1: \subsection{Introduction}
2: The double charge exchange (DCX) reaction of pions received a considerable attention in the past (see for instance Ref.~\cite{LAMPF} and references therein). The mechanism of two sequential single charge exchanges has traditionally been able to explain the main features of this reaction~\cite{Becker:1970tk,Gibbs:1977yz} at low energies although the contribution of the $A(\pi,\pi\pi)X$ reaction becomes progressively important as the energy increases~\cite{Vicente:1988iv,Alqadi:2001pe}. At higher ($\sim1~\mathrm{GeV}$) energies, the sequential mechanism becomes insufficient to account for the reaction cross section~\cite{Abramov:2002nz,Krutenkova:2005nh}.
3: Extensive experimental studies performed at LAMPF obtained high precision data for doubly differential cross sections on $^3He$~\cite{Yuly:1997ja} and heavier nuclei ($^{16}O$, $^{40}Ca$, $^{208}Pb$)~\cite{Wood:1992bi} in the region of $\ekin=120-270 \MeV$.
4:
5: H\"ufner and Thies~\cite{huefnerThies} explored for the first time the applicability of the Boltzmann equation in $\pi A$ collisions and achieved qualitative agreement with data on single and double charge exchange. Their method to solve the Boltzmann equation was based upon an expansion of the pion one-body distribution function in the number of collisions. There, in contrast to our work, the Boltzmann equation is not solved with a test-particle ansatz but by reformulating it into a set of coupled differential equations which can then be solved in an iterative manner. However, this approach was based on simplifying assumptions of averaged cross sections and averaged potentials.
6: The work by Vicente et al.~\cite{Vicente:1988iv} was based upon the cascade model described in~\cite{osetSimulation}. There, a microscopic model for $\pi N$ scattering was used as input for the pion reaction rates in the simulation. In that work~\cite{Vicente:1988iv}, pion DCX off $^{16}O$ and $^{40}Ca$ has been explored and fair quantitative agreement with data has been achieved.
7:
8: In the following, we explore DCX on heavier nuclei, comparing with the data measured by Wood et al.~\cite{Wood:1992bi}. We also address the scaling of the total cross section discussed by Gram et al.~\cite{Gram:1989qh}. To focus only on single-pion re-scattering, we consider incoming pion energies below $\ekin=180\MeV$; above that energy $2\pi$ production becomes prominent and DCX does not happen necessarily in a two-step process anymore. %Moreover, we will address a more technical issue.
9: Due to the small mean free path of the incoming pions, the process is mostly sensitive to the surface of the nucleus. In \cite{Buss:2006yk} we also addressed the issue of parallel versus full ensemble scheme in the numerical implementation of the BUU equation. There, we found that the parallel ensemble scheme is a good approximation and has therefore been applied in the following calculations.
10:
11:
12: \subsection{Influence of the density profile} \label{densChapter}
13: \begin{figure}[b]
14: \begin{center}
15: \includegraphics[]{./data/nucleus/parallel/Pb/dsigma_dOmega_Pb.eps}
16: \label{densVer}
17: \caption{Influence of the density distribution on the angular distributions for the double charge exchange process $\pi^+ Pb\to \pi^- X$ at $E_{\mathrm{kin}}=180 \MeV$. The solid line shows the result where we implement different radii of the density distributions of neutrons and protons in our model ($R_n-R_p=0.266$ fm). As a consequence, the nucleus is surrounded by a small neutron skin. For the dashed line we assumed same radii for the proton and neutron distributions.}
18: \end{center}
19: \end{figure}
20: Due to the low pion mean free path in nuclear matter~\cite{Buss:2006vh}, DCX is very sensitive to the surface properties of the nuclei. Therefore, we compared the results with a neutron skin for the $Pb$ nucleus to the results obtained without such a neutron skin; for details see \cite{Buss:2006yk}.
21: Neutron skins are very interesting because in those skins only $\pi^+$ mesons can undergo charge exchange reactions. For the positive pions this causes an enhancement of DCX processes at the surface, so the pions do not need to penetrate deeply for this reaction. Hence, the probability for their absorption is reduced. As can be observed in fig. 1, the enhancement in the total cross section for $R_n-R_p=0.266$ fm is roughly $35\%$ at $180\MeV$.
22: The accurate determination of neutron skins is relevant for different
23: areas of physics such as nuclear structure, neutron star properties,
24: atomic parity violation (PV) and heavy ion collisions \cite{Horowitz:1999fk,Horowitz:2006iv,Piekarewicz:2006vp}. The
25: Parity Radius Experiment (PREX) at JLab \cite{PREX} shall measure the neutron
26: radius with high precision ($1\%$) using PV electron scattering. We can see that the DCX cross section is very sensitive to the size of the
27: difference between the proton and the neutron radii of $Pb$. The effect is
28: specially large (more than a factor $2$) at forward angles, where our model
29: performs very well (see next Section). Indeed, without neutron skin, due to strong pion absorption in the bulk of nuclear matter, the DCX cross section is small at forward angles. The presence of a neutron skin favors DCX in peripheral reactions, where the pion propagates in practically pure neutron matter. This naturally enhances the DCX cross section at forward angles.
30: Hence, a precise measurement of DCX at forward angles, combined with a realistic theoretical analysis could be a valuable source of information on the neutron skins complementary to the one obtained with PV electron scattering. Note that for the $\pi^-$ similar arguments lead to a reduction of the cross section.
31:
32: % \begin{figure}[h]
33: % \begin{center}
34: % %\includegraphics[angle=0,width=9cm]{./BUUData/potentials/deltaPotential.eps}
35: % \includegraphics[]{./data/densities/PlotDens.eps}
36: % \label{densPlot}
37: % \caption{In the left panel we show the comparison of the Wood-Saxon (WS) density description according to eq. \ref{densWS} to the parameterization out of \cite{osetPionicAtoms}. In the right panel we show the neutron (n) and proton (p) density of the second distribution. For the lead nucleus there is a small neutron skin of approximately $0.23 \fm$ due to the larger radius of the neutron density.}
38: % \label{dens}
39: % \end{center}
40: % \end{figure}
41:
42:
43:
44: %*************************************************************************
45: %*************************************************************************
46: %*************************************************************************
47:
48:
49: \subsection{Comparison to data}
50: \begin{figure}[]
51: \begin{center}
52: %\includegraphics[angle=0,width=9cm]{./BUUData/potentials/deltaPotential.eps}
53: \includegraphics[width=0.45\textwidth]{./data/nucleus/parallel/sigma_tot.eps}
54: \end{center}
55: \caption{\new{The inclusive double charge exchange total cross section as function of the nuclear target mass at $E_{\mathrm{kin}}=120,150$ and $180 \MeV$. The lines connecting our results are meant to guide the eye; the data are taken from Ref.~\cite{Wood:1992bi} (lower panel: $E_{\mathrm{kin}}=120,150$ and $180 \MeV$, upper panel: only $180 \MeV$).}}
56: \label{total}
57: \end{figure}
58: Now we proceed to the comparison with the data measured at LAMPF by Wood et al.~\cite{Wood:1992bi}. We discuss first the total cross section and thereafter the angular distributions. \footnote{In \cite{Buss:2006yk} we also show double differential cross sections as a function of both angles and energies of the outgoing pions.}
59:
60: In \fig{\ref{total}} one can see the excellent quantitative agreement to the total cross \newla{section} data at $120$, $150$ and $180 \MeV$ for Oxygen and Calcium. For the Lead nucleus we see some discrepancies, which are, however, within the experimental accuracy. Notice that we reproduce the different $A$ \new{dependencies} of \new{both} $(\pi^+,\pi^-)$ and $(\pi^-,\pi^+)$ reactions. It is due to the fact that, when $A$ increases, the number of neutrons increases with respect to the number of protons, and this favors the $\pi^+$ induced reaction.
61:
62:
63:
64: \begin{figure}[]
65: \begin{center}
66: \includegraphics[width=0.45\textwidth]{./data/nucleus/parallel/scaling.eps}
67: \end{center}
68: \caption{The scaling of the total charge exchange cross section according to eq.~(\ref{scalingEQ}) is visualized by dividing $\sigma_{tot}$ by the factor $A^{2/3}Q(Q-1)/(A-1)$ and plotting it as a function of $A-Q$. $Q$ denotes the number of protons in the case of $(\pi^-,\pi^+)$ and the number of neutrons in $(\pi^+,\pi^-)$. The points are GiBUU results at pion kinetic energies of $180 \MeV$; the dashed line
69: denotes a function proportional $1/(A-Q)$, corresponding to the exact scaling.}
70: \label{scaling}
71: \end{figure}
72: \begin{figure}[h]
73: \begin{center}
74: %\includegraphics[angle=0,width=9cm]{./BUUData/potentials/deltaPotential.eps}
75: \includegraphics[width=0.4\textwidth]{./data/nucleus/parallel/dsigma_dOmega_Oxygen.eps}
76: \end{center}
77: \caption{Angular distributions for the double charge exchange process $\pi^\pm A\to \pi^\mp X$ at $E_{\mathrm{kin}}=120,150$ and $180 \MeV$. The data points are taken from \cite{Wood:1992bi}; only systematical errors are shown. The solid lines represents the GiBUU results at $120 \MeV$, the dashed lines at $150 \MeV$ and the dotted lines at $180 \MeV$. The bold lines represent the results while the thin lines represent a $1\sigma$ confidence level for each point based upon our statistics.}
78:
79: \label{parallel}
80: \end{figure}
81:
82: \begin{figure}[h]
83: \begin{center}
84: %\includegraphics[angle=0,width=9cm]{./BUUData/potentials/deltaPotential.eps}
85: \includegraphics[width=0.4\textwidth]{./data/nucleus/parallel/dsigma_dOmega_ca.eps}
86: \end{center}
87: \caption{Same as fig.\ref{parallel} for $^{40}Ca$.}
88: \label{parallel1}
89: \end{figure}
90:
91: \begin{figure}[h]
92: \begin{center}
93: %\includegraphics[angle=0,width=9cm]{./BUUData/potentials/deltaPotential.eps}
94: \includegraphics[width=0.45\textwidth]{./data/nucleus/parallel/dsigma_dOmega_lead.eps}
95: \end{center}
96: \caption{Same as fig.\ref{parallel} for $^{208}Pb$.}
97: \label{parallel2}
98: \end{figure}
99: In \cite{Gram:1989qh}, Gram et al. discuss a scaling law of the total cross section. They argue as follows. Since the first collision takes place predominantly at the surface, the cross section should scale with $A^{2/3}$. Furthermore they assume that DCX is mainly a two-step process, and that a pion which undergoes an elastic process at the first collision will not contribute. This is reasonable because the incoming pions loose energy in the elastic process, and their cross section for a second charge-exchange reaction is hereafter very much reduced. For a negative pion the first charge exchange reaction occurs with a probability of $Z/N$ where $Z(N)$ denotes the number of protons(neutrons). This is the case if the interaction is dominated by the $\Delta$ resonance, as it should be in this energy region. Finally, the second charge exchange process then takes place with the probability $(Z-1)/(A-1)$ since, in the isospin limit, the $\pi^0$ interacts equally well with protons and neutrons.
100: Putting these considerations together and extending them to the $\pi^+$ case, the cross section for DCX is expected to scale according to
101: \begin{eqnarray}
102: \sigma_{tot}\sim A^{2/3} \frac{Q}{A-Q}\frac{Q-1}{A-1} \; ,
103: \label{scalingEQ}
104: \end{eqnarray}
105: where $Q$ denotes the number of protons in the case of $\pi^-$ induced and the number of neutrons in $\pi^+$ induced DCX.
106:
107: Gram et al. \cite{Gram:1989qh} find good agreement of this scaling law with experimental data. This scaling is fulfilled within our GiBUU simulations as can be seen in \fig{\ref{scaling}}. Nevertheless, one may wonder why this scaling law works in a process which is so sensitive to the neutron skin on heavy nuclei, as has been shown in fig. 1. Since the first collision takes place on the surface, a neutron skin causes an enhancement in the $A(\pi^+,\pi^-)X$ reaction while $A(\pi^-,\pi^+)X$ is suppressed. This effect leads to a deviation from the scaling. However there are also Coulomb forces which are not negligible. The Coulomb force enhances $A(\pi^-,\pi^+)X$ by attracting the negative projectiles and repelling the positive products, which therefore have a smaller path in the nucleus and undergo less absorption. And, due to similar arguments, the reaction $A(\pi^+,\pi^-)X$ is suppressed. We find that this effect counteracts the one from the neutron skin restoring the scaling. In any case\new{,} the approximate scaling exhibited by the cross section shows that the reaction is very much surface driven and can be very well understood in terms of a two-step process.
108:
109:
110:
111:
112: In fig. \ref{parallel}-\ref{parallel2} we show $d\sigma / d\Omega$ for DCX at $E_{\mathrm{kin}}=120,150$ and $180 \MeV$ on $^{16}O$, $^{40}Ca$ and $^{208}Pb$ as a function of the scattering angle $\theta$ in the laboratory frame. Our results
113: (bold lines) are shown together with their uncertainties of statistical nature (thin lines). The latter ones are well under control except at very small and very large angles, where statistics is very scarce. Again, there is a very good quantitative agreement for both $O$ and $Ca$. In the $Pb$ case, the $(\pi^-,\pi^+)$ reaction is well described, but the $(\pi^+,\pi^-)$ one is underestimated in spite of the enhancement caused by the neutron skin.
114:
115: