frac {\epsilon\lambda} 2\sqrt{|g|}g^{ab}R_{ab}(\Gamma)~.
\eer{conf}
Alternatively we can see this directly in the action \re{act3} using that in $D=2$
\ber
R_{ab}-\half g_{ab}R=0~
\eer{}
for a symmetric Ricci tensor and connection. Then,
\ber
S_{BIE}=\frac 1 \epsilon \int d^2x \Big(\sqrt{|\det (g_{ab} + \epsilon R_{ab}(\Gamma)|}
-\lambda \sqrt{|g|}\Big)= \frac 1 \epsilon \int d^2x\sqrt{|g|}\Big(1+\half \epsilon R-\lambda\Big)~,
\eer{act333}
which is the ES action \re{act2} with $\Lambda =\frac 2 \epsilon(\lambda - 1)$.
However the action \re{act2} is again equivalent to the right hand side of \re{conf} since tracing the field equations \re{ConEq}
now results in
\ber
g^{ab}R_{ab}-\left(g^{cd}R_{cd}-\Lambda\right)=0~~~~\Rightarrow \Lambda =0~.
\eer{ConEq2}
Plugging this back into the action \re{act2} (for $D=2$) again returns the mixed variable
action \re{conf}\footnote{With $\lambda\epsilon=2$.}.
Thus, unlike the case $D\ne 2$, both models depend on both metric and connection as independent
variables. Indeed, the actions \re{act2} and \re{act3} in two dimensions reproduces the conformally
invariant first order action for the bosonic string, introduced in \cite{Lindstrom:1987sg}. At first sight
this action may look strange, since we are taught in string theory to ignore the $2D$ curvature term
because it is a total derivative. This is indeed true for the curvature for the spin connection, or
equivalently when the connection is Levi-Civita. It is not true, however, for the curvature of an arbitrary
affine connection.
Briefly, \re{conf} is invariant, up to a boundary contribution, under Weyl rescaling of the metric and
transformations of the affine connection that read
\ber
\Gamma^a_{~bc}\to\Gamma^a_{~bc}+2V_{(b}\delta^a_{c)}-g_{cb}V^a~.
\eer{Gsym}
The $\Gamma$ field equations lead to
\ber
%&&\na_a(\sqrt hh^{ab})=0\\[1mm]
\na_a(\sqrt{|g|}g^{bc})=0
\eer{}
as usual, but unlike when $D\ne2$, the solution is not the Levi-Civita connection $\Gamma^{(0)}$, but
\ber
\Gamma^a_{~bc}=\Gamma^{(0)a}{}_{bc}+2U_{(b}\delta^a_{c)}-g_{cb}U^a~,
\eer{}
with $U^a$ the most general parameter of the transformation \re{Gsym}. It can then be shown that
there is a gauge choice where the model becomes the usual Nambu-Goto string.
\section{Boundary actions}
\label{bdy}
In this section we consider what boundary terms need to be added to the actions to cancel the
respective contributions \re{ESB},\re{FRB} and \re{EiBIB} and make the variational problem well
defined in the presence of a boundary.
\subsection{Variational principle}
We are concerned with both the variation of the metric $\delta g$ and the independent
connection $\delta\Gamma$. When the full set of field equations hold, $\Gamma$ gets
expressed in terms of derivatives of the metric. In this sense the system is reminiscent of
a Hamiltonian system $H(q,p)$ with the metric corresponding to the coordinates $q$
and $\Gamma$ to the momenta $p$. The field equation that gives $\Gamma=\Gamma(g,\pa g)$
then roughly corresponds to $\pa H/\pa p=\dot q$ which may be solved for $p=p(q,\dot q)$.
As in the variational principle for $H$, where the derivation of the canonical equations requires $\delta q=0$
on the boundary of the integration volume while no conditions are required for $\delta p$, we
shall require $\delta g=0$ on the boundary but leave $\delta\Gamma$ free. This choice can
also be motivated by the fact that on-shell the connection becomes the Levi-Civita connection
for the metric and its variation does not vanish on the boundary as well as by the wish to find
a dual action reproducing the boundary terms for both the dual models.
Apart from the extra variation of $\Gamma$, our derivation follows that
%We shall make use of the following relations from the derivation in \cite{WiKi}
of the Gibbons-Hawking-York (GHY) boundary term \cite{York:1972sj,Gibbons:1976ue}
described in \cite{Feng:2017ygy}.
The boundary is defined by
\ber
x^a = x^a(y^i) \,, a = 1, \dots, D \qquad i = 1, \dots,D-1 \,,
\eer{}
which leads to the tangential vectors to the boundary
\ber
e^a_i=\Big(\frac{\partial x^a}{\partial y^i}\Big)_{\pa{}{\cal M}}
\eer{}
and a normal $n_a$
\ber
e^a_in_a=0~.
\eer{}
The induced metric is
\ber
\gamma_{ij}=g_{ab}e^a_ie^b_j~.~~~%\makebox{and}~~~%\tilde\gamma_{ab}=g_{ab}-n_an_b~,
\eer{}
It is related to the full metric as
\ber
g_{ab}=\gamma_{ab}-n_an_b~,~~~g^{ab}=\gamma^{ab}+n^an^b~.
\eer{}
Gauss' divergence theorem reads
\ber
\int_{\cal M}d^Dx\pa_{a}(\sqrt{|g|}{V^a})= \int_{\pa{}\cal M} d^{D-1}y\sqrt{\gamma}~\!n_a{V^a}~.
\eer{}
In the variational principle we shall keep the normal to the boundary ${\pa{}{\cal M}}$, as well as its partial derivatives,
constant
\ber
\delta n_a=0~,~~~\delta n^a =0~,~~\delta n_a,_b=0~,~~~\delta n^a,_b =0~.
\eer{}
See sec.4 of \cite{Feng:2017ygy} for details.
We also require that the
metric is held constant when confined to the boundary
\ber
\delta g_{ab}{|_{\pa{}\cal M}}=0~,~~\delta g^{ab}{|_{\pa{}\cal M}}=0~.
\eer{delg}
This means that the induced metric $\gamma_{ij}$ is fixed during the variation. It also implies that,
although $\delta\pa_{c}g_{ab}$ does not vanish on ${\pa{}\cal M}$, the tangential derivatives must also vanish:
\ber
\delta g_{ab ,c }e_{i}^{c}=0~.
\eer{}
It follows that
\ber
\gamma^{ij}e_{i}^{a }e_{j}^{b}\delta g_{c b ,a}=0~.
\eer{}
\subsubsection{Gibbons-Hawking-York}
The GHY boundary action for GR is
\ber
\half \int_{\pa{}\cal M}d^{D-1}y \sqrt{\gamma} K~,
\eer{}
where
\ber
K=g^{ab}K_{ab}=g^{ab}\na_{a}n_b
\eer{TR}
is the trace of the extrinsic curvature (second fundamental form) $K_{ab}$.
There are various definitions of $K_{ab}$. One involving the Lie derivative reads
\ber
K_{ab}=\half{\cal L}_n\gamma_{ab}=\half n^c\na_c\gamma_{ab}+\gamma_{c(a}\na_{b)}n^c~.
\eer{}
This has a trace which is only equivalent to that used in \re{TR} when the connection is the Levi-Civita one.
\subsubsection{Eddington-Schr\"odinger}
For a general $\Gamma$ the boundary term will have to be modified. To this end we note that
\ber
\half \int_{\pa{}\cal M}d^{D-1}y \sqrt{\gamma} (g^{ab}\na_{a}n_b+\na_{a}n^a)=: \half\int_{\pa{}\cal M}d^{D-1}y\sqrt{\gamma} ~\hat K
\eer{bdyact1}
reduces to the GHY boundary action when the connection is metric. Further varying the connection
\ber
\half\delta\int_{\pa{}\cal M}d^{D-1}y \sqrt{\gamma} ~\hat K
=\half\int_{\pa{}\cal M}d^{D-1}y \sqrt{\gamma} ~ \delta\hat K~,
\eer{}
we find
\ber
\delta\hat K=\delta\big(g^{ab}K_{ab}+\na_an^b\big)=-2\delta \Gamma^{[ca]}_{~~~a}n_c
\eer{}
so that \re{bdyact1} precisely cancels the boundary contribution \re{ESB} that arises in the
$\Gamma $ variation of the ES bulk action \re{act2}.
\subsubsection{$f(R)$}
Similarly, we can cancel the boundary contribution \re{FRB} that arises in the
$\Gamma $ variation of the $f(R)$ bulk action \re{act1} by the varying the following boundary action
\ber
\half\int_{\pa{}\cal M}d^{D-1}y \sqrt{\gamma} f'(R)~\hat K~,
\eer{bdyact2}
provided that we also set $\delta R=0$ on the boundary. When the connection is metric, this reduces to the boundary term in \cite{Guarnizo:2010xr}. The vanishing of the variation of $R$ is required also there and in \cite{Dyer:2008hb}. In both these references the connection is the Levi-Civita connection, so the implications of $\delta R=0$ merits to be studied in more detail.
\subsubsection{Born-Infeld-Einstein gravity}
Formally, we can cancel the boundary contribution \re{EiBIB} that arises in the
$\Gamma $ variation of the BIE bulk action \re{act3} by varying the following boundary action
\ber
\half\int_{\pa{}\cal M}d^{D-1}y \sqrt{\gamma^h} ~\hat K^h~,
\eer{bdyact3}
i.e., the same boundary action as \re{bdyact1} but
with the induced metric $\gamma^h$ and $\hat K^h$ now defined with respect to the metric \re{metric2} so that
\ber
\gamma^h{}_{ij}= h_{ab}e^a_ie^b_j= (g_{ab} + \epsilon R_{ab})e^a_ie^b_j~,~~~\hat K^h=(h^{ab}\na_{a}n_b+\na_{a}n^a)~,~~~n^a=h^{ab}n_b~.
\eer{indmetric2}
The conditions \re{delg} become
\ber
\delta h_{ab}{|_{\pa{}\cal M}}=\delta (g_{ab} + \epsilon R_{ab}){|_{\pa{}\cal M}}=0~,~~~\delta h^{ab}{|_{\pa{}\cal M}}=0~.
\eer{}
Since \re{delg} also holds, this means that we must set
\ber
\delta R_{ab}{|_{\pa{}\cal M}}=0~,
\eer{}
in analogy to the $f(R)$ case in \re{bdyact2} described above. This is consistent with the equivalence of BIE theory with $f(R)$ for the special case \re{fRspec}.
\subsubsection{The bosonic string}
Finally we discuss the boundary terms for the $D=2$ first order bosonic string. This differs from \re{bdyact1} by a term proportional to $U^a$
\ber
\half\int_{\cal \pa{}M}dy\sqrt{\gamma}\big(\hat K+2U^an_a\big)~.
\eer{}
Here $\hat K$ takes care of the boundary contributions from the bulk variation of $R$ as before, while the action is still invariant under conformal symmetry of $\Gamma$
\ber
\Gamma^{a}_{~~bc} \to \Gamma^{a}_{~~bc} +2V_{(b}\delta^a_{c)}-g_{bc}V^a~.
\eer{confsy}
This is because
\ber
K\to K-g^{ab}\big(2V_{(a}\delta^c_{b)}-g_{ab}V^c
\big)n_c=K
\eer{}
while
\ber
U_a \to U_a+2V_{(a}\delta^b_{b)}-g_{ab}V^b=U_a+2V_a
\eer{2ndBdy}
and from the $R$ term in the bulk
\ber
\half \sqrt{|g|} R\to \half \sqrt{|g|} R-\half\pa_{a}\big(4\sqrt{|g|} V^a\big)
\eer{}
which gives a boundary contribution
\ber
-\sqrt{\gamma}V^a n_a
\eer{}
that will cancel the term from \re{2ndBdy}. So the $U$ boundary term ensures the invariance under \re{confsy} and by the same token takes care of a general variation of $U$ to make the variational principle well defined.
\section{Conclusions}
\label{conc}
We have studied first order models for $f(R)$ theories and BIE theory with metric and connection as independent variables, and compared them to the famous ES duality. We have shown that, when the metric is eliminated, certain Weyl invariant $f(R)$ models result in dual models that are equivalent to those of ES. Similarly, for BIE we showed that eliminating the metric again leads to the ES result. In the form where the connection is the only variable, BIE is thus equivalent to Weyl invariant $f(R)$ in its dual form.
These constructions hold when the space-time dimension $D>2$ but break down for
$D=2$. The models are then still equivalent but now equal to the first order bosonic string.
We have discussed boundary terms for all these theories in general (no specific form of $f(R)$). After describing our variational principle, we showed that the necessary boundary terms carry over from the connection form to the metric form under duality.
Open problems are to include torsion in the discussion as done for e.g. metric modified gravity in \cite{Olmo:2011uz}, to understand if and how the Weyl invariance of our $f(R)$ model can be related to symmetries of the BIE ation.
\bigskip
\noindent{\bf Acknowledgments}\\
The research of U.L. is supported in part by the 2236 Co-Funded
Scheme2 (CoCirculation2) of T\"UB{\.I}TAK (Project No:120C067)\footnote{\tiny However
the entire responsibility for the publication is ours. The financial support received from
T\"UB{\.I}TAK does not mean that the content of the publication is approved in a scientific
sense by T\"UB{\.I}TAK.}.
\begin{thebibliography}{6666}
%\cite{Lindstrom:1975ry}
\bibitem{Lindstrom:1975ry}
U.~Lindstr\"om,
``Comments on the Jordan-Brans-Dicke Scalar Field Theory of Gravitation,''
\href{https://dx.doi.org/10.1007/BF02727640}{{\em Nuovo Cim. B} {\bfseries 32} (1976), 298-302}.
%10 citations counted in INSPIRE as of 14 Oct 2022
%\cite{Lindstrom:1976pq}
\bibitem{Lindstrom:1976pq}
U.~Lindstr\"om,
``The Palatini Variational Principle and a Class of Scalar-Tensor
Theories,''
\href{https://dx.doi.org/10.1007/BF02726289}{{\em Nuovo Cim.
B} {\bfseries 35} (1976), 130-136}.
%13 citations counted in INSPIRE as of 14 Oct 2022
%\cite{Schrodinger:2011gqa}
\bibitem{Schrodinger:2011gqa}
E.~Schr\"odinger,
``Space-Time Structure,'' \\
\href{https://www.cambridge.org/tr/academic/subjects/physics/
theoretical-physics-and-mathematical-physics/space-time-
structure?format=PB}{{\em Cambridge University Press}, 2011,
ISBN 978-0-511-89221-9, 978-0-521-31520-3}.
%4 citations counted in INSPIRE as of 27 Sep 2022
\bibitem{Eddington}
A.~S.~Eddington,
``The Mathematical Theory of Relativity,''\\
\href{https://www.cambridge.org/tr/academic/subjects/physics/
cosmology-relativity-and-gravitation/mathematical-theory-
relativity?format=PB}{{\em 2nd ed. Cambridge Univ. Press.}
(1924)}.
%\cite{Olmo:2011uz}
\bibitem{Olmo:2011uz}
G.~J.~Olmo,
``Palatini Approach to Modified Gravity: f(R) Theories and
Beyond,''
\href{https://dx.doi.org/10.1142/S0218271811018925}{{\em Int.
J. Mod. Phys. D} {\bfseries 20}
(2011), 413-462};
\href{http://arxiv.org/abs/1101.3864}{{\ttfamily arXiv:1101.3864
[gr-qc]}}.
%535 citations counted in INSPIRE as of 27 Sep 2022
%\cite{Lindstrom:1987cv}
\bibitem{Lindstrom:1987cv}
U.~Lindstr\"om,
``First Order Actions for Gravitational Systems, Strings and
Membranes,''
\href{https://dx.doi.org/10.1142/S0217751X88001028}{{\em Int.
J. Mod. Phys. A} {\bfseries 3}
(1988), 2401}.
%34 citations counted in INSPIRE as of 02 Oct 2022
%\cite{Lindstrom:1987ps}
\bibitem{Lindstrom:1987ps}
U.~Lindstr\"om, M.~Ro\v cek and P.~van Nieuwenhuizen,
``A Weyl Invariant Rigid String,''
\href{https://dx.doi.org/10.1016/0370-2693(87)91362-1}{{\em
Phys. Lett. B} {\bfseries199}
(1987), 219}.
%29 citations counted in INSPIRE as of 01 Oct 2022
%\cite{Lindstrom:1987fr}
\bibitem{Lindstrom:1987fr}
U.~Lindstr\"om and M.~Ro\v cek,
``Bosonic and Spinning Weyl Invariant Rigid Strings,''
\href{https://dx.doi.org/10.1016/0370-2693(88)90081-0}{{\em
Phys. Lett. B} {\bfseries 201}
(1988), 63-66}.
%11 citations counted in INSPIRE as of 01 Oct 2022
%\cite{Lindstrom:1988zm}
\bibitem{Lindstrom:1988zm}
U.~Lindstr\"om and G.~Theodoridis,
``Weyl Invariant Super p-branes,''\\
\href{https://dx.doi.org/10.1016/0370-2693(88)90638-7}{{\em
Phys. Lett. B} {\bfseries 208}
(1988), 407-410}.
%15 citations counted in INSPIRE as of 01 Oct 2022
%\cite{Lindstrom:1988az}
\bibitem{Lindstrom:1988az}
U.~Lindstr\"om and M.~Ro\v cek,
``A Super-Weyl-Invariant Spinning Membrane,''\\
\href{https://dx.doi.org/10.1016/0370-2693(89)91419-6}{{\em
Phys. Lett. B} {\bfseries 218}
(1989), 207}.
%32 citations counted in INSPIRE as of 01 Oct 2022
%\cite{Lindstrom:1987sg}
\bibitem{Lindstrom:1987sg}
U.~Lindstr\"om and M.~Ro\v cek,
``A Gravitational First Order Action for the Bosonic String,''
\href{https://dx.doi.org/10.1088/0264-9381/4/4/002}{{\em Class.
Quant. Grav.} {\bfseries 4}
(1987), L79}.
%30 citations counted in INSPIRE as of 27 Sep 2022
%\cite{Banados:2010ix}
\bibitem{Banados:2010ix}
M.~Banados and P.~G.~Ferreira,
``Eddington's theory of gravity and its progeny,''\\
\href{https://dx.doi.org/10.1103/PhysRevLett.105.011101}{{\em
Phys. Rev. Lett.} {\bfseries 105}
(2010), 011101};
\href{https://journals.aps.org/prl/abstract/10.1103/
PhysRevLett.113.119901}{{\em
Erratum: Phys. Rev. Lett.} {\bfseries 113} (2014) no.11,
119901};
\href{http://arxiv.org/abs/1006.1769}{{\ttfamily arXiv:1006.1769
[astro-ph.CO]}}.
%313 citations counted in INSPIRE as of 01 Oct 2022
%\cite{Nascimento:2019qor}
\bibitem{Nascimento:2019qor}
J.~R.~Nascimento, G.~J.~Olmo, P.~J.~Porf\'\i{}rio,
A.~Y.~Petrov and A.~R.~Soares,
``Nonlinear $\sigma$-models in the Eddington-inspired Born-
Infeld Gravity,''\\
\href{https://dx.doi.org/10.1103/PhysRevD.101.064043}{{\em
Phys. Rev. D} {\bfseries 101}
(2020) no.6, 064043};
\href{http://arxiv.org/abs/1912.10779}{{\ttfamily
arXiv:1912.10779 [hep-th]}}.
%9 citations counted in INSPIRE as of 01 Oct 2022
%\cite{Deser:1998rj}
\bibitem{Deser:1998rj}
S.~Deser and G.~W.~Gibbons,
``Born-Infeld-Einstein actions?,''\\
\href{https://dx.doi.org/10.1088/0264-9381/15/5/001}{{\em
Class. Quant. Grav.} {\bfseries 15}
(1998), L35-L39};
\href{http://arxiv.org/abs/hep-th/9803049}{{\ttfamily arXiv:hep-th/
9803049 [hep-th]}}.
%250 citations counted in INSPIRE as of 27 Sep 2022
\bibitem{York:1972sj}
J.~W.~York, Jr.,
``Role of conformal three geometry in the dynamics of
gravitation,''\\
\href{https://dx.doi.org/10.1103/PhysRevLett.28.1082}{{\em
Phys. Rev. Lett.} {\bfseries 28}
(1972), 1082-1085}.
%1055 citations counted in INSPIRE as of 04 Oct 2022
%\cite{Gibbons:1976ue}
\bibitem{Gibbons:1976ue}
G.~W.~Gibbons and S.~W.~Hawking,
``Action Integrals and Partition Functions in Quantum Gravity,''
\href{https://dx.doi.org/10.1103/PhysRevD.15.2752}{{\em Phys.
Rev. D} {\bfseries 15} (1977), 2752-2756}.
%2956 citations counted in INSPIRE as of 04 Oct 2022
%\cite{Feng:2017ygy}
\bibitem{Feng:2017ygy}
J.~C.~Feng and R.~A.~Matzner,
``The Weiss Variation of the Gravitational Action,''
\href{https://dx.doi.org/10.1007/s10714-018-2420-2}{{\em Gen.
Rel. Grav. } {\bfseries 50}
(2018) no.8, 99};
\href{http://arxiv.org/abs/1708.04489}{{\ttfamily
arXiv:1708.04489 [gr-qc]}}.
%8 citations counted in INSPIRE as of 06 Oct 2022
%\cite{Guarnizo:2010xr}
\bibitem{Guarnizo:2010xr}
A.~Guarnizo, L.~Castaneda and J.~M.~Tejeiro,
``Boundary Term in Metric f(R) Gravity: Field Equations in the
Metric Formalism,''
\href{https://dx.doi.org/10.1007/s10714-010-1012-6}{{\em Gen.
Rel. Grav. } {\bfseries 42}
(2010), 2713-2728};
\href{http://arxiv.org/abs/1002.0617}{{\ttfamily arXiv:1002.0617
[gr-qc]}}.
%75 citations counted in INSPIRE as of 06 Oct 2022
%\cite{Dyer:2008hb}
\bibitem{Dyer:2008hb}
E.~Dyer and K.~Hinterbichler,
``Boundary Terms, Variational Principles and Higher Derivative
Modified Gravity,''\\
\href{https://dx.doi.org/10.1103/PhysRevD.79.024028}{{\em
Phys. Rev. D} {\bfseries 79}
(2009), 024028};
\href{http://arxiv.org/abs/0809.4033}{{\ttfamily arXiv:0809.4033
[gr-qc]}}.
%150 citations counted in INSPIRE as of 06 Oct 2022
\end{thebibliography}
\end{document}