0710.5432/em.tex
1: 
2: \documentclass[11pt, twoside, fleqn]{article}
3: 
4: 
5: % ------- HEADER -----------------------------------------------------
6: 
7: 
8: % Packages
9: 
10: \usepackage{a4wide}
11: \usepackage[fleqn]{amsmath}
12: \usepackage{amssymb, exscale}
13: \usepackage{epsfig}
14: \usepackage{titlesec}
15: \usepackage{cite}
16: \usepackage[marginal]{footmisc} 
17: \usepackage[overload]{textcase} 
18: \usepackage{subfigure}
19: \usepackage{color}
20: 
21: \usepackage{latexsym}
22: \usepackage{graphics}
23: \usepackage{pst-node}
24: \usepackage{epsf}
25: \usepackage{exscale}
26: 
27: 
28: \usepackage[singlelinecheck=false,margin=\parindent,labelfont={small,sc},textfont={small,sf}]{caption}
29: 
30: 
31: %\usepackage[]{showlabels}\def\showlabelfont{\tiny\tt}
32: %note that the \usepackage{showlabels} command must appear after amsmath
33: 
34: % hopefully we can dispose of this at one point
35: \allowdisplaybreaks[1]
36: 
37: % Fonts
38: 
39: \newfont{\bigrm}{cmr10 at 12pt}
40: \newfont{\Bigrm}{cmr12 at 16pt}
41: \newfont{\hermes}{cmtt10 at 10pt}
42: 
43: \makeatletter
44: 
45: % Need small letters for numbering
46: \renewcommand\thesection{\oldstylenums{\arabic{section}}}
47: \renewcommand\thesubsection{\thesection.\oldstylenums{\arabic{subsection}}}
48: \renewcommand\thesubsubsection{\thesubsection.\oldstylenums{\arabic{subsubsection}}}
49: \renewcommand\thefigure{\oldstylenums{\arabic{figure}}}
50: \renewcommand\thetable{\oldstylenums{\arabic{table}}}
51: \renewcommand{\theequation}{\thesection.\oldstylenums{\arabic{equation}}}
52: \@addtoreset{equation}{section}
53: \renewcommand\citeform[1]{\oldstylenums {#1}}
54: 
55: % differentiate in outline between section and appendix
56: \newcommand{\sectionname}{Section}
57: \newcommand{\marktotoc}[1]{\renewcommand{\sectionname}{#1}}
58: \renewenvironment{appendix}
59: {\renewcommand{\sectionname}{Appendix}%
60:   \addtocontents{toc}{\protect\marktotoc{Appendix}}%
61:   \setcounter{section}{0}%
62:   \renewcommand\thesection{{\sc\alph{section}}}%
63: }
64: {\addtocontents{toc}{\protect\marktotoc{Section}}}
65: 
66: 
67: 
68: % aux for layout and title
69: \newcommand\secformat[1]{\sc{\MakeTextUppercase{#1}}}
70: \newcommand\subsecformat[1]{\sc{\MakeTextLowercase{#1}}}
71: \newcommand\subsubsecformat[1]{{\it #1}}
72: 
73: % add up for layout of title 
74: \titleformat{\section}[block]{}{\thesection\quad}{1em}{\secformat}[]
75: \titleformat{\subsection}[block]{}{\thesubsection\quad}{1em}{\subsecformat}[]
76: \titleformat{\subsubsection}[block]{}{\thesubsubsection\quad}{1em}{\subsubsecformat}[]
77: 
78: \titlespacing*{\section}{0pt}{2\baselineskip}{2\baselineskip}
79: \titlespacing*{\subsection}{0pt}{\baselineskip}{\baselineskip}
80: \titlespacing*{\subsubsection}{0pt}{\baselineskip}{\baselineskip}
81: 
82: % general stuff
83: \frenchspacing
84: \setlength{\parskip}{0ex plus 0ex minus 0ex}
85: \setlength{\itemsep}{0ex plus 0ex minus 0ex}
86: \renewcommand{\thefootnote}{\oldstylenums{\arabic{footnote}}}
87: \renewcommand{\footnoterule}{}
88: 
89: \makeatother
90: 
91: 
92: % Indentation of formulae
93: \setlength{\mathindent}{20pt}
94: 
95: 
96: 
97: % mathematics 
98: \newcommand{\relsp}[1]{\mathrel{\phantom{{#1}}}{}}
99: \newcommand{\binsp}[1]{\mathbin{\phantom{{{}#1{}}}}}
100: \newcommand{\myfrac}[2]{\mbox{\large{$\frac{#1}{#2}$}}}
101: \newcommand{\diag}{\mathrm{diag}}
102: \renewcommand{\Re}{\mathrm{Re}\;}
103: \renewcommand{\Im}{\mathrm{Im}\;}
104: \newcommand{\relphantom}[1]{\mathrel{\phantom{#1}}}
105: \newcommand{\hhv}[2]{\smash{\overset{\leftrightarrow}{v}}^{#1}_{#2}}
106: \newcommand{\SU}{\mathrm{SU}}
107: %\renewcommand{\L}{\mathcal{L}}
108: \newcommand{\T}{\mathrm{T}}
109: \renewcommand{\l}{\ell}
110: \newcommand{\Lr}[1]{L^{\mathrm{r}}_{#1}}
111: \newcommand{\lr}[1]{l^{\mathrm{r}}_{#1}}
112: \newcommand{\Cr}[1]{C^{\mathrm{r}}_{#1}}
113: \newcommand{\clr}[1]{c^{\mathrm{r}}_{#1}}
114: \newcommand{\Hr}[1]{H^{\mathrm{r}}_{#1}}
115: \newcommand{\hr}[1]{h^{\mathrm{r}}_{#1}}
116: \newcommand{\kr}[1]{k^\mathrm{r}_{#1}}
117: \newcommand{\Kr}[1]{K^\mathrm{r}_{#1}}
118: \newcommand{\bj}{\mathbf{j}}
119: \newcommand{\bp}{\mathbf{p}}
120: \newcommand{\bra}{\langle}
121: \newcommand{\ket}{\rangle}
122: \renewcommand{\d}{\partial}
123: \newcommand{\dd}{\mathrm{d}}
124: \newcommand{\ord}[1]{O\!\left(#1\right)}
125: \newcommand{\comm}[2]{[#1,#2]}
126: \newcommand{\anticomm}[2]{\{#1,#2\}}
127: \newcommand{\sprod}[2]{(#1,#2)}
128: \newcommand{\Tr}{\operatorname{Tr}}
129: \newcommand{\Cl}{\operatorname{Cl_2}}
130: \newcommand{\tr}{\operatorname{tr}}
131: \newcommand{\ri}{\mathrm{i}}
132: 
133: % interpunction
134: \newcommand{\fs}{\; .}
135: \newcommand{\mc}{\; ;}
136: \newcommand{\co}{\; ,}
137: \newcommand{\nn}{\nonumber}
138: 
139: % fonts
140: \newcommand{\cred}[1]{\textcolor{red}{#1}}
141: \newcommand{\tocheck}[1]{\hermes\cred{#1}\rm}
142: \newcommand{\osn}[1]{\oldstylenums{#1}}
143: \def\query#1{\marginpar{\begin{flushleft}\footnotesize#1\end{flushleft}}}%
144: 
145: %abbreviations
146: \newcommand{\eom}{\sc eom\rm}
147: \newcommand{\lecs}{\sc lec\rm s}
148: \newcommand{\lec}{\sc lec\rm}
149: \newcommand{\chpt}{$\chi$\sc{pt}\rm}
150: \newcommand{\cpt}[1]{$\chi$\sc pt\rm${}^\gamma_{#1}$}
151: \newcommand{\opr}{\oldstylenums{1}\sc{pr}\rm}
152: \newcommand{\qcd}{\sc qcd\rm}
153: \newcommand{\lo}{\sc lo\rm}
154: \newcommand{\nlo}{\sc nlo\rm}
155: \newcommand{\nnlo}{\sc nnlo\rm}
156: \newcommand{\Mpl}{M_+}
157: \newcommand{\mpip}{M_{\pi^+}}
158: \newcommand{\Fpi}{F_\pi}
159: \newcommand{\Mpi}{M_\pi}
160: \newcommand{\Mk}{M_K}
161: \newcommand{\muP}{\mu_\pi}
162: \newcommand{\muPp}{\mu_{\pi^+}}
163: \newcommand{\muK}{\mu_K}
164: \newcommand{\muKp}{\mu_{K^+}}
165: \newcommand{\muE}{\mu_\eta}
166: \newcommand{\tree}
167: {\hspace*{1ex}{\raisebox{1.7ex}{\scriptsize $\circ$}\hspace{-2ex}}}
168: \newcommand{\treel}
169: {\hspace*{1.0ex}{\raisebox{1.7ex}{\scriptsize $\circ$}\hspace{-1.2ex}}}
170: 
171: \def\query#1{\marginpar{\begin{flushleft}\tiny#1\end{flushleft}}}%
172: 
173: 
174: 
175: % --------------------------------------------------------------------
176: 
177: \begin{document}
178: 
179: \thispagestyle{empty}
180: 
181: \begin{flushright}
182: IFIC/07--53
183: \vspace{\baselineskip}\par
184: \end{flushright}
185: 
186: % --------------------------------------------------------------------
187: 
188: \vspace*{0.8cm}
189: \begin{center}
190: {\Bigrm Electromagnetic low--energy constants in {\Large \chpt}}
191: \vskip 0.6cm
192: {\bigrm Christoph Haefeli$^1$, 
193: Mikhail A.~Ivanov$^2$, and
194: Martin Schmid$^3$ }
195: \vskip 3ex
196: {\bigrm\it 
197: $^1$Departament de F\'{\i}sica Te\`orica,
198: IFIC, Universitat de Val\`encia -- CSIC,}
199: \\
200: {\bigrm\it Apt. Correus 22085, E--46071 Val\`encia, Spain}
201: \\[2ex]
202: {\bigrm\it $^2$Laboratory of Theoretical Physics,
203: Joint Institute for Nuclear Research,}
204: \\
205: {\bigrm\it 141980 Dubna (Moscow region), Russia}
206: \\[2ex]
207: {\bigrm\it $^3$Institute for Theoretical Physics, University of Bern,}
208: \\
209: {\bigrm\it Sidlerstr. 5, 3012 Bern,  Switzerland } 
210: \end{center}
211: 
212: 
213: \vskip 0.8cm
214: \hrule
215: \vskip 2.0ex
216: {\small
217: \setlength{\baselineskip}{14pt}
218: \noindent
219: {\bf Abstract}
220: \vskip 1.0ex\noindent
221: We investigate three--flavour chiral perturbation theory including virtual
222: photons in a limit where the strange quark mass is much larger than the
223: external momenta and the up and down quark masses, and where the external
224: fields are those of two--flavour chiral perturbation theory. In particular we
225: work out the strange quark mass dependence of the electromagnetic two--flavour 
226: low--energy constants $\,C$ and $k_i$. We expect that these relations will be
227: useful for a more precise determination of the electromagnetic low--energy
228: constants. 
229: }
230: \vskip 2.0ex
231: \hrule
232: \vspace*{1cm}
233: 
234: \noindent
235: {\it Keywords:} Chiral symmetries, Chiral perturbation theory, Chiral
236: Lagrangians\\
237: 
238: \noindent
239: \sc{pacs}\rm: 11.30.Rd{}, {12.39.Fe}{},  {13.40.Dk}{},  {13.40.Ks}{}
240: 
241: 
242: \setlength{\baselineskip}{17pt}
243: 
244: % ------ SECT. INTRODUCTION ---------------------------------------------
245: \newpage
246: 
247: \section{Introduction}
248: 
249: \noindent
250: Chiral perturbation theory (\chpt) \cite{Weinberg,GL:Ann,GL:NPB} is the
251: effective theory of \qcd{} at low energies. It relies on an effective
252: Lagrangian whose coupling constants are the chiral low--energy constants
253: (\lecs{}). They are independent of the light quark masses and encode the
254: influence of the heavy degrees of freedom that are not contained in the
255: Lagrangian explicitly. For many phenomenological applications the predictivity
256: of \chpt{} depends on realistic estimates of these \lecs{}. An up--to--date
257: account of our knowledge about the \lecs{} can be found in the recent
258: conference reports of Ecker \cite{Ecker:lecs} and Bijnens \cite{Bijnens:lecs}.
259: 
260: In this article we are concerned with the electromagnetic \lecs{} of
261: \chpt{} in the (natural parity) meson sector including virtual photons. 
262: In the following we
263: abbreviate the effective theory with three flavours with
264: \cpt{3}~\cite{Urech,Neufeld}, and
265: accordingly  for two flavours with
266: \cpt{2}~\cite{Knecht.Urech,Steininger}. Recently, quite some progress has 
267: been achieved estimating the \nlo{} electromagnetic \lecs{} in \cpt{3}
268: \cite{Anant.Mouss,Baur.Urech,Bijnens.Prades,MoussallamGauge,K_i}, while
269: little is known about the pertinent \lecs{} in \cpt{2}, to the
270: best of our knowledge.  
271: In such a situation the following strategy may be pursued
272: \cite{Gasser:atom,Sazdjian-I}:
273: if one limits the external momenta to values small compared to the kaon and
274: eta mass and treats $m_u,m_d$ as small in comparison to $m_s$
275: %
276: \begin{equation}
277:   |p^2| \ll M_K^2 \co \qquad m_u,m_d \ll m_s \co
278: \end{equation}
279: %
280: the degrees of freedom of the kaons and the eta freeze. In this region one may work
281: out relations among the \lecs{} in \cpt{2} and \cpt{3}, which allow one to
282: estimate the electromagnetic \lecs{} in \cpt{2} through the knowledge of the
283: ones in \cpt{3}. 
284: The purpose of the present article is to
285: systematically provide \emph{all} relations between the
286: $\ord{e^2,e^2 p^2,e^4}$ 
287: \lecs{} in \cpt{2} and \cpt{3} at one--loop order. The calculation is performed along
288: the lines outlined in \cite{Nyffeler,Schmid:PhD,matchingI}. Briefly, the
289: method consists in a non--trivial matching between the three--flavour versus
290: the two--flavour generating functional of the effective theory.
291: \footnote{After we had performed these matching relations, we were informed
292:   by Marc Knecht that the relations at $\ord{e^2}$ and $\ord{e^2p^2}$ had also
293:   been derived by Nehme \cite{NehmePhD}. We agree with the
294:   relations given there.} 
295: 
296: We briefly comment on related work in the literature. Analogous relations
297: between the two--flavour and the three--flavour \lecs{} in the strong sector
298: have been provided by Gasser and
299: Leutwyler in \cite{GL:NPB}. Recently, we have worked out the same relations to
300: the next higher order (at two--loops) in the perturbative expansion
301: \cite{matchingI}, see also \cite{Mouss:Sigma,Schweizer:B} for earlier
302: contributions of such relations at two--loops. Finally, analogous work was
303: performed at one--loop accuracy in the baryonic sector in \cite{meissner_frink}.
304: 
305: The remainder of the article is organised as follows. After setting the
306: notation in Sec.~\ref{sec:ChPT_photons}, we give some details on the
307: derivation of the matching relations in Sec.~\ref{sec:method}. 
308: Section~\ref{sec:numerics}  contains a numerical analysis of
309: the matching relations, leading to estimates of the electromagnetic \nlo{}
310: \lecs{} in \cpt{2}. 
311: 
312: 
313: 
314: % --- ChPT WITH PHOTONS ----------------------------------------------------
315: 
316: \section{ Including virtual photons in  \chpt{}}
317: \label{sec:ChPT_photons}
318: 
319: A general procedure to construct the effective theory with photons in the
320: mesonic sector for three light flavours (\cpt{3}) has been proposed by Urech
321: \cite{Urech}. The two--flavour effective theory \cpt{2} may
322: be constructed along the same lines \cite{Knecht.Urech,Steininger}. We set our
323: notation, following the nomenclature of the \lecs{}
324: introduced by Urech \cite{Urech} for \cpt{3}, and by Knecht and Urech
325: \cite{Knecht.Urech} for \cpt{2}.\\ 
326: 
327: \noindent
328: The basic building block of the chiral Lagrangian is the Goldstone matrix
329: field  $u(\phi)$ which transforms under a chiral rotation $g=(g_L,g_R)\in
330: \SU(n)\times\SU(n)$ as 
331: %
332: \[u(\phi)\stackrel{g}{\longrightarrow} u(\phi')=g_Ru(\phi)h(g,\phi)^{-1} =
333: h(g,\phi)u(\phi)g_L^{-1}\co\]
334: %
335: where $h$ is called the compensator field. The mesonic Lagrangian then consists
336: of operators $X$ that either transform as
337: %
338: \begin{equation}
339: \label{eq:chirtrans}
340: X\stackrel{g}{\longrightarrow} h(g,\phi)Xh(g,\phi)^{-1}\co
341: \end{equation}
342: %
343: or remain invariant under chiral transformations. As a result, (products of)
344: traces of products of chiral operators $X$ are chiral invariant. 
345: The elementary building blocks of the effective Lagrangian that have the
346: transformation property of Eq.~(\ref{eq:chirtrans}) and furthermore contain 
347: the external vector $v_\mu$, axial $a_\mu$ (both traceless), scalar $s$, and
348: pseudoscalar $p$ sources are given by
349: % 
350: \begin{equation}
351: \label{eq:buildblockI}
352: \begin{split}
353:   u_\mu &= \ri\left[u^\dag (\d_\mu-\ri r_\mu)u-u (\d_\mu-\ri
354:     l_\mu)u^\dag\right]
355: \co\\
356:   \chi_\pm &= u^\dag\chi u^\dag\pm u\chi^\dag u\co
357: \end{split}
358: \end{equation}
359: %
360: where 
361: %
362: \begin{equation}
363:   \begin{split}
364:   r_\mu &= v_\mu+a_\mu+Q_RA_\mu\co\qquad l_\mu = v_\mu-a_\mu+Q_LA_\mu\co\\
365:   \chi &= 2B_0(s+ip)\co
366:   \end{split}
367: \end{equation}
368: %
369: with $A_\mu$ the photon field and $Q_{L,R}$ spurion sources with the
370: transformation properties
371: %
372: \begin{align*}
373:   Q_R\stackrel{g}{\longrightarrow} g_R\,Q_R\,g_R^\dag\co\qquad
374:   Q_L\stackrel{g}{\longrightarrow} g_L\,Q_L\,g_L^\dag\fs 
375: \end{align*}
376: %
377: They are also contained in the building blocks 
378: %
379: \begin{align}
380:   q_R = u^\dag Q_R u\co\qquad q_L=u Q_L u^\dag\co
381: \end{align}
382: %
383: which transform according to Eq.~(\ref{eq:chirtrans}).
384: Below, we will consider 
385: constant sources $Q_R = Q_L = Q$ only, 
386: and of phenomenological interest are the cases with two $(Q=Q_2)$, as well as
387: three light flavours $(Q=Q_3)$,
388: %
389: \begin{equation}
390: \label{eq:charge}
391:  Q_2=\tfrac{e}{3}\diag(2,-1)\co\qquad Q_3=\tfrac{e}{3}\diag(2,-1,-1)\fs
392: \end{equation}
393: %
394: Note that $Q_3$ is traceless, while $Q_2$ is not. For three flavours the
395: leading order Lagrangian reads in Euclidean space--time, 
396: %
397: \begin{equation}
398:   \begin{split}
399:     \mathcal{L}_2^{(3)} &= \frac{F_0^2}{4}\bra u\cdot u-\chi_+\ket-C_0\bra q_Lq_R\ket
400:     +\tfrac{1}{4}F_{\mu\nu}F_{\mu\nu}+\tfrac{1}{2}(\d_\mu A_\mu)^2
401:     \co
402: \end{split}
403: \end{equation}
404: %
405: where the superscript ${(3)}$ labels the number of flavours. Further, 
406: $u\cdot u \equiv u_\mu u_\mu$, $F_{\mu\nu} = \d_\mu A_\nu - \d_\nu A_\mu$
407: is the fieldstrength of the photon, and the gauge fixing term is put in the
408: Feynman gauge, as is customary in \cpt{2,3}. 
409: The symbol $\bra\cdot\ket$ denotes the trace of the flavour matrix enclosed.  
410: For two flavours the leading order Lagrangian amounts to have the same form
411: alike for three flavours, with the difference of restricting the $u$ fields to
412: elements of $\SU(2)$, and similar for the sources. Furthermore, the \lecs{}
413: $F_0$, $B_0$, and $C_0$ are to be replaced with $F$, $B$, and $C$,
414: respectively. To distinguish two-- from  three--flavour fields,
415: we decorate the former ones with a superscript $^\pi$. In summary, 
416: %
417: \begin{equation}
418:   \begin{split}
419:     \label{eq:lagr2}
420:     \mathcal{L}_2^{(2)} &= \frac{F^2}{4}\bra u^\pi\cdot u^\pi-\chi_+^\pi\ket
421:     -C \bra q_L^\pi q_R^\pi \ket
422:     +\tfrac{1}{4}F^\pi_{\mu\nu}F^\pi_{\mu\nu}
423:     +\tfrac{1}{2}(\d_\mu A^\pi_\mu)^2
424:     \fs
425: \end{split}
426: \end{equation}
427: %
428: % --- TABLE: STRONG MONOMIALS ----------------------------------------
429: %
430: \begin{table}[!ht]
431: \begin{center}
432: \def\arraystretch{1.5}
433: \begin{tabular}{rll}
434: %
435: \hline\hline
436: $j$ & $\binsp{-}x_j$ \hspace*{3cm} & $\binsp{-}X_j$ \\
437: \hline\\[-3ex]
438:  1  
439: & 
440: {} $ -\tfrac{1}{4}\bra u^\pi\cdot u^\pi\ket^2 $
441: &
442: {} $ -\bra u\cdot u\ket^2 $
443: \\
444: 2
445: &
446: {} $ -\tfrac{1}{4}\bra u^\pi_\mu u^\pi_\nu\ket^2 $
447: &
448: {} $ -\bra u_\mu u_\nu\ket^2 $
449: \\
450: 3
451: &
452: {} $ -\tfrac{1}{16}\bra\chi^\pi_+\ket^2 $
453: &
454: {} $ -\bra(u\cdot u)^2\ket $
455: \\
456: 4
457: &
458: $ \binsp{-} \tfrac{i}{4}\bra u^\pi_\mu\chi^\pi_{-\mu}\ket $
459: &
460: $ \binsp{-} \bra u\cdot u\ket\bra\chi_+\ket $
461: \\
462: 5
463: &
464: $ \binsp{-} \tfrac{1}{2}\bra {f^\pi_-}^2\ket $
465: &
466: $ \binsp{-} \bra u\cdot u\,\chi_+\ket $
467: \\
468: 6
469: &
470: {} $ -\tfrac{i}{4}\bra f^\pi_{+\mu\nu}[\,u^\pi_\mu\,,\,u^\pi_\nu\,]\ket $
471: &
472: {} $ -\bra\chi_+\ket^2 $
473: \\
474: 7
475: &
476: $ \binsp{-} \tfrac{1}{16}\bra\chi^\pi_-\ket^2 $
477: &
478: {} $ -\bra\chi_-\ket^2 $
479: \\
480: 8
481: &
482: {} $ -\tfrac{1}{8}(\det\chi^\pi_++\det\chi^\pi_-) $
483: &
484: {} $ -\tfrac{1}{2}\bra\chi_+^2+\chi_-^2\ket $
485: \\
486: 9
487: &
488: $ \binsp{-} \bra {\tilde{f}_+^{\pi\,2}}+{f^\pi_-}^2\ket $
489: &
490: $ \binsp{-} \tfrac{i}{2}\bra f_{+\mu\nu}[\,u_\mu\,,\,u_\nu\,]\ket $
491: \\
492: 10
493: &
494: {} $ -\tfrac{1}{16}\bra{\chi^\pi_+}^2-{\chi^\pi_-}^2\ket $
495: &
496: {} $ -\tfrac{1}{4}\bra f_+^2-f_-^2\ket $
497: \\
498: 11
499: &
500: {} $ -\tfrac{1}{4}\bra f_{+\mu\nu}^\pi\ket^2 $
501: &
502: {} $ -\tfrac{1}{2}\bra f_+^2+f_-^2\ket $
503: \\
504: 12
505: &
506: &
507: {} $ -\tfrac{1}{4}\bra\chi_+^2-\chi_-^2\ket $
508: \\[1ex]
509: \hline\hline\\[-1ex]
510: \end{tabular}
511: \caption{
512: \setlength{\baselineskip}{14pt}
513: Basis of the strong operators at order $p^4$ 
514: in Euclidean metric for two $(x_j)$ and for three flavours $(X_j)$.}
515: \label{tab:x_j} 
516: \end{center}
517: \end{table}
518: %
519: For the \nlo{} Lagrangian $\mathcal{L}_4$ we need the following additional
520: building blocks:
521: %
522: \begin{equation}
523: \label{eq:buildblockII}
524:   \begin{split}
525:     f_{\pm\mu\nu}&=ul_{\mu\nu}u^\dag\pm u^\dag r_{\mu\nu}u\co\\
526:     \chi_{\pm\mu}&=\nabla_\mu\chi_\pm-\tfrac{\ri}{2}\anticomm{\chi_\mp}{u_\mu}
527: \co\\
528:     q_{R\mu} & = \nabla_\mu q_R - \tfrac{\ri}{2}\comm{u_\mu}{q_R}\co\\
529:     q_{L\mu} & = \nabla_\mu q_L + \tfrac{\ri}{2}\comm{u_\mu}{q_L}\co
530:   \end{split}
531: \end{equation}
532: %
533: where we have introduced the field strengths
534: %
535: \begin{align*}
536:     y_{\mu\nu}&=\d_\mu y_\nu-\d_\nu y_\mu-\ri\comm{y_\mu}{y_\nu}
537: \co\qquad
538: y\,\in\, \{r,l\}
539: \co
540: \end{align*}
541: %
542: and the covariant derivative $\nabla_\mu$ in terms of the chiral connection
543: $\Gamma_\mu$,
544: %
545: \begin{align*}
546:   \nabla_\mu X&=\d_\mu X+[\Gamma_\mu,X]\co\\ 
547:   \Gamma_\mu&=\tfrac{1}{2}\left[u^\dag(\d_\mu-\ri r_\mu)u
548:   +u(\d_\mu-\ri l_\mu)u^\dag\right]\fs
549: \end{align*}
550: %
551: It is worth noting that $f_{+\mu\nu}$ for two--flavours is not traceless, since
552: the charge matrix $Q_2$ is not. To be in line with the basis operators
553: introduced by Gasser and Leutwyler \cite{GL:Ann}, it is convenient to
554: introduce in addition the traceless operator $\tilde{f}_{+\mu\nu}^\pi$,
555: %
556: \begin{equation}
557: \tilde{f}_{+\mu\nu}^\pi=f_{+\mu\nu}^\pi-\tfrac{1}{2}\bra f_{+\mu\nu}^\pi\ket
558: \fs
559: \end{equation}
560: %
561: The \nlo{} Lagrangians $\mathcal{L}_4$ then read
562: %
563: \begin{align}
564:   \mathcal{L}_4^{(2)} &= \sum_{j=1}^{11}l_jx_j+F^2\sum_{j=1}^{11}k_j w_j
565: +F^4\sum_{j=12}^{14}k_j w_j\co\\
566:   \mathcal{L}_4^{(3)} &= \sum_{j=1}^{12}L_jX_j+F_0^2\sum_{j=1}^{14}K_j W_j
567: +F_0^4\sum_{j=15}^{17}K_j W_j 
568: \fs
569: \end{align}
570: %
571: In the following, we call the
572: operators $x_j,X_j$ and the \lecs{} $l_j,L_j$ \emph{strong operators} and
573: \emph{strong \lecs{}}, as these operators do not vanish when switching off the
574: electromagnetic coupling constant, with the exception of
575: $x_{11}$. Accordingly, we denote the operators/\lecs{}
576: $w_j,W_j/k_j,K_j$ \emph{electromagnetic operators/\lecs{}}. The strong
577: operators $x_j,X_j$ have been introduced by Gasser and Leutwyler
578: \cite{GL:Ann,GL:NPB} and for convenience, we reproduce them here in
579: Tab.~\ref{tab:x_j}. The coefficients  $h_i$ and $H_i$ of the contact terms --
580: introduced by Gasser and Leutwyler in \cite{GL:Ann,GL:NPB} -- are related to
581: our \lecs{} as
582: %
583: \begin{equation}
584:   \label{eq:contact}
585:   \begin{split}
586:     h_1-h_3 &= l_8\co&\qquad h_2 &= l_9\co\qquad
587:     h_1+h_3=l_{10}\co\\ 
588:     H_1&=L_{11}\co&\qquad H_2&=L_{12}\fs
589:   \end{split}
590: \end{equation}
591: 
592: % --- TABLE: ELECTROMAGNETIC MONOMIALS --------------------------------
593: %
594: \begin{table}[!h]
595: \begin{center}
596: \def\arraystretch{1.5}
597: \begin{tabular}{rll}
598: %
599: \hline\hline
600: $j$ & $\binsp{-}w_j$ \hspace*{5cm} & $\binsp{-}W_j$ \\
601: \hline\\[-3ex]
602:  1  &  $ \binsp{-}\tfrac{1}{2}\bra u^\pi\!\cdot u^\pi\ket\bra
603:  q^\pi_R{}^2+q^\pi_L{}^2\ket $  &
604:        $ \binsp{-}\tfrac{1}{2}\bra u\!\cdot u\ket\bra q_R^2+q_L^2\ket $ \\
605: 
606:  2  &  $ \binsp{-}\bra u^\pi\!\cdot u^\pi\ket\bra q^\pi_R q^\pi_L\ket $  &
607:        $ \binsp{-}\bra u\!\cdot u\ket\bra q_R q_L\ket $ \\
608: 
609:  3  &  ${}-\bra u^\pi_\mu q^\pi_R\ket^2-\bra u^\pi_\mu q^\pi_L\ket^2 $ &
610:        ${}-\bra u_\mu q_R\ket^2-\bra u_\mu q_L\ket^2 $ \\
611: 
612:  4  &  $ \binsp{-}\bra u^\pi_\mu q^\pi_R\ket\bra u^\pi_\mu q^\pi_L\ket $ &
613:        $ \binsp{-}\bra u_\mu q_R\ket\bra u_\mu q_L\ket $ \\
614: 
615:  5  &  $ {}-\tfrac{1}{2}\bra\chi^\pi_+\ket\bra q^\pi_R{}^2+q^\pi_L{}^2\ket $ &
616:        $ \binsp{-}\bra u\!\cdot u (q_R^2+q_L^2)\ket $ \\  
617: 
618:  6  &  ${}-\bra\chi^\pi_+\ket\bra q^\pi_R q^\pi_L\ket$ &
619:        $ \binsp{-}\bra u\!\cdot u\anticomm{q_R}{q_L}\ket $  \\
620: 
621:  7  &  ${}-\tfrac{1}{2}\bra\chi^\pi_+(q^\pi_R+q^\pi_L)\ket\bra
622:  q^\pi_R+q^\pi_L\ket $ &  
623:        ${}-\tfrac{1}{2}\bra\chi_+\ket\bra q_R^2+q_L^2\ket $ \\ 
624: 
625:  8  &  ${}-\bra\chi^\pi_-\comm{q^\pi_R}{q^\pi_L}\ket $ &
626:        ${}-\bra\chi_+\ket\bra q_R q_L\ket $ \\
627: 
628:  9  &  $ \binsp{-}\ri\bra
629:  u^\pi_\mu(\comm{q^\pi_{R\mu}}{q^\pi_R}-\comm{q^\pi_{L\mu}}{q^\pi_L})\ket $ &  
630:        ${}-\bra\chi_+(q_R^2+q_L^2)\ket $ \\
631: 
632:  10 &  $ \binsp{-}\bra q^\pi_{R\mu}q^\pi_{L\mu}\ket $ &
633:        ${}-\bra\chi_+\anticomm{q_R}{q_L}\ket  $ \\
634: 
635:  11 &  $ \binsp{-}\bra q^\pi_R\!\cdot q^\pi_R+q^\pi_L\!\cdot q^\pi_L\ket $ &
636:        ${}-\bra\chi_-\comm{q_R}{q_L}\ket $ \\ 
637: 
638:  12 &  ${}-\tfrac{1}{4}\bra q^\pi_R{}^2+q^\pi_L{}^2\ket^2 $ &
639:        $ \binsp{-}\ri\bra u_\mu(\comm{q_{R\mu}}{q_R}-\comm{q_{L\mu}}{q_L})\ket
640:        $ \\ 
641: 
642:  13 &  ${}-\tfrac{1}{2}\bra q^\pi_R q^\pi_L\ket\bra
643:  q^\pi_R{}^2+q^\pi_L{}^2\ket $ &  
644:        $ \binsp{-}\bra q_{R\mu}q_{L\mu}\ket $ \\
645: 
646:  14 &  ${}-\bra q^\pi_R q^\pi_L\ket^2 $ & 
647:        $ \binsp{-}\bra q_R\!\cdot q_R+q_L\!\cdot q_L\ket $ \\
648: 
649:  15 & & $ {}-\bra q_R q_L\ket^2 $ \\
650: 
651:  16 & & $ {}-\tfrac{1}{2}\bra q_R q_L\ket\bra q_R^2+q_L^2\ket $ \\
652: 
653:  17 & & $ {}-\tfrac{1}{4}\bra q_R^2+q_L^2\ket^2 $\\[1ex]
654: \hline\hline\\[-1ex]
655: \end{tabular}
656: \caption{
657: \setlength{\baselineskip}{14pt}
658: Basis of the electromagnetic operators at order $e^2p^2$ and $e^4$ 
659: in Euclidean metric for two $(w_j)$ and for three flavours $(W_j)$.
660: The abbreviation $q_A\!\cdot q_A = q_{A\mu}q_{A\mu}$ for
661: $A\in\{R,L\}$ is adopted.}
662: \label{tab:w_j} 
663: \end{center}
664: \end{table}
665: %
666: \noindent
667: We are using the set of the \nlo{} electromagnetic operators from Knecht and
668: Urech \cite{Knecht.Urech} for two, and from Urech 
669: \cite{Urech} for three flavours, cf. Tab.~\ref{tab:w_j}. 
670: We do not consider the odd--intrinsic parity
671: sector which accounts for the axial anomaly, see
672: e.g.~\cite{WZW,Kaiser,MoussAnomaly}. 
673: For the counting we rely on the standard \cpt{} assignment 
674: with $e^2 \sim{} \ord{p^2}$ and we make use of the
675: convention to write $\ord{e^4,e^2p^2,p^4}$ as $\ord{p^4}$, and similarly for
676: the Landau symbol at order $p^6$. 
677: \clearpage
678: 
679: 
680: % --- MATCHING -------------------------------------------------------------
681: 
682: \section{Integrating out the strange quark}
683: \label{sec:method}
684: 
685: This section is devoted to give some details on the derivation of the main
686: results presented below in Eq.~(\ref{eq:results}). We will follow the steps
687: outlined in \cite{Nyffeler,Schmid:PhD}.
688: 
689: % - - - generating functional - - - - - - - - - - - - - - - - - - - - - - -
690: 
691: \subsection{Generating functional}
692: \label{sec:genfunc}
693: 
694: We start by considering the generating functional $Z$ of \cpt{3}
695: \cite{GL:NPB}, 
696: %
697: \begin{align}
698:   \mathrm{e}^{- Z[v,a,s,p,Q_{L,R}]} &= 
699:   \mathcal{N} \int[\dd u][\dd A_\mu]\, \mathrm{e}^{ - \int \dd^d x\,
700:     \mathcal{L}_{\mathrm{eff}}^{(3)} }
701:   \co\\[1ex]
702:   \mathcal{L}_{\mathrm{eff}}^{(3)} &= \mathcal{L}_2^{(3)} + \mathcal{L}_4^{(3)} + \ldots
703:   \fs
704: \end{align}
705: %
706: It may be evaluated in a low--energy
707: expansion in the number of loops,
708: %
709: \begin{equation}
710:   Z = Z_0 + Z_1 + \ldots
711:   \co
712: \end{equation}
713: %
714: where $Z_0$ $(Z_1)$ collects the tree--level (one--loop)
715: contributions.  They are given by
716: %
717: \begin{align}
718:   \label{eq:z0}
719:   Z_0 &= \bar{S}_2 \co
720:   \\
721:   \label{eq:z1}
722:   Z_1 &= \bar{S}_4 + \tfrac{1}{2}\ln\frac{\det D}{\det D^0}
723:   \co
724: \end{align}
725: %
726: where $\bar{S}_n$ denotes the classical action
727: %
728: \begin{equation}
729: \label{eq:barSn}
730:   \bar{S}_n = \int \dd^d x\, 
731:   \mathcal{L}_n^{(3)}(u^\text{cl},A^\text{cl},v,a,s,p,Q_{L,R})
732:   \co
733: \end{equation}
734: %
735: the Goldstone Boson fields $u^\text{cl}$ and the photon field $A^\text{cl}_\mu$ being evaluated
736: at the solution of the classical equation of motion (\eom{}), 
737: %
738: \begin{equation}
739:   \label{eq:EOM}
740:   \begin{split}
741:     \nabla_\mu
742:     u^\text{cl}_\mu+\tfrac{\ri}{2}\tilde{\chi}^\text{cl}_-+2\ri\tfrac{C_0}{F_0^2}\comm{q_R^\text{cl}}{q_L^\text{cl}}&=0\co\\
743:     \Delta A^\text{cl}_\mu-\tfrac{F_0^2}{2}\bra u^\text{cl}_\mu(q_R^\text{cl}-q_L^\text{cl})\ket&=0\co
744:   \end{split}
745: \end{equation}
746: where $\tilde{\chi}_-^\text{cl}$ denotes the traceless part of $\chi_-^\text{cl}\fs$
747: %
748: The Green's function of the differential operator $D$ [$D^0$] of
749: Eq.~(\ref{eq:z1}) is the [free] propagator $G(x,y)$\,$[G_0(x,y)]$,
750: %
751: \begin{equation}
752:  D_{AC}(x) G_{CB}(x,y) = \delta_{AB}\, \delta^{(d)}(x-y)
753: \fs
754: \end{equation}
755: %
756: The explicit form of $D$ was first given by Urech\footnote{Capital flavour indices
757:   $A,B,C,\ldots$ run from $1,\ldots,12$, lower case flavour indices from
758:   $1,\ldots,8$, they span the meson flavour space, and greek indices
759:   $\rho,\sigma,\ldots$ from $1,\ldots,4$ for the photon field components. The
760:   symbols $\lambda^a$ stand for the Gell--Mann matrices.} \cite{Urech},  
761: %
762: \begin{equation}
763:   \label{eq:diffop}
764:   D(x)=-\Sigma^2(x)+\Lambda(x)\co\qquad \Sigma_\mu(x)=\d_\mu^x+Y_\mu(x)\co
765: \end{equation}
766: %
767: with\footnote{To ease notation, we will drop
768:   from now on the label $_\text{cl}$ for the fields that satisfy the \eom{}.}
769: %
770: \begin{equation}
771: \begin{split}
772: \label{eq:loopvert}
773:   Y_\mu &=
774:   \begin{pmatrix}
775:     \hat\Gamma_\mu^{ab} & X_\mu^{a\rho}\\
776:     X_\mu^{\sigma b} & 0
777:   \end{pmatrix},\qquad\Lambda=
778:   \begin{pmatrix}
779:     \sigma^{ab} & \tfrac{1}{2}\gamma^{a\rho}\\
780:     \tfrac{1}{2}\gamma^{\sigma b} & \rho\delta^{\sigma\rho}
781:   \end{pmatrix}\co\\[1ex]
782:   \hat{\Gamma}_\mu^{ab}&=-\tfrac{1}{2}\bra\comm{\lambda^a}{\lambda^b}\Gamma_\mu\ket\co\\
783:   X_\mu^{a\rho} &= -X^{\rho a}_\mu =
784:   -\tfrac{F_0}{4}\delta_\mu^\rho\bra\lambda^a(q_R-q_L)\ket\co\\[1ex]
785:     \sigma^{ab}&=\binsp{+}\tfrac{1}{8}\bra\comm{\lambda^a}{u_\mu}\comm{\lambda^b}{u_\mu}\ket+\tfrac{1}{8}\bra\anticomm{\lambda^a}{\lambda^b}\chi_+\ket\\
786:     &\relsp{=}-\tfrac{C_0}{4F_0^2}\bra\big(\comm{\lambda^a}{q_R+q_L}\comm{\lambda^b}{q_R+q_L}
787:  -\comm{\lambda^a}{q_R-q_L}\comm{\lambda^b}{q_R-q_L}\big)\ket\\
788:     &\relsp{=}-\tfrac{F_0^2}{4}\bra\lambda^a(q_R-q_L)\ket\bra\lambda^b(q_R-q_L)\ket\co\\[1ex]
789:     \gamma^a_\mu &=
790:     \tfrac{F_0}{2}\bra\lambda^a\left\{\nabla_\mu(q_R-q_L)+i\comm{u_\mu}{q_R+q_L}\right\}\ket\co\\
791:     \rho &= \tfrac{3}{8}F_0^2\bra(q_R-q_L)^2\ket\fs
792: \end{split}
793: \end{equation}
794: %
795: \noindent
796: The generating functional for two flavours $z$ is defined analogously
797: to the one with three flavours. For later purposes we explicitly introduce its
798: low--energy expansion up to one--loop,
799: %
800: \begin{equation}
801:   z = \bar{s}_2 + \bar{s}_4
802:    + \tfrac{1}{2}\ln\frac{\det d}{\det d^{\,0}}
803:    + \ldots
804: \co
805: \end{equation}
806: %
807: where $\bar{s}_{2,4}$ and the operator $d$ are the two--flavour equivalent of
808: Eq.~(\ref{eq:barSn}) and Eq.~(\ref{eq:diffop}), and the ellipsis stand for
809: two--loop corrections and higher. 
810: For a state-of-the-art evaluation of the two--flavour functional and more
811: details we refer the reader to Schweizer \cite{Schweizer}. 
812: 
813: 
814: 
815: 
816: 
817: % - - - matching - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
818: 
819: \subsection{matching}
820: 
821: We impose now the following
822: constraints on the three--flavour functional:
823: %
824: \begin{itemize}
825: \item[i)]  the external sources of \cpt{3} are restricted to the two--flavour
826:   subspace. The generating functionals shall depend on the \emph{same}
827:   external sources;
828: \item[ii)] $m_{u,d} \ll m_s$,\\
829: since the \lecs{} of \cpt{2,3} are independent of $m_{u,d}$, 
830: we will work in the chiral limit for the up and down quark
831: masses, i.e.~$m_{u,d}=0$, for simplicity;
832: \item[iii)] external momenta are restricted to values below the threshold of
833:   the massive fields, $|p^2| \ll M_K^2$.
834: \end{itemize}
835: %
836: We will refer to the limit that satisfies i), ii) and iii) as \emph{the
837: two--flavour limit} of the three--flavour theory. In this limit the
838: three--flavour functional reduces to the two--flavour functional, i.e. both
839: theories yield the same Green's functions in the low--energy region, 
840: %
841: \begin{equation}
842: \label{eq:match_genF}
843:   Z = z
844:   \co
845: \end{equation}
846: %
847: provided the \lecs{} of both theories are accordingly matched. 
848: In the following, we will solve this equation for the \lecs{}. Both sides
849: receive non--local contributions that are associated to the propagation of
850: massless pions and photons. Once the matching is fully worked out, these contributions
851: cancel against each other. At order $p^4$, this has  
852: been discussed in detail by Nyffeler and Schenk
853: \cite{Nyffeler} and further details shall also be given elsewhere 
854: \cite{matchingII}. To find the relations among the \lecs{} it suffices 
855: therefore to work out the \emph{local} parts of the generating
856: functionals. At order $p^4$ we have in the two--flavour limit, 
857: %
858: \begin{equation}
859:   \label{eq:match_genF2}
860:   \bar{S}_2 + \bar{S}_4
861:   + \tfrac{1}{2}\ln\frac{\det D}{\det D^0}\Big|_\text{local} = 
862:   \bar{s}_2 + \bar{s}_4
863: \fs
864: \end{equation}
865: %
866: The lhs. of Eq.~(\ref{eq:match_genF2}) is now being worked out. 
867: We start with the tree--level
868: contributions, and proceed with the one--loop corrections. 
869: 
870: 
871: % - - - tree level: solution of the eom in the two-flavour limit - - - - - -
872: 
873: \subsubsection{Tree level: solution of the {\footnotesize EOM} in the two--flavour limit}
874: 
875: In view of the Eqs.~(\ref{eq:z0}\,,\,\ref{eq:z1}) we need to solve the
876: \eom{}
877: in the two--flavour 
878: limit. Due to the absence of strangeness containing external sources
879: [restriction i)] as well as
880: strangeness conservation, the following ansatz 
881: for the Goldstone Boson fields will turn out to be fruitful, 
882: %
883: \begin{equation}
884:   \label{eq:solfields}
885:   {u}={u}^\pi \mathrm{e}^{\tfrac{\ri}{2F_0}{\eta}\lambda_8}
886:   \co
887: \end{equation}
888: %
889: where in $u^\pi$ only the pions contribute non--trivially. 
890: Below, we will frequently identify without further notice
891: $3\times 3$ matrices that have only non--vanishing elements in their upper
892: left $2\times 2$ block with the $2\times 2$ matrices from the two--flavour
893: theory. Inserting the ansatz Eq.~(\ref{eq:solfields}) into the building blocks
894: Eqs.~(\ref{eq:buildblockI}) yields
895: %
896: \begin{equation}
897: \label{eq:transl1}
898: \begin{split}
899:       {u}_\mu&={u}^\pi_\mu-\tfrac{1}{F_0}\lambda^8\d_\mu{\eta}
900: \co\\[1ex]
901:     q_{L,R} &= q_{L,R}^\pi-e_{33}\bra q_{L,R}^\pi\ket
902: \co\\
903:     { \chi}_\pm&= \tfrac{B_0}{B}\chi_\pm^\pi\cos\alpha
904:      - \ri\tfrac{B_0}{B}\chi_\mp^\pi\sin\alpha+4B_0m_se_{33}
905:     \begin{cases}
906:        \;\cos 2\alpha & \!\! (+) \\
907:       \ri\sin 2\alpha & \! \! (-)
908:     \end{cases}
909: \co\\
910: \alpha &= \eta/(\sqrt{3}F_0) 
911: \co\qquad
912: e_{33} = \text{diag}(0,0,1)
913: \fs
914: \end{split}
915: \end{equation}
916: %
917: And similar for the building blocks of
918: Eqs.~(\ref{eq:buildblockII}),
919: %
920: \begin{equation}
921: \label{eq:transl2}
922: \begin{split}
923:     {f}_{+\mu\nu}&= f^\pi_{+\mu\nu}-e_{33}\bra f^\pi_{+\mu\nu}\ket
924: \co\qquad
925:     &{f}_{-\mu\nu}= {f}^\pi_{-\mu\nu}
926: \co\\
927:     q_{A\mu} &= q_{A\mu}^\pi
928: \co\qquad 
929: &A\in\{R,L\}
930: \fs
931: \end{split}
932: \end{equation}
933: %
934: Next, we write down the \eom{} of the $\eta$ field, 
935: %
936: \begin{align}
937: \label{eq:eometa}
938: \left(\Delta-{M}_{\eta}^2\right){\eta}
939: &=
940: \frac{F_0}{4\sqrt{3}}\left[\tfrac{B_0}{B}\bra\chi_+^\pi\ket\sin\alpha
941: +\ri\tfrac{B_0}{B}\bra\chi_-^{\pi}\ket\cos\alpha
942: +8B_0m_s\sin 2\alpha\right]-{M}_\eta^2{\eta}
943: \co
944: \end{align}
945: %
946: where $M_\eta^2 = \tfrac{4}{3} B_0m_s$ is the eta mass squared at tree level
947: at $m_{u,d}=0$ [similarly we will use $M_K^2 = B_0m_s$ below]. The \eom{} 
948: may be solved recursively for small $\alpha$. Note that the sum of the last
949: two terms in Eq.~(\ref{eq:eometa}) is of order $\alpha^3$. The differential
950: equation suggests a counting in which every occurrence of an $\eta$ particle
951: counts as order $p^2$ in the two--flavour limit,
952: %
953: \begin{align}
954:   \label{eq:eta}
955:   \eta &= -\frac{\ri\sqrt{3}F_0}{16B m_s}\bra{\chi}^\pi_-\ket + \ord{p^4}
956:   \fs
957: \end{align}
958: %
959: As a result, we obtain a systematic low--energy expansion of the $\SU(3)$
960: building blocks. 
961: %
962: \begin{figure}[!b]
963:   \centering
964:   \epsfig{file=./figures/detmix,height=7\baselineskip,bbllx=129,bblly=634,bburx=300,bbury=716}
965:   \caption{
966:     \setlength{\baselineskip}{14pt}
967:     (Left:) 
968:     Diagrammatic illustration for a specific contribution to
969:     $\ln\det D_K$ in Eq.~(\ref{eq:det}). The vertices denote insertions from
970:     external fields, related to the operator $D$. Only kaons flow through the
971:     loop. The whole determinant
972:     consists of a sum of such diagrams, ordered by an increasing number of
973:     insertions, being equivalent to the low--energy expansion introduced in
974:     the text.
975:     (Right:)
976:     Pion--eta mixing at order $p^4$. In the two--flavour limit this diagram
977:     yields to local terms only.
978:   }
979:   \label{fig:det.mix}
980: \end{figure}
981: %
982: 
983: Before proceeding, we add a remark: to be precise, the pions of
984: \cpt{3} differ from their two--flavour equivalent, since
985: they satisfy different \eom{}s. Indeed, in the two--flavour limit, we find
986: %
987: \begin{equation}
988: \label{eq:su3eom}
989:   \nabla_\mu
990: u^\pi_\mu+\tfrac{\ri}{2}\tfrac{B_0}{B}\tilde{\chi}^\pi_-\cos\alpha+\tfrac{1}{2}\tfrac{B_0}{B}\tilde{\chi}^\pi_+\sin\alpha+2\ri\tfrac{C_0}{F_0^2}\comm{q_R^\pi}{q_L^\pi}=0\co
991: \end{equation}
992: %
993: to be compared with the \eom{} of \cpt{2}, 
994: %
995: \begin{equation}
996:   \nabla_\mu
997: u^\pi_\mu+\tfrac{\ri}{2}\tilde{\chi}^\pi_-+2\ri\tfrac{C}{F^2}\comm{q_R^\pi}{q_L^\pi}=0\fs
998: \end{equation}
999: %
1000: However, expanding the trigonometric functions in Eq.~(\ref{eq:su3eom}), one
1001: observes that the difference is of order $p^4$; hence, it affects the matching
1002: relations only beyond the accuracy we are working. Also at the level of
1003: $\bar{S}_2$ -- which is the only one to matter for the \eom{} -- the \lecs{} of
1004: both theories coincide due to the matching condition (\ref{eq:match_genF}). These are the reasons why we
1005: do not distinguish between the pions of both theories in this article. However, if one wishes to
1006: carry out the matching beyond one--loop order, this issue requires a
1007: considerable deeper examination, see also  \cite{matchingII}. Similar remarks
1008: apply for the photon.  
1009: 
1010: By now, it is straight forward to evaluate the tree level
1011: diagrams of 
1012: $\bar{S}_{2,4}$ in the two--flavour limit using
1013: Eqs.~(\ref{eq:transl1}\,,\,\ref{eq:transl2}\,,\,\ref{eq:eta}). To avoid
1014: overflowing formulae, for $\bar{S}_4$
1015: we only show the reduction of the electromagnetic
1016: operators  explicitly. We find,
1017: %
1018: \begin{equation}
1019: \label{eq:treeResults}
1020: \begin{split}
1021: \bar{S}_2 &= 
1022: \int\dd^d x\,\Big[
1023: \tfrac{F_0^2}{4} \bra u^\pi\cdot u^\pi \ket
1024: -\tfrac{F_0^2}{4} (B_0/B) \bra \chi_+^\pi \ket
1025: -C_0 \bra q_L^\pi q_R^\pi \ket
1026: +\tfrac{1}{4}F^\pi_{\mu\nu}F^\pi_{\mu\nu}
1027: +\tfrac{1}{2}(\d_\mu A^\pi_\mu)^2
1028: \\
1029: &
1030: \hspace*{1.5cm}
1031: + \tfrac{F_0^2}{8\Mk^2} (B_0/B)^2\, x_7
1032: \Big] 
1033: + \ord{p^6}
1034: \co
1035: \\[1ex]
1036: \bar{S}_4 &= 
1037: F_0^2\int\dd^d x\,\Big\{
1038: -4\Mk^2K_{8}\bra q_L^\pi q_R^\pi \ket
1039: +(\tfrac{6}{5}K_{1} + \tfrac{1}{5}K_{2} + K_{5})\,w_{1}
1040: + (K_{2}+K_{6})\,w_{2}
1041: \\
1042: &
1043: \hspace*{1.5cm}
1044: + K_{3}\,w_{3}
1045: + K_{4}\,w_{4}
1046: +(\tfrac{6}{5}K_{7}+\tfrac{1}{5}K_{8}+\tfrac{4}{5}K_{9}-\tfrac{1}{5}K_{10}
1047: )\,w_{5}
1048: +(K_{8}+K_{10})\,w_{6}
1049: \\
1050: &
1051: \hspace*{1.5cm}
1052: +(K_{9}+K_{10})\,w_{7}
1053: +K_{11}\,w_{8}
1054: +K_{12}\,w_{9}
1055: +K_{13}\,w_{10}
1056: +K_{14}\,w_{11}
1057: \\
1058: &
1059: \hspace*{1.5cm}
1060: +F_0^2\big[(\tfrac{1}{25}K_{15}+\tfrac{6}{25}K_{16}+\tfrac{36}{25}K_{17})\,w_{12}
1061: +(\tfrac{2}{5}K_{15}+\tfrac{6}{5}K_{16})\,w_{13}
1062: +K_{15}\,w_{14}\big]
1063: \Big\}
1064: \\[0.5ex]
1065: &+\bar{S}_4\big|_\mathrm{strong} 
1066: + \ord{p^6}
1067: \fs
1068: \end{split}
1069: \end{equation}
1070: %
1071: 
1072: 
1073: 
1074: % - - - one-loop calculation in the two-flavour limit  - - - - - - - - - - -
1075: 
1076: \subsubsection{Loops: the determinant in the two--flavour limit}
1077: 
1078: The determinant of the differential operator $D$ covers all one--loop diagrams
1079: of the generating functional. Its evaluation in terms of an expansion in
1080: external fields in the two--flavour limit may be worked out as follows. To
1081: begin with, we note that the contributions from the massless (pions 
1082: and photon) and from the massive fields (kaons and eta) may be separated
1083: \cite{Nyffeler,Schmid:PhD},  
1084: %
1085: \begin{equation}
1086: \begin{split}
1087:   \label{eq:det}
1088: \ln\det D &= \ln\det D_{\ell}+\ln\det D_\eta+\ln\det D_{K}
1089: +\ln\det(1-D_{\pi}^{-1}D_{\pi\eta}D_\eta^{-1} D_{\eta\pi})
1090: \fs
1091: \end{split}
1092: \end{equation}
1093: %
1094: The first determinant $\ln\det D_\ell$ involves contributions from pions
1095: and photons only, it is a purely non--local object and for the matching of the
1096: \lecs{} needs not to be considered any further. The operators $D_\eta$, $D_K$
1097: are related to heavy particles only: their determinant describe tadpoles with
1098: insertions where only particles of identical masses run in the loop, either
1099: etas or kaons, cf. Fig.~\ref{fig:det.mix} (left). Diagrams of this type are
1100: efficiently calculated with heat--kernel methods. It
1101: results in an expansion in terms of local quantities involving an increasing
1102: number of derivatives, which corresponds to an expansion in powers of
1103: momenta. The last determinant on the rhs. of Eq.~(\ref{eq:det}) is a more
1104: complicated object. Since it involves the operator $D_\pi$ of the massless
1105: modes, it is not a purely local object. However, its non--locality in the
1106: low--energy expansion only shows up at order $p^6$ and can therefore be
1107: neglected for our analysis. This is due to the symmetric operator
1108: $D_{\pi\eta}$, which mediates $\pi-\eta$ mixing and is of order $p^2$. At
1109: order $p^4$, the operators $D_\pi^{-1}$ and $D_\eta^{-1}$ may be replaced by
1110: their free propagators. As a result, the diagram of Fig.~\ref{fig:det.mix}
1111: (right) is the only contribution from the last term in Eq.~(\ref{eq:det}) at
1112: order $p^4$. And the low--energy expansion of this diagram yields to local
1113: terms only. The photon does not show up in this mixing term, because the eta
1114: is not charged. We find
1115: %
1116: \begin{equation}
1117:   \tfrac{1}{2}\ln(\det D/\det D^0)\big|_\mathrm{local} = 
1118: \int\! \dd^dx\,
1119:  \Big(\mathcal{L}^\eta_{1\,\mathrm{loop}}
1120: +\mathcal{L}^K_{1\,\mathrm{loop}}
1121: +\mathcal{L}^{\pi\eta}_{1\,\mathrm{loop}}\Big)+\ord{p^6}
1122: \co
1123: \end{equation}
1124: %
1125: \begin{equation}
1126: \begin{split}
1127: \label{eq:loops}
1128: \mathcal{L}^\eta_{1\,\mathrm{loop}} 
1129: &=
1130: \tfrac{1}{24}F_1(M_\eta^2) 
1131: \big[ \bra{\chi}_+^\pi\ket + M_K^{-2}\,x_7 \big]
1132:   +\tfrac{1}{36}F_2(M_\eta^2)\,x_3
1133: \co
1134: \\[1.5ex]
1135: \mathcal{L}^K_{1\,\mathrm{loop}} 
1136: &=
1137: -\tfrac{1}{4}F_1(M_K^2)
1138: \big[ \bra{u}_\mu^\pi{u}_\mu^\pi-{\chi}^\pi_+\ket 
1139:      -8\tfrac{C_0}{F_0^2}\bra q_R^\pi q_L^\pi \ket\big]
1140: \\
1141: &
1142: \hspace*{-0.9cm}
1143: +\tfrac{1}{48}F_2(M_K^2)\Big(x_1+2x_2+12x_4-2x_5-4x_6
1144: -12x_7+x_9+24x_{10}-18x_{11}
1145: \\
1146: &
1147: \hspace*{-0.9cm}
1148: +\tfrac{C_0}{F_0^2}
1149: \Big[\tfrac{48}{5}w_1+12w_2+\tfrac{12}{5}w_5+12w_6+36w_7
1150: -6w_8
1151: +\tfrac{C_0}{F_0^2}\big(
1152: \tfrac{2496}{25}w_{12}+\tfrac{816}{5}w_{13}+48w_{14}\big)\Big]\Big)
1153: \co
1154: \\[1.5ex]
1155: \mathcal{L}^{\pi\eta}_{1\,\mathrm{loop}} 
1156: &=
1157:  -\tfrac{1}{6}F_2^1(M^2_\eta)\big(x_7+x_8-x_{10}\big)
1158: \co
1159: \end{split}
1160: \end{equation}
1161: %
1162: where $F_n^l(m^2)$ denote loop integrals, 
1163: %
1164: \begin{equation}
1165: F_n^l(m^2)=
1166: \int\!\frac{\dd^dq}{(2\pi)^d}\,\frac{1}{(m^2+q^2)^{n-l}(q^2)^l}
1167: \co\qquad
1168: n>l\geq0
1169: \co\qquad
1170: F_n(m^2) \equiv F_n^0(m^2)
1171: \co
1172: \end{equation}
1173: %
1174: which are well--known, cf. e.g. Ref.~\cite{GasserSainio}. 
1175: The renormalisation is carried out in the $\overline{\mathrm{MS}}$--scheme,
1176: where the 
1177: \lecs{} $c_i\in\{l_i,k_i,L_i,K_i\}$ are splitted
1178: into a divergent and a finite part as follows,
1179: %
1180: \begin{equation}
1181: \begin{split}
1182: c_i & = \alpha_i \lambda + c_i^\mathrm{r}(\mu,d)\co
1183: \qquad
1184: c_i^\mathrm{r}(\mu) \, \equiv \, c_i^\mathrm{r}(\mu,4)  \co\\[1ex]
1185: \lambda &= \frac{\mu^{d-4}}{16\pi^2} 
1186: \Big\{\frac{1}{d-4}-\tfrac{1}{2} [\ln 4\pi + \Gamma'(1) + 1] \Big\}.
1187: \end{split}
1188: \end{equation}
1189: %
1190: The coefficients $\alpha_i\in\{\gamma_i,\sigma_i,\Gamma_i,\Sigma_i\}$  are
1191: given in Refs.~\cite{GL:Ann},\cite{Knecht.Urech},\cite{GL:NPB} and
1192: \cite{Urech}\footnote{For $\Sigma_{15,16,17}$ consult
1193:   Eq.~(12) in \cite{Steininger}.}.
1194: 
1195: 
1196: % - - - results - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
1197: 
1198: \subsection{Results}
1199: 
1200: Collecting the results of the tree [Eq.~(\ref{eq:treeResults})] and one--loop
1201: [Eq.~(\ref{eq:loops})] analysis allows us to determine
1202: the relationship among the three-- and two--flavour
1203: \lecs{} via Eq.~(\ref{eq:match_genF2}). The results are
1204: %
1205: % \begin{equation}
1206: \begin{align}
1207: \label{eq:results}
1208: %\begin{split}
1209:    C&=C_0\left[1-4\mu_K\right]
1210:  +4 M_K^2 F_0^2 \Kr{8} 
1211: \co\nonumber
1212: \\[0.5ex]
1213:   \kr{1}&= \tfrac{6}{5}\Kr{1}
1214:  +\tfrac{1}{5}\Kr{2}+\Kr{5}-\tfrac{2}{5}Z_0\nu_K
1215: \co\nonumber
1216: \\[0.5ex]
1217:   \kr{2}&= \Kr{2}+\Kr{6}
1218: -\tfrac{1}{2}Z_0\nu_K
1219: \co\nonumber
1220: \\[0.5ex]
1221:   \kr{3}&= \Kr{3}
1222: \co\nonumber
1223: \\[0.5ex]
1224:   \kr{4}&= \Kr{4}
1225: \co\nonumber
1226: \\[0.5ex]
1227:   \kr{5}&= \tfrac{6}{5}K_7
1228: +\tfrac{1}{5}\Kr{8}+\tfrac{4}{5}\Kr{9}-\tfrac{1}{5}\Kr{10}
1229: -\tfrac{1}{10}Z_0\nu_K
1230: \co\nonumber
1231: \\[0.5ex]
1232:   \kr{6}&= \Kr{8}+\Kr{10}
1233: -\tfrac{1}{2}Z_0\nu_K
1234: \co\nonumber
1235: \\[0.5ex]
1236:    k_7&= \Kr{9}+\Kr{10}
1237: -\tfrac{3}{2}Z_0\nu_K
1238: \co\nonumber
1239: \\[0.5ex]
1240:   \kr{8}&= \Kr{11}-2Z_0(2\Lr{4}
1241: +\Lr{5})+\tfrac{1}{4}Z_0\nu_K
1242: \co\nonumber
1243: \\[0.5ex]
1244:   \kr{9}&= \Kr{12}
1245: \co\nonumber
1246: \\[0.5ex]
1247:    k_{10}&= K_{13}
1248: \co\nonumber
1249: \\[0.5ex]
1250:    k_{11}&= K_{14}
1251: \co\nonumber
1252: \\[0.5ex]
1253:   \kr{12}&= \tfrac{1}{25}\Kr{15}
1254: +\tfrac{6}{25}\Kr{16}+\tfrac{36}{25}\Kr{17}-\tfrac{104}{25}Z_0^2\nu_K
1255: \co\nonumber
1256: \\[0.5ex]
1257:   \kr{13}&= \tfrac{2}{5}\Kr{15}
1258: +\tfrac{6}{5}\Kr{16}-\tfrac{34}{5}Z_0^2\nu_K
1259: \co\nonumber
1260: \\[0.5ex]
1261:   \kr{14}&= \Kr{15}-2Z_0^2\nu_K
1262: \co\nonumber
1263: \\[0.5ex]
1264:    l_{11}&=\tfrac{3}{2}\Lr{10}+3\Lr{11}
1265: +\tfrac{3}{4}\nu_K 
1266: \co
1267: %\end{split}
1268: %\end{equation}
1269: \end{align}
1270: %
1271: where we introduced the abbreviations, 
1272: %
1273: \begin{equation}
1274: \begin{split}
1275:   Z_0 &= C_0/F_0^4 \co \qquad
1276:   \mu_K\,=\,\frac{M_K^2}{32\pi^2F_0^2}\ln\frac{M_K^2}{\mu^2}
1277: \co \qquad
1278:   \nu_K\,=\,\frac{1}{32\pi^2}\left(\ln\frac{M_K^2}{\mu^2}+1\right) 
1279: \fs
1280: \end{split}
1281: \end{equation}
1282: %
1283: These relations are the main results of our article and deserve a few
1284: comments:
1285: %
1286: \begin{itemize}
1287: \item[--] 
1288:   we only display the matching relations for
1289:   the electromagnetic \lecs{} as well as the strong \lec{} $l_{11}$. The ones
1290:   for the remaining strong \lecs{} may be established along the very same lines
1291:   and may be found in \cite{GL:NPB,matchingI};
1292: \item[--]
1293:   at first glance it might come as a surprise that strong three--flavour
1294:   \lecs{}  show up in the matching of the electromagnetic \lecs{},
1295:   cf. $\kr{8}$. Along the derivation presented here it is the \eom{} which
1296:   links strong and electromagnetic operators and needs to be used to project
1297:   the operators $X_4$ and $X_5$ (in the two--flavour limit) into the
1298:   two--flavour 
1299:   basis; 
1300: \item[--]
1301:   briefly, we go back again to Eq.~(\ref{eq:loops}). We remark that
1302:   only the kaon loop contributes non--trivially to the matching of the
1303:   electromagnetic \lecs{}. Having in mind that along the method presented
1304:   here, we had performed the matching for the strong \lecs{} $l_i$ at
1305:   two--loops 
1306:   already before \cite{matchingI,matchingII}, the relations in
1307:   Eq.~(\ref{eq:results}) come at almost no additional cost. Exactly here lies the beauty
1308:   of this approach: once the framework is set, all relations
1309:   of the \lecs{} are obtained in one strike at a reasonable amount of effort;
1310: \item[--] 
1311:   one verifies that the dependence on the scale $\mu$ of the left and
1312:   right hand side of Eq.~(\ref{eq:results}) is the same;
1313: \item[--]
1314:   specific linear combinations for electromagnetic matching relations were
1315:   presented earlier in Ref.~\cite{Gasser:atom}. We completely agree with
1316:   the relations given there. Some of the relations were also given in
1317:   Ref.~\cite{Sazdjian-I}; see Ref.~\cite{Sazdjian-II} for comments why some
1318:   of the relations in Ref.~\cite{Sazdjian-I} were given erroneously;
1319: \item[--]
1320:   the approach that was advocated in
1321:   Refs.~\cite{Sazdjian-I,Gasser:atom,NehmePhD} relied on an analysis of
1322:   physical 
1323:   observables -- e.g. the charged pion mass or $\pi\pi$--scattering -- both in
1324:   \cpt{3} as well as in \cpt{2}. The functional relationship among (linear
1325:   combinations of) the \lecs{} emerges then from a comparison of the two
1326:   representations in a large $m_s$ expansion. The agreement between the results
1327:   found in the literature and those presented here based on the generating
1328:   functional provides a thorough and welcome check on both calculations.
1329: \end{itemize}
1330: 
1331: 
1332: % --- NUMERICAL ANALYSIS ---------------------------------------------------
1333: 
1334: \section{Numerical analysis}
1335: \label{sec:numerics}
1336: 
1337: The relations derived in Eq.~(\ref{eq:results}) are useful to obtain
1338: constraints on the pertinent \lecs{}. In
1339: the strong sector for instance, already in the early days of \chpt{} Gasser and
1340: Leutwyler made use of such relations to convert information about the values
1341: of the two--flavour \lecs{} to determine three--flavour \lecs{}; e.g.~an
1342: estimate for $L_2$ was obtained in this way via $l_2$. 
1343: The here presented formulae may by used in an analogous fashion: to the best
1344: of our knowledge only estimates for the \lecs{} in \cpt{3} are known 
1345: \cite{Anant.Mouss,Baur.Urech,Bijnens.Prades,MoussallamGauge,K_i}. The matching
1346: relations may thus be inferred to obtain estimates on
1347: the values of the two--flavour coupling constants. Two remarks shall be
1348: pointed out which are relevant in the numerical determination of the \lecs{}:  
1349: first, some of the \lecs{} depend on the gauge
1350: \cite{BijnensGauge,MoussallamGauge}. The values inferred below are evaluated in
1351: the Feynman gauge. Second, due to ultraviolet
1352: divergences generated by photon loops, the splitting of the Hamiltonian of
1353: $\text{\qcd{}} + \gamma$ into a strong and an electromagnetic piece is
1354: ambiguous. This ambiguity must be reflected also in the effective theory in
1355: the \lecs{} \cite{Scimemi}. Estimates of their sizes should therefore take
1356: this into account. The authors of Ref.~\cite{Scimemi} have discussed the
1357: problem in detail on the basis of fieldtheoretic models and come up with a
1358: proposal how the ambiguity may be addressed systematically within these
1359: models. Still, this delicate issue has not been investigated in the literature
1360: for the \cpt{2,3} \lecs{} yet and is beyond the scope we are aiming at here. 
1361: 
1362: %
1363: % --- TABLE: Kr[i] ---------------------------------------------------------
1364: %
1365: \begin{table}[t]
1366: \begin{center}
1367: \def\arraystretch{1.1}
1368: \begin{tabular}{lrclr}
1369: \hline\hline\\[-1ex]
1370: $\Kr{1}$
1371: &
1372: $- 2.7$
1373: &
1374: \hspace*{1cm}
1375: &
1376: $K_7$
1377: &
1378: $0$
1379: \\
1380: $\Kr{2}$
1381: &
1382: $0.7$
1383: &&
1384: $\Kr{8}$
1385: &
1386: $0$
1387: \\
1388: $\Kr{3}$
1389: &
1390: $2.7$
1391: &&
1392: $\Kr{9}$
1393: &
1394: --
1395: \\
1396: $\Kr{4}$
1397: &
1398: $1.4$
1399: &&
1400: $\Kr{10}$
1401: &
1402: $4.0$
1403: \\
1404: $\Kr{5}$
1405: &
1406: $11.6$
1407: &&
1408: $\Kr{11}$
1409: &
1410: $1.3$
1411: \\
1412: $\Kr{6}$
1413: &
1414: $2.8$
1415: &&
1416: $\Kr{12}$
1417: &
1418: $-4.2$
1419: \\
1420: &&&
1421: $K_{13}$
1422: &
1423: $4.7$
1424: \\[1ex]
1425: \hline\hline\\[-1ex]
1426: \end{tabular}
1427: \caption{
1428: Values of electromagnetic {\lecs{}} in
1429:   {\cpt{3}} in units of $10^{-3}$ at the scale $\mu=M_\rho=0.77\,\mathrm{GeV}$,
1430:   in the Feynman gauge. The values for $\Kr{1},\ldots, \Kr{6}$ are
1431:   invoked from \cite{Anant.Mouss}, and $K_7,\ldots,K_{13}$ from
1432:   \cite{MoussallamGauge}, see
1433:   also text for further details. 
1434: }   
1435: \label{tab:Kr}
1436: \end{center}
1437: \end{table}
1438: 
1439: %
1440: % --- TABLE: kr[i] ---------------------------------------------------------
1441: %
1442: \begin{table}[b]
1443: \begin{center}
1444: \def\arraystretch{1.1}
1445: \begin{tabular}{lrclr}
1446: \hline\hline\\[-1ex]
1447: $\kr{1}$
1448: &
1449: $8.4$
1450: &
1451: \hspace*{1cm}
1452: &
1453: $\kr{6}$
1454: &
1455: $3.9$
1456: \\
1457: $\kr{2}$
1458: &
1459: $3.4$
1460: &&
1461: $k_7$
1462: &
1463: $3.7+\Kr{9}\cdot 10^3$
1464: \\
1465: $\kr{3}$
1466: &
1467: $2.7$
1468: &&
1469: $\kr{8}$
1470: &
1471: $-1.4$
1472: \\
1473: $\kr{4}$
1474: &
1475: $1.4$
1476: &&
1477: $\kr{9}$
1478: &
1479: $-4.2$
1480: \\
1481: $\kr{5}$
1482: &
1483: $-0.8+4/5\Kr{9}\cdot 10^3$
1484: &&
1485: $k_{10}$
1486: &
1487: $4.7$
1488: \\[1ex]
1489: \hline\hline\\[-1ex]
1490: \end{tabular}
1491: \caption{Values of electromagnetic low--energy constants in
1492:   {\cpt{2}} in 
1493:   units of $10^{-3}$ at the scale $\mu=M_\rho=0.77\,\mathrm{GeV}$, in the
1494:   Feynman gauge. From general dimensional arguments, one might attribute an
1495:   uncertainty of $1/(16\pi^2)\approx 6.3\cdot 10^{-3}$ to each \lec{}.}
1496: \label{tab:kr}
1497: \end{center}
1498: \end{table}
1499: %
1500: 
1501: We proceed with the numerical analysis of Eq.~(\ref{eq:results}). At the
1502: accuracy we are working, we may identify $F_0$ with the pion decay constant
1503: $F_\pi=92.4\,\mathrm{MeV}$ and  $M_K^2 = B_0 m_s$ with
1504: %
1505: \begin{equation}
1506:   B_0 m_s = M_{K^+}^2 - M_{\pi^+}^2/2 \simeq (485 \,\mathrm{MeV})^2
1507: \fs
1508: \end{equation}
1509: %
1510: We further set $Z_0 = 0.91$ (obtained from Ref.~\cite[Eq.~(42)]{MoussallamGauge}
1511: with $M_V = 0.77\,\mathrm{GeV}$, $z\equiv M_A^2/M_V^2 = 2$). For the values of
1512: the three--flavour \nlo{} electromagnetic \lecs{} we will stick
1513:  for $\Kr{1},\ldots,\Kr{6}$ to \cite{Anant.Mouss}, and  for
1514: $K_7,\ldots,\Kr{12}$ to \cite{MoussallamGauge}, summarised here in 
1515: Tab.~\ref{tab:Kr}. The \lec{} $\Kr{9}$
1516: remained undetermined in \cite{MoussallamGauge}, as it yet suffers from a
1517: reliable estimate (consult
1518: \cite{MoussallamGauge} for the details). As a result we will not give a
1519: numerical estimate for $\kr{5}$ and $k_7$. For the \lecs{} $\Kr{11},\Kr{12}$
1520: (from Eq.(59) and (61) in \cite{MoussallamGauge}) we furthermore set $\mu_0 =
1521: 1\,\mathrm{GeV}$ for the \qcd{} scale, the parameters $M_V$ and $z$ are as
1522: introduced above. The coupling constants $\Kr{14},\ldots,\Kr{17}$ are
1523: associated to contact operators and/or operators at order $e^4$ and are not
1524: considered in this section. Furthermore, $\Lr{4} = 0$ and $\Lr{5} = 1.5\cdot
1525: 10^{-3}$, taken from the $\ord{p^4}$ fit in \cite{BijnensFit}. All \lecs{} are
1526: evaluated at the scale $\mu=0.77\,\mathrm{GeV}$, and in the Feynman gauge
1527: $\xi=1$. The results for the so obtained electromagnetic two--flavour \lecs{} 
1528: are finally summarised in Tab.~\ref{tab:kr}. As an illustration we also show in
1529: Fig.~\ref{fig:kr1kr2.eps} the strange quark mass dependence of $\kr{1}$ and
1530: $\kr{2}$. We observe that the two--flavour \lecs{} only show 
1531: a very moderate strange quark mass dependence in the neighbourhood of the
1532: physical point. This pattern is due to its solely logarithmic strange quark
1533: mass dependence at this order of the matching. Note 
1534: that the matching relations only apply over a certain range for
1535: the strange quark mass: the formulae break down for $m_s \rightarrow 0$, as
1536: the expansion performed here requires that all external momenta are much
1537: smaller than $m_s$. Remarkably, the pertinent chiral logarithm becomes dominant
1538: numerically only for very small $m_s$. 
1539: On the other hand as one increases the strange
1540: quark mass, higher order contributions in the matching expansion become more
1541: dominant and start to spoil the behaviour of the chiral logarithm. This was
1542: discussed in more detail in \cite{matchingI} for the strong \lecs{} $l_i$. 
1543: 
1544: %
1545: % --- FIGURE 2 ------------------------------------------------------------
1546: %
1547: \begin{center}
1548: \begin{figure}[t]
1549: \begin{minipage}[]{1.0\linewidth}
1550: \epsfig{file=./figures/kr1.eps,width=0.48\linewidth}
1551: \hfill
1552: \epsfig{file=./figures/kr2.eps,width=0.48\linewidth}\\
1553: \end{minipage}
1554: \caption{Strange quark mass dependence of the electromagnetic two--flavour
1555:   {\lecs{}} $\kr{1}$ (left) and $\kr{2}$ (right) in units of $10^{-3}$  at the
1556:   scale $\mu=M_\rho=0.77\,\mathrm{GeV}$. The physical value of $m_s$ 
1557:   corresponds to $M_K^2 = B_0m_s \approx (485\,\mathrm{MeV})^2$. 
1558: }
1559: \label{fig:kr1kr2.eps}       
1560: \end{figure}
1561: \end{center}
1562: 
1563: We consider it difficult to assign reliable errors to the estimates of the
1564: \lecs{} in Tab.~\ref{tab:kr}. The determinations in
1565: \cite{MoussallamGauge,Anant.Mouss,Baur.Urech} -- from where we invoked the
1566: $\Kr{i}$ -- are model dependent, for which reliable estimates of uncertainties
1567: are always a delicate affair. Moreover, the scale dependence in various
1568: \lecs{} can be strongly correlated. We shall therefore refrain from assigning
1569: individual errors to the estimates in Tab.~\ref{tab:kr}. To be conservative,
1570: one might attribute an uncertainty of $1/(16\pi^2)\approx 6.3\cdot 10^{-3}$ to
1571: each \lec{} $\kr{i}$, stemming from general dimensional arguments. The size of
1572: this uncertainty compared to the values in Tab.~\ref{tab:kr} indicates that
1573: the entries of the table 
1574: are yet only a rough order of magnitude estimate.  
1575: 
1576: In the near future, a more precise determination of (some combinations of)
1577: electromagnetic \lecs{} may also be expected from lattice
1578: \qcd{}. In this respect we mention two recent studies that address the
1579: electromagnetic splitting of pseudoscalar meson masses \cite{RBC,JLQCD}.
1580: 
1581: 
1582: 
1583: % --- CONCLUSIONS ----------------------------------------------------------
1584: 
1585: \section{Conclusions}
1586: \label{sec:conclusions}
1587: 
1588: In summary we have worked out the strange quark mass dependence of the
1589: two--flavour electromagnetic \lecs{} $C,k_i$ at next--to--leading order. The
1590: calculation relied on a non--trivial matching between the three--flavour 
1591: and the two--flavour generating functional of \chpt{} 
1592: including virtual photons and amounts to 16 relations among the \lecs{} of
1593: \cpt{2} and \cpt{3}.\\  
1594: 
1595: \noindent
1596: These relations are useful to obtain constraints and further information on
1597: the pertinent \lecs{}. As an application we have used these relations to obtain
1598: numerical estimates for the values of the two--flavour electromagnetic
1599: \lecs{}.  
1600: 
1601: 
1602: % --- ACKNOWLDEDGMENTS -----------------------------------------------------
1603: 
1604: \subsection*{Acknowledgements}
1605: 
1606: We thank J\"urg Gasser for suggesting us to undertake this work, and are
1607: grateful for many enlightening discussions and comments on the
1608: manuscript. We further thank Marc Knecht, Bachir Moussallam and Hagop Sazdjian
1609: for useful discussions and/or comments on the manuscript. C.H. and M.A.I. wish
1610: to express their gratitude to the members of the Institute of Theoretical
1611: Physics, University of Bern, for their kind hospitality during the final stages of
1612: this work. This work was supported  by the Swiss National Science Foundation,
1613: by the Ministerio de Educaci\'on y Ciencia under the project FPA2004-00996, by
1614: Generalitat Valenciana GVACOMP2007-156, and by EU MRTN-CT-2006-035482
1615: (FLAVIA{\it net}).   
1616: 
1617: 
1618: 
1619: 
1620: 
1621: 
1622: % --- REFERENCES -----------------------------------------------------------
1623: 
1624: \begin{thebibliography}{99}
1625: \setlength{\baselineskip}{14pt}
1626: 
1627: \bibitem{Weinberg}
1628:   S.~Weinberg,
1629:   %``Phenomenological Lagrangians,''
1630:   Physica A {\bf 96} (1979) 327.
1631:   %%CITATION = PHYSA,A96,327;%%
1632: 
1633: \bibitem{GL:Ann}
1634:   J.~Gasser and H.~Leutwyler,
1635:   %``Chiral Perturbation Theory To One Loop,''
1636:   Annals Phys.\  {\bf 158} (1984) 142.
1637:   %%CITATION = APNYA,158,142;%%
1638: 
1639: \bibitem{GL:NPB}
1640:   J.~Gasser and H.~Leutwyler,
1641:   %``Chiral Perturbation Theory: Expansions In The Mass Of The Strange Quark,''
1642:   Nucl.\ Phys.\  B {\bf 250} (1985) 465.
1643:   %%CITATION = NUPHA,B250,465;%%
1644: 
1645: \bibitem{Ecker:lecs}
1646:   G.~Ecker,
1647:   %``Chiral low-energy constants,''
1648:   arXiv:hep-ph/0702263.
1649:   %%CITATION = HEP-PH/0702263;%%
1650: 
1651: \bibitem{Bijnens:lecs}
1652:   J.~Bijnens,
1653:   %``Quark Mass dependence at Two Loops for Meson Properties,''
1654:   arXiv:0708.1377 [hep-lat].
1655:   %%CITATION = ARXIV:0708.1377;%%
1656: 
1657: \bibitem{Urech}
1658:   R.~Urech,
1659:   %``Virtual photons in chiral perturbation theory,''
1660:   Nucl.\ Phys.\  B {\bf 433} (1995) 234
1661:   [hep-ph/9405341].
1662:   %%CITATION = NUPHA,B433,234;%%
1663: 
1664: \bibitem{Neufeld}
1665:   H.~Neufeld and H.~Rupertsberger,
1666:   %``The Electromagnetic Interaction In Chiral Perturbation Theory,''
1667:   Z.\ Phys.\  C {\bf 71} (1996) 131
1668:   [hep-ph/9506448];
1669:   %%CITATION = ZEPYA,C71,131;%%
1670:   H.~Neufeld and H.~Rupertsberger,
1671:   %``Isospin breaking in chiral perturbation theory and the decays eta $\to$ pi
1672:   %lepton neutrino and tau $\to$ eta pi neutrino,''
1673:   Z.\ Phys.\  C {\bf 68}, 91 (1995).
1674:   %%CITATION = ZEPYA,C68,91;%%
1675: 
1676: \bibitem{Knecht.Urech}
1677:   M.~Knecht and R.~Urech,
1678:   %``Virtual photons in low energy pi pi scattering,''
1679:   Nucl.\ Phys.\  B {\bf 519} (1998) 329
1680:   [hep-ph/9709348].
1681:   %%CITATION = NUPHA,B519,329;%%
1682: 
1683: \bibitem{Steininger}
1684:   U.~G.~Meissner, G.~Muller and S.~Steininger,
1685:   %``Virtual photons in SU(2) chiral perturbation theory and electromagnetic
1686:   %corrections to pi pi scattering,''
1687:   Phys.\ Lett.\  B {\bf 406} (1997) 154
1688:   [Erratum-ibid.\  B {\bf 407} (1997) 454]
1689:   [hep-ph/9704377].
1690:   %%CITATION = PHLTA,B406,154;%%
1691: 
1692: \bibitem{Anant.Mouss}
1693:   B.~Ananthanarayan and B.~Moussallam,
1694:   %``Four-point correlator constraints on electromagnetic chiral parameters
1695:   %and resonance effective Lagrangians,''
1696:   JHEP {\bf 0406} (2004) 047
1697:   [hep-ph/0405206].
1698:   %%CITATION = JHEPA,0406,047;%%
1699: 
1700: \bibitem{Baur.Urech}
1701:   R.~Baur and R.~Urech,
1702:   %``Resonance contributions to the electromagnetic low energy constants of
1703:   %chiral perturbation theory,''
1704:   Nucl.\ Phys.\  B {\bf 499}, 319 (1997)
1705:   [hep-ph/9612328].
1706:   %%CITATION = NUPHA,B499,319;%%
1707: 
1708: \bibitem{Bijnens.Prades}
1709:   J.~Bijnens and J.~Prades,
1710:   %``Electromagnetic corrections for pions and kaons: Masses and
1711:   %polarizabilities,''
1712:   Nucl.\ Phys.\  B {\bf 490}, 239 (1997)
1713:   [hep-ph/9610360];
1714:   %%CITATION = NUPHA,B490,239;%% 
1715: 
1716: \bibitem{MoussallamGauge}
1717:   B.~Moussallam,
1718:   %``A sum rule approach to the violation of Dashen's theorem,''
1719:   Nucl.\ Phys.\  B {\bf 504}, 381 (1997)
1720:   [hep-ph/9701400];
1721:   %%CITATION = NUPHA,B504,381;%%
1722: 
1723: \bibitem{K_i}
1724:   V.~E.~Lyubovitskij, T.~Gutsche, A.~Faessler and R.~Vinh Mau,
1725:   %``pi N scattering and electromagnetic corrections in the
1726:   % perturbative  chiral quark model,''
1727:   Phys.\ Lett.\  B {\bf 520}, 204 (2001)
1728:   [hep-ph/0108134];
1729:   %%CITATION = PHLTA,B520,204;%%
1730:   V.~E.~Lyubovitskij, T.~Gutsche, A.~Faessler and R.~Vinh Mau,
1731:   %``Electromagnetic couplings of the ChPT Lagrangian from the perturbative
1732:   %chiral quark model,''
1733:   Phys.\ Rev.\  C {\bf 65}, 025202 (2002)
1734:   [hep-ph/0109213].
1735:   %%CITATION = PHRVA,C65,025202;%%
1736: 
1737: \bibitem{Gasser:atom}
1738:   J.~Gasser, V.~E.~Lyubovitskij, A.~Rusetsky and A.~Gall,
1739:   %``Decays of the pi+ pi- atom,''
1740:   Phys.\ Rev.\  D {\bf 64} (2001) 016008
1741:   [hep-ph/0103157].
1742:   %%CITATION = PHRVA,D64,016008;%%
1743: 
1744: \bibitem{Sazdjian-I}
1745:   H.~Jallouli and H.~Sazdjian,
1746:   %``Relativistic effects in the pionium lifetime,''
1747:   Phys.\ Rev.\  D {\bf 58}, 014011 (1998)
1748:   [Erratum-ibid.\  D {\bf 58}, 099901 (1998)]
1749:   [hep-ph/9706450].
1750:   %%CITATION = PHRVA,D58,014011;%%
1751: 
1752: \bibitem{Nyffeler}
1753:   A.~Nyffeler and A.~Schenk,
1754:   %``Effective field theory of the linear O(N) sigma model,''
1755:   Annals Phys.\  {\bf 241} (1995) 301
1756:   [hep-ph/9409436].
1757:   %%CITATION = APNYA,241,301;%%
1758: 
1759: \bibitem{Schmid:PhD}
1760: M.~Schmid, {\it Strangeless $\chi PT$ at large $m_s$}, PhD thesis, 
1761: University of  Bern, January 2007.
1762: 
1763: \bibitem{matchingI}
1764:   J.~Gasser, C.~Haefeli, M.~A.~Ivanov and M.~Schmid,
1765:   %``Integrating out strange quarks in ChPT,''
1766:   Phys.\ Lett.\  B {\bf 652} (2007) 21
1767:   [arXiv:0706.0955 [hep-ph]].
1768:   %%CITATION = PHLTA,B652,21;%%
1769: 
1770: \bibitem{NehmePhD}
1771:   A.~Nehme, {\it ``La Brisure d'Isospin dans les Interactions M\'eson--M\'eson \`a
1772:     Basse Energie''}, PhD thesis, CPT Marseille, July 2002.
1773: 
1774: \bibitem{Mouss:Sigma}
1775:   B.~Moussallam,
1776:   %``Flavor stability of the chiral vacuum and scalar meson dynamics,''
1777:   JHEP {\bf 0008}, 005 (2000)
1778:   [hep-ph/0005245].
1779:   %%CITATION = JHEPA,0008,005;%%
1780: 
1781: \bibitem{Schweizer:B}
1782:   R.~Kaiser and J.~Schweizer,
1783:   %``The expansion by regions in pi K scattering,''
1784:   JHEP {\bf 0606}, 009 (2006)
1785:   [hep-ph/0603153].
1786:   %%CITATION = JHEPA,0606,009;%%
1787: 
1788: \bibitem{meissner_frink}
1789:   M.~Frink and U.~G.~Meissner,
1790:   %``Chiral extrapolations of baryon masses for unquenched three-flavor
1791:   %lattice simulations,''
1792:   JHEP {\bf 0407}, 028 (2004)
1793:   [hep-lat/0404018].
1794:   %%CITATION = JHEPA,0407,028;%%
1795: 
1796: \bibitem{WZW}
1797:   J.~Wess and B.~Zumino,
1798:   %``Consequences of anomalous Ward identities,''
1799:   Phys.\ Lett.\  B {\bf 37} (1971) 95;
1800:   %%CITATION = PHLTA,B37,95;%%
1801:   E.~Witten,
1802:   %``Global Aspects Of Current Algebra,''
1803:   Nucl.\ Phys.\  B {\bf 223} (1983) 422.
1804:   %%CITATION = NUPHA,B223,422;%%
1805: 
1806: \bibitem{Kaiser}
1807:   R.~Kaiser,
1808:   %``Anomalies and WZW-term of two-flavour QCD,''
1809:   Phys.~Rev.~{\bf D63}, 76010 (2001)
1810:   [hep-ph/0011377].
1811:   %%CITATION = HEP-PH/0011377;%%
1812: 
1813: \bibitem{MoussAnomaly}
1814:   B.~Ananthanarayan and B.~Moussallam,
1815:   %``Electromagnetic corrections in the anomaly sector,''
1816:   JHEP {\bf 0205} (2002) 052
1817:   [hep-ph/0205232].
1818:   %%CITATION = JHEPA,0205,052;%%
1819: 
1820: \bibitem{Schweizer}
1821:   J.~Schweizer,
1822:   %``Explicit generating functional for pions and virtual photons,''
1823:   JHEP {\bf 0302}, 007 (2003)
1824:   [hep-ph/0212188].
1825:   %%CITATION = JHEPA,0302,007;%%
1826: 
1827: \bibitem{matchingII}
1828:   J.~Gasser, C.~Haefeli, M.~A.~Ivanov and M.~Schmid,
1829:   in preparation.
1830: 
1831: \bibitem{GasserSainio}
1832:     J.~Gasser and M.~E.~Sainio,
1833:   %``Two-loop integrals in chiral perturbation theory,''
1834:   Eur.\ Phys.\ J.\  C {\bf 6} (1999) 297
1835:   [hep-ph/9803251].
1836:   %%CITATION = EPHJA,C6,297;%%
1837: 
1838: \bibitem{Sazdjian-II}
1839:   H.~Sazdjian,
1840:   %``The pionium lifetime in generalized chiral perturbation theory,''
1841:   Phys.\ Lett.\  B {\bf 490}, 203 (2000)
1842:   [hep-ph/0004226].
1843:   %%CITATION = PHLTA,B490,203;%%
1844: 
1845: \bibitem{BijnensGauge}
1846:   J.~Bijnens,
1847:   %``Violations Of Dashen's Theorem,''
1848:   Phys.\ Lett.\  B {\bf 306}, 343 (1993)
1849:   [hep-ph/9302217];
1850:   %%CITATION = PHLTA,B306,343;%%
1851: 
1852: \bibitem{Scimemi}
1853:   J.~Gasser, A.~Rusetsky and I.~Scimemi,
1854:   %``Electromagnetic corrections in hadronic processes,''
1855:   Eur.\ Phys.\ J.\  C {\bf 32} (2003) 97
1856:   [hep-ph/0305260].
1857:   %%CITATION = EPHJA,C32,97;%%
1858: 
1859: \bibitem{BijnensFit}
1860:   G.~Amoros, J.~Bijnens and P.~Talavera,
1861:   %``QCD isospin breaking in meson masses, decay constants and quark mass
1862:   %ratios,''
1863:   Nucl.\ Phys.\  B {\bf 602}, 87 (2001)
1864:   [hep-ph/0101127].
1865:   %%CITATION = NUPHA,B602,87;%%
1866: 
1867: \bibitem{RBC}
1868:   T.~Blum, T.~Doi, M.~Hayakawa, T.~Izubuchi and N.~Yamada,
1869:   %``Determination of light quark masses from the electromagnetic splitting of
1870:   %pseudoscalar meson masses computed with two flavors of domain wall
1871:   %fermions,''
1872:   arXiv:0708.0484 [hep-lat].
1873:   %%CITATION = ARXIV:0708.0484;%%
1874: 
1875: \bibitem{JLQCD}
1876:   E.~Shintani, H.~Fukaya, S.~Hashimoto, H.~Matsufuru, J.~Noaki, T.~Onogi and
1877:   N.~Yamada 
1878:   [JLQCD Collaboration],
1879:   %``Pion mass difference from vacuum polarization,''
1880:   arXiv:0710.0691 [hep-lat].
1881:   %%CITATION = NONE,,;%%
1882: 
1883: 
1884: \end{thebibliography}
1885: 
1886: 
1887: 
1888: 
1889: 
1890: \end{document}