1: \documentclass[aps,pra,showpacs]{revtex4}

2: \usepackage{graphicx}

3: \usepackage{comment}

4: %\usepackage{showkeys}

5: \begin{document}

6: 

7: \title{Frustrated total internal reflection and the illusion of

8: superluminal propagation}

9: 

10: \author{Vera L. Brudny}

11: \email{vera@df.uba.ar}

12: \affiliation{Departamento de F\'{\i}sica, Facultad de Ciencias Exactas

13: y Naturales, Universidad de Buenos Aires, Cd. Universitaria,

14: Pabell\'on 1, C1428 EHA Buenos Aires, Argentina}

15: \author{W. Luis Moch\'an}

16: \email{mochan@fis.unam.mx}

17: \affiliation{Instituto de Ciencias F\'{\i}sicas, Universidad Nacional

18:   Aut\'onoma de M\'exico, Apartado Postal 48-3, 62251 Cuernavaca,

19:   Morelos, M\'exico} 

20: \date{\today}

21: 

22: \begin{abstract}

23: 

24: 

25: We analyze the propagation of a pulse across a vacuum gap separating 

26: opposite flat parallel faces of two

27: transparent dielectrics by means of an explicitly causal and

28: retarded propagator constructed directly from the free-space wave

29: equation. Nevertheless, our approach yields apparently superluminal

30: propagation for the case of frustrated total internal reflection

31: (FTIR), that is, a transmitted wave packet appears on the far side of

32: the gap at the same time that the corresponding incident packet

33: crosses the front one.  Thus, in this example superluminality is just

34: an illusion, being consistent with both casuality and classical

35: electrodynamics.  We study the origin of the apparent superluminality

36: in this case, which is inherent to light pulse propagation in free

37: space and does not depend on the particulars of light-matter

38: interaction, and find that it is due to propagation from the lateral

39: wings of the incident pulse to the central part of the

40: transmitted pulse. Thus, notwithstanding their similarities, FTIR is

41: not equivalent to 1D tunneling. We propose experiments to test our

42: explanation of superluminality using opaque screens to block part of

43: the wavefront, although we demonstrate that the

44: propagation of smooth finite pulses constrained to be made up

45: completely of evanescent Fourier components is

46: indistinguishable from truly superluminal propagation, i.e., it may be

47: completely accounted for using an explicitely superluminal and acausal

48: propagator as well as the causal subluminal one.

49: 

50: \end{abstract}

51: 

52: \pacs{

53: 42.25.Bs, %Wave propagation, transmission and absorption 

54: 41.20.Jb, %	Electromagnetic wave propagation; radiowave propagation 

55: 03.65.Xp, %	Tunneling, traversal time, quantum Zeno dynamics

56: 42.25.Gy %	Edge and boundary effects; reflection and refraction

57: }

58: 

59: \maketitle

60: 

61: \begin{comment}

62: 42.25.Fx 	Diffraction and scattering

63: 03.30.+p 	Special relativity

64: 42.79.Fm 	Reflectors, beam splitters, and deflectors

65: 42.79.Bh 	Lenses, prisms and mirrors

66: 78.68.+m 	Optical properties of surfaces

67: 03.65.Pm 	Relativistic wave equations

68: 01.55.+b 	General physics

69: \end{comment}

70: 

71: 

72: \section{Introduction}

73: 

74: The physics of light propagation is a topic of active research, due to

75: its relevance to several technological applications and basic

76: research. Recent research on new materials has shown that it is

77: possible to exercise an extraordinary control on the propagation of

78: light pulses, which in turn has generated new interest in both

79: practical and fundamental questions on light propagation. Moreover,

80: despite the fact that Maxwell completed the formulation of the

81: classical theory of electromagnetism in 1864 and Einstein's special

82: theory of relativity was presented in 1905, some hot controversies

83: continue to arise on the subject of superluminal propagation

84: \cite{wang07,winful07,pereyra,ranfagni07,winful05,winful03,buttiker03,winful03-1,buttiker03-1,winful03-2}

85: and its possible implications for both classical and quantum

86: information theory \cite{qi,de_angelis}.  Recent reviews on the

87: subject can be found in Refs. \cite{winful06,boyd02}

88: 

89: 

90: 

91: One consequence of the special theory of relativity is that no signal

92: can cause an effect outside the light cone of its source.  Violation of this

93: principle of relativistic causality leads to paradoxes such as that of

94: an effect preceding its cause \cite{azbel,garrison1998}.  When dealing

95: with light propagation in a material characterized by a given dispersion

96: relation $\omega(\vec k)$ between the frequency $\omega$ and the wave vector

97: $\vec k$, several {\em velocities} may be

98: defined, such as the {\em phase velocity} $v_\phi = \omega/k$ and the {\em

99:   group velocity} $v_g=\nabla_{\vec k}\omega$. It is

100: recognized that under certain conditions \cite{jackson} both

101: velocities can exceed the speed of light in 

102: vacuum $c$. This does not

103: contradict the postulates of the special theory of relativity, for it

104: has been recognized since the works of Sommerfeld \cite{sommerfeld}

105: and Brillouin \cite{brillouin-wp} that in order not to violate the

106: principle of causality it is the {\em information velocity} that must

107: not exceed $c$.  The question that arises then is what is an

108: appropriate and operative definition of {\em information

109:   velocity}. There have been several discussions and proposals on this

110: subject, but the question is not yet settled

111: \cite{diener96,diener97,kuzmich,stenner,stenner05,ranfagni06}.

112: 

113: 

114: The vast majority of the published work concerning superluminal pulse

115: propagation deals with light propagation in material media, and the

116: usual analysis attempts to explain how the interaction of

117: electromagnetic radiation with the medium affects the propagation of

118: light pulses in such a way that they appear to travel

119: superluminally.  In a pioneering paper \cite{icsevgi}, Icsevgi and

120: Lamb performed a theoretical investigation of the propagation of

121: intense laser pulses through a laser amplifier. Apparent superluminal

122: light propagation has been reported in gain-assisted systems

123: \cite{wang00,janowicz,huang} as well as in birefringent crystals

124: \cite{solli03,brunner,halvorsen}, composite media and photonic

125: crystals \cite{kulkarni,safian} and dispersive media

126: \cite{bigelow,talukder}.

127: 

128: There have also been claims of evidence of superluminal propagation in

129: free space \cite{mugnai00,mugnai05} and during {\em optical tunneling} in

130: frustrated total internal reflection (FTIR) configurations

131: \cite{carey00,mochan01,carey01,shaarawi}. Optical tunneling has been

132: studied by several authors, both theoretically and experimentally

133: \cite{oe,barbero,reiten,balcou97,resch}. Some of the observed results

134: are still subject of debatable interpretation \cite{winful06} and do

135: not close the subject of whether there are possibilities for

136: superluminal transmission of information in such systems. We have

137: therefore chosen to address this subject in way that leads to

138: straightforward interpretation of the results while resorting only to

139: classical electromagnetic theory. We claim that although some results

140: may {\em appear} to indicate superluminal propagation, there is no

141: real superluminal transfer of information.

142: 

143: In this paper we analyze mathematically the

144: propagation of a pulse across a vacuum gap separating opposite flat

145: parallel faces of two transparent

146: dielectrics by means of an explicitly causal and

147: retarded propagator constructed directly from the free-space wave

148: equation. Our results yield indeed an apparent

149: superluminal propagation corresponding to the conditions of FTIR,

150: but they show explicitly that it is consistent with both casuality and

151: with classical electrodynamics. Our example shows

152: superluminality effects inherent to light pulse propagation in free

153: space which therefore does

154: not depend on the particulars of light-matter interaction.  The {\em

155:   illusion} of 

156: superluminality consists of transmitted pulses arriving to the

157: far side of the gap in synchrony with the crossing of the front

158: surface by the incident pulse.  We explain this illusion of

159: superluminal behavior in terms of a causal, subluminal propagation, 

160: taking into account the spatial extent of the incident pulse along its

161: transverse as well as its longitudinal directions and we propose

162: experiments to demonstrate the retarded nature of propagation in

163: FTIR. Nevertheless, we find that for constrained pulses fully made up of

164: evanescent Fourier components, subluminal and superluminal

165: propagation in FTIR experiments are indistinguishable.

166: 

167: 

168: This paper is organized as follows. We first introduce a propagator

169: that will allow the analyzis of electromagnetic pulse propagation

170: across a vacuum gap

171: (Section \ref{secpropagator}). This propagator is both casual and

172: retarded and complies with the classical electromagnetic theory.  We

173: then study the propagation across the gap of pulses that arrive as

174: plane waves with well defined angles of incidence. In Section

175: \ref{secnormal} we study the case of

176: subcritical angles, yielding non-evanescent transmitted waves. In

177: Section \ref{secevan} we study the case of hypercritical angles,

178: yielding evanescent waves. We conclude that the propagation of a

179: light pulse in a FTIR configuration may appear superluminal and

180: acausal but that it is actually subluminal and that propagation

181: has to account necessarily for the lateral wings of the incident

182: pulse. In Section \ref{secscreens} we suggest  experiments that

183: might demonstrate  the actual causal and subluminal nature of the

184: apparent superluminal behavior by using sharp opaque screens that

185: block parts of the incident wave so that its extent becomes finite

186: along both its propagation and its transverse

187: directions. Nevertheless, as the borders of these screens  

188: produce propagating diffracted waves,

189: in Section \ref{secsmooth} we eliminate them and we study incident

190: pulses that are finite along several spatial

191: directions but that have a smooth profile. We obtain that if they are 

192: comprised of evanescent Fourier components only, they appear

193: to propagate 

194: superluminally through the vacuum gap, even though their behavior is determined

195: by our causal retarded propagator. In Section \ref{secevprop} we construct an

196: alternative acausal, superluminal propagator, and prove that it is exactly

197: equivalent to the causal and retarded propagator when applied to fully

198: evanescent finite pulses. Thus, for such constrained pulses, it is

199: impossible to distinguish causal subluminal from acausal and

200: superluminal propagation;  the illusion of superluminality

201: appears to be 

202: not only a matter of interpretation of the {\em result} of the

203: propagation, but may be also present in the description of the

204: propagation process itself. We present our conclusions in Section

205: \ref{secconclusion}.

206: 

207: 

208: \section{Propagator}\label{secpropagator}

209: Consider two transparent dielectrics occupying the regions $z\le0$ and

210: $z\ge d$. In this section we obtain the causal and retarded propagator

211: that 

212: describes the motion of a pulse across a vacuum gap $0<z<d$ spanning from the

213: interface at $z=0$ to that at $z=d$ (Fig. \ref{normal}).

214: \begin{figure}

215: \includegraphics{fig1}%{pr.1}

216: \caption{\label{normal}Pulse with arbitrary profile incident at an

217: angle $\theta_i<\theta_c$ onto the surface 

218: $z=0$ of a dielectric with index of refraction $n=2$. The pulse is

219: transmitted into a vacuum gap and into a second dielectric at

220: $z=d$. The speed of propagation within each media, $c/n$ and $c$, are

221: indicated, as well as the speed of propagation of the wavefronts along

222: the interfaces $v_\parallel$.} 

223: \end{figure}

224: As we want to consider explicitly the angle of incidence onto the

225: interface $z=0$, we cannot treat the problem beforehand as if it were 1D. 

226: For simplicity, we will assume full translational symmetry along the

227: $y$ direction, so that our problem becomes 2D.

228: Thus, we start with the scalar wave equation

229: \begin{equation}\label{greenseq}

230: (\nabla^2-\frac{1}{c^2}\frac{\partial^2}{\partial

231: t^2})G_0(t,x,z;t',x',z') =  \delta(t-t') \delta(x-x') \delta(z-z')

232: \end{equation}

233: with a unit singular point source fired

234: at time $t'$ at position $(x',z')$, which is solved by the

235: explicitly causal 

236: and retarded free space Green's function 

237: \cite{inmorse}

238: \begin{equation}\label{G0}

239: G_0(t,x,z;t',x',z')=\frac{c}{2\pi}\frac{\Theta\left(c(t-t')-\sqrt{(x-x')^2+(z-z')^2}\right)}{\sqrt{c^2(t-t')^2-(x-x')^2-(z-z')^2}},

240: \end{equation}

241: where $\Theta(\ldots)$ is the Heaviside unit step function. Using

242: image theory we can construct a Green's function that obeys 

243: Dirichlet boundary conditions at the surface $z=0$,

244: \begin{equation}\label{G}

245: G(t,x,z;t',x',z')=G_0(t,x,z;t',x',z')-G_0(t,x,z;t',x',-z').

246: \end{equation}

247: In the half-space $z>0$ Green's theorem yields the solution

248: \cite{inmorse}

249: \begin{equation}\label{propagate}

250: \phi(t,x,z)=\int dx' \int dt'\,P(t,x,z;t',x',0^+) \phi(t',x',0^0)

251: \end{equation}

252: of the homogeneous scalar wave equation 

253: that is outgoing as $z\to\infty$ and is null in the remote past, where

254: $\phi(t',x',0^+)$ denotes its previous values on the boundary $z=0^+$,

255: and 

256: \begin{equation}\label{propagator}

257: P(t,x,z;t',x',0^+)=\frac{\partial}{\partial

258: z'}\left. G(t,x,z;t',x',z')\right|_{z'=0^+}

259: \end{equation}

260: is the causal retarded propagator of the problem. We may identify the

261: field $\phi$ with the component $E_y$ of the electric field $\vec E$

262: in the case of a TE or $s$ polarized incoming wave, and with the

263: component $B_y$ of the magnetic field $\vec B$ in the case of TM or

264: $p$ polarization. The propagator

265: (\ref{propagator}) does not account for the presence of the two

266: dielectrics bounding the air gap. Thus, the field (\ref{propagate})

267: has no

268: information about the multiple reflections at the boundary of the

269: gap. In principle, 

270: these can be incorporated by reflecting the field at

271: the interfaces $z=0,d$ using the appropriate Fresnel coefficient

272: and propagating it back and forth across the gap with the propagator

273: (\ref{propagator}) for the $z=0$ surface and a similar one for the

274: $z=d$ surface. The total field would then be the sum of all the

275: multiply reflected fields and would have information about the

276: electromagnetic properties of the reflecting surfaces. In this paper

277: we will restrict ourselves to an analysis of the first crossing of the

278: air gap $0^+\to d^-$, and thus our results will be unrelated to the

279: nature of the bounding media.

280: 

281: Substituting 

282: Eq. (\ref{G0}) into (\ref{G}) and (\ref{propagator}) we obtain

283: \begin{equation}\label{P}

284: P(t,x,z;t',x',0^+)=-\frac{c}{\pi} \frac{\partial}{\partial z} 

285: \frac{\Theta\left(c(t-t')-\sqrt{(x-x')^2+z^2}\right)}{\sqrt{c^2(t-t')^2-(x-x')^2-z^2}}.

286: \end{equation}

287: The field $\phi$ can then be written in terms of an ancillary

288: function 

289: \begin{equation}\label{phivspsi}

290: \phi(t,x,z)=-\frac{\partial}{\partial z}\psi(t,x,z),

291: \end{equation}

292: where

293: \begin{equation}\label{psi}

294: \psi(t,x,z) =\frac{c}{\pi} \int dx' \int dt'\,

295: \frac{\phi(t',x',0^+)}{\sqrt{c^2(t-t')^2-(x-x')^2-z^2}},

296: \end{equation}

297: plays the role of a potential and the integration is performed within

298: the region 

299: $c(t-t')>\sqrt{(x-x')^2+z^2}$. Clearly, the procedure above yields a

300: causal (sub)luminal propagation from the $z=0^+$ plane to any point in

301: the $z>0$ vacuum.

302: 

303: \section{Non-Evanescent wave transmission}\label{secnormal}

304: 

305: We consider now that an arbitrarily shaped pulse impinges {\em at a

306:   well defined angle} $\theta_i$ on the inside surface $z=0^-$ of a

307: homogeneous non-dispersive dielectric with index of refraction $n$

308: (Fig. \ref{normal}).  The incident pulse is therefore described by an

309: arbitrary function $\phi_i(t,x,z)=f_i[t-(n/c)\hat n_i\cdot\vec\rho]$

310: of a single variable $t-(n/c)\hat n_i\cdot\vec\rho$, where $\hat

311: n_i\equiv (\sin\theta_i,\cos\theta_i)$ is a unit vector pointing along

312: the angle of incidence $\theta_i$ and $\vec \rho\equiv(x,z)$. Notice

313: that in this case the incident wavefronts have an infinite extension

314: in the direction normal to $\hat n_i$. At $z=0^+$ and after being

315: transmitted into vacuum, the outgoing field can therefore be written

316: as

317: \begin{equation}\label{phivsf}

318: \phi(t,x,0^+)=f_t(t-x/v_\parallel)

319: \end{equation}

320: where $f_t$ is related to the arbitrary function $f_i$ and the Fresnel

321: amplitude for transmission from the dielectric into vacuum. As we are

322: concerned only with propagation across the vacuum gap, we will take $f_t$

323: as given and we will disregard its relation with $f_i$, which would

324: involve the dielectric properties of the incident medium.

325: 

326: 

327: The

328: intersection of the pulse with the interface $z=0$ 

329: is therefore seen to travel along $x$ with velocity

330: $v_\parallel=c/(n\sin\theta_i)$ (not to be confused with the parallel

331: component of the incident velocity $\hat n_i c/n$). For incidence angles 

332: smaller than the critical angle $\theta<\theta_c=\sin^{-1}(1/n)$,

333: $v_\parallel>c$ and we have normal transmission, while for

334: $\theta>\theta_c$, $v_\parallel<c$ and 

335: total internal reflection ensues.

336: 

337: In order to set up a reference with which to compare the evanescent

338: case, in this section we employ our propagator to 

339: study a non-evanescent plane pulse. Thus,

340: we consider here the case $\theta<\theta_c$ and we substitute

341: Eq. (\ref{phivsf}) 

342: into Eq. (\ref{psi}) to obtain

343: \begin{equation}\label{psinormal}

344: \psi(t,x,z)=\frac{c}{\pi} \int dt''\, f_t(t'') \int dx'\,

345: \left(c^2(t-t''-x'/v_\parallel)^2 - (x-x')^2-z^2\right)^{-1/2}

346: \end{equation}

347: after changing integration variables from $t'$ to $t''\equiv

348: t'-x'/v_\parallel$. The integration (\ref{psinormal}) has to be

349: performed over the region $c(t-t'')-\mu x'>\sqrt{(x-x')^2+z^2}$, where

350: $\mu\equiv c/v_\parallel$. Thus, $t''$ has an upper bound

351: \begin{equation}\label{tmax}

352: t_m=t-(\mu x - \nu z)/c,

353: \end{equation}

354: where $\nu\equiv\sqrt{1-c^2/v_\parallel^2}$, 

355: and for each value of $t''<t_m$, $x'$ is bounded by the limits 

356: \begin{equation}\label{xmasmenos}

357: x'_\pm=-(1/\nu^2)\left(x-\mu c(t-t'') \pm\sqrt{[\mu

358:     x-c(t-t'')]^2-\nu^2 z^2}\right).

359: \end{equation} 

360: 

361: A simple interpretation of Eqs. (\ref{tmax}) and (\ref{xmasmenos}) can

362: be obtained with the help of Fig. \ref{cono1}.

363: \begin{figure}

364: \includegraphics{fig2}%{conos1.0}

365: \caption{\label{cono1}Contributions to the non-evanescent wave observed

366: at an event ${\cal E}=(t,x,z)$. Some world lines

367: $(ct',x'=v_\parallel(t'-t''),0)$  of source points labeled by

368: fixed values of $t''$ and thus moving on the $z=0$ plane along

369: with the incident pulse are shown. The positions $x'_\pm$ denote the

370: intersections of the past light cone of $\cal E$ with one of the world

371: lines plotted. We consider the case $v_\parallel>c$. The

372: coordinate and time axes are indicated.} 

373: \end{figure}

374: Consider an event ${\cal E}=(ct,x,z)$ defined by the observation of

375: the field at a given position $(x,z)$ with $z>0$ and at a given time

376: $t$. Causality requires that only events within the past light-cone of

377: $\cal E$ are able to influence it.  Notice that a given value of $t''$

378: denotes a point that moves along $x$ keeping a fixed position with

379: respect to the

380: intersection of the incident pulse with the $z=0$ interface. In

381: Fig. \ref{cono1} we show the world lines $(ct',x',0)$ of a few such

382: points. Since each of them moves with speed $v_\parallel>c$, most of

383: its world line lies outside the past light cone of $\cal E$. Only if

384: $t''<t_m$ can it actually cross the light cone, entering and leaving

385: at positions $x'_-$ and $x'_+$ respectively.

386: 

387: 

388: After the change of variables from $x'$ to

389: \begin{equation}\label{eta}

390: \eta\equiv \frac{x'-[x-\mu c(t-t'')]}{\sqrt{[\mu

391: x-c(t-t'')]^2-\nu^2 z^2}}, 

392: \end{equation}

393: Eq. (\ref{psinormal}) simplifies to

394: \begin{equation}\label{psivseta}

395: \psi(t,x,z)=\frac{c}{\pi \nu} \int_{-\infty}^{t_m} dt''\,f_t(t'')\int_{-1}^1\frac{d\eta}{\sqrt{1-\eta^2}}.

396: \end{equation}

397: The integration over $\eta$ is immediate, so that substituting

398: Eq. (\ref{psivseta}  in Eq.(\ref{phivspsi}) we obtain finally

399: \begin{equation}

400: \phi(t,x,z)=f_t(t-\hat n_t\cdot\vec\rho/c).

401: \end{equation}

402: where $\hat n_t=(\mu,\nu)$.  Thus, as illustrated in

403: Fig. \ref{normal}, the transmitted wave is a pulse with the same

404: profile as the incident field and propagating with speed $c$ at the

405: well defined angle $\theta_t=\sin^{-1}\mu=\cos^{-1}\nu$, in accordance

406: with Snell's law as could have been expected.

407: 

408: \section{Evanescent wave transmission}\label{secevan}

409: 

410: We consider now the case $\theta_i>\theta_c$, for which

411: $v_\parallel<c$ and the transmitted wave becomes evanescent. In this

412: case we substitute 

413: Eq. (\ref{phivsf}) into Eq. (\ref{psi}) to obtain

414: \begin{equation}\label{psievan}

415: \psi(t,x,z)=\frac{c}{\pi}\int dt'' f_t(t-x/v_\parallel+t'')\int dx''

416: \left(c^2(t''+x''/v_\parallel)^2 - (x'')^2-z^2\right)^{-1/2},

417: \end{equation}

418: after introducing the variables $x''\equiv x'-x$ and

419: $t''=t'-t-x''/v_\parallel$. As shown in Fig. \ref{cono2}, for any

420: observation event $\cal E$ and any value of $t''$, there is exactly

421: one intersection $x''_-$ between the  past light cone of $\cal E$ and

422: the world line $(ct',x''=v_\parallel(t'-t''-t),0)$, where now

423: $x''_-=-\gamma \beta [\gamma c t'' + \sqrt{z^2+(\gamma\beta c

424: t'')^2}]$ and we introduced the definitions $\beta\equiv

425: v_\parallel/c$ and $\gamma\equiv1/\sqrt{1-v_\parallel^2/c^2}$. We have

426: assumed that $v_\parallel>0$. 

427: As the world

428: line leaves the past light cone at $x''_-$, in

429: Eq. (\ref{psievan}) the integration over $t''$ 

430: is unconstrained and that over $x''$ extends from $-\infty$ to $x''_-$.

431: \begin{figure}

432: \includegraphics{fig3}%{conos2.0}

433: \caption{\label{cono2}Contributions to the evanescent wave observed at

434:   an event ${\cal E}=(ct,x,z)$ as in Fig. \ref{cono1}, but for

435:   $v_\parallel<c$. The position $x''_-$ denotes the intersections of

436:   the past light cone of $\cal E$ with one of the world lines

437:   $x''=v_\parallel(t'-t''-t)$ corresponding to $t''$.}

438: \end{figure}

439: Another change of variable, from $x''$ to 

440: \begin{equation}\label{etaagain}

441: \eta=\frac{1}{\gamma\beta}\frac{x''+\gamma^2\beta c

442: t''}{\sqrt{z^2+(\gamma\beta c t'')^2}}, 

443: \end{equation}

444: yields

445: \begin{equation}\label{psianew}

446: \psi(t,x,z)=\frac{\gamma\beta c}{\pi}\int_{-\infty}^\infty dt''\,

447: f_t(t-x/v_\parallel+t'') \int_{-\infty}^{-1}

448: \frac{d\eta}{\sqrt{\eta^2-1}}.

449: \end{equation}

450: 

451: The integrals over $x''$ and over $\eta$ in 

452: Eqs. (\ref{psievan}) and (\ref{psianew}) respectively yield an

453: infinite value for 

454: $\psi$. This divergence is not unlike that commonly found for the

455: electromagnetic potentials produced by infinitely  extended

456: sources. For example, when calculating the electric field 

457: produced by a uniformly charged plane one cannot simply obtain the

458: corresponding potential by integrating the Coulomb kernel over the whole

459: surface. However, in that case the electric field may be obtained

460: either by deriving the Coulomb kernel first and integrating afterwards

461: or else, by

462: truncating the integrations at a finite distance, deriving the resulting

463: potential to obtain the field and 

464: afterwards taking the limit of an infinite  

465: surface.

466: %\cite{inberkeley}. 

467: Here we follow the later procedure.

468: Thus, we set a finite lower integration limit

469: $x''_L$ in Eq. (\ref{psievan}), corresponding to a lowest point

470: $\eta_L$  in Eq. (\ref{psianew}), and we take the limit

471: $x''_L\to-\infty$, $\eta_L\to-\infty$ after obtaining the field $\phi$.

472: 

473: As $\psi$ depends on $z$ only through $\eta_L$, substituting

474: Eq. (\ref{psianew}) in (\ref{phivspsi}) we obtain

475: \begin{equation}\label{phivsetaL}

476: \phi(t,x,z)=\frac{\gamma\beta c}{\pi}\int_{-\infty}^\infty dt''\,

477: f_t(t-x/v_\parallel+t'')

478: \frac{1}{\sqrt{\eta_L^2-1}}\frac{\partial\eta_L}{\partial z}. 

479: \end{equation}

480: In the limit $x''_L\to-\infty$ we evaluate 

481: \begin{equation}\label{detadz}

482: \xi_L \equiv \frac{1}{\sqrt{\eta_L^2-1}}\frac{\partial\eta_L}{\partial

483: z}\to  \frac{z}{z^2+(\gamma\beta c t'')^2},

484: \end{equation}

485: and we obtain finally

486: \begin{equation}\label{phievan}

487: \phi(t,x,z)=\frac{1}{\pi}\int_{-\infty}^\infty dt''\,

488: f_t(t-x/v_\parallel+t'') \frac{\gamma |v_\parallel| z}{z^2+(\gamma

489:   v_\parallel t'')^2}.

490: \end{equation}

491: Taking the absolute value of $v_\parallel$ in the numerator of

492: Eq. (\ref{phievan}) allows its use also for $v_\parallel<0$. 

493: 

494: To grasp the meaning of Eq. (\ref{phievan}) we evaluate it for an

495: infinitely sharp pulse

496: \begin{equation}\label{sharp}

497: f_t(\tau)\equiv f_0 \delta(\tau).

498: \end{equation}

499: Substitution into Eq. (\ref{phievan}) yields

500: \begin{equation}\label{lorentz}

501: \phi(t,x,z)=\frac{f_0}{\pi}  \frac{\gamma |v_\parallel| z}{z^2+

502: \gamma^2(x-v_\parallel t)^2}.

503: \end{equation}

504: Surprisingly,  at any time $t$ the pulse transmitted at a distance $z$

505: from the 

506: interface is given by a Lorentzian of width

507: $z/\gamma$ centered in front of the actual position

508: $x=v_\parallel t$ of the incident pulse on the $z=0$ surface

509: (Fig. \ref{figlorentz}).

510: \begin{figure}

511: \includegraphics{fig4}%{pr.7}

512: \caption{\label{figlorentz}Infinitely sharp wavefront (heavy solid

513: line) incident at an angle $\theta=34.5^\circ>\theta_c$ upon the

514: surface of a dielectric with index of refraction $n=2$. The pulse

515: widens and diminishes as it is transmitted across an air gap and into a

516: second dielectric. The thin lines indicate the nominal pulse width and

517: the dashed line its center. 

518: }

519: \end{figure}

520: Thus, the propagation seems to be instantaneous in the direction

521: normal to the surface, and actually, part of the pulse seems to travel

522: backwards in time \cite{carey00,resch}, as at a position $(x,z)$ it

523: becomes appreciable at times $t<x/v_\parallel$, that is, before the

524: incoming pulse reaches the corresponding position $(x,0)$. However,

525: our deduction of Eq. (\ref{sharp}) shows that it is completely

526: consistent with a causal and retarded propagation, and that the field

527: at a $(x,z)$ at time $t$ is not produced instantaneously by the

528: incoming field at $(x,0)$, but arises from previously excited

529: positions $(x',0)$ with $x'<x+x''_-$.

530: 

531: It is interesting to note that, according to Eq. (\ref{lorentz}), the

532: height of the transmitted pulse is inversely proportional to

533: the distance $z$ from the surface, instead of decaying exponentially

534: as usually found for evanescent waves. However, Eq. (\ref{lorentz})

535: describes the propagation into vacuum of a single infinitely sharp

536: incident wavefront. In a wavetrain made up of a succession of incident

537: pulses, the regions excited by neighboring pulses overlap each other,

538: as the width of each transmitted pulse increases in

539: proportion to $z$, and therefore their corresponding fields

540: interfere. This interference is at 

541: the origin of the 

542: exponential decay of periodic waves, as can be verified by choosing

543: \begin{equation}\label{periodico}

544:   f_t(\tau)=A e^{-i\omega \tau}

545: \end{equation}

546: and substituting into

547: Eq. (\ref{phievan}), which yields

548: \begin{equation}\label{periodic}

549: \phi(t,x,z)=A e^{i(Qx-\omega t)}\int_{-\infty}^\infty \frac{dt''}{\pi}

550: e^{-i\omega t''} \frac{\gamma|v_\parallel|z}{z^2+(\gamma v_\parallel

551: t'')^2}. 

552: \end{equation}

553: A simple contour integration closing the integration path with an

554: infinite semicircle on the lower half complex $t''$ plane yields the

555: familiar result

556: \begin{equation}\label{exponential}

557: \phi(t,x,z)=A e^{i(Qx-\omega t)-\kappa z},

558: \end{equation}

559: where $Q=\omega/v_\parallel=n\sin\theta_i\omega/c$ is the parallel

560: component of the wave vector and

561: $\kappa=\omega/(\gamma|v_\parallel|)=\sqrt{Q^2-\omega^2/c^2}=1/l$ the

562: inverse of the decay length $l$. As $\omega$ increases, the distance

563: along $x$ between successive maxima and minima decreases, yielding

564: larger interference effects and a shorter decay length.

565: 

566: 

567: \section{Screens}\label{secscreens}

568: 

569: The results of the previous section suggest an experiment

570: that could confirm that propagation of evanescent waves in the FTIR

571: geometry is not superluminal nor acausal. The experiment could be

572: performed simply by

573: partially covering the surface of the first interface with a couple of

574: opaque screens as shown in Fig. \ref{screens}.

575: \begin{figure}

576: \includegraphics{fig5}%{pr.2}

577: \caption{\label{screens}Pulse as in Fig. \ref{normal} but impinging on

578:   the surface of the first dielectric at an angle

579:   $\theta_i>\theta_c$. The surface is covered by semi-infinite opaque

580:   screens $S_a$ and $S_b$ with edges at $x_a$ and $x_b$. The leading

581:   wavefront of the incoming pulse is about to reach $x_a$. $\cal F$

582:   denotes the foremost wavefront.}

583: \end{figure}

584: If transmission were indeed superluminal, we would expect a non-null

585: transmitted field $\phi(t,x_a,d)$ across the gap in front of the edge

586: $x_a$ of the first screen $S_a$ as soon as the leading wavefront $\cal

587: F$ of the incident pulse reaches $x_a$. Similarly, we would expect

588: that the field $\phi(t,x_b,d)$ would be modified as soon as $\cal F$

589: reaches the edge $x_b$ of the second screen $S_b$. 

590: 

591: To calculate the field corresponding to Fig. \ref{screens} we go back

592: to Eq. (\ref{psievan}). The screens confine the integration region to

593: the interval $x_a<x'=x+x''<x_b$. This inequality has to be obeyed

594: together with the previous constriction $x'<x+x''_-$. These conditions

595: can only be satisfied by those wavefronts which have reached $x_a$ and

596: left behind the first screen before $t-T_a$, where

597: $T_i=(1/c)\sqrt{(x-x_i)^2+z^2}$, $i=a,b$, is the minimum time required

598: to reach $(x,z)$ from $(x_i,0)$ moving at speed $c$. Thus, only those

599: points on the wavefront labeled by $t''>T''_a$ can contribute to

600: (\ref{psievan}), where

601: \begin{equation}\label{T''}

602: T''_i\equiv

603: \frac{x-x_i}{v_\parallel}-\frac{1}{c}\sqrt{(x_i-x')^2+z^2},\quad i=a,b.

604: \end{equation}

605: For those wavefronts which have left $S_a$ by time $T_a$

606: but have not reached $S_b$ at time $T_b$, namely, those with

607: $T''_b<t''<T''_a$, the integral over $x''$ in Eq. (\ref{psievan}) has

608: to be performed from $x''_a=x+x_a$ up to $x''_-$. Finally, for those

609: wavefronts which have already been blocked by $S_b$ by time $T_b$,

610: namely, those with $t''<T''_b$, the upper limit of integration has to

611: be replaced by $x''_b=x+x_b$. Therefore,

612: \begin{equation}\label{psiscreens}

613: \psi(t,x,z)=\frac{\gamma\beta c}{\pi}\int_{T''_b}^{T''_a} dt''\,

614: f_t(t-x/v_\parallel+t'') \int_{\eta_a}^{-1}

615: \frac{d\eta}{\sqrt{\eta^2-1}} + \frac{\gamma\beta

616:   c}{\pi}\int_{-\infty}^{T''_b} dt''\, f_t(t-x/v_\parallel+t'')

617: \int_{\eta_a}^{\eta_b} \frac{d\eta}{\sqrt{\eta^2-1}}

618: \end{equation}

619: where we used the change of variables (\ref{etaagain}) and substituted

620: $x''\to x+x_i$ in it to define the limits $\eta_i$. Notice that $\psi$

621: depends on $z$ 

622: only through the integration limits $\eta_i$, so that substituting

623: Eq. (\ref{psiscreens}) in (\ref{phivspsi}) we obtain

624: \begin{equation}\label{psivsxi}

625: \psi(t,x,z)=\frac{\gamma\beta c}{\pi}\int_{-\infty}^\infty dt''\,

626: f_t(t-x/v_\parallel+t'') [\xi_a\Theta(T''_a-t'') - \xi_b

627:   \Theta(T''_b-t'')],

628: \end{equation}

629: where

630: \begin{equation}\label{xi}

631: \xi_i\equiv

632: \frac{1}{\sqrt{\eta_i^2-1}}\frac{\partial\eta_i}{\partial

633: z}\equiv \xi_L \zeta_i,

634:  \end{equation}

635: $\xi_L$ is given by Eq. (\ref{detadz})and $\zeta_i\equiv\zeta(x''_i,t'')$ with

636: \begin{equation}\label{zeta}

637: \zeta(x'',t'')=-\,\frac{x''+\gamma^2\beta c

638: t''}{\sqrt{(x'')^2+2\gamma^2 

639: \beta c x'' t'' - \gamma^2\beta^2 (z^2-c^2 (t'')^2)}}.

640: \end{equation}

641: Substituting Eqs. (\ref{etaagain}) and (\ref{xi}) in (\ref{psivsxi})

642: we finally obtain

643: \begin{equation}\label{phiscreened}

644: \phi(t,x,z)=\frac{1}{\pi}\int_{-\infty}^\infty dt''\, 

645: f_t(t-x/v_\parallel+t'') \frac{\gamma v_\parallel z}{z^2+(\gamma

646: v_\parallel t'')^2} (1-C(t'')),

647: \end{equation}

648: where 

649: \begin{equation}\label{C}

650: C(t'')=1+[\zeta_b\Theta(T''_b-t'') - \zeta_a \Theta(T''_a-t'')].

651: \end{equation}

652: Notice that the field $\phi$ in the presence of screens

653: (Eq. (\ref{phiscreened})) is given by an expression similar to that

654: corresponding to the field in the absence of screens

655: (Eq. (\ref{phievan})) but with a correction term $C$ due to the

656: diffraction by the screen.

657: 

658: As in the previous section, we consider again the case of a sharp

659: incident pulse  Eq. (\ref{sharp}). Substituting in (\ref{phiscreened})

660: we obtain

661: \begin{equation}\label{lorentzscreened}

662: \phi(t,x,z)=\frac{\gamma v_\parallel z}{z^2+\gamma^2(x-v_\parallel t)^2}

663: \times \left\{

664: \begin{array}{ll}

665: 0&\mbox{if $ct<x_a/\beta+\sqrt{(x-x_a)^2+z^2}$},\\

666: \zeta(x_a-x,x/v_\parallel-t)-\zeta(x_b-x,x/v_\parallel-t)&\mbox{if

667: $ct>x_b/\beta+\sqrt{(x-x_b)^2+z^2}$},\\ 

668: \zeta(x_a-x,x/v_\parallel-t)&\mbox{otherwise}.

669: \end{array}

670: \right.

671: \end{equation}

672: We remark that the field is zero until the time $x_a/v_\parallel$ when

673: the incident pulse shows up from behind the screen $S_a$, and this

674: information has had enough time $(1/c)\sqrt{(x_a-x)^2+z^2}$ to

675: propagate from the screen's edge $(x_a,0)$ to the observation point

676: $(x,z)$. Similarly, the information that the pulse has hidden behind

677: screen $S_b$ does not reach the observation point until the time

678: $x_b/v_\parallel+(1/c)\sqrt{(x-x_b)^2+z^2}$. The field has

679: singularities due to the passage of the incident pulse through the

680: screen edges, that propagate at speed $c$ from the events

681: $(x_a/\beta,x_a,0)$ and $(x_b/\beta,x_b,0)$. Notice that

682: $\zeta(x_i-x,0)\to 1$ as $x\to\infty$. Thus, if we follow the incident

683: pulse, i.e., we take $x\approx v_\parallel t$, then $\phi\to 0$

684: asymptotically after the pulse hides behind $S_b$. Furthermore, if the

685: screens are very far apart we recover the field (\ref{lorentz})

686: between the screens.

687: 

688: The features above are illustrated in Fig. \ref{pantalla}

689: \begin{figure}

690: \includegraphics{fig6}%{rejilla.0}

691: \caption{\label{pantalla}Field isolines of $\phi(t,x,d)$ produced by

692:   an unit delta function input pulse that propagates along the $z=0$

693:   surface with speed $v_\parallel=c/\sqrt{2}$ and is blocked at

694:   $x<x_a=-3d$ and $x>x_b=3d$ by opaque screens (shaded regions). The

695:   world line $x=v_\parallel t$ of the incident pulse is indicated by a

696:   dashed line. The times $t_a$ and $t_b$ when the incident pulse cross

697:   the edges of each screen are indicated by the horizontal dot-dashed

698:   lines. The singularities of the transmitted field are indicated by

699:   the thick solid hyperbolas.}

700: \end{figure}

701: which show the transmitted field at the plane $z=d$. Notice the delay

702: $d/c$ after the incident pulse crosses $x_a$ at $t_a=x_a/v_\parallel$

703: before a non-null field first appears across the gap at $(x_a,d)$, and

704: a similar delay after the incident pulse crosses $x_b$ at

705: $t_b=x_b/v_\parallel$ before the field starts to be extinguished at

706: $(x_b,d)$. Furthermore, notice that for some time the field penetrates

707: a small distance $\approx d$ beyond $x_b$ as if there were no

708: screen. The field is singular at the hyperbolas with vertices at

709: $x=x_a$, $t=t_a+d/c$ and at $x=x_b$, $t=t_b+d/c$ given by the

710: intersection of the $(ct,x,d)$ hyperplane and the future light cone of

711: the events $(c t_a,x_a,0)$, $(c t_b,x_b,0)$. Thus, we have shown

712: observable consequences of the fact that evanescent waves in FTIR do

713: not propagate superluminally nor acausally in the direction normal to

714: the dielectric-vacuum interfaces, but with retardation and

715: obliquely. A graphical approach to the results of this section and an

716: animation illustrating them may be found in

717: Refs. \onlinecite{mochan01} and Ref. \onlinecite{oe} respectively.

718: 

719: 

720: \section{A smooth transverse profile}\label{secsmooth}

721: 

722: Perfectly opaque screens such as those considered in the previous

723: section introduce sharp discontinuities in the pulse at $z=0$. The

724: truncated pulse no longer has a well defined propagation direction

725: $\theta_i$ but may still be represented by a superposition of pulses

726: with varying propagation directions. A sharp truncation leads to the

727: presence of subcritical incident angles $\theta_i<\theta_c$, and

728: therefore to the presence of both, evanescent and non-evanescent

729: transmitted fields. It has been argued \cite{carey01} that the

730: retardation effects discussed in the previous section may be due only

731: to the non-evanescent contributions,  known to be

732: subluminal.  The comparatively slow subluminal contributions would be

733: unable to affect the arrival of the superluminal signals if the later

734: were actually present. However, any small non-evanescent wave would

735: dominate the transmitted signal after a wide enough gap. Thus, it is

736: interesting to study the propagation of pulses with a finite

737: transverse extension but with a smooth lateral cutoff and built up

738: completely from hypercritical $\theta_i>\theta_c$ evanescent

739: contributions.

740: 

741: To explore the propagation of the smoothly truncated pulses discussed

742: above, we consider an incoming field given by a Fourier integral

743: \begin{equation}\label{fourier}

744: \phi(t,x,0^+)=\int\frac{d\omega}{2\pi}\int\frac{d Q}{2\pi}\, f_{\omega Q}

745: e^{i(Q x - \omega t)},

746: \end{equation}

747: where $f_{\omega Q}$ is the amplitude for each parallel component of

748: the wave vector $Q$ and frequency $\omega$. We can change integration

749: variable from $Q$ to the parallel velocity $v\equiv \omega/Q$,

750: \begin{equation}\label{vfourier}

751: \phi(t,x,0^+)=\int\frac{d\omega}{2\pi}\int\frac{d

752: v}{2\pi}\, f_{\omega v} 

753: e^{-i\omega (t-x/v)}.

754: \end{equation}

755: where we introduced the velocity dependent amplitude $f_{\omega v}

756: \equiv (\omega/v^2) f_{\omega,\omega/v}$.  The incident field

757: (\ref{vfourier}) will give rise to evanescent waves exclusively as

758: long as all non-null components $f_{\omega v}$ have $v<c$.  At this

759: point we could integrate first Eq. (\ref{vfourier}) with respect to

760: $\omega$, obtaining thus a superposition of plane pulses, each of

761: which may be propagated across the air gap according to

762: Eq. (\ref{phievan}). Alternatively, we may propagate each

763: monochromatic component using Eq. (\ref{exponential}) and afterward

764: perform the integrations in Eq. (\ref{vfourier}). We follow the later

765: approach and write

766: \begin{equation}\label{expprop}

767: \phi(t,x,z)=\int\frac{d\omega}{2\pi}\int\frac{d v}{2\pi}\, f_{\omega

768: v} e^{-\omega[z/(\gamma v)+i(t-x/v)]}, 

769: \end{equation}

770: where $\gamma=1/\sqrt{1-v^2/c^2}$ as in Sec. \ref{secevan}. For

771: simplicity we assume that we may factor $f_{\omega v}=f_\omega f_v$

772: into frequency and velocity dependent amplitudes, $f_\omega$ and

773: $f_v$ respectively; $f_\omega$ controls the time duration of the

774: pulse, or equivalently, its longitudinal extent, while $f_v$ controls

775: its transverse extent. We further assume a narrow Gaussian velocity

776: distribution of width $\Delta v$ around a nominal velocity $v_0<c$, 

777: \begin{equation}\label{gaussian}

778: f_v= \frac{\sqrt{2\pi}}{\Delta v} e^{-u^2/2\Delta v},

779: \end{equation}

780: where $u\equiv v-v_0$. We assume $\Delta v$ is small enough that the

781: non-evanescent contributions to the field may be neglected and the

782: exponent in Eq. (\ref{expprop}) may be linearized in $u$. Thus, the

783: transmitted field becomes 

784: \begin{equation}\label{transm}

785: \phi(t,x,z)\approx \int\frac{d\omega}{2\pi}\, f_\omega

786: e^{-\omega[z/(\gamma_0 v_0)+i(t-x/v_0)]} \int\frac{d

787: u}{\sqrt{2\pi}\Delta v}\,  e^{-u^2/2\Delta v^2 +\omega u[z/\gamma_0

788: -i x]/v_0^2}, 

789: \end{equation}

790: where $\gamma_0=1/\sqrt{1-v_0^2/c^2}$. The integration over $u$

791: is immediate and yields

792: \begin{equation}\label{transm1}

793: \phi(t,x,z)= \int\frac{d\omega}{2\pi}\, f_\omega

794: e^{-\omega[z/(\gamma_0 v_0)+i(t-x/v_0)] - [\Delta v \omega(x+i

795: z/\gamma_0)]^2/2 v_0^4}.

796: \end{equation}

797: 

798: In Fig. \ref{figgauss} we illustrate the results of applying

799: Eq. (\ref{transm1}) to a pulse with a Gaussian frequency

800: distribution,

801: \begin{equation}

802: f_\omega = A \frac{\sqrt{2\pi}}{\Delta\omega}

803: e^{-(\omega-\omega_0)^2/2\Delta\omega^2},

804: \end{equation}

805: of area $A$ and width $\Delta\omega$ centered at $\omega_0$

806: ($\omega_0=16c/d$, $\Delta\omega=2c/d$, $v_0=0.7c$, $\Delta v=0.15c$).

807: \begin{figure}

808: %\includegraphics{pr.6} \parbox[t]{4in}{%

809: %\makebox[0in]{\includegraphics[width=2in]{gauss0.eps}%

810: %\includegraphics[width=2in]{gauss1.eps}}%

811: %\makebox[0in]{\includegraphics{pr.4}}\\

812: %\makebox[0in]{\includegraphics{pr.5}}%

813: %}

814: \includegraphics{fig7.eps}

815: \caption{\label{figgauss}Intensity of a Gaussian pulse localized in

816: both space and time ($x$ and $t$) incident on the front face of an air

817: gap of width $d$ (left) and after after crossing it (right). The

818: nominal velocity along $x$ is $v_0=0.7c$ with width $\Delta

819: v=0.15c$. The nominal frequency is $\omega_0=16 c/d$ with width $\Delta

820: \omega =2 c/d$}

821: \end{figure}

822: The pulse is seen to form on the $z=0$ surface at $x\approx -5d$ and

823: at time $t\approx7d/c$, it propagates along the $x\approx v_0 t$ line

824: for a while, peaks at $x=0$ at time $t=0$ and disappears at $x\approx

825: 5d$, $t\approx7d/c$. Its maximum duration $\tau$ and size $L$ are

826: $c\tau\approx L\approx d$ for a fixed position and fixed observation

827: time respectively, and it contains altogether about six nodes.

828: 

829: Surprisingly, after crossing the gap, the pulse looks essentially the

830: same! It appears at the back  $z=d$ face of the air gap at roughly the

831: same time and the same position as at the front $z=0$ face. It also

832: peaks at the origin at $t=0$ and disappears from the back surface in

833: concordance to the incident pulse on the front face. Thus, it truly

834: appears to propagate instantaneously. The main difference between the

835: incident and transmitted pulse is that the intensity of the later is

836: suppressed by 12 orders of magnitude. Another interesting difference is

837: that the number of visible nodes in the transmitted pulse has

838: decreased to about 4. This is a consequence of the fact that in FTIR,

839: the Fourier components with higher frequencies are more damped than

840: those with lower frequencies. Finally, a more subtle difference is

841: that the speed of propagation along the back face is slightly but

842: noticeably larger than that on the front face. This is due to the fact

843: that plane waves incident at angles closer to $\theta_c$ have a larger

844: penetration length than waves incident at larger angles. Thus, the

845: angle of propagation of the transmitted pulse is smaller than that of

846: the 

847: incident wave \cite{balcou97}.

848: 

849: %\section{Superluminal behavior of pulses}\label{secbehavior}

850: 

851: In Sec. \ref{secevan} we have argued that the transmission of

852: evanescent plane pulses across an air gap under FTIR conditions is

853: fully consistent with a retarded and causal propagation along oblique

854: directions. We have strengthened our argument by showing that there is

855: a

856: delay before a perturbation, such as blocking part of the incident

857: wavefront, can produce an effect on the pulse transmitted across the

858: gap. Furthermore, by truncating an incident plane pulse 

859: producing an abrupt transverse profile, we showed that the transmitted

860: pulse is shifted along the surface in the direction of propagation. 

861: However, in Sec. \ref{secsmooth} we showed through an example that if

862: the pulse has a smooth transverse profile, such that all its

863: Fourier components are evanescent, it is transmitted as if it were

864: indeed 

865: superluminal. To understand this result, in Fig. \ref{misaligned} we

866: show schematically a pulse  smoothly truncated along its transverse

867: direction, built from narrow plane components propagating along well

868: defined directions $\theta>\theta_c$. 

869: \begin{figure}

870: \includegraphics{fig8}%{pr.3}

871: \caption{\label{misaligned}Evanescent transmission of a smoothly

872: truncated pulse made up initially of a

873: superposition of narrow plane wavefronts (heavy solid lines) with a

874: distribution of angles above 

875: $\theta_c$. As it crosses the gap, each wavefront widens (shaded bands

876: around dashed lines, as in Fig. \ref{figlorentz}). The ellipses

877: represent 

878: schematically the 

879: contour levels of 

880: the pulse and are centered at the region where the different

881: contributions add coherently in phase, i.e., the regions where the

882: centers of each component coincide. Snapshots are taken at three

883: different times. The solid arrows illustrate the nominal propagation

884: of the peak. The heavy arrow indicates the apparent instantaneous

885: tunneling across the air gap. 

886: The dashed arrows illustrate the actual (sub)luminal propagation.} 

887: \end{figure}

888: The incident field peaks at the regions

889: where the directional components add in phase, indicated in the figure

890: by elliptical regions around the crossing point of the different

891: incident wavefronts. As each of the components crosses the gap,

892: it is widened according to Eq. (\ref{lorentz}). The peak of the

893: transmitted pulse appears in the regions of largest overlap between

894: the different widened transmitted components, i.e., at the regions

895: where their centers coincide. The peak of the transmitted pulse

896: at the back face of the air gap is seen to appear at the same time as

897: the peak of 

898: the incident pulse reaches the front surface. Thus, it would seem as

899: if the peak 

900: tunneled instantaneously in the direction normal to the gap. However,

901: in the previous sections we have shown that each of the

902: transmitted components of the transmitted wave originates causally

903: from regions in the lateral wings of the incident pulse. Thus, the

904: peak of the transmitted field does not actually originate from the

905: peak of the incident field; it is formed by

906: contributions from the lateral wings of the incident field, which

907: reach the front face of the air gap first. The lateral wings of each

908: component have enough time to cross the gap traveling at speed

909: $c$ and combine to form the relatively small transmitted peak right at

910: the time when 

911: the larger incident peak reaches the front surface. Similarly, the

912: different 

913: components of the field produced by the peak of the incident pulse get

914: out of step as they cross the air gap and, therefore, do not

915: contribute to the peak of the transmitted pulse, but rather, to its

916: lateral wing. This is illustrated by the dashed arrows in

917: Fig. \ref{misaligned}. Thus, it seems that the physical

918: propagation (retarded and  causal) can not be

919: distinguished from the nonphysical propagation (superluminal and

920: non-causal) as long as we only consider smooth incident pulses which

921: contain only propagation directions above the critical angle

922: \cite{carey00,carey01}. 

923: 

924: \section{Evanescent propagator}\label{secevprop}

925: 

926: In Sec. \ref{secsmooth} we found that a particular pulse made up of only

927: evanescent components seemed to propagate instantaneously across the

928: air gap, in contrast to the abruptly truncated pulses considered in

929: Sec. \ref{secscreens} for which retardation effects have observable

930: consequences. This was explained graphically in  Fig. \ref{misaligned}

931: %Sec. \ref{secbehavior} 

932: for incident pulses built up from narrow

933: directional components, each of which is widened as it is transmitted

934: across the gap. To show that this behavior is generic, we start from

935: the Fourier decomposition of an arbitrary field

936: \begin{equation}\label{Fourier}

937: \phi(t,x,z)=\int\frac{d Q}{2\pi} \int\frac{d\omega}{2\pi}

938: e^{i(Q x - \omega t)} \phi_{\omega,Q}(z), 

939: \end{equation}

940: where

941: \begin{equation}\label{Fourierinv}

942: \phi_{\omega,Q}(z)=\int d t\int d x\, e^{-i(Q x - \omega t)}

943: \phi(t,x,z). 

944: \end{equation}

945: The condition that $\phi(t,x,z)$ is made up exclusively of evanescent

946: waves is equivalent to stating that the integration region in

947: Eq. (\ref{fourier}) is given by $|\omega|<|Q|c$, i.e.,

948: $\phi_{\omega,Q}=0$ 

949: if $|\omega|>|Q|c$. Using Eq. (\ref{exponential}) we propagate  each

950: Fourier component from $z=0^+$ to $z>0$ as 

951: \begin{equation}\label{phiwQz}

952: \phi_{\omega, Q}(z)=e^{-\kappa z} \phi_{\omega, Q}(0^+),

953: \end{equation}

954: where $\kappa=\sqrt{Q^2-\omega^2/c^2}$. Thus, we can combine

955: Eqs. (\ref{Fourier}), (\ref{Fourierinv}) and (\ref{phiwQz}) to obtain

956: \begin{equation}\label{propagate1}

957: \phi(t,x,z)=\int dx' \int dt'\,P'(t,x,z;t',x',0^+) \phi(t',x',0^+),

958: \end{equation}

959: where

960: \begin{equation}\label{propagator1}

961: P'(t,x,z;t',x',0^+)\equiv P'(t-t',x-x',z)=\int\frac{dQ}{2\pi}

962: \int\frac{d\omega}{2\pi} e^{i[Q(x-x')-\omega(t-t')]-\kappa

963:   z} 

964: \end{equation}

965: and the integration region is given by $|\omega|<|Q|c$.

966: Comparing Eq. (\ref{propagate1}) with (\ref{propagate}) we find that

967: $P'$ is a propagator that can be used in the same way as the

968: propagator $P$ defined in Eq. (\ref{P})  

969: to find the value of

970: the field at 

971: $z>0$ given its values at $z=0^+$, provided the field is built up of evanescent

972: components only, as in the examples of the two previous sections. We

973: remark that our original causal, retarded and subluminal propagator

974: $P$ was able to propagate any arbitrary outgoing field. However,

975: by adding constrains to the field, we gain freedom in our choice of

976: propagator, as we can chose arbitrarily its effect on fields that do

977: not obey the constrain. Thus, if we can find any function $P''$ such that

978: \begin{equation}\label{constrain}

979: \int dx'\int dt'\,[P''(t,x,z;t',x',0^+)-P(t,x,z;t',x', 0^+)] f(t'-x'/v) =0,

980: \end{equation}

981: for an arbitrary flat pulse $f$ moving along $x$ with any velocity

982: $-c<v<c$, then we could employ $P''$ instead of $P$ to propagate an

983: arbitrary evanescent pulse. $P'$ above is just one of the

984: many possible choices of a propagator for evanescent pulses.

985: 

986: To proceed, we make a change of variable $\omega\to Qv$ to write

987: Eq. (\ref{propagator1}) as

988: \begin{equation}\label{prop2}

989: %P'(t,x,z)=\int\frac{dQ}{2\pi}\int_{-c}^c\frac{dv}{2\pi}\, |Q| e^{iQ(x-vt)}

990: %e^{-Qz/\gamma},

991: %P'(t,x,z;t',x')=\int\frac{dQ}{2\pi}\int_{-c}^c\frac{dv}{2\pi}\, |Q|

992: %e^{iQ[(x-x')-v(t-t')]} 

993: %e^{-Qz/\gamma},

994: P'(\tau,\xi,z)=\int\frac{dQ}{2\pi}\int_{-c}^c\frac{dv}{2\pi}\, |Q| e^{iQ(\xi-v\tau)}

995: e^{-Qz/\gamma},

996: \end{equation}

997: where $\gamma=1/\sqrt{1-v^2/c^2}$, $\xi=x-x'$, $\tau=t-t'$ and we

998: perform the integration over 

999: $Q$,

1000: \begin{equation}\label{prop3}

1001: %P'(t,x,z)=\frac{1}{2\pi^2} \int_{-c}^c dv\,\frac{(z/\gamma)^2-(x-vt)^2}

1002: %{[(z/\gamma)^2+(x-vt)^2]^2}.

1003: P'(\tau,\xi,z)=\frac{1}{2\pi^2} \int_{-c}^c

1004: dv\,\frac{(z/\gamma)^2(\xi-v\tau)^2} 

1005: {[(z/\gamma)^2+(\xi-v\tau)^2]^2}.

1006: \end{equation}

1007: Notice that $P'(t,x,z;t',x',0^+)$ is symmetric under the interchange

1008: $\xi\leftrightarrow -\xi$ and also under the interchange 

1009: $\tau\leftrightarrow -\tau$, i.e., $P'(t,x,z;t',x') = P'(t,x',z;t',x) =

1010: P'(t',x,z;t,x') = P'(t',x',z;t,x)$. Thus, the evanescent propagator

1011: $P'$ {\em is superluminal and acausal}.

1012: 

1013: To finish the calculation of $P'$ we make another change of

1014: integration variable $v=c \sin\alpha$ to write

1015: \begin{equation}\label{prop4}

1016: %P'(t,x)=\frac{c}{2\pi^2}\int_{-\pi/2}^{\pi/2} d\alpha \cos\alpha

1017: %\frac{(z\cos\alpha)^2-(x-c t \sin\alpha)^2}

1018: %{[(z\cos\alpha)^2+(x-c t \sin\alpha)^2]^2},

1019: P'(\tau,\xi)=\frac{c}{2\pi^2}\int_{-\pi/2}^{\pi/2} d\alpha \cos\alpha

1020: \frac{(z\cos\alpha)^2-(\xi-c \tau \sin\alpha)^2}

1021: {[(z\cos\alpha)^2+(\xi-c \tau \sin\alpha)^2]^2},

1022: \end{equation}

1023: and we perform the integration

1024: \begin{equation}\label{evprop}

1025: %P'(t,x)=\frac{c}{4\pi^2}\frac{1}{(\rho^2-c^2 t^2)^{3/2}} \left[ z \log

1026: %\left( \frac{(c^2 t^2-z^2-z\sqrt{\rho^2-c^2 t^2})^2-c^2 t^2 x^2} {(c^2

1027: %t^2-z^2+z\sqrt{\rho^2-c^2 t^2})^2-c^2 t^2 x^2}\right)

1028: %-4\sqrt{\rho^2-c^2 t^2}\right],

1029: P'(\tau,\xi)=\frac{c}{4\pi^2}\frac{1}{s^3} \left[ z \log

1030: \left( \frac{(c^2 \tau^2-z^2-z s)^2-c^2 \tau^2

1031:   \xi^2} {(c^2 

1032: \tau^2-z^2+z s)^2-c^2 \tau^2 \xi^2}\right)

1033: -4 s\right],

1034: \end{equation}

1035: where

1036: %$\rho=\sqrt{x^2+z^2}$ 

1037: $\rho=\sqrt{\xi^2+z^2}$ 

1038: is the spatial distance from the source to the

1039: observation point and $s^2 = \rho^2-c^2 \tau^2$ is the squared

1040: space-time interval.

1041: This expression may be simplified to

1042: \begin{equation}\label{evprop1}

1043: P'(\tau,\xi)= \frac{c}{2\pi^2}\times \left\{ 

1044: \begin{array}{ll}

1045:  z \log(|z+s|/|z-s|)/s^3 -2/s^2 &\mbox{if $s^2 > 0$},\\

1046: -2z\arctan(|s|/z)/|s|^3-2/|s|^2

1047: &\mbox{if $s^2 < 0$},\\

1048: \end{array}

1049: \right..

1050: \end{equation}

1051: 

1052: The evanescent propagator is displayed in Fig. \ref{figpropagador}. 

1053: \begin{figure}

1054: %\input{propagador.tex}

1055: \includegraphics{fig9}

1056: \caption{\label{figpropagador} Propagator $P'(t,x,d;t',x',0^+)$ normalized

1057: to $c/d^2$ as a

1058: function of the displacement $x-x'$ and the delay $t-t'$. Distance is

1059: measured in units of $d$ and time in units of $d/c$. For aid in

1060: visualization, the height of the propagator was truncated at $P'=0.1

1061: c/d^2$. }

1062: \end{figure}

1063: The figure shows explicitly the temporal and spatial symmetry, and

1064: thus the superluminality and non-causality of $P'$. Notice that $P'$

1065: has singularities at the projected light-lines $x-x'=\pm c(t-t')$ which

1066: converge at the origin $x=x'$, $t=t'$. Thus, propagation is largest

1067: for instantaneous propagation in the direction normal to the air

1068: gap. 

1069: %There are also singularities for $|x|=|ct|=|z|$. ????

1070: 

1071: We have found that the propagation of evanescent pulses can be

1072: described with either the exact propagator $P$ of the problem, which is

1073: causal, retarded and subluminal, or with an {\em evanescent}

1074: propagator $P'$ which is superluminal and non-causal. Both yield

1075: exactly the same transmitted pulse when the incident pulse contains

1076: only evanescent components. Thus, it seems to be impossible to

1077: distinguish superluminal from subluminal propagation in experiments

1078: performed with purely evanescent pulses \cite{carey01}.

1079: 

1080: An explanation for the curious conclusion found in this section can be

1081: obtained by going back to Eq. (\ref{Fourier}) which we rewrite as

1082: \begin{equation}

1083: \phi(t,x,z)=\int\frac{d Q}{2 \pi} \phi_Q(t,z).

1084: \end{equation}

1085: Notice that for each

1086: finite wave vector $Q$, the time dependent Fourier component

1087: $\phi_Q(t,z)$ has a strictly finite spectrum

1088: $-|Q|c<\omega<|Q|c$. Thus, $\phi_Q(t,z)$ is an analytical function of

1089: $t$ with no singularities and can be analytically

1090: continued to an arbitrary time $t_2$ from its values in an arbitrarily small

1091: neighborhood of any other arbitrary time $t_1$. Therefore,

1092: $\phi_Q(t,z)$ is perfectly 

1093: {\em predictable} in principle. Using antropomorphic language, we may

1094: say that at time $t_1$ the system {\em knows} from

1095: the present values of $\phi(t\approx t_1,x,0^+)$ what its future values

1096: $\phi(t\approx t_2 \ge t_1+z/c, x, 0^+)$ will be, and thus it 

1097: can {\em use} that knowledge to build subluminally a transmitted pulse

1098: $\phi(t\approx t_2,x,z)$ that will mimic $\phi(t\approx t_2, x, 0^+)$,

1099: giving the impression that superluminal transmission has taken

1100: place at time $t_2$. The validity of the argument above in the

1101: presence of thermal or quantum noise has to be investigated.

1102: 

1103: \section{Conclusions}\label{secconclusion}

1104: 

1105: To study the propagation of a light pulse through a vacuum gap between

1106: two parallel dielectrics in a FTIR configuration we constructed a propagator

1107: derived directly from the wave equation resulting from Maxwell's

1108: equations. This  

1109: propagator is retarded and complies with the relativistic causality

1110: principle inherent to classical electromagnetism. Therefore, it can

1111: only account for {\em causal, (sub)luminal}  

1112: propagation of light pulses. However, when this propagator is used to

1113: study the propagation of wave packets through the gap, we find

1114: apparent superluminal behavior, that is, a wave packet might appear on

1115: the far side of the gap at the same time that the incident packet

1116: reaches the front one.  Therefore this illusion of

1117: superluminality, present within classical electromagnetic

1118: theory even in vacuum, is fully consistent with relativistic

1119: causality. We showed explicitly that propagation  in FTIR actually takes place

1120: subluminally between the lateral wings of the incident pulse and the

1121: central peak  of the transmitted pulse, and we proposed simple

1122: experiments that could verify this statement. Thus, although FTIR has

1123: many similitudes to 1D tunneling, its correct physical interpretation

1124: requires a 2D or 3D analysis. On the other hand, we

1125: constructed an explicitly superluminal and acausal propagator that

1126: yields identical results as the retarded causal one when applied to

1127: smooth pulses made up of evanescent contributions only. Thus, there is

1128: a class of pulses for which superluminal and subluminal propagation

1129: would be indistinguishable. 

1130: 

1131: \begin{comment}

1132: An accurate definition of the information velocity may be key for

1133: solving the controversy in the interpretation of the superluminal

1134: behavior of pulses. Unfortunately, there is no universally agreed-upon

1135: definition of this concept.  However, we believe that the approach

1136: presented in this paper is both elegant and clarifying, and can in

1137: principle by applied to any configuration, including the propagation

1138: of light pulses in material media if the response of the material is

1139: properly accounted for in a dispersion relation. What we have shown

1140: here is that it is simply a matter of {\em how the story is told}, and

1141: that there is not {\em a different story} to be told, which may

1142: challenge our view of classical and quantum physics.

1143: \end{comment}

1144: 

1145: 

1146: \begin{acknowledgments}

1147: We acknowledge partial support from 

1148: UBACYT and CONICET (VLB) and from DGAPA-UNAM under

1149: project IN111306 (WLM). VLB is a member of CONICET.  

1150: \end{acknowledgments}

1151: 

1152: % Create the reference section using BibTeX:

1153: %\bibliography{pr,master}

1154: 

1155: \begin{thebibliography}{7}

1156: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

1157: \expandafter\ifx\csname bibnamefont\endcsname\relax

1158:   \def\bibnamefont#1{#1}\fi

1159: \expandafter\ifx\csname bibfnamefont\endcsname\relax

1160:   \def\bibfnamefont#1{#1}\fi

1161: \expandafter\ifx\csname citenamefont\endcsname\relax

1162:   \def\citenamefont#1{#1}\fi

1163: \expandafter\ifx\csname url\endcsname\relax

1164:   \def\url#1{\texttt{#1}}\fi

1165: \expandafter\ifx\csname urlprefix\endcsname\relax\def\urlprefix{URL }\fi

1166: \providecommand{\bibinfo}[2]{#2}

1167: \providecommand{\eprint}[2][]{\url{#2}}

1168: 

1169: 

1170: 

1171: \bibitem[{\citenamefont{Wang et~al.}(2007)\citenamefont{Wang and Xiong}}]{wang07}

1172: \bibinfo{author}{\bibfnamefont{Zhi-Yong}~\bibnamefont{Wang}} \bibnamefont{and}

1173:   \bibinfo{author}{\bibfnamefont{Cai-Dong}~\bibnamefont{Xiong}},

1174:   \bibinfo{journal}{Phys. Rev. A} \textbf{\bibinfo{volume}{75}},

1175:   \bibinfo{pages}{042105} (\bibinfo{year}{2007}).

1176: 

1177: \bibitem[{\citenamefont{Winful}(2007)\citenamefont{Winful}}]{winful07}

1178:  \bibinfo{author}{\bibfnamefont{H. G.}~\bibnamefont{Winful}},

1179:   \bibinfo{journal}{Phys. Rev. A} \textbf{\bibinfo{volume}{76}},

1180:   \bibinfo{pages}{057803} (\bibinfo{year}{2007}).

1181: 

1182: \bibitem[{\citenamefont{Pereyra et~al.}(2007)\citenamefont{Pereyra and Simanjuntak}}]{pereyra}

1183: \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Pereyra}} \bibnamefont{and}

1184:   \bibinfo{author}{\bibfnamefont{H.P.}~\bibnamefont{Simanjuntak}},

1185:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{75}},

1186:   \bibinfo{pages}{056604} (\bibinfo{year}{2007}).

1187: 

1188: \bibitem[{\citenamefont{Ranfagni et~al.}(2007)\citenamefont{Ranfagni et al}}]{ranfagni07}

1189: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Ranfagni}},

1190:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Viliani}},

1191:   \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Ranfagni}},

1192:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Mignani}},

1193:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Ruggeri}} \bibnamefont{and} 

1194:   \bibinfo{author}{\bibfnamefont{A.M.}~\bibnamefont{Ricci}},

1195:   \bibinfo{journal}{Phys. Lett. A} \textbf{\bibinfo{volume}{370}},

1196:   \bibinfo{pages}{370} (\bibinfo{year}{2007}).

1197: 

1198: 

1199: \bibitem[{\citenamefont{Winful}(2005)\citenamefont{Winful}}]{winful05}

1200: \bibinfo{author}{\bibfnamefont{H. G.}~\bibnamefont{Winful}},

1201:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{72}},

1202:   \bibinfo{pages}{046608} (\bibinfo{year}{2005}).

1203: 

1204: \bibitem[{\citenamefont{Winful}(2003)\citenamefont{Winful}}]{winful03}

1205: \bibinfo{author}{\bibfnamefont{H. G.}~\bibnamefont{Winful}},

1206:   \bibinfo{journal}{Phys. Rev. Lett} \textbf{\bibinfo{volume}{90(2)}},

1207:   \bibinfo{pages}{23901} (\bibinfo{year}{2003}).

1208: 

1209: 

1210: \bibitem[{\citenamefont{Buttiker et~al.}(2003)\citenamefont{Buttiker and

1211:   Washburn}}]{buttiker03}

1212: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Buttiker}} \bibnamefont{and}

1213:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Washburn}},

1214:   \bibinfo{journal}{Nature} \textbf{\bibinfo{volume}{422}},

1215:   \bibinfo{pages}{271} (\bibinfo{year}{2003}).

1216: 

1217: 

1218: \bibitem[{\citenamefont{Windul}(2003-1)\citenamefont{Winful}}]{winful03-1}

1219: \bibinfo{author}{\bibfnamefont{Herbert G.}~\bibnamefont{Winful}},

1220:   \bibinfo{journal}{Nature} \textbf{\bibinfo{volume}{424}},

1221:   \bibinfo{pages}{628} (\bibinfo{year}{2003}).

1222: 

1223: \bibitem[{\citenamefont{Buttiker et~al.}(2003)\citenamefont{Buttiker and  Washburn}}]{buttiker03-1}

1224: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Buttiker}} \bibnamefont{and}

1225:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Washburn}},

1226:   \bibinfo{journal}{Nature} \textbf{\bibinfo{volume}{424}},

1227:   \bibinfo{pages}{638} (\bibinfo{year}{2003}).

1228: 

1229: \bibitem[{\citenamefont{Winful}(2003-2)\citenamefont{Winful}}]{winful03-2}

1230: \bibinfo{author}{\bibfnamefont{H. G.}~\bibnamefont{Winful}},

1231:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{68}},

1232:   \bibinfo{pages}{016615} (\bibinfo{year}{2003}).

1233: 

1234: \bibitem[{\citenamefont{Peres et~al.}(2004)\citenamefont{Peres and Terno}}]{qi}

1235: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Peres}} \bibnamefont{and}

1236:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Terno}},

1237:   \bibinfo{journal}{Rev. Mod. Phys.} \textbf{\bibinfo{volume}{76}},

1238:   \bibinfo{pages}{93} (\bibinfo{year}{2004}).

1239: 

1240: \bibitem[{\citenamefont{De Angelis et~al}(2007)\citenamefont{De Angelis et al}}]{de_angelis}

1241: \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{De Angelis}},

1242: \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Nagali}},

1243: \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Sciarrino}} \bibnamefont{and}

1244:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{De Martini}},

1245:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{99}},

1246:   \bibinfo{pages}{193601} (\bibinfo{year}{2007}).

1247: 

1248: 

1249: \bibitem[{\citenamefont{Winful}(2006)\citenamefont{Winful}}]{winful06}

1250:  \bibinfo{author}{\bibfnamefont{H. G.}~\bibnamefont{Winful}},

1251:   \bibinfo{journal}{Physics Reports} \textbf{\bibinfo{volume}{436}},

1252:   \bibinfo{pages}{1} (\bibinfo{year}{2006}).

1253: 

1254: \bibitem[{\citenamefont{Boyd et~al.}(2002)\citenamefont{Boyd and Gauthier}}]{boyd02}

1255: \bibinfo{author}{\bibnamefont{R. W. Boyd and D.J. Gauthier}}, \emph{\bibinfo{title}{Progress in Optics, Vol. 43, E. Wolf (ed.)}} 

1256: (\bibinfo{publisher}{Elsevier},

1257:   \bibinfo{address}{Amsterdam}, \bibinfo{year}{2002}).

1258: 

1259: 

1260: 

1261: \bibitem[{\citenamefont{Garrison et~al}(1998)\citenamefont{Garrison,Mitchell, Chiao and

1262:  Bolda}}]{garrison1998}

1263: \bibinfo{author}{\bibfnamefont{J. C.}~\bibnamefont{Garrison}},

1264:   \bibinfo{author}{\bibfnamefont{M. W.}~\bibnamefont{Mitchell}},

1265:   \bibinfo{author}{\bibfnamefont{R. Y.}~\bibnamefont{Chiao}} \bibnamefont{and}

1266:   \bibinfo{author}{\bibfnamefont{E. L.}~\bibnamefont{Bolda}},

1267:   \bibinfo{journal}{Phys. Lett. A} \textbf{\bibinfo{volume}{245}},

1268:   \bibinfo{pages}{19} (\bibinfo{year}{1998}).

1269: 

1270: 

1271: \bibitem[{\citenamefont{Azbel'}(1994)\citenamefont{Azbel}}]{azbel}

1272: \bibinfo{author}{\bibfnamefont{Mark Ya.}~\bibnamefont{Azbel'}},

1273:   \bibinfo{journal}{Solid State Commun.} \textbf{\bibinfo{volume}{91(6)}},

1274:   \bibinfo{pages}{439} (\bibinfo{year}{1994}).

1275: 

1276: 

1277: \bibitem[{\citenamefont{Jackson}(Jackson)}]{jackson}

1278: \bibinfo{author}{\bibnamefont{J. D. Jackson}}, \emph{\bibinfo{title}{Classical Electrodynamics}} 

1279: (\bibinfo{publisher}{Wiley}, \bibinfo{address}{New York},

1280:   \bibinfo{year}{1975}), 

1281:   \bibinfo{edition}{2nd.} ed.

1282: 

1283: \bibitem[{\citenamefont{Sommerfeld}(1914)\citenamefont{Sommerfeld}}]{sommerfeld}

1284: \bibinfo{author}{\bibfnamefont{A.} \bibnamefont{Sommerfeld}},

1285:     \bibinfo{journal}{Ann. Physik} \textbf{\bibinfo{volume}{44}},

1286:   \bibinfo{pages}{177} (\bibinfo{year}{1914}).

1287: English translation available in Chap. II of \cite{brillouin-wp}.

1288: 

1289: \bibitem[{\citenamefont{Brillouin1960}(Brillouin)}]{brillouin-wp}

1290: \bibinfo{author}{\bibnamefont{L. Brillouin}}, \emph{\bibinfo{title}{Wave Propagtion and Group Velocity}} 

1291: (\bibinfo{publisher}{Academic Press},

1292:   \bibinfo{address}{New York}, \bibinfo{year}{1960}).

1293: 

1294: \bibitem[{\citenamefont{Diener}(1996)\citenamefont{Diener}}]{diener96}

1295:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Diener}},

1296:   \bibinfo{journal}{Phys. Lett. A} \textbf{\bibinfo{volume}{223}},

1297:   \bibinfo{pages}{327} (\bibinfo{year}{1996}).

1298: 

1299: \bibitem[{\citenamefont{Diener}(1997)\citenamefont{Diener}}]{diener97}

1300:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Diener}},

1301:   \bibinfo{journal}{Phys. Lett. A} \textbf{\bibinfo{volume}{235}},

1302:   \bibinfo{pages}{118} (\bibinfo{year}{1997}).

1303: 

1304: 

1305: \bibitem[{\citenamefont{Kuzmich et~al.}(2000)\citenamefont{Kuzmich, Dogariu, Wang, Milonni and Chao}}]{kuzmich}

1306: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Kuzmich}},

1307: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Dogariu}},

1308: \bibinfo{author}{\bibfnamefont{L.J.}~\bibnamefont{Wang}},

1309:   \bibinfo{author}{\bibfnamefont{P.W.}~\bibnamefont{Milonni}} \bibnamefont{and}

1310:   \bibinfo{author}{\bibfnamefont{R.Y.}~\bibnamefont{Chiao}},

1311:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{86}},

1312:   \bibinfo{pages}{3925} (\bibinfo{year}{2001}).

1313: 

1314: \bibitem[{\citenamefont{Stenner et~al.}(2003)\citenamefont{Stenner, Gauthier and Neifeld}}]{stenner}

1315: \bibinfo{author}{\bibfnamefont{M.D.}~\bibnamefont{Stenner}},

1316:   \bibinfo{author}{\bibfnamefont{D.J.}~\bibnamefont{Gauthier}} \bibnamefont{and}

1317:   \bibinfo{author}{\bibfnamefont{M.A.}~\bibnamefont{Neifeld}},

1318:   \bibinfo{journal}{Nature} \textbf{\bibinfo{volume}{425}},

1319:   \bibinfo{pages}{695} (\bibinfo{year}{2003}).

1320: 

1321: \bibitem[{\citenamefont{Stenner et~al.}(2005)\citenamefont{Stenner, Gauthier and Neifeld}}]{stenner05}

1322: \bibinfo{author}{\bibfnamefont{M.D.}~\bibnamefont{Stenner}},

1323:   \bibinfo{author}{\bibfnamefont{D.J.}~\bibnamefont{Gauthier}} \bibnamefont{and}

1324:   \bibinfo{author}{\bibfnamefont{M.A.}~\bibnamefont{Neifeld}},

1325:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{94}},

1326:   \bibinfo{pages}{053902} (\bibinfo{year}{2005}).

1327: 

1328: 

1329: \bibitem[{\citenamefont{Ranfagni et~al.}(2006)\citenamefont{Ranfagni et al}}]{ranfagni06}

1330: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Ranfagni}},

1331:   \bibinfo{author}{\bibfnamefont{P.}~\bibnamefont{Fabeni}},

1332:   \bibinfo{author}{\bibfnamefont{G.P.}~\bibnamefont{Pazzi}},

1333:   \bibinfo{author}{\bibfnamefont{A.M.}~\bibnamefont{Ricci}},

1334:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Trinci}},  

1335:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Mignani}},

1336:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Ruggeri}} \bibnamefont{and}

1337:   \bibinfo{author}{\bibfnamefont{F.}~\bibnamefont{Cardone}},

1338:   \bibinfo{journal}{Phys. Lett. A} \textbf{\bibinfo{volume}{352}},

1339:   \bibinfo{pages}{473} (\bibinfo{year}{2006}).

1340: 

1341: 

1342: 

1343: \bibitem[{\citenamefont{Icsevgi et~al.}(1969)\citenamefont{Icsevgi and Lamb}}]{icsevgi}

1344: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Icsevgi}} \bibnamefont{and}

1345:   \bibinfo{author}{\bibfnamefont{W.E.}~\bibnamefont{Lamb}},

1346:   \bibinfo{journal}{Phys. Rev.} \textbf{\bibinfo{volume}{185(2)}},

1347:   \bibinfo{pages}{517} (\bibinfo{year}{1969}).

1348: 

1349: 

1350: \bibitem[{\citenamefont{wang et~al.}(2000)\citenamefont{Wang, Kuzmich and Dogariu}}]{wang00}

1351: \bibinfo{author}{\bibfnamefont{L.J.}~\bibnamefont{Wang}},

1352:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Kuzmich}} \bibnamefont{and}

1353:   \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Dogariu}},

1354:   \bibinfo{journal}{Nature (London)} \textbf{\bibinfo{volume}{406}},

1355:   \bibinfo{pages}{277} (\bibinfo{year}{2000}).

1356: 

1357: \bibitem[{\citenamefont{Janowicz et~al.}(2006)\citenamefont{Janowicz and Mostowski}}]{janowicz}

1358: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Janowicz}} \bibnamefont{and}

1359:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Mostowski}},

1360:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{73}},

1361:   \bibinfo{pages}{046613} (\bibinfo{year}{2006}).

1362: 

1363: \bibitem[{\citenamefont{Huang et~al.}(2008)\citenamefont{Huang, Hang and Deng}}]{huang}

1364:   \bibinfo{author}{\bibfnamefont{G.}~\bibnamefont{Huang}},

1365: \bibinfo{author}{\bibfnamefont{Ch.}~\bibnamefont{Hang}} \bibnamefont{and}

1366:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Deng}},

1367:   \bibinfo{journal}{Phys. Rev. A} \textbf{\bibinfo{volume}{77}},

1368:   \bibinfo{pages}{011803(R)} (\bibinfo{year}{2008}).

1369: 

1370: \bibitem[{\citenamefont{solli et~al.}(2003)\citenamefont{Solli et al}}]{solli03}

1371: \bibinfo{author}{\bibfnamefont{D.R.}~\bibnamefont{Solli}},

1372: \bibinfo{author}{\bibfnamefont{C.F.}~\bibnamefont{McCormick}},

1373: \bibinfo{author}{\bibfnamefont{C.}~\bibnamefont{Ropers}},

1374: \bibinfo{author}{\bibfnamefont{J.J.}~\bibnamefont{Morehead}},

1375:   \bibinfo{author}{\bibfnamefont{R.Y.}~\bibnamefont{Chiao}} \bibnamefont{and}

1376:   \bibinfo{author}{\bibfnamefont{J.M.}~\bibnamefont{Hickmann}},

1377:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{91}},

1378:   \bibinfo{pages}{143906} (\bibinfo{year}{2003}).

1379: 

1380: \bibitem[{\citenamefont{Brunner et~al.}(2004)\citenamefont{Brunner,Scarani, Wegm\"uller, Legr\'e and Gisin}}]{brunner}

1381: \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Brunner}},

1382:   \bibinfo{author}{\bibfnamefont{V.}~\bibnamefont{Scarani}},

1383:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Wegm\"uller}},

1384:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Legr\'e}} \bibnamefont{and}

1385:   \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Gisin}},

1386:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{93}},

1387:   \bibinfo{pages}{203902} (\bibinfo{year}{2004}).

1388: 

1389: \bibitem[{\citenamefont{halvorsen et~al.}(2008)\citenamefont{Halvorsen and Leinaas}}]{halvorsen}

1390:   \bibinfo{author}{\bibfnamefont{T.G.}~\bibnamefont{Halvorsen}} \bibnamefont{and}

1391:   \bibinfo{author}{\bibfnamefont{J.M.}~\bibnamefont{Leinaas}},

1392:   \bibinfo{journal}{Phys. Rev. A} \textbf{\bibinfo{volume}{77}},

1393:   \bibinfo{pages}{023808} (\bibinfo{year}{2008}).

1394: 

1395: \bibitem[{\citenamefont{Kulkarni et~al.}(2004)\citenamefont{Kulkarni et al}}]{kulkarni}

1396: \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Kulkarni}},

1397:   \bibinfo{author}{\bibfnamefont{N.}~\bibnamefont{Seshadri}},

1398:   \bibinfo{author}{\bibfnamefont{V.S.C.}~\bibnamefont{Manga Rao}} \bibnamefont{and}

1399:   \bibinfo{author}{\bibfnamefont{S.}~\bibnamefont{Dutta Gupta}},

1400:   \bibinfo{journal}{J. Mod. Opt.} \textbf{\bibinfo{volume}{51}},

1401:   \bibinfo{pages}{549} (\bibinfo{year}{2004}).

1402: 

1403: 

1404: \bibitem[{\citenamefont{safian et~al.}(2006)\citenamefont{Safian, Sarris, Mojahedi}}]{safian}

1405:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Safian}},

1406:   \bibinfo{author}{\bibfnamefont{C.D.}~\bibnamefont{Sarris}} \bibnamefont{and}

1407:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Mojahedi}},

1408:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{73}},

1409:   \bibinfo{pages}{066602} (\bibinfo{year}{2006}).

1410: 

1411: \bibitem[{\citenamefont{Bigelow et~al.}(2006)\citenamefont{Bigelow, Lepeshkin, Shin and Boyd}}]{bigelow}

1412:   \bibinfo{author}{\bibfnamefont{M.S.}~\bibnamefont{Bigelow}},

1413: \bibinfo{author}{\bibfnamefont{N.N.}~\bibnamefont{Lepeshkin}}, 

1414: \bibinfo{author}{\bibfnamefont{H.}~\bibnamefont{Shin}} \bibnamefont{and}

1415:   \bibinfo{author}{\bibfnamefont{R.W.}~\bibnamefont{Boyd}},

1416:   \bibinfo{journal}{J. Phys.: Condens. Matter} \textbf{\bibinfo{volume}{18}},

1417:   \bibinfo{pages}{3117} (\bibinfo{year}{2006}).

1418: 

1419: \bibitem[{\citenamefont{Talukder et~al.}(2005)\citenamefont{Talukder, Haruta, Tomita}}]{talukder}

1420:   \bibinfo{author}{\bibfnamefont{A.I.}~\bibnamefont{Talukder}},

1421:   \bibinfo{author}{\bibfnamefont{T.}~\bibnamefont{Haruta}} \bibnamefont{and}

1422:   \bibinfo{author}{\bibfnamefont{M.}~\bibnamefont{Tomita}},

1423:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{94}},

1424:   \bibinfo{pages}{223901} (\bibinfo{year}{2005}).

1425: 

1426: \bibitem[{\citenamefont{Mugnai et~al.}(2000)\citenamefont{Mugnai, Ranfagni, Ruggeri}}]{mugnai00}

1427:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Mugnai}},

1428: \bibinfo{author}{\bibfnamefont{A.}~\bibnamefont{Ranfagni}} \bibnamefont{and}

1429:   \bibinfo{author}{\bibfnamefont{R.}~\bibnamefont{Ruggeri}},

1430:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{84}},

1431:   \bibinfo{pages}{4830} (\bibinfo{year}{2000}).

1432: 

1433: \bibitem[{\citenamefont{Mugnai et~al.}(2005)\citenamefont{Mugnai and Mochi}}]{mugnai05}

1434: \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Mugnai}} \bibnamefont{and}

1435:   \bibinfo{author}{\bibfnamefont{I.}~\bibnamefont{Mochi}},

1436:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{73}},

1437:   \bibinfo{pages}{016606} (\bibinfo{year}{2006}).

1438: 

1439: \bibitem[{\citenamefont{Carey et~al.}(2000)\citenamefont{Carey, Zawadzka, Jaroszynski, and Wynne}}]{carey00}

1440: \bibinfo{author}{\bibfnamefont{J.~J.} \bibnamefont{Carey}},

1441:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Zawadzka}},

1442:   \bibinfo{author}{\bibfnamefont{D.~A.} \bibnamefont{Jaroszynski}}

1443:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Wynne}},

1444:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{84}},

1445:   \bibinfo{pages}{1431} (\bibinfo{year}{2000}).

1446: 

1447: \bibitem[{\citenamefont{Mochan and Brudny}(2001)}]{mochan01}

1448: \bibinfo{author}{\bibnamefont{W.L.}~\bibnamefont{Moch\'an}} \bibnamefont{and}

1449:   \bibinfo{author}{\bibnamefont{V.L.}~\bibnamefont{Brudny}}, 

1450:   \bibinfo{journal}{Phys.Rev.Lett} \textbf{\bibinfo{volume}{87}},

1451:   \bibinfo{pages}{119101} (\bibinfo{year}{2001}).

1452: 

1453: \bibitem[{\citenamefont{Carey et~al.}(2001)\citenamefont{Carey, Zawadzka,  Jaroszynski and Wynne}}]{carey01}

1454: \bibinfo{author}{\bibfnamefont{J.~J.} \bibnamefont{Carey}},

1455:   \bibinfo{author}{\bibfnamefont{J.}~\bibnamefont{Zawadzka}},

1456:   \bibinfo{author}{\bibfnamefont{D.~A.} \bibnamefont{Jaroszynski}}

1457:   \bibnamefont{and} \bibinfo{author}{\bibfnamefont{K.}~\bibnamefont{Wynne}},

1458:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{87}},

1459:   \bibinfo{pages}{119102} (\bibinfo{year}{2001}).

1460: 

1461: \bibitem[{\citenamefont{Shaarawi et~al.}(2002)\citenamefont{Shaarawi, Tawfik and Besieris}}]{shaarawi}

1462:   \bibinfo{author}{\bibfnamefont{A. M.}~\bibnamefont{Shaarawi}},

1463: \bibinfo{author}{\bibfnamefont{B. H.}~\bibnamefont{Tawfik}} \bibnamefont{and}

1464:   \bibinfo{author}{\bibfnamefont{I. M.}~\bibnamefont{Besieris}},

1465:   \bibinfo{journal}{Phys. Rev. E.} \textbf{\bibinfo{volume}{66}},

1466:   \bibinfo{pages}{046626} (\bibinfo{year}{2002}).

1467: 

1468: \bibitem[{\citenamefont{Brudny and Mochan}(2001)}]{oe}

1469: \bibinfo{author}{\bibnamefont{V.L.}~\bibnamefont{Brudny}} \bibnamefont{and}

1470:   \bibinfo{author}{\bibnamefont{W.L.}~\bibnamefont{Moch\'an}}, 

1471:   \bibinfo{journal}{Optics Express} \textbf{\bibinfo{volume}{19(11)}} 

1472:   \bibinfo{pages}{561}(\bibinfo{year}{2001}).

1473: 

1474: \bibitem[{\citenamefont{Barbero et~al.}(2000)\citenamefont{Barbero et al}}]{barbero}

1475:   \bibinfo{author}{\bibfnamefont{A. P.}~\bibnamefont{Barbero}}

1476:   \bibinfo{author}{\bibfnamefont{H. E.}~\bibnamefont{Hern\'andez-Figueroa}} \bibnamefont{and}

1477:   \bibinfo{author}{\bibfnamefont{E.}~\bibnamefont{Recami}},

1478:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{62(6)}},

1479:   \bibinfo{pages}{8628} (\bibinfo{year}{2000}).

1480: 

1481: \bibitem[{\citenamefont{Reiten et~al.}(2001)\citenamefont{Reiten et al}}]{reiten}

1482:   \bibinfo{author}{\bibfnamefont{M. T.}~\bibnamefont{Reiten}}

1483:   \bibinfo{author}{\bibfnamefont{D.}~\bibnamefont{Grischkowsky}} \bibnamefont{and}

1484:   \bibinfo{author}{\bibfnamefont{R. A.}~\bibnamefont{Cheville}},

1485:   \bibinfo{journal}{Phys. Rev. E} \textbf{\bibinfo{volume}{64}},

1486:   \bibinfo{pages}{036604} (\bibinfo{year}{2001)}.

1487: 

1488: \bibitem[{\citenamefont{Balcou et~al.} (1997)\citenamefont{Balcou and Dutriaux}}]{balcou97}

1489:   \bibinfo{author}{\bibfnamefont{Ph.}~\bibnamefont{Balcou}} \bibnamefont{and}

1490:   \bibinfo{author}{\bibfnamefont{L.}~\bibnamefont{Dutriaux}},

1491:   \bibinfo{journal}{Phys. Rev. Lett.} \textbf{\bibinfo{volume}{78(5)}},

1492:   \bibinfo{pages}{851} (\bibinfo{year}{1997}).

1493: 

1494: \bibitem[{\citenamefont{Resch et~al.}(2001)\citenamefont{Resch, Lundeen and Steinberg}}]{resch}

1495:   \bibinfo{author}{\bibfnamefont{K.J.}~\bibnamefont{Resch}},

1496: \bibinfo{author}{\bibfnamefont{J.S.}~\bibnamefont{Lundeen}} \bibnamefont{and}

1497:   \bibinfo{author}{\bibfnamefont{A.M.}~\bibnamefont{Steinberg}},

1498:   \bibinfo{journal}{IEEE J. Quantum Electron.} \textbf{\bibinfo{volume}{37(6)}},

1499:   \bibinfo{pages}{794} (\bibinfo{year}{2001}).

1500: 

1501: 

1502: 

1503: \bibitem[{\citenamefont{Morse and Feshbach}(alguno)}]{inmorse}

1504: \bibinfo{author}{\bibnamefont{Phillip M.}~\bibnamefont{Morse}} \bibnamefont{and}

1505:   \bibinfo{author}{\bibnamefont{Herman}~\bibnamefont{Feshbach}}, \emph{\bibinfo{title}{Methods of Theoretical Physics}} 

1506: (\bibinfo{publisher}{Mc.Graw-Hill},

1507:   \bibinfo{address}{New York}, \bibinfo{year}{1953}).

1508: 

1509: %\bibitem[{\citenamefont{purcell}(alguno)}]{inberkeley}

1510: %\bibinfo{author}{\bibnamefont!{E.M.}\bibnamefont{purcell}}, \emph{\bibinfo{title}{Electricity and Magnetism (Berkely Physics Course, Vol. 2)}}

1511: %(\bibinfo{publisher}{alguien}, \bibinfo{address}{alguna},

1512: %  \bibinfo{year}{alguno}), p. \bibinfo{pages}{alguna},

1513: %  \bibinfo{edition}{2nd.} ed.

1514: 

1515: 

1516: \end{thebibliography}

1517: \end{document}

1518: 

1519: 

1520: %

1521: % ****** End of file template.aps ******

1522: 

1523: 

1524: % If in two-column mode, this environment will change to single-column

1525: % format so that long equations can be displayed. Use

1526: % sparingly.

1527: %\begin{widetext}

1528: % put long equation here

1529: %\end{widetext}

1530: 

1531: % \begin{figure}

1532: % \includegraphics{}%

1533: % \caption{\label{}}

1534: % \end{figure}

1535: 

1536: % Surround figure environment with turnpage environment for landscape

1537: % figure

1538: % \begin{turnpage}

1539: % \begin{figure}

1540: % \includegraphics{}%

1541: % \caption{\label{}}

1542: % \end{figure}

1543: % \end{turnpage}

1544: