1: \section{Extraction of the $\rho$ Mass Spectra}
2: \label{sec:simulations}
3:
4: \subsection{Simulations}
5: \label{sec:simBUU}
6: To simulate the expected physics processes, a realistic model was employed and corrected for the CLAS acceptance. The events were generated using a code based on a semi-classical Boltzmann-Uehling-Uhlenbeck (BUU) transport model~\cite{mue02,bra99} that was developed at the University of Giessen. This model treats nuclear effects such as the shadowing of the photon-induced reactions, as well as modeling of the Fermi motion, Pauli blocking, Coulomb interaction, final state interactions, and collisional broadening. Additional medium effects, such as decreasing the mass of the vector mesons~\cite{hatsuda}, can be incorporated on demand. In photoproduction reactions, the meson is formed throughout the nuclear volume, unlike reactions with hadronic beams, where the meson is expected to be produced close to where the beam particle entered the nucleus.
7:
8: The model treats the photon-nucleus reactions as a two-step process. In the first step, the incoming photons react with a single nucleon taking into account the effects of shadowing. In the second step, the produced particles are propagated explicitly through the nucleus allowing for final-state interactions. This step is governed by the semi-classical BUU transport equations. A complete treatment of the $\pair$ pair production from $\gamma A$ reactions at Jefferson Lab energies can be found in Ref.~\cite{bra99}. The output is the $\pair$ mass spectrum from 7 decays: the direct vector meson decays $\rho \rightarrow \pair$, $\omega \rightarrow \pair$, and $\phi \rightarrow \pair$, as well as 4 Dalitz decays: $\Delta \rightarrow N \pair$, $\eta \rightarrow \pair \gamma$, $\pi^{0} \rightarrow \pair \gamma$, and $\omega \rightarrow \pi^{0} \pair$.
9:
10: The simulations were employed in two ways. The first application was to calculate the detector acceptance. Fig.~\ref{fig:accZ} shows the $\pair$ detection acceptance for the $\rho$ decay as a function of the invariant mass for carbon targets at different $z$-locations in CLAS. The smooth, slowly-varying acceptance does not distort the shape of the invariant mass spectrum.
11:
12: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
13: \begin{figure}[htpb]
14: \includegraphics[width=7.5cm]{acc-c-50bins.eps}
15: \caption{\small{CLAS acceptance as a function of the $\pair$ invariant mass for $\cnuc$ targets at different $z$-locations in CLAS. The individual targets are located at $z = -12.0, -7.0, 2.0, and 3.0$~cm. The results are shown for the $\rho$ channel and summed over all photon energies.}}
16: \label{fig:accZ}
17: \end{figure}
18: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
19:
20: The second application was to remove the $\omega$- and $\phi$-meson contributions from the total invariant mass spectrum for each target. The $\omega$- and $\phi$-mesons have long lifetimes ($c \tau$ = 23.4~fm and 44.4~fm, respectively~\cite{pdg}) and momenta greater than 0.8~GeV; therefore, their probabilities of decaying inside the nucleus are low. The $\omega$ and $\phi$ channels in the simulation have been treated for many-body nuclear effects as they escape the nucleus, while changes due to chiral symmetry restoration have not been included. A comparison of the simulated $\omega$- and $\phi$-meson line shapes without and with the modifications of Ref.~\cite{hatsuda} shows negligible differences. The $\omega$ mass shape contained both the direct and the Dalitz decay channels. These two channels were combined by fixing the branching ratios in the BUU generator. A single $\omega$-meson distribution was used in the global fit. The acceptance-corrected BUU mass shapes for the $\rho$, $\omega$, and $\phi$ mesons were scaled to match the experimental mass spectra. Fig.~\ref{fig:buufits} shows the result of the fits to the mass distributions from the $\dnuc$, $\cnuc$, and $\tife$ data after the combinatorial background was subtracted. The dot-dashed, dashed, and dotted curves are the fits for the $\rho$, $\omega$, and $\phi$ mesons, respectively. The solid curve is the sum of the three fit contributions. The normalized $\omega$- and $\phi$-meson shapes were subtracted. A substantial contribution from the $\rho$ meson is present in our mass spectra unlike in the KEK analysis (see Fig.~1 in Ref.~\cite{kek-new}).
21: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
22: \begin{figure}[htpb]
23: \mbox{\includegraphics[width=7.0cm]{p-run10-paper-bw.eps}}
24: \mbox{\includegraphics[width=7.0cm]{c-run10-paper-bw.eps}}
25: \mbox{\includegraphics[width=7.0cm]{fe-run10-paper-bw.eps}}
26: \caption{\small{(color online) Result of the fits to the $\pair$ invariant mass spectrum obtained for the $\dnuc$ (top), $\cnuc$ (middle), and $\tife$ (bottom) data. The curves are Monte-Carlo calculations by the BUU model~\cite{effenberger1,effenberger2} for various vector meson decay channels.}}
27: \label{fig:buufits}
28: \end{figure}
29: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
30:
31: \subsection{The $\rho$ Mass Spectra}
32: \label{sec:rhoSpec}
33: After subtracting the combinatorial background and the $\omega$- and $\phi$-meson contributions, all that remains in the $\pair$ invariant mass spectra were the $\rho$ mass distributions (see Fig.~\ref{fig:simplefits}) for the various targets. The exact functional form for the mass spectrum has been obtained by calculating the cross section for $\rho$-meson production including the leptonic decay width~\cite{effenberger2} and is given by:
34: \begin{equation}
35: A(\mu) = \frac{2}{\pi} \frac{\mu^2 \Gamma(\mu)}{(\mu^2-M_{\rho}^2)^2+\mu^2\Gamma^2(\mu)},
36: \label{eq:rho}
37: \end{equation}
38: where $\Gamma(\mu)$ is the width of the resonance, $M_{\rho}$ is the $\rho$ pole mass, and $\mu$ is the $\pair$ invariant mass. A good approximation for the function in Eq.~\ref{eq:rho} is a Breit-Wigner function divided by $\mu^{3}$. In the Vector Dominance Model, the photon propagator has the form of $1/q^{2} = 1/\mu^{2}$. This term contributes as $1/\mu^{4}$ in the cross section, while the phase space also gives a factor of $\mu$, resulting in a $1/\mu^{3}$ factor~\cite{guo,bra99,oconnell}. Indeed the fits to the Breit-Wigner/$\mu^{3}$, rather than a simple Breit-Wigner function, describe the data very well (see Fig.~\ref{fig:rhoDcomp}). For example, for the $\dnuc$ target, a simple Breit-Wigner fit gives a $\chi^{2}$ per number of degrees of freedom ($\chi^{2}_{ndf}$) equal to 3.9, while a Breit-Wigner/$\mu^{3}$ gives a $\chi^{2}_{ndf}=1.08$ (see Fig.~~\ref{fig:rhoDcomp}). A similar effect is seen with the $\cnuc$ and $\tife$ data. For the $\cnuc$ spectrum, the fits without and with the $1/\mu^{3}$ factor produce $\chi^{2}_{ndf}$ of 2.92 and 1.23, respectively. For the $\tife$ data, the $\chi^{2}_{ndf}$ are 3.02 and 1.35, respectively. The sensitivity of the fits to the $1/\mu^{3}$ factor indicates that the systematic uncertainties in the background subtraction procedure are insignificant. Similar results are obtained for the heavier targets $\cnuc$ and $\tife$, with larger uncorrelated background (proportional to the atomic number of the target).
39:
40: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
41: \begin{figure}[htpb]
42: \mbox{\includegraphics[width=8.0cm]{d2fites2.eps}}
43: \mbox{\includegraphics[width=8.0cm]{cfites2.eps}}
44: \mbox{\includegraphics[width=8.0cm]{ffites2.eps}}
45: \caption{\small{Individual Breit-Wigner/$\mu^{3}$ fits to the $\rho$ mass spectra for $\dnuc$ (top), $\cnuc$ (middle), and $\tife$ (bottom)}}
46: \label{fig:simplefits}
47: \end{figure}
48: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
49:
50: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
51: \begin{figure}[htpb]
52: \includegraphics[width=7.5cm]{rhod2comp-bw.eps}
53: \caption{\small{ Fits to the $\rho$ mass spectrum for the deuterium data using a Breit-Wigner (dashed) and a Breit-Wigner/$\mu^{3}$ function (solid).}}
54: \label{fig:rhoDcomp}
55: \end{figure}
56: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
57: