1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: % Dilatonic wormholes:
3: % construction, operation, maintenance and collapse to black holes
4: % Sean A. Hayward, Sung-Won Kim and Hyunjoo Lee
5: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
6:
7: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}
8: \usepackage{graphicx}
9: \usepackage{dcolumn}
10: \usepackage{bm}
11:
12: \begin{document}
13:
14: \draft
15: \title{Dilatonic wormholes:
16: construction, operation, maintenance and collapse to black holes}
17: \author{Sean A. Hayward}\email{hayward@mm.ewha.ac.kr}
18: \author{Sung-Won Kim}\email{sungwon@mm.ewha.ac.kr}
19: \author{Hyunjoo Lee}\email{hyunjoo@mm.ewha.ac.kr}
20: \affiliation{Department of Science Education, Ewha Womans University, Seoul
21: 120-750, Korea}
22:
23: \date{\today}
24:
25: \begin{abstract}
26: The CGHS two-dimensional dilaton gravity model is generalized to include a
27: ghost Klein-Gordon field, i.e.\ with negative gravitational coupling. This
28: exotic radiation supports the existence of static traversible wormhole
29: solutions, analogous to Morris-Thorne wormholes. Since the field equations are
30: explicitly integrable, concrete examples can be given of various dynamic
31: wormhole processes, as follows. (i) Static wormholes are constructed by
32: irradiating an initially static black hole with the ghost field. (ii) The
33: operation of a wormhole to transport matter or radiation between the two
34: universes is described, including the back-reaction on the wormhole, which is
35: found to exhibit a type of neutral stability. (iii) It is shown how to maintain
36: an operating wormhole in a static state, or return it to its original state, by
37: turning up the ghost field. (iv) If the ghost field is turned off, either
38: instantaneously or gradually, the wormhole collapses into a black hole.
39: \end{abstract}
40: \pacs{04.70.Bw, 04.20.-q, 04.50.+h}
41:
42: \maketitle
43:
44: \section{Introduction}
45:
46: The theoretical existence of space-time wormholes has intrigued experts and
47: public alike. Wheeler \cite{W} speculated that quantum fluctuations in
48: space-time topology can occur, so that the smooth space-time of classical
49: Einstein gravity becomes, at the Planck scale, a continual foam of short-lived
50: interconnections. If such a wormhole could be expanded to macroscopic size, it
51: could provide a short cut between otherwise distant regions, like a bookworm
52: tunnelling between different pages of an atlas. Certainly all the standard
53: stationary black-hole solutions have wormhole ($R\times S^2$) spatial topology,
54: describing two Alice-through-the-looking-glass universes joined by the famous
55: Einstein-Rosen bridge \cite{ER}, or wormhole throat. In such cases, the two
56: universes are not causally connected, so that it is impossible to travel
57: between them, any attempt leading instead into the black hole \cite{K}.
58: However, this is only just so; a light ray can be sent along the past boundary
59: of one universe to the future boundary of the other, just escaping the black
60: hole. Thus it is not difficult to imagine a small modification which would
61: yield a traversible wormhole.
62:
63: Morris \& Thorne \cite{MT} popularized traversible wormholes as a respectable
64: theoretical possibility, studying static, spherically symmetric cases in
65: detail. The spatial topology is the same as in the black-hole cases, but the
66: throat or minimal surface is preserved in time, so that observers can pass
67: through it in either direction, travelling freely between the two universes.
68: According to the Einstein equation, such a geometry requires matter which does
69: not satisfy a classical positive-energy property, the weak energy condition.
70: However, this condition can be violated quantum-mechanically, e.g.\ in the
71: Casimir effect, or in alternative gravity theories. Such negative-energy matter
72: was dubbed exotic matter by Morris \& Thorne. This provoked renewed interest in
73: traversible wormholes, as reviewed by Visser \cite{V}. The possible
74: astrophysical existence of wormholes has been taken seriously enough for
75: searches of observational data \cite{TRA}.
76:
77: The Morris-Thorne framework begs development in at least two important
78: respects, quite apart from generalizing beyond static, spherically symmetric
79: cases. Firstly, the exotic matter was not modelled, but simply assumed to exist
80: in exactly the configuration needed to support the wormhole. Secondly, the
81: back-reaction of the transported matter on the wormhole was ignored. If the
82: wormhole turns out to be unstable to such back-reaction, it would be useless
83: for actual transport. Many physicists' instinctive reaction is that negative
84: energy, unbounded below, will lead to instability. Another practical question
85: is: if the negative-energy source fails, i.e.\ the exotic matter is not
86: maintained, what happens to the wormhole? Again, a pessimistic reaction is that
87: the negative energy densities are likely to create naked singularities. An
88: alternative prediction was that the wormhole would collapse to a black hole
89: \cite{wh}. The same reference introduced a framework for studying dynamic
90: wormholes, in which both wormholes and black holes are locally defined by the
91: same geometrical objects, trapping horizons, with one key difference: black
92: holes have achronal horizons and wormholes have temporal horizons, so that they
93: are respectively one-way and two-way traversible. This also indicated that a
94: wormhole could be constructed from a black hole using exotic matter.
95:
96: To find definite answers to the above questions, it is necessary to specify the
97: exotic matter model. Various studies have been based on alternative gravity
98: theories or semi-classical quantum field theory, e.g.\ \cite{HPS,KL}. However,
99: the difficulty of solving the field equations tends to obscure issues of
100: principle. Here we propose a toy model, intended to describe the essential
101: dynamics of a wormhole, but explicitly integrable, so that concrete answers to
102: the above questions are easily found. A similar approach in the context of
103: black-hole evaporation involved the CGHS two-dimensional dilaton gravity model
104: \cite{CGHS}, which was expected to share similar features with spherically
105: symmetric black-hole evaporation by Hawking radiation. Certainly classical
106: black-hole dynamics is qualitatively similar, for instance in regard to cosmic
107: censorship, but such properties are much easier to prove in the toy model
108: \cite{cc} than in the corresponding realistic model, the spherically symmetric
109: Einstein-Klein-Gordon system \cite{C}. Here we propose a simple generalization
110: of the CGHS model to include a ghost Klein-Gordon field, i.e.\ with the
111: gravitational coupling taking the opposite sign to normal. This is a specific
112: model for the exotic matter, or more accurately exotic radiation, as we take
113: the massless field for simplicity.
114:
115: The article is organised as follows. \S II describes the model and how its
116: general solution can be constructed from initial data. \S III reviews the
117: static black-hole solution and introduces the static wormhole solution, each of
118: which depend on one parameter, mass or horizon radius. \S IV describes what
119: happens to an initially static wormhole when the ghost field supporting it is
120: turned off and allowed to disperse. \S V shows how to construct a static
121: wormhole by irradiating an initially static black hole with the ghost field. \S
122: VI describes the operation of a wormhole for transport or signalling, using a
123: normal Klein-Gordon field to model the matter or radiation, including the
124: back-reaction of the field on the wormhole. \S VII shows how to maintain the
125: operating wormhole in a static state, or return it to its original state. \S
126: VIII concludes.
127:
128: \section{Dilaton gravity}
129:
130: The CGHS two-dimensional dilaton gravity of Callan {\it et al.} \cite{CGHS} is
131: generalized by the action
132: \begin{equation}\label{action}
133: \int_S \mu
134: \left[ e^{-2\phi } \left( R + 4 (\nabla \phi )^2 + 4\lambda ^2 \right) -
135: \frac{1}{2} (\nabla f)^2
136: + \frac{1}{2}(\nabla g)^2 \right]
137: \end{equation}
138: where $S$ is a 2-manifold, $\mu$, $R$ and $\nabla$ are the area form, Ricci
139: scalar and covariant derivative of a Lorentz 2-metric on $S$, $\lambda$
140: represents a negative cosmological constant, $\phi$ is a scalar dilaton field,
141: $f$ is a scalar field representing matter and $g$ is a ghost scalar field with
142: negative gravitational coupling. The last term is added to the CGHS action in
143: order that g provides the negative energy densities needed to maintain a
144: wormhole \cite{MT,V,wh,HV1,HV2,IH,wd}.
145:
146: \begin{figure}
147: \includegraphics[width=9cm,height=13cm,angle=0]{initialdata.eps}
148: \vspace{-80mm}
149: \caption{Initial data, taking $x^\pm_0=0$.} \label{initialdata}
150: \end{figure}
151:
152: By choosing future-pointing null coordinates $(x^+ , x^- )$, the line element
153: may be written as
154: \begin{equation}\label{metric}
155: ds^2 = -2 e^{2\rho} dx^+ dx^-
156: \end{equation}
157: where the conventional factor of 2 differs from that of earlier references
158: \cite{CGHS,cc}. One component of the field equations is
159: \begin{equation}\label{dilaton eqn}
160: \partial_+ \partial_- \phi = \partial_+ \partial_- \rho
161: \end{equation}
162: where $\partial_\pm=\partial/\partial x^\pm$, so the coordinate
163: freedom $x^\pm\mapsto\hat x^\pm(x^\pm)$ can be used to take
164: \begin{equation}\label{fix}
165: \rho = \phi.
166: \end{equation}
167: The remaining coordinate freedom is just
168: \begin{equation}\label{freedom}
169: x^\pm\mapsto e^{\pm b}x^\pm+c^\pm
170: \end{equation}
171: where the constants $c^\pm$ fix the origin and the constant $b$ refers to
172: relative rescalings of $x^\pm$. It is convenient to transform the dilaton field
173: $\phi$ to
174: \begin{equation}\label{r}
175: r=2e^{-2\phi}.
176: \end{equation}
177: Then the remaining field equations are the evolution equations
178: \begin{eqnarray}
179: \label{fequation}
180: \partial_+ \partial_- f &=& 0 \\
181: \label{gequation}
182: \partial_+ \partial_- g &=& 0 \\
183: \label{requation}
184: \partial_+ \partial_- r &=& -4 \lambda^2
185: \end{eqnarray}
186: and the constraints
187: \begin{equation}\label{constraints0}
188: \partial_\pm \partial_\pm r = ( \partial_\pm g )^2 -
189: (\partial_\pm f )^2 .
190: \end{equation}
191: The evolution equations (\ref{fequation}-\ref{requation}) have the
192: general solutions
193: \begin{eqnarray}
194: \label{fsolution}
195: f(x^+,x^-) &=& f_+ (x^+) + f_- (x^-) \\
196: \label{gsolution}
197: g(x^+,x^-) &=& g_+ (x^+) + g_- (x^-) \\
198: \label{rsolution}
199: r(x^+,x^-) &=& r_+ (x^+) + r_- (x^-) - 4\lambda ^2 x^+ x^-.
200: \end{eqnarray}
201: The constraints (\ref{constraints0}) are preserved by the evolution equations
202: in the $\partial_\mp$ directions, and so may be reduced to
203: \begin{equation}\label{constraints}
204: \partial_\pm \partial_\pm r_\pm = {G_\pm}^2 - {F_\pm}^2
205: \end{equation}
206: where the null derivatives
207: \begin{eqnarray}
208: F_\pm&=&\partial_\pm f_\pm\\
209: G_\pm&=&\partial_\pm g_\pm
210: \end{eqnarray}
211: are convenient variables. The constraints are manifestly integrable for $r_\pm$
212: given initial data
213: \begin{eqnarray}\label{data}
214: (f_+,g_+)\qquad&&\hbox{on $x^-=x_0^-$}\\
215: (f_-,g_-)\qquad&&\hbox{on $x^+=x_0^+$}\\
216: (r,\partial_+r,\partial_-r)\qquad&&\hbox{at $x^+=x_0^+$, $x^-=x_0^-$}
217: \end{eqnarray}
218: for constants $x_0^\pm$ (Fig.~\ref{initialdata}). Then the general procedure is
219: to specify this initial data, integrate the constraints (\ref{constraints}) for
220: $r_\pm$, then the general solution follows as
221: (\ref{fsolution}-\ref{rsolution}).
222:
223: Finally, we note from the vacuum constraints (\ref{constraints0}) that $r$ is
224: analogous to the areal radius in spherically symmetric Einstein gravity
225: \cite{cc,1st}, correctly reproducing the expansion of a light wave in flat
226: space-time, $\partial_\pm\partial_\pm r=0$.
227:
228: \section{Static black-hole and wormhole solutions}
229:
230: In vacuum, $f=g=0$, the general solution to the field equations is \cite{CGHS}
231: \begin{equation}\label{vacuumsol}
232: r = 2m - 4\lambda^2 x^+ x^-
233: \end{equation}
234: where the origin has been fixed using (\ref{freedom}). The constant $m$ may be
235: interpreted as the mass of the space-time, whose global structure has been
236: described previously \cite{cc}. For $m>0$ there is a static black hole,
237: analogous to the Schwarzschild black hole (Fig.~\ref{staticbhwh}(a)). The case
238: $m=0$ is flat and the case $m<0$ contains an eternal naked singularity,
239: analogous to the negative-mass Schwarzschild solution. Henceforth we take
240: $m>0$. Note that the original reference \cite{CGHS} defined mass as $\lambda
241: m$, in which case $\lambda r$ would be analogous to radius, but for $\lambda>0$
242: this makes no essential difference in the following.
243:
244: \begin{figure}
245: \includegraphics[width=8cm,height=12cm,angle=0]{staticbhwh.eps}
246: \vspace{-15mm}
247: \caption{Penrose conformal diagrams of (a) the static black hole,
248: (b) the static wormhole.
249: Both space-times are divided into two universes (unshaded regions),
250: but observers can travel freely between them via the wormhole,
251: whereas the black hole (upper shaded region) swallows any such attempt.}
252: \label{staticbhwh}
253: \end{figure}
254:
255: Next, we consider the case that $f=0$ but $g \neq 0 $, finding the solution
256: \begin{eqnarray}\label{staticwhsol}
257: r &=& a + 2\lambda^2 (x^+ - x^- )^2 \\
258: g &=& 2\lambda (x^+ - x^- )
259: \end{eqnarray}
260: where the origin has again been fixed. A coordinate transformation $x^\pm=(t\pm
261: z)/\sqrt{2}$ shows that the solution is static:
262: \begin{equation}
263: ds^2 = \frac{2(dz^2-dt^2)}{a+4\lambda^2 z^2} .
264: \end{equation}
265: If $a>0$, this represents a traversible wormhole with analogous global
266: structure to a Morris-Thorne wormhole, with a throat $r=a$ at $z=0$, joining
267: two regions with $r>a$, a $z>0$ universe and a reflected $z<0$ universe
268: (Fig.~\ref{staticbhwh}(b)). If $a < 0$, there is an eternal naked singularity
269: $r=0$, while if $a=0$, the space-time has constant negative curvature, as can
270: be seen by calculating the Ricci scalar \cite{cc}
271: \begin{equation}
272: R=r^{-1}\partial_+r\partial_-r -\partial_+\partial_-r.
273: \end{equation}
274: Henceforth we take $a>0$.
275:
276: In summary, the model naturally contains both static black holes and static
277: traversible wormholes. Before proceeding, we note an important feature of both
278: cases, the trapping horizons, defined by $\nabla r\cdot\nabla r=0$, or
279: equivalently $\partial_+r=0$ or $\partial_-r=0$ \cite{wh,cc,1st,bhd}. In the
280: black hole, they coincide with the event horizons $r=2m$ at $x^-=0$ and $x^+=0$
281: respectively. In the wormhole, there is a double trapping horizon,
282: $\partial_+r=\partial_-r=0$, at the throat $r=a$. This illustrates how trapping
283: horizons of different type may be used to locally define both black holes and
284: wormholes \cite{wh}. Locating the trapping horizons, or equivalently the
285: trapped regions where $\nabla r\cdot\nabla r<0$, is a key feature of the
286: following analysis of dynamic situations.
287:
288: \section{Wormhole collapse}
289:
290: One can now study what happens to a static wormhole if its negative-energy
291: source fails. We consider first that the supporting ghost field $g$ is switched
292: off suddenly from both sides of the wormhole, then that $g$ is instead
293: gradually reduced to zero.
294:
295: \subsection{Sudden collapse}
296:
297: We set the initial data so that there is a static wormhole, with $g$ then
298: switched off suddenly from both sides of the wormhole (Fig.~\ref{sud}(a)):
299:
300: \begin{figure}
301: \includegraphics[width=7cm,height=10cm,angle=0]{sud.eps}
302: \vspace{-10mm}
303: \caption{(a) The ghost field $g$ is switched off suddenly
304: from both sides of the wormhole. (b) Conformal diagram: when the wormhole's
305: negative-energy source fails suddenly at $x^\pm = 0$, it immediately collapses
306: into a black hole.}
307: \label{sud}
308: \end{figure}
309:
310: \begin{eqnarray}\label{suddenGeq}
311: G_\pm = \left\{
312: \begin{array}{ll}
313: \pm 2\lambda\,, & x^\pm < 0 \\
314: 0\,, &x^\pm \geq 0.
315: \end{array}\right.
316: \end{eqnarray}
317: Taking $f=0$, the constraints (\ref{constraints}) are
318: \begin{eqnarray}\label{suddenreq}
319: \partial_\pm \partial_\pm r_\pm =
320: \left\{
321: \begin{array}{ll}
322: 4 \lambda^2\,, &x^\pm < 0 \\
323: 0\,, & x^\pm \geq 0.
324: \end{array}\right.
325: \end{eqnarray}
326: Integrating twice, we obtain
327: \begin{eqnarray}\label{suddenr}
328: r_\pm=
329: \left\{
330: \begin{array}{ll}
331: 2 \lambda^2 (x^\pm )^2 + \frac{a}{2}\,, &x^\pm < 0\\
332: \frac{a}{2}\,, &x^\pm \geq 0
333: \end{array}\right.
334: \end{eqnarray}
335: where the constants of integration are determined by continuity at $x^\pm = 0$.
336: The resulting solution from (\ref{rsolution}) is
337: \begin{eqnarray}\label{suddensol}
338: r =
339: \left\{
340: \begin{array}{ll}
341: a + 2 \lambda^2 (x^+ - x^- )^2\,, &x^+ < 0 ~{\rm and}~ x^- < 0\\
342: a + 2 \lambda^2 x^+ (x^+ - 2x^- )\,, &x^+ < 0 ~{\rm and}~ x^- \geq 0 \\
343: a - 2 \lambda^2 x^- (2x^+ - x^- )\,, & x^+ \geq 0 ~{\rm and}~ x^- < 0 \\
344: a - 4 \lambda^2 x^+ x^-\,, &x^+ \geq 0 ~{\rm and}~ x^- \geq 0.
345: \end{array}\right.
346: \end{eqnarray}
347: One may recognize the solution in the final region as a static black hole
348: (\ref{vacuumsol}) with mass $m=a/2$. The double trapping horizon of the static
349: wormhole bifurcates into the event horizons of the black hole, all with
350: $r=a=2m$ (Fig.~\ref{sud}(b)). Thus the wormhole immediately collapses into a
351: black hole.
352:
353: The Schwarzschild-like relationship between mass and throat radius is no
354: accident; there is a definition of active gravitational mass-energy \cite{cc}
355: \begin{equation}\label{energy}
356: E=\frac{r}2\left(1-\frac{\nabla r\cdot\nabla r}{4\lambda^2r^2}\right)
357: \end{equation}
358: which evaluates as $r/2$ on any trapping horizon. Thus a wormhole with a throat
359: $r=a$ has an effective mass $a/2$. Here and elsewhere, it is useful to recall
360: the analogy with spherically symmetric Einstein gravity, where $r$ corresponds
361: to areal radius and there is a similar definition of active gravitational
362: mass-energy \cite{1st}.
363:
364: \subsection{Gradual collapse}
365:
366: Again starting with a static wormhole, we now reduce the ghost field gradually
367: to zero in the simplest way, linearly (Fig.~\ref{grad}(a)):
368: \begin{eqnarray}\label{gradualGeq}
369: G_\pm=
370: \left\{
371: \begin{array}{ll}
372: \pm 2 \lambda\,, &x^\pm < 0\\
373: \mp \alpha x^\pm \pm 2 \lambda\,, &0 \leq x^\pm < x_0 \\
374: 0\,, &x_0 \leq x^\pm
375: \end{array} \right.
376: \end{eqnarray}
377: where $\alpha$ is a constant and $x_0 = 2\lambda/\alpha$. Here $\alpha\to0$
378: recovers the static wormhole and $\alpha\to\infty$ recovers sudden collapse. By
379: similar calculations, again taking $f=0$,
380: \begin{eqnarray}\label{gradualreq}
381: \partial_\pm \partial_\pm r_\pm =
382: \left\{
383: \begin{array}{ll}
384: 4\lambda^2\,, &x^\pm < 0\\
385: \alpha^2 {x^\pm}^2 - 4 \lambda \alpha x^\pm + 4\lambda^2\,,
386: &0 \leq x^\pm < x_0 \\
387: 0\,, &x_0 \leq x^\pm
388: \end{array}\right.
389: \end{eqnarray}
390: integrate to
391: \begin{eqnarray}\label{gradualr}
392: r_\pm=
393: \left\{
394: \begin{array}{ll}
395: 2\lambda^2 {x^\pm}^2 + \frac{a}{2}\,, &x^\pm < 0\\
396: \frac{\alpha^2}{12} {x^\pm}^4 - \frac{2\alpha\lambda}{3} {x^\pm}^3 + 2\lambda^2 {x^\pm}^2
397: + \frac{a}{2}\,,
398: &0 \leq x^\pm < x_0 \\
399: \frac{8 \lambda^3}{3\alpha} x^\pm - \frac{4\lambda^4}{3\alpha^2} +
400: \frac{a}{2}\,,
401: &x_0 \leq x^\pm.
402: \end{array}\right.
403: \end{eqnarray}
404: and the solution
405: \begin{widetext}
406: \begin{eqnarray}\label{gradualsol}
407: r=
408: \left\{
409: \begin{array}{ll}
410: a + 2 \lambda^2 (x^+ - x^- )^2\,,
411: & x^+ < 0 ~{\rm and}~ x^- < 0\\
412: a + 2 \lambda^2 (x^+ - x^- )^2 - \frac{2\alpha\lambda}{3} {x^- }^3
413: + \frac{\alpha^2}{12} {x^- }^4\,,
414: &x^+ < 0 ~{\rm and}~ 0 \leq x^- < x_0 \\
415: a + 2 \lambda^2(x^+ - x^- )^2 - \frac{2\alpha\lambda}{3} {x^+ }^3
416: + \frac{\alpha^2}{12} {x^+ }^4\,,
417: &0 \leq x^+ < x_0 ~{\rm and}~ x^- < 0 \\
418: a - \frac{4\lambda^4}{3\alpha^2} - 4 \lambda^2 x^+ x^-
419: + \frac{8\lambda^3}{3\alpha} x^- + 2\lambda^2 {x^+} ^2\,,
420: &x^+ < 0 ~{\rm and}~ x_0 \leq x^- \\
421: a + 2 \lambda^2 (x^+ - x^- )^2 - \frac{2\alpha\lambda}{3} ( {x^+}^3 + {x^- }^3)
422: + \frac{\alpha^2}{12} ( {x^+}^4 + {x^- }^4)\,,
423: &0 \leq x^+ < x_0 ~{\rm and}~ 0 \leq x^- < x_0 \\
424: a - \frac{4\lambda^4}{3\alpha^2} - 4 \lambda^2 x^+ x^-
425: + \frac{8\lambda^3}{3\alpha} {x^+ } + 2\lambda^2 {x^-
426: }^2\,,
427: & x_0 \leq x^+ ~{\rm and}~ x^- < 0\\
428: a - \frac{4\lambda^4}{3\alpha^2} - 4 \lambda^2 x^+ x^- + \frac{8\lambda^3}{3\alpha} x^-
429: + 2\lambda^2 {x^+ }^2
430: - \frac{2\alpha\lambda}{3} {x^+ }^3 + \frac{\alpha^2}{12} {x^+
431: }^4\,,
432: &0 \leq x^+ < x_0 ~{\rm and}~ x_0 \leq x^-\\
433: a - \frac{4\lambda^4}{3\alpha^2} - 4 \lambda^2 x^+ x^- + \frac{8\lambda^3}{3\alpha} x^+
434: + 2\lambda^2 {x^- }^2
435: - \frac{2\alpha\lambda}{3} {x^- }^3 + \frac{\alpha^2}{12} {x^-
436: }^4\,,
437: & x_0 \leq x^+ ~{\rm and}~ 0 \leq x^- < x_0 \\
438: a - \frac{8\lambda^4}{9\alpha^2} - 4 \lambda^2 \left(x^+ - \frac{2\lambda}{3\alpha} \right)
439: \left(x^- - \frac{2\lambda}{3\alpha}
440: \right)\,,
441: & x_0 \leq x^+ ~{\rm and}~ x_0 \leq x^-.
442: \end{array}\right.
443: \end{eqnarray}
444: \end{widetext}
445:
446: \begin{figure}
447: \includegraphics[width=8cm,height=13cm,angle=0]{grad.eps}
448: \vspace{-32mm}
449: \caption{(a) The ghost field $g$ is gradually reduced to zero.
450: (b) Conformal diagram: the wormhole throat bifurcates and the resulting
451: non-static wormhole again eventually becomes a black hole.
452: Shading in these diagrams indicates trapped regions,
453: where $\partial_+r\partial_-r>0$.}
454: \label{grad}
455: \end{figure}
456:
457: The solution in the final region is again recognizable as a black hole
458: (\ref{vacuumsol}), but with reduced mass and shifted by
459: $\frac{2\lambda}{3\alpha}=\frac{x_0}3$ in $x^\pm$ compared with the sudden
460: case. The trapping horizons $\partial_\pm r=0$ are located at
461: \begin{eqnarray}\label{gr-hor}
462: x^\mp=
463: \left\{
464: \begin{array}{ll}
465: x^\pm\,, &x^\pm < 0\\
466: x^\pm - \frac{1}{x_0}{x^\pm}^2
467: +\frac{1}{3x_0^2}{x^\pm}^3\,, &0 \leq x^\pm < x_0 \\
468: \frac{x_0}{3}\,, &x_0 \leq x^\pm.
469: \end{array} \right.
470: \end{eqnarray}
471: The final expressions coincide with the locations of the event horizons of the
472: black hole. The middle expressions are cubic curves which smoothly join the
473: initial wormhole throat to the final event horizons (Fig.~\ref{grad}(b)). Thus
474: the double trapping horizon of the wormhole has again bifurcated, eventually
475: forming a black hole. In between, the geometry may be characterized as a
476: non-static wormhole, as observers may still traverse it for a certain time,
477: after which the attempt leads only into the black hole.
478:
479: \section{Wormhole construction}
480:
481: We next study how to convert a black hole into a traversible wormhole by
482: irradiating it with the ghost field. This is not simply the time reverse of the
483: wormhole collapses described above, which would represent the creation of a
484: wormhole from a white hole, as one can see by inverting the conformal diagrams
485: (Fig.~\ref{sud}(b),\ref{grad}(b)). Instead, one needs to begin the irradiation
486: at some positive value $x_0$ of the Kruskal-like coordinates $x^\pm$
487: (Fig.~\ref{staticbhwh}(a)). Moreover, in order to close up a future trapped
488: region by merging its trapping horizons, the negative energy densities required
489: must be larger than those required to maintain the static wormhole which
490: finally forms \cite{wh}. In other words, $|G_\pm|$ must first reach a maximum
491: greater than $2\lambda$. The simplest way to achieve this is to set initial
492: data as double step functions (Fig.~\ref{cons}(a))
493: \begin{eqnarray}\label{constructGeq}
494: G_\pm =
495: \left\{
496: \begin{array}{ll}
497: 0\,, &x^\pm < x_0 \\
498: \pm 2\beta\lambda\,, &x_0 \leq x^\pm < x_1 \\
499: \pm 2 \lambda\,, &x_1 \leq x^\pm
500: \end{array}\right.
501: \end{eqnarray}
502: where $\beta>1$ is a constant, to be determined by the requirement that the
503: trapping horizons merge at $x^\pm=x_1$, for which we find
504: $x_0=x_1(1-\beta^{-2})$. Then taking $f=0$, the constraints (\ref{constraints})
505: are
506: \begin{eqnarray}\label{constructreq}
507: \partial_\pm \partial_\pm r_\pm =
508: \left\{
509: \begin{array}{ll}
510: 0\,, &x^\pm < x_0 \\
511: 4 \lambda^2\beta^2\,,
512: & x_0 \leq x^\pm < x_1 \\
513: 4 \lambda^2\,, & x_1 \leq x^\pm.
514: \end{array}\right.
515: \end{eqnarray}
516: Integrating twice, assuming a black hole of mass $m$ in the initial region,
517: leads to
518: \begin{eqnarray}\label{constructr}
519: r_\pm =
520: \left\{
521: \begin{array}{ll}
522: m \,, &x^\pm < x_0 \\
523: 2\lambda^2\beta^2(x^\pm-x_0)^2 + m\,, &x_0 \leq x^\pm < x_1\\
524: 2 \lambda^2 ({x^\pm}^2 - x_0 x_1) + m\,, &x_1 \leq x^\pm
525: \end{array}\right.
526: \end{eqnarray}
527: and
528: \begin{widetext}
529: \begin{eqnarray}\label{constructsol}
530: r=
531: \left\{
532: \begin{array}{ll}
533: 2m - 4\lambda^2 x^+ x^- \,,
534: &x^+ < x_0 ~{\rm and}~ x^- < x_0 \\
535: 2m - 4\lambda^2 x^+ x^- + 2\lambda^2\beta^2\left((x^+-x_0)^2+(x^--x_0)^2\right)\,,
536: & x_0 \leq x^+ < x_1 ~{\rm and}~ x_0 \leq x^- < x_1 \\
537: 2\lambda^2 (x^+ - x^-)^2 + 2m - 4\lambda^2 x_0 x_1 \,,
538: & x_1 \leq x^+ ~{\rm and}~ x_1 \leq x^-
539: \end{array}\right.
540: \end{eqnarray}
541: \end{widetext}
542: where we have omitted the less relevant regions. One may recognize the solution
543: in the final region as a static wormhole (\ref{staticwhsol}) with throat radius
544: $2m-4\lambda^2x_0x_1$. Thus we require $2\lambda^2x_0x_1<m$. By choice of the
545: parameters $(x_0,x_1)$, we have constructed a static wormhole with any throat
546: radius less than $2m$, the radius of the original black hole. The trapping
547: horizons $\partial_\pm r=0$ are located at
548: \begin{eqnarray}\label{construct-hor}
549: x^\mp=
550: \left\{
551: \begin{array}{ll}
552: 0\,, &x^\pm < x_0\\
553: \beta^2(x^\pm-x_0) \,, &x_0 \leq x^\pm < x_1 \\
554: x^\pm\,, &x_1 \leq x^\pm
555: \end{array} \right.
556: \end{eqnarray}
557: which are straight line segments. Thus the ghost radiation causes the trapping
558: horizons of the initial black hole to shrink towards each other, eventually
559: merging to form the throat of the final static wormhole (Fig.~\ref{cons}(b)).
560: The trapped region composing the black hole simply evaporates. This classically
561: unexpected behaviour is, of course, due to the negative energy densities. As in
562: the wormhole collapse case, the trapping horizons can be smoothed off by taking
563: smoother profiles for $G_\pm$, but the details are not particularly
564: illuminating. In summary, static wormholes have been constructed by irradiating
565: a black hole with ghost radiation.
566:
567: \begin{figure}
568: \includegraphics[width=7cm,height=10cm,angle=0]{cons.eps}
569: \vspace{-2mm}
570: \caption{(a) Irradiating a vacuum black hole with the ghost field $g$.
571: (b) Conformal diagram: the initially static black hole, as in
572: (Fig.~\ref{staticbhwh}(a)), becomes a dynamic wormhole, eventually reaching a
573: static state as in (Fig.~\ref{staticbhwh}(b)). The black hole has been
574: converted into a traversible wormhole.}
575: \label{cons}
576: \end{figure}
577:
578: One can also regard this as analogous to black-hole evaporation, with the ghost
579: radiation modelling the ingoing negative-energy Hawking radiation, suggesting
580: that the final state of black hole evaporation might be a stationary wormhole
581: \cite{wh,HPS,KL}. This would naturally resolve the information-loss puzzle, as
582: there is no singularity in which information is lost; everything that fell into
583: the black hole eventually re-emerges.
584:
585: \section{Wormhole operation}
586:
587: If a wormhole is actually used to transport a parcel or person between the two
588: universes, the transported matter will affect the wormhole by changing the
589: gravitational field. In principle this occurs even if the wormhole is merely
590: used for signalling. We will study the dynamical effects of such back-reaction
591: by using the field $f$ to model the matter or radiation. The use of
592: Klein-Gordon radiation rather than more realistic matter is for simplicity
593: only; any source of mass-energy would have more or less similar gravitational
594: effects. In the current model, the constraints (\ref{constraints}) show
595: explicitly that increasing ${F_\pm}^2$ has an equivalent gravitational effect
596: to reducing ${G_\pm}^2$, which we have already shown can cause collapse to a
597: black hole. Thus it is to be expected that too much transport would destroy the
598: wormhole. A worse possibility is that the wormhole might be unstable to the
599: slightest perturbation and start to collapse immediately. We investigate this
600: below, giving the first concrete examples of wormhole operation including
601: back-reaction.
602:
603: \begin{figure}
604: \includegraphics[width=8cm,height=11cm,angle=0]{opr1.eps}
605: \vspace{-7mm}
606: \caption{(a) A step pulse of positive-energy radiation is beamed
607: through the wormhole.
608: (b) Conformal diagram: the wormhole becomes non-static but, for a
609: small-energy pulse, remains traversible for a long time.}
610: \label{opr1}
611: \end{figure}
612:
613: Again taking the simplest case, we consider a step pulse of positive-energy
614: radiation (Fig.~\ref{opr1}(a)):
615: \begin{eqnarray}\label{step}
616: F_+ &=&
617: \left\{
618: \begin{array}{ll}
619: 0\,,~ &x^+ < 0 \\
620: \Delta\,,& 0 \leq x^+ < x_0 \\
621: 0\,, & x_0 \leq x^+
622: \end{array}\right. \\
623: F_- &=& 0 \nonumber
624: \end{eqnarray}
625: with $G_\pm = \pm 2\lambda$. The energy of the pulse may be defined as the
626: change in the gravitational energy (\ref{energy}) due to the pulse, which
627: evaluates at $x^-=0$ as
628: \begin{equation}
629: \epsilon=\frac14(\Delta x_0 )^2.
630: \end{equation}
631: In the following, we assume that the energy of the pulse should not be too
632: large, corresponding to the anticipated limit on how much mass-energy can be
633: sent through the wormhole without causing it to collapse into a black hole.
634: Specifically we find $\epsilon<a/2$ to avoid an $r=0$ singularity in the middle
635: region. Thus the pulse energy should be less than the effective mass of the
636: wormhole. Inserting the conversion factor $c^2/G$ from length to mass, a
637: one-metre wormhole could transport over a hundred Earth masses. This is hardly
638: a practical limitation, once we establish stability.
639:
640: The constraints (\ref{constraints})
641: \begin{eqnarray}
642: \partial_+ \partial_+ r_+ &=&
643: \left\{
644: \begin{array}{ll}
645: 4\lambda^2\,, &x^+ < 0 \\
646: 4\lambda^2 - \Delta^2\,,& 0 \leq x^+ < x_0 \\
647: 4 \lambda^2\,, & x_0 \leq x^+
648: \end{array}\right. \\
649: \partial_- \partial_- r_- &=&4 \lambda^2\nonumber
650: \end{eqnarray}
651: integrate to
652: \begin{widetext}
653: \begin{eqnarray}
654: r_+ &=&
655: \left\{
656: \begin{array}{ll}
657: 2\lambda^2 {x^+}^2 + \frac{a}{2}\,, &x^+ < 0 \\
658: 2\lambda^2 {x^+}^2 - \frac{\Delta^2}{2}{x^+}^2
659: + \frac{a}{2}\,,~ & 0 \leq x^+ < x_0 \\
660: 2\lambda^2 {x^+}^2 + \Delta^2 x_0 \left( \frac{1}{2}x_0 - x^+ \right)
661: + \frac{a}{2}\,,
662: ~& x_0 \leq x^+
663: \end{array}\right.\\
664: r_- &=&2\lambda^2 {x^-}^2 + \frac{a}{2}\nonumber
665: \end{eqnarray}
666: and the solution is
667: \begin{eqnarray}
668: r =
669: \left\{
670: \begin{array}{ll}
671: a + 2\lambda^2 \left({x^+} - {x^-} \right)^2\,, &x^+ < 0 \\
672: a + 2\lambda^2 \left({x^+} - {x^-} \right)^2- \frac{\Delta^2}{2}{x^+}^2
673: \,,~ & 0 \leq x^+ < x_0 \\
674: a + 2\lambda^2 \left({x^+} - {x^-} \right)^2
675: + \Delta^2 x_0 \left( \frac{1}{2}x_0 - x^+ \right) \,, ~& x_0 \leq x^+.
676: \end{array}\right.\,
677: \end{eqnarray}
678: \end{widetext}
679: The locations of the trapping horizons $\partial_+r=0$ and $\partial_-r=0$ are
680: given respectively by
681: \begin{eqnarray}
682: x^- &=&
683: \left\{
684: \begin{array}{ll}
685: x^+ \,, &x^+ < 0 \\
686: x^+ - \frac{\Delta^2}{4\lambda^2} x^+
687: \,,~ & 0 \leq x^+ < x_0 \\
688: x^+ - \frac{\Delta^2}{4\lambda^2} x_0
689: \,, ~& x_0 \leq x^+
690: \end{array}\right.\\
691: x^+ &=& x^- \,.\nonumber
692: \end{eqnarray}
693: Thus the double trapping horizon of the initially static wormhole bifurcates
694: when the radiation arrives (Fig.~\ref{opr1}(b)). After the pulse has passed,
695: the two trapping horizons run parallel in the $x^\pm$ coordinates, forming a
696: non-static traversible wormhole. If the pulse energy is small, $\epsilon\ll a$,
697: the wormhole persists in an almost static state for a long time, $t\sim
698: x_0a/\epsilon$. Nevertheless, even for an arbitrarily weak pulse, eventually a
699: spatial $r=0$ singularity develops, similar to that of the static black hole.
700: Observers close enough to the singularity can no longer traverse the wormhole,
701: so it constitutes a black hole with two event horizons. Thus the static
702: wormhole exhibits a type of neutral stability, neither strictly stable in that
703: it does not return to its initial state, nor strictly unstable in that there is
704: no sudden runaway. Note that black holes are also neutrally stable in this
705: sense; perturbing a black hole by dropping positive-energy matter into it
706: increases the area of the trapping horizon finitely, by the first and second
707: laws of black-hole dynamics \cite{cc,1st,bhd}.
708:
709: \section{Wormhole maintenance}
710:
711: Keeping an operating wormhole viable indefinitely, defying its natural fate as
712: a black hole, requires additional negative energy to balance the transported
713: matter. The simplest way to maintain the wormhole is just to set initial data
714: \begin{equation}
715: G_\pm = \pm\sqrt{4\lambda^2 + {F_\pm}^2}
716: \end{equation}
717: so that the source terms in the constraints (\ref{constraints}) cancel to the
718: static wormhole values. Thus the wormhole remains static. The additional
719: negative-energy radiation has balanced the positive-energy radiation, leaving
720: the gravitational field unchanged.
721:
722: Alternatively, the wormhole may be maintained by beaming in additional
723: negative-energy radiation before or after the positive-energy pulse. We take
724: the same step pulse in $F_+$ (\ref{step}) and precede it with a compensating
725: pulse in the ghost field (Fig.~\ref{opr2}(a)):
726: \begin{eqnarray}\label{addG}
727: G_+ &=&
728: \left\{
729: \begin{array}{ll}
730: 2\lambda\,,~ &x^+ < -x_0 \\
731: \sqrt{4\lambda^2 + \Delta^2}\,,~ & -x_0 \leq x^+ < 0 \\
732: 2\lambda\,, & 0 \leq x^+
733: \end{array}\right.\,\\
734: G_- &=& -2\lambda\, .\nonumber
735: \end{eqnarray}
736: Again we require small pulse energy, $\epsilon<a/2$, to avoid a singularity.
737: The constraints (\ref{constraints}) become
738: \begin{eqnarray}
739: \partial_+ \partial_+ r_+ &=&
740: \left\{
741: \begin{array}{ll}
742: 4\lambda^2\,, &x^+ < -x_0 \\
743: 4\lambda^2 + \Delta^2\,,& -x_0 \leq x^+ < 0 \\
744: 4\lambda^2 - \Delta^2\,,& 0 \leq x^+ < x_0 \\
745: 4 \lambda^2\,, & x_0 \leq x^+
746: \end{array}\right. \\
747: \partial_- \partial_- r_- &=&4 \lambda^2\nonumber
748: \end{eqnarray}
749: which integrate to
750: \begin{widetext}
751: \begin{eqnarray}
752: r_+ &=&
753: \left\{
754: \begin{array}{ll}
755: 2\lambda^2 {x^+}^2 + \frac{a}{2}\,, &x^+ < -x_0 \\
756: 2\lambda^2 {x^+}^2 + \Delta^2 x^+ \left(\frac{1}{2}x^+ + x_0 \right)
757: -\frac{1}{2}\Delta^2 {x_0}^2 + \frac{a}{2} \, ,
758: ~ & -x_0 \leq x^+ < 0 \\
759: 2\lambda^2 {x^+}^2 - \Delta^2 x^+ \left(\frac{1}{2}x^+ - x_0 \right)
760: -\frac{1}{2}\Delta^2 {x_0}^2 + \frac{a}{2} \, ,
761: ~ & 0 \leq x^+ < x_0 \\
762: 2\lambda^2 {x^+}^2 + \frac{a}{2}\,, &x_0 \leq x^+
763: \end{array}\right.\\
764: r_- &=&2\lambda^2 {x^-}^2 + \frac{a}{2}\nonumber
765: \end{eqnarray}
766: and the solution follows as
767: \begin{eqnarray}
768: r=
769: \left\{
770: \begin{array}{ll}
771: a + 2\lambda^2 \left({x^+} - {x^-} \right)^2\,, & x^+ < -x_0 \\
772: a + 2\lambda^2 \left({x^+} - {x^-} \right)^2
773: + \frac{1}{2}\Delta^2 \left({x^+}^2 + 2x_0 x^+ - {x_0}^2 \right) \,,
774: ~ & -x_0 \leq x^+ < 0 \\
775: a + 2\lambda^2 \left({x^+} - {x^-} \right)^2
776: - \frac{1}{2}\Delta^2 \left({x^+}^2 - 2x_0 x^+ + {x_0}^2 \right) \,,
777: ~& 0 \leq x^+ < x_0 \\
778: a + 2\lambda^2 \left({x^+} - {x^-} \right)^2 \,, ~& x_0 \leq x^+.
779: \end{array}\right.
780: \end{eqnarray}
781: \end{widetext}
782: The locations of the trapping horizons are
783: \begin{eqnarray}
784: x^- &=&
785: \left\{
786: \begin{array}{ll}
787: x^+ \,, &x^+ < -x_0 \\
788: x^+ + \frac{\Delta^2}{4\lambda^2}\left( x^+ + x_0 \right) \,,
789: ~ & -x_0 \leq x^+ < 0 \\
790: x^+ - \frac{\Delta^2}{4\lambda^2}\left( x^+ - x_0 \right) \,,
791: ~& 0 \leq x^+ < x_0 \\
792: x^+ \,, ~& x_0 \leq x^+
793: \end{array}\right.\\
794: x^+ &=& x^- \nonumber
795: \end{eqnarray}
796: which again are straight line segments. The double horizon bifurcates when the
797: ghost pulse arrives, temporarily opening up a trapped region, but the two
798: horizons subsequently merge to form a static wormhole again
799: (Fig.~\ref{opr2}(b)). This is not unexpected, since the energy of the ghost
800: pulse, $\epsilon'=-\Delta^2{x_0}^2/4$, balances that of the other pulse:
801: \begin{equation}
802: \epsilon+\epsilon'=0.
803: \end{equation}
804: In fact, the final state is identical to the initial state in this symmetric
805: case. The wormhole can also be returned to a different static state, with
806: different throat radius, by less symmetric double pulses. The examples show
807: that there is no practical problem of fine-tuning the ghost field to keep the
808: wormhole static. An almost static wormhole is still traversible and can be
809: adjusted at any time to bring it closer to staticity, or to change its size.
810: This is essentially due to the neutral stability.
811:
812: \begin{figure}
813: \includegraphics[width=7cm,height=10cm,angle=0]{opr2.eps}
814: % \vspace{-40mm}
815: \caption{(a) The step pulse of positive-energy radiation is balanced by a
816: preceding pulse of negative-energy radiation.
817: (b) Conformal diagram: the wormhole returns to its original static state.}
818: \label{opr2}
819: \end{figure}
820:
821: \section{Conclusion}
822:
823: Space-time wormholes remain in the realms of science fiction and theoretical
824: physics. By the standards of either genre, they are not so far-fetched,
825: differing from experimentally established physics by only one step, the
826: existence of negative energy densities in sufficient concentrations. Here we
827: have not addressed this issue but, assuming a positive answer, have
828: investigated the behaviour of the resulting wormholes, evolving dynamically
829: according to field equations. In particular, we have found detailed answers to
830: the following practical questions. (i) How can one construct a traversible
831: wormhole? By bathing a black hole in exotic radiation. (ii) Is an operating
832: wormhole stable under the back-reaction of the transported matter? In this
833: case, neutrally stable. (iii) How can a wormhole be maintained indefinitely for
834: transport or signalling? By balance of positive and negative energy. (iv) What
835: happens if the negative-energy source fails? The wormhole collapses into a
836: black hole.
837:
838: This was mostly predicted by a general theory of wormhole dynamics \cite{wh},
839: but here we have given concrete examples, by virtue of the exact solubility of
840: the field equations of two-dimensional dilaton gravity. Despite the supposedly
841: unphysical nature of a ghost Klein-Gordon field, fears about instability,
842: runaway processes and naked singularities proved to be unfounded. More
843: realistic situations may differ in some respects, such as wormhole stability,
844: which may be affected by backscattering and will presumably depend on the
845: exotic matter model. However, the same principles and methods should apply,
846: such as energy balance and tracking of trapping horizons. Indeed, apart from
847: the inclusion of exotic matter or radiation, the methods are the same as those
848: used to analyze black-hole dynamics \cite{1st,bhd}. In particular, the explicit
849: examples of dynamic interconversion of black holes and wormholes should assuage
850: objections that they are fundamentally different objects. Rather, wormholes and
851: black holes have similar physics and a unified theory.
852:
853: \medskip\noindent
854: S.A.~Hayward was supported by Korea Research Foundation grant
855: KRF-2001-015-DP0095, S-W.~Kim by Korea Research Foundation grant
856: KRF-2000-041-D00128 and H.~Lee by KOSEF grant R01-2000-00015.
857:
858: \begin{references}
859: \bibitem{W}J A Wheeler, {Ann. Phys.} {\bf 2}, 604 (1957).
860: \bibitem{ER}A Einstein \& N Rosen, {Phys. Rev.} {\bf 48}, 73 (1935).
861: \bibitem{K}M D Kruskal, {Phys. Rev.} {\bf 116}, 1743 (1960).
862: \bibitem{MT}M S Morris \& K S Thorne, {Am. J. Phys.} {\bf 56}, 395 (1988).
863: \bibitem{V}M Visser, Lorentzian Wormholes: from Einstein to Hawking (AIP Press 1995).
864: \bibitem{TRA}D F Torres, G E Romero \& L A Anchordoqui, {Phys. Rev.} {\bf D58}, 123001 (1998).
865: \bibitem{wh} S A Hayward, {Int. J. Mod. Phys.} {\bf D8}, 373 (1999).
866: \bibitem{HPS}D Hochberg, A Popov \& S V Sushkov, {Phys. Rev. Lett.}
867: {\bf 78}, 2050 (1997).
868: \bibitem{KL} S-W Kim and H Lee, {Phys. Lett.} {\bf B458}, 245
869: (1999).
870: \bibitem{CGHS} C Callan, S Giddings, J Harvey, and A Strominger, Phys. Rev. D {\bf 45},
871: R1005 (1992).
872: \bibitem{cc}S A Hayward, {Class. Quantum Grav.} {\bf 10}, 985 (1993).
873: \bibitem{C}D Christodoulou, {Ann. Math.} {\bf 149}, 183 (1999).
874: \bibitem{HV1}D Hochberg \& M Visser, {Phys. Rev.} {\bf D58}, 044021 (1998).
875: \bibitem{HV2}D Hochberg \& M Visser, {Phys. Rev. Lett.} {\bf 81}, 746 (1998).
876: \bibitem{IH}D Ida \& S A Hayward, {Phys. Lett.} {\bf A260}, 175 (1999).
877: \bibitem{wd}S A Hayward, Wormhole dynamics (in preparation).
878: \bibitem{1st}S A Hayward, {Class. Quantum Grav.} {\bf 15}, 3147 (1998).
879: \bibitem{bhd}S A Hayward, {Phys. Rev.} {\bf D49}, 6467 (1994).
880: \end{references}
881:
882: \end{document}
883: