1: \documentclass[aps,prd,showpacs,amssymb,amsmath,nofootinbib,preprint]{revtex4}

2: 

3: \usepackage{slashed}

4: 

5: \newcommand{\md}{\mathrm{d}}

6: \newcommand{\me}{\mathrm{e}}

7: \newcommand{\mi}{\mathrm{i}}

8: \newcommand{\mca}{\mathcal{A}}

9: \newcommand{\mcl}{\mathcal{L}}

10: \newcommand{\mco}{\mathcal{O}}

11: \newcommand{\mcs}{\mathcal{S}}

12: \newcommand{\mbr}{\mathbb{R}}

13: \newcommand{\mbz}{\mathbb{Z}}

14: \newcommand{\half}{\tfrac{1}{2}}

15: \newcommand{\thhalf}{\tfrac{3}{2}}

16: \newcommand{\quart}{\tfrac{1}{4}}

17: 

18: \DeclareMathOperator{\diag}{diag}

19: 

20: %\bibliographystyle{unsrt}

21: 

22: \begin{document}

23: 

24: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

25: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

26: 

27: \title{Kink modes and effective four dimensional fermion and Higgs brane models}

28: \author{Damien P. George}

29: \email{d.george@physics.unimelb.edu.au}

30: \author{Raymond R. Volkas}

31: \email{r.volkas@physics.unimelb.edu.au}

32: \affiliation{School of Physics, Research Centre for High Energy

33: Physics,\\The University of Melbourne, Victoria 3010, Australia}

34: \date{April 19, 2007}

35: 

36: \begin{abstract}

37: 

38: In the construction of a classical smoothed out brane world model

39: in five dimensions, one uses a dynamically generated domain wall

40: (a kink) to localise an effective four dimensional theory.  At the

41: level of the Euler-Lagrange equations the kink sets up a potential

42: well, a mechanism which has been employed extensively to obtain localised,

43: four dimensional, massless chiral fermions.  We present the

44: generalisation of this kink trapping mechanism for both scalar and

45: fermionic fields, and retain all degrees of freedom that were

46: present in the higher dimensional theory.  We show that a kink

47: background induces a symmetric modified P\"oschl-Teller potential

48: well, and give explicit analytic forms for all the bound modes and

49: a restricted set of the continuum modes.  We demonstrate that it

50: is possible to confine an effective four dimensional scalar field

51: with a quartic potential of arbitrary shape.  This can be used to

52: place the standard model electroweak Higgs field on the brane, and

53: also generate nested kink solutions.  We also consider the limits

54: of the parameters in the theory which give thin kinks and

55: localised and de-localised scalar and fermionic fields.

56: 

57: \end{abstract}

58: 

59: \pacs{11.27.+d, 11.10.Kk}

60: 

61: \maketitle

62: 

63: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

64: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

65: 

66: \section{Introduction}

67: 

68: Our physical universe is described extremely well by the standard

69: model and general relativity, both of which are expressed in one

70: time and three space dimensions.  But these models are incomplete,

71: and a large amount of work has been done addressing the question:

72: can we improve our model of the universe by augmenting it with one

73: or more extra spatial dimensions?  Extra dimensions give extra

74: degrees of freedom and these have been used to tackle a variety of

75: problems present in the existing models.

76: 

77: There has been a vast investment of research into serious extra

78: dimensional models: Universal Extra

79: Dimensions~\cite{appelquist2000}, the Arkani-Hamed Dimopoulos and

80: Dvali model~\cite{add,antoniadis1990,aadd} and the Randall-Sundrum

81: model~\cite{rs1,rs2} to name a few.  The last of these

82: papers~\cite{rs2} demonstrated that

83: large, even infinite, extra dimensions are not ruled out by

84: experiment.  In the large extra dimensional scenarios, one

85: generically has the concept of a brane, to where the energy which

86: we observe in our four dimensional world is concentrated, or

87: confined.  The majority of the Randall-Sundrum (RS) models assume

88: the existence of a higher theory (e.g. string theory) which

89: provides such a brane along with a confinement mechanism.  The

90: effective low energy description of this set-up consists simply of

91: the combination of a five dimensional bulk action, and a four

92: dimensional action placed by hand at a specific location in the

93: extra dimension.  The branes in these models are infinitely thin

94: and Dirac delta distributions are used to couple the five and four

95: dimensional sectors.

96: 

97: If our four dimensional universe is embedded in higher dimensions

98: in such a way, then given enough energy, it should be possible to

99: probe the structure of our brane.  Theoretical predictions then

100: require a more sophisticated description of the dynamics of the

101: brane than just a delta distribution.  Ideally, one would like

102: to give a dynamical explanation for the formation of a brane, and

103: the ensuing localisation, using classical field theoretic ideas.

104: The canonical example of this comes from the work done initially

105: by Rubakov and Shaposhnikov in~\cite{rubakov1983} (see

106: also~\cite{akama1982,visser1985}).  One has a five

107: dimensional scalar field $\Phi(x^\mu,w)$ in a potential with

108: $\mbz_2$ symmetry that admits a topologically stable solution

109: $\phi_c(w) \sim \tanh (m w)$.  The mass parameter $m$ controls the

110: tension or inverse width of this domain wall defect.  A massless

111: five dimensional

112: fermion with a Yukawa coupling to this kink has a Dirac equation

113: which admits a separated solution

114: $\Psi(x^\mu,w) \sim \psi_L(x^\mu) \cosh^{-a}(m w)$. The parameter

115: $a$ is proportional to the five dimensional Yukawa coupling

116: constant and $\psi_L$

117: is a massless left-handed four dimensional fermion.  In the

118: limit of a thin brane, the extra dimensional factor of $\Psi$

119: becomes a delta distribution, giving essentially the RS

120: localisation mechanism, but with known dynamical origins.

121: 

122: Certainly, if one is just interested in the exact thin brane

123: limit, then all degrees of freedom freeze out except for the

124: massless zero mode $\psi_L$.  But if the brane is dynamically

125: generated, this limit will only be approximate, and at high enough

126: energies the excited states (essentially the Kaluza-Klein modes

127: of the infinite extra dimension) will become phenomenologically

128: important.  Gaining an understanding of the behaviour of these

129: excitations is therefore a prerequisite for any realistic model

130: building attempts.

131: 

132: Even with a full grasp of the dynamics of the confinement of

133: massless chiral modes, one is still not in a position to

134: dynamically embed the standard model in an extra dimension.  In

135: any realistic interacting model, there will be gauge symmetries

136: that need to be broken, most notably the electroweak symmetry.

137: The most straightforward idea is to localise the standard Higgs

138: field to the brane, and then proceed in the usual manner of

139: spontaneous symmetry breaking.  Such a mechanism requires an

140: analysis of the five dimensional couplings needed to produce an

141: effective four dimensional quartic potential, and a compatible

142: way of confining gauge fields.

143: 

144: In this paper we address some of these model building issues by

145: demonstrating that it is possible to confine an effective Higgs

146: field to a kink.  We also present the full spectrum of modes of

147: the kink, a coupled scalar field and a coupled fermion field in

148: the canonical kink scenario.  Along with this mode analysis, we

149: also investigate the relevant limits of the parameters in the

150: model which yield a thin kink, and show that we can obtain just

151: what is needed for an embedding of the standard model.  We do

152: not address the issue of gauge field confinement; one promising

153: idea is the Dvali-Shifman mechanism~\cite{dvali1996}.  Gravity

154: will also be ignored for the sake of simplicity in this initial

155: pass through the problem.

156: 

157: We begin in Section~\ref{sec:kink} by presenting the toy model

158: which supports a scalar kink and determine the full spectrum of

159: its associated modes.  We discuss the different limits of this

160: model which give the scenarios with no kink, a thick kink and a

161: thin kink.  In Section~\ref{sec:scalar} we add a scalar field to

162: the kink model, and show that the kink sets up a symmetric

163: modified P\"oschl-Teller potential for the extra dimensional

164: component of the scalar field.  We determine the modes of this

165: potential and use them to obtain an effective four dimensional

166: action, discussing in detail the thin kink limit.  In

167: Section~\ref{sec:fermion} we analyse a fermion coupled to the

168: kink, present the full mode decomposition, and show that in the

169: thin kink limit, the massless left-handed mode is the only

170: surviving dynamical field.  We also present an action that

171: contains this massless four dimensional mode coupled to a five

172: dimensional field.  We conclude and discuss further work in

173: Section~\ref{sec:conc}.  Appendix~\ref{app:smpt} contains

174: analytic solutions of the potential well set up by the kink which

175: are used extensively throughout the analysis.

176: Appendix~\ref{app:freeze} details the dynamics of the translation

177: symmetry of the kink, and we show how the zero mode of translation

178: freezes out in the thin kink limit.

179: 

180: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

181: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

182: 

183: \section{The kink and its limits}

184: \label{sec:kink}

185: 

186: All subsequent work will be in $4+1$ dimensional Minkowski space

187: with metric $g_{M N} = \diag (1,-1,-1,-1,-1)$.  Capital Latin

188: letters index the full space, Greek letters index the $3+1$

189: dimensional subspace, the extra dimensional coordinate is $w$ and

190: the embedding is $x^M = (x^\mu,w)$.

191: We begin with an action

192: describing a real five dimensional scalar field $\Phi$, given by

193: %

194: \begin{align}

195: \label{eq:phi-act-5d}

196: \mcs_\Phi &= \int \md^5 x \left[ \half \partial^M \Phi \partial_M \Phi - V(\Phi) \right] \\

197: & \text{with} \quad V(\Phi) = \frac{a}{4m} \left( \Phi^2 - \frac{m^3}{a} \right)^2,

198: \end{align}

199: %

200: where $a$ is a dimensionless constant\footnote{The constant $a$

201: is the five dimensional analogue of $\lambda$ in the four

202: dimensional potential

203: $V(\phi)=\tfrac{\lambda}{4}(\phi^2-\tfrac{m^2}{\lambda})^2$.}

204: and $m$ is the mass of $\Phi$.  From this action we find the

205: Euler-Lagrange equation for $\Phi$ to be

206: %

207: \begin{equation}

208: \label{eq:phi-kink}

209: \partial^M \partial_M \Phi - m^2 \Phi + \frac{a}{m} \Phi^3 = 0.

210: \end{equation}

211: 

212: With the aim of producing an effective four dimensional theory, 

213: $\Phi$ will initially be taken to depend only on the extra

214: dimensional coordinate $w$; this behaviour is denoted as

215: $\phi_c(w)$.  A topologically stable solution to

216: Eq.~\eqref{eq:phi-kink} is then

217: %

218: \begin{equation}

219: \label{eq:phi-clas}

220: \phi_c(w) = \sqrt\frac{m^3}{a} \tanh \left( \frac{mw}{\sqrt2} \right),

221: \end{equation}

222: %

223: which is the classical kink solution interpolating between the

224: $\mbz_2$ degenerate minima of the potential $V$.  This kink has

225: constant energy per unit volume at every spatial point in the

226: $3+1$ dimensional subspace, given by

227: %

228: \begin{equation}

229: \label{eq:phi-energy}

230: \varepsilon_{\phi_c}

231:     = \int_{-\infty}^{\infty} \md w \left[ \half (\partial_w \phi_c)^2 + V(\phi_c) \right]

232:     = \frac{2\sqrt2 m^4}{3 a}.

233: \end{equation}

234: 

235: We are interested in the behaviour of models where the classical

236: background is a thin kink, meaning $m$ is very large.  To make

237: this more precise, we write our parameters as

238: %

239: \begin{align*}

240: a &= \tilde{a} \Lambda^\alpha, &

241: m &= \tilde{m} \Lambda^\mu,

242: \end{align*}

243: %

244: with $\tilde{a}$ and $\tilde{m}$ finite, and consider the limit

245: $\Lambda \rightarrow \infty$.  In such a limit the kink energy

246: density is $\varepsilon_{\phi_c} \sim \Lambda^{4\mu-\alpha}$

247: and must remain finite, giving the constraint $\alpha = 4\mu$.

248: This in turn means the amplitude of the kink is

249: $|\phi_c| \sim \Lambda^{-\half\mu}$.  The single parameter $\mu$

250: now describes all possible limiting scenarios of the theory, with

251: $\mu=0$ corresponding to no limit being taken.  If $\mu<0$ the

252: potential $V$ disappears, $\phi_c \rightarrow 0$\footnote{The

253: relevant limit is

254: $\lim_{m\rightarrow0} \tanh(mw/\sqrt2) / \sqrt{m}= 0$.}

255: and the action describes a massless, freely propagating, five

256: dimensional scalar field.  The case $\mu>0$ is the

257: more interesting thin kink limit.  Here, the width of $\phi_c$

258: tends to zero and to keep the energy density finite, the height

259: also vanishes.  We will refer extensively to these limits in the

260: following sections.

261: 

262: By assuming that $\Phi$ depends only on $w$ we have of course lost

263: a lot of the dynamics of the full theory.  First, since $\phi_c$

264: breaks translational invariance along $w$, we expect a zero mode

265: which can act to translate the kink.  Second, if we have a thick

266: kink, we expect there to be massive modes associated with

267: arbitrary deformations of the kink.  We now proceed to incorporate

268: these dynamics.

269: 

270: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

271: 

272: \subsection{Kink modes and the effective model}

273: 

274: The classical kink background $\phi_c$ breaks the five dimensional

275: Poincar\'e symmetry, leaving a four dimensional Poincar\'e

276: subgroup.  This makes it natural to decompose a field into a sum

277: of products of an extra dimensional component and a $3+1$

278: dimensional component.  For $\Phi$, we want this expansion to be

279: made about the kink solution, and so we take

280: %

281: \begin{equation}

282: \label{eq:phi-expand}

283: \Phi(x^\mu,w) = \phi_c(w) + \sum_i \phi_i(x^\mu) \eta_i(w),

284: \end{equation}

285: %

286: where $\eta_i$ are a fixed orthonormal basis of the extra

287: dimension, $\phi_i$ are four dimensional dynamical fields and the

288: sum over $i$ can in general be a combination of discrete and

289: continuous modes.  We would like to determine a basis $\eta_i$

290: such that the equations of motion for the $\phi_i$ describe

291: massive scalar fields.  This can be done in the standard way by

292: taking the action for $\Phi$ given by~\eqref{eq:phi-act-5d},

293: substituting the expansion~\eqref{eq:phi-expand}, using the fact

294: that $\phi_c$ satisfies~\eqref{eq:phi-kink}, discarding terms

295: $\mco(\phi_i \eta_i)^3$ and higher and using integration by parts.

296: The effective second order action is then

297: %

298: \begin{equation*}

299: \mcs_\Phi^{\mco(2)} = \int \md^5 x \left[

300:     \frac{a}{4m} \phi_c^4 - \frac{m^5}{4a}

301:     - \half \phi_i \eta_i \left( \partial^\mu \partial_\mu - \partial_w^2

302:                                  + \frac{3a}{m} \phi_c^2 - m^2 \right) \phi_j \eta_j

303:     \right],

304: \end{equation*}

305: %

306: with implicit sum $i,j$ over the modes.  For the $\phi_i$ to

307: satisfy the massive Klein-Gordon equation with mass $\lambda_i$ we

308: require

309: %

310: \begin{equation*}

311: \left( -\frac{\md^2}{\md w^2} + \frac{3a}{m} \phi_c^2 - m^2 \right) \eta_i = \lambda_i^2 \eta_i.

312: \end{equation*}

313: %

314: We can use the known form of $\phi_c$ to get

315: %

316: \begin{equation}

317: \label{eq:eta-defn}

318: \left( -\frac{\md^2}{\md z^2} + 6 \tanh^2 z - 2 \right) \eta_i = \left( \frac{2 \lambda_i^2}{m^2} \right) \eta_i,

319: \end{equation}

320: %

321: where $z=mw/\sqrt2$.  This differential equation is a

322: Schr\"odinger equation with a symmetric modified P\"oschl-Teller

323: potential.  Analytic solutions are known in terms of

324: hypergeometric functions, and in general there are both bound and

325: continuum solutions.  In Appendix~\ref{app:smpt} we present these

326: solutions expressed in terms of regular functions, along with

327: their normalisation coefficients, for a more general form of the

328: potential.  For the kink modes at hand,

329: Eq.~\eqref{eq:eta-defn} is Eq.~\eqref{eq:smpt} with

330: $l=2$, and so there are two bound modes and a continuum

331: %

332: \begin{align*}

333: \lambda_0^2 &= 0 &

334:     \eta_0(w) &= E_0 \cosh^{-2} z, \\

335: \lambda_1^2 &= \thhalf m^2 &

336:     \eta_1(w) &= E_1 \sinh z \cosh^{-2} z, \\

337: \lambda_q^2 &= \half (q^2 + 4) m^2 &

338:     \eta_q(w) &= E_q \me^{\mi q z} \left( 3 \tanh^2 z - (q^2+1) - 3 \mi q \tanh z \right).

339: \end{align*}

340: %

341: The bound $\eta_{0,1}$ are square integrable normalised

342: by~\eqref{eq:norm-sqint} and the continuum $\eta_q$ are delta

343: function normalised by~\eqref{eq:norm-delta}, the normalisation

344: constants being

345: %

346: \begin{align*}

347: E_0 &= \sqrt\frac{3m}{4\sqrt{2}}, &

348: E_1 &= \sqrt\frac{3m}{2\sqrt{2}}, &

349: E_q &= \sqrt{\frac{m}{2 \pi \sqrt2 (q^2+1)(q^2+4)}}.

350: \end{align*}

351: %

352: 

353: Armed with the basis $\eta_i$, we return to the analysis of the

354: full dynamics of the kink.  Expanding the original

355: action~\eqref{eq:phi-act-5d} with $\Phi$ decomposed in the

356: $\eta_i$ basis and integrating over the extra dimension gives

357: %

358: \begin{equation}

359: \label{eq:phi-act-4d}

360: \mcs_\Phi = \int \md^4 x \left[ -\varepsilon_{\phi_c} + \mcl_\phi \right],

361: \end{equation}

362: %

363: where the $\phi$ kinetic, mass and self coupling terms are

364: %

365: \begin{equation}

366: \label{eq:phi-lag-4d}

367: \begin{aligned}

368: \mcl_\phi &=

369:     \half \partial^\mu \phi_0 \partial_\mu \phi_0

370:     + \half \partial^\mu \phi_1 \partial_\mu \phi_1

371:     - \tfrac{3}{4} m^2 \phi_1^2 \\

372: & \quad

373:     + \int_{-\infty}^{\infty} \md q \left[

374:         \half \partial^\mu \phi_q^* \partial_\mu \phi_q

375:         - \quart (q^2 + 4) m^2 \phi_q^* \phi_q \right] \\

376: & \quad

377:     - \kappa^{(3)}_{i j k} \phi_i \phi_j \phi_k

378:     - \kappa^{(4)}_{i j k l} \phi_i \phi_j \phi_k \phi_l.

379: \end{aligned}

380: \end{equation}

381: %

382: The fields $\phi_{0,1}$ are the real valued scalars associated

383: with the two bound state modes $\eta_{0,1}$.  The integration over

384: $q$ is over the complex valued continuum modes $\phi_q$ associated

385: with $\eta_q$.  Note that $\phi_{-q}=\phi_q^*$ and

386: $\eta_{-q}=\eta_q^*$ and so this integral is real.  The effective

387: cubic and quartic self interaction couplings are

388: %

389: \begin{align*}

390: \kappa^{(3)}_{i j k} &= \frac{a}{m} \int_{-\infty}^\infty \md w \left[ \phi_c \eta_i \eta_j \eta_k \right], \\

391: \kappa^{(4)}_{i j k l} &= \frac{a}{4m} \int_{-\infty}^\infty \md w \left[ \eta_i \eta_j \eta_k \eta_l \right].

392: \end{align*}

393: %

394: For brevity, the indices $i,j,k,l$ label bound modes, continuum

395: modes or a mixture of both and the sum over these labels is

396: implied in Eq.~\eqref{eq:phi-lag-4d}.  The couplings $\kappa$

397: can be computed as their integrands are known; some are zero due

398: to parity, some are non-zero.

399: 

400: Equations~\eqref{eq:phi-act-4d} and~\eqref{eq:phi-lag-4d} are

401: exact manipulations of the original five dimensional $\Phi$

402: model~\eqref{eq:phi-act-5d}, and provide a description in a

403: basis useful for investigating the effective four dimensional

404: behaviour.  The bound modes are reminiscent of the Kaluza-Klein

405: modes one obtains in compact extra dimensions, but in the case at

406: hand the gaps in the mass spectrum are not uniform.  Furthermore,

407: the continuum modes do not have analogues in the Kaluza-Klein

408: model and are not strictly four dimensional, but instead form a

409: pseudo-five dimensional field with reduced degrees of freedom.

410: 

411: Regarding the renormalisability of our model, we note that because

412: the original five dimensional action~\eqref{eq:phi-act-5d}

413: contains non-renormalisable terms, we do not have any reason

414: to stop writing down potentials at quartic order.  But the

415: manipulations which bring us to the four dimensional level given

416: by Eq.~\eqref{eq:phi-lag-4d}, do in fact leave us with a

417: renormalisable theory if we truncate the action to just the bound

418: states.  This renormalisability of the bound states may be a

419: useful criterion for restricting the types of terms that one begins

420: with in the five dimensional (or higher dimensional) action.

421: 

422: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

423: 

424: \subsection{Limiting behaviour of the effective model}

425: 

426: Now that we have such a reformulation of the kink model, we are in

427: a position to analyse the full dynamics of the system for the

428: three different limits of the mass $m$.  Following our previous

429: parameterisation, for $\mu<0$ there is no kink and the basis

430: $\eta_i$ is replaced by a standard complex Fourier expansion.  All

431: the dynamical components are packaged together by the Fourier

432: transform and it is no longer sensible to perform the $w$

433: integral.  Instead one should consider $\Phi$ as a free, massless,

434: five dimensional field.

435: 

436: For the thick kink case when $\mu=0$, the spectrum consists of the

437: energy density of the integrated kink, a zero mode, a massive

438: bound mode and a continuum of massive complex scalars.  These

439: effective four dimensional fields are self-coupled and coupled

440: amongst each other via cubic and quartic interactions.  In

441: particular, the zero mode $\phi_0$ and massive bound mode $\phi_1$

442: each have a potential:

443: %

444: \begin{align*}

445: V_0(\phi_0) &= \frac{9 \sqrt2 \, a}{140} \phi_0^4, \\

446: V_1(\phi_1) &= \thhalf m^2 \phi_1^2

447:                + \frac{3\pi}{32} \sqrt{\frac{3a}{2\sqrt2}} \, m \, \phi_1^3

448:                + \frac{9\sqrt2 \, a}{280} \phi_1^4,

449: \end{align*}

450: %

451: which are due to the non-zero values of $\kappa^{(4)}_{0000}$,

452: $\kappa^{(3)}_{111}$ and $\kappa^{(4)}_{1111}$.  Similarly, we can

453: compute the coupling potential amongst these bound modes, to get

454: %

455: \begin{equation*}

456: V_{0,1}(\phi_0,\phi_1) =

457:     \frac{9\pi}{64} \sqrt{\frac{3a}{2\sqrt2}} \, m \, \phi_0^2 \phi_1

458:     + \frac{9\sqrt2 \, a}{70} \, \phi_0^2 \phi_1^2.

459: \end{equation*}

460: 

461: Note that while $\phi_0$ has no mass term it does have a non-zero

462: potential, making it energetically unfavourable to excite the

463: field, even though it costs zero energy to translate the kink.  We

464: can account for this unexpected result by recalling that

465: $\eta_0$ corresponds to \emph{infinitesimal} translations of

466: $\phi_c$.  Adding any small but finite multiple of $\eta_0$

467: to $\phi_c$ will, to first order, perform a translation, but to

468: higher order it will \emph{deform} the kink.  The energy cost of

469: these higher order deformations are described by the potential

470: $V_0$.  For completeness we point out that while $V_1$ has a cubic

471: term, the potential has only one extremum, which is a minimum at

472: $\phi_1=0$.

473: 

474: We now move on to the thin kink limit where $\mu>0$ and both $m$

475: and $a$ tend to infinity.  The kink energy density remains finite,

476: but the masses of the bound mode $\phi_1$ and the continuum modes

477: $\phi_q$ go like $m$.  Recall that these massive modes correspond

478: to deformations in the kink and as the kink gets thinner it also

479: gets stiffer, requiring more energy for a given deformation.  The

480: dynamics of the massive modes are thus frozen out, as they are no

481: longer able to deform the kink without possessing infinite energy.

482: This argument also applies to the translation zero mode $\phi_0$,

483: which gets frozen out in the thin kink limit, despite remaining

484: massless, because its

485: corresponding potential $V_0$ becomes infinitely steep.  This

486: can be understood from the arguments give above: for an infinitely

487: stiff kink, the higher order deformations due to the zero mode are

488: forbidden and the only physical resolution is to remove the

489: dynamics of this mode.  So in the thin kink limit, we are left

490: with only the kink energy density $\varepsilon_{\phi_c}$ in the

491: effective four dimensional action.  Appendix~\ref{app:freeze}

492: discusses this result in more depth and includes an identification

493: of the mode responsible for finite translations.  Also

494: see~\cite{shaposhnikov2005} for a discussion of the translational

495: zero mode in the presence of gravity.

496: 

497: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

498: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

499: 

500: \section{Adding a scalar field}

501: \label{sec:scalar}

502: 

503: We have so far performed an analysis of the kink and its modes in

504: isolation.  As stated previously, we aim to use the properties of

505: the domain wall to dynamically trap five dimensional fields to a

506: brane and create an effective four dimensional model.  We can

507: achieve this if the kink is coupled to a different five

508: dimensional field and projects out a set of modes with the lowest

509: mode separated from the rest by a significant mass gap.  Then if

510: we are at energies where only the lowest bound state can be

511: excited, the degree of freedom of propagation along the extra

512: dimension has been lost and the bound mode is confined.

513: 

514: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

515: 

516: \subsection{Scalar modes}

517: 

518: The simplest place to start is to take another five dimensional

519: scalar field $\Xi(x^M)$ and couple it to the kink field.  The

520: action describing this model is

521: $\mcs_{\Phi+\Xi} = \mcs_\Phi + \mcs_\Xi$ where

522: %

523: \begin{align}

524: \label{eq:xi-act-5d}

525: \mcs_\Xi &= \int \md^5 x \left[

526:     \half \partial^M \Xi \partial_M \Xi

527:     - \frac{a b(b+1)}{4 m} \Phi^2 \Xi^2

528:     - W(\Xi)

529:     \right] \\

530: & \text{with} \quad W(\Xi) = \frac{n^2}{2} \Xi^2 + \frac{c}{4n} \Xi^4.

531: \end{align}

532: %

533: The parameters $b$ and $c$ are dimensionless, while the

534: dimensionful parameter $n$ is the mass of $\Xi$.  We follow the

535: analysis for the kink modes and perform the general separation

536: %

537: \begin{equation*}

538: \Xi(x^M) = \sum_i \xi_i(x^\mu) k_i(w),

539: \end{equation*}

540: %

541: where the sum over $i$ can again be a combination of discrete and

542: continuous parts.  To obtain a suitable basis $k_i$, we look at

543: the linearised equation of motion for $\Xi$ with $\Phi = \phi_c$

544: %

545: \begin{equation*}

546: \left( \partial^\mu \partial_\mu - \partial_w^2 + \frac{a b(b+1)}{2 m} \phi_c^2 + n^2 \right) \xi_i k_i = 0.

547: \end{equation*}

548: %

549: We want $\xi_i$ to satisfy the four dimensional Klein-Gordon

550: equation with mass $\delta_i$.  This leads to

551: %

552: \begin{equation*}

553: \left( -\frac{\md^2}{\md z^2} + b(b+1) \tanh^2 z \right) k_i = \left( \frac{2(\delta_i^2 - n^2)}{m^2} \right) k_i

554: \end{equation*}

555: %

556: where $z=mw/\sqrt2$ as before.  This Schr\"odinger equation has the same

557: form as the one obtained for the kink modes.  We see that the kink

558: sets up a symmetric modified P\"oschl-Teller potential well which

559: traps not only its own modes, but also those of a coupled scalar

560: field.  Looking to Appendix~\ref{app:smpt} we see that the basis

561: $k_i$ contains $\lceil b \rceil$ bound modes\footnote{We use the

562: standard notation $\lceil . \rceil$ for the ceiling function.} and

563: a continuum.  The masses of the bound states are

564: %

565: \begin{align*}

566: \delta_0^2 &= n^2 + \half b m^2, \\

567: \delta_1^2 &= n^2 + \half (3b-1) m^2, \\

568: \delta_2^2 &= n^2 + \half (5b-4) m^2, \\

569: \vdots \\

570: \delta_i^2 &= n^2 + \half ((2i+1)b-i^2) m^2,

571: \end{align*}

572: %

573: and for the continuum we have

574: \begin{align*}

575: \delta_q^2 &= n^2 + \half (q^2 + b(b+1)) m^2,

576: \end{align*}

577: %

578: where $q\in\mbr$ labels the continuum modes.  We will not give

579: explicit forms of the functions $k_i$; they are easily determined

580: from Appendix~\ref{app:smpt}.  Unlike the modes of the kink, this

581: spectrum of masses does not in general include a zero mode and the

582: bottom of the spectrum is dependent on the parameters $n$, $b$ and

583: $m$.  We also have the freedom to change the sign of $n^2$ in the

584: original action and dial up any positive, zero, or negative value

585: of $\delta_0^2$.

586: 

587: As before, we use the basis $k_i$ to expand $\Xi$ in the original

588: action~\eqref{eq:xi-act-5d} and integrate over the extra

589: dimension.  Including the kink sector, the effective four

590: dimensional action is then

591: %

592: \begin{equation*}

593: \mcs_{\Phi+\Xi} = \int \md^4 x \left[ -\varepsilon_{\phi_c} + \mcl_\phi + \mcl_\xi \right],

594: \end{equation*}

595: %

596: where the kink-only parts are given previously and the scalar

597: Lagrangian is

598: %

599: \begin{equation}

600: \label{eq:xi-lag-4d}

601: \begin{aligned}

602: \mcl_\xi &=

603:     \sum_{i=0}^{\lceil b-1 \rceil} \left[ \half \partial^\mu \xi_i \partial_\mu \xi_i

604:     - \half \delta_i^2 \xi_i^2 \right]

605:     + \int_{-\infty}^{\infty} \md q \left[ \half \partial^\mu \xi_q \partial_\mu \xi_q

606:     - \half \delta_q^2 \xi_q^2 \right] \\

607: & \quad

608:     - g^{(3)}_{ijk} \phi_i \xi_j \xi_k

609:     - g^{(4)}_{ijkl} \phi_i \phi_j \xi_k \xi_l

610:     - \tau_{ijkl} \xi_i \xi_j \xi_k \xi_l.

611: \end{aligned}

612: \end{equation}

613: %

614: The Yukawa and self coupling factors are

615: %

616: \begin{align*}

617: g^{(3)}_{ijk}  &= \frac{ab(b+1)}{2m} \int_{-\infty}^\infty \md w \left[ \phi_c \eta_i k_j k_k \right], \\

618: g^{(4)}_{ijkl} &= \frac{ab(b+1)}{4m} \int_{-\infty}^\infty \md w \left[ \eta_i \eta_j k_k k_l \right], \\

619: \tau_{ijkl}    &= \frac{c}{4n}       \int_{-\infty}^\infty \md w \left[ k_i k_j k_k k_l \right].

620: \end{align*}

621: 

622: With this expanded four dimensional action, we are ready to

623: analyse the various limits of the model with the scalar field.

624: 

625: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

626: 

627: \subsection{The thin kink with a scalar field}

628: 

629: We are considering the combined five dimensional action

630: $\mcs_{\Phi+\Xi} = \mcs_\Phi + \mcs_\Xi$ and the various limits

631: that arise through $m$, the mass of the kink.  As discussed

632: previously, we have three scenarios which are characterised by the

633: sign of $\mu$.  In the case $\mu<0$ there is no kink and we are

634: left with two interacting five dimensional fields $\Phi$ and

635: $\Xi$.  The thick kink scenario, $\mu=0$, has many interacting

636: four dimensional scalar fields, the details given by the mass

637: spectra of $\phi_i$ and $\xi_i$ and the couplings $\kappa$, $g$

638: and $\tau$.  We will not dwell on these two cases, but instead

639: concentrate our attention on the thin kink limit and explore the

640: parameter space of $b$, $c$ and $n$.

641: 

642: It was shown previously that the thin kink limit leaves only the

643: energy density of the domain wall in the effective four

644: dimensional action.  The dynamics of all the scalar modes $\phi_i$

645: are removed and so the $g^{(3)}$ and $g^{(4)}$ Yukawa terms

646: in~\eqref{eq:xi-lag-4d} are eliminated\footnote{It is shown in

647: Appendix~\ref{app:freeze} that the translation symmetry manifests

648: as a linear combination of $\phi_{0,1,q}$ which can assume a

649: non-zero constant value with zero energy cost, even in the thin

650: kink limit.  It is possible to show that this contribution to the

651: Yukawa terms is counteracted by shifting the $k_i$ basis to align

652: with the shifted kink profile, and redefining the fields

653: $\xi_i$.}.  With $m\rightarrow\infty$

654: and $b$ finite, the masses of all the $\xi_i$ modes will also tend

655: to infinity and the scalar $\Xi$ becomes completely frozen out.

656: To leave some remnant of $\Xi$ in the model we have two choices:

657: either take $b$ to zero to counter $m^2$, or choose $n^2$ such

658: that it cancels $\half b m^2$.

659: 

660: For the first choice, let $n$ be finite and

661: $b=\tilde b \Lambda^\beta$.  Then

662: $\delta_0^2 \sim \Lambda^{\beta+2\mu}$ and the mode $\xi_0$ has

663: finite mass if $\beta+2\mu\le0$.  Since $b\rightarrow0$ in the

664: limit $\Lambda\rightarrow\infty$, there are in fact no bound

665: modes, the basis $k_i$ is not valid and we must consider $\Xi$ as

666: a five dimensional field.  The effective action for such a limit

667: of the parameters is\footnote{This result uses

668: $\tanh^2(mw/\sqrt2) \rightarrow 1$ as $m\rightarrow\infty$, which

669: ignores the fact that the distribution vanishes on a set of

670: measure zero at the origin.}

671: %

672: \begin{equation}

673: \label{eq:xi-eff-5d}

674: \mcs_{\Phi+\Xi}^\text{5D} = \int \md^4 x \left[ -\varepsilon_{\phi_c} \right]

675:     + \int \md^5 x \left[ \half \partial^M \Xi \partial_M \Xi - \quart b m^2 \Xi^2 - W(\Xi) \right].

676: \end{equation}

677: %

678: We see that the five dimensional field $\Xi$ has nothing dynamical

679: to couple to, and just picks up an addition to its $\Xi^2$ term.

680: If we have the strict inequality $\beta+2\mu<0$, this addition to

681: the mass will be zero.

682: 

683: The second choice which keeps some part of $\Xi$ alive is to

684: change the sign of $n^2$ and fine tune it to exactly cancel the

685: infinite term in $\delta_0^2$.  An exact cancellation would

686: render the mode $\xi_0$ massless; we can be more general and allow

687: a finite mass to remain by choosing

688: %

689: \begin{equation}

690: \label{eq:nsq-cancel}

691: n^2 = -\half b m^2 + n_0^2,

692: \end{equation}

693: %

694: where $n_0$ is finite.  Let us briefly comment on the situation

695: where $b\rightarrow0$, but not quickly enough to counter $m$

696: (thus $-2\mu<\beta<0$) and so we must choose $n^2$ as

697: in~\eqref{eq:nsq-cancel}.  In this case there are again no bound

698: modes and $\Xi$ is a five dimensional field with equivalent

699: physics as described by~\eqref{eq:xi-eff-5d}, except the

700: quadratic term of the generated potential is replaced by

701: $\half n_0^2 \Xi^2$.

702: 

703: We can now restrict our analysis to the case where $\beta\ge0$

704: and $n^2$ is of the form given by~\eqref{eq:nsq-cancel}.  As we

705: have a non-zero $b$, there is at least one bound mode, and in fact

706: only the lowest bound mode will have finite mass.  The dynamics of

707: the higher bound modes and the continuum will not be part of the

708: model as their corresponding masses are infinite.  With only

709: $\xi_0$ alive, the quartic coupling terms reduce to just the

710: one with the factor $\tau_{0000}$.  Putting all these pieces

711: together we arrive at the four dimensional action

712: %

713: \begin{align*}

714: \mcs_{\Phi+\Xi}^\text{4D} &= \int \md^4 x \left[

715:     -\varepsilon_{\phi_c}

716:     + \half \partial^\mu \xi_0 \partial_\mu \xi_0

717:     - W_0(\xi_0)

718:     \right] \\

719: & \text{with} \quad W_0(\xi_0) =

720:     \half n_0^2 \xi_0^2

721:     + \frac{c}{4\sqrt{2\pi}}

722:         \frac{\sqrt2 \, \Gamma^2(b+\half)\Gamma(2b)}{\sqrt b \, \Gamma^2(b)\Gamma(2b+\half)}

723:         \xi_0^4.

724: \end{align*}

725: %

726: For large values of $b$ the potential simplifies to

727: %

728: \begin{equation*}

729: W_0(\xi_0) \sim \half n_0^2 \xi_0^2 + \frac{c}{4\sqrt{2\pi}} \xi_0^4.

730: \end{equation*}

731: 

732: This analysis shows that in the thin kink limit, a five

733: dimensional coupled scalar field is projected down to a single,

734: localised, four dimensional scalar field $\xi_0$.  As the

735: parameters $n_0$ and $c$ are arbitrary, one can generate a

736: phenomenologically suitable potential for $\xi_0$, in particular

737: the sign of $n_0^2$ can be changed to yield a potential which

738: encourages a non-zero vacuum expectation value.  We mention two

739: uses of this.  Most obviously this mechanism can be used to

740: localise the standard model electroweak Higgs field to a brane, and it should

741: be no trouble to arrange Yukawa couplings to fermion fields for

742: mass generation.  Secondly, the effective potential $W_0(\xi_0)$

743: has the same form as the kink potential $V(\Phi)$ and can thus

744: support a domain wall solution, leading to the idea of nested

745: brane worlds.  Beginning with a six dimensional model with the two

746: scalar fields $\Phi$ and $\Xi$ and a suitable potential, one can

747: use $\Phi$ to generate a domain wall and an effective five

748: dimensional action and then use the lowest projected mode of $\Xi$

749: in the same way to generate an effective four dimensional action.

750: The method used does not depend on the dimensionality and one is

751: free to generate an arbitrary number of nestings.  Of course, this

752: mechanism just deals with the particle content.  The non-trivial

753: exercise is to check that gravity can be broken down in similar

754: stages and made to reproduce four dimensional general relativity;

755: this will not be attempted here.

756: 

757: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

758: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

759: 

760: \section{Adding a fermion field}

761: \label{sec:fermion}

762: 

763: In the previous section we performed a full analysis of the modes

764: of a five dimensional scalar field coupled to a kink.  We showed

765: that in the thin kink limit, an effective four dimensional scalar

766: field remains and could be potentially useful for model building.

767: In direct analogy with this analysis we now consider a five

768: dimensional massless fermion with Yukawa coupling to $\Phi$, find

769: a suitable basis for decomposition, and investigate the limiting

770: behaviour.  This is a generalisation of the well known and heavily

771: used result that the kink supports a zero mode fermion, as first

772: discussed in the brane context in~\cite{rubakov1983}.  For a

773: partially analytic analysis of massive fermion modes confined to a

774: thick brane in the presence of gravity, see~\cite{ringeval2002}.

775: 

776: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

777: 

778: \subsection{Fermion modes}

779: 

780: Fermions in five dimensions are four component spinors and their

781: Dirac structure is described by $\Gamma^M$ with

782: $\{\Gamma^M,\Gamma^N\}=g^{M N}$.  Specifically

783: $\Gamma^\mu = \gamma^\mu$ and $\Gamma^5=\mi \gamma^5$ where

784: $\gamma^{\mu,5}$ are the usual gamma matrices in the Dirac

785: representation.  Our action for a massless fermion coupled to the

786: kink is $\mcs_{\Phi+\Psi} = \mcs_\Phi + \mcs_\Psi$ where

787: %

788: \begin{equation}

789: \label{eq:psi-act-5d}

790: \mcs_\Psi = \int \md^5 x \left[

791:     \overline\Psi \mi \Gamma^M \partial_M \Psi

792:     - \sqrt{\frac{ad^2}{2m}} \Phi \overline\Psi \Psi

793:     \right].

794: \end{equation}

795: %

796: The kink parameters $a$ and $m$ are the same as before and $d$ is

797: a dimensionless coupling parameter.  As with the scalar field

798: $\Xi$, we expect the extra dimensional behaviour of $\Psi$ to be

799: quite different to the four dimensional part.  Also, because of

800: the Dirac structure of the fifth gamma matrix

801: $\Gamma^5 = \mi \gamma^5$, we expect left- and right-handed

802: projections of the four dimensional part to behave differently.

803: Thus we choose the general expansion

804: %

805: \begin{equation}

806: \label{eq:psi-expand}

807: \Psi(x^\mu,w) = \sum_i \psi_{L i}(x^\mu) f_{L i}(w)

808:               + \sum_i \psi_{R i}(x^\mu) f_{R i}(w),

809: \end{equation}

810: %

811: where the $f_{L i}$ and $f_{R i}$ are a fixed basis, the $\psi_i$

812: are dynamical, $\gamma^5 \psi_{L i} = -\psi_{L i}$,

813: $\gamma^5 \psi_{R i} = \psi_{R i}$ and the sum over $i$ can be

814: both discrete and continuous.  To obtain the defining equations

815: for the basis functions $f_i$, we impose that the $\psi_i$ satisfy

816: the massive Dirac equation by

817: $\mi \slashed\partial \psi_{L i} = \sigma_i \psi_{R i}$

818: and $\mi \slashed\partial \psi_{R i} = \sigma_i \psi_{L i}$.

819: Then, substituting the expansion~\eqref{eq:psi-expand} into the

820: Euler-Lagrange equation for $\Psi$ we arrive at

821: %

822: \begin{equation}

823: \begin{aligned}

824: & \psi_{L i} \left( -\partial_w f_{L i} + f_{R i} \sigma_i - \sqrt{\frac{ad^2}{2m}} \phi_c f_{L i} \right) \\

825: & + \psi_{R i} \left( \partial_w f_{R i} + f_{L i} \sigma_i - \sqrt{\frac{ad^2}{2m}} \phi_c f_{R i} \right)

826:     = 0.

827: \end{aligned}

828: \label{eq:f-defn1}

829: \end{equation}

830: 

831: Since left and right Dirac components are independent and the

832: $\psi_i$ are arbitrary fields, both of the two parenthesised

833: factors in Eq.~\eqref{eq:f-defn1} must be zero.  Hence the

834: $f_{L i}$ and $f_{R i}$ must satisfy a set of two first order

835: coupled ordinary differential equations.  We turn these equations

836: into two uncoupled second order equations:

837: %

838: \begin{align*}

839: \left( -\frac{\md^2}{\md z^2} + d(d+1) \tanh^2 z - d \right) f_{L i} &= \frac{2 \sigma_i^2}{m^2} f_{L i}, \\

840: \left( -\frac{\md^2}{\md z^2} + d(d-1) \tanh^2 z + d \right) f_{R i} &= \frac{2 \sigma_i^2}{m^2} f_{R i}.

841: \end{align*}

842: %

843: As with the scalar field, we see that the kink sets up a symmetric

844: modified P\"oschl-Teller potential well which traps the extra

845: dimensional component of the fermion field.  This has been noted

846: in a similar context of fat branes in~\cite{hung2004}, and in the

847: context of a two dimensional Dirac equation

848: in~\cite{decastro2006}.  We use Appendix~\ref{app:smpt} of the

849: current work to obtain the solutions; the bound modes come in

850: pairs given by

851: %

852: \begin{align*}

853: (\sigma_0^d)^2 &= 0 && \Biggl\{

854:     \begin{aligned}

855:         f_{L 0}^d(w) &= F_{L 0}^d \cosh^{-d} z \\

856:         f_{R 0}^d(w) &= 0,

857:     \end{aligned} \\

858: (\sigma_1^d)^2 &= \half (2d-1) m^2 && \Biggl\{

859:     \begin{aligned}

860:         f_{L 1}^d(w) &= F_{L 1}^d \sinh z \cosh^{-d} z \\

861:         f_{R 1}^d(w) &= F_{R 1}^d \cosh^{-d+1} z,

862:     \end{aligned} \\

863: \vdots \\

864: (\sigma_n^d)^2 &= \half (2nd-n^2) m^2 && \Biggl\{

865:     \begin{aligned}

866:         f_{L n}^d(w) &= \frac{1}{\sigma_n^d} \left( \frac{dm}{\sqrt2} \tanh z - \frac{\md}{\md w} \right) f_{L, n-1}^{d-1} \\

867:         f_{R n}^d(w) &= f_{L, n-1}^{d-1}(w).

868:     \end{aligned} \\

869: \end{align*}

870: %

871: These bound state modes are valid for all positive values of

872: $d$ and there are $\lceil d \rceil$ sets of modes.  For the

873: continuum, the solutions can be found in terms of standard

874: functions when $d$ is a positive integer.  They are

875: %

876: \begin{align*}

877: (\sigma_q^1)^2 &= \half (q^2 + 1) m^2 && \Biggl\{

878:     \begin{aligned}

879:         f_{L q}^1 &= F_{L q}^1 \me^{\mi q z} \left( \tanh z - \mi q \right) \\

880:         f_{R q}^1 &= F_{R q}^1 \me^{\mi q z},

881:     \end{aligned} \\

882: (\sigma_q^2)^2 &= \half (q^2 + 4) m^2 && \Biggl\{

883:     \begin{aligned}

884:         f_{L q}^2 &= F_{L q}^2 \me^{\mi q z} \left( 3 \tanh^2 z - (q^2 + 1) - 3 \mi q \tanh z \right) \\

885:         f_{R q}^2 &= F_{R q}^2 \me^{\mi q z} \left( \tanh z - \mi q \right),

886:     \end{aligned} \\

887: \vdots \\

888: (\sigma_q^d)^2 &= \half (q^2 + d^2) m^2 && \Biggl\{

889:     \begin{aligned}

890:         f_{L q}^d &= \frac{1}{\sigma_n^d} \left( \frac{dm}{\sqrt2} \tanh z - \frac{\md}{\md w} \right) f_{L q}^{d-1} \\

891:         f_{R q}^d &= f_{L q}^{d-1}(w).

892:     \end{aligned}

893: \end{align*}

894: %

895: The normalisation coefficients $F$ can be computed using

896: Appendix~\ref{app:smpt}.

897: 

898: Using these basis functions, we expand the original

899: action~\eqref{eq:psi-act-5d} and integrate over the extra

900: dimension.  The effective four dimensional action, including the

901: kink dynamics, is

902: %

903: \begin{equation*}

904: \mcs_{\Phi+\Psi} = \int \md^4 x \left[ -\varepsilon_{\phi_c} + \mcl_\phi + \mcl_\psi \right],

905: \end{equation*}

906: %

907: where the Lagrangian for the expanded $\Psi$ is

908: %

909: \begin{equation}

910: \label{eq:psi-lag-4d}

911: \begin{aligned}

912: \mcl_\psi &=

913:     \overline{\psi_{L 0}} \mi \slashed\partial \psi_{L 0}

914:     + \sum_{i=1}^{\lceil d-1 \rceil} \overline{\psi_i} (\mi \slashed\partial - \sigma_i) \psi_i

915:     + \int_{-\infty}^{\infty} \md q \left[ \overline{\psi_q} (\mi \slashed\partial - \sigma_q) \psi_q \right] \\

916: & \quad

917:     - h_{i j k} \phi_i \overline{\psi_{L j}} \psi_{R k}

918:     - h_{i k j}^* \phi_i^* \overline{\psi_{R j}} \psi_{L k}.

919: \end{aligned}

920: \end{equation}

921: %

922: We have condensed the notation using

923: $\psi_i=\psi_{L i}+\psi_{R i}$ and similarly for $\psi_q$.  For

924: brevity in the Yukawa terms, the implicit sum over $i$ denotes a

925: sum over the bound modes $i=0,1$ and an integral over the

926: continuum; similarly the sum over $j$ and $k$ denotes the sum over

927: bound and continuum fermion modes.  The effective dimensionless

928: Yukawa coupling is

929: %

930: \begin{equation*}

931: h_{i j k} = \sqrt{\frac{a d^2}{2m}} \int_{-\infty}^\infty \md w \left[ \eta_i f_{L j}^* f_{R k} \right].

932: \end{equation*}

933: %

934: We can compute these $h$.  Those of importance are the

935: couplings between the bound kink modes and the bound fermion

936: modes, the first few being

937: %

938: \begin{align*}

939: h_{0 i 0} &= 0 \qquad \text{(for all $i$)}, \\

940: h_{0 0 1} &= \sqrt{\frac{3a}{8\sqrt2}}

941:              \left(\frac{d-\half}{d-1}\right)^{\thhalf}

942:              \frac{\Gamma^2(d-\half)}{\Gamma^2(d-1)}, \\

943: h_{0 1 1} &= h_{1 0 1} = 0, \\

944: h_{1 1 1} &= \sqrt{\frac{3a}{8\sqrt2}}

945:              \frac{(d-\half)^{\half}}{d-1}

946:              \frac{\Gamma^2(d-\half)}{\Gamma^2(d-1)}.

947: \end{align*}

948: 

949: This analysis includes the well known chiral zero mode

950: localisation when all fermion modes except $\psi_{L0}$ are removed

951: from~\eqref{eq:psi-lag-4d}.  In this reduced model there are no

952: Yukawa couplings between $\psi_{L0}$ and any of the kink modes

953: $\phi_i$ due to the chirality of the fermion mode.

954: 

955: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

956: 

957: \subsection{The thin kink with a fermion field}

958: 

959: With the expanded effective four dimensional action for $\Psi$, we

960: can now consider the limits of the kink.  With no kink ($\mu<0$) the basis

961: $f_i$ is not valid and we instead obtain a model containing the

962: coupled five dimensional fields $\Phi$ and $\Psi$.  The thick kink

963: scenario ($\mu=0$) contains four dimensional interacting scalar fields

964: $\phi_i$ and $\psi_i$.  The thin kink limit ($\mu>0$) is what we are most

965: interested in, where the remnant of the kink sector is just the

966: energy density.  Then the Yukawa term in~\eqref{eq:psi-act-5d}

967: reduces to $(dm/\sqrt2) \tanh^2 z \, \overline\Psi \Psi$ and we have

968: only the parameter $d$ left to play with.  There are four

969: scenarios, which we classify using the previous parameterisation

970: of the limit and write $d=\tilde d\Lambda^\delta$.

971: 

972: First, if $\delta=-\mu$ then $d\rightarrow0$, there are no bound

973: fermion modes and the fermion is left as a five dimensional field

974: with the action

975: %

976: \begin{equation*}

977: \mcs_{\Phi+\Psi}^\text{5D} =

978:     \int \md^4 x \left[ -\varepsilon_{\phi_c} \right]

979:     + \int \md^5 x \left[ \overline\Psi \mi \Gamma^M \partial_M \Psi

980:                           - \frac{dm}{\sqrt2} \theta(w) \overline\Psi \Psi \right],

981: \end{equation*}

982: %

983: where $\theta(w)$ is the step function.  The fermion in this case

984: has an unusual mass term that changes sign across the kink.

985: Second, if $\delta<-\mu$, we have the same situation as in the

986: first case, except the mass term disappears.  Third, if

987: $-\mu<\delta<0$, $d$ is not going to zero fast enough to counter

988: $m$ and the unusual mass term becomes infinite.  The five

989: dimensional field $\Psi$ is thus frozen out and we are left with

990: just the kink energy density in our effective theory.

991: 

992: The fourth case has $\delta\ge0$, so $d>0$ and there is at least

993: one bound fermion mode.  Because the masses of these modes go like

994: $m$, all modes are frozen out except the zero mode $\psi_{L0}$.

995: None of the Yukawa couplings $h$ are relevant because they couple

996: $\psi_{L0}$ to higher fermion modes which are not dynamical.

997: The effective action is then simply

998: %

999: \begin{equation*}

1000: \mcs_{\Phi+\Psi}^\text{4D} =

1001:     \int \md^4 x \left[ -\varepsilon_{\phi_c} + \overline{\psi_{L0}} \mi \slashed\partial \psi_{L0} \right],

1002: \end{equation*}

1003: %

1004: which contains the dynamics of a single, four dimensional, chiral,

1005: massless fermion.  Thus we obtain the result of Rubakov and

1006: Shaposhnikov in the thin kink limit with $d>0$.  This is of

1007: obvious importance in model building where chiral zero modes are

1008: the starting point for the particle content.

1009: 

1010: In all the limits of the models we have considered so far, the

1011: dynamical fields are either exclusively four dimensional or

1012: exclusively five dimensional.  It is possible to construct an

1013: action which in the thin kink limit contains interacting four

1014: and five dimensional fields.  Because the dynamics of the kink

1015: are frozen out, two extra fields are required from the outset: one

1016: which couples appropriately to the kink and provides a four

1017: dimensional zero-mode, and another which remains five dimensional.

1018: To demonstrate this idea, we take the original action for the kink

1019: and the fermion and add the scalar field $\Xi$ with a coupling to

1020: the fermion only.  The action is then

1021: %

1022: \begin{equation*}

1023: \mcs_\text{all} =

1024:     \mcs_\Phi

1025:     + \mcs_\Psi

1026:     + \int \md^5 x \left[ \half \partial^M \Xi \partial_M \Xi - s \, \Xi (\overline\Psi \Psi^c + \overline{\Psi^c} \Psi) \right],

1027: \end{equation*}

1028: %

1029: where $s$ is a dimensionful coupling constant and the charge

1030: conjugate field is defined as $\Psi^c = \Gamma^2\Gamma^4\Psi^*$.

1031: We choose the kink-fermion coupling $d>0$ such that there is at

1032: least one bound mode and follow the thin kink analysis performed

1033: above.  All the massive fermion modes are frozen out and the

1034: $\Xi$ coupling term becomes

1035: %

1036: \begin{equation}

1037: \label{eq:delta-couple}

1038: s \, \Xi (\overline\Psi \Psi^c + \overline{\Psi^c} \Psi) \xrightarrow{m\rightarrow\infty} s \, (f_{L0})^2 \Xi (\overline{\psi_{L0}} \psi_{L0}^c + \overline{\psi_{L0}^c} \psi_{L0}).

1039: \end{equation}

1040: %

1041: The extra dimensional factor from the fermion zero mode becomes a

1042: delta distribution in the thin kink limit

1043: %

1044: \begin{equation*}

1045: (f_{L0})^2 =

1046:     \frac{m \, \Gamma(d+\half)}{\sqrt{2\pi} \, \Gamma(d)} \cosh^{-2d} \left( \frac{m w}{\sqrt2} \right)

1047:     \xrightarrow{m\rightarrow\infty} \delta(w).

1048: \end{equation*}

1049: %

1050: The $w$ integral can then be performed over the Yukawa term given

1051: by~\eqref{eq:delta-couple} which reduces the action to

1052: %

1053: \begin{equation*}

1054: \begin{aligned}

1055: \mcs_\text{all}^\text{4D/5D} &=

1056:     \int \md^4 x \left[

1057:         -\varepsilon_{\phi_c}

1058:         + \overline{\psi_{L0}} \mi \slashed\partial \psi_{L0} 

1059:         - s \, \Xi(w=0) (\overline{\psi_{L0}} \psi_{L0}^c + \overline{\psi_{L0}^c} \psi_{L0}) \right] \\

1060: &\quad

1061:     + \int \md^5 x \left[

1062:         \half \partial^M \Xi \partial_M \Xi \right].

1063: \end{aligned}

1064: \end{equation*}

1065: 

1066: This is a dynamically generated model describing a four dimensional

1067: zero mode $\psi_{L0}$ coupled at the extra dimensional point $w=0$

1068: to a five dimensional field $\Xi$.  The situation can easily be

1069: reversed to have $\Psi$ five dimensional and coupling at $w=0$ to

1070: the four dimensional ground state mode of $\Xi$.  Extensions to

1071: multiple four and five dimensional fields are also easily

1072: obtained.

1073: 

1074: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1075: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1076: 

1077: \section{Conclusion}

1078: \label{sec:conc}

1079: 

1080: Our universe may be embedded in a large extra spatial dimension

1081: and it seems sensible to look for a field theoretic description of

1082: such a scenario.  We would like such a model to explain the

1083: dynamics behind the confinement of higher dimensional fields to

1084: our four dimensional subspace.  We have demonstrated in this paper

1085: that the domain wall defect (the kink) is an attractive candidate

1086: for such dynamical confinement as it can localise not only

1087: massless chiral fermions but also scalars with arbitrary quartic

1088: potentials.

1089: 

1090: We have analysed in detail the kink and its modes and discussed

1091: the behaviour of these degrees of freedom in the limits of no

1092: kink, a thick kink and a thin kink.  Due to the quartic potential

1093: which sets up the kink profile, the Kaluza-Klein spectrum consists

1094: of a massless bound mode, a massive bound mode and a massive

1095: continuum.  In the thin kink limit all the massive modes freeze

1096: out as their mass becomes infinite.  The zero mode also freezes

1097: out in this limit due to its divergent quartic self coupling.

1098: This leaves just the kink energy density in the effective four

1099: dimensional action.

1100: 

1101: A full analysis of a second scalar field coupled to the kink was

1102: performed.  We showed that this field is trapped by the kink in a

1103: symmetric modified P\"oschl-Teller potential well and we gave the

1104: mass spectrum for this trapping.  In the thin kink limit, this

1105: extra scalar field could either freeze out completely, be a free

1106: five dimensional field, or have just its ground state mode in an

1107: effective four dimensional action.  We showed that in this latter

1108: case, an arbitrary quartic potential could be generated for this

1109: mode and it could thus be used as a standard model electroweak

1110: Higgs field.

1111: 

1112: Our final analysis was that of a fermion coupled to the kink.  It

1113: has been known for some time that such a model admits a massless

1114: chiral mode.  We generalised this result by giving the full

1115: spectrum of fermion modes in the presence of the kink.  These

1116: four dimensional fields consist of a left-handed zero mode, a

1117: certain number of pairs of left- and right-handed modes and a

1118: continuum of pairs.  In the thin kink limit, we showed that the

1119: fermion can either freeze out completely, be five dimensional with

1120: or without a mass term, or reproduce the result of a localised

1121: four dimensional, massless, left-handed mode.  We also

1122: demonstrated that with the kink and fermion and an extra scalar

1123: field, one can arrange things such that the left-handed zero mode

1124: couples to a five dimensional field at a single point in the extra

1125: dimension.

1126: 

1127: The results of the present work gives one the tools necessary to

1128: write down the standard model, without gauge fields, on a brane in

1129: five dimensions.  With the details of the spectrum of the

1130: decomposed five dimensional fields, one can compute the

1131: phenomenology of the interactions between the massless and

1132: massive modes and also use experimental data to put an upper bound

1133: on the width of the brane.  Further work would focus on the

1134: addition of gauge fields and their localisation, as well as the

1135: inclusion of gravity.

1136: 

1137: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1138: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1139: 

1140: \begin{acknowledgments}

1141: We thank Mark Trodden for very fruitful discussions.

1142: DPG was supported by the Puzey bequest to the University of

1143: Melbourne and RRV by the Australian Research Council.

1144: \end{acknowledgments}

1145: 

1146: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

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

1148: 

1149: \appendix

1150: 

1151: \numberwithin{equation}{section}

1152: 

1153: \section{Symmetric modified P\"oschl-Teller potential mode solutions}

1154: \label{app:smpt}

1155: 

1156: In this appendix we give analytic solutions to the symmetric

1157: modified P\"oschl-Teller potential.  The non-symmetric potential

1158: was first studied by Rosen and Morse in the context of molecular

1159: dynamics~\cite{rosen1932}.  They presented solutions in terms

1160: of hypergeometric functions.  Later work by Nieto~\cite{nieto1978}

1161: computed explicit forms of the bound mode solutions, including

1162: normalisation coefficients, in terms of regular functions.

1163: Rajaraman~\cite{rajaraman1982} gives unnormalised solutions for

1164: the bound and continuum modes for a specific case of the

1165: potential, and the hypergeometric forms of the solutions are

1166: again explored in~\cite{hung2004}.  Bound state solutions are

1167: also expressed in terms of Gegenbauer polynomials

1168: in~\cite{decastro2006}.  As far as we are aware, exact closed form

1169: solutions for the continuum modes and their normalisation factors

1170: have not been presented in the literature.  We present these forms

1171: in this Appendix, along with simple expressions for the bound

1172: states, simple recurrence relations for higher modes,

1173: normalisation coefficients, and the closure relation.

1174: 

1175: The time-independent Schr\"odinger equation with the symmetric

1176: version of the potential takes the form

1177: %

1178: \begin{equation}

1179: \label{eq:smpt}

1180: \left( -\frac{\md^2}{\md x^2} + l(l+1) \tanh^2 x - l \right) \psi_n = E_n \psi_n.

1181: \end{equation}

1182: %

1183: If $l=0$ then the solutions are just plane waves.  For $l>0$ there

1184: are a set of bound modes followed by continuum modes.  The bound

1185: solutions are

1186: %

1187: \begin{align*}

1188: E_0^l &= 0 &

1189: \psi_0^l(x) &= A_0^l \cosh^{-l} x, \\

1190: E_1^l &= 2l-1 &

1191: \psi_1^l(x) &= A_1^l \sinh x \cosh^{-l} x, \\

1192: E_2^l &= 4l-4 &

1193: \psi_2^l(x) &= A_2^l \left( \frac{2l-2}{2l-1} \cosh^{-l+2} x - \cosh^{-l} x \right), \\

1194: \vdots \\

1195: E_n^l &= 2nl-n^2 &

1196: \psi_n^l(x) &= \frac{1}{\sqrt{E_n^l}} \left( l \tanh x - \frac{\md}{\md x} \right) \psi_{n-1}^{l-1}(x).

1197: \end{align*}

1198: %

1199: The square integrable ortho-normalisation condition is

1200: %

1201: \begin{equation}

1202: \label{eq:norm-sqint}

1203: \int_{-\infty}^{\infty} \psi_n^l(x) \psi_{n'}^l(x) \; \md x = \delta_{n n'},

1204: \end{equation}

1205: %

1206: and the normalisation coefficients are

1207: %

1208: \begin{align*}

1209: A_0^l &= \sqrt\frac{\Gamma(l+\half)}{\sqrt{\pi} \, \Gamma(l)}, &

1210: A_1^l &= \sqrt{2l-2} A_0^l, &

1211: A_2^l &= \sqrt{(2l-1)(l-2)} A_0^l.

1212: \end{align*}

1213: %

1214: These bound mode solutions are valid for for all positive real

1215: values of $l$.  There are $\lceil l \rceil$ bound

1216: modes\footnotemark[2] and so the mode index takes the values

1217: $n = 0,1,\ldots,\lceil l-1 \rceil$.  For the continuum we have

1218: found forms for the solutions in terms of regular functions for

1219: the case where $l$ is a positive integer.  Instead of the discrete

1220: bound mode index $n$, the continuum is indexed with a continuous

1221: label $p\in\mbr$.  They take the form

1222: \begin{align*}

1223: E_p^1 &= p^2 + 1 &

1224: \psi_p^1(x) &= A_p^1 \me^{\mi p x} \left( \tanh x - \mi p \right), \\

1225: E_p^2 &= p^2 + 4 &

1226: \psi_p^2(x) &= A_p^2 \me^{\mi p x} \left( 3 \tanh^2 x - (p^2 + 1) - 3 \mi p \tanh x \right), \\

1227: \vdots \\

1228: E_p^l &= p^2 + l^2 &

1229: \psi_p^l(x) &= \frac{1}{\sqrt{E_p^l}} \left( l \tanh x - \frac{\md}{\md x} \right) \psi_p^{l-1}(x).

1230: \end{align*}

1231: %

1232: The delta distribution ortho-normalisation condition is

1233: %

1234: \begin{equation}

1235: \label{eq:norm-delta}

1236: \int_{-\infty}^{\infty} \psi_p^l(x) \psi_{p'}^l(x)^* \; \md x = \delta(p-p'),

1237: \end{equation}

1238: %

1239: and the normalisation coefficients are

1240: %

1241: \begin{align*}

1242: A_p^1 &= \frac{1}{\sqrt{2\pi}} \frac{1}{\sqrt{p^2+1}}, &

1243: A_p^2 &= \frac{1}{\sqrt{p^2+4}} A_p^1.

1244: \end{align*}

1245: %

1246: These continuum modes are valid only for $l=1,2,3,...$; for other

1247: values of $l$ one must resort to the hypergeometric form.  Note

1248: that the bound modes are orthogonal to the continuum modes.

1249: Finally, we state the closure relation:

1250: %

1251: \begin{equation*}

1252: \sum_{n=0}^{\lceil l-1 \rceil} \psi_n^l(x) \psi_n^l(x')

1253:   + \int_{-\infty}^{\infty} \psi_p^l(x) \psi_p^l(x')^* \; \md p = \delta(x-x').

1254: \end{equation*}

1255: 

1256: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1257: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1258: 

1259: \section{Freezing out of the kink translation dynamics}

1260: \label{app:freeze}

1261: 

1262: The thin kink limit, as parameterised by the width going to zero

1263: while keeping the energy density finite, still retains the

1264: continuous translation symmetry of the original thick kink.  It is

1265: unexpected then, that in this limit, the symmetry has no

1266: observable consequence at the effective four dimensional level.

1267: The physical reason for this is that a thin kink is infinitely

1268: stiff and cannot be perturbed; the propagating scalar fields have

1269: been frozen out.  With no dynamical degrees of freedom left, the

1270: translation symmetry is manifest by the zero energy cost of a

1271: specific constant field configuration which has no physical

1272: consequence.  This constant configuration is equivalent to

1273: shifting the mode expansion basis to align it with the translated

1274: kink profile.  We discuss this in detail in the following

1275: sections.

1276: 

1277: \subsection{No dynamical degrees of freedom}

1278: 

1279: We proceed to show that there are no effective four dimensional

1280: degrees of freedom associated with a kink background in the thin

1281: kink limit.  We do this by looking at the action for a general

1282: field expanded about the kink, and show that the potential of such

1283: a field has coefficients which tend to infinity in the thin kink

1284: limit.

1285: 

1286: Because the kink modes $\eta_{0,1,q}$ derived in

1287: Section~\ref{sec:kink} form a basis of extra dimensional profiles,

1288: the corresponding fields $\phi_{0,1,q}$ form a basis of four

1289: dimensional scalar degrees of freedom.  So any dynamical

1290: behaviour of the kink that manifests itself in four dimensions as

1291: a scalar field, must be some linear combination of the fields

1292: $\phi_{0,1,q}$.  Let this arbitrary kink degree of freedom be

1293: $\psi_0(x^\mu)$ so we can write

1294: %

1295: \begin{equation}

1296: \label{eq:psi-comb}

1297: \psi_0 = \alpha_0 \phi_0 + \beta_0 \phi_1 + \int \md q [\gamma_0^*(q) \phi_q].

1298: \end{equation}

1299: 

1300: To determine the effective action for $\psi_0$ we will need to

1301: invert Eq.~\eqref{eq:psi-comb} to express $\phi_{0,1,q}$ in

1302: terms of $\psi_0$.  To do this properly, we need a complete

1303: basis which includes $\psi_0$; call it $\psi_{0,n,q}$, where

1304: $n$ labels the extra bound modes and $q$ labels the extra

1305: continuum modes needed to complete the basis.  Then the inverse

1306: of~\eqref{eq:psi-comb} is

1307: %

1308: \begin{align*}

1309: \phi_0 &= \alpha_0 \psi_0

1310:     + \sum_n \alpha_n \psi_n

1311:     + \int \md r [ \alpha_r \psi_r ], \\

1312: \phi_1 &= \beta_0 \psi_0

1313:     + \sum_n \beta_n \psi_n

1314:     + \int \md r [ \beta_r \psi_r ], \\

1315: \phi_q &= \gamma_0(q) \psi_0

1316:     + \sum_n \gamma_n(q) \psi_n

1317:     + \int \md r [ \gamma_r(q) \psi_r ].

1318: \end{align*}

1319: %

1320: Substituting this expansion in the action~\eqref{eq:phi-act-4d}

1321: gives the full action for the new fields $\psi_{0,n,q}$.  We are

1322: only interested in the terms that describe $\psi_0$ as a

1323: self-interacting field that does not couple to any of the other

1324: fields.  Keeping just these terms, the effective action is

1325: %

1326: \begin{align*}

1327: S_{\psi_0} = \int \md^4 x \left[

1328:     \half A^\text{(kin)} \partial^\mu \psi_0 \partial_\mu \psi_0

1329:     - \frac{m^2}{2} A^{(2)} \psi_0^2

1330:     - m \sqrt{a} \, A^{(3)} \psi_0^3 - a A^{(4)} \psi_0^4 \right],

1331: \end{align*}

1332: %

1333: where

1334: %

1335: \begin{align*}

1336: A^\text{(kin)} &= \alpha_0^2 + \beta_0^2 + \int \md q \left[ \gamma_0^*(q) \gamma_0(q) \right], \\

1337: A^{(2)} &= \thhalf \beta_0^2 + \int \md q \left[ \half (q^2 + 4) \gamma_0^*(q) \gamma_0(q) \right], \\

1338: A^{(3)} &= \text{(terms with at least one factor of $\beta_0$ or $\gamma_0(q)$)}, \\

1339: A^{(4)} &= \frac{9 \sqrt2}{140} \alpha_0^4 + \text{(terms with at least one factor of $\beta_0$ or $\gamma_0(q)$)}.

1340: \end{align*}

1341: 

1342: In the thin kink limit\footnote{Recall the limit is

1343: $m\rightarrow\infty$ and $a\rightarrow\infty$  but $m^4/a$

1344: finite.} the mass of $\psi_0$ will tend to infinity unless

1345: $A^{(2)}$ vanishes.

1346: For $A^{(2)}$ to be identically zero, we require $\beta_0 = 0$ and

1347: $\gamma_0(q) = 0$.  Then because $A^\text{(kin)}=1$, we require

1348: $\alpha_0 = 1$ (hence $\psi_0 = \phi_0$) and so the quartic

1349: coefficient is $9 \sqrt2 \, a / 140$ which freezes out $\psi_0$ in

1350: the thin kink limit.  Alternatively, we can choose $\beta_0$

1351: and $\gamma_0(q)$ such that $A^{(2)}$ is not identically zero, but

1352: in the thin kink limit tends to zero in order to keep

1353: $m^2 A^{(2)}$ finite.  Then in this limit

1354: $\beta_0 \rightarrow 0$ and $\gamma_0(q) \rightarrow 0$ and

1355: so $\alpha_0 \rightarrow 1$.  The quartic coefficient $A^{(4)}$ is

1356: then dominated by the $\alpha_0^4$ term which, as before, tends to

1357: infinity in the thin kink limit and freezes out $\psi_0$.  Thus

1358: there is no linear combination of the fields $\phi_{0,1,q}$, and

1359: hence no scalar degree of freedom, whose dynamics survive in the

1360: thin kink limit.

1361: 

1362: \subsection{Manifestation of translation symmetry}

1363: 

1364: The energy density $\varepsilon_{\phi_c}$ of the kink

1365: configuration $\phi_c(w)$ is independent of the location of the

1366: kink along the $w$-dimension.  Thus, for a linear combination of

1367: the basis $\eta_{0,1,q}$ which just translates the kink profile,

1368: the corresponding linear combination of $\phi_{0,1,q}$ should have

1369: zero energy density.  This statement needs to be made more

1370: precise, as the previous section shows that there is no linear

1371: combination of $\phi_{0,1,q}$ that has a vanishing potential.

1372: Instead, there is a linear combination with a potential that

1373: vanishes only at a specific value of the field, corresponding to a

1374: shift of the kink.

1375: 

1376: Consider the expansion

1377: %

1378: \begin{equation}

1379: \label{eq:phi-expand-shift}

1380: \Phi(x^M) = \phi_c(w) + \psi_0(x^\mu) s_0(w),

1381: \end{equation}

1382: %

1383: where $s_0$ is the profile of a finite shift of the kink.  For a

1384: given fixed shift of the kink $\Delta w$, we want to have the

1385: relation

1386: %

1387: \begin{equation}

1388: \label{eq:shift-defn}

1389: \psi_0(x^\mu) s_0(w) = \phi_c(w+\Delta w) - \phi_c(w).

1390: \end{equation}

1391: %

1392: For this to be true, $\psi_0$ must be an arbitrary constant,

1393: which we fix by normalising the profile

1394: $\int \md w \, s_0^2(w) = 1$.  This gives

1395: %

1396: \begin{equation*}

1397: s_0(w) = \sqrt{m} \, S \sigma_0(w),

1398: \end{equation*}

1399: %

1400: with

1401: %

1402: \begin{align*}

1403: & S = \frac{1}{2^{5/4}} \sqrt{\frac{\tanh \Delta z}{\Delta z - \tanh \Delta z}}

1404: & \text{and}

1405: && \sigma_0(w) = \frac{\tanh \Delta z \, \cosh^{-2} z}{1 + \tanh \Delta z \, \tanh z}.

1406: \end{align*}

1407: %

1408: We have $z = mw/\sqrt2$ and $\Delta z = m \Delta w / \sqrt2$. For

1409: the combination $\psi_0 s_0$ to actually shift $\phi_c$ by an

1410: amount $\Delta w$, as given by~\eqref{eq:shift-defn}, it is

1411: required that

1412: %

1413: \begin{equation}

1414: \label{eq:shift-psi}

1415: \psi_0(x^\mu) = \frac{m}{\sqrt a \, S}.

1416: \end{equation}

1417: 

1418: We now want to treat the field $\psi_0$ as a dynamical four

1419: dimensional scalar degree of freedom.  It will correspond to

1420: finite translations of the kink profile and we are interested in

1421: its behaviour in the thin kink limit.  One can write $\psi_0$ as

1422: a linear combination of $\phi_{0,1,q}$, which can be determined

1423: from the linear combination needed to write $s_0$ in terms of

1424: $\eta_{0,1,q}$.  It is then possible to determine the effective

1425: action for $\psi_0$ by substituting this combination in the

1426: Lagrangian~\eqref{eq:phi-lag-4d}.  A simpler way to obtain the

1427: same result is to substitute~\eqref{eq:phi-expand-shift} into the

1428: original five dimensional action~\eqref{eq:phi-act-5d} and

1429: integrate out the extra dimension.  The result is

1430: %

1431: \begin{equation*}

1432: \mcs_{\psi_0} = \int \md^4 x \left[

1433:     -\varepsilon_{\phi_c} 

1434:     + \half \partial^\mu \psi_0 \partial_\mu \psi_0

1435:     - V_{\psi_0} (\psi_0) \right],

1436: \end{equation*}

1437: %

1438: with the effective potential

1439: %

1440: \begin{equation*}

1441: V_{\psi_0} (\psi_0) = \left( \frac{5}{2 \tanh^2 \Delta z} - \frac{5 \tanh \Delta z}{6 (\Delta z - \tanh \Delta z)} - \frac{3}{2} \right) \psi_0^2 \left( \sqrt{a} S \psi_0 - m \right)^2.

1442: \end{equation*}

1443: 

1444: Note that the potential for $\psi_0$ has two minima: one at

1445: $\psi_0=0$ and the other at precisely the value corresponding to

1446: the finite translation, given by~\eqref{eq:shift-psi}.  The

1447: energy density associated with $\psi_0$ is

1448: %

1449: \begin{equation}

1450: \label{eq:shift-psi-energy}

1451: \varepsilon_{\psi_0} = \half (\partial_t \psi_0)^2 + \half (\nabla \psi_0)^2 + V_{\psi_0}(\psi_0).

1452: \end{equation}

1453: %

1454: Any dynamical behaviour of $\psi_0$ has non-zero, positive energy

1455: density.  This energy density will vanish only when $\psi_0=0$ or

1456: $\psi_0$ is the constant~\eqref{eq:shift-psi} at all locations in

1457: the four dimensional space-time.  This is the manifestation of the

1458: translation symmetry of the kink.  The field $\psi_0$, or

1459: equivalently a specific linear combination of $\phi_{0,1,q}$,

1460: can assume this constant value at each point $x^\mu$ with zero

1461: energy cost.  It is possible to show that the fields assuming this

1462: value is equivalent to shifting the original basis $\eta_{0,1,q}$

1463: by the amount $\Delta w$.

1464: 

1465: In the thin kink limit, the potential $V_{\psi_0}$ becomes

1466: infinitely steep and the dynamics of $\psi_0$ are frozen out.  The

1467: only remnant of the translation symmetry is the fact that $\psi_0$

1468: is allowed to be a certain constant value everywhere with no

1469: energy cost.  But this has no effect on the four dimensional

1470: effective action and so the translation symmetry is hidden at this

1471: level.

1472: 

1473: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1474: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1475: 

1476: %\bibliography{kme}

1477: 

1478: \begin{thebibliography}{17}

1479: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

1480: \expandafter\ifx\csname bibnamefont\endcsname\relax

1481:   \def\bibnamefont#1{#1}\fi

1482: \expandafter\ifx\csname bibfnamefont\endcsname\relax

1483:   \def\bibfnamefont#1{#1}\fi

1484: \expandafter\ifx\csname citenamefont\endcsname\relax

1485:   \def\citenamefont#1{#1}\fi

1486: \expandafter\ifx\csname url\endcsname\relax

1487:   \def\url#1{\texttt{#1}}\fi

1488: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi

1489: \providecommand{\bibinfo}[2]{#2}

1490: \providecommand{\eprint}[2][]{\url{#2}}

1491: 

1492: \bibitem[{\citenamefont{Appelquist et~al.}(2001)\citenamefont{Appelquist,

1493:   Cheng, and Dobrescu}}]{appelquist2000}

1494: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Appelquist}},

1495:   \bibinfo{author}{\bibfnamefont{H.-C.} \bibnamefont{Cheng}}, \bibnamefont{and}

1496:   \bibinfo{author}{\bibfnamefont{B.~A.} \bibnamefont{Dobrescu}},

1497:   \bibinfo{journal}{Phys. Rev. D} \textbf{\bibinfo{volume}{64}},

1498:   \bibinfo{pages}{035002} (\bibinfo{year}{2001}), \eprint{hep-ph/0012100}.

1499: 

1500: \bibitem[{\citenamefont{Arkani-Hamed et~al.}(1998)\citenamefont{Arkani-Hamed,

1501:   Dimopoulos, and Dvali}}]{add}

1502: \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Arkani-Hamed}},

1503:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Dimopoulos}},

1504:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Dvali}},

1505:   \bibinfo{journal}{Phys. Lett. B} \textbf{\bibinfo{volume}{429}},

1506:   \bibinfo{pages}{263} (\bibinfo{year}{1998}), \bibinfo{note}{hep-ph/9803315}.

1507: 

1508: \bibitem[{\citenamefont{Antoniadis}(1990)}]{antoniadis1990}

1509: \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Antoniadis}},

1510:   \bibinfo{journal}{Phys. Lett. B} \textbf{\bibinfo{volume}{246}},

1511:   \bibinfo{pages}{377} (\bibinfo{year}{1990}).

1512: 

1513: \bibitem[{\citenamefont{Antoniadis et~al.}(1998)\citenamefont{Antoniadis,

1514:   Arkani-Hamed, Dimopoulos, and Dvali}}]{aadd}

1515: \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Antoniadis}},

1516:   \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Arkani-Hamed}},

1517:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Dimopoulos}},

1518:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Dvali}},

1519:   \bibinfo{journal}{Phys. Lett. B} \textbf{\bibinfo{volume}{436}},

1520:   \bibinfo{pages}{257} (\bibinfo{year}{1998}), \bibinfo{note}{hep-ph/9804398}.

1521: 

1522: \bibitem[{\citenamefont{Randall and Sundrum}(1999{\natexlab{a}})}]{rs1}

1523: \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Randall}} \bibnamefont{and}

1524:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Sundrum}},

1525:   \bibinfo{journal}{Phys. Rev. Lett} \textbf{\bibinfo{volume}{83}},

1526:   \bibinfo{pages}{3370} (\bibinfo{year}{1999}{\natexlab{a}}),

1527:   \bibinfo{note}{hep-ph/9905221}.

1528: 

1529: \bibitem[{\citenamefont{Randall and Sundrum}(1999{\natexlab{b}})}]{rs2}

1530: \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Randall}} \bibnamefont{and}

1531:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Sundrum}},

1532:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{83}},

1533:   \bibinfo{pages}{4690} (\bibinfo{year}{1999}{\natexlab{b}}),

1534:   \bibinfo{note}{hep-th/9906064}.

1535: 

1536: \bibitem[{\citenamefont{{Rubakov} and {Shaposhnikov}}(1983)}]{rubakov1983}

1537: \bibinfo{author}{\bibfnamefont{V.~A.} \bibnamefont{{Rubakov}}}

1538:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{M.~E.}

1539:   \bibnamefont{{Shaposhnikov}}}, \bibinfo{journal}{Phys. Lett. B}

1540:   \textbf{\bibinfo{volume}{125}}, \bibinfo{pages}{136} (\bibinfo{year}{1983}).

1541: 

1542: \bibitem[{\citenamefont{Akama}(1982)}]{akama1982}

1543: \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Akama}},

1544:   \bibinfo{journal}{Lecture Notes in Physics, Berlin Springer Verlag}

1545:   \textbf{\bibinfo{volume}{176}}, \bibinfo{pages}{267} (\bibinfo{year}{1982}).

1546: 

1547: \bibitem[{\citenamefont{Visser}(1985)}]{visser1985}

1548: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Visser}},

1549:   \bibinfo{journal}{Phys. Lett. B} \textbf{\bibinfo{volume}{159}},

1550:   \bibinfo{pages}{22} (\bibinfo{year}{1985}), \eprint{hep-th/9910093}.

1551: 

1552: \bibitem[{\citenamefont{Dvali and Shifman}(1997)}]{dvali1996}

1553: \bibinfo{author}{\bibfnamefont{G.~R.} \bibnamefont{Dvali}} \bibnamefont{and}

1554:   \bibinfo{author}{\bibfnamefont{M.~A.} \bibnamefont{Shifman}},

1555:   \bibinfo{journal}{Phys. Lett. B} \textbf{\bibinfo{volume}{396}},

1556:   \bibinfo{pages}{64} (\bibinfo{year}{1997}), \eprint{hep-th/9612128}.

1557: 

1558: \bibitem[{\citenamefont{Shaposhnikov et~al.}(2005)\citenamefont{Shaposhnikov,

1559:   Tinyakov, and Zuleta}}]{shaposhnikov2005}

1560: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Shaposhnikov}},

1561:   \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Tinyakov}}, \bibnamefont{and}

1562:   \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Zuleta}},

1563:   \bibinfo{journal}{JHEP} \textbf{\bibinfo{volume}{0509}}, \bibinfo{pages}{062}

1564:   (\bibinfo{year}{2005}), \bibinfo{note}{hep-th/0508102}.

1565: 

1566: \bibitem[{\citenamefont{Ringeval et~al.}(2002)\citenamefont{Ringeval, Peter,

1567:   and Uzan}}]{ringeval2002}

1568: \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Ringeval}},

1569:   \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Peter}}, \bibnamefont{and}

1570:   \bibinfo{author}{\bibfnamefont{J.~P.} \bibnamefont{Uzan}},

1571:   \bibinfo{journal}{Phys. Rev. D} \textbf{\bibinfo{volume}{65}},

1572:   \bibinfo{pages}{044016} (\bibinfo{year}{2002}).

1573: 

1574: \bibitem[{\citenamefont{Hung and Tran}(2004)}]{hung2004}

1575: \bibinfo{author}{\bibfnamefont{P.~Q.} \bibnamefont{Hung}} \bibnamefont{and}

1576:   \bibinfo{author}{\bibfnamefont{N.-K.} \bibnamefont{Tran}},

1577:   \bibinfo{journal}{Phys. Rev. D} \textbf{\bibinfo{volume}{69}},

1578:   \bibinfo{pages}{064003} (\bibinfo{year}{2004}).

1579: 

1580: \bibitem[{\citenamefont{{de Castro} and Hott}(2006)}]{decastro2006}

1581: \bibinfo{author}{\bibfnamefont{A.~S.} \bibnamefont{{de Castro}}}

1582:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Hott}},

1583:   \bibinfo{journal}{Phys. Lett. A} \textbf{\bibinfo{volume}{351}},

1584:   \bibinfo{pages}{379} (\bibinfo{year}{2006}).

1585: 

1586: \bibitem[{\citenamefont{{Rosen} and {Morse}}(1932)}]{rosen1932}

1587: \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{{Rosen}}} \bibnamefont{and}

1588:   \bibinfo{author}{\bibfnamefont{P.~M.} \bibnamefont{{Morse}}},

1589:   \bibinfo{journal}{Phys. Rev.} \textbf{\bibinfo{volume}{42}},

1590:   \bibinfo{pages}{210} (\bibinfo{year}{1932}).

1591: 

1592: \bibitem[{\citenamefont{{Nieto}}(1978)}]{nieto1978}

1593: \bibinfo{author}{\bibfnamefont{M.~M.} \bibnamefont{{Nieto}}},

1594:   \bibinfo{journal}{Phys. Rev. A} \textbf{\bibinfo{volume}{17}},

1595:   \bibinfo{pages}{1273} (\bibinfo{year}{1978}).

1596: 

1597: \bibitem[{\citenamefont{Rajaraman}(1982)}]{rajaraman1982}

1598: \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Rajaraman}},

1599:   \emph{\bibinfo{title}{Solitons and instantons : an introduction to solitons

1600:   and instantons in quantum field theory}} (\bibinfo{publisher}{North-Holland

1601:   Pub. Co.}, \bibinfo{year}{1982}).

1602: 

1603: \end{thebibliography}

1604: \end{document}

1605: