1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: \section{ANALYSIS METHOD}
3: \label{sec:Analysis}
4:
5: Due to the presence of multiple neutrinos, the \btn \xspace decay mode
6: lacks the kinematic constraints which are usually exploited in $B$ decay
7: searches in order to reject both continuum and $B\overline{B}$ backgrounds.
8: The strategy adopted for this analysis is to reconstruct exclusively
9: the decay of one of the $B$ mesons in the event, referred to as ``tag'' $B$,
10: and to compare the remaining particle(s) in the event, referred as the
11: ``signal side'', with the signature
12: expected for the decay \btn. In order to avoid experimenter bias, the
13: signal region in data is not examined (``blinded'') until the selection is
14: optimized based on MC simulation.
15:
16: The tag $B$ is reconstructed in the set of decay modes \btodszlnu, where
17: $\ell$ is $e$ or $\mu$. The $D^{*0}$ is reconstructed in $D^{0}\piz$ and
18: $\Dz \gamma$ modes. The $\Dz$ is reconstructed in four decay modes:
19: $K^{-}\pi^{+}$, $K^{-}\pi^{+}\pi^{-}\pi^{+}$, $K^{-}\pi^{+}\pi^{0}$, and
20: $K_{s}^{0}\pi^{+}\pi^{-}$. The $K_{s}^{0}$ is reconstructed only in the
21: mode $K_{s}^{0} \rightarrow \pi^{+}\pi^{-}$.
22: On the signal side the \btn \xspace signal is searched for in
23: both leptonic and hadronic $\tau$ decay modes:
24: $\tautoenunu$, $\tautomununu$, $\tautopinu$, $\tautopipiznu$,
25: $\tautothreepinu$. The branching fractions of the above $\tau$ decay
26: modes are listed in Table \ref{tab:TauDecayModes}.
27: Most of the kinematic variables used for
28: event selection and background rejection are measured in the
29: CM frame.
30:
31:
32: \begin{table}[h]
33: \caption{\label{tab:TauDecayModes} Branching fractions for the $\tau$ decay modes used in the \btn\ search~\cite{ref:pdg2004}.}
34: \begin{center}
35: \begin{tabular}{|l|c|}
36: \hline
37: Decay Mode & Branching Fraction (\%) \\
38: \hline\hline
39: $\tautoenunu$ & 17.84 $\pm$ 0.06 \\ \hline
40: $\tautomununu$ & 17.36 $\pm$ 0.06 \\ \hline
41: $\tautopinu$ & 11.06 $\pm$ 0.11 \\ \hline
42: $\tautopipiznu$ & 25.42 $\pm$ 0.14 \\ \hline
43: $\tautothreepinu$ & 9.16 $\pm$ 0.10 \\ \hline
44: \end{tabular}
45: \end{center}
46: \end{table}
47:
48:
49:
50: \subsection{TAG $B$ RECONSTRUCTION}
51: \label{sec:TagReco}
52:
53: The tag $B$ reconstruction proceeds as follows. First we reconstruct the
54: $\Dz$ candidates in the above four decay modes using tracks
55: and/or a $\piz$. The tracks
56: are required to meet particle identification criteria consistent
57: with the particle hypothesis, and are required to converge at a common vertex.
58: The $\piz$ candidate is required to have invariant mass between
59: 0.115--0.150 \gev/$c^2$ and its daughter photon candidates must
60: have a minimum energy of 30 \mev.
61: The mass of the reconstructed $\Dz$ candidates in
62: $K^{-}\pi^{+}$, $K^{-}\pi^{+}\pi^{-}\pi^{+}$, and $K_{s}^{0}\pi^{+}\pi^{-}$
63: modes are required to be within 40 \mev\ of the nominal mass
64: \cite{ref:pdg2004}.
65: In the $K^{-}\pi^{+}\pi^{0}$ decay mode
66: the mass is required to be within 70 \mev\ of the nominal mass
67: \cite{ref:pdg2004}.
68:
69:
70: The $D^{*0}$ candidates are reconstructed by combining the
71: $\Dz$ candidates with a soft $\piz$ or $\gamma$,
72: whose momentum in the CM frame is less than 0.45 \gev/c.
73: The mass difference between $D^{*0}$ and $D^{0}$ ($\Delta M$) is
74: restricted to be within 0.13--0.17 \gev/$c^2$ and 0.12--0.17 \gev/$c^2$
75: for $D^{0} \piz$ and $D^{0} \gamma$ modes, respectively.
76: We further require that the photon used in $D^{*0} \to \Dz \gamma$
77: reconstruction has a minimum energy of 100 \mev.
78:
79:
80: Finally $D^{*0} \ell$ candidates are reconstructed by combining
81: $D^{*0}$ with an identified electron
82: or muon of a momentum above 1.0 \gev/c in the CM frame.
83: The $D^{*0}$ and $\ell$ candidates are required to meet at a common vertex.
84: An additional kinematic constraint is imposed on the reconstructed
85: $D^{*0} \ell$ candidates:
86: Assuming that the massless neutrino is the only missing particle, we
87: calculate the cosine of the angle between the $D^{*0} \ell$ candidate
88: and the $B$ meson,
89: %
90: \begin{equation}
91: \cos\theta_{B-D^{*0}\ell} = \frac{2 E_{B} E_{D^{*0}\ell} - m_{B}^{2} - m_{D^{*0}
92: \ell}^{2}}{2|\vec{p}_{B}||\vec{p}_{D^{*0}\ell}|}.
93: \end{equation}
94: %
95: Here ($E_{D^{*0}\ell}$, $\vec{p}_{D^{*0}\ell}$) and
96: ($E_{B}$, $\vec{p}_{B}$) are the
97: four-momenta in the CM frame, and $m_{D^{*0}\ell}$ and $m_{B}$
98: are the masses of the $D^{*0}\ell$ candidate and $B$ meson, respectively.
99: $E_{B}$ and the magnitude of $\vec{p}_{B}$ are calculated
100: from the beam energy: $E_{B} = E_{\rm{beam}} = E_{\rm{CM}}/2$ and
101: $ | \vec{p}_{B} | = \sqrt{E_{\rm{beam}}^{2} - m_{B}^{2} }$.
102: Correctly reconstructed candidates
103: populate the range [-1,1], whereas combinatorial backgrounds
104: can take unphysical values outside this range.
105: We retain events in the interval
106: $-1.1 < \cos\theta_{B-D^{*0}\ell} < 1.1$, to take into account
107: the detector energy and momentum resolution.
108:
109:
110: If more than one suitable $D^{*0} \ell$ candidate is
111: reconstructed in an event, the best candidate is
112: selected based on the $\Dz$ mass and $\Delta M$.
113: A two dimensional likelihood function is formed by taking
114: the product of the $\Dz$ mass and $\Delta M$ distributions
115: obtained from MC simulation. We select the candidate
116: with the largest likelihood.
117:
118:
119: The following additional cuts are applied on the selected
120: best candidate.
121: The $\Delta M$ value for the best candidate is
122: required to be between 0.135--0.150 \gev/$c^2$ and
123: 0.130--0.155 \gev/$c^2$ for
124: candidates reconstructed in $\Dz \piz$ and
125: $\Dz \gamma$ modes, respectively.
126: The angle between the $\Dz$ and the
127: soft $\piz$ or $\gamma$ from $D^{*0}$ decay in the
128: CM frame is restricted to be less than
129: $60^{\circ}$ and $90^{\circ}$ for $D^{*0}$
130: reconstructed in $\Dz \piz$ and $\Dz \gamma$ modes,
131: respectively.
132: The sum of the charge of all the particles in the event (net charge)
133: must be equal to zero for events with selected candidates.
134:
135:
136: At this stage of the selection, the observed yield in data and
137: the predicted yield in the MC simulation agree to
138: within approximately 6\%.
139: This discrepancy is corrected by scaling the yield and efficiency
140: obtained from MC simulation.
141: The scale factor of 0.937 is used to correct
142: the tag $B$ reconstruction efficiency in the signal MC simulation.
143: The systematic error associated with this correction will
144: be described in Sec. \ref{sec:Systematics}. The corrected
145: tag reconstruction efficiency in the signal MC simulation is
146: (1.818 $\pm$ 0.074)$\times 10^{-3}$. Figure
147: \ref{fig:dmdstarl-Ds0El} shows the $\Delta M$ distributions
148: of the selected best $D^{*0} e \nu$ candidates.
149:
150: \input{figuretex/dmdstarl_ElTag_2fig}
151:
152: \subsection{SELECTION OF \btn \xspace DECAYS}
153: \label{sec:SigSelection}
154:
155: After the tag $B$ reconstruction, in the signal side
156: the $\tau$ from the $\btn$ decay is identified in one of the following modes:
157: $\tautoenunu$, $\tautomununu$, $\tautopinu$, $\tautopipiznu$, or
158: $\tautothreepinu$.
159: We select events with one or three signal-side track(s).
160: The event is rejected if any of the signal-side tracks
161: fail the following selection criteria:
162: it must have at least 12 DCH hits, its momentum transverse to the
163: beam axis, $p_{\rm{T}}$, is greater than 0.1 \gev/c, and
164: its point of closest approach to the interaction point is
165: less than 10.0~\cm\ along the beam axis and less than 1.5~\cm\ transverse
166: to the beam axis. Figure \ref{fig:nextrkroedstarlD0gammaBestCandEl}
167: shows the distribution of the number of signal-side tracks before
168: this cut.
169: The invariant mass of a signal-side $\piz$ candidate
170: must be between 0.10--0.16 \gev/$c^2$,
171: the shower shape of the daughter photon candidates must be consistent with
172: an electromagnetic shower shape and the photons
173: must have a minimum energy of 50 \mev.
174:
175:
176: The most powerful variable for separating signal
177: and background is the remaining neutral energy ($E_{\rm{extra}}$),
178: calculated by adding the CM energy of the photons that are not
179: associated with either the tag $B$ or the $\piz$ candidate from
180: $\tautopipiznu$ signal decay. The photon candidates contributing
181: to the $E_{\rm{extra}}$ variable have minimum cluster energies of 20 \mev.
182: For signal events
183: the neutral clusters contributing to $E_{\rm{extra}}$ can only
184: come from processes like beam-background, hadronic split-offs, and
185: bremsstrahlung. Therefore the signal events peak at
186: low $E_{\rm{extra}}$ values and the background events, which contain
187: additional sources of neutral clusters, are distributed
188: towards higher $E_{\rm{extra}}$ values (see fig. \ref{fig:Eextra-BestCand}).
189: The $E_{\rm{extra}} < 0.3$ \gev region is defined as the signal region.
190: This $E_{\rm{extra}} < 0.35$ \gev region, which is slightly larger than
191: the signal region, is kept blinded in on-resonance data until the selection
192: is optimized.
193:
194: \input{figuretex/nExtraTrk_Eextra_BestCand}
195:
196: The different signal modes are distinguished by the following
197: selection criteria.
198: The $\tautoenunu$, $\tautomununu$, $\tautopinu$, and $\tautopipiznu$
199: signal modes, all of which contain one charged track, are separated by
200: particle identification.
201: Both the $\tautopinu$ and the $\tautopipiznu$ modes
202: contain a pion signal track and are characterized by the
203: number of signal-side $\piz$ mesons. The $\tautothreepinu$ mode
204: contains three signal-side tracks. These signal selection
205: requirements are as follows.
206:
207:
208: \begin{itemize}
209:
210: \item{Particle identification:}
211:
212: \begin{itemize}
213: \item For the $\tautoenunu$ selection the track must be
214: identified as an electron and not identified as a muon or a kaon.
215:
216: \item For the $\tautomununu$ selection the track must be
217: identified as a muon and not identified as an electron or a kaon.
218:
219: \item For the $\tautopinu$ and $\tautopipiznu$ selection we require that
220: the track is not identified as an electron or a muon or a kaon.
221:
222: \item For the $\tautothreepinu$ selection each of the tracks
223: must be identified as a pion and not identified as
224: an electron or a muon or a kaon.
225:
226: \end{itemize}
227:
228: \item{Signal-side $\piz$ multiplicity:}
229:
230: \begin{itemize}
231:
232: \item For the $\tautopinu$ selection we require the event to contain
233: no signal-side $\piz$.
234:
235: \item For the $\tautopipiznu$ selection we require that the
236: event contains at least one signal-side $\piz$.
237:
238: \end{itemize}
239:
240: \item{$E_{\rm{extra}}$ requirement:}
241:
242: \begin{itemize}
243: \item For all the signal modes $E_{\rm{extra}}$ must be less than 0.3 \gev.
244: \end{itemize}
245:
246: \end{itemize}
247:
248:
249: Background consists primarily of $B^{+}B^{-}$ events in which the tag
250: $B$ meson has been correctly reconstructed. The recoil side contains
251: one or three track(s) and the additional particles which are not
252: reconstructed by the tracking detectors or calorimeters. Typically these
253: events contain one or more $K_{L}^{0}$ and/or neutrinos, and frequently
254: also additional charged or neutral particles which pass outside of the
255: tracking and calorimeter acceptance. Background events also contain
256: $B^{0}\overline{B^{0}}$ events. The continuum background contributes to
257: hadronic $\tau$ decay modes. In addition some excess events in data,
258: most likely from two-photon processes which are not modeled in MC
259: simulation, are also seen. These backgrounds can be suppressed by the
260: following constraints on the kinematics of the $\btn$ decays.
261:
262: \begin{itemize}
263:
264: \item{Missing mass:} The missing mass is calculated as follows.
265: %
266: \begin{equation}
267: M_{\rm{miss}} = \sqrt{ (E_{\FourS}-E_{\rm{vis}})^2 - ( \vec{p}_{\FourS} - \vec{p}_{\rm{vis}} )^2 }.
268: \end{equation}
269: %
270: Here ($E_{\FourS}$, $\vec{p}_{\FourS}$) is the four-momenta of the $\FourS$,
271: known from the beam energies. The quantities $E_{\rm{vis}}$ and $\vec{p}_{\rm{vis}}$ are the
272: total visible energy and momentum of the event which are calculated by adding the
273: energy and momentum, respectively, of all the reconstructed
274: charged tracks and photons in the event.
275:
276: \begin{itemize}
277:
278: \item For the $\tautoenunu$ and $\tautomununu$ selections, events with
279: missing mass less than 4 \gev/$c^2$ are rejected.
280:
281: \item For the $\tautopinu$ and $\tautopipiznu$ selections,
282: the missing mass is required to be greater than 3 \gev/$c^2$.
283:
284: \item For the $\tautothreepinu$ selection, the missing mass is
285: required to be greater than 2 \gev/$c^2$.
286:
287: \end{itemize}
288:
289:
290:
291: \item{Maximum CM momentum of the $\tau$ daughter:}
292:
293: The following maximum CM momentum requirements are applied to
294: the $\tau$ daughter particles.
295:
296: \begin{itemize}
297:
298: \item The electron candidate from the $\tautoenunu$ decay must have a CM momentum of less than 1.4 \gev/c.
299: The CM momentum requirement is not applied to the
300: $\tautomununu$ selection because of the following reason:
301: The momentum spectrum of the lepton from $\tau$ decays peaks below 1 \gev/c.
302: The particle identification efficiency for low momentum muons is lower than
303: that for low momentum electrons. Therefore, applying the maximum momentum
304: cut reduces the selection efficiency of the $\tautomununu$ mode
305: significantly.
306:
307:
308: \item For the three hadronic $\tau$ decay modes,
309: the $\pi$ from $\tautopinu$, the $\pi \piz$ combination from
310: $\tautopipiznu$, or the $3 \pi$ combination from $\tautothreepinu$
311: must all have CM momenta less than 2.7 \gev/c.
312:
313: \end{itemize}
314:
315: \end{itemize}
316:
317: The $\tautopipiznu$ and $\tautothreepinu$ decays proceed via intermediate
318: resonances. For these modes further background rejection can be
319: achieved by applying the following requirements on the intermediate mesons.
320:
321: \begin{itemize}
322:
323: \item{$\rho^{+}$ selection:}
324:
325: The $\tautopipiznu$ decay proceeds via an intermediate $\rho^{+}$
326: state. The signal-side track is combined with a signal-side $\piz$ to form
327: the $\rho^{+}$ candidate.
328: In events with more than one signal-side $\piz$, the candidate with
329: invariant mass closest to the nominal $\piz$ mass \cite{ref:pdg2004}
330: is chosen. The invariant mass of the
331: reconstructed $\rho^{+}$ is required to be within
332: 0.55--1.00 \gev/$c^2$. A quantity similar to $\cos \theta_{B-D^{*0} \ell}$,
333: which is defined in section \ref{sec:TagReco}, can be
334: reconstructed for $\tau \to \rho \nu$ as follows:
335: %
336: \begin{equation}
337: \cos\theta_{\tau-\rho} = \frac{2 E_{\tau} E_{\rho} - m_{\tau}^{2} - m_{\rho}^{2}}{2|\vec{p}_{\tau}||\vec{p}_{\rho}|},
338: \end{equation}
339: %
340: where ($E_{\tau}$, $\vec{p}_{\tau}$) and
341: ($E_{\rho}$, $\vec{p}_{\rho}$) are the
342: four-momenta in the CM frame, $m_{\tau}$ and $m_{\rho}$
343: are the masses of the $\tau$ and $\rho$ candidate, respectively.
344: %
345: The quantities $|\vec{p}_{\tau}|$ and $E_{\tau}$ are calculated
346: assuming that the $\tau$ is from the $\btn$ decay, and
347: that the $B^{+}$ is almost at rest in the CM frame.
348: %
349: Candidates outside of $-1.1 < \cos\theta_{\tau-\rho} < 1.1$
350: are excluded.
351:
352:
353: \item{$\rho^{0}$ and $a_{1}^{+}$ selection:}
354:
355: The $\tau$ decays to three charged tracks via two intermediate resonances:
356: $\tau^{+} \to a_{1}^{+} \nutb$,
357: $a_{1}^{+} \to \rho^{0} \pi^{+}$, and
358: $\rho^{0} \to \pi^{+} \pi^{-}$.
359: The $\pi^{-}\pi^{+}$ combination with an invariant mass closest to the
360: nominal $\rho^{0}$ mass \cite{ref:pdg2004}
361: is selected as the best $\rho^{0}$ candidate.
362: The invariant mass of the selected $\rho^{0}$ must be within
363: 0.55--1.00 \gev/$c^2$. The CM momentum of the
364: selected $\rho^{0}$ candidate is required to be greater than 0.5 \gev/c.
365: The invariant mass of the three signal tracks must be within
366: 1.0--1.6 \gev/$c^2$. The total CM momentum of the
367: three tracks has to be greater than 1.0 \gev/c. The
368: three tracks are also required to converge to a common vertex and the
369: candidates are rejected if the vertex fit probability is less than 0.1\%.
370: For $\tau \to a_{1} \nu$ decay the quantity,
371: %
372: \begin{equation}
373: \cos\theta_{\tau-a_{1}} = \frac{2 E_{\tau} E_{a_{1}} - m_{\tau}^{2} - m_{a_{1}}^{2}}{2|\vec{p}_{\tau}||\vec{p}_{a_{1}}|},
374: \end{equation}
375: %
376: is obtained using the similar procedure used for calculating
377: $\cos \theta_{\tau-\rho}$. Here ($E_{\tau}$, $\vec{p}_{\tau}$) and
378: ($E_{a_1}$, $\vec{p}_{a_1}$) are the
379: four-momenta in the CM frame, $m_{\tau}$ and $m_{a_1}$
380: are the masses of the $\tau$ and $a_1$ candidate, respectively.
381: Candidates not satisfying
382: $-1.1 < \cos\theta_{\tau-a_{1}} < 1.1$ are excluded.
383:
384: \end{itemize}
385:
386:
387: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
388:
389: The signal selection criteria for all five signal modes are summarized in Table
390: \ref{tab:SigSelSummary}.
391:
392: \begin{table}[!htb]
393: \caption{The selection criteria for different signal modes are listed in this table. The symbols $P^{*}_{x}$ and $M_{x}$, used in the table, correspond to the CM momentum and invariant mass of $x$, respectively.}
394: \vspace{-.3cm}
395: \begin{center}
396: \footnotesize
397: \begin{tabular}{|c|c|c|c|c|} \hline
398: $\tautoenunu$ & $\tautomununu$ & $\tautopinu$ & $\tautopipiznu$ & $\tautothreepinu$ \\ \hline \hline
399: \multicolumn{4}{|c|}{One signal-side track} & Three signal-side tracks \\ \hline
400: \multicolumn{5}{|c|}{ track quality requirements for each signal track } \\ \hline
401: electron & muon & \multicolumn{2}{|c|}{not electron} & pion \\
402: not muon & not electron & \multicolumn{2}{|c|}{not muon} & not electron \\
403: not kaon & not kaon & \multicolumn{2}{|c|}{not kaon} & not muon \\
404: & & \multicolumn{2}{|c|}{} & not kaon \\ \hline
405: none & none & No signal-side $\pi^{0}$ & Non-zero signal-side $\pi^{0}$ & none \\ \hline
406: \multicolumn{5}{|c|}{ $E_{\rm{extra}} < 0.3$ $\gev$} \\ \hline
407: \multicolumn{2}{|c|}{Missing Mass $>$ 4 $\gev$} & \multicolumn{2}{|c|}{Missing Mass $>$ 3 $\gev$} & Missing Mass $>$ 2 $\gev$ \\\hline
408: $P^{*}$ of & none & $P^{*}$ of & $P^{*}$ of & $P^{*}$ of \\
409: signal-side e & & signal-side $\pi$ & signal-side & of signal-side \\
410: track $<$ 1.4 $\gev$ & & track $<$ 2.7 $\gev$ & $\pi \pi^{0} < $ 2.7 $\gev$ & 3 $\pi < 2.7$ $\gev$ \\ \hline
411: none & none & none & $\rho^{+}$ selection: & $a_{1}^{+}$ selection: \\
412: & & & 0.55 $< M_{\rho^{+}}< $ 1 $\gev$ & 0.55 $ < M_{\pi^+ \pi^-} <$ 1.0 $\gev$ \\
413: & & & $-1.1 < \cos\theta_{\tau-\rho} < 1.1$ & $P^{*}_{\pi^+ \pi^-} >$ 0.5 $\gev$ \\
414: & & & & 1.0 $ < M_{3 \pi} < $ 1.6 $\gev$ \\
415: & & & & $P^{*}_{3 \pi} >$ 1.0 $\gev$ \\
416: & & & & Vertex prob. of 3 tracks $> 1 \%$ \\
417: & & & & $-1.1 < \cos\theta_{\tau-a_{1}} < 1.1$ \\ \hline
418: \end{tabular}
419: \label{tab:SigSelSummary}
420: \end{center}
421: \end{table}
422:
423: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
424:
425: \subsubsection{SIGNAL EFFICIENCY}
426: \label{sec:SigEff}
427:
428:
429: The ``signal-side selection efficiencies''
430: for the $\tau$ decay modes are
431: determined from signal MC simulation and summarized
432: in Table \ref{tab:Comb-Final-SigSelEff}.
433: For each $\tau$ selection mode, the signal-side
434: efficiency ($\varepsilon_{i}$, where $i$ $\equiv$ selection mode)
435: is computed as the ratio of the number of events surviving
436: the requirements of that selection mode
437: to the number of events where a tag $B$ meson is reconstructed.
438: In the computation of the total
439: signal-side efficiency for each selection
440: we take into account the cross-feed from other $\tau$ decay modes reported in
441: Table \ref{tab:Comb-Final-SigSelEff}.
442:
443: The selection efficiency for $\tautomununu$ is low compared to that of the
444: $\tautoenunu$ mode, because of the fact that the momentum spectrum
445: of the signal muons peaks below 1 \gev/c, where the muon detection
446: efficiency is low. Since no minimum momentum requirement and no tight pion
447: identification criteria are applied to the
448: $\tautopinu$ signal selection, electron and muon signal tracks
449: that fail particle identification requirement get selected in this mode.
450: Any true $\tautopipiznu$ signal events, with a missed $\piz$
451: also get included in $\tautopinu$ selection mode.
452: Therefore the $\tautopinu$ selection
453: mode has the highest signal efficiency.
454: From MC estimation we expect $\sim$10 signal events at 112.5 $\rm{fb^{-1}}$,
455: assuming $\mathcal{B}(\btn)=10^{-4}$.
456:
457:
458: \begin{table}[!htb]
459: \caption{Efficiency of the different selections (columns) for the most abundant $\tau$ decay modes (rows). The last two rows show the total efficiency for each selection weighted by the decay branching fractions, and the total efficiency. The errors are statistical only. The total efficiency for each selection is $\varepsilon_{i} = \Sigma_{j=1}^{8} \varepsilon_{i}^{j} f_{j}$, where $\varepsilon_{i}^{j}$ is the efficiency of the selection $i$ for the simulated $\tau$ decay mode $j$. The index $j$ corresponds to the different $\tau$ decay mode in the MC simulation, and $f_{j} = \mathcal{B}(\tau \to j)$ are the $\tau$ branching fractions from Ref. \cite{ref:pdg2004}. If the efficiency is zero a 90\% upper limit is quoted.}
460: \begin{center}
461: \input{table-contents/Comb-Final-SigSelEff}
462: \end{center}
463: \label{tab:Comb-Final-SigSelEff}
464: \end{table}
465:
466: \subsubsection{EXPECTED BACKGROUND FROM MONTE CARLO SIMULATION}
467: \label{sec:MCBkgEst}
468:
469: To obtain the background estimation from the MC simulation, \BB\ and
470: $\epem \to$ \uubar, \ddbar, \ssbar, \ccbar, and \tautau\ events are
471: scaled to equivalent luminosity in data. The estimated background
472: in different selection modes is listed in Table
473: \ref{tab:SummaryMCBkgEstSigYield}.
474: The three modes $\tautoenunu$, $\tautomununu$, $\tautopinu$ are
475: relatively clean compared to the other two selection
476: modes. Signal to background ratios in $\tautoenunu$,
477: $\tautomununu$, and $\tautopinu$ selection modes are
478: 10:59, 10:75, and 10:85 respectively,
479: assuming $\mathcal{B}(\btn) = 10^{-4}$.
480: For the $\tautopipiznu$ and $\tautothreepinu$ modes the
481: signal to background ratio is 10:354 and 10:513, respectively.
482:
483:
484: \begin{table}[!htb]
485: \caption{Expected final raw background and signal yield at \onlumi \xspace estimated from background and signal MC simulation. No systematic correction is applied on simulated events. The listed errors are statistical only. If the efficiency is zero a 90\% upper limit on the expected background is quoted.}
486: \vspace{-.30cm}
487: \begin{center}
488: \input{table-contents/final-Comb-MCBkgEst-afterAllCuts}
489: \label{tab:SummaryMCBkgEstSigYield}
490: \end{center}
491: \end{table}
492:
493: % ----------------------------------------------------------------------
494: \subsubsection{BACKGROUND ESTIMATION FROM $E_{\rm{extra}}$ SIDE BAND IN DATA}
495: \label{sec:EextraSBExtrapolation}
496:
497:
498: The $E_{\rm{extra}} < 0.3$ \gev\ region is defined as the ``signal
499: region'' and the $0.35 < E_{\rm{extra}} < 1.0$ \gev\ region
500: is defined as the ``side band''.
501: The $E_{\rm{extra}}$ shape in the MC distribution is used to extrapolate the data
502: side band to the signal region.
503:
504:
505: The number of MC events in signal region ($N_{\rm{Sig}}^{\rm{MC}}$)
506: and side band ($N_{\rm{SideB}}^{\rm{MC}}$) are counted
507: and their ratio ($R_{\rm{MC}}$) is obtained.
508:
509: \begin{eqnarray*}
510: R_{\rm{MC}} & = & \frac{N_{\rm{Sig}}^{\rm{MC}}}{N_{\rm{SideB}}^{\rm{MC}}}
511: \end{eqnarray*}
512:
513: \noindent Using the number of data events in the side band ($N_{\rm{SideB}}^{\rm{data}}$)
514: and the ratio $R_{\rm{MC}}$, the number of expected background events in the
515: signal region in data ($N_{\rm{SideB}}^{\rm{data}}$) is estimated.
516:
517: \begin{eqnarray*}
518: N_{\rm{SideB}}^{\rm{data}} & = & N_{\rm{SideB}}^{\rm{data}} \cdot R_{\rm{MC}}
519: \end{eqnarray*}
520:
521: \noindent The background estimation for the different selection modes from
522: the $E_{\rm{extra}}$ side band extrapolation are shown in Table
523: \ref{tab:BkgEst_EleftSB_Extrapolation}. The number of estimated
524: background events in the signal region from the data side band extrapolation
525: are in agreement with the background estimation from MC simulation within
526: statistical uncertainty.
527:
528:
529: \begin{table}[!htb]
530: \caption{Background estimation in the signal region
531: ($E_{\rm{extra}} <$ 0.3 \gev) for the different selection modes.}
532: \vspace{-1.0cm}
533: \begin{center}
534: \include{BkgEst_EleftSB_Extrapolation_table}
535: \label{tab:BkgEst_EleftSB_Extrapolation}
536: \end{center}
537: \end{table}
538:
539:
540:
541:
542: % ----------------------------------------------------------------------
543: \section{VALIDATION OF $E_{\rm{extra}}$ SIMULATION}
544: \label{sec:EextraValidation}
545:
546: The $E_{\rm{extra}}$ distribution in signal and background MC simulation
547: are validated using various control samples.
548: We compare the $E_{\rm{extra}}$ distributions between on-resonance data and
549: MC simulation, in both signal region and side band, using
550: the control samples. Agreement between the distributions would provide
551: validation of the $E_{\rm{extra}}$ modeling in the simulation.
552:
553:
554: \subsection{$E_{\rm{extra}}$ IN THE SIGNAL MONTE CARLO SIMULATION}
555: \label{sec:SigMCEextraValidation}
556:
557: The ``double-tagged'' events, for which
558: both of the $B$ mesons are reconstructed in tagging modes,
559: $B^{-} \rightarrow D^{*0} \ell^{-} \overline{\nu}_{\ell}$ vs
560: $B^{+} \rightarrow \overline{D}^{*0} \ell^{+} \nu_{\ell}$,
561: are used as a control sample to validate the $E_{\rm{extra}}$
562: simulation. Due to the large branching fraction and
563: high tagging efficiency for these events, a sizable
564: sample of such events
565: is reconstructed in the on-resonance dataset.
566: These double-tag events contain very little background due to the
567: full reconstruction of the event.
568:
569: To select double-tag events
570: we require that the two tag $B$ candidates do not share
571: any tracks or neutrals. If there are more than two such
572: non-overlapping tag $B$ candidates in the event then the
573: best two are selected using
574: the same best candidate selection criteria, as described in
575: Sec. \ref{sec:TagReco}.
576: After selecting the two tag $B$ candidates only
577: the events with no extra charged tracks are selected.
578:
579: The $E_{\rm{extra}}$ for the double-tagged sample is calculated by summing the
580: CM energy of the photons which are not associated with
581: either of the tag $B$ candidates. The sources of neutrals contributing
582: to the $E_{\rm{extra}}$ distribution in double-tagged events
583: are similar to those contributing to the $E_{\rm{extra}}$ distribution
584: in the signal MC simulation. Therefore the agreement of
585: the $E_{\rm{extra}}$ distribution between data and MC simulation for the
586: double-tagged sample, in figure \ref{fig:DoubleTageCNSum},
587: is used as a validation of the $E_{\rm{extra}}$ simulation in the signal
588: MC.
589:
590: \input{figuretex/DoubleTageCNSum-BestCand}
591:
592:
593: \subsection{$E_{\rm{extra}}$ IN THE BACKGROUND MONTE CARLO SIMULATION}
594: \label{sec:BkgMCEextraValidation}
595:
596: The following two background control samples are used to study the
597: agreement between data and simulation in the $E_{\rm{extra}} < 0.35$ \gev\
598: region.
599:
600: \begin{itemize}
601:
602: \item Events with two remaining signal-side tracks
603:
604: \item Events with non-zero net charge
605:
606: \end{itemize}
607:
608: The procedure from Sec. \ref{sec:EextraSBExtrapolation},
609: the background estimation from the $E_{\rm{extra}}$ side band, is used
610: for this test. In Table \ref{tab:SideBand_Extrapolation}
611: we show the comparison of the number of expected data
612: events ($N_{\rm{SideB}}^{\rm{data}}$) in the signal region with the observed number of data
613: events ($N_{\rm{Sig}}^{\rm{obs}}$) in the signal region.
614: The agreement between
615: the above two quantities provides validation of background estimation in the
616: low $E_{\rm{extra}}$ region.
617:
618:
619: \begin{table}[!htb]
620: \caption{Test of background estimation in the low $E_{\rm{extra}}$ region from various control samples. Signal selection cuts for different $\tau$ decay modes are applied on the control samples. The expected and observed number of events in the signal region agree within error for all the samples ($P(\chi^2)=0.15$ between the entries in the last two columns. Here $\chi^2 = 10.77$, and the number of degrees of freedom is 7.).}
621: \vspace{-1.0cm}
622: \begin{center}
623: \include{SideBand_Extrapolation_table}
624: \label{tab:SideBand_Extrapolation}
625: \end{center}
626: \end{table}
627:
628:
629:
630:
631:
632:
633:
634: