1: \section{Event Reconstruction and Selection}
2: \label{sec:strategy}
3:
4: The analysis proceeds in three steps: first, one of the two $B$ mesons is
5: fully reconstructed in hadronic decays (\breco), second, for the
6: recoiling $B$ meson (\bsig) we only reconstruct a charged lepton, electron or
7: muon, and then we select the exclusive decays \bet\ and \betp.
8: In order to minimize the systematic uncertainties due to the \breco selection
9: and lepton identification, the exclusive branching fractions are measured
10: relative to the inclusive semileptonic branching fraction.
11:
12: \subsection{Full Reconstruction of Hadronic $B$ Decays}
13: \label{sec:strategyBreco}
14: To reconstruct a sample of $B$ mesons, the hadronic decays $B^+
15: \rightarrow \overline{D}^0 Y^{+}, \overline{D}^{*0} Y^{+}$ are selected.
16: The system $Y^{+}$ consists of hadrons with a total charge of $+1$, composed
17: of $n_1\pi^{\pm}\, n_2K^{\pm}\, n_3\KS\, n_4\piz$, where $n_1 + n_2 \leq
18: 5$, $n_3 \leq 2$, and $n_4 \leq 2$. We reconstruct
19: $\Dstarzb \ra \Dzb\piz, \Dzb\gamma$ and $\Dzb\ra K^+\pi^-$,
20: $K^+\pi^-\piz$, $K^+\pi^-\pi^-\pi^+$, $\KS\pi^+\pi^-$ and
21: $\KS \ra \pi^+\pi^-$.
22: The kinematical consistency of \breco\ candidates
23: is checked with two variables,
24: the beam energy-substituted mass $\mes = \sqrt{s/4 - \vec{p}^{\,2}_B}$ and
25: the energy difference $\Delta E = E_B - \sqrt{s}/2$. Here $\sqrt{s}$ is the
26: total energy in the \FourS center-of-mass (CM) frame, and $\vec{p}_B$ and $E_B$
27: denote the momentum and energy of the \breco\ candidate in the same
28: frame.
29: We require $|\Delta E|<3\sigma_{\Delta E}$, where $\sigma_{\Delta E}=10$
30: to $35\mev$, depending on the decay mode, is the resolution
31: on $\Delta E$ for signal \breco\ events.
32: On average, we reconstruct one signal $\breco$ candidate in 200 \BpBm events.
33:
34: The combinatorial background from \BB\ events and
35: $e^+e^-\to q\bar{q}$ ($q=u,d,s,c$) events is subtracted by performing an
36: unbinned maximum likelihood fit
37: to the \mes\ distribution, using the following threshold function \cite{argus}:
38: \begin{equation}
39: \label{eq:argus}
40: \frac{dN}{d\mes} \propto \mes \sqrt{1-x^2} \exp \left(- \xi (1 - x^2)\right),
41: \end{equation}
42:
43: \noindent for the background (where $x = \mes/m_{\rm max}$, $m_{\rm max}$ is the endpoint
44: of the curve and $\xi$ is a free parameter determined by the fit to the \mes distribution).
45: A Gaussian function corrected for radiation losses \cite{crystallball} peaked
46: at the \B meson mass is used to describe the signal.
47:
48: \begin{figure}[!b]
49: \begin{centering}
50: \epsfig{file=ps/fig01.eps,width=8cm}
51: \caption{ Fit to the \breco \mes distribution for events with a fully
52: reconstructed \Bp decay, after semileptonic selection has been applied
53: The fitted curve (solid line) to the data points is the
54: sum of a radiation loss corrected Gaussian and a threshold function
55: (dashed line) described by Eq.~\ref{eq:argus}.
56: \label{fig:mesfitsemilep}}
57: \end{centering}
58: \end{figure}
59:
60: \subsection{Selection of Semileptonic $B$ Decays}
61:
62: The semileptonic selection identifies a charged lepton
63: with a momentum $p^{*}_{\ell}$ in the \bsig rest frame
64: greater than $0.5\gevc$ for electrons and $0.8\gevc$ for muons.
65: Electron candidates are identified using a likelihood method whose
66: efficiency is about 93$\%$ and the
67: hadron misidentification rate is less than 0.1$\%$.
68: Muons are identified with an efficiency of about 75$\%$ and the hadron
69: misidentification rate is about 3$\%$.
70: We also require the lepton and the \breco\ candidate to have opposite charge
71: and that the lepton track has not been used to reconstruct the \breco\
72: candidate.
73: Tracks are assumed to be pions if they are not identified as either
74: a muon or an electron.
75: The number of events after the semileptonic selection is obtained with
76: the \mes fit described in Section \ref{sec:strategyBreco}.
77: The fit result on data is shown in Fig.~\ref{fig:mesfitsemilep}.
78:
79: The distributions of the lepton momentum, $p^{*}_{\ell}$, computed in the
80: recoiling \B rest frame, at this stage of the selection, are shown
81: in Fig.~\ref{fig:pcms}.
82:
83: \subsection{Selection of $B\to \eta \ell\nu$ and
84: $B\to \eta^{'}\ell\nu$ Decays}
85:
86: The \bet (\betp) decay of \bsig\
87: is reconstructed by combining $\eta$ (\etap ) candidates with the charged
88: lepton.
89: We reconstruct $\eta$ candidates in three decay modes:
90: $\eta \rightarrow \gamma \gamma$
91: ($BF = 39.4\%$), $\eta \rightarrow \pip \pim \piz$ ($BF = 22.6\%$) and
92: $\eta \rightarrow \piz \piz \piz$ ($BF = 32.5\%$). The \piz candidates used
93: to build the $\eta$ are defined as pairs of photons, each with an energy in
94: the laboratory frame $E_{\gamma}>30\mev$,
95: in the invariant mass window $110<m_{\gamma\gamma}<160\mevcc$.
96: For the $\eta \rightarrow \piz \piz \piz$ channel one of
97: the three reconstructed \piz mesons should satisfy additional requirements
98: based on the shape of the neutral clusters of the electromagnetic
99: calorimeter and a tighter cut on the invariant
100: mass of the \piz ($115<m_{\gamma\gamma}<150\mevcc$). The aim of these
101: additional cuts is the reduction of the combinatorial background.
102:
103: We reconstruct $\etapr$ candidates in two decay modes:
104: $\etapr \rightarrow \rho^{0} \gamma$
105: (\rm BF = $29.5\%$) and $\etapr \rightarrow \eta \pip \pim$ (\rm BF =
106: $44.3\%$). The $\rho^{0}$ candidates used to build the \etap are reconstructed
107: as pairs of charged pions with opposite charge while the $\eta$ candidates
108: are selected as described above. In the $\etapr \rightarrow \rho^{0} \gamma$
109: channel we apply a cut on the momentum of the
110: $\gamma$ at $p^{*}_\gamma>0.35\gevc$ to remove the background from \brhoz
111: and \bclnu decays.
112:
113: \begin{figure}[!b]
114: \begin{centering}
115: \epsfig{file=ps/fig02a.eps,width=8cm}
116: \epsfig{file=ps/fig02b.eps,width=8cm}
117: \caption{ Distributions of the electron and muon momenta, $p^*_{\ell}$,
118: computed in the rest frame of the recoiling $B$, for data points
119: and Monte Carlo (histogram) for \bet (left) and \betp (right), after
120: applying all cuts of the
121: semileptonic selection except for the request on $p^*_{\ell}$, for events
122: in which at least one $\eta$($\etap$) candidate has been found.
123: The distributions are normalized to the same area. The ratio between data
124: and MC simulation is shown below the histogram.
125: \label{fig:pcms}}
126: \end{centering}
127: \end{figure}
128:
129: After these selection criteria, the dominant background is due to
130: \bclnu semileptonic decays with either a real or combinatorial $\eta^{(')}$.
131: A good rejection variable against these events is the missing four
132: momentum of the event
133: \begin{equation}
134: \label{eq:pmiss}
135: p_{miss} = p_{\Upsilon(4S)} - p_{\breco} - p_{\eta(\etap )} - p_\ell,
136: \end{equation}
137: \noindent where $p_{\Upsilon(4S)}$ is the sum of
138: the four-momenta of the colliding beams, $p_{\breco}$ is the measured
139: four-momentum of the \breco, $p_{\eta(\etap )}$ is the measured
140: four-momentum of the $\eta$ or $\etap$ and $p_{\ell}$ is the
141: measured four-momentum of the lepton. For signal events the only missing
142: particle should be a single undetected neutrino, while for background
143: events the missing momentum and energy in the event are due to other
144: undetected or poorly measured particles. Thus, in signal events the
145: resulting missing mass squared, defined as $\mmiss$ = $p_{miss}^2$, peaks
146: at zero while for background events it tends to have larger values,
147: and provides a
148: discrimination of signal and background.
149:
150: To select the decay modes of interest, the following additional selection
151: criteria are applied:
152:
153: \begin{itemize}
154:
155: \item a cut on the invariant mass of the $\eta$ and $\etap$ candidates, different for each mode;
156:
157: \item event charge balance: $Q_{tot} = Q_{\breco} + Q_{\bsig} = 0$.
158: This condition rejects preferentially $b \ra c \ell \nu$ events, since their
159: higher charge multiplicity implies a larger number of lost charged tracks;
160:
161: \item a cut on the squared missing mass, $|\mmiss|<0.5 \gev^2/c^4$;
162:
163: \item the only tracks allowed to be present in the recoil are
164: the charged lepton and the tracks used to reconstruct the $\eta$ or \etap candidate;
165:
166: \item for the \bet channel, the missing mass squared calculated assuming a
167: \bpiz decay is required to be $|\mmiss|_{\piz} > 1.5 \gev^2/c^4$.
168: This condition rejects \bpiz events, which are the main \bulnu background
169: source.
170:
171: \end{itemize}
172:
173: The selection criteria described above have been optimized by minimizing the
174: expected statistical error of the measurement and are
175: summarized in Table ~\ref{tab:cuts}.
176: After all cuts have been applied we have 10-15$\%$ signal events with more
177: than one $\eta$ (\etap) candidates for event.
178: When several candidates remain in an event after all the cuts, the one
179: with \mmiss closest to zero is chosen.
180: The selection efficiencies $\epsilon_{sel}^{excl}$ as estimated from the
181: Monte Carlo simulation are reported in Table~\ref{tab:fitres}.
182: The number of events after all analysis cuts are obtained with the fit to the
183: \mes distribution. The fit results on data are shown in
184: Fig.~\ref{fig:mesfitexcl}.
185:
186: \begin{figure}[h!]
187: \begin{centering}
188: \epsfig{file=ps/fig03a.eps,width=7cm}
189: \epsfig{file=ps/fig03b.eps,width=7cm}
190: \caption{ Fit to the \mes distribution for \betagen (left) and
191: \betapgen (right) candidates, after all selection criteria have been
192: applied. The fitted curve (solid line) to the data points is the
193: sum of a radiation loss corrected Gaussian and a threshold function
194: (dashed line) described by Eq.~\ref{eq:argus}.
195: \label{fig:mesfitexcl}}
196: \end{centering}
197: \end{figure}
198:
199: \begin{table}[!hp]
200: \begin{center}
201: \caption{Summary of event selection for \bet and \betp.
202: }
203: \vspace{0.1in}
204: \begin{tabular}{|l|l|l|}
205: \hline
206: Selection & \bet & \betp \\
207: \hline\hline
208: \piz mass & \multicolumn{2}{c|}{$110<m_{\gamma\gamma}<160 \mevcc$} \\
209: & \multicolumn{2}{c|}{$115<m_{\gamma\gamma}<150 \mevcc$ ($\eta \rightarrow \piz \piz \piz$)} \\
210: $\eta$ mass &\multicolumn{2}{c|}{} \\
211: ($\eta \rightarrow \gamma \gamma$ ) & \multicolumn{2}{c|}{$505<m_{\eta}< 585\mevcc$ } \\
212: ($\eta \rightarrow \pip \pim \piz$) & \multicolumn{2}{c|}{$530<m_{\eta}< 560\mevcc$ } \\
213: ($\eta \rightarrow \piz \piz \piz$) & \multicolumn{2}{c|}{$510<m_{\eta}< 580\mevcc$ } \\ \hline
214:
215: \etap mass & & \\
216: ($\etapr \rightarrow \rho^{0} \gamma$) & & $930<m_{\etap}<980 \mevcc$ \\
217: ($\etapr \rightarrow \eta \pip \pim$, $\eta \rightarrow \gamma \gamma$ ) & & $940<m_{\etap}<970 \mevcc$ \\
218: ($\etapr \rightarrow \eta \pip \pim$, $\eta \rightarrow \pip \pim \piz$) & & $935<m_{\etap}<975 \mevcc$ \\
219: ($\etapr \rightarrow \eta \pip \pim$, $\eta \rightarrow \piz \piz \piz$) & & $910<m_{\etap}<1000\mevcc$ \\
220: $\rho^{0}$ mass & & $595<m_{\pip\pim}<955\mevcc$ \\
221: $\gamma$ momentum & & $p^{*}_\gamma>0.35\gevc$ \\ \hline
222: Lepton momentum & \multicolumn{2}{c|}{$p_{el}^* > 0.5 \gevc$, $p_{\mu}^* > 0.8 \gevc$}\\
223: Number of leptons & \multicolumn{2}{c|}{$N_{lepton} = 1$}\\
224: Charge conservation & \multicolumn{2}{c|}{$ Q_{tot}= 0$} \\
225: Number of tracks & \multicolumn{2}{c|}{no additional charged tracks} \\
226: Charge correlation & \multicolumn{2}{c|}{$Q_{b(reco)} Q_{\ell} <0$} \\
227: Missing mass squared & \multicolumn{2}{c|}{$|\mmiss|<0.5 \gev^2/c^4$} \\ \hline
228: \bpiz rejection & $|\mmiss(\piz)| > 1.5 \gev^2/c^{4}$ & \\ \hline
229: \end{tabular}
230: \label{tab:cuts}
231: \end{center}
232: \end{table}
233:
234:
235:
236: