1: \section{Analysis}
2: \label{sec:analysis}
3:
4: \subsection{Monte Carlo simulation}
5: \label{sec:analysis-mc}
6:
7: We used JETSET7.3\cite{L73} for $e^{+}e^{-} \rightarrow
8: q\bar{q}$ event generations. The used parameters in JETSET7.3 were
9: tuned-up for hadronic events using the event shape
10: data \cite{Ohnishi}. In the hadronization,
11: we used the LUND symmetric function for the fragmentation function
12: for light quarks (u,d,s) with $a = 0.413$ and $b = 0.9$.
13: For heavy quarks (c,b), a function by
14: Peterson {\it et al.}\cite{Peterson} was used
15: with $\epsilon_{c} = 0.05$ and $\epsilon_{b} = 0.01$.
16: For the standard model parameters, we used $\sin^{2}\theta_{W} = 0.2315,
17: M_{Z^{0}} = 91.187$ GeV/c$^{2}$, and $\Gamma_{Z^{0}} = 2.490$ GeV\cite{PDG96}.
18:
19: %\paragraph{Simulator}
20:
21: For a detector simulation, we used the TOPAZ detector
22: simulator, which simulated the behaviors of the particles
23: in the TOPAZ detector: such as the energy loss,
24: multiple scattering, decay-in-flights and the detector signals.
25: For the simulations of particle's interactions with
26: the detector material, EGS4\cite{EGS4} was used for
27: electromagnetic processes and GHEISHA-7\cite{GHEI7} for nuclear
28: interactions. We used 314,463 hadronic events for the studies described
29: in this paper.
30:
31: To test the validity of the hadronic event selection and the Monte-Carlo
32: simulation, we examined the general features of hadronic events.
33: Figure \ref{fig:hadron} shows the distributions of the momentum (a)
34: and the polar angle (b) of the track,
35: and the polar angle of the thrust axis (c).
36: The data and the Monte-Carlo results are in good agreement, except for a dip
37: observed at $\theta=90^{\circ}$ in the polar-angle distribution.
38: This dip is due to the effect of the central membrane of TPC.
39: The track reconstruction inefficiency at the membrane is 0.4\% for
40: hadronic events and 0.8\% for inclusive muon events. The effect of this
41: inefficiency is negligible on the cross section and asymmetry results.
42:
43: \begin{figure}
44: \input epsf
45: \begin{center}
46: \leavevmode
47: \epsfxsize=15.0cm
48: \epsfbox{fig4.eps}
49: \caption{Hadronic events; momentum (a) and
50: polar angle (b) of good tracks, and angular distribution of
51: thrust axis (c). The dots with error bars are the data;
52: the histograms are Monte-Carlo results.}
53: \label{fig:hadron}
54: \end{center}
55: \end{figure}
56:
57: \subsection{Background estimation}
58:
59: Possible background sources to prompt muons are
60: hadron punch-throughs, muons from decay-in-flights of
61: light hadrons (mainly $\pi^{\pm}$ and $K^{\pm}$) and accidental hits
62: in MDC due to the beam backgrounds or cosmic rays. To estimate the
63: rate of accidental hits, we applied the muon identification criteria to the
64: electron (positron) tracks in Bhabha events, obtained in the same experimental
65: period. No track was identified as a muon. From this result, the accidental
66: hit rate was considered to be negligible.
67:
68: \begin{figure}
69: \input epsf
70: \begin{center}
71: \leavevmode
72: \epsfxsize=10.0cm
73: \epsfbox{fig5.eps}
74: \caption{Punch-through rate as a function of momentum derived through the pions
75: from $\tau$-pairs.
76: The dots with error bars are the data. The dotted and dashed lines are
77: Monte Carlo simulations before and after tuning the parameters in GHEISHA.}
78: \label{fig:scale-factor-mdc}
79: \end{center}
80: \end{figure}
81:
82: The other background sources were evaluated using Monte-Carlo
83: simulations. Since the pion and kaon lifetimes and decay modes are well
84: established, the decay-in-flights of light hadrons can be precisely
85: calculated in the simulation. On the other hand,
86: the rate of hadron punch-throughs depends on nuclear interactions with
87: detector materials.
88: Therefore, it must be verified experimentally. For this purpose,
89: we studied the punch-through of pions by using 1,202 charged pions from
90: $\tau \rightarrow \pi\pi\pi\nu$ and $\tau \rightarrow
91: \rho\nu \rightarrow \pi^{\pm}\pi^{0}\nu$ in the
92: $e^{+}e^{-} \rightarrow \tau^{+}\tau^{-}$ events.
93: Figure \ref{fig:scale-factor-mdc} shows the experimental ratio together
94: with the Monte-Carlo expectation as a function of the momentum. As
95: can be clearly seen in this figure, it was found that the
96: Monte-Carlo prediction significantly underestimated the rate,
97: especially in the high-momentum region
98: with the default parameters. To correct this difference,
99: we tuned up the most influential
100: parameter, which is the total cross section of pions (kaons) interacting
101: with materials in the GHEISHA routine. We introduced a correction
102: factor to scale the total cross section as a function of the momentum, as
103: listed in Table~\ref{tab:scale}. In the high-momentum region of $p >
104: 10$ GeV/c, the correction factor was determined through the pion punch-through
105: rate in the $\tau$ pair events. For $p< 3$ GeV/c, the correction factor
106: was estimated through energy deposits of hadrons in BCL, and it was
107: fitted to a function. These two correction factors were linearly
108: connected in the intermediate momentum regions, i.e. $3 < p < 10$ GeV/c.
109:
110: The results of the tuned-up Monte-Carlo simulation are also shown in
111: Figure \ref{fig:scale-factor-mdc}. The discrepancy between the
112: Monte-Carlo results and the data was reduced significantly
113: with the tuned-up Monte-Carlo procedure.
114:
115: \begin{table}[h]
116: \caption{Correction factor for the cross section for each momentum region.}
117: \label{tab:scale}
118: \begin{center}
119: \begin{tabular}{ll}
120: \hline \hline
121: Momentum region (GeV/c)& correction factor $f(p)$\\
122: \hline
123: $p<3$ & $f(p) = -0.013p^{2} + 0.139p + 0.826$ \\
124: $3 \leq p<10$ & $f(p) = -0.054p + 1.295$ \\
125: $10\leq p$ & $f(p) = 0.726$ \\
126: \hline \hline
127: \end{tabular}
128: \end{center}
129: \end{table}
130:
131: In the inclusive muon sample, the fractions of decay-in-flight and
132: hadron punch-through were estimated to be 25\% and 23\%, respectively,
133: by using the tuned-up Monte-Carlo program.
134: Figure \ref{fig:inclusive-mu} shows the momentum spectrum (a) and polar-angle
135: distribution (b).
136: The background from light quarks was obtained by a Monte-Carlo
137: simulation using the total number of hadronic events for normalization.
138:
139: \begin{figure}
140: \input epsf
141: \begin{center}
142: \leavevmode
143: \epsfxsize=15.0cm
144: \epsfbox{fig6.eps}
145: \caption{Distributions of momentum (a) and polar angle (b) of inclusive muons.
146: The dots with error bars are data and the histograms are Monte-Carlo results.
147: The open regions show the contributions from c and b quarks.
148: The shaded regions are the background from light quarks.}
149: \label{fig:inclusive-mu}
150: \end{center}
151: \end{figure}
152:
153: \subsection{Flavor separation}
154:
155: Inclusive muon events can be categorized into
156: four groups by the sources: (A) prompt muons from primary b-decays
157: ($b \rightarrow \mu$), (B) cascade muons from b $\rightarrow$ c($\bar{c}$)
158: decays ($b \rightarrow c \rightarrow \mu$), (C) prompt muons from
159: primary c-decays ($c \rightarrow \mu$), and (D) backgrounds from
160: decay-in-flights and punch-throughs ($others$).
161: In this analysis, we included the $b \rightarrow \tau \rightarrow \mu$ mode
162: in category (A).
163: Table \ref{tab:composition} gives the results of Monte-Carlo
164: studies, giving the fraction of each components in the inclusive muon events.
165:
166: We used the thrust axis to determine the primary quark direction.
167: The accuracy of the quark direction determined by this method was estimated
168: to be $5^{\circ}$, based on a study by a Monte-Carlo simulation.
169: We define the angle of the quark direction by
170: $-Q\cos\theta_{trust}$, where $Q$ is the charge of the tagged muon
171: and $\theta$ is the angle of the thrust axis with respect to the beam axis.
172: To separate the quark flavors, we used the muon transverse momentum
173: with respect to the jet axis, $p_{T}^{jet}$. Jets were reconstructed
174: using the JADE jet clustering algorithm\cite{JET} with a scaled
175: invariant mass cut, $Y_{cut}(= M_{ij}/E_{vis}) = 0.04$.
176: Muons from b-quarks have a larger $p_{T}^{jet}$ compared with those
177: from c-quarks, due to the heavy b-quark mass.
178: Using the $p_{T}^{jet}$ cut at 0.8 GeV/c, we classified the inclusive
179: muon events into a b-enriched sample ($p_{T}^{jet} \ge 0.8$ GeV/c) and
180: a c-enriched sample ($p_{T}^{jet}<0.8$ GeV/c).
181: We chose $p_{T}^{jet} = 0.8$ GeV/c, because the systematic
182: uncertainty was the smallest at 0.8 GeV/c.
183: The purity of the c- and b-quarks as a function of the $p_{T}$ cut
184: was studied by a Monte-Carlo simulation.
185: Figure \ref{fig:ptcut} shows the b, c-quark purities as a
186: function of the $p_{T}$ cut. For b-quarks, the purity was calculated by
187: summing up all of the b-quarks with $p_{T}$ greater than a given $p_{T}$ cut,
188: divided by the total number of inclusive muon tracks
189: in the same $p_{T}$ region.
190: For c-quarks, summing was done for $p_{T}$ less than
191: a given $p_{T}$ cut.
192:
193: \newcommand{\lw}[1]{\smash{\lower2.0ex\hbox{#1}}}
194: \begin{table}[h]
195: \caption{Percentage of muons from each source in the inclusive muon sample.}
196: \label{tab:composition}
197: \begin{center}
198: \begin{tabular}{llll}
199: \hline \hline
200: \lw{Source} & \multicolumn{3}{c}{\rm Fraction$(\%)$} \\
201: & $p^{jet}_{T} < 0.8$ GeV/c & $p^{jet}_{T} \geq 0.8$ GeV/c & total \\
202: \hline
203: $b \rightarrow \mu$ & 9.7 & 41.9 & 20.1 \\
204: $b \rightarrow c \rightarrow \mu$ & 5.8 & 5.2 & 5.6 \\
205: $c \rightarrow \mu$ & 31.6 & 13.2 & 25.7 \\
206: $others$ & 52.8 & 39.6 & 48.5 \\
207: \hline \hline
208: \end{tabular}
209: \end{center}
210: \end{table}
211:
212: \begin{figure}
213: \input epsf
214: \begin{center}
215: \leavevmode
216: \epsfxsize=15.0cm
217: \epsfbox{fig7.eps}
218: \caption{ $p^{jet}_{T}$ distribution of inclusive muons (a) and
219: purities of b and c-quarks as a function of $p^{jet}_{T}$ cut (b).
220: In (a), dots with errors show the data.
221: Histograms are Monte Carlo: $b \rightarrow \mu$(open),
222: $b \rightarrow c \rightarrow \mu$(hatched),
223: $c \rightarrow \mu$(double-hatched) and
224: $others$(triple-hatched), respectively.
225: In (b), solid curve shows the b-quark purity
226: and dotted curve shows the c-quark purity.}
227: \label{fig:ptcut}
228: \end{center}
229: \end{figure}
230:
231: \subsection{Simultaneous fit for b and c-quark}
232:
233: In a previous paper\cite{shimonaka}, we derived only the b-quark parameters
234: by assuming the standard-model parameters for the c-quark because of low
235: statistics. This time, we
236: performed a four-parameter fit of ($R_{b\bar{b}}, R_{c\bar{c}},
237: A_{FB}^{b}, A_{FB}^{c}$) to $p$ and $-Q\cos\theta_{thrust}$
238: ($p_{T}^{jet} \ge 0.8$ GeV/c and $<0.8$ GeV/c) distributions.
239: Definitions of $R_{q\bar{q}}$, and $A_{FB}$ are given by
240: %
241: \begin{equation}
242: R_{q\bar{q}}\equiv\sigma(e^{+}e^{-}\rightarrow q\bar{q})/
243: \sigma(e^{+}e^{-}\rightarrow hadrons)
244: \end{equation}
245: and
246: \begin{equation}
247: A_{FB}\equiv{\sigma_{F}-\sigma_{B}\over\sigma_{F}+\sigma_{B}}
248: \end{equation}
249: where $\sigma$ represents the lowest order (SU(3) $\times$ SU(2)$_{L}$
250: $\times$ U(1)) cross-section and
251: $\sigma_{F}(\sigma_{B})$ is the cross-section for quark to
252: travel forward (backward) with respect to the $e^{-}$ direction.
253: %
254: The $\chi^{2}$ of the fit was defined by
255: %
256: \begin{eqnarray}
257: \chi^{2}_{total} & = & \chi^{2}_{\cos\theta(p_{T}^{jet} \ge 0.8 GeV/c)}
258: + \chi^{2}_{\cos\theta(p_{T}^{jet}<0.8 GeV/c)}
259: + \chi^{2}_{p}\nonumber\\
260: & = & \sum_{i=1}^{8}\frac{(N^{\mu}_{i} -
261: \tilde{N}^{\mu}_{i})^{2}}{\sigma_{N^{\mu}_{i}}^{2}}
262: + \sum_{j=1}^{8}\frac{(N^{\mu}_{j} -
263: \tilde{N}^{\mu}_{j})^{2}}{\sigma_{N^{\mu}_{j}}^{2}}
264: + \sum_{k=1}^{6}\frac{(N^{\mu}_{k} -
265: \tilde{N}^{\mu}_{k})^{2}} {\sigma_{N^{\mu}_{k}}^{2}},
266: \end{eqnarray}
267: %
268: where $N^{\mu}_{i}$ and $\sigma_{N^{\mu}_{i}}$ are the number of inclusive
269: muon events and the statistical error for each bin, respectively.
270: $\tilde{N}^{\mu}_{i}$ is
271: the number of expected events for each bin, which is defined by
272: %
273: \begin{eqnarray}
274: \tilde{N}^{\mu}_{i} & = & N^{exp}_{had}\{R_{b\bar{b}}\cdot
275: 2Br^{\rm sum}(b \rightarrow \mu)F(A_{FB}^{b})_{i}\cdot
276: C^{b \rightarrow \mu}_{i} \nonumber\\
277: & + & R_{b\bar{b}}\cdot2Br(c \rightarrow \mu)
278: (1+\alpha)F(-\frac{1-\alpha}{1+\alpha}A_{FB}^{b})_{i}\cdot C^{b \rightarrow c
279: \rightarrow \mu}_{i} \nonumber\\
280: & + & R_{c\bar{c}}\cdot2Br(c \rightarrow
281: \mu)F(-A_{FB}^{c})_{i}\cdot C^{c \rightarrow \mu}_{i}\} \nonumber\\
282: & + & N^{others}_{i}
283: \end{eqnarray}
284: %
285: for $-Q\cos\theta_{thrust}$ bins, and
286: %
287: \begin{eqnarray}
288: \tilde{N}^{\mu}_{k} & = & N^{exp}_{had}\{R_{b\bar{b}}\cdot
289: 2Br^{\rm sum}(b \rightarrow \mu)\cdot W^{b \rightarrow \mu}_{k} \nonumber\\
290: & + & R_{b\bar{b}}\cdot2Br(c \rightarrow \mu)(1+\alpha)\cdot
291: W^{b \rightarrow c \rightarrow \mu}_{k} \nonumber\\
292: & + & R_{c\bar{c}}\cdot2Br(c \rightarrow \mu)\cdot
293: W^{c \rightarrow \mu}_{k}\} \nonumber\\
294: & + & N^{others}_{k}
295: \end{eqnarray}
296: %
297: for $p$ bins, where $N^{exp}_{had}$ is the total number of
298: hadronic events.
299: $Br^{\rm sum}(b \rightarrow \mu)$ is the combined branching ratio of
300: $Br(b \rightarrow \mu)$ and $Br(b \rightarrow \tau \rightarrow \mu)$.
301: $Br(b \rightarrow \mu)$ and $Br(c \rightarrow \mu)$
302: are the branching ratios of $b \rightarrow \mu$ and $c \rightarrow \mu$,
303: which are equivalent to the branching ratios for muons semileptonically decayed
304: from b and c hadrons, respectively.
305: $Br(b \rightarrow c\bar{c}s)$ is denoted as $\alpha$
306: in the above equations.
307: The used branching ratios are listed in Table \ref{tab:br}.
308: $F(A_{FB}^{q})_{i}$ is the polar-angle distribution function, integrated
309: in each $\cos \theta$ bin ($i$-th bin), which is written as
310: %
311: \begin{equation}
312: F(A_{FB}^{q})_{i} = \int_i \frac{3}{8}(1 + \cos^{2}\theta +
313: \frac{8}{3}A_{FB}^{q}\cos\theta) d\cos\theta.
314: \end{equation}
315: %
316: \begin{table}[h]
317: \caption{Branching ratios used in the fits.}
318: \label{tab:br}
319: \begin{center}
320: \begin{tabular}{|r|r|}
321: \hline \hline
322: Mode & Branching ratio(\%) \\
323: \hline
324: $Br^{\rm sum}(b \rightarrow \mu)$ & $11.3 \pm 0.5$ \\
325: $Br(b \rightarrow \mu)$\cite{PDG98} & $10.8 \pm 0.5$ \\
326: $Br(b \rightarrow \tau \rightarrow \mu)$\cite{b-tau} & $0.45$ \\
327: $Br(c \rightarrow \mu)$\cite{C-mu} & $9.0 \pm 0.7$ \\
328: $Br(b \rightarrow c\overline{c}s)$\cite{CCS} & $21.9 \pm 3.7$ \\
329: \hline \hline
330: \end{tabular}
331: \end{center}
332: \end{table}
333: %
334: $C_{i}^{mode}$ is the correction factor for the each $\cos \theta$ bin
335: from the {\it decay mode}, given by
336: %
337: \begin{eqnarray}
338: C^{mode}_{i} =
339: \frac{(1+\delta)^{q\bar{q}}_{i}}{(1+\delta)^{had}_{total}}\frac{\eta^{q\bar{q}}
340: _{i}}{\eta^{had}}
341: \varepsilon^{\mu}_{i},
342: \end{eqnarray}
343: %
344: where $(1+\delta)_{i}^{q\bar{q}}$ is the combined correction factor of the
345: QED (initial state photon radiation) and QCD (parton shower) radiative
346: corrections for the $i$-th bin of $e^{+}e^{-} \rightarrow q\bar{q}$.
347: Similarly, $(1+\delta)_{total}^{had}$ is a radiative correction
348: factor for the total hadronic cross section.
349: These correction factors were estimated through a Monte-Carlo
350: simulation based on the LUND event generator, JETSET7.3.
351: $\eta_{i}^{q\bar{q}}$ is the acceptance correction for the
352: $i$-th bin of each mode, $\eta^{had}$ is that for hadronic events,
353: and $\varepsilon^{\mu}_{i}$ is the muon identification efficiency
354: for the $i$-th bin ($\cos \theta$ bin). $W^{mode}_{k}$ is the correction
355: factor for each $p$ bin, which is given by
356: %
357: \begin{eqnarray}
358: W^{mode}_{k} =
359: \frac{(1+\delta)^{q\bar{q}}_{k}}{(1+\delta)^{had}_{total}}
360: \frac{\eta^{q\bar{q}}_{k,mode}}{\eta^{had}}\varepsilon^{\mu}_{k}.
361: \end{eqnarray}
362: %
363: Table \ref{tab:num-corr} gives the correction factors
364: for each bin. $N_{i}^{others}$ is the number of background events
365: for the $i$-th bin, which is also listed in Table \ref{tab:num-corr}.
366: The numbers were derived through Monte Carlo simulations.
367:
368: As a result of the four-parameter fit (Figure \ref{fig:flavor-fit} and
369: \ref{fig:cont}), we obtained
370: $R_{b\bar{b}} = 0.13 \pm 0.02$, $R_{c\bar{c}} = 0.36 \pm 0.05$,
371: $A_{FB}^{b} = -0.20 \pm 0.16$ and $A_{FB}^{c} = -0.17 \pm 0.14$,
372: with $\chi^{2}/\mbox{D.O.F} = 14.93/18$. The errors are statistical only.
373: The correlation coefficients obtained from the fit is given in Table
374: \ref{tab:coefficients}. The 1$\sigma$ contours of the fit for $R_{cc}$
375: v.s. $R_{bb}$ and $A_{c}$ v.s. $A_{b}$ are shown in Figure \ref{fig:cont}.
376:
377: \begin{table}[h]
378: \caption{Number of events and correction factors in each bin.}
379: \label{tab:num-corr}
380: \begin{center}
381: (a) \hspace{0.5cm}Correction factor for each $\cos \theta$ bin:
382: $p_{T}^{Jet} < 0.8$ (GeV/c)\\
383: \begin{tabular}{cccccc}
384: \hline \hline
385: $\cos\theta$ bin & $N^{\mu}_{i}$ & $N^{others}_{i}$
386: &{\rm$C^{b \rightarrow\mu}_{i}$}
387: &{\rm$C^{b \rightarrow b \rightarrow \mu}_{i}$}
388: &{\rm$C^{c \rightarrow \mu}_{i}$}\\
389: \hline
390: -0.8 $\sim$ -0.6 & 20.$\pm$ 4.5& 14.0$\pm$1.1& 0.049& 0.034& 0.038\\
391: -0.6 $\sim$ -0.4 & 133.$\pm$11.5& 65.0$\pm$2.4& 0.178& 0.180& 0.271\\
392: -0.4 $\sim$ -0.2 & 142.$\pm$11.9& 76.2$\pm$2.6& 0.222& 0.187& 0.319\\
393: -0.2 $\sim$ 0.0 & 152.$\pm$12.3& 76.9$\pm$2.6& 0.183& 0.137& 0.288\\
394: 0.0 $\sim$ 0.2 & 154.$\pm$12.4& 76.6$\pm$2.6& 0.234& 0.149& 0.273\\
395: 0.2 $\sim$ 0.4 & 132.$\pm$11.5& 76.5$\pm$2.6& 0.201& 0.176& 0.313\\
396: 0.4 $\sim$ 0.6 & 144.$\pm$12.0& 60.5$\pm$2.3& 0.209& 0.128& 0.253\\
397: 0.6 $\sim$ 0.8 & 24.$\pm$ 4.9& 9.8$\pm$0.9& 0.064& 0.029& 0.044\\
398: \hline \hline
399: \end{tabular}
400: \\
401: \vspace*{1cm}
402: (b) \hspace{0.5cm}Correction factor for each $\cos \theta$ bin:
403: $p_{T}^{Jet} \geq 0.8$ (GeV/c)\\
404: \begin{tabular}{cccccc}
405: \hline \hline
406: $\cos\theta$ bin & $N^{\mu}_{i}$ & $N^{others}_{i}$
407: &{\rm$C^{b \rightarrow\mu}_{i}$}
408: &{\rm$C^{b \rightarrow c \rightarrow \mu}_{i}$}
409: &{\rm$C^{c \rightarrow \mu}_{i}$} \\
410: \hline
411: -0.8 $\sim$ -0.6 & 22.$\pm$4.7& 4.8$\pm$0.7 & 0.106& 0.042& 0.014\\
412: -0.6 $\sim$ -0.4 & 62.$\pm$7.9& 22.0$\pm$1.4 & 0.326& 0.062& 0.046\\
413: -0.4 $\sim$ -0.2 & 80.$\pm$8.9& 26.3$\pm$1.5 & 0.449& 0.057& 0.070\\
414: -0.2 $\sim$ 0.0 & 64.$\pm$8.0& 24.8$\pm$1.5 & 0.476& 0.063& 0.062\\
415: 0.0 $\sim$ 0.2 & 64.$\pm$8.0& 26.3$\pm$1.5 & 0.443& 0.073& 0.050\\
416: 0.2 $\sim$ 0.4 & 66.$\pm$8.1& 25.2$\pm$1.5 & 0.411& 0.080& 0.060\\
417: 0.4 $\sim$ 0.6 & 54.$\pm$7.3& 17.4$\pm$1.2 & 0.334& 0.037& 0.040\\
418: 0.6 $\sim$ 0.8 & 15.$\pm$3.9& 5.7$\pm$0.7 & 0.125& 0.014& 0.008\\
419: \hline \hline
420: \end{tabular}
421: \\
422: \vspace*{1cm}
423: (c) \hspace{0.5cm}Correction factor for each momentum bin \\
424: \begin{tabular}{cccccc}
425: \hline \hline
426: Momentum bin (GeV/c) & $N^{\mu}_{k}$ & $N^{others}_{k}$
427: &{\rm $W^{b \rightarrow \mu}_{k}$}
428: &{\rm $W^{b \rightarrow c \rightarrow \mu}_{k}$}
429: &{\rm $W^{c \rightarrow \mu}_{k}$} \\
430: \hline
431: 2.5 $\sim$3.0 & 181.$\pm$ 13.5 & 95.0 $\pm$2.9 & 3.048 & 2.902 & 2.998\\
432: 3.0 $\sim$4.0 & 281.$\pm$ 16.7 & 116.9$\pm$3.2 & 4.921 & 3.764 & 4.903\\
433: 4.0 $\sim$6.0 & 325.$\pm$ 18.0 & 124.6$\pm$3.3 & 4.572 & 1.573 & 2.627\\
434: 6.0 $\sim$10.0& 309.$\pm$ 17.6 & 130.5$\pm$3.4 & 2.600 & 0.402 & 0.936\\
435: 10.0$\sim$16.0& 165.$\pm$ 12.8 & 98.2 $\pm$2.9 & 0.865 & 0.045 & 0.145\\
436: 16.0$\sim$30.0& 68. $\pm$ 8.2 & 43.2 $\pm$1.9 & 0.091 & 0.007 & 0.017\\
437: \hline \hline
438: \end{tabular}
439: \end{center}
440: \end{table}
441:
442: \begin{table}[h]
443: \caption{The correlation coefficients for the parameters in the
444: four-parameter fit.}
445: \label{tab:coefficients}
446: \begin{center}
447: \begin{tabular}{l|llll}
448: \hline \hline
449: & $R_{b\bar{b}}$ & $R_{c\bar{c}}$ &
450: $A_{FB}^{b}$ & $A_{FB}^{c}$ \\
451: \hline
452: $R_{b\bar{b}}$ & 1 & -0.77 & 0.01 & -0.17 \\
453: $R_{c\bar{c}}$ & & 1 & -0.37 & 0.16 \\
454: $A_{FB}^{b}$ & & & 1 & 0.44 \\
455: $A_{FB}^{c}$ & & & & 1 \\
456: \hline
457: \hline
458: \end{tabular}
459: \end{center}
460: \end{table}
461:
462: \begin{figure}
463: \input epsf
464: \begin{center}
465: \leavevmode
466: \epsfxsize=15.0cm
467: \epsfbox{fig8.eps}
468: \caption{Results of the four-parameter fit:
469: $\cos\theta_{thrust}$ for the c-enriched sample($p_{T}^{jet}<0.8$ GeV/c) (a)
470: and the b-enriched sample($p_{T}^{jet}\ge0.8$ GeV/c) (b), and
471: momentum distributions (c). The dots with error bars are the data.
472: The histograms are the results from the four-parameter fit:
473: $b \rightarrow \mu$(open), $b \rightarrow c \rightarrow \mu$(hatched),
474: $c \rightarrow \mu$(double-hatched) and
475: $others$(triple-hatched), respectively.}
476: \label{fig:flavor-fit}
477: \end{center}
478: \end{figure}
479:
480: \begin{figure}
481: \input epsf
482: \begin{center}
483: \leavevmode
484: %\epsfxsize=15.0cm
485: \epsfbox{fig9.eps}
486: \caption{Results of the four-parameter fit for $R_{q\bar{q}}$ (a)
487: and $A_{FB}^{q}$ (b).
488: The cross bar indicates the central value and the ellipse shows 1$\sigma$
489: contour. The errors are only statistical. The points show the
490: standard-model prediction.}
491: \label{fig:cont}
492: \end{center}
493: \end{figure}
494:
495: \subsection{Systematic errors for $R_{q\bar{q}}$ and $A^{q}_{FB}$}
496:
497: \hspace*{12pt}Possible sources of systematic errors are listed in
498: Table \ref{tab:systematic}. The largest systematic error comes from
499: the uncertainty in the probabilities of mis-identified hadrons as muons.
500: This is caused by the light quark background in the heavy
501: quark samples. In order to estimate the effect of hadron mis-identifications,
502: we performed the same analysis by changing the correction factor of the
503: effective pion-nucleus cross section by $\pm 1 \sigma$ of the measured error
504: of fake rate. The effect of the cuts on
505: muon identification was studied by varying the cut values
506: on momentum $p$, $\Delta d$, and $\Delta d/\sigma_{track}$ by
507: $\pm 0.5$ GeV/c, $\pm 5$ cm, and $\pm 0.5 \sigma_{track}$, respectively,
508: and examined the change in the obtained cross sections and asymmetries.
509: We studied the effect of the $p_{T}^{jet}$ cut by shifting the
510: cut value by $\pm 0.1$ GeV/c. The effect of the uncertainly in the MDC
511: detector acceptance was checked by varying the $|\cos\theta_{track}|$ cut
512: from 0.6 to 0.58. Based on these studies, we selected the cut values that
513: minimize the systematic errors. The error due to the semileptonic branching
514: ratios for b and c hadrons were obtained by changing the branching
515: ratios by $\pm 1 \sigma$ of the quoted number.
516: Changes of fragmentation parameters cause the changes in the muon
517: identification efficiency and flavor separation. The error due to this effect
518: was estimated by changing the fragmentation parameters by
519: $\pm 1 \sigma$ of the quoted number.
520: \begin{table}[h]
521: \caption{Summary of systematic errors}
522: \label{tab:systematic}
523: \begin{center}
524: \begin{tabular}{lllll}
525: \hline \hline
526: Source & $\Delta R_{b\bar{b}}/R_{b\bar{b}}$ &
527: $\Delta R_{c\bar{c}}/R_{c\bar{c}}$ &
528: $\Delta A_{FB}^{b}/A_{FB}^{b}$ &
529: $\Delta A_{FB}^{c}/A_{FB}^{c}$ \\
530: \hline
531: hadron mis-ID & 4.9\%& 12.1\%& 3.8\%& 13.3\%\\
532: muon ID & $<$0.1\%& 0.6\%& 1.5\%& 0.9\%\\
533: $p^{jet}_{T}$ cut & 0.9\%& 0.6\%& 1.5\%& 0.2\%\\
534: MDC acceptance & $<$0.1\%& $<$0.1\%& 0.5\%& 0.9\%\\
535: branching ratio & 2.8\%& 5.0\%& 2.1\%& 1.2\%\\
536: fragmentation parameter & 1.7\%& 3.8\%& $<$0.1\%& $<$0.1\%\\
537: \hline
538: total & 6.1\%& 13.7\%& 4.9\%& 13.8\%\\
539: \hline \hline
540: \end{tabular}
541: \end{center}
542: \end{table}
543:
544: