eta' meson production in proton-proton collisions
Extracted source text ยท arXiv nucl-th/9908077
\documentstyle[preprint,prc,aps]{revtex}
%\documentstyle[prc,aps]{revtex}
\begin{document}
\draft
\title{$\eta\prime$ meson production in proton-proton collisions}
\author{K. Nakayama$^{a,b}$, H. F. Arellano$^c$, J. W. Durso$^{a,d}$, and
J. Speth$^a$}
\address{$^a$Institut f\"ur Kernphysik, Forschungszentrum-J\"ulich,
D-52425, J\"ulich, Germany \\
$^b$Department of Physics and Astronomy, University of Georgia, Athens,
GA 30602, USA \\
$^c$Departamento de F\'{\i}sica, Facultad de Ciencias F\'{\i}sicas y
Matem\'aticas, Universidad de Chile, Blanco Encalada 2008, Santiago, Chile \\
$^d$Physics Department, Mount Holyoke College, South Hadley, MA 01075, USA}
\maketitle
\begin{abstract}
The $pp\rightarrow pp\eta^\prime$ reaction is investigated within a relativistic
meson-exchange model of hadronic interactions. We explore the role of nucleonic
and mesonic, as well as the $N^*$ resonance currents, in producing $\eta^\prime$
mesons. In order to learn more about the production mechanisms, new measurements
in the energy region far from the threshold are required.
\end{abstract}
PACS: 13.60Le, 25.10+s, 25.40-h
\newpage
%------------------------------------------------
\section{ Introduction}
%------------------------------------------------
With the advent of particle accelerators in the few $GeV$ energy region,
heavy meson production in hadronic collisions has attracted increasing
attention in the past few years. In particular, heavy meson production
in nucleon-nucleon ($NN$) collisions at near-threshold energies is of
special interest, not only because it is suited for extracting information
on a few lowest-order multipole amplitudes, but also because it is considered to
provide important information on the short distance behavior of the $NN$
interaction. Due to the large momentum transfer between the initial
and final nucleons, these reactions at near-threshold energies necessarily
probe the $NN$ interaction at short distances. In the present work
we concentrate on the $\eta^\prime$ meson production in proton-proton
($pp$) collisions.
Among heavy mesons, the $\eta^\prime$ meson is of particular interest for various
reasons. The $\eta^\prime$ meson is thought to couple
strongly to gluons via the QCD anomaly coupling $\eta^\prime \rightarrow g + g$
\cite{Gluons}. Also, an unexpectedly large branching ratio measured recently for the
inclusive decay of beauty particles, $B \rightarrow \eta ^{\prime }+X$ \cite{CLEO}
has been interpreted as possible evidence for the strong coupling of $\eta^\prime$
meson to the gluonic components\cite{Gluonic}. It would then be conceivable that the
$pp \rightarrow pp\eta^\prime$ reaction might probe the gluon content of the
$\eta^\prime$ meson via its coupling to gluons emitted from the quarks exchanged
between two interacting nucleons. This mechanism would be complementary to the vector
meson-exchange current mechanism for producing $\eta^\prime$ mesons.
One of the properties of the $\eta ^{\prime }$ meson of
extreme importance is its yet-poorly-known coupling strength to the nucleon.
This has attracted much attention in connection with the so-called
''nucleon-spin crisis'' in polarized deep inelastic lepton scattering
\cite{EMC88}. The $NN\eta ^{\prime }$ coupling constant, $g_{NN\eta ^{\prime }}$,
may be related (through the axial vector coupling using the Goldberger-Treiman
relation) to the quark helicity contribution to the spin of the proton
\cite{Efremov}. Therefore one can argue that $g_{NN\eta ^{\prime }}$ tells us
about the total spin of the nucleon carried by its constituents; conversely, the
quark contribution to the spin of the nucleon would tell us about
$g_{NN\eta ^{\prime }}$. So far there is no direct experimental measurement of
$g_{NN\eta ^{\prime }}$. Recently, cross sections for the $pp \rightarrow
pp\eta ^{\prime }$ reaction near threshold have been measured by the COSY-11
\cite{Moskal} and SPESIII \cite{Hibou} collaborations. This reaction may offer
an opportunity to determine this coupling from a direct emission process of
$\eta^\prime$ by a proton. Of course, other production mechanisms, such as
meson exchange and nucleon resonance currents, must be taken into account
before a quantitative determination of $g_{NN\eta^\prime}$ is possible. The major
aim of the present work is to explore the roles of various production mechanisms.
The theory of $\eta ^{\prime }$ production in $pp$ collisions is still in its
early stage of development \cite{Sibirtsev,Bernard,Gedalin,Baru}. In this
work we investigate this reaction using a relativistic meson-exchange model of
hadronic interactions. In section 2 we outline the formalism for calculating the
production amplitude. The final state interaction is known to play a crucial role
in the production of particles near threshold energies in $NN$ collisions
\cite{Watson}; therefore our formalism includes the $pp$ final state interaction
explicitly. The Coulomb correction in the $pp$ final state interaction is also known
to be important when these protons have small energies \cite{Miller,Herrmann1}. In the
present work it is treated exactly and reduces the calculated cross section by as much
as a factor of two for energies close to threshold. The initial state interaction is
taken into account via a reduction factor \cite{Hanhart} determined from the available
phase shifts and inelasticities in the threshold incident energy region. In section 3 our
$\eta ^{\prime }$ meson production currents are constructed. We consider the nucleonic,
mesonic and resonance currents. The roles of these currents are explored in section 4.
Our results are summarized in Section 5.
%------------------------------------------------
\section{Formalism}
%------------------------------------------------
The formalism used in the present work is essentially the same as that
employed in Refs.\cite{Nak1,Nak2} for studying the production of vector mesons.
It is based on a relativistic meson-exchange model of hadronic interactions
in which the transition amplitude is calculated in the Distorted Wave Born
Approximation, taking explicitly into account effects of the final state
interaction. We write the transition amplitude describing the
$p+p\rightarrow p+p+\eta^{\prime }$ process as
%
\begin{equation}
M = < \phi_f | (T^{(-)\dagger}_f iG_f + 1) J | \phi_i > \ ,
\label{amplitude}
\end{equation}
%
where $\phi _{i,f}$ denotes the four-component unperturbed $pp$
wave function in the initial ($i$) and final ($f$) state. $T_{f}^{(-)}$ is
the final state $pp$ T-matrix. $G_{f}$ stands for the two-nucleon propagator
and $J$ is the $\eta ^{\prime }$-emission current defined in the next
section.
The T-matrix used in our calculation is generated by solving a
three-dimensional reduced Bethe-Salpeter equation (the Blankenbecler-Sugar
equation) for a relativistic one-boson-exchange $NN$ potential $V$, i.e.,
%
\begin{equation}
T = V + ViG_{BBS}T \ ,
\label{scateqn}
\end{equation}
%
where $G_{BBS}$ denotes the Blankenbecler-Sugar (BBS) two-nucleon propagator.
In this work we employ a slightly modified version \cite{Haidenbauer} of the
Bonn B $NN$ model as defined in Table A.1 of Ref.\cite{MHE87} for constructing
the potential $V$. This modification has been made in order to reproduce the $pp$
low-energy parameters rather than the $pn$ low-energy parameters while preserving
other features of the Bonn B interaction model for describing $pp$ scattering
\cite{Haidenbauer}. It should also be mentioned that each nucleon-nucleon-meson
($NNM$) vertex in the $NN$ potential is modified by a form factor of either
monopole or dipole form. We refer to \cite{MHE87} for further details.
Furthermore we note that the two-nucleon propagator $G_{f}$ appearing in
Eq.(\ref{amplitude}) is, for consistency, also chosen to be the BBS propagator,
$G_{f}=G_{f;BBS}$.
In Eq.(\ref{amplitude}) the initial state interaction is neglected. For heavy
meson production such as $\eta ^{\prime }$, its effect on the production cross
sections near threshold may be taken into account by multiplying the calculated
cross sections using Eq.(\ref{amplitude}) by a factor \cite{Hanhart}
%
\begin{equation}
\lambda_\alpha = \eta_\alpha \cos ^2(\delta_\alpha ) +
\frac{1}{4}[1-\eta_\alpha]^2
\label{isifactor}
\end{equation}
%
where $\alpha$ stands for the quantum numbers specifying the corresponding initial
$NN$ state and $\delta _{\alpha }$ and $\eta _{\alpha }$ denote the corresponding
$NN$ phase-shift and inelasticity, respectively, at the nucleon incident energy.
Although this reduction factor is only meant to account for the gross effect of the
initial state interaction, it works well when confronted with a calculation where
the initial state interaction has been treated explicitly \cite{lee}. In the near
threshold energy region the relevant initial $pp$ state is $\alpha=\ ^3P_0$. For
$\eta^\prime$ production the threshold incident energy is about $T_{lab}=2.404~GeV$.
Using the values $\delta _{(^3P_0)}=-50{{}^\circ}$ and $\eta _{(^3P_0)}=0.75$ from
a partial wave analysis \cite{said}, we obtain the reduction factor of
$\lambda_\alpha=0.33$ for the cross section. An alternative way to account
for effects of the initial state interaction is to absorb these effects in the
form factors at the $\eta ^{\prime }$ meson production vertices as has been done
in Refs.\cite{Nak1,Nak2}. We choose the former method in the present work.
%------------------------------------------------
\section{Production Currents}
%------------------------------------------------
Within our hadronic model of strong interactions, the $\eta ^{\prime }$-emission
current $J$ in Eq.(\ref{amplitude}) consists of a sum of the baryonic and mesonic
currents. The baryonic current is further divided into the nucleonic and nucleon
resonance ($N^*$) currents, so that the total current is given by
%
\begin{equation}
J = J_{nuc} + J_{res} + J_{mec} \ .
\label{curr}
\end{equation}
%
The individual currents are illustrated diagrammatically in Fig.~~\ref{fig1}. In the
following subsections we construct each of these currents.
%------------------------------------------------
\subsection{The Nucleonic Current}
%------------------------------------------------
The nucleonic current is defined as
%
\begin{equation}
J_{nuc} = \sum_{j=1,2}\left ( \Gamma_j iS_j U + U iS_j \Gamma_j \right ) \ ,
\label{nuc_cur}
\end{equation}
%
with $\Gamma _{j}$ denoting the $NN\eta ^{\prime }$ vertex and $S_{j}$ the nucleon
(Feynman) propagator for nucleon $j$. The summation runs over the two interacting
nucleons, 1 and 2. $U$ stands for the meson-exchange $NN$ potential. It is, in
principle, identical to the potential $V$ appearing in the $NN$ scattering equation,
except that here meson retardation effects (which are neglected in the potential
entering in Eq.(\ref{scateqn})) are kept as given by the Feynman prescription.
The structure of the $NN\eta ^{\prime }$ vertex, $\Gamma _{j}$, in
Eq.(\ref{nuc_cur}) is derived from the Lagrangian density
%
\begin{equation}
{\cal L}(x) = - g_{NN\eta^\prime} \bar\Psi_N(x) \gamma_5 \left[
\left( i\lambda + {1-\lambda \over 2m_N}\gamma^\mu\partial_\mu \right)
\eta^\prime(x) \right] \Psi_N(x) \ ,
\label{NNps}
\end{equation}
%
where $g_{NN\eta ^{\prime }}$ denotes the $NN\eta ^{\prime }$ coupling constant and
$\lambda $ is the parameter controlling the pseudoscalar(ps) - pseudovector(pv)
admixture. $\eta ^{\prime }(x)$ and $\Psi_N(x)$ stand for the $\eta ^{\prime }$ and
nucleon field, respectively; $m_{N} $ denotes the nucleon mass.
As mentioned in the introduction, the coupling constants $g_{NN\eta ^{\prime}}$ and
the ps-pv mixing parameter $\lambda$ are poorly known at present. The predictions
for $g_{NN\eta ^{\prime}}$ range anywhere from 1.9 to 7.5
\cite{Dumbrajs,Zhang,Bernard}; an estimate based on the dispersion method even gives
$g_{NN\eta^\prime}$ consistent with zero \cite{Grein}. Zhang et al.\cite{Zhang} in
their analysis of the photoproduction on protons of $\eta^\prime$ mesons used
$\lambda=1$ for the ps-pv mixing parameter. Bernard et al. \cite{Bernard}, in their
analysis of the $pp \rightarrow pp\eta^\prime$ process, extracted a value of
$\lambda = 0.4 \pm 0.1$ in conjunction with the value of $g_{NN\eta^\prime} = 2.5 \pm
0.7$ determined from a recent measurements of $\Delta\Sigma$ in deep inelastic
lepton-nucleon scattering \cite{Altarelli}. The latter quantity is related to
$g_{NN\eta ^{\prime}}$ \cite{Efremov}. SU(3) symmetry, together with the OZI rule
\cite{OZI}, relates the $NN\eta^\prime$ coupling to the $NN\eta $ coupling:
$g_{NN\eta ^{\prime }}=g_{NN\eta }$ tan($\alpha_{P}$), where $\alpha _{P}\equiv
\theta _{P}-\theta _{P(ideal)}\simeq -45{{}^\circ}$ denotes the deviation from the
pseudoscalar ideal mixing angle. With the value of $g_{NN\eta }=6.14$ used in $NN$
scattering analysis \cite{MHE87}, we then have $g_{NN\eta ^{\prime }}\simeq 6.1$
which is close to the upper end of the predicted range mentioned above. The value
of $g_{NN\eta }=6.14$ together with the $\eta -\eta^\prime$ mixing angle of
$\theta _{P}\simeq -9.7{{}^\circ}$, as suggested by the quadratic mass formula, and
the $NN\pi $ coupling constant of $g_{NN\pi }=13.45$ leads to the ratio $D/F\simeq
1.43$. This is not too far from the value of $D/F\cong 1.73$ extracted from a systematic
analysis of semileptonic hyperon decays \cite{DFratio}. In the present work we use
the value of $g_{NN\eta ^{\prime }}=6.1$. We shall consider both of the extreme values
of the parameter $\lambda$ in Eq.(\ref{NNps}), i.e., $\lambda=0$ and $1$. Note that
$\eta^\prime$ is {\em not} a Goldstone boson. Consequently there is, {\em a priori},
no constraint on the ps-pv admixture.
The $NN\eta^\prime$ vertex derived from Eq.(\ref{NNps}) should be provided with an
off-shell form factor. Following Ref.\cite{Nak2}, it is assumed to be of the form
%
\begin{equation}
F_N(l^2) = \frac{\Lambda_N^4} {\Lambda_N^4 + (l^2-m_N^2)^2} \ ,
\label{formfactorN}
\end{equation}
%
where $l^{2}$ denotes the four-momentum squared of either the incoming or outgoing
off-shell nucleon. We also introduce the form factor given by Eq.(\ref{formfactorN})
at those $NNM$ vertices appearing next to the $\eta^{\prime }$-production vertex,
where the (intermediate) nucleon and the exchanged mesons are off their mass shell
(see Fig.~\ref{fig1}). Therefore, the corresponding form
factors are given by the product $F_{N}(l^{2})F_{M}(q_{M}^{2})$, where $M$ stands for
each of the exchanged mesons between the two interacting nucleons. The form factor
$F_{M}(q_{M}^{2})$ accounts for the off-shellness of the exchanged meson and is
taken consistently with the $NN$ potential used for generating the T-matrix. In
Ref.\cite{Nak2} cutoff parameters in the range of $\Lambda _{N} = 1.17-1.55 ~GeV$
have been used in the description of $\phi$- and $\omega$-meson productions in $pp$
collisions. In order to reduce the number of free parameters, here, we adopt the value
of $\Lambda_N=1.2 ~GeV$, which is also the value adopted at the $NN^*\eta^\prime$
vertex in the construction of the resonance current in the next subsection.
%------------------------------------------------
\subsection{The Resonance Current}
%------------------------------------------------
Currently there are no well established nucleon resonances that decay into $N\eta^\prime$.
In Ref.\cite{Zhang} the strong peak in
the total cross section of the $\eta^\prime$ photoproduction off protons close to
threshold has been attributed to the $D_{13}(2080)$ resonance. On the other hand,
a recent multipole analysis of the $\eta^\prime$ photoproduction \cite{Plotzke}
indicates that this reaction is dominated by an $S_{11}$ and a $P_{11}$ resonance.
In Ref.\cite{Plotzke}, apart from the resonance dominance assumption---which probably
leads to an overestimation of the total decay width of the resonances considered,
it is also assumed that these resonances decay only into $p\eta^\prime$ and
$p\pi^o$. The latter decay channel also effectively accounts for all other
decay channels that were not considered explicitly in the analysis due to lack of
information. Although this may not affect appreciably the description of the
$\eta^\prime$ photoproduction reaction, the $pp \rightarrow pp\eta^\prime$ process
can be very sensitive to such an assumption. Since the pion couples strongly to the
nucleon, an overestimation of the $p\pi^o$ partial decay width is likely to lead to a
considerable overestimation of the cross section due to these resonances in the
$pp \rightarrow pp\eta^\prime$ reaction. In addition to all these uncertainties
in the extraction of the resonance parameters, there is also the
possibility of a threshold cusp effect, as discussed in Ref.\cite{Hoehler}, that
might explain the observed behavior of the photoproduction cross section close to
threshold in the absence of any resonance. All these issues require further careful
considerations. With this background in mind we will explore the
role of the $S_{11}(1897)$ and $P_{11}(1986)$ resonances as determined in
Ref.\cite{Plotzke}.
The resonance current, in analogy to the nucleonic current, is written as
%
\begin{equation}
J_{res} = \sum_{j=1,2}\sum_{N^*}\left ( \Gamma_{jN^*\eta^\prime} iS_{N^*} U_{N^*} +
\tilde U_{N^*} iS_{N^*} \Gamma_{jN^*\eta^\prime} \right ) \ .
\label{res_cur}
\end{equation}
%
Here $\Gamma _{jN^*\eta^\prime}$ stands for the $NN^*\eta ^{\prime }$ vertex
function involving the nucleon $j$. $S_{N^*}(p) = ({\mbox{$ \not \! p$}} + m_{N^*}) /
(p^2 - m_{N^*}^2 + im_{N^*}\Gamma_{N^*})$ is the $N^*$ resonance propagator, with
$m_{N^*}$ and $\Gamma_{N^*}$ denoting the mass and width of the resonance,
respectively. The summation runs over the two interacting nucleons, $j=1$ and $2$,
and also over the resonances considered, i.e., $N^* = S_{11}(1897),~P_{11}(1986)$.
In the above equation $U_{N^*}$ ($\tilde U_{N^*}$) stands for the
$NN\rightarrow NN^*$ ($NN^*\rightarrow NN$) meson-exchange transition potential.
It is given by
%
\begin{equation}
U_{N^*} = \Gamma_{NN^*\pi}(q_\pi) i\Delta_\pi(q_\pi^2) \Gamma_{NN\pi}(q_\pi) \ ,
\label{transpot}
\end{equation}
%
where $\Delta_\pi(q_\pi^2)$ denotes the (Feynman) propagator of the exchanged pion
with four-momentum $q_\pi$ and $\Gamma_{NN\pi}(q_\pi)$ the $NN\pi$ vertex. The
latter is taken consistently with the $NN$ potential $V$ appearing in
Eq.(\ref{scateqn}). We note that there is also an additional contribution to
$U_{N^*}$ from the $\eta^\prime$ exchange in Eq.(\ref{transpot}). However this is
negligible compared to the pion-exchange contribution.
Following Ref.\cite{Benmer}, the $NN^*\eta^{\prime }$ and $NN^*\pi$ vertices in
Eqs.(\ref{res_cur},\ref{transpot}) are obtained from the interaction Lagrangian
densities
%
\begin{eqnarray}
{\cal L}^{(\pm)}_{NN^*\eta^\prime}(x) & = & - g_{NN^*\eta^\prime}
\bar \Psi_N(x) \left\{ \left[ i\lambda\Gamma^{(\pm)} +
\left({1 - \lambda\over m_{N^*}\pm m_N}\right)\Gamma^{(\pm)}_\mu \partial^\mu
\right]\eta^\prime(x)\right\} \Psi_{N^*}(x) + h.c. \nonumber \\
{\cal L}^{(\pm)}_{NN^*\pi}(x) & = & - g_{NN^*\pi}
\bar \Psi_N(x) \left\{ \left[ i\lambda\Gamma^{(\pm)} +
\left({1 - \lambda\over m_{N^*}\pm m_N}\right)\Gamma^{(\pm)}_\mu \partial^\mu
\right]\vec\tau \cdot \vec\pi(x)\right\} \Psi_{N^*}(x) + h.c. \ ,
\label{NN^*ps}
\end{eqnarray}
%
where $\vec\pi(x)$ and $\Psi_{N^*}(x)$ denote the pion and nucleon resonance field,
respectively. The upper and lower signs refer to the even($+$) and odd($-$)
parity resonance, respectively. The operators $\Gamma^{(\pm)}$ and
$\Gamma^{(\pm)}_\mu$ in the above equation are given by
%
\begin{eqnarray}
\Gamma^{(+)} = \gamma_5 & \ , & \Gamma^{(+)}_\mu = \gamma_\mu\gamma_5 \nonumber \\
\Gamma^{(-)} = 1 & \ , & \Gamma^{(-)}_\mu = \gamma_\mu \ .
\label{NRpsa}
\end{eqnarray}
%
The parameter $\lambda$ in Eq.(\ref{NN^*ps}) controls the admixture of the two
types of couplings: ps ($\lambda=1$) and pv ($\lambda=0$) in the case of an even
parity resonance and, scalar ($\lambda=1$) and vector ($\lambda=0$) in the case of
an odd parity resonance. On-shell, both choices of the parameter $\lambda$ are
equivalent. Hereafter the mixing parameter $\lambda$ for the $NN^*\pi$ vertex is
fixed to be $\lambda=0$. We consider the choices $\lambda=0$ and $1$ in the
$NN^*\eta^\prime$ vertex, however.
The coupling constants $g_{NN^*\eta^\prime}$ and $g_{NN^*\pi}$ are determined from the
extracted decay widths (and masses) of the resonances from the $\eta^\prime$
photoproduction on protons \cite{Plotzke},
%
\begin{eqnarray}
(m_{S_{11}}\ ,\ \Gamma_{S_{11}}) & = & (1.897\pm0.050^{+0.030}_{-0.002}\ ,\
0.396\pm0.155^{+0.035}_{-0.045})\ GeV \nonumber \\
(m_{P_{11}}\ ,\ \Gamma_{P_{11}}) & = & (1.986\pm0.026^{+0.010}_{-0.030}\ ,\
0.296\pm0.100^{+0.060}_{-0.010})\ GeV \ ,
\label{resmw}
\end{eqnarray}
%
with partial decay widths
%
\begin{eqnarray}
\Gamma(S_{11} \rightarrow p\eta^\prime) = 0.05\Gamma_{S_{11}} & \ ,\ &
\Gamma(S_{11} \rightarrow p\pi^o) = 0.95\Gamma_{S_{11}} \nonumber \\
\Gamma(P_{11} \rightarrow p\eta^\prime) = 0.25\Gamma_{S_{11}} & \ ,\ &
\Gamma(P_{11} \rightarrow p\pi^o) = 0.75\Gamma_{S_{11}} \ .
\label{partdecay}
\end{eqnarray}
%
Using these values we obtain
%
\begin{eqnarray}
g_{NS_{11}\eta^\prime} = 2.9 &\ ,\ & g_{NS_{11}\pi} = 2.4 \nonumber \\
g_{NP_{11}\eta^\prime} = 11.7 &\ ,\ & g_{NP_{11}\pi} = 5.2 \ .
\label{NRpscoupl}
\end{eqnarray}
%
Here we assume the coupling constants to be positive. We mention in advance
that the interference of the resonance current with the rest of the currents
(nucleonic and mesonic currents) is practically negligible.
In complete analogy to the nucleonic current, we introduce the off-shell form
factors at each vertex involved in the resonance current. We adopt the same
form factor given by Eq.(\ref{formfactorN}), with $m_N$ replaced by $m_{N^*}$
at the $NN^*\eta^\prime$ and $NN^*\pi$ vertices, in order to account for the
off-shellness of the $N^*$ resonances. We note that Zhang et al. \cite{Zhang}
have also employed this form for the form factor at the $NN^*\eta^\prime$
vertex in their study of $\eta^\prime$ photoproduction. In order to account
for the off-shellness of the exchanged pion (see Eq.(\ref{transpot})), the
$NN^*\pi$ vertex is also multiplied by an extra form factor of $F_\pi(q_\pi^2) =
(\Lambda_\pi^2 - m_\pi^2) / (\Lambda_\pi^2 - q_\pi^2)$, the same form used at
the $NN\pi$ vertex in the construction of the $NN$ potential $V$ entering in
Eq.(\ref{scateqn}), with the cutoff parameter of $\Lambda_\pi=0.8 ~GeV$. We also
use the same form factor at the $NN\pi$ vertex in Eq.(\ref{transpot}).
%------------------------------------------------
\subsection{The Mesonic Current}
%------------------------------------------------
For the meson-exchange current we consider the contribution from the
$vv\eta^\prime$ vertex with $v$ denoting either a $\rho $ or $\omega $ meson.
As we shall show later, this gives rise to the dominant meson-exchange current.
The $vv\eta^\prime$ vertex required for constructing the meson-exchange current
is derived from the Lagrangian densities
%
\begin{eqnarray}
{\cal L}_{\rho\rho\eta^\prime}(x) & = & - \frac{g_{\rho\rho\eta^\prime}}{2 m_\rho}
\varepsilon_{\alpha\beta\nu\mu} \partial^\alpha \vec \rho^\beta(x) \cdot
\partial^\nu \vec \rho^\mu(x) \eta^\prime(x) \nonumber \\
{\cal L}_{\omega\omega\eta^\prime}(x) & = & - \frac{g_{\omega\omega\eta^\prime}}
{2 m_\omega}\varepsilon_{\alpha\beta\nu\mu} \partial^\alpha \omega^\beta(x)
\partial^\nu \omega^\mu(x) \eta^\prime(x) \ ,
\label{vveta'}
\end{eqnarray}
%
where $\varepsilon _{\alpha \beta \nu \mu }$ is the Levi-Civita antisymmetric
tensor with $\varepsilon _{0123}=+1$. The vector meson-exchange current is then
given by
%
\begin{equation}
J_{vv\eta^\prime} = \sum_{v=\rho ,\omega}\left\{[\Gamma^\alpha_{NNv}(k_v)]_1
iD_{\alpha\beta}(k_v)
\Gamma^{\beta\mu}_{vv\eta^\prime}(k_v, k^\prime_v)
iD_{\mu\nu}(k^\prime_v) [\Gamma^\nu_{NNv}(k^\prime_v)]_2
\right\} \ ,
\label{vveta'_cur}
\end{equation}
%
where $D_{\alpha \beta }(k_{v })$ and $D_{\mu\nu}(k^\prime_v)$ stand for the (Feynman)
propagators of the two exchanged vector mesons (either the $\rho $ or $\omega $ mesons as
$v =\rho $ or $\omega $) with four-momentum $k_{v }$ and $k^\prime_v$, respectively. The
vertices $\Gamma $ involved in the above equation are self-explanatory. The $NNv $ vertex
$\Gamma _{NNv }^{\mu}(v =\rho ,\omega )$ is taken consistently with those in the potential
used for constructing the $NN$ T-matrix in Eq.(\ref{scateqn}).
The coupling constant $g_{vv\eta^\prime}$ is determined from a systematic analysis of the
radiative decay of pseudoscalar and vector mesons in conjunction with
vector-meson dominance. This is done following Refs.\cite{Durso,Nak2}, with the aid of an
effective Lagrangian with SU(3) flavor symmetry and imposition of the OZI rule. The
parameters of this model are the angle $\alpha_V (\alpha _{P})$, which measures the
deviation from the vector(pseudoscalar) ideal mixing angle, and the coupling constant of
the effective SU(3) Lagrangian. They are determined from a fit to radiative decay of
pseudoscalar and vector mesons. The parameter values determined in this way in
Ref.\cite{Durso} (model B), however, overpredict the measured radiative decay width of
the $\eta^\prime$ meson\cite{PDG}. Therefore, we have readjusted slightly the value of
the coupling constant of the SU(3) Lagrangian in order to reproduce better the measured
width. We have $\alpha _{V}\cong 3.8{{}^{\circ }}$ and $\alpha _{P}\cong -45{{}^{\circ }}$,
as given by the quadratic mass formula, and the coupling constant of the effective SU(3)
Lagrangian of $G=7$ in units of $1/\sqrt{m_{v }m_{v }^{\prime }}$, where $m_{v }$ and
$m_v^\prime$ stand for the mass of the two vector-mesons involved. The sign of the
coupling constant $G$ is consistent with the sign of the $\rho\pi\gamma$ and
$\omega\pi\gamma$ coupling constants taken from an analysis of the pion photoproduction
data in the $\sim 1~GeV$ energy region \cite{Garc}. With these parameter values we obtain
%
\begin{eqnarray}
g_{\rho\rho\eta^\prime} & = & - G \sin(\alpha_P) = 4.95 \nonumber \\
g_{\omega\omega\eta^\prime} & = & -G\left( \sqrt{2}\sin^2(\alpha_V)\cos(\alpha_P)
+ \cos^2(\alpha_V)\sin(\alpha_P) \right) = 4.90 \ .
\label{meccoupl}
\end{eqnarray}
%
The $vv \eta ^{\prime }$ vertex $(v =\rho ,\omega )$ in Eq.(\ref{vveta'_cur}), where the
exchanged vector mesons are both off their mass shells, is accompanied by a form factor.
Following Ref.\cite{Nak2}, we assume the form
%
\begin{equation}
F_{vv\eta^\prime}(k_v^2, {k^\prime_v}^2) =
\left( \frac{\Lambda_v^2 - m_v^2} {\Lambda_v^2 - k_v^2} \right)
\left ( \frac{\Lambda_v^2} {\Lambda_v^2 - {k^\prime_v}^2} \right ) \ .
\label{formfactorM}
\end{equation}
%
It is normalized to unity at $k_{v }^{2}=m_{v }^{2}$ and $k_{v }^{\prime \text{ }2}=0$,
consistent with the kinematics at which the value of the coupling constant
$g_{vv\eta^\prime}$ was determined. We adopt the cutoff parameter value of $\Lambda _{v}
= 1.45 ~GeV$ as determined in Ref.\cite{Nak2} from the study of the $\omega$ and
$\phi$ meson production in $pp$ collisions.
Another potential candidate for mesonic current is the $\sigma\eta\eta ^{\prime }$-exchange
current, whose coupling constant may be estimated from the decay width of $%
\eta ^{\prime }$ into an $\eta $ and two pions. We take the Lagrangian densities
%
\begin{eqnarray}
{\cal L}_{\sigma\eta\eta^\prime}(x) & = &
{g_{\sigma\eta\eta^\prime}\over \sqrt{m_\eta m_{\eta^\prime}}}
\sigma(x)\partial_\mu\eta(x)\partial^\mu\eta^\prime(x) \nonumber \\
{\cal L}_{\sigma\pi\pi}(x) & = &
{g_{\sigma\pi\pi}\over 2m_\pi}
\sigma(x)\partial_\mu\vec\pi(x)\cdot\partial^\mu\vec\pi(x) \ ,
\label{sigmapsps}
\end{eqnarray}
%
where $\sigma (x)$ and $\eta (x)$ stand for the $\sigma $ and $\eta $ meson field,
respectively, and $m_\pi$, $m_{\eta }$ and $m_{\eta ^{\prime }}$ stand for the masses of
the $\pi$, $\eta $ and $\eta ^{\prime }$ meson. With the coupling constant of
$g_{\sigma \pi \pi} \sim 1.3$, extracted from Ref.\cite{Schutz} in conjunction with
the $NN\sigma$ coupling constant used in the $NN$ potential $V$, and assuming that the
measured decay width of $\Gamma (\eta ^{\prime }\rightarrow \eta +2\pi ^{o})=42KeV$
\cite{PDG} is entirely due to the $\sigma $ meson intermediate state, we obtain an
(upper) estimate of $\left| g_{\sigma \eta \eta ^{\prime }} \right| \sim 1.14$. The
sign of this coupling constant is not fixed. We shall consider both possibilities.
The $\sigma\eta\eta^\prime$ current reads
%
\begin{equation}
J_{\sigma\eta\eta^\prime} = \left\{[\Gamma_{NN\sigma}]_1
i\Delta_\sigma(q_\sigma^2)
\Gamma_{\sigma\eta\eta^\prime}(q_\sigma , q_\eta)
i\Delta_\eta(q_\eta^2) [\Gamma_{NN\eta}(q_\eta)]_2
\right\} + (1 \leftrightarrow 2) \
\label{sigmaetaeta'_cur}
\end{equation}
%
in our previously-defined notation. The vertices
$\Gamma_{NN\sigma}$ and $\Gamma_{NN\eta}$ are taken consistently with those
in the $NN$ potential $V$ in Eq.(\ref{scateqn}). The $\sigma\eta\eta^\prime$
vertex, $\Gamma_{\sigma\eta\eta^\prime}(q_\sigma , q_\eta)$, is provided with
the same form factor given by Eq.(\ref{formfactorM}), except for the replacement
$m_v \rightarrow m_\eta$.
The total mesonic current is then given by
%
\begin{equation}
J_{mec} = J_{vv\eta^\prime} + J_{\sigma\eta\eta^\prime} \ .
\label{mec_cur}
\end{equation}
There are, of course, other possible mesonic currents, such as the $\omega \phi
\eta ^{\prime }$- and $\phi \phi \eta ^{\prime }$-exchange currents, that
contribute to $\eta ^{\prime }$ meson production in $pp$ collisions.
Their contributions can be estimated in a systematic way following Ref.\cite{Nak2}
and are found to be negligible.
%------------------------------------------------
\section{Numerical Results and Discussion}
%------------------------------------------------
Once all the ingredients are specified, the total cross section for the
reaction $p+p\rightarrow p+p+\eta ^{\prime }$ can be calculated. We will
first analyze each current separately, exploring the role of individual
contributions and the uncertainties associated with them, and then combine
them later in the section.
The $N^*$ resonance current contribution to the total cross section as a function
of excess energy $Q$ is shown in Fig.~\ref{fig2}. The excess energy is defined as
$Q\equiv \sqrt{s}-\sqrt{s_{o}}$, where $\sqrt{s}$ denotes the total center-of-mass
energy of the system and $\sqrt{s_{o}} \equiv 2m_{N}+m_{\eta^{\prime }}$, its
$\eta ^{\prime }$-production threshold energy. Here we also display the recent
experimental data from the COSY-11 \cite{Moskal} and SPESIII \cite{Hibou}
collaborations. The mixing parameter $\lambda$ in Eq.(\ref{NN^*ps}) for the
$NN^*\pi$ vertices is fixed to be $\lambda=0$. Hereafter $N^*$ stands for both the
$S_{11}(1897)$ and $P_{11}(1986)$ resonances unless otherwise indicated. The upper
panel shows the results with $\lambda =0$ in the $NN^*\eta^\prime$ vertices. We see
that the cross section is largely dominated by the $S_{11}(1897)$ contribution
(dash-dotted line). The $P_{11}(1986)$ resonance leads to an energy dependence (dashed
line) of the cross section steeper than that of the $S_{11}(1897)$ resonance, although
in the excess energy region considered its contribution is practically negligible. The
solid line corresponds to the total contribution, which is close to the data.
The lower panel corresponds to the choice of $\lambda=1$ at
the $NN^*\eta^\prime$ vertices. Compared to the case of $\lambda=0$, we see that the
mixing parameter has practically no influence on the $S_{11}(1897)$ contribution. This
is due to the fact that this resonance is practically on its mass shell in the
near-threshold energy region, so that both choices of the parameter $\lambda$ are
equivalent. Although still very small compared to the $S_{11}(1897)$ contribution, the
$P_{11}(1986)$ resonance contribution increases substantially in the low excess energy
region compared to the $\lambda=0$ case. This is because this resonance is off-shell
and the ps coupling mixes efficiently the positive and negative energy states. The pv
coupling ($\lambda=0$) suppresses this mixing. The interference with the $S_{11}(1897)$
contribution is now destructive. In this case the total contribution (solid curve)
is slightly reduced compared to that in the upper panel. These results are consistent
with the recent analysis of $\eta^\prime$ photoproduction \cite{Plotzke}. We recall that
the coupling constants at the $NN^*\eta^\prime$ and $NN^*\pi$ vertices entering in the
resonance current have been determined from the decay widths extracted in Ref.\cite{Plotzke}.
Although we find the consistency between our result above for $pp \rightarrow
pp\eta^\prime$ and the $\eta^\prime$ photoproduction analysis of Ref.\cite{Plotzke},
care must be taken in drawing any premature conclusions about the basic $\eta^\prime$
production mechanisms in these reactions. As mentioned in subsection III.B, given the
assumptions made in Ref.\cite{Plotzke} in the extraction of the partial decay widths,
the corresponding coupling constants, $g_{NN^*\eta^\prime}$ and $g_{NN^*\pi}$, are
subject to considerable uncertainties. In particular, the resonance dominance assumption
does not allow any other possible production mechanism. As we shall show below,
the vector-meson exchange current yields also a cross section comparable to the
data. This indicates that the resonance current contribution shown in Fig.\ref{fig1} is
probably overestimated. Another source of uncertainty in the resonance current arises
from the off-shell form factors introduced at the $NN^*\pi$ vertex. Although the dominant
$S_{11}(1897)$ resonance contribution is insensitive to the form factor at the
$NN\eta^\prime$ vertex because the resonance is nearly on its mass shell near the threshold
energy, the corresponding exchanged pion (see Fig.\ref{fig1}) is far off-shell. This
makes the resonance contribution very sensitive to the form factor involving this pion.
In the present calculation we have employed the same monopole form factor used at the
$NN\pi$ vertex entering in the $NN$ potential $V$ in Eq.(\ref{scateqn}) with the cutoff
parameter of $\Lambda_\pi=0.8 ~GeV$. One might then argue that, within the resonance
dominance assumption, the $\gamma p \rightarrow p\eta^\prime$ and $pp \rightarrow
pp\eta^\prime$ processes can be described consistently---provided that one uses a soft form
factor at the $NN^*\pi$ vertex. Whatever the arguments, it is clear that further
investigation is required in order to determine better the relevant coupling
constants before a more definite conclusion can be drawn about the role of the nucleon
resonance current in the $pp \rightarrow pp\eta^\prime$ reaction. Of course one should
also keep in mind the possibility that the strong peak in the cross section observed in
$\eta^\prime$ photo-production close to threshold may be due to a cusp effect
\cite{Hoehler} and {\em not} due to a resonance.
We turn now to investigate the mesonic current contribution, which is displayed in
Fig.~\ref{fig3}. Among the exchange currents considered, the
$\rho\rho\eta^\prime$ exchange gives rise to the dominant contribution (dashed line).
The $\omega\omega\eta^\prime$- and $\sigma\eta\eta^\prime$-exchange current
contributions (dotted and dash-dotted line, respectively) are of similar magnitude.
However they are smaller by about a factor of five compared to the
$\rho\rho\eta^\prime$-exchange contribution. The slightly different energy dependence
of the cross section resulting from the $\rho\rho\eta^\prime$- and
$\omega\omega\eta^\prime$-exchange currents is due to the tensor coupling in the
$NN\rho$ vertex. This coupling is absent in the $NN\omega$ vertex used here. As
mentioned in subsection III.C, the sign of the $\sigma\eta\eta^\prime$-exchange current
relative to $vv\eta^\prime$-exchange currents ($v=\rho, \omega$) is not fixed; therefore
we consider both possibilities. Assuming the negative sign for
$g_{\sigma\eta\eta^\prime}$, the interference with the $vv\eta^\prime$-exchange
current is destructive, yielding the total contribution represented by the solid
line. If $g_{\sigma\eta\eta^\prime}$ is positive, then the interference is
constructive and gives the total contribution shown by the dotted line. Although
this latter choice leads to an overprediction of the data, we cannot exclude it
at this stage because, as we shall show below, the nucleonic current can interfere
destructively with the mesonic current so that the combined contribution may be
smaller than that of mesonic current alone. Moreover, the off-shell form factor
at the $\eta^\prime$-production vertices in the mesonic current is not
well-determined at present. As mentioned in subsection III.C, we employ the same
form factors at the $\pi\rho\omega$ and $\pi\rho\phi$ vertices as in the study of
$\omega$- and $\phi$-meson production in $pp$ collisions \cite{Nak2}, which are
themselves subject to considerable uncertainties. Indeed, a reduction of about
$10\%$ in the value of the cutoff parameter can bring the short-dashed curve in
Fig.~\ref{fig3} onto the data. Therefore neither sign of $g_{\sigma\eta\eta^\prime}$
can be excluded from the present analysis. In this regard, measurements of cross
sections at higher excess energies may be useful in deciding between the two choices
of the sign in the $\sigma\eta\eta^\prime$ coupling.
In the upper panel of Fig.~\ref{fig4}, the nucleonic current contribution (dashed
curve) is shown with pv coupling at the $NN\eta^\prime$ vertex. The contribution
from the mesonic current with the choice of positive sign for
$g_{\sigma\eta\eta^\prime}$ (dashed curve in Fig.~\ref{fig3}) is also shown here as
the dash-dotted curve. As we can see, the nucleonic current is about two orders of
magnitude smaller than the mesonic current. Their interference is constructive and
results in the total contribution shown by the solid curve, which overpredicts the data.
The lower panel of Fig.~\ref{fig4} shows the same results, except that ps coupling is
used at the $NN\eta^\prime$ vertex in the nucleonic current contribution. Compared to
the case of pv coupling, we observe a considerable enhancement of the nucleonic current.
This is due to the strong admixture of the positive and negative energy nucleon states
provided by the ps coupling; with pv coupling this mixing is suppressed. Furthermore
the interference with the mesonic current is now destructive, resulting in a total
contribution (solid line) that is substantially smaller than the mesonic current
contribution alone.
Fig.~\ref{fig4a} shows the same results as in Fig.~\ref{fig4}, except that here the
mesonic current contribution corresponds to the choice of negative sign for
$g_{\sigma\eta\eta^\prime}$ (solid curve in Fig.~\ref{fig3}). We see a similar feature
to that observed in Fig.~\ref{fig4}, yielding net cross sections that underpredict
the data. In particular, we see a strong destructive interference between the mesonic
and nucleonic currents in the case of the ps coupling at the $NN\eta^\prime$ vertex
(lower panel).
We observe that the effect of the nucleonic current may be overestimated here because we
have used a rather large $NN\eta^\prime$ coupling constant, as discussed in subsection III.A.
In any case, given the fact that the nucleonic current is small compared to other currents,
a quantitative determination of the $NN\eta^\prime$ coupling constant using the $pp
\rightarrow pp\eta^\prime$ reaction is unlikely to be possible---at least not before the
other dominant currents are better understood.
The above considerations expose the role of individual currents as well as the
current uncertainties associated with them in the description of the $pp \rightarrow
pp\eta^\prime$ reaction. We combine now all the currents. In the following we consider
all the possibilities for the mesonic-plus-nucleonic current contribution as shown in
Figs.~\ref{fig4} and \ref{fig4a}. For each scenario considered we adjust the coupling
constants $g_{NN^*\eta^\prime}$ and $g_{NN^*\pi}$ in the resonance current so as to
reproduce the data, taking $\lambda=1$ in the $NN^*\eta^\prime$ vertices. The results
are shown in Fig.~\ref{fig5}. For the case of the mesonic-plus-nucleonic current shown
in the upper panel of Fig.~\ref{fig4}, the corresponding resonance contribution is set
to zero because the mesonic-plus-nucleonic current already overpredicts the data and
the resonance current can only enhance the cross section further. We recall that the
interference between the resonance and other currents is practically negligible.
Therefore we have here adjusted the ps-pv mixing parameter in the $NN\eta^\prime$ vertex
to $\lambda=0.7$ in order to reproduce the data, and the corresponding result is shown in
the solid curve of Fig.~\ref{fig5}. For the choice of the mesonic-plus-nucleonic current
as shown in the lower panel of Fig.~\ref{fig4}, the resonance current is required in order
to reproduce the data. We have adjusted this current contribution by multiplying the
product of the coupling constants $g_{NN^*\eta^\prime} g_{NN^*\pi}$ for both
$N^*=S_{11}(1897)$ and $P_{11}(1986)$ resonances by an arbitrary reduction factor of $1/2$.
The result is shown by the dash-dotted line. For the scenario of the mesonic-plus-nucleonic
current shown in the upper panel of Fig.~\ref{fig4a}, the reduction factor of the product
$g_{NN^*\eta^\prime}g_{NN^*\pi}$ required in the resonance current is $0.625$. The
corresponding result is shown by the dotted curve. Finally, for the scenario of the lower
panel in Fig.~\ref{fig4a}, no reduction of the resonance current is required in order to
reproduce the data, since this corresponds to an extreme case where the destructive
interference between the nucleonic and mesonic currents leads to a mesonic-plus-nucleonic
current contribution that is nearly negligible. The short-dashed curve is the corresponding
result. Fig.~\ref{fig5} reveals the different energy dependences arising from each scenario
considered. We see that if one wants to learn about the production mechanism from the total
cross section then one has to go to much higher energy region than that covered by the
presently available data.
At this point we note that, in view of the latter scenario discussed in Fig.\ref{fig5} above,
it is not surprising that the approach of Ref.\cite{Sibirtsev} describes the $pp \rightarrow
pp\eta^\prime$ data. In Ref.\cite{Sibirtsev} the $\pi N \rightarrow \eta^\prime N$ amplitude,
extracted from the measured cross section, has been used as the production mechanism. The
$\pi N \rightarrow \eta^\prime N$ amplitude obtained in this way must include the nucleon
resonance contribution that we have considered in this work, which leads to a strong $\pi N
\rightarrow \eta^\prime N$ transtion. We emphasize, however, that one has to be careful in
drawing any premature conclusion here. First, as discussed before, our nucleon resonance current
is subject to considerable uncertainties. Second, although the calculation of Ref.\cite{Sibirtsev}
is based on the measured $\pi N \rightarrow N \eta^\prime$ cross section, it requires an
introduction of an off-shell form factor at the $NN\pi$ vertex in order to account for the
off-shellness of the exchanged pion. As we have pointed out, the exchanged pion is far off-shell
in the present reaction and consequently the resulting cross sections are very sensitive to the
corresponding form factor at the $NN\pi$ vertex. We note that in Ref.\cite{Sibirtsev} a monopole
form factor with the cutoff parameter of $\Lambda_\pi = 1.3~GeV$, as used in the Bonn $NN$ potential,
has been employed. There are, however, a number of indications that the $NN\pi$ form factor is much
softer ($\Lambda_\pi \sim 0.8~GeV$) than that used in the Bonn potential \cite{Lattice,Bockman}. This
would reduced the contribution of the one-pion exchange to the $pp \rightarrow pp\eta^\prime$ process
substantially. The above consideration gives an idea of the kind of uncertanties involved in the
theoretical predictions.
We now focus our attention on the excess energy region close to threshold where
the data exist. Fig.~\ref{fig6} shows the results of Fig.~\ref{fig5} in the
range of excess energy up to $Q=10~MeV$. Here, only one total current contribution
(solid curve in Fig.~\ref{fig5}) is shown since all four scenarios considered
lead to almost indistinguishable results at the scale used in the limited energy
domain considered in the figure. We note that the energy dependence of the cross
section is not reproduced. In particular, after adjusting the ps-pv mixing parameter
or the reduction factor of $g_{NN^*\eta^\prime} g_{NN^*\pi}$ in order to reproduce the
higher energy data points, as has been done in Fig.~\ref{fig5},
we are not able to describe the lowest two data points. We observe that the data have
considerable uncertainties in the excess energy, presumably due to background subtraction
that becomes more difficult for energies very close to the threshold. Nevertheless the
results indicate the need for some mechanism(s) that introduces an additional energy
dependence of the total cross section very close to threshold. The energy dependence of
the total cross section near threshold is basically determined by the final state $NN$
interaction \cite{Watson}. This is illustrated by the dotted line in Fig.~\ref{fig6},
which corresponds to the result calculated taking into account only the
three-body phase space-plus-final state interaction; it has been normalized to the solid
curve in the low excess energy region. We see that the production currents start to modify
the energy dependence given by the phase space-plus-final state interaction only for
energies above $Q=4-5~MeV$. Therefore the predicted energy dependence in the near-
threshold region is unlikely to be modified by the {\it basic} production mechanism(s).
In fact, all the individual currents considered in this work lead practically to
the same energy dependence in the restricted energy range considered in Fig.~\ref{fig6}.
The fact that the energy dependence given by the three-body phase space-plus-final
state interaction does not reproduce the observed energy dependence in $pp\rightarrow
pp\eta^\prime$ process has been a subject of attention since the publication of the
COSY-11 data \cite{Moskal,Bernard,Baru}. In Fig.~\ref{fig7} we show the measured cross
section data for $\pi^o$, $\eta$ and $\eta^\prime$ production in $pp$ collisions
together with the corresponding energy dependence predicted by the phase
space-plus-final state interaction only (solid curves). Apart from the fact that
in $\eta^\prime$ production the deviation starts a little lower in excess energy than
in $\pi^o$ production, we see no indication for any peculiar feature in $\eta^\prime$
production.
The most trivial source of influence on the energy dependence of the cross section
close to threshold is the Coulomb force \cite{Miller}, which has not been considered so
far in our calculations. We follow the Gell-Mann and Goldberger two-potential formalism
\cite{GellMann} to include the Coulomb force in the final state interaction. The
perturbed wave function is calculated exactly by solving the Schr\"odinger equation in
coordinate space with the nuclear-plus-Coulomb potential. For the nuclear potential
we use the Paris potential \cite{Lacomb}. The half-off-shell (pure) Coulomb T-matrix
is obtained in analytic form \cite{Dilin}.
We note that care must be taken in order to use consistently a $NN$ T-matrix
constructed within a non-relativistic formulation---such as the Paris T-matrix---in a
relativistic approach. We have followed Ref.\cite{Herrmann} in order to use the Paris
T-matrix in the present relativistic approach. Also, in the following, the production
current has been kept the same as that used in conjunction with the Bonn B potential,
except for a readjustment of the reduction factor of the product $g_{NN^*\eta^\prime}
g_{NN^*\pi}$ in the nucleon resonance current that was needed to reproduce the data.
We are aware that by using the Paris potential instead of the Bonn B potential we may
loose the consistency between the production current and the final state interaction,
however the
energy dependence of the cross section is practically the same---when the Coulomb force
is switched off---as that given by the Bonn B potential over a wide range of excess energy.
The result of including the Coulomb effect as described above is shown in Fig.~\ref{fig8}
by the solid line. The dotted line corresponds to the result when the Coulomb force is
switched off. As can be seen, the Coulomb effect is considerable; in particular, it reduces
the cross section by almost $40\%$ at $Q=1.5~MeV$.
The picture that emerges from the above considerations is that there is no
obvious indication of a need for other mechanisms than those already
considered here in order adequately to describe the existing data.
A close comparison of the COSY-11 \cite{Moskal} and SPESIII \cite{Hibou} data, however,
seems to indicate different trends in the energy dependence. In this connection, new
measurements at higher energies, and even remeasurement of some of the data
near $Q\sim 4~MeV$, are extremely important for resolution of this issue. The new data at
higher energies will certainly impose more stringent constraints on the parameters
of our model. They would also tell us whether a proper description of the $pp \rightarrow
pp\eta^\prime$ requires the inclusion of the $p\eta^\prime$ final state interaction, as has
been argued recently~\cite{Baru}. The $p\eta^\prime$ final state interaction as discussed
in Ref.\cite{Baru} can be viewed as the higher order terms in the production current
treated in the present work. In fact, the $\eta ^{\prime }$-emission current defined in
the previous section (cf. Eqs.(\ref{nuc_cur},\ref{res_cur},\ref{mec_cur})) is, in part,
just the Born term of a more general current that can be obtained by using the T-matrix
amplitude for the $M+p\rightarrow \eta^\prime +p$ transition, where $M$ denotes any meson
of interest. This can easily be seen if we disregard the nucleon labeled 2 in Fig.~\ref{fig1};
the current then becomes nothing other than the Born term of the $M+p\rightarrow \eta^\prime+p$
transition amplitude. The T-matrix amplitude for the $M+p\rightarrow \eta ^{\prime }+p$
process may be separated into the so-called pole and non-pole terms according to Pearce
and Afnan \cite{Pearce}. By using the physical baryon mass and physical $NB\eta^\prime$
($B=N,N^*$) and $NNM$ vertices in the baryonic current, the pole term of the $M+p
\rightarrow \eta^\prime+p$ T-matrix amplitude is fully accounted for in the present work.
What is taken in the Born approximation is, therefore, the non-pole part of the T-matrix
only.
%----------------
\section{Summary}
%----------------
We have investigated the reaction $pp \rightarrow pp\eta^\prime$ within a
relativistic meson-exchange model of hadronic interactions. We find that the
$S_{11}(1897)$ resonance as determined in a recent multipole analysis of
$\eta^\prime$ photoproduction \cite{Plotzke} gives rise to a contribution that
reproduces the measured cross section and allows for no other production
mechanisms. This is consistent with the analysis of Ref.\cite{Plotzke}, where
the nucleon resonance parameters have been extracted under the resonance dominance
assumption. Whether this is already the entire physics of the production mechanism is an
open question that requires further investigation. Indeed, in Ref.\cite{Plotzke}, in
addition to the assumption of resonance dominance, only the $p\eta^\prime$ and $p\pi^o$ decay
channels have been explicitly considered; the $p\pi^o$ decay channel accounts for
{\em all} possible decay channels
other than the $p\eta^\prime$. The coupling constants $NN^*\eta^\prime$ and
$NN^*\pi$ obtained from the decay widths extracted under these assumptions are,
therefore, subject to considerable uncertainties which, in turn, lead to
corresponding uncertainties in the resonance current contribution to the cross
section. A more accurate determination of the relevant resonance parameters is
called for before a more definite conclusion can be drawn about the role of the
$N^*$ resonances in the $pp \rightarrow pp\eta^\prime$ reaction. We should also
keep in mind that the strong peak in the cross section close to threshold in
$\eta^\prime$ photoproduction may be due to a cusp effect \cite{Hoehler}, and not
to the resonance assumed in Ref.\cite{Plotzke}. All these
issues associated with the resonance require further careful investigation.
The mesonic current gives rise to cross sections which are comparable to the
measured values. It is dominated by the $\rho\rho\eta^\prime$-exchange current.
Depending on the choice of the sign of the $\sigma\eta\eta^\prime$ coupling constant,
the calculated cross sections overpredict or underpredict the data. However neither
choice can be excluded at present---mainly due to uncertainties associated with the
baryonic current. Data in higher energy regions than currently available may provide
the necessary constraint to fix the sign of this coupling, as the different signs lead
to different energy dependences.
Since the nucleonic current is found to be small, it is unlikely that the $pp
\rightarrow pp\eta^\prime$ reaction can be used quantitatively to determine the
$NN\eta^\prime$ coupling constant---at least not before the other dominant currents
are under better control.
As expected from earlier calculations \cite{Miller}, the Coulomb force in the final state
interaction is found to play a crucial role in explaining the observed energy dependence
of the cross section close to threshold. The basic production mechanisms start to
influence the energy dependence of the cross section only for excess energies above
$\sim 5~MeV$. For lower energies the energy dependence is determined by the three-body
phase space-plus-final state interaction \cite{Watson}. New data at higher energies will
certainly impose constraints on the production mechanisms. In addition
to more exclusive observables than the total cross section, a combined analysis of
production processes using both electromagnetic and hadronic probes should
help to disentangle the different production mechanisms.
%---------------------
{\bf Acknowledgment}
%---------------------
We are grateful to J. Haidenbauer and S. Krewald for useful discussions. We also thank
Ulf Mei\ss ner and J. Haidenbauer for critical readings of the manuscript.
H.F.A. acknowledges the partial support from FONDECYT, grant No. 1970508.
\vfill \eject
\begin{thebibliography}{99}
\bibitem{Gluons}
B. Diekmann,
Phys. Rep. {\bf 159}, 99 (1988);
J. Jousset et al.,
Phys. Rev. {\bf D41}, 1389 (1990);
P. Ball, J.-M. Frere and M. Tygat,
Phys. Lett. {\bf B365}, 367 (1996).
\bibitem{CLEO}
CLEO Collaboration, B. H. Behrens et al.,
Phys. Rev. Lett. {\bf 80}, 3710 (1998);
T. E. Browder et al.,
Phys. Rev. Lett. {\bf 81}, 1786 (1998).
\bibitem{Gluonic}
D. Atwood and A. Soni,
Phys. Lett. {\bf B405}, 150 (1997);
Phys. Rev. Lett. {\bf 79}, 5206 (1997);
W-S. Hou and B. Tseng,
Phys. Rev. Lett. {\bf 80}, 434 (1998);
M. R. Ahmady, E. Kou, and A.Sugamoto,
Phys. Rev. {\bf D58}, 14014 (1998);
For a most recent determination of the $\eta^\prime$ gluonic admixture,
see E. Kou, hep-ph/9908214.
\bibitem{EMC88}
J. Ashman et al.,
Phys. Lett. {\bf B206}, 364 (1988).
\bibitem{Efremov}
A. V. Efremov, J. Soffer, and N. A. T\"ornqvist,
Phys. Rev. Lett. {\bf 64}, 1495 (1990);
Phys. Rev. {\bf D44}, 1369 (1991);
S. Narison, G. M. Shore, and G. Veneziano,
Nucl. Phys. {\bf B546}, 235 (1999) and references therein.
\bibitem{Moskal}
P. Moskal et al.,
Phys. Rev. Lett. {\bf 80}, 3202 (1998).
\bibitem{Hibou}
F. Hibou et al.,
Phys. Lett. {\bf B438}, 41 (1998).
\bibitem{Sibirtsev}
A. Sibirtsev and W. Cassing,
Eur. Phys. J. {\bf A2}, 333 (1998).
\bibitem{Bernard}
V. Bernard, N. Kaiser and U. Mei\ss ner,
Eur. Phys. J. {\bf A4}, 259 (1999).
\bibitem{Gedalin}
E. Gedalin, A. Moalem and L. Razdolskaja,
Nucl. Phys. {\bf A650}, 471 (1999).
\bibitem{Baru}
V. Baru, J. Haidenbauer, C. Hanhart, A. Kudryavtsev and J. Speth,
Eur. Phys. J. {\bf A}., in press.
\bibitem{Watson}
K. Watson,
Phys. Rev. {\bf 88}, 1163 (1952).
For a recent discussion, see Ref.\protect\cite{Hanhart}.
\bibitem{Miller}
G. A. Miller and P. U. Sauer,
Phys. Rev. {\bf C44}, R1725 (1991).
\bibitem{Herrmann1}
V. Herrmann, K. Nakayama, H. F. Arellano, and O. Scholten,
Nucl. Phys. {\bf A582}, 568 (1995).
\bibitem{Hanhart}
C. Hanhart and K. Nakayama,
Phys. Lett. {\bf B454}, 176 (1999).
\bibitem{Plotzke}
R. Pl\"otzke et al.,
Phys. Lett. {\bf B444}, 555 (1998).
\bibitem{Nak1}
K. Nakayama, A. Szczurek, C. Hanhart, J. Haidenbauer and J. Speth
Phys. Rev. {\bf C57}, 1580 (1998).
\bibitem{Nak2}
K. Nakayama, J.W. Durso, J. Haidenbauer, C. Hanhart and J. Speth,
Phys. Rev. {\bf C}, in press; nucl-th/9904040.
\bibitem{Haidenbauer}
J. Haidenbauer, private communication; Following Phys. Rev. {\bf C40}, 2465 (1989),
the Bonn B potential can be constrained to the low-energy $pp$ scattering length
with only a minor readjustment of its $NN\sigma$ coupling constant in the total
isospin $T=1$ state. See also Table 1 in Ref.\protect\cite{Herrmann1}, where
another version of the Bonn potential has been modified in a similar way.
\bibitem{MHE87}
R. Machleidt, Adv. Nucl. Phys. {\bf 19}, 189 (1989).
\bibitem{lee} M. Batinic, A. Svarc and T.--S.H. Lee,
{\it Phys.Scripta} {\bf 56}, 321 (1997).
\bibitem{said}
extracted from the VIRGINIA TECH PARTIAL-WAVE ANALYSES ON-LINE
(http://said.phys.vt.edu/said\_branch.html)
\bibitem{Dumbrajs}
O. Dumbrajs, et al.
Nucl. Phys. {\bf B216}, 277 (1983).
\bibitem{Zhang}
J. F. Zhang, N. Mukhopadhyay and M. Benmerrouche,
Phys. Rev. {\bf C52}, 1134 (1995).
\bibitem{Grein}
W. Grein and P. Kroll,
Nucl. Phys. {\bf A338}, 332 (1980).
\bibitem{Altarelli}
G. Altarelli, R. D. Ball, S. Forte and G. Ridolfi,
Nucl. Phys. {\bf B496}, 337 (1997).
\bibitem{OZI}
S. Okubo, Phys. Lett. {\bf 5}, 165 (1963);
G. Zweig, CERN Report No. TH412, 1964;
J. Iizuka, Prog. Theor. Phys. Suppl. {\bf 37 \& 38}, 21 (1966).
\bibitem{DFratio}
F. E. Close and R. G. Roberts,
Phys. Lett. {\bf B316}, 165 (1993);
P. G. Ratcliffe,
Phys. Lett. {\bf B365}, 383 (1996);
X. Song, P. K. Kabir, and J. S. McCarthy,
Phys. Rev. {\bf D54}, 929 (1996).
\bibitem{Hoehler} G. H\"ohler,
$\pi N$ Newsletter {\bf 14}, 168 (1998).
\bibitem{Benmer}
M. Benmerrouche and N. C. Mukhopadhyay,
Phys. Rev. {\bf D51}, 3237 (1995).
\bibitem{Durso}
J.W. Durso,
Phys. Lett. {\bf B184}, 348 (1987).
\bibitem{PDG}
Particle Data Group,
Eur. Phys. J. {\bf C3}, 1 (1998).
\bibitem{Garc}
H. Garcilazo and E. Moya de Guerra,
Nucl. Phys. {\bf A562}, 521 (1993).
\bibitem{Schutz}
C. Sch\"utz, J. W. Durso, K. Holinde and J. Speth,
Phys. Rev. {\bf C49}, 2671 (1994).
\bibitem{Lattice}
K. F. Liu, S. J. Dong, T. Draper, and W. Wilcox,
Phys. Rev. Lett. {\bf 74}, 2172 (1995).
\bibitem{Bockman}
R. Bockmann, C. Hanhart, O. Krehl, S. Krewald, and J. Speth,
Phys. Rev. {\bf C}, in press; nucl-th/9905043.
\bibitem{GellMann}
M. Gell-Mann and M. L. Goldberger,
Phys. Rev. {\bf 91}, 398 (1953).
\bibitem{Lacomb}
M. Lacombe, B. Loiseau, J. M. Richard, R. Vinh Mau, J. C\^ot\'e,
P. Pir\`es and R. de Tourreil,
Phys. Rev. {\bf C21}, 861 (1980).
\bibitem{Dilin}
Dilinskii and Mukhamedzhanov,
Soviet J. Nucl. Phys. {\bf 3}, 180 (1966).
\bibitem{Herrmann}
V. Herrmann and K. Nakayama,
Phys. Rev. {\bf C46}, 2199 (1992).
\bibitem{Pearce}
B. C. Pearce and I. R. Afnan,
Phys. Rev. {\bf C34}, 991 (1986).
\bibitem{iucf90}
H. O. Meyer et al.,
Phys. Rev. Lett. {\bf 65}, 2846 (1990).
Corrected data according to Ref.\protect\cite{iucf92}.
\bibitem{iucf92}
H. O. Meyer et al.,
Nucl. Phys. {\bf A539}, 633 (1992).
\bibitem{celsiuspi}
A. Bondar et al.,
Phys. Lett. {\bf B356}, 8 (1995).
\bibitem{spes3eta2}
E. Chiavassa et al.,
Phys. Lett. {\bf B322}, 270 (1994).
\bibitem{celsiuseta}
H. Cal\'en et all.,
Phys. Lett. {\bf B366}, 39 (1996).
\end{thebibliography}
\vglue 0.5cm
\begin{figure}[h]
\vspace{20cm}
\special{psfile=fig1.eps
hoffset=10 voffset=-10 hscale=65 vscale=65}
\caption{$\eta^\prime$ meson production currents included in
the present study. Upper row: nucleonic current $J_{nuc}$,
$M = \pi, \eta, \rho, \omega, \sigma, a_o$.
Middle row: nucleon resonance current $J_{res}$, $N^*=S_{11}(1897),~P_{11}(1986)$.
Lower row: mesonic current $J_{mes}$, $v = \rho, \omega$.}
\label{fig1}
\end{figure}
\newpage
\vglue 0.5cm
\begin{figure}[h]
\vspace{15cm}
\special{psfile=fig2_new.eps
hoffset=20 voffset=5 hscale=60 vscale=80}
\caption{$N^*$ resonance current contribution to the total cross section for the
reaction $pp\rightarrow pp\eta^\prime$ as a function of excess energy $Q$.
The upper panel corresponds to the results with the choice $\lambda=0$ in
Eq.(\protect\ref{NN^*ps}) at the $NN^*\eta^\prime$ vertices, the lower
panel to those with the choice $\lambda=1$. The dashed-dotted curves correspond
to the $S_{11}(1897)$ resonance contribution, the dashed curves to the
$P_{11}(1986)$ resonance contribution. The solid curves are the total contributions.
The experimental data are from Refs.~\protect\cite{Moskal,Hibou}.}
\label{fig2}
\end{figure}
\newpage
\vglue 2cm
\begin{figure}[h]
\vspace{10cm}
\special{psfile=fig3.eps
hoffset=20 voffset=5 hscale=60 vscale=65}
\caption{Mesonic current contribution to the total cross section for the
reaction $pp\rightarrow pp\eta^\prime$ as a function of excess energy $Q$.
The dash-dotted curve correspond to the $\sigma\eta\eta^\prime$-exchange
and the dotted curve to the $\omega\omega\eta^\prime$-exchange current
contribution. The dashed curve represents the $\rho\rho\eta^\prime$-exchange
current contribution. The solid curve is the total contribution corresponding
to choosing the negative sign for the $\sigma\eta\eta^\prime$ coupling constant,
the short-dashed curve to the positive sign. The experimental data are from
Refs.~\protect\cite{Moskal,Hibou}.}
\label{fig3}
\end{figure}
\newpage
\vglue 1cm
\begin{figure}[h]
\vspace{15cm}
\special{psfile=fig4_new.eps
hoffset=20 voffset=5 hscale=60 vscale=80}
\caption{Nucleonic-plus-mesonic current contributions to the total cross section
for the reaction $pp\rightarrow pp\eta^\prime$ as a function of excess energy $Q$.
The upper panel corresponds to the results with the pv coupling ($\lambda=0$) at
the $NN\eta^\prime$ vertex in the nucleonic current, the lower
panel to those with ps coupling ($\lambda=1$). The dashed-dotted curve in both
panels represents the total mesonic current contribution shown by the dashed curve
in Fig.~\protect\ref{fig3}. The nucleonic current contributions are represented by the
dashed curves and the corresponding total contributions by the solid curves.
The experimental data are from Refs.~\protect\cite{Moskal,Hibou}.}
\label{fig4}
\end{figure}
\newpage
\vglue 1cm
\begin{figure}[h]
\vspace{15cm}
\special{psfile=fig4a_new.eps
hoffset=20 voffset=5 hscale=60 vscale=80}
\caption{Same as Fig.~\protect\ref{fig4}, except for the total mesonic current
contribution, which corresponds to the result represented by the solid curve in
Fig.~\protect\ref{fig3}}.
\label{fig4a}
\end{figure}
\newpage
\vglue 1cm
\begin{figure}[h]
\vspace{12cm}
\special{psfile=fig5_new.eps
hoffset=20 voffset=5 hscale=60 vscale=70}
\caption{The total cross section for the reaction $pp\rightarrow pp\eta^\prime$
as a function of excess energy $Q$. The solid line represents the mesonic-plus-nucleonic
current contribution corresponding to the upper panel of Fig.~\protect\ref{fig4},
except for the value of the ps-pv mixing parameter at the $NN\eta^\prime$ vertex, which
is taken to be $\lambda=0.7$. The dotted line represents the result corresponding to the
mesonic-plus-nucleonic current shown in the lower panel of Fig.~\protect\ref{fig4}; it
includes also the nucleon resonance current, whose product of the coupling constants
$g_{NN^*\eta^\prime} g_{NN^*\pi}$ is multiplied by an arbitrary reduction factor of $1/2$.
The dash-dotted and dashed curves are the same as the dotted curve, except that they
correspond to the mesonic-plus-nucleonic current shown in the upper and lower panels of
Fig.~\protect\ref{fig4a}, respectively. The corresponding reduction factors of the product
$g_{NN^*\eta^\prime} g_{NN^*\pi}$ are $0.625$ and $1$. The experimental data are from
Refs.~\protect\cite{Moskal,Hibou}.}
\label{fig5}
\end{figure}
\newpage
\vglue 2cm
\begin{figure}[h]
\vspace{10cm}
\special{psfile=fig6_new.eps
hoffset=20 voffset=5 hscale=60 vscale=65}
\caption{Same as Fig.~\protect\ref{fig5} for excess energy below 10 $MeV$. Only one
total contribution (solid curve) is shown since all the curves shown in
Fig.~\protect\ref{fig5} are almost indistinguishable at the scale used here. The
dotted line corresponds to the result with only the phase space-plus-final state
interaction.}
\label{fig6}
\end{figure}
\newpage
\vglue 1cm
\begin{figure}[h]
\vspace{15cm}
\special{psfile=fig7.eps
hoffset=20 voffset=5 hscale=60 vscale=85}
\caption{The total cross section for $\pi^o$-, $\eta$- and $\eta^\prime$-meson
production in $pp$ collisions as a function of excess energy $Q$. The curves represent
the corresponding energy dependence given by the phase space-plus-final state
interaction with {\em no} Coulomb force. In the upper panel the experimental
data are from Refs.\protect\cite{iucf90,iucf92,celsiuspi}; in the middle panel from
Refs.\protect\cite{Hibou,spes3eta2,celsiuseta}; in the lower panel from
Refs.\protect\cite{Moskal,Hibou}.}
\label{fig7}
\end{figure}
\newpage
\vglue 2cm
\begin{figure}[h]
\vspace{10cm}
\special{psfile=fig8.eps
hoffset=20 voffset=5 hscale=60 vscale=65}
\caption{The total cross section for the reaction $pp\rightarrow pp\eta^\prime$
as a function of excess energy $Q$. The solid curve corresponds to the total current
contribution including the Coulomb force in the final state interaction.
The dotted line corresponds to that when the Coulomb force is switched off.
Here the Paris T-matrix has been used as the final state interaction.
The experimental data are from Refs.~\protect\cite{Moskal} (circle) and
\protect\cite{Hibou} (triangle).}
\label{fig8}
\end{figure}
\end{document}