1: \Bz mesons are exclusively reconstructed by combining two charged \Dstar
2: candidates reconstructed in a number of \Dstar and $D$ decay modes.
3: Events are pre-selected by requiring that there be
4: three or more charged tracks and that the normalized second Fox-Wolfram
5: moment~\cite{ref:fox} of the event be less than 0.6. We also require
6: that the cosine of the angle between the reconstructed $B$ direction
7: and the thrust axis of the rest of the event be less than 0.9.
8:
9: Charged kaon candidates are required to be inconsistent with the pion
10: hypothesis, as inferred from the Cherenkov ring
11: measured by the DIRC and the
12: \dedx as measured by the SVT and DCH.
13: There are two exceptions to this: tighter kaon
14: identification is applied to one of the charged kaons in decay $D^+
15: \to \Km \Kp \pip$,
16: and no particle identification requirements are made
17: for the kaon from the decay
18: $\Dz \to \Km \pip$.
19:
20: $\KS \to \pip\pim$ candidates are required to have an invariant mass
21: within 25\mevcc of the nominal \KS mass. The opening angle between the
22: flight direction and the momentum vector of the \KS candidate is
23: required to be less than 200\mrad, and the transverse flight distance
24: from the primary event vertex must be greater than 2\,mm.
25:
26: Neutral pion candidates are formed from
27: pairs of photons in the EMC with energy above 30\mev,
28: an invariant mass within 20\mevcc\ of the nominal \piz\ mass,
29: and a summed energy greater than 200\mev.
30: A mass-constraint fit is then applied to these \piz\ candidates.
31: The \piz from \Dstptopiz decays (``soft'' \piz),
32: however, is required to have an invariant mass within
33: 35\mevcc\ of the nominal \piz\ mass and momentum in the \FourS frame
34: of $70 < p^* < 450\mevc$, with no requirement on the summed
35: photon energy.
36:
37: The decay modes of the \Dz and \Dp used in
38: this analysis were selected
39: by an optimization of \SsqovSpB based on Monte Carlo simulations,
40: where $S$ is the expected number of signal events and $B$ is the
41: expected number of background events.
42: The \Dz and \Dp modes used and their branching fractions are
43: summarized in Table~\ref{dzbr}. \Dz (\Dp) meson candidates are required to
44: have an invariant mass within 20\mevcc of the nominal \Dz (\Dp) mass.
45:
46: \begin{table}
47: \caption{\label{dzbr} \Dz and \Dp decay modes and branching
48: fractions~\cite{pdg}. The branching fraction for $\KS \to \pipi$ is
49: included for modes containing a \KS.}
50: \begin{center} \begin{tabular}{lc}
51: \\ \hline \hline
52: Decay Mode & Branching Fraction (\%) \\[0.5ex]
53: \hline
54: $\Dz \to \Km \pip$ & $3.83 \pm 0.09$ \\
55: $\Dz \to \Km \pip \piz$ & $13.9 \pm 0.9$ \\
56: $\Dz \to \Km \pip \pip \pim$ & $7.49 \pm 0.31$ \\
57: $\Dz \to \KS \pip \pim$ & $1.85 \pm 0.14$ \\
58: \hline
59: Total \Dz Branching Fraction & 27.1 \\
60: \hline \hline
61: Decay Mode & Branching Fraction (\%) \\[0.5ex]
62: \hline
63: $\Dp \to \Km \pip \pip$ & $9.0 \pm 0.6$ \\
64: $\Dp \to \KS \pip$ & $0.99 \pm 0.09$ \\
65: $\Dp \to \Km \Kp \pip$ & $0.87 \pm 0.07$ \\
66: \hline
67: Total \Dp Branching Fraction & 10.9 \\
68: \hline \hline
69: \end{tabular} \end{center}
70: \end{table}
71:
72: The \Dstarp mesons
73: are reconstructed in their decays \Dstptopip and \Dstptopiz.
74: We include for this analysis the decay combinations \Dstarp\Dstarm decaying
75: to (\Dz\pip,\ \Dzb\pim) or (\Dz\pip,\ \Dm\piz), but not (\Dp\piz,\
76: \Dm\piz) due to
77: the smaller branching fraction and larger expected backgrounds.
78: The branching fractions for these modes are summarized in Table \ref{dstarbr}.
79: \Dz and \Dp candidates are subjected to a mass-constraint fit and
80: then combined with soft pion candidates. A vertex fit is performed
81: that includes the position of the beam spot to improve the angular
82: resolution of the soft pion.
83:
84: \begin{table}
85: \caption{\label{dstarbr}\Dstar decay modes and branching fractions~\cite{pdg}.}
86: \begin{center} \begin{tabular}{lcc}
87: \hline \hline
88: Particle & Decay Mode & Branching Fraction (\%) \\[0.5ex]
89: \hline
90: \Dstarp & \Dstptopip & $67.7 \pm 0.5$ \\
91: & \Dstptopiz & $30.7 \pm 0.5$ \\
92: Total Visible \Dstarp Branching Fraction & & 98.4 \\
93: %\hline
94: %\Dstarz & \Dstztopiz & $61.9 \pm 2.9$ \\
95: \hline \hline
96: \end{tabular} \end{center}
97: \end{table}
98:
99: To select \Bz candidates with well reconstructed \Dstar and $D$
100: mesons, we construct a $\chi^2$ that includes all measured \Dstar
101: and $D$ masses:
102:
103: \begin{eqnarray*}
104: \chisqM =&
105: \left(\displaystyle \frac{m_D - m_{D_{PDG}}}{\sigma_{m_D}}\right)^2
106: + \left(\displaystyle \frac{m_{\Db} - m_{\Db_{PDG}}}{\sigma_{m_{\Db}}}\right)^2 \\
107: &\quad + \left(\displaystyle\frac{\Delta m_{\Dstar} - \Delta
108: m_{\Dstar_{PDG}}}{\sigma_{\Delta m}}\right)^2
109: + \left(\displaystyle\frac{\Delta m_{\overline{D}^*} - \Delta
110: m_{\Dstar_{PDG}}}{\sigma_{\Delta m}}\right)^2
111: \end{eqnarray*}
112: where the subscript $PDG$ refers to the nominal value, and $\Delta m$
113: is the $\Dstar - D$ mass difference. For $\sigma_{m_D}$ we use
114: values computed for each $D$ candidate, while for $\sigma_{\Delta m}$
115: we use fixed values of 0.83\mevcc for \Dstptopip and 1.18\mevcc
116: for \Dstptopiz. A requirement that $\chisqM < 20$ is applied to all \Bz
117: candidates. In events with more than one \Bz candidate, we chose the
118: candidate with the lowest value of \chisqM.
119:
120: A $B$ meson candidate is characterized by two kinematic variables. We
121: use the energy-substituted mass, \mes, defined as
122: $$\mes \equiv \sqrt{{E_{Beam}^{*2}} - {p_B^*}^2}$$
123: and the difference of the $B$ candidate's energy from the beam energy,
124: \DeltaE,
125: $$\DeltaE \equiv E_{B}^* - E_{Beam}^* $$
126: %where the star indicates variables evaluated in the center-of-mass
127: %frame.
128: where $E_{B}^*$ ($p_B^*$) are the energy (momentum) of the \B\ candidate
129: in the center-of-mass frame and $E_{Beam}^*$ is one-half of the
130: center-of-mass energy.
131: The signal region in the \DeltaE {\it vs.} \mes plane is defined to be
132: $|\DeltaE| < 25\mev$ and $5.273 < \mes < 5.285\gevcc$.
133: The width of this region corresponds to approximately $\pm 2.5\sigma$
134: in both \DeltaE and \mes.
135:
136: To estimate the contribution from background in the signal region, we
137: define a sideband in the \DeltaE {\it vs.} \mes plane as
138: $$ |\DeltaE| < \DelEHiSide $$
139: $$ \mesLowSide < \mes < \mesMidSide $$
140: and
141: $$ \DelELowSide < |\DeltaE| < \DelEHiSide $$
142: $$ 5.26 < \mes < \mesHiSide $$
143: We parameterize the shape of the background in the \DeltaE {\it vs.}
144: \mes plane as the product of an ARGUS
145: function~\cite{ref:argus} in \mes and a first-order polynomial in \DeltaE.
146: %<<<<<<< Analysis.tex
147: Based on this parameterization we estimate that the ratio
148: of the number of background events in the signal region
149: to the number in the sideband region is
150: $(\fsideVal \pm \fsideErr)\times 10^{-2}$.
151: The uncertainty is derived from the observed variation of this ratio
152: under alternative assumptions
153: for the background shape in \mes\ and \DeltaE.
154: %, as well as the
155: %statistical uncertainties on the shape parameters
156: %of the ARGUS background function and first-order polynomial.
157: %The error is assigned based on variation of the shape
158: %parameterization, both by changing the type of parameterization used,
159: %and by varying the parameters describing the ARGUS function and the
160: %first-order polynomial.
161: %=======
162: %The parameterization of the shape of the background through the signal
163: %region estimates that the area of the signal region to that
164: %of the sideband is $(\fsideVal \pm \fsideErr)\times 10^{-2}$.
165: %The error is assigned based on variation of the background shape
166: %parameterization in \mes and \DeltaE, both by changing the type of parameterization used
167: %and by varying the parameters describing the Argus function and the
168: %first-order polynomial. \textit{Ie.}, to establish the systematic, we do trial fits in \mes to both
169: %constant and linear polynomials in addition to the ARGUS, and in \DeltaE we do trial fits
170: %to quadratic and exponential functions in addition to linear. The differences in the resulting
171: %estimates for the background within the signal region are used to establish the systematic
172: %due to the background parametrization.
173: %>>>>>>> 1.3
174:
175:
176: Figure~\ref{fig:mesdeltae} shows the events in the \DeltaE {\it vs.}
177: \mes plane after all selection criteria have been applied. The small
178: box in the figure indicates the signal region defined above, and the
179: sideband is the entire plane excluding the region bounded by the
180: larger box outside the signal region. There are a total of 38 events
181: located in the signal region, with 363 events in the sideband region.
182: The latter, together with the effective ratio of areas of the signal
183: region to the sideband region, implies an expected number of
184: background events in the signal region of $6.24 \pm 0.33(stat) \pm
185: 0.36 (syst)$.
186: %, giving a signal purity of 83.6\%.
187: The quoted systematic
188: uncertainty comes from the
189: background shape variation discussed previously. Figure~\ref{fig:mes}
190: shows a projection of the data on to the \mes axis after requiring
191: $|\DeltaE| < 25\mev$.
192:
193: \begin{figure}[!htb]
194: \begin{center}
195: \includegraphics[height=7cm]{offplot_data_mESdeltaE.eps}
196: \caption{Distribution of events in the \DeltaE {\it vs.} \mes plane.
197: The small box indicates the signal region, while the sideband region
198: is everything outside the larger box.
199: }
200: \label{fig:mesdeltae}
201: \end{center}
202: \end{figure}
203:
204: \begin{figure}[!htb]
205: \begin{center}
206: \includegraphics[height=7cm]{offplot_data_mES.eps}
207: \caption{Distribution of events in \mes plane with a cut of $|\DeltaE|
208: < 25\,\mev$ applied. The curve represents a fit to the distribution
209: of the sum of a Gaussian to model the signal and an ARGUS
210: function~\cite{ref:argus} to model the background shape.
211: }
212: \label{fig:mes}
213: \end{center}
214: \end{figure}
215:
216: We use a detailed Monte Carlo simulation of the \babar\ detector to
217: determine the efficiency for reconstructing the signal. This,
218: together with the total number of \BB pairs produced during data
219: collection, allows us to determine a preliminary branching fraction for
220: \Bztodstdst to be
221:
222: $$\BRbztodstdst$$
223:
224: The dominant systematic uncertainty in this measurement comes from our
225: level of understanding of the charged particle tracking efficiency (9.4\%).
226: The high charged particle multiplicity in this decay mode makes this
227: measurement particularly sensitive to tracking efficiency. Uncertainties
228: were assigned on a per track basis for $\pi$, $K$ and slow $\pi$,
229: and were added linearly due to large correlations. The imprecisely
230: known partial-wave content of the \Bztodstdst final state
231: is another source of systematic uncertainty (6.6\%). This was estimated by
232: calculating the change in the reconstruction efficiency for different
233: final angular states in Monte Carlo.
234: Other significant systematic uncertainties arise due to the
235: uncertainties on the ${\Dstar}^+$, \Dz and
236: $D^+$ branching fractions (5.6\%) and the differences in mass
237: resolutions between Monte Carlo and data (4.1\%).
238: The total systematic uncertainty from all sources is 14.5\%.
239:
240: