1: %%version of 1903. some minor changes
2: \documentclass[aps,prd,preprint,showpacs,floats]{revtex4}
3: \usepackage[dvips]{graphicx}
4: \usepackage{bm}
5: \begin{document}
6: \title{Power Corrections in Charmless B Decays }
7: \author{T. N. Pham}
8: \email[E-mail address: ]{Tri-Nang.Pham@cpht.polytechnique.fr}
9: \author{Guohuai Zhu}
10: \email[E-mail address: ]{Guohuai.Zhu@cpht.polytechnique.fr}
11: \affiliation{Centre de Physique Theorique, \\
12: Centre National de la Recherche Scientifique, UMR 7644, \\
13: Ecole Polytechnique, 91128 Palaiseau Cedex, France}
14: \date{\today}
15: \begin{abstract}
16: In this paper, we focus on the role of power corrections in QCD
17: factorization(QCDF) method in charmless two-body nonleptonic $B$ meson decays.
18: We use the ratio of the branching fraction of $B^+ \to \pi^+ K^{\ast 0}$ to
19: that of $B^0 \to \pi^- \rho^+$, for which the theoretical uncertainties are
20: greatly reduced, to show clearly that the power corrections in charmless B
21: decays are probably large. With other similar ratios considered, for
22: example, for the $B^0 \to K^- \rho^+$ decay, it is very
23: likely that, among various sources of power corrections, annihilation
24: topology plays an indispensable role at least for penguin dominated
25: $\rm PV$ channels. We also consider some selective
26: ratios of direct CP asymmetries.
27: Among these, we find that, if power corrections other than the
28: chirally enhanced power corrections and annihilation topology
29: were negligible, QCDF would
30: predict the direct CP asymmetry of $B \to \pi^+ \pi^-$ to be
31: about 3 times larger than that of
32: $B \to \pi^\pm K^\mp$, with opposite sign. Experimentally any significant
33: deviation from this prediction
34: would suggest either new physics or possibly the importance of long-distance
35: rescattering effects.
36:
37: \end{abstract}
38: \pacs{13.25.Hw 12.38.Bx}
39: \maketitle
40: \section{Introduction}
41: With the excellent performance of KEK and SLAC B factories, a great number
42: of charmless $B$ decay channels have been measured with high precision. It
43: is even more exciting that, with the expected larger data sample in the
44: near future, CP violations in many decay channels might be soon
45: reachable. However in most cases, our ignorance of strong dynamics stands
46: in the way of
47: either identifying the signal of new physics or extracting the
48: fundamental Cabibbo-Kobayashi-Maskawa (CKM) angles from the observables.
49:
50: Theorists have tried hard to have a better understanding on the strong
51: dynamics in non-leptonic $B$ decays, in which QCD factorization method
52: (QCDF)\cite{QCDF} is a recent progress.
53: In the heavy quark limit, QCDF has shown
54: that the hadronic matrix elements of $B$ decays have simpler
55: non-perturbative structures, while the QCD corrections are perturbatively
56: calculable, at least at one-loop order. However the $1/m_b$ power
57: corrections, especially annihilation contributions and chirally enhanced
58: power corrections \cite{QCDF1,DYZ}, are phenomenologically important. It
59: is even worse that
60: factorization generally breaks down beyond the leading power expansion. As
61: a result, the model-dependent parametrization has to be introduced in
62: QCDF to account for these formally power-suppressed contributions. For
63: some decay channels, these power corrections could even
64: compete with the
65: leading power contributions in some parameter space which makes the
66: reliability of QCDF predictions in doubt. Very recently, a new effective
67: theory, called soft-collinear effective theory (SCET), has been applied to
68: charmless $B$ decays \cite{SCET}, in which the power corrections can be
69: studied in a systematic way, though more non-perturbative operators have
70: to be introduced. However since this method is still under development and
71: several issues remain to be resolved, we shall not discuss SCET further in
72: this paper.
73:
74: Thus it is of great interest to investigate the role of power corrections in
75: charmless $B$ decays, and to determine if the power corrections were
76: really important,
77: whether they were mainly from chirally enhanced corrections
78: and annihilation topology or from other sources of power corrections,
79: such as long-distance rescatterings.
80: In this paper, we
81: shall discuss these problems in a transparent way based on experimental
82: measurements.
83:
84: \section{QCD factorization for charmless $B$ decays}
85: To make the paper self-contained, we will recapitulate the main point
86: of QCDF in the following. One may find more details in Refs
87: \cite{QCDF,QCDF1,DYZ,QCDF2,DYZPV}.
88:
89: There are three distinct scales: $M_W \gg m_b \gg \Lambda_{QCD}$ involved
90: in charmless $B$ decays. To go beyond naive model estimation, the physics of
91: different scales should be separated from each other. It is known that,
92: with the operator product expansion and renormalization group equation,
93: the effective Lagrangian can be obtained, in which short-distance effects
94: involving large virtual momenta of the loop corrections from the scale
95: $M_W$ down to $\mu ={\cal O}(m_b)$ are cleanly integrated into the Wilson
96: coefficients. Then the amplitude for the decay $B \to M_1 M_2$ can be
97: expressed as \cite{Buras}:
98: \begin{equation}
99: {\cal A}(B \to M_1 M_2)=\frac{G_F}{\sqrt{2}} \sum_{i=1}^6 \sum_{q=u,c}
100: \lambda_q C_i (\mu) \langle M_1 M_2 \vert Q_i (\mu) \vert B \rangle ~,
101: \end{equation}
102: where $\lambda_q$ is a CKM factor, $C_i (\mu)$ is the Wilson coefficient
103: which is perturbatively calculable from first principles. The effective
104: operators $Q_{1,2}$ and $Q_{3,...,6}$ are tree level and QCD penguin
105: operators, respectively. We have neglected electroweak penguin operators
106: $Q_{7,...,10}$ because their effects are numerically small in most decay
107: channels. $\langle M_1 M_2 \vert Q_i (\mu) \vert B \rangle$ is a hadronic
108: matrix element which contains the physics effects from the scale $\mu =
109: {\cal O}(m_b)$ down to $\Lambda_{\rm QCD}$. Since the perturbative and
110: nonperturbative effects related to $m_b$ and $\Lambda_{\rm QCD}$ still
111: entangle with each other, it is highly nontrivial to estimate the hadronic
112: matrix elements. But in the heavy quark limit, QCDF shows that the above
113: hadronic matrix elements can be factorized into hard radiative corrections
114: and simpler nonperturbative structures which can be parametrized by the
115: form factors and meson light-cone distribution amplitudes (LCDAs).
116:
117: But for phenomenological applications, the power corrections in $1/m_b$,
118: especially the chirally enhanced corrections and annihilation
119: contributions should be taken into account. Chirally enhanced corrections
120: arise from two-body twist-3 LCDAs of the final state mesons. Unfortunately the
121: factorization breaks down for hard spectator scattering diagrams because
122: the twist-3 LCDA does not approach zero in the endpoint region which
123: introduces a logarithmic divergence. A similar divergence also appears in
124: the annihilation contributions. Phenomenologically, Beneke {\it et al.}
125: \cite{QCDF1} adopted a model parametrization for the endpoint
126: divergence:
127: \begin{equation}
128: X_{A,H}=\int \limits_0^1 \frac{dx}{x} = \log \frac{m_B}{\Lambda_h}
129: (1+\rho_{A,H} e^{i \phi_{A,H} } ) ~~~~~(0 \le \rho_{A,H} \le 1 ) ~,
130: \end{equation}
131: where $X_A$ ($X_H$) denotes the divergent terms from annihilation
132: topology (hard spectator scattering). The corresponding price to pay
133: is model-dependence uncertainties and
134: also large numerical uncertainties from $X_{A,H}$.
135: With the above discussions, the decay amplitudes can be written as
136: \begin{equation}
137: {\cal A}(B \to M_1 M_2)=\frac{G_F}{\sqrt{2}} \sum_{p=u,c} v_p \left (
138: \sum_{i=1}^6 a_i^p \langle M_1 M_2 \vert O_i \vert B \rangle_f +
139: \sum_{j} f_B f_{M_1}f_{M_2} b_j \right ),
140: \end{equation}
141: where $\langle M_1 M_2 \vert O_i \vert B \rangle_f$ is the factorized
142: hadronic matrix element which has the same definition as that in the naive
143: factorization approach. For the complete expressions of QCD
144: coefficients $a_i$ and annihilation parameters $b_j$, one may refer to
145: Ref. \cite{QCDF2}.
146:
147:
148: \section{Are power corrections large ?}
149:
150: It is clear that chirally enhanced power corrections should be large, at
151: least in $B$ decays to two light pseudoscalar mesons, because although
152: they
153: are suppressed by the factor (taking the pion as an illustration)
154: \begin{equation}
155: r_\chi^\pi = \frac{2 m_\pi^2}{m_b (m_u + m_d)} \mbox{~,}
156: \end{equation}
157: which is formally of the order of $\Lambda_{QCD}/m_b$, numerically this
158: factor
159: is of order one. Experimentally, the large branching ratios of $B \to \pi K$
160: decays also strongly support this statement. For example, the ratio
161: \begin{equation}
162: \frac{{\cal B}(B^+ \to \pi^+ K^0)}{{\cal B}(B^+ \to \pi^+ K^{\ast 0})}=
163: \frac{(21.8 \pm 1.4) \times 10^{-6}}{(9.0 \pm 1.3) \times 10^{-6}}=2.4 \pm
164: 0.4
165: \end{equation}
166: would be theoretically smaller than $1$ without chirally enhanced
167: contributions, say $r_{\chi}^K a_6$ term. The experimental data in the
168: above equation are from ref. \cite{HFAG1}.
169:
170: However it is not very clear about the role of annihilation topology
171: in $B$ decays. It was once believed to be quite small in most $B$ decay
172: channels, because of the power suppression \cite{Ali}. The importance of
173: annihilation contributions was first noticed in the perturbative QCD
174: method for charmless $B$ decays\cite{pQCD}. Recent phenomenological
175: analyses
176: \cite{DYZGA,QCDF2,aleksan,cottingham} based on QCDF also suggest
177: substantial
178: contributions from annihilation topology. However since in QCDF,
179: there are many
180: parameters involved in the global analysis of experimental data, it would
181: be interesting to show the importance of annihilation topology in a
182: transparent way. In the following we will try to do it with various
183: ratios. We find that annihilation topology very likely plays a significant
184: role, at least in penguin dominated
185: $B \to PV$ decays, where $P$ denotes light pseudoscalar meson and $V$
186: denotes light vector meson.
187:
188: Let us first consider the ratio of the decay amplitude of $B^+
189: \to \pi^+ K^{\ast 0}$ to that of $B^0 \to \rho^+ \pi^-$. If the power
190: corrections were negligible, this ratio would be theoretically very
191: clean where the form factors cancel out, furthermore it is
192: almost independent ot the CKM angle $\gamma$ and the strange-quark mass:
193: \begin{equation} \label{pik}
194: \left \vert \frac{{\cal A}(B^+ \to \pi^+ K^{\ast 0})}{{\cal A}(B^0 \to
195: \rho^+ \pi^-)}\right \vert \simeq
196: \left \vert \frac{ V_{cb}V_{cs}}{V_{ub}V_{ud}} \right \vert
197: \frac{f_{K^\ast}}{f_\rho} \left \vert
198: \frac{a_4^c (\pi K^\ast)+r_\chi^{K^\ast} a_6^c (\pi K^\ast) }{a_1^u}
199: \right \vert \mbox{~,}
200: \end{equation}
201: where the penguin contributions to $B^0 \to \rho^+ \pi^-$ decay
202: and the term proportional to $V_{ub}V_{us}$ in the numerator are
203: neglected. This should be a reasonable approximation up to a few percent
204: level. In QCDF, $\vert (a_4^c (\pi K^\ast)+r_\chi^{K^\ast} a_6^c (\pi
205: K^\ast) ) /a_1^u \vert$ should be about or less than 0.04 (With the
206: default parameters in ref. \cite{QCDF2}, it is $0.03$)
207: and $f_{K^\ast}/f_\rho$ is very close to unity. While for the CKM matrix
208: elements, there is a useful inequality \cite{Pham,Buras1}:
209: \begin{equation}\label{inequality}
210: \left \vert \frac{V_{ub}}{V_{cb}}\right \vert = \lambda \sin \beta
211: \sqrt{1+\frac{\cos^2 \alpha}{\sin^2 \alpha}} \ge \lambda \sin \beta ~.
212: \end{equation}
213: The current measured value of $\sin 2\beta$ from BaBar and Belle gives
214: \cite{HFAG}
215: \begin{equation}
216: \sin 2\beta= 0.736 \pm 0.049 \mbox{~,}
217: \end{equation}
218: from which we obtain $\sin \beta = 0.402 \pm 0.033$. In principle we could
219: get another solution $\sin \beta \simeq 0.9$ which however is inconsistent
220: with the direct $\vert V_{ub} \vert$ measurements. So given
221: $\lambda=0.224$, we will get an interesting lower limit
222: \begin{equation}\label{inequality2}
223: \left \vert \frac{V_{ub}}{V_{cb}} \right \vert \ge \lambda \sin \beta =
224: 0.090 \pm 0.007 > 0.078 ~~~~~ \mbox{( 90\% C.L.) .}
225: \end{equation}
226: With this lower limit, there arises a clear discrepancy
227: between theory and experiments:
228: \begin{equation}\label{pikrhopi}
229: 0.53 > \left \vert \frac{{\cal A}(B^+ \to \pi^+ K^{\ast
230: 0})}{{\cal A}(B^0 \to \rho^+ \pi^-)}\right \vert = \left [
231: \frac{\tau(B^0)}{\tau(B^+)} \frac{{\cal B}(B^+ \to \pi^+ K^{\ast
232: 0})}{{\cal B}(B^0 \to \rho^+ \pi^-)} \right ]^{1/2} = 0.77 \pm 0.09~,
233: \end{equation}
234: where the left hand side is from theoretical estimation Eqs.
235: (\ref{pik}), (\ref{inequality}), (\ref{inequality2}) and the right
236: hand side is obtained from the following experimental measurements:
237: ${\cal B}(B^0 \to \rho^+ \pi^-)=13.9 \pm 2.7$ \cite{laget},
238: ${\cal B}(B^+ \to \pi^+ K^{\ast 0})=9.0 \pm 1.3$ \cite{HFAG1} and
239: $\tau(B^+)/ \tau(B^0)=1.083 \pm 0.017$ \cite{PDG}.
240: Notice that $\cos^2 \alpha$ is of the order of a few percent, it seems
241: appropriate to neglect it in Eq. (\ref{inequality}). Then the left hand
242: side of Eq. (\ref{pikrhopi}) would be reduced further to $0.46 \pm 0.04$
243: if we fix $(a_4+r_\chi a_6)/a_1$ to be $0.04$. In fact, $(a_4+r_\chi
244: a_6)/a_1$ for
245: $\pi K^\ast$ channels is typically less than 0.04 in most of the QCDF
246: parameter space. For example, if using the default parameters in ref.
247: \cite{QCDF2}, we will get $0.35$ for this ratio.
248: Since the chirally enhanced corrections
249: for this ratio are not expected to be large because
250: $a_6 (\pi^+ K^\ast)$ vanishes at tree level, this is a strong indication
251: that annihilation topology and/or other sources of power corrections might
252: play an important role at least in $B \to PV$ decays.
253:
254: If we turn to another ratio, the branching fraction of $B^0 \to K^+
255: \rho^-$ to that of $B^0 \to \rho^- \pi^+$, there is similar disagreement,
256: although this time it is not well established considering
257: the large theoretical uncertainties. Assuming
258: negligible annihilation contributions, the form factors cancel out again
259: for this ratio. However there are significant dependences on angle
260: $\gamma$, $V_{ub}/V_{cb}$ and especially the strange-quark mass $m_s$ in
261: the numerator because of the factor $r_\chi^K$:
262: \begin{equation}
263: \left \vert \frac{{\cal A}(B^0 \to K^+ \rho^-)}{{\cal A}(B^0 \to
264: \rho^- \pi^+)}\right \vert \simeq
265: \frac{f_K}{f_\pi} \left \vert \lambda e^{i \gamma}+\left \vert
266: \frac{V_{cb}}{V_{ub}} \right \vert
267: \frac{a_4^c (\rho K)-r_{\chi}^K a_6^c (\rho K) }{a_1^u}
268: \right \vert \mbox{~,}
269: \end{equation}
270: where as usual, the small penguin terms in the denominator has been
271: neglected. If we take the default parameters in ref. \cite{QCDF2}, namely
272: $\gamma=70^\circ$, $V_{ub}/V_{cb}=0.09$ and $a_4^c (\rho K)-r_\chi^K a_6^c
273: (\rho K)=0.037+0.003 i$
274: with the strange-quark mass set to be $90$ MeV, the theoretical
275: prediction for the ratio
276: ${\cal B}(B^0 \to K^+ \rho^-)/{\cal B}(B^0 \to \rho^- \pi^+)$ is
277: $0.38$, which is again significantly smaller than the experimental
278: measurements $1.01 \pm 0.34$ \cite{HFAG1,laget}. Of course, when the
279: theoretical uncertainties, especially the uncertainty of the strange-quark
280: mass, are taken into account, the disagreement is not that impressive. For
281: example, the ratio can rise up to 0.69 when $m_s$ is lowered to $70$ MeV.
282:
283: However when combining with the $B^+ \to \pi^+ K^{\ast 0}$ decay, this
284: result shows that there
285: is a tendency that, the penguin-dominant $B \to PV$
286: decay amplitudes are consistently underestimated without annihilation
287: contributions. When they are included,
288: $B^+ \to \pi^+ K^{\ast 0}$ is dominated by $f_{K^\ast}F^{B \pi}m_B^2 a_4
289: +b_3(V, P)$ while $B^0 \to K^+ \rho^-$ is dominated by
290: $f_K A^{B\rho}_0 m_B^2 (a_4-r_\chi^K a_6)+b_3(P,V)$ with
291: \begin{equation}
292: b_3 (M_1, M_2)=\frac{C_F}{N_c^2}\{ C_3 A_1^i (M_1, M_2) + C_5
293: A^i_3 (M_1, M_2) + (C_5+N_c C_6) A^f_3 (M_1, M_2) \}~.
294: \end{equation}
295: Since the penguin
296: terms $a_4 \simeq -0.03$ and $a_4-r_\chi^K a_6 \simeq 0.037$ are of
297: opposite sign, the key observation is that, the annihilation terms
298: $b_3 (V, P)$ and $b_3(P, V)$ are also roughly of the opposite sign,
299: because numerically $(C_5+N_c C_6) A^f_3$ is the dominant
300: term in $b_3$ while $A^f_3(P,V)=-A^f_3(V,P)$ if the annihilation
301: parameter $X_A$ are the same for both channels. So with the inclusion of
302: annihilation topology, QCDF can easily enhance both ratios without fine
303: tuning. For example, using the default parameters given in ref.
304: \cite{QCDF2} but letting the annihilation parameter $\rho_A =1$
305: we can see from Fig. 1 that the ratios are then consistent with
306: experimental observations. We believe that this is a strong indication
307: that annihilation topology probably plays an important role, at least in
308: penguin-dominated $B \to PV$ decays.
309: \begin{figure}[htb]
310: \begin{center}
311: \unitlength 1mm
312: \begin{picture}(160,78)
313: \put(0,0){\includegraphics{fig1l.eps}}
314: \put(80,0){\includegraphics{fig1r.eps}}
315: \end{picture}
316: \caption{ The ratios ${\cal B}(B^+ \to \pi^+ K^{\ast 0})/
317: {\cal B}(B^0 \to \rho^+ \pi^-)$ (left plot) and
318: ${\cal B}(B^0 \to K^+ \rho^-)/{\cal B}(B^0 \to \rho^- \pi^+)$ (right plot)
319: versus the weak annihilation phase
320: $\phi_A$. The default parameters of ref. \cite{QCDF2} are used but letting
321: the annihilation parameter $\rho_A=1$. The dashed lines show the
322: ratios without weak annihilation contributions. The gray areas denote
323: the experimental measurements with $1 \sigma$ error. }
324: \end{center}
325: \end{figure}
326:
327: Similar disagreement can be found in the ratio of
328: $B^0 \to \omega K^0$ over $B^0 \to \pi^+ \rho^-$. But we will not go into
329: details here because $B^0 \to \omega K^0$ decay is quite similar to
330: $B^0 \to K^+ \rho^-$ decay. It may also be interesting to discuss
331: $B \to \phi K$ decays, but again they are quite similar
332: to $B \to \pi K^\ast$ decays if there is no new physics contributions
333: in this decay.
334:
335: \section{Direct CP violations}
336: If one agrees that annihilation terms are very likely important in
337: charmless $B$ decays, one negative point is that, these intractable terms
338: will bring large uncertainties to the theoretical predictions, not
339: only in the branching fractions, but also in direct CP asymmetry .
340: In the following
341: we will try to see whether we can get some interesting relations with a
342: few selective ratios of direct CP asymmetry (DCPVs).
343:
344: Let us first consider the ratio of DCPVs
345: $A_{CP}(\pi^+ \pi^-)$ over $A_{CP}(\pi^- K^+)$. Belle has claimed large
346: DCPV observed in $B^0 \to \pi^+ \pi^-$ decay while BaBar has not confirmed
347: it yet, but both of them are close to a measurement on $A_{CP}(\pi^- K^+)$
348: \cite{Belle,BaBar}
349: \begin{equation}
350: A_{\pi\pi}=\left \{ \begin{array}{ll}
351: 0.58 \pm 0.15 \pm 0.07 & \mbox{(Belle)}~, \\
352: 0.19 \pm 0.19 \pm 0.05 & \mbox{(BaBar)}~;
353: \end{array} \right.
354: A_{\pi K}=\left \{ \begin{array}{ll}
355: (-8.8 \pm 3.5 \pm 1.8)\% & \mbox{(Belle)}~, \\
356: (-10.7 \pm 4.1 \pm 1.2)\% & \mbox{(BaBar)}~.
357: \end{array} \right.
358: \end{equation}
359: Since QCDF generally predicts small DCPVs for charmless $B$ decays, at
360: first glance it seems to contradict the Belle data badly. But notice that
361: annihilation topology and other sources of power corrections could bring
362: large uncertainties, it is of interest to have further investigations. In
363: QCDF, we could define the tree and penguin amplitudes as:
364: \begin{eqnarray}
365: T_{\pi \pi}&=&f_\pi F^{B\pi}m_B^2 (a_1^u + a_4^u +r_\chi a_6^u)+
366: f_B f_\pi f_\pi (b_1+b_3+2 b_4) \equiv f_\pi F^{B\pi}m_B^2 T~,
367: \nonumber \\
368: P_{\pi \pi}&=&f_\pi F^{B\pi}m_B^2 (a_4^c+r_\chi a_6^c)+f_B f_\pi f_\pi
369: (b_3+2 b_4) \equiv f_\pi F^{B \pi}m_B^2 P ~,\nonumber \\
370: T_{\pi K}&=&f_K F^{B\pi}m_B^2 (a_1^u + a_4^u +r_\chi a_6^u)+
371: f_B f_\pi f_K b_3 \equiv f_K F^{B\pi}m_B^2 {\tilde T}~, \nonumber \\
372: P_{\pi K}&=&f_K F^{B\pi}m_B^2 (a_4^c+r_\chi a_6^c)+f_B f_\pi f_K
373: b_3 \equiv f_\pi F^{B \pi}m_B^2 {\tilde P}~.
374: \end{eqnarray}
375: Then the DCPVs can be expressed as
376: \begin{eqnarray}
377: A_{\pi \pi} &\equiv& \frac{{\cal B}({\bar B}^0 \to \pi^+ \pi^-)-
378: {\cal B}(B^0 \to \pi^+ \pi^-)}{{\bar B}^0 \to \pi^+ \pi^-)+
379: {\cal B}(B^0 \to \pi^+ \pi^-)}=\frac{4 \vert V_{ub}V_{ud}
380: V_{cb}V_{cd} T P\vert \sin \gamma \sin \delta }
381: {2 {\cal B}(B \to \pi^+ \pi^- )} ~\mbox{,} \nonumber \\
382: A_{\pi K} &\equiv& \frac{{\cal B}({\bar B}^0 \to \pi^+ K^-)-
383: {\cal B}(B^0 \to \pi^- K^+)}{{\bar B}^0 \to \pi^+ K^-)+
384: {\cal B}(B^0 \to \pi^- K^+)}=-\frac{4 \vert V_{ub}V_{us}
385: V_{cb}V_{cs} {\tilde T} {\tilde P}\vert \sin \gamma \sin {\tilde \delta} }
386: {2 {\cal B}(B \to \pi^+ K^- )} ~\mbox{,}
387: \end{eqnarray}
388: where $\delta=\delta_P - \delta_T$ is the strong phases difference between
389: the penguin and tree amplitudes. It is easy to see that many factors
390: cancel out for the ratio
391: \begin{equation}\label{DCPV}
392: \frac{A_{\pi\pi}}{A_{\pi K}}=-\frac{f_\pi^2}{f_K^2}
393: \frac{{\cal B}(B \to \pi^+ K^- )}
394: {{\cal B}(B \to \pi^+ \pi^- )} \left \vert \frac{TP}
395: {{\tilde T} {\tilde P}} \right \vert
396: \frac{\sin \delta}{\sin {\tilde \delta}}
397: \simeq (-2.7 \pm 0.3)
398: \frac{\sin \delta}{\sin {\tilde \delta}} ~,
399: \end{equation}
400: where the ratio $TP/{\tilde T} {\tilde P}$ has been taken to be $1$, which
401: is a reasonable approximation in QCDF at about $10$ percent level
402: uncertainty. The experimental data on relevant branching ratios
403: \cite{HFAG1} have been used in the above equation, and only the
404: experimental uncertainties are included in the error estimation.
405:
406: At first sign, it is amazing to realize that one would expect very
407: naturally a larger DCPV
408: for $\pi^+ \pi^-$ decay compared with $\pi^- K^+$ decay, if the strong
409: phases of both channels are not too small and of similar magnitude.
410: Notice that the DCPV measurements already told us that at least the strong
411: phase of $\pi^+ \pi^-$ channel should not be small and QCDF predicts quite
412: similar strong phases for both channels if other sources of power
413: corrections are negligible.
414: Actually, it is quite plausible that CP-violating asymmetry in
415: $\pi^+ \pi^-$ decay should be bigger than that for $\pi^- K^+$ by a
416: factor of $3-4$ since the $\pi^+ \pi^-$ decay rate is smaller than
417: the $\pi^- K^+$ decay rate by a similar factor $3-4$, given the fact
418: that in QCDF the CP-violating parts of the tree-penguin interference
419: terms are almost equal as shown above.
420: The current experimental value \cite{Belle,BaBar} for this
421: ratio is
422: \begin{equation}
423: \frac{A_{\pi\pi}}{A_{\pi K}}=\frac{0.42 \pm 0.13}{-0.10 \pm 0.03}
424: =-4.2 \pm 1.8 ~,
425: \end{equation}
426: which is still consistent, within $1 \sigma$ error, with the theoretical
427: estimation of $-2.7 \pm 0.3$ under the assumption that the strong phases
428: $\delta$ and ${\tilde \delta}$ are the same. It should be interesting to
429: keep an eye on this ratio and any significant deviation from theoretical
430: estimation would suggest either different strong phases between $\pi \pi$
431: and $\pi K$ decays or New physics effects affecting one of the channels.
432: It is also interesting to note that inelastic
433: $B \to DD \to \pi \pi$ and $B \to DD_{s} \to K \pi$ rescatterings or
434: charming penguin contributions could
435: also produce a large strong phase, but since these contributions are
436: related by $SU(3)$ symmetry and CKM factor, the interference terms
437: would be essentially equal and our relation for the CP asymmetry applies
438: and we expect large CP asymmetry in $B \to \pi\pi$ as found
439: recently \cite{Isola,Ciuchini}.
440:
441:
442: With the same reasoning, we can get similar relations for other
443: decay channels such as:
444: \begin{eqnarray}
445: \frac{A_{CP}(B^0 \to \rho^+ \pi^-)}{A_{CP}(B^0 \to K^{\ast +} \pi^-)}
446: &\simeq& -\frac{{\cal B}(B^0 \to K^{\ast +} \pi^-)}{{\cal B}(B^0 \to
447: \rho^+ \pi^-)}\frac{f_\rho^2}{f_{K^\ast}^2}\frac{\sin \delta_{\pi \rho}}
448: {\sin \delta_{\pi K^\ast}}\mbox{~,} \nonumber \\
449: \frac{A_{CP}(B^0 \to \rho^- \pi^+)}{A_{CP}(B^0 \to \rho^- K^+ )}
450: &\simeq& -\frac{{\cal B}(B^0 \to \rho^- K^+ )}{{\cal B}(B^0 \to
451: \rho^- \pi^+)}\frac{f_\pi^2}{f_{K}^2}\frac{\sin \delta_{\rho \pi}}
452: {\sin \delta_{\rho K}}\mbox{~.}
453: \end{eqnarray}
454: If the above pairs of strong phases are roughly the same, which is true in
455: QCDF, The DCPVs of the penguin-dominated decays would be about 1.5 times
456: larger than those of their tree-dominated counter parts. Precise
457: measurements on these ratios could help us to get some insight into the
458: strong dynamics of $B$ decays.
459:
460: \section{Conclusion}
461: In this paper, we have used the ratios of the branching fraction of $B^+
462: \to \pi^+ K^{\ast 0}$ to that of $B^0 \to \pi^- \rho^+$, and of
463: $B^0 \to K^+ \rho^-$ to that of $B^0 \to \pi^- \rho^+$, to show clearly
464: and with greatly reduced theoretical uncertainties,
465: that power corrections in charmless B
466: decays are probably large. The key observation is that QCDF predicts
467: the annihilation terms for $B^+\to \pi^+ K^{\ast 0}$ and $B^0 \to K^+ \rho^-$
468: are almost the same magnitude but opposite in sign. The
469: result is that both these two branching fractions could be enhanced by
470: the annihilation contribution to accommodate the experimental data
471: without fine tuning. Assuming the presence of the annihilation
472: contribution, we then derive relations between CP asymmetries for
473: a few selective decay channels and shows that QCDF would
474: predict the direct CP asymmetry
475: of $B \to \pi^+ \pi^-$ to be about 3 times larger than that of
476: $B \to \pi^\pm K^\mp$ with opposite sign, which is consistent, within one
477: sigma error, to the
478: current experimental data $-4.2 \pm 1.8$.
479: Any significant deviation from this prediction
480: would suggest either new physics or possibly the importance
481: of long-distance
482: rescattering effects. We also discuss other similar direct CP ratios which
483: might help us to get some insight into the strong dynamics of charmless B
484: decays.
485:
486: \section*{Note Added}
487: After finishing this paper, we were informed that Eq. (\ref{DCPV}), the
488: relation between the direct CP asymmetries and CP-averaged branching
489: ratios of $B^0 \to \pi^+ \pi^-$ and $B^0 \to \pi^- K^+$, has been derived
490: previously by R. Fleischer \cite{fleischer}.
491:
492: \begin{thebibliography}{99}
493: \bibitem{QCDF}
494: M. Beneke, G. Buchalla, M. Neubert and C.T. Sachrajda, \prl
495: {\bf 83}, 1914 (1999); Nucl. Phys. B {\bf 591}, 313 (2000).
496:
497: \bibitem{QCDF1}
498: M. Beneke, G. Buchalla, M. Neubert and C.T. Sachrajda, Nucl. Phys. B
499: {\bf 606}, 245 (2001).
500:
501: \bibitem{DYZ}
502: D.S. Du, D.S. Yang and G.H. Zhu, Phys. Rev. D {\bf 64}, 014036 (2001);
503: Phys. Lett. B {\bf 509}, 263 (2001).
504:
505: \bibitem{SCET}
506: C.W. Bauer, D. Pirjol, I.Z. Rothstein and I.W. Stewart, hep-ph/0401188.
507:
508: \bibitem{QCDF2}
509: M. Beneke and M. Neubert, Nucl. Phys. B {\bf 675}, 333 (2003).
510:
511: \bibitem{DYZPV}
512: D.S. Du, H.J. Gong, J.F. Sun, D.S. Yang and G.H. Zhu, Phys. Rev. D
513: {\bf 65}, 074001 (2002); Phys. Rev. D
514: {\bf 65}, 094025 (2002); {\bf 66}, 079904(E) (2002).
515:
516: \bibitem{Buras}
517: For a review, see G. Buchalla, A.J. Buras and M.E. Lautenbacher, Rev. Mod.
518: Phys. {\bf 68}, 1125 (1996).
519:
520: \bibitem{HFAG1}
521: J. Alexander, P. Chang and J. Smith, Heavy Flavor Averaging Group,
522: September 2003.
523:
524: \bibitem{Ali}
525: A. Ali, G. Kramer and C.D. L\" u, Phys. Rev. D {\bf 58}, 094009 (1998).
526:
527: \bibitem{pQCD}
528: Y.Y. Keum, H.N. Li and A.I. Sanda, Phys. Rev. D {\bf 63}, 054008 (2001);
529: Phys. Lett. B {\bf 504}, 6 (2001).
530:
531: \bibitem{DYZGA}
532: D.S. Du, J.F. Sun, D.S. Yang and G.H. Zhu, Phys. Rev. D {\bf 67}, 014023
533: (2003).
534:
535: \bibitem{aleksan}
536: R. Aleksan, P.F. Giraud, V. Morenas, O. Pene and A.S. Safir, Phys. Rev. D
537: {\bf 67}, 094019 (2003).
538:
539: \bibitem{cottingham}
540: N.de Groot, W.N. Cottingham and I.B. Whittingham, Phys. Rev. D {\bf 68},
541: 113005 (2003).
542:
543: \bibitem{Pham}
544: T.N. Pham, invited talk at 2nd Workshop on the CKM Unitarity
545: Triangle, Durham, April 2003, hep-ph/0306271.
546:
547: \bibitem{Buras1}
548: A.J. Buras, F.Parodi and A. Stocchi, JHEP {\bf 0301}, 029 (2003).
549:
550: \bibitem{HFAG}
551: H. Evans, A. H\" ocker, O. Long, M. Shapiro and Y. Sakai, Heavy Flavor
552: Averaging Group, Summer 2003.
553:
554: \bibitem{laget}
555: A. H\" ocker, M. Laget, S. Laplace and J.V. Wimmersperg-Toeller, Using
556: Flavor Symmetry to Constraint $\alpha$ from $B \to \rho \pi$, preprint LAL
557: 03-17.
558:
559: \bibitem{PDG}
560: The Particle Data Group, K. Hagiwara {\it et al.}, Phys. Rev. D
561: {\bf 66}, 010001 (2002).
562:
563: \bibitem{Belle}
564: Belle Collaboration, K. Abe {\it et al}., hep-ex/0401029;
565: T. Tomura {it et al}., hep-ex/0305036.
566: \bibitem{BaBar}
567: Ch. Y\`eche (BaBar Collaboration), presented at the International
568: Europhysics Conference on High Energy Physics, Aachen, Germany, July
569: 17-23, 2003.
570:
571: \bibitem{Isola} C. Isola, M. Ladisa, G. Nardulli, T. N. Pham and
572: P. Santorelli, Phys. Rev. {\bf D64}, 014029 (2001);ibid {\bf D65},
573: 094005 (2002).
574:
575: \bibitem{Ciuchini} M. Ciuchini, E. Franco, G. Martinelli, M. Pierini, and
576: L. Silvestrini, Phys. Lett. {\bf B 515}, 33 (2001).
577:
578: \bibitem{fleischer}
579: R. Fleischer, Phys. Lett. B {\bf 459}, 306 (1999);
580: A.J. Buras, R. Fleischer, S. Recksiegel and F. Schwab, hep-ph/0402112 and
581: other references cited therein.
582:
583: \end{thebibliography}
584:
585:
586: \end{document}
587: