1: \documentclass[secnumarabic,amssymb, amsmath,12pt,
2: nofootinbib,tightenlines, nobibnotes, aps, prl]{revtex4}
3: \usepackage{graphicx}
4: \begin{document}
5: %\renewcommand{\baselinestretch}{1.3}
6: \newcommand{\be}{\begin{equation}}
7: \newcommand{\ee}{\end{equation}}
8: \newcommand{\ba}{\begin{eqnarray}}
9: \newcommand{\ea}{\end{eqnarray}}
10: \newcommand{\bea}{\begin{eqnarray*}}
11: \newcommand{\eea}{\end{eqnarray*}}
12: \newcommand{\ts}{\textstyle}
13: %\hoffset=-1.5cm \voffset=-1.5cm
14: %\textheight=20.0cm \textwidth=15.5cm
15:
16: \bigskip
17: \vspace{2cm}
18: \title{Structure-dependent radiative corrections to $\phi \to
19: K^+K^-/K_LK_S$ decays}
20: \vskip 6ex
21: \author{F. V. Flores-Ba\'ez$^1$}
22: \author{G. L\'opez Castro$^2$\footnote{On sabbatical leave from:
23: Departamento de F\'\i sica, Cinvestav, A.P. 14-740, 07000 M\'exico D.F.}}
24: \affiliation{$^1$Departamento de F\'{\i}sica, Cinvestav,
25: Apartado Postal 14-740, 07000 M\'exico D.F., M\'exico \\
26: $^2$Instituto de F\'{\i}sica, Universidad Nacional Aut\'onoma de M\'exico,
27: 04510 M\'exico D.F, M\'exico}
28: \bigskip
29:
30: %\date{6th March 2005}
31: \bigskip
32:
33: \bigskip
34:
35: \begin{abstract}
36: Current predictions for the ratio of $\phi \to K^+K^-/K_LK_S$ decay
37: rates exceed the corresponding experimental value in about five standard
38: deviations. By far, the dominant sources of isospin breaking to this
39: ratio are the phase-space (52$\%$) and the electromagnetic radiative
40: (4.3$\%$, computed within scalar QED) corrections. Here we estimate the
41: effects of the electromagnetic structure of kaons and other
42: model-dependent contributions into the radiative corrections.
43: \end{abstract}
44:
45: \maketitle
46: \bigskip
47:
48: \section{Introduction}
49:
50: Precise knowledge of the ratio for $P^+P^-/P^0\overline{P}^0$ production
51: ($P=K,\ D$ or a $B$ pseudoscalar meson) is very important for
52: measurements of branching fractions and determination of fundamental
53: parameters at the
54: $\phi,\ \psi(3770)$ and $\Upsilon(4S)$ resonances \cite{ratio}.
55: It has long been known that standard theoretical calculations
56: overestimate the isospin breaking corrections to the ratio of $\phi \to
57: K^+K^-/K_LK_S$ decay rates \cite{Bramon:2000qe}. Thus, the isospin
58: breaking corrections induced by the mass difference of kaons and the
59: electromagnetic radiative corrections, change the ratio
60: \be
61: R \equiv \frac{\Gamma(\phi\to K^+K^-)}{\Gamma(\phi\to K^0\overline{K}^0)}
62: \ ,
63: \ee
64: from unity to about 1.59 \cite{Bramon:2000qe}. This value lies
65: 4.7$\sigma$'s
66: above the corresponding experimental value \cite{Yao:2006px}:
67: \be
68: R_{exp}= 1.45 \pm 0.03\ ,
69: \ee
70: quoted by the PDG from an {\it overall fit} to measured branching
71: fractions of $\phi$ decays. This large discrepancy is still missing a
72: convincing
73: explanation.
74:
75: Large isospin breaking corrections to the $\phi \to
76: K^+K^-/K^0\overline{K}^0$ ratio are naturally
77: expected since both decay modes occurs near threshold in a $p$-wave.
78: However, it is rather difficult to identify additional contributions
79: (beyond conventional phase-space and radiative corrections) which
80: may render theory
81: and experiment into agreement. For instance, calculations of the isospin
82: breaking to the ratio of $\phi K^+K^-$ and $\phi K^0\overline{K}^0$
83: coupling constants, carried out in the context of effective hadronic
84: interactions, may increase further the theoretical prediction up to
85: $R^{theory}=1.62$
86: \cite{Bramon:2000qe,Benayoun:2001qz}. Related to this problem, the effects
87: of strong scattering phases on the isospin breaking effects to
88: $P^+P^-/P^0\overline{P}^0$ ($P=K,\ D$ or $B$ meson states) production in
89: $e^+e^-$ annihilation near
90: threshold where considered in Ref. \cite{Dubynskiy:2007xw}.
91: Finally, we should mention that a non conventional mechanism to solve this
92: discrepancy which proposes corrections to the Fermi Golden rule formula
93: of decay rates was discussed in ref. \cite{Fischbach:2001ie}.
94:
95: In this paper we revisit the calculation of radiative corrections to
96: the $\phi \to K^+K^-,\ K^0\overline{K}^0$ decays. The corrections of
97: order $\alpha$ to the decay into charged kaons were calculated long ago by
98: Cremmer and Gourdin \cite{Cremmer:1969er} using scalar QED. Since
99: the electromagnetic structure of kaons is ignored in the framework
100: of scalar QED, no corrections to the neutral mode are induced in this
101: case. In the present paper we focus on the electromagnetic form factors
102: of kaons and compute their effects in the observable $R$ defined in eq.
103: (1). One should note that this calculation was first considered in
104: \cite{Bramon:2000qe}, where a correction of order $2\times
105: 10^{-3}$ was found for the decay rate into charged kaons. In
106: that paper, the correction to the neutral mode is mention to be
107: negligible. We find good agreement with the findings of ref.
108: \cite{Bramon:2000qe} and provide and independent test of the model by
109: considering that the dominant contributions come from the region of
110: validity of the vector meson dominance model. In addition, we
111: also consider the effects of the sub-leading contributions due to
112: hard-photon emission. Since structure-dependent effects have been
113: found to be important in other calculations of radiative corrections, for
114: instance in $\tau \to \pi\nu_{\tau}$ and $\pi \to \mu\nu_{\mu}$ decays
115: \cite{Decker:1994ea}, it is worth seeing if they can be important in
116: the $\phi$ decays of our interest.
117:
118: \section{Isospin breaking corrections to $R$ in scalar QED}
119:
120: We start by defining the tree-level amplitude for the
121: $\phi(q,\eta) \to K(p)\overline{K}(p')$ decays:
122: \be
123: {\cal M}_0(K\overline{K}) =ig^j(p-p')\cdot \eta,\ \
124: \ee
125: where $g^+$ ($g^0$) refers to the strong coupling constant for the
126: $K^+K^-$ ($K^0\overline{K}^0$) final state, and $\eta^{\mu}$ is the
127: polarization four-vector of the $\phi$ vector meson ($q\cdot \eta=0$).
128: The corresponding ratio of decay rates reads:
129: \be
130: R^{theory}_0= \left(\frac{g^+}{g^0} \right)^2\cdot \frac{v_+^3}{v_0^3} \ .
131: \ee
132: where $v_+$ ($v_0$) denotes the velocity of the charged (neutral) kaons in
133: the rest frame of $\phi$ meson. In the limit of isospin symmetry,
134: $g^+=g^0$ and
135: $m_{K^+}=m_{K^0}$, thus the ratio of decay rates becomes $R^{theory}_0=1$.
136:
137: The largest source of isospin breaking corrections to $R^{theory}_0$
138: arise from the mass difference of charged and neutral kaons
139: [$(v_+/v_0)^3=1.5225$]. The next most important breaking effect comes from
140: radiative corrections. The corrections of $O(\alpha)$ to the
141: $\phi \to K^+K^-$ decay rate, $\delta_{QED}=
142: 2\delta^v_{point}+\delta^r$, were calculated in Ref.
143: \cite{Cremmer:1969er} in
144: the context of scalar QED. They include the sum of virtual corrections
145: ($\delta^v_{point}$) to the non-radiative amplitude and the real
146: ($\delta^r$) photonic corrections. The
147: virtual corrections are divergent for infrared photons but are finite in
148: the ultraviolet region owing to the Ward identity satisfied by the vertex
149: and self-energy corrections.
150: The real photon corrections $\delta^r$ contains an infrared divergent
151: piece $\delta^r_I$ and a regular contribution $\delta^r_R$, namely
152: $\delta^r=\delta^r_I+\delta^r_R$. For consistency $\delta^r_I$ is
153: computed by summing over the longitudinal and
154: transverse degrees of freedom of a massive photon \cite{Coester}.
155: The sum of virtual and soft-photon corrections, $2\delta^v_{point}+\delta^r_i$
156: is explicitly free from infrared divergences, as it should be. The explicit
157: expressions of $\delta_{point}^v$ and $\delta_I^r$ can be found in Ref.
158: \cite{baez2007}.
159:
160: The calculation of the regular term $\delta^r_R$ can be done in a
161: numerical way. It can receive contributions from intermediate states
162: other than $K^{\pm}$ mesons (for instance, $\phi \to K^+K^{*-} \to
163: K^+K^-\gamma$). However, these model-dependent corrections are expected to
164: be very small either for charged (which we found to be $-7.1
165: \times 10^{-8}$
166: for the contribution of $K^*$ intermediate state) or
167: neutral \cite{KLOE} channels because
168: $\omega^{max}$ is very small compared to the masses of other hadrons.
169: Thus we obtain:
170: \be
171: \delta^r_R= 7.96 \times 10^{-5} \ .
172: \ee
173:
174: When we include the phase space corrections, $(v^+/v^-)^3$, and the
175: radiative corrections of scalar QED, $\delta_{QED}=2\delta^v_{point}+
176: \delta^r_I+\delta^r_R = 0.04315$, one gets:
177: \be
178: R^{theory}=R^{theory}_0\left(1+\frac{}{}\delta_{QED}\right)= 1.588
179: \ ,
180: \ee
181: which is about 5$\sigma$'s above the experimental value shown in Eq. (2).
182:
183: \section{Structure-dependent effects in radiative corrections to $R$}.
184:
185: Measurements of the electromagnetic interactions of kaons at low
186: \cite{ff_low}
187: and intermediate \cite{cmd2} energies exhibit an structure which
188: can be well described within a vector dominance model. As usual, we
189: define the kaon electromagnetic vertex $\gamma^*(k) \to
190: K(p)\overline{K}(p')$:
191: \be
192: -ieF_{K}(k^2)(p-p')_{\mu}\ ,
193: \ee
194: where $k^2=(p+p')^2$ is the squared momentum of the virtual photon.
195:
196: Following refs. \cite{cmd2,Decker:1994ea} we write the form factors
197: in the vector dominance model as:
198: \be
199: F_K(k^2)= \sum_{V=\rho,\omega,\phi} \frac{g_{VK\bar{K}}}{f_V}\cdot
200: \frac{m_V^2}{\hat{m}_V^2-k^2} \ ,
201: \ee
202: where $em_V^2/f_V$ denotes the photon--vector-meson couplings, $m_V$ is
203: the mass of
204: intermediate vector meson and $\hat{m}_V^2=m_V^2-im_V\Gamma_V
205: \theta(k^2-k^2_{threshold})$ for timelike $k^2$ ($k^2_{threshold}$ is
206: the square of virtual momentum that allows to generate imaginary parts
207: in the one-loop corrections to the $V$ meson propagator). At $k^2=0$, the
208: form
209: factors are normalized to the electric charges of kaons and the following
210: condition must be satisfied:
211: \ba
212: && \sum_{V=\rho,\omega,\phi} \frac{g_{VK^+K^-}}{f_V} =1 \\
213: && \sum_{V=\rho,\omega,\phi} \frac{g_{VK^0\overline{K}^0}}{f_V} =0 \ .
214: \ea
215: A simplifying assumption (also used in the experimental analysis
216: of Ref. \cite{cmd2}) is implemented by using the SU(3)-invariant
217: Lagrangian for the $VPP'$ interaction which gives the couplings:
218: \ba
219: g_{\rho K^{+}K^{-}}&=&-g_{\rho K^{0}\bar{K^{0}}}=
220: \frac{1}{2}G_{VPP'}\nonumber \\
221: g_{\omega K^{+}K^{-}}&=& g_{\omega
222: K^{0}\bar{K^{0}}}=\frac{\sqrt{3}}{2}G_{VPP'}\sin \theta_V \\
223: g_{\phi K^{+}K^{-}}&=& g_{\phi
224: K^{0}\bar{K^{0}}}=\frac{\sqrt{3}}{2}G_{VPP'}\cos \theta_V \nonumber \ ,
225: \ea
226: where $\theta_V$ is the $\omega-\phi$ mixing angle (we will use here
227: $\tan\theta_V=1/\sqrt{2}$ as in Ref. \cite{cmd2}) and $G_{VPP'}$ is the
228: SU(3)-invariant strong coupling constant. Solving Eqs. (9,10) with the
229: constraints given in (11) leads to:
230: \ba
231: T^+_{\rho}&=&-T^0_{\rho}=\frac{1}{2}\nonumber \\
232: T^+_{\omega}&=&T^0_{\omega}= \frac{f_{\phi}}{2\left[f_{\phi}+
233: \sqrt{2}f_{\omega} \right]} \nonumber\\
234: T^+_{\phi}&=&T^0_{\phi}=\frac{f_{\omega}}{\sqrt{2}\left[f_{\phi}+
235: \sqrt{2}f_{\omega} \right]}\ ,
236: \ea
237: where we have introduced the notation $T^{+,0}_V\equiv
238: g_{VK^{+,0}K^{-,0}}/f_V$ for $V=\rho,\ \omega,\ \phi$ mesons.
239: Finally, if we specify the values of the electromagnetic couplings
240: ($T_{\omega}=0.1790,\ T_{\phi}=0.3210$, using the measured rates of
241: $\omega,\ \phi\to e^+e^-$ decays) the form factors do not contain
242: further free parameters.
243:
244: Now, let us write the form factors by separating explicitly the point
245: and structure-dependent contributions:
246: \ba
247: F_{K^+}(k^2)&=& 1 + \sum_{V=\rho,\omega,
248: \phi}T^+_V \left(\frac{k^2+m_V^2-\hat{m}_V^2}{\hat{m}_V^2-k^2}\right) \
249: ,\\
250: F_{K^0}(k^2)&=& 0+ \sum_{V=\rho,\omega,\phi}
251: T^0_V\left(\frac{k^2+m_V^2-\hat{m}_V^2}{\hat{m}_V^2-k^2}\right) \ .
252: \ea
253: Note that for small values of $k^2$ ($k^2 \leq k^2_{threshold}$), the
254: numerator of the second term in the r.h.s of the above equations is
255: linear in $k^2$
256:
257: The Feynman diagrams contributing to the virtual corrections within
258: this meson dominance model are displayed in Figure 1 for the generic $\phi
259: \to K\overline{K}$ decay. The contributions of self energies
260: should be added to these virtual corrections.
261: Since the structure-dependent piece of form factors --second term in
262: the r.h.s. of Eqs. (13)-(14)-- falls linearly as $k^2$ approaches zero ,
263: their
264: contributions to radiative corrections are free from infrared
265: divergences. As in the corresponding case of scalar QED, the result is
266: also free of ultraviolet divergences.
267:
268: \begin{figure}
269: \includegraphics[width=12cm]{phi1.eps}
270: \vspace{-0.0cm}
271: \caption{Feynman graphs for virtual corrections to $\phi \to
272: K\overline{K}$ decays within the vector meson dominance model.}
273: \end{figure}
274:
275:
276: When we insert the form factors in the calculation of the virtual
277: corrections, the expressions for the one-loop amplitudes are the same as
278: in the point case but with an additional factor $|F_K(k^2)|^2$ in the
279: integrand over virtual momenta. The structure-dependent pieces of the
280: radiative corrected amplitudes (due to $T_V^{+,0}\not = 0$) become:
281: \ba
282: {\cal M}_{\small SD}^v ({\small K^+K^-})\!\! &=&\!\! {\cal
283: M}_0({\small K^+K^-})
284: \!\left(\frac{\alpha}{4\pi}\right)\! \left\{
285: [4.37-i0.44]\!\! \frac{}{}
286: (T^+_{\rho})^2+[4.39+i0.04]\!\! (T^+_{\omega})^2 +
287: \![5.84+i0.02]\!(T^+_{\phi})^2
288: \right.
289: \nonumber
290: \\ && \ \ \ \ \ \ \ \ \left. + [10.25-i0.34]T^+_{\rho}T^+_{\phi}
291: +
292: [9.05-i0.40]T^+_{\rho}T^+_{\omega}
293: + [10.27-i0.03]T^+_{\omega}T^+_{\phi} \right. \nonumber \\
294: && \ \ \ \ \ \ \ \ \left. +[- 6.30+i0.53] T^+_{\rho} \frac{}{}
295: +[-6.31+i0.03]T^+_{\omega}+[-7.94+i0.01]T^+_{\phi} \right\} \\
296: &=& {\cal M}_0(K^+K^-)\times \delta^+_{VMD}\ ,
297: \ea
298: and
299: \ba
300: {\cal M}_{\small SD}^v (K^0\overline{K}^0)\!\! &=&\!\! {\cal M}_0
301: (K^0\overline{K}^0)\! \left(\frac{\alpha}{4\pi} \right)\!
302: \left\{ [4.34-i0.46] (T^0_{\rho})^2 \frac{}{} +[4.36+i0.02]
303: (T^0_{\omega})^2
304: + [5.81+i0.01](T^0_{\phi})^2 \right.
305: \nonumber
306: \\ && \ \ \ \ \ \ \ \ \left. + [10.16-i0.34]T^0_{\rho}T^0_{\phi}
307: +
308: [8.96-i0.40]T^0_{\rho}T^0_{\omega} \frac{}{}
309: + [10.18-i0.03]T^0_{\omega}T^0_{\phi} \right\} \\
310: &=& {\cal M}_0(K^0\overline{K}^0) \times \delta^0_{VMD}\ .
311: \ea
312: where ${\cal M}_0(K\overline{K})$ denote the amplitudes at the tree-level
313: defined in Eq. (3).
314: In the above results, the terms linear in $T_i^{+,0}$ appear from the
315: interference between the point and the structure-dependent terms in the
316: square of Eqs. (13),(14). Note that the imaginary parts arising
317: from the finite decay width of vector mesons give a small
318: contribution in radiative corrections. Finally, when
319: we insert the numerical values for
320: the couplings constants in the above expressions, we get:
321: \ba
322: \delta^+_{VMD} &=& -1.13 \times 10^{-3}\ ,\\
323: \delta^0_{VMD} &=& -1.37 \times 10^{-5} .
324: \ea
325: Thus, once we include the effects induced by the electromagnetic structure
326: of kaons, Eq. (5) gets modified to:
327: \ba
328: R^{theory}&=&
329: R_0^{theory}\left(1+\delta_{QED}\frac{}{}
330: +2[\delta^+_{VMD}-\delta^0_{VMD}]\right) \nonumber \\
331: &=& 1.585\ .
332: \ea
333: Therefore, the structure-dependent effects in virtual radiative
334: corrections are tiny but larger than hard-photon contributions (eq. 5).
335:
336: In view of the above result, one may wonder how appropriate is using the
337: meson dominance model, Eqs. (13),(14), in the full range of virtual photon
338: momenta $k$. To address this
339: question we introduce a modified photon propagator according to (see for
340: example Ref. \cite{Decker:1994ea}):
341: \be
342: \frac{1}{k^2} \to \frac{1}{k^2}\cdot \frac{\mu^2}{\mu^2-k^2}\ .
343: \ee
344: where $\mu$ is a cutoff scale which suppresses the contributions of high
345: $k^2$ values. Clearly, in the limit $\mu^2 \to \infty$ one should recover
346: the previous results, while for finite values of
347: $\mu^2$ the
348: contributions of very high $k^2$ get suppressed.
349: \begin{figure}
350: \includegraphics[width=11cm, angle=270]{phi4.eps}
351: \vspace{-0.0cm}
352: \caption{Structure-dependent virtual corrections $2\delta^+_{VMD}$ to the
353: decay rate of $\phi \to K^+K^-$ decays as a
354: function of the cutoff scale $\mu$.}
355: \end{figure}
356:
357: Using in loop calculations the modified photon propagator of Eq. (22),
358: we can compute again the structure-dependent parts of radiative
359: corrections. Following the recipe given in Eqs. (111-113) of reference
360: \cite{Decker:1994ea}, we can easily compute these corrections in the case
361: of $\phi \to K^+K^-$ decays for several values of the cutoff $\mu$. Our
362: results, displayed in Figure 2, show that the values of
363: $\delta^+_{VMD}(\mu)$ reach the previous value already for low values of
364: $\mu$, {\it
365: i.e.} the most important contributions arise in the region of validity of
366: the model (intermediate energies).
367:
368: \section{Summary and conclusions}
369:
370: We have computed the radiative corrections to the ratio $R=\Gamma(\phi
371: \to K^+K^-)/\Gamma(\phi\to K_LK_S)$ by taking into account the
372: electromagnetic structure of the $K$ mesons within a vector-meson
373: dominance model. Our results are in agreement with previous calculations
374: of ref. \cite{Bramon:2000qe} for the charged kaon channel. Although tiny,
375: the structure-dependent corrections are larger than hard real photon
376: corrections. We have shown that the main contribution of
377: structure-dependent corrections arise from the region of energies where
378: the vector-meson dominance of kaon form factors is expected to hold.
379:
380: Clearly, structure-dependent corrections do not resolve the discrepancy
381: between theory and the experimental value of $R$ (see eq. 2 above). It
382: is worth noticing,
383: however, that a weighted average of direct measurements of $\phi \to
384: K\overline{K}$ decay rates gives $R_{exp}=1.49\pm 0.05$
385: \cite{Yao:2006px}, which is only 1.9$\sigma$ below the theoretical
386: prediction (eq. 21). This weighted average of direct measurements is
387: dominated by results of the CMD2 Collaboration \cite{cmd2}. It may happen
388: that this result of direct measurements will turn out to be more reliable
389: than the {\it indirect} value obtained from a constrained fit which
390: requires that the
391: sum of dominant decay modes saturates the total decay width of the $\phi$
392: meson \cite{note}.
393:
394: If the experimental value shown in Eq. (2) is confirmed by new
395: measurements, still another possibility to solve the discrepancy
396: can be provided by the short-distance effects in radiative
397: corrections. Short-distance corrections can be induced by
398: highly virtual photons coupled to $q\bar{q}$ quark pairs ($q=u,\ d,\
399: s$), in other words, by the quark components of the
400: photon wavefunction \cite{Anisovich:1996hh}. Since the light
401: $q\bar{q}$ pair required to produce $K^+K^-/K^0\overline{K}^0$ in
402: $\phi$ decays is $u\bar{u}/d\bar{d}$, additional isospin breaking
403: correction to $R$ may be induced. In the case of radiative corrections
404: to semileptonic weak decays, short-distance effects provide a
405: universal
406: correction (they do not break isospin symmetry) to all semileptonic
407: processes \cite{sdc} since tdhey affect only the underlying quark decay
408: process (isospin breaking arises only from long-distance corrections).
409: Thus, although short-distance corrections could eventually contribute
410: to isospin breaking corrections in $R$ we restrict here only to long-
411: and intermediate-distance radiative corrections.
412:
413: \
414: {\bf Acknowledgements}
415:
416: The authors acknowledge partial financial support from Conacyt.
417:
418: \bigskip
419:
420:
421: \begin{thebibliography}{40}
422: \bibitem{ratio}
423: S.~B.~Athar {\it et al.} [CLEO Collaboration],
424: Phys.\ Rev.\ D {\bf 66}, 052003 (2002)
425: [arXiv:hep-ex/0202033]; B.~Aubert {\it et al.} [BABAR Collaboration],
426: Phys.\ Rev.\ D {\bf 69}, 071101 (2004)
427: [arXiv:hep-ex/0401028]; M.~Ablikim {\it et al.} [BES Collaboration],
428: Phys.\ Rev.\ Lett.\ {\bf 97}, 121801 (2006)
429: [arXiv:hep-ex/0605107].
430: \bibitem{Bramon:2000qe}
431: A.~Bramon, R.~Escribano, J.~L.~Lucio M. and G.~Pancheri,
432: Phys.\ Lett.\ B {\bf 486}, 406 (2000)
433: [arXiv:hep-ph/0003273].
434: \bibitem{Yao:2006px}
435: W.~M.~Yao {\it et al.} [Particle Data Group],
436: J.\ Phys.\ G {\bf 33}, 1 (2006).
437: \bibitem{Benayoun:2001qz}
438: M.~Benayoun and H.~B.~O'Connell,
439: Eur.\ Phys.\ J.\ C {\bf 22}, 503 (2001)
440: [arXiv:nucl-th/0107047].
441: \bibitem{Dubynskiy:2007xw}
442: S.~Dubynskiy, A.~Le Yaouanc, L.~Oliver, J.~C.~Raynal and M.~B.~Voloshin,
443: Phys.\ Rev.\ D {\bf 75}, 113001 (2007)
444: [arXiv:0704.0293 [hep-ph]].
445: \bibitem{Fischbach:2001ie}
446: E.~Fischbach, A.~W.~Overhauser and B.~Woodahl,
447: Phys.\ Lett.\ B {\bf 526}, 355 (2002)
448: [arXiv:hep-ph/0112170].
449: \bibitem{Cremmer:1969er}
450: E.~Cremmer and M.~Gourdin,
451: Nucl.\ Phys.\ B {\bf 9}, 451 (1969).
452: \bibitem{Decker:1994ea}
453: R.~Decker and M.~Finkemeier,
454: Nucl.\ Phys.\ B {\bf 438}, 17 (1995)
455: [arXiv:hep-ph/9403385].
456: \bibitem{Coester}
457: F. Coester, Phys. Rev. {\bf 83}, 798 (1951); T. Kinoshita and A. Sirlin,
458: Phys. Rev. {\bf 113}, 1652 (1959).
459: \bibitem{baez2007}
460: F. V. Flores-Ba\'ez, G. L\'opez Castro and G. Toledo S\'anchez, Phys.
461: Rev. {\bf D76}, 096010 (2007).
462: \bibitem{KLOE}
463: F. Ambrosino {\it et al.} [KLOE Collaboration], arXiv:0707.4148v2
464: [hep-ex] have found that $B(\phi \to K^0\overline{K}^0\gamma) \leq 1.8
465: \times 10^{-8}$ at the
466: 90\% c.l.
467: \bibitem{ff_low}
468: S.~R.~Amendolia {\it et al.},
469: Phys.\ Lett.\ B {\bf 178}, 435 (1986); E.~B.~Dally {\it et al.},
470: Phys.\ Rev.\ Lett.\ {\bf 45}, 232 (1980).
471: \bibitem{cmd2}
472: R. R. Akhmetshin et al, [CMD2 Collaboration], Phys. Lett. {\bf B364}, 199
473: (1995).
474: \bibitem{note}
475: Constrained fits used to obtain the semileptonic rates of $K$ mesons in
476: editions of the PDG prior to 2004 underestimated their values by about
477: 5\%. Later results for these branching fractions obtained by KLOE,
478: KTeV, BNL E865, ISTRA and NA48 experiments have confirmed that weigthed
479: averages of direct measurements and indirect values obtained from
480: constrained fits can lead to different results (see for instance
481: \cite{Yao:2006px}).
482: \bibitem{Anisovich:1996hh}
483: V.~V.~Anisovich, D.~I.~Melikhov and V.~A.~Nikonov,
484: Phys.\ Rev. \ D {\bf 55}, 2918 (1997);
485: A.~V.~Anisovich, V.~V.~Anisovich, L.~G.~Dakhno, V.~A.~Nikonov and
486: A.~V.~Sarantsev,
487: Phys.\ Atom.\ Nucl.\ {\bf 68}, 1830 (2005)
488: [Yad.\ Fiz.\ {\bf 68}, 1892 (2005)].
489: \bibitem{sdc}
490: A.~Sirlin,
491: Rev.\ Mod.\ Phys.\ {\bf 50}, 573 (1978)
492: [Erratum-ibid.\ {\bf 50}, 905 (1978)];
493: A.~Sirlin,
494: Nucl.\ Phys.\ B {\bf 196}, 83 (1982);
495: W.~J.~Marciano and A.~Sirlin,
496: Phys.\ Rev.\ Lett.\ {\bf 61}, 1815 (1988).
497:
498: \end{thebibliography}
499:
500:
501:
502: \end{document}
503:
504: