1: \documentclass[aps, preprint]{revtex4}

2: \usepackage{latexsym}

3: \usepackage{epsfig}

4: %\usepackage{graphicx}

5: 

6: \begin{document}

7: 

8: \title{Effective theory approach to brane world black holes}

9: \author{Paul L. McFadden}

10: \email{p.l.mcfadden@damtp.cam.ac.uk}

11: \author{Neil G. Turok}

12: \email{n.g.turok@damtp.cam.ac.uk}

13: \affiliation{D.A.M.T.P., C.M.S., Wilberforce Road, 

14:   Cambridge, CB3 0WA, UK.}

15: \date{\today}

16: 

17: \newcommand{\nc}{\newcommand}

18: \nc{\rnc}{\renewcommand}

19: \nc{\eg}{\textit{e.g. }}

20: \nc{\dx}{\mathrm{d} ^4 x}

21: \nc{\D}{\partial}

22: \rnc{\d}{\mathrm{d}}

23: \nc{\tr}{\mathrm{Tr}}

24: \nc{\gdxdx}{ g_{\mu \nu}(x) \mathrm{d}x^\mu \mathrm{d}x^\nu}

25: \nc{\ndxdx}{ \eta _{\mu \nu}(x) \mathrm{d}x^\mu \mathrm{d}x^\nu}

26: \nc{\tgdxdx}{\tilde{g}_{\mu \nu}(x) \mathrm{d}x^\mu \mathrm{d}x^\nu}

27: \nc{\dxdx}{\mathrm{d}x^\mu \mathrm{d}x^\nu}

28: \nc{\tz}{\tilde{z}}

29: \nc{\ty}{\tilde{y}}

30: \nc{\g}{g_{\mu \nu}}

31: \nc{\gi}{g^{\mu \nu}}

32: \nc{\gh}{\hat{g}_{\mu\nu}}

33: \nc{\psih}{\hat{\psi}}

34: \nc{\Phih}{\hat{\Phi}}

35: \nc{\dPhi}{\delta \Phi}

36: \nc{\dpsi}{\delta\psi}

37: \nc{\Boxhat}{\hat{\Box}}  % Don't forget sqrt{-\hat{g}}'s in text still

38: \nc{\gpm}{g^\pm _{\mu \nu}}

39: \nc{\gp}{g^+ _{\mu \nu}}

40: \nc{\gm}{g^- _{\mu \nu}}

41: \nc{\tg}{\tilde{g}_{\mu \nu}}

42: \rnc{\[}{\begin{equation}}

43: \rnc{\]}{\end{equation}}

44: \nc{\bea}{\begin{eqnarray}}

45: \nc{\eea}{\end{eqnarray}}

46: \nc{\ie}{\textit{i.e. }}

47: \nc{\dw}{\mathrm{d}\Omega _2^2}

48: \rnc{\j}{\mathcal{J}}

49: \rnc{\tt}{\rightarrow} % i.e.  \tt = 'tends to'

50: \rnc{\inf}{\infty}

51: \rnc{\l}{L}

52: \nc{\RN}{Reissner-Nordstr{\"o}m }

53: \rnc{\S}{Schr{\"o}dinger }

54: \nc{\e}{\epsilon}

55: \nc{\te}{\tilde{\e}}

56: 

57: \begin{abstract}

58: 

59: We derive static spherically-symmetric vacuum solutions

60: of the low-energy effective action for the two brane Randall-Sundrum model.

61: The solutions with non-trivial radion belong to a one-parameter family

62: describing traversable wormholes between the branes and a black hole,

63: and were first discovered in the context of Einstein gravity with a 

64: conformally-coupled scalar field.

65: From a brane world perspective, a distinctive feature of all the

66: solutions with non-trivial radion is a 

67: brane intersection about which the bulk geometry is conical but the

68: induced metrics on the branes are regular.  

69: Contrary to earlier claims in the literature, we show these solutions are 

70: stable under monopole perturbations.

71: 

72: 

73: \end{abstract}

74: 

75: \maketitle

76: 

77: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

78: 

79: \section{Introduction}

80: 

81: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

82: 

83: 

84: The brane world provides a fresh perspective on the nature of gravity.

85: We will consider the simplest possible scenario possessing a warped bulk

86: geometry, the two brane Randall-Sundrum model \cite{RS1}.

87: This consists of a pair of four-dimensional positive- and negative-tension $Z_2$ branes 

88: bounding a five-dimensional bulk with negative cosmological constant.

89: In addition to the graviton, the spectrum of low-energy gravitational

90: degrees of freedom includes a massless scalar field, the radion,

91: pertaining to the separation of the brane pair along the extra dimension.

92: The low-energy dynamics may then be described by a four-dimensional

93: effective theory \cite{Ekpyrotic, KS, KSnutshell, Us}, amounting to a specific

94: scalar-tensor theory of gravity. 

95: 

96: In this article, we derive static spherically-symmetric vacuum 

97: solutions of the brane world effective theory.  In addition to the well-known 

98: black string solution with trivial radion \cite{Chamblin}, we find a 

99: one-parameter family of solutions with non-trivial radion.  These solutions

100: describe traversable wormholes between the branes, and a black hole. 

101: For the latter, the geometry induced on the branes is equivalent to that

102: of the extremal \RN solution, even though there is no electromagnetic

103: charge present.   

104: The singularity for this solution is time-like, in contrast to the space-like

105: singularity of the black string, raising new possibilities for the

106: fate of infalling matter and the endstate of gravitational collapse

107: on the brane.

108: The solutions with non-trivial radion are found to possess a brane intersection

109: at a finite radius, about which the bulk geometry is conical but the induced 

110: metrics on the branes are nevertheless regular.

111: 

112: One method of solving the brane world effective theory is to re-cast it in the form 

113: of Einstein gravity with a conformally-coupled scalar field \cite{Us}, for which the

114: static spherically-symmetric vacuum solutions are already known 

115: \cite{BBM, Bekenstein1, Bekenstein2, WH_orig, Barcelo}.  However, here we will instead

116: pursue a more direct method based around a solution-generating symmetry transformation.

117: We start in Section II by considering the simple case of tensionless 

118: branes compactified on an $S^1/Z_2$ orbifold, for which an exact

119: solution of the bulk geometry may be found. 

120: Then in Section III we proceed to the physically relevant scenario of

121: branes with their canonical  

122: Randall-Sundrum tensions.  Finally, in Section IV we analyse the stability of the

123: effective theory solutions under monopole 

124: perturbations.  We show that both the black hole and wormhole

125: solutions are stable, contrary to earlier 

126: claims in the literature on Einstein gravity with a conformally-coupled scalar field. 

127: 

128: 

129: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

130: \section{Tensionless branes}

131: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

132: 

133: For tensionless vacuum branes compactified on an $S^1/Z_2$ orbifold, the bulk warp is

134: absent and so the ground state is independent of the fifth dimension $Y$.

135: Ignoring gauge fields, we introduce the bulk ansatz

136: \[

137: \label{bulk_ansatz}

138: \d s_5^2 = \gdxdx + \Phi^2(x)\d Y^2,

139: \]

140: where $x^\mu$, $\mu=0,1,2,3$, parametrise the four conventional

141: dimensions.  

142: Inserting this ansatz into the five-dimensional pure

143: Einstein-Hilbert action and integrating over $Y$, we obtain the

144: four-dimensional effective action

145: \[

146: \label{M_EFT}

147: S=m_P^2\int \dx \sqrt{-g}\Phi R ,

148: \]

149: where $m_P$ denotes the Planck mass.

150: The corresponding equations of motion are

151: \bea

152: \label{M_eom}

153: \Phi R_{\mu\nu} &=& \nabla_\mu\nabla_\nu \Phi \\

154: \label{Box_Phi}

155: \Box \Phi &=& 0 .

156: \eea

157: Since the branes are located at constant $Y$, the metric $\g$

158: appearing in the effective theory may be identified with 

159: the induced metric on the branes.

160: 

161: Any five-dimensional Ricci-flat metric provides a solution of the 

162: four-dimensional effective theory.

163: For example, starting with the five-dimensional black string, we obtain 

164: the four-dimensional Schwarzschild solution with constant $\Phi$.

165: Alternatively, we could take the product of Euclidean Schwarzschild and a

166: flat time dimension as our five-dimensional metric.  This gives a four-dimensional

167: effective theory solution with non-trivial radion:

168: \bea

169: \label{g_soln}

170: \d s_4^2 &=&  -\d t^2 +(1-\frac{2m}{R})^{-1}\d R^2+R^2\dw \\

171: \label{Phi_soln}

172: \Phi &=& \sqrt{1-\frac{2m}{R}}  .

173: \eea

174: At $R=2M$, the size of the fifth dimension shrinks to zero

175: and the branes intersect.

176: The bulk geometry about this point is conical, as may be seen by

177: setting $R=2m+\e^2/8m$ and expanding to leading order.  This gives

178: \[

179: \d s_5^2 = -\d t^2 +  4m^2 \dw + \d \e^2  +\frac{\e^2}{16m^2}\d Y^2.

180: \]

181: For $2(Y^+ -Y^-)=8\pi m$ the geometry is regular at $R=2m$, however 

182: in general there is a conical singularity.

183: Nevertheless, all the curvature components remain finite and a straightforward

184: analytic continuation to $\e<0$ is feasible.  The bulk geometry then describes

185: a double cone wherein the fifth dimension collapses down to zero size before 

186: opening up again on the other side.

187: 

188: From the perspective of the four-dimensional effective theory, this

189: amounts to an analytic continuation of $\Phi$ to values less than

190: zero.  Changing coordinates to

191: $R=m(1+y/2+1/2y)$, (\ref{g_soln}) and (\ref{Phi_soln}) become

192: \bea

193: \label{M_metric}

194: \d s^2 &=& -\d t^2 + \frac{M^2}{y^4}(1+y)^4 (\d y^2 +

195: y^2 \dw ) \\

196: \label{M_Phi_soln}

197: \Phi &=& \frac{1-y}{1+y} ,

198: \eea

199: where $M=m/2$ and $y$ takes values in the range $0\le y<\inf$.  The brane

200: intersection has been mapped to $y=1$.

201: As $\Phi$ passes through zero, the effective action

202: (\ref{M_EFT}) changes sign but the equations of motion remain

203: unchanged. 

204: 

205: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

206: \section{Randall-Sundrum Branes}

207: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

208: 

209: For Randall-Sundrum branes, the effective action may also be written in 

210: the form (\ref{M_EFT}) (see Appendix A).  However, due to the presence

211: of a bulk warp, the induced metrics on the positive-tension (plus) and 

212: negative-tension (minus) branes are no longer identical.  

213: Instead, they are given in terms of the effective theory metric $\g$ and the

214: radion $\Phi$ by

215: \[

216: \label{new_g_eqns}

217: \g^{(1)}=\frac{1}{4}(1+\Phi)^2\g , \ \ \ \ \ \ \ 

218: \g^{(2)}=\frac{1}{4}(1-\Phi)^2\g ,

219: \]

220: where brane one is the plus brane and brane two the minus brane for

221: $\Phi>0$, and the converse holds for $\Phi<0$.  At a brane intersection, 

222: $\Phi$ changes sign and a plus brane is continued into a minus brane,

223: and vice versa.  In this fashion, the direction of the bulk warp is 

224: preserved.

225: 

226: The action (\ref{M_EFT}) has the symmetry

227: \[

228: \g \tt \beta \g, \ \ \  \ \ \ \ \Phi \tt \Phi/\beta,

229: \]

230: for any constant $\beta>0$.  In the tensionless case, this just amounts to a 

231: trivial rescaling of the coordinates.  However, in the Randall-Sundrum case, we see

232: from (\ref{new_g_eqns}) that this symmetry has a non-trivial effect on the brane metrics 

233: (provided $\Phi$ is itself non-trivial).

234: Thus, given any one solution of the effective action (\ref{M_EFT}) with non-trivial $\Phi$, 

235: we may generate a full one-parameter family of solutions.

236: 

237: Applying this procedure to (\ref{M_metric}) and (\ref{M_Phi_soln}), 

238: we find the one-parameter family of solutions

239: \bea

240: \label{BM_metric}

241: \d s^2 &=& \beta \left(-\d t^2 + \frac{M^2}{y^4}(1+y)^4 (\d y^2 +

242: y^2 \dw )\right) \\

243: \label{BM_Phi_soln}

244: \Phi &=& \frac{1}{\beta}\left(\frac{1-y}{1+y}\right) .

245: \eea

246: The metric on the first brane is 

247: \[

248: \label{WH}

249: \d s^2 = \frac{(1+y_0 y)^2}{1-y_0 ^2}\left( -\frac{\d t^2}{(1+y)^2} + \frac{M^2}{y^4}(1+y)^2(\d y^2 + y^2 \dw)\right) ,

250: \]

251: where $y_0 = (\beta-1)/(\beta+1)$, and satisfies $|y_0|<1$.

252: It is easy to see that the metric on the second brane is identical to that on the first, up to an inversion

253: of the $y$ coordinate:  this is because the effective theory metric (\ref{BM_metric}) is preserved under $y\tt1/y$, 

254: whereas from (\ref{BM_Phi_soln}), $\Phi\tt -\Phi$.  Hence from (\ref{new_g_eqns}), a $y$ coordinate

255: inversion transforms brane one into brane two, and vice versa.

256: 

257: The nature of the solution is determined by the parameter $\beta$:

258: for $\beta>1$, the conformal factor $(1\pm\Phi)^2/4$ is always greater

259: than zero, and so the causal structure of the brane metrics is

260: identical to that of (\ref{M_metric}). 

261: For $\beta \le 1$ however, the conformal factor possesses zeroes and

262: so we obtain different causal structures.

263: 

264: Explicitly, for $0<\beta<1$ ($y_0<0$), the solution (\ref{WH}) is defined in the range $0<y<1/|y_0|$ and

265: describes a naked singularity at $y=1/|y_0|$.  As we are

266: interested in solutions that are regular we will not pursue this case

267: further.  

268: For $\beta>1$ ($y_0>0$), the solution is defined in the range $0\le y<\inf$, and

269: describes a traversable wormhole in which the flat asymptotic region 

270: at $y=0$ is joined to a second such region at $y\tt\inf$ by an Einstein-Rosen

271: throat at $y=1$ with no event horizon.

272: In the case where $\beta=1$ ($y_0=0$), it is useful to substitute $y=M/(r-M)$ giving

273: \[

274: \label{BBMB}

275: \d s^2 = -(1-\frac{M}{r})^2 \d t^2 + (1-\frac{M}{r})^{-2} \d r^2 +

276: r^2 \dw .

277: \]

278: This geometry is equivalent to that of the extremal \RN black hole, even 

279: though no electric charge is present.  

280: The Carter-Penrose diagram is given in Figure (\ref{Penrose}).

281: \begin{figure}[t]

282: \begin{center}

283: \includegraphics[width=4cm, keepaspectratio=]{RN.eps}

284: \end{center}

285: \caption{Conformal diagram for the BBMB black hole.  Dotted lines

286:   represent trajectories where $r=\mathrm{const.}$  The vertical

287:   dashed line represents the central time-like singularity.}

288: \label{Penrose}

289: \end{figure}

290: 

291: Historically, the solutions (\ref{WH}) and (\ref{BBMB}) were first discovered in the context

292: of Einstein gravity with a conformally-coupled scalar field; in particular,

293: the black hole solution (\ref{BBMB}) is known as the BBMB black hole after its

294: discoverers Bocharova, Bronnikov, Melnikov \cite{BBM}, and independently

295: Bekenstein \cite{Bekenstein1, Bekenstein2}.

296: The traversable wormhole solution was first derived in \cite{WH_orig} and has been discussed 

297: more recently in \cite{Barcelo}.

298: That the same solutions re-appear in the present context simply reflects the fact

299: that it is possible to re-write the brane world effective theory in a different conformal 

300: gauge (see Appendix A), in which it takes the form of gravity with a conformally-coupled scalar field.

301: The conformal scalar field is given by

302: \[ 

303: \label{psi_soln}

304: \psi = \sqrt{6}\left(\frac{1-\Phi}{1+\Phi}\right)=

305: \sqrt{6}\left(\frac{y+y_0}{1+yy_0}\right), 

306: \] 

307: and so takes the form

308: \[

309: \label{BBMB_psi}

310: \psi=\sqrt{6}\frac{M}{r-M}

311: \]

312: for the BBMB black hole.

313: However, deriving the solutions of gravity with a conformally-coupled scalar involves

314: performing a conformal transformation on the solutions of gravity with a \textit{minimally-coupled}

315: scalar, a more involved calculation than the solution-generating method presented

316: above.

317: 

318: We now turn to consider the bulk geometry of these solutions.  In the case of branes

319: with tension, it is not possible to reconstruct the exact bulk geometry 

320: from a solution of the four-dimensional effective theory alone.  However, 

321: the behaviour of the radion does provide us with an understanding of

322: the interbrane separation.  (Specifically, this is given by

323: $d=2L\tanh^{-1}{|\Phi|}$).  In the

324: case of the traversable wormhole ($\beta>1$), where $\Phi \tt \pm 1/\beta$ for the flat asymptotics

325: $y\tt 0$ and $y\tt\inf$, we see that the interbrane separation is asymptotically constant.

326: In the case of the BBMB black hole ($\beta=1$), the interbrane separation becomes infinite

327: both at radial infinity ($y=0$) and at the event horizon $r=M$ ($y=1$).

328: From a brane world perspective, this of course presents no phenomenological difficulties.  Similarly,

329: in the context of gravity with a conformally-coupled scalar field, Bekenstein has argued that the 

330: divergence of the scalar field (\ref{BBMB_psi}) at the event horizon does not lead to 

331: any physical pathologies \cite{Bekenstein2}.

332: 

333: For both the wormhole and the BBMB solution, $\Phi$ vanishes at $y=1$, indicating that

334: the fifth dimension has shrunk to zero size and the branes are intersecting.

335: Since the bulk geometry is locally flat in the neighbourhood of any given point, an

336: approximation to the true bulk geometry about the collision point is provided by neglecting the

337: brane tensions and using the the Ricci-flat bulk ansatz (\ref{bulk_ansatz}) from the

338: previous section.  Substituting (\ref{BM_metric}) and

339: (\ref{BM_Phi_soln}) into (\ref{bulk_ansatz}), and expanding

340: about $y=1+\e/4M\sqrt{\beta}$, we obtain to leading order 

341: \[

342: \d s_5^2 = -\beta \d t^2 +  16M^2\beta \dw + \d \e^2 

343: +\frac{\e^2}{64M^2\beta^3}\d Y^2.

344: \]

345: The bulk geometry about the collision point is again conical (see Figure (\ref{dblcone})), 

346: but now the conical deficit angle depends on $\beta$ as well as $M$.  

347: Thus, given a particular asymptotic brane

348: separation at spatial infinity (fixing $\beta$), there is one specific value of the mass $M$

349: that removes the conical singularity in the bulk geometry.

350: \begin{figure}[t]

351: \begin{center}

352: \includegraphics[width=8cm, keepaspectratio=]{dblcone.eps}

353: \end{center}

354: \caption{The bulk geometry about the brane intersection.  The dashed

355:   line indicates that the fifth dimension is orbifolded on an

356:   $S^1/Z_2$.}

357: \label{dblcone}

358: \end{figure}

359: 

360: Since the brane geometry for the BBMB solution (\ref{BBMB}) is equivalent

361: to that of the extremal \RN solution, it is interesting to note that 

362: there have already been numerical attempts 

363: to find the bulk geometry corresponding to a \textit{fully-general}

364: \RN geometry on the brane \cite{Chamblin2}.

365: This work was motivated by the observation that the \RN geometry

366: solves the projected Einstein equations on the brane \cite{Dadhich},

367: equivalent to the Hamiltonian constraint equations of general

368: relativity, thus providing suitable initial data for evolution into the bulk.  

369: However, the four-dimensional effective theory approach considered here

370: is far more constraining than that of solving the projected Einstein 

371: equations.   We have found that in fact only the \textit{extremal} \RN 

372: geometry is feasible, and even then the bulk geometry will possess a conical 

373: singularity for general values of the mass parameter.

374: Our approach additionally allows the ``tidal charge'' parameter

375: of \cite{Dadhich} to be re-interpreted as the conformal scalar charge of

376: the BBMB black hole (which is equal to its mass \cite{Bekenstein1}).

377: 

378: Finally, we observe that it is possible to obtain black hole solutions for

379: which the brane separation at the event horizon is finite by 

380: detuning the brane tensions slightly from their Randall-Sundrum values.

381: This follows from recent work \cite{Martinez} in which a black

382: hole solution was found for conformal scalar gravity with a cosmological

383: constant and a quartic self-interaction term.  

384: From \cite{Us}, 

385: we see that this is simply the brane

386: frame effective theory for the case where both branes carry additional

387: cosmological constant terms.  For example, if we work in the plus

388: brane gauge, then the Lagrangian density for a cosmological term on the minus

389: brane produces a quartic self-interaction term for $\psi^-$  since

390: \[

391: \sqrt{-g^-}\Lambda^-=\sqrt{-g^+}(\psi^-)^4{\Lambda^-\over 36}. 

392: \]

393: If the cosmological constants satisfy $\Lambda^+=-\Lambda^->0$,  

394: the resulting field equations were shown to have a static solution

395: with extremal Reissner-Nordstr{\"o}m-de Sitter geometry and a non-trivial

396: conformal scalar field.  

397: In the case that cosmological constants vanish, this solution reduces

398: as expected to the  

399: BBMB solution (\ref{BBMB}).  However, for nonzero cosmological

400: constants, the solution possesses both an inner, an event, and a cosmological

401: horizon.  The brane separation at the event horizon is now finite,

402: as it is at the cosmological horizon.  The simple pole of the conformal scalar

403: field, corresponding to infinite brane separation, 

404: is then hidden from view between the inner horizon and the event horizon (which

405: coalesce in the BBMB limit of $\Lambda^\pm\tt0$).

406: 

407: 

408: 

409: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

410:  

411: \section{Stability}

412: 

413: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

414: 

415: \subsection{General case}

416: 

417: In this section we analyse the stability of the BBMB and traversable wormhole solutions

418: under linear monopole perturbations of the metric and the radion.  

419: Such perturbations are a distinctive feature

420: of configurations with a scalar field, and moreover, are the most

421: likely to display any instabilities that may be present.  This is

422: because the effective potentials for higher-order multipoles generally

423: contain centrifugal barrier terms.

424: 

425: Starting with the effective action in the form (\ref{M_EFT}), for which the equations

426: of motion are (\ref{M_eom}) and (\ref{Box_Phi}), we will take (\ref{BM_metric}) and 

427: (\ref{BM_Phi_soln}) as the background solution for $\g$ and $\Phi$ respectively.

428: The perturbed metric $\gh$ and scalar field $\Phih$ may then be written

429: \bea

430: \gh (y,t) &=& \g (y) + \delta \g (y,t) \\

431: \Phih (y,t) &=& \Phi (y) + \dPhi (y,t).

432: \eea

433: It is convenient to analyse the perturbations in a gauge in which the scalar

434: field perturbations decouple from those of the metric.  This gauge is

435: \bea

436: \delta (\sqrt{-g}g^{yy}) &=& 0 \\

437: \delta g_{yt} &=& 0,

438: \eea

439: where spherical symmetry additionally implies $\delta g_{y\theta} = \delta g_{y\phi}=0$.

440: A simple calculation suffices to show that this choice of gauge is always permitted.

441: Perturbing the scalar field equation (\ref{Box_Phi}), we find

442: \[

443: \label{pert_scalar_eqn}

444: \D_\mu (\delta (\sqrt{-g}g^{\mu y})\Phi'

445: +\sqrt{-g}\gi\D_\nu\dPhi)=0,

446: \]

447: where a prime denotes differentiation with respect to $y$.

448: Yet with our choice of gauge, 

449: \bea

450: \delta(\sqrt{-g}g^{\mu y})&=& \delta^\mu_y

451: g^{yy}\delta\sqrt{-g}-\sqrt{-g}g^{\mu\lambda}g^{yy}\delta g_{\lambda

452:   y} \nonumber \\

453: &=& \delta^\mu_y \delta(\sqrt{-g}g^{yy}) \nonumber \\

454: &=& 0

455: \eea

456: and so the scalar field perturbations do indeed decouple from those of the metric.  

457: In fact, the scalar field perturbations are the sole dynamical degrees of freedom:

458: once their behaviour has been determined, that of the metric perturbations follows 

459: automatically as we will see.

460: 

461: To put the perturbed scalar field equation into Schr{\"o}dinger form

462: we change variables to

463: \bea

464: \label{x_eqn}

465: x &=& M\left(\frac{1}{y}-y-2\ln{y}\right) \\

466: \dPhi &=& \frac{y}{M}(1+y)^{-2}\chi e^{i\omega t},

467: \eea

468: where $\chi$ is a function of $x$.  Then, 

469: \[

470: \D_x = - \frac{y^2}{M(1+y)^2}\D_y

471: \]

472: and (\ref{pert_scalar_eqn}) assumes the form

473: \[

474: (-\D_x^2+V)\chi = \omega^2 \chi ,

475: \]

476: where the potential 

477: \[

478: V = \frac{2y^3}{M^2 (1+y)^6}.

479: \]

480: Since this potential is positive definite over the entire region of interest $0<y<\inf$,

481: there are no bound states satisfying the boundary conditions 

482: $\chi\tt0$ as $y\tt 0$ and $y\tt\inf$.  Thus $\omega^2>0$ and there are no 

483: exponentially-growing unstable modes.   

484: Since the parameter $\beta$ has disappeared from the perturbed scalar field equation, 

485: this conclusion holds for both the wormhole and the BBMB solutions.

486: 

487: To analyse the corresponding metric perturbations, let

488: \bea

489: \delta g_{tt} &=& a(y) e^{i\omega t}g_{tt} \\

490: \delta g_{\theta\theta} &=& c(y)e^{i\omega t} g_{\theta\theta}. 

491: \eea

492: We then use the equations of motion in the form

493: \[

494: \Phi G_{\mu\nu} = (\nabla_\mu\nabla_\nu-\g\Box)\Phi

495: \]

496: (by taking the trace and using (\ref{Box_Phi}) 

497: one may check that this is equivalent to (\ref{M_eom})).

498: Perturbing the $yt$ equation to linear order, we find

499: \[

500: a(y)=\frac{ -(y^2+1)c+y(y^2-1)c'-\beta y(y+1)^2\dPhi'}{y^2-y+1}.

501: \]

502: We may then substitute this into the $yy$ equation.  After using

503: the perturbed scalar field equation, we obtain the following second order

504: O.D.E. for $c$, sourced by $\dPhi$ and $\dPhi'$:

505: \bea

506: 0 &=& y^4(y^2-1)(y^2-y+1)c''-2y^3(y^3-3y^2+1)c' \nonumber \\

507: &&  +(y^2-1)\left(2y^3+M^2\omega^2(y+1)^4(y^2-y+1)\right)c \nonumber

508: \\

509: && 4\beta y^3(y^2-y+1)\dPhi + 6\beta y^4(y^2-1)\dPhi'.

510: \eea

511: At $y=1$ the coefficient of $c''$ vanishes leaving us with the

512: boundary condition $c'=-2\beta\dPhi$ at this point.

513: Everywhere else, the equation is regular.  We conclude that the metric

514: perturbations are well-behaved, and that both the BBMB and wormhole solutions are

515: stable under monopole perturbations.

516: 

517: We note that this result contradicts an earlier result

518: in the literature \cite{Bronnikov_WH}.  There, the

519: stability of the wormhole solution is examined in the Einstein frame 

520: conformal gauge (see \cite{Us}), in which the action takes the form of gravity with a

521: minimally-coupled scalar field.  However, in this conformal gauge the scalar

522: field diverges at the brane intersection leading to a singular Schr{\"o}dinger potential

523: in the perturbed scalar field equation.  To deal with this singular potential

524: correctly, it is necessary to impose boundary conditions at the singularity (as

525: we will see in greater detail in the next section).  As this is not done, the analysis

526: of \cite{Bronnikov_WH} is incomplete.  

527: In contrast, the analysis presented

528: above utilises the $\Phi R$ conformal gauge in which the perturbed scalar field

529: equation is manifestly non-singular.

530:  

531:  

532: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

533: \subsection{BBMB black hole}

534: 

535: 

536: It has likewise been claimed in the literature on gravity with a conformally-coupled

537: scalar field that the BBMB solution is unstable to monopole perturbations

538: \cite{Bronnikov_BH}.

539: As this stands in contradiction to the results of the previous section, we will now

540: re-examine the stability of the BBMB black hole in the brane frame

541: conformal gauges (see Appendix A),

542: in which the action takes the form of gravity with a conformally-coupled scalar field.

543: 

544: The background metric $\g$ for the BBMB solution is given by (\ref{BBMB}) and the

545: corresponding conformal scalar field by (\ref{BBMB_psi}).  For $r>2M$, the scalar

546: field takes values in the range $0<\psi<\sqrt{6}$ and so we are in the plus brane

547: conformal gauge.  Then at $r=2M$, $\psi=\sqrt{6}$ and the branes intersect.  Nonetheless,

548: there is a smooth continuation to $r<2M$ for which $\psi$ takes values

549: greater than $\sqrt{6}$, 

550: indicating that we have matched onto an interior solution in the minus brane conformal gauge.

551: 

552: However, the action in the minus brane frame possesses ghosts, as

553: the relative sign between

554: the gravitational and kinetic terms in (\ref{S-}) is incorrect.  

555: These ghosts make the analysis of perturbations in the region interior

556: to the brane intersection much more subtle.

557: We will find that the gauge choice employed in \cite{Bronnikov_BH}

558: permits an unphysical influx of scalar charge across the

559: event horizon.

560: By contrast, if the perturbations are analysed using the decoupled gauge

561: introduced in the preceding section this problem is avoided.  In the decoupled

562: gauge, there is no influx of scalar charge across the event horizon,

563: and the BBMB black is found to be manifestly stable.

564: (Additionally, use of the decoupled gauge leads to a regular \S potential for the perturbed

565: scalar field equation, unlike the gauge choice used in

566: \cite{Bronnikov_BH}).

567: Such subtleties were not encountered in the previous section as the

568: $\Phi R$ conformal gauge employed there is automatically free from ghosts. 

569: 

570: Working in the brane frame gauges and dropping the plus or

571: minus labels, the vacuum equations of motion are given by

572: \bea

573: \label{full_eqn}

574: (\Box-\frac{1}{6}R) \psi &=& 0 \\

575: \label{conf_eom}

576: G_{\mu\nu}&=&T_{\mu\nu},

577: \eea

578: where the energy-momentum tensor for the conformal scalar field $\psi$

579: is 

580: \[

581: T_{\mu\nu} = \D_\mu \psi \D_\nu\psi - \frac{1}{2}\g

582: (\D\psi)^2  +\frac{1}{6}[\g\Box

583: -\nabla_\mu\nabla_\nu+G_{\mu\nu}]\psi^2 .

584: \]

585: Taking the trace of (\ref{conf_eom}) we find that $R=0$, 

586: allowing us to re-arrange (\ref{full_eqn}) and (\ref{conf_eom}) into the form

587: \bea

588: \label{boxpsi=0}

589: \Box \psi &=& 0 \\

590: \label{conf_E_eqns}

591: (1-\frac{1}{6}\psi^2)R_{\mu\nu}

592: &=&\frac{2}{3}\D_\mu\psi\D_\nu\psi-\frac{1}{6}\g (\D\psi)^2 

593:  -\frac{1}{3}\psi\nabla_\mu\nabla_\nu\psi .

594: \eea

595: The vanishing of the left-hand side of (\ref{conf_E_eqns}) when

596: $\psi=\sqrt{6}$ corresponds to the divergence of the effective

597: Newton's constant, $G_N\sim (1-\psi^2/6)^{-1}$.  

598: 	

599: The perturbed metric $\gh$ and scalar field $\psih$ may be written as

600: \bea

601: \label{pert_ansatz}

602: \gh (r,t) &=& \g(r)+\delta \g (r,t) \\ 

603: \psih (r,t) &=& \psi(r)+\dpsi(r,t).

604: \eea

605: We will first choose to describe the perturbations in the decoupled gauge:

606: \bea

607: \label{our_gauge}

608: \delta (\sqrt{-g}g^{rr}) &=& 0 \\

609: \delta g_{rt} &=& 0 .

610: \eea

611: Defining

612: \[

613: \label{x_def}

614: x = \frac{r(r-2M)}{r-M} + 2M\ln{\left(\frac{r}{M}-1\right)},

615: \]

616: so that

617: \[

618: \D_x = \left(1-\frac{M}{r}\right)^2\D_r

619: \] 

620: (recalling that $y=M/(r-M)$ we see that this $x$ is identical to (\ref{x_eqn})),

621: we find upon perturbing (\ref{boxpsi=0}) that 

622: \[

623: \delta\ddot{\psi} = \dpsi_{,xx}+2\frac{r_{,x}}{r}\dpsi_{,x},

624: \]

625: where the dots indicate differentiation with respect to $t$.

626: This may be put in \S form

627: by writing $\dpsi = (\chi/r)e^{i\omega t}$, where $\chi$ is a function

628: of $x$, giving

629: \[

630: \label{SE}

631: (-\D ^2 _x + V)\chi = \omega ^2 \chi

632: \]

633: where now

634: \[

635: \label{V}

636: V = \frac{2M}{r^3}(1-M/r)^3.

637: \]

638: This potential is positive definite over the region of interest

639: $M<r<\inf$.

640: Consequently there are no bound states satisfying the necessary

641: boundary conditions, $\chi\tt0$ for $x\tt\pm\inf$, and hence no

642: exponentially-growing instabilities. 

643: 

644: The behaviour of the corresponding metric perturbations may then be

645: determined using the perturbed Einstein equations,

646: as shown in Appendix B.  It may be checked that they are everywhere

647: well-behaved, thus completing the analysis.

648: 

649: We now turn to the earlier analysis of \cite{Bronnikov_BH}.  In this work, the

650: perturbations are studied in the gauge  

651: \[

652: \label{B_gauge}

653: \delta g_{\theta\theta}=0 

654: \]

655: instead of (\ref{our_gauge}).  

656: With this choice, the metric perturbations no longer

657: decouple from those of the scalar field in the perturbed scalar field

658: equation.  

659: Writing

660: \bea

661: \label{Bronnikov_a}

662: \delta g_{tt} &=& a(r) e^{i\omega t}g_{tt} \\

663: \label{Bronnikov_b}

664: \delta g_{rr} &=& b(r)e^{i\omega t} g_{rr}, 

665: \eea

666: the perturbed scalar field equation is now

667: \[

668: \label{scalar_eom_1}

669: 0=2r^4\delta\ddot{\psi}+(r-M)^2\Big(

670: \sqrt{6}M(a'-b')-2(r-M)(2\dpsi'+(r-M)\dpsi'')\Big).  

671: \]

672: The derivatives $a'$ and $b'$ may then be eliminated using the

673: perturbed Einstein equations, as shown in Appendix C. 

674: Following \cite{Bronnikov_BH}, we re-cast the result in \S form by setting

675: \[

676: \dpsi = e^{i\omega t}\chi\cdot\frac{(u^3+M^3)}{u^2r\sqrt{u^2-M^2}},

677: \]

678: where we have used the shorthand $u=r-M$, and

679: $\chi$ is a function of $x$ (defined in (\ref{x_def})). 

680: %(Note this transformation is singular at $r=2M$).

681: This yields a \S equation identical to (\ref{SE}), but with a 

682: different effective potential defined implicitly through \cite{error}

683: \[

684: \label{V_B}

685: V(x)=\frac{u^4}{r^4}\Big[\frac{2M}{r^2u}-\frac{M^2}{(u^2-M^2)^2}

686:  -\frac{6M^2r^2}{(u^3+M^3)^2} \Big].

687: \]

688: This potential decays as $V \sim 2m|x|^{-3}$ for $x\tt

689: +\inf$ (spatial infinity) and $x\tt-\inf$ (the event horizon).  

690: Significantly, the potential possesses a double negative pole located

691: at $r=2M$: in $x$-coordinates, $V \sim -1/4x^2$ close to $x=0$.

692: As we have seen, this singularity arose from back-substituting 

693: (\ref{conf_E_eqns}) to eliminate the metric perturbations from 

694: the perturbed scalar field equation.  (In the

695: decoupled gauge this step was not necessary and so there we obtained a regular \S problem).

696: 

697: It is claimed in \cite{Bronnikov_BH} that the

698: quantum-mechanical boundary-value problem corresponding to the potential (\ref{V_B})

699: (\ie $\chi\tt0$ as $x\tt \pm \inf$) has a spectrum of eigenvalues unbounded from below,

700: and hence there exists an infinite number of exponentially-growing instabilities.

701: However, this is no longer the case once appropriate boundary

702: conditions have been imposed at the singularity:

703: consider the general form of the solution about $x=0$, 

704: \[

705: \label{log_soln}

706: \chi (x) \sim A\sqrt{|x|}+B\sqrt{|x|}\ln{|x|},

707: \]

708: for arbitrary constants $A$ and $B$.

709: Since $\dpsi \sim \chi/\sqrt{|x|}$, we must impose the boundary condition $B=0$ at $x=0$,

710: discarding the logarithmic solution that would otherwise lead to the divergence of $\dpsi$.

711: We are then free to re-scale $\chi$ so as to set $A=1$ without loss of generality.

712: 

713: Using these boundary conditions, we may numerically

714: solve the perturbed scalar field equation by shooting towards the event horizon

715: and towards spatial infinity.  Performing this in the original $r$-coordinates

716: (equation (\ref{r_coord_version}) of Appendix C) for a range of

717: imaginary frequencies $\Omega=i\omega$ corresponding to an $e^{\Omega t}$ time dependence, 

718: we may scan the system for unstable modes.

719: Imposing the additional boundary condition that $\dpsi$ vanish at

720: spatial infinity we find that there is only 

721: \textit{one} unstable mode, with frequency $\Omega = 0.219$ %04.

722: (see Figure (\ref{mode_fig})). 

723: \begin{figure}[top]

724: \begin{center}

725: \includegraphics[width=8cm, keepaspectratio=]{psifig.eps}

726: \includegraphics[width=8cm, keepaspectratio=]{afig.eps}

727: \includegraphics[width=8cm, keepaspectratio=]{bfig.eps}

728: %\includegraphics[width=8cm, keepaspectratio=]{gttfig.eps}

729: %\includegraphics[width=8cm, keepaspectratio=]{grrfig.eps}

730: \end{center}

731: \caption{Radial profile of the unstable mode with $\Omega = 0.219$.}

732: \label{mode_fig}

733: \end{figure}

734: 

735: For values of $\Omega$ smaller than this the solution always diverges

736: to negative infinity at the event horizon and at spatial infinity, and for  

737: values of $\Omega$ greater than this, the solution always diverges to

738: positive infinity at both boundaries.  (Note the potential (\ref{V_B}) is

739: even under $x\tt -x$ and for bound states in one dimension there is no

740: degeneracy, so the eigenmodes must have a definite parity.  Hence a

741: divergence as $x\tt\inf$ implies a similar divergence for $x\tt-\inf$).

742: 

743: We are then faced with a paradox: in the decoupled gauge (\ref{our_gauge}) the BBMB

744: solution is manifestly stable, whereas in the alternative gauge

745: (\ref{B_gauge}) there appears to be one unstable mode satisfying all the

746: requisite boundary conditions.

747: 

748: To resolve the paradox, we re-interpret the scalar field equation

749: (\ref{boxpsi=0}) as a conservation law:

750: \[

751: \label{cons_law}

752: \D_\mu \j^\mu = 0,

753: \]

754: where the current density

755: \[

756: \j^\mu=\sqrt{-\hat{g}}\hat{g}^{\mu\nu}\D_\nu \psih .

757: \]

758: By Stokes's theorem, the only way in which the scalar

759: charge contained in a given spacetime volume can increase is if there is a

760: corresponding influx of current across the boundary.  Specifically,

761: consider the spacetime volume bounded by the event horizon and spatial

762: infinity, and the two arbitrary times $t_1$ and $t_2$.

763: Applying Stokes's theorem then gives

764: \[

765: 0=\left[ \int\d\theta\d\phi\d r \j^t\right]^{t_2}_{t_1}

766: +\left[\int\d\theta\d\phi\d t \j^r\right]^{r=\inf}_{r=M}

767: \]

768: as $\j^\theta$ and $\j^\phi$ vanish by spherical symmetry.

769: Evaluating $\j^t$ we find

770: \bea

771: \j^t &=& \sqrt{-g}g^{tt}\delta\dot{\psi}-\sqrt{-g}g^{tt}g^{rr}\delta

772:          g_{tr}\psi' \nonumber \\

773: &=& -r^2\sin{\theta}(1-\frac{M}{r})^{-2}\delta\dot{\psi}

774: \eea

775: as in both gauges $\delta g_{tr}=0$.  For an instability with 

776: exponential time dependence $\dpsi=\dpsi(r)e^{\Omega t}$, this becomes

777: \[

778: \label{charge}

779: \left[ \int\d\theta\d\phi\d r \j^t\right]^{t_2}_{t_1} =

780: -4\pi\Omega\left(e^{\Omega t_2}-e^{\Omega t_1}\right)\int_M^\inf \d r

781: (1-\frac{M}{r})^{-2}r^2\dpsi(r) . 

782: \]

783: Similarly,

784: \[

785: \j^r = \sqrt{-g}g^{rr}(\psi'+\dpsi')+\delta (\sqrt{-g}g^{rr})\psi'.

786: \]

787: Working first of all in the gauge where $\delta (\sqrt{-g}g^{rr})=0$,

788: we find that

789: \[

790: \left[\int\d\theta\d\phi\d t \j^r\right]^{r=\inf}_{r=M} = 

791: \frac{4\pi}{\Omega}\left(e^{\Omega t_2}-e^{\Omega

792:  t_1}\right)\left[(r-M)^2\dpsi'\right]^{r=\inf}_{r=M} ,

793: \]

794: which vanishes since $\dpsi'$ tends to zero exponentially rapidly at

795: both boundaries. 

796: (The solutions of the \S equation (\ref{SE}) vanish as

797: $e^{-|\Omega x|}$ as $x\tt\pm\inf$, and here $\dpsi\sim \chi/r$).

798: Thus, in the decoupled gauge, there is no influx of scalar current across the boundaries.

799: 

800: This is not the case for the gauge choice (\ref{B_gauge}).  

801: In this gauge, using (\ref{Bronnikov_a}) and

802: (\ref{Bronnikov_b}) we find 

803: \[

804: \left[\int\d\theta\d\phi\d t \j^r\right]^{r=\inf}_{r=M} = 

805: \frac{2\pi\sqrt{6}M}{\Omega}\left(e^{\Omega t_2}-e^{\Omega

806:  t_1}\right)\left[b-a\right]^{r=\inf}_{r=M} .

807: \]

808: The physical metric perturbations $\delta g_{tt}$ and $\delta g_{rr}$

809: must vanish at spatial infinity, hence by (\ref{Bronnikov_a}) and

810: (\ref{Bronnikov_b}) we see that $a$ and $b$ must likewise vanish.

811: Furthermore, $b$ must also vanish at the event horizon in order for

812: $\delta g_{rr}$ to remain finite there.  (This also follows from

813: (\ref{B_b_eqn}), as $\dpsi$ and $\dpsi'$ vanish exponentially at the horizon).

814: However, $a$ is \textit{not} required to vanish at the horizon: its value there

815: is determined solely by the radial integral of (\ref{B_a_eqn}).

816: For the unstable mode discussed above, $a$ is nonzero at the horizon

817: as may be seen from Figure (\ref{mode_fig}).

818: Thus, the exponential growth of this mode is supported by the influx

819: of scalar current across the event horizon.  Equating with

820: (\ref{charge}), we deduce

821: \[

822: a(M)=-\sqrt{\frac{2}{3}}\frac{\Omega^2}{M}\int_M^\inf \d r

823: (1-\frac{M}{r})^{-2}r^2\dpsi(r) ,

824: \] 

825: as is confirmed numerically.

826: 

827: We are now in a position to understand why the two perturbation

828: analyses performed in different gauges yield conflicting results.  One

829: gauge permits an influx of scalar current across the horizon; the

830: other forbids it. Exponentially-growing instabilities are only

831: possible in the former case.

832: However, since such fluxes are not physically realistic, we

833: must instead adopt the latter gauge and conclude that the BBMB black

834: hole is stable under monopole perturbations.  This conclusion is

835: supported by our calculations in the $\Phi R$ conformal gauge.

836: 

837: 

838: 

839: 

840: 

841: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

842: 

843: \section{Conclusions}

844: 

845: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

846: 

847: 

848: 

849: In this article we have studied a one-parameter family of solutions of

850: the brane world effective theory corresponding to traversable

851: wormholes and a black hole. 

852: While these solutions were already known in the context of gravity with a

853: conformally-coupled scalar field, our solution-generating method is

854: novel and our stability analysis corrects that of earlier work.

855: 

856: From a brane world perspective, the most pressing challenge ahead is

857: to find the exact bulk geometry corresponding to the effective theory

858: solutions we have found here.  

859: In addition to knowing the induced

860: metrics on the branes, a further clue is provided by the conical nature

861: of the bulk geometry about the brane intersection.

862: In fact, one possible strategy would be to take the initial data provided by

863: the brane metrics and solve for the bulk geometry order by order in a

864: power series expansion about the brane intersection 

865: (an analogous cosmological example of this is provided in \cite{Tolley}).

866: 

867: Ultimately, we may be led to question the fate of gravitational

868: collapse on the brane: since the central singularity of the BBMB black hole is

869: time-like, might it not be possible for matter to avoid the

870: singularity and instead pass through  

871: into a second causally-disconnected region of spacetime?

872: 

873: 

874: \textit{Acknowledgements:} We thank PPARC for support.

875: 

876: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

877: 

878: \appendix

879: 

880: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

881: \section{Four-dimensional Effective Theory}

882: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

883: 

884: 

885: In \cite{Us}, the form of the brane world low-energy effective theory was shown

886: to be highly constrained by symmetry considerations.  

887: The underlying five-dimensional theory, being pure gravity, is naturally invariant under

888: diffeomorphisms acting on the bulk.  However, a specific subset of

889: these bulk diffeomorphisms happen to have the effect of performing a conformal

890: transformation on the four-dimensional effective theory, whilst preserving the

891: induced metrics on the branes.

892: Consequently, the effective theory must itself be conformally invariant.  

893: This constraint, along with a

894: knowledge of the static solutions of the theory, is

895: sufficient to fully determine the form of the effective action up to

896: two derivatives.  

897: 

898: In the absence of matter, the effective action takes the form 

899: \bea

900: \label{EFT}

901: S &=& m_P^2\int \dx \sqrt{-g}\left(-\psi^+ \Delta \psi^+ + \psi^- \Delta

902: \psi^-\right) ,

903: \eea

904: where $\Delta \equiv \Box - R/6$ and $\psi^\pm(x)$ are two conformally-coupled 

905: scalar fields related to the positions of the branes in

906: the extra dimension, $Y^\pm(x)$, by $\psi^\pm=e^{Y^\pm/L}$. 

907: In this expression $L$ is the bulk anti-de Sitter length 

908: and $\psi^+>\psi^-$ as $Y^+>Y^-$.

909: The separation between the branes is then given by $d =Y^+-Y^-= L

910: \ln {(\psi^+/\psi^-)}$.

911: The induced metrics on the branes, $\gpm$, are

912: given in terms of the effective theory metric $\g$ by 

913: \[

914: \label{g_eqns}

915: \gpm = \frac{(\psi^\pm)^2}{6}\g .

916: \]

917: As required, the effective action is invariant under conformal

918: transformations

919: \[

920: \label{conf_transf}

921: \g \tt \Omega^2(x) \g, \ \ \ \ \ \ \psi^\pm \tt \Omega^{-1}(x)\psi^\pm.

922: \]

923: The brane metrics $\gpm$ remain invariant by (\ref{g_eqns}).

924: 

925: We will find it useful to fix the conformal gauge in two different ways:

926: the \textit{$\Phi R$} gauge and the brane frame gauges. 

927: The $\Phi R$ gauge is obtained by setting $\psi^++\psi^-=\sqrt{6}$ in (\ref{EFT}).  Writing

928: $\psi^+-\psi^-=\sqrt{6}\Phi$, the effective action in this gauge reads

929: \[

930: S = m_P^2\int\dx \sqrt{-g}\Phi R,

931: \]

932: and the induced brane metrics are

933: \[

934: \label{M_g_eqns}

935: \g^\pm = \frac{1}{4}(1\pm\Phi)^2\g.

936: \]

937: To allow for the possibility of brane collisions, we can generalise the formalism by re-labelling

938: the plus and minus branes as branes one and two, so that $\Phi=(\psi^{(1)}-\psi^{(2)})/\sqrt{6}$

939: now takes both positive and negative values.  The brane metrics are then given by (\ref{new_g_eqns}), 

940: with brane one being the plus brane for $\Phi>0$ and the minus brane for $\Phi<0$.

941: The interbrane separation is $d=2L\tanh^{-1}{|\Phi|}$.

942: 

943: The brane frame gauges are obtained by setting either of $\psi^+$ or

944: $\psi^-$ to $\sqrt{6}$, identifying the effective theory metric $\g$ appearing the

945: action with either the plus or minus brane metric, $\gp$ or $\gm$, respectively.  

946: In the plus brane gauge, $\psi^+=\sqrt{6}$, the action reads

947: \[

948: \label{S+}

949: S_+=m_P^2 \int\dx\sqrt{-g^+}\left( (1-\frac{1}{6}(\psi^-)^2)R -(\D\psi^-)^2\right),

950: \]

951: \ie gravity plus a conformally-coupled scalar field $\psi^-$ taking

952: values in the range $0\le \psi^-\le \sqrt{6}$. 

953: In the minus brane gauge, $\psi^-=\sqrt{6}$, the action is

954: \[

955: \label{S-}

956: S_-=-m_P^2\int\dx\sqrt{-g^-}\left( (1-\frac{1}{6}(\psi^+)^2)R -(\D\psi^+)^2\right),

957: \]

958: where $\sqrt{6}\le \psi^+<\inf$.  As (\ref{S+}) and (\ref{S-}) are equivalent up to a sign, 

959: the corresponding vacuum equations of motion are identical; the only difference is

960: the range of the conformal scalar field.  At a brane collision, where $\psi=\sqrt{6}$,

961: the solution then smoothly matches from the plus brane gauge to the minus brane gauge 

962: and vice versa.

963: 

964: 

965: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

966: 

967: \section{Metric Perturbations in Decoupled Gauge}

968: 

969: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

970: 

971: To analyse the behaviour of the metric perturbations in the decoupled gauge

972: (\ref{our_gauge})

973: we must evaluate the perturbed Einstein equations (\ref{conf_E_eqns}) to

974: linear order.  

975: This is easily accomplished with the help of a

976: standard computer algebra package.  Writing

977: \bea

978: \delta g_{tt} &=& a(r) e^{i\omega t}g_{tt} \\

979: \delta g_{\theta\theta} &=& c(r)e^{i\omega t} g_{\theta\theta}, 

980: \eea

981: it follows from

982: the gauge constraints and spherical symmetry that $\delta g_{rr} =

983: \left(a(r)+2c(r)\right)e^{i\omega t} g_{rr}$, $\delta g_{\phi\phi}

984: =c(r)e^{i\omega t} 

985: \delta g_{\phi\phi}$ and all the other

986: components of $\delta \g$ are zero.

987: The $tr$ constraint equation may then be solved for $a$:

988: \bea

989: \label{a_eqn}

990: a = -\frac{1}{3r(3M^2-3Mr+r^2)}&&\Big[3r(2M^2-2Mr+r^2)c

991:   +3r^2(M-r)(r-2M)c' \nonumber \\

992: && +\sqrt{6}M(M-r)(M-2r)\dpsi +\sqrt{6}rM(M-r)^2\dpsi'\Big] .

993: \eea

994: This equation is everywhere regular.  At the

995: event horizon, $r=M$, we find $a=c$ and as $r\tt\inf$ we find $a=-c+rc'$.  

996: Substituting (\ref{a_eqn}) into the $\theta\theta$ Einstein equation then

997: provides a second order O.D.E. for $c$, sourced by $\dpsi$ and $\dpsi'$:

998: \bea

999: \label{c_eqn}

1000: 0&=& 3r^3(r-M)^4(r-2M)(r^2-3Mr+3M^2)c'' \nonumber \\

1001: && -6Mr^2(r-M)^3(r^3-6Mr^2+9M^2r-3M^3)c' \nonumber \\

1002: && +3r^3(r-2M)[2M(r-M)^3 +\omega^2r^4(r^2-3Mr+3M^2)]c \nonumber \\

1003: && +6\sqrt{6}M(r-M)^4(r^3-2Mr^2+M^2r-M^3)\dpsi \nonumber \\

1004: &&+ 2\sqrt{6}Mr(r-M)^5(r^2-3M^2)\dpsi' .

1005: \eea

1006: At $r=2M$ the coefficient of $c''$ vanishes and we obtain 

1007: \[

1008: \label{bc1}

1009: 2\sqrt{6}Mc'+3\dpsi+2M\dpsi'=0,

1010: \]

1011: thus furnishing a boundary condition for $c'$ at this point.

1012: Elsewhere, the differential equation is regular:

1013: As $r\tt\inf$, (\ref{c_eqn}) reduces to $c''+\omega^2 c=0$ at leading order, 

1014: and at $r=M$ it becomes $c=0$.  

1015: Thus, once the perturbations of the scalar field have been determined,

1016: it is a simple matter to solve for the metric perturbations $a$ and

1017: $c$.

1018: 

1019: 

1020: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1021: 

1022: \section{Metric Perturbations in $\delta g_{\theta\theta}=0$ Gauge}

1023: 

1024: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1025: 

1026: Here we wish to evaluate the metric perturbations in the gauge

1027: (\ref{B_gauge}).  Substituting (\ref{Bronnikov_a}) and

1028: (\ref{Bronnikov_b}) into (\ref{conf_E_eqns}) and expanding to linear order,

1029: the $tr$ constraint equation yields

1030: \[

1031: \label{B_b_eqn}

1032: b=\frac{\sqrt{2}M(r-M)\left(

1033: (M-2r)\dpsi-r(r-M)\dpsi'\right)}{\sqrt{3}r(3M^2-3Mr+r^2)} .

1034: \]

1035: Substituting this into the $\theta\theta$ equation we find

1036: \bea

1037: \label{B_a_eqn}

1038: a' &=& \frac{\sqrt{2}M}{\sqrt{3}r^2(2M-r)(3M^2-3Mr+r^2)^2}\cdot

1039:  \nonumber \\

1040: && \Big[(-12M^5+21M^4r-42M^3r^2

1041:  +42M^2r^3-20Mr^4+4r^5)\dpsi \nonumber \\

1042: && -r(r-M)(12M^4+9M^3r -28M^2r^2+18Mr^3-4r^4)\dpsi' \nonumber \\

1043: && +r^2(r-M)^2(r-2M)(3M^2-3Mr+r^2)\dpsi''\Big] . 

1044: \eea

1045: (The perturbed Einstein equations in this gauge depend only on $a'$,

1046: and not $a$).

1047: Note that in order for $a'$ to be finite at $r=2M$ 

1048: it is necessary that $3\dpsi + 2\dpsi'=0$ at this point.  Eliminating 

1049: $a'$ and $b'$ from the perturbed scalar field equation (\ref{scalar_eom_1}), we find

1050: \[

1051: \label{r_coord_version}

1052: (\mathcal{A}(r)+\omega^2

1053: \mathcal{B}(r))\dpsi+\mathcal{C}(r)\dpsi'+\mathcal{D}(r)\dpsi''=0 ,

1054: \]

1055: where

1056: \bea

1057: \mathcal{A}(r) &=& -6 M^2 (r-M)^2 (-r^3+2Mr^2-M^2r+M^3)  \\

1058: \mathcal{B}(r) &=& r^6 (r-2M) (r^2-3Mr+3M^2) \\

1059: \mathcal{C}(r) &=& 2 r (r-M)^3 (r^4-5Mr^3 +10M^2r^2-6M^3r-3M^4)  \\

1060: \mathcal{D}(r) &=& r^2 (r-M)^4 (r-2M) (r^2-3Mr+3M^2) . 

1061: \eea

1062: This equation is singular when

1063: $\mathcal{D}(r)$, the coefficient of $\dpsi''$, vanishes at $r=2M$.

1064: The general solution about this point is given by the Taylor series 

1065: \[

1066: \label{Taylor_soln}

1067: \dpsi(2M+\e) = \mathcal{P}_1(\e)+j\ln{\e}\cdot\mathcal{P}_2(\e) ,

1068: \]

1069: where $r=2M+\e$ and the polynomials

1070: \bea

1071: \mathcal{P}_1(\e) &=&

1072: k-\left(\frac{j+3k}{2M}\right)\e

1073: +\left(\frac{19j+12k+32(k-j)M^2\omega^2}{8M^2}\right)\e^2 +O(\e^3)  \\

1074: \mathcal{P}_2(\e) &=&

1075: 1-\left(\frac{3}{2M}\right)\e+\left(\frac{3+8M^2\omega^2}{2M^2}\right)\e^2

1076: +O(\e^3), 

1077: \eea

1078: for arbitrary constants $j$ and $k$.  Since the single pure logarithmic term

1079: diverges as $\e\tt0$, we must set $j=0$ for $\dpsi$ to be finite. 

1080: (This is the equivalent condition in $r$-coordinates of setting $B=0$

1081: in (\ref{log_soln})). 

1082: The remaining solution branch then

1083: satisfies $3\dpsi+2\dpsi'=0$ at $r=2M$, ensuring that $a'$ is finite.

1084: By re-scaling $\dpsi$ we can without loss of generality set $k=1$,

1085: fixing the form of $\dpsi$ about $r=2M$.  With these boundary conditions

1086: we can then numerically evolve the solution towards infinity and towards

1087: the event horizon.

1088: 

1089: 

1090: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1091: 

1092: \begin{thebibliography}{x}

1093: 

1094: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

1095: 

1096: %\cite{Randall:1999vf}

1097: \bibitem{RS1}

1098: L.~Randall and R.~Sundrum,

1099: %``An alternative to compactification,''

1100: Phys.\ Rev.\ Lett.\  {\bf 83}, 4690 (1999);

1101: %[arXiv:hep-th/9906064];

1102: %%CITATION = HEP-TH 9906064;%%

1103: ibid. 3370. %[arXiv:hep-th/9905221]. 

1104: %%CITATION = HEP-TH 9905221;%%

1105: 

1106: %\cite{McFadden:2004se}

1107: \bibitem{Us}

1108: P.~L.~McFadden and N.~Turok,

1109: %``Conformal symmetry of brane world effective actions,''

1110: arXiv:hep-th/0409122.

1111: %%CITATION = HEP-TH 0409122;%%

1112: 

1113: %\cite{Kanno:2002ia}

1114: \bibitem{KS}

1115: S.~Kanno and J.~Soda,

1116: %``Radion and holographic brane gravity,''

1117: Phys.\ Rev.\ D {\bf 66}, 083506 (2002).

1118: %[arXiv:hep-th/0207029].

1119: %%CITATION = HEP-TH 0207029;%%

1120: %  ibid. 043526.

1121: %  % (hep-th/0205188).

1122: 

1123: %\cite{Khoury:2001wf}

1124: \bibitem{Ekpyrotic}

1125: J.~Khoury, B.~A.~Ovrut, P.~J.~Steinhardt and N.~Turok,

1126: %``The ekpyrotic universe: Colliding branes and the origin of the hot big

1127: %bang,''

1128: Phys.\ Rev.\ D {\bf 64}, 123522 (2001).

1129: %[arXiv:hep-th/0103239].

1130: %%CITATION = HEP-TH 0103239;%%

1131: 

1132: %\cite{Kanno:2003vf}

1133: \bibitem{KSnutshell}

1134: S.~Kanno and J.~Soda,

1135: %``Braneworld Kaluza-Klein corrections in a nutshell,''

1136: Phys.\ Lett.\ B {\bf 588}, 203 (2004).

1137: %[arXiv:hep-th/0312106].

1138: %%CITATION = HEP-TH 0312106;%%

1139: 

1140: %\cite{Chamblin:1999by}

1141: \bibitem{Chamblin}

1142: A.~Chamblin, S.~W.~Hawking and H.~S.~Reall,

1143: %``Brane-world black holes,''

1144: Phys.\ Rev.\ D {\bf 61}, 065007 (2000).

1145: %[arXiv:hep-th/9909205].

1146: %%CITATION = HEP-TH 9909205;%%

1147: 

1148: \bibitem{WH_orig} K. Bronnikov, Acta Phys. Polon. B \textbf{4}, 251 (1973).

1149: 

1150: %\cite{Barcelo:1999hq}

1151: \bibitem{Barcelo}

1152: C.~Barcelo and M.~Visser,

1153: %``Traversable wormholes from massless conformally coupled scalar fields,''

1154: Phys.\ Lett.\ B {\bf 466}, 127 (1999).

1155: %[arXiv:gr-qc/9908029].

1156: %%CITATION = GR-QC 9908029;%%

1157: 

1158: \bibitem{BBM} N. Bocharova, K. Bronnikov and V. Melnikov,

1159:   Vestn. Mosk. Univ. Fiz. Astron. \textbf{6}, 706 (1970).

1160: 

1161: %\cite{Bekenstein:1974sf}

1162: \bibitem{Bekenstein1}

1163: J.~D.~Bekenstein,

1164: %``Exact Solutions Of Einstein Conformal Scalar Equations,''

1165: Annals Phys.\  {\bf 82}, 535 (1974).

1166: %%CITATION = APNYA,82,535;%%

1167: 

1168: %\cite{Bekenstein:1975ts}

1169: \bibitem{Bekenstein2}

1170: J.~D.~Bekenstein,

1171: %``Black Holes With Scalar Charge,''

1172: Annals Phys.\  {\bf 91}, 75 (1975).

1173: %%CITATION = APNYA,91,75;%%

1174: 

1175:   

1176: %\cite{Chamblin:2000ra}

1177: \bibitem{Chamblin2}

1178: A.~Chamblin, H.~S.~Reall, H.~a.~Shinkai and T.~Shiromizu,

1179: %``Charged brane-world black holes,''

1180: Phys.\ Rev.\ D {\bf 63}, 064015 (2001).

1181: %[arXiv:hep-th/0008177].

1182: %%CITATION = HEP-TH 0008177;%%

1183: 

1184: 

1185: %\cite{Dadhich:2000am}

1186: \bibitem{Dadhich}

1187: N.~Dadhich, R.~Maartens, P.~Papadopoulos and V.~Rezania,

1188: %``Black holes on the brane,''

1189: Phys.\ Lett.\ B {\bf 487}, 1 (2000).

1190: %[arXiv:hep-th/0003061].

1191: %%CITATION = HEP-TH 0003061;%%

1192: 

1193: 

1194: %\cite{Martinez:2002ru}

1195: \bibitem{Martinez}

1196: C.~Martinez, R.~Troncoso and J.~Zanelli,

1197: %``De Sitter black hole with a conformally coupled scalar field in  four

1198: %dimensions,''

1199: Phys.\ Rev.\ D {\bf 67}, 024008 (2003).

1200: %[arXiv:hep-th/0205319].

1201: %%CITATION = HEP-TH 0205319;%%

1202: 

1203: %\cite{Bronnikov:2002sf}

1204: \bibitem{Bronnikov_WH}

1205: K.~A.~Bronnikov and S.~Grinyok,

1206: %``Instability of wormholes with a nonminimally coupled scalar field,''

1207: Grav.\ Cosmol.\  {\bf 7}, 297 (2001).

1208: %[arXiv:gr-qc/0201083].

1209: %%CITATION = GR-QC 0201083;%%

1210: 

1211: %\cite{Bronnikov:1978mx}

1212: \bibitem{Bronnikov_BH}

1213: K.~A.~Bronnikov and Y.~N.~Kireev,

1214: %``Instability Of Black Holes With Scalar Charge,''

1215: Phys.\ Lett.\ A {\bf 67}, 95 (1978).

1216: %%CITATION = PHLTA,A67,95;%%

1217: 

1218: 

1219: \bibitem{error} Note the erroneous sign appearing in equation

1220:   (4) of \cite{Bronnikov_BH}.

1221: 

1222: %\cite{Tolley:2003nx}

1223: \bibitem{Tolley}

1224: A.~J.~Tolley, N.~Turok and P.~J.~Steinhardt,

1225: %``Cosmological perturbations in a big crunch / big bang space-time,''

1226: Phys.\ Rev.\ D {\bf 69}, 106005 (2004).

1227: %[arXiv:hep-th/0306109].

1228: %%CITATION = HEP-TH 0306109;%%

1229: 

1230: 

1231: 

1232: 

1233: \end{thebibliography}

1234: 

1235: \end{document}

1236: