1: \documentclass[11pt]{article}

2: \def\baselinestretch{1.1}

3: \usepackage{amsfonts}

4: \usepackage{amsbsy} 

5: \usepackage{epsfig}

6: \usepackage{latexsym}

7: \input amssym.def

8: \input amssym.tex

9: 

10: %%%%%%%%%%%%%%%% Definitions %%%%%%%%%%%%%%%%%%%

11: \def\bequ{\begin{equation}}

12: \def\eequ{\end{equation}}

13: \def\barr{\begin{array}}

14: \def\earr{\end{array}}

15: \def\half{{1\over 2}}

16: \def\ben{\begin{equation}}

17: \def\een{\end{equation}}

18: \def\bena{\begin{eqnarray}}

19: \def\eena{\end{eqnarray}}

20: \def\mathgrave{\mathaccent"7012} 

21: \def\mathacute{\mathaccent"7013} 

22: \newcommand{\sect}[1]{\setcounter{equation}{0}\section{#1}}

23: \renewcommand{\theequation}{\arabic{section}.\arabic{equation}}

24: \topmargin 0pt

25: \advance \topmargin by -\headheight

26: \advance \topmargin by -\headsep

27: \textheight 8.9in

28: \oddsidemargin 0in

29: \evensidemargin \oddsidemargin

30: \marginparwidth 0.5in

31: \textwidth 6.5in

32: \advance\hoffset by -3mm  % A4 is narrower.

33: \advance\voffset by  8mm  % A4 is taller.

34: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

35: 

36: \def\vol{\rm Vol}

37: \def\const{\rm constant}

38: \def\herm{\rm Herm}

39: \def\bR{\Bbb R}

40: \def\bC{\Bbb C}

41: \def\bH{\Bbb H}

42: \def\bO{\Bbb O}

43: \def\bK{\Bbb K}

44: \def\bZ{\Bbb Z}

45: \def\bE{\Bbb E}

46: \def\bP{\Bbb P}

47: \def\bI{\Bbb I}

48: \def\bW{\Bbb W}

49: \def\b1{e^0}

50: \def\ir{{\rm irr}}

51: 

52: 

53: 

54: \begin{document}

55: \hfuzz=100pt

56: \title{{\Large \bf{Born-Infeld Theory and Stringy Causality}}} 

57: \author{G W Gibbons\footnote{Email address: gwg1@damtp.cam.ac.uk} \ \& C A R Herdeiro \footnote{Email address: car26@damtp.cam.ac.uk}

58: \\ (D.A.M.T.P.)

59: \\Department of Applied Mathematics and Theoretical Physics,

60: \\ Centre for Mathematical Sciences,

61: \\ University of Cambridge, 

62: \\ Wilberforce Road,

63: \\ Cambridge CB3 0WA,

64:  \\ U.K.}

65: 

66: \date{August 2000}

67: 

68: \maketitle

69: \centerline{DAMTP-2000-71}

70: 

71: \begin{abstract} 

72: Fluctuations around a non-trivial solution of Born-Infeld theory have a limiting speed given not by the Einstein metric but the Boillat metric. The Boillat metric is S-duality invariant and conformal to the open string metric. It also governs the propagation of scalars and spinors in Born-Infeld theory. We discuss the potential clash between causality determined by the closed string and open string light cones and find that the latter never lie outside the former. Both cones touch along the principal null directions of the background Born-Infeld field. We consider black hole solutions in situations in which the distinction between bulk and brane is not sharp such as space-filling branes and find that the location of the event horizon and the thermodynamic properties do not depend on whether one uses the closed or open string metric. Analogous statements hold in the more general context of non-linear electrodynamics or effective quantum-corrected metrics. We show how Born-Infeld action to second order might be obtained from higher-curvature gravity in Kaluza-Klein theory. Finally we point out some intriguing analogies with Einstein-Schr\"odinger theory.

73: \end{abstract}

74: 

75: 

76: 

77: 04.50.+h; 04.70.-s; 11.25.-w; 11.10.Lm

78: 

79: 

80: 

81: 

82: 

83: 

84: 

85: \sect{Introduction}

86: 

87: A striking feature of much recent work on open string

88: states in String/M-Theory is the considerable insights

89: afforded by the Born-Infeld \cite{BI} approximation.

90: Reciprocally, String/M-Theory has provided a rationale

91: for some of the hitherto mysterious and only partially understood

92: properties of this remarkable theory.

93: 

94: The existence of a limiting electric field strength, which 

95: was originally the {\it raison d'etre} of Born-Infeld

96: theory, now finds a dynamical justification

97: in the increasingly copious 

98: production of electrically  charged open string states

99: as one approaches the critical value \cite{porratti}.   

100: More subtly, the electric-magnetic duality symmetry

101: of Born-Infeld theory, a non-linear generalization of Hodge

102: duality first recognized by Schr\"odinger \cite{S1}, 

103: may be viewed as a special case of S-duality. 

104: In fact electric-magnetic duality

105: is a special case of Born-Reciprocity \cite{B}, a transformation 

106: which acts as a rotation in phase space

107: $

108: (p,q) \rightarrow 

109: (p \cos \theta + q \sin \theta, q \cos \theta- p\sin \theta)

110: $. In non-linear electrodynamics, the phase space variables might be considered to be 

111: $\bf B$ and ${\bf D}= \partial L / \partial {\bf E}$, which are canonically

112: conjugate variables in the sense of the Poisson bracket\footnote{We use the convention $\{x^i,p_j\}_{P.B.}=\delta^i_j$.}

113: \ben

114: \Bigl \{B_i({\bf x}), D_j({\bf y}) \Bigr \}_{P.B} =-\epsilon_{ijk}\partial_k \delta({\bf x}-{\bf y}).

115: \label{canbd}

116: \een

117: Born reciprocity applied to string theory

118: gives rise to T-duality \cite{V}. According to Hull and Townsend \cite{hull} 

119: T and S duality are included in the more general U-duality symmetry.

120: This leads naturally to the question of whether

121: Born-Infeld theory is the only

122: non-linear electrodynamic theory admitting electric-magnetic duality.

123: It is not \cite{GR}.

124: 

125: Another striking feature of Born-Infeld theory, 

126: is that it admits BIon solutions. These are exact solutions of the full

127: non-linear theory with distributional sources with finite total

128: energy. They can now be understood in terms of strings ending

129: on D-branes \cite{G, CM}. 

130: 

131: Schr\"odinger

132: recognized yet another remarkable property of Born-Infeld theory:

133:  viewed as a non-linear  optical theory,  

134: Born-Infeld theory exhibits uncommon properties with respect 

135: to the scattering of light by light. He constructed exact wave like

136: solutions of the full non-linear equations

137: representing light pulses with

138: solitonic properties. They  pass through one another without

139: scattering \cite{S2,S3}. This can be also understood from a string theory

140: point of view \cite{G}.

141: 

142: Some time after  Shr\"odinger's work

143: it was realized that the propagation

144: of Born-Infeld fluctuations around a background solution

145: has exceptional causal

146: properties \cite{B13,P} and that the theory 

147:  also admits exact solutions

148: exhibiting these exceptional  properties \cite{BB}. From the String Theory perspective, this relates to the recent interest in open string theory 

149: in a constant background Kalb-Ramond potential $B_{\mu \nu}$

150: and thus with gauge theory in a flat non-commutative

151: spacetime \cite{seiwit}. The  Kalb-Ramond potential

152: $B_{\mu \nu}$ appears in the Born-Infeld action

153:  in the combination $F_{\mu \nu} + B_{\mu \nu}$

154: and so from the Born-Infeld point of view, a constant $B_{\mu \nu}$ field

155: may be regarded as a background solution of Born-Infeld theory.

156: Some of the open string states propagating around the constant background

157: $B_{\mu \nu}$  field may be identified (at least in the abelian case)

158:  with fluctuations of the Born-Infeld theory.

159: Thus we need to understand the causal structure of their

160: propagation. It turns out that this is governed by a metric,

161:  $G_{\mu \nu}=g_{\mu \nu} -B_{\mu \alpha} g^{\alpha \beta } B_{\beta \nu}$\footnote{$B_{\mu \nu}$ may be taken to stand for the background field

162: or for the constant Kalb-Ramond 2-form. In string language we are using units

163: in which $2 \pi \alpha ^\prime=1$ }, which {\sl differs}

164:  from the usual spacetime metric

165:  $g_{\mu \nu}$ \cite{seiwit}. In fact we have two light-cones: the usual

166: light-cone given by $g_{\mu \nu}$ which  governs the

167: propagation of closed string states such as the graviton

168: and that given by 

169: $G_{\mu \nu}$ which governs the propagation of open string states such as the Born-Infeld photon. In general, the former lies outside the latter

170: except in two privileged directions corresponding

171: to the two principal null directions or eigenvectors

172: of the background two-form field. To put things provocatively,

173: gravitons almost always travel faster than light. 

174: 

175: One immediate

176: consequence is that if the closed string metric admits no closed

177: timelike curves then neither will the open string metric.

178: Another obvious consequence is that if the closed string metric contains

179: an event horizon   then the open string metric will

180: also contain an event horizon which lies inside or 

181: on the closed string event horizon. 

182: 

183: 

184: The comments in the last two paragraphs encode the global theme of this paper. Within many physical theories, two non-conformal metric structures arise, 

185: \bequ

186: g_{\mu \nu} \ \ \ \ \  and  \ \ \ \ \ g_{\mu \nu} + S_{\mu \nu},

187: \label{bimet}

188: \eequ

189: where $S_{\mu \nu}$ is some symmetric two-tensor. Geometrically they represent two sets of light-cones. Questions regarding causality or the existence of event horizons might then become particularly subtle. It is our purpose to discuss such issues in several contexts. For the aforementioned open versus closed string theory causal structures,  there is some advantage in  placing these

190: properties  in the general context of non-linear

191: electrodynamic theories, just as in the case of electric-magnetic

192: duality. Particularly so because

193: a quite separate strand of recent research

194: has been concerned with analogues of black holes, closed timelike

195: curves and circular null-geodesics in  non-linear electrodynamics \cite{novello}. There is also   considerable interest

196: in black holes in theories of non-linear electrodynamics coupled

197: to Einstein gravity (some recent references are \cite{beato}). 

198: 

199: 

200: 

201: 

202: 

203: In section 2 we look at general non-linear electrodynamics in a four-dimensional flat background. Hence $g_{\mu \nu}$ is the Minkowski metric. The study of the propagation of fluctuations of the electromagnetic field in a given electromagnetic background introduces naturally a second metric structure: the Boillat metric $A_{\mu \nu}^{Boillat}$. We emphasize the special properties of Born-Infeld theory in at least four senses: the existence of both electric-magnetic and Legendre duality and the absence of both bi-refringence and shocks.

204: 

205: 

206: In section 3, we specialize to the case when the electrodynamics theory is Born-Infeld. One can deal with a general curved background. So the natural background geometry is described by $g_{\mu \nu}$, the closed string metric. But open strings propagating in a non-trivial electromagnetic field or Kalb-Ramond potential see a different metric: the open string metric $G_{\mu \nu}$, conformal to $A_{\mu \nu}^{Boillat}$. We shall then see that if the closed string metric is static and the

207: Born-Infeld field is pure electric or pure magnetic

208: then the open string metric cannot have a non-singular

209: event horizon distinct from the one given by the closed string metric, because on it the electric field $\bf E$ must either equal its limiting value

210: or the magnetic field $\bf B$ must diverge.  Note that the metric $G_{\mu \nu}$ is not invariant, even up to a conformal

211: factor, under Hodge duality $\delta B_{\mu \nu}=\star B_{\mu \nu}$

212: but, as we shall see, it {\sl is}  invariant up to a conformal factor under electric-magnetic duality rotations.

213: 

214: In section 4 we show that scalar and spinor fluctuations around a Born-Infeld background are governed by the open string metric.

215: 

216: 

217: In section 5 the bi-metric theme takes another perspective. Recently \cite{her} examples have been given of how dimensional reduction can alter the causal structure of stringy black holes. Considering a trivial dilaton field, the relation between the lower dimensional metric $g_{\mu \nu}$ and the higher dimensional one $\hat{g}_{\mu \nu}$ is of the form (\ref{bimet}), with $\hat{g}_{\mu \nu}=g_{\mu \nu}+A_{\mu}A_{\nu}$. With this motivation, we look to higher-order Kaluza-Klein theory.  We notice that it is possible to obtain Born-Infeld theory to second order and still avoid ghosts, as long as the higher dimensional graviton is only excited along the compact dimensions. We similarly show that the effective theory for QED, the Euler-Heisenberg theory, may be obtained in this fashion, and discuss some properties of the theory obtained by starting with an Einstein-Hilbert plus Gauss-Bonnet action in higher dimensions \cite{lemos}. This gives an application of the general concepts discussed in section 2.

218: 

219: 

220: In section 6 we start by reviewing the results of section 2 considering the gravitational effects of the electromagnetic background field. One is then led to consider besides the usual Einstein metric $g_{\mu \nu}$ an effective co-metric of the form $g^{\mu \nu}+AR^{\mu \nu}$. Another possible origin for such effective metric is quantum renormalization of the propagator of test fields in a fixed background. We then discuss the universality of black holes event horizons and thermodynamic properties, by applying a result derived in the context of quantum renormalized metrics to the case of the Boillat metric for non-linear electrodynamics coupled to gravity.

221: 

222: 

223: 

224: In section 7, we review an old attempt of Einstein and Schr\"odinger to construct a unified theory of gravity and electromagnetism (see \cite{deser} for original references). One then introduces a metric which has an antisymmetric part. The symmetric part $g_{\mu \nu}$ and the inverse of the symmetric part of the inverse of the full metric, $A_{\mu \nu}^{\rm Eins-Schro}$ have remarkable similarities with the closed and open string metric, as first noticed by Boillat \cite{B11}. We discuss some exact solutions found by Papapetrou \cite{papape}.

225: 

226: We close with a discussion.

227: 

228: 

229: 

230: 

231: \sect{Causality in Non-Linear Electrodynamics}

232: 

233: \subsection{Characteristics and effective geometry}

234: 

235: 

236: We consider a general Lagrangian $L=L(x,y)$  

237: depending on the Lorentz invariants $x={ 1\over4}  F_{\mu \nu} F^{\mu \nu}$

238: and $y={ 1\over4}  F_{\mu \nu} \star F^{\mu \nu}$ .\footnote{We use a mainly minus metric signature and, contrary to Boillat and some other references

239: who use the opposite sign we choose $L$ to have the standard sign such that for Maxwell theory $L=-x$. Subscripts indicate partial differentiation.} These are the only independent Lorentz invariants in four spacetime dimensions. 

240: The energy momentum tensor is given by      

241: \ben

242: T_{\mu \nu}= -L_x T^{\rm Maxwell}_{\mu \nu}  + { 1\over 4} T g_{\mu \nu},

243: \een

244: where the trace and the Maxwell energy-momentum tensor are given by

245: \ben

246: T\equiv T^\mu_\mu =-4(L-xL_x-yL_y), \ \ \ \ \ \ \ \ \ \ \ \

247: T^{\rm Maxwell}_{\mu \nu}= -F_{\mu \alpha} F_{\nu \beta} g^{\alpha \beta}+x g_{\mu \nu}.

248: \een

249: 

250: Since both $ T^{\rm Maxwell}_{\mu \nu}$ and $g_{\mu \nu}$ (with mainly minus signature) 

251: satisfy the dominant energy condition, and the set   

252: of energy momentum tensors

253: satisfying the dominant energy condition is  a convex cone, a sufficient requirement 

254: for $T_{\mu\nu}$ to satisfy the dominant energy condition is that 

255: $L_x <0$ and  $T\ge0$.  An argument of Hawking and Ellis \cite{hawell} then shows that 

256: propagation in the full non-linear theory is causal in the sense that if at time zero all fields vanish

257: outside some compact set,

258: then they will vanish outside the future of that set.

259: In general one expects the fields to advance into empty space with no background field 

260:  at the speed of light and this expectation is supported by the observation

261: (originally due to Schr\"{o}dinger \cite{S2}) 

262: that any solution of Maxwell's linear electrodynamics

263:  with vanishing invariants,   

264: $x=y=0$,  will also be an exact solution of non-linear electrodynamics.

265: Among such so-called self-conjugate solutions are the usual

266: plane wave solutions which have unit speed.  

267: 

268: If a background field is present however these arguments require 

269: re-examination. One approach might be to look at the energy momentum tensor

270: of the fluctuations. We shall not do this here but begin by considering 

271: the {\it characteristics},

272:  which by definition are hypersurfaces

273: along  which {\sl weak} discontinuities

274: propagate. Assuming $F_{\mu \nu}$ to be discontinuous across the  $S(x^\mu)= {\rm constant }$ the characteristics are given by \cite{B6,B13,P,GDP}

275: \ben

276: (T_{\rm Maxwell}^ {\mu \nu}+\mu g^{\mu \nu})\partial _\mu S \partial _\nu S =0.

277: \label{charac}

278: \een

279: This has the form of a relativistic Hamilton-Jacobi equation for massless particles with effective co-metric $T_{\rm Maxwell}^ {\mu \nu}+\mu g^{\mu \nu}$, and where $S$ would be the action function. This effective metric also governs the propagation of 

280: weak, but not necessarily discontinuous fluctuations

281: around a background. Later we will turn

282: to the propagation of shocks and the behavior

283: of fully non-linear fluctuations. The function $\mu=\mu(x,y)$ satisfies

284: \ben

285: \varpi \mu^2 +\mu +\omega -\varpi (x^2+y^2)=0,

286: \label{quadra}

287: \een

288: where,

289: \ben

290: \varpi ={ L_{xx}L_{yy}-L_{xy}^2 \over L_{x}(L_{xx}+L_{yy})},

291: \een

292: and 

293: \ben

294: \omega={L_x+x(L_{xx}-L_{yy})+2yL_{xy} \over L_{xx}+L_{yy} } . 

295: \een

296: 

297: 

298: In the general case

299: the characteristics exhibits bi-refrigence: $\mu(x,y)$, which for convenience we parameterize as $\mu(x,y)= x + \zeta_\pm (x,y),

300: $ can take {\sl two values}, depending upon the polarization state and the background field. Thus there are, in general, two metrics.

301: The interpretation of the quantities $\zeta_\pm$ is that they

302: correspond to critical electric field strengths above which

303: the theory breaks down. For exceptional

304: theories the two values of $\zeta_\pm$ coincide and there

305: is a single light-cone  and no bi-refringence. Exceptional theories fall into two classes. The first has $\varpi=0$. This happens for instance

306: if the Lagrangian $L$ is independent of $y$, which includes Maxwell's theory as a special case. But not all theories with $\varpi=0$ are exceptional in this sense. In fact, although (\ref{quadra}) still encodes relevant information for this case, it does not contain \textit{all} the information anymore. One example will be given in section 5.

307: 

308: 

309: For $\varpi \neq 0$, the only exceptional theory is Born-Infeld \cite{B13}. The latter is also very special in that $\zeta_\pm$ is a constant

310: independent of $x$ and $y$. It is the only theory for which this is true.

311: We shall use  units in which this constant

312: is taken to be one. 

313: 

314: The condition that the theory admit electric-magnetic duality rotations

315: is rather weaker. It suffices that ${\bf B}\cdot {\bf E}={\bf D}\cdot {\bf H}$ \cite{GR}, which implies that the Lagrangian satisfy the  first order Hamilton-Jacobi type equation

316: \ben

317: y(L_x^2-L_y^2)-2xL_{x}L_{y}=y.

318: \een

319:  

320:  The characteristics or wave surfaces

321: may be thought of as null

322: hypersurfaces of a metric whose null geodesics 

323: correspond to  the {\it rays}.  Note that the characteristics and the rays

324: depend only a conformal equivalence class of metrics, defined by (\ref{charac}). A particular

325: choice of conformal representative used by Boillat, which

326: we shall refer to as the Boillat metric and co-metric,

327: is given by 

328: \ben

329: A^{\rm Boillat}_{\mu \nu}= { 1\over \sqrt{\mu^2-x^2-y^2}} 

330: \bigl (\mu g_{\mu \nu}

331: -T_{\mu \nu} ^ {\rm Maxwell} \bigr ) 

332: \label{boime}

333: \een

334: \ben

335: C_{\rm Boillat} ^{\mu \nu}= { 1\over \sqrt{\mu^2-x^2-y^2}}

336:  \bigl (

337: \mu g^{\mu \nu}+T^{\mu \nu} _{\rm Maxwell}  \bigr)  , 

338: \label{boicome}

339: \een

340: so that\footnote{Throughout this paper all indices will 

341: be raised or lowered using the usual Einstein

342: or closed string metric $g_{\mu \nu}$ {\sl with the exception}

343: of the open string metric $G_{\mu \nu}$ whose inverse 

344: is denoted by 

345: $G^{\mu \nu}$ in accordance with string theory conventions.}  

346: $A^{\rm Boillat} _{\mu \alpha} C_{\rm Boillat} ^{\alpha \nu}=\delta ^\nu _\mu$. As we shall see in detail

347: later, in the case of Born-Infeld theory, the open-string

348: metric $G_{\mu \nu}$ and the Boillat metric  $A^{\rm Boillat}_{\mu \nu}$

349: are conformal. 

350: 

351: Because

352: \ben

353: A^{\rm Boillat}_{\mu \nu}= { \mu -x \over \sqrt{\mu^2-x^2-y^2}} 

354: \bigl (g_{\mu \nu} -{ 1 \over \mu -x } F_{\mu \alpha} g^{\alpha \beta} F_{\beta \nu}

355: \bigr ), 

356: \een

357: the Boillat metric has a remarkable expression as a sort of

358: square root:

359: \ben

360: A ^{\rm Boillat}= { \mu-x \over \sqrt{\mu^2-x^2-y^2}} 

361: (g+ { 1\over \sqrt{\mu-x}} F) g^{-1} (g-  { 1\over \sqrt{\mu-x}}F).

362: \label{boisqu}

363: \een

364: It follows easily that

365: \ben

366: \sqrt {-\det A^{\rm Boillat} _{\mu \nu}} =\sqrt{-\det g_{\mu \nu}},

367: \label{elem}

368: \een

369: in other words, the Boillat metric and the spacetime metric induce the 

370: same volume element. The two principal null vectors common to both cones

371: are annihilated 

372: by  $g+ F/ \sqrt{\mu-x} $ or by  $ g- F / \sqrt{\mu-x}$ .           

373: 

374: 

375: We record for later use that if the background Einstein metric $g$ is flat, then

376: up to the conformal factor $ 1/ \sqrt{\mu^2-x^2-y^2} $

377: the Boillat metric is

378: \ben

379: (\mu -x) (dt^2-d{\bf x}^2 ) -{\bf E}^2 dt^2 + ({\bf E}\cdot d {\bf x} )^2

380: \pm 2 {\bf E} \times {\bf B} \cdot d {\bf x} dt -{\bf B}^2 d{\bf x}^2 +({\bf

381: B} \cdot d{\bf x} )^2.   \label{keymetric}

382: \een

383:   

384: 

385: 

386: In the generic case, one may diagonalise the Boillat

387: metric with respect

388: to the usual spacetime metric $g_{\mu \nu}$. 

389: This gives the speeds of propagation of the fluctuations in the associated inertial frame.

390: In this frame the Poynting vector ${\bf E}\times {\bf H}=

391: -L_x {\bf E}\times {\bf B}$ vanishes. The  velocities, i.e. the ratio of spacelike to timelike eigenvalues, turn out to be 

392: \ben

393: (1,{\mu -\sqrt{x^2+y^2} \over \mu+ \sqrt{x^2 +y^2}},

394: {\mu -\sqrt{x^2+y^2} \over \mu+ \sqrt{x^2 +y^2}}). 

395: \label{velo}

396: \een

397: Thus in general

398: there are two directions in which the Boillat-cone touches the usual

399: Einstein light-cone, corresponding to the first component of (\ref{velo}). These are the principal null directions of 

400: $F_{\mu \nu}$. Note that $F_{\mu \nu}$ and any duality rotation of it have the same principal null directions. In Figure \ref{cones1}, we represent the light cones for the effective plus Einstein geometry. The left cones illustrate the case with bi-refringence; we then have the Einstein plus two effective geometry cones. For the causal case, the Einstein cone will be $C1$. All the cones touch in two points, along the principal null directions of $F_{\mu \nu}$. The cones in the centre of Figure \ref{cones1} illustrate the exceptional case (like the Born-Infeld or open string theory case) where the effective geometry only possesses one light cone.

401: 

402: 

403: It may happen that the two principal null directions coincide. This occurs if and only if $x=0=y$. In this case the metric takes the form,

404: \bequ

405: A_{\mu \nu}^{Boillat}=g_{\mu \nu}-l_{\mu}l_{\nu}, 

406: \eequ

407: where $l_{\mu}$ is parallel to the principal null direction. The characteristic cone touches the Einstein cone along a single generator. This degenerate case is illustrated for exceptional theories in Figure \ref{cones1} right. For a generic electromagnetic field the principal null directions will coincide on a submanifold of dimension (and also co-dimension) two, $\mathcal{N}$. The complement $\mathcal{M} \setminus \mathcal{N}$ in the spacetime manifold $\mathcal{M}$, may not be simply connected. This gives rise to ambiguities in defining the `complexion' $\frac{1}{2}\arctan{y/x}$ of the electromagnetic field. In many ways, particularly if it is timelike, $\mathcal{N}$ behaves rather like a cosmic string \cite{penrin}.

408: 

409: 

410: 

411: 

412: 

413: 

414: 

415: 

416: 

417: 

418: In string theory, if a dilaton $\Phi$

419: is present,

420: one distinguishes between the Einstein metric $g_{\mu \nu}$

421: and the (closed)-string metric $e^{-2 \Phi} g_{\mu \nu}$. However both

422: have the same (Einstein) light-cone, i.e., they are conformal. This is because the dilaton

423: is a state of the closed string. It seems therefore, at least at the level of approximation we are considering, that there are just two causal structures

424: and two sets of cones: the 

425: open and the closed. Of course from the strict string theory point of view

426: one refers  brane and the other to bulk propagation but we have in mind

427: situations where the distinction is not sharp, such as for example

428: in the case of space-filling branes, or when considering

429: gravitons confined to, or at least moving parallel

430: to, the surface of a brane.  

431: 

432: 

433: \begin{figure}

434: \begin{picture}(0,0)(0,0)

435: \end{picture}   

436: \centering\epsfig{file=cones2.eps,width=16cm}   

437: \caption{Cones for the Einstein geometry and effective geometry describing the propagation of fluctuations in a non-trivial background $F_{\mu \nu}$ field. If condition (\ref{bradyon}) is obeyed, $C1$ is the Einstein cone, $C2$ (and $C3$ for the non-exceptional case) the effective geometry cones. The cone on the right represents the exceptional degenerate case.}

438: \label{cones1}

439: \end{figure} 

440: 

441: 

442: 

443: 

444: 

445: A sufficient condition that the Boillat-cone does not lie outside

446: the usual Einstein light-cone, i.e. that the speeds never exceed unity is that both $\mu$'s must satisfy 

447: \ben

448: \mu >\sqrt{x^2 +y^2}\equiv r.

449: \label{bradyon}

450: \een

451: In terms of the coefficients in (\ref{quadra}) this requirement reads $\omega<-r$, $-1/(2r)<\varpi<0$. Specialized to the Born-Infeld case (\ref{bradyon}) yields positive the quantity under the square root in the Born-Infeld action.

452:  

453: 

454: 

455: \subsection{Wave surfaces and Ether drift}

456: Boillat \cite{B13} has calculated the wave-front produced by waves moving

457: outwards from a point source with respect to an inertial frame

458:  in which the Poynting vector 

459: ${\bf E}\times {\bf H}=-L_x {\bf E}\times {\bf B}$

460:  does not vanish. If

461: \bequ 

462: w={ 1\over  2} ({\bf E}^2 + {\bf B}^2), \ \ \ a={ \sqrt{\mu+\sqrt{x^2 +y^2} \over \mu+w}}, \ \ \ 

463: b=\sqrt{ \mu-\sqrt{x^2 +y^2} \over \mu+w },

464: \eequ

465: and $c=ab$, so that

466: $1>a\ge b >c$, he finds that it is given by the family of  ellipsoids (in cartesian coordinates $(x,y,z)$)

467: \ben

468: {x^2 \over a^2} + {y^2 \over b^2} + {(z-tv_{\rm drift})^2 \over c^2}= t^2, 

469: \een

470: where the drift velocity is given by

471: \ben

472: {\bf v}_{\rm drift}= {{\bf E} \times {\bf B} \over \mu +w }.

473: \label{drift}

474: \een

475: Since $v^2_{\rm drift}=(1-a^2)(1-b^2)$, the drift velocity is always

476: less than one. Therefore, the presence of the background electromagnetic field causes the drift of the origin of disturbances and establishes preferred directions in spacetime; in a sense plays the role of `ether'.

477: 

478:     

479: 

480:   

481: 

482: 

483: 

484: \subsection{Convexity of the Hamiltonian Function}

485: 

486: The energy density or Hamiltonian density

487: $T_{00}$ should be considered as a function

488: $H({\bf D},{\bf B})$  of the canonically conjugate variables 

489: $({\bf D}, {\bf B})$ (in the sense of (\ref{canbd})). Their time evolution is obtained by taking the curl 

490: of $({\bf H},-{\bf E})$ where the constitutive relation

491:  $({\bf E}, {\bf H})=

492: (\partial H / \partial {\bf D} , \partial H / \partial

493: {\bf B})$ holds.

494: In other words $({\bf E}, {\bf H})$ and $({\bf D}, {\bf B})$ are related

495: by a Legendre transformation and in this sense one may regard the variables  $({\bf E}, {\bf H})$ as canonically conjugate to the variables $({\bf D}, {\bf B})$.

496: Of course this is a  different sense 

497: of canonically conjugate than that in which $\bf B$ and $\bf D$ are

498: canonically conjugate. It is a covariant sense in which one thinks

499: of the space of  Faraday tensors $F_{\mu \nu}$ (possibly subject to the

500: closure constraint $\partial_{[\mu } F_{\nu \tau]}=0$)

501: as the covariant

502: configuration space rather than the non-covariant configuration space

503: of magnetic induction fields

504:  $\bf B$ subject to the constraint ${\rm div}{\bf B}=0$.

505: 

506: 

507: The Legendre transformation

508: will be well defined and invertible if and only if the Hamiltonian

509: density $H({\bf B},{\bf D})$

510: is a convex function of it's arguments. In other words

511: the matrix of second derivatives or  Hessian is positive definite.

512: Note that in general $H$ may be defined

513: only in a portion of the six-dimensional space of possible ${\bf D}$ and

514:  $\bf B$ 's and the Legendre transform may only map into

515: part of  six-dimensional space of possible ${\bf E}$ and

516:  $\bf H$ 's. Thus for example, in the case of Born-Infeld theory

517: \ben

518: H=\sqrt{ (1 + {\bf B}^2)(1 + {\bf D} ^2) -({\bf B} \cdot {\bf D})^2 }-1,

519: \een

520: which is, in fact, defined for all $\bf B$ and $\bf D$.

521: However the inverse Legendre transformation is effected

522:  by means of the function

523: \ben

524: 1-\sqrt{ (1-{\bf H} ^2 ) ( 1-{\bf E}^2) - ({\bf E} \cdot {\bf H} )^2 }.

525: \een

526: which is defined only over the domain

527: of   $({\bf E}, {\bf H})$ given by

528: \ben

529: {\bf E}^2 + {\bf H}^2 < 1+ ({\bf E}\times {\bf H})^2. 

530: \een

531: Born-Infeld theory is one with the same constant upper

532: bound for both the electric (at zero magnetic field)

533: and magnetic field  strengths (at zero electric field). 

534: 

535: 

536: 

537: \begin{figure}

538: \begin{picture}(0,0)(0,0)

539: \end{picture}   

540: \centering\epsfig{file=convexity.eps,width=10cm}   

541: \caption{The Gibbs surface is the Lagrangian function. Inverting the constitutive relations, to find ${\bf E}={\bf E}({\bf B}, {\bf D})$, corresponds to finding the ${\bf E}$ coordinate of the contact point  of the Gibbs surface with a plane with slope ${\bf D}$ along a line of constant ${\bf B}$. Convexity is necessary for the inverse constitutive relations to be well defined.}

542: \label{gibbs}

543: \end{figure} 

544: 

545: 

546: 

547: 

548: Of course it should be borne in mind that

549: singling out a particular pair of variables 

550: is rather artificial. The underlying invariant

551: geometric structure is

552: the 12-dimensional  symplectic vector space 

553: with symplectic form 

554: $d{\bf B }\cdot \wedge d {\bf H} + d{\bf D} \cdot \wedge  d {\bf E}$

555: and a Lagrangian 

556: submanifold which defines the constitutive relations.

557: If one wishes one may pass to a 13 dimensional contact manifold

558: with contact form $dL-{\bf D}\cdot d {\bf E} +{\bf H} \cdot d{\bf E}$.

559: Then the constitutive relation provides a Legendre submanifold,

560: which of course on projection onto the $L$ coordinate

561: gives back  the Lagrangian submanifold.

562: One may instead  perform a projection onto any pair of the 12 vector

563: coordinates to obtain a ``Gibbs surface'' in a seven-dimensional space.

564: Picking for example the pair $({\bf E}, {\bf B})$

565: the Gibbs surface is given by

566: \ben

567: L=1-\sqrt{1-{\bf E}^2 + {\bf B} ^2 -({\bf B}\cdot {\bf E})^2 }.

568: \een

569: This is defined only in the domain 

570: $D\subset \bR^6$ connected to the origin

571:  for which $\det(g+F) <0$, that is

572:  ${\bf E}^2 -{\bf B} ^2 +({\bf B}\cdot {\bf E})^2 <1$. 

573: Geometrically for example, to find ${\bf E}$ as a function of ${\bf D}$ and ${\bf B}$  one brings up a 6-plane parallel to the ${\bf B}$ axis whose slope is given by ${\bf D}$ until it touches the Gibbs surface. The point of contact defines ${\bf E}$. If the Gibbs surface is convex there will be only one such contact point. This is illustrated in figure \ref{gibbs}.

574: 

575: 

576: Convexity will guarantee that all these projections are well defined

577: over the relevant domain and that, the surface has no folds for example as it would if the system exhibited

578: some sort of  hysteresis phenomenon.

579: For a general non-linear electrodynamic theory

580: the Hessian will only be positive definite over some 

581: domain in  $({\bf B}, {\bf D})$ space. Outside that domain

582: the constitutive relation is just that: a relation rather than

583: a function.  

584:   

585: 

586: The components of the Hessian are just the electric permitivities

587: and magnetic permeabilities. They  govern the behaviour 

588: of small disturbances around a background. Thus the background will be stable

589: as long as the Hessian is positive definite.

590: The equations for small fluctuations will also be hyperbolic

591: as long as the Hessian is positive definite \cite{B18}. 

592: 

593: 

594: 

595: 

596: 

597: 

598: 

599:     

600: \subsection{Shock Waves and Exceptionality}

601: 

602: In Maxwell theory, in flat space $\bE^{3,1}$, there exist traveling

603: wave solutions of the form

604: \ben

605: F_{\mu \nu}= f(S) F^0_{\mu \nu}, 

606: \een

607: where $f$ is an arbitrary function of it's argument, $S={\bf n}\cdot {\bf x}- v t$, and $\bf n$ is a constant unit 3-vector.

608: For fixed $\bf n$ these

609: represent a train of parallel waves moving with unit speed

610: in a fixed direction.  The arbitrary function $f$ allows us to pick

611: the profile of the wave train arbitrarily. One may even choose it to

612: be discontinuous. The amplitude

613: of the wave is constant on a family of wave surfaces $S={\rm constant}$

614: which correspond to a family of spacetime parallel null hyperplanes

615: whose intersection with any surface  of constant time 

616: gives a family of parallel 2-planes in $\bE^3$. Because they

617: move at the speed of light, wave trains cannot be brought to rest by means of a Lorentz

618: transformation.

619: 

620: In non-linear theories

621: in flat space one may, by analogy, adopt   the ansatz

622: \ben

623: F_{\mu \nu}= F^0_{\mu \nu}(f(S)), 

624: \een   where 

625: $F^0_{\mu \nu}$ will now in general depend on the arbitrary function $f$ 

626: and where

627: \ben

628: S={\bf n}\cdot {\bf x}-v({\bf n},S)t.

629: \een

630: Now  we get a family of hyperplanes  $S=\const$ in $\bE^{3,1}$

631: but they are no-longer parallel, although their intersections

632: with any surface of constant time still

633: gives a family of parallel 2-planes in $\bE^3$. The wave train

634: therefore moves with in a constant direction but not with constant

635: speed. They may slow down or speed up in the sense that

636: a hyperplane which passes a given point in space at a later times may

637: have a smaller or greater speed $v({\bf n},S)$. 

638: The hyperplanes will thus in general intersect (see figure \ref{shofig}). At these locations

639: the ansatz breaks down.

640: Neighbouring hyperplanes will envelope  a caustic hypersurface

641: obtained by eliminating $S$ from the equations

642: \ben

643: S={\bf n}.{\bf x}-v(S)t, \ \ \ \ \ \ \ \ \ 1=-v^\prime(S)t,

644: \een      

645: where $^\prime$ indicates differentiation with respect to $S$. Exceptional waves are those for which

646: \ben

647: v^\prime(S)=0.

648: \een

649: If all waves are exceptional, i.e. if $v^\prime=0$ $\forall S$,

650:  then parallel hyperplanes are possible. If $v\ne 1$ these can be

651:  brought to rest by means of

652: a Lorentz transformation. One then has stationary solutions \footnote{We say stationary

653: rather than static because the Poynting vector may not vanish} depending upon

654: two arbitrary functions $f_1(z)$ and $f_2(z)$

655: of a single spatial coordinate, $z$ say. We shall give concrete examples in the next section for the Born-Infeld case. 

656: 

657: 

658: \begin{figure}

659: \begin{picture}(0,0)(0,0)

660: \end{picture}   

661: \centering\epsfig{file=shocks.eps,width=14cm}   

662: \caption{Family of hyperplanes describing the propagation of wave fronts. If the hyperplanes intersect (right figure) the theory will be singular. Regularity (left figure) arises for exceptional theories only, like Born-Infeld, which have no shock formation.}

663: \label{shofig}

664: \end{figure} 

665: 

666: 

667: 

668: 

669: 

670: 

671: 

672: To understand the the physical significance of exceptionality, in the sense

673: of the absence of shock waves, one should consider

674: non-exceptional theories which do admit shock waves.

675: As theories they are essentially incomplete. One needs

676: extra physical assumptions to render the evolution beyond the shock. This typically may come from some underlying

677: more fundamental theory. Thus the predictions

678: of  classical theory 

679: admitting shocks, or indeed other singularities,

680: cannot be trusted in situations where they arise or are about to arise.

681: In this sense such theories ``predict their own demise", 

682: something that is often said of classical general relativity.

683: By contrast a classical theory, such as Yang-Mills theory,

684: for which the evolution of regular finite energy initial data remains

685: non-singular for all times \cite{moncrief} is certainly complete

686: as a theory, even though, because of quantum mechanics one

687: does not trust every classical prediction. To check the reliability

688: of a classical prediction we must check to see how 

689: it might be effected by quantum effects.

690: Generally speaking, we expect classical Yang-Mills theory 

691: to be useful in the weak coupling limit and when dealing

692: with very massive excitations such as magnetic monopoles. 

693: 

694: Classical general relativity is known to admit singularities

695: as a consequence of gravitational collapse. Only for weak

696: data do we expect non-singular evolution  for all time \cite{christ}.

697: There exist fully non-linear non-singular

698: solutions of general relativity depending upon

699: two arbitrary functions propagating at unit speed.

700: These are the pp-waves. They may be generalized to

701: propagate in an Anti-de-Sitter background \cite{GRu}. In some ways AdS is analogous to a background $B$ field. But pp-waves wave-fronts in AdS are null hypersurfaces of the AdS metric. This is in contrast with 

702: Born-Infeld theory and other non-linear electrodynamical theories, where there are 

703: plane wave solutions

704: traveling in some non-flat background spacetime

705: at a slower speed. Again, by contrast with

706: Born-Infeld theory, the collision of two pp-waves

707: gives rise to a spacetime singularity \cite{Griffiths}.

708: Such waves definitely cannot pass through one-another.

709: In this respect Born-Infeld theory resembles Yang-Mills theory

710: more than it does general relativity \cite{moncrief}.

711: One is tempted to speculate that it may

712: be a complete classical theory. Even if this is so,

713: any of its  classical predictions is subject

714: to quantum correction unless there is some

715: reason, such as supersymmetry, for believing that the quantum 

716: corrections vanish.

717:  

718: 

719: 

720: 

721: 

722:  

723: 

724: \subsection{Covariant Legendre Transformation}

725: 

726: We introduce here a dual notation via the fields $P_{\mu \nu}$ and $N_{\mu \nu}\equiv \star P_{\mu \nu}$. This notation has the advantage of making Legendre self-duality of Born-Infeld theory manifest.

727: 

728: In the dual notation, the field equations of \textit{any} non-linear theory of electrodynamics are

729: \ben

730: \nabla _\mu  P^{\mu \nu}=0,

731: \een

732: or, in form language, $d\star P=0$, where  the field $P^{\mu \nu}$ is defined by

733: \ben

734: dL=-\frac{1}{2}P^{\mu \nu}dF_{\mu \nu}.

735: \label{dualm}

736: \een

737: $P_{\mu \nu}$ coincides with $F_{\mu \nu}$ for Maxwell's theory. In general it reads

738: \ben

739: P_{\mu \nu}= - \bigl ( L_x F_{\mu \nu} + L_y \star F_{\mu \nu} \bigr ).

740: \een

741: The components of $P_{\mu\nu}$ are just $\bf D$ and $\bf H$. Using this two-form, the energy-momentum tensor can be cast in a form identical to $T_{\mu \nu}^{Maxwell}$

742: \ben

743: T_{\mu \nu}=-P_{\mu \alpha} g^{\alpha \beta}  F_{\nu \beta} -g_{\mu \nu} L.

744: \een

745: The formulation of the theory in terms of $P_{\mu \nu}$ is dual to the $F_{\mu \nu}$ formulation in the sense of a Legendre transformation. In fact if one takes the Legendre transform with respect to $\hat L$ by

746: \ben

747: {\hat L}=-{ 1\over 2} P^{\mu \nu}F_{\mu \nu}-L,

748: \label{lhat}

749: \een

750: one has

751: \ben

752: d{\hat L}=-{1 \over 2} F_{\mu \nu}d P^{\mu \nu},

753: \een 

754: in analogy to (\ref{dualm}). For the special case of a purely electric configuration in flat space, $\hat{L}$ is the ordinary Hamiltonian. Introducing the Hodge dual field $N_{\mu \nu}\equiv \star P_{\mu \nu}$, and defining $s\equiv { 1\over 4} N^{\mu \nu}N_{\mu \nu}= -{ 1\over 4} P^{\mu \nu}P_{\mu \nu}$ and $t \equiv { 1\over 4} \star N^{\mu \nu}N_{\mu \nu}=-{ 1\over 4}P^{\mu \nu} \star P_{\mu \nu}$ then the theory is specified by giving  $\hat L$ as a function of $s$ and $t$. Then we have

755: \bequ

756: F_{\mu \nu}=\hat{L}_s P_{\mu \nu} +\hat{L}_t\star P_{\mu \nu}.

757: \label{const2}

758: \eequ

759: The energy momentum in tensor in terms of the dual variables follows from (\ref{lhat}) and (\ref{const2}):

760: 

761: \ben

762: T_{\mu \nu}=\hat{L}_s T_{\mu \nu}^{Maxwell}[P]-g_{\mu \nu}(s\hat{L}_s+t\hat{L}_t-\hat{L}), \ \ \ \  T_{\mu \nu}^{Maxwell}[P]=  -P_{\mu \alpha}g^{\alpha \beta}P_{\nu \beta} -s g_{\mu \nu}.

763: \een

764: Sufficient conditions for the dominant energy condition to hold 

765: are

766: \ben

767: {\hat L}_s >0, \ \ \ \ \ \ \  s{\hat L}_s+ t{\hat L}_t-\hat{L}\ge 0.

768: \een

769: In the case of Born-Infeld theory one has $t=-y$ by electric-magnetic duality invariance and expressing also $x$ in terms of $(s,t)$ one gets

770: \ben

771: -{\hat L}=1-\sqrt{1+2s-t^2} \ \ \ \ \Leftrightarrow  \ \ \ \ L(F_{\mu \nu})=-\hat{L}(N_{\mu \nu}). 

772: \een

773: For Legendre self-dual theories like Born-Infeld, the equations describing propagation of perturbations (\ref{charac}), (\ref{quadra}), will have exactly the same form in terms of the variables $(x,y)$ as they do in terms of the variables  $(s,t)$. 

774: 

775: 

776: \sect{Born-Infeld/String theory}

777: 

778: 

779: 

780: 

781: \subsection{Open and Closed string metrics}

782: 

783: The  open string metric $G_{\mu \nu}$

784: is  usually obtained as follows \cite{seiwit}.\footnote{We will always use $F_{\mu \nu}$ for the gauge field, but it might represent the Kalb-Ramond potential $B_{\mu \nu}$.} 

785: One starts with the matrix $g+F$ whose components are $g_{\mu \nu} + F_{\mu \nu}$. Then one inverts to obtain a matrix with components

786: \ben

787: \Bigr ({ 1 \over g+F} \Bigl )^{\mu \nu}= G^{\mu \nu} + \theta ^{\mu \nu},

788: \een

789: where $G^{\mu \nu}$ is symmetric and $\theta^{\mu \nu}$ is antisymmetric.

790: Let $G_{\mu \nu}$ be the inverse of $G^{\mu \nu}$, i.e. $G_{\alpha \mu}G^{\mu \beta}=\delta_{\alpha}^{\beta}$. Calculation reveals that  

791: \bequ

792: G^{\mu \nu}= (G^{-1})^{\mu \nu}= \left((g-F)^{-1} g 

793: (g+F)^{-1}\right)^{\mu \nu}, 

794: \label{trouble}

795: \eequ

796: which is conformal to the inverse of (\ref{boisqu}) specialized to the Born-Infeld case. Then one checks that

797: \ben

798: G_{\mu \nu}= g_{\mu \nu}-F_{\mu \alpha}g^{\alpha \beta}F_{\beta \nu} 

799: \label{open}.

800: \een

801: 

802: A slightly more involved calculation shows that

803: 

804: 

805: 

806: \ben

807: \theta^{\mu \nu}=- { 1\over 1+2x-y^2} 

808: \bigl ( F^{\mu \nu}-y \star F^{\mu \nu} \bigr )

809: =- { 1\over \sqrt{1+2x-y^2} } P^{\mu \nu},

810: \label{noncom} 

811: \een

812: where $P^{\mu \nu}$ is the dual Maxwell field in the sense of (\ref{dualm}).

813: In verifying (\ref{noncom}) the following four dimensional matrix identities are useful (where {\bf 1} stands for the identity matrix):

814: \ben

815: g^{-1}F g^{-1} \star F=-y {\bf 1},

816: \een

817: \ben

818: g^{-1}F g^{-1}F -g^{-1}\star Fg^{-1} \star F=-2x{\bf 1},

819: \een

820: and hence

821: \ben

822: \bigl (g^{-1}F -yg^{-1} \star F\bigr )\bigl (g^{-1}F+ { 1\over y}g^{-1} \star F \bigr )=-(1+2x-y^2){\bf 1}. 

823: \een

824: 

825: Comparing (\ref{open}) with (\ref{boime}) one sees that the open string metric is equal, up to a conformal factor, to the Boillat metric governing the propagation of fluctuations around a Born-Infeld background:

826: \bequ

827: G_{\mu \nu}=\sqrt{1+2x-y^2} A_{\mu \nu}^{Boillat}.

828: \label{gboi}

829: \eequ

830: Relations (\ref{gboi}) an (\ref{noncom}) translate the stringy quantities $G_{\mu \nu}$ and $\theta_{\mu \nu}$ into pure non-linear electrodynamics language, i.e. the metric describing fluctuations around a fixed background and the dual Maxwell field.

831: 

832: An essential requirement on the causal structure defined by the open string metric is to be invariant under electric-magnetic duality rotations. To examine this we recall that the stress tensor of Born-Infeld theory, which is known to be invariant \cite{GR}, is given by

833: \ben

834: g^{\mu \nu} -T^{\mu \nu }_{\rm Born-Infeld} =

835: {2 \over \sqrt{-\det g}} {\partial \sqrt {-\det ( g+F)} \over \partial g_{\mu \nu}} 

836: \een

837: But 

838: \ben

839: \delta \sqrt {-\det ( g+F)}= 

840: {1 \over 2} \sqrt {-\det ( g+F)} \Bigl ({1 \over g+F} \Bigr ) ^{\mu \nu}

841:  \delta g_{\mu \nu}.

842: \een

843: Thus 

844: \ben

845: g^{\mu \nu}-T^{\mu \nu}_{\rm Born-Infeld}  ={\sqrt{-\det (g+F)} \over \sqrt{-\det g}} G^{\mu \nu} \label{tensor}.  

846: \een

847: Since the left hand side of (\ref{tensor}) is invariant so is the right hand side. Notice that the scalar $ \sqrt {-\det( g+F)} / \sqrt{-\det g} $ is not invariant but its change merely induces a conformal transformation in $G^{\mu \nu}$ and hence in $G_{\mu \nu}$, preserving the causal structure. It is worthwhile noticing that the right hand side of (\ref{tensor}) coincides with the Boillat co-metric 

848: \ben

849: g^{\mu \nu}-T^{\mu \nu}_{\rm Born-Infeld}  

850:  = C^{\mu \nu} _{\rm Boillat},

851: \label{Boillat}  

852: \een

853: which is therefore completely invariant under electric-magnetic duality rotations.

854: It is easily seen from (\ref{elem}) that the determinant of both sides of (\ref{Boillat}) equals $\det g^{\mu \nu}$.

855: Thus we get the remarkable result that

856: \ben

857: \det (\delta ^\mu_\nu - T_{\rm Born-Infeld} \thinspace ^{\mu} _{\nu} )=1.

858: \een

859: 

860: In the next subsections we illustrate the results above by analyzing the geometries seen by open strings in several special cases. We use the simplest exact solutions to Born-Infeld theory: plane wave solutions and spherically symmetric solutions.

861: 

862: 

863: 

864: 

865: 

866: \subsection{Exact plane wave solutions}

867: Boillat \cite{B14} found an exact stationary solution to Born-Infeld theory given in terms of two arbitrary functions $f_{1,2}$ of only one of the cartesian coordinates, say $z$, with an electric field and a magnetic induction given by 

868: \ben

869: \barr{c}

870: {\bf E}= \cosh \alpha {\bf i} + ( \cosh \alpha \sinh \beta f_1(z) -\sinh \alpha f_2(z)) {\bf k},

871: \\\\

872: {\bf B}= (\cosh \alpha f_2(z)-\sinh \alpha \sinh \beta f_1(z)) {\bf i}

873: -\cosh \beta f_1(z){\bf j}  + \sinh \alpha {\bf k},

874: \label{domwal}

875: \earr

876: \een

877: where $\alpha, \beta$ are arbitrary constants. The magnetic field and electric induction are easily obtained via the constitutive relations. The two Lorentz-invariants are 

878: \ben

879: -2x=1-f_1^2-f_2^2, \ \ \ \ \ \ y=f_2,

880: \een

881: so that the Born-Infeld lagrangian equals $1-|f_1|$. The Poynting

882: vector ${\bf P}={\bf E} \times {\bf H}$ is given by

883: \ben

884: \barr{c}

885: 2|f_1|{\bf P}=\left(f_1^2\cosh \alpha \sinh 2\beta -2f_1f_2 \sinh \alpha \cosh \beta\right) {\bf i} +

886: \\\\ +\left(2f_1f_2\sinh \beta \cosh 2\alpha  -\sinh 2\alpha (f_1^2\sinh^2\beta+f_2^2+1) \right) {\bf j} -2\cosh \alpha \cosh \beta f_1 {\bf k}.

887: \label{poynt}

888: \earr

889: \een

890: One might wonder if these stationary solutions may be interpreted as domain wall solutions. That is can one choose the asymptotic values of the

891: arbitrary functions $f_1$ and $f_2$ so as to interpolate between

892: two stable ``ground states"? One would then expect to have a static family of domain walls, that is, a non trivial solution for which the Poynting vector would be zero. However, this is not allowed by $P_z$ in (\ref{poynt}): $f_1$ would need to be zero for which case ${\bf D}, {\bf H}$ blow up. Therefore one finds no domain walls, just as in Maxwell's theory.

893: 

894: By performing a Lorentz transformation on (\ref{domwal}), one gets the general

895: fully non-linear sub-luminal plane wave solution. It presents no shocks, in accordance to section 2.4., since it propagates with constant speed. In general we do not expect superposition to hold in non-linear electrodynamics and therefore such plane waves propagating on top of some background solution should not solve the equations of motion anymore. However, the plane waves obtained by boosting (\ref{domwal}) may be superimposed to a background field and still yield a solution to Born-Infeld, as shown in \cite{BB}. Therein a background magnetic field along the x-axis and electric field along the y-axis are considered: ${\bf B}= B {\bf i}$ , ${\bf E}=E{\bf j}$ so that ${\bf E} \times {\bf B}= -EB {\bf k}$. If we set

896: \ben

897: v_\pm = {-EB \pm \sqrt{1-E^2+B^2} \over 1+B^2}= { 1-E^2 \over EB \pm\sqrt{1-E^2 +B^2}}, 

898: \een

899: one checks that plane waves traveling in the z-direction can be superimposed to the background field. The waves can do so in two polarization states, with the electric and magnetic field given by:

900: \begin{itemize}

901: \item  $({\bf e}, {\bf b} ) = ( v_\pm {\bf j}, -{\bf i} )  f_\parallel (z-v_\pm t)$ for the parallel polarization state, 

902: \item

903: $({\bf e}, {\bf b} ) = ( v_\pm {\bf i}, {\bf j} )

904:  f_\perp (z-v_\pm t) $ for the perpendicular polarization state,

905: 

906: \end{itemize}

907: where $f_\parallel$ and $f_\perp$ are arbitrary functions of their argument. Note that there is a net drift in the $z$ direction

908: \ben

909: v_{\rm drift}=  { 1\over 2} (v_+ + v_-) = -{ EB \over 1+B^2}, 

910: \een

911: in agreement with (\ref{drift}). This drift effect may be understood

912: as a consequence of Lorentz-invariance.

913: If $B^2 >E^2$ and one performs a Lorentz boost with velocity $u=E/B$ one may

914: pass to a frame in which the electric field vanishes

915: and   the magnetic field becomes

916: equal to $B_0=\sqrt{B^2-E^2}$.

917: Now the velocity $v_0$ in the this frame is symmetric with respect to 

918: reversing the z-direction and is given by

919: \ben

920: v_0={ 1\over \sqrt {1+B_0^2}}.

921: \een

922: One may check that $v_\pm$, $v_0$ and $u$ satisfy

923: the usual relative velocity addition formula

924: \ben

925: v_\pm= {u \pm v_0 \over 1 \pm  u v_0}.

926: \een

927: If $ E^2 > B^2$ one may reduce the magnetic field to zero.

928: The electric field in the de-boosted frame will be $E_0=\sqrt{E^2-B^2}$

929: and $v_0= \sqrt{1-E_0^2}$.

930: In these two case the open string metrics are (using (\ref{keymetric})),

931: \ben

932: ds^2_{\rm open}=  dt^2 -dx^2 - ( 1 + B_0 ^2) ( dy^2  +dz ^2 ) 

933: \label{bispeb}

934: \een 

935: and 

936: \ben

937: ds^2_{\rm open}= (1-E^2_0) (dt^2-dy^2) -dx^2 -dz^2 

938: \een 

939: In terms of the electric induction $D_0$ the latter is

940: \ben

941: ds^2_{\rm open}= { 1\over 1+D^2_0}(dt^2-dy^2) -dx^2  -dz^2.

942: \een  

943: which illustrates invariance of the open  string metric up 

944: to a conformal factor under  the discrete electric-magnetic duality transformation

945: $({\bf B} , {\bf D}) \rightarrow (-{\bf D},{\bf B})$. The general metric may be obtained using a Lorentz transformation.  

946: 

947: 

948: 

949: 

950: 

951: 

952: 

953: \subsection{BIons and other Static solutions}

954: 

955: Non-linear electrodynamic theories 

956: typically admit static finite energy solutions with distributional sources.

957: Because they have sources and also 

958: have (albeit mild) singularities, these solutions

959: are not solitons in the usual sense of the word. In \cite{G} they were called

960: BIons.

961: For the electrically charged BIon of Born-Infeld

962: theory we find the open string metric to be

963: \ben

964: ds^2_{\rm open}= { r^4 \over 1+ r^4} (dt^2-dr^2) -r^2(d\theta ^2 +\sin ^2 \theta d \phi^2 ). 

965: \label{BIele}

966: \een

967: The scattering of null 

968: geodesics is most conveniently represented as geodesics of the 

969: optical metric

970: \ben

971: ds_{\rm optical}^2= dr^2 + {1+ r^4 \over r^2} (d\theta ^2 +\sin ^2 \theta d \phi^2),

972: \een

973: which is easily seen to admit a 2-sphere of circular geodesics

974: at $r=1$ 

975: surrounding an infinite redshift infinite area naked singularity at 

976: finite proper distance situated at $r=0$. Such geodesics correspond to null geodesics of (\ref{BIele}).

977: 

978: The open string metric for the magnetically charged BIon is different:

979:  \ben

980: ds^2_{\rm open}= dt^2-dr^2 -{ 1+ r^4 \over r^2} (d\theta ^2 +\sin ^2 \theta d \phi^2 ). 

981: \een 

982: However the optical metrics are identical and in fact the two metrics

983: are conformally related.

984: 

985: This example may be generalized to any static configuration

986: in Minkowski spacetime. By static one means that the Poynting vector vanishes,

987: so that ${\bf E} \times {\bf B}=0$. The open string metric is

988: then also static and given by

989: \ben

990: ds^2_{\rm open}= (1-{\bf E}^2) dt^2 -d{\bf x}^2 +({\bf E}\cdot d{\bf x})^2

991: -{\bf B}^2 d{\bf x}^2 +({\bf B} \cdot d{\bf x} )^2.  

992: \een

993: Since $|{\bf E}|\le 1$ we have $G_{00}\ge 0$, moreover $G_{00}=0 $ implies

994: that $|{\bf E}|=1$. 

995: Thus any static event horizon of the Boillat metric which is not

996: an event horizon of the Einstein metric  must be singular, just as in the 

997: case of a single BIon solution.

998: 

999:  

1000: 

1001: Now consider what happens

1002: if the closed string metric $g$ 

1003: ceases to be flat but remains static .

1004: One has

1005: \ben

1006: G_{00}= g_{00}+F_{i0} F_{j0} g^{ij}=g_{00}(1+2x).

1007: \een

1008: Clearly, as long as $x >-1/2$ the sign of $G_{00}$ is

1009: determined entirely by the sign of $g_{00}$. Thus unless

1010: the electric field reaches the critical value, there

1011: can be no open-string static event horizon

1012: which is not also a closed string event horizon.

1013: 

1014: This result is really obvious from the viewpoint of

1015: electric-magnetic duality because we could instead have considered

1016: a purely magnetic field. In this case 

1017: \ben

1018: G_{00}=g_{00},

1019: \een

1020: and the magnetic field has no effect on that part

1021: of the metric which governs the location of event horizons. Actually, these results do not depend upon the detailed form of the

1022: open string metric obtained from Born-Infeld theory,

1023: nor upon electric-magnetic duality.

1024: They hold quite generally, as may be seen directly from

1025: the general expression (\ref{keymetric}) for the metric.

1026: At an event horizon we need

1027: \ben

1028: {\bf E}^2=\zeta_\pm,

1029: \een

1030: That is the electric field attains it limiting value

1031: at which point the theory breaks down. 

1032: 

1033: 

1034: 

1035: 

1036: 

1037: 

1038: 

1039: 

1040: 

1041: 

1042: 

1043: 

1044: 

1045: 

1046: 

1047: 

1048: 

1049: \subsection{The Boillat metric and Bionic scattering}

1050: 

1051: In this subsection we apply the fore-going theory to the 

1052: problem of scattering off the supersymmetric BIon or

1053: spike solution of the Dirac-Born-Infeld equations of motion \cite{G,CM}. This has been the subject of a number of detailed studies (see \cite{kastor, park} and references therein). Physically the solution represents a fundamental string attached to a D-brane.

1054: It is static and the transverse displacement of the brane is given by a scalar

1055: field $\phi({\bf x})$. The metric $g$ induced on the brane

1056: is thus 

1057: \bequ

1058: ds^2= dt^2 - (d{\bf x})^2 -(\nabla \phi \cdot d {\bf x} )^2. 

1059: \eequ

1060: Using the Bogomol'nyi conditions

1061: 

1062: \bequ

1063: {\bf E}= \pm \nabla \phi,

1064: \eequ

1065: where

1066: \bequ

1067: \nabla ^2\phi=0, 

1068: \eequ

1069: the induced   metric becomes

1070: 

1071: \bequ

1072: ds^2= dt^2 -(d{\bf x} )^2 -({\bf E} \cdot d{\bf x} )^2.

1073: \eequ

1074: The open string 

1075: metric then becomes

1076: \bequ

1077: ds_{open}^2 =\frac{dt^2}{ 1+ {\bf E}^2} -(d{\bf x})^2.

1078: \eequ

1079: This metric generates the same classical scattering as the Lagrangian (77) of \cite{park}.  In the case of a single susy BIonic spike we get

1080: \bequ

1081: ds_{open}^2= \left({r ^4 \over 1+r^4}\right)dt^2 - dr^2 -r^2 ( d \theta ^2 + \sin ^2  \theta

1082: d \phi ^2 ). 

1083: \eequ

1084: All the information about the classical scattering is now contained in $G_{\mu \nu}$.

1085: 

1086: 

1087: 

1088: 

1089: 

1090: 

1091: 

1092: \sect{Other spins}

1093: Born-Infeld actions maybe extended to include scalars and spinors. In this section we shall investigate the characteristics of these fields around a background constant $B$ field. 

1094: 

1095: \subsection{Scalars}

1096: 

1097: The Born-Infeld action with a single scalar field reads

1098: \bequ

1099: S=\int d^4x\left(\sqrt{-det(g_{\mu \nu})}-\sqrt{-det(g_{\mu \nu}+F_{\mu \nu}-\partial_{\mu}\phi\partial_{\nu}\phi)} \right).

1100: \eequ

1101: Expanding around a background $F$-field and retaining only quadratic terms in $\phi$ we get the S-duality invariant expression 

1102: \bequ

1103: S\simeq S_{BI}+\frac{1}{2}\int d^4x \sqrt{-g}C^{\mu \nu}_{Boillat}\partial_{\mu}\phi \partial_{\nu}\phi.

1104: \eequ

1105: $S_{BI}$ is the usual Born-Infeld action. Therefore, as expected,  scalar perturbations propagate according to the characteristics of the open string metric. 

1106: 

1107: \subsection{Spinors}

1108: 

1109: Consider a general Dirac action of the form

1110: \bequ

1111: S_D=\frac{i}{2}\int d^4x  \mu  \left(\bar{\Psi}\gamma_{\alpha}a^{\alpha \beta}\nabla_{\beta}\Psi +  \dots \right),

1112: \eequ

1113: where $\mu$ is a scalar density, $a^{\mu \nu}=(a)^{\mu \nu}$ are the components of a contravariant second rank tensor which need be neither symmetric or antisymmetric and the ellipsis denotes other possible terms in fermions but with no derivatives. 

1114: 

1115: The gamma matrices $\gamma$ generate the Clifford algebra associated with the closed string metric $g$ 

1116: \bequ

1117: \{\gamma_{\alpha},\gamma_{\beta}\}=2g_{\alpha \beta}.

1118: \eequ

1119: 

1120: The characteristics of this system are easily seen to be given by the co-metric\bequ

1121: a^{\alpha \mu}g_{\alpha \beta}a^{\beta \nu},

1122: \label{chara}

1123: \eequ

1124: which are the components of $a^t g a$. Note that we could rewrite the action as

1125: \bequ

1126: S_D=\frac{i}{2}\int d^4x  \mu  \left(\bar{\Psi}\Gamma^{\alpha}\nabla_{\alpha}\Psi + \dots\right) ,

1127: \eequ

1128: where

1129: \bequ

1130: \Gamma^{\alpha}=\gamma_{\beta}a^{\beta \alpha}.

1131: \label{newgam}

1132: \eequ

1133: In the case of Born-Infeld theory, it is natural to take $a=(g+B)^{-1}$, in which case use of (\ref{trouble}) shows that the characteristics as determined by (\ref{chara}) are given by the open string metric. Moreover, the gamma matrices introduced in (\ref{newgam}) generate the Clifford algebra associated with the metric $a^t g a$

1134: \bequ

1135: \{\Gamma^{\alpha},\Gamma^{\beta}\}=2a^{\alpha \mu}g_{\alpha \beta}a^{\beta \nu}.

1136: \eequ

1137: For the Born-Infeld case this is the open string Clifford algebra.

1138: 

1139: 

1140: Consider the Born-Infeld-Volkov-Akulov action which arises when one supersymmetrizes the Born-Infeld action. 

1141: \bequ

1142: S_{DBIVA}=\int d^4x\left(\sqrt{-det(g_{\mu \nu})}-\sqrt{-det(g_{\mu \nu}+F_{\mu \nu}+B_{\mu \nu}-\partial_{\mu}\phi\partial_{\nu}\phi+i\bar{\Psi}\gamma_{\mu}\nabla_{\nu}\Psi)} \right).

1143: \eequ

1144: In the absence of $B,\phi$ and $A$ fields this reduces to the Volkov-Akulov action \cite {VA}. Expanding to quadratic order in fermions gives a spinor action of the form $S_D$, with $\mu=\sqrt{-\det(g+B)}$ and $a=(g+B)^{-1}$. Thus, as one might have anticipated, the fermions characteristic cone is also given by the open string metric. 

1145: 

1146: 

1147: 

1148: 

1149: 

1150: 

1151: 

1152: 

1153: 

1154: 

1155: 

1156: 

1157: \subsection{Gravitons}

1158: It is not clear whether gravitons propagating in an external $B$ field would have their characteristics modified, since these are closed string modes and propagate on the bulk. However, in the light of the fact that we now believe that gravity can be localized on the brane \cite{randsun}, one might be tempted to speculate that under some circumstances that should happen. If that is the case the obvious guess for the characteristics would be the open string metric. Indeed precisely this happens in Einstein-Schr\"odinger theory. This is a unified theory in which the usual symmetric

1159: Einstein metric is replaced by an arbitrary $4\times4$ 

1160: (or more generally in $n$ spacetime dimensions an $n\times n$) tensor field $a$

1161: which we write  suggestively as:

1162: \ben

1163: (a^{-1}) _{\alpha \beta}= a_{\alpha \beta} =g_{\alpha \beta}+B_{\alpha \beta}.

1164: \een

1165: Lichnerowicz and Maurer-Tison \cite{Licher, maurer} showed that some of the small fluctuations have characteristics given 

1166: by the symmetric part of the co-metric, i.e. ${ a } ^{(\mu \nu)}\equiv G^{\mu \nu}_{Eins-Schro}$,  in striking analogy to the open string/Born-Infeld case. Therefore the properties of these characteristics are the same as the ones presented in section 2 and 3. However, the theory exhibits a kind of bi-refringence, due to the existence of a second set of (co-)cones for small fluctuations, given by \cite{maurer}\footnote{We would like to thank M.Clayton for poining out to us the existence of this second light cone.}

1167: \bequ

1168: 2\frac{\det{g_{\alpha \beta}}}{\det{(g_{\alpha \beta}+B_{\alpha \beta})}} g^{\mu \nu}-G^{\mu \nu}_{Eins-Schro}.

1169: \eequ

1170: If we define $T^{\mu  \nu}_{Eins-Schro}$ by an expression similar to (\ref{tensor}) replacing $G^{\mu \nu}$ by $G^{\mu \nu}_{Eins-Schro}$ and $F$ by $B$, our two co-metrics are conformal to, respectively:

1171: \bequ

1172: g^{\mu \nu}-T^{\mu  \nu}_{Eins-Schro}, \ \ \ \ \  \left(2\sqrt{\frac{\det{g_{\alpha \beta}}}{\det{(g_{\alpha \beta}+B_{\alpha \beta})}}}-1\right)g^{\mu \nu}+T^{\mu  \nu}_{Eins-Schro}.

1173: \eequ

1174: Just as in the BI case the first set of light cones will lie inside the Einstein cones. But because of the opposite sign in $T^{\mu  \nu}_{Eins-Schro}$, the second set of light cones will be outside both the first set of cones and the Einstein light cones. The former result was pointed out in \cite{maurer} while the latter appears to confirm the pathological properties of this theory, in that some fluctuations are tachyonic with respect to the Einstein metric. Presumably these fluctuations can carry negative energy  

1175: 

1176: 

1177:  We shall return to this theory in section 7. Before doing so we should recall that Einstein-Schr\"odinger theory appears to break invariance under the gauge transformation $B\rightarrow B+dA$ and for this reason it has been claimed to admit negative energy states \cite{deser}.

1178: 

1179: 

1180: \subsection{Gravitinos?}

1181: This is a short subsection because as yet we have no consistent supergravity brane solution and thus as far as we know no consistent theory of a gravitino propagating on a brane. However if such a theory exists and the gravitino propagation is affected by a background $B$ field then there is an obvious suggestion for the characteristics. 

1182: 

1183: 

1184: 

1185: 

1186: 

1187: 

1188: 

1189: \sect{Non-Linear Electrodynamics from $U(1)$ Kaluza-Klein Theory}

1190: Kaluza-Klein theory stems from the fact that the Ricci scalar for the $(D+1)$ dimensional ansatz 

1191: \begin{equation}

1192: d\hat{s}^2=ds^2+(dy+A_{\mu}dx^{\mu})^2,

1193: \label{kklein}

1194: \end{equation}

1195: is $\hat{R}=-x$, i.e., the Maxwell Lagrangian. Here $ds^2$ is the D-dimensional Minkowski metric and $y$ the coordinate along the extra dimension. It is known, however, that the truncation of Kaluza-Klein theory to pure electromagnetism is not consistent. In fact, considering a trivial scalar field implies $x=constant$ via the scalar equation of motion. We will not be concerned about this point in what follows, but rather study some properties of the electrodynamical theory that arises from considering the lowest order in $\alpha'$ tree level string theory corrections to the Einstein-Hilbert action in dimensions higher than four. Full study of such Kaluza-Klein theory must be performed by considering also gravitational and scalar excitations.

1196: 

1197: With the ansatz (\ref{kklein}) the curvature invariants of second order in $D+1$ dimensions are (excluding a possible Chern-Simons term):

1198: \begin{equation}

1199: \begin{array}{c}

1200: \hat{R}^2=x^2, \ \ \ \ \ \ \hat{R}_{MN}\hat{R}^{MN}=x^2+\frac{1}{2}\partial_{\beta}F_{\alpha}^{ \ \ \beta}\partial_{\mu}F^{\alpha \mu} +\frac{1}{4}F_{\mu \nu}F^{\nu \alpha}F_{\alpha \beta}F^{\beta \mu}, 

1201: \\\\ \hat{R}_{MNPQ}\hat{R}^{MNPQ}=6x^2+\frac{5}{8}F_{\mu \nu}F^{\nu \alpha}F_{\alpha \beta}F^{\beta \mu}+\partial_{\alpha}F_{\mu \sigma}\partial^{\alpha}F^{\mu \sigma}.

1202: \label{rsq}

1203: \end{array}

1204: \end {equation}

1205: The most general parity conserving term quadratic in the curvature is then

1206: \begin{equation}

1207: \begin{array}{c}

1208: \hat{R}_{MNPQ}\hat{R}^{MNPQ}+a\hat{R}_{MN}\hat{R}^{MN}+b\hat{R}^2=

1209: \\\\ =(6+a+b)x^2+\frac{5+2a}{8}F_{\mu \nu}F^{\nu \alpha}F_{\alpha \beta}F^{\beta \mu}+\frac{4+a}{4}\partial_{\alpha}F_{\mu \nu}\partial^{\alpha}F^{\mu \nu}+...,

1210: \label{genqua}

1211: \end{array}

1212: \end{equation}

1213: where the dots stand for total derivatives. Hence, the terms with derivatives of the field strength in (\ref{rsq}) cancel (up to total derivatives) in the combination $\hat{R}_{MNPQ}\hat{R}^{MNPQ}-4\hat{R}_{MN}\hat{R}^{MN}$, thus avoiding ghosts in the propagation of the electromagnetic field. Actions with such derivative terms have nevertheless been considered in the past, as in the Bopp-Podolsky action \cite{bopp}. We will require the cancellation of ghosts and therefore consider the dimensional reduction of an action of the type

1214: \begin{equation}

1215: S=\frac{1}{16\pi G_{D+1}}\int d^{D+1}\hat{x} \sqrt{\hat{g}} \left(\hat{R}+\Upsilon (\hat{R}_{MNPQ}\hat{R}^{MNPQ}-4\hat{R}_{MN}\hat{R}^{MN}+b\hat{R}^2)\right).

1216: \label{actiondp1}

1217: \end{equation}

1218: Specializing to $D=4$, where one can use the identity $F_{\mu \nu}F^{\nu \alpha}F_{\alpha \beta}F^{\beta \mu}=8x^2+4y^2$, we get the lagrangian

1219: \begin{equation}

1220: {\mathcal{L}_{KK}}= -x + \Upsilon \left((b-1)x^2-\frac{3}{2}y^2\right).

1221: \label{lagkk}

1222: \end{equation}

1223: One notices the absence of an $xy$ parity breaking term to this order. In principle one could bring such term into the theory by including a Chern-Simons term. In $D+1=5$, two such possible terms are

1224: \ben

1225: S_{CS}=\int tr(\hat{R}_{AB}\wedge \hat{R}^{B}_{\ \ C}\wedge X), \  \ or \ \ S_{CS}=\int tr(\hat{R}_{AB}\wedge \hat{R}^{B}_{\ \ C}\wedge \hat{w}^{C}_{\ \ D}),

1226: \een 

1227: for some one form field $X$, or using the one form connection $\hat{w}$. The second and most natural possibility gives, however, terms of order higher than the ones considered in ${\mathcal{L}_{KK}}$. For the first possibility, the most natural choice of $X$ is as being dual to the fiber direction $\partial / \partial y$; then the first possibility contributes only to the ghosts. Hence we will not consider them anymore.

1228: 

1229: By arranging the constants $b$ and $\Upsilon$ in (\ref{lagkk}), one can recover several interesting cases which analyse in the following subsections. 

1230: 

1231: \subsection{Gauss-Bonnet electromagnetism}

1232: This theory is obtained for $b=1$ \cite{lemos} (see also earlier work in \cite{kerner}). As pointed out in \cite{zwiebach}, an analysis for linear perturbations of the gravitational field shows that ghost cancellation requires $\hat{R}^2$ to enter the combination (\ref{genqua}) as

1233: 

1234: \begin{equation}

1235: \Omega_{2}=\hat{R}_{MNPQ}\hat{R}^{MNPQ}-4\hat{R}_{MN}\hat{R}^{MN}+\hat{R}^2 \stackrel{\mathrm{4D}}{=}\frac{1}{4}\hat{R}_{MN}^{ \ \ \ \ AB}\hat{R}_{PQ}^{ \ \ \ \ CD}\eta^{MNPQ}\eta_{ABCD}.

1236: \end{equation}

1237: $\eta$ is the Levi-Civita tensor, not density. This is the Gauss-Bonnet combination. The last equality, which holds in four dimensions where the four-form $\eta$ is the volume form, shows it is the second Euler density, a topological term in four dimensions but dynamical in higher dimensions. We recall that the first Euler density, topological in two dimensions but dynamical in higher is just the Ricci scalar:

1238: \begin{equation}

1239: \Omega_1=\hat{R}\stackrel{\mathrm{2D}}{=}\frac{1}{2}\hat{R}_{MN}^{\ \ \ \ AB}\eta^{MN}\eta_{AB}.

1240: \end{equation}

1241: The last equality holds in two dimensions, where the two-form $\eta$ is the volume form. The Gauss-Bonnet combination is usually referred to as describing the first order string theory corrections to general relativity \cite{zwiebach}. The first order (in $\alpha'$) stringy gravitational action can then be written exclusively in terms of Euler densities (that does not seem to be the case already at third order):

1242: \begin{equation}

1243: S^{(1)}=\frac{1}{16\pi G_{D+1}}\int d^{D+1}\hat{x} \sqrt{-\hat{g}}(\Omega_1 +\Upsilon \Omega_2),

1244: \end{equation}

1245: with $\Upsilon \propto \alpha'$. That $S^{(1)}$ is the correct effective action relies on two arguments.  Matching the amplitude for the scattering of three on-shell gravitons in bosonic closed string theory only fixes the $(\hat{R}_{MNPQ})^2$ term; the $(\hat{R}_{MN})^2$ and $\hat{R}^2$ do not contribute to the on-shell amplitude. These are fixed by the no-ghost requirement, since one does not see any ghosts in the string spectrum. But for purely electromagnetic excitations within a Kaluza-Klein context, the no-ghost requirement is more relaxed and makes sense to consider an arbitrary $\hat{R}^2$ coefficient.

1246:  

1247: In non-covariant language, the Gauss-Bonnet lagrangian is described by

1248: \bequ

1249: \mathcal{L}_{GB}=\frac{1}{2}\left({\bf E}^2-{\bf B}^2-3\Upsilon ({\bf E} \cdot {\bf B})^2\right).

1250: \eequ

1251: 

1252: \subsubsection{Properties of Gauss-Bonnet electromagnetism}

1253: 

1254: The constitutive relations are very simple and easily invertible. ${\bf E}$ and ${\bf H}$ may be expressed in terms of ${\bf B}$ and ${\bf D}$ as 

1255: \bequ

1256: {\bf E}={\bf D}+\frac{3\Upsilon({\bf B} \cdot {\bf D})}{1-3\Upsilon {\bf B}^2}{\bf B}, \ \ \ \ \  {\bf H}={\bf B}+\frac{3\Upsilon({\bf B} \cdot {\bf D})}{1-3\Upsilon {\bf B}^2}{\bf D}+\left(\frac{3\Upsilon({\bf B} \cdot {\bf D})}{1-3\Upsilon {\bf B}^2}\right)^2{\bf B}.

1257: \eequ

1258: It follows from the constitutive relations that ${\bf B} \cdot {\bf E}\neq  {\bf D} \cdot {\bf H}$. Therefore this theory does not admit electric-magnetic duality.

1259: 

1260: The Hamiltonian becomes

1261: \bequ

1262: \mathcal{H}_{GB}=\frac{1}{2}\left({\bf E}^2+{\bf B}^2-3\Upsilon ({\bf E} \cdot {\bf B})^2\right)=\frac{1}{2}\left({\bf D}^2+{\bf B}^2+\frac{3\Upsilon({\bf B}\cdot {\bf D})^2}{1-3\Upsilon{\bf B}^2}\right).

1263: \eequ

1264: The dominant energy condition for the theory is obeyed if $\Upsilon<0$. Hence the first expression for the Hamiltonian shows the energy is positive, whereas the one in terms of the canonically conjugate variables ${\bf B}$ and ${\bf D}$ imposes no upper-bound on the magnitude of the magnetic induction. However, the expressions for ${\bf B}$ and ${\bf D}$ in terms of ${\bf E}$ and ${\bf H}$,

1265: \bequ

1266: {\bf B}={\bf H}-\frac{3\Upsilon({\bf E} \cdot {\bf H})}{1+3\Upsilon {\bf E}^2}{\bf E}, \ \ \ \ \  {\bf D}={\bf E}-\frac{3\Upsilon({\bf E} \cdot {\bf H})}{1+3\Upsilon {\bf E}^2}{\bf H}+\left(\frac{3\Upsilon({\bf E} \cdot {\bf H})}{1+3\Upsilon {\bf E}^2}\right)^2{\bf E},

1267: \eequ

1268: \textit{do} constrain the value of the electric field to be bounded by

1269: \bequ

1270: {\bf E}^2=\frac{1}{3|\Upsilon|}.

1271: \label{limite}

1272: \eequ

1273: Another way to see this is by using our analysis of section 2. For the Gauss-Bonnet electromagnetic theory (\ref{quadra}) becomes

1274: \bequ

1275: \mu =x-\frac{1}{3\Upsilon},

1276: \label{quadragb}

1277: \eequ

1278: from where we can read immediately the limiting field value (\ref{limite}), in agreement with  the discussion following (\ref{quadra}). What happens to the light cones in this limit? Considering ${\bf B}=0$, we see from (\ref{velo}) that the Boillat light cone collapses in the two non-principal directions, manifesting the breakdown of the theory. Moreover, beyond such limit, the causality inequality (\ref{bradyon}) is violated.

1279: 

1280: Yet another manifestation of the limiting electric field can be seen by studying the convexity of the Hamiltonian function, as discussed in section 2.3. The latter property is equivalent to the positive-definiteness of the following six dimensional quadratic form (of the variables ${\bf b}$, ${\bf d}$):

1281: \bequ

1282: \barr{l}

1283: \displaystyle{{\bf b}^2 + {\bf d}^2+ \frac{3\Upsilon}{1-3\Upsilon {\bf B}^2}\left[({\bf D} \cdot {\bf b})^2+ ({\bf B} \cdot {\bf d})^2+2({\bf b} \cdot {\bf d})({\bf D} \cdot {\bf B})  +2({\bf D} \cdot {\bf b})({\bf B} \cdot {\bf d})\right] +\frac{27\Upsilon^3}{(1-3\Upsilon {\bf B}^2)^3}\times } 

1284: \\\\ \displaystyle{\times ({\bf b} \cdot {\bf B})^2({\bf D} \cdot {\bf B})^2 +     \frac{9\Upsilon^2}{2(1-3\Upsilon {\bf B}^2)^2}({\bf D} \cdot {\bf B})\left[4({\bf D} \cdot {\bf b})({\bf B} \cdot {\bf b})+4({\bf B} \cdot {\bf d})({\bf B} \cdot {\bf b})+({\bf D} \cdot {\bf B}){\bf b}^2\right]  .}

1285: \earr

1286: \eequ

1287: Again, for vanishing magnetic induction both eigenvalues will be positive if and only if the electric field is smaller than (\ref{limite}).

1288: 

1289: Since (\ref{quadra}) reduces to (\ref{quadragb}), which has a unique solution for $\mu$ one might think that Gauss-Bonnet electromagnetism admits no bi-refringence. However, as discussed in section 2, when $\varpi$ in (\ref{quadra}) vanishes, the information contained in (\ref{quadra}) might be incomplete. As shown in \cite{lemos} this theory exhibits bi-refringence, with one cone given by the Boillat cone with (\ref{quadragb}) and the second coinciding with the Minkowski light cone. So, the middle illustration in Figure \ref{cones1} is the one to bear in mind, but now, the Minkowski light cone is degenerate; it represents both the background geometry and one of the effective geometries describing the propagation of fluctuations.

1290: 

1291: 

1292: 

1293: 

1294: 

1295: 

1296: 

1297: 

1298: 

1299: 

1300: 

1301: 

1302: 

1303: 

1304: \subsection{The Born-Infeld theory to second order}

1305: The \textit{Born-Infeld} case, ${\mathcal{L}_{KK}}={\mathcal{L}_{BI}}^{(2)}$, $b=-1/2$. The action matches the Born-Infeld action to second order. To match the constants $\Upsilon$ with $\beta$ one must remember that in the Kaluza-Klein ansatz one should replace $A_{\mu}\rightarrow \zeta A_{\mu}$ where $\zeta$ is a constant with dimension length (we are using quantum units, i.e., $c=\hbar=1$). If we write down ${\mathcal{L}_{BI}}$ as

1306: \begin{equation}

1307: {\mathcal{L}_{BI}}=\frac{1}{\beta^2}\left(\sqrt{-g}-\sqrt{-det(g_{\mu \nu}+\beta F_{\mu\nu})}\right),

1308: \end{equation}

1309: the constants match as 

1310: \bequ

1311: \beta^2=-\frac{3\Upsilon \zeta^4 L}{16\pi G_5},

1312: \eequ

1313: where $L$ is the perimeter of the compact dimension (constant since we considered a trivial dilaton).

1314: 

1315: \subsection{The Euler-Heisenberg action}

1316: The \textit{Euler-Heisenberg} case, ${\mathcal{L}_{KK}}={\mathcal{L}_{EH}}$, $b=1/7$. This is the effective action to QED due to one-loop corrections \cite{heisen}. The constant $\Upsilon$ should then be

1317: \begin{equation}

1318: \frac{\Upsilon \zeta^4 L}{16\pi G_5}=-\frac{28\alpha^2}{135m_{e}^4},

1319: \end{equation}

1320: were $\alpha$ is the fine structure constant.

1321: 

1322: 

1323: 

1324: 

1325: 

1326: 

1327: 

1328: 

1329: 

1330: 

1331: 

1332: 

1333: 

1334: 

1335: 

1336: 

1337: 

1338: 

1339: 

1340: 

1341: 

1342: 

1343: 

1344: 

1345: 

1346: 

1347: 

1348: \sect{Metric independence of Black Hole properties}

1349: In this section we consider the propagation of fluctuations of fields (including the electromagnetic) in curved backgrounds. Our main theme will be that even though there maybe more than one metric present in theory, many properties of black holes and their thermodynamic behaviour are metric invariant. In this sense we find that the event horizon and its properties have a universality which goes beyond the universality implied by the equivalence principle.  

1350: 

1351: \subsection{Causality and the Strong Energy Condition}

1352: 

1353: The presence of gravity as a background field is expected to induce changes in the propagation of electromagnetic fluctuations, in the same way a background electromagnetic field does. In fact the former is a consequence of the latter via the Einstein equations. Let us start by using the Boillat metric presented in section 2 to ask when such propagation is causal. The Boillat co-metric (\ref{boicome})

1354: is conformal to

1355: \ben

1356: \left((\mu -x)L_x-yL_y+L\right)g^{\mu \nu} -\left( T^{\mu \nu} -

1357: { T\over 2} g^{\mu \nu} \right).

1358: \een

1359: From now on in this section  we assume that $g$ is the Einstein metric,

1360: rather than some conformal multiple, such as the closed string metric.

1361: This is because we wish to assume 

1362: that the  Einstein equations hold.

1363: Then  the  Boillat co-metric is conformal to

1364: \ben

1365: g_{\rm effective}^{\mu \nu}= g^{\mu \nu} + A R^{\mu \nu},\label{effective}

1366: \label{geff}

1367: \een

1368: where $R^{\mu \nu} $ is the Ricci tensor and

1369: \ben

1370: {1\over 8\pi G_N A}= (x-\mu) L_x+yL_y-L.

1371: \een

1372: If $T_{\mu \nu}$ satisfies the strong energy condition and the Einstein equations hold, then

1373: $R^{\alpha \beta}p_\alpha p_\beta \ge 0$ for all co-vectors

1374: lying inside or on the Einstein co-cone. Thus if $A\ge 0$, the

1375:  Boillat co-cone

1376: lies outside or on the Einstein co-cone. Passing back to the original

1377: Einstein and Boillat cones, remembering that duality reverses inclusions

1378: we see that the strong energy condition together with the requirement

1379: that $A \ge0 $ is a sufficient condition that the Boillat cone

1380: lies inside the Einstein cone. In these circumstances  small

1381: disturbances travel no faster than gravitons. 

1382: 

1383: \subsection{Stationary Event Horizons and the Touching Theorem}

1384: 

1385: Before discussing the even horizon given by the co-metric (\ref{geff}) it seems worth recalling that quantum mechanical effects renormalize the propagation equations in a background gravitational field. For scalars $\phi$ and spinors $\psi$, additional terms appear in the effective action of the form

1386: \bequ

1387: \frac{A}{2}R^{\alpha \beta}\partial_{\alpha}\phi \partial_{\beta}\phi, \ \ \ \ \ \  \frac{iA}{2}R^{\alpha \beta}\bar{\psi}\gamma_{\alpha}\nabla_{\beta}\psi,

1388: \eequ

1389: for some coefficients $A$. 

1390: In the case of scalars these give an effective metric of the form (\ref{geff}). In the case of spinors the discussion given in section 4.2 applies. In the notation used there one has

1391: \bequ

1392: a^{\alpha \beta}=g^{\alpha \beta}+\frac {A}{2}R^{\alpha \beta}

1393: \eequ

1394: and from equation (\ref{chara}) it follows that 

1395: \bequ

1396: g^{\mu \nu}_{effective}=g^{\mu \nu}+AR^{\mu \nu}+\frac{A^2}{4}R^{\mu}_{ \ \alpha}R^{\alpha \nu}.

1397: \eequ

1398: In perturbative calculations one neglects the last third term. The second was computed by Ohkuwa within the Weinberg-Salam model \cite{ohk} yielding

1399: \bequ

1400: A=-\frac{11}{192\pi^2}\frac{e^2 \hbar}{M_{W}^2\sin^2\theta_W c^3},

1401: \eequ

1402: where $\theta_{W}$ is the Weinberg angle and $M_{W}$ is the $W$-boson mass.\footnote{Notice that the different sign in \cite{ohk} is due to the opposite convention for the Riemann tensor.} Since $A$ is negative the effective cones lie outside the Einstein cone. Physically however it is not clear that this implies the neutrino speeds faster than light, because the approximation of retaining only first derivatives in the effective action may break down.

1403:  

1404: 

1405: 

1406: The case of photons is more complex and it involves the Riemann tensor \cite{drum}.\footnote{It maybe of interest to note that if one has as many scalar as spinor degrees of freedom with the same mass going around the loop then the Drummond-Hathrell correction vanishes.}

1407: 

1408: Work on the causality properties of such effective metrics (\cite{shore} and references therein) uncovered a striking result  which is also relevant in the context of non-linear electrodynamics.

1409: 

1410: 

1411: 

1412: 

1413: If the Einstein metric $g$ contains a stationary event horizon

1414: $\cal H$ with null generators $l_\alpha$ and the weak energy condition holds,

1415: $T^{\alpha \beta} l_\alpha l_\beta \ge 0$, then Hawking has shown that

1416: restricted to $\cal H$ 

1417: \ben

1418: R^{\alpha \beta} l_\alpha l_\beta=0.

1419: \een

1420: It follows that 

1421: \ben

1422: g_{\rm effective}^{\alpha \beta}l_\alpha l_\beta=0.

1423: \een

1424: Thus the null generator of the horizon

1425: lies on the effective co-cone. Passing to

1426: the dual space we see that the Einstein cone and the effective

1427: cone actually touch along the null generator of the horizon.

1428: In the case that $A\ge 0$ the effective cone will touch from the inside.

1429: This makes the existence of another effective event horizon outside

1430: the Einstein event horizon unlikely. 

1431: 

1432: 

1433: 

1434:    

1435: 

1436: 

1437: \begin{figure}[t]

1438: \begin{picture}(0,0)(0,0)

1439: \end{picture}   

1440: \centering\epsfig{file=event.eps,width=8cm}   

1441: \caption{The touching theorem.}

1442: \end{figure} 

1443: 

1444: This `Touching Theorem' shows that the concept of an absolute event horizon is more `absolute' than one might have thought. After all because quantum fluctuations will in general affect different particles differently and because the effective metric they see clearly depends upon their couplings one might have imagined that in quantum theory different particles would have different effective event horizons, in contradiction with the classical equivalence principle. However we have seen that to the order we have been working this is not so. All particles see the same event horizon. In other words, the concept of a black hole remains universal in the quantum theory.

1445: 

1446: 

1447: 

1448: 

1449: \subsection{The surface gravity and the universality

1450: of the Hawking temperature}

1451: 

1452: 

1453: As well as the location of the event horizon one might ask whether the thermodynamic properties, such as the temperature, are universal. Because more than one metric is involved, this is not immediately obvious. 

1454: In the case of static solutions, the simplest way of obtaining

1455: the surface gravity $\kappa$ and hence the Hawking temperature 

1456: $T_H= {\kappa/( 2 \pi)}$ is by setting $t=\sqrt{-1} \tau$, $\tau$ real 

1457: and calculating the period $\beta= (T_H)^{-1} = { 2 \pi / \kappa}

1458: $ required to remove

1459: the potential conical singularity at the horizon.

1460: It is clear that there will be no

1461: conical singularity in one metric if and only if there is no conical

1462: singularity in the other metric. Thus we get the same period

1463: $\beta$ for both metrics.

1464: 

1465: If the timelike

1466: Killing vector, which is of course a Killing vector

1467: of both  metrics, is normalized to have unit magnitude at infinity

1468: with respect the Einstein metric, 

1469: then this calculation yields the temperature in Einstein units

1470: as judged by closed observers at infinity. If a background

1471: dilaton $\Phi$ is non-zero then this must be rescaled to get the 

1472: temperature in closed string units. Similarly if the background

1473: Kalb-Ramond field is non-vanishing we must rescale to get the temperature

1474: in open string units. For previous work on the universality of the thermodynamic properties of black holes in generally covariant theories including arbitrary higher derivative interactions see \cite{jacobson}.

1475: 

1476: 

1477: \subsection{Black Hole in a magnetic Field in Einstein-Maxwell theory}

1478: 

1479: One stimulus for this work is the current activity on physics in an external $B$ field. It is worth recalling therefore the properties of a black hole immersed  in an external magnetic field according to Einstein-Maxwell theory. The main point we wish to make is that the thermodynamic properties of the black hole are unaffected by the magnetic field passing through it. This is perhaps not unexpected if one believes that the thermodynamics has its origin in microscopic degrees of freedom whose number and nature are essentially unchanged by external fields. To be concrete the metric is \cite{ernst}

1480: \bequ

1481: ds^2=\Lambda(r,\theta)^2\left(-(1-\frac{2M}{r})dt^2+\frac{dr^2}{1-\frac{2M}{r}}+r^2d\theta^2\right)+\Lambda(r,\theta)^{-2}r^2\sin^2\theta d\phi^2,

1482: \eequ

1483: where $M$ is the analogue of the ADM mass for asymptotically Melvin boundary conditions and

1484: \bequ

1485: \Lambda(r,\theta)=1+\frac{1}{4}B_0^2r^2\sin^2\theta.

1486: \eequ

1487: The Hawking temperature $T_H$ and the area of the event horizon $A_H$ are easily seen to be the same as for the Schwarzchild solution. For some other comments on `non-commutative black holes' see \cite{garhas}. 

1488: 

1489: 

1490: 

1491: 

1492: 

1493: 

1494: 

1495: \sect{Einstein-Schr\"{o}dinger Theory}

1496: 

1497: It is well known that there are many similarities between Born-Infeld theory and the Einstein-Schr\"odinger theory of gravity. In section 4.3 we discussed that the characteristics and therefore the causal structure relevant for fluctuations is analogous to the one for the open string. We now specialize the discussion to some black hole solutions found by Papapetrou.

1498: 

1499: The connection in this theory is not the usual Levi-Civita connection, but rather computed from the relation

1500: \bequ

1501: a_{\alpha \beta, \mu} - a_{\nu \beta}\Gamma^{\nu}_{\alpha \mu}-a_{\alpha \nu}\Gamma^{\nu}_{\mu \beta}=0.

1502: \label{connec}

1503: \eequ

1504: The notation is the one of section 4.3. The Ricci tensor is computed by an expression formally identical to the one in General Relativity, but has both a symmetric piece $R_{(\alpha \beta)}$ and an antisymmetric one $R_{[\alpha \beta]}$. In analogy to the dual Maxwell tensor introduced in section 2.5 we define $P^{\mu \nu}= (a)^{[\mu \nu]}$. The vacuum field equations read

1505: \bequ

1506: R_{(\alpha \beta)}=0, \ \ \ \ \partial_{\beta}(\frak{P}^{\alpha \beta})=0, \ \ \  R_{[[\alpha \beta], \mu]}=0.

1507: \eequ

1508: The contravariant second rank tensor density $\frak{P}^{\mu \nu}=\sqrt{-\det{a^{-1}}} P^{\mu \nu}$. These are similar to the usual Einstein equations, equation of motion and Bianchi identities in non-linear electrodynamics, provided one thinks of $F_{\mu \nu}$ as being analogous $R_{[\mu \nu]}$. We will avoid the issue of positivity of energy in this theory.

1509: 

1510: 

1511: It is perhaps worth remarking that every Ricci flat K\"{a}hler metric including Calabi-Yau spaces, provides a Euclidean solution to this theory. In fact, if $g$ is K\"{a}hler, choosing for $B$ a multiple of the K\"{a}hler form, which is covariantly constant, the Levi-Civita connection of $g$ will solve (\ref{connec}). Hence all equations of motion are obeyed. For these solutions $G^{\mu \nu}_{\rm Eins-Schro} \propto g^{\mu \nu}$ and so there is no ambiguity as to which metric to use. For the analogous phenomenon in Born-Infeld theory see \cite{garhas}.

1512: 

1513: The last comment is not true in general. Spherically symmetric solutions were found by Papapetrou \cite{papape}, corresponding to electrically' and `magnetically' charged spherically symmetric objects. The most general electrical solution reads

1514: \ben

1515: (a^{-1})_{\mu \nu}dx^{\mu}dx^{\nu}= ( 1+ {Q^2 \over r^4}) ( 1-{2MG_N \over r}) dt^2 -{dr^2 \over

1516:  (1-{2MG_N \over r})} - r^2( d\theta ^2 +\sin ^2\theta d\phi^2)

1517:  +{Q \over r^2} dt \wedge dr.

1518: \een

1519: A short calculation reveals that

1520: \ben ds^2_{\rm Eins-Schro}=  

1521: (1-{ 2G_NM \over r} ) dt^2 -{dr^2 \over (1+ {Q^2 \over r^4} )(1-{2G_NM \over r}) } - r^2( d\theta ^2 +\sin ^2\theta d\phi^2).

1522: \een

1523: 

1524: It is striking that some of our previous findings concerning the invariance under the change of metric of the black hole properties still hold in this theory.  For example, in general the causal structure of $g$ and $G^{\rm Eins-Schro}_{\mu \nu}$  differ

1525: but both agree about the location of the event horizon $r=2G_NM$

1526: and its surface gravity which is 

1527: \ben

1528: \kappa= { 1\over 4G_NM } (1+ {Q^2 \over 16 G_N M^4} ).

1529: \een

1530: Note, while the area of the event horizon is given by the same formula in terms of the mass as it is in the Schwarzchild solution, a black hole with $Q\neq 0$ is hotter than the Schwarzchild hole

1531: with the same mass. The hotter temperature is ascribable to the fact that the factor $ (1+ Q^2 / r^4) $ in $g_{00}$

1532:  is blue-shifting rather than redshifting. 

1533: 

1534: 

1535: A magnetic solution found by Papapetrou reads

1536: \bequ

1537: (a^{-1})_{\mu \nu}dx^{\mu}dx^{\nu}=( 1-{2MG_N \over r}) dt^2 -{dr^2 \over

1538:  (1-{2MG_N \over r})} - r^2( d\theta ^2 +\sin ^2\theta d\phi^2)+B_0 r^2 \sin\theta d\theta \wedge d\phi.

1539: \label{papab}

1540: \eequ

1541: The Einstein-Schr\"odinger metric then becomes

1542: \bequ

1543: ds^2_{Eins-Schro}=( 1-{2MG_N \over r}) dt^2 -{dr^2 \over

1544:  (1-{2MG_N \over r})} - (1+B_0^2)r^2( d\theta ^2 +\sin ^2\theta d\phi^2).

1545: \eequ

1546: The physical meaning of the two form in (\ref{papab}) is not clear. It is spherically symmetric and of constant magnitude. It is striking that the metric has a similar form to that of (\ref{bispeb}). The location of the horizon and the surface gravity are independent of $B_0$ and indeed the $r-t$ metric is identical to that of a Schwarzchild black hole.

1547: 

1548: 

1549: 

1550: \section{Conclusions}

1551: The background geometry determined by a gravitational theory might not be the relevant one seen by fluctuations of some test field. This is true even at the classical level, but quantum effects can also renormalize the geometry describing the propagation of fluctuations. One quite interesting example of such distinction was uncovered in work on string propagation in a background $B$ field \cite{seiwit}: in this setup open and closed string fluctuations move, in general, at different velocities. Gravitons and Born-Infeld photons see different light cones. The discussion in sections 2 and 3 shows that the  latter causal structure is the Boillat causal structure, studied long ago in the context of non-linear electrodynamics. Moreover, the $\theta^{\mu \nu}$ parameter describing the non-commutativity of spacetime in the duality established in \cite{seiwit} is just the dual Maxwell tensor of Born-Infeld theory.

1552: 

1553: The open string metric is intrinsically connected to the Born-Infeld action, as we showed in section 4 by including scalars and fermions in a Born-Infeld type action, and showing the characteristics are determined by the open string metric. At this point a question requires more thorough understanding. In the context of string theory, the Born-Infeld action describes brane dynamics. The brane world scenario motivated by \cite{randsun} tries to bind gravitons to the brane. The difficulty is, of course, that gravitons are closed rather than open string modes. But if this program is successful, either the brane graviton sees different light cones from the other spin brane fields or, if it is governed by the open string metric, the question arises to what effective field theory describes such gravitons. The Einstein-Schr\"odinger theory seems to have exactly the characteristics we would then be looking for. But it seems to suffer from instabilities \cite{deser}.

1554: 

1555: In section 5 we looked at higher order gravity and Kaluza-Klein theory. To lowest order in $\alpha'$, the abelian truncation of the effective open string theory (Maxwell's theory) is obtained by Kaluza-Klein compactification of the effective closed string theory (Einstein's gravity). But this does not seem to hold to the next order in $\alpha'$: the Gauss-Bonnet contribution to the effective closed string theory gives upon Kaluza-Klein reduction what we named as `Gauss-Bonnet electromagnetism', distinct from both the Euler-Heisenberg theory and the Born-Infeld theory to this order.\footnote{We remark that the Born-Infeld theory to second order does not coincide with the Euler Heisenberg theory, which might be puzzling. But at least in the supersymmetric case they do coincide to this order \cite{GR}.} We remarked however that for purely electromagnetic excitations, the no-ghost requirement is weaker and by an appropriate choice of couplings one can obtain the latter two theories to this order.  It would be interesting to consider also the full Kaluza-Klein theory, with all excitations present. This would give non-minimal gravitational-electromagnetic couplings, violating the equivalence principle. Gravitational bi-refringence and dispersion effects might also be present, although the latter seem only to occur at even higher order in $\alpha'$ \cite{myers}.

1556: 

1557: In section 6 we made use of a result know in the literature as the `Touching Theorem' to show that the propagation of fluctuations in non-linear electrodynamics coupled to gravity will see a universal event horizon and black hole temperature. Such comment also holds for one-loop corrected propagators in curved spacetime. In this way black holes don't seem to `leak'. We also noticed that similar invariance is seen for a black hole immersed in a magnetic field in Einstein-Maxwell theory. It would perhaps be interesting to explicitly look at such black holes in a `Melvin Universe' for the case of non-linear electrodynamics.

1558: 

1559: 

1560: 

1561: 

1562: 

1563: \section *{Acknowledgments}

1564: G.W.G. thanks Michael Green, Koji Hashimoto and Richard Kerner for helpful discussions. We would also like to thank M.Clayton and T.Jacobson for helpful correspondence. Some of the ideas in this paper surfaced long ago in discussions with  Malcolm Perry. C.H. is supported by FCT (Portugal) through grant no. PRAXIS XXI/BD/13384/97. This work is also supported by the PPARC grant PPA/G/S/1998/00613.

1565: 

1566: 

1567: 

1568: 

1569: \begin{thebibliography}{99}

1570: 

1571: \bibitem{BI}M. Born, L.Infeld, {\em Foundations of the New Field Theory}, Proc. Roy. Soc. {\bf 144A} (1934) 425.

1572: 

1573: \bibitem{hawell}S.Hawking, G.Ellis, {\em The large scale structure of space-time}, Cambridge University Press (1973), section 4.3.

1574: 

1575: \bibitem{S1}E.Schr\"odinger, {\em Contributions to Born's New Theory of the Electromagnetic Field}, Proc. Roy. Soc.  {\bf 150A} (1935) 465. 

1576: 

1577:  \bibitem{S2}E.Schr\"odinger, {\em Non-Linear Optics}, Proc. Roy. Irish. Acad. {\bf A 47} (1942) 77.

1578: 

1579: \bibitem{S3}E.Schr\"odinger, {\em A new exact solution in non-linear optics 

1580: (two-wave-system)}, Proc. Roy. Irish. Acad. {\bf A 49} (1943) 59.

1581: 

1582: \bibitem{CM}C.Callan, J.Maldacena, {\em Brane Dynamics From the Born-Infeld Action}, Nucl. Phys. {\bf B513} (1998) 198, hep-th/9708147.

1583: 

1584: \bibitem{hull}C.Hull, P.Townsend, {\em Unity of Superstring Dualities}, Nucl. Phys. {\bf B438} (1995) 109,  hep-th/9410167. 

1585: 

1586: 

1587: \bibitem{porratti}C.Bachas, M.Porratti {\em  Pair Creation of Open strings in an electric field}, Phys. Lett. {\bf B296} (1992) 77, hep-th/9209032.

1588: 

1589: \bibitem{GR}G.W.Gibbons, D.Rasheed, {\em Electric-magnetic duality rotations in non-linear electrodynamics}, Nucl. Phys. {\bf B454} (1995) 185.

1590:  

1591: 

1592: \bibitem{B}M.Born, {\em Reciprocity theory of elementary particles}, Rev. Mod. Phys {\bf 21} (1949) 463.

1593: 

1594: 

1595: 

1596: 

1597: \bibitem{G}G.W.Gibbons, {\em Born-Infeld particles and Dirichlet p-branes}, Nucl. Phys. {\bf B514} (1998) 603.                                                       

1598: 

1599: \bibitem{christ}D.Christodoulou, S.Klainerman, {\em The global nonlinear stability of Minkowski space}, Princeton Mathematical Series, {\bf 41}, Princeton University Series 1993.

1600: 

1601: 

1602: \bibitem{B6}G.Boillat, {\em Vari\'et\'es caract\'eristiques ou surfaces d'ondes en \'electrodynamiques non lin\'eaire}, C. R. Acad. Sci. Paris {\bf 262} (1966) 1285. 

1603: 

1604: 

1605: \bibitem{B11}G.Boillat, {\em Surfaces d'ondes compar\'es de la theorie

1606: d'Einstein-Schroedinger et de l'\'electrodynamique

1607: non lin\'eaire; champs absolus}, C. R. Acad. Sci. Paris {\bf 264} (1967) 113.

1608: 

1609: 

1610: 

1611: \bibitem{B13}G.Boillat, {\em Nonlinear Electrodynamics: Lagrangians and Equations of Motion}, J. Math. Phys. {\bf 11} (1970) 941.    

1612: 

1613: \bibitem{B14}G.Boillat, {\em Exact Plane-Wave Solution of Born-Infeld Electrodynamics}, Lett. al Nuovo Cimento {\bf 4} (1972) 274.

1614: 

1615: 

1616: \bibitem{B15}G.Boillat, {\em Shock relations in Non-Linear Electrodynamics}, Phys. Lett. {\bf 40A} (1972) 9.

1617: 

1618: 

1619:  

1620: \bibitem{BB}Z.Bialynicka-Birula, {\em Solitary Waves in Born-Infeld Electrodynamics}, Bulletin de l'Academie Polonaise des Sciences {\bf 27} (1979) 41.

1621: 

1622: \bibitem{B18}G.Boillat, {\em Convexit\'e et hyperbolicit\'e

1623: en \'electrodynamique non-lin\'eaire}, C. R. Acad. Sci. Paris {\bf 290} (1980)

1624: 259.

1625: 

1626: 

1627: \bibitem{GRu}G.W.Gibbons, R.Ruback, {\em Classical Gravitons and Their Stability in Higher Dimensions}, Phys. Lett. {\bf 171B} (1986) 390.

1628: 

1629: \bibitem{Griffiths}J.Griffiths, {\em Colliding plane waves in General Relativity}, Oxford Mathematical Monographs, Oxford University Press 1991.

1630: 

1631: 

1632: \bibitem{P}J.F.Pleba\~nski, {\em Lectures on Non-linear Electrodynamics}, Published by Nordisk Institut for Teoretisk Atomfysik (NORDITA) 1970.

1633: 

1634: \bibitem{shore}G.Shore, {\em "Faster than Light" Photons in Gravitational Fields - Causality, Anomalies and Horizons}, Nucl. Phys. {\bf B460} (1996) 379.

1635: 

1636: 

1637: \bibitem{GDP}S.A.Guti\'errez, A.L.Dudley and J.F.Pleba\~nski,

1638: {\em Signals and discontinuities in general realtivistic nonlinear

1639: electrodynamics}, J. Math. Phys. {\bf 22} (1981) 2835.

1640: 

1641: 

1642: 

1643: \bibitem{moncrief}D.Eardley, V.Moncrief {\em The global existence of Yang-Mills-Higgs fields in 4-dimensional Minkowski space. I. Local existence and smoothness properties}, Comm. Math. Phys. {\bf 83} (1982) 171; {\em The global existence of Yang-Mills-Higgs fields in 4-dimensional Minkowski space. II. Completion of proof.}, Comm. Math. Phys. {\bf 83} (1982) 193.

1644: 

1645: 

1646: \bibitem{zwiebach}B.Zwieback, {\em Curvature Squared terms and String Theories}, Phys. Lett. {\bf 156B} (1985) 315.

1647: \bibitem{lemos}J.Lemos, R.Kerner, {\em The Born-Infeld Electromagnetism in Kaluza-Klein Theory}, J. Gravit. and Cosmology {\bf 6} (2000) 49, hep-th/9907187.

1648: 

1649: \bibitem{kerner}R.Kerner {\em Infinite Dimensional Lie Algebras and Quantum Field theory}, Proceedings of the Varna Summer School 1987, ed. H.Doebner, J.Henning, T.Palev, World Scientific; {\em Non-linear electrodynamics in Kaluza-Klein theory}, C. R. Acad. Sci. II , {\bf 304} (1987) 621.

1650: 

1651: \bibitem{bopp}F.Bopp, {\em Eine lineare Theorie des Elektrons}, Ann. Phys. (Leipzig) {\bf 38} (1940) 345; H.Podolsky, {\em A Generalized electrodynamics; Part I- Non-Quantum}, Phys. Rev. {\bf 62} (1941) 68.

1652: 

1653: \bibitem{papape}A.Papapetrou, {\em Static spherically symmetric solutions in the unitary field theory}, Proc. Roy. Irish Acad. {\bf 52} (1948) 69.

1654: 

1655: \bibitem{seiwit}N.Seiberg, E.Witten, {\em String Theory and Noncommutative Geometry}, J.H.E.P. 9909 (1999) 32, hep-th/9908142. 

1656: 

1657: 

1658: \bibitem{drum}I.Drummond, S.Hathrell, {\em QED vacuum polarization in a background gravitational field and its effect on the velocity of photons}, Phys. Rev. {22D} (1980) 343.

1659: 

1660: \bibitem{penrin}R.Penrose, W.Rindler, {\em Spinors and Space-Time}, Cambridge Monographs on Mathematical Physics, Cambridge University Press, 1987, Vol.1.

1661: 

1662: \bibitem{V}G.Veneziano, {\em A stringy nature needs just two constants}, Europhys. Lett. {\bf 2} (1986) 199.

1663: 

1664: \bibitem{ernst}F.Ernst, {\em Black Holes in a magnetic Universe}, J.Math. Phys. {\bf 17} (1976) 54.

1665: 

1666: \bibitem{her}C.A.R.Herdeiro, {\em Special Properties of Five Dimensional BPS Rotating Black Holes}, Nucl. Phys. {\bf B582} (2000) 363, hep-th/0003063.

1667: 

1668: 

1669: 

1670: \bibitem{novello}M.Novello, V.Lorenci, J.Salim, R.Klippert, {\em Geometrical aspects of light propagation in nonlinear electrodynamics}, gr-qc/9911085.

1671: 

1672: 

1673: 

1674: \bibitem{beato}E. Ay\'{o}n-beato, A.Garcia, {\em Regular Black Hole in General Relativity Coupled to Non-linear Electrodynamics}, gr-qc/9911046; H.Yajima, T.Tamaki, {\em Black Hole solutions in Euler-Heisenberg theory}, gr-qc/0005016.

1675: 

1676: 

1677: \bibitem{Licher}A.Lichnerowicz, {\em Th\'eories relativistes de la gravitation et de l'electromangn\'etisme}, Masson, Paris 1955; {\em Compatibilit\'e des \'equations de la th\'eorie unitaire du champ d'Einstein}, J.Rational Mech. Anal. {\bf 3} (1954) 487.

1678: 

1679: \bibitem{maurer}F.Maurer-Tison, {\em Sur les Vari\'et\'es caract\'eristiques de la th\'eorie unitaire du champ d'Einstein}, C. R. Acad. Sci. Paris {\bf 242} (156) 1127; {\em Th\'eorie unitaire et \'electromagn\'etisme dans la mati\`ere } C. R. Acad. Sci. Paris {\bf 242} (156) 3042; see Maths Reviews MR 17,907c; MR 18,262b.

1680: 

1681: 

1682: \bibitem{randsun}L.Randall, R.Sundrum, {\em An alternative to compactification}, Phys. Rev. Lett. {\bf 83} (1999) 3370, hep-th/9906064.

1683: 

1684: \bibitem{VA}V.Akulov, D.Volkov, {\em Is the Neutrino a goldstone particle?}, Phys. Lett. {\bf B46} (1973) 109.

1685: 

1686: \bibitem{deser}T.Damour, S.Deser, J.McCarthy, {\em Nonsymmetric Gravity Theories: Inconsistencies and a Cure}, Phys. Rev. {\bf D47} (1993) 1541, gr-qc/9207003.

1687: 

1688: \bibitem{ohk}Y.Ohkuwa, {\em Effect of a Background Gravitational Field on the Velocity of Neutrinos}, Prog. Theo. Phys. {\bf 65} (1981) 1058.

1689: 

1690: \bibitem{myers}R.Lafrance, R.Myers, {\em Gravity's Rainbow: Limits for the applicability of the equivalence principle}, Phys. Rev. {\bf D51} (1995) 2584, hep-th/9411018.

1691: 

1692: \bibitem{heisen}W.Heisenberg, H.Euler, {\em Folgerungen aus der Diracschen Theorie des Positrons}, Z.Physik, {\bf 98} (1936) 714.

1693: 

1694: \bibitem{kastor}D.Kastor, J.Traschen, {\em Dynamics of the DBI spike soliton}, Phys. Rev. {\bf D61} (2000) 024035, hep-th/9906237.

1695: 

1696: \bibitem{park}D.Park, S.Tamarian, H.Muller-Kirsten, J.Zhang, {\em D-branes and their absorptivity in Born-Infeld theory}, hep-th/0005156.

1697: 

1698: \bibitem{garhas}G.W.Gibbons, K.Hashimoto, {\em Non-linear electrodynamics in curved backgrounds}, hep-th/0007019.

1699: 

1700: \bibitem{jacobson}T.Jacobson, G.Kang, R.Myers, {\em On Black Hole entropy}, Phys. Rev. {\bf D49} (1994) 6587.

1701: 

1702: \end{thebibliography}

1703: 

1704: 

1705: 

1706: \end{document}

1707: 

1708: 

1709: 

1710: 

1711: 

1712: 

1713: