0810.4084/cdc.tex
1: 
2:  \section{CDC performance}
3:  \label{sec:cdc}
4: 
5: \newcommand \ie{{\it i.e.}}
6: 
7: % What we can do for CDC hardware against high background is very
8: % limited. Since we can not find any better gas than the current one
9: % (Helium-ethane 50\%-50\%), the remaining thing is to have smaller cells to
10: % reduce the hit rate and also to reduce the maximum drift time.
11: % The cell size cannot be smaller than we can manipulate; the effective
12: % background reduction factor due to smallness of cell size is only about
13: % 2/3. Therefore, 20 times more background will be 13 times, effectively;
14: % the rest should be covered by software efforts.
15: % We study background effects on tracking in the following subsections.
16: 
17:  There are three requirements to be able to operate the new CDC with
18:  higher beam background.
19:  First, the dead time for the readout electronics must be reduced.
20:  At present, we are using charge to time conversion chips
21:  and multi hit TDCs for measurements of the drift time and the energy
22:  loss.  The conversion takes some time (800 -- 2200\,nsec) and the drift time
23:  cannot be measured during that time. This contributes to the dead time. The new
24:  system records the timing with a pipelined TDC and the charge with an FADC
25:  separately. The dead time for the new system depends on the shaping time for
26:  the new ASIC chip and will be around 200\,nsec. This is significantly
27:  better than in the present system. Second, the cell size must be reduced.
28:  This means that the number of cells in one layer increases and the maximum
29:  drift time decreases. The number of cells for the first super layer
30:  (where the most serious background is expected) increases from 64 to 160 (A
31:  small cell chamber was installed instead of the cathode chamber in
32:  2003. It has 128 cells for just two layers in order to test the small
33:  cell principle and has been working well for 5 years.) Finally, we will
34:  increase the number of stereo layers in each super layer to obtain higher
35:  efficiency for three dimensional tracking. A detailed wire configuration is
36:  described in the LoI~\cite{bib:LoI}.  We studied
37:  the effects of background on tracking performance including these
38:  improvements included in the upgrade; the results are described in the
39:  subsections that follow.
40: 
41: %\subsection{Track reconstruction efficiency}
42: \subsection{Tracking efficiency and mass resolution}
43: We approximately estimate the tracking performance in sBelle using the 
44: Geant3-based detector simulator for Belle. The same CDC and SVD as used in Belle 
45: are assumed except for the dead time of the readout electronics for the CDC, 
46: which is set to 200\,nsec.
47: The high beam background environment expected at sBelle
48: is simulated by superimposing real random 
49: triggered events (taken with the Belle detector) on each simulated 
50: signal event.
51: 
52: We assume three beam background levels in the simulation study:
53: \begin{itemize}
54: \item nominal ($1 \times$ bkg): 1 times bkg both in the CDC and SVD,
55: where bkg is the nominal level of beam background observed in the Belle 
56: detector;
57: \item moderately-high ($5 \times$ bkg): 
58: 3 times bkg in the CDC and 1 times bkg in the SVD; 
59: \item highest ($20 \times$ bkd): 
60: 13 times bkg in the CDC and 1 times bkg in the SVD.
61: \end{itemize}
62: \noindent
63: Here note that 
64: the background level in the highest (moderately-high) case is 20 
65: (5) times nominal, which approximately corresponds to 13 (3) times bkg in the 
66: sBelle CDC, whose cell sizes are smaller than those of the Belle CDC.
67: The beam background in the SVD is assumed to be nominal in all cases, because 
68: readout electronics with much shorter ($\sim 1/16$) 
69: shaping time than in the Belle SVD will be used for the sBelle SVD.
70: 
71: The track-finder in the CDC used in the study is one that has been
72: recently updated. It has higher tracking efficiency than that used to
73: process data so far in Belle, especially for events with high charged multiplicity. 
74: In addition, the stand-alone track-finder for the SVD, which 
75: became available recently is used. This finder aims at efficient
76: tracking of low momentum particles such as the slow pion from $D^\ast\to
77: D^0\pi$ decay.
78:  
79: The physics channels studied are,
80: $ B^0~\to~J/\psi K^0_S$ ($J/\psi~\to~\mu^+\mu^-;~K^0_S~\to~\pi^+\pi^-$),
81: which is an example of a typical channel, and $ B^0 \to D^{*+} D^{*-}$ 
82: ($~D^{*\pm} \to D^0\pi^\pm_s;~D^0 \to K3\pi$), 
83: which is an example of the most difficult channels in terms of tracking.
84: 
85: \subsubsection{Result for $B^0 \to J/\psi K^0_S$ }
86: Summarized in Table~\ref{cdc_eff1} are the reconstruction efficiencies and 
87: relative efficiencies for the three background cases.~\footnote{The event selection is performed with the following criteria: 
88: $-60~<~(M(\mu^+\mu^-)-M_{J/\psi})~<~36~{\rm{MeV}}/c^2$, 
89: $ |M(\pi^+\pi^-)~-~M_{K^0_S}|~<~16~{\rm{MeV}}/c^2 $,
90: $ 5.270~<~M_{\rm{bc}}~<~5.290~{\rm{GeV}}/c^2$ and $|\Delta E|~<~40$~MeV. 
91: These are somewhat simpler criteria than those applied in the data analysis in 
92: Belle~\cite{bib:JpsiKs_analysis}.}
93: %Here the event selection is performed with the following criteria: 
94: %$-60~<~(M(\mu^+\mu^-)-M_{J/\psi})~<~36~{\rm{MeV}}/c^2$, 
95: %$ |M(\pi^+\pi^-)~-~M_{K^0_S}|~<~16~{\rm{MeV}}/c^2 $,
96: %$ 5.270~<~M_{\rm{bc}}~<~5.290~{\rm{GeV}}/c^2$ and $|\Delta E|~<~40$~MeV. 
97: %These are simpler criteria than those applied in the data analysis in 
98: %Belle~\cite{bib:jpsiks_analysis}.
99: 
100: \begin{table}[hbtp]
101: \caption{\label{cdc_eff1}
102: Reconstruction efficiencies and relative efficiencies
103: for $B^0\to J/\psi K^0_S$ for the three background cases.
104: The efficiency in Belle is also shown in the last row for 
105: comparison.}
106: \begin{center}
107: \begin{tabular}{|c|c|c|c|} \hline
108: bkg level  & eff. (\%)   & eff. ratio - 1 & eff. ratio - 1 \\ 
109:                &             & wrt $1\times$ bkg (\%) 
110:                              & wrt Belle (\%)                \\ \hline\hline
111: $ 1\times$ bkg &  58.7       &  $\equiv 0$   & $+11.3$       \\ \hline
112: $ 5\times$ bkg &  57.7       &    $-1.7$     & $+~9.4$       \\ \hline
113: $20\times$ bkg &  53.6       &    $-8.8$     & $+~1.5$       \\ \hline\hline
114: $ 1\times$ bkg (Belle)&  52.7       &    -          & $\equiv 0$ \\ \hline
115: \end{tabular}
116: \end{center}
117: \end{table}
118: 
119: As shown in column 3 of the table, 
120: the reconstruction efficiency in the highest (moderately-high) background case
121: drops only by $\sim 9~(2)\%$ with respect to the nominal background case.
122: Furthermore as shown in column 4, 
123: nearly the same efficiency as in Belle can be maintained even 
124: in the highest background case, thanks to the hardware (reduced dead time and 
125: smaller cell size) and software improvements.
126: Note that the efficiencies in the table are those derived from the CDC 
127: track-finder, with the SVD stand-alone track-finder turned off. 
128: The difference in efficiency between the 
129: %SVD track-finder off and on is very small, {$\sim\it O(0.1\%)$} level, 
130: SVD track-finder off and on is very small, $\sim 0.1 - 0.2\%$ level, 
131: because the final state does not include slow particles.
132: 
133: Summarized in Table~\ref{cdc_resol1} are the mass and $\Delta E$ resolutions.
134: %The degradation of the resolutions is not large, only a $\sim 4(J/\psi)
135: % - $9(K^0_S)$\% level with respect to $1 \times bkg$ case.
136: The degradations of the resolutions due to background are not large, 
137: at the few \% level.
138: 
139: \begin{table}[hbtp]
140: \caption{\label{cdc_resol1}
141: Mass and $\Delta E$ resolutions for $B^0\to J/\psi K^0_S$ for the three
142: background cases. Here the resolution is simply defined as $\sigma$ of a 
143: single-Gaussian fit to the distribution in question.
144: The resolutions in Belle are also shown in the last row for comparison.}
145: \begin{center}
146: \begin{tabular}{|c|c|c|c|} \hline
147: bkg level  & $M(\mu^+\mu^-)$ & $M(\pi^+\pi^-)$  & $\Delta E$  \\ 
148:                & (MeV/$c^2$)     & (MeV/$c^2$)  & (MeV)       \\ \hline\hline
149: %$ 1\times$ bkg &  8.87            &    2.21       &  7.24     \\ \hline
150: %$ 5\times$ bkg &  8.86            &    2.22       &  7.36     \\ \hline
151: %$20\times$ bkg &  9.10            &    2.29       &  7.47     \\ \hline\hline
152: %  Belle        &  8.76            &    2.10       &  7.14     \\ \hline
153: $ 1\times$ bkg &  8.9            &    2.2       &  7.2     \\ \hline
154: $ 5\times$ bkg &  8.9            &    2.2       &  7.4     \\ \hline
155: $20\times$ bkg &  9.1            &    2.3       &  7.5     \\ \hline\hline
156: $ 1\times$ bkg (Belle)&  8.8            &    2.1       &  7.1     \\ \hline 
157: \end{tabular}
158: \end{center}
159: \end{table}
160: 
161: 
162: \subsubsection{Result for $B^0 \to D^{*+}D^{*-} $}
163: The final state includes ten charged particles, 
164: two of which are low momentum pions from 
165: $D^\ast$'s, hence this is regarded as one of the most difficult channels 
166: in term of charged particle tracking.
167: 
168: Summarized in Table~\ref{cdc_eff2} are the reconstruction efficiencies and 
169: relative efficiencies for the three background cases.~\footnote{
170: The event selection is performed with the following criteria: 
171: $ |M(K3\pi) - M_{D^0}|~<~16~{\rm{MeV}}/c^2 $,
172: $ |\delta M - \delta M_{nominal}| < 3~{\rm{MeV}}/c^2 $ 
173: ($\delta M \equiv  M(K3\pi\pi_s) - M(K3\pi)$),
174: $ |M_{\rm{bc}} - M_{B^0}| < 10.5~{\rm{MeV}}/c^2$ and 
175: $|\Delta E|~<~40$~MeV. 
176: These are somewhat simpler criteria than those applied in the Belle data 
177: analysis~\cite{bib:D*D*_analysis}.}
178: 
179: \begin{table}[hbtp]
180: \caption{\label{cdc_eff2}
181: Same as Table~\ref{cdc_eff1} but for $B^0\to D^{*+}D^{*-}$.}
182: \begin{center}
183: \begin{tabular}{|c|c|c|c|} \hline
184: bkg level  & eff. (\%)   & eff. ratio - 1 & eff. ratio - 1 \\ 
185:                &             & wrt $1\times$ bkg (\%) 
186:                              & wrt Belle (\%)            \\ \hline\hline
187: $ 1\times$ bkg &  7.3       &  $\equiv 0$ & $+119$       \\ \hline
188: $ 5\times$ bkg &  6.7       &    $ -9$    & $+~99$       \\ \hline
189: $20\times$ bkg &  4.4       &    $-41$    & $+~30$       \\ \hline\hline
190: $ 1\times$ bkg (Belle)    &  3.3       &    -        & $\equiv 0$   \\ \hline
191: \end{tabular}
192: \end{center}
193: \end{table}
194: 
195: As shown in column 4 of the table, the reconstruction efficiency for 
196: the nominal (and 5 times) background case is  
197: approximately twice that of Belle, thanks to the hardware and software 
198: improvements. The efficiency in the highest 
199: background case, however, significantly drops (column 3) in contrast to the
200: $J/\psi K^0_S$ channel. This is because many (2.5 times more) 
201: low momentum charged particle tracks must be reconstructed 
202: in this channel compared to $J/\psi K^0_S$. 
203: Nevertheless, an efficiency higher than in Belle can be maintained (column 4).
204: 
205: 
206: \begin{figure}[h]
207: \begin{center}
208: %\includegraphics[width=0.7\linewidth]{epsfiles/CDC_performance/dstdst_eff.eps}
209: \includegraphics[width=0.56\linewidth]{epsfiles/CDC_performance/dstdst_eff.eps}
210: \caption{\label{cdc_eff2fig}
211: Reconstruction efficiencies for $B^0\to D^{*+}D^{*-}$ with the SVD 
212: stand-alone track-finder on and off.}
213: \end{center}
214: \end{figure}
215: \noindent 
216: Shown in Fig.~\ref{cdc_eff2fig} are the reconstruction efficiencies with the 
217: SVD stand-alone track-finder on,  
218: along with the efficiencies with the SVD finder off (the same 
219: efficiencies as in Table~\ref{cdc_eff2}). 
220: As seen in the figure, 
221: the SVD finder improves the efficiency significantly, because this finder 
222: is very efficient for the low momentum pions from $D^\ast$'s.
223: 
224: Summarized in Table~\ref{cdc_resol2} are the mass and $\Delta E$ resolutions.
225: No degradation is found (within fitting errors) 
226: either in the $D$ mass or in the $D^\ast - D$ mass 
227: difference resolution, whereas there is a $\sim 10\%$ degradation
228: in the $\Delta E$ resolution. 
229: 
230: \begin{table}[hbtp]
231:  \caption{\label{cdc_resol2}
232:  Same as Table~\ref{cdc_resol1} but for $B^0\to D^{*+}D^{*-}$. 
233:  These are results obtained without the SVD track-finder.
234: $\delta M \equiv  M(K3\pi\pi_s) - M(K3\pi)$.
235: }
236:  \begin{center}
237:   \begin{tabular}{|c|c|c|c|} \hline
238:    bkg level  & $M(K3\pi)$          & $\delta M $   & $\Delta E$  \\ 
239:    & (MeV/$c^2$)     & (MeV/$c^2$)  & (MeV)    \\ \hline\hline
240:    $ 1\times$ bkg &  5.1            &    0.72       &  7.2     \\ \hline
241:    $ 5\times$ bkg &  5.0            &    0.72       &  7.5     \\ \hline
242:    $20\times$ bkg &  5.1            &    0.70       &  7.9     \\ \hline\hline
243:    $ 1\times$ bkg (Belle) &  4.9            &    0.73       &  7.1 \\ \hline 
244:   \end{tabular}
245:  \end{center}
246: \end{table}
247: 
248: 
249: \subsection{Summary of CDC performance}
250: The CDC will be upgraded to handle much higher beam background.
251: Quick simulation studies including the effects of the upgraded components, such as the
252: smaller cell size, optimized time windows, as well as the new tracking
253: algorithm and the SVD standalone tracker, have been carried out.
254: We found that the new CDC will maintain a track reconstruction
255: performance that is similar to the present CDC.
256: