1: \documentclass[12pt]{article}

2: \textwidth 6.5in

3: \oddsidemargin 0in

4: \evensidemargin 0in

5: \textheight 8.6in

6: \topmargin -0.5in

7: 

8: %\documentclass[prd,11pt,nofootinbib,superscriptaddress]{revtex4}

9: 

10: \usepackage{epsfig}

11: \usepackage{amssymb}

12: %\usepackage{bbm}

13: %\usepackage{graphicx}

14: %\usepackage{psfig}

15: %\usepackage{feynmf}

16: %\usepackage{axodraw}

17: 

18: \newcommand{\la}{{\lambda}}

19: \newcommand{\ie}{{\it i.e.}}

20: \newcommand{\eg}{{\it e.g.}}

21: \newcommand{\mrm}[1]{\mbox{\rm #1}}

22: \newcommand{\be}{\begin{equation}}

23: \newcommand{\ee}{\end{equation}}

24: \newcommand{\ba}{\begin{eqnarray}}

25: \newcommand{\ea}{\end{eqnarray}}

26: \newcommand{\bi}{\begin{itemize}}

27: \newcommand{\ei}{\end{itemize}}

28: \newcommand{\bn}{\begin{enumerate}}

29: \newcommand{\en}{\end{enumerate}}

30: \newcommand{\bc}{\begin{center}}

31: \newcommand{\ec}{\end{center}}

32: \newcommand{\ul}{\underline}

33: \newcommand{\bT}{\bar{T}}

34: \def\bY{{\bf Y}}

35: \def\bA{{\bf A}}

36: \def\bB{{\bf B}}

37: \def\bP{{\bf P}}

38: \def\bU{{\bf U}}

39: \def\bV{{\bf V}}

40: \def\bm{{\bf m}}

41: \def\bM{{\bf M}}

42: \def\bk{{\cal {\bf K}}_\nu}

43: \def\tl{{\tilde{L}}}

44: \def\tm{{\tilde{m}}}

45: \def\te{{\tilde{e^c}}}

46: \def\tel{{\tilde{e}}}

47: \def\td{{\tilde{d^c}}}

48: \def\tq{{\tilde{Q}}}

49: \def\tu{{\tilde{u^c}}}

50: \def\mcirc{{\stackrel{o}{m}}}

51: \def\dem{\delta m^2}

52: %%%%%%%%%%%%%%%%% Uli's macros

53: \def\unity{{\hbox{1\kern-.8mm l}}}

54: \newcommand{\ov}{\overline}

55: \newcommand{\ta}{\tilde{a}}

56: \newcommand{\tH}{\tilde{H}}

57: \newcommand{\tn}{\tilde{N}}

58: \newcommand{\lt}{\left}

59: \newcommand{\rt}{\right}

60: \newcommand{\no}{\nonumber}

61: \newcommand{\nn}{\nonumber}

62: \newcommand{\ds}{\displaystyle}

63: \newcommand{\eps}{\epsilon}

64: %\newcommand{\to}{\rightarrow}

65: \newcommand{\eq}[1]{eq.(\ref{#1})}

66: \newcommand{\Eq}[1]{eq.(\ref{#1})}

67: \newcommand{\rfn}[1]{(\ref{#1})}

68: \newcommand{\imag}{{\rm Im}\,}

69: \newcommand{\real}{{\rm Re}\,}

70: \newcommand{\mev}{\mbox{MeV}}

71: \newcommand{\gev}{\mbox{GeV}}

72: \newcommand{\ga}{\gamma}

73: %\newcommand{\ov}[1]{\overline{#1}}

74: \def\Ord{\buildrel{\scriptscriptstyle <}\over{\scriptscriptstyle\sim}}

75: \def\OOrd{\buildrel{\scriptscriptstyle >}\over{\scriptscriptstyle\sim}}

76: 

77: \def\tv#1{\vrule height #1pt depth 5pt width 0pt}

78: \def\tvbas#1{\vrule height 0pt depth #1pt width 0pt}

79: 

80: \newcommand{\gsim}{\lower.7ex\hbox{$\;\stackrel{\textstyle>}{\sim}\;$}}

81: \newcommand{\lsim}{\lower.7ex\hbox{$\;\stackrel{\textstyle<}{\sim}\;$}}

82: 

83: \def\mysection#1{\noindent {\bf #1} }

84: 

85: \begin{document}

86: 

87: \begin{titlepage}

88: \begin{flushright}

89: CERN-PH-TH/2006-159 \\  CPHT-RR-nnn.0806 \\ LPT-ORSAY-06-55 \\

90: {\tt hep-th/0608054}

91: \end{flushright}

92: 

93: \vskip.5cm

94: \begin{center}

95: {\huge \bf Dual realizations of dynamical symmetry breaking} \\

96: \vskip.1cm

97: \end{center}

98: \vskip0.2cm

99: 

100: \begin{center}

101: {\bf {Emilian Dudas}$^{a,b,c}$ and  {Chlo\'e Papineau}$^{c,b}$}

102: \end{center}

103: \vskip 8pt

104: 

105: \begin{center}

106: $^{a}$ {\it CERN Theory Division, CH-1211, Geneva 23, Switzerland} \\

107: \vspace*{0.1cm}

108: $^{b}$ {\it Centre de Physique Th\'eorique~\footnote{Unit{\'e} mixte du CNRS et de l'EP, UMR 7644.}, Ecole Polytechnique, 91128 Palaiseau Cedex, France} \\

109: \vspace*{0.1cm}

110: $^{c}$ {\it

111:  LPT~\footnote{Unit{\'e} mixte du CNRS, UMR 8627.},

112: B{\^a}t. 210, Univ. de Paris-Sud, 91405 Orsay Cedex, France} \\

113: \vspace*{0.3cm} {\tt Emilian.Dudas@cpht.polytechnique.fr,

114: chloe.papineau@th.u-psud.fr}

115: \end{center}

116: 

117: \vglue 0.3truecm

118: 

119: \begin{abstract}

120: \vskip 3pt \noindent We show the infrared equivalence between a recently proposed model containing a six dimensional scalar

121: field with a four-dimensional localized Higgs type potential and the

122: four-dimensional Nambu-Jona-Lasinio (NJL) model. In the dual NJL description, the fermions are localized at the origin

123: of a large two-dimensional compact space. Due to a  classical running effect above the compactification scale,

124: the four-fermion coupling of the NJL model increases from the cutoff scale down to the compactification scale, 

125: providing the large Fermi coupling needed for the dynamical symmetry breaking. We also present a string theory embedding of our field-theory construction. On more general grounds, our results suggest that 4d models with dynamical symmetry breaking can be given a higher

126: dimensional description in terms of field theories with nontrivial boundary conditions in the internal space.  

127: 

128: \end{abstract}

129: 

130: \end{titlepage}

131: 

132: \newpage

133: 

134: %\renewcommand{\thefootnote}{(\arabic{footnote})}

135: 

136: 

137: %%%%%%%%%%%%%%%%

138: \section{Introduction and Conclusions}

139: 

140: Dimensional transmutation and generation of a small scale is a remarkable result common to many quantum

141: field theories, most notably the four-dimensional QCD and the two-dimensional Gross-Neveu model. The

142: effect is also realized in two-dimensional quantum mechanics with a deep (delta-like) attractive potential

143: and in six-dimensional scalar models\footnote{For gravitational aspects of codimension two models, see

144: e.g. \cite{6d}.} with 4d localized scalar potential and a large but compact

145: transverse space. The last example, put forward in \cite{dpr}, was analyzed from the point of the

146: quantum-mechanical problem in the case of a perturbative coupling $\mu$ appearing as the (dimensionless)

147: localized parameter interpreted as a mass term in 4d. It was shown that for a 6d scalar field with Dirichlet boundary condition

148: on a large two-dimensional compact space taken for simplicity to be a disk, there is a phase transition with a very light (compared

149: to the compactification scale) particle for a small critical value $\mu_c \ll 1$. The parameter $\mu$ was shown to

150: run between the cutoff scale $\Lambda$ and the compactification scale $R^{-1}$, such that precisely at the critical point, 

151: $\mu$ becomes large at $R^{-1}$. A very similar phenomenon of appearance of a light state close to a critical point where 

152: an (four-fermion) interaction becomes strong is in the

153: 4d Nambu-Jona-Lasinio (NJL) model \cite{njl}.  The purpose of the present paper is to study closer this analogy and argue that

154: the 6d model studied in \cite{dpr} and the 4d NJL model are, in a sense which will be defined in detail later on, dual descriptions

155: of the same physics. 

156: Our starting point is to provide an explicit framework in which the 4d localized potential is generated,

157: by a Yukawa interaction of the bulk scalar field with N 4d localized fermions. In the large N limit,

158: integrating out the fermions produces precisely the potential needed for the symmetry breaking.

159: Alternatively, we show that integrating out the bulk scalar leads to a dual 4d NJL model with chiral symmetry breaking, where the

160: Fermi coupling is generated at tree-level by scalar bulk exchange. We show that the critical Fermi

161: coupling calculated by NJL methods in the large N limit agrees with the bulk 6d calculation of the

162: critical coupling calculated as a problem with nontrivial boundary conditions.     

163:    The 6d $\leftrightarrow$ 4d duality we study exchanges some quantum and classical natures of the symmetry breaking phenomenon.

164: In the bulk 6d picture, the quantum (Yukawa) interactions are completely encoded in a boundary condition, the localized 

165: scalar potential,  

166: whereas the symmetry breaking can be studied as a quantum-mechanical problem with nontrivial boundary condition and

167: can be understood as a result of a classical running effect in the transverse 2d space.  

168: In the 4d NJL picture, the symmetry breaking is provided by the nonperturbative self-consistent gap equation \cite{njl,nambu,bhl}, 

169: but in addition the four-fermion coupling has a classical logarithmic running between the cutoff and the compactification scale.

170: When in the bulk picture $\mu = \mu_c$ at $\Lambda$, the four-fermion coupling in the NJL picture at the compactification

171: scale reaches the critical value for the dynamical symmetry breaking $G (R^{-1}) = G_c$. 

172:   Our main interest in this equivalence is that, whereas a consistent treatment of the 

173: NJL model involves nonperturbative techniques like the large $N$ expansion or going below 4d and using UV nontrivial fixed points

174: and $1/\epsilon$ techniques, the bulk analysis is essentially classical\footnote{Throughout the paper by ``classical''

175: we mean classical from the point of view of quantum field theory, i.e. no quantum interaction. The treatment is still

176: quantum mechanical.} and does not need, in principle, any nonperturbative techniques.

177:  

178: The structure of the paper is as follows. In section 2 we review the six dimensional model worked out in \cite{dpr} and argue

179: that, in addition to the perturbative critical coupling $\mu_c$ we found there, there are other critical points corresponding

180: to large values $\mu_c^{(n)} > 1$. In section 2.1 we study a similar setup in which the nontrivial boundary condition in the compact

181: space is replaced by a bulk mass and Neumann boundary condition. In Section 3 we define the 6d model on an orbifold space instead of

182: a disk, which is more suitable for a microscopic (string theory) realization. Section 4 contains the main arguments concerning the

183: infrared equivalence between the 6d scalar model with nontrivial boundary condition and the 4d NJL model. Section 5 generalizes

184: the previous section to a (softly broken) supersymmetric theory. 

185: %Section 6 describes briefly some phenomenological 

186: %considerations independent of the explicit model analyzed in the rest of the paper,

187: %in particular the possibility to increase the compositeness scale of the NJL model above the compactification scale, in which

188: %case the top and the Higgs mass are lighther than in the standard top condensation application of the NJL model \cite{nambu,bhl}.      

189: Section 6 provides an explicit string theory realization of the present setup in an orientifold of type IIB strings with D-branes.

190: %%%%%%%%%%%%%%%%

191: \section{Six dimensional phase transition : perturbative and nonperturbative critical couplings}

192: 

193: Recently, \cite{dpr} addressed the problem of spontaneous symmetry breaking in a 6d scalar model 

194: with 4d localized Higgs potential. The corresponding action

195: reads\footnote{We are using a $(+,-,-,-,-,-)$ metric. The index $M$

196: denotes bulk coordinates and runs from $0,1,2,3,4,5$, while

197: $\mu=0,1,2,3$ denotes brane coordinates. We'll use either $x^{4,5}$

198: or $y_{1,2}$ to denote the two extra dimensions.} : 

199: \begin{eqnarray}

200: && S \ = \ \int d^4 x d^2 y \ \biggl[ {1 \over 2} (\partial_M

201: \phi)^2 \ - \  V_\delta (\phi) \biggr] \ , \nonumber \\

202: && V_\delta (\phi) \ = \ \left(-\frac{\mu^2}{2}\phi^2 +

203: \frac{\lambda}{4} \phi^4 \right) \cdot \delta^2 ({ y})  \ .

204: \label{npt1} 

205: \end{eqnarray}

206:  The scalar field $\phi$ has dimension two

207: and therefore $\mu^2$ is dimensionless. The scalar potential is

208: localized at the origin of the compact space. 

209: We resolve the singularity at ${ y}=0$ by introducing a disk $r < \epsilon$ supporting the potential,

210: \ba

211: && V(\phi) \ = \ \frac{1}{\pi \epsilon^2}\left(-\frac{\mu^2}{2}\phi^2 +

212: \frac{\lambda}{4} \phi^4 \right)  \quad {\rm for} \ 0 < r < \epsilon \ ,  \label{npt01} \\

213: && V(\phi) \ = \ 0  \quad \quad \quad \quad \quad \quad \quad \quad \quad \quad

214: {\rm for} \ \epsilon  < r < R \ .  \nonumber

215: \ea

216: According

217: to \cite{gw,dgv,dpr}, there is a "classical" running of the

218: tachyonic mass parameter 

219: \be 

220: {1 \over \mu^2 (Q)} = {1 \over \mu^2

221: (\Lambda)} - {1 \over 2 \pi} \ln ({\Lambda \over Q}) \ .

222: \label{pt18} 

223: \ee 

224: It was shown in

225: \cite{dpr} for a compact two-dimensional space, chosen to be a

226: disk of radius $R$, with 4d localized Higgs potential and Dirichlet boundary condition

227: \be

228: \phi |_{r =R} \ = \ 0 \ , \label{npt02}

229: \ee

230: that this model has a phase transition for a small

231: critical value 

232: \be {\mu_c^2 \over 2 \pi} \ln (R \Lambda) = 1 \ ,

233: \label{npt2} 

234: \ee 

235: where $\Lambda$ is a UV cutoff defined in connection with the resolution of the delta

236: singularity $\Lambda = 1 / \epsilon$. So the phase transition happens precisely when the renormalized value

237: $\mu_c^2 (R^{-1}) \rightarrow \infty$ blows up at the compactification scale. 

238: The running interpretation breaks down close to the

239: phase transition point. The classical running of $\mu$ induces also a running

240: for the self-coupling $\lambda$, according to the RG equation

241: \be

242: Q {d \lambda \over d Q} \ = \ - {2 \over \pi} \mu^2 \lambda \ , \label{pt018}

243: \ee

244: which, by using (\ref{pt18}), readily  integrates to

245: \be

246: \lambda (Q) \ = \ { \lambda (\Lambda) \over (1 - {\mu^2 \over 2 \pi} \ln {\Lambda \over Q})^4} 

247: \ . \label{pt019}

248: \ee

249: Notice that at the phase transition point $\mu = \mu_c$, 

250: \be

251: \mu (R^{-1}) \ \rightarrow \ \infty \quad , \quad \lambda (R^{-1}) \ \rightarrow \ \infty \ . \label{pt020} 

252: \ee

253: 

254: We will argue later on in section 4 that in a dual 4d theory which turns out to be a NJL theory, the conditions (\ref{pt020})

255: have the interpretation of compositeness conditions of \cite{bhl}. Close to the critical coupling, however, the running interpretation

256: breaks down and actually the higher-dimensional 6d and also the 4d physics turn out to be perturbative.

