1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: \documentclass[12pt]{elsart}
3: \usepackage{amsmath,amssymb}
4: \usepackage{graphicx,psfrag,here,epsfig}
5: \usepackage{pstricks}
6: \usepackage{pst-node}
7: \usepackage{epsf}
8: \usepackage{citesort}
9: \allowdisplaybreaks[1]
10:
11: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
12: \begin{document}
13: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
14: \newcommand{\co}{\; \; ,}
15: \def\words#1{\mbox{\small{\,#1}}}
16: \def\bea{\begin{eqnarray}}
17: \def\eea{\end{eqnarray}}
18: \def\eq{\begin{eqnarray}}
19: \def\en{\end{eqnarray}}
20: \def\be{\begin{equation}}
21: \def\ee{\end{equation}}
22: \newcommand{\ed}{\end{document}}
23:
24: \newcommand{\nnnl}{\nonumber\\}
25: \newcommand{\fs}{\, . \,}
26: \def\query#1{\marginpar{\begin{flushleft}\footnotesize#1\end{flushleft}}}%
27:
28: \runauthor{Colangelo, Gasser, Kubis and Rusetsky}
29: \renewcommand{\theequation}{\arabic{equation}}
30: \begin{frontmatter}
31:
32: \begin{flushright}
33: HISKP-TH-06/09
34: \end{flushright}
35:
36: \title{\Large\bf Cusps in \boldmath{$K\rightarrow 3\pi$} decays}
37:
38: \author[Bern]{G.~Colangelo,}
39: \author[Bern]{J.~Gasser,}
40: \author[Bonn]{B.~Kubis,}
41: \author[Bonn]{A.~Rusetsky\thanksref{Tbilisi}}
42:
43: \address[Bern]{Institute for Theoretical Physics, University of Bern,
44: Sidlerstr. 5, CH-3012 Bern, Switzerland}
45: \address[Bonn]{Helmholtz-Institut f\"ur Strahlen- und Kernphysik,
46: Universit\"at Bonn, Nussallee~14-16, D-53115 Bonn, Germany}
47:
48: \thanks[Tbilisi]{On leave of absence from:
49: High Energy Physics Institute, Tbilisi State University,
50: University St.~9, 380086 Tbilisi, Georgia.}
51:
52:
53: \begin{abstract}
54: The pion mass difference generates a pronounced cusp in
55: $K\rightarrow3\pi$ decays. As has recently been pointed out by Cabibbo
56: and Isidori,
57: an accurate measurement of the cusp may allow one to pin down the S-wave
58: $\pi\pi$ scattering lengths to high precision. Here, we present and
59: illustrate an effective field theory framework that allows one to determine
60: the structure of this cusp in a straightforward manner. The
61: strictures imposed by analyticity and unitarity are respected
62: automatically.
63: \end{abstract}
64:
65: \begin{keyword}
66: Chiral symmetries\sep Analytic properties of the $S$-matrix
67: \sep Decays of $K$-mesons \sep Meson-meson interactions
68:
69: \PACS 11.30.Rd\sep 11.55.Bq\sep 13.20.Eb \sep 13.75.Lb
70: \end{keyword}
71:
72:
73: \end{frontmatter}
74:
75:
76: \noindent{\bf 1.}
77: The S-wave $\pi\pi$ scattering lengths $a_0,a_2$
78: have been predicted with
79: percent level accuracy some time ago
80: \cite{Colangelo:2000jc,Colangelo:2001df},
81: and first steps for an experimental verification of this prediction
82: have been performed in Ref.~\cite{Pislak:2003sv,Adeva:2005pg}.
83: Recently, it has been pointed out by
84: Cabibbo and Isidori \cite{Cabibbo:2004gq,Cabibbo:2005ez}
85: that isospin violating effects
86: generate a pronounced cusp in $K\rightarrow 3\pi$ decays whose experimental
87: investigation may allow one to
88: determine the combination $a_0-a_2$ with high precision. A
89: first analysis of data based on this proposal
90: has appeared \cite{Batley}.
91: (The strong impact of the unitarity cusp on $\pi^0\pi^0$ scattering
92: close to threshold was already mentioned in~\cite{MMS}.)
93: In order for this program to be carried through successfully,
94: one needs to determine
95: the structure of the cusp with a precision that matches the experimental
96: accuracy. In view of the large amount of data available
97: \cite{Batley}, this
98: is a considerable task. In the present letter, we present a method which
99: -- we believe -- has the potential to achieve this goal.
100:
101: In \cite{Cabibbo:2004gq,Cabibbo:2005ez}, the structure of the singularity
102: at the cusp is investigated using unitarity, analyticity and cluster
103: decomposition properties of the $S$-matrix. In addition, an approximation
104: scheme is used, which consists in expanding the decay amplitude in powers
105: of $\pi\pi$ scattering lengths. The latest work \cite{Cabibbo:2005ez}
106: retains effects up to order (scattering lengths)$^2$ and omits explicit
107: electromagnetic effects. Here, we present a Lagrangian framework, which
108: automatically satisfies unitarity and analyticity constraints and, in
109: addition, allows one to include electromagnetic contributions in a standard
110: manner. Specifically, we use a non-relativistic framework that has already
111: proven to be useful in the description of bound states
112: \cite{nrqfta,nrqft1,nrqft2,nrqft3,nrqft4,nrqft5,nrqft6,nrqft7,nrqft8,nrqft9,nrqft10,nrqft11,nrqft12,nrqft13,nrqft14,nrqft15,nrqftz}.
113: In this framework -- in contrast to relativistic field theory -- an
114: expansion in powers of scattering lengths emerges automatically from the
115: loop expansion. Moreover, it is a scheme that provides a proper power
116: counting. [The low-energy expansion proposed here is closely related to
117: early work performed in the sixties by many authors (for a review see
118: \cite{anisovich}), who used $S$-matrix methods to investigate the
119: production of particles -- in parti\-cu\-lar also in $K\to3\pi$ decays --
120: in the threshold region. The method presented here may be considered an
121: effective field theory realisation of these approaches.]
122:
123: The strategy that we follow in this letter is the following. First we
124: write down the most general non-relativistic Lagrangian relevant for this
125: decay and determine all
126: four-pion couplings therein through a matching procedure in terms of the
127: threshold parameters of $\pi\pi$ scattering. In the next step, we evaluate
128: the $K\to 3\pi$ decay amplitudes to two loops. This results in an explicit
129: representation of the $S$-matrix elements which is valid in the whole decay
130: region and slightly beyond. We propose to analyse the experimental data
131: with the use of this representation. We plan to include real and virtual
132: photon corrections at a later stage. On the other hand, we do keep the pion
133: and kaon masses at their physical values, which is a fully consistent
134: procedure in this framework. This guarantees that the various branch
135: points and cusps occur at the proper place in the Mandelstam plane.
136:
137: We display the results without a detailed derivation,
138: which will be provided in a forthcoming publication~\cite{big}.
139:
140: \begin{sloppypar}
141: \noindent{\bf 2.}
142: We consider the neutral and charged decay modes
143: $K^+(p_K)\to\pi^0(p_1)\pi^0(p_2)\pi^+(p_3)$ and
144: $K^+(p_K)\to\pi^+(p_1)\pi^+(p_2)\pi^-(p_3)$. The kinematical variables
145: are defined as usual: $s_i=(P_K-p_i)^2$ with $p_i^2=M_i^2\, ,~i=1,2,3$, where
146: $M_{\pi^+}\doteq M_\pi$ and $M_{\pi^0}$ denote the masses of the charged and
147: neutral pions, respectively, and $\Delta_\pi=M_\pi^2-M_{\pi^0}^2\neq 0$.
148: In the centre-of-mass (CM) frame $P_K=(M_K,{\bf 0})$, with $M_K$ the
149: charged kaon mass,
150: \end{sloppypar}
151: \be\label{eq:kin1}
152: p_i^0=\frac{M_K^2+M_i^2-s_i}{2M_K}\, ,\quad
153: {\bf p}_i^2=\frac{\lambda(M_K^2,M_i^2,s_i)}{4M_K^2}\, ,
154: \ee
155: where $\lambda(x,y,z)=x^2+y^2+z^2-2xy-2xz-2yz$ is the triangle
156: function. Below we also use the velocities $v_{jk}$ and kinetic energies $T_i$,
157: \be\label{eq:kin2}
158: v_{jk}^2(s_i)=\frac{\lambda(s_i,M_j^2,M_k^2)}{s_i^2}\, ,\quad
159: T_i=p_i^0-M_i\,\,.
160: \ee
161: \noindent {\bf 3.}
162: A non-relativistic approach to describe decays $K\to 3\pi$ can be
163: justified, if the typical kinetic energies of the decay products
164: are much smaller than the masses. This can be achieved by considering a world
165: where the strange quark mass is taken to be smaller than its actual value.
166: Then, a consistent counting scheme arises, if one
167: introduces a formal parameter $\epsilon$ and counts $T_i$ as a term
168: of order $\epsilon^2$, the pion momenta as order $\epsilon$,
169: whereas the pion and kaon masses are counted as $O(1)$.
170: From $\sum_iT_i=M_k-\sum_iM_i$, one concludes that the difference
171: $M_K-\sum_iM_i$ is then a quantity of order $\epsilon^2$ as well.
172: The pion mass difference $\Delta_\pi$ is also counted as
173: $O(\epsilon^2)$. The effective field theory
174: framework, which we construct below, enables one
175: to obtain a systematic expansion of the amplitudes in $\epsilon$.
176: For sufficiently small $m_s$, the expansion in $\epsilon$ is expected to work
177: very well.
178:
179: Together with $\epsilon$, our theory has another expansion parameter,
180: na\-me\-ly a characteristic size of the $\pi\pi$ threshold parameters,
181: which we denote generically as $a$. In particular, the amplitudes in the
182: non-relativistic framework are given in form of an expansion in several
183: low-energy couplings $C_i,D_i$, which can be expressed in terms of the
184: threshold parameters of the relativistic $\pi\pi$ scattering amplitude. We
185: expect the expansion in $a$ to converge rapidly because of the smallness of
186: the scattering lengths. These two expansions are correlated \cite{big}:
187: because one-loop integrals are of order $\epsilon$, adding a pion loop
188: generated by a four-pion vertex increases both the order in $a$ and in
189: $\epsilon$ by one. A consistent power counting is achieved: to a given
190: order in $a$ and in $\epsilon$, a well-defined finite number of diagrams
191: contribute.
192:
193: Increasing now $m_s$ to its physical value again,
194: convergence in the $\epsilon$-expansion
195: is not {\it a priori} evident,
196: because $T_i/M_i$ can become as large as 0.4, and the
197: corresponding maximal momentum ${\bf {|p|}}$ is then
198: not much smaller than the pion mass.
199: However, let us note the non-relativistic framework is only
200: used to correctly reproduce
201: the non-analytic behaviour
202: of the decay amplitudes in the kinematical variables $s_1,s_2,s_3$, and to
203: thus provide a parametrisation consistent with unitarity and analyticity
204: -- a trivial polynomial part in the amplitudes can be removed by a
205: redefinition of the couplings in the Lagrangian.
206: In addition, from the analysis of the experimental data one
207: knows~\cite{Cabibbo:2005ez} that in the
208: whole physical region the real part of the decay
209: amplitude can be well approximated by a polynomial in $s_1,s_2,s_3$ with
210: a maximum degree $2$. We interpret this fact as an experimental indication
211: for a good convergence of the $\epsilon$-expansion for the
212: quantities one is interested in.
213:
214: We now proceed with the construction of the
215: non-relativistic Lagrangian framework. In the decay amplitudes,
216: we shall restrict ourselves to terms up to and including
217: $O(\epsilon^2,a\epsilon^3,a^2\epsilon^2)$.
218:
219: \noindent {\bf 4.}
220: It is convenient to formulate the non-relativistic approach in a manner
221: that describes the two-particle subsystems in a manifestly covariant
222: way~\cite{big}. We start with the $\pi\pi$ interaction and consider the
223: following five channels in $\pi^a\pi^b\to\pi^c\pi^d$: $(ab;cd)=$
224: (1)~$(00;00)$, (2)~$(+0;+0)$, (3)~$(+-;00)$, (4)~$(+-;+-)$, (5)~$(++;++)$.
225: The Lagrangian takes the form
226: \be\label{eq:l_pipi}
227: {\mathcal L}_{\pi\pi}=2\sum_\pm\Phi_\pm^\dagger W_\pm\bigl(i\partial_t-
228: W_\pm\bigr)\Phi_\pm
229: +2\Phi_0^\dagger W_0\bigl(i\partial_t- W_0\bigr)\Phi_0
230: +\sum_{i=1}^5{\mathcal L}_{i}\, ,
231: \ee
232: where $\Phi_i$ is the non-relativistic pion field operator, $
233: W_\pm=\sqrt{M_\pi^2-\triangle}$, $ W_0=\sqrt{M_{\pi^0}^2-\triangle}$, with
234: $\triangle$ the Laplacian, and
235: \eq\label{eq:l_1-5}
236: \hspace*{-15mm}{\mathcal L}_{i}&=&x_iC_i
237: \bigl(\Phi_c^\dagger\Phi_d^\dagger\Phi_a\Phi_b+h.c.\bigr)
238: +x_iD_i\Bigl\{\bigl( W_c\Phi_c^\dagger W_d\Phi_d^\dagger\Phi_a\Phi_b
239: +\Phi_c^\dagger\Phi_d^\dagger W_a\Phi_a W_b\Phi_b
240: \nonumber\\[1mm]
241: \hspace*{-1.mm}&+&\nabla\Phi_c^\dagger\nabla\Phi_d^\dagger\Phi_a\Phi_b
242: +\Phi_c^\dagger\Phi_d^\dagger\nabla\Phi_a\nabla\Phi_b
243: -h_i\Phi_c^\dagger\Phi_d^\dagger\Phi_a\Phi_b\bigr)+h.c.\Bigr\}+\ldots\,\,,
244: \en
245: with $h_1=2M_{\pi^0}^2$, $h_2=2M_\pi M_{\pi^0}$,
246: $h_3=3M_\pi^2-M_{\pi^0}^2$, $h_4=h_5=2M_\pi^2$.
247: The ellipsis stands for terms of order $\epsilon^4$ as well as for
248: P-wave contributions, which occur at order
249: $\epsilon^2$ in the $\pi\pi$ amplitude. They
250: do not enter the $K\to 3\pi$ matrix elements
251: at the order of accuracy considered, so we omit them here.
252: The low-energy constants $C_i,D_i$ are matched to the physical scattering
253: lengths below.
254: To simplify the resulting expressions, we have furthermore
255: introduced the scaling $x_1=x_5=1/4$, $x_2=x_3=x_4=1$.
256: Finally, note that we omit local 6-pion couplings as well.
257: Their contribution to the $K\to 3\pi$ amplitude is purely imaginary in
258: the non-relativistic framework, and of order $\epsilon^4$,
259: see also ~\cite{Cabibbo:2005ez,big}.
260:
261:
262: The pion propagator is given by
263: \be\label{eq:propnew}
264: i\langle 0|T\Phi_a(x)\Phi_b^\dagger(y)|0\rangle= \delta_{ab}
265: \int\frac{d^4p}{(2\pi)^4}\,
266: \frac{{\rm e}^{-ip(x-y)}}{2w_a({\bf p})(w_a({\bf p})-p_0-i0)}\, ,\quad
267: a,b=\pm,0\, ,
268: \ee
269: and $w_\pm({\bf p})=\sqrt{M_\pi^2+{\bf p}^2}$,
270: $w_0({\bf p})=\sqrt{M_{\pi^0}^2+{\bf p}^2}$.
271: The loops are evaluated by using the following prescription:
272: one expands the square root in a series,
273: calculates the emerging integrals using dimensional regularisation,
274: and sums up the series. In this manner
275: one ensures that the power counting
276: in the non-relativistic theory is not destroyed by the loop corrections.
277:
278: All loops in $\pi\pi$ scattering can be expressed through the basic integral
279: \begin{equation}\label{eq:F}
280: \hspace*{-4.mm}J_{ab}(P^2)=\int\frac{d^Dl}{i(2\pi)^D}\,
281: \frac{1}{2w_a({\bf l})2w_b({\bf P}-{\bf l})}\,
282: \frac{1}{(w_a({\bf l})-l_0)(w_b({\bf P}-{\bf l})-P_0+l_0)}\, ,
283: \end{equation}
284: with $P^2=P_0^2-{\bf P}^2$. In the limit $D\to 4$,
285: \vspace*{-.1cm}
286: \begin{equation}\label{eq:functionJ}
287: J_{ab}(P^2)=\frac{i}{ 16\pi} v_{ab}(P^2) \,\, ,
288: \end{equation}
289:
290: \vspace*{-.6cm}
291:
292: which is a quantity of order $\epsilon$.
293:
294: \noindent{\bf 5.}
295: The couplings $C_i,D_i$ can be expressed in terms of
296: the threshold parameters of the underlying relativistic theory.
297: In the isospin symmetry
298: limit, the expansion of the relativistic amplitude reads
299: \be
300: {\rm Re}\,\bar T_i(s,t)=
301: \bar A_i\Bigl\{1+\frac{\bar r_i}{4M_\pi^2}\,\bigl(s-4M_\pi^2\bigr)\Bigr\}
302: +\ldots\,\,.
303: \ee
304: The ellipsis stands for higher orders in $\epsilon$,
305: as well as for P-wave
306: contributions. The bar indicates the isospin symmetric limit,
307: at $M_\pi=139.57$ MeV. In terms of the standard dimensionless scattering
308: lengths $a_0$ and $a_2$, one has
309: \bea\label{eq:isospin}
310: 3\bar A_1&=&{N(a_0+2a_2)}{}\co\quad
311: 2\bar A_2={Na_2}\co\quad
312: 3\bar A_3={N(a_2-a_0)}\co\nnnl
313: 6\bar A_4&=&{N(2a_0+a_2)}\co\quad
314: \bar A_5=Na_2 \; \; ; \qquad N=32 \pi\,\,,
315: \eea
316: with $a_0-a_2=0.265\pm0.004$ \cite{Colangelo:2000jc,Colangelo:2001df}.
317: The products $\bar A_i \bar r_i$ denote effective ranges.
318: Still in the isospin symmetry limit, the couplings $C_i,D_i$ are related to these
319: threshold parameters in the following manner,
320: \be\label{eq:pipimatching1}
321: 2\bar C_i=\bar A_i\,\,\,,\qquad 8 M_\pi^2 \bar D_i=\bar A_i\bar r_i\,\,.
322: \ee
323: Taking isospin breaking into account, one
324: finds at leading order in chiral perturbation theory \cite{knechturech}
325: \be\label{eq:pipimatching2}
326: 2C_{1,2,5}=\bar A_{1,2,5}(1-\eta),\,\,
327: 2C_{3}=\bar A_3(1+\eta/3),\,\,
328: 2C_{4}=\bar A_4(1+\eta),
329: \ee
330: where
331: $\eta=\Delta_\pi/M_{\pi}^2=6.5\times 10^{-2}$. Isospin breaking corrections in
332: $D_i$ do not contribute at the order considered here.
333:
334:
335: \begin{sloppypar}
336: \noindent{\bf 6.}
337: It remains to display the $K\rightarrow 3\pi$
338: Lagrangian,
339: \eq\label{eq:lag_K}
340: \hspace*{-4.mm}{\mathcal L}_K&=&2K^\dagger W_K\bigl(i\partial_t- W_K\bigr)K
341: \nonumber \\[1mm]
342: \hspace*{-4.mm}&+&\frac{1}{2}\, G_0\,\bigl(K^\dagger\Phi_+\Phi_0^2+h.c.\bigr)
343: +\frac{1}{2}\, G_1\,\bigl(K^\dagger( W_+-M_\pi)\Phi_+\Phi_0^2+h.c.\bigr)
344: \\[1mm]
345: \hspace*{-4.mm}&+&\frac{1}{2}\, H_0\,\bigl(K^\dagger\Phi_-\Phi_+^2+h.c.\bigr)
346: +\frac{1}{2}\, H_1\,\bigl(K^\dagger( W_- -M_\pi)\Phi_-\Phi_+^2+h.c.\bigr)
347: +\ldots\, , \nonumber
348: \en
349: where $K$ denotes the non-relativistic field for the $K^+$ meson,
350: $ W_K=\sqrt{M_K^2-\triangle}$, and the ellipsis stands for the
351: higher-order terms in $\epsilon$. Note that all couplings $G_i$, $H_i$ are
352: assumed to be real. Their contribution to the decay matrix elements
353: at tree level is provided below, in the amplitudes $A_{N,C}$.
354: \end{sloppypar}
355:
356: The complete Lagrangian of the theory is
357: ${\mathcal L}_K+{\mathcal L}_{\pi\pi}$.
358: The tree-level expressions
359: for the amplitudes, generated by ${\mathcal L}_K$, are
360: modified by final state interactions of the pions, generated by loops
361: evaluated with ${\mathcal L}_{\pi\pi}$. We use the notation
362: \bea\label{eq:defexpand}
363: {\mathcal M}_{00+}={\mathcal M}_N^{\words{tree}}
364: +{\mathcal M}_N^{\words{1-loop}}
365: +{\mathcal M}_N^{\words{2-loops}}+\ldots\,\,\,\,
366: [K^+\to\pi^0\pi^0\pi^+]\,\,,\nonumber\\[1mm]
367: {\mathcal M}_{++-}={\mathcal M}_C^{\words{tree}}
368: +{\mathcal M}_C^{\words{1-loop}}
369: +{\mathcal M}_C^{\words{2-loops}}+\ldots\,\,\,\,
370: [K^+\to\pi^+\pi^+\pi^-]\,\,
371: \eea
372: for the decay amplitudes and the Condon-Shortley phase convention for
373: the pions. Our amplitudes are
374: normalised such that the decay rates are given by
375: \be
376: d\Gamma=\frac{1}{2M_K}(2\pi)^4\delta^{(4)}(P_f-P_i)
377: {|\mathcal M|}^2\prod_{i=1}^3
378: \frac{d^3{\bf p}_i}{2(2\pi)^3p_i^0}\,.
379: \ee
380:
381: \noindent{\bf 7.}
382: We now display the tree and one-loop results and modify the notation for
383: the couplings $C_i,D_i$ in order to make the formulae more transparent:
384: \be
385: (C_1,C_2,C_3,C_4,C_5)=(C_{00},C_{+0},C_{x},C_{+-},C_{++})\,,
386: \ee
387: and analogously for the $D_i$. We find
388: \eq\label{eq:1loop1}
389: {\mathcal M}_N^{\words{tree}} &=& A_N(s_3)\,\,,\nonumber\\[1mm]
390: {\mathcal M}_N^{\words{1-loop}}
391: &=&B_{N1}(s_3)J_{+-}(s_3)+B_{N2}(s_3)J_{00}(s_3)
392: \nonumber\\[1mm]
393: &+&\{B_{N3}(s_1)J_{+0}(s_1) + (s_1\leftrightarrow s_2)\}\,\,,
394: \nonumber\\[1mm]
395: {\mathcal M}_C^{\words{tree}}&=& A_C(s_3)\,\,,
396: \nonumber\\[1mm]
397: {\mathcal M}_C^{\words{1-loop}}
398: &=&
399: B_{C1}(s_3)J_{++}(s_3)
400: \nonumber\\[1mm]
401: &+&\bigl\{B_{C2}(s_1)J_{+-}(s_1)
402: +B_{C3}(s_1)J_{00}(s_1)+(s_1\leftrightarrow s_2)\bigr\}\,\,,
403: \en
404: where\footnote{To render the formulae more compact, we keep some terms
405: in Eqs.~(\ref{eq:1loop2}), (\ref{eq:1loop3}) that contribute at
406: order $a\epsilon^5$.}
407: \eq\label{eq:1loop2}
408: A_N(s_3)&=&
409: G_0+G_1\bigl(p_3^0-M_\pi\bigr)\, , \nonumber
410: \\[1mm]
411: B_{N1}(s_3)&=&2\bigl(C_x+D_x(s_3-\bar s_x)\bigr)
412: \biggl\{H_0+H_1\Bigl(\frac{p_1^0+p_2^0}{2}-M_\pi\Bigr)\biggr\}\, ,
413: \nonumber\\[1mm]
414: B_{N2}(s_3)&=&\bigl(C_{00}+D_{00}(s_3-\bar s_{00})\bigr)
415: \Bigl\{ G_0+G_1\bigl(p_3^0-M_\pi\bigr)\Bigr\}\, ,
416: \\[1mm]
417: B_{N3}(s_1)&=&2\bigl(C_{+0}+D_{+0}(s_1-\bar s_{+0})\bigr)
418: \biggl\{ \! G_0+G_1\Bigl(
419: \frac{p_2^0+p_3^0}{2}\Bigl(1+\frac{\Delta_\pi}{s_1}\Bigr)- \! M_\pi \!
420: \Bigr) \! \biggr\}\,,
421: \nonumber
422: \en
423: and
424: \eq\label{eq:1loop3}
425: A_C(s_3)&=&
426: H_0+H_1\bigl(p_3^0-M_\pi\bigr)\, , \nonumber
427: \\[1mm]
428: B_{C1}(s_3)&=&\bigl(C_{++}+D_{++}(s_3-\bar s_{++})\bigr)
429: \Bigl\{ H_0+H_1\bigl(p_3^0-M_\pi\bigr)\Bigr\}\, ,
430: \nonumber\\[1mm]
431: B_{C2}(s_1)&=&2\bigl(C_{+-}+D_{+-}(s_1-\bar s_{+-})\bigr)
432: \biggl\{ H_0+H_1\Bigl(\frac{p_2^0+p_3^0}{2}-M_\pi\Bigr)\biggr\}\, ,
433: \nonumber\\[1mm]
434: B_{C3}(s_1)&=&\bigl(C_x+D_x(s_1-\bar s_x)\bigr)
435: \Bigl\{ G_0+G_1\bigl(p_1^0-M_\pi\bigr)\Bigr\}\, .
436: \en
437: In the above expressions, $\bar s_i$ denotes the physical threshold in
438: the $i$ channel and $p_i^0$ are given by Eq.~(\ref{eq:kin1}).
439: Note that, according to this equation, the masses in the
440: relation of $p_i^0$ to $s_i$ differ in the neutral and charged channels.
441:
442: \noindent {\bf 8.}
443: There are two topologically distinct two-loop graphs that describe
444: pion-pion rescattering in the final state, see Fig.~\ref{fig:2loop_text}.
445: At the order of accuracy we are working, it is sufficient to consider the case
446: of non-derivative couplings. In this case, the contributions of
447: both diagrams depend only on the variable $s$, where
448: \begin{equation}
449: Q^\mu=(q_1+q_2)^\mu \,\,,\,\,Q^2=s\,\, .
450: \end{equation}
451: The diagram
452: \begin{figure}[t]
453: \begin{center}
454: \includegraphics[width=9.cm]{figure1.eps}
455: \end{center}
456: \caption{Two topologically distinct non-relativistic two-loop graphs
457: describing the final-state $\pi\pi$ rescattering in the decay $K\to 3\pi$, with
458: $Q^\mu=(q_1+q_2)^\mu$.}
459: \label{fig:2loop_text}
460: \end{figure}
461: in Fig.~\ref{fig:2loop_text}B, apart from a factor
462: containing coupling constants, is given by a product of two one-loop
463: diagrams which were already calculated in Eq.~(\ref{eq:F}).
464: The non-trivial contribution from Fig.~\ref{fig:2loop_text}A
465: is proportional to
466: \eq\label{eq:twoloop_F}
467: {\mathcal M}(s)
468: &=&\int\frac{d^D l}{i(2\pi)^D}\,\frac{d^D k}{i(2\pi)^D}\,
469: \nonumber\\[1mm]
470: &\times&
471: \frac{1}{2w_a({\bf l}+{\bf k})}\,
472: \frac{1}{w_a({\bf l}+{\bf k})-M_K+l^0+k^0}\,
473: \frac{1}{2w_b({\bf l})}\,
474: \frac{1}{w_b({\bf l})-l^0}\,
475: \nonumber\\[1mm]
476: &\times&\frac{1}{2w_c({\bf k})}\,
477: \frac{1}{w_c({\bf k})-k^0}\,
478: \frac{1}{2w_d({\bf Q}-{\bf k})}\,
479: \frac{1}{w_d({\bf Q}-{\bf k})-Q^0+k^0}\, .
480: \en
481: A detailed discussion of this integral will be provided in \cite{big}.
482: Here we simply note that
483: the most general
484: representation for diagram
485: Fig.~\ref{fig:2loop_text}A can be written in the form
486: \be\label{eq:mostgeneral}
487: {\mathcal M}(s)
488: =F\bigl(M_a,M_b,M_c,M_d;s\bigr)+P(s)J_{cd}(s)+P'(s)\, ,
489: \ee
490: where $F$ is ultraviolet-finite and
491: contains the full non-analytic behaviour of the two-loop diagram
492: in the low-energy domain. Further, $J_{cd}$ is the one-loop function
493: Eq.~(\ref{eq:functionJ}),
494: and $P,\; P'$ are real polynomials.
495: We have suppressed the dependence of
496: $F$ on the mass $M_i$ generated by ${\bf Q}^2$, see below.
497:
498: In the following, we use a simplified form of $F$, where
499: part of its imaginary part is dropped -- this omission affects
500: the decay width at order $a^3$ only and is therefore of no relevance here.
501: We use the integral representation \cite{big}
502: \eq\label{eq:functionF}
503: \hspace*{-.8cm}
504: &&F\bigl(M_a,M_b,M_c,M_d;s\bigr)=\nonumber\\[1mm]
505: &&\frac{{\mathcal N}_0}{64\pi^3\sqrt{s}}
506: \int_0^1\frac{dy}{\sqrt{y}}\,\frac{dg(y,s)}{dy}\,
507: \Bigl(\ln{g(y,s)}-\ln{g(y,\bar s)}\Bigr)+O(\epsilon^4)\, ,
508: \en
509: where
510: \eq
511: \hspace*{-.4cm}&&{\mathcal N}_0=\frac{M_K}{2\sqrt{s_0}}\,
512: \biggl(1-\frac{(M_a-M_b)^2}{s_0}\biggr)^{1/2}
513: \frac{1}{\bigl(2(M_K^2+M_c^2)-(M_a+M_b)^2-s_0\bigr)^{1/2}}\, ,
514: \nonumber\\[1mm]
515: \hspace*{-.4cm}&&s_0=M_K^2+M_c^2-2M_K\biggl( M_c^2+\frac{{\bf Q}^2(1+\delta)^2}{4}\biggr)^{1/2}\, ,
516: \nonumber\\[1mm]
517: \hspace*{-.4cm}&&g(y,s)=-(1-y)q_0^2-y\Delta^2+\frac{y(1-y){\bf Q}^2(1+\delta)^2}{4(1+y{\bf Q}^2/s)}-i0\, ,
518: \nonumber\\[1mm]
519: \hspace*{-.4cm}&&q_0^2=\frac{\lambda(s,M_c^2,M_d^2)}{4s}\, ,\quad\quad
520: \bar s=(M_c+M_d)^2\, ,
521: \nonumber\\[1mm]
522: \hspace*{-.4cm}&&\Delta^2=\frac{\lambda(M_K^2,M_c^2,(M_a+M_b)^2)}{4M_K^2}\, ,\quad\quad
523: \delta=\frac{M_c^2-M_d^2}{s}\, .
524: \en
525: Approaching threshold from above, we find
526: \be
527: F\bigl(M_a,M_b,M_c,M_d;s\bigr)=-\frac{{q_0}}{128\pi^2(M_c+M_d)}
528: \frac{\lambda^{1/2}(\bar s_0,M_a^2,M_b^2)}{\bar s_0}+O(q_0^2)\, ,
529: \ee
530: where $\bar s_0$ denotes $s_0$ at $q_0^2=0$.
531: This last relation shows that $F$ is of order $\epsilon^2$.
532:
533: \noindent {\bf 9.}
534: Our prescription for the representation of the decay amplitudes at $O(a^2)$
535: is as follows: we evaluate the contributions from all the graphs displayed
536: in Fig.~\ref{fig:2loop_rep_n} and Fig.~\ref{fig:2loop_rep_c}. Further, in
537: the graphs of the type
538: Fig.~\ref{fig:2loop_text}A, we retain only the non-analytic piece $F$,
539: whereas the polynomials $P,\; P^\prime$ are included in the tree-level couplings
540: $G_i,H_i$. This choice of a particular representation of $F$ is equivalent
541: to a renormalisation prescription.
542:
543: \begin{figure}[t]
544: \begin{center}
545: \includegraphics[width=13.5cm]{figure2.eps}
546: \end{center}
547: \caption{Two-loop graphs contributing to the decay $K^+\to\pi^0\pi^0\pi^+$
548: in the non-relativistic effective theory. The graphs obtained by a
549: permutation of identical particles in the final state are not
550: shown.}
551: \label{fig:2loop_rep_n}
552: \end{figure}
553:
554: With this convention, we find for the amplitudes at order $a^2\epsilon^2$
555: \be\label{eq:rep}
556: {\mathcal M}_I^{\words{2-loops}}
557: ={\mathcal M}_I^A(s_1,s_2,s_3)
558: +{\mathcal M}_I^B(s_1,s_2,s_3)\,\,;\,\, I=N,C\,\,,
559: \ee
560: where
561: \eq
562: \hspace*{-2.mm}{\mathcal M}_N^A
563: &=&4H_0C_{+-}C_xF_+\bigl(M_\pi,M_\pi,M_\pi,M_\pi;s_3\bigr)
564: \nonumber\\[1mm]
565: &+&2G_0C_x^2 F_+\bigl(M_{\pi^0},M_{\pi^0},M_\pi,M_\pi;s_3\bigr)
566: \nonumber\\[1mm]
567: &+&2H_0C_{++}C_xF_+\bigl(M_\pi,M_\pi,M_\pi,M_\pi;s_3\bigr)
568: \nonumber\\[1mm]
569: &+&4G_0C_{00}C_{+0}F_+\bigl(M_\pi,M_{\pi^0},M_{\pi^0},M_{\pi^0};s_3\bigr)
570: \nonumber\\[1mm]
571: &+&\Bigl\{4H_0C_xC_{+0}F_0\bigl(M_\pi,M_\pi,M_\pi,M_{\pi^0};s_1\bigr)
572: \nonumber\\[1mm]
573: &+&4G_0C_{+0}^2F_0\bigl(M_{\pi^0},M_\pi,M_{\pi^0},M_\pi;s_1\bigr)
574: \nonumber\\[1mm]
575: &+&2G_0C_{00}C_{+0}F_0\bigl(M_{\pi^0},M_{\pi^0},M_\pi,M_{\pi^0};s_1\bigr)
576: +(s_1\leftrightarrow s_2)\Bigr\}\,\,, \nonumber \\
577: && \nonumber \\
578: \label{eq:0B}
579: {\mathcal M}_N^B
580: &=&4H_0C_xC_{+-}J^2_{+-}(s_3)
581: +G_0C_{00}^2J^2_{00}(s_3)
582: \nonumber\\[1mm]
583: &+&2\bigl[G_0C_x^2+H_0C_xC_{00}\bigr]J_{+-}(s_3)J_{00}(s_3)
584: \nonumber\\[1mm]
585: &+&\Bigl\{4G_0C_{+0}^2J^{\,\,2}_{+0}(s_1)
586: +(s_1\leftrightarrow s_2)\Bigr\}\, ,
587: \en
588: \eq
589: \hspace*{-5.mm}{\mathcal M}_C^A
590: &=&2G_0C_xC_{++} F_-\bigl(M_{\pi^0},M_{\pi^0},M_\pi,M_\pi;s_3\bigr)\, ,
591: \nonumber\\[1mm]
592: &+&4H_0C_{+-}C_{++} F_-\bigl(M_\pi,M_\pi,M_\pi,M_\pi;s_3\bigr)
593: \nonumber\\[1mm]
594: &+&\Bigl\{4H_0C_{+-}^2 F_+\bigl(M_\pi,M_\pi,M_\pi,M_\pi;s_1\bigr)
595: \nonumber\\[1mm]
596: &+&2G_0C_xC_{+-} F_+\bigl(M_{\pi^0},M_{\pi^0},M_\pi,M_\pi;s_1\bigr)
597: \nonumber\\[1mm]
598: &+&2H_0C_{++}C_{+-} F_+\bigl(M_\pi,M_\pi,M_\pi,M_\pi;s_1\bigr)
599: \nonumber\\[1mm]
600: &+&4G_0C_{+0} C_x F_+\bigl(M_\pi,M_{\pi^0},M_{\pi^0},M_{\pi^0};s_1\bigr)
601: + (s_1\leftrightarrow s_2)\Bigr\}\co \nonumber \\
602: && \nonumber \\
603: \label{eq:+B}
604: {\mathcal M}_C^B
605: &=&H_0C_{++}^2J^2_{++}(s_3)\nonumber\\[1mm]
606: &+&\Bigl\{ 4H_0C_{+-}^2J^2_{+-}(s_1)+G_0C_xC_{00}
607: J^2_{00}(s_1)
608: \nonumber\\[1mm]
609: &+&2\bigl[H_0C_x^2+G_0C_x C_{+-}\bigr]J_{+-}(s_1)J_{00}(s_1)
610: +(s_1\leftrightarrow s_2)\Bigr\}\,\,.
611: \en
612: \begin{sloppypar}
613: Here, $F_i(\ldots;s)$ stands for the integral $F(\ldots;s)$,
614: evaluated at ${\bf
615: Q}^2=\lambda(M_K^2,M_{\pi^i}^2,s)/4M_K^2$, with $i=\pm,0$.
616: \end{sloppypar}
617:
618: \begin{figure}[t]
619: \begin{center}
620: \includegraphics[width=13.5cm]{figure3.eps}
621: \end{center}
622: \caption{Two-loop graphs contributing to the decay $K^+\to\pi^+\pi^+\pi^-$
623: in the non-relativistic effective theory. The graphs obtained by a
624: permutation of identical particles in the final state are not
625: shown.}
626: \label{fig:2loop_rep_c}
627: \end{figure}
628:
629: \noindent{\bf 10.}
630: The decay amplitudes depend on the four {\em real} $K\to 3\pi$ coupling
631: constants $H_i,G_i$ and on the threshold parameters for $\pi\pi$
632: scattering. Combining the tree and one-loop result
633: Eqs.~(\ref{eq:1loop1})--(\ref{eq:1loop3}) with the two-loop contributions
634: Eqs.~(\ref{eq:rep})--(\ref{eq:+B}), we obtain the neutral and charged decay
635: amplitudes up to and
636: including terms of order $\epsilon^2,a\epsilon^3$ and $a^2\epsilon^2$,
637: expressed in terms of the one- and two-loop integrals $J$ and $F$ displayed
638: in Eqs.~(\ref{eq:functionJ}) and (\ref{eq:functionF}), respectively. This
639: representation is valid in the whole decay region, and is the main result
640: of this article.
641:
642:
643: The decay amplitude $K^+\to\pi^0\pi^0\pi^+$ obeys what we refer to as the
644: {\it threshold theorem}: the coefficient of the leading non-analytic piece,
645: which is proportional to $v_{+-}(s_3)$, is given by a product of two
646: factors, the decay amplitude $K^+\to\pi^+\pi^+\pi^-$ and the scattering
647: amplitude $\pi^+\pi^-\to\pi^0\pi^0$, both evaluated at threshold. Thus,
648: the heuristic argumentation of Ref.~\cite{Cabibbo:2004gq}, which serves as
649: a cornerstone of the whole method, is confirmed in the effective field
650: theory framework. This threshold theorem has its analogue in hadronic
651: atoms, viz., in the modification of energy levels and decay widths through
652: hadronic interactions
653: \cite{nrqft6,nrqft7,nrqft8,nrqft9,nrqft10,nrqft11,nrqft12,nrqft13,nrqft14,nrqft15,nrqftz}.
654: Of course, aside from the determination of the leading term in $v_{+-}$,
655: our approach also allows a systematic evaluation of higher-order
656: contributions $v_{+-}^3,v_{+-}^5\ldots$\,.
657:
658: \noindent {\bf 11.}
659: We now compare the content of this letter with the recent work of Cabibbo
660: and Isidori~\cite{Cabibbo:2005ez} (CI), who have proposed an alternative
661: representation for the $K\to 3\pi$ decay amplitudes.
662: Assuming certain analytic properties of the decay amplitudes, CI derive a
663: representation of the amplitudes up to and including terms of order $a^2$,
664: using analyticity, unitarity, and cluster decomposition properties of the
665: $S$ matrix.
666:
667: Conceptual aspects of our methods have already been compared in the
668: introduction. Here we add that, firstly, we do provide an explicit
669: representation of the decay amplitudes that is valid in the whole decay
670: region, including all powers of velocities generated by the graphs
671: considered. Secondly, we note that our explicit two-loop calculation
672: confirms the expected \cite{Cabibbo:2005ez} analytic properties of the
673: amplitudes in the vicinity of the $\pi^+\pi^-$ threshold. On the other
674: hand, away from this threshold, but still in the physical decay region,
675: branch points develop \cite{big}, contrary to the expectations spelled out
676: in \cite{Cabibbo:2005ez}.
677:
678: After these general remarks we compare the amplitudes in more detail at one
679: and two loops. In the actual calculations in Ref.~\cite{Cabibbo:2005ez},
680: an approximation has been used: the angular integrals have been replaced by
681: averages, where the integrand is evaluated at a certain value of
682: $\cos\theta$. As CI note, this approximation is exact, if the integrand is
683: at most a linear function in $\cos\theta$. Since this is true at
684: $O(a\epsilon^2)$, one expects that our results at this order algebraically
685: agree with CI. We have checked that this is indeed the case (up to a few
686: typos \cite{isidoriprivate}).
687:
688: At two loops, our results are algebraically different from those of
689: Ref.~\cite{Cabibbo:2005ez}. The reason can be traced back to the following.
690: At two loops, the discontinuity cannot, in general, be obtained from an
691: integration over $\cos\theta$ without further ado: the path of integration
692: has to be deformed into the complex plane -- it does not simply run from
693: $\cos\theta=-1$ to $\cos\theta=1$ along the real axis \cite{anisovich}. In
694: fact, near the pseudothreshold $s_3=(M_K-M_\pi)^2$, the deformed path runs
695: to infinity, thus generating an infinity in the discontinuity there. In the
696: case where all internal masses are equal to the charged pion mass,
697: integrating $\cos\theta$ along the real axis does generate the correct
698: discontinuity up to $s_3=(M_K^2-M_\pi^2)/2$. In the case where two neutral
699: pions are running in the inner loop, the situation is more complicated,
700: because an anomalous threshold develops in the lower half plane. Still,
701: integrating along the path mentioned generates the correct discontinuity
702: near threshold. [These difficulties do not arise in our approach
703: -- the function $F$ in Eq.~(\ref{eq:functionF}) is smooth on the upper rim
704: of the cut, while its discontinuity develops the singularities mentioned,
705: at the positions predicted by the Landau equations
706: \cite{landau,smatrix,big}.]
707:
708: A replacement of the angular integrals by an average
709: \cite{Cabibbo:2005ez}, where the integrand is evaluated at a certain value
710: of $\cos\theta$, can therefore be a reasonable approximation only in the
711: vicinity of the threshold $s_3=4M_\pi^2$. Indeed, our result agrees (up to
712: a typo~\cite{isidoriprivate}) with Ref.~\cite{Cabibbo:2005ez} to
713: the order considered here, at threshold. Away from threshold, the
714: expressions differ.
715:
716: Finally, we shortly comment on the (revised) article by Gamiz
717: et al. \cite{gamizetal}, which appeared only very recently. It is the aim
718: of that article to provide an error analysis of the procedure proposed in
719: \cite{Cabibbo:2004gq,Cabibbo:2005ez}. The authors investigate the process
720: $K\rightarrow 3\pi$ in the framework of chiral perturbation theory, and
721: approximate two-loop graphs by retaining their discontinuity only -- an
722: approach which generates fake singularities in the transition amplitude, as
723: just mentioned. An ad hoc prescription is invoked in \cite{gamizetal} to
724: avoid these singularities when investigating the cusp, see e.g. their
725: comment after Eq.~(4.20).
726:
727: \noindent {\bf 12.}
728: In summary, we have investigated $K\to 3\pi$ decays within a
729: non-relativistic effective Lagrangian framework. The amplitudes are
730: calculated in a systematic double expansion in the kinetic energies of the
731: decay products (which we count as terms of order $\epsilon^2$), and in the
732: threshold parameters of elastic $\pi\pi$ scattering (which are generically
733: denoted by $a$). We provide an explicit representation of the amplitudes
734: at order $\epsilon^2, a\epsilon^3, a^2\epsilon^2$ -- valid in the whole
735: decay region -- in terms of the (real) $K\to 3\pi$ coupling constants
736: $G_i,H_i$ and of the threshold parameters $a$.
737:
738: Our amplitudes agree with the ones of Cabibbo and
739: Isidori \cite{Cabibbo:2005ez}, up to and including terms of
740: order $a\epsilon^3$.
741: On the other hand, at order $a^2$, they differ
742: away from threshold. We propose to repeat the
743: analysis of the experimental data of the NA48/2 collaboration
744: \cite{Batley} with our representation of the amplitudes, for
745: several reasons: i) In view of the aimed precision, one ought to examine
746: the importance of the mentioned differences in the determination of
747: $\pi\pi$ scattering lengths. ii) It would be useful to extend the fit to
748: the full decay region, and to the charged decay modes $K^+\rightarrow
749: \pi^+\pi^+\pi^-$ as well, in order to determine a maximal set of $\pi\pi$
750: threshold parameters. iii) As was already pointed out in
751: \cite{Cabibbo:2005ez}, cusps also occur at the border of the Dalitz plot.
752: Investigating data in those regions may allow one to determine different
753: combinations of scattering lengths.
754:
755:
756: It remains to investigate the importance of higher orders in the low-energy
757: expansion, and to apply radiative corrections, which can be evaluated
758: in the field-theoretical framework used here in a standard manner.
759: The effects generated by the $\pi^+\pi^-$ bound state at the $\pi^+\pi^-$
760: threshold can also be investigated within
761: the same approach (see, e.g.,
762: \cite{nrqft6,nrqft7,nrqft8,nrqft9,nrqft10,nrqft11,nrqft12,nrqft13,nrqft14}).
763: We plan to include these effects in forthcoming publications.
764:
765: \begin{sloppypar}
766: {\it Acknowledgements.}
767: It is a pleasure
768: to thank N.~Cabibbo, J.~Donoghue, G.~Isidori and J.~Prades for enjoyable discussions.
769: Partial financial support under the EU Integrated Infrastructure
770: Initiative Hadron Physics Project (contract number RII3-CT-2004-506078)
771: and DFG (SFB/TR 16, ``Subnuclear Structure of Matter'') is gratefully
772: acknowledged. This work was supported by the Swiss
773: National Science Foundation, by RTN, BBW-Contract No. 01.0357,
774: and EC-Contract HPRN--CT2002--00311 (EURIDICE).
775: One of us (J.G.) is grateful to the Alexander von Humboldt-Stiftung and to
776: the Helmholtz-Gemeinschaft for the award of a prize
777: that allowed him to stay at the HISKP at the University of Bonn,
778: where part of this work was performed.
779: He also thanks the HISKP for the warm hospitality during these stays.
780: \end{sloppypar}
781:
782:
783:
784: \begin{thebibliography}{99}
785:
786: \bibitem{Colangelo:2000jc}
787: G.~Colangelo, J.~Gasser and H.~Leutwyler,
788: %``The pi pi S-wave scattering lengths,''
789: Phys.\ Lett.\ B {\bf 488} (2000) 261
790: [arXiv:hep-ph/0007112].
791: %%CITATION = HEP-PH 0007112;%%
792:
793:
794: \bibitem{Colangelo:2001df}
795: G.~Colangelo, J.~Gasser and H.~Leutwyler,
796: %``pi pi scattering,''
797: Nucl.\ Phys.\ B {\bf 603} (2001) 125
798: [arXiv:hep-ph/0103088].
799: %%CITATION = HEP-PH 0103088;%%
800:
801: \bibitem{Pislak:2003sv}
802: S.~Pislak {\it et al.},
803: %``High statistics measurement of K(e4) decay properties,''
804: Phys.\ Rev.\ D {\bf 67} (2003) 072004
805: [arXiv:hep-ex/0301040].
806: %%CITATION = HEP-EX 0301040;%%
807:
808:
809: \bibitem{Adeva:2005pg}
810: B.~Adeva {\it et al.} [DIRAC Collaboration],
811: %``First measurement of the pi+ pi- atom lifetime,''
812: Phys.\ Lett.\ B {\bf 619} (2005) 50
813: [arXiv:hep-ex/0504044].
814: %%CITATION = HEP-EX 0504044;%%
815:
816:
817:
818: \bibitem{Cabibbo:2004gq}
819: N.~Cabibbo,
820: Phys.\ Rev.\ Lett.\ {\bf 93} (2004) 121801
821: [arXiv:hep-ph/0405001].
822: %%CITATION = HEP-PH 0405001;%%
823:
824: \bibitem{Cabibbo:2005ez}
825: N.~Cabibbo and G.~Isidori,
826: JHEP {\bf 0503} (2005) 021
827: [arXiv:hep-ph/0502130].
828: %%CITATION = HEP-PH 0502130;%%
829:
830: \bibitem{Batley}
831: J.~R.~Batley {\it et al.} [NA48/2 Collaboration],
832: %``Observation of a cusp-like structure in the pi0 pi0 invariant mass
833: %distribution
834: %from K+- $\to$ pi+- pi0 pi0 decay and determination of the pi pi
835: %scattering lengths,''
836: Phys.\ Lett.\ B {\bf 633} (2006) 173
837: [arXiv:hep-ex/0511056].
838: %%CITATION = HEP-EX 0511056;%%
839:
840:
841: \bibitem{MMS}
842: U.-G.~Mei{\ss}ner, G.~M\"uller and S.~Steininger,
843: %``Virtual photons in SU(2) chiral perturbation theory and electromagnetic
844: %corrections to pi pi scattering,''
845: Phys.\ Lett.\ B {\bf 406} (1997) 154
846: [Erratum-ibid.\ B {\bf 407} (1997) 454]
847: [arXiv:hep-ph/9704377].
848: %%CITATION = HEP-PH 9704377;%%
849:
850: \bibitem{nrqfta}
851: W.~E.~Caswell and G.~P.~Lepage,
852: %``Effective Lagrangians For Bound State Problems
853: %In QED, QCD, And Other Field
854: %Theories,''
855: Phys.\ Lett.\ B {\bf 167} (1986) 437.
856: %%CITATION = PHLTA,B167,437;%%
857: \bibitem{nrqft1}
858: P.~Labelle and K.~Buckley,
859: arXiv:hep-ph/9804201.
860: %%CITATION = HEP-PH 9804201;%%
861: \bibitem{nrqft2}
862: X.~Kong and F.~Ravndal,
863: Phys.\ Rev.\ D {\bf 59} (1999) 014031.
864: %%CITATION = PHRVA,D59,014031;%%
865: \bibitem{nrqft3}
866: X.~Kong and F.~Ravndal, Phys.\ Rev.\ D {\bf 61}
867: (2000) 077506 [arXiv:hep-ph/9905539].
868: %%CITATION = HEP-PH 9905539;%
869: \bibitem{nrqft4}
870: B.~R.~Holstein,
871: Phys.\ Rev.\ D {\bf 60} (1999) 114030 [arXiv:nucl-th/9901041].
872: %%CITATION = NUCL-TH 9901041;%%
873: \bibitem{nrqft5}
874: D.~Eiras and J.~Soto,
875: Phys.\ Rev.\ D {\bf 61} (2000) 114027 [arXiv:hep-ph/9905543].
876: %%CITATION = HEP-PH 9905543;%%
877: \bibitem{nrqft6}
878: A.~Gall, J.~Gasser, V.~E.~Lyubovitskij and A.~Rusetsky,
879: Phys.\ Lett.\ B {\bf 462} (1999) 335 [arXiv:hep-ph/9905309].
880: %%CITATION = HEP-PH 9905309.%%
881: \bibitem{nrqft7}
882: J.~Gasser, V.~E.~Lyubovitskij, A.~Rusetsky and A.~Gall,
883: Phys.\ Rev.\ D {\bf 64} (2001) 016008 [arXiv:hep-ph/0103157].
884: %%CITATION = HEP-PH 0103157.%%
885: \bibitem{nrqft8}
886: J.~Gasser, V.~E.~Lyubovitskij and A.~Rusetsky,
887: Phys.\ Lett.\ B {\bf 471} (1999) 244 [arXiv:hep-ph/9910438].
888: %%CITATION = HEP-PH 9910438
889: \bibitem{nrqft9}
890: J.~Schweizer,
891: Phys.\ Lett.\ B {\bf 587} (2004) 33 [arXiv:hep-ph/0401048].
892: %%CITATION = HEP-PH 0401048.%%
893: \bibitem{nrqft10}
894: J.~Schweizer,
895: Eur.\ Phys.\ J.\ C {\bf 36} (2004) 483 [arXiv:hep-ph/0405034].
896: %%CITATION = HEP-PH 0405034.%%
897: \bibitem{nrqft11}
898: V.~E.~Lyubovitskij and A.~Rusetsky,
899: Phys.\ Lett.\ B {\bf 494} (2000) 9 [arXiv:hep-ph/0009206].
900: %%CITATION = HEP-PH 0009206.%%
901: \bibitem{nrqft12}
902: J.~Gasser, M.~A.~Ivanov, E.~Lipartia, M.~Moj\v{z}i\v{s} and A.~Rusetsky,
903: Eur.\ Phys.\ J.\ C {\bf 26} (2002) 13 [arXiv:hep-ph/0206068].
904: %%CITATION = HEP-PH 0206068.%%
905: \bibitem{nrqft13}
906: U.-G.~Mei{\ss}ner, U.~Raha and A.~Rusetsky,
907: Eur.\ Phys.\ J.\ C {\bf 35} (2004) 349
908: [arXiv:hep-ph/0402261].
909: %%CITATION = HEP-PH 0402261.%%
910: \bibitem{nrqft14}
911: U.-G.~Mei{\ss}ner, U.~Raha and A.~Rusetsky,
912: Eur.\ Phys.\ J.\ C {\bf 41} (2005) 213
913: [arXiv:nucl-th/0501073]; erratum {\it ibid.} C {\bf 45} (2006) 545.
914: %%CITATION = NUCL-TH 0501073.%%
915: \bibitem{nrqft15}
916: U.-G.~Mei{\ss}ner, U.~Raha and A.~Rusetsky,
917: %``Isospin-breaking corrections in the pion deuteron scattering length,''
918: arXiv:nucl-th/0512035.
919: %%CITATION = NUCL-TH 0512035.%%
920: \bibitem{nrqftz}
921: U.-G.~Mei{\ss}ner, U.~Raha and A.~Rusetsky,
922: %``Kaon nucleon scattering lengths from kaonic deuterium experiments,''
923: arXiv:nucl-th/0603029.
924: %%CITATION = NUCL-TH 0603029.%%
925:
926: \bibitem{anisovich}
927: V.~V.~Anisovich and A.A.~Ansel'm,
928: Sov.~Phys.~Uspekhi \ {\bf 9} (1966) 117
929: [Usp.~Fiz.~Nauk \ {\bf 88} (1966) 287].
930:
931:
932: \bibitem{big}
933: J.~Gasser, B.~Kubis and A.~Rusetsky, in preparation.
934:
935: \bibitem{knechturech}
936: M.~Knecht and R.~Urech,
937: %``Virtual photons in low energy pi pi scattering,''
938: Nucl.\ Phys.\ B {\bf 519} (1998) 329
939: [arXiv:hep-ph/9709348].
940: %%CITATION = HEP-PH 9709348.%%
941:
942: \bibitem{isidoriprivate}
943: G.~Isidori, private communication.
944:
945: \bibitem{landau}
946: L.~D.~Landau, Nucl.\ Phys. \ {\bf 13} (1959) 181.
947:
948: \bibitem{smatrix}
949: R.~Eden, P.~V.~ Landshoff, D.~I.~Olive and J.~C.~Polkinghorne,
950: The analytic S-matrix (Cambridge University Press, 1966).
951:
952: \bibitem{gamizetal}
953: E.~Gamiz, J.~Prades and I.~Scimemi,
954: %``K $\to$ 3pi final state interactions
955: %at NLO in CHPT and Cabibbo's proposal
956: %to measure a(0)-a(2),''
957: arXiv:hep-ph/0602023.
958: %%CITATION = HEP-PH 0602023.%%
959:
960: \end{thebibliography}
961:
962:
963: \ed
964: