Renormalization of ${}^1S_0$ NN scattering amplitude in effective field theory
Extracted source text ยท arXiv nucl-th/0108005
%revised version (re)submitted to PLB
%\documentstyle[preprint,aps,tighten,epsf,floats]{revtex}
\documentstyle[aps,tighten,epsf,floats]{revtex}
%\documentstyle[12pt]{article}
%%%%%%%%%%%%% Uncomment for A4 paper %%%%%%%%%%%%%
%\setlength{\textheight}{25.0cm}
%\setlength{\oddsidemargin}{-.5cm}
%\setlength{\topmargin}{-1.5cm}
%\setlength{\leftmargin}{-5.5cm}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\setlength{\textwidth}{16cm}
\newcommand{\nn}{\nonumber\\}
\newcommand{\NNN}{$N\!N\!N$}
\newcommand{\NNNNNN}{$N\!N\!N\rightarrow N\!N\!N$}
\newcommand{\bk}{\bar{k}}
\newcommand{\bq}{\bar{q}}
\newcommand{\bp}{\bar{p}}
\newcommand{\bu}{\bar{u}}
\newcommand{\tPhi}{\tilde{\Phi}}
\newcommand{\tphi}{\tilde{\phi}}
\newcommand{\tX}{\tilde{X}}
\newcommand{\tY}{\tilde{Y}}
\newcommand{\tT}{\tilde{T}}
\newcommand{\tv}{\tilde{v}}
\newcommand{\ttt}{\tilde{t}}
\newcommand{\td}{\tilde{d}}
\newcommand{\tj}{\tilde{j}}
\newcommand{\tG}{\tilde{G}}
\newcommand{\la}{\langle}
\newcommand{\ra}{\rangle}
\newcommand{\pslash}{\not\hspace{-0.7mm}p}
\newcommand{\ben}{\begin{displaymath}}
\newcommand{\een}{\end{displaymath}}
\newcommand{\be}{\begin{equation}}
\newcommand{\ee}{\end{equation}}
\newcommand{\bea}{\begin{eqnarray}}
\newcommand{\eea}{\end{eqnarray}}
\newcommand{\eqn}[1]{\label{#1}}
\newcommand{\eq}[1]{Eq.~(\ref{#1})}
\newcommand{\eqs}[1]{Eqs.~(\ref{#1})}
\newcommand{\fign}[1]{\label{#1}}
\newcommand{\fig}[1]{Fig.~\ref{#1}}
\newcommand{\bPsi}{\bar{\Psi}}
\newcommand{\bPhi}{\bar{\Phi}}
\newcommand{\bphi}{\bar{\phi}}
\newcommand{\bfp}{{\bf p}}
\newcommand{\bfP}{{\bf P}}
\newcommand{\bfq}{{\bf q}}
\newcommand{\bfk}{{\bf k}}
\begin{document}
% \draft command makes pacs numbers print
%\draft
\begin{center}
{\large \bf Renormalization of ${ }^1S_0$ NN scattering amplitude in Effective
Field Theory}
\end{center}
% repeat the \author\address pair as needed
\begin{center}
{ J. Gegelia}
\end{center}
\begin{center}
{\it INFN - Sezione di Ferrara, Via Paradiso 12,
44100 Ferrara (FE), Italy \\
and \\
High Energy Physics Institute of TSU, University Str.. 9, Tbilisi 380086,
Georgia}
\end{center}
\begin{center}
{ G. Japaridze}
\end{center}
\begin{center}
{\it Center for Theoretical Studies of Physical Systems, Clark Atlanta
University, \\
Atlanta, GA 30314, USA}
\end{center}
\date{\today}
%\maketitle
\begin{abstract}
Cutoff regularized subleading order ${ }^1S_0$ NN potential of effective field
theory(EFT) is iterated using Lippmann-Schwinger equation. It is shown that
the scattering amplitudes
calculated in cutoff and subtractively renormalized EFT are equal up to the
accuracy of performed calculations.
Non-perturbative renormalization, where part of divergences are absorbed into
two contact interaction coupling constants with subsequent removal of
regularization is also performed. Cutoff and dimensional regularizations
both lead to finite but different results within this scheme.
\end{abstract}
% insert suggested PACS numbers in braces on next line
{\it PACS numbers}: 03.65.Nk, 11.10.Gh, 12.39.Fe, 13.75.Cs.
%\pacs{
\medskip
\medskip
\medskip
There has been significant development in effective field theory (EFT)
approach to problems of multi nucleon systems originated by Weinberg's
papers \cite{Weinberg:1990rz}, \cite{Weinberg:1991um}.
(for recent developments see refs. \cite{Beane:2000fx},\cite{Bedaque:2000kn}).
In Weinberg's approach power counting is formulated for effective
potential and the scattering amplitude is calculated by substituting this
effective
potential into Lippman-Schwinger (or Schr\" odinger) equation.
In ref. \cite{Kaplan:1996xu} it has been argued that Weinberg's approach
suffers
from formal inconsistences. While this conclusion has been questioned
\cite{Park:1999cu}, \cite{Gegelia:1998ee} a new explicitly consistent power
counting (KSW power counting) has been suggested in ref. \cite{Kaplan:1998tg}.
In this approach pions are included perturbatively and it allows to perform
all calculations analytically which can be considered as a considerable
advantage. Unfortunately it has been realized that perturbation theory series
do not converge in this approach \cite{Gegelia:1998ee},\cite{Cohen:1999jr},
\cite{Fleming:2000ee}. An attempt to combine
the advantages of Weinberg's and KSW schemes has been made recently in
ref. \cite{Beane:2001bc}. In that paper it has been confirmed that Weinberg's
power counting is formally inconsistent in the ${}^1S_0$ channel NN
scattering problem. In particular, divergences which appear in leading order,
can not be absorbed by leading order counter-terms. Different approach to the
problem of non-perturbative renormalization of Lippmann-Schwinger equation
has been suggested recently in ref. \cite{Eiras:2001hu}.
In present letter, which is the continuation of the work started in refs.
\cite{Gegelia:1999iu},\cite{Gegelia:1999ja}, we follow Weinberg's
approach and iterate sub-leading order ${ }^1S_0$ NN potential in the
framework of Lippman-Schwinger equation. We note that the above mentioned
inconsistency of Weinberg's power counting manifests itself only when it is
applied to
unrenormalized diagrams and counter-terms separately. We advocate the point of
view that the power counting should be applied to renormalized diagrams
\cite{Park:1999cu},\cite{Gegelia:1998ee},\cite{Gegelia:1998gn},
\cite{Lepage:1999kt}.
Let us emphasize that we do not attempt to resolve the above mentioned formal
inconsistency of Weinberg's power counting, we believe that it can not be
resolved within this scheme. We share the point of view that this problem
is irrelevant: if one applies the power counting to renormalized diagrams then
there are no formal inconsistencies, renormalized diagrams are of the order
which has been assigned by power counting and the approach is successful in
studies of processes in few nucleon systems
\cite{vanKolck:1999mw},\cite{Epelbaum:2000dj},\cite{Epelbaum:2001mx},
\cite{Epelbaum:2001fm}.
This scheme can be implemented using BPHZ renormalization procedure.
%Within this renormalization scheme power counting can be consistently applied
%to the Lagrangian of the EFT (to decide which terms should be kept within
%given order of calculations) as well as to Feynman diagrams.
In BPHZ approach no counter-terms are included into the Lagrangian but
instead the Feynman rules are accompanied by prescriptions for subtracting
loop diagrams \cite{Collins:1984xc}. The subtracted diagrams depend on
renormalization point(s). This dependence is cancelled by the renormalization
point dependence of renormalised coupling constants so that the physical
quantities are independent of the choice of subtraction point(s). The above
mentioned problem of inconsistency that higher order counter terms contribute
into the renormalization of low order diagrams translates now into the
following problem: low order diagrams give large contributions into beta
functions of coupling constants of higher order (counter) terms.
There have been concerns that even if for some value of the renormalization
point(s) the higher order coupling constants are tuned to be small, the slight
change of this renormalization point(s) would result in large values for
coupling constants and therefore power counting would be badly violated.
Note that these concerns are
based on perturbative estimations of beta-functions in the area where
perturbation theory is not actually applicable. Hence there is no reason to
expact that the assumption of naturalness which is an ingredient of Weinberg's
power counting is violated. To make any reliable
conclusions about renormalization group behaviour of renormalized coupling
constants non-perturbative analysis of this behaviour are required. We are not
aware of any such analysis of
renormalized coupling constants in EFT with explicitly included pions. On the
other hand the considerable success of cutoff EFT, which is based on
Weinberg's approach
\cite{vanKolck:1999mw,Epelbaum:2000dj,Epelbaum:2001mx,Epelbaum:2001fm}
suggests that renormalized coupling constants should be well-behaved.
Below we use Weinberg's power counting with BPHZ
renormalization procedure and compare the results of subtractively
renormalized and cutoff EFT for ${ }^1S_0$
NN scattering amplitude. We find that the results of two schemes coincide
up to the order of the accuracy of given calculations. We also explore the
possibility of
non-perturbative renormalization via absorbing part of divergences into the
two available ``bare'' coupling constants with successive removal of
regularization.
In this letter we quote the results, details of calculations will be
presented elsewhere.
We iterate subleading order potential substituting it into Lippman-Schwinger
equation. To perform subleading order analysis we include two-derivative
contact
interaction part of the potential for ${ }^1S_0$ $NN$-scattering as it is
enhanced in comparison with the two-pion exchange part \cite{Kaplan:1996xu}
(note that to renormalize the amplitude obtained by iterating leading order
potential no contact terms with
derivatives are required \cite{Kaplan:1996xu}, \cite{Beane:2001bc})
and therefore the sub-leading order potential has the following form:
\be
V(\mbox{\boldmath $p'$},\mbox{\boldmath $p$} )=C+C_2\left(
\mbox{\boldmath $p'$}^2+\mbox{\boldmath $p$}^2\right)+
V_\pi \left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right)
\label{potential}
\ee
where $C$ and $C_{2}$ are (``bare'') coupling constants, and
\be
V_\pi \left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right) =
-\frac{4\pi\alpha_\pi }
{\left( \mbox{\boldmath $p'$}-\mbox{\boldmath $p$}\right)^2+m_\pi^2},
\ \ \ \
\alpha_\pi =\frac{g_A^2m_\pi^2}{8\pi f_\pi^2}.
\label{vpi}
\ee
We need to extract the $S$-wave part from the solution to the
Lippman-Schwinger equation for the $NN$ scattering amplitude $T$:
\be
T \left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right)=
V \left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right)+
\frac {M}{(2\pi)^3}\int d^3\mbox{\boldmath $q$} \
V \left( \mbox{\boldmath $p'$},\mbox{\boldmath $q$} \right) \
G(\mbox{\boldmath $q$}) \
T \left( \mbox{\boldmath $q$},\mbox{\boldmath $p$}
\right)\equiv V \left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right)+
V \left( \mbox{\boldmath $p'$},\mbox{\boldmath $q$} \right)
\otimes G(\mbox{\boldmath $q$})\otimes
T \left( \mbox{\boldmath $q$},\mbox{\boldmath $p$}
\right),
\label{lschweq}
\ee
where
$$
G(\mbox{\boldmath $q$})=\frac{1}{\mbox{\boldmath $k$}^2-
\mbox{\boldmath $q$}^2+i\epsilon },
$$
integration over "$\mbox{\boldmath $q$}$ - variable" $\frac {M}{(2\pi)^3}\int
d^3\mbox{\boldmath $q$}$ is
denoted by $\otimes$, $M$ is the nucleon mass and
$\mbox{\boldmath $p$}$ is on-mass-shell value of three-momenta. For physical
amplitude $\mbox{\boldmath
$p'$}=\mbox{\boldmath $p$}$.
When substituted into Lippman-Schwinger equation (\ref{lschweq}), potential
$V \left( \mbox{\boldmath $p'$},\mbox{\boldmath$p$} \right)$ generates
divergences.
To regularize potential we
introduce factorized form-factor into the contact interaction part of
$V \left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right)$
and write:
\be
V_{R}(\mbox{\boldmath $p'$},\mbox{\boldmath $p$} )=\left[ C+C_2\left(
\mbox{\boldmath $p'$}^2+\mbox{\boldmath $p$}^2\right)\right]
\chi (\Lambda ,\mbox{\boldmath $p'$})\chi (\Lambda ,\mbox{\boldmath $p$}) +
V_\pi \left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right),
\label{potentialreg}
\ee
where $\chi $ is a regularization function.
For definiteness let us use
\be
\chi (\Lambda ,\mbox{\boldmath $p$})=\frac {\Lambda^4}{\left( \Lambda^2+
\mbox{\boldmath $p$}^2\right)^2}.
\label{ff}
\ee
Using $V_{R}(\mbox{\boldmath $p'$},\mbox{\boldmath $p$} )$ instead of
$V(\mbox{\boldmath $p'$},\mbox{\boldmath $p$} )$, we
rewrite the Lippman-Schwinger equation (\ref{lschweq}) into the matrix form:
\begin{eqnarray}
\nonumber
\left(
\begin{array}{cc}
T\left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$}\right), & 0 \\
0, & 0
\end{array}
\right)=\chi \left( \Lambda ,\mbox{\boldmath $p'$} \right)
\left(
\begin{array}{cc}
1, & p'^2 \\
0, & 0
\end{array}
\right) \
\left(
\begin{array}{cc}
C, & C_2 \\
C_2, & 0
\end{array}
\right) \
\left(
\begin{array}{cc}
1, & 0 \\
p^2, & 0
\end{array}
\right)\chi\left( \Lambda ,\mbox{\boldmath $p$}\right)+
\left(
\begin{array}{cc}
V_\pi\left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right), & 0 \\
0, & 0
\end{array}
\right)+ \\
\nonumber
\frac{M}{(2\pi )^3}\int d^3\mbox{\boldmath $q$}
\left\{ \chi \left( \Lambda ,\mbox{\boldmath $p'$} \right)
\left(
\begin{array}{cc}
1, & p'^2 \\
0, & 0
\end{array}
\right) \
\left(
\begin{array}{cc}
C, & C_2 \\
C_2, & 0
\end{array}
\right) \
\left(
\begin{array}{cc}
1, & 0 \\
q^2, & 0
\end{array}
\right)\chi\left( \Lambda ,\mbox{\boldmath $q$}\right)+
\left(
\begin{array}{cc}
V_\pi\left( \mbox{\boldmath $p'$},\mbox{\boldmath $q$} \right), & 0 \\
0, & 0
\end{array}
\right)\right\} \times \\
\left(
\begin{array}{cc}
G(\mbox{\boldmath $q$}), & 0 \\
0, & 0
\end{array}
\right) \
\left(
\begin{array}{cc}
T\left( \mbox{\boldmath $q$},\mbox{\boldmath $p$}\right), & 0 \\
0, & 0
\end{array}
\right)
\label{mlschweq}
\end{eqnarray}
\noindent
We solve (\ref{mlschweq}) for the ${ }^1S_0$ amplitude and obtain:
\be
T_0(p)=T_\pi+\frac{\left[ C+2p^2C_2 +C_2^2G_3(p)
\right] V_1(p)^2-2C_2V_1(p)V_2(p)\left[ 1+C_2G_2(p)\right]+
C_2^2G_1(p)V_2^2(p)}{\left[ 1+C_2G_2(p)\right]^2-\left[ C+2p^2C_2 +C_2^2G_3(p)
\right] G_1(p)},
\label{invt}
\ee
where $T_\pi$ is ${ }^1S_0$ wave solution of Lippman- Schwinger equation (\ref{lschweq}) for
$V \left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right)=
V_\pi \left( \mbox{\boldmath $p'$},\mbox{\boldmath $p$} \right)$. Functions
$G_i(p)$ are given by infinite series as follows:
\begin{eqnarray}
\nonumber
G_1(p)=\chi^2\left( \Lambda ,\mbox{\boldmath $q$}\right)\otimes G\left( \mbox{\boldmath $q$}\right)+
\chi \left( \Lambda ,\mbox{\boldmath $q$}_1\right)\otimes G\left( \mbox{\boldmath $q$}_1\right)\otimes
V_\pi\left( \mbox{\boldmath $q$}_1,\mbox{\boldmath $q$}_2 \right)\otimes G\left( \mbox{\boldmath $q$}_2\right)\otimes
\chi \left( \Lambda ,\mbox{\boldmath $q$}_2\right)+ \\
\chi \left( \Lambda ,\mbox{\boldmath $q$}_1\right)\otimes G\left( \mbox{\boldmath $q$}_1\right)\otimes
V_\pi\left( \mbox{\boldmath $q$}_1,\mbox{\boldmath $q$}_2 \right) \otimes G\left( \mbox{\boldmath $q$}_2\right)\otimes
V_\pi\left( \mbox{\boldmath $q$}_2,\mbox{\boldmath $q$}_3 \right)\otimes G\left( \mbox{\boldmath $q$}_3\right)\otimes
\chi \left( \Lambda ,\mbox{\boldmath $q$}_3\right)+ \cdots
\label{g1exp}
\end{eqnarray}
\begin{eqnarray}
\nonumber
G_2(p)= \chi^2\left( \Lambda ,\mbox{\boldmath $q$}\right)\otimes 1+
\chi \left( \Lambda ,\mbox{\boldmath $q$}_1\right)\otimes V_\pi\left( \mbox{\boldmath $q$}_1,\mbox{\boldmath $q$}_2 \right)
\otimes G\left( \mbox{\boldmath $q$}_2\right)\otimes
\chi \left( \Lambda ,\mbox{\boldmath $q$}_2\right)+ \\
\chi \left( \Lambda ,\mbox{\boldmath $q$}_1\right)\otimes
V_\pi\left( \mbox{\boldmath $q$}_1,
\mbox{\boldmath $q$}_2 \right)\otimes
G\left( \mbox{\boldmath $q$}_2\right)V_\pi\left( \mbox{\boldmath $q$}_2,
\mbox{\boldmath $q$}_3 \right)\otimes
G\left( \mbox{\boldmath $q$}_3\right)\otimes
\chi \left( \Lambda ,\mbox{\boldmath $q$}_3\right)+ \cdots
\label{g2exp}
\end{eqnarray}
\begin{eqnarray}
\nonumber
G_3(p)=
\chi^2\left( \Lambda ,\mbox{\boldmath $q$}\right)\otimes
G^{-1}\left( \mbox{\boldmath $q$}\right)+
\chi \left( \Lambda ,\mbox{\boldmath $q$}_1\right)\otimes
V_\pi\left( \mbox{\boldmath $q$}_1,
\mbox{\boldmath $q$}_2 \right)\otimes
\chi \left( \Lambda ,\mbox{\boldmath $q$}_2\right)+ \\
\chi \left( \Lambda ,\mbox{\boldmath $q$}_1\right)\otimes
V_\pi\left( \mbox{\boldmath $q$}_1,
\mbox{\boldmath $q$}_2 \right) \otimes
G\left( \mbox{\boldmath $q$}_2\right)\otimes
V_\pi\left( \mbox{\boldmath $q$}_2,
\mbox{\boldmath $q$}_3 \right)\otimes
\chi \left( \Lambda ,\mbox{\boldmath $q$}_3\right)+ \cdots
\label{g3exp}
\end{eqnarray}
\noindent
and the functions $V_1(\mbox{\boldmath $p$})$, $V_2(\mbox{\boldmath $p$})$
satisfy integral equations:
\be
V_1(\mbox{\boldmath $p$})=\chi \left( \Lambda ,\mbox{\boldmath $p$}\right)+
V_1\left( \mbox{\boldmath $q$}\right)\otimes
G\left( \mbox{\boldmath $q$}\right)\otimes V_\pi\left( \mbox{\boldmath $q$},
\mbox{\boldmath $p$} \right),
\label{v1eq}
\ee
%$$
%V_1(p)=\chi \left( \Lambda ,\mbox{\boldmath $p$}\right)+
%\frac{M}{(2\pi)^3}\int d^3\mbox{\boldmath $q$} \
%\chi\left( \Lambda ,\mbox{\boldmath $q$}\right)
%G\left( \mbox{\boldmath $q$}\right)V_\pi\left( \mbox{\boldmath $q$},
%\mbox{\boldmath $p$} \right) +
%$$
%\be
%\left[ \frac{M}{(2\pi)^3}\right]^2 \int d^3\mbox{\boldmath $q$}_1
%d^3\mbox{\boldmath $q$}_2 \
%\chi \left( \Lambda ,\mbox{\boldmath $q$}_1\right)
%G\left( \mbox{\boldmath $q$}_1\right)V_\pi\left( \mbox{\boldmath $q$}_1,
%\mbox{\boldmath $q$}_2 \right)
%G\left( \mbox{\boldmath $q$}_2\right)
%\chi \left( \Lambda ,\mbox{\boldmath $q$}_2\right)
%V_\pi\left( \mbox{\boldmath $q$}_2,
%\mbox{\boldmath $p$} \right)+ \cdots
%\label{v1exp}
%\ee
\be
V_2(\mbox{\boldmath $p$})=
\chi \left( \Lambda ,\mbox{\boldmath $q$}\right)\otimes
V_\pi\left( \mbox{\boldmath $q$},
\mbox{\boldmath $p$} \right)+
V_2 \left( \mbox{\boldmath $q$}\right)\otimes
G\left( \mbox{\boldmath $q$}\right)\otimes
V_\pi\left( \mbox{\boldmath $q$},
\mbox{\boldmath $p$} \right).
\label{v2eq}
\ee
%$$
%V_2(p)=\frac{M}{(2\pi)^3}\int d^3\mbox{\boldmath $q$} \
%\chi \left( \Lambda ,\mbox{\boldmath $q$}\right)
%V_\pi\left( \mbox{\boldmath $q$},
%\mbox{\boldmath $p$} \right)+
%\left[ \frac{M}{(2\pi)^3}\right]^2 \int d^3\mbox{\boldmath $q$}_1
%d^3\mbox{\boldmath $q$}_2 \
%\chi \left( \Lambda ,\mbox{\boldmath $q$}_1\right)
%V_\pi\left( \mbox{\boldmath $q$}_1,
%\mbox{\boldmath $q$}_2 \right)
%G\left( \mbox{\boldmath $q$}_2\right)
%V_\pi\left( \mbox{\boldmath $q$}_2,
%\mbox{\boldmath $p$} \right)+
%$$
%\be
%\left[ \frac{M}{(2\pi)^3}\right]^3 \int d^3\mbox{\boldmath $q$}_1
%d^3\mbox{\boldmath $q$}_2 d^3\mbox{\boldmath $q$}_3\
%\chi \left( \Lambda ,\mbox{\boldmath $q$}_1\right)
%V_\pi\left( \mbox{\boldmath $q$}_1,
%\mbox{\boldmath $q$}_2 \right)
%G\left( \mbox{\boldmath $q$}_2\right)V_\pi\left( \mbox{\boldmath $q$}_2,
%\mbox{\boldmath $q$}_3 \right)
%G\left( \mbox{\boldmath $q$}_3\right)
%V_\pi\left( \mbox{\boldmath $q$}_3,
%\mbox{\boldmath $p$} \right)+ \cdots
%\label{v2exp}
%\ee
Expression (\ref{invt}) for the amplitude remains valid in the framework of
dimensional regularization
provided that integrals are regulated using this regularization and
$\chi $-function is substituted by $1$.
To give physical meaning to Eq. (7) we need to get rid off divergences, which
show up in $\Lambda\to\infty $ limit.
Note that the amplitude can be made finite using
non-perturbative renormalization, i.e. renormalizing two available ``bare''
parameters $C$ and $C_2$ and taking $\Lambda\to\infty $ limit in
non-perturbative expression of $T_0(p)$.
Let us describe this procedure .
We fix $C$ and $C_2$ from two conditions of finiteness of
the amplitude at $p^2=0$ and $p^2=-\mu^2$ ($p=i\mu$). As $T_\pi (p)$ does not
contain divergences, we require the finiteness of
$a(p)\equiv 1/\left[ T_0(p)-T_\pi (p)\right] $:
\be
a(0)=\frac{\left[ 1+C_2G_2(0)\right]^2-\left[ C+C_2^2G_3(0)
\right] G_1(0)}{\left[ C+C_2^2G_3(0)
\right] V_1(0)^2-2C_2V_1(0)V_2(0)\left[ 1+C_2G_2(0)\right]
+C_2^2G_1(0)V_2^2(0)},
\label{rencon1}
\ee
\be
a(\mu )=
\frac{\left[ 1+C_2G_2(i\mu )\right]^2-\left[ C-2\mu^2C_2
+C_2^2G_3(i\mu )
\right] G_1(i\mu )}{\left[ C-2\mu^2C_2 +C_2^2G_3(i\mu )
\right] V_1(i\mu )^2-2C_2V_1(i\mu )V_2(i\mu )\left[ 1+C_2G_2(i\mu )\right]
+C_2^2G_1(i\mu )V_2^2(i\mu )}
\label{rencon2}
\ee
In (\ref{rencon1}) - (\ref{rencon2}) $a(0)$ and $a(\mu^2 )$ are
finite.
Analyzing expressions (\ref{g1exp})-(\ref{v2eq}) it is straightforward to
obtain following relations:
\begin{eqnarray}
\label{g1}
G_1(p)&=&a_0 \Lambda +a_1 \ln \Lambda +a_2(p)+0\left( 1/\Lambda \right) \\
\label{g2}
G_2(p)&=&b_0 \Lambda^3 +b_1 \Lambda^2 +b_2(p)\Lambda +b_3(p) \ln \Lambda
+b_4(p) +0\left( 1/\Lambda \right) \\
\label{g3}
G_3(p)&=&c_0 \Lambda^5 +c_1 \Lambda^4 +c_2(p)\Lambda^3+c_3(p)\Lambda^2
+c_4(p)\ln \Lambda +c_5(p) +0\left( 1/\Lambda \right) \\
\label{v1}
V_1(p)&=&V_1(p)+0\left( 1/\Lambda \right) \\
\label{v2}
V_2(p)&=&v \Lambda +d(p)+0\left( 1/\Lambda \right),
\end{eqnarray}
where $a_0, a_1, b_0, b_1, c_l, c_1, v$ are constants independent of
$\Lambda $ and
$a_2(p), b_2(p), b_3(p), c_2(p)-c_5(p), V_{1}(p), d(p)$ are some functions of
$p$ also
independent of $\Lambda $.
Substituting expansion (\ref{g1})-(\ref{v2}) into expressions for $a(0)$ and
$a(\mu)$ we solve the system of two equations for $C$ and $C_2$.
%and
%solving the system of two equations for $C$ and $C_2$ we observe that
%for $\Lambda\to\infty $ the emerging quadratic equation for $C_2$ has
%solutions only if ({\bf where do we use further requirement
%(20)? NOWHERE, tu ginda amovagdot! What the hell is $e_{0}$ in (20)? IT IS
%$a_0$})
%\be
%D= \Lambda^5 \left\{ a_0^2\left[ c_2(i\mu )-c_2(0)+2b_0\mu^2\right]
%\left[ a_2(0)-a_2(i\mu )+a(0) V_1^2(0)-a(i\mu ) V_1^2(i\mu )\right]\right\}
%\geq 0.
%\label{discr}
%\ee
Substituting these solutions for $C$ and $C_2$ into expression for the
amplitude
(\ref{invt}) and taking into account expansions (\ref{g1})-(\ref{v2}) we
obtain
in the limit $\Lambda\to\infty $ amplitude which is finite:
\be
T_0(p)=T_\pi (p)+\frac{1}{\alpha(p)},
\label{nprampl}
\ee
where
\begin{eqnarray}
\nonumber
& & \alpha (p)=\biggl\{ \left( a(\mu )V_1^2(\mu )+a_2(\mu ) \right)
\left( c_2(p)-c_2(0)-4b_0p^2 \right)-
\left( a(0)V_1^2(0)+a_2(0) \right)\ \times
\\
& &\left( c_2(p)-c_2(\mu)
-4b_0p^2-2 b_0\mu^2\right)+
a_2(p)\left( c_2(0)-c_2(\mu )-2b_0\mu^2\right) \biggr\}
/ \left[ \left( c_2(\mu )-c_2(0)+2b_0\mu^2\right)V_1^2(p)\right]
\label{treninv}
\end{eqnarray}
When using the same renormalization condition and dimensional regularization
we
obtain finite result which differs from (\ref{treninv}) in analogy with
the case of pionless
EFT \cite{Beane:1998pk}. This difference between two results is due to the
fact that dimensional
regularization discards all power-law divergences, i.e. dimensional
regularization is in fact not only regularization in standard sense but it
also subtracts power-law divergences.
As it was argued in \cite{Gegelia:1998xr}, the above described renormalization
scheme is
not consistent with EFT. To carry out the consistent renormalization procedure
it is necessary to absorb {\it all} divergences into coupling
constants or in other words, subtract all divergent integrals. To do so
let us start from the
observation that $G_{2}$, $G_{3}$ and $V_{2}$ can be expressed as follows:
\begin{eqnarray}
\nonumber
G_2(p)&=&\Delta_2 +\Delta_\pi G_1(p), \\
\nonumber
G_3(p)&=&\Delta_3 +\Delta_\pi G_2(p), \\
V_2(p)&=&\Delta_\pi \ V_1(p),
\label{grelations}
\end{eqnarray}
where
\begin{equation}
\Delta_2=
\chi^2\left( \Lambda ,\mbox{\boldmath $q$}\right)\otimes 1, \ \
\Delta_3=
\chi^2\left( \Lambda ,\mbox{\boldmath $q$}\right)\otimes
G^{-1}\left( \mbox{\boldmath $q$}\right), \ \
\Delta_\pi =
\chi \left( \Lambda ,\mbox{\boldmath $q$}\right)\otimes
V_\pi\left( \mbox{\boldmath $q$},
\mbox{\boldmath $p$} \right).
\label{deltadefs}
\ee
Relations (\ref{grelations}) are correct up to terms which are
vanishing after renormalization is performed and regularization is removed ,
i.e. after sub-divergences and overall divergences are substracted. Note that
these relations
become exact in dimensional regularization provided that in
(\ref{deltadefs}) $\chi $-factors are substituted by 1.
Using relations (\ref{grelations}) we express amplitude (\ref{invt}) as
\be
T_0(p)=T_\pi (p)+\frac{V_1^2(p)\ \left[ C -\Delta_\pi \ C_2 \left( 2
+\Delta_2 C_2\right) + 2\ p^2\ C_2 +\Delta_3 C_2^2
\right]}
{1 +2\Delta_2 C_2+ \Delta_2^2 C_2^2 -G_1(p)\left[ C -\Delta_\pi \ C_2 \left( 2
+\Delta_2 C_2\right) + 2\ p^2\ C_2 +\Delta_3 C_2^2 \right]}
\label{tsimplified}
\ee
To renormalize amplitude (\ref{tsimplified}) following standard BPHZ
renormalization scheme we subtract divergent parts from
$G_1(p)$, $\Delta_2$, $\Delta_3$ and $\Delta_\pi$ and substitute $C$ and
$C_2$ by their renormalized values $C_R$ and $C_{2R}$
%({\bf ? Which values, in
%what scheme are we now,
%where this renormalised constants came from?})
and obtain:
\be
T_0^R(p)=T_\pi (p)+\frac{V_1^2(p)\ \left[ C_R -\Delta_\pi^R \ C_{2R} \left( 2
+F_2 C_{2R}\right) + 2\ p^2\ C_{2R} +\left( F_{31}+F_{32} \ p^2\right) C_{2R}^2
\right]}
{1 +2F_2 C_{2R}+ F_2^2 C_{2R}^2 -G_1^R(p) \left[ C_R -\Delta_\pi^R \ C_{2R}
\left( 2 +F_2 C_{2R}\right) + 2\ p^2\ C_{2R} +\left( F_{31}+F_{32}
\ p^2\right)
C_{2R}^2 \right]}
\label{trenormalized}
\ee
Here $F_2$, $F_{31}$ and $F_{32}$ are (arbitrary) finite parts. Note that
arbitrary finite parts are also present in $\Delta_\pi^R$ and $G_1^R(p)$, the
particular choice for the finite parts corresponds to the particular choice of
renormalization
condition.
Expression (\ref{trenormalized}) agrees with the one for the renormalized
amplitude, obtained in the framework of dimensional regularization
\cite{Kaplan:1996xu}.
Amplitude obtained in \cite{Kaplan:1996xu} can be reproduced from
Eq. (\ref{trenormalized}) by
appropriate choice of finite parts of $G_1(p)^R$ and $\Delta_\pi^R$, and
by the conditions $F_2=F_{31}=F_{32}=0$.
Hence, as it was expected, dimensional and cutoff regularizations lead to the
same
results provided that the same renormalization condition is used.
\medskip
\medskip
Let us now turn to the cut-off EFT.
In this approach we absorb part of divergences into the
``bare'' coupling constants $C$ and $C_2$ but
instead of considering $\Lambda\rightarrow \infty $ limit, we
choose the value of cutoff parameter $\Lambda $ in such a way that the
amplitude of cut-off theory reproduces the above calculated subtractively
renormalized amplitude up to sub-sub-leading order.
%$$
%\frac{1 - G_1(p)\ C}{V_1^2(p)\ C}+ \frac{ 2\left( -p^2\ V_1(p)+ \ V_2(p)+
%G_2(p)\ V_1(p)\ C - G_1(p)\ V_2(p)\ C\right)\ C_2}{V_1^3(p)\ C^2}
%$$
We determine $C$ and $C_2$ from two conditions, the first of which is to
absorbs all terms in Eq. (\ref{tsimplified}) for $p=0$, which are divergent
if the $\Lambda\rightarrow \infty$ limit is taken. This condition can be
written as:
\begin{eqnarray}
\nonumber
\frac
{1 +2\Delta_2 C_2+ \Delta_2^2 C_2^2 -G_1(p)\left[ C -\Delta_\pi \ C_2 \left( 2
+\Delta_2 C_2\right) + 2\ p^2\ C_2 +\Delta_3 C_2^2 \right]}
{V_1^2(0)\ \left[ C -\Delta_\pi \ C_2 \left( 2
+\Delta_2 C_2\right) + 2\ p^2\ C_2 +\Delta_3 C_2^2
\right]}= \\
\frac
{1 +2\Delta_{2R} C_{2R}+ \Delta_{2R}^2 C_{2R}^2 -G_1^R(p)\left[ C_R
-\Delta_{\pi R} \ C_{2R} \left( 2
+\Delta_{2R} C_{2R}\right) +\Delta_{3R} C_{2R}^2 \right]}
{V_1^2(0)\ \left[ C_R -\Delta_{\pi R} \ C_{2R} \left( 2
+\Delta_{2R} C_{2R}\right) +\Delta_{3R} C_{2R}^2
\right]}
\label{rencond}
\end{eqnarray}
\noindent
Another condition can be obtained by expanding Eq. (\ref{tsimplified}) in
powers
of $C_2$ and requiring the finiteness of the first order
term in $\Lambda\rightarrow \infty $ limit. Taking into account
Eq. (\ref{rencond}), this requirement
can be expressed as
\begin{equation}
%\nonumber
\frac{C_2}{C^2}= \frac{C_{2R}}{C_R^2}
\label{c2}
\end{equation}
Substituting the solutions to Eqs. (\ref{rencond})-(\ref{c2}) into the
(unrenormalized) solution of Lippman-Schwinger equation given by
Eq. (\ref{invt})
and expanding in powers of $C_{2R}$ we obtain:
\begin{eqnarray}
\nonumber
T_0(p)=T_\pi (p)+\frac{V_1^2(p)\ C_R}{1
+ C_R\ \left( \Delta_1 - G_1(p)\ \right)}
+ \frac{2\ V_1(p)\ \left( p^2\ V_1(p)-\Delta_\pi^R \ V_1(p)
-C_R \ \Delta_2^R\ V_1(p) \right) C_{2R}}
{\left( 1 + C_R \left[ \Delta_1 - G_1(p)\right] \right)^2} \\
+ \frac{2\ V_1(p)\ \left( \Delta_\pi V_1(p) -V_2(p) \right) C_{2R}}
{\left( 1 + C_R \left[ \Delta_1 - G_1(p)\right] \right)^2}
+ \frac{2\ V_1(p)C_{2R} C_R \left\{ \Delta_2 V_1(p)
-G_2(p) V_1(p)+G_1(p) V_2(p)
\right\}}{V_1(0)\ \left( 1+\Delta_1\
C_R \right)\ \left( 1
+ C_R \left( \Delta_1 - G_1(p)\right)\ C_R\right)^2 }+\cdots ,
\label{cutofft}
\end{eqnarray}
where $\Delta_1$ contains divergent part of $G_1$ as well as finite part,
the latter depending on the choice of renormalization scheme.
Note that fourth and fifth terms of Eq. (\ref{cutofft}) are
suppressed by an additional factor of $1/\Lambda$ in comparison with
third term. Higher order terms in $C_{2R}$ which are
indicated by dots are suppressed by higher orders of small parameter
(three-momentum $p$, pion mass or the renormalization scale which we choose of
the order of $p$). As our potential includes only leading and subleading order
terms,
calculated amplitude can be trusted up to this order, provided that formally
higher order contributions are indeed suppressed. Note that these higher
order terms also contain factors like $p^4 \Lambda^2 C_{2R}^2$.
To suppress this kind of contributions as well as terms with
inverse powers
of $\Lambda $, we need to choose cutoff $\Lambda $ of the
order of
large scale parameter $Q$ which governs the behavior of $C_R$, $C_{2R}$
\cite{Park:1999cu}, \cite{Gegelia:1999iu}. With this choice of cutoff parameter
one can easily see that two amplitudes given by Eq. (\ref{trenormalized}) and
Eq. (\ref{cutofft}) coincide up to
the order of accuracy of our calculations the difference being of higher
order.
\medskip
\medskip
\medskip
In this work we have iterated cutoff regulated ${ }^1S_0$
partial wave NN
potential of subleading order EFT and wrote a closed expression for the
scattering amplitude. We have
performed non-perturbative renormalization via absorbing part of
divergences into two available coupling constants and taking the removed
regularization limit. Within this renormalization scheme cutoff and
dimensional
regularizations both lead to finite results but these results differ from each
other. This difference between two results is due to the
fact that dimensional regularization discards all power-law divergences. We
argue that this kind of renormalization is not consistent with EFT approach.
We advocate the point of view that Weinberg's power counting should be
applied to renormalized diagrams
\cite{Park:1999cu},\cite{Gegelia:1998ee},\cite{Gegelia:1998gn},
\cite{Lepage:1999kt}, i.e. one should draw all (an infinite
number of them for NN-scattering) relevant diagrams, renormalize and sum them
up. This summation is performed by solving Lippman-Schwinger equation (or
Schr\" odinger equation). In fact we have solved the regularized equation and
performed renormalization by subtracting divergences in the solution. For the
case considered in this work it is not difficult to demonstrate that expanding
the renormalized solution to the Lippmann-Schwinger equation we reproduce
separate renormalized Feynman diagrams.
%, i.e. solving the equation and
%renormalizing the solution by subtracting divergencies corresponds to
%first renormalizing diagrams and after summing them up.
Of course in general the solutions to
Lippmann-Schwinger equation (or Schr\" odinger equation) can contain more
contributions than just the sum of (an infinite number of) perturbative
diagrams, but note that there is no power counting for such
non-perturbative contributions.
We have implemented the renormalization procedure
using BPHZ procedure. Within this scheme no counter-terms are included in the
Lagrangian and therefore it is free of formal inconsistences that to remove
divergences from diagrams of given order calculations one needs to include
contributions of counter-terms which are of higher order. Let us emphasize
that we have not attempted to solve this problem of formal inconsistences. We
argue that this issue is irrelevant and it can be avoided by applying power
counting to renormalized diagrams. The usage of BPHZ renormalization proves to
be very useful in dealing with this problem.
The main result of our work reads as follows.
While the power counting is consistently formulated in subtractively
renormalised EFT, in most cases we are not able to solve regularized equations
and therefore it
is technically very difficult (or rather impossible) to implement subtractive
renormalization scheme. In practice most of practically important calculations
are performed using cutoff EFT. For the potential considered in this work
we have compared the amplitude of cutoff EFT with
subtractively renormalized amplitude. The cutoff amplitude is equal
to subtractively renormalized amplitude up to (including) sub-leading order,
i.e. the difference between these two is of higher order than the order of
the accuracy of our calculations (note that subtractively
renormalized subleading order calculations reproduce the phase shift analyses
for ${ }^1S_0$ NN scattering reasonably well \cite{Gegelia:1999ja}).
We consider the present work as a demonstration of a general feature of
cutoff EFT that it
reproduces the results of subtractively renormalized theory up to the order of
the accuracy of given calculations. This suggests that the
considerable success of Weinberg's approach implemented in cutoff EFT
\cite{vanKolck:1999mw},\cite{Epelbaum:2000dj},\cite{Epelbaum:2001mx},
\cite{Epelbaum:2001fm} is by far not accidental.
%{\bf ACKNOWLEDGEMENTS}
%\newpage
%\begin{thebibliography}{99}
\begin{references}
%\cite{Weinberg:1990rz}
\bibitem{Weinberg:1990rz}
S.~Weinberg,
%``Nuclear Forces From Chiral Lagrangians,''
Phys.\ Lett.\ B {\bf 251}, 288 (1990).
%%CITATION = PHLTA,B251,288;%%
%\cite{Weinberg:1991um}
\bibitem{Weinberg:1991um}
S.~Weinberg,
%``Effective chiral Lagrangians for nucleon - pion interactions and
%nuclear forces,''
Nucl.\ Phys.\ B {\bf 363}, 3 (1991).
%%CITATION = NUPHA,B363,3;%%
%\cite{Beane:2000fx}
\bibitem{Beane:2000fx}
S.~R.~Beane, P.~F.~Bedaque, W.~C.~Haxton, D.~R.~Phillips and M.~J.~Savage,
%``From hadrons to nuclei: Crossing the border,''
nucl-th/0008064.
%%CITATION = NUCL-TH 0008064;%%
%\cite{Bedaque:2000kn}
\bibitem{Bedaque:2000kn}
P.~F.~Bedaque, M.~J.~Savage, R.~.~Seki and U.~van Kolck,
``Nuclear physics with effective field theory. II. Proceedings, INT Workshop,
Seattle, USA, February 25-26, 1999,''
{\it Singapore, Singapore: World Scientific (2000) 368 p}.
%\cite{Kaplan:1996xu}
\bibitem{Kaplan:1996xu}
D.~B.~Kaplan, M.~J.~Savage and M.~B.~Wise,
%``Nucleon-Nucleon Scattering from Effective Field Theory,''
Nucl.\ Phys.\ B {\bf 478}, 629 (1996)
[nucl-th/9605002].
%%CITATION = NUCL-TH 9605002;%%
%\cite{Park:1999cu}
\bibitem{Park:1999cu}
T.~Park, K.~Kubodera, D.~Min and M.~Rho,
%``The power of effective field theories in nuclei: The deuteron, N N
%scattering and electroweak processes,''
Nucl.\ Phys.\ A {\bf 646}, 83 (1999)
[nucl-th/9807054].
%%CITATION = NUCL-TH 9807054;%%
%\cite{Gegelia:1998ee}
\bibitem{Gegelia:1998ee}
J.~Gegelia,
%``Are pions perturbative in effective field theory?,''
nucl-th/9806028.
%%CITATION = NUCL-TH 9806028;%%
%\cite{Kaplan:1998tg}
\bibitem{Kaplan:1998tg}
D.~B.~Kaplan, M.~J.~Savage and M.~B.~Wise,
%``A new expansion for nucleon nucleon interactions,''
Phys.\ Lett.\ B {\bf 424}, 390 (1998)
[nucl-th/9801034].
%%CITATION = NUCL-TH 9801034;%%
%\cite{Cohen:1999jr}
\bibitem{Cohen:1999jr}
T.~D.~Cohen and J.~M.~Hansen,
%``Low energy theorems for nucleon nucleon scattering,''
Phys.\ Rev.\ C {\bf 59}, 13 (1999)
[nucl-th/9808038].
%%CITATION = NUCL-TH 9808038;%%
%\cite{Fleming:2000ee}
\bibitem{Fleming:2000ee}
S.~Fleming, T.~Mehen and I.~W.~Stewart,
%``NNLO corrections to nucleon nucleon scattering and perturbative pions,''
Nucl.\ Phys.\ A {\bf 677}, 313 (2000)
[nucl-th/9911001].
%%CITATION = NUCL-TH 9911001;%%
%\cite{Beane:2001bc}
\bibitem{Beane:2001bc}
S.~R.~Beane, P.~F.~Bedaque, M.~J.~Savage and U.~v.~Kolck,
%``Towards a perturbative theory of nuclear forces,''
nucl-th/0104030.
%%CITATION = NUCL-TH 0104030;%%
%\cite{Eiras:2001hu}
\bibitem{Eiras:2001hu}
D.~Eiras and J.~Soto,
%``Renormalizing the Lippmann-Schwinger equation for the one pion exchange potential,''
nucl-th/0107009.
%%CITATION = NUCL-TH 0107009;%%
%\cite{Gegelia:1999iu}
\bibitem{Gegelia:1999iu}
J.~Gegelia,
%``Issues of regularization and renormalization in effective field theories
%of nucleon nucleon scattering,''
J.\ Phys.\ {\bf G25}, 1681 (1999)
[nucl-th/9805008].
%%CITATION = NUCL-TH 9805008;%%
%\cite{Gegelia:1999ja}
\bibitem{Gegelia:1999ja}
J.~Gegelia,
%``Nucleon-Nucleon Scattering and Effective Field Theory: Including Pions
%Non-perturbatively,''
Phys.\ Lett.\ B {\bf 463}, 133 (1999)
[nucl-th/9908055].
%%CITATION = NUCL-TH 9908055;%%
%\cite{Gegelia:1998gn}
\bibitem{Gegelia:1998gn}
J.~Gegelia,
%``EFT and N N scattering,''
Phys.\ Lett.\ B {\bf 429} (1998) 227.
%%CITATION = PHLTA,B429,227;%%
%\cite{Lepage:1999kt}
\bibitem{Lepage:1999kt}
G.~P.~Lepage,
%``Conference summary,''
{\it In *Seattle 1999, Nuclear physics with effective field theory* 353-365}.
%\cite{vanKolck:1999mw}
\bibitem{vanKolck:1999mw}
U.~van Kolck,
%``Effective field theory of nuclear forces,''
Prog.\ Part.\ Nucl.\ Phys.\ {\bf 43}, 337 (1999)
[nucl-th/9902015].
%%CITATION = NUCL-TH 9902015;%%
%\cite{Epelbaum:2000dj}
\bibitem{Epelbaum:2000dj}
E.~Epelbaum, W.~Glockle and U.~Meissner,
%``Nuclear forces from chiral Lagrangians using the method of unitary transformation. II: The two-nucleon system,''
Nucl.\ Phys.\ A {\bf 671}, 295 (2000)
[nucl-th/9910064].
%%CITATION = NUCL-TH 9910064;%%
%\cite{Epelbaum:2001mx}
\bibitem{Epelbaum:2001mx}
E.~Epelbaum, H.~Kamada, A.~Nogga, H.~Witala, W.~Glockle and U.~Meissner,
%``The three- and four-nucleon systems from chiral effective field theory,''
Phys.\ Rev.\ Lett.\ {\bf 86}, 4787 (2001)
[nucl-th/0007057].
%%CITATION = NUCL-TH 0007057;%%
%\cite{Epelbaum:2001fm}
\bibitem{Epelbaum:2001fm}
E.~Epelbaum, U.~G.~Meissner, W.~Glockle and C.~Elster,
%``Resonance saturation for four-nucleon operators,''
nucl-th/0106007.
%%CITATION = NUCL-TH 0106007;%%
%\cite{Collins:1984xc}
\bibitem{Collins:1984xc}
J.~C.~Collins,
``Renormalization. An Introduction To Renormalization, The Renormalization
Group, And The Operator Product Expansion,''
{\it Cambridge, Uk: Univ. Pr. ( 1984) 380p}.
%\cite{Beane:1998pk}
\bibitem{Beane:1998pk}
S.~R.~Beane, T.~D.~Cohen and D.~R.~Phillips,
%``The potential of effective field theory in N N scattering,''
Nucl.\ Phys.\ A {\bf 632}, 445 (1998)
[nucl-th/9709062].
%%CITATION = NUCL-TH 9709062;%%
%\cite{Gegelia:1998xr}
\bibitem{Gegelia:1998xr}
J.~Gegelia,
%``Chiral perturbation theory approach to N N scattering problem,''
In *Adelaide 1998, Nonperturbative methods in quantum field theory* 30-35.
nucl-th/9802038.
%%CITATION = NUCL-TH 9802038;%%
%\end{thebibliography}
\end{references}
\end{document}