257:   

258: We now review and slightly update the arguments of \cite{dpr} by arguing that there are actually

259: additional but large critical couplings $\mu_c^{(n)} \ge \sqrt{4 \pi}$, 

260: defined by the presence of a 4d massless

261: mode in the spectrum. 

262: Assuming that this exists, slightly below it $\mu \le \mu_c$,  

263: in the background $\phi_c = 0$, the field eqs. for a 4d mode of mass $M^2 = p^2$, are  

264: \ba

265: &&  \Delta^{(2)} \phi \ + \ \frac{\mu^2}{\pi \epsilon^2} \phi=

266: 0 \; , \;\;\; r < \epsilon \ , \nonumber \\

267: &&   \Delta^{(2)} \phi \ + p^2 \phi \ = \ 0 \; , \;\;\; r > \epsilon \ ,  \label{npt3}

268: \ea

269: where we neglected the mass $p^2$ inside the brane (this is a very good approximation for all masses much lighter than

270: the cutoff $\Lambda = 1/ \epsilon$). The solutions of (\ref{npt3}) with Dirichlet boundary condition 

271: (\ref{npt02}) and for $p^2 \ll R^{-2}$ are 

272: \ba

273: && \phi (r) \ = \ f_0 \ J_0 ({\mu r \over \sqrt{\pi} \epsilon}) \quad , \qquad r < \epsilon \ , \nonumber \\ 

274: &&  \phi(r) \ = \ a \ \left[ \ln \frac{R}{r} - \frac{p^2 r^2}{4} \ln  \frac{R}{r} + \frac{p^2}{4}

275: (R^2 - r^2) \right]  \quad , \qquad  r > \epsilon \ .

276: \label{npt4}

277: \ea

278: The zero mass solutions $p^2=0$ define the whole set of critical couplings.

279: The matching conditions of the logarithmic derivative of the wave function at $r = \epsilon$ then give

280: \be

281: {\mu_c \ J'_0 ({\mu_c \over \sqrt{\pi}}) \over \sqrt{\pi} \ J_0 ({\mu_c \over \sqrt{\pi}})} \ \ln {R \over \epsilon}

282: \ = \ -1 \ . \label{npt5} 

283: \ee

284: For small $\mu_c$, eq. (\ref{npt5}) has the unique solution (\ref{npt2}), which is indeed small provided 

285: that $R^{-1} \ll \Lambda$. 

286: Equation (\ref{npt5}), however, has an {\it infinite discrete set} of solutions, as can easily be shown by a numerical

287: plot. The peculiarity of the perturbative solution (\ref{npt2}) is that the wave function  (\ref{npt4}) inside

288: the brane $r < \epsilon$, and actually also the wave functions of the massive modes below the cutoff $\Lambda$, are almost constant. 

289: On the contrary, the wavefunctions corresponding to the ``nonperturbative'' critical couplings 

290: (in the sense $\mu_c^2 / 4 \pi >1 $) 

291: have substantial variation inside the brane. Slightly below the critical coupling(s) $\mu_c$, the zero

292: mode becomes massive, with a mass $M$ given approximately by

293: \be

294: M^2 \ = \ p_{\mu} p^{\mu}  \ \simeq \ {2 (\mu_c^2 - \mu^2) \over \ \pi R^2} \  (\ln {R \over \epsilon})^2  \ 

295: \ . \label{npt6}  

296: \ee

297: The mass  (\ref{npt6}) reduces to the one computed in \cite{dpr} in the case of the perturbative critical

298: coupling (\ref{npt2}). 

299: Close to the other critical couplings $\mu \simeq \mu_c^{(n)}$ the light mode can also be described in a 4d effective field theory.

300: The 6d field is decomposed according to $\phi (x,r) = \sigma (x) \chi (r)$, where  

301: \ba

302: && \chi (r) \ = \ \sqrt{2 \over \pi R^2} \ln {\left(R \over \epsilon\right)} \ \frac{J_0 ({\mu  r \over \sqrt{\pi} \epsilon})}{J_0({\mu \over \sqrt{\pi}})} \quad , \qquad r < \epsilon \ , \nonumber \\ 

303: &&  \chi(r) \ = \ \sqrt{2 \over \pi R^2}  \ \ln \frac{R}{r} \quad , \qquad  r > \epsilon \ .

304: \label{npt7}

305: \ea

306: The effective 4d potential for $\sigma$ is given by

307: \be

308: V_{eff} (\sigma) \ = \ {m_4^2 \over 2} \sigma^2 + {\lambda_4 \over 4} \sigma^4 \ , \label{npt8}

309: \ee

310: where the mass parameter and the coupling are given by

311: \ba

312: && m_4^2 \ = \ - {2 \mu^2 \over \epsilon^2}  \int_0^{\epsilon} r dr \chi^2 + 

313:    {2 \mu^2 \over \epsilon^2}  \int_0^{\epsilon} r dr (\chi')^2  + 

314: 2 \pi \int_{\epsilon}^R r dr  (\chi')^2 \ , \nonumber \\

315: && \lambda_4 \ = \ {2  \lambda \over \epsilon^2} \ \int_0^{\epsilon} r dr \chi^4 (r) \ = \ 

316: {8 \lambda \over \pi^2 \mu^2 R^4} {1 \over J_0^4 ({\mu / \sqrt{\pi}})} \ (\ln \frac{R}{\epsilon})^4 \

317: \int_0^{\mu \over \sqrt{\pi}} x dx J_0^4 (x) \ , \label{npt9}

318: \ea

319: where the derivative in $\chi'$ is wrt the argument of the Bessel functions. Very close to $\mu_c^{(n)}$, the resulting mass

320: coincides with $M^2$ in (\ref{npt6}), showing the validity of the 4d description.   

321: The presence of light 4d modes close to the large critical couplings $\mu_c^{(n)}$ is a signature of a UV-IR mixing, where

322: the UV physics changes the masses in the IR. While in a microscopic theory in which $\mu$ is generated dynamically, large values

323: ask presumably for nonperturbative effects, from the bulk 2d viewpoint, $\mu$ changes only the boundary conditions of the

324: scalar field and its consequences can be treated exactly quantum-mechanically. On the other hand, the explicit values of $\mu_c^{(n)}$

325: depend on the way we regularize the origin of the 2d space and thus on the UV physics. Therefore, the physical

326: consequences of the large critical couplings are probably highly sensitive on the UV physics. This is not the case

327: for the small critical coupling (\ref{npt2}), whose value is insensitive to the regularization procedure and therefore

328: of the UV physics, as we explicitly check by using a different regularization in section 3.   

329: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

330: \subsection{Phase transition with bulk mass and boundary Higgs potential}

331: 

332: A natural question arising in the present setup is what happens if one replaces the positive contribution 

333: to the mass coming from the Dirichlet boundary condition by a bulk mass $m$, keeping the 4d localized Higgs-type

334: potential.  The action describing this case is given by

335: \begin{eqnarray}

336: && S \ = \ \int d^4 x d^2 y \ \biggl[ {1 \over 2} (\partial_M

337: \phi)^2 \ - {1 \over 2} m^2 \phi^2 \ - \  V_\delta (\phi) \biggr] \ , \nonumber \\

338: && V_\delta (\phi) \ = \ \left(-\frac{\mu^2}{2}\phi^2 +

339: \frac{\lambda}{4} \phi^4 \right) \cdot \delta^2 ({ y})  \ ,

340: \label{bulkmass1} 

341: \end{eqnarray}

342: where the field $\phi$ has now Neumann boundary condition

343: \be

344: \partial_r \phi |_{r = R} \ = \ 0 \ . \label{bulkmass2}

345: \ee

346: We are working in the unbroken phase $\phi_c = 0$, in which case the field equations for a 4d field of mass

347: $p^2$ are

348: \ba

349: &&  \Delta^{(2)} \phi \ + \ \frac{\mu^2}{\pi \epsilon^2} \phi=

350: 0 \; , \;\;\; r < \epsilon \ , \nonumber \\

351: &&   \Delta^{(2)} \phi \ + (p^2-m^2) \phi \ = \ 0 \; , \;\;\; r > \epsilon \ .  \label{bulkmass3}

352: \ea

353: By defining $q^2 = m^2- p^2$, we find the solutions of (\ref{bulkmass3}) with Neumann boundary conditions (\ref{bulkmass2})

354: and for $q^2 \ll R^{-2}$ to be 

355: \ba

356: && \phi (r) \ = \ f_0 \ J_0 ({\mu r \over \sqrt{\pi} \epsilon}) \quad , \qquad r < \epsilon \ , \nonumber \\ 

357: &&  \phi(r) \ = \ a \ \left[ 1 + \frac{q^2 R^2}{2} \ln  \frac{R}{r} + \frac{q^2}{4}

358: r^2 \right]  \quad , \qquad  r > \epsilon \ .

359: \label{bulkmass4}

360: \ea

361: Matching conditions at $r = \epsilon$ for zero mass solutions $p^2=0$ define the critical couplings in this case to be

362: given by the solutions of 

363: \be

364: \left[ 1 - {\mu_c \ J_1 ({\mu_c \over \sqrt{\pi}}) \over \sqrt{\pi} \ J_0 ({\mu_c \over \sqrt{\pi}})} \ 

365: \ln {R \over \epsilon} \right] \ {m^2 R^2 \over 2} \ = \ 

366: {\mu_c \ J_1 ({\mu_c \over \sqrt{\pi}}) \over \sqrt{\pi} \ J_0 ({\mu_c \over \sqrt{\pi}})} \ . \label{bulkmass5}

367: \ee

368: Analogously to the case discussed in the previous section, eq. (\ref{bulkmass5}) has an 

369: infinity but discrete number of solutions $\mu_c^{(n)}$, out of which only {\it one} is perturbative $\mu_c \ll 1$. 

370: In this perturbative case, similarly to (\ref{npt2}), there is a classical running interpretation of the critical coupling

371: \be

372: \pi R^2 m^2 \ = \ {\mu_c^2 \over 1 - {\mu_c^2 \over 2 \pi} \ln {R \over \epsilon}} \ = \ \mu_c^2 (R^{-1}) \ ,

373: \label{bulkmass6}

374: \ee  

375: where $ \mu_c^2 (R^{-1})$ is the renormalized value of the (perturbative) critical coupling at the compactification 

376: scale $Q = R^{-1}$. From  (\ref{bulkmass6}) it follows that for small bulk mass $m^2 R^2 \ll 1$, $\mu$ stays perturbative

377: at all energies above the compactification scale, whereas for large masses $m^2 R^2 \ge 1$, 

378: $\mu$ enters strong coupling regime if there is a light 4d mode in the spectrum.    

379: 

380: Slightly below the critical couplings, the light 4d mass is given by

381: \be

382: p^2 \ = \ m^2 \ - \ {2 \over R^2} \   {{\mu \ J_1 ({\mu \over \sqrt{\pi}}) / 

383: (\sqrt{\pi} \ J_0 ({\mu \over \sqrt{\pi}}))}

384: \over  1 - \ln {R \over \epsilon} {\mu \ J_1 ({\mu \over \sqrt{\pi}}) / (\sqrt{\pi} \ J_0 ({\mu \over \sqrt{\pi}}))} \ 

385: } \ . \label{bulkmass7}

386: \ee  

387: Very close to the perturbative critical coupling (\ref{bulkmass6}) , the light mass becomes

388: \be

389: p^2 \ = \ m^2 -  {1 \over \pi R^2} \ {\mu^2 \over 1 - {\mu^2 \over 2 \pi} \ln {R \over \epsilon}} 

390:  = \ m^2 -  {1 \over \pi R^2} \mu^2 (R^{-1}) \label{bulkmass8}

391: \ee

392: and has again a transparent interpretation in terms of the classical running between the compactification scale

393: and the cutoff. 

394:  

395: There are also light 4d states for  $q^2 \sim R^{-2}$ or larger. However, we are especially interested in 

396: the case of small bulk masses $m^2 R^2 \ll 1$, for reasons to be explained in the dual NJL 

397: formulation of a supersymmetric extension of this model.

398: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

399: \section{ Symmetry breaking phase transition in orbifolds}

400: 

401: More standard and easy to handle spaces in string theory are orbifolds. We will

402: consider in the following a compactification on the orbifold

403: $T^2/\mathbb{Z}_2$ and check as a warmup the properties of the phase

404: transition. The orbifold acts as the reflection $(y_1,y_2)

405: \rightarrow (-y_1,-y_2) $. This orbifold has four fixed points. The fixed points and their corresponding

406: $\mathbb{Z}_2$ coordinate transformations are summarized as:

407: %

408: \begin{equation}

409:      \label{orbifoldfix}

410: \begin{array}{cccc}

411: y_1 \to -y_1 ~& y_1 \to -y_1 + 2 \pi R_1~ &~ y_1 \to -y_1 ~&~

412: y_1 \to -y_1 + 2 \pi R_1 \\

413: y_2 \to -y_2 ~&~ y_2 \to -y_2  ~&~ y_2 \to -y_2 + 2 \pi R_2

414: ~&~ y_2 \to -y_2 + 2 \pi R_2 \\

415: (0,0)~&~ (\pi R_1,0) ~&~ (0, \pi R_2)~&~ (\pi R_1, \pi R_2)

416: \label{pt2} \ .

417: \end{array}

418: \end{equation}

419: %

420: In complex notation, the action of $\mathbb{Z}_2$ on the compact space

421: is a  two-dimensional $\pi$ rotation,

422: $Z_2(y_1 +iy_2)$ = $e^{i \pi}(y_1 + i y_2)$.

423: 

424: The field equation is free in the bulk and has a delta function

425: source at the origin, suitably replaced by a mass distribution  

426: %

427: \be

428:     \label{sc1}

429: \partial_M \partial^M \Phi + {\partial V \over \partial \Phi} \, \delta^2 ({\bf y})= \ 0 \

430: . \label{pt3} \ee

431: Let us now proceed to study the mass spectrum for a scalar field

432:  with antiperiodic boundary conditions in the $y_1$-direction\footnote{This is the analog of the

433: Dirichlet boundary condition on the disk imposed in \cite{dpr}. As will become clear from our discussion, different boundary conditions, for

434: example $\Phi (y_1 ,y_2 + 2 \pi R_2) = - \Phi (y_1,y_2)$ or $\Phi (y_1 + 2 \pi R_1,y_2 + 2 \pi R_2) = - \Phi (y_1,y_2)$ lead to the same critical

435: coupling (\ref{npt2}). }

436: \be 

437: \Phi (y_1 + 2 \pi R_1,y_2) \ = \ - \ \Phi (y_1,y_2) \ .

438: \label{pt4} 

439: \ee

440: 

441: If the scalar field $\Phi$ is even under the orbifold action,

442: it can be decomposed on a complete basis formed by the cosine functions:

443: %

444:  \begin{equation}

445: \Phi(x,{\bf y})

446:  =

447:  \sum_{(k_1,k_2) \in \mathcal{I}}

448:  \frac{1}{\sqrt{2 \pi^2 R_1 R_2 }}

449: {\cos \left( \frac{k_1+1/2}{R_1} y_1 +  \frac{k_2}{R_2} y_2 \right)}

450:  \, \phi_{(k_1,k_2)} (x) \ . \label{pt5}

451:  \end{equation}

452: The indices $k_{1,2}$ belong to the set $\mathcal{I}$

453: %

454: \begin{equation}

455:     \label{pt6}

456: \mathcal{I} = \left\{ (0;0), (1\ldots \infty; 0), (0,-1;1\ldots

457: \infty), (1\ldots \infty; 1\ldots \infty), (1\ldots \infty; -\infty

458: \ldots -1) \right\}\ .

459: \end{equation}

460: %

461: In the unbroken vacuum, the quadratic part of the scalar action

462: takes the following form after integration over the two extra

463: dimensions

464: %

465: \begin{equation}

466:     \label{pt7}

467: \mathcal{L} = \mathcal{L}_{kin} - \frac{1}{2} \sum_{(k_1,k_2) \in

468: \mathcal{I}} \left( \frac{(k_1+1/2)^2}{R_1^2} + \frac{k_2^2}{R_2^2}

469: \right) \phi_{(k_1,k_2)}^2 + {\bar{\mu}^2} \left( \sum_{(k_1,k_2)

470: \in \mathcal{I}} \,  \, \phi_{(k_1,k_2)} \right)^2 \ ,

471: \end{equation}

472: %

473: where

474: %

475: \begin{equation}

476: \bar{\mu}^2 \equiv \frac{\mu^2}{4 \pi^2 R_1 R_2} \  \label{pt8}

477: \end{equation}

478: %

479: is the naive (volume suppressed) four dimensional lightest scalar

480: mass. The mass term of the 4d action  is

481: %

482: \begin{equation}

483: \mathcal{L}_{\rm mass} =

484: - {1 \over 2}

485: \sum_{(k_1,k_2),(p_1,p_2) \in \mathcal{I}}

486:  \phi_{(k_1,k_2)} \,

487: \mathcal{M}^2_{(k_1,k_2),(p_1,p_2)} \,

488:  \phi_{(p_1,p_2)} \ , \label{pt9}

489: \end{equation}

490: %

491: with the mass matrix given by

492: %

493: \begin{equation}

494:     \label{massmatrix1}

495: \mathcal{M}^2_{(k_1,k_2),(p_1,p_2)}  = - 2 {\bar \mu}^2 + \left(

496: \frac{(k_1+1/2)^2}{R_1^2} + \frac{k_2^2}{R_2^2} \right)

497: \delta_{k_1,p_1} \delta_{k_2,p_2} \ . \label{pt10}

498: \end{equation}

499: %

500: The diagonalization of this  mass matrix defines the physical mass

501: eigenstates.

502: 

503: Let us now try to find the eigenvalues and eigenvectors of the

504: mass matrix~(\ref{massmatrix1}). We use the techniques used in 5d models in \cite{ddg,bfz} and in 6d models in  \cite{dgv}.

505: The characteristic equation

506: is given

507: by

508: %

509: \begin{equation}

510:     \label{charectersticeq}

511:  \mathcal{M}^2 \Psi_m  \ = \  m^2 \Psi_m \ , \label{pt11}

512: \end{equation}

513: %

514: where $m^2$ represents the eigenvalues and $\Psi$ is the eigenvector

515: in the basis $\left| k_1,k_2 \right. \rangle_{(k_1,k_2) \in \mathcal{I}}$, i.e

516: $\Psi_{(k_1,k_2)} = \langle k_1, k_2 | \Psi_m \rangle$ .

517: The matrix equation~(\ref{charectersticeq}) is equivalent to the infinite set of explicit equations for every

518: $(k_1,k_2) \in \mathcal{I}$

519: %

520: \begin{equation}

521: 2 \bar{\mu}^2 \Psi' = \left( - m^2 + \frac{(k_1+1/2)^2}{R_1^2} +

522: \frac{k_2^2}{R_2^2} \right) \Psi_{(k_1,k_2)} \ , \label{pt12}

523: \end{equation}

524: %

525: where $\Psi'$ is independent of $(k_1,k_2)$. 

526: The solution of the equations~(\ref{pt12}) is given by

527: %

528: \begin{equation}

529: \Psi_{(k_1,k_2)} \ =  \ \frac{\mathcal{N}}{ - m^2 + (k_1+1/2)^2/R_1^2 +

530: k_2^2/R_2^2}

531:  \ , \label{pt13}

532: \end{equation}

533: %

534: where $\mathcal{N}$ is a normalization constant independent of

535: $(k_1,k_2)$. Putting this solution back in the equation~(\ref{pt12})

536: and using the fact that 

537: \be 

538: \sum_{(k_1,k_2) \in \mathcal{I}} \ = \

539: {1 \over 2} \sum_{k_1,k_2 = \ -\infty}^{\infty} \ ,  

540: \ee

541: we obtain the eigenvalue equation

542: %

543: \begin{equation}

544:     \label{eigenvalue}

545: \frac{1}{\bar{\mu}^2} = \sum_{k_1, k_2 = -\infty}^\infty

546: \frac{1}{ - m^2 + (k_1+1/2)^2/R_1^2 + k_2^2/R_2^2 } \ , \label{pt14}

547: \end{equation}

548: or equivalently

549: \begin{eqnarray}

550: && {1 \over \mu^2} \ = \ D (p^2 = m^2 , y_1=y_2=0) \quad , \quad {\rm where} \nonumber \\

551: && D (p^2  , y_1,y_2) \ = \ {1 \over 4 \pi^2 R_1 R_2}  \sum_{k_1, k_2 = -\infty}^\infty

552: \frac{\cos [(k_1+1/2) y_1/R_1 + k_2 y_2/R_2] }{ - p^2 + (k_1+1/2)^2/R_1^2 + k_2^2/R_2^2 } \ \label{pt014} 

553: \end{eqnarray}

554: is the propagator in a mixed, 4d momentum and 2d position, representation.  

555: We want to find an estimate for the lightest solution, $m^2$, of the

556: eigenvalue equation (\ref{pt14}). The procedure we are using is

557: similar to the one used in \cite{dgv} and we only give the result

558: here. It is clear from (\ref{pt14}) that there is a critical

559: coupling, defined for arbitrary radii by 

560: \be 

561: {4 \pi^2 R_1 R_2 \over \mu_c^2} \ = \ \sum_{ |k_i| < k_i^{max}}

562: \frac{1}{(k_1+1/2)^2/R_1^2 + k_2^2/R_2^2 } \ , \label{pt15} 

563: \ee

564: which signals a second order phase transition, where the lightest

565: mass $m^2$ changes sign. For equal and large radii and by cutting the sums at $k_i^{max} = R \Lambda$, we find $\mu_c$

566: to be exactly equal to the value (\ref{npt2}) worked out in \cite{dpr}. 

567: A puzzle arises however in this KK approach to the phase transition. Indeed, whereas we accurately describe the perturbative critical

568: coupling (\ref{npt2}), eq. (\ref{pt15}) does not contain the nonperturbative couplings $\mu_c^{(n)}$ present in  (\ref{npt5}). We

569: believe that this is due to the way the logarithmic divergence is handled in (\ref{pt15}), or equivalently, to the ``brane

570: resolution'' for nonperturbative values of $\mu^2$. Indeed, as we already mentioned, in this case wave functions oscillate

571: significantly inside the brane and the regularization procedure becomes more subtle.  

572: 

573: Very close and slightly below the phase transition we can linearize

574: the mass equation (\ref{pt14}) in order to find for the lightest mode

575: \be

576:  m^2 \ \simeq \

577: {4 \pi^2 \over \alpha R_1 R_2} \ {\mu_c^2 - \mu^2 \over \mu_c^4}

578: \simeq \ {4 \pi^2 \over \alpha R_1 R_2} ( {1 \over \mu^2} - {1 \over

579: \mu_c^2}) \ ,

580:  \label{pt16}

581:  \ee 

582: where 

583: \be 

584: \alpha \ = \ R_1^{-2} R_2^{-2} \

585: \sum_{k_1, k_2 = -\infty}^\infty \frac{1}{[(k_1+1/2)^2/R_1^2 +

586: k_2^2/R_2^2]^2 } \ . 

587: \label{pt17} 

588: \ee 

589: The mass (\ref{pt16}) is not

590: exactly the same as the one worked out in \cite{dpr}. The reason is

591: that the geometries in \cite{dpr} and in the present section are

592: different and affect the IR physics, in particular physical masses. UV

593: physics, however is the same for both geometries ; in

594: particular the value (at the leading order in $\mu^2$) of the

595: perturbative critical coupling $\mu_c$ defining the phase transition is the same as in (\ref{npt2}). Since the regularizations

596: used on the disk and in (\ref{pt15}) are different, this shows the regularization independence of the perturbative critical 

597: coupling.   

598: 

599: Above the critical value, the scalar gets a vev (more precisely, a profile in the compact space) 

600: \be

601: \phi_c (y_1,y_2 ) \ = \ \phi_0 {\cal N} \sum_{k_1,k_2} {\cos [{(k_1+1/2) y_1 / R_1} + {k_2 y_2 / R_2}]

602:  \over  {(k_1+1/2)^2 / R_1^2} +  {k_2^2 / R_2^2}} \ , \label{pt015}

603: \ee

604: where according to \cite{dpr}

605: \be

606: \phi_0^2 \ = \ {\mu^2 - \mu_c^2 \over \lambda}  

607: \ee

608: and ${\cal N}$ is a normalization constant such that at the origin (more precisely, at the regularized mass distribution),

609: $\phi_c (0,0 ) = \phi_0$.  

610: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

611: \section{Localized matter and dual description : the NJL model}

612: 

613: An immediate question is how to generate in a natural way the

614: localized scalar potential needed for the symmetry breaking. The

615: simplest idea is to add (Weyl for definiteness, but the situation is

616: similar for Dirac fermions) $N$ fermions $\chi_i$ on the boundary,

617: with Yukawa couplings 

618: \be

619: g \chi_i \chi_i \phi (y=0) \ + \ {\rm h.c.} \ , \label{pt021}

620: \ee

621: to the (now complex) bulk field. The model has, in addition to a global $SO(N)$ symmetry, a continuous chiral $U(1)$

622: symmetry under which $\phi$ is charged. 

623: 

624: \subsection{The bulk picture}

625: One-loop quantum corrections generated by the fermion loops automatically

626: generate a scalar potential of the appropriate form (\ref{npt1}),

627: plus higher-order terms. The continuous chiral

628: symmetry will now be spontaneously broken at the phase transition.

629: 

630: In the large $N$ limit, the leading induced scalar potential is \be

631: V_{\rm eff} (\phi) \ = \ - \ N \int {d^4 p \over (2 \pi)^4} \  \ \ln

632: (p^2 + 4 g^2 |\phi|^2)  \ , \label{pt19}

633:  \ee

634: which can be expanded in powers of $\phi$ as

635:  \be

636: V_{\rm eff} (\phi) \ = \ - 4 N g^2 |\phi|^2 \int {d^4 p \over (2

637: \pi)^4} {1 \over p^2 } + 8 N  g^4 |\phi|^4 \int {d^4 p \over (2

638: \pi)^4} {1 \over p^4 } \ + \ \cdots \ . \label{pt20}

639:  \ee

640: We therefore induced radiatively, to the leading order in an $1/N$

641: expansion, a potential of the form (\ref{npt1}) with \be \mu^2 \ = \

642: {N g^2 \over 4 \pi^2} \ \ \Lambda^2  \ . \label{pt21} \ee As usual

643: \cite{cw} the power expansion in $\phi$ has severe IR divergences,

644: which are however resummed in the effective potential (\ref{pt19}).

645: Then the condition defining the symmetry breaking phase is \be {N

646: g^2 \over 4 \pi^2} \ \Lambda^2 \

647: > \ {4 \pi \over \ln (R^2 \Lambda^2)} \ , \label{pt22} \ee whereas

648: the perturbative expansion used in \cite{dpr}, for $\mu^2 \ll 1$,

649: translates here into \be {N g^2 \over 4 \pi^2} \ \Lambda^2  \ll 1 \ .

650: \label{pt23} \ee For $\langle \phi \rangle \not=0$ the brane

651: fermions $\chi_i$ acquire a mass and the chiral symmetry is

652: spontaneously broken, with the imaginary part of $\phi$ being the Goldstone boson.  

653: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

654: \subsection{Dual picture : the NJL model}

655: There is a dual description in which the bulk field $\phi$ is

656: integrated out at tree-level and the chiral symmetry breaking is

657: entirely described in terms of nonperturbative brane dynamics. The

658: resulting brane lagrangian has the simple form \be S_{\rm brane} \ =

659: - i \ \chi_i \sigma^{\mu} \partial_{\mu} {\bar \chi}_i

660:  \ + \ G \chi_i\chi_i {\bar \chi}_j {\bar \chi}_j \ ,

661: \label{pt24}

662:  \ee

663:  with

664:  \be

665: G \ = \ {g^2 \over 2 \pi^2 R_1 R_2} \sum_{(k_1,k_2) \in \mathcal{I}}

666: \  {1 \over \frac{(k_1+1/2)^2}{R_1^2} + \frac{k_2^2}{R_2^2}} \

667: \simeq {g^2 \over 4 \pi} \ \ln (R^2 \Lambda^2), \label{pt25}

668:  \ee

669:  where in the last equality we considered equal and large radii

670:  $R_1 = R_2 = R \gg M_*^{-1}$, where $M_*$ is the 6d fundamental (Planck) scale. Therefore the "dual"

671:  lagrangian (\ref{pt24}) is the Nambu-Jona-Lasinio model \cite{njl}, in which the chiral symmetry is dynamically

672: broken by the fermion condensate $\langle \chi_i \chi_i \rangle$

673: for values of the four-fermion coupling above the critical value

674: \be G \ > \ G_c \quad , \quad {\rm where} \ G_c^{-1} \ = \ {N

675: \Lambda^2 \over 4 \pi^2} \ \ . \label{pt26} \ee By using (\ref{pt21}) and

676: (\ref{pt25}), we find that the condition (\ref{pt26}) is precisely

677: the same  as the condition for the broken phase derived in the

678: "bulk" approach (\ref{pt22}).

679: 

680: Whereas in the deep IR the 6d bulk model is equivalent to the 4d NJL model, their UV behaviour is different\footnote{For earlier

681: ideas of the role of extra dimensions in dynamical symmetry breaking, see \cite{dobrescu}. For a recent extensive review on strong dynamics,

682: see e.g. \cite{hills}.}. In particular, due to the cumulative effects of the KK states, the four-fermion

683: coupling $G$ has a logarithmic running \be G (Q) \ \simeq {g^2 \over

684: 4 \pi} \ \  \ln (\Lambda^2/Q^2) \label{pt27} \ee from the cutoff

685: scale $\Lambda$ to the compactification scale $R^{-1}$,

686: as illustrated in Fig.(\ref{fig:running}). So $G$

687: increases in the IR and can generate dynamical chiral symmetry

688: breaking. Even for couplings $g^2 \ll 4 \pi$ such that a perturbative treatment is available, the non-decoupling of heavy KK states 

689: generates a large four-fermion coupling in the infrared which drives the symmetry breaking. 

690: 

691: 	\vspace{0.3cm}

692: \begin{figure}[htbp]

693:     \begin{center}

694: \centerline{

695:        \epsfig{file=runningG.eps,width=0.5\textwidth}

696:        }

697:           \caption{{\footnotesize

698: Classical running of the four-fermion coupling induced by the cumulative effect of the Kaluza-Klein states. The coupling

699: becomes strong $G=G_c$ at the compactification scale and induces chiral symmetry breaking.

700: }}

701: \label{fig:running}

702:     \end{center}

703: \end{figure}

704: 

705: 

706: In the bulk formulation, the mass parameter was

707: generated through quantum loops and the phase transition had a

708: "classical" (quantum mechanical) nature. In the NJL formulation, the

709: four-fermion coupling is generated classically while the phase

710: transition is generated in a nonperturbative fashion through the quantum gap equation 

711: \be 1 \ =

712: \ 4 N  G \int {d^4 p \over (2 \pi)^4} {1 \over p^2 + m^2} \ = \ 4 \ N \ G \ D (x=0,m^2)

713:  \ . \label{pt28} 

714: \ee 

715: 

716: In the previous expression, $D (x,m^2)$ is the 4d propagator in the position representation of a scalar field,  

717: and $m$ is the dynamical fermion mass. 

718: 

719: The counterpart of the NJL quantum gap

720: equation (\ref{pt28}) in the bulk formulation is the classical gap

721: equation (\ref{pt14})-(\ref{pt014}).  The similarity of the bulk equation (\ref{pt014}) and the NJL gap equation 

722: (\ref{pt28}) is transparent.  

723: 

724: Notice that the dual NJL description is valid in principle only below the phase transition. In this case and when the classical 

725: running interpetation is valid, the connection between the Higgs-localized 

726: scalar potential and the NJL model can also be understood in terms 

727: of the compositeness condition of \cite{bhl}. There, it was argued, by introducing a composite scalar field $H$, 

728: that the kinetic term $Z_H |\partial H|^2$ vanishes, $Z_H \rightarrow 0$, at the compositeness scale $\Lambda_C$. By a rescaling  

729: of the kinetic term to the canonical form, this is equivalent of imposing the boundary conditions $m_H \rightarrow \infty$, 

730: $\lambda_H \rightarrow \infty$,

731: where the parameters $m_H,\lambda_H$ are defined from the effective scalar potential $V = - m_H^2 |H^2| + \lambda_H |H|^4$.

732: In our 6d model, the running between the UV cutoff and the compactification scale $R^{-1}$ produces $\mu^2$ and $\lambda$ to

733: diverge at $R^{-1}$ precisely at the critical point. With a rescaling $\phi \rightarrow Z^{1/2} \phi$ of the scalar to go from the convention of \cite{dpr} to

734: the normalization of \cite{bhl}, keeping the scalar mass fixed, 

735: the wave function is

736: \be

737: Z (Q) \ = \ 1 - {\mu^2 \over 2 \pi} \ln {\Lambda \over Q} \ . \label{pt29}

738: \ee

739: At the phase transition $\mu = \mu_c$, we get

740: \be

741: Z (Q = R^{-1} , \mu=\mu_c) \ = \ 0 \quad , \quad \lambda (Q=R^{-1}) \ = \ \lambda (\Lambda) \ 

742: Z^{-2} (Q = R^{-1} , \mu=\mu_c) \ \rightarrow \infty \ .  \label{pt30}

743: \ee

744: The first of these conditions is similar to the one-loop (large N) induced wave function displayed

745: in \cite{bhl}, whereas the one for scalar self-coupling is different, since the UV physics is different. 

746: A natural candidate for the compositeness scale in our case is therefore $\Lambda_C = 1/R$.      

747: 

748: Analogously to the scalar parameters $\mu , \lambda$, the Yukawa coupling $g$ gets an induced running which can be easily

749: integrated :

750: \be

751: Q {d g \over d Q} \ = \ - {\mu^2 \over 2 \pi} g \quad , \quad g (Q) \ = \ 

752: { g (\Lambda) \over 1 - {\mu^2 \over 2 \pi} \ln {\Lambda \over Q}} 

753: \ . \label{pt024}

754: \ee

755: The compositeness conditions have therefore to be supplemented with

756: \be

757: g (Q = R^{-1}) \ \rightarrow \ \infty \ , \label{pt025}

758: \ee

759: in analogy with the compositeness condition for the top quark in the top condensation scenario \cite{bhl}. 

760: Whereas the first and the third conditions (\ref{pt30})-(\ref{pt025}) are indeed similar to the compositeness conditions in \cite{bhl},

761: close to the critical coupling $\mu_c$ the running interpretation breaks down and the 4d physics is 

762: actually completely perturbative. Indeed,

763: defining the 4d effective theory as in \cite{dpr} by

764: \ba

765: && \phi (x^{\mu},r) \ = \ \sqrt{2 \over \pi R^2} \ \ln {R \over \epsilon} \ \sigma (x^{\mu}) \quad , \quad r < \epsilon \ , \nonumber \\ 

766: && \phi (x^{\mu},r) \ = \ \sqrt{2 \over \pi R^2} \ \ln {R \over r} \ \sigma (x^{\mu})   \quad , \quad \epsilon < r < R \ , \label{pt031} 

767: \ea

768: we find, at $\mu \simeq \mu_c$, the 4d scalar self-coupling and Yukawa coupling to the fermions to be

769: \be

770: \lambda_4 \ = \ {64 \pi^2 \over \mu_c^8} \ {\lambda \over R^4} \quad , \quad  g_4 \ = \ {2 \sqrt{2\pi} \over \mu_c^2 } \ {g \over R}

771: \ . \label{pt032}

772: \ee 

773: Since $1/\mu_c^2 \sim \ln (\Lambda R)$, the couplings (\ref{pt032}) are log enhanced compared to their naive (volume suppressed) values,

774: but are still perturbative and under control. 

775: 

776: Whereas the NJL 4d description and the bulk 6d one with appropriate boundary conditions are equivalent in the IR, the 6d picture can sometimes

777: be simpler to use in order to describe the symmetry breaking pattern. For example, consider two set of fermions living in two different 

778: fixed points,

779: $(0,0)$ and $(0,\pi R_2)$ and interacting both with (relatively large) Yukawa coupling to the same 6d bulk scalar, which has nontrivial boundary conditions     

780: (\ref{pt4}). In the NJL picture, there are three four-fermionic interactions coming from the two sets of fermions, with specific four-fermion

781: couplings. In the bulk picture, localized loops of fermions generate localized scalar potentials at both fixed points.

782: The scalar action in this case is

783: \begin{eqnarray}

784: && S \ = \ \int d^4 x d^2 y \ \biggl[ {1 \over 2} (\partial_M

785: \phi)^2 \ - \  V_{\delta,1} (\phi) -   V_{\delta,2} (\phi) \biggr] \ , \nonumber \\

786: && V_{\delta,1} (\phi) \ = \ \left(-\frac{\mu_1^2}{2}\phi^2 +

787: \frac{\lambda_1}{4} \phi^4 \right) \cdot \delta ({ y_1})  \delta ({ y_2})  \ , \nonumber \\

788: && V_{\delta,2} (\phi) \ = \ \left(-\frac{\mu_2^2}{2}\phi^2 +

789: \frac{\lambda_2}{4} \phi^4 \right) \cdot \delta ({ y_1})  \delta ({ y_2 - \pi R_2})  \ . \label{pt31} 

790: \end{eqnarray}

791: It is a straightforward exercise to work out the equation defining the mass eigenstates (\ref{pt11}) in the unbroken phase. We find

792: \ba

793: && \left[ {1 \over {\bar \mu}_1^2} -  \sum_{k_i = -\infty}^\infty \frac{1}{m^2 + (k_1+1/2)^2/R_1^2 + k_2^2/R_2^2 } \right]

794: \times \left[ {1 \over {\bar \mu}_2^2} -  \sum_{k_i = -\infty}^\infty \frac{1}{m^2 + (k_1+1/2)^2/R_1^2 + k_2^2/R_2^2 } \right] \nonumber \\

795: && = \ \left( \sum_{k_i = -\infty}^\infty \frac{(-1)^{k_2}}{m^2 + (k_1+1/2)^2/R_1^2 + k_2^2/R_2^2 } \right)^2 \ . \label{pt32}

796: \ea   

797: Putting different fermions in different positions in the compact space can also provide a geometrical understanding of various values of their

798: Yukawa couplings via the wave function profile of the bulk scalar.  

799: 

800: 

801: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

802: \section{Supersymmetric extension and coupling to localized chiral fields}

803: 

804: The basic mechanism we used for generating a sizable running uses the

805: logarithmic terms coming from the renormalization of the localized mass term. This implies

806: in particular that the corrections to the mass should be forbidden in

807: the bulk. The natural way to implement this is to have supersymmetry

808: in the bulk. Since the potential has to be generated on the brane,

809: we add boundary chiral fields in supersymmetric multiplets, with

810: supersymmetry softly broken on the boundary.

811: 

812: There are two inequivalent ways to supersymmetrize, by using either bulk hypermultiplets or bulk vector multiplets.   

813: Let us start  with the first case and consider a bulk hypermultiplet in 6d, which in 4d ${\cal N}=1$ supersymmetric

814: language has two chiral (super)fields $\Phi_{1,2} = (\phi_{1,2} ,

815: \psi_{1,2})$, and $N$ localized matter superfields $A_i = (z_i ,

816: \chi_i)$, with $i = 1 \cdots N $, where we denoted in parenthesis

817: the scalar and fermionic components of the multiplets. By using the

818: ${\cal N}=1 $ superfield formalism of Ref. \cite{agw}, the action is

819: given by \ba  S \ = && \int d^2 y d^4 \theta \left( \Phi_1^{\dagger}

820: \Phi_1 + \Phi_2^{\dagger} \Phi_2 \right) + ( \int d^2 y d^2 \theta \

821: \Phi_1 \

822: (\partial_5 + i \partial_6) \ \Phi_2 + {\rm h.c.} ) \nonumber \\

823: && + \int d^4 \ \theta A_i^{\dagger} A_i  (1 - \Sigma^2 \theta^2

824: {\bar \theta}^2)  + \left( \int d^2 \theta \ g \ \Phi_1 (y=0) \ A_i

825: A_i + {\rm h.c.} \right) \ , \label{se1} \ea

826: 

827: where $\Sigma$ is a scalar soft mass term for the boundary fields. In order to write the

828: component lagrangian, we first solve for the auxiliary fields \ba &&

829: F_{\phi,1}^{\dagger} \ = \ - (\partial_5 + i \partial_6) \phi_2 - g

830: z_i z_i

831: \delta^2 (y) \ , \nonumber \\

832: &&  F_{\phi,2}^{\dagger} \ = \  (\partial_5 + i \partial_6) \phi_1

833: \quad , \quad F_{A_i}^{\dagger}  = - 2 g \phi_1 z_i \ . \label{se2}

834: \ea

835: 

836: After eliminating auxiliary fields, the component lagrangian is \ba

837: S = &&\int d^2 y \left[ |\partial_M \phi_1|^2 + |\partial_M \phi_2|^2 - i

838: \psi_1 \sigma^{\mu} \partial_{\mu} {\bar \psi}_1 - i \psi_2

839: \sigma^{\mu} \partial_{\mu} {\bar \psi}_2 + (\psi_1 (\partial_5+i \partial_6) \psi_2 + {\rm h.c.}) 

840: \right] \nonumber \\

841: && - \biggl[  S_{\rm kin} (z_i,\chi_i) + g \phi_1 \chi_i \chi_i + 2 g z_i

842: \psi_1 \chi_i + 4 g^2 |\phi_1|^2 |z_i|^2  \biggr. \nonumber \\

843: && + g  \biggl. \left\{ z_i z_i (\partial_5 - i \partial_6) {\bar \phi}_2 +

844: {\rm h.c.} \right\} + \Sigma^2 |z_i|^2 + g^2 z_i^2 {\bar z}_j^2

845: \delta^2 (0) \biggr] \bigg|_{y=0} \ . \label{se3} \ea

846: 

847: 

848: Analogously to the non supersymmetric case we impose nontrivial

849: boundary conditions \be \Phi_{1,2} (y_1+ 2 \pi R_1, y_2) \ = \ - \

850: \Phi_{1,2} (y_1, y_2) \ . \label{se4} \ee  The $\mathbb{Z}_2$ orbifold has a

851: nontrivial action on the bulk hypermultiplet fields. A consistent

852: assignement is $\Phi_1$ to be even and $\Phi_2$ to be odd.

853:  The KK expansions in this case are

854:  \ba

855: && \Phi_1 (x,{\bf y}) \ = \ \sum_{(k_1,k_2) \in \mathcal{I}}

856:  \frac{1}{\sqrt{2 \pi^2 R_1 R_2 }} \

857: {\cos \left( \frac{k_1+1/2}{R_1} y_1 +  \frac{k_2}{R_2} y_2 \right)}

858:  \, \Phi_1^{(k_1,k_2)} (x) \ , \nonumber \\

859: && \Phi_2 (x,{\bf y}) \ = \ \sum_{(k_1,k_2) \in \mathcal{I}}

860:  \frac{1}{\sqrt{2 \pi^2 R_1 R_2 }} \

861: {\sin \left( \frac{k_1+1/2}{R_1} y_1 +  \frac{k_2}{R_2} y_2 \right)}

862:  \, \Phi_2^{(k_1,k_2)} (x) \ . \label{se5}

863:  \ea

864: 

865: The second possibility of supersymmetrization is to add bulk 6d vector multiplets, which in 4d ${\cal N} = 1 $ language are described

866: by vector $V$ and chiral $\phi$ supermultiplets, both in the adjoint representation of a gauge group $G$. 

867: In order to be able to couple $\phi$ to the localized matter, we need to choose $Z_2$, the orbifold action, such that $V$ is 

868: odd and therefore has no zero modes, whereas $\phi$ is even and can therefore couple to boundary chiral multiplets. 

869: More precisely, we can start from a nonabelian gauge group and give a nontrivial action of the orbifold on the gauge

870: degrees of freedom

871: \be

872: V (-y_1,-y_2) \ = \ P^{\dagger} \ V (y_1,y_2) \ P \quad , \quad  

873: \phi (-y_1,-y_2) \ = \ - \ P^{\dagger} \ \phi (y_1,y_2) \ P \ , \label{se05}

874: \ee

875: where $P$ is a matrix in the adjoint representation of the gauge group, such that $P^2 = 1$. The surviving (even) gauge group generators

876: $T_a $ satisfy $[T_a,P] = 0 $, whereas the remaining ones $T_{\alpha}$ are projected out. The complementary states 

877: $\phi_{\alpha} $ from the adjoint scalar

878: $\phi$ survive at the fixed points and can consistently be coupled to the localized matter. By imposing nontrivial 

879: boundary conditions on $\phi_{\alpha}$ we generate a setup where $\phi_{\alpha}$ plays the role of the scalar field with 

880: localized Higgs potential and can trigger the phase transition. While this can be an interesting alternative, we will not pursue 

881: this possibility further on.   

882: 

883: %

884: %By defining the complex derivative $\partial = \partial_5 + i

885: %\partial_6$, the action of the model is described by

886: %\ba

887: %&& S \ = \ \left[ \int d^2 y d^2 \theta \ tr W^{\alpha} W_{\alpha} \ + \ {\rm h.c.} + \int d^2y d^4 \theta \ |\sqrt{2} V - \phi|^2 

888: %\ - \ |\partial V |^2 \right] \nonumber \\

889: %&& + \ \left[ \int d^2 \theta \ h \ A_i \ (\partial  - \phi) \ B_i \ + \ {\rm h.c.} \right] \ , \label{se6} 

890: %\ea

891: %where $A_i,B_i$ are localized chiral fields and $i$ is a flavor index. We introduced two different type of localized

892: %chiral multiplets in the view of the nonabelian generalization, where the gauge invariance   

893: % of (\ref{se6} asks for the doubling of the number of chiral multiplets. Notice that that coupling   

894: %$A_i \ \phi \ B_i$ can  be seen as the remnant of an ${\cal N} =2$ supersymmetric coupling between a hypermultiplet

895: %$(A_i,B_i)$ and the vector multiplet $(V,\phi)$. The main difference, however, is that due to the orbifold action, $V$

896: %is odd and has no zero modes, therefore the localized matter has no minimal coupling (no $U(1)$ charge). 

897: %Analogously to the previously discussed cases, we impose nontrivial boundary conditions

898: %\be

899: %(V , \phi) (y_1 + 2 \pi R_1, y_2) \ = \ - \ (V , \phi) (y_1 , y_2) \ , \label{se7}

900: %\ee

901: %which generate a supersymmetric mass for the bulk vector multiplet. The KK expansion of $\phi$ is then similar

902: %to the one of $\Phi_1$ in (\ref{se5}), whereas the one of $V$ is then similar

903: %to the one of $\Phi_2$.   

904: 

905: %The generalisation to the nonabelian case is straightforward but interesting. In this case, $A_i$ and $B_i$ transforms  

906: %in the fundamental representation of an orthogonal or symplectic gauge group,

907: %or fundamental (antifundamental) representation for $A_i$ ($B_i$) in the case of unitary bulk gauge group. The main interest

908: %in this case is that, whereas gauge fields have no zero modes and therefore there is no gauge group in four dimensions, bulk

909: %gauge symmetry is realized as the flavor symmetry in the four dimensional NJL model, through the gauge invariance of the

910: %nonabelian generalisation of the boundary coupling (\ref{se6}). 

911:    

912: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

913: \subsection{Bulk picture}

914: 

915: The leading quantum corrections in $1/N$ come at one-loop with the

916: chiral (super)fields $A_i$ running in the loop. There is an

917: induced effective potential for $\phi_1$ which can be computed in

918: the standard way \cite{cw}. The result is \be V_{\rm eff} (\phi_1) \

919: = \ N \int {d^4 p \over (2 \pi)^4} \ \{ \ln (p^2 + \Sigma^2 + 4 g^2

920: |\phi_1|^2) - \ln (p^2 + 4 g^2 |\phi_1|^2) \} \ , \label{bp1}

921:  \ee

922:  which can be expanded in powers of $\phi_1$ as

923:  \be

924: V_{\rm eff} (\phi_1) \ = \ - 4 N \Sigma^2 g^2 |\phi_1|^2 \int {d^4 p

925: \over (2 \pi)^4} {1 \over p^2 (p^2+\Sigma^2)} + 8 N \Sigma^2 g^4

926: |\phi_1|^4 \int {d^4 p \over (2 \pi)^4} {2 p^2 + \Sigma^2 \over p^4

927: (p^2+\Sigma^2)^2} \ + \ \cdots \ . \label{bp2}

928:  \ee

929: We therefore induced radiatively, in the leading order in an $1/N$

930: expansion, a potential of the form (\ref{npt1}) with \be \mu^2 \ = \

931: {N g^2 \over 4 \pi^2} \ \Sigma^2 \ \ln {\Lambda^2 \over \Sigma^2} \

932: . \label{bp3} \ee In a first approximation, we can consider this as

933: the bare coupling in the model of \cite{dpr}. Then the condition

934: defining the symmetry breaking phase, for large and equal radii

935: $R_1=R_2 \gg {1 \over \Lambda}$, is \be {N g^2 \over 4 \pi^2} \ \Sigma^2 \ \ln

936: {\Lambda^2 \over \Sigma^2} \

937: > \ {4 \pi \over \ln (R^2 \Lambda^2)} \ , \label{bp4} \ee whereas

938: the perturbativity condition translates here into \be {N g^2 \over 4 \pi^2} \ \Sigma^2 \ \ln

939: {\Lambda^2 \over \Sigma^2} \ll 1 \ . \label{bp5} \ee For

940: $\langle \phi_1 \rangle \not=0$ the brane fermions $\chi_i$ acquire a

941: mass and the chiral symmetry is spontaneously broken. Finally, fermion loops also induce localized operators of the 

942: form $|(\partial_5 + i \partial_6) \phi_2|^2$. Their effect is to

943: renormalize the KK masses of the odd field $\phi_2$, but this effect has no relevance for our present discussion.

944: 

945: %In the case of the bulk vector multiplets, the computation is very similar, with a localized potential generated

946: %for $\phi$ by loops of $A_i,B_i$. Since the computation is very similar to the one we just performed, we don't

947: %present it here. In the symmetry breaking phase, the flavor symmetry, which in this case is the original

948: %bulk gauge symmetry, is spontaneously broken.  

949: 

950: Notice that for natural values of (the dimensionful coupling) $g$, eq. (\ref{bp4}) can be satisfied only for large values 

951: of the soft breaking parameter $\Sigma$. This is easy to interpret in the dual NJL description, to which we now turn. 

952: 

953: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

954: \subsection{Dual description : the softly supersymmetric NJL model}

955: 

956: There is a dual description in which the bulk fields $\Phi_i$ are

957: integrated out at tree-level and the chiral symmetry breaking is

958: entirely described in terms of nonperturbative brane dynamics. Since

959: the bulk plus the interaction action is supersymmetric, the

960: integrating out procedure gives rise to a supersymmetric effective

961: action, to add to the brane lagrangian with softly broken

962: supersymmetry. There are some subtleties in proving that the

963: integration out leads to a well-defined four-dimensional

964: action without ill-defined (i.e.$\delta^2 (0)$) terms. Analogously

965: to former studies in 5d \cite{mp}, it can be checked that

966: the singular terms cancel out as they should. The resulting brane

967: lagrangian has the simple form 

968: \be 

969: S_{\rm brane} \ = \ \int d^4

970: \theta \{  A_i^{\dagger} A_i (1 - \Sigma^2 \theta^2 {\bar \theta}^2

971: ) \ + \ G \ A_i A_i  A_j^{\dagger} A_j^{\dagger}  \} \ ,

972: \label{njl1}

973: \ee

974:  where

975:  \be

976: G \ = \ {g^2 \over 2 \pi^2 R_1 R_2} \sum_{(k_1,k_2) \in \mathcal{I}}

977: \  {1 \over \frac{(k_1+1/2)^2}{R_1^2} + \frac{k_2^2}{R_2^2}} \ .

978: \label{njl2}

979:  \ee

980: Therefore the "dual" lagrangian is a softly-broken  supersymmetric

981: version \cite{njlsusy1, njlsusy2} of the Nambu-Jona-Lasinio model

982: \cite{njl}. The dynamics of the softly broken supersymmetric version

983: of the NJL model in the large $N$ expansion was investigated in

984: detail in \cite{njlsusy2}. It was found there that chiral symmetry

985: is dynamically broken by the fermion condensate $\langle \chi_i

986: \chi_i \rangle$, for values of the four-fermion coupling above the

987: critical value \be G \ > \ G_c \quad , \quad {\rm where} \ G_c^{-1}

988: \ = \ {N \Sigma^2 \over 4 \pi^2} \ \ln {\Lambda^2 \over \Sigma^2} \

989: . \label{njl3} \ee

990:  

991:  By using (\ref{njl2}), we find that the condition

992: (\ref{njl3}) is precisely the same  as the condition for the broken

993: phase derived in the previous section, which for equal and large

994: radii is displayed in (\ref{bp4}). 

995: 

996: As in the 4d supersymmetric NJL model, there is a naturalness problem in the 6d construction.

997: In the 4d SUSY NJL model, the symmetry breaking occurs for

998: values of the four-fermion interaction $G \gsim 1/\Sigma^2 \gg 1/ \Lambda^2 $ much larger than its natural value. For small supersymmetry

999: breaking $\Sigma R \ll 1$, this generates strong four-fermion interactions well below the compositeness scale $\Lambda_c = R^{-1}$.

1000: As transparent in (\ref{njl2}), the natural scale  of the strong four-fermion interactions for our 6d explicit realization 

1001: is actually $R^{-1}$, unless $g$ is much larger than its natural value. The reason is simpler to understand

1002: by rewriting (\ref{njl1}) in an appropriate form to compare with the Minimal Supersymmetric Standard Model (MSSM) \cite{njlsusy1}

1003: \ba

1004: S_{\rm brane} \ = && \ \int d^4

1005: \theta \{  A_i^{\dagger} A_i (1 - \Sigma^2 \theta^2 {\bar \theta}^2) + H_1^{\dagger} H_1 \} \ + \nonumber \\

1006: && \ \left( \int d^2 \theta H_2 \ ( m H_1 - g \ A_i A_i ) \ + \ {\rm h.c.} \right)   \ . \label{njl4}

1007: \ea        

1008: Since $H_2$ is a Lagrange multiplier in (\ref{njl4}), the two lagrangians (\ref{njl1}) and (\ref{njl4}) are equivalent

1009: for $G = g^2 / m^2$.  

1010: As explained in \cite{njlsusy1,njlsusy2},  $H_2$ acquires a kinetic term $ Z_2 H_2^{\dagger} H_2$ at one-loop in the large $N$ expansion,

1011: which vanishes at the compositeness scale $Z_2 (Q = \Lambda_c) \ = \ 0$. Below $\Lambda_c$, both $H_1$ and $H_2$ are dynamical fields

1012: and are to be identified with the two Higgs doublets of the MSSM. In the language of (\ref{njl4}), the naturalness problem is that

1013: in order to keep the Higgs mass light, the supersymmetric mass term should be of order $m \sim \Sigma$, which is nothing but the 

1014: reincarnation of the so-called $\mu$-problem of MSSM

1015: in the SUSY NJL case. In the 6d case, the analog of the action (\ref{njl4}) is (\ref{se1}), the analog of $H_2$ is $\Phi_1$, whereas

1016: the analog of the supersymmetric mass $m$ is the mixing term $\Phi_1 (\partial_5+i \partial_6) \Phi_2$ in (\ref{se1}). Upon KK expansion, 

1017: we get $m \sim (1 / 2R_1)$. Equivalently, this can be seen from the nontrivial boundary conditions (\ref{se4}) via the KK expansion (\ref{se5}).

1018: Since the symmetry breaking only occurs for $m \sim \Sigma$, the model also requires large supersymmetry breaking scale, as already

1019: anticipated. 

1020: 

1021: If we believe in the equivalence between the 6d scalar model and the 4d NJL model also close to the nonperturbative critical couplings

1022: of section 2, then the corresponding Yukawa coupling $g$ can be large and produce a larger Fermi coupling $G$. However it will be hard to argue

1023: reliably for very large Yukawas. 

1024: Maybe a simpler road is to use a small bulk mass $m R \ll 1$, which will generate a large Fermi coupling

1025: $G \sim g^2 / R^2 m^2$.      

1026: Another solution in order to get symmetry breaking compatible with small supersymmetry breaking scale $\Sigma$ is to start with small

1027: supersymmetry mass generated by boundary conditions. In other words, we need

1028: \be

1029: (\Phi_1 + i \Phi_2  ) (y_1 + 2 \pi R_1 , y_2) \ = \ e^{2 \pi i \omega} (\Phi_1 + i \Phi_2  ) (y_1 , y_2) \ , \label{njl5}

1030: \ee     

1031: with $\omega \ll 1$, in which case $m \sim (\omega /R_1)$. Whereas from first principles in string theory $\omega$ is quantized and cannot

1032: be very small, in analogy with known 5d examples \cite{bfz}, $\omega \ll 1$ can actually be realized by starting

1033: with periodic boundary conditions and adding small supersymmetric mass terms for bulk fields localized at the fixed points

1034: $(\pi R_1,0)$ and/or $(\pi R_1, \pi R_2)$. Supersymmetry is broken softly with $\Sigma \Lambda \ll 1$ only at the origin

1035: $(0,0)$. After re-diagonalization of the mass matrix, this is equivalent to starting with nontrivial

1036: boundary conditions (\ref{njl5}) and no localized mass terms. This is technically natural in the sense that a small supersymmetric

1037: mass term in the fixed points is protected by supersymmetry.   

1038:   

1039: 

1040: %%%%

1041: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1042: \section{String theory realization}

1043: 

1044: It is legitimate to ask if it is possible to realize the field theory construction we did provide in \cite{dpr} and

1045: in this paper from a string theory framework. The answer is positive and we provide here one possible construction for the 

1046: (softly broken) supersymmetric case of section 5\footnote{ Other realizations of 

1047: dynamical symmetry breaking  can be found in \cite{harvey} for string realizations of a nonlocal NJL version and 

1048: \cite{qcd} (\cite{alex})

1049: for a string (field-theory) realization of the chiral symmetry breaking in QCD.}. The main requirement for the string construction is to

1050: provide a large and flat two-dimensional compact space, so there should be no localized source which would curve the large 2d space. 

1051: Indeed, our field theory analysis was done in flat space. 

1052: This situation can be realized in orientifold constructions, where orientifold planes cancel the sources provided by the

1053: branes. The basic ingredients of the construction are that the field $\phi$ arises from D$5$ branes wrapping our large 2d space,

1054: whereas the localized fermions arise at the intersection between the D5 branes and a different set of branes, the intersection

1055: being four dimensional and generating chiral fermions.  

1056: We use orientifolds of type IIB string theory with D5-branes wrapping different coordinates of the internal space.

1057: Our example is based on the orientifold projection $\Pi' = \Pi \ \pi_4 \pi_5 \pi_8 \pi_9$, where $\Pi$ is the left-right world-sheet interchange,

1058: $\pi_4$, $\pi_5$ are parity  operations in two compact coordinates $x_4$ and $x_5$, to be identified with the 

1059: two large dimensions $(y_1,y_2)$ in our field theoretical construction and $\pi_8$, $\pi_9$ are parity  operations in two internal

1060: noncompact coordinates $x_8$ and $x_9$. The basic building block for brane configuration we consider is then the following

1061: \ba

1062: && {\rm coord.} \quad \ 0 \qquad 1 \qquad 2 \qquad 3 \qquad 4 \qquad  5 \qquad  6 \qquad  7 \qquad  8 \qquad  9 

1063: \qquad \ \nonumber \\

1064: && {\rm D}5_1 \  \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} 

1065: \qquad 0 \qquad 0 \qquad 0 \qquad 0 \nonumber \\

1066: && {\rm D}5_2  \  \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad 0 \qquad 0

1067: \qquad {\rm x}  \qquad {\rm x} \qquad 0 \qquad 0 \nonumber \\

1068: && {\rm O}5_2  \  \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad 0 \qquad 0

1069: \qquad {\rm x} \qquad {\rm x} \qquad 0 \qquad 0

1070: \  \label{st1}

1071: \ea  

1072: In (\ref{st1}), crosses ${\rm x}$ denote coordinates parallel to the branes, whereas $0$ denotes orthogonal coordinates.

1073: Notice that D$5_2$ branes and O$5_2$ planes are orthogonal to the 

1074: large 2d space $(x_4,x_5)$. In order to keep the 2d space flat we 

1075: need a configuration with D$5_2$ branes on top of the O$5_2$ planes, with locally zero tension and charge. 

1076: We add a Wilson line $\langle W_4 \rangle$ on the D$5_1$ branes in the compact $x_4$ coordinate, which 

1077: has the effect of breaking 

1078: the gauge group and giving masses to fields charged under $W_4$. There are in particular

1079: four charged scalar fields $\phi_i$, which get a mass from this Hosotani mechanism and will be identified with the master

1080: field(s) of our field theory model. Notice first of

1081: all that this field lives indeed in six dimensions, in the bulk of our large 2d compact space and it corresponds to a hypermultiplet

1082: $\phi = (\phi_1, \phi_2)$ from the 4d viewpoint, as in section 5. The mass of $\phi$ 

1083: is positive and it corresponds to the mass generated by boundary conditions

1084: analyzed in \cite{dpr} and in sections 2, 4 and 5 of the present paper. Alternatively, using a Wilson line in one of the last 

1085: four coordinates $x_6 \cdots x_9$ is equivalent to considering a bulk supersymmetric mass as in the section 3.  

1086: For simplicity, we can consider the last three coordinates $x_7 \cdots x_9$ as being noncompact,

1087: whereas $x_6$ is a circle and will be used to break supersymmetry a la Scherk-Schwarz. 

1088: 

1089: The D$5_2$ branes gauge fields are nondynamical in four dimensions and play the role of global symmetries.

1090: The D$5_1$ brane degrees of freedom, on the other hand, are dynamical and contain in particular gauge fields and the field(s) $\phi$.

1091: The D$5_1$-D$5_2$ sector, after additional orbifold projections to be discussed below, 

1092: contains massless ${\cal N}=1$ chiral multiplets localized in  four dimensions, to be identified with the 4d chiral 

1093: multiplets $A_i$ in section 5. 

1094: At the effective low energy action level , the setup is similar to the one considered in section 5, with couplings of the 

1095: form $A  \phi A $ and one expects the arguments presented there to apply

1096: and generate dynamical symmetry breaking.   

1097:   Non-trivial boundary conditions in the compact coordinate $x_6$ a la Scherk-Schwarz break supersymmetry at tree-level in the 

1098: D$5_2$ sector, whereas the $D5_1$ branes, being orthogonal to the $x_6 $ coordinate, feel the breaking only through radiative

1099: corrections \cite{ads1}. Notice that the dynamics in the large bulk coordinates $x_4,x_5$ is supersymmetric at that stage.

1100: At one-loop, supersymmetry breaking propagates in the D$5_1$- D$5_2$ sector and generates

1101: the localized 4d soft terms that were used in section 5. 

1102: 

1103: The setup preserve until now ${\cal N}=2$ supersymmetry in 4d spontaneously broken to ${\cal N}=0$ by the Scherk-Schwarz deformation, 

1104: so additional ingredients are needed in order to generate chirality. The standard internal spaces used in this respect are the

1105: Calabi-Yau spaces or the orbifolds. We choose here the second possibility. We introduce additional

1106: $Z_2$ and $Z_3$ orbifold operations acting on the internal coordinates as 

1107: \ba

1108: && Z_2 \ (z_1,z_2,z_3) \ = \  \ (- z_1,- z_2, z_3) \quad , \quad  Z_3 \ (z_1,z_2,z_3) \ = \  \ (e^{2 \pi i \over 3} z_1, z_2, 

1109: e^{-{2 \pi i \over 3}} z_3) \ , \nonumber \\

1110: && {\rm where} \qquad z_1 \ = \ {x_4 + i x_5 \over \sqrt{2}} \quad , \quad   z_2 \ = \ {x_6 + i x_7 \over \sqrt{2}} \quad ,

1111: \quad z_3 \ = \ {x_8 + i x_9 \over \sqrt{2}} \ \label{st2}

1112: \ea

1113: are the three complex internal coordinates. The resulting orientifold, which is dual to the so-called $Z_3 \times Z_2$ or 

1114: $Z_6'$ type I orbifold in the literature \cite{afiv} after performing four T-dualities in $x_6,x_7,x_8,x_9$, 

1115: reduces supersymmetry down to ${\cal N}=1$ in 4d. The 4d type I $Z_3 \times Z_2$ orbifold has D9 brane / O9 planes 

1116: and one set of  D5 branes / O5 planes, wrapping the third internal torus. After the four T-dualities, the D9 branes (O9 planes)

1117: become our D$5_1$ branes  (O$5_1$ planes) and the D5 (O5) branes become our  D$5_2$ branes  (O$5_2$ planes). Our Wilson line

1118: is in the type I orbifold a Wilson line on the D9 branes.  

1119: 

1120: In order to break completely supersymmetry, as already announced we are adding a Scherk-Schwarz operation in the compact

1121: coordinate $x_6$, compatible with the two orbifold operations. Our Scherk-Schwarz deformation

1122: is a $2 \pi$ rotation in $x_6$ and one 4d spacetime coordinate. The corresponding worldsheet current anticommutes with the

1123: $Z_2$ orbifold projection and commutes with the $Z_3$ one, as required by the consistency of the string construction \cite{ss}. 

1124: A last subtlety, explained in detail in \cite{ads1} is that due to the Scherk-Schwarz operation, the  O$5_1$ planes, which are

1125: perpendicular to the $x_6$ coordinate used in the supersymmetry breaking, are

1126: actually pairs of  O$5_1$ orientifold - ${\overline O5}_1$ antiorientifold planes, situated at $x_6=0$ and $x_6 = \pi R_6$, respectively.

1127: If the radius $R_6 \gg l_s$ is large enough, the closed string tachyon is massive and the timescale for the instability can 

1128: be large enough. The D$5_1$ branes should be at (or close to) the point $x_6=0$, such that the strings  D$5_1$ -${\bar O5}_1$,

1129: which break supersymmetry, to be very massive.  

1130: 

1131: A possible objection to the present setup is that the boundary conditions generated mass and/or the bulk hypermultiplet mass $m$ are 

1132: not constant but field-dependent, given by the vev of the Wilson 

1133: line on the D$5_1$  branes. Since the setup in non-supersymmetric, the Wilson line acquires a potential and its

1134: vev will be dynamically fixed. This last point needs further investigation which is however beyond the goals of the present paper. 

1135: This objection is also valid for the field-theory construction, in that we assumed that the radii of the large

1136: 2d space were stabilized by additional dynamics. 

1137: To conclude, the setup presented in this section does realize the (softly) supersymmetric model work out in section 5, where the

1138: supersymmetry breaking is soft and comes from a Scherk-Schwarz deformation in an extra space coordinate. 

1139: 

1140: The original non-superymmetric 

1141: setup of \cite{dpr} and in sections 2,3 of the present paper can be in principle also realized for smaller values of 

1142: $R_6 \gsim l_s$. However in this case

1143: we expect severe tachyonic instabilities in the system, which need to be suppressed in order for the string picture to be a viable

1144: description of the field-theory construction.     

1145: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1146: \vskip 0.5cm

1147: %%%%%%%%%%%%%%%%%%%%%%%%%%%

1148: 

1149: \mysection{Acknowledgements}: We thank T. Gherghetta, A. Pomarol and V.~Rubakov

1150: for useful discussions. E.D. thanks the Galileo Institute of Theoretical Physics and INFN

1151: for partial support during the completion of this work.

1152: Work partially supported by the CNRS PICS \#~2530

1153: and 3059, RTN contracts MRTN-CT-2004-005104 and MRTN-CT-2004-503369

1154: and the European Union Excellence Grant, MEXT-CT-2003-509661.

1155: 

1156: 

1157: %%%%%%%%%%%%%%%%%%%%%%%%%%%

1158: \begin{thebibliography}{99}

1159: %%%%%%%%%%%%%%%%%%%%%%%%%%%

1160: 

1161: \bibitem{6d}

1162: The literature on gravitational aspects of codimension two models is vast. Some references are

1163: A.~G.~Cohen and D.~B.~Kaplan,

1164:   %``Solving the hierarchy problem with noncompact extra dimensions,''

1165:   Phys.\ Lett.\ B {\bf 470} (1999) 52

1166:   [arXiv:hep-th/9910132];

1167:   %%CITATION = HEP-TH 9910132;%%

1168: T.~Gherghetta and M.~E.~Shaposhnikov,

1169:   %``Localizing gravity on a string-like defect in six dimensions,''

1170:   Phys.\ Rev.\ Lett.\  {\bf 85} (2000) 240

1171:   [arXiv:hep-th/0004014];

1172: I.~Navarro,

1173:   %``Codimension two compactifications and the cosmological constant  problem,''

1174:   JCAP {\bf 0309} (2003) 004

1175:   [arXiv:hep-th/0302129]; 

1176:   %%CITATION = HEP-TH 0302129;%%

1177: Y.~Aghababaie, C.~P.~Burgess, S.~L.~Parameswaran and F.~Quevedo,

1178:   Nucl.\ Phys.\ B {\bf 680} (2004) 389

1179:   [arXiv:hep-th/0304256];

1180:   %%CITATION = HEP-TH 0304256;%%

1181: G.~W.~Gibbons, R.~Guven and C.~N.~Pope,

1182:   %``3-branes and uniqueness of the Salam-Sezgin vacuum,''

1183:   Phys.\ Lett.\ B {\bf 595} (2004) 498

1184:   [arXiv:hep-th/0307238];

1185:   %%CITATION = HEP-TH 0307238;%%

1186: M.~Peloso, L.~Sorbo and G.~Tasinato

1187:   %%CITATION = HEP-TH 0004014;%%,

1188:   %``Standard 4d gravity on a brane in six dimensional flux compactifications,''

1189:   Phys.\ Rev.\ D {\bf 73} (2006) 104025

1190:   [arXiv:hep-th/0603026].

1191:   %%CITATION = HEP-TH 0603026;%%

1192: 

1193: \bibitem{dpr}

1194:   E.~Dudas, C.~Papineau and V.~A.~Rubakov,

1195:   %``Flowing to four dimensions,''

1196:   JHEP {\bf 0603} (2006) 085

1197:   [arXiv:hep-th/0512276].

1198:   %%CITATION = HEP-TH 0512276;%%

1199:   

1200:   

1201: \bibitem{njl}

1202:   Y.~Nambu and G.~Jona-Lasinio,

1203:   %``Dynamical Model Of Elementary Particles Based On An Analogy With

1204:   %Superconductivity. I,''

1205:   Phys.\ Rev.\  {\bf 122} (1961) 345.

1206:   %%CITATION = PHRVA,122,345;%%

1207:  

1208:   

1209:  \bibitem{nambu}

1210:  Y.~Nambu,

1211:   %``Quasisupersymmetry, Bootstrap Symmetry Breaking And Fermion Masses,''

1212: EFI-88-62-CHICAGO

1213: {\it Invited talk to appear in Proc. of 1988 Int. Workshop  New Trends in Strong Coupling 

1214: Gauge Theories, Nagoya, Japan, Aug 24-27, 1988};

1215: V.~A.~Miransky, M.~Tanabashi and K.~Yamawaki,

1216:   %``Dynamical Electroweak Symmetry Breaking With Large Anomalous Dimension And

1217:   %T Quark Condensate,''

1218:   Phys.\ Lett.\ B {\bf 221} (1989) 177.

1219:   %%CITATION = PHLTA,B221,177;%%

1220: 

1221: \bibitem{bhl}

1222:   W.~A.~Bardeen, C.~T.~Hill and M.~Lindner,

1223:   %``Minimal Dynamical Symmetry Breaking Of The Standard Model,''

1224:   Phys.\ Rev.\ D {\bf 41} (1990) 1647.

1225:   %%CITATION = PHRVA,D41,1647;%%

1226: 

1227: 

1228: \bibitem{gw}

1229:   W.~D.~Goldberger and M.~B.~Wise,

1230:   %``Renormalization group flows for brane couplings,''

1231:   Phys.\ Rev.\ D {\bf 65} (2002) 025011

1232:   [arXiv:hep-th/0104170]; 

1233:   %%CITATION = HEP-TH 0104170;%%

1234: K.~A.~Milton, S.~D.~Odintsov and S.~Zerbini,

1235:   %``Bulk versus brane running couplings,''

1236:   Phys.\ Rev.\ D {\bf 65} (2002) 065012

1237:   [arXiv:hep-th/0110051];

1238:   %%CITATION = HEP-TH 0110051;%%

1239: F.~del Aguila, M.~Perez-Victoria and J.~Santiago,

1240:   %``Bulk fields with general brane kinetic terms,''

1241:   JHEP {\bf 0302} (2003) 051

1242:   [arXiv:hep-th/0302023].  

1243:   %%CITATION = HEP-TH 0302023;%%

1244: 

1245: \bibitem{dgv}

1246:   E.~Dudas, C.~Grojean and S.~K.~Vempati,

1247:   %``Classical running of neutrino masses from six dimensions,''

1248:   arXiv:hep-ph/0511001.

1249:   %%CITATION = HEP-PH 0511001;%%

1250: 

1251:   

1252: \bibitem{ddg}

1253:  K.~R.~Dienes, E.~Dudas and T.~Gherghetta,

1254:   %``Invisible axions and large-radius compactifications,''

1255:   Phys.\ Rev.\ D {\bf 62} (2000) 105023

1256:   [arXiv:hep-ph/9912455].

1257:   %%CITATION = HEP-PH 9912455;%%

1258:   

1259:   

1260:   

1261: \bibitem{bfz}

1262: J.~Bagger, F.~Feruglio and F.~Zwirner,

1263:   %``Brane induced supersymmetry breaking,''

1264:   JHEP {\bf 0202} (2002) 010

1265:   [arXiv:hep-th/0108010].

1266:   %%CITATION = HEP-TH 0108010;%%

1267: 

1268:  

1269: \bibitem{cw}

1270:   S.~R.~Coleman and E.~Weinberg,

1271:   %``Radiative Corrections As The Origin Of Spontaneous Symmetry Breaking,''

1272:   Phys.\ Rev.\ D {\bf 7} (1973) 1888.

1273:   %%CITATION = PHRVA,D7,1888;%%

1274:   

1275:   

1276:   \bibitem{dobrescu}

1277:   B.~A.~Dobrescu,

1278:   Phys.\ Lett.\ B {\bf 461} (1999) 99

1279:   [arXiv:hep-ph/9812349];

1280:   %%CITATION = HEP-PH 9812349;%%

1281: N.~Arkani-Hamed, H.~C.~Cheng, B.~A.~Dobrescu and L.~J.~Hall,

1282:   %``Self-breaking of the standard model gauge symmetry,''

1283:   Phys.\ Rev.\ D {\bf 62} (2000) 096006

1284:   [arXiv:hep-ph/0006238];

1285:   %%CITATION = HEP-PH 0006238;%%

1286: M.~Hashimoto, M.~Tanabashi and K.~Yamawaki,

1287:   Phys.\ Rev.\ D {\bf 64} (2001) 056003

1288:   [arXiv:hep-ph/0010260].

1289:   %%CITATION = HEP-PH 0010260;%%

1290: V.~Gusynin, M.~Hashimoto, M.~Tanabashi and K.~Yamawaki,

1291:   Phys.\ Rev.\ D {\bf 65} (2002) 116008

1292:   [arXiv:hep-ph/0201106].

1293:   %%CITATION = HEP-PH 0201106;%%

1294:   

1295: 

1296:   

1297: \bibitem{agw}

1298: N.~Marcus, A.~Sagnotti and W.~Siegel,

1299:   %``Ten-Dimensional Supersymmetric Yang-Mills Theory In Terms Of

1300:   %Four-Dimensional Superfields,''

1301:   Nucl.\ Phys.\ B {\bf 224} (1983) 159;

1302:   %%CITATION = NUPHA,B224,159;%%

1303:   N.~Arkani-Hamed, T.~Gregoire and J.~G.~Wacker,

1304:   %``Higher dimensional supersymmetry in 4D superspace,''

1305:   JHEP {\bf 0203} (2002) 055

1306:   [arXiv:hep-th/0101233].

1307:   %%CITATION = HEP-TH 0101233;%%

1308:   

1309:   

1310:   

1311: \bibitem{mp}

1312:   E.~A.~Mirabelli and M.~E.~Peskin,

1313:   %``Transmission of supersymmetry breaking from a 4-dimensional boundary,''

1314:   Phys.\ Rev.\ D {\bf 58} (1998) 065002

1315:   [arXiv:hep-th/9712214];

1316: E.~Dudas, T.~Gherghetta and S.~Groot Nibbelink,

1317:   %``Vector / tensor duality in the five dimensional supersymmetric

1318:   %Green-Schwarz mechanism,''

1319:   Phys.\ Rev.\ D {\bf 70} (2004) 086012

1320:   [arXiv:hep-th/0404094].

1321:   %%CITATION = HEP-TH 0404094;%

1322:   

1323:   

1324: 

1325:   \bibitem{njlsusy1}

1326:   W.~Buchmuller and S.~T.~Love,

1327:   %``Chiral Symmetry And Supersymmetry In The Nambu-Jona-Lasinio Model,''

1328:   Nucl.\ Phys.\ B {\bf 204} (1982) 213;

1329:   %%CITATION = NUPHA,B204,213;%%

1330:  W.~Buchmuller and U.~Ellwanger,

1331:   %``On The Structure Of Composite Goldstone Supermultiplets,''

1332:   Nucl.\ Phys.\ B {\bf 245} (1984) 237.

1333:   %%CITATION = NUPHA,B245,237;%%

1334: 

1335: \bibitem{njlsusy2}

1336: T.~E.~Clark, S.~T.~Love and W.~A.~Bardeen,

1337:   %``The Top Quark Mass In A Supersymmetric Standard Model With Dynamical

1338:   %Symmetry Breaking,''

1339:   Phys.\ Lett.\ B {\bf 237} (1990) 235;

1340:   %%CITATION = PHLTA,B237,235;%

1341:  M.~Carena, T.~E.~Clark, C.~E.~M.~Wagner, W.~A.~Bardeen and K.~Sasaki,

1342:   %``Dynamical symmetry breaking and the top quark mass in the minimal

1343:   %supersymmetric standard model,''

1344:   Nucl.\ Phys.\ B {\bf 369} (1992) 33;

1345:   %%CITATION = NUPHA,B369,33;%

1346: P.~Binetruy, E.~A.~Dudas and F.~Pillon,

1347:   %``The Vacuum structure in a supersymmetric gauged Nambu-Jona-Lasinio model,''

1348:   Nucl.\ Phys.\ B {\bf 415} (1994) 175

1349:   [arXiv:hep-ph/9304278].

1350:   %%CITATION = HEP-PH 9304278;%%

1351: 

1352: 

1353:   

1354: \bibitem{harvey}

1355: E.~Antonyan, J.~A.~Harvey, S.~Jensen and D.~Kutasov,

1356:   %``NJL and QCD from string theory,''

1357:   arXiv:hep-th/0604017;

1358:   %%CITATION = HEP-TH 0604017;%%

1359: A.~Parnachev and D.~A.~Sahakyan,

1360:   %``Chiral phase transition from string theory,''

1361:   arXiv:hep-th/0604173.

1362:   %%CITATION = HEP-TH 0604173;%%

1363:   

1364:   

1365:   

1366: \bibitem{qcd}

1367:  M.~Kruczenski, D.~Mateos, R.~C.~Myers and D.~J.~Winters,

1368:   %``Towards a holographic dual of large-N(c) QCD,''

1369:   JHEP {\bf 0405} (2004) 041

1370:   [arXiv:hep-th/0311270];

1371:   %%CITATION = HEP-TH 0311270;%%

1372: T.~Sakai and S.~Sugimoto,

1373:   %``Low energy hadron physics in holographic QCD,''

1374:   Prog.\ Theor.\ Phys.\  {\bf 113} (2005) 843

1375:   [arXiv:hep-th/0412141].

1376:   %%CITATION = HEP-TH 0412141;%%

1377: 

1378: \bibitem{alex}

1379: J.~Erlich, E.~Katz, D.~T.~Son and M.~A.~Stephanov,

1380:   %``QCD and a holographic model of hadrons,''

1381:   Phys.\ Rev.\ Lett.\  {\bf 95} (2005) 261602

1382:   [arXiv:hep-ph/0501128];

1383:   %%CITATION = HEP-PH 0501128;%%

1384: L.~Da Rold and A.~Pomarol,

1385:   %``Chiral symmetry breaking from five dimensional spaces,''

1386:   Nucl.\ Phys.\ B {\bf 721} (2005) 79

1387:   [arXiv:hep-ph/0501218].

1388:   %%CITATION = HEP-PH 0501218;%%

1389:   

1390:   

1391: 

1392:   

1393: \bibitem{ads1}

1394: I.~Antoniadis, E.~Dudas and A.~Sagnotti,

1395: %``Supersymmetry breaking, open strings and M-theory,''

1396: Nucl.\ Phys.\ B {\bf 544} (1999) 469 [arXiv:hep-th/9807011];

1397: %%CITATION = HEP-TH 9807011;%%

1398: I.~Antoniadis, G.~D'Appollonio, E.~Dudas and A.~Sagnotti,

1399: %``Partial breaking of supersymmetry, open strings and M-theory,''

1400: Nucl.\ Phys.\ B {\bf 553} (1999) 133 [arXiv:hep-th/9812118].

1401: 

1402: 

1403: 

1404: \bibitem{afiv}

1405: G.~Zwart,

1406:   %``Four-dimensional N = 1 Z(N) x Z(M) orientifolds,''

1407:   Nucl.\ Phys.\ B {\bf 526} (1998) 378

1408:   [arXiv:hep-th/9708040];

1409:   %%CITATION = HEP-TH 9708040;%%

1410: G.~Aldazabal, A.~Font, L.~E.~Ibanez and G.~Violero,

1411:   %``D = 4, N = 1, type IIB orientifolds,''

1412:   Nucl.\ Phys.\ B {\bf 536} (1998) 29

1413:   [arXiv:hep-th/9804026].

1414:   %%CITATION = HEP-TH 9804026;%%

1415:   

1416:   

1417:   \bibitem{ss}

1418: R.~Rohm,

1419: %``Spontaneous Supersymmetry Breaking In Supersymmetric String Theories,''

1420: Nucl.\ Phys.\ B {\bf 237} (1984) 553;

1421: %%CITATION = NUPHA,B237,553;%%

1422: S.~Ferrara, C.~Kounnas, M.~Porrati and F.~Zwirner,

1423: %``Superstrings With Spontaneously Broken Supersymmetry And Their Effective

1424: %Theories,''

1425: Nucl.\ Phys.\ B {\bf 318} (1989) 75.

1426: 

1427: 

1428: 

1429:   

1430: \bibitem{hills}

1431:   C.~T.~Hill and E.~H.~Simmons,

1432:   %``Strong dynamics and electroweak symmetry breaking,''

1433:   Phys.\ Rept.\  {\bf 381} (2003) 235

1434:   [Erratum-ibid.\  {\bf 390} (2004) 553]

1435:   [arXiv:hep-ph/0203079].

1436:   %%CITATION = HEP-PH 0203079;%%

1437: 

1438: 

1439: 

1440: 

1441: 

1442: 

1443: 

1444: 

1445: 

1446: 

1447: 

1448: 

1449: 

1450: %\bibitem{bachas} C.~P.~Bachas,

1451: %  JHEP {\bf 9811} (1998) 023

1452: %  [arXiv:hep-ph/9807415];

1453: %  %%CITATION = HEP-PH 9807415;%%

1454: %N.~Arkani-Hamed, S.~Dimopoulos and J.~March-Russell,

1455: %  arXiv:hep-th/9908146;

1456: %  %%CITATION = HEP-TH 9908146;%%

1457: %I.~Antoniadis and C.~Bachas,

1458: %  %``Branes and the gauge hierarchy,''

1459: %  Phys.\ Lett.\ B {\bf 450} (1999) 83

1460: %  [arXiv:hep-th/9812093].

1461: %  %%CITATION = HEP-TH 9812093;%%

1462: 

1463: %\bibitem{kazakov}

1464: %D.~I.~Kazakov,

1465: %  JHEP {\bf 0303} (2003) 020

1466: %  [arXiv:hep-th/0209100];

1467: %  %%CITATION = HEP-TH 0209100;%%

1468: %K.~R.~Dienes, E.~Dudas and T.~Gherghetta,

1469: %  Phys.\ Rev.\ Lett.\  {\bf 91} (2003) 061601

1470: %  [arXiv:hep-th/0210294].

1471: %  %%CITATION = HEP-TH 0210294;%%

1472: 

1473: 

1474: 

1475: 

1476: 

1477: 

1478: 

1479: 

1480: 

1481: 

1482:   

1483: 

1484: 

1485: \end{thebibliography}

1486: 

1487: \end{document}

1488: 

1489: 

1490: 

1491: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1492: \section{Phenomenological applications}

1493: 

1494: We presented evidence that the model we had exhibit in \cite{dpr} and

1495: in the present paper is equivalent, in the IR region below the

1496: compositeness scale $\Lambda_c = 1/R$, with the NJL model in

1497: the large N expansion, at least in the large $N$ approximation.  

1498: The would-be zero-mode of the 6d scalar field is the bound state of

1499: the NJL model. Very close to the point of the phase transition, the

1500: tachyonic mass term becomes large at the compactification

1501: (compositeness) scale, which is equivalent, after rescaling of the

1502: scalar field, to the compositeness condition proposed in \cite{bhl}.    

1503: At energies $\mu > 1/R$, the zero mode becomes part of the

1504: Kaluza-Klein tower of the scalar field, interacting with the localized

1505: 4d fermions.  KK states lighter than $\mu$ become part of the

1506: low-energy theory, whereas the ones with masses $ \mu < m_{KK} < \Lambda $    

1507: are heavy but generate a (running) four-fermion interaction 

1508: $G \sim \ln (\Lambda / \mu)$. Our construction gives therefore a

1509: rationale to the compositeness condition of the NJL model.    

1510: 

1511: If we add the (non-supersymmetric) Standard Model fields at the same 

1512: fixed point where the

1513: induced scalar potential is generated and if the scalar is identified

1514: with the Higgs field, at energies below $1/R$ the model has precisely

1515: the same predictions as the NJL top condensation model \cite{nambu,bhl}. In particular, the top Yukawa coupling becomes

1516: nonperturbative at the compositeness scale $h_t (\Lambda_c = R^{-1}) \rightarrow \infty$. 

1517: As a result, the top mass is probably too heavy and the model is ruled out \cite{bhl}. Various

1518: model building avenues are possible, along the lines of the 4d

1519: constructions summarized in \cite{hills}. In the 6d picture, however,

1520: it is clear that bulk supersymmetry, softly broken on the brane, is a crucial ingredient 

1521: in the construction. A natural generalisation, therefore, is to add

1522: MSSM in 6d, with a softly broken SUSY localized at the origin of the

1523: 2d compact space. Below the compositeness scale $1/R$, the theory

1524: becomes equivalent to the supersymmetric generalisation of the 

1525: top condensation NJL model \cite{njlsusy1,njlsusy2}. This is a viable possibility, with $\omega /R \ll 1/R \equiv \Lambda_c \ll \Lambda$,

1526: since the top quark mass in the supersymmetric top condensation scenario is compatible with the experimentally measured value

1527: for compositeness scale sufficiently high, in particular for intermediate values $R^{-1} \sim 10^{10}-10^{13} GeV$.  

1528: The caveat in this case, already mentioned, is that $\omega \ll 1$ is needed in order to generate the electroweak scale.         

1529: We stress again, however, that supersymmetry makes such small values stable under radiative corrections. 

1530:    

1531: A different possibility is that the bulk scalar fields $\Phi_i$ are singlets under the SM gauge group, whereas the full spectrum

1532: of MSSM fields live at $(y_1,y_2)$, close (but not exactly

1533: at ) to the fixed point $(\pi R_1,0)$, whereas SUSY is softly broken at the origin $(0,0)$. The supersymmetric localized interaction

1534: {\bf below $\omega = 1/2$}

1535: \be

1536: \int d^2 \ \theta \ \lambda' \ \Phi_1 (y_1,y_2) \ H_1 \  H_2 \ \label{pa1}

1537: \ee   

1538: generate a $\mu$ term of MSSM equal to\footnote{We denote the $\mu$-term of MSSM by $m$ in what follows, in order to distinguish from

1539: the Higgs localized mass parameter $\mu$ in (\ref{pt1}). We also use for simplicity $R_1 = R_2 = R$}

1540: \be

1541: m \ = \ \lambda' \ \Phi_{1,c} (y_1,y_2)  \ \sim \ \lambda' { y_1-\pi R \over R} \phi_0 \ , \label{pa2}

1542: \ee

1543: for $y_1$ close to $\pi R_1$. Soft terms for $H_1,H_2$ are induced at one-loop from the origin of the compact 

1544: two-dimensional space

1545: to the point where the MSSM fields live, transmitted by the bulk fields $\Phi_i$. The size of the soft terms can be 

1546: estimated along the lines

1547: of \cite{ahhnw}. Due to the spatial separation, the result is UV finite and of the order 

1548: \be

1549: m_{soft}^2 \ \sim \ { (y_1-\pi R)^2 \over R^2} \times {\lambda'^2 \over R^2} \ , \label{pa3}

1550: \ee

1551: allowing electroweak symmetry breaking. A more detailed analysis of the resulting phenomenology would be interesting.  

1552: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1553: 

1554: \section{String theory realization}

1555: 

1556: It is legitimate to ask if it is possible to realize the field theory construction we provide in \cite{dpr} and

1557: in this paper in this theory. The answer is positive and we provide here one possible construction for both the

1558: nonsupersymmetric and the (softly broken) supersymmetric case. The main requirement for the string construction is to

1559: provide a large and flat two-dimensional compact space, so there should be no source in the large 2d space. Indeed, our

1560: field theory analysis was done in flat space. 

1561: This situation can be realized in orientifold constructions, where orientifold planes cancel the sources provided by the

1562: branes.  We use type IIB orientifolds of string theory with D-branes of various dimensionalities.

1563: Let us start with the non-supersymmetric example. It is an orientifold of type IIB string theory based on

1564: the orientifold projection $\Pi' = \Pi \ \pi_4 \pi_5 \ (-1)^{F_L}$, where $\Pi$ is the left-right world-sheet interchange,

1565: $\pi_4$, $\pi_5$ are parity  operations in two compact coordinates $x_4$ and $x_5$, to be identified with the 

1566: two large dimensions in our field theoretical construction, and  $(-1)^{F_L}$ is the spacetime fermion number carried 

1567: by the left movers. The brane configuration we consider is then the following

1568: \ba

1569: && \quad \ \qquad 0 \qquad 1 \qquad 2 \qquad 3 \qquad 4 \qquad  5 \qquad  6 \qquad  7 \qquad  8 \qquad  9 

1570: \qquad \ \nonumber \\

1571: && D5 \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} 

1572: \qquad 0 \qquad 0 \qquad 0 \qquad 0 \nonumber \\

1573: && {\overline D5} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x}

1574: \qquad 0 \qquad 0 \qquad 0 \qquad 0 \nonumber \\

1575: && D7 \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad 0 \qquad 0

1576: \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \nonumber \\

1577: && O7 \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad 0 \qquad 0

1578: \qquad {\rm x} \qquad {\rm x} \qquad {\rm x} \qquad {\rm x}

1579: \  \label{st1}

1580: \ea  

1581: In (\ref{st1}) crosses ${\rm x}$ denote coordinates parallel to the branes, whereas $0$ denotes orthogonal coordinates.

1582: Notice that D7 branes and O7 planes are orthogonal to the large 2d space and in order to keep the 2d space flat we 

1583: need a configuration with D7 planes on top of the O7 planes, with locally zero tension and charge. 

1584: We separate the D5 branes and D${\bar 5}$ antibranes by a distance $d$ in one of the last four coordinates, say $x_9$. The lightest

1585: open string excitation stretched between them is the field $\phi$ which generated the symmetry breaking. Notice first of

1586: all that this field lives indeed in six dimensions, in the bulk of out large 2d compact space. If the distance

1587: $d$ is larger than the string length $l_s$, the mass of $\phi$ is positive and it corresponds to the bulk positive mass

1588: analyzed in section 3, whereas for $d < l_s$ the bulk $\phi$ mass is tachyonic due to the change of the GSO projection in the

1589: D5-D${\bar 5}$ sector. For simplicity, we can consider the last four coordinates $x_6 \cdots x_9$ as being noncompact,

1590: therefore the D7 branes gauge field is nondynamical in four dimensions and plays the role of a global symmetry.

1591: The D5 brane degrees of freedom, on the other hand, are dynamical and contain in particular gauge fields (containing

1592: the electroweak gauge fields in the top condensation version of the NJL model) and the field $\phi$. 

1593: The D5-D7 sector contains massless Weyl fermions $\psi_L $from the R sector localized in four dimensions, to be identified 

1594: with the 4d fermions in section 4.1, whereas the NS sector contains only massive states. The D${\bar 5}$-D7 sector 

1595: contains massless Weyl fermions of opposite chirality $\psi_R$. The field $\phi$ has couplings of the form 

1596: $\phi {\bar \psi}_L \psi_R + {\rm h.c.}$, which are localized in four dimensions. At the effective low energy action level , 

1597: the setup is very similar to the one considered in section 4 and one expects the arguments presented there to apply

1598: and generate dynamical symmetry breaking.  

1599: There are a couple of points which would require further clarification for the string embedding above to be completely satisfactory.

1600: First of all, the fermions $\psi_L$ and  $\psi_R$ live in different points in $x_9$, so the interaction $$\phi {\bar \psi}_L \psi_R$

1601: is nonlocal. However, from the 6d perspective of section 4 the interaction is local.    

1602: Secondly, due to the orientifold projection (more precisely, the D5-O7 amplitude) the D5 open sector contains complex tachyons in the

1603: symmetric representation of the gauge group.    

1604: 

1605: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1606: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1607: \section{NJL model and holography}

1608: 

1609: Having giving arguments fora  dual description of a 4d NJL model in terms of a higher-dimensional scalar theory with non-trivial boundary conditions,

1610: it is interesting to ask if there is a way to realize this setup within the AdS/CFT correspondence \cite{malda} and gain new insights from holography.

1611: The answer is positive and we will try to present more evidence in this section. Our starting observation for this section is that

1612: a setup with a bulk scalar mass and an appropriate localized scalar potential realize holographically the NJL model, under certain conditions

1613: to be explained later on.  

1614: The setup is the RS1 model \cite{rs1}, with a compact extra dimension

1615: $ 0 <y < \pi R$, with the UV brane at $y=0$ and the IR brane at $y = \pi R$,  

1616: and its holographic interpretation \cite{tonyreview}, supplemented with a (massive) bulk scalar field and and a localized scalar potentiel. 

1617: The action is given by

1618: \be

1619: S \ = \ \int d^4 x \ dy \ \sqrt{g} \left[ {1 \over 2} M_5^3 R \ - \ \Lambda + {1 \over 2} (\partial \phi)^2 - {m^2 \over 2} \phi^2

1620: \ - \ \delta (y) (T_0 + V (\phi)) \  -\  \delta (y- \pi R) T_1   \right] \ , \label{hol1}

1621: \ee   

1622: where $\Lambda$ is the bulk cosmological constant, the brane tensions $T_0$, $T_1$ are tuned as in RS1 in order to get a flat 4d space and 

1623: \be

1624:  V (\phi) \ = \ - \ {\mu^2 \over 2} \phi^2 \ + \ {\lambda \over 4} \phi^4 \ . \label{hol2}

1625: \ee 

1626: The background metric in the unbroken vacuum is the $AdS_5$ metric

1627: \be

1628: d s^2 \ = \ e^{- 2 k y} \ \eta_{\mu \nu} dx^{\mu}  dx^{\nu} \ - \ d y^2 \ , \label{hol3}

1629: \ee 

1630: where $k$ is the $AdS_5$ radius. Defining the 4d mass as usual as $ \eta_{\mu \nu} dx^{\mu}  dx^{\nu} \phi = - M^2 \phi$, the solution to the

1631: scalar field equation in the bulk is

1632: \be

1633: \phi (y) \ = \ e^{2 k y} \ \left[ c_1 \ J_{\alpha} ({M \over k} e^{ky}) \ + \  c_2 \ N_{\alpha} ({M \over k} e^{ky})   \right] \ , \label{hol4}

1634: \ee 

1635: where $c_1,c_2$ are integration constants and

1636: \be

1637: \alpha \ = \ \sqrt{ 4 + {m^2 \over k^2}  } \ . \label{hol5}

1638: \ee

1639: In the unbroken phase, the boundary conditions to be imposed on the solution (\ref{hol4}) are

1640: \ba

1641: && (\partial_y  \phi + \mu^2 \phi)|_{y=0} \ = \ 0 \ , \nonumber \\

1642: && \partial_y  \phi |_{y=\pi R} \ = \ 0 \ . \label{hol6}

1643: \ea

1644: In the unbroken phase $\langle \phi \rangle = 0$, the field $\phi$ has a zero mode $M=0$ if

1645: \be

1646: \mu^2 \ = \ (\alpha - 2 ) \ k \ , \label{hol7} 

1647: \ee 

1648: in which case (\ref{hol4}) reduces to

1649: \be

1650: \phi (y) \ \sim \ c_1 \ e^{(2 + \alpha) k y} \ + \   c_2 \ e^{(2 - \alpha) k y} \ . \label{hol8}

1651: \ee

1652: >From the AdS/CFT dictionary, the coefficients $c_i$ are connected with the condensate $\langle {\cal O} \rangle$ and the UV value of $\lambda$,

1653: respectively

1654: \be

1655: c_1 \ \sim \ \langle {\cal O} \rangle \quad , \quad c_2 \sim \lambda \ . \ \label{hol08}

1656: \ee  

1657: As we will check explicitly in a moment, with our boundary conditions the second term is always the dominant one, such that in a first

1658: approximation $\phi (y) \sim  \exp ({(2 - \alpha) k y})$. In this case, the wave function of the zero mode $\Phi (x,y) = \phi_0 (y) \phi (x)$, 

1659: defined with respect to the 4d Minkowski metric, is given by

1660: \ba

1661: && \sqrt{g} \ g^{\mu \nu} \partial_{\mu} \Phi  \partial_{\nu} \Phi \ \equiv \ \phi_0^2 (y) \ \eta^{\mu \nu} \partial_{\mu} \phi (x)  

1662: \partial_{\nu} \phi (x) \ , \nonumber \\

1663: &&  \phi_0 (y) \ \sim e^{(1-\alpha) k y} \ . \label{hol9}

1664: \ea

1665: >From the profile of $\phi_0$ we see the zero mode is localized on the UV brane for $\alpha >1$, whereas it is localized on the

1666: IR brane for $\alpha <1$. On the other hand, by the AdS/CFT correspondence \cite{malda,tonyreview} the field $\phi$ couple to the

1667: an operator ${\cal O}$ from the CFT sector via

1668: \be

1669: \lambda \ \phi (x,y=0) \ {\cal O} \quad , \quad {\rm dim} \ {\cal O} \ = \ 2 + \alpha  \ . \label{hol10} 

1670: \ee    

1671: In the NJL model, the composite field couples to the fermionic current $ {\cal O} = J_5 = {\bar \psi} (1-\gamma_5) \psi$, which has dimension three.

1672: Like in the previous sections, we want to identify the would-be zero mode of the bulk scalar with the NJL composite

1673: state and therefore we

1674: take $\alpha = 1$ in what follows. Consequently, the zero-mode $\phi$ has a flat profile. We interpret then the condition defining 

1675: the massless 4d mode in this case 

1676: \be

1677: m_c^2 \ = \ -3 \ k^2 \quad , \quad \mu_c^2 \ = \ - \ k \label{hol11}

1678: \ee   

1679: as the critical point of the phase transition. In order to provide more evidence for this interpretation, we will show that :

1680: 

1681: i) when $m^2 \ \gsim \ m_c^2$, i.e. $\alpha \ \gsim \ 1$, the 4d mass$^2$ is positive $M^2 > 0$ and the would-be zero mode 

1682: is localized towards the UV brane. This is the symmetric phase (no condensate) and the localization towards the UV brane is consistent with the fact that there should be no  bound state. Since ${\rm dim} \ {\cal O} > 3$ in this case, the coupling $\lambda$ 

1683: in (\ref{hol10}) becomes irrelevant. 

1684: 

1685: ii)  when $m^2 \ \lsim \ m_c^2$, i.e. $\alpha \ \lsim \ 1$, the 4d mass$^2$ is negative $M^2 < 0$ and the would-be zero mode is 

1686: localized towards the IR brane. This is the symmetry breaking phase where the condensate should form. The localization towards 

1687: the IR brane is consistent with the 

1688: fact that a  bound state appears at the phase transition.  Since ${\rm dim} \ {\cal O} < 3$ in this case, the coupling 

1689: $\lambda$ in (\ref{hol10}) becomes relevant above the critical point.

1690: 

1691: In order to support the claims i) and ii), we first compute the wave function of the would-be zero mode for $\alpha$ slighly larger than one,

1692: the 4d mass and its localisation. By expanding (\ref{hol4}) for $\alpha \gsim 1$ and keeping the leading terms for masses

1693: $M \ll  k \exp (-k \pi R)$, we find

1694: \be

1695: \phi (y) \ = \ c \ \left[  e^{(2-\alpha) k y} \ + \ {1 \over \alpha-1} \left({M \over 2 k}\right)^2 \ 

1696: e^{(4-\alpha) k y} \right] \ , \label{hol12}

1697: \ee

1698: whereas the second boundary condition in (\ref{hol6}) gives the 4d mass to be given by

1699: \be

1700: M^2 \ \simeq \ 4 k^2 \ (\alpha-1) \ > \ 0 \ . \label{hol13}

1701: \ee  

1702: According to (\ref{hol08}), the condensate is indeed zero in this phase.

1703: On the other hand, for $\alpha$ slightly smaller than one, keeping again the leading terms in the expansion in $\alpha$ we find

1704: \be

1705: \phi (y) \ = \ c' \ \left[ e^{(2-\alpha) k y} \ - \ {(4-\alpha) \over (\alpha-1)(\alpha+2)} \ e^{(2+\alpha) k y} \ 

1706: \left({M \over 2k}\right)^2 \ e^{2 (1-\alpha)k \pi R }  \right] \ . \label{hol14} 

1707: \ee

1708: The 4d mass in this case given again by the boundary conditions (\ref{hol6}) is

1709: \be

1710: M^2 \ = \ - 4 k^2 \ {(1-\alpha)^2 (\alpha+2) \over (4-\alpha)(1+\alpha)} \ e^{- 2 (1-\alpha) k \pi R} \ < \ 0 \ . \label{hol15}

1711: \ee 

1712: According to (\ref{hol08}), the condensate is nonzero in this phase $\langle {\cal O} \rangle \sim 1 - \alpha$ and vanishes precisely at

1713: the phase transition.

1714: 

1715: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1716: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1717: 

1718: \section{Phenomenological considerations}

1719: 

1720: We did present evidence that the higher-dimensional model put forward in \cite{dpr} and

1721: in the present paper is equivalent, in the IR region below the

1722: compactification scale $1/R$, with the NJL model in

1723: the large N expansion.  The would-be zero-mode of the 6d scalar field is the bound state of

1724: the NJL model. Below the point of the phase transition, the

1725: tachyonic mass term becomes large at the compactification

1726: (compositeness) scale, which is similar, after rescaling of the

1727: scalar field, to the compositeness condition proposed in \cite{bhl}. The 6d physics is however perturbative and also the reduction to 4d.     

1728: At energies $Q > 1/R$, the zero mode becomes part of the

1729: Kaluza-Klein tower of the scalar field, interacting with the localized

1730: 4d fermions.  KK states lighter than $Q$ become part of the

1731: low-energy theory, whereas the ones with masses $ Q < m_{KK} < \Lambda $    

1732: are heavy, but generate a four-fermion interaction 

1733: $G \sim \ln (\Lambda / Q)$, which runs above $R^{-1}$. Our construction gives therefore a

1734: rationale to the compositeness condition of the NJL model.    

1735: 

1736: If we add the (non-supersymmetric) Standard Model fields at the same 

1737: fixed point where the

1738: induced scalar potential is generated and if the scalar is identified

1739: with the Higgs field, at energies below $1/R$ and by using the compositentess conditions, the model would have

1740: similar predictions as the NJL top condensation model \cite{nambu,bhl}. Close to the phase transition, however, the

1741: running interpretation is not reliable, whereas the 6d analysis is perturbative. For large radius and if we run up to the

1742: 6d fundamental scale, we get small 4d couplings $\lambda_4, g_4$. The main obstacle in applying our scenario to the top condensation

1743: version of the NJL model \cite{nambu,bhl} is that if the the electroweak gauge fields and the Higgs live in 6d, the large volume

1744: will generate too small gauge couplings and top Yukawa in 4d. One possibility is to argue that the two radii $R$ are actually small such that

1745: 4d couplings are not suppressed,

1746: but the running continues well above the 6d fundamental scale \cite{bachas} by producing the large logs we need for the top condensation

1747: scenario. An alternative is to use the higher-dimensional UV fixed points generated by bulk interactions, invoked in \cite{kazakov}, 

1748: in order to raise the 4d couplings in the infrared.      

1749: In these cases, by using the improved 

1750: compositeness conditions and the Standard Model RGE's below $\Lambda_c$ the predictions become the same as in the NJL model. 

1751: In particular, the top Yukawa coupling becomes nonperturbative at the compositeness scale $h_t (\Lambda_c = R^{-1}) 

1752: \rightarrow \infty$. 

1753: As a result, the top mass is probably too heavy and the model is ruled out \cite{bhl}. Various

1754: standard model building avenues are possible, along the lines of the 4d

1755: constructions summarized in \cite{hills}. A qualitatively new possibility arises if we consider 

1756: the case where  the compositeness scale is above the compactification scale $\Lambda_c > 1/R$ and therefore

1757: the runnings above the compactification scale becomes higher dimensional. This possibility could be realized

1758: close to the  nonperturbative

1759: critical couplings $\mu_c^{(n)}$ discussed in section 2,3, but independently of these considerations, from a phenomenological viewpoint

1760: this is valuable possibility to consider. 

1761:  In this case, the running of the top Yukawa coupling $h_t$, written for simplicity by keeping 

1762: only the strong gauge coupling $g_3$, becomes

1763: \be

1764: 16 \pi^2 \ {d h_t \over d \ln Q} \ = \ \ h_t \ \left( {9 \over 2} N_H^{KK} h_t^2 - 8 N_g^{KK} g_3^2 \right) \ , \label{pheno1} 

1765: \ee  

1766: where $N_H^{KK}$ ($N_g^{KK}$) is the number of the Higgs (gluons) KK states lighter than $\Lambda_c^{(n)}$. In this case, the top mass

1767: descreases faster in the IR and the model is not ruled out anymore. It remains to be seen if some specific phenomenological 

1768: predictions can be made in this case.

1769:   

1770: In the 6d picture it is clear that bulk supersymmetry, softly broken on the brane, is a crucial ingredient 

1771: in the construction. A natural generalisation, therefore, is to add

1772: MSSM in 6d, with a softly broken SUSY localized at the origin of the

1773: 2d compact space and to consider again the perturbative critical coupling. Below $1/R$, the theory

1774: becomes equivalent to the supersymmetric generalisation of the 

1775: top condensation NJL model \cite{njlsusy1,njlsusy2}. This is a viable possibility, since the top quark mass in the supersymmetric top 

1776: condensation scenario is compatible with the experimentally measured value

1777: for compositeness scale $\Lambda_c$ sufficiently high.       

1778: Finally, another possibility is that the bulk scalar fields $\Phi_i$ are singlets under the SM gauge group 

1779: interacting with a localized Higgs 

1780: field $H$. Symmetry breaking in the $\Phi$ sector triggers electroweak symmetry breaking for the Higgs. In the dual NJL 

1781: type setup, this will correspond to a condensation involving a fermion different than the top \cite{hills}.  

1782: 

1783: