1: \documentclass{elsart}
2: \usepackage{graphicx}
3: \usepackage{cite}
4: %%\usepackage[colorlinks]{hyperref}
5:
6: %
7: %==============================================================================
8: % Definitions
9: %==============================================================================
10:
11: \def\ifm#1{\relax\ifmmode#1\else$#1$\fi}
12: \def\x{\ifm{\times}}
13: \def\BR{\hbox{BR}} \def\prl{Phys. Rev. Lett}
14: \def\up#1{$^{#1}$} \def\dn#1{$_{#1}$}
15: \def\DAF{DA\char8NE} \def\sig{\ifm{\sigma}}
16: \def\f{\ifm{\phi}} \def\ff{\f--factory}
17: \def\ab{\ifm{\sim}} \def\x{\ifm{\times}}
18: \def\gam{\ifm{\gamma}} \def\pic{\ifm{\pi^+\pi^-}}
19: \def\pt#1;#2;{\ifm{#1\x10^{#2}}} \def\epm{\ifm{e^+e^-}}
20: \def\to{\ifm{\rightarrow}} \def\sig{\ifm{\sigma}} \def\plm{\ifm{\pm}}
21: \def\ks{\ifm{K_S}} \def\kl{\ifm{K_L}}
22: \def\po{\ifm{\pi^0}} \def\pio{\ifm{\pi^0\pi^0}}
23: \def\sig{\ifm{\sigma}} \def\plm{\ifm{\pm}} \def\kpm{\ifm{K^\pm}}
24: \def\kpkm{\ifm{K^+K^-}} \def\L{\ifm{{\mathcal L}}}
25: \def\C{\ifm{C}} \def\P{\ifm{P}} \def\T{\ifm{T}}
26: \def\dif{{\rm d\hspace{.3mm}}}
27: %\def\ctwoppmas{\ifm{\sum_{i=1}^2\left(\frac{\Delta m_i^2}{\sigma_{\pi_i}^2}\right)
28: %+\frac{(E_{K_S}-\sum_{i}E_{\gamma_i})^2}{\sigma_E^2}+\frac{(P_x^{K_S}-\sum_iP_x^
29: % {\gamma_i})^2}{\sigma_{P_x}^2}}}
30: \def\ctwoppmas{\ifm{\frac{\Delta m_1^2}{\sigma_{m}^2}+\frac{\Delta m_2^2}{\sigma_{m}^2}
31: +\frac{(E_{K_S}-\sum_{i}E_{\gamma_i})^2}{\sigma_E^2}+\frac{(P_x^{K_S}-\sum_iP_x^
32: {\gamma_i})^2}{\sigma_{P_x}^2}}}
33: \def\ctwoppimp{\ifm{\frac{(P_y^{K_S}-\sum_iP_y^
34: {\gamma_i})^2}{\sigma_{P_y}^2}+\frac{(P_z^{K_S}-\sum_iP_z^
35: {\gamma_i})^2}{\sigma_{P_z}^2}+\frac{(\pi-\vartheta_{\pi\pi}^{*})^2}{\sigma_{\vartheta_{\pi\pi}^{*}}^2}}}
36: \def\ctwoppcos{\ifm{\frac{(cos\pi-cos\vartheta_{\pi\pi}^{*})^2}{\sigma_{cos\vartheta_{\pi\pi}^{*}}^2}}}
37: \def\xc{10\up2\x} \def\xm{10\up3\x} \def\xcc{10\up4\x}
38:
39: %%%%%%%%%%%%%%%
40: \makeatletter
41: %%%%This defines the \eqalign PLAIN macro, which
42: %%%%Latex, replaced with an absurd ones.
43: \newdimen\z@ \z@=0pt % can be used both for 0pt and 0
44: \newskip\z@skip \z@skip=0pt plus0pt minus0pt
45: \def\m@th{\mathsurround=\z@}
46: \def\ialign{\everycr{}\tabskip\z@skip\halign} % initialized \halign
47: \def\eqalign#1{\null\,\vcenter{\openup\jot\m@th
48: \ialign{\strut\hfil$\displaystyle{##}$&$\displaystyle{{}##}$\hfil
49: \crcr#1\crcr}}\,}
50:
51: \def\section{\@startsection {section}{1}{\z@}%
52: {-1.5ex plus -1ex minus -.2ex}%
53: {1.3ex plus .2ex}{\normalsize\bf}}
54: \def\subsection{\@startsection{subsection}{2}{\z@}%
55: {-.8ex plus -.6ex minus -.2ex}
56: {.5ex plus .2ex}{\normalsize}}
57: \parskip .4pc \@plus .1\p@ % Extra vertical space between paragraphs.
58: \makeatother
59: %%%%%%%%%%%%%%%
60:
61: \newcommand{\aff}[2]
62: {Dipartimento di Fisica dell'Universit\`a #1 e Sezione INFN, #2, Italy.}
63: \newcommand{\affd}[1]
64: {Dipartimento di Fisica dell'Universit\`a e Sezione INFN, #1, Italy.}
65: %
66: % Particles
67: %-----------
68: \newcommand{\ep}{\mbox{$e^{+}$}}
69: \newcommand{\el}{\mbox{$e^{-}$}}
70: %%\newcommand{\pio}{\mbox{$\pi^{0}$}}
71: \newcommand{\pim}{\mbox{$\pi^{-}$}}
72: \newcommand{\pip}{\mbox{$\pi^{+}$}}
73: \newcommand{\mup}{\mbox{$\mu^{+}$}}
74: \newcommand{\mum}{\mbox{$\mu^{-}$}}
75: %\newcommand{\ks}{\mbox{$K_{S}$}}
76: %\newcommand{\kl}{\mbox{$K_{L}$}}
77: \newcommand{\kp}{\mbox{$K^{+}$}}
78: \newcommand{\km}{\mbox{$K^{-}$}}
79: %\newcommand{\ko}{\mbox{$K^{o}$}}
80: \newcommand{\ok}{\mbox{$\overline{K^o}$}}
81: %\newcommand{\ao}{\mbox{$a_{o}$}}
82: %
83: %\newcommand{\fo}{\mbox{$f\!_{o}$}}
84: %
85: % CP related stuff
86: %
87: \newcommand{\cpv}{\mbox{$C\!P$}}
88: \newcommand{\eep}{\mbox{$\epsilon'\!/\epsilon$}}
89: \newcommand{\reep}{\mbox{$\Re(\epsilon'\!/\epsilon)$}}
90: \newcommand{\ieep}{\mbox{$\Im(\epsilon'\!/\epsilon)$}}
91: %\newcommand{\3pio}{\mbox{$K_{S} \rightarrow 3 \pi^{0}$}}
92: %
93: % Useful for radiative phi decays
94: %
95: \newcommand{\subfo}{\mbox{$_{f\!_{o}}$}}
96: \newcommand{\phiaog}{\mbox{$\phi\rightarrow a_{o}\gamma$}}
97: \newcommand{\phifog}{\mbox{$\phi\rightarrow f\!_{o}\gamma$}}
98: \newcommand{\qqbar}{\mbox{$q\overline{q}$}}
99: \newcommand{\uubar}{\mbox{$u\overline{u}$}}
100: \newcommand{\ddbar}{\mbox{$d\overline{d}$}}
101: \newcommand{\ssbar}{\mbox{$s\overline{s}$}}
102: \newcommand{\kkbar}{\mbox{$K\overline{K}$}}
103: \newcommand{\qqqq}{\mbox{$qq\overline{qq}$}}
104: %
105: % Others...
106: %
107: \newcommand{\br}{\mbox{${\rm BR}$}}
108: %\newcommand{\etal}{{\it et al.}}
109: \newcommand{\dafne}{\mbox{DA$\Phi$NE}}
110: %
111: % Italics for italian text
112: %
113: \newcommand{\quark}{{\it quark}}
114: \newcommand{\barr}{{\it barrel}}
115: \newcommand{\ecap}{{\it endcap}}
116: \newcommand{\cluster}{{\it cluster}}
117: %
118: %
119: %
120: %==============================================================================
121: %
122: \newcommand{\ttbs}{\char'134}
123: \newcommand{\AmS}{{\protect\the\textfont2
124: A\kern-.1667em\lower.5ex\hbox{M}\kern-.125emS}}
125:
126: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
127: \begin{document}
128: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
129:
130: %==============================================================================
131: \begin{frontmatter}
132: %==============================================================================
133:
134: %\title{\boldmath A direct search for the decay $K_s\longrightarrow 3\pi^0$ with
135: %the KLOE detector at \DAF}%%$\Phi$NE}
136: \title{\boldmath A direct search for the \C\P-violating decay $K_S\,$~\to~$\,3\pi^0$
137: with the KLOE detector at \DAF}%%$\Phi$NE}
138:
139: %-----------------------------------------------------------------------------
140: % Authors & Institutions
141: %------------------------------------------------------------------------------
142: %
143: \collab{The KLOE Collaboration}
144: %
145: \author[Na]{F.~Ambrosino},
146: \author[Frascati]{A.~Antonelli},
147: \author[Frascati]{M.~Antonelli},
148: \author[Roma3]{C.~Bacci},
149: %\author[Roma3]{M.~Barva},
150: \author[Frascati]{P.~Beltrame},
151: \author[Frascati]{G.~Bencivenni},
152: \author[Frascati]{S.~Bertolucci},
153: \author[Roma1]{C.~Bini},
154: \author[Frascati]{C.~Bloise},
155: \author[Roma1]{V.~Bocci},
156: \author[Frascati]{F.~Bossi},
157: \author[Frascati,Virginia]{D.~Bowring},
158: \author[Roma3]{P.~Branchini},
159: %\author[Moscow]{S.~A.~Bulychjov},
160: \author[Roma1]{R.~Caloi},
161: \author[Frascati]{P.~Campana},
162: \author[Frascati]{G.~Capon},
163: \author[Na]{T.~Capussella},
164: %\author[Roma2]{G.~Carboni},
165: \author[Roma3]{F.~Ceradini},
166: %\author[Pisa]{F.~Cervelli},
167: \author[Frascati]{S.~Chi},
168: \author[Na]{G.~Chiefari},
169: \author[Frascati]{P.~Ciambrone},
170: \author[Virginia]{S.~Conetti},
171: \author[Frascati]{E.~De~Lucia},
172: \author[Roma1]{A.~De~Santis},
173: \author[Frascati]{P.~De~Simone},
174: \author[Roma1]{G.~De~Zorzi},
175: \author[Frascati]{S.~Dell'Agnello},
176: \author[Karlsruhe]{A.~Denig},
177: \author[Roma1]{A.~Di~Domenico},
178: \author[Na]{C.~Di~Donato},
179: \author[Pisa]{S.~Di~Falco},
180: \author[Roma3]{B.~Di~Micco},
181: \author[Na]{A.~Doria},
182: \author[Frascati]{M.~Dreucci},
183: %\author[Roma3]{A.~Farilla},
184: \author[Frascati]{G.~Felici},
185: \author[Karlsruhe]{A.~Ferrari},
186: \author[Frascati]{M.~L.~Ferrer},
187: \author[Frascati]{G.~Finocchiaro},
188: \author[Frascati]{C.~Forti},
189: \author[Roma1]{P.~Franzini},
190: \author[Roma1]{C.~Gatti},
191: \author[Roma1]{P.~Gauzzi},
192: \author[Frascati]{S.~Giovannella},
193: \author[Lecce]{E.~Gorini},
194: \author[Roma3]{E.~Graziani},
195: \author[Pisa]{M.~Incagli},
196: \author[Karlsruhe]{W.~Kluge},
197: \author[Moscow]{V.~Kulikov},
198: \author[Roma1]{F.~Lacava},
199: \author[Frascati]{G.~Lanfranchi},
200: \author[Frascati,StonyBrook]{J.~Lee-Franzini},
201: \author[Karlsruhe]{D.~Leone},
202: %\author[Frascati,Moscow]{M.~Martemianov},
203: \author[Frascati]{M.~Martini\corauthref{cor1}},
204: % \footnote{Corresponding author: Matteo Martini
205: %INFN - LNF, Casella postale 13, 00044 Frascati (Roma),
206: %Italy; tel. +39-06-94032696, e-mail matteo.martini@lnf.infn.it},
207: \author[Na]{P.~Massarotti},
208: %\author[Frascati,Moscow]{M.~Matsyuk},
209: \author[Frascati]{W.~Mei},
210: \author[Na]{S.~Meola},
211: %\author[Roma2]{R.~Messi},
212: \author[Frascati]{S.~Miscetti\corauthref{cor2}},
213: % \footnote{Corresponding author: Stefano Miscetti
214: %INFN - LNF, Casella postale 13, 00044 Frascati (Roma),
215: %Italy; tel. +39-06-94032771, e-mail stefano.miscetti@lnf.infn.it},
216: \author[Frascati]{M.~Moulson},
217: \author[Karlsruhe]{S.~M\"uller},
218: \author[Frascati]{F.~Murtas},
219: \author[Na]{M.~Napolitano},
220: \author[Roma3]{F.~Nguyen},
221: \author[Frascati]{M.~Palutan},
222: \author[Roma1]{E.~Pasqualucci},
223: %\author[Frascati]{L.~Passalacqua},
224: \author[Roma3]{A.~Passeri},
225: \author[Frascati,Energ]{V.~Patera},
226: \author[Na]{F.~Perfetto},
227: \author[Roma1]{L.~Pontecorvo},
228: \author[Lecce]{M.~Primavera},
229: \author[Frascati]{P.~Santangelo},
230: \author[Roma2]{E.~Santovetti},
231: \author[Na]{G.~Saracino},
232: %\author[StonyBrook]{R.~D.~Schamberger},
233: \author[Frascati]{B.~Sciascia},
234: \author[Frascati,Energ]{A.~Sciubba},
235: \author[Pisa]{F.~Scuri},
236: \author[Frascati]{I.~Sfiligoi},
237: %\author[Frascati,Novos]{A.~Sibidanov},
238: \author[Frascati]{T.~Spadaro},
239: %\author[Roma3]{E.~Spiriti},
240: %\author[Frascati,Georgia]{M.~Tabidze},
241: \author[Roma1]{M.~Testa},
242: \author[Roma3]{L.~Tortora},
243: \author[Roma1]{P.~Valente},
244: \author[Karlsruhe]{B.~Valeriani},
245: \author[Frascati]{G.~Venanzoni},
246: \author[Roma1]{S.~Veneziano},
247: \author[Lecce]{A.~Ventura},
248: \author[Roma3]{R.~Versaci},
249: %\author[Na]{I.~Villella},
250: \author[Frascati,Beijing]{G.~Xu}
251: %
252: \address[Beijing]{Permanent address: Institute of High Energy
253: Physics of Academica Sinica, Beijing, China.}
254: \address[Frascati]{Laboratori Nazionali di Frascati dell'INFN,
255: Frascati, Italy.}
256: \address[Karlsruhe]{Institut f\"ur Experimentelle Kernphysik,
257: Universit\"at Karlsruhe, Germany.}
258: \address[Lecce]{\affd{Lecce}}
259: \address[Moscow]{Permanent address: Institute for Theoretical
260: and Experimental Physics, Moscow, Russia.}
261: \address[Na]{Dipartimento di Scienze Fisiche dell'Universit\`a
262: ``Federico II'' e Sezione INFN,
263: Napoli, Italy}
264: %\address[Novos]{Permanent address: Budker Institute of Nuclear Physics,
265: %Novosibirsk, Russia.}
266: \address[Pisa]{\affd{Pisa}}
267: \address[Energ]{Dipartimento di Energetica dell'Universit\`a
268: ``La Sapienza'', Roma, Italy.}
269: \address[Roma1]{\aff{``La Sapienza''}{Roma}}
270: \address[Roma2]{\aff{``Tor Vergata''}{Roma}}
271: \address[Roma3]{\aff{``Roma Tre''}{Roma}}
272: \address[StonyBrook]{Physics Department, State University of New
273: York at Stony Brook, USA.}
274: %\address[Georgia]{Permanent address: High Energy Physics Institute,
275: %Tbilisi State University, Tbilisi, Georgia}
276: \address[Virginia]{Physics Department, University of Virginia, USA.}
277: %
278: \corauth[cor1]{cor1}{\small $^1$ Corresponding author: Matteo Martini
279: INFN - LNF, Casella postale 13, 00044 Frascati (Roma),
280: Italy; tel. +39-06-94032896, e-mail matteo.martini@lnf.infn.it}
281:
282: \corauth[cor2]{cor2}{\small $^2$ Corresponding author: Stefano Miscetti
283: INFN - LNF, Casella postale 13, 00044 Frascati (Roma),
284: Italy; tel. +39-06-94032771, e-mail stefano.miscetti@lnf.infn.it}
285: %
286: %==============================================================================
287: \begin{abstract}
288: %==============================================================================
289: We have searched for the decay $K_S \to 3 \pi^{0}$ with the
290: KLOE experiment at DA\char8NE using data from $e^+ e^-$
291: collisions at a center of mass energy
292: $W\sim m_{\phi}$ for an integrated luminosity \L\ = 450 pb$^{-1}$.
293: The search has been performed with a pure \ks\ beam obtained by
294: tagging with \kl\ interactions in the calorimeter and
295: detecting six photons. We find an upper limit for the branching
296: ratio of $1.2 \times 10^{-7}$ at 90\% C.L.
297: \begin{keyword}
298: $e^+e^-$ collisions \sep DA\char8NE \sep KLOE \sep rare
299: $K_S$ decays \sep \C\P \sep \C\P\T
300: %% PACS codes here, in the form: \PACS code \sep code
301: %\PACS 1234 \sep 1234
302: \end{keyword}
303: \end{abstract}
304: \end{frontmatter}
305: %==============================================================================
306: \parindent 6mm%%1\z@ % Indentation of each paragraph.
307:
308: \section{Introduction}
309: The decay $\ks\to 3\pi^0$ violates \C\P\ invariance.
310: The parameter $\eta_{000}$, defined as the ratio of
311: \ks\ to \kl\ decay amplitudes, can be written as:
312: $\eta_{000} = A(\ks\to3\po)/A(\kl\to3\po)=
313: \epsilon + \epsilon'_{000}$, where $\epsilon$ quantifies
314: the \ks\ \C\P\ impurity and $\epsilon'_{000}$ is
315: due to a direct \C\P-violating term.
316: Since we expect $\epsilon'_{000}\;\ll\;\epsilon$
317: \cite{dambrosio}, it follows that $\eta_{000}\!\ab\!\epsilon$.
318: In the Standard Model, therefore,
319: BR($\ks\to 3\pi^0$) \ab \pt1.9;-9; to an accuracy of a few \%,
320: making the direct observation
321: of this decay quite a challenge.
322:
323: The best upper limit on BR($\ks\to 3\pi^0$) from a search
324: for the decay was obtained by the SND experiment at
325: Novosibirsk. They find
326: BR($\ks\to 3\pi^0$)\pt\,\leq1.4;-5; at 90\% C.L.
327: \cite{SND3pi0}.
328: CPLEAR has pioneered the method of searching for
329: interference between \ks\ and \kl\ decays. Interference
330: results in the appearance of a
331: term $\Re\,(\eta_{000})\cos(\Delta m\,t)-\Im\,(\eta_{000})\sin(\Delta m\,t)$
332: in the decay intensity. $\Re\,(\eta_{000})$
333: and $\Im\,(\eta_{000})$ are obtained from a fit,
334: without discriminating between \kl\ or $\ks\to\ 3\po$ decays. In this
335: way CPLEAR finds $\eta_{000}=(0.18\pm0.15)+i\,(0.15\pm 0.20)$
336: \cite{CPLEAR3pi0}.
337: The NA48 collaboration~\cite{NA48} has recently
338: reached much higher sensitivity. By fitting the
339: $\ks/\kl\to 3\pi^0$ interference pattern at small decay
340: times, they find
341: $\Re\,(\eta_{000})=-0.002\;\pm\;0.011_{\rm stat}\;\pm\;0.015_{\rm sys}$
342: and $\Im\,(\eta_{000})=-0.003\;\pm\;0.013_{\rm stat}\;\pm\;0.017_{\rm sys}$,
343: corresponding to BR($\ks\to 3\pi^0$)\pt\,\leq7.4;-7; at 90\% C.L.
344: The sensitivity to \C\P\T\ violation via unitarity \cite{Zhou} is now
345: limited by the error in
346: $\eta_{+-} = A(\kl \to \pi^+\pi^-)/A(\ks \to \pi^+\pi^-)$.
347:
348: We report in the following an improved limit from a
349: {\it direct search} for the 3\po\ decays of the \ks.
350: Apart from the interest in confirming the Standard Model, knowledge
351: of $\eta_{000}$ allows tests of the validity of
352: \C\P\T\ invariance using unitarity.
353:
354:
355: \section{\DAF\ and KLOE}
356: The data were collected with the KLOE detector~[\citen{kloe1}--\citen{kloe4}]
357: at DA$\Phi$NE~\cite{dafne},
358: the Frascati $\phi$ factory. DA$\Phi$NE is an
359: $e^+e^-$ collider
360: operated at a center-of-mass energy $W\sim1020$ MeV,
361: the mass of the $\phi$ meson. Positron and electron
362: beams of equal energy collide at an angle of $\pi-$
363: 0.025 rad, producing $\phi$ mesons nearly at rest
364: ($p_\phi$ \ab\ 12.5 MeV). $\phi$ mesons decay 34\% of the
365: time into nearly collinear $K^{0}\overline K^0$ pairs.
366: Because $J^{PC}(\phi)=1^{--}$, the kaon pair is in a
367: \C-odd antisymmetric state, so that the final state is
368: always \ks-\kl. Detection of a \kl\ signals the
369: presence of a \ks\ of known momentum and direction.
370: We say that detection of a \kl\ ``tags'' the \ks.
371:
372: The KLOE detector consists of a large cylindrical
373: drift chamber (DC), surrounded by a
374: lead/scintillating-fiber electromagnetic calorimeter
375: (EMC). A superconducting coil around the calorimeter
376: provides a 0.52 T field.
377: The drift chamber, 4~m in diameter and 3.3~m long,
378: is described in Ref. \citen{kloe1}. The momentum
379: resolution is $\sigma(p_{\perp})/p_{\perp}\approx 0.4\%$.
380: Two track vertices are reconstructed with a spatial
381: resolution of \ab\ 3~mm.
382: The calorimeter, described in Ref. \citen{kloe2},
383: is divided into a barrel and two endcaps, for a
384: total of 88 modules, and covers 98\% of the solid
385: angle. The modules are read out at both ends by
386: photomultipliers providing energy deposit and arrival
387: time information.
388: The readout segmentation provides
389: the coordinates transverse to the fiber plane.
390: The coordinate along the fibers is
391: obtained by the difference between the arrival times of the
392: signals at either end.
393: %The modules are segmented in
394: %five planes in the direction normal to the particle path.
395: %The \ab\ 4.4\x\ 4.4 cm\up2 readout
396: %segmentation provides the $\{r,\phi\}$ coordinates
397: %of single energy deposits. The $z$ coordinate is
398: %obtained by the difference between the arrival times of the
399: %signals at either end.
400: Cells close in time and space are grouped into
401: calorimeter clusters. The energy and time resolutions
402: are $\sigma_E/E = 5.7\%/\sqrt{E\ {\rm(GeV)}}$ and
403: $\sigma_T = 54\ {\rm ps}/\sqrt{E\ {\rm(GeV)}}\oplus50\ {\rm ps}$,
404: respectively.
405:
406: The KLOE trigger, described in Ref. \citen{kloe4},
407: uses calorimeter and chamber information. For this
408: analysis, only the calorimeter signals are used.
409: Two energy deposits above threshold ($E>50$ MeV for the
410: barrel and $E>150$ MeV for the endcaps) are required.
411: Recognition and rejection of cosmic-ray events is also
412: performed at the trigger level. Events with two energy
413: deposits above a 30 MeV threshold in two of the outermost
414: calorimeter planes are rejected.
415:
416: During 2002 data taking, the maximum luminosity
417: reached by DA$\Phi$NE was \pt7.5;31; cm\up{-2}s\up{-1},
418: and in September 2002, \DAF\ delivered 91.5 pb\up{-1}.
419: We collected data in 2001-2002 for an integrated
420: luminosity \L\ = 450 pb$^{-1}$. A total of 1.4 billion $\phi$ mesons
421: were produced, yielding 450 million \ks-\kl\ pairs.
422: Assuming BR($\ks\to 3\pi^0$) = \pt1.9;-9;, \ab1 signal
423: event is expected to have been produced.
424: %The entire collected sample (2001-2002) amounts to 450 pb$^{-1}$,
425: %equivalent to 1.4 billion (450 million) $\phi$ decays ($K_S$,$K_L$ pairs).
426: %Assuming 100\% efficiency we should have produced 0.9
427: %$K_S \to 3 \pi^0$ standard model events.
428: %\subsection{\ks-\kl\ at \DAF}
429:
430: The mean decay lengths of the $K_S$ and $K_L$ are
431: $\lambda_S \sim 0.6$ cm and $\lambda_L \sim 340$ cm
432: at DA$\Phi$NE.
433: About 50\% of $K_L$'s reach the calorimeter before
434: decaying. The $K_L$ interaction in the calorimeter
435: (``\kl\ crash'') is identified by requiring a cluster
436: with energy greater than 100 MeV that is not associated
437: to any track and whose time corresponds to a velocity
438: in the $\phi$ rest frame, $\beta^*$, of
439: %$\beta=R/(cT)$
440: \ab\ 0.2. The \kl-crash provides a very
441: clean $K_S$ tag. The average value of the center-of-mass
442: energy, $W$, is obtained with a precision of 30 keV
443: for each 100 nb$^{-1}$ running period (of duration $\sim 1$ hour)
444: using large-angle Bhabha events. The value of $W$ and
445: the $K_L$-crash cluster position allows us to establish,
446: for each event, the trajectory of the $K_S$ with an
447: angular resolution of 1$^{\circ}$ and a momentum
448: resolution better than 2 MeV.
449:
450: Because of its very short lifetime, the displacement
451: of the $K_S$ from the $\phi$ decay position
452: is negligible. We therefore identify as \ks\ decay
453: photons neutral particles that travel with $\beta=1$
454: from the interaction point (IP) to the EMC. Each cluster
455: is required to satisfy the condition
456: $|T-R/c|<{\rm min}(3\sigma_T, 2\ {\rm ns})$,
457: where $T$ is the photon flight time and $R$ the path
458: length; $\sigma_T$ also includes a contribution from
459: the finite bunch length (2--3 cm), which introduces
460: a dispersion in the collision time.
461:
462: In order to retain a large control sample for the
463: background while preserving high efficiency for the
464: signal, we keep all photons satisfying $E_\gamma>$ 7 ~MeV and
465: $|\cos(\theta)|<$ 0.915. The photon detection efficiency
466: is \ab90\% for $E_\gamma$ = 20 MeV, and reaches
467: 100\% above 70 MeV. The signal is searched for by requiring six
468: prompt photons after tagging.
469:
470: The normalization is provided by counting the $\ks\to 2\pi^0$ events
471: in the same tagged sample.
472:
473: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
474: \section{Monte Carlo simulation}
475: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
476: The response of the detector to the decay of interest and
477: the various backgrounds is studied using the
478: Monte Carlo (MC) program GEANFI \cite{NIM}.
479: GEAN\-FI accounts for
480: changes in machine operation and background conditions,
481: following the machine conditions run by
482: run, and has been calibrated with Bhabha scattering events and other
483: processes. The response of the EMC to \kl\ interactions
484: is not simulated but has been obtained from a large
485: sample of \kl-mesons tagged by identifying
486: $\ks\to\pic$ decays.
487: This not only gives accurate representation of the
488: EMC response to the
489: $K_L$ crash, but also results in an effective 40\%
490: increase in MC statistics.
491: The \kl-crash
492: efficiency cancels in the final 3\po/2\po\ ratio to
493: better than 1\% and we assign a 0.9\% systematic error
494: to the final result due to this source.
495:
496: Backgrounds are obtained from MC $\f \to \ks\kl$ events
497: corresponding to an integrated luminosity \L\ = 900 pb\up{-1}.
498: We also use a MC sample of $\f\to\kpkm$
499: events for \L\ = 450 pb$^{-1}$ and a MC sample of radiative \f\
500: decays for \L\ = 2250 pb$^{-1}$. A sample of \ab\ $340\,000$
501: $\ks\to 3\pi^0$
502: MC events is used to obtain the signal efficiency.
503: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
504: \section{Photon counting for data and Monte Carlo}
505: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
506: To test how well the MC reproduces the observed photon
507: multiplicity after tagging, we determine the fraction
508: of events of given multiplicity, $N_{\gamma}=k$,
509: defined as
510: $F(k)= N_{{\rm ev}}(N_{\gamma}=k)/\sum_{i=3}^{6}{N_{{\rm ev}}(N_{\gamma}=i)}.$
511: As shown in Table \ref{tab:fraction}, there is a
512: significant discrepancy between data and Monte Carlo
513: for events with multiplicity five and six.
514: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
515: \begin{table*}[hb]
516: \begin{center}
517: \newcommand{\m}{\hphantom{$-$}}
518: \newcommand{\cc}[1]{\multicolumn{1}{c}{#1}}
519: \renewcommand{\arraystretch}{1.1} % enlarge line spacing
520: \begin{tabular}{@{}lcccc}
521: \hline
522: & Data 2001 & MC 2001 & Data 2002 & MC 2002 \\ \hline
523: $F$(3) &30.95\plm0.16 &30.31\plm0.11 &30.79\plm0.12 &30.06\plm0.08 \\
524: $F$(4) &67.35\plm0.23 &67.93\plm0.17 &67.93\plm0.18 &68.15\plm0.12 \\
525: $F$(5) & 1.55\plm0.01 & 1.80\plm0.01 & 1.19\plm0.01 & 1.66\plm0.01 \\
526: $F$(6) & 0.15\plm0.01 & 0.14\plm0.01 & 0.08\plm0.01 & 0.13\plm0.01 \\ \hline
527: \hline
528: \end{tabular}\\[2pt]
529: \caption{Measured values of $F$ for data and Monte Carlo samples, in percent. }
530: \label{tab:fraction}
531: \end{center}
532: \end{table*}
533: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
534: These samples
535: are dominated by $K_S \to 2 \pi^0$ decays plus
536: additional clusters due either to shower fragmentation
537: ({\em split clusters}) or the accidental coincidence
538: of machine background photons ({\em accidental clusters}).
539: To understand this discrepancy,
540: we have measured the probability, $P_A(1,2)$,
541: of having one, or more than one,
542: accidental cluster passing our selection
543: %requirements.we did this
544: by extrapolating
545: the rates measured in an out-of-time window,
546: $(-68 \leq T \leq -14)$~ns,
547: that is earlier than the bunch crossing.
548: In Table \ref{tab:probacci_split},
549: we list the average values of these probabilities.
550: %\def\xc{10\up2\x} \def\xm{10\up3\x} \def\xcc{10\up4\x}
551: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
552: \begin{table*}[hb]
553: \begin{center}
554: \newcommand{\m}{\hphantom{$-$}}
555: \newcommand{\cc}[1]{\multicolumn{1}{c}{#1}}
556: %\renewcommand{\tabcolsep}{1.5pc} % enlarge column spacing
557: \renewcommand{\arraystretch}{1.2} % enlarge line spacing
558: \begin{tabular}{@{}lcccc}
559: \hline
560: $F(K)$ & Data 2001 & MC 2001 & Data 2002 & MC 2002 \\ \hline
561: \xc$P_A$(1) &0.75\plm0.30 &1.03\plm0.16 &0.38\plm0.17 &0.89\plm0.08 \\
562: \xc$P_A$(2) &0.14\plm0.05 &0.16\plm0.03 &0.07\plm0.02 &0.10\plm0.03 \\
563: \xm$P_S(1)$ &3.6 \plm 0.2 &3.8 \plm 0.3 &3.7 \plm 0.2 & 3.3 \plm 0.1 \\
564: \xcc$P_S(2)$ &1.5 \plm 0.4 &1.5 \plm 0.3 &0.9 \plm 0.2 & 1.7 \plm 0.2 \\ \hline
565: \hline
566: \end{tabular}\\[2pt]
567: \caption{Measured values of the probabilities
568: $P_A$ and $P_S$.}
569: \label{tab:probacci_split}
570: \end{center}
571: \end{table*}
572: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
573: The observed discrepancy has been traced to an
574: understood problem with the procedure for the selection
575: of machine-background clusters.
576:
577: The MC-true fraction of
578: events with a given multiplicity, $f_{MC}$, is obtained
579: by ignoring clusters due to machine background and
580: counting at most one cluster per simulated particle
581: incident on the calorimeter.
582: Using the fractions $f_{MC}$, together with the
583: values of $P_A$ obtained as discussed above,
584: we fit the observed $F(k)$ distribution to get the
585: probability for a cluster to generate fragments, $P_{S}$
586: (see Table \ref{tab:probacci_split}).
587: This fit accurately reproduces the observed
588: fractions in the multiplicity bins five and six.
589: %The observed fractions in the $N_{\gamma}=5, 6$ are
590: %well reproduced.
591: More details on these measurements
592: can be found in Ref. \citen{NotaAcci}.
593: The results of this study demonstrate the need for
594: careful calibration of the background composition
595: when comparing data and MC samples.
596:
597:
598: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
599: \section{ Data analysis}
600: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
601: $\ks\to 3\pi^0$ candidates consist of a \kl\ crash plus six photons.
602: In our data sample of \L\ = 450 pb\up{-1}, we find $39\,538$ events,
603: essentially all background. After removing background, we
604: obtain the branching ratio by normalizing to the number of
605: $\ks\to 2\pi^0$ events. The latter are found by asking for three
606: to five prompt photons plus the \kl-crash.
607:
608: According to the MC, the six-photon sample is dominated (95\%)
609: by $\ks\to 2\pi^0$ decays plus two additional photon clusters.
610: These clusters are due to fragmented or split showers
611: (2S, 1S+1A, 34\%) and to accidental photons from machine background
612: (2A, 64\%).
613: About 2\% of the background events are due to false $K_L$-crash
614: tags from $\f\to\ks\kl\to\pic$, 3\po\ events.
615: In such events, charged pions from \ks\ decays interact in the low-beta
616: insertion quadrupoles\footnote{The first quadrupoles are located
617: approximately 45 cm on either side of the IP.}, ultimately
618: simulating the \kl-crash signal, while \kl\ decays close to
619: the IP produce six photons.
620: Similarly, $\f\to\kpkm$ events give a false signal (\ab\ 1\%),
621: as well as $\f\to\eta\gam\to 3\pi^0\gam$ events (\ab\ 0.3\%).
622: The cuts described in the following make these contaminations
623: negligible.
624:
625: To reduce the background, we first perform a kinematic fit
626: with 11 constraints: energy and momentum conservation, the
627: kaon mass and the velocity of the six photons.
628: The $\chi^2$ distribution of the fit to data and MC background
629: is shown in Fig. \ref{fig:chi2fit}.
630: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
631: \begin{figure}[htb]
632: \begin{center}
633: \includegraphics[width=0.5\textwidth]{chifit.eps}
634: \vglue-15mm\end{center}
635: \caption{Distribution of $\chi^2$ for
636: the tagged six-photon sample for data (points),
637: MC background (solid line), and $10\,000$ events of MC signal (dashed line).}
638: \label{fig:chi2fit}
639: \end{figure}
640: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
641: In the same plot, we also show the expected shape for
642: signal events.
643: Cutting at a reasonable $\chi^2$ value
644: ($\chi^{2}/\rm{11} < 3$)
645: retains 71\% of the signal while considerably
646: reducing the background from false $K_L$-crash events
647: (33\%), in which the direction of the $K_S$
648: and $K_L$ are not correlated.
649: However, this cut is not as effective on
650: the 2S, 2A background, due to
651: the soft energy spectrum of the fake clusters.
652: In order to gain rejection power over the background for
653: events with split and accidental clusters,
654: we look at the correlation between the following two
655: $\chi^2$-like estimators:
656: \begin{itemize}
657: \item $\zeta_2$, defined as
658: \begin{eqnarray}
659: \zeta_2 & = & \ctwoppmas+ \nonumber \\
660: && \;\;\; + \ctwoppimp \nonumber \\
661: \nonumber
662: \end{eqnarray}
663: selecting the four out of six photons
664: that provide the
665: best kinematic agreement with the $K_S\to 2\pi^0$ decay
666: hypothesis. This variable is quite insensitive to fake
667: clusters. It is constructed using the two values of
668: $\Delta m$ = $m_i-m_{\pi^0}$ (where $m_i$ is the
669: invariant mass of a photon pair),
670: the opening angle between $\pi^0$'s in the $K_S$ rest
671: frame, and 4-momentum conservation.
672: The resolutions on these quantities have been
673: evaluated using a control sample of events with a
674: $K_L$-crash and four prompt photons.
675: \item $\zeta_3$, defined as
676: $$\zeta_3=\frac{\Delta m_1^2}{\sigma_{m}^2}+
677: \frac{\Delta m_2^2}{\sigma_{m}^2}+
678: \frac{\Delta m_3^2}{\sigma_{m}^2}$$
679: %using the three best recontructed mass that minimize
680: %$\Delta m_i = (m_i -m_{\po})$, weighted by
681: %the mass resolution calculated for the decay
682: %(on average, $\sigma_{\pi} \sim$ 18 MeV).
683: where the pairing of the six photons into $\pi^0$'s is
684: performed by minimizing
685: this variable.
686: $\zeta_3$ is close to zero for a $\ks\to 3\pi^0$ event and
687: large for six-photon background events.
688: \end{itemize}
689: %The distributions of these variables have been
690: %obtained using the sample of data with a $K_L$-crash,
691: %and six or four prompt photons respectively.
692: For each estimator, the photon pairing
693: with smallest $\zeta$ value is kept.
694: Figure~\ref{fig:scatterplots}a shows the distribution of
695: events in the $\zeta_2$-$\zeta_3$ plane
696: for the MC background. Most of the events are
697: concentrated at low values of $\zeta_2$,
698: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
699: \begin{figure}[hb]
700: \begin{center}
701: \includegraphics[width=1.\textwidth]{c2c3.eps}
702: \end{center}
703: \vspace{-.3cm}
704: \caption{Scatter plot of $\zeta_2$ vs.
705: $\zeta_3$ plane for the tagged six-photon sample:
706: a) MC background, b) MC signal.}
707: \label{fig:scatterplots}
708: \end{figure}
709: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
710: as expected for $\ks\to 2\pi^0$ events plus some additional
711: isolated energy deposits in the EMC.
712: A clear signal/background separation is achieved as
713: can be seen by comparing the background and signal
714: distributions in Figs.~\ref{fig:scatterplots}a and
715: \ref{fig:scatterplots}b.
716: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
717: \begin{figure}[htb]
718: \begin{center}
719: \begin{tabular}{cc}
720: \includegraphics[width=.8\textwidth]{4c2c3.eps}
721: \end{tabular}
722: \end{center}
723: \caption{Scatter plots of $\zeta_2$ vs. $\zeta_3$ for
724: the tagged six-photon sample:
725: data (a), MC sample with two split clusters (b),
726: two accidental clusters (c), and false $K_L$-crash events (d).}
727: \label{fig:chi2chi3fit}
728: \end{figure}
729: %%%%%%%%%%%%%%%%%%%%%
730: We subdivide the $\zeta_2$-$\zeta_3$ plane into the six regions
731: B1, B2, B3, B4, B5, and S as indicated in Fig.
732: \ref{fig:scatterplots}a. Region S, with the largest
733: signal-to-background value, is the ``signal'' box.
734:
735: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
736: \begin{table*}[htb]
737: \newcommand{\m}{\hphantom{$-$}}
738: \renewcommand{\arraystretch}{1.2} % enlarge line spacing
739: \begin{tabular}{@{}lcccccc}
740: \hline
741: & B1 & B2 & {\bf S} & B3 & B5 & B4 \\ \hline
742: data & $452\pm 21$ & $10132\pm 101$ &
743: {\boldmath $282\pm 17$} & $5037\pm 71$ &
744: $326\pm 18$ & $22309\pm 149$ \\
745: MC & $419\pm 19$ & $9978\pm 104$ &
746: {\boldmath $282\pm 13$} & $4816\pm 43$ &
747: $380\pm 10$ & $22682\pm 190$\\
748: \hline
749: \end{tabular}\\[2pt]
750: \caption{Comparison between data and MC expectations
751: in the different regions of the $\zeta_2$-$\zeta_3$ plane
752: for the entire sample
753: with a $K_L$-crash and six prompt photons. The boxes are defined as in
754: Fig.~\ref{fig:scatterplots}a.}
755: \label{tab:compare}
756: \end{table*}
757: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
758: The scatter plot in the $\zeta_2$-$\zeta_3$ plane for
759: the data is shown in Fig.~\ref{fig:chi2chi3fit}a.
760: %From the data-MC comparison
761: %of the probabilities $P_A$ and $P_S$,
762: Our MC simulation does not accurately reproduce the
763: absolute number of 2S and 2A background events. This
764: is also true of the predicted number of false
765: $K_L$-crash events. However, the simulation does
766: describe the kinematical properties of these events
767: quite well.
768: The two-dimensional $\zeta_2$-$\zeta_3$ distribution
769: allows us to calibrate the contributions from the different backgrounds.
770: The MC shapes for each of the three categories are
771: shown in Figs.~\ref{fig:chi2chi3fit}b-d.
772: We perform a binned likelihood fit of a linear combination
773: of these shapes to the data, excluding the signal-box
774: region. From the fit we find the composition of
775: the six-photon sample to be (37.9\plm1.0)\%, (57.4\plm1.3)\%,
776: and (4.7\plm0.3)\% for the 2S, 2A, and false $K_L$-crash
777: categories, respectively.
778:
779: As a check, we compare data and MC for the projected
780: distribution in $\zeta_3$ for the three bands in $\zeta_2$,
781: as shown in
782: Figs.~\ref{fig:chi2chi3}a-b. Excellent agreement is observed.
783: The large peak at low values of $\zeta_3$ in the central band
784: is due to the false $K_L$-crash events.
785: As a final test, we compare data and MC in the signal box
786: and the five surrounding control regions. The agreement is
787: better than 10\% in all regions, as seen from Table \ref{tab:compare}.
788: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
789: \begin{figure}[!t]
790: \begin{center}
791: \includegraphics[width=.6\textwidth]{3proj.eps}
792: \end{center}
793: \caption{
794: Distributions in $\zeta_3$ for the tagged six-photon sample.
795: Plots on the left are for events in the central band in
796: $\zeta_2$;
797: plots on the right are for events in all other regions of
798: the plane.
799: The plots in the top row are for the entire sample,
800: before any cuts are made. The plots in the bottom row are after
801: the application of the track veto.
802: In all cases, black points represent data; solid line represents MC.}
803: \label{fig:chi2chi3}
804: \end{figure}
805:
806:
807: Although cutting on $\chi^2$ substantially suppresses
808: the false $K_L$-crash background, we reduce this background
809: to a negligible level by vetoing events with tracks coming
810: from the IP. This effectively eliminates events in which
811: the false $K_L$-crash is due to a $\ks\to \pic$ decay with
812: the pion secondaries interacting in the quadrupoles. The
813: effect on the signal region can be appreciated by comparison
814: of Figs.~\ref{fig:chi2chi3}a and \ref{fig:chi2chi3}c.
815: Moreover, in order to improve the quality of the photon
816: selection using $\zeta_2$, we cut on the variable
817: $\Delta E=(m_\phi/2-\sum E_i)/\sigma_{E}$, where $i$ = 1--4 stands
818: for the four chosen photons
819: in the $\zeta_2$ estimator
820: and $\sigma_E$ is the appropriate
821: resolution. For $\ks\to 2\pi^0$ decays plus two background
822: clusters, we expect $\Delta E$ \ab\ 0,
823: while for $\ks\to\pio\po$, $\Delta E\geq m_{\pi^0}/\sigma_{E}$.
824:
825: Before opening the signal box,
826: we refine our cuts on $\chi^2$, $\zeta_2$, $\zeta_3$,
827: and $\Delta E$
828: %to minimize the upper limit on the branching ratio
829: %following
830: using the optimization procedure described in Ref.~\citen{METODO}.
831: %we optimize our
832: %search~\cite{METODO} using the MC signal and background samples
833: %and varying the cuts on $\chi^2$, $\zeta_2$, $\zeta_3$,
834: %and $\Delta E$.
835: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
836: \begin{figure}[!t]
837: \begin{center}
838: \includegraphics[width=.8\textwidth]{cb_sb_fineana-n.eps}
839: \end{center}
840: \caption{Distributions of $\zeta_3$ for
841: the central band 12.1$<\zeta_2<$60 (a),
842: the side-bands $\zeta_2<$12.1, $\zeta_2> $60 (b),
843: after all cuts.
844: Points represent data, solid line MC.}
845: \label{fig:results}
846: \end{figure}
847: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
848: We end up choosing $\chi^2<40.4$ and $\Delta E>$1.7.
849: The signal box is defined by 12.1 $<\zeta_2<$ 60 and $\zeta_3<$ 4.6.
850:
851: Figures \ref{fig:results}a and \ref{fig:results}b show the $\zeta_3$ distributions
852: for the central band and the sidebands in $\zeta_2$.
853: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
854:
855: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
856: \begin{table*}[t]
857: \begin{center}
858: \newcommand{\m}{\hphantom{$-$}}
859: \renewcommand{\arraystretch}{1.1} % enlarge line spacing
860: \begin{tabular}{@{}lcccccc}
861: \hline
862: & B1 & B2 & {\bf S} & B3 & B4 & B5 \\ \hline
863: data & 0 & 4 \plm\ 2&{\bf 2.0 \plm\ 1.4}&520 \plm\ 23&3 \plm\ 2&326 \plm\ 18\\
864: MC &0&3.2 \plm\ 0.8&{\bf3.1 \plm\ 0.8}&447 \plm\ 10&2.5 \plm\ 0.8&389 \plm\ 10\\ \hline
865: \end{tabular}\\[2pt]
866: \end{center}
867: \caption{Same as Table \ref{tab:compare}, after all cuts.}
868: \label{tab:compare_fin}
869: \end{table*}
870: In Table~\ref{tab:compare_fin}, we also list
871: the number of events obtained in each of the six regions
872: %in the optimized signal box and in the
873: %surrounding control regions
874: of the $\zeta_2$-$\zeta_3$ plane
875: at this final stage of the analysis.
876: In Figs.~\ref{fig:scatter_end}a-b we show the
877: $\zeta_2$-$\zeta_3$ scatter plots
878: for data and Monte Carlo.
879: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
880: \begin{figure}[htb]
881: \begin{center}
882: \includegraphics[width=.6\textwidth]{sbox.eps}
883: \end{center}
884: \caption{Distribution of $\zeta_2$ vs $\zeta_3$
885: %at the end
886: %of the analysis chain
887: after cuts: MC background
888: 900 pb$^{-1}$ (a), data 450 pb$^{-1}$ (b).}
889: \label{fig:scatter_end}
890: \end{figure}
891: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
892: The rectangular region illustrates
893: the boundaries of the optimized signal box.
894: Seventeen MC events are counted
895: in this
896: region before applying the data-MC scale factors
897: resulting from the calibration procedure described above.
898: Contributions to the scale factors include the fact
899: that the simulated integrated luminosity is greater than
900: that for the data set (\x2), the increased
901: $K_L$-crash efficiency in the simulation (\x1.4),
902: and the increased probability of having accidental or
903: split clusters in the simulation (on average,
904: \ab\x1.9).% for a total of \dots.
905: %%%%%%%%%%%%%%%
906:
907: The selection efficiency at each step of the analysis has been studied
908: using the MC.
909: After tagging, the efficiency for the six-photon selection is
910: $(47.8 \pm 0.1_{\rm{stat}})$\%.
911: Including all cuts, we estimate a total efficiency
912: of $\epsilon_{3\pi} = (24.4 \pm 0.1_{\rm{stat}} $)\%. At the
913: end of the analysis chain, we have two candidates
914: with an expected background of
915: $B^{\rm{exp}} = 3.13 \pm 0.82_{\rm{stat}}$.
916:
917:
918: In the same tagged sample, we also count
919: events with
920: photon multiplicities of three, four, or five.
921: The corresponding efficiency
922: is (91.8\plm0.2\dn{\rm stat})\% for $\ks\to\pio$ events.
923: The residual background contamination is estimated to be
924: (0.77 $\pm 0.24_{\rm{stat+sys}})\%$ and
925: (0.65$\pm 0.10_{\rm{stat+sys}})\%$
926: in the 2001 and 2002 running periods respectively.
927: Subtracting
928: the background and correcting for the efficiency,
929: we count \pt3.78;7; $\ks\to\pio$ events. We use this number to normalize
930: the number of signal events when obtaining the
931: branching ratio.
932:
933: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
934: \section{Systematic uncertainties}
935: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
936: Systematics arise from uncertainties in estimation
937: of the acceptance, backgrounds, and the analysis efficiency.
938: The evaluation of the systematic uncertainties is described
939: in detail in Ref. \citen{NOTA}.
940:
941: Concerning the acceptance of the event selection for both
942: the 2\po\ and 3\po\ samples, we estimate the systematic
943: errors in photon counting by comparing data and MC
944: values for the $P_A$ and $P_S$ probabilities described above.
945: %As described in Ref.~\citen{NOTA},
946: The photon
947: reconstruction efficiency
948: for both data and MC is evaluated using a large sample of
949: $\f\to\pic\po$, \po\to\gam\gam events.
950: The momentum of one of the photons is estimated from
951: tracking information and position of the other cluster.
952: The candidate photon is searched for within a search
953: cone.
954: The efficiency is
955: parameterized as a function of the photon energy.
956: Systematics related to this correction are obtained
957: from the variation of the efficiency as a function
958: of the width of the search cone.
959: %Varying the photon acceptance criteria we evaluate the
960: %systematics uncertainty.
961: The results are listed in Table \ref{tab:syst_effi}
962: under the heading {\em cluster}.
963: %%%%
964: \begin{table*}[!htb]
965: \begin{center}
966: \newcommand{\m}{\hphantom{$-$}}
967: \newcommand{\cc}[1]{\multicolumn{1}{c}{#1}}
968: \renewcommand{\arraystretch}{1.1} % enlarge line spacing
969: \begin{tabular}{@{}lcc}
970: \hline
971: & $\Delta \alpha / \alpha$
972: ($K_S \to 2 \pi^0$) &
973: $\Delta \alpha / \alpha$ ($K_S \to 3 \pi^0$)
974: \\ \hline
975: Cluster & 0.16 \% & 0.70 \% \\
976: Trigger & 0.08 \% & 0.08 \% \\
977: Background filter & 0.20 \% & 0.08 \% \\ \hline \hline
978: Total & 0.27 \% & 0.71 \% \\
979: \hline
980: \end{tabular}\\[2pt]
981: \caption{Systematic acceptance uncertainties,
982: $\Delta\alpha$, for the 2\po\ and 3\po\ event selection criteria.}
983: \label{tab:syst_effi}
984: \end{center}
985: \end{table*}
986: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
987: The total uncertainty is smaller for the normalization
988: sample since an inclusive
989: selection criterion is used in this case.
990:
991: The normalization sample also suffers
992: a small ($0.4\%$) loss due to the
993: use of a filter during data reconstruction to reject
994: cosmic rays, Bhabha fragments from the low-beta quadrupole,
995: and machine background events. This loss
996: is estimated using the MC. We correct for it and
997: add a 0.2\% systematic error
998: to the selection efficiency.
999: The trigger and cosmic-ray veto efficiencies
1000: have been estimated with data for the normalization
1001: sample and extrapolated by MC to the signal sample.
1002: These efficiencies are very close to unity and
1003: the related systematics are negligible.
1004:
1005: For the tagged six-photon sample, we have investigated
1006: the uncertainties related to the estimate of the
1007: background in the signal box after all cuts,
1008: $B^{\rm{exp}}$. % for the six photon sample.
1009: We have first considered the error related to the
1010: calibration of the MC background composition
1011: by propagating the errors on the scale
1012: factors obtained from the fit. %reported in Tab.~\ref{tab:syst_bkg_ana}.
1013: This corresponds to a relative error of $2.4\%$ on $B^{\rm{exp}}$.
1014: Moreover, we have investigated the extent to which the
1015: track-veto efficiency influences the residual false $K_L$-crash
1016: contamination. To do so, we examine the data-MC ratio, $R_{\beta}$,
1017: of the sidebands in $\beta^{*}$ for events rejected by
1018: this veto, since for true $K_L$'s $\beta^*$ peaks at \ab\ 0.2 while
1019: false $K_L$-crashes are broadly distributed in $\beta^*$.
1020: We obtain $R_{\beta}=1.10 \pm 0.01$.
1021: Knowing that in the MC only 24\% of the fakes survive the
1022: veto, we find a fractional error of 32\%
1023: on the fake background. Since false $K_L$-crash
1024: events account for 15\% of the total background,
1025: the error on $B^{\rm{exp}}$ from data-MC differences
1026: in the track veto efficiency is 4.6\%.
1027:
1028: A control sample of $K_S \to 2 \pi^{0}$
1029: with four prompt photons has been used to compare
1030: the energy scale and
1031: resolution of the calorimeter in data and in the MC.
1032: % have
1033: %also been compared between data and
1034: %MC by using the pulls of the kinematic fit
1035: %in a control sample of $K_S \to 2 \pi^{0}$ with four prompt photons.
1036: The distributions of the $m$ and $\Delta E$ variables have
1037: also been compared by fitting them with Gaussians.
1038: By varying the mass and energy resolution by $\pm 1 \sigma$
1039: in the definitions of $\zeta_2$ and $\zeta_3$, we
1040: observe a relative change of 6.6\% in the background
1041: estimate. Similarly, correcting
1042: for small differences in the energy scale for data and MC, we
1043: derive a systematic uncertainty of 6.7\% on $B^{\rm{exp}}$.
1044:
1045: Finally, we have tested the effect of the cut on
1046: $\chi^2$ by constructing the ratio between
1047: the cumulative distributions for data and MC. An error
1048: of 5\% is obtained. A summary of all the
1049: systematic errors on the background estimate is given
1050: in Table~\ref{tab:syst_bkg_ana}.
1051: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1052: \begin{table*}[htb]
1053: \begin{center}
1054: \newcommand{\m}{\hphantom{$-$}}
1055: \newcommand{\cc}[1]{\multicolumn{1}{c}{#1}}
1056: \renewcommand{\arraystretch}{1.1} % enlarge line spacing
1057: \begin{tabular}{@{}lcc}
1058: \hline
1059: & $\Delta B^{\rm{exp}} / B^{\rm{exp}}$ &
1060: $\Delta \epsilon_{\rm ana} / \epsilon_{\rm ana}$ \\ \hline
1061: Background composition & 2.4\% & - \\
1062: Track veto & 4.8\% & 0.2\% \\
1063: Energy resolution & 6.6\% & 0.5\% \\
1064: Energy scale & 6.7\% & 1.0\% \\
1065: $\chi^2$ & 5.0\% & 1.8\% \\ \hline \hline
1066: Total & 11.5\% & 2.1\% \\
1067: \hline
1068: \end{tabular}\\[2pt]
1069: \caption{Systematic uncertainties on the expected background
1070: and analysis efficiency, $\epsilon_{\rm ana}$.}
1071: \label{tab:syst_bkg_ana}
1072: \end{center}
1073: \end{table*}
1074: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1075: Adding in quadrature all sources we obtain a total
1076: systematic error of 12\%
1077: on the background estimate. %$B^{exp} = 0.37$.
1078:
1079: To determine the systematics related
1080: to the analysis cuts for the signal, we have
1081: first evaluated the effect of the track veto.
1082: Using the MC signal sample, we estimate a
1083: vetoed event fraction of (3.7 $\pm$ 0.1)\%.
1084: The data-MC ratio of the cumulative distributions
1085: for the track-vetoed events in the tagged six-photon
1086: sample is $R_{TV}=1.06\pm0.03$, which translates into a
1087: 0.2\% systematic error on the track-veto efficiency.
1088:
1089: Because of the similarity of the $\chi^2$
1090: distributions
1091: for the tagged four- and six-photon samples, as confirmed
1092: by MC studies, an estimate of the systematic error
1093: associated with the application of the $\chi^2$
1094: cut can be obtained from the data-MC comparison
1095: of the cumulative $\chi^2$ distributions for the
1096: four-photon sample.
1097: The systematic error arising from data-MC discrepancies
1098: in the $\chi^2$ distribution is
1099: estimated to be 1.8\% by this comparison.
1100: %A control sample of $K_S \to 2 \pi^{0}$ with four
1101: %prompt photons, and the Monte Carlo comparison
1102: %between these events and the signal events, allows
1103: %us to quote a 1.8\% systematic error connected to
1104: %the application of the $\chi^2$ cut.
1105:
1106: Moreover, the efficiency changes related to
1107: differences between the calorimeter resolution and energy
1108: scale for data and MC events have
1109: been studied in a manner similar to
1110: that previously described for the evaluation of the
1111: systematics on
1112: the background. The systematic
1113: uncertainties on the analysis efficiency are
1114: summarized in Table~\ref{tab:syst_bkg_ana}. Adding all sources
1115: in quadrature we quote a total systematic error of 2.1\%
1116: on the estimate of the analysis efficiency.
1117:
1118: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1119: \section{Results}
1120: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1121: At the end of the analysis, we find 2 events in the signal box
1122: with an estimated background of
1123: $B^{\rm{exp}}=3.13 \pm 0.82_{\rm{stat}} \pm 0.37_{\rm{syst}}$.
1124: To derive an upper limit on the number of signal counts,
1125: we build the background probability distribution function,
1126: taking into account our finite MC statistics and
1127: the uncertainties on the MC calibration factors. This
1128: function is folded with a Gaussian of width equivalent
1129: to the entire systematic uncertainty on the background.
1130: Using the Neyman construction described in Ref.~\citen{FC},
1131: we limit the number of $\ks\to3\po$ decays observed to 3.45
1132: at 90\% C.L., with a total reconstruction efficiency
1133: of $(24.36\pm 0.11_{\rm{stat}}\pm 0.57_{\rm{sys}})\%$.
1134: In the same tagged sample, we count
1135: \pt3.78;7; $\ks\to\pio$ events. This number
1136: is used for normalization.
1137: Finally, using the value BR($\ks\to\pio$) = 0.3105 \plm\ 0.0014~\cite{pdg_2004} we obtain:
1138: \begin{equation}
1139: \br(\ks\to3\po) \leq 1.2 \times 10^{-7} \; \;\;\; {\rm at}\;\; \; 90\%\;\;\; {\rm C.L.},
1140: \end{equation}
1141: which represents an improvement by a factor of $\sim$6
1142: with respect to the best
1143: previous limit \cite{NA48}, and
1144: by a factor of 100 with respect to the best limit obtained with
1145: a direct search~\cite{SND3pi0}. % (performed with a direct search).
1146:
1147: The limit on the BR can be directly translated into
1148: a limit on $|\eta_{000}|$:
1149: \begin{equation}
1150: \eqalign{
1151: |\eta_{000}| = &\left|{A( K_S\to3\pi^0)\over A( K_L\to3\pi^0)}\right|=\cr
1152: &\kern2cm=\sqrt{{\tau_L\over\tau_S}\:{{\rm BR}(\ks\to3\po)\over{\rm BR}(\kl\to3\po)}}
1153: <0.018\ {\rm at}\ 90\%\ {\rm C.L.}\cr}
1154: \end{equation}
1155:
1156: This result describes a circle of radius 0.018 centered
1157: at zero in the $\Re({\eta_{000}})$, $\Im{(\eta_{000})}$ plane and
1158: represents a limit $2.5$ times smaller
1159: than the result of Ref.~\citen{NA48}.
1160: As follows from the discussion in that reference, our
1161: result confirms that the sensitivity of the \C\P\T\ test
1162: from unitarity is now limited by the uncertainty on $\eta_{+-}$.
1163:
1164:
1165: %==============================================================================
1166: \section*{Acknowledgments}
1167: %==============================================================================
1168: %
1169: We thank the DA$\Phi$NE team for their efforts in maintaining
1170: low-background running
1171: conditions and their collaboration during all data taking.
1172: We would like to thank our technical staff:
1173: G.F.Fortugno for his dedicated work to ensure efficient operations of
1174: the KLOE computing facilities;
1175: M.Anelli for his continuous support of the gas system and the safety of the
1176: detector;
1177: A.Balla, M.Gatta, G.Corradi and G.Papalino for the maintenance of the
1178: electronics;
1179: M.Santoni, G.Paoluzzi and R.Rosellini for the general support the
1180: detector;
1181: C.Piscitelli for his help during major maintenance periods.
1182:
1183: This work was supported in part by DOE grant DE-FG-02-97ER41027;
1184: by EURODA$\Phi$NE contract FMRX-CT98-0169;
1185: by the German Federal Ministry of Education and Research (BMBF) contract 06-KA-957;
1186: by Graduiertenkolleg `H.E. Phys. and Part. Astrophys.' of Deutsche Forschungsgemeinschaft,
1187: Contract No. GK 742;
1188: by INTAS, contracts 96-624, 99-37;
1189: %and by TARI, contract HPRI-CT-1999-00088.
1190: and by the EU Integrated Infrastructure Initiative Hadron Physics
1191: Project under contract number RII3-CT-2004-506078.
1192:
1193: \begin{thebibliography}{99}
1194:
1195: \bibitem{dambrosio} G.~D'Ambrosio {\it et al.}, \C\P\ and \C\P\T\
1196: measurements at \DAF, in
1197: {\em The second DA\char8NE handbook}, L.~Maiani {\it et al.} editors, Frascati, 63 (1995).
1198:
1199: \bibitem{SND3pi0} SND Collaboration, M.~N.~Achasov {\it et al.},
1200: Phys. Lett. B 459 (1999), 674.
1201:
1202: \bibitem{CPLEAR3pi0} CPLEAR Collaboration, A.~Angelopoulos {\it et al.},
1203: Phys. Lett. B 425 (1998), 391.
1204:
1205: \bibitem{NA48} NA48 Collaboration, A.~Lai {\it et al.},
1206: Phys. Lett. B610 (2005), 165.
1207:
1208: \bibitem{Zhou} G.B.~Thomson and Y.~Zou, Phys. Rev. D 51 (1995), 1412.
1209: %L. Maiani in
1210: %{\em The Second or First?? DA\char8NE Hfandbook}, in
1211: %{\em The second DA\char8NE handbook}, L.~Maiani {\it et al.}
1212: %editors, Frascati, 63 (1995),
1213:
1214: \bibitem{kloe1} KLOE Collaboration, M.~Adinolfi {\it et al.}, Nucl. Inst. Meth. A 488 (2002), 51.
1215:
1216: \bibitem{kloe2} KLOE Collaboration, M.~Adinolfi {\it et al.}, Nucl. Inst. Meth. A 482 (2002), 364.
1217:
1218: \bibitem{kloe3} KLOE Collaboration, M.~Adinolfi {\it et al.}, Nucl. Inst. Meth. A 483 (2002), 649.
1219:
1220: \bibitem{kloe4} KLOE Collaboration, M.~Adinolfi {\it et al.}, Nucl. Inst. Meth. A 492 (2002), 134.
1221:
1222: \bibitem{dafne} S.~Guiducci, Status Report on \DAF, in: P.~Lucas, S.~Weber (Eds.),
1223: Proceedings of the 2001 Particle Accelerator Conference, Chicago, IL.,
1224: USA, 2001.
1225:
1226: \bibitem{NIM} KLOE Collaboration, F.~Ambrosino {\it et al.},
1227: Nucl. Inst. Meth. A 453 (2004), 403.
1228:
1229: %\bibitem{Tommy} T.~Spadaro, performance of the KLOE calorimeter
1230: %with DA$\Phi$NE data, in: C.~Cecchi {\it et al.}, Proc.
1231: %\emph{Eleventh Int. Conf. on Calorimetry in Particle Physics},
1232: %Perugia, Italy, 2004.
1233:
1234: \bibitem{NotaAcci} M.~Martini and S.~Miscetti, Determination of the probability
1235: of accidental coincidence between machine background and collision events and
1236: fragmentation of electromagnetic showers, KLOE note 201 (2005),
1237: http://www.lnf.infn.it/kloe.
1238:
1239: \bibitem{METODO} J.F.~Grivaz and F.~Le~Diberder, LAL 92-37 (1992).
1240:
1241: \bibitem{NOTA}M.~Martini and S.~Miscetti, A direct search for $\ks\to3\pi^0$
1242: decay, KLOE note 200 (2005),
1243: http://www.lnf.infn.it/kloe.
1244:
1245: \bibitem{FC}G.~J.~Feldman and R.~Cousins, Phys. Rev. D57 (1998), 57.
1246:
1247: \bibitem{pdg_2004}S. Eidelman et al., Phys. Lett. B 592 (2004).
1248:
1249: \end{thebibliography}
1250: \end{document}
1251:
1252:
1253: