hep-ph0410061/WWD.tex
1: \documentclass[12pt,a4paper]{article}
2: 
3: % You can change the spacing by this parameter.
4: \newcommand{\blst}{1.24}
5: \renewcommand{\baselinestretch}{\blst}
6: % For (10pt,11pt,12pt), 1.5- and double-spacings are obtained by
7: % (1.25, 1.21, 1.24) and (1.67, 1.62, 1.66), respectively.
8: 
9: \usepackage{amsmath}
10: \usepackage{amssymb}
11: \usepackage{a4wide}
12: \usepackage{epsfig}
13: 
14: %%% Just in case one wants to change the notations:
15: \newcommand{\y}{\ensuremath{y}}
16: \newcommand{\z}{\ensuremath{z}}
17: \newcommand{\ig}{\ensuremath{ig}}
18: 
19: %%% For shorthand:
20: \newcommand{\piR}{\ensuremath{\pi R}}
21: \newcommand{\wt}[1]{\ensuremath{\widetilde{#1}}}
22: \newcommand{\dL}[1]{\ensuremath{\overleftarrow{\partial_{#1}}}}
23: \newcommand{\tr}{\ensuremath{\,\text{tr}}}
24: \newcommand{\omst}{\ensuremath{\overline{\omega}}}
25: \newcommand{\Mhat}{\ensuremath{\hat{M}}}
26: \newcommand{\mhat}{\ensuremath{\hat{m}}}
27: \newcommand{\vhat}{\ensuremath{\hat{v}}}
28: \newcommand{\N}{\ensuremath{\mathcal{N}}}
29: \newcommand{\Pd}{\ensuremath{\mathcal{P}}}
30: \newcommand{\F}{\ensuremath{\mathcal{F}}}
31: \newcommand{\J}{\ensuremath{\mathcal{J}}}
32: \newcommand{\Y}{\ensuremath{\mathcal{Y}}}
33: 
34: 
35: \begin{document}
36: %%% Title page
37: \begin{titlepage}
38:   \begin{center}
39:     \hfill
40:     \begin{minipage}{6cm}
41:       \begin{flushright}
42:         TUM-HEP-560/04
43:       \end{flushright}
44:       \end{minipage}\\
45:     \bigskip
46:     \vspace{3\baselineskip}
47:   
48:     {\Large\bf Wilson Lines in Warped Space: Dynamical Symmetry Breaking 
49: and Restoration} \\
50:     \bigskip
51:     \bigskip
52:   
53:     {\bf  Kin-ya Oda} \\
54:     \smallskip
55:     {\small\sl  
56:       Physikalisches Institut der Universit\"at Bonn,
57:       D-53115 Bonn, Germany} \\
58:     \medskip
59:     {\bf Andreas Weiler} \\
60:     \smallskip
61:     {\small\sl
62:       Physik Department, Technische Universit\"at M\"unchen,
63:       D-85748 Garching, Germany} \\
64:     \bigskip
65:   
66:     {\tt odakin@th.physik.uni-bonn.de} \\
67:     {\tt Andreas.Weiler@ph.tum.de} \\
68:     \bigskip
69:     \vspace*{.5cm}
70:   
71:     {\bf Abstract} \\
72:     \end{center}
73:   \noindent
74:   The dynamics of Wilson lines integrated along a warped extra dimension has been unknown.
75:   We study a five dimensional $SU(N)$ pure gauge theory with Randall-Sundrum warped compactification on $S^1/Z_2$.
76:   We clarify the notion of large gauge transformations
77:   that are non-periodic on the covering space for this setup.
78:   We obtain Kaluza-Klein expansions of gauge and ghost fields
79:   for the most general twists and background gauge field configurations,
80:   which break the gauge symmetry at classical level in general.
81:   We calculate the one-loop effective potential and find that
82:   the symmetry corresponding to the subgroup allowing continuous Wilson lines is dynamically restored.
83:   The presented method can be directly applied to include extra fields.
84:   The connection to dynamical Scherk-Schwarz supersymmetry breaking in warped space is discussed.
85:   \bigskip
86:   \bigskip
87:   \end{titlepage}
88: 
89: 
90: 
91: 
92: 
93: %%%%
94: \section{Introduction}
95: Warped compactification not only provides
96: a beautiful explanation how the large hierarchy
97: $m_\text{weak}/M_\text{Planck}$ is generated
98: from the exponential profile of its metric~\cite{Randall:1999ee}
99: but also is itself a quite general consequence of string theory
100: due to the fact that D-branes 
101: generically provide sources for warping
102: (See~\cite{WarpinString} and references therein).
103: %
104: In the original Randall-Sundrum model~\cite{Randall:1999ee},
105: five dimensional spacetime is compactified
106: on the orbifold $S^{1}/Z_{2}$, along which the
107: normalization of the four dimensional metric is exponentially scaled.
108: %
109: Orbifold compactification is a powerful tool in string theory to get
110: three generations and especially to reduce the rank of the gauge group
111: when combined with continuous Wilson lines~\cite{OrbwWil}. 
112: %
113: The values of these Wilson lines and the resulting gauge symmetry breaking pattern
114: must be determined dynamically
115: via the Hosotani mechanism~\cite{BunchofHosotani} 
116: when supersymmetry breaking is taken into account.
117: %
118: Studies of Wilson line dynamics on simpler orbifolds have been started
119: recently~\cite{Hosotanietal,Haba:2003ux}. 
120: %
121: These are also applied to gauge-Higgs unification models~\cite{GHunif}.
122: 
123: 
124: The orbifold compactification on $S^{1}/Z_{2}$
125: with radius $R=M_\text{GUT}^{-1}$ provides
126: a simple mechanism to solve the doublet-triplet splitting problem
127: in $SU(5)$ grand unified theories
128: (GUT's)~\cite{BunchofKawamura}. 
129: %
130: Realistic models along this line have been proposed
131: to implement low energy supersymmetry (SUSY) that is
132: broken at the weak scale~\cite{OrbGUT}.  
133: %
134: Once one considers an orbifold GUT,
135: it is tempting to think that SUSY is also broken by the compactification
136: via the Scherk-Schwarz mechanism~\cite{Scherk:1978ta}.
137: %
138: However, in such an orbifold model
139: the Scherk-Schwarz parameter, the amount of $SU(2)_{R}$ twist,
140: must be put to an extremely small value of order $m_\text{weak}/M_\text{GUT}$
141: at the classical level~\cite{Barbieri:2001yz}.
142: %
143: Furthermore, quantum corrections lead
144: either to restored supersymmetry $m_\text{SUSY}=0$
145: or to its violation of the order of $m_\text{SUSY}\simeq R^{-1}$,
146: unless one introduces an extra source of supersymmetry breaking
147: at the orbifold fixed point~\cite{vonGersdorff:2002tj}. 
148: This is generally true for symmetry breaking via the Hosotani mechanism in flat space.
149: 
150: 
151: Therefore it is natural to generalize the above considerations to 
152: a gauge theory in the bulk of the Randall-Sundrum geometry
153: where one can make use of the exponential hierarchy in its metric.
154: %
155: Warped SUSY GUT's are constructed in Refs.~\cite{Pomarol:2000hp,Goldberger:2002pc}
156: with SUSY breaking assumed to be by boundary conditions 
157: or by an extra source at the orbifold fixed point, respectively.  
158: %
159: Recently, it has been proposed
160: in the framework of superconformal gravity
161: that supersymmetry breaking by boundary conditions can be consistently defined and
162: equivalent to Wilson line breaking,
163: % 
164: if the compensator multiplet has vanishing gauge coupling and
165: the warping is generated from the vacuum configuration of
166: the bulk scalars~\cite{Abe:2004ar}.\footnote{
167:     See also Ref.~\cite{WarpedSB} for discussions on
168:     the Scherk-Schwarz breaking in the usual warped setup.}
169: %
170: Eventually, the vacuum expectation value (vev) 
171: of the $SU(2)$ 
172: gauge field that is related to the Scherk-Schwarz twist
173: must be determined dynamically by quantum corrections to the effective potential.
174: %
175: To that end, it is
176: important to solve the dynamics of Wilson lines in
177: warped space, which has not been explored so far.\footnote{
178:     \renewcommand{\baselinestretch}{1}\footnotesize
179:     In Ref.~\cite{Contino:2003ve} a Wilson line in warped space is considered
180:     in the context of the AdS/Conformal Field Theory (CFT) correspondence,
181:     where the analysis is confined to a (potentially false) vacuum
182:     that corresponds to imposing both diagonal twists
183:     and vanishing background field configurations.
184:     See Ref.~\cite{KinyaAndi} for further discussions.}
185: % See Ref. [17] for a treatment of Wilson lines in warped space in the context
186: % of the AdS/Conformal Field Theory (CFT) correspondence, where the contribution
187: % from a bulk fermion to the two point function of $A_y$ is computed assuming both diagonal twists % and vanishing background fields.
188: % The authors find a negative contribution to the mass squared of A_5 and therefore that
189: % the symmetry could be broken.
190: % For the determination of the mass of A_5 it is however necessary
191: % to calculate the effective potential as it is done in this paper.
192: \renewcommand{\baselinestretch}{\blst}\normalsize
193: 
194: 
195: In this paper we study the dynamics of Wilson lines of $SU(N)$ pure gauge theory
196: in the bulk of Randall-Sundrum geometry.
197: %
198: In the course of this calculation,
199: we derive for the first time
200: Kaluza-Klein expansions for the most general twists and background gauge field
201: configurations and calculate the corresponding one-loop effective potential.
202: %
203: In Section~\ref{flat_sec}, we briefly review the Hosotani mechanism on the orbifold $S^{1}/Z_{2}$.
204: %
205: In Section~\ref{KK_sec}, we obtain the Kaluza-Klein (KK) expansions of
206: the five dimensional gauge and ghost fields
207: with most general twists
208: in the presence of a gauge field background.
209: %
210: In Section~\ref{eff_pot_sec}, we calculate the effective potential for
211: the extra dimensional component of the background gauge field.
212: %
213: In the last section we summarize and discuss our result.
214: 
215: 
216: 
217: %%%%
218: \section{\Large Wilson lines on flat $S^{1}/Z_{2}$}
219: \label{flat_sec}
220: We briefly review how twists and background gauge field configurations
221: are related by large gauge transformations.
222: We consider a five dimensional $SU(N)$ gauge theory
223: compactified on the orbifold $S^{1}/Z_{2}$,
224: which is obtained from the simply-connected space $R^{1}: -\infty<y<\infty$
225: by modding with $S^{1}$ and $Z_{2}$ identifications
226:   $y      \sim y+2\piR$ and
227:   $y      \sim -y$, where $R$ is the compactification radius.
228: %
229: Under these identifications, the gauge fields $A_{M}$ ($M=0,\dots,3;4$) are in general
230: twisted by global $SU(N)$ transformations
231: %
232: \begin{align}
233:   A_M(-y)      &= \pm P_0 A_M(y)      P_0^{-1}, \nonumber\\
234:   A_M(\piR+y)  &= \pm P_1 A_M(\piR-y) P_1^{-1}, &
235:   A_M(y+2\piR) &= UA_M(y)U^{-1},
236:   \end{align}
237: where
238:   the extra $\pm$~sign is positive for four dimensions and
239:     negative for the extra dimension.\footnote{
240:     In principle local identifications are possible
241:     but we assume them global for simplicity in this paper.
242:     }
243: (We use $\mu$ for $0,\dots,3$ and 
244: $y$ for both $x^{4}$ and index ``4'' such as $A_{\y}=A_{4}$.)
245: %
246: Note that the consistency conditions $U=P_1P_0$ and $P_0^2=P_1^2=1$ are imposed.
247: %
248: Starting from the most general twists 
249: we can always choose the following basis~\cite{Haba:2003ux}  
250: %
251: \begin{align}
252:   \setcounter{MaxMatrixCols}{20}
253:   \begin{array}{rclccrrrr}
254:       P_0 &=& \text{blockdiag}\,(\sigma_3,         &\ldots,&\sigma_3,         &I_r,&I_s, &-I_t,&-I_u),\\
255:       P_1 &=& \text{blockdiag}\,(\sigma_{\theta_1},&\ldots,&\sigma_{\theta_q},&I_r,&-I_s,&I_t, &-I_u),
256:     \end{array}
257:   \setcounter{MaxMatrixCols}{10}
258:   \label{bc_mat}
259:   \end{align}
260: where
261:   $I_{r}$ is $r\times r$ unit matrix,
262:   $\sigma_{a}$ ($a=1,2,3$) are Pauli matrices, and
263:   $\sigma_{\theta}
264:     =\sigma_{3}\cos{\theta}+\sigma_{1}\sin{\theta}
265:     =e^{-i\theta\sigma_{2}}\sigma_{3}
266:     =\sigma_{3}e^{i\theta\sigma_{2}}$.
267: ($2q+r+s+t+u=N$.)
268: The $A_{\y}^{a}$ either within a block of $\pm I$
269: or connecting different blocks
270: does not have a zero-mode, a mode having vanishing KK mass,
271: and the dynamics of the corresponding Wilson line is trivial~\cite{Hall:2001tn}.\footnote{
272:     \renewcommand{\baselinestretch}{1}\footnotesize
273:     Of course when, say, $r<u$, we can combine $I_r$ and a part of $-I_u$
274:     to form $r$ additional $\sigma_\theta$ blocks with $\theta=0$.
275:     }
276: \renewcommand{\baselinestretch}{\blst}\normalsize
277: Therefore we can concentrate
278: on a $SU(2)$ subblock with twists
279:   $P_{0} = \sigma_{3}$ and
280:   $P_{1} = \sigma_{\theta}$ without loss of generality.
281: In general, only $A_{\y}^{(2)}$ ($A_{M}=A_{M}^{(a)}\sigma_{a}/2$)
282: has even $Z_{2}$ parities, hence a zero mode background:
283:   $gA^{(2)}_{\y}{}^{c} \equiv v$.
284: The KK expansions 
285: are given by
286: \begin{align}
287:   A^{(2)}_{\mu}(x,y) &=   \sum_{n=1}^{\infty}A^{(2)}_{\mu n}(x){\sin{ny\over R}\over\sqrt{\piR}}, &
288:   A^{(2)}_{\y}(x,y)  &= A^{(2)}_{\y0}(x){1\over\sqrt{2\piR}}
289:                          +\sum_{n=1}^{\infty}A^{(2)}_{\y n}(x) {\cos{ny\over R}\over\sqrt{\piR}}, \\
290:   A^{\pm}_{\mu}(x,y) &=      \sum_{n=-\infty}^{\infty}A_{\mu n}(x){e^{\pm i(m_{n}+v)y}\over N_{n}}, &
291:   A^{\pm}_{\y}(x,y)  &= \pm i\sum_{n=-\infty}^{\infty}A_{\y n}(x) {e^{\pm i(m_{n}+v)y}\over N_{n}},
292:                         \label{flat_Apm}
293:   \end{align}
294: where
295:   $A^{\pm}_{M} = {A^{(3)}_{M}\pm iA^{(1)}_{M}\over\sqrt{2}}$,  
296:   $A_{Mn}(x)$ are real fields,
297:   $m_{n}       = {n\over R}+{\theta\over\piR}-v$ are KK masses, and
298:   $N_{n}       = \sqrt{2\piR}\,m_{n}$ are normalization constants.
299: (Here we present a different form from literature in Eq.~\eqref{flat_Apm}
300: to make the orthogonality of wave functions transparent.)
301:     
302: 
303:   
304: Consider the following large gauge transformation that is non-periodic on the covering space $R^{1}$:
305: \begin{align}
306:   \ig A_{M}(y) &\rightarrow \ig\wt{A}_{M}(y) = \left[\Omega\left(\ig A_{M}-\dL{M}\right)\Omega^{-1}\right](y) &
307:   \text{with}\quad
308:   \Omega(y)    &=           \exp\left[i\varphi y{\sigma_{2}\over 2}\right],
309:   \label{large_gtf}
310:   \end{align}
311: which results in
312:   $g\wt{A}_{\y}^{(2)}{}^{c} = v-{\varphi/\piR}$.
313: We find that the shift of the background is canceled by the transformation of
314: $A^{\pm}_{M}\rightarrow \wt{A}^{\pm}_{M}=e^{\mp i\varphi y/\piR}A^{\pm}_{M}$
315: leaving its KK masses invariant: $\wt{m}_{n}=m_{n}$.
316: Now the new fields are twisted by matrices
317:   $\wt{P}_{0} = \sigma_{3}$ and
318:   $\wt{P}_{1} = \sigma_{\theta-\varphi} \equiv \sigma_{\tilde{\theta}}$.
319: Above, the two sets of twists $P_{i}$ and $\wt{P}_{i}$ with the corresponding backgrounds are
320: equivalent under the large gauge transformation,
321: by which one may e.g.\ diagonalize the twist $\tilde{\theta}=0$
322: or remove the background $\wt{A}_{\y}^{(2)}{}^{c}=0$, but not both.
323: %
324: If we choose to take the former (or latter) gauge,
325: different values of $\wt{A}_{\y}^{(2)}{}^{c}$ (or $\tilde{\theta}$) 
326: correspond to physically different vacua
327: which are degenerate at the classical level.
328: Quantum corrections determine whether the symmetry is dynamically
329: broken or restored, depending on 
330: the matter content~\cite{Hosotanietal,Haba:2003ux}. 
331: 
332: 
333: %%%%
334: \section{Kaluza-Klein expansions}
335: \label{KK_sec}
336: We consider a $SU(N)$ gauge theory in the bulk of
337: the Randall-Sundrum geometry~\cite{Randall:1999ee},
338: which is a five dimensional Anti de Sitter (AdS) space compactified on $S^{1}/Z_{2}$
339: with the metric:
340: \begin{align}
341:   G_{MN}dx^Mdx^N = e^{-2\sigma(y)}\eta_{\mu\nu}dx^\mu dx^\nu+dy^2,
342:   \end{align}
343: where
344:   $\eta_{\mu\nu}$ is the Lorentzian metric and
345:   $\sigma(y)$ is defined by $\sigma(y)=k|y|$ at $-\piR<y\leq\piR$,
346:   with $k$ being the inverse AdS curvature radius.
347: Elsewhere on the covering space $R^{1}$,
348: we define $\sigma$ by the periodicity condition $\sigma(y+2\piR)=\sigma(y)$.
349: For later use we also define\footnote{
350:     \renewcommand{\baselinestretch}{1}\footnotesize
351:     When there arises an ambiguity at the orbifold fixed point, say, around $y=0$,
352:     we can use the regularized form $\sigma(y)=k\delta\log\cosh(y/\delta)$
353:     with an infinitesimal $\delta=+0$ to check the expression.
354:     For our purpose we can use
355:       $\epsilon(y)     = \theta(y)-\theta(-y)$, 
356:       $\epsilon'(y)    = 2\left[\delta(y)-\delta(y-\piR)\right]$ and
357:       $\epsilon(y)^{2} = 1$
358:     at $-\piR<y\leq\piR$.
359:     \label{sig_ep}
360:     }
361: \renewcommand{\baselinestretch}{\blst}\normalsize
362: $\epsilon(y)=\sigma'(y)/k$ and $z(y)=e^{\sigma(y)}$.
363: %
364: We call the orbifold fixed points at $y=0$ and $y=\piR$ ultraviolet (UV)
365: and infrared (IR) branes, respectively, and write $\z_{0}=\z(0)=1$ and $\z_{1}=\z(\piR)=e^{\pi kR}$.
366: %
367: In this paper we assume that the radion is already stabilized e.g.\  by
368: the Goldberger-Wise mechanism~\cite{Goldberger:1999uk}.
369: 
370: We employ the background field method,
371: separating the gauge field into classical and quantum parts
372: $A_{M} = A_{M}^{c}+A_{M}'$,
373: and take the following gauge fixing\footnote{
374:     The choice $t=4$ and $\xi=1$ makes
375:     $f$ manifestly invariant under five dimensional diffeomorphisms.
376:     }
377: \begin{align}
378:   S_{f} &= -{1\over\xi}\int d^{4}x\int_{-\piR}^{\piR}dy\sqrt{-G}\tr\left[ff\right], &
379:   \text{with}\quad
380:   f     &= \z^{2}\eta^{\mu\nu}D_{\mu}^{c}A_{\nu}'
381:            +\xi \z^{t}D_{\y}^{c}\z^{-t}A_{\y}',
382:   \end{align}
383: % 
384: where
385:   $D_{M}$ is the gauge covariant derivative and
386:   the superscript $c$ indicates that the gauge field is replaced by its classical part,
387:   e.g.\ $D_{M}^{c}A_{N}'=\partial_{M}A_{N}'+\ig[A_{M}^{c},A_{N}']$.
388: We consider the pure gauge background $F_{MN}^{c}=0$,
389: being a classical potential minimum,
390: and assume $A_{\mu}^{c}=0$
391: since it can be gauged away towards spatial infinity.
392: %
393: When we choose $\xi=1$ and $t=2$, the quadratic terms for
394: the gauge and ghost fields simplify: 
395: \begin{align}\label{quadraticaction}
396:   S &= \int d^{4}x\int_{-\piR}^{\piR}dy\tr\left[
397:           \eta^{\mu\nu}A_\mu'(\Box+\Pd_{4})A_\nu'
398:          +A_\y'\z^{-2}(\Box+\Pd_{\y})A_\y'
399:          +2\z^{-2}\omst'(\Box+\Pd_{4})\omega'
400:          \right],
401:   \end{align}
402: where
403:   $\Pd_{4}=D_\y^c\z^{-2}D_\y^c$,
404:   $\Pd_{\y}=D_\y^cD_{\y}^{c}\z^{-2}$, and
405:   we have put $\omega^{c}=\omst^{c}=0$.
406: (The surface terms vanish consistently for the given boundary conditions below.)
407: 
408: 
409: Following the same argument as in the flat case,
410: we concentrate on a $SU(2)$ subblock with twists
411:   $P_{0} = \sigma_{3}$ and
412:   $P_{1} = \sigma_{\theta}$, without loss of generality.
413: Again a zero mode resides only in $A_{\y}^{(2)}$ and we can write
414:   $gA_{\y}^{(2)}{}^{c}(y) = v\z^{2}$.
415: (The derivation of the form of zero mode $\z^{2}$ is given below.)
416: %
417: To obtain the KK expansions, we follow the strategy of Ref.~\cite{Gherghetta:2000qt}.
418: %
419: First, we solve the bulk KK equations at $0<y<\piR$ in terms of $\z$ neglecting all the boundary effects.
420: Second, we put the boundary conditions at $\z=\z_{0}$ and $\z_{1}$
421: on the obtained ``downstairs'' solution
422: to make it consistent with the $Z_{2}$ twists
423: so that the ``upstairs'' field on the covering space
424: is well-defined, i.e. continuous everywhere.
425: 
426: 
427: 
428: Let us start with $A_{\mu}^{(2)}$ and $A_{\y}^{(2)}$ which have definite odd and even $Z_{2}$ parities.
429: We obtain the following KK expansions
430: \begin{align}
431:   A_\mu^{(2)}(x,y)
432:     &= \sum_{n=1}^\infty A_{\mu n}^{(2)}(x)\epsilon
433: {\z[J_1(\Mhat_n\z)+B_{n}Y_1(\Mhat_n\z)]\over N_n}, &
434:   A_\y^{(2)}(x,y)
435:     &= \sum_{n=0}^\infty A_{\y n}^{(2)}(x)
436:  {f_{n}(\z)\over \N_n},
437:   \end{align}
438: where
439:   $N_n$, $\N_{n}$ are normalization constants,
440:   $B_{n} = -{J_1(\Mhat_n\z_0)\over Y_1(\Mhat_n\z_0)}= -{J_1(\Mhat_n\z_1)\over Y_1(\Mhat_n\z_1)}$, and
441:   the downstairs KK wave functions for vectoscalar
442:     $f_{n}(\z)=\z^{2}[J_{0}(\Mhat_{n}\z)+B_{n}Y_{0}(\Mhat_{n}\z)]$
443:     are defined for later use.
444: %
445: The KK masses $M_{n}=k\Mhat_{n}$ are determined by zeros of the KK mass function:
446: $J_{1}(\Mhat\z_{1})Y_{1}(\Mhat\z_{0})-Y_{1}(\Mhat\z_{1})J_{1}(\Mhat\z_{0})$,
447: which we find is exactly the same for both $A_{\mu}^{(2)}$ and $A_{\y}^{(2)}$.
448: 
449: 
450: 
451: In order to get the KK expansions of $A_{M}^{\pm}$, it is convenient to perform
452: the large (background) gauge transformation~\eqref{large_gtf},
453: which again results in the new twists
454:   $\wt{P}_{0} = \sigma_{3}$,
455:   $\wt{P}_{1} = \sigma_{\theta-\varphi} \equiv \sigma_{\tilde{\theta}}$
456: and the background
457: \begin{align}
458:   g\wt{A}_{\y}^{(2)}{}^{c}
459:     &= v\z^{2}-{\varphi\over\piR}
460:      = \left(v-{2ka^{2}\varphi\over 1- a^{2}}\right)\z^{2}
461:        -\frac{\varphi}{\pi R}\sum_{n=1}^{\infty}{(1,f_{n})\over\N_{n}^{2}}f_{n}(\z),
462:        \label{with_higher_modes}
463:   \end{align}
464: where
465:   $a=\z_{0}/\z_{1}=e^{-k\piR}\ll 1$ and
466:   $(1,f_{n})=\int_{\z_{0}}^{\z_{1}}2{d\z\over k\z}\z^{-2}f_{n}(\z)$.
467: Unlike in the flat case, all the higher KK modes are generated, as can be seen in the last step of
468: Eq.~\eqref{with_higher_modes}.
469: %
470: However, these higher modes all vanish
471: when integrated along the extra dimension in the Wilson line 
472: due to the orthogonality conditions of KK wave functions
473: and one would expect that these modes can be gauged away.
474: 
475: 
476: 
477: To see this, consider the following background gauge transformation
478:   $\check\Omega(y) = \exp\left[i\F(y){\sigma_{2}\over 2}\right]$,
479: which we require to be normal in the sense that $\F(y)$ is continuous everywhere and periodic $\F(y+2\piR)=\F(y)$.
480: When $\F(y)$ is odd $\F(-y)=-\F(y)$, twists $P_{i}$ are left
481: invariant under this transformation, while
482: gauge fields transform as
483:   $gA_{\y}^{(2)} \rightarrow g\wt{A}^{(2)}_\y  = gA^{(2)}_\y-\F'(y)$.
484: Again, this shift is canceled by the transformation of $A^{\pm}_{M}$ in its KK mass.
485: Let us take
486: \begin{align}
487:   \F(y)    &= k\sum_{n=1}^{\infty}\varphi_{n}F_{n}(y), &
488:   \text{with}\quad
489:   F_{n}(y) &= \int_{0}^{y}dy' f_{n}\left(z(y')\right),
490:   \end{align}
491: where the summation is over all the non-zero modes.
492: %
493: By definition, $F_{n}(y)$ is odd and its derivative is $f_{n}(y)$.
494: Due to the downstairs boundary conditions, we find that $F_{n}(y)$ vanishes at both boundaries,
495: i.e., the transformation $\check\Omega(y)$ is continuous everywhere on the covering space
496: as well as periodic, as promised.
497: To summarize, all the non-zero mode can always be removed
498: by taking $\varphi_{n}$ appropriately,
499: without changing the twists $P_{i}$.
500: 
501: 
502: 
503: Now we choose $\varphi=\theta$ to diagonalize the twist $\tilde{\theta}=0$
504: and gauge away all the resulting non-zero mode background.
505: We then have definite $Z_{2}$ parity for all $\wt{A}_{M}^{(a)}$.
506: Hereafter we omit the tilde for notational simplicity.
507: The KK expansions are obtained as\footnote{
508:     To derive Eq.~\eqref{KK_Apm}
509:     we have used $\epsilon^{2}=1$ in $A_{\y}^{(1)}$,
510:     which can be justified similarly as above.
511:     }
512: \begin{align}
513:   A^{\pm}_{\mu}(x,y)
514:     &= \sum_{n=0}^{\infty}A_{\mu n}(x)
515:        E^{\pm i\epsilon{\vhat\z^{2}/ 2}}{\chi_{1,n}^{\pm}(\z)\over N_{n}}, &
516:   A^{\pm}_{\y}(x,y)
517:     &= \sum_{n=0}^{\infty}A_{\y n}(x)
518:        \epsilon E^{\pm i\epsilon{\vhat\z^{2}/ 2}}{\chi_{0,n}^{\pm}(\z)\over\N_{n}},
519:        \label{KK_Apm}
520:   \end{align}
521: %
522: where
523:   $A_{Mn}(x)$ are real fields,
524:   $\vhat={v/k}-{2\theta a^{2}/(1-a^{2})}$ is the dimensionless vev,
525:   $E^{\pm i\epsilon w}\equiv\cos w \pm i\epsilon\sin w$,
526:   $N_{n}$ and $\N_{n}$ are normalization constants, and
527:   $\chi_{\nu,n}^{\pm}(\z)=\z^{2-\nu}[\alpha^{\pm}_{n}J_{\nu}(\mhat_{n}\z)+\beta^{\pm}_{n}Y_{\nu}(\mhat_{n}\z)]$
528:   are the downstairs KK wave functions
529:   with $\mhat_n \equiv m_n/k$ and $\alpha_n^\pm,\beta_n^\pm$ being dimensionless KK masses and complex constants, respectively.
530:   (Note $\alpha^{\pm}_{n}={\alpha^{(3)}_{n}\pm i\alpha^{(1)}_{n}\over\sqrt{2}}$ etc.)
531: %  
532: The boundary conditions on the downstairs fields can be summarized as
533: \begin{align}
534:   M(\mhat_{n})\vec{V}=0,
535:   \end{align}
536: with
537: $\vec{V}=(\alpha^{(3)}_{n}\,\beta^{(3)}_{n}\,\alpha^{(1)}_{n}\,\beta^{(1)}_{n})^{T}$
538: for $A_{\mu}$,
539: $\vec{V}=(\alpha^{(1)}_{n}\,\beta^{(1)}_{n}\,\alpha^{(3)}_{n}\,\beta^{(3)}_{n})^{T}$
540: for $A_{\y}$, and
541: \begin{align*}
542:   \lefteqn{M(\mhat_{n})=}\nonumber\\
543:   &
544:       \begin{pmatrix}
545:         \J_{C}(1)        & \Y_{C}(1)        & \mp \J_{S}(1) & \mp \Y_{S}(1) \\
546:         \J_{C}(0)        & \Y_{C}(0)        & \mp \J_{S}(0) & \mp \Y_{S}(0) \\
547:         \pm\sin\frac{\vhat\z_{1}^{2}}{2} \, J_{\nu}(\mhat_{n}\z_{1})  & \pm\sin\frac{\vhat\z_{1}^{2}}{2}\, Y_{\nu}(\mhat_{n}\z_{1})  &
548:            \cos\frac{\vhat\z_{1}^{2}}{2}\, J_{\nu}(\mhat_{n}\z_{1})  &    \cos\frac{\vhat\z_{1}^{2}}{2}\, Y_{\nu}(\mhat_{n}\z_{1}) 
549:         \\
550:         \pm\sin\frac{\vhat\z_{0}^{2}}{2}\, J_{\nu}(\mhat_{n}\z_{0})  & \pm\sin\frac{\vhat\z_{0}^{2}}{2}\, Y_{\nu}(\mhat_{n}\z_{0})  &
551:            \cos\frac{\vhat\z_{0}^{2}}{2}\, J_{\nu}(\mhat_{n}\z_{0})  &    \cos\frac{\vhat\z_{0}^{2}}{2}\, Y_{\nu}(\mhat_{n}\z_{0}) 
552:         \end{pmatrix},
553:   \end{align*}
554: where
555:   upper and lower signs as well as $\nu=1$ and 0 are for $A_{\mu}$ and $A_{\y}$, respectively,
556:     $\begin{pmatrix}\J_{C}(i)\\ \J_{S}(i)\end{pmatrix}
557:         = \begin{pmatrix} \cos\frac{\vhat\z_{i}^{2}}{2} & -\sin\frac{\vhat\z_{i}^{2}}{2}\\
558:                           \sin\frac{\vhat\z_{i}^{2}}{2} &  \cos\frac{\vhat\z_{i}^{2}}{2}\end{pmatrix}
559:           \begin{pmatrix} \nu J_{\nu}(\mhat\z_{i})+\mhat_{n}\z_{i}J'_{\nu}(\mhat\z_{i}) \\
560:                           \vhat\z_{i}^{2} J_{\nu}(\mhat\z_{i}) \end{pmatrix}$,
561:   and $\Y_{C}$ and $\Y_{S}$ are defined similarly to $\J_{C}$ and $\J_{S}$ with $J$ replaced by $Y$.
562:   ($J_{\nu}$ and $Y_{\nu}$ are Bessel functions of order $\nu$.)
563: The KK mass function for $A^{\pm}$ is obtained from the determinant of the boundary condition matrix:
564: \begin{align}\label{massfunction}
565:   N(\mhat)  &= \det M(\mhat),
566:   \end{align}
567: and $N(\mhat) = 0$ determines the KK masses.
568: We find that the $N$'s are exactly the same for $A_{\mu}$ and $A_{\y}$
569: and that its dependence on $\vhat$ is only through the term $\frac{2}{\pi^2}\cos[\vhat(\z_{1}^{2}-\z_{0}^{2})]$.
570: %
571: The KK expansions for the ghost field is obtained similarly to
572: $A_{\mu}$.  (Recall that the antighost field $\omst$ is not
573: necessarily the complex conjugate of the ghost field $\omega$.)
574: 
575: In Fig.~\ref{fig_nfunc} we plot $N(ax)$ as a function of $x=\mhat/a$
576: within half a period of $\vhat$, i.e.\ for $\vhat(\z_{1}^{2}-\z_{0}^{2})=0,\pi/2,\pi$ from above to below
577: when $a=10^{-15}$.
578: \begin{figure}
579:   \begin{center}
580:     \leavevmode
581:     \epsfig{file=nfunc.eps,width=10cm}
582:     \caption{KK mass function $N(ax)$ vs $x$
583:              with $a=10^{-15}$ for ${\vhat/ a^{2}}=0,{\pi\over 2},\pi$.
584:              The zeros of $N$ correspond to the KK masses.
585:              Note that a massless mode appears only for $\vhat=0$ mod $2\pi$.}
586:     \label{fig_nfunc}
587:     \end{center}
588:   \end{figure}
589: The points where the curve crosses the $x$-axis give the values of the corresponding KK masses,
590: with even and odd modes appearing in alternating order.
591: We can see that there appears an extra massless mode for $\vhat=0$, as expected, and
592: that the dependence on the variation of $\vhat$ is strongest for this
593: would-be zero mode,
594: whose mass can
595: be easily determined from Eq.~\eqref{massfunction} to be
596: \begin{align}
597:   m_0 &= ka\sqrt{1-\cos(\vhat/a^2) \over k \pi R},
598:   \end{align}
599: in the approximation $a\ll 1$. For the maximal breaking with $\cos(\vhat/a^2) = -1$,
600:     we find $m_0\approx 0.24\,ka$ for $a=10^{-15}$.
601: 
602: 
603: %%%%
604: \section{One loop effective potential}
605: \label{eff_pot_sec}
606: Following convention,
607: we perform the dimensional reduction
608: in the coordinate frame where the warp factor is unity at the UV brane.\footnote{
609:     The physical effective potential from the point of view of the IR brane
610:     will be enhanced by $a^{-4}$~\cite{Garriga:2000jb}.
611:     }
612: The contribution of a pair of $Z_{2}$ even and odd gauge fields $A^{(3)}_{M}$ and $A^{(1)}_{M}$ is
613: \begin{align}
614:   V_\text{eff}
615:     &= {\mu^{4-d}\over 2}\int{d^dp\over(2\pi)^d}\sum_{n=0}^{\infty}\log(p^2+m_n^2)
616:      = -{1\over 2}{(ka)^4\over(4\pi)^2}\left(4\pi\mu^2\over k^2a^2\right)^{\varepsilon/2}
617:        \Gamma\left(-2+{\varepsilon\over 2}\right)\sum_{n=0}^{\infty} x_n^{4-\varepsilon},
618:   \end{align}
619: per degree of freedom, where
620:   $d=4-\varepsilon$ is the number of dimensions with $\varepsilon$ being infinitesimal,
621:   $\mu$ is an arbitrary scale, and
622:   $x_n=m_n/ka$ is the dimensionless KK mass.
623: The infinite sum over KK masses can be evaluated utilizing
624: zeta function regularization techniques~\cite{Garriga:2000jb,Toms:2000bh,Goldberger:2000dv}
625: \begin{align}
626:   v_\text{eff}(\vhat)
627:     &\equiv -\Gamma\left(-2+{\varepsilon\over 2}\right)\sum_{n=0}^{\infty} x_n^{4-\varepsilon}
628:      =      -\Gamma\left(-2+{\varepsilon\over 2}\right)
629:              \int_C{dx\over 2\pi i}x^{4-\varepsilon}{N'(ax)\over N(ax)},
630:   \end{align}
631: where $C$ is a contour encircling all the poles on the positive real
632: axis counter-clockwise.
633: Note that these are the only poles in the right half plane
634: since there is a one-to-one correspondence between the zeros of the
635: KK mass function~\eqref{massfunction} and the eigenvalues of
636: the operators $\mathcal{P}_4$ and $\mathcal{P}_\y$ in Eq.~\eqref{quadraticaction}
637: which are Hermitian with respect to our boundary conditions.
638: 
639: 
640: After a few manipulations, we find
641: \begin{align}
642:   v_\text{eff}(\vhat)
643:     &= I_\text{IR}+{I_\text{UV}\over a^{4-\varepsilon}}
644:        +2\int_{0}^{\infty}dx\,x^{3-\varepsilon}\log\Bigg[\nonumber\\
645:     &\quad\mbox{}
646:          1-{1\over 2}\left({K_{0}(x)I_{0}(ax)\over I_{0}(x)K_{0}(ax)}
647:                           +{K_{1}(x)I_{1}(ax)\over I_{1}(x)K_{1}(ax)}
648:                           -{K_{0}(x)I_{1}(ax)\over I_{0}(x)K_{1}(ax)}
649:                           -{K_{1}(x)I_{0}(ax)\over I_{1}(x)K_{0}(ax)}\right)\nonumber\\
650:     &\quad\mbox{}
651:          +{K_{0}(x)K_{1}(x)I_{0}(ax)I_{1}(ax)\over I_{0}(x)I_{1}(x)K_{0}(ax)K_{1}(ax)}
652:          -{\cos\left({\vhat\over a^{2}}(1-a^{2})\right)\over 2ax^{2}I_{0}(x)I_{1}(x)K_{0}(ax)K_{1}(ax)}\Bigg],
653:          \nonumber\\
654:     &\simeq I_\text{IR}+{I_\text{UV}\over a^{4-\varepsilon}}
655:        +2\int_{0}^{\infty}dx\,x^{3-\varepsilon}\log\bigg[
656:          1-{I_{0}(x)K_{1}(x)-K_{0}(x)I_{1}(x)-{1\over x}\cos{\vhat\over a^{2}}  \over
657:            2I_{0}(x)I_{1}(x)\left(\gamma+\log{ax\over 2}\right)}\bigg],
658:   \label{eq_Veff_vec}
659:   \end{align}
660: where divergent integrals $I_\text{IR}$ and $I_\text{UV}$ are independent of $v$ and $a$
661: and can be absorbed in the renormalization of the IR- and UV-brane tensions, respectively,
662: as in Ref.~\cite{Goldberger:2000dv}.
663: ($I_{\nu}$ and $K_{\nu}$ are the modified Bessel functions.)
664: We find that the effective potential is a periodic function of $\vhat$ with the period $2\pi a^{2}/(1-a^{2})$
665: as is expected from the shape of the KK mass function.
666: In the last line of Eq.~\eqref{eq_Veff_vec}, the small $a$ limit is taken,
667: assuming that
668: $\vhat$ is within the first period, i.e.\ $\vhat/a^{2}=O(1)$,
669: without loss of generality.\footnote{
670:     \renewcommand{\baselinestretch}{1}\footnotesize
671:     For large $x$, the integrand goes to zero and a small $a$ expansion can be performed
672:     with converging coefficients.}
673: \renewcommand{\baselinestretch}{\blst}\normalsize
674: (One might find it suggestive that the scale of the period of $v$ is of the order of $ka^{2}$,
675: which roughly corresponds to the order of the observed value of the cosmological constant $\simeq \text{meV}$.)
676: 
677: 
678: In Fig.~\ref{fig_eff_pot}, we plot $v_\text{eff}$ as a function of $\vhat/a^{2}$ when $a=10^{-15}$.
679: \begin{figure}
680:   \begin{center}
681:     \leavevmode
682:     \epsfig{file=eff_pot.eps,width=10cm}
683:     \caption{Normalized effective potential $v_\text{eff}$ vs $\vhat/a^{2}$
684:              with $a=10^{-15}$.}
685:     \label{fig_eff_pot}
686:     \end{center}
687:   \end{figure}
688: Contribution from the ghost loop is equal to Eq.~\eqref{eq_Veff_vec}
689: multiplied by $-2$.
690: The final result including gauge and ghost field contributions is therefore:
691: \begin{align}
692:   V_\text{eff} &= {3\over 32\pi^{2}}(ka)^{4}v_\text{eff}.
693:   \end{align}
694:   Note that if we include extra adjoint bulk fermions,
695:   they would contribute with opposite signs to Eq.~\eqref{eq_Veff_vec}
696:   and that if we add more than required to make the theory supersymmetric,
697:   the potential of Fig.~\ref{fig_eff_pot} would be flipped upside down,
698:   realizing a dynamical symmetry breaking vacuum
699:   which corresponds to the maximal twist $\theta=\pi/2$ in the $A_{\y}^{c}=0$ gauge.
700: This vacuum breaks $SU(2)$ completely and hence provides a rank reduction of the gauge symmetry.
701: We find that the symmetry breaking scale is of the order of $ka\simeq \text{TeV}$ for this case.
702: 
703: %%%%
704: \section{Summary and discussions}
705: We have studied the $SU(N)$ pure gauge theory in the bulk of the Randall-Sundrum geometry
706: and have obtained Kaluza-Klein expansions of gauge and ghost fields 
707: under the presence of the gauge field background with most general twists $P_{i}$.
708: %
709: We find that four dimensional gauge, vectoscalar and ghost fields have exactly the same
710: KK masses.
711: %
712: During the course of this calculation we have clarified the notion of a large gauge transformation
713: that is non-periodic on the covering space and
714: how it is consistently realized in the warped background.
715: %
716: The effective potential for the background $A_{\y}^{c}$ is obtained.
717: We find that a gauge symmetry corresponding to a continuous Wilson line,
718: i.e.\ a $SU(2)$ subgroup of Eq.~\eqref{bc_mat},
719: which is completely broken for finite $\theta$ at the classical level,
720: is dynamically restored to $U(1)$.
721: 
722: It is straightforward to apply our method to include other fields
723: with or without extra boundary masses and especially to
724: supersymmetrize our setup, where the symmetry breaking scales due to
725: continuous Wilson lines will be of the order of $k a\simeq\text{TeV}$
726: according to the analysis presented here.
727: %
728: Since we find that the dynamics of Wilson lines in warped space can be controlled
729: in quite a parallel manner to that in flat space
730: if one treats the large gauge transformation carefully,
731: we expect that 
732: the Hosotani mechanism will lead to a dynamical supersymmetry breaking
733: of the order of weak scale in the setup discussed in the introduction.
734: %
735: It is also interesting to pursue the AdS/CFT correspondence
736: generalizing the analysis of Ref.~\cite{Contino:2003ve}
737: to our setup,
738: as the Wilson line on the AdS side
739: will correspond to a quantity that is integrated
740: all the way from UV to IR on the CFT side.
741: %
742: The techniques developed here can also be applied to the gauge-Higgs unification models in warped space.
743: %
744: These points will be presented in separate publications~\cite{KinyaAndi}.
745: 
746: 
747: 
748: \bigskip
749: \noindent{\bf Acknowledgments:}
750: We wish to acknowledge the helpful communication
751: with H.\ Abe, A.\ J.\ Buras, M.\ Drees, D.\ Ida, T.\ Kobayashi, A.\ Pomarol, 
752: A.\ Poschenrieder, M.\ Quiros, Y.\ Sakamura and A.\ Wingerter.
753: K.O.\ thanks N.\ Haba, Y.\ Hosotani and S.\ Komine for useful discussions.
754: The work of K.O.\ is partly supported by the SFB375 of the Deutsche
755: Forschungs\-gemein\-schaft and A.W. is partly supported by the German
756: Bundesministerium f\" ur Bildung und Forschung under the contract
757: 05HT4WOA/3 and the DFG Project Bu.\ 706/1-2.
758: 
759: 
760: \renewcommand{\baselinestretch}{1}\normalsize
761: 
762: \input{WWD.bbl}
763: %Above WWD.bbl file is made from the following command with paper.bib file.
764: %\bibliography{paper}
765: %\bibliographystyle{utphys}
766: 
767: 
768: \end{document}
769: