1: \subsection{Event Selection and Classification}
2: \label{subsec:analysis1_selection}
3: % Editors: Lisa
4:
5: % MRD sample, classify events based on track multiplicity at
6: % vertex, $\Delta\theta_p$, reconstructed neutrino energy bins.
7: % Resulting number of events in data, and plots related to event
8: % classification variables.
9:
10:
11: The fiducial volume of SciBar is defined to be 2.6 m in both the x and y directions and 1.35 m in the beam direction, for a total fiducial mass of 9.38 tons. To select charged-current events, we identify the muon produced in the interaction. We search for tracks starting in the fiducial volume of SciBar and in time with the beam that are matched with a track in the MRD or with hits in the first layer of the MRD. For track matching, the MRD track is required to start in the first layer of the MRD and stop inside the MRD. The distance between the extrapolation of the SciBar track and the actual MRD track starting point must be less than 20 cm, and the angular difference of the tracks must be less than 0.5 radians in both projections. For matching a SciBar track to first layer MRD hits, the distance between the extrapolation of the SciBar track and the hits in the MRD must be less than 20 cm. This SciBar-MRD matched track is identified as the muon for the event. If there is more than one SciBar-MRD matched track, the most energetic one is defined as the muon track. The MRD matching requirement imposes a muon momentum ($p_{\mu}$) threshold of 450 MeV/$c$.
12: The sample of events in which a SciBar-MRD track is found is our CC sample. We veto events with hits in the first layer of SciBar to eliminate events due to incoming particles produced by neutrino interactions in the other near detectors. According to MC simulation, 96\% of the events in the CC sample are true CC interactions. In addition, approximately 70\% of true CC events that produce a muon with momentum of at least 450 MeV/$c$ are successfully reconstructed in the CC sample, the main inefficiency being due to unsuccessful muon track reconstruction in the SciBar detector.
13:
14: The 3D angle of the muon with respect to the $z$-axis ($\theta_{\mu}$) can be calculated using the slopes of the track in each 2D projection.
15: The energy of the muon is reconstructed by the range and expected energy deposit per unit length in SciBar, the EC, and the MRD. In Equation \ref{emu},
16: \begin{align}
17: \label{emu}
18: E_{\mu} &= E_{\mu}^{SciBar} + E_{\mu}^{EC} + E_{\mu}^{MRD}\notag\\
19: &= L_{SciBar}(\frac{dE}{dx})_{SciBar} + \frac{dE_{0,EC}}{\cos\theta_\mu} + E_{\mu}^{MRD},
20: \end{align}
21: $E_{\mu}^{SciBar}$, $E_{\mu}^{EC}$, and $E_{\mu}^{MRD}$ are the energy deposited in each detector. $L_{SciBar}$ is the muon's range through SciBar; $\displaystyle(\frac{dE}{dx})_{SciBar}$ is set to 2.10 MeV/cm. $dE_{0,EC}$, set to 90 MeV, is the most probable value for the energy deposited in the EC for a muon crossing the EC planes perpendicularly, as estimated from the MC simulation~\cite{Brun:1987ma}.
22: $E_{\mu}^{MRD}$ is calculated from a range to energy lookup table based on GEANT3~\cite{Brun:1987ma} MC and includes the muon's mass. The average muon momentum and muon angle resolutions are 90 MeV/$c$ and 1.4 degrees, respectively.
23:
24: The upstream endpoint of the SciBar-MRD track is the reconstructed event vertex. We apply a vertex matching cut to other tracks in each event. We search for tracks starting in the fiducial volume that have an edge that is no more than 4.8 cm from the reconstructed event vertex and are in coincidence with the SciBar-MRD track within 100 ns. The 4.8 cm cut corresponds to 3$\sigma$ in vertex resolution in the $z$ dimension and more than 5$\sigma$ in the $x$ and $y$ dimensions. Figure \ref{fig:NtrackAtVtx} shows the distribution of vertex-matched tracks for data and MC. For the MC, the contributions from CCQE, CCp$\pi^{+}$, CCn$\pi^{+}$, and other nonQE interactions are shown separately.
25:
26: \begin{figure}[htp]
27: \begin{center}
28: \includegraphics[width=0.45\textwidth]{figures/NtrackAtVtx.eps}
29: \end{center}
30: \caption{\label{fig:NtrackAtVtx} (Color online) Number of vertex-matched tracks. Most events are 1-track, i.e. only the SciBar-MRD track. $\chi^2$/d.o.f = 3.46/4 considering the systematic errors discussed in Section \ref{sec:systematics}.}
31: \end{figure}
32:
33: Since the CCQE interaction is a two-body interaction, the direction of the proton can be calculated given the momentum and direction of the muon. For 2-track events, we define an angle called $\Delta \theta_{p}$ which is the angle between the expected proton track (calculated assuming the event was CCQE) and the observed second track. Figure \ref{fig:Dtheta} shows the distribution of $\Delta\theta_{p}$ for data and MC. For the MC, the contributions from CCQE, CCp$\pi^{+}$, CCn$\pi^{+}$, and other nonQE interactions are shown separately. There is some apparent discrepancy between data and MC in Figure \ref{fig:Dtheta}; however, when systematic uncertainties are taken into account (see Section \ref{sec:systematics}), the data and MC are consistent. 2-track events with $\Delta \theta_{p}$ less than 20 degrees are considered QE-like, and all other 2-track events are considered nonQE-like. The cut value of 20 degrees is chosen to maximize the purity squared times efficiency of selecting CC1$\pi^{+}$ events in the 2-track nonQE-like sample.
34:
35: \begin{figure}[htp]
36: \begin{center}
37: \includegraphics[width=0.45\textwidth]{figures/Dtheta.eps}
38: \end{center}
39: \caption{\label{fig:Dtheta} (Color online) $\Delta \theta_{p}$, angle between expected proton track assuming CCQE and the observed second track for the 2-track sample. $\chi^2$/d.o.f = 32.58/30 considering the systematic errors discussed in Section \ref{sec:systematics}.}
40: \end{figure}
41:
42: SciBar has the capability to distinguish protons from muons and pions using $dE/dx$. The MIP confidence level (MIPCL) is related to the probability that a particle is a minimum ionizing particle based on the energy deposition. The confidence level per layer is the fraction of events in the muon $dE/dx$ distribution (obtained from cosmic muons) with larger energy deposition than what is observed in that layer. The total confidence level, MIPCL, is obtained by assuming the confidence level at each layer is independent and calculating the combined probability. This variable is considered for vertex-matched tracks other than the SciBar-MRD track. Tracks with MIPCL less than 0.04 are considered proton-like and all other tracks are considered pion-like. A MIPCL cut at 0.04 maximizes the purity squared times efficiency of selecting CC1$\pi^{+}$ events in the 2-track nonQE-like pion-like sample.
43: Figure \ref{fig:MuCL} shows the distribution of MIPCL for the second track in the 2-track nonQE-like sample for data and MC. For the MC, the distributions for protons, pions, and other particles are shown separately.
44:
45: \begin{figure}[htp]
46: \begin{center}
47: \includegraphics[width=0.45\textwidth]{figures/PID_nonQE_log.eps}
48: \end{center}
49: \caption{\label{fig:MuCL} (Color online) MIP Confidence Level for the second track in the 2-track nonQE sample in log scale. $\chi^2$/d.o.f = 6.36/20 considering the systematic errors discussed in Section \ref{sec:systematics}.}
50: \end{figure}
51:
52: For this analysis, we consider only 1- and 2-track events. 2-track events are classified as QE- or nonQE-like based on $\Delta\theta_{p}$. 2-track nonQE-like events are classified as pion- or proton-like based on the MIPCL of the second track. Thus there are four samples of events: 1-track, 2-track QE, 2-track nonQE pion, and 2-track nonQE proton. Table \ref{tab:EventSamples} shows the number of data events in each sample and the composition as predicted by our nominal MC simulation. Contamination from backgrounds that are not neutrino-induced is negligible. The combined efficiency for selecting CC1$\pi^{+}$ events in these samples is 44\%.
53:
54: \input{tables/evtsamples.tex}
55: