Monopole Excitation to Cluster States
Extracted source text ยท arXiv nucl-th/0703045
%%% Physics Review manuscript
\documentclass[preprint,aps]{revtex4}
\usepackage[dvipdfm]{graphicx}
\def\Vec#1{\mbox{\boldmath $#1$}}
%\usepackage{txfonts}
\begin{document}
\title{Monopole Excitation to Cluster States}
\author{
Taiichi Yamada,
Yasuro Funaki$^{1}$,
Hisashi Horiuchi$^{2,3}$, \\
Kiyomi Ikeda$^{1}$, and Akihiro Tohsaki$^{2}$
}
\affiliation{
Laboratory of Physics, Kanto Gakuin University, Yokohama 236-8501, Japan,}
\affiliation{
$^{1}$Nishina Center for Accelerator-based Science,
The Institute of Physical and Chemical Research (RIKEN), Wako 351-0098,
Japan,}
\affiliation{
$^{2}$Research Center for Nuclear Physics, Osaka University, Osaka 567-0047,
Japan,}
\affiliation{
$^{3}$International Institute for Advanced Studies, Kizugawa 619-0225,
Japan}
\date{\today}
\begin{abstract}
We discuss strength of monopole excitation of the ground state to cluster
states in light nuclei. We clarify that the monopole excitation to cluster
states is in general strong as to be comparable with the single particle
strength and shares an appreciable portion of the sum rule value in spite
of large difference of the structure between the cluster state and the
shell-model-like ground state. We argue that the essential reasons
of the large strength are twofold. One is the fact that the clustering
degree of freedom is possessed even by simple shell model wave functions.
The detailed feature of this fact is described by the so-called
Bayman-Bohr theorem which tells us that SU(3) shell model wave function
is equivalent to cluster model wave function. The other is the ground
state correlation induced by the activation of the cluster degrees of
freedom described by the Bayman-Bohr theorem. We demonstrate, by deriving
analytical expressions of monopole matrix elements, that the order of
magnitude of the monopole strength is governed by the first reason, while
the second reason plays a sufficient role in reproducing the data up to
the factor of magnitude of the monopole strength. Our explanation is made
by analysing three examples which are the monopole excitations to
the $0^+_2$ and $0^+_3$ states in $^{16}$O and the one to the
$0^+_2$ state in $^{12}$C. The present results imply that the measurement
of strong monopole transitions or excitations is in general very useful
for the study of cluster states.
\end{abstract}
\pacs{23.20.-g, 21.60.Gx, 21.60.Cs}
\maketitle
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Introduction}\label{introduction}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The monopole transitions from cluster states to ground states in light
nuclei are rather large in comparison with the single particle strength.
For example in $^{16}$O the monopole matrix elements $M({\rm E0})$ between the
ground state and the first and second excited $0^+$ states at
$E_x$ = 6.05 MeV and 12.05 MeV which are known to have $^{12}$C+$\alpha$
cluster structure~\cite{supple,hoik,suzuki} are $3.55\pm0.21$~fm$^2$
and $4.03\pm0.09$~fm$^2$~\cite{ajzen86}, respectively.
Also in $^{12}$C the $M({\rm E0})$ value between the ground state and the first
excited $0^+$ state at $E_x$ = 7.66 MeV (so-called Hoyle state~\cite{Hoyle})
which is known to have a $3\alpha$ cluster structure~\cite{supple} is
$5.4\pm0.2$~fm$^2$~\cite{ajzen86}.
A rough estimate of the single particle strength
$\langle u_f(r) | r^2 | u_i(r) \rangle$ is $(3/5)R^2\sim5.4~{\rm fm}^2$
for $p$- and $sd$-shell nuclei ($R\sim3.0$ fm). This estimation formula
is obtained under the uniform-density approximation of $u(r)
\sim \sqrt{3/R^3}$ for $u_f(r)$ and $u_i(r)$ with $R$ standing for the
nuclear radius. The energy weighted strengths of the above mentioned
monopole transitions give an appreciable portion of the sum rule values:
in $^{16}$O they are about 3 $\%$ and 8 \% for $0^+_2$ and $0^+_3$,
respectively, and in $^{12}$C about 16 \% for $0^+_2$ (see Appendix~\ref{sumrule}).
Recently Kawabata and his collaborators have studied the excited states
of $^{11}$B by performing ${^{11}{\rm B}}(d, d')$ reaction and they
concluded that the third $3/2^-$ state at $E_x$ = 8.56 MeV has a
$2\alpha + t$ cluster structure~\cite{kawabata}. Among many reasons for
this conclusion, one is a large monopole strength for the third $3/2^-$
state which is of similar value to the monopole strength for the
second $0^+$ state in $^{12}$C, and another is that the AMD
(antisymmetrized molecular dynamics) calculation~\cite{kawabata}
as well as the $2\alpha+t$ OCM (orthogonality condition model)
calculation~\cite{yamada07} have reproduced the large monopole strength
and have assigned loosely bound $2\alpha + t$ cluster structure to the
third $3/2^-$ state.
The single particle estimate of the monopole transition is based on
the assumption that the excited state has a one-particle one-hole excitation
from the ground state. However, the cluster structure is very different
from the shell-model-like structure of the ground state, and its state is
described as a superposition of many-particle many-hole configurations when
it is expanded by shell model configurations. This means that in the
excited state with a cluster structure, the component of a one-particle
one-hole excitation from the ground state configuration is expected to be
very small. Therefore the observation of rather large monopole strengths
for cluster states which are comparable with single particle strength
looks {\it not} to be easy to explain. The $^{12}$C+$\alpha$ OCM
calculation~\cite{suzuki} for $^{16}$O and $3\alpha$ RGM
(resonating group method) calculation~\cite{kami,uega}, however, have
reproduced rather well the experimental data of the monopole transitions.
No explicit and detailed analyses of the reason why the cluster models
reproduce plausibly the experimental data have been presented so far
as long as we know. There should exist underlying physics in the
monopole transition strengths in light nuclei.
The purpose of this paper is to clarify the basic reasons why monopole
transition strength between a cluster state and the ground state in light
nuclei is generally rather large in comparison with the single particle
strength and shares an appreciable portion of the sum rule value,
in spite of the large difference of structure between the initial and
final states. We analyse the above-mentioned three cases of monopole
transisions in $^{16}$O and $^{12}$C, namely the monopole transitions
between the ground state and the first and second excited $0^+$ states
in $^{16}$O, and the monopole transition between the ground state and
the first excited $0^+$ state in $^{12}$C. By using these analyses
we will show that there are two basic reasons for the generally large
strength of monopole transitions. The first reason is the fact that the
clustering degree of freedom is possessed even by simple shell model
wave functions. The detailed feature of this fact is described by
the so-called Bayman-Bohr theorem~\cite{BB}. This theorem tells us
that the SU(3) shell model wave function~\cite{elliott58} describing
the ground state is in most cases equivalent to the cluster-model wave
function discussed by Wildermuth and Kanellopoulos~\cite{Wildermuth}.
Thus we can see what kinds of clustering degrees of freedom are embedded
in the ground state. For example the doubly closed-shell wave function
of the $^{16}$O ground state (total quanta $N_{TOT}=12$) which is just
the SU(3) shell model wave function with $(\lambda, \mu)=(0, 0)$
is equivalent to a $^{12}$C + $\alpha$ cluster-model wave function
with $N_{TOT}=12$. This means that the ground-state wave function
of $^{16}$O originally has a $^{12}$C+$\alpha$ clustering degree of
freedom. The second reason is the ground state correlation induced
by the activation of the cluster degrees of freedom described by the
Bayman-Bohr theorem. In the case of the above example of the $^{16}$O
ground state, the ground state correlation is due to the $^{12}$C +
$\alpha$ clustering degrees of freedom. As was explained, the first
and second excited $0^+$ states of $^{16}$O are the cluster states
with $^{12}$C + $\alpha$ structure. These cluster states are formed
just by the excitation of the $^{12}$C + $\alpha$ clustering degree
of freedom which is already existent in the ground state. Therefore it
is quite reasonable that the large strength of the monopole transition
between the ground state and the first and second excited $0^+$
states is explained by the above-mentioned first and second reasons.
We will demonstrate, by deriving analytical expressions of monopole
matrix elements, that the order of magnitude of the monopole strength
is governed by the first reason, while the second reason plays a
sufficient role in reproducing the data up to the factor of
magnitude of the monopole strength.
In the present paper we discuss the details of the first and second
reasons for $^{16}$O and $^{12}$C. In the case of $^{16}$O, we
make use of the microscopic $^{12}$C + $\alpha$ cluster wave function,
while in the case of $^{12}$C, we discuss the problem by using the
so-called THSR wave function~\cite{tohsaki,funaki}. Our results
mean that the measurement of strong monopole transitions provides
us in general with a very useful tool for the experimental study
of cluster states as has been practiced in Ref.~\cite{kawabata}.
The present paper is organized as follows. In Sec.~\ref{formulation}
we derive analytical expressions of the monopole matrix elements
between the ground state and $^{12}$C+$\alpha$ cluster states in
$^{16}$O and those between the ground state and $3\alpha$ cluster state
in $^{12}$C, by using the Bayman-Bohr theorem.
In Sec.~\ref{gr-st-correl} we discuss the effect of the ground
state correlation on the monopole transitions from the $^{12}$C$ +
\alpha$ cluster states in $^{16}$O, and those from the $3\alpha$ cluster
state in $^{12}$C. In Sec.~\ref{summary} we give discussions and summary.
%\newpage
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Monopole Transition and Bayman-Bohr Theorem} \label{formulation}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Monopole transition from two-cluster states in $^{16}$O}\label{subsection:Bayman_Bohr}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
We discuss the following two observed values of the monopole transition
matrix element in $^{16}$O:~One is $M({\rm E0}) = 3.55 \pm 0.21$ fm$^2$
between the ground state ($0^+_1$) and the first excited $0^+$ state
($0^+_2$) at $E_x$ = 6.05 MeV, and the other is $M({\rm E0}) = 4.03 \pm 0.09$
fm$^2$ between the ground state and the excited $0^+$ state ($0^+_3$)
at $E_x$ = 12.05 MeV. These excited $0^+$ states are known to have
$^{12}$C+$\alpha$ structures~\cite{supple,hoik,suzuki}.
In this section, we explain that the order of magnitude of these $M({\rm E0})$
values comparable with the single nucleon strength is explained to
come from the fact that the doubly closed shell wave function already contains
in it the $^{12}$C+$\alpha$ clustering degree of freedom. For this purpose
we derive analytical expressions of these monopole matrix elements by the
use of the Bayman-Bohr theorem.
The nuclear SU(3) model or Elliott model~\cite{elliott58} is known to
describe well ground states of light nuclei.
The ground state of $^{16}$O has a doubly closed shell structure of
$0s$ and $0p$ orbits which belongs to the SU(3) irreducible
representation $(\lambda,\mu)=(0,0)$.
This doubly closed shell model wave function with the nucleon size
parameter $\nu_N=M\omega/2\hbar$ ($M$: nucleon mass)
is equivalent to a cluster wave function of $^{12}$C + $\alpha$
configuration, according to the Bayman-Bohr theorem,
\begin{eqnarray}
&& \frac{1}{\sqrt{16!}} \det |(0s)^4 (0p)^{12} | = N_g \frac{1}
{\sqrt{{}_{16}C_4}}
{\cal A} \{ [{\cal R}_4(\Vec{r},3\nu_N) \phi({}^{12}{\rm C})]_{(0,0)}
\phi(\alpha) \}\phi_G(\Vec{r}_G), \label{eq:clshl} \\
&& [{\cal R}_4(\Vec{r},3\nu_N) \phi({}^{12}{\rm C})]_{(\lambda,\mu)=(0,0)}
= \sum_{L=0,2,4} C_L \ [{\cal R}_{4L}(\Vec{r},3\nu_N)
\phi_L({}^{12}{\rm C})]_{J=0}, \label{eq:alc} \\
&& \phi_G(\Vec{r}_G)=\left(\frac{32\nu_N}{\pi}\right)^{3/4}
\exp\left(-16\nu_N\Vec{r}_G^2\right),\ \ \Vec{r}_{\rm G} =
\frac{1}{16} \sum_{i=1}^{16} \Vec{r}_i \\
&& C_L = \langle (4,0)L, (0,4)L ||(0,0)0 \rangle, \ \ {}_{16}C_4 =
\frac{16!}{12!4!}.
\end{eqnarray}
Here $\phi(\alpha)$ and $\phi_L({}^{12}{\rm C})$ stand for the internal wave
function of $\alpha$ cluster with the $(0s)^4$ configuration and
internal wave function of $^{12}$C with angular momentum $L$, respectively.
$\phi_G$ denotes the center-of-mass wave function of $^{16}$O, which can be
separated from the internal wave function as is written in
Eq.~(\ref{eq:clshl}). The relative wave function between the $\alpha$ and
$^{12}$C clusters is presented by the harmonic oscillator wave function
${\cal R}_{NLm}(\Vec{r},\beta) = R_{NL}(r,\beta) Y_{Lm}({\hat{\Vec{r}}})$
with the oscillator quanta $N=4$ [nodal number $n = (N-L)/2$] and
size parameter $\beta=3\nu_N$, where $\Vec{r}$ is the relative coordinate
between the center-of-masses of $\alpha$ and $^{12}$C clusters.
It is noted that ${\cal R}_{4L}(\Vec{r},3\nu_N)$ and
$\phi_L({}^{12}{\rm C})$ belong to the SU(3) irreducible representations
$(\lambda,\mu)=(4,0)$ and (0,4), respectively. Equation~(\ref{eq:alc})
means that these representations are coupled to the SU(3) scalar
representation $(\lambda,\mu)=(0,0)$. ${\cal A}$ is the nucleon
antisymmetrizer between $^{12}$C and $\alpha$ cluster, $N_g$ is the
normalization constant, and $C_L$ is the reduced Clebsch-Gordan
coefficient of SU(3) group for the SU(3) vector coupling $(4,0) \times
(0,4) \rightarrow (0,0)$.
The doubly closed shell model wave function of $^{16}$O has the total
number of the oscillator quanta $N_{TOT}=12$ and is only one possible
wave function allowed for $N_{TOT}=12$. Since all three wave functions
of $^{16}$O,
${\cal A} \{ [{\cal R}_{4L}(\Vec{r},3\nu_N) \phi_L({}^{12}{\rm C})]_{J=0}
\phi (\alpha) \}$ for $L$=0, 2, and 4, have the total quanta
$N_{TOT} = 12$, they are necessarily equivalent to the doubly closed shell
model wave function $\Phi_{CS}$ and represent the internal wave function
of the $^{16}$O ground state (see also Appendix~\ref{Baymann_Bohr}),
\begin{eqnarray}
|{0^+_1}\rangle &=& \Phi_{CS} \equiv {\frac{1}{\sqrt{16!}}
\det |(0s)^4 (0p)^{12} |}
\times [\phi_G(\Vec{r}_G)]^{-1} \label{eq:clshl0} \\
&=& N_{g0}\frac{1}{\sqrt{{}_{16}C_4}}
{\cal A} \{ [{\cal R}_{40}(\Vec{r},3\nu_N)
\phi_{L=0}({}^{12}{\rm C})]_{J=0} \phi(\alpha) \} \label{eq:clshla}
\\
&=& N_{g2} \frac{1}{\sqrt{{}_{16}C_4}}
{\cal A} \{ [{\cal R}_{42}(\Vec{r},3\nu_N)
\phi_{L=2}({}^{12}{\rm C})]_{J=0} \phi(\alpha) \} \label{eq:clshlb}
\\
&=& N_{g4} \frac{1}{\sqrt{{}_{16}C_4}}
{\cal A} \{ [{\cal R}_{44}(\Vec{r},3\nu_N)
\phi_{L=4}({}^{12}{\rm C})]_{J=0} \phi(\alpha) \}, \label{eq:clshlc}
\end{eqnarray}
where $N_{g0}$, $N_{g2}$, and $N_{g4}$ denote the normalization constants.
It is important to recognize the implication of these relations of
Eqs.~(\ref{eq:clshl}), and (\ref{eq:clshla})$\sim$(\ref{eq:clshlc}).
They imply that the ground state of $^{16}$O can
be excited not only through single particle degrees of freedom by promoting
nucleons from $0s$ and $0p$ orbits to higher orbits, but also through cluster
degrees of freedom by exciting the $^{12}$C$-$$\alpha$ relative motion from
${\cal R}_{4L}(\Vec{r},3\nu_N)$ state to higher nodal states.
The latter characteristic is an essential point to understand
why the monopole transition matrix elements to cluster states are in general
large.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsubsection{Monopole transition between $0_1^+$ and $0_2^+$ states}\label{sec:ME0_16O_21}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The $0_2^+$ state of $^{16}$O is known to have a loosely bound
$^{12}$C+$\alpha$ structure, in which the dominant component of
$^{12}$C is the ground state~\cite{supple,hoik,suzuki}.
Thus we express the $0_2^+$ wave function as
\begin{equation}
|0_2^+ \rangle = N_I \frac{1}{\sqrt{{}_{16}C_4}} {\cal A} \{ \chi_0(\Vec{r})
\phi_{L=0}({}^{12}{\rm C}) \phi(\alpha) \}, \label{eq:mysta}
\end{equation}
where $N_I$ represents the normalization constant. By expanding
$\chi_0(\Vec{r})$ in terms of harmonic oscillator functions, we have
\begin{eqnarray}
|0^+_2 \rangle &=& \sum_{N=6}^\infty \eta_N \Phi_N, \\
\Phi_N &=& \frac{1}{\sqrt{\tau_{0,N}}} \frac{1}{\sqrt{{}_{16}C_4}}
{\cal A} \{ R_{N0}(r,3\nu_N) Y_{00}(\hat{\Vec{r}})
\phi_{L=0}({}^{12}{\rm C}) \phi(\alpha) \}, \\
\tau_{0,N} &\equiv& \langle {\cal R}_{N0}(\Vec{r},3\nu_N)
\phi_{L=0}({}^{12}{\rm C}) \phi(\alpha) | {\cal A}
\{ {\cal R}_{N0}(\Vec{r},3\nu_N) \phi_{L=0}({}^{12}{\rm C})
\phi(\alpha) \} \rangle.
\end{eqnarray}
It should be noted that $\Phi_N$ are normalized. Also it should be noted
that $\Phi_{N=4}$ is just the doubly closed shell wave function as is
seen in Eq.~(\ref{eq:clshla}), $|0_1^+ \rangle = \Phi_{N=4}$.
Since both $0_1^+$ and $0_2^+$ states have the total isospin $T=0$,
the monopole transition matrix element $M({\rm E0})$ is
\begin{eqnarray}
&& M({\rm E0},0_2^+ - 0_1^+) = \langle 0_1^+ | \sum_{i=1}^{16}\frac{1}{2}
(1+\tau_{3i}) (\Vec{r}_i - \Vec{r}_G)^2 | 0_2^+ \rangle \nonumber \\
&& \ \ = \langle 0_1^+ | \frac{1}{2} \sum_{i=1}^{16}
(\Vec{r}_i - \Vec{r}_G)^2 | 0_2^+ \rangle \nonumber \\
&& \ \ = \eta_6 \langle \Phi_{N=4} | \frac{1}{2} \sum_{i=1}^{16}
(\Vec{r}_i - \Vec{r}_G)^2 | \Phi_{N=6} \rangle.
\end{eqnarray}
Here $\Vec{r}_G$ stands for the total center-of-mass coordinate,
$\Vec{r}_{\rm G} = (1/16) \sum_{i=1}^{16} \Vec{r}_i$.
The last equality is because
$\sum_{i=1}^{16} (\Vec{r}_i - \Vec{r}_G)^2 \Phi_{N=4}$ can not have
more than 2$\hbar \omega$ excitation than $\Phi_{N=4}$. Then we have
\begin{eqnarray}
&& M({\rm E0},0_2^+ - 0_1^+) = \frac{\eta_6}{2} \frac{1}
{\sqrt{\tau_{0,4}\tau_{0,6}}} \nonumber \\
&& \ \ \ \times \langle {\cal R}_{40}(\Vec{r},3\nu_N)
\phi_{L=0}({}^{12}{\rm C}) \phi(\alpha) | {\cal A}
\{ (\sum_{i=1}^{16}(\Vec{r}_i - \Vec{r}_{\rm G})^2)
{\cal R}_{60}(\Vec{r},3\nu_N)
\phi_{L=0}({}^{12}{\rm C}) \phi(\alpha) \} \rangle \nonumber \\
&& \ \ = \frac{\eta_6}{2} \frac{1}{\sqrt{\tau_{0,4}\tau_{0,6}}}
\langle {\cal R}_{40}(\Vec{r},3\nu_N)
\phi_{L=0}({}^{12}{\rm C}) \phi(\alpha) | {\cal A}
\{ \frac{12 \times 4}{16} \Vec{r}^2 {\cal R}_{60}(\Vec{r},3\nu_N)
\phi_{L=0}({}^{12}{\rm C}) \phi(\alpha) \} \rangle. \label{eq:matr}.
\end{eqnarray}
In obtaining Eq.~(\ref{eq:matr}), we first used the identity
\begin{equation}
\sum_{i=1}^{16}(\Vec{r}_i - \Vec{r}_G)^2 = \sum_{i \in {^{12}{\rm C}} }
(\Vec{r}_i - \Vec{r}_C)^2
+ \sum_{i \in \alpha }(\Vec{r}_i - \Vec{r}_\alpha)^2 +
\frac{12 \times 4}{16} \Vec{r}^2, \label{eq:monop}
\end{equation}
where $\Vec{r}_{\rm C}$ and $\Vec{r}_\alpha$ express the center-of-mass
coordinate of $^{12}$C and $\alpha$, respectively.
We then used the following relations,
\begin{eqnarray}
&& \langle {\cal R}_{40}(\Vec{r},3\nu_N) \phi_{L=0}({}^{12}{\rm C})
\phi(\alpha) | {\cal A} \{ {\cal R}_{60}(\Vec{r},3\nu_N) \
( \sum_{i \in {}^{12}{\rm C} } (\Vec{r}_i - \Vec{r}_{\rm C})^2 )
\phi_{L=0}({}^{12}{\rm C})\ \phi(\alpha) \} \rangle = 0,
\label{eq:matra} \\
&& \langle {\cal R}_{40}(\Vec{r},3\nu) \phi_{L=0}({}^{12}{\rm C})
\phi(\alpha) | {\cal A} \{ {\cal R}_{60}(\Vec{r},3\nu_N) \
\phi_{L=0}({}^{12}{\rm C})\ (\sum_{i \in \alpha }(\Vec{r}_i
- \Vec{r}_\alpha)^2 ) \phi(\alpha) \} \rangle = 0. \label{eq:matrb}
\end{eqnarray}
These relations can be easily proved by counting the total numbers of
oscillator quanta of the bra and ket functions.
First, the number of the oscillator quanta of
${\cal R}_{60}(\Vec{r},3\nu_N)$ is larger
than that of ${\cal R}_{40}(\Vec{r},3\nu_N)$ by 2.
Second, the number of the oscillator quanta of
$( \sum_i (\Vec{r}_i - \Vec{r}_{\rm C})^2 ) \phi_{L=0}({}^{12}{\rm C})$
can not be smaller than that of
$\phi_{L=0}({}^{12}{\rm C})$ because $\phi_{L=0}({}^{12}{\rm C})$ has the
smallest number of the oscillator quanta in the $^{12}$C ($N=Z=6$) system.
Similarly, the number of the oscillator quanta of
$(\sum_i (\Vec{r}_i - \Vec{r}_\alpha)^2 ) \phi(\alpha)$
can not be smaller than that of $\phi(\alpha)$.
Therefore in each of Eqs.~(\ref{eq:matra}) and (\ref{eq:matrb}),
the ket function has larger total number of the oscillator
quanta than that of the bra function at least by 2, which leads to
the orthogonality of the bra and ket functions.
Now we expand $(12 \times 4/16) \Vec{r}^2 {\cal R}_{60}(\Vec{r},3\nu_N)$ in
Eq.~(\ref{eq:matr}) in terms of the harmonic oscillator function
\begin{equation}
\frac{12 \times 4}{16} \Vec{r}^2 {\cal R}_{60}(\Vec{r},3\nu_N) = \sum_N
\langle {\cal R}_{N0}(\Vec{r},3\nu_N) | \frac{12 \times 4}{16} \Vec{r}^2 |
{\cal R}_{60}(\Vec{r},3\nu_N) \rangle \ {\cal R}_{N0}(\Vec{r},3\nu_N)
\label{eq:expand}.
\end{equation}
By inserting Eq.~(\ref{eq:expand}) into Eq.~(\ref{eq:matr}) we obtain
\begin{eqnarray}
M({\rm E0},0_2^+ - 0_1^+) &=& \frac{\eta_6}{2} \frac{1}
{\sqrt{\tau_{0,4}\tau_{0,6}}} \langle
{\cal R}_{40}(\Vec{r},3\nu_N) | \frac{12 \times 4}{16}
\Vec{r}^2 | {\cal R}_{60}(\Vec{r},3\nu_N) \rangle \nonumber \\
&& \ \ \times \langle {\cal R}_{40}(\Vec{r},3\nu_N)
\phi_{L=0}({}^{12}{\rm C}) \phi(\alpha) | {\cal A}
\{ {\cal R}_{40}(\Vec{r},3\nu_N) \phi_{L=0}({}^{12}{\rm C})
\phi(\alpha) \} \rangle \nonumber \\
&=& \frac{\eta_6}{2} \sqrt{\frac{\tau_{0,4}}{\tau_{0,6}}}
\langle {\cal R}_{40}(\Vec{r},3\nu_N) |
\frac{12 \times 4}{16} \Vec{r}^2 | {\cal R}_{60}(\Vec{r},3\nu_N) \rangle
\nonumber \\
&=& \frac{\eta_6}{2} \sqrt{\frac{\tau_{0,4}}{\tau_{0,6}}}
\langle R_{40}(r,3\nu_N) | \frac{12 \times 4}{16} r^2 | R_{60}(r,3\nu_N)
\rangle. \label{eq:ezeroa}
\end{eqnarray}
Here we note the following relation
\begin{equation}
\langle R_{40}(r,3\nu_N) | \frac{12 \times 4}{16} r^2 | R_{60}(r,3\nu_N)
\rangle
= \langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle,
\end{equation}
where $R_{N0}(r,\nu_N)$ is the harmonic oscillator radial function
of single nucleon with the nucleon size parameter $\nu_N$.
It is noted here that the matrix elements for calculating the single
particle E0 matrix element in $^{16}$O are
$\langle R_{00}(r,\nu_N) | r^2 | R_{20}(r,\nu_N) \rangle$ and
$\langle R_{11}(r,\nu_N) | r^2 | R_{31}(r,\nu_N) \rangle$
which are a few times smaller than the present
$\langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle$ as shown below,
\begin{eqnarray}
\langle R_{00}(r,\nu_N) | r^2 | R_{20}(r,\nu_N) \rangle &=&
\sqrt{\frac{3}{8}} \frac{1}{\nu_N}, \ \ \
\langle R_{11}(r,\nu_N) | r^2 | R_{31}(r,\nu_N) \rangle =
\sqrt{\frac{5}{8}} \frac{1}{\nu_N}, \label{mt_of_r2}\\
\langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle &=&
\sqrt{\frac{21}{8}} \frac{1}{\nu_N}.
\end{eqnarray}
The reason why the number of oscillator quanta of the relative wave function
is higher than those of the single particle wave functions is due to the
Fermi statistics of nucleons.
The final analytical formula of $M({\rm E0},0_2^+ - 0_1^+)$ is expressed as follows,
\begin{eqnarray}
M({\rm E0},0_2^+ - 0_1^+)
= \frac{1}{2} \sqrt{ \frac{\tau_{0,4}}{\tau_{0,6}}} \eta_6
\langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle.
\label{eq:form1}
\end{eqnarray}
This analytical expression of $M({\rm E0},0_2^+ - 0_1^+)$
is our desired result. It explains clearly why $M({\rm E0},0_2^+ - 0_1^+)$
has a comparable magnitude as the single nucleon E0 matrix element.
The factor $\langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle$
is a few times larger than the single nucleon E0 matrix elements
$\langle R_{00}(r,\nu_N) | r^2 | R_{20}(r,\nu_N) \rangle$ and
$\langle R_{11}(r,\nu_N) | r^2 | R_{31}(r,\nu_N) \rangle$, while the
factor $\eta_6$ works to make the E0 value smaller.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsubsection{Monopole transition between $0_1^+$ and $0_3^+$ states}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The $0_3^+$ state of $^{16}$O at $E_x$ = 12.05 MeV is known to have also
a $^{12}$C + $\alpha$ structure like the $0_2^+$ state~\cite{supple,suzuki}.
The $^{12}$C cluster in the $0_3^+$ state, however, is not mainly in its
ground state like in $0_2^+$ state but dominantly in its excited $2^+$ state
at $E_x$ = 4.44 MeV. Thus we can express the $0_3^+$ wave function in a
good approximation as
\begin{equation}
|0_3^+ \rangle = N_{II} \frac{1}{\sqrt{{}_{16}C_4}} {\cal A}
\{ [\chi_2(\Vec{r}) \phi_{L=2}({}^{12}{\rm C})]_{J=0} \phi(\alpha) \}.
\end{equation}
Like in the case of $0_2^+$ state, we expand $\chi_2(\Vec{r})$ in terms of
harmonic oscillator wave functions and we obtain
\begin{eqnarray}
|0^+_3 \rangle &=& \sum_{N=6}^\infty \zeta_N \Psi_N, \\
\Psi_N &=& \frac{1}{\sqrt{\tau_{2,N}}} \frac{1}{\sqrt{{}_{16}C_4}}
{\cal A} \{ [{\cal R}_{N2}(\Vec{r},3\nu_N)
\phi_{L=2}({}^{12}{\rm C})]_{J=0} \phi(\alpha) \}, \\
\tau_{2,N} &\equiv& \langle [{\cal R}_{N2}(\Vec{r},3\nu_N)
\phi_{L=2}({}^{12}{\rm C})]_{J=0} \phi(\alpha) | {\cal A}
\{ [{\cal R}_{N2}(\Vec{r},3\nu_N) \phi_{L=2}({}^{12}{\rm C})]_{J=0}
\phi(\alpha) \} \rangle.
\end{eqnarray}
It should be noted that $\Psi_N$ are normalized. Also it should be noted
that $\Psi_{N=4}$ is just the doubly closed shell wave function as is
seen in Eq.~(\ref{eq:clshlb}), $|0_1^+ \rangle = \Psi_{N=4}$.
The calculation of the monopole transition matrix element
$M({\rm E0},0_3^+ - 0_1^+)$ can be made in the same manner as that of
$M({\rm E0},0_2^+ - 0_1^+)$ in the previous section, although we use
Eq.~(\ref{eq:clshlb}) for the $0^+_1$ state of $^{16}$O,
\begin{eqnarray}
&& M({\rm E0},0_3^+ - 0_1^+)= \langle 0_1^+ |\frac{1}{2}
\sum_{i=1}^{16}(\Vec{r}_i - \Vec{r}_{\rm G})^2 | 0_3^+ \rangle \nonumber \\
&& \ \ = \frac{1}{2} \sqrt{ \frac{\tau_{2,4}}{\tau_{2,6}}}
\zeta_6
\langle R_{42}(r,\nu_N) | r^2 | R_{62}(r,\nu_N) \rangle, \label{eq:form2}
\end{eqnarray}
where
\begin{eqnarray}
\langle R_{42}(r,\nu_N) | r^2 | R_{62}(r,\nu_N) \rangle =
\frac{3}{2} \frac{1}{\nu_N}.
\end{eqnarray}
The analytical expression of $M({\rm E0},0_3^+ - 0_1^+)$ in Eq.~(\ref{eq:form2})
is our another desired result. Like in the case of $M({\rm E0},0_2^+ - 0_1^+)$,
it explains clearly why $M({\rm E0},0_3^+ - 0_1^+)$ has also a comparable magnitude
as the single nucleon E0 matrix element.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsubsection{Wave function which absorbs total monopole strength from the
doubly closed shell}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The wave function $\Phi_{(2,0)}$ which absorbs total monopole strength from
the doubly closed shell wave function $\Phi_{CS}$ is given by
\begin{eqnarray}
\Phi_{(2,0)} &=& N_{(2,0)} (1 - |\Phi_{CS} \rangle \langle \Phi_{CS}|) O_M
\Phi_{CS}, \label{eq:Phi_20} \\
O_M &=& \frac{1}{2} \sum_{i=1}^{16} (\Vec{r}_i - \Vec{r}_G)^2,
\end{eqnarray}
where $N_{(2,0)}$ is the normalization constant and is presented as
\begin{eqnarray}
\frac{1}{N_{(2,0)}} & = & \sqrt{ \langle \Phi_{CS}| O_M^2 |\Phi_{CS} \rangle - \langle \Phi_{CS}| O_M |\Phi_{CS} \rangle^2 }
\label{eq:normalization_Phi_20}
= \sqrt{\frac{69}{32}}\frac{1}{\nu_N}.
\end{eqnarray}
Any wave function $\Phi$ which is orthogonal to both $\Phi_{CS}$ and
$\Phi_{(2,0)}$ has zero monopole strength from $\Phi_{CS}$, namely
$\langle \Phi | O_M | \Phi_{CS} \rangle = 0$. This fact is easily derived
from the orthogonality of $\Phi$ to $\Phi_{CS}$ and $\Phi_{(2,0)}$.
%It is easy to show that the monopole strength of the wave function
%$\Phi_{(2,0)}$, which is the total monopole strength from $\Phi_{CS}$,
%is given by
Then, the monopole strength of the wave function $\Phi_{(2,0)}$ from $\Phi_{CS}$
is given by
\begin{eqnarray}
\langle \Phi_{(2,0)} | O_M | \Phi_{CS} \rangle = \frac{1}{N_{(2,0)}}
= \sqrt{\frac{69}{32}}\frac{1}{\nu_N} = \frac{1.47}{\nu_N}.
\label{eq:monopole_strength_from_CS}
\end{eqnarray}
Reminding of the relation of
$\langle \Phi_{CS}| O_M^2 |\Phi_{CS} \rangle = { \sum_{k} \left| \langle \Phi_{k} | O_{M} | \Phi_{CS} \rangle \right|^{2} }$
($\{ \Phi_{k} \}$ denoting a complete set of wave functions) in Eq.~(\ref{eq:normalization_Phi_20}),
one finds that the monopole strength in Eq.~(\ref{eq:monopole_strength_from_CS}) corresponds
to the squared root of the non-energy-weighted sum rule
($\sum_{k\not={CS}} |\langle \Phi_{k} | O_{M} | \Phi_{CS} \rangle |^2$)
of the monopole operator $O_{M}$ with respect to $\Phi_{CS}$,
i.e.~exhausting the total monopole strength from $\Phi_{CS}$.
Let us denote by $\Phi_{(2,0)}^{cl}$ the $^{12}$C + $\alpha$ cluster wave
function which absorbs the total monopole strength from $\Phi_{CS}$ within
the $^{12}$C + $\alpha$ cluster model space. $\Phi_{(2,0)}^{cl}$ is not
equal to $\Phi_{(2,0)}$. It is because the monopole operator of $^{12}$C
cluster, $(1/2)\sum_{i \in {^{12}{\rm C}}}(\Vec{r}_i - \Vec{r}_C)^2$,
and that of the $\alpha$ cluster, $(1/2)\sum_{i \in \alpha}
(\Vec{r}_i - \Vec{r}_\alpha)^2$, which are contained in the total
monopole operator $O_M$ as seen in Eq.~(\ref{eq:monop}) do excite the
$^{12}$C and $\alpha$ clusters when $O_M$ operates on $\Phi_{CS}$.
These excitations of clusters imply that the wave function $\Phi_{(2,0)}$
contains components out of the $^{12}$C + $\alpha$ cluster model space.
The explicit form of $\Phi_{(2,0)}^{cl}$ is given as
\begin{eqnarray}
&& \Phi_{(2,0)}^{cl} = N_{(2,0)}^{cl} \frac{1}{\sqrt{{}_{16}C_4}}
{\cal A} \{ [{\cal R}_6(\Vec{r},3\nu_N) \phi({}^{12}{\rm C})]_{(2,0)}
\phi(\alpha) \}, \\
&& [{\cal R}_6(\Vec{r},3\nu_N) \phi({}^{12}{\rm C})]_{(2,0)} =
\sum_{L=0,2,4} \langle (6,0)L, (0,4)L ||(2,0)0 \rangle
[{\cal R}_{6L}(\Vec{r},3\nu_N) \phi_L({}^{12}{\rm C})]_{J=0}.
\end{eqnarray}
Here $\langle (6,0)L, (0,4)L ||(2,0)0 \rangle$ is the reduced Clebsch-
Gordan coefficient of the SU(3) group for the SU(3) vector coupling
$(6, 0) \times (0, 4) \rightarrow (2, 0)$. This relation is proved as
follows. Since the nucleon coordinate $\Vec{r}_i$ is the sum of the
creation ($\Vec{a}^\dagger_i$) and annihilation($\Vec{a}_i$) operators
of oscillator quanta,
$\Vec{r}_i \propto \Vec{a}^\dagger_i + \Vec{a}_i$,
the monopole operator $O_M$ consists of three parts,
$O_M = O^{(2,0)}_M + O^{(0,2)}_M + O^{(0,0)}_M$. The number of the
oscillator quanta is raised by 2 by $O^{(2,0)}_M$, lowered by 2
by $O^{(0,2)}_M$, and kept unchanged by $O^{(0,0)}_M$. The superfix
$(\lambda,\mu)$ of the operator $O^{(\lambda,\mu)}_M$ expresses
its SU(3) tensor character. Thus the 2$\hbar \omega$-excited
wave function created by operating $O_M$ on $\Phi_{CS}$
necessarily has the SU(3) symmetry (2, 0). Within the
$^{12}$C + $\alpha$ cluster model space, $\Phi_{(2,0)}^{cl}$
is the only one wave function which is 2$\hbar \omega$-excited and
has (2, 0) symmetry. Thus $\Phi_{(2,0)}^{cl}$ absorbs all the
monopole strength from $\Phi_{CS}$ and other excited wave functions
orthogonal to $\Phi_{(2,0)}^{cl}$ all have zero monopole strength.
The monopole strength of $\Phi_{(2,0)}^{cl}$ is given by (see Ref.~\cite{hecht})
\begin{eqnarray}
\langle \Phi_{(2,0)}^{cl} | O_M | \Phi_{CS} \rangle
= \sqrt{\frac{45}{32}}\frac{1}{\nu_N} = \frac{1.19}{\nu_N}.
\end{eqnarray}
This magnitude of $\langle \Phi_{(2,0)}^{cl} | O_M | \Phi_{CS}
\rangle$ is about 80~\% of the total monopole strength
$\langle \Phi_{(2,0)} | O_M | \Phi_{CS} \rangle$. We now know,
from the studies in previous subsections, that the reason of this
large value is just because of the $^{12}$C + $\alpha$ clustering
character embedded in the doubly closed shell wave function
which is described by the Bayman-Bohr theorem. Namely,
$\langle \Phi_{(2,0)}^{cl} | O_M | \Phi_{CS} \rangle$ can be
expressed as
\begin{eqnarray}
&& \langle \Phi_{(2,0)}^{cl} | O_M | \Phi_{CS} \rangle
= \sum_{L=0,2,4} E_L
\frac{1}{2} \sqrt{\frac{\tau_{L,4}}{\tau_{L,6}}}
\langle R_{4L}(r,\nu_N) | r^2 | R_{6L}(r,\nu_N) \rangle, \\
&& \frac{1}{2} \sqrt{ \frac{\tau_{0,4}}{\tau_{0,6}}}
\langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle
= \frac{0.784}{\nu_N}, \label{eq:mono0} \\
&& \frac{1}{2} \sqrt{ \frac{\tau_{2,4}}{\tau_{2,6}}}
\langle R_{42}(r,\nu_N) | r^2 | R_{62}(r,\nu_N) \rangle
= \frac{1.03}{\nu_N}, \label{eq:mono2} \\
&& \frac{1}{2} \sqrt{ \frac{\tau_{4,4}}{\tau_{4,6}}}
\langle R_{44}(r,\nu_N) | r^2 | R_{64}(r,\nu_N) \rangle
= \frac{0.946}{\nu_N}. \label{eq:mono4}
\end{eqnarray}
The definition of $\tau_{L,N}$ which is already given for
$L=$ 0 and 2 is as follows
\begin{eqnarray}
\tau_{L,N} &\equiv& \langle [{\cal R}_{NL}(\Vec{r},3\nu_N)
\phi_L({}^{12}{\rm C})]_{J=0} \phi(\alpha) | {\cal A}
\{ [{\cal R}_{NL}(\Vec{r},3\nu_N) \phi_L({}^{12}{\rm C})]_{J=0}
\phi(\alpha) \} \rangle.
\end{eqnarray}
The coefficient $E_L$ is expressed as follows
\begin{eqnarray}
E_L &=& N_{(2,0)}^{cl} \sqrt{\tau_{L,6}} \langle (6,0)L, (0,4)L ||
(2,0)0 \rangle, \\
\left(\frac{1}{N_{(2,0)}^{cl}}\right)^2 &=&
\langle [{\cal R}_6(\Vec{r},3\nu_N) \phi({}^{12}{\rm C})]_{(2,0)}
\phi(\alpha) | {\cal A} \{ [{\cal R}_6(\Vec{r},3\nu_N)
\phi({}^{12}{\rm C})]_{(2,0)} \phi(\alpha) \} \rangle \\
&=& \frac{112}{81} = 1.38.
\end{eqnarray}
The values of $\langle (6,0)L, (0,4)L ||(2,0)0 \rangle$ are
$\sqrt{1/10}$, $\sqrt{3/7}$, and $\sqrt{33/70}$, for $L$ = 0, 2,
and 4, respectively.
The value of $(1/N_{(2,0)}^{cl})^2$ is given in Ref.~\cite{horic}
with the notation $\mu^6_{(2,0)}$.
In Eqs.~(\ref{eq:mono0}) $\sim$ (\ref{eq:mono4}) we used the values of
$\tau_{L,N}$ calculated by the use of their analytical
expressions presented in Refs.~\cite{suzuki,horic}. The values of
$\tau_{0,N}$ and $\tau_{2,N}$ are given in Table I.
As we already emphasized, each term
$(1/2) \sqrt{\tau_{L,4}/\tau_{L,6}}
\langle R_{4L}(r,\nu_N) | r^2 | R_{6L}(r,\nu_N) \rangle$
is all large comparable with single nucleon strength.
The monopole matrix elements between the $0_2^+$ and $0_3^+$
states and the $0_1^+$ state can be calculated by using
$\Phi_{(2,0)}^{cl}$ as follows
\begin{eqnarray}
&& M({\rm E0},0_2^+ - 0_1^+) = \langle 0_2^+ |
\Phi_{(2,0)}^{cl} \rangle
\langle \Phi_{(2,0)}^{cl} | O_M | \Phi_{CS} \rangle
= \langle 0_2^+ | \Phi_{(2,0)}^{cl} \rangle \sqrt{\frac{45}{32}}
\frac{1}{\nu_N}, \label{eq:SmonoI} \\
&& M({\rm E0},0_3^+ - 0_1^+) = \langle 0_3^+ |
\Phi_{(2,0)}^{cl} \rangle
\langle \Phi_{(2,0)}^{cl} | O_M | \Phi_{CS} \rangle
= \langle 0_3^+ | \Phi_{(2,0)}^{cl} \rangle \sqrt{\frac{45}{32}}
\frac{1}{\nu_N}. \label{eq:SmonoII}
\end{eqnarray}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Monopole transition from three-cluster state in $^{12}$C}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The calculation of the monopole transition from three-cluster state in
$^{12}$C can be made essentially in the same way as in the case of
two-cluster state. We explain this point by calculating the monopole
transition matrix element from the second $0^+$ state at $E_x = 7.66$ MeV
to the ground state.
The experimental data is $M({\rm E0},0_2^+ - 0_1^+)=5.4\pm0.2$ fm$^{2}$.
In the previous section we described the ground state ($0^+_1$) of
$^{12}$C by the SU(3) shell model wave function
$\phi_{L=0}({}^{12} {\rm C})$ which belongs to
the SU(3) irreducible representation $(\lambda,\mu)$ = (0,4).
This wave function is known of course to be a rather good approximation.
According to the Bayman-Bohr theorem the internal wave function of
the $^{12}$C ground state can be expressed in terms of
the $3\alpha$ cluster wave function,
\begin{eqnarray}
|0_1^+ \rangle &=& {|(0s)^4 (0p)^8 (0,4) J=0 \rangle}_{internal} \nonumber\\
&=& {\widehat N}_g
\sqrt{\frac{4!4!4!}{12!}} {\cal A} \{ {\widehat g}_{(04)0}
(\Vec{s},\Vec{t}) \phi(\alpha_1)\phi(\alpha_2)\phi(\alpha_3) \},
\end{eqnarray}
where $\Vec{s}$ and $\Vec{t}$ are the Jacobi coordinates defined by
\begin{equation}
\Vec{s} = \Vec{X}_2 - \Vec{X}_1,\ \ \ \Vec{t} = \Vec{X}_3 -
\frac{\Vec{X}_2 + \Vec{X}_1}{2}, \ \ \
\Vec{X}_k = \frac{1}{4} \sum_{i \in \alpha_k} \Vec{r}_i,
\end{equation}
and ${\cal A}$ is antisymmetrizer among nucleons belonging to different
$\alpha$ clusters.
The relative wave function ${\widehat g}_{(04)0}(\Vec{s},\Vec{t})$ is
expressed as follows
\begin{eqnarray}
{\widehat g}_{(04)0}(\Vec{s},\Vec{t}) &=& \sum_{L=0,2,4}
\langle (4,0)L, (4,0)L ||
(0,4) J=0 \rangle {\cal R}_{4,4,L}^{8,J=0}(\Vec{s},\Vec{t}), \\
{\cal R}^{N,J=0}_{N_1,N_2,L}(\Vec{s},\Vec{t}) &\equiv&
[{\cal R}_{N_1 L}(\Vec{s},2\nu_N)
{\cal R}_{N_2 L}(\Vec{t},\frac{8}{3}\nu_N)]_{J=0}, \ \ (N_1 + N_2 = N).
\label{eq:expb}
\end{eqnarray}
where ${\cal R}_{NL}(\Vec{u},\beta)$ stands for the harmonic oscillator
function of the size parameter
$\beta$ of the coordinate $\Vec{u}$ with the oscillator quantum number
$N$ and angular momentum $L$.
The SU(3) symmetry $(0,4)$ for $(0s)^4(0p)^8$ configuration is equivalent
to the spatial symmetry $[44]$ for $(0s)^4(0p)^8$ configuration.
Since there is only one state with $J=0$
for the $(0s)^4(0p)^8 [44]$ configuration, the following identities hold
(see also Appendix~\ref{Baymann_Bohr}),
\begin{eqnarray}
|0_1^+ \rangle &=& {\widehat N}_{g0} \sqrt{\frac{4!4!4!}{12!}} {\cal A}
\{ {\cal R}_{4,4,L=0}^{8,J=0}(\Vec{s},\Vec{t})
\phi(\alpha_1)\phi(\alpha_2)\phi(\alpha_3) \} \label{eq:c12a} \\
&=& {\widehat N}_{g2} \sqrt{\frac{4!4!4!}{12!}} {\cal A}
\{ {\cal R}_{4,4,L=2}^{8,J=0}(\Vec{s},\Vec{t})
\phi(\alpha_1)\phi(\alpha_2)\phi(\alpha_3) \} \label{eq:c12b} \\
&=& {\widehat N}_{g4} \sqrt{\frac{4!4!4!}{12!}} {\cal A}
\{ {\cal R}_{4,4,L=4}^{8,J=0}(\Vec{s},\Vec{t})
\phi(\alpha_1)\phi(\alpha_2)\phi(\alpha_3) \}. \label{eq:c12c}
\end{eqnarray}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsubsection{Monopole transition between the ground and Hoyle states}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The second $0^+$ state ($0^+_2$) is known to have $3\alpha$ structure
\cite{supple}, and so we express its wave function as
\begin{equation}
|0_2^+ \rangle = {\widehat N}_H \sqrt{\frac{4!4!4!}{12!}} {\cal A}
\{ {\widehat \chi}_H(\Vec{s},\Vec{t}) \phi(\alpha_1) \phi(\alpha_2)
\phi(\alpha_3) \}.
\end{equation}
In the expansion of the relative wave function
${\widehat \chi}_H(\Vec{s},\Vec{t})$
in terms of the harmonic oscillator wave functions, the number of the
total oscillator quanta of these oscillator wave functions is larger
than 8 which is the number of total oscillator quanta of relative wave
function of the ground state.
Just in the same manner as in the previous section, we can express
the monopole transition matrix element $M({\rm E0},0_2^+ - 0_1^+)$
as follows,
\begin{eqnarray}
&& M({\rm E0},0_2^+ - 0_1^+) = \langle 0_1^+ | \frac{1}{2} \sum_{i=1}^{12}
(\Vec{r}_i - \Vec{r}_G)^2 | 0_2^+ \rangle \nonumber \\
&& \ = \frac{1}{2} \sum_L \frac{{\widehat N}_H}
{{\widehat N}_{gL}}
\langle {\cal R}^{8,J=0}_{4,4,L}(\Vec{s},\Vec{t}) |
( 2 \Vec{s}^2 + \frac{8}{3} \Vec{t}^2 ) |
{\widehat \chi}_H(\Vec{s},\Vec{t}) \rangle. \label{eq:ezeroc}
\end{eqnarray}
Here we used the following relation,
\begin{eqnarray}
\sum_{i=1}^{12}(\Vec{r}_i - \Vec{r}_G)^2 = \sum_{k=1}^3
\sum_{i\in\alpha_k} (\Vec{r}_i - {\bf X}_k)^2 + 2 \Vec{s}^2
+ \frac{8}{3}\Vec{t}^2. \label{eq:sqrms_12C}
\end{eqnarray}
It is noted that the first term in Eq.~(\ref{eq:sqrms_12C}) does not
contribute to the monopole transition matrix element like in the
previous case of $^{16}$O.
The second $0^+$ state in $^{12}$C which is known as the Hoyle state
has been studied by many authors with 3$\alpha$ cluster model and its
structure is now regarded as being mainly composed of weakly
interacting $3\alpha$ clusters mutually in $S$-wave~\cite{supple,hori,kami,uega}.
Therefore we write ${\widehat \chi}_H(\Vec{s},\Vec{t})$ as follows
\begin{equation}
{\widehat \chi}_H(\Vec{s},\Vec{t}) = {\widetilde \chi}_H(s,t)
Y_{00}(\Vec{\hat s}) Y_{00}(\Vec{\hat t}).
\end{equation}
As we already mentioned, the expansion ${\widetilde \chi}_H(s,t)$
in terms of the harmonic oscillator function does not contain
the components whose numbers of oscillator quanta are less than or
equal to 8,
\begin{eqnarray}
{\widetilde \chi}_H(s,t) &=& \sum_{N_1,N_2} D_{H,N_1,N_2}
R_{N_1,0}(s,2\nu_N)
R_{N_2,0}(t,\frac{8}{3}\nu_N), \label{eq:hoexp} \\
D_{H,N_1,N_2} &=& 0 \ \ \ {\rm for} \ \ N_1+N_2 \le 8.
\end{eqnarray}
Substituting Eq.~(\ref{eq:hoexp}) into Eq.~(\ref{eq:ezeroc}), we have
the following simple result
\begin{equation}
M({\rm E0},0_2^+ - 0_1^+) = \frac{1}{2} \frac{{\widehat N}_H}{{\widehat N}_{g0}}
( D_{H,6,4} + D_{H,4,6} ) \langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N)
\rangle.
\label{eq:monopole_Hoyle}
\end{equation}
This analytical expression explains why $M({\rm E0},0_2^+ - 0_1^+)$
has a comparable magnitude as the single nucleon E0 matrix element.
The factor $\langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle$,
which appears also in the case of $^{16}$O,
is a few times larger than the single nucleon E0 matrix elements
$\langle R_{00}(r,\nu_N) | r^2 | R_{20}(r,\nu_N) \rangle$ and
$\langle R_{11}(r,\nu_N) | r^2 | R_{31}(r,\nu_N) \rangle$, while the
other factors work to make the E0 value smaller. The reason why
we have this formula of Eq.~(\ref{eq:monopole_Hoyle}) is just the
3$\alpha$ clustering character of the shell model wave function
of the $^{12}$C ground state which is described by the Bayman-Bohr
theorem.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsubsection{Description of the Hoyle state as a 3$\alpha$ condensate}
\label{subsection:hoyle_state}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
Recently the structure of the Hoyle state has been studied from a new
point of view that this state is the Bose-condensed state of $3\alpha$
particles~\cite{tohsaki,yamada,funaki}.
It has been demonstrated that both of the 3$\alpha$ wave functions of
Refs.~\cite{kami} and \cite{uega} which are the full solutions of
3$\alpha$ Resonating Group Method (RGM) equation of motion have large
overlaps close to 100 \%
with the 3$\alpha$ Bose-condensed wave functions~\cite{funaki}.
Therefore we here adopt as ${\widehat \chi}_H(\Vec{s},\Vec{t})$ the
following form
\begin{eqnarray}
{\widehat \chi}_H(\Vec{s},\Vec{t}) &=& ( 1 - P )\
{\widehat \chi}_{HG}(\Vec{s},\Vec{t}), \\
{\widehat \chi}_{HG}(\Vec{s},\Vec{t}) &\equiv&
\left( \frac{8\gamma}{\sqrt{3} \pi} \right)^{\frac{3}{2}}
\exp \{ -4\gamma \sum_{k=1}^3 ({\bf X}_k - \Vec{r}_G)^2 \} \nonumber \\
&=& \left( \frac{4\gamma}{\pi} \right)^{\frac{3}{4}}
\left(\frac{16\gamma}{3\pi}\right)^{\frac{3}{4}}
\exp \{ - \gamma ( 2\Vec{s}^2 + \frac{8}{3} \Vec{t}^2 ) \}, \\
P &\equiv& \sum_{N \le 8} \sum_{N_1+N_2=N} \sum_L
|{\cal R}^{N,J=0}_{N_1,N_2,L}(\Vec{s},\Vec{t}) \rangle \langle
{\cal R}^{N,J=0}_{N_1,N_2,L}(\Vec{s},\Vec{t})|,
\end{eqnarray}
where $\gamma$ denotes the width parameter which characterizes the
$3\alpha$ condensate wave function.
$P$ is the projection operator onto the state of SU(3) relative motion
of the ground state and the states forbidden by the antisymmetrization.
Then, the analytical expression of the monopole transition matrix element
in Eq.~(\ref{eq:monopole_Hoyle}) is given as follows:
\begin{eqnarray}
M({\rm E0},0_2^+ - 0_1^+) &=& \sqrt{\frac{7}{6}} \sqrt{\frac{\langle F_4 \rangle}
{\langle F_5 \rangle}} \xi_5 \langle R_{40}(r,\nu_N) | r^2 |
R_{60}(r,\nu_N) \rangle, \label{eq:form3} \\
\xi_5 &\equiv& \sqrt{\frac{\langle F_5 \rangle}{\langle F_5 \rangle +
\sum_{n=6}^\infty \displaystyle{ \left( \frac{\nu_N-\gamma}
{\nu_N+\gamma} \right)^{2(n-5)} }
\langle F_n \rangle}},
\end{eqnarray}
where the definitions of $F_n$ and $\langle F_n \rangle$ are
\begin{eqnarray}
&& {\widehat \chi}_{H}(\Vec{s},\Vec{t}) = \left( \frac{2\sqrt{\nu_N\gamma}}
{\nu_N+\gamma} \right)^3 \sum_{n=5}^\infty \left( \frac{\nu_N-\gamma}
{\nu_N+\gamma} \right)^n F_n(\Vec{s},\Vec{t}), \\
&& F_n(\Vec{s},\Vec{t}) = \sum_{n_1+n_2=n}
\sqrt{\frac{(2n_1+1)!!(2n_2+1)!!} {(2n_1)!!(2n_2)!!}}
{\cal R}^{2n,J=0}_{{2n_1},{2n_2},L=0}(\Vec{s},\Vec{t}), \\
&& \langle F_n \rangle = \langle F_n(\Vec{s},\Vec{t}) \phi(\alpha_1)
\phi(\alpha_2) \phi(\alpha_3) | {\cal A} \{ F_n(\Vec{s},\Vec{t})
\phi(\alpha_1) \phi(\alpha_2) \phi(\alpha_3) \} \rangle.
\end{eqnarray}
We see that the dependence of $M({\rm E0},0_2^+ - 0_1^+)$ on the parameter
$\gamma$ is contained only in the factor $\xi_5$.
The derivation of the above analytical expression of $M({\rm E0},0_2^+ - 0_1^+)$
is given in Appendix~\ref{Hoyle}.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Ground State Correlations}\label{gr-st-correl}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
We first study in this section the numerical values of the monopole
matrix elements in $^{16}$O and $^{12}$C, by using the formulae
obtained in the previous section. As is expected from the analytical
forms, the numerical values are shown to have the same order of
magnitude as the observed values which are comparable with the
single nucleon strength. However, the calculated values are
found to be smaller by a few times in $^{16}$O and by several
times in $^{12}$C. Therefore we next study in this section
the effect of the ground state correlation on the magnitude
of the monopole matrix elements. We will see that the ground state
correlation largely improves the reproduction of the observed values
up to the factor of magnitude. The ground state correlation we consider
is due to the activation of the clustering degree of freedom which
is described by the Bayman-Bohr theorem.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Monopole transition matrix elements in $^{16}$O}\label{results:16O}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
We calculate $M({\rm E0},0_2^+ - 0_1^+)$ and $M({\rm E0},0_3^+ - 0_1^+)$ in $^{16}$O by
using their analytical expressions given in Eqs.~(\ref{eq:form1}) and
(\ref{eq:form2}). The nucleon size parameter $\nu_N$
($\nu_N=0.151$ fm$^{-2}$) is chosen so as to reproduce the experimental rms
radius of $^{16}$O whose wave function is described with the doubly closed
shell configuration $\Phi_{CS}$. We also use the expressions given in
Eq.~(\ref{eq:mono0}) and (\ref{eq:mono2}). The realistic values of
$\eta_6$ and $\zeta_6$, of course, should be obtained from the structure
calculation. A representative structure calculation is that of
Ref.~\cite{suzuki} in which the $^{12}$C + $\alpha$ OCM was adopted. For
the sake of the study of this paper, we repeated the same calculation as
Ref.~\cite{suzuki}. According to the results, the $0^+_2$ wave function of
$^{16}$O has the predominant component of $^{12}$C($0^+$)+$\alpha$ channel
and small components of $^{12}$C($2^+$)+$\alpha$ and $^{12}$C($4^+$)+$\alpha$
channels, as mentioned already. Similarly, the $0^+_3$ wave function of
$^{16}$O has the large component of $^{12}$C($2^+$)+$\alpha$ channel and small
components of $^{12}$C($0^+$)+$\alpha$ and $^{12}$C($4^+$)+$\alpha$ channels.
By using the results in Table I of the first paper of Ref.~\cite{suzuki}, we
get approximate values for $\eta_6$ and $\zeta_6$, and they are
$\sqrt{0.144}=0.379$ and $\sqrt{0.325}=0.570$, respectively. We explain
the details in Appendix~\ref{suzuki}. Then, the monopole matrix elements
are estimated to be
\begin{eqnarray}
M({\rm E0},0_2^+ - 0_1^+) &=& 0.379 \times \frac{0.784}{\nu_N} = 1.97~{\rm fm}^2,
\label{eq:result1} \\
M({\rm E0},0_3^+ - 0_1^+) &=& 0.570 \times \frac{1.03}{\nu_N} = 3.88~{\rm fm}^2,
\label{eq:result2}
\end{eqnarray}
where the formulae in Eqs.~(\ref{eq:mono0}) and (\ref{eq:mono2}) are used.
The result for $M({\rm E0},0_2^+ - 0_1^+)$ gives about 60~\% of the experimental
value, and the result for $M({\rm E0},0_3^+ - 0_1^+)$ amounts to 96~\% of the
experimental one.
In estimating the above values for $M({\rm E0},0_2^+ - 0_1^+)$ and $M({\rm E0},0_3^+ - 0_1^+)$,
we used Eqs.~(\ref{eq:form1}) and (\ref{eq:form2}), respectively.
These formulae are based on the assumption that the wave functions of the
$0_2^+$ and $0_3^+$ states are purely of the structure of $^{12}$C($0^+$) +
$\alpha$ and $^{12}$C($2^+$) + $\alpha$, respectively. However, of course,
the OCM calculation of Ref.~\cite{suzuki} shows the small contamination
of other channel components than the dominant component. So we here also
give values which take into account these small contamination of other
non-dominant channel components. Such values can be obtained by using the
formulae given in Eqs.~(\ref{eq:SmonoI}) and (\ref{eq:SmonoII}).
The values $\langle 0_2^+ | \Phi_{(2,0)}^{cl} \rangle $ and
$\langle 0_3^+ | \Phi_{(2,0)}^{cl} \rangle $ are given in Table I of the
second paper of Ref.~\cite{suzuki}, and they are 0.171 and 0.321,
respectively. Then, the monopole matrix elements are estimated to be
\begin{eqnarray}
M({\rm E0},0_2^+ - 0_1^+) &=& 0.171 \times \frac{1.19}{\nu_N} = 1.35~{\rm fm}^2,
\label{eq:result3} \\
M({\rm E0},0_3^+ - 0_1^+) &=& 0.321 \times \frac{1.19}{\nu_N} = 2.53~{\rm fm}^2.
\label{eq:result4}
\end{eqnarray}
Now this result for $M({\rm E0},0_2^+ - 0_1^+)$ gives about 38~\% of the
experimental value, while the result for $M({\rm E0},0_3^+ - 0_1^+)$ gives about
63~\% of the experimental one. These values are smaller than the former
values in Eqs.~(\ref{eq:result1}) and (\ref{eq:result2}), but still they
are different from the observed values only by a few factors. The reason
why the former values in Eqs.~(\ref{eq:result1}) and (\ref{eq:result2}) are
larger than the corresponding present values is that the magnitudes of the
$\Phi_{(2,0)}^{cl}$ component in the pure channel wave functions,
$^{12}$C($0^+$) + $\alpha$ for $0_2^+$ and $^{12}$C($2^+$) + $\alpha$ for
$0_3^+$, are larger than those of the OCM wave functions for the $0_2^+$
and $0_3^+$ states, respectively. We give detailed explanation in the
Appendix~\ref{suzuki}.
As we see above the calculated values of the monopole matrix elements
surely reproduce the order of magnitude of the observed values which are
comparable with single nucleon strength. But when compared with the
data in detail, they are a few times smaller than the observed values.
We below show that when we take into account the ground state correlation
theoretical values are improved as to attain the reproduction of the data
within 10 $\sim$ 20 \% accuracy.
The ground state correlation we consider is the one caused by the activation
of the clustering degree of freedom described by Bayman-Bohr theorem.
In previous sections we demonstrated that the clustering degree of freedom
described by Bayman-Bohr theorem is the very reason why the monopole
strengths of excited cluster states are so large as to be comparable with
single nucleon strength. However, we only considered the clustering degree
of freedom rather in a static way. Namely we did not consider the dynamical
effect of the clustering degree of freedom which excites the ground state
configuration toward including higher quantum configurations. We know that
the clustering degree of freedom described by Bayman-Bohr theorem has the
physical reality because we observe many excited cluster states which are
formed by exciting the clustering degree of freedom embedded in the ground
state. Therefore taking into account the ground state correlation caused
by the clustering degree of freedom described by Bayman-Bohr theorem is
very natural and should be studied.
In order to study the effect of the ground state correlation we make use
of the $^{12}$C + $\alpha$ OCM calculation.
We repeat the same calculation as Ref.~\cite{suzuki}.
Of course we adopt the same effective nuclear force.
We express by $|0_2^+ \rangle$ and $|0_3^+ \rangle$ the obtained
OCM wave functions of the $0_2^+$ and $0_3^+$ states, respectively.
Next, by $|0_1^+; N \rangle$ we express the wave function of the ground state
($0_1^+$) which is calculated within the limited cluster model space
where the highest number of the total oscillator quanta of the basis
wave function is $N$ ($N$ = 4, 6, $\cdots$, 30 ).
The wave function $|0_1^+; N=4 \rangle$ is just the closed shell wave function
without any ground state correlation.
As $N$ becomes larger the wave function $|0_1^+; N \rangle$ contains
more amount of ground state correlation.
Since $|0_2^+ \rangle$ and $|0_3^+ \rangle$ are not orthogonal to
$|0_1^+; N \rangle$, we construct the orthogonalized wave functions
as follows
\begin{eqnarray}
|\widetilde{0_k^+} \rangle = N_k (1 - |0_1^+; N \rangle
\langle 0_1^+; N| ) |0_k^+ \rangle, \ \ ( k=2, 3),
\end{eqnarray}
where $N_k$ is the normalization constant. We calculate the monopole matrix
element between $|\widetilde{0_k^+} \rangle$ and $|0_1^+; N \rangle$
\begin{eqnarray}
M_N({\rm E0}, 0_1^+ - 0_k^+) = \langle 0_1^+; N | \frac{1}{2} \sum_{i=1}^{16}
( \Vec{r}_i - \Vec{r}_G )^2 |\widetilde{0_k^+} \rangle,
\label{eq:monopole_N_dependence}
\end{eqnarray}
and study the dependence on $N$ of this quantity.
Figure~\ref{fig:1} shows the calculated results of $M_N({\rm E0}, 0_1^+ - 0_k^+), ( k=2, 3)$
as a function $N$.
The values of $M_N({\rm E0}, 0_1^+ - 0_k^+)$ for $N=4$ are the monopole matrix
elements without ground state correlation which we discussed in detail in the above.
They are $M_{N=4}({\rm E0}, 0_1^+ - 0_2^+)$ = 1.39 fm$^2$ and
$M_{N=4}({\rm E0}, 0_1^+ - 0_3^+)$ = 2.36 fm$^2$.
These values are very close to the values given in Eqs.~(\ref{eq:result3}) and (\ref{eq:result4})
for which the normalization constants $N_k$ are not accounted.
We clearly see that the values of $M_N({\rm E0}, 0_1^+ - 0_k^+)$ grow almost monotonously as $N$
becomes larger.
At $N$ = 30, the values are already converged, and they are
$M_{N=30}({\rm E0}, 0_1^+ - 0_2^+)$ = 4.01 fm$^2$ and $M_{N=30}({\rm E0}, 0_1^+ - 0_3^+)$ = 3.56 fm$^2$.
These converged values are very close to the observed values, and the difference
from the observed values is only within 13~\%.
Thus we have shown that by taking into account the ground state correlation
theoretical values are improved as to attain the reproduction of the data
within a factor of $1.13$.
An important reason why the ground state correlation enhances the monopole strengths
is explained as follows.
We study the deviation of the ground-state wave function $|0^{+}_{1} \rangle$ obtained
by the $^{12}$C+$\alpha$ OCM from the doubly closed shell wave function.
For this purpose we define a modified doubly closed shell model wave function
$\Phi_{0^+}(\beta)$ ($\beta$ denoting the size parameter of the $^{12}$C+$\alpha$
relative wave function)
and calculate the squared overlap of it
with the OCM ground state wave function
$|0^{+}_{1} \rangle$ obtained with the full model space,
\begin{eqnarray}
&& {\Phi_{0^+}(\beta)}
= N_{g0}(\beta) \frac{1}{\sqrt{{}_{16}C_4}}
{\cal A} \{ [{\cal R}_{40}(\Vec{r},\beta)
\phi_{L=0}({}^{12}{\rm C})]_{J=0} \phi(\alpha) \}, \label{eq:clshla_beta} \\
&& P(\beta/(3\nu_{N})) = \left| \langle {\Phi_{0^+}(\beta)} | {0^+_1} \rangle \right|^{2},
\label{eq:overlap}
\end{eqnarray}
where $N_{g0}(\beta)$ is the normalization constant.
When $\beta$ is equal to $3\nu_{N}$, the wave function $\Phi_{0^+_1}(\beta=3\nu_{N})$
is equivalent to the doubly closed shell model wave function in Eq.~(\ref{eq:clshla}), which
originally has the $\alpha$ cluster degree of freedom or sort of like a seed of $\alpha$
clustering, as discussed in Sec.~\ref{subsection:Bayman_Bohr}.
For $\beta < 3\nu_{N}$, ${\Phi_{0^+}(\beta)}$ expresses a wave function
in which the $^{12}$C+$\alpha$ relative motion or
the seed of $\alpha$ clustering is swollen in comparison with those in the doubly closed shell
model wave function.
Thus, the study of the dependence of the squared overlap $P$ on
the parameter $\beta/(3\nu_{N})$ gives a rough indication on the degree of the deviation
from the doubly closed wave function, i.e.~the degree of $\alpha$ clustering activated
in the ground state.
We found that $P$ has the maximum value of 0.958 at $\beta/(3\nu_{N})=0.847$
[vs.~$P=0.890$ at $\beta/(3\nu_{N})=1$].
The result of $\beta/(3\nu_{N})=0.847$ means that in the ground state wave function
$|0^{+}_{1} \rangle$ the $^{12}$C+$\alpha$ relative motion or the seed of $\alpha$
clustering is definitely swollen in comparison with those in the doubly closed shell model wave function.
Thus the ground state correlation makes the structure of the ground state closer
to the $^{12}$C+$\alpha$ cluster structure in the $0^{+}_{2}$ and $0^{+}_{3}$ states,
and then the monopole strengths become significantly larger in comparison with those
with no ground state correlation.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\subsection{Monopole transition matrix elements in $^{12}$C}\label{results:12C}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The analytical expression of the monopole transition matrix element
$M({\rm E0},0^+_1-0^+_2)$ is demonstrated in Eq.~(\ref{eq:form3}), which
depends on the nucleon size parameter $\nu_N$ and width parameter of
the Hoyle state $\gamma$. The expression of the monopole matrix element
consists of three parts like the case of $^{16}$O, and the dominant
part is the radial integral referring the relative motions among three
$\alpha$ clusters,
${\langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle}
= \sqrt{21/8}/\nu_N$. The strength of the radial integral is a few times
larger than the single particle monopole strengths,
$\langle 0s |r^2| 1s \rangle$ and $\langle 0p |r^2| 1p \rangle$.
In the present study we use the value $\nu_N=0.168\ {\rm fm}^{-2}$, which
reproduces the observed rms radius of $^{12}$C with the SU(3) shell model
wave function in Eqs.~(\ref{eq:c12a})$\sim$(\ref{eq:c12c}).
The monopole matrix element in Eq.~(\ref{eq:form3}) which is expressed
as $M({\rm E0},0_2^+ - 0_1^+)=\xi_5\times 0.882/\nu_N$ ${\rm fm}^2$ contains
$\xi_5$ which depends on $\gamma$.
In Table~\ref{tab:2}, we display $\xi_5$ and $M({\rm E0},0_2^+ - 0_1^+)$ calculated
at several $\gamma$ values. According to Ref.~\cite{funaki}, we should use
the value of $\gamma \approx 0.018$ fm$^{-2}$. With this value of $\gamma$,
the Bose-condensate wave function has a very large (almost 100~\%) overlap
with the full solution of the $3\alpha$ RGM for the Hoyle state and gives
an rms radius 3.8 fm for the Hoyle state.
For $\gamma$ = 0.018 fm$^{-2}$, we obtain $\xi_5=0.191$, which is relatively
small in comparison with $\eta_6=0.379$ and $\zeta_6=0.570$ in the case of
$^{16}$O in Sec.~\ref{results:16O}. This leads to
\begin{eqnarray}
M({\rm E0},0_2^+ - 0_1^+) = 0.19 \times \frac{0.882}{\nu_N} = 1.3 {\rm fm}^2,
\end{eqnarray}
which is of the same order of magnitude as the observed
value ($5.4 \pm 0.2$ ${\rm fm}^2$~\cite{ajzen86}) but reproduces only about
25 \% in comparison with that. This value of about 25 \% is a little bit
smaller in contrast to that of the $^{16}$O case (see the previous subsection
\S\ref{results:16O}) in which our simple estimates are larger than about
40\% of the experimental data.
We should note that in more realistic situation the description of the ground
state adopted here for $^{12}$C using the SU(3) shell model is not necessarily
good and a deviation from the SU(3) shell model representation should be
taken into account~\cite{supple}.
According to the structure study of $^{12}$C with the $3\alpha$ orthogonality
condition model (OCM)~\cite{yamada}, the SU(3)$(\lambda,\mu)=(0,4)$ component
with the lowest oscillator quantum ($N_{TOT}=8$) is only about
60~\% in the ground state.
%Fig.2 (a) demonstrates the
%distribution of the harmonic oscillator basis component of the ground state
%obtained with the $3\alpha$ orthogonality condition model (OCM)~\cite{yamada}.
%We see that the SU(3)$(\lambda,\mu)=(0,4)$ component with the lowest
%oscillator quantum ($N_{TOT}=8$) is only about 60~\% in the ground state.
The smallness of the SU(3)$(\lambda,\mu)=(0,4)$ component with the lowest
oscillator quantum ($N_{TOT}=8$) is in contrast to the $^{16}$O case, in
which the ground state is well described by the doubly closed shell model
wave function $(0s)^4(0p)^{12}$:~The SU(3)$(\lambda,\mu)=(0,0)$ component
with the lowest quantum ($N_{TOT}=12$) of the $^{16}$O ground state is as
large as about 90\%~\cite{suzuki}.
Here we demonstrate the effect of the ground state correlation to the
monopole matrix element by adopting the following wave function for the
ground state~\cite{funaki}:
\begin{equation}
\Psi_G({\widetilde \gamma},\nu_N)={\cal N}_G \sqrt{\frac{4!4!4!}{12!}}
{\cal A}[\exp\{-{\widetilde \gamma}(2\Vec{s}^2+\frac{8}{3}\Vec{t}^2) \}
\phi(\alpha_1)\phi(\alpha_2)\phi(\alpha_3)],\label{eq:Psi_G}
\end{equation}
where ${\cal N}_G$ is normalization constant. This wave function is
called THSR wave function and depends on two
parameters ${\widetilde \gamma}$ and $\nu_N$. We impose the condition
that this wave function reproduces the observed rms radius, 2.47 fm, of
the ground state. Then, the ratio ${\widetilde \gamma}/\nu_N$ is the only
parameter which describes the property of the ground state.
It is noted that $\Psi_G({\widetilde \gamma},\nu_N)$ with ${\widetilde \gamma}/\nu_N=1$
agrees with the SU(3) $(\lambda,\mu)=(0,4)$ shell model wave function
in Eqs.~(\ref{eq:c12a})$\sim$(\ref{eq:c12c}), demonstrating directly
that the wave function originally has sort of like a seed of $3\alpha$
clustering.
Taking the ${\widetilde \gamma}$ value a little smaller than $\nu_N$,
$\Psi_G({\widetilde \gamma},\nu_N)$ deviates from the SU(3) shell model
wave function and the slightly more relaxed spatial clustering of
$3\alpha$ clusters than the SU(3) wave function is induced in the ground
state, i.e.~$\Psi_G({\widetilde \gamma},\nu_N)$ expresses sort of like
a generalized SU(3) wave function in which the seed of $3\alpha$
clustering are slightly swollen in comparison with the original SU(3) wave function.
This is the ground state correlation taking into account here,
which is similar to the case in $^{16}$O as discussed in Sec.~\ref{results:16O}.
In Ref.~\cite{funaki}, it is reported that the ground
state wave function of the $3\alpha$ RGM calculation of
Refs.~\cite{kami} and \cite{uega} can be well approximated by this kind of
wave function.
The amount of the $3\alpha$-like ground state correlation,
thus, can be characterized by the ratio ${\widetilde \gamma}/\nu_N$,
which should be less than or equal to unity.
In the $3\alpha$ cluster model~\cite{supple,funaki,kami,uega,yamada},
the nucleon size parameter $\nu_N$ is usually chosen to reproduce
the rms radius of $\alpha$ cluster, $\nu_N=0.275~{\rm fm}^{-2}$ which
is larger than that for the SU(3) shell
model wave function ($\nu_N=0.168~{\rm fm}^{-2}$) shown above.
The estimation of ${\widetilde \gamma}/\nu_N$ for the ground state wave
function given by Ref.~\cite{funaki} is as small as
${\widetilde \gamma}/\nu_N\sim0.29$.
This small value indicates that
the ground state of $^{12}$C has a significant amount of the $3\alpha$
correlation.
Below we change the value of the parameter
${\widetilde \gamma}/\nu_N$ from $1.0$ down to $0.27$.
The wave function of the Hoyle state is constructed so as to be orthogonal
to the ground state wave function $\Psi_G$ in Eq.~(\ref{eq:Psi_G}) and
is given as follows:
\begin{eqnarray}
&&\Psi_H(\gamma,{\widetilde \gamma},\nu_N)={\cal N}_H (1-P)
\sqrt{\frac{4!4!4!}{12!}} {\cal A} [\exp \{ -\gamma (2\Vec{s}^2 +
\frac{8}{3}\Vec{t}^2) \}\phi(\alpha_1)\phi(\alpha_2)\phi(\alpha_3)],
\label{eq:phi_H_modify}\\
&& P \equiv | \Psi_G({\widetilde \gamma},\nu_N) \rangle \langle
\Psi_G({\widetilde \gamma},\nu_N) |,
\end{eqnarray}
where ${\cal N}_H$ is normalization constant.
The width parameter $\gamma$ is determined so as to reproduce the rms
radius of the Hoyle state, 3.8 fm. The Hoyle state of $^{12}$C ($0^+_2$)
is known to have a dilute $3\alpha$ condensate structure with the nuclear
radius of about 4 fm.
%The spatially extended $3\alpha$ cluster structure
%or $3\alpha$-gas-like structure of the Hoyle state can be seen in
%Fig.2 (b), where we demonstrate the distribution of the component
%of harmonic oscillator basis in the Hoyle-state wave function which is
%obtained by $3\alpha$ OCM calculation~\cite{yamada}.
%We see that harmonic oscillator bases up to about 100 quanta are needed
%to describe the $3\alpha$-gas-like structure of the Hoyle state.
This exotic structure of the Hoyle state was found to be described
simply~\cite{funaki} with a single $\alpha$-condensate wave function
given in Eq.~(\ref{eq:phi_H_modify}). The monopole matrix element given as
\begin{eqnarray}
{M({\rm E0},0^+_1-0^+_2})=
{\langle \Psi_G({\widetilde \gamma},\nu_N) | \frac{1}{2}\sum_{i=1}^{12}
(\Vec{r}_i-\Vec{r}_G)^2 | \Psi_H(\gamma,{\widetilde \gamma},\nu_N)
\rangle},
\label{eq:E0_HG}
\end{eqnarray}
depending only on the parameter ${\widetilde \gamma}/\nu_N$.
Table~\ref{tab:3} shows the values of the monopole matrix elements
[Eq.~(\ref{eq:E0_HG})] calculated at several ${\widetilde \gamma}/\nu_N$
values. We see that the monopole matrix element increases as the ratio
${\widetilde \gamma}/\nu_N$ decreases from unity, namely as the
$3\alpha$-like correlation becomes stronger in the ground state.
This can be reasonably understood from the fact that the ground state wave
function $\Psi_G$ with stronger $3\alpha$-like correlation has larger
$3\alpha$-cluster component which makes larger the overlap with the Hoyle
state wave function $\Psi_H$ with the dilute $3\alpha$ cluster structure,
and then the monopole matrix element becomes larger.
At the value of ${\widetilde \gamma}/\nu_N\sim0.27$, the monopole matrix
element is about $4.0$ fm$^2$, which is about three times larger than that
for ${\widetilde \gamma}/\nu_N=1$, and is closer to the observed value
$5.4\pm0.2$ fm$^2$. It is noted that ${\widetilde \gamma}/\nu_N\sim0.27$
gives the nucleon size parameter $\nu_N\sim0.26~{\rm fm}^{-2}$ which
corresponds to the value used usually in the microscopic $3\alpha$
cluster model calculations~\cite{supple,funaki,kami,uega,yamada}.
Without the ground state correlation the calculated monopole
value is smaller than the observed value by a factor of 4.15 but
now with inclusion of the ground state correlation the calculated monopole
value changed to be smaller only by a factor of 1.35 than the observed
value.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Discussions and Summary}\label{summary}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The monopole transitions from cluster states to ground states in light
nuclei are rather large which is comparable with the single particle strength.
The single particle estimate of the monopole transition is based on
the assumption that the excited state has a one-particle one-hole excitation
from the ground state.
However, the cluster structure is very different
from the shell-model-like structure of the ground state, and its state is
described as a superposition of many-particle many-hole configurations when
it is expanded by shell model configurations.
This means that in the excited state with a cluster structure,
the component of a one-particle one-hole excitation from the ground state
configuration is expected to be very small.
Therefore the observation of rather large monopole strengths
for cluster states which are comparable with single particle strength
looks {\it not} to be easy to explain.
Under this kind of understanding it has been often regarded that the monopole
transition occurs through the mixing of shell model wave function
$|\rm{shell} \rangle$ and the cluster model wave function $|\rm{cluster} \rangle$
\begin{eqnarray}
&& |\rm{Ground} \rangle = \alpha |\rm{shell} \rangle + \beta
|\rm{cluster} \rangle, \\
&& |\rm{Excited} \rangle = -\beta |\rm{shell} \rangle + \alpha
|\rm{cluster} \rangle.
\end{eqnarray}
Since it is assumed that the monopole operator $O_M$ does not connect
$|\rm{cluster} \rangle$ and $|\rm{shell} \rangle$,
$\langle \rm{cluster} | {\it O_M} | \rm{shell} \rangle$ = 0,
the monopole matrix element is considered to come
from the diagonal matrix elements (for example see Ref.~\cite{bertsch}),
\begin{eqnarray}
\langle \rm{Excited} | {\it O_M} |\rm{Ground} \rangle = \alpha \beta (
\langle \rm{cluster} | {\it O_M} |\rm{cluster} \rangle - \langle
\rm{shell} | {\it O_M} |\rm{shell} \rangle ).
\end{eqnarray}
Our explanation of the strong monopole transition between ground state and
excited cluster states is quite different from this explanation. We
insist that the order of magnitude of the strong monopole transition is
given by the matrix element $\langle \rm{cluster} | {\it O_M} | \rm{shell}
\rangle \ne$ 0,
\begin{eqnarray}
\langle \rm{Excited} | {\it O_M} |\rm{Ground} \rangle \approx
\langle \rm{cluster} | {\it O_M} |\rm{shell} \rangle, \ ({\rm for}
\ {\rm order} \ {\rm of} \ {\rm magnitude}).
\end{eqnarray}
Our argument is based on the Bayman-Bohr theorem which says that the SU(3)
shell-model wave function which describes rather well the structure of the
ground state of light nuclei is equivalent in most cases to cluster model wave
function.
The implication of this theorem is that the clustering degree of
freedom is already embedded even in the shell model wave function.
In the present study the monopole excitation of the ground state to cluster
states is understood as just the excitation of the inter-cluster relative motion in the
ground state to the inter-cluster relative motion in excited cluster states.
This resembles the monopole excitation of the single nucleon motion.
Our understanding was explicitly shown to be true by deriving the analytical
expressions of the monopole matrix elements.
In this paper we analyzed the monopole transitions in $^{16}$O between the
ground state and $^{12}$C + $\alpha$ cluster states ($0^+_2$ and $0^+_3$)
together with the one in $^{12}$C between the ground state and $3\alpha$
cluster state ($0^+_2$: Hoyle state). According to the Bayman-Bohr theorem,
the doubly closed shell model wave function of $^{16}$O, which has the
SU(3)($\lambda \mu$)=(00) symmetry and total quanta $N_{\rm TOT}=12$, is
equivalent to the $^{12}$C ($0^+$) + $\alpha$ cluster wave function as
well as $^{12}$C ($2^+$) + $\alpha$ with orbital angular momentum of
the inter-cluster relative motion $L_r=0$ and $2$, respectively.
The number of oscillator quanta of the inter-cluster relative motion
$N_r$ is 4. Similarly, the ground state wave function of $^{12}$C with
SU(3) ($\lambda\mu$) = (04) and $N_{\rm TOT}=8$ is equivalent to the
$3\alpha$ cluster wave function. The number of oscillator quanta of the
inter-cluster relative motion with respect to each of two Jacobi
coordinates is $N_r$ = 4. On the other hand, the $0^+_2$ [$0^+_3$]
state of $^{16}$O has a $^{12}$C($0^+$)+$\alpha$ cluster structure
[$^{12}$C($2^+$)+$\alpha$] with the relative orbital angular momentum
$L_r=0$ [$L_r=2$]. The $0^+_2$ state of $^{12}$C has a $3\alpha$ cluster
structure with $S$-wave relative angular momenta referring to two
Jacobi coordinates for $3\alpha$ clusters.
The analytical expressions of the monopole matrix elements we derived
for the above transitions in $^{16}$O and $^{12}$C are composed of three
factors. The most important factor is the radial integrals with harmonic
oscillator wave functions $\langle N_r=6,L_r | r^2 | N_r=4,L_r \rangle$
with $L_r=0$ or 2. The values of these integrals are a few times larger
than the single particle monopole transition matrix elements in {\it p}-shell
nuclei, ${\langle 1s | r^2 | 0s \rangle}$ =
${\langle N_r=2,L_r=0 | r^2 | N_r=0, L_r=0 \rangle}$ and
${\langle 1p | r^2 | 0p \rangle}$ =
${\langle N_r=3,L_r=1 | r^2 | N_r=1,L_r=1\rangle}$.
The second factor is
the amplitude (not squared amplitude) of the $2\hbar \omega$ - excited
harmonic oscillator wave function in the cluster states which is denoted as
$\eta_6$ for $M({\rm E0},0^+_1-0^+_2)$ and $\zeta_6$ for $M({\rm E0},0^+_1-0^+_3)$
in $^{16}$O and $\xi_5$ for $M({\rm E0},0^+_1-0^+_2)$ in $^{12}$C. They are
not so small; $\eta_6$ = 0.38, $\zeta_6$ = 0.57, and $\xi_5$ = 0.19.
The third factor is due to the antisymmetrization among nucleons, which
is denoted as $\sqrt{\tau_{0,4}/\tau_{0,6}}$ or
$\sqrt{\tau_{2,4}/\tau_{2,6}}$ in $^{16}$O and
$\sqrt{\langle F_4 \rangle / \langle F_5 \rangle}$ in
$^{12}$C. Since the quantities with strong antisymmetrization effect
are contained in the form of ratio, the third factor has magnitude close
to unity. As is expected from the analytical expressions, the calculated
numerical values of the monopole matrix elements were shown to have the same
order of magnitude as the observed values which are comparable with the
single nucleon strength.
Although the calculated values of the monopole matrix elements without
ground state correlation surely reproduce the order of magnitude of the
observed values, when compared with the data in detail, they are a few times
smaller than the observed values. In the case of $^{16}$O, two kinds of
theoretical values of $M({\rm E0},0^+_1-0^+_2)$ are 60 \% and 38 \% of the
observed values, respectively,
while those of $M({\rm E0},0^+_1-0^+_3)$ are 96 \% and 63 \%.
In the case of $^{12}$C, theoretical value of $M({\rm E0},0^+_1-0^+_2)$ is 25 \%
of the observed value. Therefore we next investigated the effect of the
ground state correlation to the monopole matrix elements.
The ground state correlation we considered was the one caused by the
activation of the clustering degree of freedom described by Bayman-Bohr
theorem. In the calculation of the monopole strength without ground state
correlation, we only considered the clustering degree of freedom rather
in a static way. Namely we did not consider the dynamical effect of the
clustering degree of freedom which excites the ground state configuration
toward including higher quantum configurations. We know that the
clustering degree of freedom described by Bayman-Bohr theorem has the
physical reality because we observe many excited cluster states which are
formed by exciting the clustering degree of freedom embedded in the ground
state. Therefore taking into account the ground state correlation caused
by the clustering degree of freedom described by Bayman-Bohr theorem is
very natural and should be studied.
The investigation of the effect of the ground state correlation to the
monopole strength in $^{16}$O was made in the framework of the $^{12}$C + $\alpha$ OCM.
It is because, in discussing the monopole strength without ground state correlation
in Sec.~\ref{formulation}, we used the results of the $^{12}$C + $\alpha$ OCM
in Ref.~\cite{suzuki}.
We repeated the same calculation as one in Ref.~\cite{suzuki}.
We found that 1)~increasing the amount of the ground state correlation,
the monopole strengths are growing almost monotonously, and
2)~at a full amount of the ground state correlation, the monopole strengths
are reproduced within a factor of $1.13$ in comparison with the observed values.
The reason why the ground state correlation enhances the monopole strengths
was discussed with a simple approach.
In the case of $^{12}$C, the investigation of the effect of the ground state
correlation to the monopole strength was performed by expressing the ground state
with the so-called THSR wave function~\cite{tohsaki}.
This wave function has two parameters ${\widetilde \gamma}$ and $\nu_N$.
When ${\widetilde \gamma}$ = $\nu_N$,
the wave function is just equal to the SU(3) wave function with
$N_{TOT}=8$ and $(\lambda, \mu) = (0, 4)$.
As we make the ratio ${\widetilde \gamma} / \nu_N$ smaller than unity, the wave
function contains more amount of the ground state correlation.
We found that at a full amount of the ground state correlation,
the monopole strength is reproduced within a factor of $1.35$ in comparison
with the observed value.
The implication of the Bayman-Bohr theorem has been misunderstood such that
the cluster model description is rather unnecessary, since a cluster model
wave function is equivalent to a shell model wave function.
The existence of cluster states especially as excited states is well
established these days.
Thus, the implication of the Bayman-Bohr theorem should be understood
straightforwardly as follows.
If the ground state is well described by an SU(3) shell model wave
function equivalent to a cluster model wave function, the ground state
possesses two different characters simultaneously, shell-model-state
character and cluster-model-state character.
This means that the ground state has mean-field degree of freedom
and clustering degree of freedom simultaneously.
Both of them can be excited, when the nucleus is stimulated by an external field.
The monopole excitation to excited cluster states demonstrates us directly
the evidence that the clustering degree of freedom is embedded in the ground state.
In this paper we showed that the clustering degree of freedom embedded
in the ground state can reproduce the order of magnitude of the monopole strength
even without taking into account the ground state correlation.
Moreover it was demonstrated that, if we take into account the
ground state correlation activating the clustering degree of freedom
described by the Bayman-Bohr theorem,
the monopole strengths are reproduced within a factor of $1.13$ in $^{16}$O
and within a factor of $1.35$ in $^{12}$C, in comparison with the observed values.
Our present study ascertains that the monopole transition between cluster
and ground states in light nuclei is generally strong as to be comparable
to the single particle strength.
The measurement of strong monopole transitions or excitations,
therefore, is in general very useful for the study of cluster states.
One of the authors (Y.~F.) is grateful for the financial assistance
from the Special Postdoctoral Researcher Program of RIKEN.
%\newpage
%%%%%%%%%%%%%%%%%%%%%%%%%%
\appendix
%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{The energy-weighted sum rule of monopole transition by the use of
the Jacobi coordinate} \label{sumrule}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
We discuss here the energy-weighted sum rule of the monopole transition.
The sum rule is written as follows
\begin{eqnarray}
&& \sum_k |\langle k | \frac{1}{2} \sum_{i=1}^A
(\Vec{r}_i - \Vec{r}_G )^2 | g \rangle|^2 ( E_k - E_g ) =
\frac{\hbar^2}{2m} A R^2_{\rm rms}, \\
&& R^2_{\rm rms} = \frac{1}{A} \langle g | \sum_{i=1}^A
(\Vec{r}_i - \Vec{r}_G )^2 | g \rangle,
\end{eqnarray}
where $| g \rangle$ and $E_g$ stand for the ground state and
its energy, respectively, $| k \rangle$ and $E_k$ represent
the $k$-th excited state and its energy, respectively, and
$\Vec{r}_G$ stands for the center-of-mass coordinate.
In $^{16}$O, the observed value of $R_{\rm rms}$ is 2.67 fm and then
the energy-weighted sum rule value $(\hbar^2/2m) 16 R^2_{\rm rms}$
is 2361 fm$^4 \cdot$MeV.
In the case of the $0^+_2$ state at 6.05 MeV which has
$M({\rm E0},0^+_1-0^+_2)_{\rm exp}$ = 3.55 fm$^2$, the energy-weighted monopole
transition strength is $(3.55)^2 \times 6.05 = 76.3$ fm$^4 \cdot$MeV.
This value is 3.2 \% of the energy-weighted sum rule value.
In the case of the $0^+_3$ state at 12.05 MeV which has
$M({\rm E0},0^+_1-0^+_3)_{\rm exp}$ = 4.03 fm$^2$, the energy-weighted
monopole transition strength is $(4.03)^2 \times 12.05 = 196$
fm$^4 \cdot$MeV. This value is 8.3 \% of the energy-weighted
sum rule value.
The sum of the energy-weighted monopole transition strengths
of $0^+_2$ and $0^+_3$ states is 11.5 \% of the energy-weighted
sum rule value.
In $^{12}$C, the observed value of $R_{\rm rms}$ is 2.37 fm and then
the energy-weighted sum rule value $(\hbar^2/2m) 12 R^2_{\rm rms}$
is 1395 fm$^4 \cdot$MeV.
In the case of the $0^+_2$ state at 7.66 MeV which has
$M({\rm E0},0^+_1-0^+_2)_{\rm exp}$ = 5.4 fm$^2$,
the energy-weighted monopole transition strength is
$(5.4)^2 \times 7.66 = 223$ fm$^4 \cdot$MeV.
This value is 16 \% of the energy-weighted sum rule value.
These percentage values show that the strength of the monopole
transition or excitation to cluster states shares an appreciable
portion of the energy-weighted sum rule value.
The formula of the energy-weighted sum rule of monopole transition is
obtained by calculating the double commutator of the monopole transition
operator $O_M$ and the system Hamiltonian $H$
\begin{equation}
[ O_M, [ H, O_M ]], \ \ \ O_M = \frac{1}{2} \sum_{i=1}^A
(\Vec{r}_i - \Vec{r}_G )^2.
\end{equation}
The calculation of the double commutator looks tedious due to
the existence of the center-of-mass coordinate $\Vec{r}_G$ but
it can be made very easily by using the Jacobi coordinate.
For the Hamiltonian $H$ with momentum-independent interaction, $H$ can be
replaced by the kinetic energy operator $K$
\begin{equation}
[ O_M, [ H, O_M ]] = [ O_M, [ K, O_M ]], \ \ \ K = \frac{-\hbar^2}{2m}
\sum_{i=1}^A \left( \frac{\partial}{\partial \Vec{r}_i} \right)^2
- \frac{-\hbar^2}{2Am} \left( \frac{\partial}{\partial \Vec{r}_G} \right)^2.
\end{equation}
Now we introduce the normalized Jacobi coordinates as
\begin{eqnarray}
\Vec{x}_j &=& \sqrt{\frac{j}{j+1}} \left( \frac{1}{j} \sum_{i=1}^j
\Vec{r}_i - \Vec{r}_{j+1} \right), \ \ \ j = 1 \sim A-1, \\
\Vec{x}_A &=& \sqrt{\frac{1}{A}} \sum_{i=1}^A \Vec{r}_i =
\sqrt{A} \Vec{r}_G.
\end{eqnarray}
One can easily check that the linear transformation from
$\{ \Vec{r}_i, i=1 \sim A \}$ to $\{ \Vec{x}_j, j=1 \sim A \}$ is unitary.
Therefore we have
\begin{eqnarray}
&& \sum_{i=1}^A \Vec{r}_i^2 = \sum_{j=1}^A \Vec{x}_j^2, \\
&& \sum_{i=1}^A (\Vec{r}_i - \Vec{r}_G)^2 = \sum_{i=1}^A \Vec{r}_i^2 -
A \Vec{r}_G^2 = \sum_{j=1}^{A-1} \Vec{x}_j^2, \\
&& \sum_{i=1}^A \left( \frac{\partial}{\partial \Vec{r}_i} \right)^2
= \sum_{j=1}^A \left( \frac{\partial}{\partial \Vec{x}_j} \right)^2, \\
&& \sum_{i=1}^A \left( \frac{\partial}{\partial \Vec{r}_i} \right)^2
- \frac{1}{A} \left( \frac{\partial}{\partial \Vec{r}_G} \right)^2
= \sum_{j=1}^{A-1} \left( \frac{\partial}{\partial \Vec{x}_j} \right)^2,
\end{eqnarray}
Thus we have
\begin{eqnarray}
O_M &=& \frac{1}{2} \sum_{j=1}^{A-1} \Vec{x}_j^2, \\
K &=& \frac{-\hbar^2}{2m} \sum_{j=1}^{A-1} \left( \frac{\partial}
{\partial \Vec{x}_j} \right)^2,
\end{eqnarray}
When we use the above expressions of $O_M$ and $K$ by the normalized Jacobi
coordinates, we can easily obtain the following result
\begin{equation}
[ O_M, [ H, O_M ]] = [ O_M, [ K, O_M ]] = \frac{\hbar^2}{m}
\sum_{j=1}^{A-1} \Vec{x}_j^2 = \frac{\hbar^2}{m} \sum_{i=1}^A
(\Vec{r}_i - \Vec{r}_G)^2.
\end{equation}
We thus have the formula of the energy-weighted sum rule of monopole
transition as follows
\begin{eqnarray}
&& \frac{\hbar^2}{2m} \langle g | \sum_{i=1}^A (\Vec{r}_i - \Vec{r}_G)^2 | g
\rangle = \frac{1}{2} \langle g | [ O_M, [ H, O_M ]] | g \rangle \\
&& \ \ = \sum_k |\langle k | \frac{1}{2} \sum_{i=1}^A
(\Vec{r}_i - \Vec{r}_G )^2 | g \rangle|^2 ( E_k - E_g ).
\end{eqnarray}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Description of $^{16}$O and $^{12}$C ground states with SU(3) wave function} \label{Baymann_Bohr}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
The total number of the oscillator quanta $N_{TOT}$ possessed by the
doubly closed shell wave function of $^{16}$O is $N_{TOT}=12$.
For $N_{TOT}=12$, it is possible to construct many $^{12}$C+$\alpha$
cluster wave functions with various SU(3) symmetry $(\lambda,\mu)$,
${\cal A} \{ [{\cal R}_4(\Vec{r}) \phi({}^{12}{\rm C})]_{(\lambda,\mu)} \phi (\alpha) \}$.
These wave functions, however, become all zero except for $(\lambda,\mu)=(0,0)$,
because of the nature of the doubly closed wave function.
Using the following relation
\begin{equation}
[{\cal R}_{4L}(\Vec{r}) \phi_L({}^{12}{\rm C})]_{J=0} = \sum_{(\lambda,\mu)}
\langle (4,0)L, (0,4)L ||(\lambda,\mu)0 \rangle
[{\cal R}_4(\Vec{r}) \phi({}^{12}{\rm C})]_{(\lambda,\mu)},
\end{equation}
we have
\begin{eqnarray}
&& {\cal A} \{ [{\cal R}_{4L}(\Vec{r}) \phi_L({}^{12}{\rm C})]_{J=0}
\phi (\alpha) \} \nonumber \\
&& \hspace{1.5cm} = \sum_{(\lambda,\mu)}
\langle (4,0)L, (0,4)L ||(\lambda,\mu)0 \rangle
{\cal A} \{ [{\cal R}_4(\Vec{r}) \phi({}^{12}{\rm C})]_{(\lambda,\mu)}
\phi (\alpha) \} \nonumber \\
&& \hspace{1.5cm} = \langle (4,0)L, (0,4)L ||(0,0)0 \rangle
{\cal A} \{ [{\cal R}_4(\Vec{r}) \phi({}^{12}{\rm C})]_{(\lambda,\mu)=(0,0)}
\phi (\alpha) \},
\end{eqnarray}
for $L=0$, 2, and 4.
This relation is an explanation of the equalities of Eqs.~(\ref{eq:clshla})$\sim$(\ref{eq:clshlc}).
Similar argument holds for the ground state of $^{12}$C.
Although there can be constructed many 3$\alpha$ cluster wave functions
with various SU(3) symmetry $(\lambda,\mu)$,
${\cal A} \{ [{\cal R}_{N_1}(\Vec{s},2\nu_N)
{\cal R}_{N_2}(\Vec{t},(8/3)\nu_N)]_{(\lambda,\mu)} \phi(\alpha_1) \phi(\alpha_2)
\phi(\alpha_3) \}$ for $N_{TOT} = N_1 + N_2 =8$ which is the lowest
number of the total oscillator quanta for $^{12}$C, only one wave
function with $(\lambda,\mu)=(0,4)$ is non-vanishing which is possible
for $N_1 = N_2 = 4$~\cite{hori,kato}.
Therefore we have the following relations
\begin{eqnarray}
&& {\cal A} \{ [{\cal R}_{N_1 L}(\Vec{s},2\nu_N) {\cal R}_{N_2 L}(\Vec{t},(8/3)\nu_N)
]_{J=0} \phi(\alpha_1) \phi(\alpha_2) \phi(\alpha_3) \}
\nonumber \\
&& \hspace{0.4cm} = \delta_{N_1,4} \delta_{N_2,4}
\langle (4,0)L, (4,0)L ||(0,4)0 \rangle \nonumber \\
&& \hspace{0.4cm} \times {\cal A} \{ [ {\cal R}_{N_1}(\Vec{s},2\nu_N)
{\cal R}_{N_2}(\Vec{t},(8/3)\nu_N)]_{(\lambda,\mu)=(0,4)}
\phi(\alpha_1) \phi(\alpha_2) \phi(\alpha_3) \},
\end{eqnarray}
for $L=0$, 2, and 4.
This relation is an explanation of the equalities of
Eqs.~(\ref{eq:c12a})$\sim$(\ref{eq:c12c}).
%\newpage
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Dependence of $M({\rm E0},0^+_1-0^+_2)$ on the width parameter $\gamma$ of
the 3$\alpha$ condensed wave function} \label{Hoyle}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
First we note
\begin{eqnarray}
&& {\cal A} \{ {\widehat \chi}_H(\Vec{s},\Vec{t}) \phi(\alpha_1)
\phi(\alpha_2) \phi(\alpha_3) \} = {\cal A}
\{ {\widehat \chi}_{HG}(\Vec{s},\Vec{t}) \phi(\alpha_1) \phi(\alpha_2)
\phi(\alpha_3) \} \nonumber \\
&& \hspace{1cm} - \langle {\cal R}^{8,J=0}_{4,4,L=0}(\Vec{s},\Vec{t})
| {\widehat \chi}_{HG}(\Vec{s},\Vec{t}) \rangle
{\cal A} \{ {\cal R}^{8,J=0}_{4,4,L=0}(\Vec{s},\Vec{t})
\phi(\alpha_1) \phi(\alpha_2) \phi(\alpha_3) \}.
\end{eqnarray}
Then we obtain
\begin{equation}
\langle {\widehat \chi}_H \rangle
= \langle {\widehat \chi}_{HG} \rangle -
\left( \langle {\cal R}^{8,J=0}_{4,4,L=0}(\Vec{s},\Vec{t}) |
{\widehat \chi}_{HG}(\Vec{s},\Vec{t}) \rangle \right)^2
\langle {\cal R}^{8,J=0}_{4,4,L=0} \rangle,
\end{equation}
with the notation
\begin{equation}
\langle G \rangle = \langle G(\Vec{s},\Vec{t})
\phi(\alpha_1) \phi(\alpha_2) \phi(\alpha_3) | {\cal A} \{
G(\Vec{s},\Vec{t}) \phi(\alpha_1) \phi(\alpha_2) \phi(\alpha_3)
\} \rangle. \label{eq:appC_G}
\end{equation}
Next we note
\begin{eqnarray}
&& {\cal A} \{ {\widehat \chi}_{HG}(\Vec{s},\Vec{t}) \phi(\alpha_1)
\phi(\alpha_2) \phi(\alpha_3) \} = \sum_{n \ge 4}^\infty \sum_{n_1+n_2=n}
\langle {\cal R}^{2n,J=0}_{2n_1,2n_2,0}(\Vec{s},\Vec{t})
| {\widehat \chi}_{HG}(\Vec{s},\Vec{t}) \rangle \nonumber \\
&& \hspace{4cm} \times {\cal A} \{ {\cal R}^{2n,J=0}_{2n_1,2n_2,0}
(\Vec{s},\Vec{t}) \phi(\alpha_1) \phi(\alpha_2) \phi(\alpha_3) \} \\
&& \hspace{0.5cm} = \left( \frac{2\sqrt{\nu_N\gamma}}
{\nu_N+\gamma} \right)^3 \sum_{n=4}^\infty \left( \frac{\nu_N-\gamma}
{\nu_N+\gamma} \right)^n {\cal A} \{ F_n(\Vec{s},\Vec{t})
\phi(\alpha_1) \phi(\alpha_2) \phi(\alpha_3) \},\label{eq:appC_HG}\\
&& F_n(\Vec{s},\Vec{t}) = \sum_{n_1+n_2=n} \sqrt{\frac{(2n_1+1)!!(2n_2+1)!!}
{(2n_1)!!(2n_2)!!}} {\cal R}^{2n,J=0}_{2n_1,2n_2,0}(\Vec{s},\Vec{t}),
\end{eqnarray}
where the following formula is used
\begin{equation}
\langle R_{2n,0}(r,\beta)| \sqrt{4\pi}
\left(\frac{2\beta'}{\pi}\right)^{\frac{3}{4}} e^{-\beta' r^2}
\rangle = \sqrt{\frac{(2n+1)!!}{(2n)!!}}
\left( \frac{2\sqrt{\beta\beta'}}{\beta+\beta'} \right)^{\frac{3}{2}}
\left( \frac{\beta-\beta'}{\beta+\beta'} \right)^n,
\end{equation}
From Eqs.~(\ref{eq:appC_G}) and (\ref{eq:appC_HG}) we obtain
\begin{equation}
\langle {\widehat \chi}_{HG} \rangle =
\left( \frac{2\sqrt{\nu_N\gamma}}{\nu_N+\gamma} \right)^6
\sum_{n=4}^\infty
\left( \frac{\nu_N-\gamma}{\nu_N+\gamma} \right)^{2n} \langle F_n \rangle.
\end{equation}
By noticing
\begin{eqnarray}
&& \langle {\cal R}^{8,J=0}_{4,4,0}(\Vec{s},\Vec{t}) |
{\widehat \chi}_{HG}(\Vec{s},
\Vec{t}) \rangle = \frac{5!!}{4!!} \left( \frac{2\sqrt{\nu_N\gamma}}
{\nu_N+\gamma} \right)^3 \left( \frac{\nu_N-\gamma}{\nu_N+\gamma}
\right)^4,
\\
&& \langle F_4 \rangle = \left( \frac{5!!}{4!!} \right)^2
\langle {\cal R}^{8,J=0}_{4,4,0} \rangle
\equiv \left( \frac{5!!}{4!!} \right)^2\frac{1}{(\widehat{N}_{g0})^2},
\end{eqnarray}
we obtain
\begin{eqnarray}
&&\langle {\widehat \chi}_H \rangle =
\left( \frac{2\sqrt{\nu_N\gamma}}{\nu_N+\gamma} \right)^6
\sum_{n=5}^\infty
\left( \frac{\nu_N-\gamma}{\nu_N+\gamma} \right)^{2n} \langle F_n \rangle
\equiv \frac{1}{(\widehat{N}_H)^2},\\
&&D_{H,2n_1,2n_2}\equiv \sqrt{\frac{(2n_1+1)!!(2n_2+1)!!}{(2n_1)!!(2n_2)!!}}
\left( \frac{2\sqrt{\nu_N\gamma}}{\nu_N+\gamma} \right)^3
\left( \frac{\nu_N-\gamma}{\nu_N+\gamma} \right)^{n_1+n_2},
\end{eqnarray}
By combining all these formulas we have
\begin{eqnarray}
&& M({\rm E0},0_2^+ - 0_1^+) = \frac{1}{2} \frac{{\widehat N}_H}
{{\widehat N}_{g0}} ( D_{H,6,4} + D_{H,4,6} ) \langle R_{40}(r,\nu_N) |
r^2 | R_{60}(r,\nu_N) \rangle \\
&& \hspace{0.5cm} = \frac{1}{2}
\sqrt{ \frac{\langle {\cal R}^{8,J=0}_{4,4,0} \rangle}
{\displaystyle{\left( \frac{2\sqrt{\nu_N\gamma}}{\nu_N+\gamma} \right)^6 }
\sum_{n=5}^\infty \displaystyle{ \left(
\frac{\nu_N-\gamma}{\nu_N+\gamma} \right)^{2n} }
\langle F_n \rangle} } \nonumber \\
&& \hspace{1cm} \times 2 \sqrt{\frac{5!!7!!}{4!!6!!}}
\left( \frac{2\sqrt{\nu_N\gamma}}{\nu_N+\gamma} \right)^3
\left( \frac{\nu_N-\gamma}{\nu_N+\gamma} \right)^5
\langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle \\
&& \hspace{0.5cm} = \sqrt{\frac{7}{6}}
\sqrt{\frac{\langle F_4 \rangle}{\langle F_5 \rangle}} \xi_5
\langle R_{40}(r,\nu_N) | r^2 | R_{60}(r,\nu_N) \rangle, \\
&& \xi_5 \equiv \sqrt{\frac{\langle F_5 \rangle}
{\langle F_5 \rangle + \sum_{n=6}^\infty
\displaystyle{ \left( \frac{\nu_N-\gamma}{\nu_N+\gamma} \right)^{2(n-5)}}
\langle F_n \rangle}}.
\end{eqnarray}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\section{Estimation of $\eta_N$ and $\zeta_N$ by Ref.~\cite{suzuki}}
\label{suzuki}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
In Ref.~\cite{suzuki} the $^{16}$O states are expressed by microscopic
$^{12}$C + $\alpha$ cluster wave functions where $^{12}$C cluster can be
excited to its first $2^+$ and $4^+$ states.
This coupled channel problem is solved by using the coupled channel
OCM (orthogonality condition model).
The wave functions $\Phi$ are expressed by the SU(3)-coupled basis of the
$^{12}$C + $\alpha$ cluster model space, which is in the case of $J=0$
\begin{eqnarray}
&& \Phi = \sum_{Nq} \xi_{Nq} \Phi_{Nq}, \\
&& \Phi_{N,q=(\lambda,\mu)} = N_{Nq} \frac{1}{\sqrt{\mu_{Nq}}}
\frac{1}{\sqrt{{}_{16}C_4}} {\cal A}
\{ [{\cal R}_N(\Vec{r},3\nu_N) \phi({}^{12}{\rm C})]_{(\lambda,\mu)}
\phi(\alpha) \}, \\
&& \mu_{N,q=(\lambda,\mu)} = \langle [{\cal R}_N(\Vec{r},3\nu_N)
\phi({}^{12}{\rm C})]_{(\lambda,\mu)} \phi(\alpha) | {\cal A}
\{ [{\cal R}_N(\Vec{r},3\nu_N) \phi({}^{12}{\rm C})]_{(\lambda,\mu)}
\phi(\alpha) \} \rangle.
\end{eqnarray}
The SU(3)-coupled basis wave functions $\Phi_{N,q=(\lambda,\mu)}$ are
ortho-normalized and are the eigenfunctions of normalization kernel
with eigenvalues $\mu_{N,q=(\lambda,\mu)}$. The values of the
expansion coefficients $\xi_{Nq}$ are given in Table I of the second
paper of Ref.~\cite{suzuki}.
Since the wave function of the $0_2^+$ state shows predominantly a
$^{12}$C($0^+$) + $\alpha$ structure, we approximated in this paper
the wave function of the $0_2^+$ state by a pure $^{12}$C($0^+$) +
$\alpha$ structure
\begin{eqnarray}
\Phi(0^+_2) &=& \sum_{N \ge 6} \eta_N \Phi_N, \\
\Phi_N &=& \frac{1}{\sqrt{\tau_{0,N}}} \frac{1}{\sqrt{{}_{16}C_4}}
{\cal A} \{ R_{N0}(r) Y_{00}(\Vec{\hat r}) \phi_{L=0}({}^{12}{\rm C})
\phi(\alpha) \}.
\end{eqnarray}
This means we adopted the following approximation for each $N$
\begin{equation}
\sum_q \xi_{Nq} \Phi_{Nq} \approx \eta_N \Phi_N.
\end{equation}
Therefore we have
\begin{equation}
\eta_N \approx \sqrt{\sum_q (\xi_{Nq})^2}.
\end{equation}
The values $\sum_q (\xi_{Nq})^2$ are tabulated in Table I of the first paper
of Ref.~\cite{suzuki}. For $N=6$, we have $\eta_6 \approx \sqrt{0.144} =
0.379$.
From Table I of the second paper of Ref.~\cite{suzuki}, we have in the case
of $N=6$ for $0_2^+$
\begin{equation}
\sum_q \xi_{6q} \Phi_{6q} = 0.338 \Phi_{6,(4,2)} + 0.171 \Phi_{6,(2,0)}.
\ \ ({\rm for}\ 0_2^+) \label{eq:ocmA}
\end{equation}
We should note here that $\Phi_{6,(2,0)}$ is nothing but $\Phi_{(2,0)}^{cl}$
which absorbs the total monopole strength of $\Phi_{CS}$ within the
$^{12}$C + $\alpha$ cluster model space. On the other hand, there holds
\begin{equation}
\Phi_6 = 0.750 \Phi_{6,(4,2)} + 0.661 \Phi_{6,(2,0)}.
\end{equation}
This relation is easily obtained by comparing $\langle \Phi_6 | O_M |
\Phi_{CS} \rangle$ with $\langle \Phi_{6,(2,0)} | O_M | \Phi_{CS} \rangle$:
\begin{eqnarray}
&& \langle \Phi_6 | O_M | \Phi_{CS} \rangle = \frac{1}{2}
\sqrt{ \frac{\tau_{0,4}}{\tau_{0,6}}} \langle R_{40}(r,\nu_N) | r^2 |
R_{60}(r,\nu_N) \rangle
= \frac{0.784}{\nu_N}, \\
&& \langle \Phi_{6,(2,0)} | O_M | \Phi_{CS} \rangle = \frac{1.19}{\nu_N}.
\end{eqnarray}
Since the monopole operator $O_M$ does not connect $\Phi_{CS}$ with
$\Phi_{6,(4,2)}$, we have
\begin{eqnarray}
\langle \Phi_{6,(2,0)} | \Phi_6 \rangle = \frac{0.784}{1.19} = 0.661.
\end{eqnarray}
By using $\eta_6 = 0.379$ we have
\begin{eqnarray}
\eta_6 \Phi_6 = 0.284 \Phi_{6,(4,2)} + 0.250 \Phi_{6,(2,0)}.
\label{eq:0-0comp}
\end{eqnarray}
The comparison of Eq.~(\ref{eq:ocmA}) with Eq.~(\ref{eq:0-0comp}) tells
us that the approximation of $\sum_q \xi_{6q} \Phi_{6q}$ with
$\eta_6 \Phi_6$ is not very good and that the monopole strength of
$\sum_q \xi_{6q} \Phi_{6q}$ is weaker than that of $\eta_6 \Phi_6$.
The arguments for the $0_3^+$ state can be made completely in the same way
as the $0_2^+$ state. We obtain following relations
\begin{eqnarray}
&& \zeta_6 \approx \sqrt{0.325} = 0.570, \\
&& \sum_q \xi_{6q} \Phi_{6q} = 0.471 \Phi_{6,(4,2)} + 0.321 \Phi_{6,(2,0)},
\ \ ({\rm for}\ 0_3^+), \\
&& \Psi_6 = 0.493 \Phi_{6,(4,2)} + 0.870 \Phi_{6,(2,0)}, \\
&& \zeta_6 \Psi_6 = 0.281 \Phi_{6,(4,2)} + 0.496 \Phi_{6,(2,0)}.
\end{eqnarray}
The approximation of $\sum_q \xi_{6q} \Phi_{6q}$ with $\zeta_6 \Psi_6$ is
not also so good and the monopole strength of $\sum_q \xi_{6q} \Phi_{6q}$
is also weaker than that of $\zeta_6 \Psi_6$.
%\newpage
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%%%% references
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{thebibliography}{9}
\bibitem{supple} Y. Fujiwara, H. Horiuchi, K. Ikeda, M. Kamimura, K. Kato,
Y. Suzuki, and E. Uegaki, Prog. Theor. Phys. Supple. {\bf 68}, 29
(1980).
\bibitem{hoik} H. Horiuchi and K. Ikeda, Prog. Theor. Phys. {\bf 40},
277 (1968).
\bibitem{suzuki} Y. Suzuki, Prog. Theor. Phys. {\bf 55}, 1751 (1976); Prog.
Theor. Phys. {\bf 56}, 111 (1976).
\bibitem{ajzen86} P. Ajzenberg-Selove, Nucl. Phys. {\bf A 460}, (1986) 1.
\bibitem{Hoyle} F. Hoyle, Astrophys. J. Supple. \textbf{1}, 121 (1954).
\bibitem{kawabata} T. Kawabata, H. Akimune, H. Fujita, Y. Fujita, M. Fujiwara,
K. Hara, K. Hatanaka, M. Itoh, Y. Kanada-En'yo, S. Kishi, K. Nakanishi,
H. Sakaguchi, Y. Simbara, A. Tamii, S. Terashima, M. Uchida, T. Wakasa,
Y. Yoshida, H. P. Yoshida, M. Yosoi, Phys. Lett. {\bf B 646}, (2007) 6.
\bibitem{yamada07} T. Yamada and Y. Funaki, to be prepared.
\bibitem{kami} M. Kamimura, Nucl. Phys. A {\bf 351}, 456 (1981).
\bibitem{uega} E. Uegaki, S. Okabe, Y. Abe, and H. Tanaka, Prog. Theor. Phys.
{\bf 57}, 1262 (1977) ; E. Uegaki, Y. Abe, S. Okabe, and H. Tanaka, Prog.
Theor. Phys. {\bf 59}, 1031 (1978) ; {\bf 62}, 1621 (1979).
\bibitem{BB} B. F. Bayman and A. Bohr, Nucl. Phys. {\bf 9}, 596 (1958/59).
\bibitem{elliott58} J. P. Elliott, Proc. Roy. Soc. (London) {\bf A245}, 128;
562.
\bibitem{Wildermuth} K. Wildermuth and Th. Kanellopoulos, Nucl. Phys. A
{\bf 7}, 150 (1958); {\bf 9}, 449 (1958/59).
\bibitem{hecht} K. T. Hecht, Phys. Rev. {\bf C16}, 2401 (1977).
\bibitem{tohsaki} A. Tohsaki, H. Horiuchi, P. Schuck, and G. R\"opke, Phys.
Rev. Lett. {\bf 87}, 192501 (2001).
\bibitem{funaki} Y. Funaki, A. Tohsaki, H. Horiuchi, P. Schuck, and
G. R\"opke, Phys. Rev. C {\bf 67}, 051306(R) (2003).
\bibitem{horic} H. Horiuchi, Prog. Theor. Phys. {\bf 51}, 745 (1974).
\bibitem{hori} H. Horiuchi, Prog.Theor. Phys. {\bf 51}, 1266 (1974).
\bibitem{yamada} T. Yamada and P. Schuck, Eur. Phys. J. A {\bf 26}, 185 (2005).
\bibitem{bertsch} G. F. Bertsch, Phys. Lett. {\bf 21}, 70 (1966).
\bibitem{kato} K. Kato and H. Bando, Prog. Theor. Phys. {\bf 53}, 692 (1975).
\end{thebibliography}
\newpage
%%%%%%%%%%%%%
% TABLE 1 %
%%%%%%%%%%%%%
\begin{table}
\caption
{Values of $\tau_{0,N}$, $\tau_{2,N}$ and $\tau_{4,N}$ which are calculated
with the analytical expression given
in Refs.~\cite{suzuki,horic}.}
\label{tab:1}
\begin{center}
\begin{tabular}{cccc}
\hline
\hline
\hspace*{10mm}$N$\hspace{10mm} & \hspace*{10mm}$\tau_{0,N}$\hspace*{10mm}
& \hspace*{10mm}$\tau_{2,N}$\hspace*{10mm}
& \hspace*{10mm}$\tau_{4,N}$\hspace*{10mm} \\
\hline
4 & 0.2963 & 1.4815 & 2.6667 \\
6 & 0.3160 & 0.7831 & 1.0243 \\
8 & 0.5615 & 0.7743 & 0.8773 \\
10 & 0.7405 & 0.8551 & 0.8985 \\
12 & 0.8564 & 0.9182 & 0.9374 \\
14 & 0.9247 & 0.9567 & 0.9654 \\
16 & 0.9619 & 0.9780 & 0.9818 \\
18 & 0.9811 & 0.9890 & 0.9907 \\
20 & 0.9908 & 0.9946 & 0.9953 \\
22 & 0.9955 & 0.9974 & 0.9977 \\
24 & 0.9979 & 0.9987 & 0.9989 \\
\hline
\hline
\end{tabular}
\end{center}
\end{table}
\newpage
%%%%%%%%%%%%%
% TABLE 2 %
%%%%%%%%%%%%%
\begin{table}
\begin{center}
\caption
{$\xi_5$ and monopole matrix element $M({\rm E0},0_2^+ - 0_1^+)$ in $^{12}$C calculated
at several values of $\gamma$.
$M({\rm E0},0_2^+ - 0_1^+)$ is given as $\xi_5\times 0.882/\nu_N$ ${\rm fm}^2$,
where the value $\nu_N=0.168$ ${\rm fm}^{-2}$ is used.
$R_{\rm rms}(0_2^+)$
is the corresponding rms radius of the Hoyle state to the adopted values
of $\nu_N$ and $\gamma$.
}
\label{tab:2}
\vspace{5mm}
\begin{tabular}{cccc}
\hline\hline
\hspace*{5mm}$\gamma$ $[{\rm fm}^{-2}]$\hspace*{5mm} & \hspace*{5mm}$\xi_5$\hspace*{5mm}
& \hspace*{5mm}$M({\rm E0}, 0_2^+ - 0_1^+)$ $[{\rm fm}^2]$\hspace*{5mm}
& \hspace*{5mm}$R_{\rm rms}(0_2^+)$ [fm]\hspace*{5mm} \\
\hline
$0.0238$ & $0.338$ & $1.775$ & $3.56$ \\
$0.0182$ & $0.252$ & $1.326$ & $3.79$ \\
$0.0143$ & $0.191$ & $1.004$ & $4.03$ \\
$0.0115$ & $0.147$ & $0.773$ & $4.28$ \\
\hline\hline
\end{tabular}
\end{center}
\end{table}
\newpage
%%%%%%%%%%%%%
% TABLE 3 %
%%%%%%%%%%%%%
\begin{table}
\begin{center}
\caption
{Dependence of the monopole matrix element in $^{12}$C on the amount of $3\alpha$-like correlation
involved in the ground state, which is characterized by ${\widetilde \gamma}/\nu_N$.
The monopole matrix element is given as
$M({\rm E0},0_2^+ - 0_1^+)$=$\langle \Psi_G({\widetilde \gamma},\nu_N)
| \sum_{i=1}^{12}(\Vec{r}_i-\Vec{r}_G)^2/2 | \Psi_H(\gamma,{\widetilde \gamma},\nu_N) \rangle$,
where $\Psi_G({\widetilde \gamma},\nu_N)$ and $\Psi_H(\gamma,{\widetilde \gamma},\nu_N)$
are the ground state and Hoyle state wave functions, respectively.
The rms radius of the ground-state wave function $\Psi_G({\widetilde \gamma},\nu_N)$ is
fixed to the experimental one ($2.47$~fm). Then, the ratio ${\widetilde \gamma}/\nu_N$ is
only the parameter to describe the property of the ground state.
For a given value of ${\widetilde \gamma}/\nu_N$, the value of $\gamma$ in
$\Psi_H(\gamma,{\widetilde \gamma},\nu_N)$ is chosen so as to reproduce the rms radius
of the Hoyle state ($3.8$ fm).
See the text for details.
}
\label{tab:3}
\vspace{5mm}
\begin{tabular}{ccc}
\hline\hline
\hspace*{5mm}${\widetilde \gamma}/\nu_N$ \hspace*{5mm} & \hspace*{5mm}$M({\rm E0}, 0_2^+ - 0_1^+)$ $[{\rm fm}^2]$ \hspace*{5mm} \\
\hline
$1.000$ & $1.326$ \\
$0.705$ & $1.810$ \\
$0.498$ & $2.473$ \\
$0.309$ & $3.597$ \\
$0.274$ & $4.035$ \\
\hline\hline
\end{tabular}
\end{center}
\end{table}
\newpage
%%%%%%%%%%%%%
% FIGURE 1 %
%%%%%%%%%%%%%
\begin{figure}
\begin{center}
\includegraphics*[scale=0.5,clip]{fig1.eps}
\caption{
Dependence of the monopole strengths $M_{N}({\rm E0})$
on the model space of the ground state wave function characterized as quanta $N$
[see Eq.~(\ref{eq:monopole_N_dependence})].
The square and circle points correspond to $M_{N}({\rm E0}; {0^+_1} \rightarrow {0^+_2})$
and $M_{N}({\rm E0}; {0^+_1} \rightarrow {0^+_3})$, respectively.
}
\label{fig:1}
\end{center}
\end{figure}
\end{document}