1: % Article on b -> penguin dominance of strangeness-changing B -> PPP decays
2: % M. Gronau and J. Rosner
3: % November 4, 2005
4: \documentclass[12pt]{article}
5: \usepackage{graphicx}
6: \textheight 8.4in
7: \voffset -0.4in
8: \textwidth 6.3in
9: \hoffset -0.6in
10: \def \b{{\cal B}}
11: \def \bea{\begin{eqnarray}}
12: \def \beq{\begin{equation}}
13: \def \bra#1{\left| #1 \right \rangle}
14: \def \eea{\end{eqnarray}}
15: \def \eeq{\end{equation}}
16: \def \ite{{\it et al.}}
17: \def \ket#1{\left \langle #1 \right|}
18: \def \lra{\leftrightarrow}
19: \def \ok{\overline{K}^0}
20: \def \s{\sqrt{2}}
21: \def \st{\sqrt{3}}
22: \def \sx{\sqrt{6}}
23: \def \vc{\vec{C}}
24: \def \vp{\vec{P}}
25: \def \vt{\vec{T}}
26: \begin{document}
27: \renewcommand{\thetable}{\Roman{table}}
28: \baselineskip 18pt
29: \begin{flushright}
30: TECHNION-PH-2005-12\\
31: EFI-05-14\\
32: hep-ph/0509155 \\
33: September 2005 \\
34: \end{flushright}
35:
36: \centerline{\bf SYMMETRY RELATIONS IN CHARMLESS $B \to PPP$ DECAYS
37: \footnote{To be published in Physical Review D}}
38: \bigskip
39: \centerline{Michael Gronau$^2$ and Jonathan L. Rosner$^3$}
40: \medskip
41: \centerline{\it $^2$Department of Physics, Technion--Israel Institute of
42: Technology,}
43: \centerline{\it Technion City, 32000 Haifa, Israel}
44: \medskip
45: \centerline{\it $^3$Enrico Fermi Institute and Department of Physics,
46: University of Chicago}
47: \centerline{\it 5640 South Ellis Avenue, Chicago, IL 60637, USA}
48: \bigskip
49: \centerline{\large Abstract}
50: \bigskip
51: Strangeness-changing decays of $B$ mesons to three-body final states
52: of pions and kaons are studied, assuming that they are dominated by a
53: $\Delta I=0$ penguin amplitude with flavor structure $\bar b \to \bar s$.
54: Numerous isospin relations for $B\to K\pi\pi$ and for underlying quasi-two-body
55: decays are compared successfully with experiment, in some cases resolving
56: ambiguities in fitting resonance parameters. The only exception is a somewhat
57: small branching ratio noted in $B^0\to K^{*0}\pi^0$, interpreted in terms of
58: destructive interference between a penguin amplitude and an enhanced
59: electroweak penguin contribution.
60: Relations for $B$ decays into three kaons are derived in terms
61: of final states involving $K_S$ or $K_L$, assuming that $\phi K$-subtracted
62: decay amplitudes are symmetric in $K$ and $\bar K$, as has been observed
63: experimentally. Rates due to nonresonant backgrounds are studied using a
64: simple model, which may reduce discrete ambiguities in Dalitz plot analyses.
65: \bigskip
66:
67: \leftline{PACS numbers: 13.25.Hw, 11.30.Cp, 12.15.Ji, 14.40.Nd}
68: \bigskip
69:
70: \centerline{\bf I. INTRODUCTION}
71: \bigskip
72:
73: The decays of $B$ mesons to charmless three-body final states provide valuable
74: information about the pattern of CP violation, as in the time-dependent
75: studies of CP asymmetries in decays to CP-eigenstates consisting of three
76: neutral pseudoscalars~\cite{Gershon:2004tk}. Data for $B^0\to K^+K^-K_S$ of
77: comparable statistical weight have been presented by the BaBar
78: \cite{Aubert:2004ta,Aubert:2005ja,Aubert:2005kd} and Belle
79: \cite{Abe:2003yt,Garmash:2003er,Chen:2005dr} Collaborations. In analyzing
80: these data it is of interest to know the CP eigenvalue of the three-body final
81: state which depends on the $K^+K^-$ angular momentum.
82: In Ref.\ \cite{Garmash:2003er} isospin symmetry was
83: utilized to relate the decays $B^+ \to K^+ K^0 \bar K^0$ (measured via
84: $B^+ \to K^+ K_S K_S$) and $B^0 \to K^+ K^- K^0$ in order to conclude that
85: the $K^+ K^-$ final state was dominated by even angular momenta.
86:
87: The question of genuine three-body decays of the $B$ meson (in contrast to
88: quasi-two-body decays which involve resonances between two of the three
89: bodies) arises in part because of the need to parametrize CP violation in $B
90: \to K \bar K K$ and to understand nonresonant contributions arising in Dalitz
91: plot studies. These contributions can be quite large, as measured in Dalitz
92: plot analyses of $B^+\to K^+K^-K^+$~\cite{Garmash:2004wa} and $B^0 \to K^+ K^-
93: K_S$~\cite{Aubert:2005kd}. They seem to be less significant in comparison with
94: quasi-two-body final states in certain $B \to K \pi \pi$
95: decays~\cite{Aubert:2004fn,Group(HFAG):2005rb}.
96:
97: In the present paper we discuss conclusions that can be drawn regarding the
98: structure of the three-body final states for $B \to PPP$ strangeness-changing
99: decays, where $P$ denotes a light pseudoscalar meson. We begin by noting some
100: relations due to isospin in the limit in which decays are dominated by a QCD
101: penguin amplitude with isospin-preserving flavor structure.
102: Smaller $\Delta I = 1$ tree and electroweak penguin amplitudes will be
103: neglected.
104: We also analyze amplitudes
105: for a nonresonant background using isospin symmetry and flavor SU(3). The
106: description of $B$ decays to a pair of charmless mesons in terms of flavor
107: SU(3) amplitudes \cite{Gronau:1994rj,Gronau:1995hn} has been able to
108: correlate decay rates and CP-violating asymmetries for a wide variety
109: of processes involving two light pseudoscalars $P$ \cite{Chiang:2004nm}
110: or one pseudoscalar and one vector ($V$) meson \cite{Chiang:2003pm}.
111:
112: We restrict our treatment for the moment in several respects. (1) We
113: consider only $|\Delta S| = 1$ transitions and assume them to be dominated
114: by a penguin amplitude with flavor structure $\bar b \to \bar s$. (2)
115: We consider only final states involving pions and kaons, in order not to
116: have to contend with octet-singlet mixing questions or posssible additional
117: flavor-singlet penguin amplitudes \cite{DGR}. (3) We do not consider $B_s$
118: mesons, since information on them has lagged considerably behind that on
119: $B^+$ and $B^0$.
120:
121: Earlier treatments of $B \to PPP$ decays, including model-dependent hadronic
122: calculations of decay rates and CP asymmetries, may be found in
123: Refs.~\cite{Eilam:1995nz,Bajc:1998bs,Cheng:2002qu,Minkowski:2004xf}.
124: Several analyses using flavor SU(3) have been performed in
125: Refs.~\cite{Gronau:2003ep,Grossman:2003qp},
126: in order to obtain model-independent bounds on deviations from the dominance
127: of a single weak phase in $B$ decays to three kaons.
128:
129: Section II reviews what is known about $B \to K \pi \pi$ and $B \to K \bar K K$
130: decay rates, pointing out certain features of resonant and nonresonant
131: contributions. Numerous isospin relations are proven for $B\to K\pi\pi$ and for
132: corresponding quasi two-body decays and are tested in Section III. Similar
133: relations hold for $B\to K\bar K K$. Assuming symmetry under the interchange
134: of $K$ and $\bar K$ momenta, as observed in the data, we prove decay rate
135: relations for processes involving $K_S$ and $K_L$. Section IV compares
136: nonresonant background amplitudes in $B \to K \pi \pi$ and $B \to K \bar K K$
137: processes using isospin symmetry and flavor SU(3) in a simple universal model.
138: Implications for Dalitz plot analyses are noted in Section V, while Section VI
139: summarizes, concluding with a few remarks about isospin-violating corrections
140: and direct CP asymmetries.
141: \bigskip
142:
143: \centerline{\bf II. EXPERIMENTAL STATUS}
144: \bigskip
145:
146: % This is Table I
147: \begin{table}
148: \caption{Summary of CP-averaged branching ratios, in units of $10^{-6}$, for
149: $B \to K \pi \pi$ and $B \to K \bar K K$ including quasi two-body
150: decays~\cite{Group(HFAG):2005rb,Abe:2005kr}.
151: Pairs of processes in the same row are related by isospin reflection,
152: except for $B^+\to \phi K^+$ and $B^0\to \phi K^0$ which are in different rows.
153: \label{tab:brs}}
154: \begin{center}
155: \begin{tabular}{l c c c} \hline \hline
156: Final state & Branching ratio & Final state & Branching ratio \\
157: in $B^+$ decay & $\times 10^{-6}$ & in $B^0$ decay & $\times 10^{-6}$ \\
158: \hline \hline
159: $K^+ \pi^+ \pi^-$ & $54.1\pm3.1$ & $K^0 \pi^+ \pi^-$ & $44.9\pm2.6$ \\
160: $K^{*0}\pi^+$ & $10.8 \pm 0.8$ & $K^{*+}\pi^-$ & $9.8 \pm 1.1$ \\
161: $K^+\rho^0$ & $4.23^{+0.56}_{-0.57}$ & $K^0\rho^0$ & $ 5.6 \pm 1.1$ \\
162: $K^+f_0(980)$ & $9.1^{+0.8}_{-1.1}~^a$ & $K^0f_0(980)$ & $6.0 \pm 0.9~^a$ \\
163: $K^*_0(1430)^0\pi^+$ & $38.2^{+4.6}_{-4.5}$ & $K^*_0(1430)^+\pi^-$ &
164: $45.1 \pm 6.1$ \\ \hline
165: $K^0 \pi^+ \pi^0$ & $< 66$ & $K^+ \pi^- \pi^0$ & $35.6^{+3.4}_{-3.3}$ \\
166: $K^{*+}\pi^0$ & $6.9 \pm 2.3$ & $K^{*0}\pi^0$ & $1.7 \pm 0.8$ \\
167: $K^0\rho^+$ & $ < 48$ & $K^+\rho^-$ & $9.9^{+1.6}_{-1.5}$ \\
168: $K^*_0(1430)^+\pi^0$ & -- & $K^*_0(1430)^0\pi^0$ & $7.9 \pm 3.1~^b$ \\ \hline
169: $K^+ \pi^0 \pi^0$ & -- & $K^0 \pi^0 \pi^0$ & -- \\ \hline
170: $K_S K_S K^+$ & $11.5\pm1.3$ & $K^+ K^- K^0$ & $24.7\pm2.3$ \\
171: ~& ~& $\phi K^0$ & $8.3^{+1.2}_{-1.0}$ \\ \hline
172: $K^+ K^- K^+$ & $30.1\pm1.9$ & $K_S K_S K_S$ & $6.2\pm0.9$ \\
173: $\phi K^+$ & $9.03^{+0.65}_{-0.63}$ & ~ &
174: \\ \hline \hline
175: \end{tabular}
176: \end{center}
177: \leftline{$^a$ Includes $\b(f_0(980) \to \pi^+\pi^-)$.}
178: \leftline{$^b$ Includes $\b(K^*_0(1430)^0 \to K^+\pi^-)$.}
179:
180: \end{table}
181:
182: The current world averages of CP-averaged branching ratios from BaBar, Belle,
183: and CLEO for the decays $B \to K \pi \pi$ and $B \to K \bar K K$ are summarized
184: in Table \ref{tab:brs} \cite{Group(HFAG):2005rb}. Averages involving $B^0 \to
185: K_S \pi^+ \pi^-$ and its sub-modes
186: include recent Belle results \cite{Abe:2005kr}. Also listed
187: are branching ratios for quasi-two-body decays for several resonances
188: contributing to these decays.
189: While no measurement exists so far for the branching ratio of
190: $B^0\to K^0\pi^0\pi^0$, a time-dependent CP asymmetry has been
191: recently reported in this process~\cite{Aubert:2005id}.
192:
193: The branching ratios quoted in Table \ref{tab:brs} for $B^{+,0}\to K^{+,0}
194: f_0(980)$
195: and $B^0\to K^*_0(1430)^0\pi^0$ include
196: decay branching ratios of the daughter scalar mesons into observed modes.
197: Using
198: $\b(K^*_0(1430)^0 \to K^+\pi^-) = 2/3$ we obtain
199: \beq
200: \b(B^0\to K^*_0(1430)^0\pi^0) = 11.9 \pm 4.7~~,
201: \eeq
202: where branching ratios here and subsequently are quoted in units of $10^{-6}$.
203:
204: Dalitz plot analyses of $B^0 \to K^+ K^- K^0$~\cite{Aubert:2005kd} and
205: $B^+ \to K^+ K^- K^+$~\cite{Garmash:2004wa} find large nonresonant
206: contributions in these decays. In addition to the $\phi K$
207: mode, where the $K^+$ and $K^-$ are in a P-wave, two sizable and comparable
208: contributions have been measured: A term peaking around $1500~{\rm MeV}/c^2$,
209: for which one finds in addition to a large solution also a small
210: solution~\cite{Aubert:2005kd,Garmash:2004wa}
211: (see Table I in Ref.~\cite{Aubert:2005kd} and Table V in Ref.~\cite{Garmash:2004wa}),
212: and a term spreading across phase space. Both terms have an S-wave behavior in
213: the $K^+$ and $K^-$ momenta~\cite{Aubert:2005kd,Garmash:2004wa}.
214: Contributions from higher waves were found consistent with zero. The decays
215: $B^{+,0}\to \chi_{c0}K^{+,0}$, also having an S-wave behavior, contribute
216: about three percent of the total branching ratios of $B^{+,0}\to
217: K^+K^-K^{+,0}$~\cite{Aubert:2005kd,Garmash:2004wa}. Ref.~\cite{Garmash:2004wa}
218: finds a second solution of about eight percent for the fraction corresponding
219: to $B^+\to \chi_{c0}K^+$
220: (see Table V in~\cite{Garmash:2004wa}).
221: All the above three S-wave contributions are symmetric under interchanging
222: $K^+$ and $K^-$.
223:
224: It is useful to subtract contributions for $B \to \phi K$ from the branching
225: ratios of $B^+ \to K^+ K^- K^+$ and $B^0 \to K^+K^- K^0$. Using values in Table
226: I and~\cite{Eidelman:2004wy} $\b(\phi\to K^+ K^-) = (49.1\pm0.6)\%$,
227: one finds
228: \bea\label{B+}
229: \b(B^+\to K^+K^-K^+)_{\phi K-{\rm subtracted}} & = & 25.7 \pm 1.9~~,\\
230: \label{B0}
231: \b(B^0\to K^+K^-K^0)_{\phi K-{\rm subtracted}} & = & 20.6 \pm 2.4~~.
232: \eea
233: An important feature of the amplitudes corresponding to these branching ratios
234: is their symmetry with respect to interchanging $K^+$ and $K^-$ momenta,
235: as they are superpositions of three S-wave
236: contributions~\cite{Aubert:2005kd,Garmash:2004wa}.
237:
238: % \bigskip
239: \newpage
240:
241: \centerline{\bf III. ISOSPIN RELATIONS}
242: \bigskip
243:
244: We assume that the dominant transition for $|\Delta S| = 1$ $B \to PPP$ decays
245: has a flavor structure $\bar b \to \bar s$, which is isospin-invariant
246: ($\Delta I = 0$). Using isospin reflection symmetry under $u \lra d$,
247: we then find that each $B^+$ decay amplitude listed in Table \ref{tab:brs}
248: is equal (up to a possible sign) to a corresponding $B^0$ decay amplitude
249: listed in the same line. Other amplitude relations follow from our assumption
250: that the final state is dominantly $I=1/2$.
251: In order to relate predictions of equal $B^+$ and $B^0$ partial widths to
252: observed branching ratios, we use the measured ratio of $B^+$ and $B^0$
253: lifetimes, $\tau_+/\tau_0 = 1.076\pm 0.008$ \cite{Group(HFAG):2005rb}.
254:
255: Relations between observed branching ratios for $B\to K\pi\pi$ and
256: corresponding quasi two-body decays follow directly from the above assumption.
257: Similar amplitude relations hold for $B\to K\bar KK$. However, in order to
258: rewrite these relations for decay rates involving $K_S$ and $K_L$ in the final
259: state one must assume a given symmetry under interchanging
260: $K$ and $\bar K$ momenta. We will use the symmetry under $K^+ \leftrightarrow
261: K^-$ of the amplitudes describing the $\phi K$-subtracted branching ratios
262: (\ref{B+}) and (\ref{B0}).
263:
264: \bigskip
265:
266: \leftline{\bf A. $B \to K \pi \pi$}
267: \bigskip
268:
269: A relation which is well-satisfied is
270: \bea
271: \b(B^+ \to K^+ \pi^+ \pi^-)&=&(\tau_+/\tau_0)\b(B^0 \to K^0 \pi^+ \pi^-)~~; \\
272: 54.1 \pm 3.1 &=& 48.3 \pm 2.8 ~~.\nonumber
273: \eea
274: The discrepancy is only $1.4 \sigma$.
275:
276: The above isospin relation should apply to corresponding quasi-two-body
277: modes contributing to these decays. Thus, the following four relations
278: hold reasonably well:
279: \bea
280: \b(B^+\to K^{*0}\pi^+) & = & (\tau_+/\tau_0)\b(B^0\to K^{*+}\pi^-)~~;\\
281: 10.8 \pm 0.8 & = & 10.6 \pm 1.2~~,\nonumber
282: \eea
283: \bea
284: \b(B^+\to K^+\rho^0) & = & (\tau_+/\tau_0)\b(B^0\to K^0\rho^0)~~;\\
285: 4.23^{+0.56}_{-0.57} & = & 6.1 \pm 1.2~~,\nonumber
286: \eea
287: \bea
288: \b(B^+\to K^+f_0(980))\b(f_0\to\pi^+\pi^-) & = &
289: (\tau_+/\tau_0)\b(B^0\to K^0f_0(980))\b(f_0\to\pi^+\pi^-);\\
290: 9.07^{+0.81}_{-1.06} & = & 6.4 \pm 0.9~~,\nonumber
291: \eea
292: \bea\label{1430-1}
293: \b(B^+\to K^*_0(1430)^0\pi^+)
294: & = & (\tau_+/\tau_0)\b(B^0\to K^*_0(1430)^+\pi^-)~~;
295: \\
296: 38.2^{+4.6}_{-4.5} & = & 48.6 \pm 6.6~~.\nonumber
297: \eea
298: The last relation disfavors a second solution, $\b(B^+\to K^*_0(1430)^0\pi^+)=
299: 8.7 \pm 2.3$ measured in~\cite{Garmash:2004wa}
300: (see Table IV there), and is in agreement with a more recent measurement
301: (see Table V in~\cite{Abe:2005ig}), $\b(B^+\to K^*_0(1430)^0\pi^+) =
302: 51.6 \pm 1.7 \pm 6.8^{+1.8}_{-3.1}$.
303:
304: A prediction satisfied only by an upper bound is
305: \bea
306: \b(B^+ \to K^0 \pi^+ \pi^0)
307: &=& (\tau_+/\tau_0)\b(B^0 \to K^+ \pi^- \pi^0)~~; \\
308: < 66 & = & 38.3^{+3.7}_{-3.6}~~.\nonumber
309: \eea
310: It should not be too difficult to obtain a value for the left-hand side; a
311: $\pi^0$ must be added to the observed final state $B^+ \to K^0 \pi^+$.
312: Corresponding predictions apply to quasi two-body decays. The
313: prediction
314: \bea\label{K*pi1}
315: \b(B^0 \to K^{*0}\pi^0) & = & (\tau_0/\tau_+)\b(B^+\to K^{*+}\pi^0)~~; \\
316: 1.7\pm 0.8 & = & 6.4 \pm 2.1~~,\nonumber
317: \eea
318: requires more data for a statistically significant test.
319:
320: Dominance of $I=1/2$ in $K^*\pi$ final states implies
321: \bea\label{K*pi2}
322: 2\b(B^0\to K^{*0}\pi^0) & = & \b(B^0\to K^{*+}\pi^-)~~;\\
323: 3.4 \pm 1.6 & = & 9.8 \pm 1.1~~,\nonumber
324: \eea
325: which is violated by $3.3\sigma$.
326: The smallness of the left-hand side may be due to its sensitivity to small
327: $\Delta I = 1$ contributions, which are present in the treatment of Ref.\
328: \cite{Chiang:2003pm}.
329: The small branching ratio measured for $B^0\to K^{*0}\pi^0$ is evidently the
330: origin of the apparent discrepancies in Eqs.~(\ref{K*pi1}) and (\ref{K*pi2}).
331:
332: The prediction
333: \bea
334: \b(B^+\to K^0\rho^+) & = & (\tau_+/\tau_0)\b(B^0\to K^+\rho^-)~~;\\
335: < 48 & = & 10.7^{+1.7}_{-1.6}~~,\nonumber
336: \eea
337: is satisfied by the upper bound,
338: while $I(K\rho)=1/2$ implies
339: \bea
340: 2\b(B^0\to K^0\rho^0) & = & \b(B^0\to K^+\rho^-)~~; \\
341: 11.3 \pm 2.2 & = & 9.9^{+1.6}_{-1.5}~~.\nonumber
342: \eea
343:
344: Similarly,
345: \beq
346: \b(B^+\to K^*_0(1430)^+\pi^0) = (\tau_+/\tau_0)\b(B^0\to K^*_0(1430)^0\pi^0) =
347: 12.8 \pm 5.0
348: \eeq
349: awaits a measurement of the left-hand-side, while $I(K^*_0(1430)\pi) = 1/2$
350: implies
351: \bea\label{1430-2}
352: 2\b(B^0\to K^*_0(1430)^0\pi^0) & = & \b(B^0\to K^*_0(1430)^+\pi^-)~~;\\
353: 23.7 \pm 9.3 & = & 45.1 \pm 6.1~~.\nonumber
354: \eea
355:
356: Finally, neither the left-hand nor right-hand side of the following prediction
357: corresponds to a current observation:
358: \beq
359: \b(B^+ \to K^+ \pi^0 \pi^0) = (\tau_+/\tau_0)\b(B^0 \to K^0 \pi^0 \pi^0)~~.
360: \eeq
361:
362: \bigskip\bigskip
363:
364: \leftline{\bf B. $B \to K \bar K K$}
365: \bigskip
366:
367: Isospin reflection symmetry implies
368: \bea\label{00+}
369: A(B^+\to K^0 \ok K^+) & = & -A(B^0\to K^+K^-K^0)~~,\\
370: \label{+-+}
371: A(B^+\to K^+K^-K^+) & = & -A(B^0\to K^0\ok K^0)~~.
372: \eea
373: In order to study $\phi K$-subtracted amplitudes, we will use their observed
374: symmetry under interchanging the $K$ and $\bar K$
375: momenta mentioned at the end of Section II~\cite{Aubert:2005kd,Garmash:2004wa}.
376: This permits writing relations for rates involving $K_S$ and $K_L$ in the final state.
377: Note that because of Bose symmetry the amplitudes in (\ref{+-+}), which involve two
378: identical $K$ mesons, are also symmetric in the two $K$ momenta.
379:
380: Using the phase convention (we neglect a tiny CP violation in $K^0$--$\ok$ mixing),
381: \beq
382: K_S \equiv (K^0 + \ok)/\s~~,~~~~~~~K_L \equiv (K^0 - \ok)/\s~~,
383: \eeq
384: a symmetric state
385: \beq
386: \bra{K^0 \ok}_{\rm sym} \equiv \left[ \bra{K^0(p_1) \ok(p_2)}
387: + \bra{\ok(p_1) K^0(p_2)} \right]/ \s~~,
388: \eeq
389: can be expressed as
390: \beq\label{KKbarSym}
391: \bra{K^0 \ok}_{\rm sym} = (\bra{K_S(p_1) K_S(p_2)}
392: - \bra{K_L(p_1) K_L(p_2)})/\s~~,
393: \eeq
394: while an antisymmetric state is given by
395: \bea
396: \bra{K^0 \ok}_{\rm anti} & \equiv & \left[ \bra{K^0(p_1) \ok(p_2)}
397: - \bra{\ok(p_1) K^0(p_2)} \right] \s\nonumber\\
398: & = &\left[ \bra{K_L(p_1) K_S(p_2)} - \bra{K_S(p_1) K_L(p_2)} \right] /\s~~.
399: \eea
400: Similarly, a state symmetric in the three momenta, $p_1, p_2, p_3$, is given by
401: \bea\label{3K}
402: & & \bra{K^0 \ok K^0}_{\rm sym} \equiv \nonumber \\
403: & & \left[ \bra{K^0 K^0 \ok}
404: + \bra{K^0 \ok K^0} + \bra{\ok K^0 K^0}\right]/ \st \\
405: & & = \frac{1}{2\s}\left [ \st \bra{K_S K_S K_S} +
406: \bra{K_S K_S K_L}_{\rm sym}
407: - \bra{K_L K_L K_S}_{\rm sym}
408: - \st \bra{K_L K_L K_L}\right ] \nonumber~~,
409: \eea
410: where dependence on the three momenta has been suppressed.
411:
412: Using Eq.~(\ref{KKbarSym}), we find
413: \bea
414: \b(B^+ \to (K^0 \ok)_{\rm sym}K^+) & = & \b(B^+ \to K_S K_S K^+) +
415: \b(B^+ \to K_L K_L K^+)\nonumber \\
416: & = & 2\b(B^+ \to K_S K_S K^+) = 23.0 \pm 2.6~~.
417: \eea
418: Eq.~(\ref{00+}) is well-satisfied for the $\phi K$-subtracted branching ratio
419: \bea\label{1KKK}
420: \b(B^+ \to (K^0 \ok)_{\rm sym}K^+) & = & (\tau_+/\tau_0) \b(B^0 \to
421: K^+K^-K^0)_{\phi K-{\rm subtracted}}~~;\\
422: 23.0 \pm 2.6 & = & 22.2 \pm 2.6~~.\nonumber
423: \eea
424:
425: Finally, in order to test the prediction (\ref{+-+}) we apply (\ref{3K})
426: \beq
427: \b(B^0 \to (K^0\ok K^0)_{\rm sym}) = \frac{8}{3}\b(B^0 \to K_SK_SK_S)
428: = 16.5 \pm 2.4~~.
429: \eeq
430: Eq.~(\ref{+-+}) then reads
431: \bea\label{2KKK}
432: \b(B^+\to K^+K^-K^+)_{\phi K-{\rm subtracted}} & = & (\tau_+/\tau_0) \b(B^0\to
433: (K^0 \ok K^0)_{\rm sym})~~;
434: \\
435: 25.7 \pm 1.9 & = & 17.8 \pm 2.6~~,\nonumber
436: \eea
437: which holds within 2.4$\sigma$.
438: A potential discrepancy may be accounted for by small $\Delta I=1$ amplitudes.
439:
440: Eq.~(\ref{3K}) also implies predictions for branching ratios involving $K_L$ in
441: the final state:
442: \bea
443: \b(B^0\to K_L K_L K_L) & = & \b(B^0 \to K_S K_S K_S) = \nonumber\\
444: 3\b(B^0\to K_L K_L K_S) & = & 3\b(B^0\to K_S K_S K_L)~~,
445: \eea
446: where subtraction of $\phi K$ contributions in the last two processes is
447: implied. We expect $\b(B^0\to K_S K_S K_L)$ to be easier to measure in
448: comparison with $\b(B^0\to K_L K_L K_S)$ and $\b(B^0\to K_L K_L K_L)$.
449:
450: The agreement in (\ref{1KKK}) and (\ref{2KKK}) supports the initial
451: suggestion~\cite{Garmash:2003er} that the $K^+ $ and $K^-$ in the respective
452: processes are in dominantly symmetric even angular momentum (S-wave) states,
453: as confirmed directly by measuring angular dependence in later experiments
454: performing full Dalitz plot analyses~\cite{Aubert:2005kd,Garmash:2004wa}.
455: A statistically significant discrepancy in Eq.~(\ref{2KKK}), implied by reduced
456: experimental errors, would provide evidence for nonzero contributions either
457: from odd angular momentum states or from a $\Delta I=1$ amplitude.
458: \bigskip
459:
460: %\newpage
461: \centerline{\bf IV. MODEL FOR A NONRESONANT BACKGROUND}
462: \bigskip
463: The measured $\phi K$-subtracted rates for $B^+\to K^+K^-K^+$
464: \cite{Garmash:2004wa} and $B^0 \to K^+K^-K^0$~\cite{Aubert:2005kd} consist each
465: of a sum of three contributions, all symmetric in the $K^+$ and $K^-$ momenta:
466: A small $\chi_{c0} K$ term, an S-wave contribution peaking around 1500
467: MeV/$c^2$, and a nonresonant background amplitude also representing an S-wave
468: in $K^+K^-$. The latter amplitude shows no significant dependence on the
469: $K^+K^+$ and $K^+K^0$ invariant masses in the two processes. Some dependence
470: on the $K^+K^-$ invariant mass is observed in $B^0\to K^+K^-K^0$ but not in
471: $B^+\to K^+K^-K^+$. In a similar analysis of the nonresonant background in
472: $B^+\to K^+\pi^+\pi^-$ some dependence was measured on the invariant masses of
473: $K^+\pi^-$ and of $\pi^+\pi^-$.
474:
475: % This is Table II
476: \begin{table}
477: \caption{Branching ratios of nonresonant background (NRB) contributions
478: for $B\to K\pi\pi$ and $B\to K\bar K K$, given as fractions
479: of total branching ratios and in units of $10^{-6}$.
480: \label{tab:NRB}}
481: \begin{center}
482: \begin{tabular}{l ccc} \hline \hline
483: Decay mode & NRB fraction ($\%$) & NRB branching ratio
484: ($\times 10^{-6}$) \\ \hline \hline
485: $B^+\to K^+\pi^-\pi^+$ & $34.0 \pm 2.2^{+2.1}_{-1.8}$~\cite{Abe:2005ig} &
486: $18.4 \pm 1.9$~$^a$ \\
487: $B^0 \to K^+\pi^-\pi^0$ & ~ & $<4.6$~\cite{Group(HFAG):2005rb} \\
488: $B^+\to K^+K^-K^+$ & $74.8 \pm 3.6$~$^b$~\cite{Garmash:2004wa}
489: & $22.5 \pm 1.8$ \\
490: $B^0 \to K^+K^-K^0$ & $70.7 \pm 3.8 \pm 1.7$~\cite{Aubert:2005kd}
491: & $17.5 \pm 1.9$ \\
492: \hline \hline
493: \end{tabular}
494: \end{center}
495: \leftline{$^a$ A much smaller nonresonant branching ratio,
496: $(2.9^{+1.1}_{-0.9})\times 10^{-6}$,
497: is quoted in~\cite{Group(HFAG):2005rb} .}
498: \leftline{$^b$ A second solution, $(65.1 \pm 5.1)\%$, is obtained
499: in~\cite{Garmash:2004wa}.}
500: \end{table}
501:
502: In the present section we will study nonresonant background amplitudes in
503: $B\to K\pi\pi$ and $B\to K\bar K K$ decays, adopting a simplified assumption
504: that these amplitudes are symmetric under interchanging the three final meson
505: momenta. This would be the case, for instance if the nonresonant amplitudes
506: were constant over the Dalitz plane; however, these amplitudes do not have to
507: be constant. We start by first presenting the data and
508: then discussing symmetry relations governing nonresonant contributions.
509:
510: Table \ref{tab:NRB} quotes measured fractions of nonresonant background (NRB)
511: contributions in $B\to K\pi\pi$ and $B\to K\bar KK$ processses. While a small
512: NRB contribution, $\b({\rm NRB})=(2.9^{+1.1}_{-0.9})\times 10^{-6}$, has been
513: measured in~\cite{Aubert:2004fn} and is quoted in~\cite{Group(HFAG):2005rb},
514: we quote in the Table a larger nonresonant fraction ($\sim 1/3$) which has been
515: measured recently by Belle~\cite{Abe:2005ig}
516: (see also Ref.~\cite{Garmash:2004wa}). As we will see, a large
517: nonresonant background in $B^+\to K^+\pi^+\pi^-$ appears to be more consistent
518: in our scheme with comparable large nonresonant contributions measured in
519: $B^+\to K^+K^-K^+$~\cite{Garmash:2004wa} and
520: $B^0\to K^+K^-K^0$~\cite{Aubert:2005kd}.
521: Two possible solutions for the fraction of a nonresonant background were
522: obtained in the first process, $(74.8 \pm 3.6)\%$ and $(65.1 \pm 5.1)\%$.
523: We quote the former value, which corresponds to a fit with lower $\chi^2$ (see
524: Table V of Ref.\ \cite{Garmash:2004wa}). These fractions and the total
525: branching ratios
526: given in Table I were used to calculate the nonresonant branching ratios.
527:
528: \begin{figure}
529: \begin{center}
530: \includegraphics
531: [height=3in]{peng3p.ps}
532: \end{center}
533: \caption{Graphs describing nonresonant background in penguin-dominated
534: $B \to P_1 P_2 P_3$ decays. The cross denotes a $\bar b \to
535: \bar s$ flavor transition. Gluons or quarks associated with the penguin
536: operator are not shown explicitly. Here $q$ denotes $u$ for a $B^+$ or
537: $d$ for a $B^0$.
538: \label{fig:peng3p}}
539: \end{figure}
540:
541: A model describing nonresonant background amplitudes in $b\to s$ dominated
542: $B\to PPP$ decays is shown in Fig.\ \ref{fig:peng3p}. The amplitudes may be
543: categorized by whether the quark pairs $q_i \bar q_i~(i=1,2)$ shown in Fig.\
544: \ref{fig:peng3p} are $u \bar u, d \bar d$ or $s \bar s$. Isospin symmetry is
545: implied by associating equal amplitudes with $u \bar u$ and $d\bar d$. This
546: symmetry assumption may be extended to flavor SU(3) by associating the same
547: amplitude with $s \bar s$. Broken SU(3) may be represented by using a
548: smaller amplitude for an $s\bar s$ pair.
549:
550: \begin{table}
551: \caption{Nonresonant background amplitudes for $B \to PPP$ decays as a
552: function of quark pairs $q_1 \bar q_1$ and $q_2 \bar q_2$.
553: \label{tab:amps}}
554: \begin{center}
555: \begin{tabular}{l c c c c} \hline \hline
556: Decaying & $q_1 \bar q_1$ & $q_2 \bar q_2$ & Final & Coefficient \\
557: $~~~~~B$ & & & state & of amplitude \\ \hline
558: $B^+$ & $u \bar u$ & $u \bar u$ & $K^+ \pi^0 \pi^0$ & $1/\s$ \\
559: $=\bar b u$ & & $d \bar d$ & $K^+ \pi^- \pi^+$ & $-1$ \\
560: & & $s \bar s$ & $K^+ K^- K^+$ & $-\s$ \\
561: & $d \bar d$ & $u \bar u$ & $K^0 \pi^+ \pi^0$ & $-1/\s$ \\
562: & & $d \bar d$ & $K^0 \pi^0 \pi^+$ & $1/\s$ \\
563: & & $s \bar s$ & $K^0 \bar K^0 K^+$ & 1 \\ \hline
564: $B^0$ & $u \bar u$ & $u \bar u$ & $K^+ \pi^0 \pi^-$ & $1/\s$ \\
565: $=\bar b d$ & & $d \bar d$ & $K^+ \pi^- \pi^0$ & $-1/\s$ \\
566: & & $s \bar s$ & $K^+ K^- K^0$ & $-1$ \\
567: & $d \bar d$ & $u \bar u$ & $K^0 \pi^+ \pi^-$ & $-1$ \\
568: & & $d \bar d$ & $K^0 \pi^0 \pi^0$ & $1/\s$ \\
569: & & $s \bar s$ & $K^0 \bar K^0 K^0$ & $\s$ \\ \hline \hline
570: \end{tabular}
571: \end{center}
572: \end{table}
573:
574: Table \ref{tab:amps} gives the contributions to the various processes in terms
575: of their coefficients. We use conventions for states defined in Refs.\
576: \cite{Gronau:1994rj} and~\cite{Gronau:1995hn}. Quark model assignments include
577: $B^+ = u \bar b,~B^0 = d \bar b$, with states containing a $\bar u$ quark
578: defined with a minus sign for convenience in isospin calculations. Thus, a
579: neutral pion is $\pi^0 = (d \bar d - u \bar u)\sqrt{2}$. The entries in Table
580: \ref{tab:amps} contain factors of $2 \cdot 1/\s = \s$ for identical particles.
581:
582: The coefficients in Table \ref{tab:amps} imply symmetry relations between decay
583: rates contributed by a nonresonant background in different processes. For
584: instance, the nonresonant branching ratio in $B^+\to K^+\pi^0\pi^0$ is
585: predicted to be half of that measured in $B^+\to K^+\pi^+\pi^-$. Relations
586: applying separately to $B\to K\pi\pi$ and $B\to K\bar K K$ decays follow
587: from isospin symmetry and are generally expected to hold in our model more
588: precisely than relations between these two types of processes which assume
589: flavor SU(3). Let us discuss some of these relations which can be tested using
590: current measurements.
591:
592: An interesting prediction follows from the two equal and opposite amplitudes
593: present in $B^+ \to K^0 \pi^+ \pi^0$. When added together, the two
594: contributions cancel. This is a key test of the S-wave nature (or any even
595: angular momentum) of the $\pi \pi$
596: system for the nonresonant amplitude. An S-wave $\pi \pi$ system with charge
597: $\pm 1$ must be in a state of $I=2$, which cannot be reached with the penguin
598: transition illustrated here. Thus the nonresonant contributions to $B^+ \to
599: K^0 \pi^+ \pi^0$ and $B^0 \to K^+ \pi^- \pi^0$ are predicted to vanish if our
600: assumptions are valid. The current upper bound of $4.6\times 10^{-6}$ on
601: the nonresonant branching ratio of the second process is indeed much
602: smaller than the other nonresonant branching ratios quoted in Table
603: \ref{tab:NRB}.
604:
605: Table \ref{tab:amps} predicts that the nonresonant decay width for $B^+\to
606: K^+K^-K^+$ is two times larger than that for $B^0\to K^+K^-K^0$. This relation
607: does not hold so well
608: (we will comment on a probable reason in the next section),
609: \bea\label{KKK}
610: \b(B^+\to K^+K^-K^+)_{\rm NRB} & = & 2(\tau_+/\tau_0)\b(B^0\to
611: K^+K^-K^0)_{\rm NRB}~~; \\
612: 22.5 \pm 1.8 & = & 37.7 \pm 4.1~~,\nonumber
613: \eea
614: where the NRB branching ratios here and subsequently are taken from
615: Table \ref{tab:NRB}.
616: This relation tests the assumption that on the right-hand-side the $K^+$ and
617: $K^0$ in the nonresonant background are in a symmetric $I=1$ state. In this
618: case the two processes involve a single isospin amplitude~\cite{Gronau:2003ep},
619: and their ratio of rates is given by the squared ratio of corresponding
620: Clebsch-Gordan coefficients, $(\s)^2 =2$. The assumption of a $K^+K^0$
621: symmetric state stands in contrast to the dependence on the $K^+K^-$ invariant
622: mass observed in the nonresonant background for $B^0\to
623: K^+K^-K^0$~\cite{Aubert:2005kd}. No such dependence was observed in the
624: process on the left-hand-side of (\ref{KKK}).
625:
626: In the SU(3) limit, one may relate the nonresonant background in the above
627: processes and the nonresonant amplitude in $B^+\to K^+\pi^+\pi^-$. Comparing
628: with $B^0\to K^+K^-K^0$, one expects
629: \bea\label{Kpp}
630: \b(B^+\to K^+\pi^+\pi^-)_{\rm NRB} & = & (\tau_+/\tau_0)\b(B^0\to
631: K^+K^-K^0)_{\rm NRB}~~;
632: \\
633: 18.4 \pm 1.9 & = & 18.8 \pm 2.0~~,\nonumber
634: \eea
635: which holds very well. Under the underlying SU(3) approximation for nonresonant
636: amplitudes, one would have expected this relation to be less precise than
637: (\ref{KKK}). Note that in both processes appearing in (\ref{Kpp}) the measured
638: nonresonnt background is not exactly symmetric under interchanging the three
639: meson momenta, as assumed in our model.
640: The agreement in (\ref{Kpp}) favors the large nonresonant background in
641: $B^+\to K^+\pi^+\pi^-$ given in Table \ref{tab:NRB} over the small value quoted
642: in Ref.~\cite{Group(HFAG):2005rb}.
643:
644: %\bigskip
645: \newpage
646:
647: \centerline{\bf V. IMPLICATIONS FOR DALITZ PLOT ANALYSES}
648: \bigskip
649:
650: The isospin relations we have quoted in the Section III are expected to be
651: valid separately for resonant and nonresonant contributions. Thus, the amount
652: of $K^*_0(1430)^0\pi^+$ in $B^+ \to K^+ \pi^+ \pi^-$, for which a two-fold
653: ambiguity appears in the analysis of Ref.~\cite{Garmash:2004wa} (see Table IV
654: there), should be the same as the amount of $K^*_0(1430)^+\pi^-$ measured in
655: $B^0 \to K^0 \pi^+ \pi^-$ [see Eq.~(\ref{1430-1})], and should be related to
656: the amount of $K^*_0(1430)^0\pi^0$ measured in $B^0 \to K^+ \pi^- \pi^0$ [see
657: Eq.~(\ref{1430-2})]. Similarly, the amount of nonresonant background in $B^+
658: \to K^+ \pi^+ \pi^-$, for which there seems to be some question in comparing
659: Refs.\ \cite{Group(HFAG):2005rb} and \cite{Abe:2005ig}, should be the same as
660: in $B^0 \to K^0 \pi^+ \pi^-$ for which no value has been quoted yet.
661: Also, if a nonresonant background is small in $B^0 \to K^+ \pi^0 \pi^-$,
662: as shown in Table \ref{tab:NRB}, we would also expect it to be small in
663: $B^+ \to K^0 \pi^+ \pi^0$.
664:
665: The isospin relation $\Gamma(B^+ \to K^+ \phi) = \Gamma(B^0 \to K^0 \phi)$
666: appears to be satisfied by present data. Thus, we expect the non-$\phi$
667: contributions in $B \to K \bar K K$ decays, which are related to one another by
668: isospin reflection, also to have equal partial widths. This is confirmed by
669: Eqs.~(\ref{1KKK}) and (\ref{2KKK}).
670:
671: Fits to Dalitz plots often involve discrete ambiguities in assigning
672: amplitudes and phases to given decay channels. This is demonstrated, for
673: instance, by the two largely different solutions for $\b(B^+ \to K^*_0(1430)^0
674: \pi^+)$ measured in Ref.~\cite{Garmash:2004wa}. Isospin symmetry, which relates
675: this process to $B^0\to K^*_0(1430)^+\pi^-$, resolves this
676: ambiguity. Similarly, the isospin relation $\Gamma(B^+\to \chi_{c0}K^+) =
677: \Gamma(B^0\to \chi_{c0}K^0)$ is useful for eliminating a two-fold ambiguity in
678: the measurement of the left-hand-side~\cite{Garmash:2004wa}. This determines
679: the fraction of $\chi_{c0}K^+$ in the total of all $B^+\to K^+K^-K^+$ decays to
680: be about three percent rather than about eight percent, both solutions being
681: permitted in~\cite{Garmash:2004wa}.
682:
683: One of the predictions of fully symmetric final states in nonresonant
684: background amplitudes is that these amplitudes should be suppressed in
685: $B^+ \to K^0 \pi^+ \pi^0$ and $B^0 \to K^+ \pi^0 \pi^-$ relative to other
686: processes under discussion. This prediction is supported by the upper bound on
687: $\b(B^0\to K^+\pi^-\pi^0)_{\rm nonres}$ in Table \ref{tab:NRB}, awaiting an
688: improvement in the upper bound.
689:
690: The violation of (\ref{KKK}) is probably related to the deviation from a fully
691: symmetric nonresonant background amplitude measured in $B^0\to
692: K^+K^-K^0$~\cite{Aubert:2005kd}. The discrepancy may be the
693: result of the fact that nonresonant backgrounds are unstable in the fits.
694: This is demonstrated by the large discrepancy between two values of the
695: nonresonant background in $B^+\to K^+\pi^+\pi^-$ measured in Refs.\
696: \cite{Aubert:2004fn} and~\cite{Abe:2005ig} using different definitions for the
697: nonresonant background. Another ambiguity in both $B^0\to K^+K^-K^0$ and
698: $B^+\to K^+K^-K^+$ is observed between a large and a very small contribution
699: peaking around 1500 MeV/$c^2$~\cite{Aubert:2005kd,Garmash:2004wa}.
700: As noted, the nonresonant background in $B^+\to K^+K^-K^+$ was found to be
701: completely symmetric~\cite{Garmash:2004wa}. Symmetry in the two identical
702: $K^+$ mesons follows from Bose symmetry.
703:
704: \bigskip
705:
706: \centerline{\bf VI. CONCLUSIONS}
707: \bigskip
708:
709: We have considered strangeness-changing $B \to PPP$ decays for $P = \pi, K$
710: under the assumption that the dominant transition is the isospin-preserving
711: penguin amplitude with flavor structure $\bar b \to \bar s$. In this
712: approximation pairs of $B^+$ and $B^0$ decay amplitudes to $K \pi \pi$ or $K
713: \bar K K$ are related to one another under the isospin reflection $u \lra d$,
714: and final states have $I=1/2$. For decays involving more than one kaon
715: relations involving final states with $K_S$ and $K_L$ hold under the assumption
716: that $\phi K$-subtracted amplitudes are symmetric under $K \leftrightarrow \bar
717: K$, as measured in processes involving $K^+$ and $K^-$.
718:
719: All the proposed isospin relations are obeyed experimentally where data exist,
720: excluding $\b(B^0\to K^{*0}\pi^0)$ which seems low relative to
721: $\frac{1}{2}\b(B^0\to K^{*+}\pi^-)$. The relations lead
722: to predictions where data are still missing.
723: This success led to our proposal to combine the study of Dalitz plots for
724: isospin-related processes, which can resolve discrete ambiguities in fitting
725: resonance parameters to given Dalitz plots.
726:
727: We have presented a model for nonresonant background amplitudes in $B\to
728: K\pi\pi$ and $B\to K\bar KK$, which are symmetric in the three outgoing meson
729: momenta. Predictions characteristic to this assumption are a suppressed
730: nonresonant background in $B^+ \to K^0 \pi^+ \pi^0$ and $B^0 \to K^+ \pi^0
731: \pi^-$ and simple relations between nonresonant branching ratios in several
732: processes. This approach has the potential for resolving some ambiguities in
733: determining nonresonant background amplitudes from fits to Dalitz plots.
734:
735: We have assumed that strangeness-changing charmless decays are dominated
736: by an isospin preserving $\bar b\to \bar s$ amplitude.
737: These decays involve also $\Delta I=1$ electroweak penguin contributions, which
738: are expected to be suppressed relative to
739: the dominant $I=0$ QCD-penguin amplitude~\cite{Gronau:1995hn,Fleischer:1993gr},
740: and small $\Delta I=1$ ``tree" amplitudes suppressed by $\lambda^2$ ($\lambda
741: \approx 0.2$).
742: The effects of these suppressed amplitudes in $B\to K\pi$ decays have been studied
743: recently in Refs.~\cite{Gronau:2005gz} and~\cite{Gronau:2005kz}, quoting
744: earlier references discussing these effects. The smallness of the effect
745: is demonstrated, for instance, by the relatively small measured deviations from
746: $\Delta I=0$ relations among $B\to K\pi$ decay rates.
747: An example is the ratio of branching ratios~\cite{Group(HFAG):2005rb,Buras:1998rb},
748: \beq\label{Rn}
749: R^{-1}_n \equiv \frac{2\b(B^0\to K^0\pi^0)}{\b(B^0\to K^+\pi^-)} = 1.22 \pm 0.11~~,
750: \eeq
751: which differs only by $2\sigma$ from the $\Delta I=0$ value of one.
752:
753: Isospin breaking effects should be considered in $B\to PPP$ when data become
754: sufficiently accurate. The first case to be studied is understanding
755: $2\b(B^0\to K^{*0}\pi^0) < \b(B^0\to K^{*+}\pi^-)$.
756: A ratio $R^{*-1}$, defined in analogy with $R^{-1}_n$,
757: \beq
758: R^{*-1}_n \equiv \frac{2\b(B^0\to K^{*0}\pi^0)}{\b(B^0\to K^{*+}\pi^-)}
759: = 0.35 \pm 0.17~~,
760: \eeq
761: is $3.9\sigma$ below one, thus presenting a larger discrepancy from $\Delta I
762: = 0$ than measured in $R^{-1}_n$. An interpretation for the small value of
763: $\b(B^0\to K^{*0}\pi^0)$ was presented in Ref.~\cite{Chiang:2003pm} in terms of
764: destructive interference between an electroweak penguin contribution and a QCD
765: penguin amplitude, the ratio of which is enhanced relative to that occurring
766: in $B\to K\pi$. An interesting and pressing question is whether such
767: enhancement can be accounted for in the Standard Model of electroweak and
768: strong interactions. While some suppression of $\b(B^0\to K^{*0}\pi^0)$ can be
769: accounted for in calculations based on QCD factorization~\cite{Beneke:2003zv},
770: central values computed for $B\to K^*\pi$ branching ratios are consistently
771: lower than the data by a factor 2.1 to 3.5.
772: Small $\Delta I =1$ contributions also may account for the $2.4 \sigma$
773: discrepancy in the relation (\ref{2KKK}).
774:
775: We have considered CP-averaged rates, disregarding in this work possible CP
776: asymmetries. The approximate relations we have derived apply separately to $B$
777: and $\bar B$ decays. While the $\Delta I=0$ terms in decay amplitudes are
778: dominated by a Cabibbo-Kobayashi-Maskawa (CKM) factor $V^*_{tb}V_{ts}$, a CKM
779: factor $V^*_{ub}V_{us}$ smaller by $\lambda^2$ is associated with ``tree"
780: contributions. The two CKM factors involve different weak phases. Direct CP
781: violation is expected if the two terms carry also different strong phases.
782:
783: Potential CP asymmetries are expected from interference of decay amplitudes for
784: $B^+\to K^+ \chi_{c0}$, where $\chi_{c0}\to \pi^+\pi^-$ and $\chi_{c0} \to
785: K^+K^-$, with ``tree" amplitudes in $B^+\to K^+\pi^+\pi^-$ and $B^+\to
786: K^+K^+K^-$, respectively. While a large strong phase difference is induced by
787: the $\chi_{c0}$ width~\cite{Eilam:1995nz},
788: the asymmetries depend also on the magnitudes of the smaller ``tree" amplitudes
789: for which calculations are model-dependent~\cite{Bajc:1998bs}. The fractions
790: of $K^+\chi_{c0}$ in $B^+\to K^+\pi^+\pi^-$
791: and $B^+\to K^+K^+K^-$ are small, at a level of three
792: percent~\cite{Garmash:2004wa,Aubert:2004fn} or smaller~\cite{Abe:2005ig}.
793:
794: The decays $B^+\to K^+\rho^0$, where $\rho^0\to\pi^+\pi^-$, amount to a
795: larger fraction of $B^+\to K^+\pi^+\pi^-$, about ten
796: percent~\cite{Garmash:2004wa,Aubert:2004fn,Abe:2005ig}.
797: Tentative evidence for a CP asymmetry
798: in these decays has been reported recently, $A_{CP}(B^+\to K^+\rho^0) =
799: 0.34 \pm 0.13 \pm 0.06^{+0.15}_{-0.20}$~\cite{Aubert:2004fn},
800: $0.30 \pm 0.11 \pm 0.03^{+0.11}_{-0.04}$~\cite{Abe:2005ig}.
801: This may be compared with a prediction, $A_{CP}(B^+\to K^+\rho^0) = 0.21 \pm
802: 0.10$, obtained in a global SU(3) fit to all $B\to PV$
803: decays~\cite{Chiang:2003pm}. A CP asymmetry at a level of 10$\%$ in the
804: processes discussed in this paper, resulting from penguins-tree interference as
805: measured in the asymmetry for $B^0\to K^+\pi^-$~\cite{Group(HFAG):2005rb},
806: would imply a small but non-negligible violation of the $\Delta I=0$ relations.
807:
808: \bigskip
809:
810: \centerline{\bf ACKNOWLEDGMENTS}
811: \bigskip
812:
813: We thank Hai-Yang Cheng, Denis Dujmic, Guy Engelhard, Guy Raz and Aaron
814: Roodman for helpful discussions. This work was performed in part while M. G.
815: visited SLAC and while J. L. R. was at the Aspen Center for Physics. The
816: research was supported in part by the
817: United States Department of Energy under Grant No.\ DE FG02 90ER40560,
818: by the Israel Science Foundation founded by the Israel Academy of Science
819: and Humanities, Grant No. 1052/04, and by the German--Israeli Foundation
820: for Scientific Research and Development, Grant No. I-781-55.14/2003.
821:
822: \begin{thebibliography}{99}
823:
824: %1
825: %\cite{Gershon:2004tk}
826: \bibitem{Gershon:2004tk}
827: T.~Gershon and M.~Hazumi,
828: %``Time-dependent CP violation in B0 $\to$ P0 P0 X0 decays,''
829: Phys.\ Lett.\ B {\bf 596}, 163 (2004)
830: [arXiv:hep-ph/0402097].
831: %%CITATION = HEP-PH 0402097;%%
832:
833: % 2
834: %\cite{Aubert:2004ta}
835: \bibitem{Aubert:2004ta}
836: B.~Aubert {\it et al.} [BABAR Collabortation],
837: %``Branching fractions and CP asymmetries in B0 $\to$ K+ K- K0(S) and B+ $\to$
838: %K+ K0(S) K0(S),''
839: Phys.\ Rev.\ Lett.\ {\bf 93}, 181805 (2004)
840: [arXiv:hep-ex/0406005].
841: %%CITATION = HEP-EX 0406005;%%
842:
843: %3
844: %\cite{Aubert:2005ja}
845: \bibitem{Aubert:2005ja}
846: B.~Aubert {\it et al.} [BABAR Collaboration],
847: %``Measurement of CP asymmetries in B0 $\to$ Phi K0 and B0 $\to$ K+ K- K0(S)
848: %decays,''
849: Phys.\ Rev.\ D {\bf 71}, 091102 (2005)
850: [arXiv:hep-ex/0502019].
851: %%CITATION = HEP-EX 0502019;%%
852:
853: %4
854: %\cite{Aubert:2005kd}
855: \bibitem{Aubert:2005kd}
856: B.~Aubert {\it et al.} [BABAR Collaboration],
857: %``Dalitz plot study of B0 $\to$ K+ K- K0(S) decays,''
858: arXiv:hep-ex/0507094.
859: %%CITATION = HEP-EX 0507094;%%
860:
861: % 5
862: %\cite{Abe:2003yt}
863: \bibitem{Abe:2003yt}
864: K.~Abe {\it et al.} [Belle Collaboration],
865: %``Measurement of time-dependent CP-violating asymmetries in B0 $\to$ Phi
866: %K0(S), K+ K- K0(S), and eta' K0(S) decays,''
867: Phys.\ Rev.\ Lett.\ {\bf 91}, 261602 (2003)
868: [arXiv:hep-ex/0308035].
869: %%CITATION = HEP-EX 0308035;%%
870:
871: %6
872: %\cite{Garmash:2003er}
873: \bibitem{Garmash:2003er}
874: A.~Garmash {\it et al.} [Belle Collaboration],
875: %``Study of B meson decays to three-body charmless hadronic final states,''
876: Phys.\ Rev.\ D {\bf 69}, 012001 (2004)
877: [arXiv:hep-ex/0307082].
878: %%CITATION = HEP-EX 0307082;%%
879:
880: %7
881: %\cite{Chen:2005dr}
882: \bibitem{Chen:2005dr}
883: K.~F.~Chen {\it et al.} [Belle Collaboration],
884: %``Time-dependent CP-violating asymmetries in b $\to$ s anti-q q
885: %transitions,''
886: Phys.\ Rev.\ D {\bf 72}, 012004 (2005)
887: [arXiv:hep-ex/0504023].
888: %%CITATION = HEP-EX 0504023;%%
889:
890: %8
891: %\cite{Garmash:2004wa}
892: \bibitem{Garmash:2004wa}
893: A. Garmash \ite~[Belle Collaboration],
894: %``Dalitz analysis of the three-body charmless decays B+ $\to$ K+ pi+ pi- and
895: %B+ $\to$ K+ K+ K-,''
896: Phys.\ Rev.\ D {\bf 71}, 092003 (2005)
897: [arXiv:hep-ex/0412066].
898: %%CITATION = HEP-EX 0412066;%%
899:
900: %9
901: %\cite{Aubert:2004fn}
902: \bibitem{Aubert:2004fn}
903: B.~Aubert {\it et al.} [BABAR Collaboration],
904: %``Amplitude analysis of B+- $\to$ pi+- pi-+ pi+- and B+- $\to$ K+- pi-+
905: %pi+-,''
906: arXiv:hep-ex/0408032;
907: %%CITATION = HEP-EX 0408032;%%
908: %\cite{Aubert:2005ce}
909: %\bibitem{Aubert:2005ce}
910: %B.~Aubert {\it et al.} [BABAR Collaboration],
911: %``Dalitz-plot analysis of the decays B+- $\to$ K+- pi-+ pi+-,''
912: arXiv:hep-ex/0507004.
913: %%CITATION = HEP-EX 0507004;%%
914:
915: %10
916: %\cite{Group(HFAG):2005rb}
917: \bibitem{Group(HFAG):2005rb}
918: K. Anikeev {\it et al.}, Heavy Flavor Averaging Group,
919: ``Averages of b-hadron properties as of winter 2005,
920: arXiv:hep-ex/0505100.
921: Updated results and references are tabulated periodically by this group:
922: {\tt http://www.slac.stanford.edu/xorg/hfag/rare.}
923: %%CITATION = HEP-EX 0505100;%%
924:
925: %11
926: %\cite{Gronau:1994rj}
927: \bibitem{Gronau:1994rj}
928: M.~Gronau, O.~F.~Hernandez, D.~London and J.~L.~Rosner,
929: %``Decays of B mesons to two light pseudoscalars,''
930: Phys.\ Rev.\ D {\bf 50}, 4529 (1994)
931: [arXiv:hep-ph/9404283].
932: %%CITATION = HEP-PH 9404283;%%
933:
934: %12
935: %\cite{Gronau:1995hn}
936: \bibitem{Gronau:1995hn}
937: M.~Gronau, O.~F.~Hernandez, D.~London and J.~L.~Rosner,
938: %``Electroweak penguins and two-body B decays,''
939: Phys.\ Rev.\ D {\bf 52}, 6374 (1995)
940: [arXiv:hep-ph/9504327].
941: %%CITATION = HEP-PH 9504327;%%
942:
943: %13
944: %\cite{Chiang:2004nm}
945: \bibitem{Chiang:2004nm}
946: C.~W.~Chiang, M.~Gronau, J.~L.~Rosner and D.~A.~Suprun,
947: %``Charmless B $\to$ P P decays using flavor SU(3) symmetry,''
948: Phys.\ Rev.\ D {\bf 70}, 034020 (2004)
949: [arXiv:hep-ph/0404073].
950: %%CITATION = HEP-PH 0404073;%%
951:
952: %14
953: %\cite{Chiang:2003pm}
954: \bibitem{Chiang:2003pm}
955: C.~W.~Chiang, M.~Gronau, Z.~Luo, J.~L.~Rosner and D.~A.~Suprun,
956: %``Charmless B $\to$ V P decays using flavor SU(3) symmetry,''
957: Phys.\ Rev.\ D {\bf 69}, 034001 (2004)
958: [arXiv:hep-ph/0307395].
959: %%CITATION = HEP-PH 0307395;%%
960:
961: %15
962: \bibitem{DGR}
963: A.~S.~Dighe, M.~Gronau and J.~L.~Rosner,
964: %``Amplitude relations for $B$ decays involving $\eta$ and $\eta'$,''
965: Phys.\ Lett.\ B {\bf 367}, 357 (1996) [arXiv:hep-ph/9509428]; {\it Erratum-ibid.} B
966: {\bf 377}, 325 (1996);
967: %%CITATION = HEP-PH 9509428;%%
968: %\bibitem{DGReta}
969: %A.~S.~Dighe, M.~Gronau and J.~L.~Rosner,
970: %``B decays involving eta and eta' in light of the B $\to$ K eta' process,''
971: Phys.\ Rev.\ Lett.\ {\bf 79}, 4333 (1997) [arXiv:hep-ph/9707521];
972: %%CITATION = HEP-PH 9707521;%%
973: %\cite{Chiang:2003rb}
974: % \bibitem{Chiang:2003rb}
975: C.~W.~Chiang, M.~Gronau and J.~L.~Rosner,
976: %``Two-body charmless B decays involving eta and eta',''
977: Phys.\ Rev.\ D {\bf 68}, 074012 (2003)
978: [arXiv:hep-ph/0306021].
979: %%CITATION = HEP-PH 0306021;%%
980:
981: %16
982: %\cite{Eilam:1995nz}
983: \bibitem{Eilam:1995nz}
984: G.~Eilam, M.~Gronau and R.~R.~Mendel,
985: %``Large CP asymmetries in B+ $\to$ eta(c) (chi(c0)) pi+- from eta(c)
986: %(chi(c0)) width,''
987: Phys.\ Rev.\ Lett.\ {\bf 74}, 4984 (1995)
988: [arXiv:hep-ph/9502293].
989: %%CITATION = HEP-PH 9502293;%%
990: %\cite{Fajfer:2004cx}
991:
992: %17
993: %\cite{Bajc:1998bs}
994: \bibitem{Bajc:1998bs}
995: B.~Bajc, S.~Fajfer, R.~J.~Oakes, T.~N.~Pham and S.~Prelovsek,
996: %``The CP violating asymmetry in B+- $\to$ M anti-M pi+- decays,''
997: Phys.\ Lett.\ B {\bf 447}, 313 (1999)
998: [arXiv:hep-ph/9809262];
999: %%CITATION = HEP-PH 9809262;%%
1000: %\cite{Fajfer:1998yc}
1001: %\bibitem{Fajfer:1998yc}
1002: S.~Fajfer, R.~J.~Oakes and T.~N.~Pham,
1003: %``The penguin operators in nonresonant B- $\to$ M anti-M pi- (M = pi-,K-,K0)
1004: %decays,''
1005: Phys.\ Rev.\ D {\bf 60}, 054029 (1999)
1006: [arXiv:hep-ph/9812313];
1007: %%CITATION = HEP-PH 9812313;%%
1008: %\cite{Fajfer:2002ct}
1009: %\bibitem{Fajfer:2002ct}
1010: S.~Fajfer, R.~J.~Oakes and T.~N.~Pham,
1011: %``CP violating phase gamma and the partial widths asymmetry in B- $\to$ pi+
1012: %pi- K- and B- $\to$ K+ K- K- decays,''
1013: Phys.\ Lett.\ B {\bf 539}, 67 (2002)
1014: [arXiv:hep-ph/0203072];
1015: %%CITATION = HEP-PH 0203072;%%
1016: %\bibitem{Fajfer:2004cx}
1017: S.~Fajfer, T.~N.~Pham and A.~Prapotnik,
1018: %``CP violation in the partial width asymmetries for B- $\to$ pi+ pi- K- and
1019: %B- $\to$ K+ K- K- decays,''
1020: Phys.\ Rev.\ D {\bf 70}, 034033 (2004)
1021: [arXiv:hep-ph/0405065].
1022: %%CITATION = HEP-PH 0405065;%%
1023:
1024: %18
1025: %\cite{Cheng:2002qu}
1026: \bibitem{Cheng:2002qu}
1027: H.~Y.~Cheng and K.~C.~Yang,
1028: %``Nonresonant three-body decays of D and B mesons,''
1029: Phys.\ Rev.\ D {\bf 66}, 054015 (2002)
1030: [arXiv:hep-ph/0205133];
1031: %%CITATION = HEP-PH 0205133;%%
1032: %\cite{Cheng:2005ug}
1033: %\bibitem{Cheng:2005ug}
1034: H.~Y.~Cheng, C.~K.~Chua and A.~Soni,
1035: %``CP-violating asymmetries in B0 decays to K+ K- K0(S(L)) and K0(S) K0(S)
1036: %K0(S(L)),''
1037: arXiv:hep-ph/0506268;
1038: %%CITATION = HEP-PH 0506268;%%
1039: %\cite{Cheng:2005nb}
1040: %\bibitem{Cheng:2005nb}
1041: H.~Y.~Cheng, C.~K.~Chua and K.~C.~Yang,
1042: %``Charmless hadronic B decays involving scalar mesons: Implications to the
1043: %nature of light scalar mesons,''
1044: arXiv:hep-ph/0508104.
1045: %%CITATION = HEP-PH 0508104;%%
1046:
1047: %19
1048: %\cite{Minkowski:2004xf}
1049: \bibitem{Minkowski:2004xf}
1050: P.~Minkowski and W.~Ochs,
1051: %``B decays into light scalar particles and glueball,''
1052: Eur.\ Phys.\ J.\ C {\bf 39}, 71 (2005)
1053: [arXiv:hep-ph/0404194].
1054: %%CITATION = HEP-PH 0404194;%%
1055:
1056: %20
1057: %\cite{Gronau:2003ep}
1058: \bibitem{Gronau:2003ep}
1059: M.~Gronau and J.~L.~Rosner,
1060: %``I-spin and U-spin in B $\to$ K K anti-K,''
1061: Phys.\ Lett.\ B {\bf 564}, 90 (2003)
1062: [arXiv:hep-ph/0304178].
1063: %%CITATION = HEP-PH 0304178;%%
1064:
1065: %21
1066: %\cite{Grossman:2003qp}
1067: \bibitem{Grossman:2003qp}
1068: Y.~Grossman, Z.~Ligeti, Y.~Nir and H.~Quinn,
1069: %``SU(3) relations and the CP asymmetries in B decays to eta' K(S), Phi K(S)
1070: %and K+ K- K(S),''
1071: Phys.\ Rev.\ D {\bf 68}, 015004 (2003)
1072: [arXiv:hep-ph/0303171];
1073: %%CITATION = HEP-PH 0303171;%%
1074: %\cite{Engelhard:2005hu}
1075: %\bibitem{Engelhard:2005hu}
1076: G.~Engelhard, Y.~Nir and G.~Raz,
1077: %``SU(3) relations and the CP asymmetry in B $\to$ K(S) K(S) K(S),''
1078: arXiv:hep-ph/0505194;
1079: %%CITATION = HEP-PH 0505194;%%
1080: %\cite{Engelhard:2005ky}
1081: %\bibitem{Engelhard:2005ky}
1082: G.~Engelhard and G.~Raz,
1083: %``Using SU(3) relations to bound the CP asymmetries in B $\to$ K K K
1084: %decays,''
1085: arXiv:hep-ph/0508046.
1086: %%CITATION = HEP-PH 0508046;%%
1087:
1088: %22
1089: %\cite{Abe:2005kr}
1090: \bibitem{Abe:2005kr}
1091: K.~Abe {\it et al.} [Belle Collaboration],
1092: %``Dalitz analysis of the three-body charmless decay B0=>Kspi+pi-,''
1093: arXiv:hep-ex/0509047.
1094: %%CITATION = HEP-EX 0509047;%%
1095:
1096: %23
1097: %\cite{Aubert:2005id}
1098: \bibitem{Aubert:2005id}
1099: B.~Aubert {\it et al.} [BABAR Collaboration],
1100: %``Measurement of CP asymmetries in B0 $\to$ K0(S) pi0 pi0 decays,''
1101: arXiv:hep-ex/0508017.
1102: %%CITATION = HEP-EX 0508017;%%
1103:
1104: %24
1105: %\cite{Eidelman:2004wy}
1106: \bibitem{Eidelman:2004wy}
1107: S.~Eidelman {\it et al.} [Particle Data Group],
1108: %``Review of particle physics,''
1109: Phys.\ Lett.\ B {\bf 592}, 1 (2004).
1110: %%CITATION = PHLTA,B592,1;%%
1111:
1112: %25
1113: %\cite{Abe:2005ig}
1114: \bibitem{Abe:2005ig}
1115: K.~Abe \ite~[Belle Collaboration],
1116: %``Search for Direct CP Violation in Three-Body Charmless B+- => K+-pi+-pi-+
1117: %Decay,''
1118: arXiv:hep-ex/0509001.
1119: %%CITATION = HEP-EX 0509001;%%
1120:
1121: %26
1122: %\cite{Fleischer:1993gr}
1123: \bibitem{Fleischer:1993gr}
1124: R.~Fleischer,
1125: %``Electroweak Penguin effects beyond leading logarithms in the B meson decays
1126: %B- $\to$ K- Phi and B- $\to$ pi- anti-K0,''
1127: Z.\ Phys.\ C {\bf 62}, 81 (1994).
1128: %%CITATION = ZEPYA,C62,81;%%
1129:
1130: %27
1131: %\cite{Gronau:2005gz}
1132: \bibitem{Gronau:2005gz}
1133: M.~Gronau and J.~L.~Rosner,
1134: %``The b $\to$ s penguin amplitude in charmless B $\to$ P P decays,''
1135: Phys.\ Rev.\ D {\bf 71}, 074019 (2005)
1136: [arXiv:hep-ph/0503131].
1137: %%CITATION = HEP-PH 0503131;%%
1138:
1139: %28
1140: %\cite{Gronau:2005kz}
1141: \bibitem{Gronau:2005kz}
1142: M.~Gronau,
1143: %``A precise sum rule among four B $\to$ K pi CP asymmetries,''
1144: %arXiv:hep-ph/0508047, to be published in Phys. Lett. B.
1145: Phys.\ Lett.\ B {\bf 627}, 82 (2005).
1146: %%CITATION = HEP-PH 0508047;%%
1147:
1148: %29
1149: %\cite{Buras:1998rb}
1150: \bibitem{Buras:1998rb}
1151: A.~J.~Buras and R.~Fleischer,
1152: %``A general analysis of gamma determinations from B $\to$ pi K decays,''
1153: Eur.\ Phys.\ J.\ C {\bf 11}, 93 (1999)
1154: [arXiv:hep-ph/9810260].
1155: %%CITATION = HEP-PH 9810260;%%
1156:
1157: %30
1158: %\cite{Beneke:2003zv}
1159: \bibitem{Beneke:2003zv}
1160: M.~Beneke and M.~Neubert,
1161: %``QCD factorization for B $\to$ P P and B $\to$ P V decays,''
1162: Nucl.\ Phys.\ B {\bf 675}, 333 (2003)
1163: [arXiv:hep-ph/0308039].
1164: %%CITATION = HEP-PH 0308039;%%
1165:
1166: \end{thebibliography}
1167: \end{document}
1168: