1: \documentclass[twocolumn,showpacs,preprintnumbers,amsmath,amssymb]{revtex4}

2: \renewcommand{\baselinestretch}{1}

3: \usepackage{graphicx}

4: \usepackage{epsfig}

5: \begin{document}

6: 

7: \title{Dynamics and instability of false vacuum bubbles}

8: 

9: \author{Anthony Aguirre \& Matthew C. Johnson}

10: \affiliation{

11: Department of Physics, University of California, Santa Cruz, 

12: California 95064, USA\\aguirre@scipp.ucsc.edu; mjohnson@physics.ucsc.edu}

13: 

14: \date{\today}

15: 

16: \begin{abstract}

17: This paper examines the classical dynamics of false vacuum regions

18: embedded in surrounding regions of true vacuum, in the thin-wall limit.

19: The dynamics of all generally relativistically allowed solutions -- most

20: but not all of which have been previously studied -- are derived,

21: enumerated, and interpreted. We comment on the relation of these solutions

22: to possible mechanisms whereby inflating regions may be spawned from

23: non-inflating ones. We then calculate the dynamics of first order

24: deviations from spherical symmetry, finding that many solutions are

25: unstable to such aspherical perturbations.  The parameter space in which

26: the perturbations on bound solutions inevitably become nonlinear is

27: mapped. This instability has consequences for the Farhi-Guth-Guven

28: mechanism for baby universe production via quantum tunneling.

29: \end{abstract}

30: 

31: \pacs{98.80.Hw, 98.80.Cq}

32: 

33: \maketitle

34: 

35: \section{Introduction}

36: 

37: Nearly two decades ago, a series of papers~\cite{GF87,FMP89,FGG90}

38: began to investigate the possibility of creating an inflationary

39: universe ``in a laboratory" -- that is, inside a surrounding region of

40: much lower vacuum energy.  The spacetime of such a ``bubble universe"

41: was modeled as a spherically symmetric de Sitter region (the false

42: vacuum) joined to a surrounding Schwarzschild geometry (the true

43: vacuum) by an infinitesimally thin domain wall.

44: 

45: These studies found that the inflating (false vacuum) region could not

46: avoid collapse unless either the null energy condition or cosmic

47: censorship were violated in the full spacetime~\cite{GF87}, but that

48: the creation of an enduring inflating region might be possible via

49: quantum tunneling~\cite{FMP89,FGG90}. In this picture, henceforth

50: denoted the Farhi-Guth-Guven (FGG) mechanism, a classically

51: constructable expanding bubble, which would classically re-collapse

52: (both from the inside and outside perspective~\cite{BGG87}), instead

53: tunnels to a new solution in which the inflating interior expands

54: forever, while an outside observer sees a black hole~\footnote{We

55: will include the means by which the constructable expanding bubble is

56: produced, be it in the lab or by some quantum or thermal fluctuation,

57: in the definition of the Farhi-Guth-Guven mechanism}. The probability

58: of this occurring can be calculated using the techniques of

59: semi-classical quantum gravity, and is extraordinarily small.

60: 

61: Though nearly-miraculous, this process has garnered new interest

62: recently, primarily because of evidence that our universe may have a

63: fundamental positive cosmological constant.  If so, it will

64: asymptotically approach everlasting equilibrium as de Sitter

65: spacetime.  Given eternity, even the most unlikely process -- such as

66: the creation of inflating bubble universes -- will eventually occur.

67: Taking this one step further, our observable universe could in fact be

68: such a bubble universe, which arose from equilibrium de Sitter space

69: and is currently returning to it.  This would realize the old idea of

70: Boltzmann that the universe is fundamentally in equilibrium, but that

71: extremely rare downward fluctuations in entropy periodically occur and

72: allow the transitory existence of non-equilibrium regions that see

73: entropy steadily increasing.

74: 

75: The classic problem with this idea was pointed out in its new context

76: by Dyson, Susskind, \& Kleban~\cite{DKS02}: if our observable universe

77: resulted from a downward entropy fluctuation that evolved with

78: increasing entropy to the present time, then it is vastly more likely

79: for this to have occurred by a fluctuation to our observable universe

80: ten minutes ago (replete with incoming photons and memories in our

81: brains to convince us that it is older) than by a fluctuation all the

82: way to the much lower entropy corresponding to

83: inflation~\footnote{This, in fact, connects to a more general question

84: of whether inflation can really be said to have ``general'' initial

85: conditions, given that it must be low entropy; see,

86: e.g.,~\cite{penrose,Hollands:2002yb,Carroll:2004pn}.}. Albrecht and

87: Sorbo~\cite{AS04}, however, argue that inflation might avoid this

88: problem by turning a tiny region of low entropy {\em density} into a

89: very large one. Their calculation shows that the FGG mechanism

90: requires a much smaller entropy fluctuation than directly creating the

91: large post-reheating region that will result from it, thus resolving

92: the paradox.

93: 

94: There may, however, be reasons to doubt that the creation of small

95: regions of false vacuum and subsequent tunneling are plausible

96: events. First, there is no regular instanton describing either the

97: nucleation of small regions of false vacuum or the gravitational

98: tunneling event; there {\em is} such an instanton describing tunneling

99: to false-vacuum, but over a huge region, larger than the true-vacuum

100: horizon (see, e.g.,~\cite{LW87,G94}). Second,

101: Banks~\cite{Banks:2002nm} has pointed out that it is hard to

102: understand the tunneling process holographically -- it appears that

103: the inflating region inside the bubble has many more states than the

104: black hole it is ``contained in" (see

105: also~\cite{Bousso:2004tv}). Third, Dutta \& Vachaspati~\cite{DV05}

106: have recently (and previously in~\cite{VT99}) argued that general

107: causality considerations preclude the formation of a small

108: false-vacuum region inside a large true-vacuum one.

109: 

110: Both in terms of the initial conditions for inflation, and also the

111: more general issue of what processes can lead to transitions between

112: vacuum states (as is important in understanding the string theory

113: ``landscape''), it is crucial to understand bubble universes, whether

114: they can form, and with what probabilities.  To this question the

115: present paper makes the following contributions: first, we organize

116: and interpret all of the thin-walled, spherically symmetric one bubble

117: spacetimes.  We then show that expanding false vacuum bubbles are

118: unstable to non-spherical perturbations; {\em if} these bubbles start

119: at a sufficiently small initial radius, then they inevitably become

120: nonlinearly aspherical before tunneling might occur in the FGG

121: mechanism. It is unclear in this case whether tunneling to an

122: inflationary universe inside the bubble can occur at all, or with what

123: probability.

124: 

125: The plan of the paper is as follows. In Sec.~\ref{intro}, the allowed

126: solutions to the junction conditions are enumerated, putting the

127: previous work into context. We provide a concise reference for all of

128: the possible spacetimes with one false-vacuum bubble and arbitrary

129: positive cosmological constant, then discuss the existing, and some new,

130: interpretations of these solutions. In Sec.~\ref{perts} we derive the

131: first order perturbation equations, and demonstrate the existence of

132: an instability in bound solutions. In Sec.~\ref{sec-results} we

133: integrate the equations to investigate the parameter space for which

134: non-linear perturbations are unavoidable, and we conclude in

135: Sec.~\ref{conclusions}.

136: 

137: \section{Junction Conditions  \label{intro}}

138: The dynamics of inflating regions has been discussed by a number of

139: authors~\cite{BGG87,BKT87,S86,APS89}\footnote{We follow the methods and

140: notations of these works, though they are not the only on the

141: subject.} using the junction condition formalism. These works study a

142: spherically symmetric region of false vacuum (high energy density) in

143: a surrounding region of true vacuum (lower energy density). The wall

144: separating the regions is assumed to be very thin compared to the

145: radius of the region of false vacuum. One can obtain the dynamics of

146: the wall by requiring metric continuity across the wall and then

147: solving Einstein's equations. Under spherical symmetry, the dynamics

148: of the problem then reduce to those of the bubble wall's radius.  This

149: radius is a gauge invariant quantity because it simply quantifies the

150: curvature of the bubble wall worldsheet, and any observer can measure

151: it by comparing the normal to the wall at two nearby points.

152: 

153: \subsection{Interior and Exterior Spacetimes}

154: 

155: Let $\Lambda_{+}$ be the cosmological constant in the true vacuum

156: region.  Then if $\Lambda_{+} > 0$, the region is Schwarzschild--de

157: Sitter spacetime with metric

158: \begin{equation}\label{gsds}

159: ds_{+}^2=-a_{\rm sds}dt^2 + a_{\rm sds}^{-1}dr^2 + r^2 d\Omega^2,

160: \end{equation}

161: \begin{equation}\label{defa}

162: a_{\rm sds}=1-\frac{2M}{r} - \frac{\Lambda_{+}}{3} r^2

163: \end{equation}

164: in the static foliation. Fixing $\Lambda_{+}$, there are then three

165: qualitatively different casual structures characterized by the value

166: of $M$ (see \cite{LR77}), due to the nature of the three roots of

167: $a_{\rm sds}(r).$

168: 

169: \begin{subequations}

170: For $3M<\Lambda_{+}^{-1/2}$, there are three distinct real roots of form:

171: \begin{equation}\label{rbh}

172: r_n=2 (\Lambda_{+})^{-1/2} \cos{\left(\frac{\theta}{3} +\frac{2\pi n}{3}\right)},

173: \end{equation}

174: where 

175: \begin{equation}

176: \cos \theta = -3M (\Lambda_{+})^{1/2},

177: \end{equation}

178: \end{subequations}

179: and  $\pi<\theta<3\pi/2$.

180: We can label them as 

181: $$r_{BH}\equiv r_0,\ \ r_{\rm neg}=r_1,\ \ r_c=r_2,$$ And the range of

182: $\theta$ means that they lie in the ranges

183: $r_{\rm neg}<0<2M<r_{BH}<3M<r_{c}$.

184: 

185: The two positive roots correspond to the black hole and cosmological

186: horizons. The conformal diagram for this spacetime is shown in

187: Fig.~\ref{sdsconf}. (See \cite{CP04} for a demonstration of the

188: explicit form of the metric in global coordinates). Surfaces of

189: constant coordinate time $t$ are drawn, with the circulating arrows

190: denoting the direction of increasing $t$. We will consider region I to

191: be the ``causal patch'' of a hypothetical observer (i.e., the region

192: lying in both the causal past and causal future of the observer's

193: world line) in what follows.

194: 

195: \begin{figure}

196: \centering

197: \includegraphics[width=8.6cm]{sdsconf.eps}

198: \caption{Conformal diagram of the Schwarzschild de

199: Sitter geometry for $3M<\Lambda_{+}$. \label{sdsconf}}

200: \end{figure}

201: 

202: For $3M=\Lambda_{+}^{-1/2}$, there are also three real roots: a double

203: positive root $r_h$ and a negative $r_{\rm neg}$, given by:

204: \begin{equation}

205: r_{h} = \Lambda_{+}^{-1/2}, \ \ r_{\rm neg}=-2 \Lambda_{+}^{-1/2}.

206: \end{equation} 

207: This mass is known as the Nariai mass, and in this spacetime there is

208: only one horizon at the positive root. The conformal diagram for this

209: spacetime is shown in Fig.~\ref{sdsconf2} \cite{LR77}. There

210: is also a time-reverse solution, starting at past null infinity and

211: ending at $r=0$. For $3M>\Lambda_{+}^{-1/2}$, there is one real negative

212: root, and therefore no horizons in the spacetime.  The conformal

213: diagram for this case is like figure \ref{sdsconf2}, but with the

214: horizon lines excised.

215: 

216: \begin{figure}

217: \centering

218: \includegraphics[width=6cm]{sdsconf2.eps}

219: \caption{Conformal diagram for the Schwarzschild de

220: Sitter geometry when $3M=\Lambda_{+}$. \label{sdsconf2}}

221: \end{figure}

222: 

223: Inside the false-vacuum region the spacetime is de Sitter space, with

224: metric

225: \begin{equation}

226:  ds_{-}^2 = -a_{\rm ds} dt^2 + a_{\rm ds}^{-1} dr^2 +r^2 d\Omega^2,

227: \end{equation}

228: \begin{equation}

229: a_{\rm ds}=1- \frac{\Lambda_{-}}{3}r^2

230: \end{equation} 

231: in the static foliation. Fig.~\ref{dsconf} shows the conformal diagram

232: for the de Sitter region. Again, surfaces of constant coordinate time

233: $t$ are shown, with the arrows denoting the direction of increasing

234: $t$. We consider region III to be the causal patch in which our

235: hypothetical false vacuum observer resides.

236: 

237: \begin{figure}

238: \centering

239: \includegraphics[width=3.5cm]{dsconf.eps}

240: \caption{Conformal diagram for the de Sitter geometry.\label{dsconf}}

241: \end{figure}

242: 

243: \subsection{Equation of Motion}\label{junctioneom}

244: The bubble wall worldsheet has metric:

245: \begin{equation}\label{wallmetric}

246: ds^2= - d\tau^2 + r(\tau)^2 d\Omega^2,

247: \end{equation}

248: where $\tau$ is the proper time in the frame of the wall, and

249: ($\theta, \phi$) are the usual angular variables.

250: 

251: The coordinates in the full $4D$ spacetime are chosen to be Gaussian

252: normal coordinates constructed in the neighborhood of the bubble wall

253: worldsheet. Three of the coordinates are $(\tau,\theta,\phi)$ on the

254: worldsheet, and the fourth, $\eta$, is defined as the proper distance

255: along a geodesic normal to the bubble worldsheet, with $\eta$

256: increasing in the direction of SdS (true vacuum).

257: 

258: The transformation from the static coordinate systems of SdS and dS to

259: the Gaussian normal system can be constructed in closed form using the

260: methods of \cite{FGG90}, and the full metric takes the form:

261: \begin{equation}

262: \label{eq-4dmetric}

263: ds^2 = g_{\tau\tau}(\tau,\eta) d\tau^2 + d\eta^2 + R(\tau,\eta)^2

264: d\Omega^2,

265: \end{equation}

266: where $\eta=0$ defines the wall and therefore

267: $g_{\tau\tau}(\tau,0)=-1$ and $R(\tau,0)=r(\tau)$.

268: 

269: The energy momentum tensor on the wall is:

270: \begin{equation}

271: \label{eq-emtens}

272: T^{\mu\nu}_{\rm wall} = -\sigma \gamma^{\mu\nu} \delta(\eta)

273: \end{equation}

274: where $\gamma^{\mu\nu}$ is the metric on the worldsheet of the wall

275: for $\mu=\nu=\tau,\theta,\phi$ and zero otherwise, and

276: $\sigma$ is the energy density of the wall.

277:  

278: Using the metric~\ref{eq-4dmetric} and the energy-momentum

279: tensor~\ref{eq-emtens} together with the contributions from the dS

280: interior and SdS exterior in Einstein's equations yields an equation

281: of motion for the bubble wall of~\cite{FGG90,BGG87} :

282: \begin{equation}

283:  \label{preisrael}

284: K^{i}_{j}(\eta_{+}) - K^{i}_{j}(\eta_{-}) = - 4 \pi \sigma r

285: \delta^{i}_{j},

286: \end{equation}

287: where $K^{i}_{j}(\eta_{\pm})$ is the extrinsic curvature tensor in SdS

288: and dS respectively. In the Gaussian normal coordinates, this takes

289: the form:

290: \begin{equation}

291: K_{ij}=\frac{1}{2} \frac{d}{d\eta}g_{ij}

292: \end{equation}

293: 

294: Evaluating this in metric~\ref{eq-4dmetric}, the $\theta\theta$ and

295: $\phi\phi$ components of Eq.~\ref{preisrael} reduce to:

296: \begin{equation}

297:  \label{israel}

298: \beta_{\rm ds} - \beta_{\rm sds} = 4 \pi \sigma r,

299: \end{equation}

300: with the definitions

301: \begin{equation}

302: \beta_{\rm ds} \equiv - a_{\rm ds} \frac{dt}{d\tau},\ \ \beta_{\rm sds} \equiv

303: a_{\rm sds} \frac{dt}{d\tau}.

304: \end{equation}

305: Here, $a$ is the metric coefficient in dS or SdS. The sign of $\beta$

306: is fixed by the trajectory because $dt/d\tau$ could potentially be

307: positive or negative (motion can be with or against the direction of

308: increasing coordinate time indicated in Fig.\ref{sdsconf} and

309: \ref{dsconf}).

310: 

311: A set of dimensionless coordinates can be defined, in which

312: Eq.~\ref{israel} can be written as the equation of motion of a

313: particle of unit mass in a one dimensional

314: potential~\cite{APS89}. Let:

315: 

316: \begin{equation} \label{ztor}

317: z=\left(\frac{L^2}{2M}\right)^{\frac{1}{3}}r,\ \ T = \frac{L^2}{2k} \tau,

318: \end{equation}

319: where M is the mass appearing in the SdS metric coefficient, and 

320: \begin{equation}

321: k=4\pi\sigma,

322: \end{equation}

323: \begin{equation}\label{Lsq}

324: L^2=\frac{1}{3}\left[\left| \left(\Lambda_{-} + \Lambda_{+} + 3k^2 \right)^2 - 4\Lambda_{+}\Lambda_{-}\right| \right]^{\frac{1}{2}}.

325: \end{equation}

326: With these definitions, Eq.~\ref{israel} becomes 

327: \begin{equation}\label{juncteom}

328: \left[\frac{dz}{d T }\right]^2=Q-V(z),

329: \end{equation}

330: where the potential $V(z)$ and energy $Q$ are 

331: \begin{equation}\label{potential}

332: V(z)=-\left[z^2+\frac{2Y}{z}+\frac{1}{z^4} \right],

333: \end{equation}

334: with

335: \begin{equation} \label{y}

336: Y=\frac{1}{3}\frac{\Lambda_{+}-\Lambda_{-}+3k^2}{L^2},

337: \end{equation}

338: and

339: \begin{equation} \label{Qtom}

340: Q=-\frac{4k^2}{\left(2M\right)^{\frac{2}{3}}L^{\frac{8}{3}}}.

341: \end{equation}

342: Note that a small negative $Q$ corresponds to a large mass, so that

343: even between $-1<Q<0$ the mass will vary by many orders of magnitude.

344: 

345: The parameter space allowed by the junction conditions is

346: characterized by the value of the cosmological constant inside and

347: outside the wall. For $0 \leq \Lambda_{+} < \Lambda_{-}$, we have $-1

348: \leq Y \leq 1$. The maximum $V_{\rm max}$ of the potential $V(z)$ then

349: satisfies $-2^{5/3}-2^{-4/3} \leq V_{max} \leq 0$. The potential

350: curves over the entire range of $Y$ are shown in Fig.\ref{potvsY}.

351: 

352: \begin{figure}

353: \centering

354: \includegraphics[width=8cm]{potvsY.eps}

355: \caption{The potential for various Y. \label{potvsY}}

356: \end{figure}

357: 

358: The interior and exterior cosmological constants can be expressed in

359: terms of $k^2$ as $\Lambda_{+}=Ak^2$ and $\Lambda_{-}=Bk^2$. With

360: these choices, the dynamics of the bubble wall are entirely determined

361: by $A$, $B$, and $Q$.

362: 

363: 

364: Let us now discuss some realistic values for the parameters in this

365: theory. The interior cosmological constant ($\Lambda_{-} =

366: M_{I}^{4}/M_{pl}^{4}$) and the bubble wall surface energy density ($k

367: = 4 \pi M_{I}^{3}/M_{pl}^{3}$) will be set by the scale of inflation

368: ($M_{I}$). The exterior cosmological constant ($\Lambda_{+} =

369: M_{\Lambda}^{4}/M_{pl}^{4}$) will be set by a scale

370: $M_{\Lambda}$. These yield

371: \begin{equation}\label{scaleAB}

372: A = \frac{M_{\Lambda}^{4} M_{\rm pl}^{2}}{(4 \pi)^2 M_{I}^{6}},\ \ 

373: B = \frac{ M_{\rm pl}^{2}}{(4 \pi)^2  M_{I}^{2}}.

374: \end{equation}

375: 

376: 

377: We will consider three representative energy scales, covering the

378: interesting range of energy scales for inflation. For weak scale

379: inflation ($100$ GeV), $k \simeq 4 \pi \times 10^{-51}$, $A \simeq 0$,

380: and $B \simeq 10^{32}$. For an inflation scale near the GUT scale

381: ($10^{14}$ GeV $\gg M_{\Lambda}$), we have $k \simeq 4 \pi \times

382: 10^{-15}$, $A \simeq 0$, and $B \simeq 10^7$. Near-Planck scale

383: inflation ($10^{17}$ GeV) yields $k \simeq 4 \pi \times 10^{-6}$, $A

384: \simeq 0$, and $B \simeq 63$. The most massive bound solution (that

385: which just reaches the top of the potential) is given by converting

386: from $Q$ to $M$ using Eq.~\ref{Qtom}. This maximal mass is very

387: different in each case, ranging from an ant-mass of $M_{\rm max}

388: \simeq 10^3 M_{\rm pl}\simeq 10^{-2}$ grams for $M_{I}=10^{17}$ GeV to

389: an Earth-mass of $M_{\rm max} \simeq 10^{33}M_{\rm pl} \simeq 10^{28}$

390: grams, for $M_{I}=100$ GeV.

391: 

392: \subsection{Allowed Solutions}\label{sols}

393: A bubble wall trajectory is characterized by $Q=$const., and there are 

394: three general types:

395: 

396: \begin{itemize}

397: \item Bound solutions with $Q < V_{\rm max}$. These solutions start at

398:   $z=0$, bounce off the potential wall and return to $z=0$.

399: \item Unbound solutions with $Q < V_{\rm max}$. These solutions

400: start at $z=\infty$, bounce off the potential wall and return to

401: $z=\infty$.

402: \item Monotonic solutions with $Q > V_{\rm max}$. These solutions

403: start out at $z=0$ and go to $z=\infty$, or execute the time-reversed motion. 

404: \end{itemize}

405: 

406: The qualitative features of a generic potential can be shown by

407: considering the four illustrative (but unrealistic; see above) cases of

408: ($A=9$, $B=15$) shown in Fig.~\ref{apspotential}, ($A=1$, $B=6$)

409: shown in Fig.~\ref{gfpotential}, ($A=1$, $B=2$) shown in Fig.~\ref{A1B2}, and ($A=2.9$, $B=3$) shown in Fig.~\ref{A2_9B3}. The important features are:

410: 

411: \begin{itemize}

412: \item 

413: As one follows a line of constant $Q$ in Fig.~\ref{apspotential}, \ref{gfpotential}, \ref{A1B2}, and \ref{A2_9B3} every intersection with the dashed line $Q_{\rm sds}$ (which is obtained by solving $a_{\rm sds} = 0$ for $Q$) 

414: represents a horizon crossing in the SdS spacetime (this could either

415: represent the black hole or cosmological horizons). 

416: \item Intersections with the dashed line $Q_{\rm ds}$ (which is obtained by solving $a_{\rm ds} = 0$ for $Q$) as one moves along a line of constant $Q$ represent the crossing of the interior dS horizon.

417: \item The vertical line on the right denotes the position at which

418: $\beta_{\rm ds}$ changes sign. $\beta_{\rm ds} > 0 $ if $t_{\rm ds}$ is

419: decreasing along the bubble wall trajectory and is negative if

420: $t_{\rm ds}$ is increasing. For there to be a $\beta_{\rm ds}$ sign change, $Y$ in Eq.~\ref{y} must be in the range $-1 \leq Y < 0$ \cite{APS89}, which yields the condition that $B > A+3$ for a sign change to occur.

421: \item The vertical dotted line on the left denotes the radius at which

422: $\beta_{\rm sds}$ changes sign. Recall that $\beta_{\rm sds} > 0$ if

423: $t_{\rm sds}$ is increasing along the bubble wall trajectory, and

424: $\beta_{\rm sds} < 0 $ if it is decreasing. For the parameters in

425: Fig.~\ref{apspotential} and \ref{A2_9B3}, or (one can show) whenever $B<3(A-1)$, the

426: sign change occurs to the right of the maximum of $V$.  When

427: $B>3(A-1)$ it occurs to the left~\footnote{Ref. \cite{APS89}

428: considered only the case where the sign change occurred to the left of

429: the maximum.}, as shown in Fig.~\ref{gfpotential} (note that this is the

430: interesting case where $\Lambda_{+} \ll \Lambda_{-}$) and \ref{A1B2}.

431: \end{itemize}

432: 

433: \begin{figure}

434: \centering

435: \includegraphics[width=8.6cm]{apspotential.eps}

436: \caption{Potential for $A=9$ and $B=15$. The two dashed lines labeled

437: $Q_{\rm sds}$ and $Q_s$ represent the exterior and interior horizon

438: crossings respectively. The vertical dotted lines denote the regions

439: in which $\beta_{\rm sds}$ and $ \beta_{\rm ds}$ are positive and

440: negative. Various trajectories are noted. \label{apspotential}}

441: \end{figure}

442: 

443: \begin{figure}

444: \centering

445: \includegraphics[width=8.6cm]{gfpotential.eps}

446: \caption{Potential for $A=1$ and $B=6$. For this choice of parameters,

447: the sign change in $\beta_{\rm sds}$ occurs to the left of the maximum

448: in the potential. Various trajectories are noted. \label{gfpotential}}

449: \end{figure}

450: 

451: \begin{figure}

452: \centering

453: \includegraphics[width=8.6cm]{A1B2.eps}

454: \caption{Potential for $A=1$ and $B=2$. For this choice of parameters,

455: the sign change in $\beta_{\rm sds}$ occurs to the left of the maximum

456: in the potential and there is no $\beta_{\rm ds}$ sign change. Various trajectories are noted. \label{A1B2}}

457: \end{figure}

458: 

459: \begin{figure}

460: \centering

461: \includegraphics[width=8.6cm]{A2_9B3.eps}

462: \caption{Potential for $A=2.9$ and $B=3$. For this choice of parameters,

463: the sign change in $\beta_{\rm sds}$ occurs to the right of the maximum

464: in the potential and there is no $\beta_{\rm ds}$ sign change. Various trajectories are noted. \label{A2_9B3}}

465: \end{figure}

466: 

467: 

468: The potential diagram contains all of the information needed to

469: determine the conformal structure of the allowed one-bubble

470: spacetimes. A complete set of the qualitatively different trajectories

471: for arbitrary interior and exterior positive definite cosmological

472: constants (with $\Lambda_{+} < \Lambda_{-}$) and $M\ge 0$ are denoted

473: in Fig.~\ref{apspotential}, \ref{gfpotential}, \ref{A1B2}, and \ref{A2_9B3}. Figures~\ref{confdiags1} and~\ref{confdiags2} display the conformal structure of these solutions \footnote{Many, but not all, of these solutions have appeared previously in the

474: literature~\cite{GF87,APS89,BKT87}}. The conformal diagrams in each

475: row are matched along the bubble wall (solid line with an arrow) and

476: the physical regions are shaded. For solutions with qualitatively similar SdS diagrams, the various options for the dS interior are listed. The naming scheme in

477: Fig.~\ref{apspotential}, \ref{gfpotential}, \ref{A1B2}, and \ref{A2_9B3} is chosen to reflect the structure of the conformal diagrams. The first numbers are the regions

478: of the SdS conformal diagram that the bubble wall passes through. The

479: numbers following the backslash are the regions of the dS conformal

480: diagram that the bubble wall passes through.

481: 

482: 

483: \begin{figure*}

484: \includegraphics[width=16.4cm]{sol1_3.eps}

485: \caption{Conformal diagrams for the allowed one-bubble spacetimes. The dS and SdS diagrams are matched across the bubble wall (line with

486: arrow), and the physical regions shaded. Solutions 3 and 4 share the same SdS diagram. Regions which

487: do not contain anti-trapped surfaces are shaded green (light), regions

488: which do are shaded blue (dark).

489: \label{confdiags1}}

490: \end{figure*}

491: 

492: \begin{figure*}

493: \includegraphics[width=17.4cm]{ST.eps}

494: \caption{Conformal diagrams for the allowed one-bubble spacetimes. The dS and SdS diagrams are matched across the bubble wall (line with arrow), and the physical regions shaded. For solutions with qualitatively similar SdS diagrams, the various options for the dS interior are listed.

495: \label{confdiags2}}

496: \end{figure*}

497: 

498: For example, consider the IV-I-II/III solution (solution 1 of Fig.~\ref{confdiags1}). These start

499: at at $r=z=0$, corresponding to the singularity in region IV of the

500: SdS conformal diagram and the $r=0$ surface in region III of the dS

501: conformal diagram. $\beta_{\rm sds}$ and $\beta_{\rm ds}$ are both

502: greater than zero over the entire trajectory. Therefore, the wall on

503: the SdS side must follow a path of increasing $t_{\rm sds}$, pushing

504: it into region I (crossing the black hole horizon). The wall on the dS

505: half must follow a path of decreasing coordinate time, and thus

506: remain in region III. The wall then reaches a turning point, falls

507: through the black hole event horizon, and ends up back at $r=z=0$ (the

508: singularity in region II of the SdS diagram, and the $r=0$ surface in

509: region III of the dS diagram). The construction of the other diagrams

510: in Fig.~\ref{confdiags1} and Fig.~\ref{confdiags2} proceeds similarly.

511: 

512: Based on the inward pressure gradient, the IV-I-II/III solution is

513: what one might expect the junction conditions to yield: a sphere of

514: false vacuum which expands and then contracts. Relativistic effects,

515: however, lead to the qualitatively different behavior exhibited by the

516: other solutions in Fig.~\ref{confdiags1} and Fig.~\ref{confdiags2}. For instance, the evolution

517: of the IV-III-II/III (solution 2) solution is qualitatively similar to the

518: IV-I-II/III solution, but is so massive that its evolution is always

519: hidden behind the black hole event horizon. The IV-I-II'/III-II (solutions

520: 3 and 4) solution is a bubble which has enough kinetic energy to escape

521: collapse by expanding through the cosmological horizon; observers

522: inside (or who travel inside from region I) the false vacuum region

523: will find themselves in an inflationary universe at late times. In the

524: time-reverse of this solution, a bubble implodes from infinity into

525: the black hole horizon, and the interior undergoes collapse.

526: 

527: Note that there are two options for the IV-I-II'/III-II solution. Solution 3 corresponds to the situation $A + 3 < B < 3 (A - 1)$, which contains both $\beta_{\rm sds}$ and $\beta_{\rm ds}$ sign changes (see Fig.~\ref{apspotential}). This causes the bubble wall to accelerate in the direction of the true vacuum from both the interior and exterior perspectives. Solution 4 corresponds to the situation $B < A + 3$, which does not contain a $\beta_{\rm ds}$ sign change (see Fig.~\ref{A2_9B3}). This causes the bubble wall to accelerate in the direction of the {\em false} vacuum from the interior perspective. The reason for the disparity is that solution 4 exists only when the interior and exterior cosmological constants are similar in magnitude, and so the wall tension can have greater influence on the dynamics.

528: 

529: The remaining solutions, shown in Fig.~\ref{confdiags2}, have interiors which approach an inflationary

530: universe at late times, but lie on the opposite side of the wormhole

531: in the exterior SdS spacetime. To an observer in region III of the SdS

532: diagram, the IV''-III-II''/IV-I-II (solutions 5) and

533: IV''-III-II''/IV-III-II solutions (solutions 6 and 7) would appear as a ``sky''

534: of false vacuum that encroaches from infinity, reaches a minimum

535: radius and then expands back out. It has been pointed out by Bousso

536: \cite{Bousso:2004tv} that at late times (in an asymptotically flat

537: spacetime) the wall trajectory according to the observer in region III

538: of the S(dS) diagram approaches that of a true vacuum

539: bubble~\cite{C77,CC77,CD80}. Amusingly, the SdS observer will think he

540: is in a true vacuum bubble surrounded by a large region of false

541: vacuum, while the dS observer will think she is in a false vacuum

542: bubble surrounded by a large region of true vacuum.

543: 

544: This symmetry between true and false vacuum bubbles is made manifest

545: in the analysis of the Coleman-De Luccia~\cite{CD80} instanton, which

546: describes the production of both true and false vacuum bubbles

547: \cite{LW87}.  These are zero energy solutions, and so we should look

548: for an $M=0$ unbound solution; this corresponds (via Eq.~\ref{Qtom})

549: to $Q\rightarrow -\infty$, and we see from Fig.~\ref{apspotential}, \ref{gfpotential}, \ref{A1B2}, and \ref{A2_9B3} that the IV''-III-II''/IV-I-II solution (solution 5) or the IV''-III-II''/IV-III-II solution (solution 6), depending on the values of $\Lambda_{+}$ and $\Lambda_{-}$, can be identified as the analytically continued false vacuum instanton.  The radius of the

550: bubble wall at the turning point is found by considering the limit as the

551: potential (Eq.~\ref{potential}) goes to $-\infty$, where on the right

552: (unbound) side of the potential hump the $z^2$ term dominates. Solving

553: for $r$ using Eq.~\ref{ztor}, we find the radius at turnaround to be

554: \begin{equation}\label{r_0inst}

555: r = 6k \left[\left|\left(\Lambda_{+} + \Lambda_{-} +3k^2 \right)^2 - 4\Lambda_{+}\Lambda_{-} \right| \right]^{-1/2},

556: \end{equation} 

557: which agrees with the previous literature \cite{BKT87} (see also

558: \cite{G94}). Since the Schwarzschild mass is zero, we are now matching

559: two pure de Sitter spacetimes across the bubble wall. The conformal

560: diagram for the exterior dS region (right) only contains the area

561: between the vertical dashed lines (which are now identified as $r=0$

562: surfaces) in solution 5 and 6 of Fig.~\ref{confdiags2}. The interior half (left)

563: of the diagram remains unchanged. It can be seen that at turnaround,

564: the bubble will be larger than both the interior horizon size and the

565: (exterior) horizon size of the region it has replaced.

566: 

567: In the IV-III-II''/III-II solution (solutions 12 and 13), the region of false

568: vacuum surrounding the observer in region III of the SdS diagram would

569: begin very small and then expand out of the cosmological horizon. This

570: solution will also have a time-reversed version in which the

571: surrounding region of false vacuum implodes. For the

572: IV''-III''-II''/IV-III-II solution (solutions 8 and 9), the ``sky'' of false

573: vacuum would forever reside outside of the horizon of a region III

574: observer. Finally, the N/III-II solution (solutions 10 and 11), which we will

575: define as solutions with mass greater than or equal to the Nariai

576: mass, will be an exploding (or imploding in the time-reversed

577: solution) bubble of false vacuum centered on $r=0$.

578: 

579: In a series of papers, Farhi et. al. \cite{GF87,FGG90} discussed the

580: application of the Penrose theorem \cite{P65} to the one-bubble

581: spacetimes discussed above. If the null energy condition (NEC) holds

582: (as it does for the postulated energy momentum tensor) and there

583: exists a non-compact Cauchy surface (as in the full SdS spacetime),

584: then the existence of a closed anti-trapped surface in the spacetime

585: implies the presence of an initial singularity. Since each point on

586: the conformal diagrams in Fig.~\ref{confdiags1} and Fig.~\ref{confdiags2} represents a

587: two-sphere, such an anti-trapped surface exists if the ingoing and

588: outgoing future-directed null rays both diverge. For example, the

589: 2-sphere represented by point $P_1$ shown in Fig.~\ref{confdiags1},

590: solution 1, is a closed anti-trapped surface. This can be seen by

591: following the null ray (null rays are denoted by the dashed lines in

592: Fig.~\ref{confdiags1}) from $r=0$ in region III of the dS diagram to

593: $P_1$ and noting that $r$ increases monotonically as $P_1$ is

594: approached (the null rays are diverging). But following the

595: future-directed null rays in the opposite direction from $r=0$ in

596: region IV of the SdS diagram, across the bubble wall, and into the

597: false vacuum region, shows that they also diverge. Thus, an initial

598: singularity is necessary for this solution to exist at and near

599: $P_1$. This spacetime also, however, contains regions without

600: anti-trapped surfaces. Following the future-directed null rays to

601: point $P_2$, for example, we see that the ingoing rays diverge, but

602: the outgoing rays converge. For examples, see solutions 1-4 in

603: Fig.~\ref{confdiags1}, where regions which contain anti-trapped surfaces

604: are shaded blue (dark) and the regions which do not are shaded green

605: (light).

606: 

607: If we cut the IV-I-II/III solution (solution 1) in the expanding phase on

608: a spacelike hypersurface at a time where the radius of the bubble wall

609: satisfies $r>r_{BH}$, then the spacetime would not necessarily contain

610: an initial singularity~\footnote{There are anti-trapped surfaces in

611: region II' of the SdS diagram, and there is a noncompact Cauchy

612: surface $C$ for it, so the Penrose theorem applies, but only indicates

613: that geodesics are incomplete in region II' because they reach its

614: edge (the past Cauchy horizon of $C$.) The region is thus extendible

615: (into regions I, III' and IV') rather than singular. Full dS has only

616: compact Cauchy surfaces so the theorem does not apply.}. We can remove

617: the initial singularity from the IV-I-II'/III-II (solution 3 and 4) solution as

618: well by cutting on the same surface. Both the IV-I-II/III (solution 1) and

619: IV-I-II'/III-II (solution 3 and 4) solutions are therefore classically

620: buildable. The IV-I-II'/III-II solution (solutions 3 and 4) is the only example

621: where it is possible to form an inflationary universe from classically

622: buildable initial conditions, but only exists when the interior and

623: exterior cosmological constant are almost equal ($B<3(A-1)$). This

624: solution might be of interest in understanding transitions between

625: nearly degenerate vacua, for example in the context of eternal

626: inflation.

627: 

628: Given the existence of a classically forbidden region in

629: Fig.~\ref{apspotential},~\ref{gfpotential},~\ref{A1B2}, and ~\ref{A2_9B3}, one might ask if

630: tunneling is allowed from one of the recollapsing solutions (solutions 1

631: and 2) to one of the expanding solutions (solutions 5-9 ) on the other

632: side of the potential hump. This event, shown in Fig.~\ref{tunnelst},

633: would constitute a violation of the NEC, and so the Penrose theorem

634: would no longer apply to the antitrapped surfaces that exist after the

635: tunneling event. Such a process is indeed apparently

636: allowed~\cite{FGG90,FMP89}, and would describe the quantum creation of

637: an inflationary universe from classically buildable initial conditions

638: (if the initial condition is the IV-I-II/III solution in solution 1).

639: 

640: This is a rather strange transition, however, as the unbound solution

641: is behind the wormhole in SdS (see Fig.~\ref{tunnelst}). An observer

642: in region I would see the bubble expand, reach its turning point, and

643: then disappear, only to be replaced by a black hole. An observer

644: inside the bubble would see the wall expanding away and -- just as it

645: is about to turn around and start collapsing -- instead disappear

646: behind the cosmological horizon. This observer will be inside an

647: inflationary universe, but forever disconnected from region I. If the

648: black hole in the SdS spacetime then evaporates, the baby universe

649: will become completely topologically disconnected.

650: 

651: 

652: \begin{figure}[h!]

653: \epsfig{file=dstunnel.eps,height=2.1cm}

654: \epsfig{file=tunnel.eps,height=2.1cm}

655: \caption{Tunneling spacetime \label{tunnelst}}

656: \end{figure}

657: 

658: Having considered one circumstance in which an NEC violation

659: precipitates the creation of an inflationary universe, one might ask

660: if there are others. Quantum fluctuations of a scalar field in de

661: Sitter space can violate the NEC, and so one might imagine that any of

662: the solutions that we have discussed which allow inflation inside the

663: bubble could be spontaneously created somewhere along their

664: trajectories. One example of this direct production of baby universes

665: is a fluctuation into one of the unbound solutions (solutions 5-9). Such

666: scenarios have been considered in the context of the stochastic

667: approach to baby universe production by Linde~\cite{L91} and in

668: reference to eternal inflation by Carroll and

669: Chen~\cite{Carroll:2004pn}. 

670: 

671: Another example of the direct fluctuation into an inflationary universe is the thermal decay of de Sitter vacua discussed by Garriga and Megevand~\cite{GM04} (Ref.~\cite{GHTW04} discusses a related mechanism). This process consists of the fluctuation from  empty de Sitter of a bubble in unstable equilibrium between expansion and collapse. This static solution is identified as the set of spacetimes which sit on top of the potentials in Fig.~\ref{apspotential}, \ref{gfpotential}, \ref{A1B2}, and \ref{A2_9B3}. The two possible configurations are shown in Fig.~\ref{thermalons}, where it can be seen that the bubble wall can lie on either side of the worm hole depending on the sign of $\beta_{\rm sds}$ at the top of the potential. It can be shown that the ``Nariai limit" in Ref.~\cite{GM04} corresponds to $B = 3 (A -1)$, where the $\beta_{\rm sds}$ sign change occurs at the max of the potential.

672: 

673: \begin{figure}[h!]

674: \includegraphics[height=4.2cm]{thermalons.eps}

675: \caption{Conformal diagrams for the thermal decay of de Sitter vacua. The upper diagram corresponds to $B < 3 (A-1)$, the lower diagram corresponds to $B > 3 (A-1)$. \label{thermalons}}

676: \end{figure}

677: 

678: With these considerations, there are at least three competing channels

679: for the formation of baby universes: direct production, the FGG

680: mechanism, and the zero-mass false vacuum instanton. It would be

681: desirable to develop a scheme to directly compare the relative

682: probabilities of each of these processes, and this is currently being explored.

683: 

684: Our full catalog of solutions is also interesting in regards to a

685: recent proposal~\cite{VT99,DV05} that false vacuum regions, assumed to

686: be larger than the interior horizon, must at all times be larger than

687: the exterior, true vacuum, horizon. The basis of this conjecture is

688: the condition that the divergence of a congruence of future directed

689: null geodesics (defined as $\theta$) must satisfy

690: \begin{equation}\label{dthetadT}

691: \frac{d\theta}{dT} \leq 0,

692: \end{equation}

693:  where $T$ is an affine parameter, if the NEC holds for all $T$. Null

694: rays in the dS and SdS spacetimes satisfy this inequality (in dS, the

695: inequality is exactly zero), but we should check that the junction

696: conditions do not violate it. One requirement imposed by

697: Eq.~\ref{dthetadT} is that the divergence of the null rays does not

698: increase at the position of the wall as they go from a true vacuum

699: region into a false vacuum one. Along any given null geodesic in the

700: bubble interior or exterior, the value of $r$ is either increasing or

701: decreasing monotonically as a function of $T$. We can therefore state

702: the condition Eq.~\ref{dthetadT} as: one cannot have a null ray along

703: which $dr/dT \leq 0$ outside the bubble and $dr/dT \geq 0$ inside the

704: bubble. Surveying the solutions in Fig.~\ref{confdiags1} and Fig.~\ref{confdiags2}, we see that this is indeed always true.

705: 

706: The authors of Ref.~\cite{VT99,DV05} intended to demonstrate that if

707: one requires the false vacuum region to be larger than the interior

708: horizon size at all times (so that inflation is unstoppable), it is

709: necessarily larger than the exterior horizon size. Although all of the

710: allowed one-bubble spacetimes satisfy the condition

711: Eq.~\ref{dthetadT}, there are only two examples in which all observers

712: agree that this requirement is met: the false vacuum instanton

713: (IV''-III-II''/IV-I-II solution (solution 5), with $M=0$) and the

714: IV-I-II'/III-II solution (solution 3 and 4, after turn-around). In every other

715: case (including the FGG spacetime in Fig.~\ref{tunnelst}), the

716: observers in region I of the SdS conformal diagram will see only a

717: {\em black hole} horizon sized volume replaced by the false vacuum

718: bubble. We therefore conjecture that if one requires the false vacuum

719: region to be larger than the interior horizon size at all times, then

720: it will replace a volume larger than the exterior horizon size

721: according to {\em only some} observers. If one relaxes this

722: requirement, then there are a diverse range of solutions which might

723: describe the spawning of an inflationary universe. For example, the

724: bubbles in solutions 3, 4 and 10-13 of Fig.~\ref{confdiags1} and~\ref{confdiags2} grow from an arbitrarily small size.

725:  

726: Having discussed the character of the various solutions, we turn now

727: to a potentially dangerous detail, which is particularly important for

728: the FGG mechanism: there exists a classical instability against

729: aspherical perturbations in the spherically symmetric solutions to the

730: junction conditions.

731: 

732: \section{Perturbations}\label{perts}

733: 

734: The solutions described in Sec.~\ref{sols} assume that the region of

735: false vacuum is spherically symmetric. The stability of these

736: solutions against aspherical perturbations has important consequences,

737: especially if one hopes to build plausible cosmologies. That there

738: might be an instability in domain walls was first discussed by Adams,

739: Freese and Widrow~\cite{AFW91}. The bubble wall can trade volume

740: energy for surface energy and wall kinetic energy locally as well as

741: globally, and so the bubble wall will become distorted if different

742: sections of the wall have different kinetic energies.  As long as the

743: local distortions of the wall remain small compared to the size of the

744: background solution's radius, this process can be formulated

745: quantitatively as perturbation theory around a background spherically

746: symmetric solution.

747: 

748: Previous authors~\cite{AFW91,GV91,G93} have considered perturbations

749: on expanding bubbles of true-vacuum~\cite{C77,CC77,CD80}, which have

750: zero total energy (surface, volume, and kinetic energies canceling),

751: and so can expand asymptotically. As was first pointed out by Garriga

752: and Vilenkin~\cite{GV91}, even though local observers on the bubble

753: wall see perturbations grow, external observers see them freeze out

754: because they do not grow faster than bubble radius.

755: 

756: The story is different for the bound (solutions 1 and 2) and unbound

757: (solutions 5-9) false vacuum bubbles: since they reach a turning point,

758: the perturbations have a chance to catch up to the bubble's expansion

759: and become nonlinear. This also presumably has implications for the thermal decay mechanism of Garriga and Megevand~\cite{GM04}, depending on the duration of time the bubble wall sits in unstable equilibrium between expansion and collapse (see discussion in Sec.~\ref{sols}). The remainder of this work will focus on the

760: instability of the bound IV-I-II/III solution (solution 1), since

761: physically plausible initial conditions may be clearly

762: formulated. There is no obvious set of initial conditions for the

763: perturbations on the unbound solutions, and so we simply observe that

764: the results we will obtain for the bound solutions apply qualitatively

765: here as well.

766: 

767: To simplify the problem, we assume that the full gravitational problem

768: described in the previous sections can be treated as motion of the bubble

769: wall in a fixed SdS background. This assumption must be validated (as

770: we do below), but we are mainly interested in the low-mass bound

771: solutions for which we might expect the gravitational contributions to

772: be small.  Assuming that a thin spherically symmetric bubble wall

773: separates an internal dS from an external SdS spacetime, we can employ

774: the action~\cite{G93,AFW91,GV91}:

775: 

776: \begin{equation}\label{matteraction}

777: S=-\sigma \int d^{3}\xi \sqrt{-\gamma}+\epsilon \int d^{4}x \sqrt{-g}, 

778: \end{equation}

779: where $\sigma$ is the surface energy density on the bubble wall,

780: $\gamma_{ab}$ ($a,b=1,2,3$) is the metric on the worldsheet of the

781: bubble wall, $\epsilon$ is the difference in volume energy density on

782: either side of the bubble wall:

783: \begin{equation}

784: \epsilon = \frac{\Lambda_{+}-\Lambda_{-}}{8 \pi},

785: \end{equation}

786:  and $g_{\alpha \beta}$ is the metric of the background spacetime.

787: 

788: \subsection{Wall Equation of Motion}\label{walleom}

789: The equation of motion resulting from Eq.~\ref{matteraction} is~\cite{G93}:

790: \begin{equation}\label{simpeom}

791: g^{a b}K_{a b}=-\frac{\epsilon}{\sigma},

792: \end{equation} 

793: where $K_{a b}$ is the extrinsic curvature tensor of the worldsheet of

794: the bubble wall,

795: \begin{equation}

796: K_{a b}=-\partial_{a}x^{\mu} \partial_{b}x^{\nu} D_{\nu}n_{\mu}, 

797: \end{equation} 

798: and $D_{\nu}$ is the covariant derivative and $n_{\mu}$ is the unit

799: normal to the bubble wall worldsheet.

800: 

801: We will use the static foliation of the SdS spacetime (see

802: Eq.~\ref{gsds}) as the coordinates $x^\mu$ for the background

803: spacetime. The world sheet is given coordinates $(\tau,\theta,\phi)$

804: as in Eq.~\ref{wallmetric}, and has metric:

805: \begin{equation}

806: \gamma_{a b}=g_{\mu \nu}\partial_{a}x^{\mu}\partial_{b}x^{\nu},

807: \end{equation}

808: with the gauge freedom in choosing $\tau$ fixed by

809: \begin{equation}

810: \frac{dt}{d\tau}\equiv t'=\frac{\sqrt{a+r'^2}}{a},

811: \end{equation}

812: so that $\gamma_{\tau\tau}=-1$. Here and henceforth primes will

813: denote derivatives with respect to $\tau$. The other non-zero

814: components of $\gamma_{a b}$ are $\gamma_{\theta \theta} = r^2$ and

815: $\gamma_{\phi \phi} = r^2 \sin^2 \theta$.

816: 

817: The first task at hand is to find the worldsheet's unit normal, which

818: by spherical symmetry has only $r$ and $t$ components.  Requiring

819: orthogonality to the worldsheet ($g_{\mu\nu} n^{\nu} \partial_{a}

820: x^{\mu}=0$) and unit norm ($g_{\mu\nu}n^{\mu}n^{\nu}=1$) yields its

821: components:

822: \begin{subequations}\label{sdsn}

823: \begin{equation}

824: n_{t}=-r',\ \ n_{r}=t'.

825: \end{equation}

826: \end{subequations}

827: The components of $K_{ab}$ are given by

828: \begin{subequations}\label{kab}

829: \begin{equation}

830: K_{\tau\tau} = \left[ r'' + \frac{1}{2}\frac{da}{dr}\right] (a + r'^2)^{-1/2},

831: \end{equation}

832: \begin{equation}

833: K_{\phi \phi} = -r a t' \sin^2 \theta = K_{\theta \theta} \sin^2 \theta.

834: \end{equation}

835: \end{subequations}

836: 

837: Substituting Eq.~\ref{kab} into Eq.~\ref{simpeom} gives the equation of motion for the bubble wall:

838: \begin{equation}\label{sdseom}

839: r''=\frac{\epsilon}{\sigma} \sqrt{a+r'^2} - \frac{2}{r} (a+r'^2) - \frac{1}{2} \frac{da}{dr}.

840: \end{equation}

841:  

842: Eq.~\ref{potential} supplies the velocity of the bubble at some position along its trajectory

843: \begin{equation}\label{v_0}

844: z'=[Q-V(z_{0})]^{1/2}.

845: \end{equation}

846: Choosing this boundary condition is effectively restricting ourselves

847: to the IV-I-II/III (solution 1) or IV-III-II/III (solution 2) solutions. Since

848: the solutions to Eq.~\ref{sdseom} approximate the dynamics of the

849: junction condition problem, we should parametrize by $A$, $B$, and

850: $Q$. This can be done by using the conversions defined in

851: Sec.~\ref{junctioneom}, and gives:

852: \begin{eqnarray}

853: \label{dimwall}

854: z'' &=& -\frac{3(B-A)}{c}\sqrt{a(-Q)+z'^2} \\ \nonumber

855: &-& \frac{2}{z} (a(-Q)+z'^2) - \frac{(-Q)}{2}\frac{da}{dz},

856: \end{eqnarray}

857: where $a$ is written in terms of $z$ as

858: \begin{equation}

859: a=1-\frac{12}{c z (-Q)} - \frac{12 A}{c^2 (-Q)} z^2,

860: \end{equation}

861: and 

862: \begin{equation}

863: c \equiv \left[|(A+B+3)^{2}-4AB|\right]^{\frac{1}{2}}.

864: \end{equation}

865: 

866: To justify the use of the simplified dynamics described above,

867: Eq.~\ref{dimwall} was numerically integrated, and the position of the

868: turning point compared to the corresponding point on the full junction

869: condition potential. Over the range of $Q$ corresponding to the bound

870: solutions, we find excellent quantitative agreement (well within 1\%)

871: between the turning points of the solutions to Eq.~\ref{dimwall} and

872: the junction condition potential. This was repeated with equally good

873: results for the weak, GUT, and Planck scale potentials and also for

874: various initial positions between the black hole radius and the

875: potential wall (turning point). This shows that to zeroth order,

876: dynamics as motion in a background is valid, and strongly suggests

877: that it will be at higher orders as well.

878: 

879: \subsection{Perturbations \label{perturbations}}

880: 

881: We are now in a position to discuss the first-order perturbations on

882: the spherically-symmetric background solutions discussed in

883: Sec.~\ref{walleom}. Physical perturbations are normal to the

884: worldsheet of the (background) bubble wall, and can be described by

885: scalar field $\phi(x)$ by taking the position of the perturbed

886: worldsheet to be

887: \begin{equation}

888: \bar{x}^{\mu}=x^{\mu} + \phi(x) n^{\mu},

889: \end{equation}

890: where $x^{\mu}$ is the spherically symmetric solution, and $n^{\mu}$

891: is the unit normal to the worldsheet. It is assumed that $\phi$ is

892: much smaller than the bubble wall radius, so that a perturbative

893: analysis can be made.

894:  

895:  The equation of motion for the perturbation field $\phi(x)$ in a

896: curved spacetime background can be derived from the action Eq.~\ref{matteraction} after expanding to second order in $\phi(x)$

897: \cite{G93}

898: \begin{equation} \label{meom}

899: \bigtriangleup \phi - \left[-R_{\mu \nu} h^{\mu \nu} + R^{(3)} - \frac{\epsilon^2}{\sigma^2} \right] \phi = 0,

900: \end{equation} 

901: where 

902: \begin{equation}

903: \bigtriangleup \phi=\frac{1}{\sqrt{-\gamma}} \partial_{a} \left(\sqrt{-\gamma} \gamma^{a b} \partial_{b} \phi \right),

904: \end{equation}

905: and $h^{\mu \nu}$ is 

906: \begin{equation}

907: h^{\mu\nu}=g^{\mu\nu}-n^{\mu}n^{\nu}.

908: \end{equation}

909: 

910: To solve the equation of motion, we can decompose $\phi(x)$ into spherical harmonics

911: \begin{equation}\label{phimodes}

912: \phi(x) = \sum_{l,m} \phi_{lm}(\tau) Y_{lm}(\theta,\phi),  

913: \end{equation}

914: and separate variables to get an equation for $\phi_{lm}(\tau)$. The

915: geometrical factors in Eq.~\ref{meom} become dependent on

916: $\theta$ or $\phi$ only at second order, so we will always be

917: able to make this decomposition.

918: 

919: $\bigtriangleup \phi$ is then given by:

920: \begin{equation}\label{triangle}

921: \bigtriangleup \phi_{lm} = -\left(\partial_{\tau}^{2} + \frac{2 r'}{r} \partial_{\tau} + \frac{l(l+1)}{r^2} \right) \phi_{lm}.

922: \end{equation}

923: The components of $h^{\mu\nu}$ are:

924: \begin{subequations}\label{h}

925: \begin{eqnarray}

926: && h^{tt}=-\frac{a+r'^2}{a^2},\ \  h^{rr}=-r'^2, \\

927: && h^{\theta\theta}=h^{\phi\phi} \sin^2\theta =\frac{1}{r^2}.

928: \end{eqnarray}

929: \end{subequations}

930: The components of the Ricci tensor are given by:

931: \begin{subequations}\label{Rt}

932: \begin{equation}

933: R_{tt}=\frac{a}{2}\partial_{r}^2 a + \frac{a}{r} \partial_{r} a= - a^2 R_{rr},

934: \end{equation}

935: \begin{equation}

936: R_{\phi\phi} = R_{\theta\theta} \sin^2 \theta = (1 - a - r \partial_{r} a).

937: \end{equation}

938: \end{subequations}

939: Contracting equations \ref{h} and \ref{Rt} gives:

940: \begin{equation}\label{hR}

941: R_{\mu\nu}h^{\mu\nu}=\frac{2(1-a)}{r^2} - \frac{3 \partial_{r}a}{r} - \frac{\partial_{r}^2 a}{2}=3 \Lambda_{+}.

942: \end{equation}

943: The Ricci scalar on the world sheet is

944: \begin{equation}

945: \label{R}

946: R^{(3)}=\frac{2}{r^2} (1 + r'^2 + 2r r''),

947: \end{equation}

948: where $r''$ is given by Eq.~\ref{sdseom}.

949: 

950: After substituting Eq.~\ref{triangle}, Eq.~\ref{hR}, and Eq.~\ref{R} into

951: Eq.~\ref{meom}, the equation of motion for $\phi_{lm}(\tau)$ is

952: \begin{widetext}

953: \begin{equation}

954: \label{sdsddphi}

955: \phi_{lm}''=\left[\frac{\epsilon^2}{\sigma^2} - \frac{4

956: \epsilon}{\sigma r} \left(a+r'^2 \right)^{1/2} + 3 \Lambda_{+} +

957: \frac{2}{r}\frac{da}{dr} + \frac{6 r'^2}{r^2} - \frac{2(1-4a)}{r^2} -

958: \frac{l(l+1)}{r^2} \right] \phi_{lm} - \frac{2r'\phi_{lm'}}{r}.

959: \end{equation}

960: In terms of the dimensionless variables of the junction condition

961: problem this reads:

962: \begin{eqnarray}

963: \label{sdsdimeom}

964: &&\Phi_{lm}'' = -\frac{2 z'}{z}\Phi_{lm}' \\ &&+ \left\{\frac{108

965: A}{c^2} + \frac{9 (A-B)^2}{c^2} + \frac{12 (B-A)}{c

966: z} \left( a(-Q)+z'^2 \right)^{1/2} + \frac{2 (-Q)}{z^2}(4a-1) + \frac{6 z'^2}{z^2} +

967: \frac{2 (-Q)}{z} \frac{da}{dz} - \frac{l(l+1)(-Q)}{z^2}

968: \right\}\Phi_{lm}, \nonumber

969: \end{eqnarray}

970: \end{widetext}

971: where $\Phi$ is the dimensionless perturbation field defined similarly

972: to $z$ (see Eq.~\ref{ztor}). The first term acts as a (anti)drag on

973: (shrinking) growing perturbations. The last term in this equation is

974: always negative, acting as a restoring force. Perturbations will grow

975: when the other terms (which are positive over most of the trajectory

976: in the expanding phase) in this equation dominate. Further, the last

977: term indicates that lower $l$ modes will experience the largest

978: growth. The full details of the solutions, however, require a

979: numerical approach, to which we now turn.

980: 

981: \section{Application to the Farhi-Guth-Guven mechanism}

982: \label{sec-results}

983: 

984: The possibility of creating an inflating false-vacuum region via

985: quantum tunneling (described in Sec.~\ref{sols}) has been investigated

986: only under the assumption of spherical symmetry, and this would be

987: grossly violated if perturbations on the bubble wall become

988: nonlinear. In this section, we investigate the circumstances under

989: which this is the case. The two basic questions at issue are: first,

990: when do perturbations go nonlinear for some given set of initial

991: perturbations, and second, what initial perturbations can be expected.

992: 

993: \subsection{Dynamics of the Perturbation Field}\label{dynamics}

994: 

995: Let us begin with the first issue. Since Eq.~\ref{sdsdimeom} is a

996: second order ODE, it can be decomposed into the sum of two linearly

997: independent solutions

998: \begin{eqnarray}\label{levolve}

999: \Phi_{\rm lm}(T)&=&\Phi_{\rm lm}(T=0) f(l,z_{0},Q,T) \nonumber \\

1000:  && + \Phi_{\rm lm}'(T=0) g(l,z_{0},Q,T).

1001: \end{eqnarray}

1002: The functions $f(l,z_{0},Q,T)$ and $g(l,z_{0},Q,T)$ can be

1003: found numerically by alternately setting $\Phi_{\rm lm}(T=0)$ and

1004: ${\Phi}_{\rm lm}'(T=0)$ to zero, then evolving the coupled

1005: Eq.~\ref{sdsdimeom} and Eq.~\ref{dimwall} numerically for a time

1006: $T$ with initial conditions for $Q$, $z_{0}$, and $l$. If the

1007: bubble is to tunnel, it will do so at the time $T_{\rm max}$, when

1008: the bubble reaches its maximum radius and begins to re-collapse.

1009: Given $f$ and $g$ at time $T_{\rm max}$, the size of the

1010: perturbations at the turning point for any $z_{0}$, $Q$, $l$,

1011: $\Phi_{\rm lm}(T=0)$, and $\Phi_{\rm lm}'(T=0)$ can be

1012: determined.  An RK4 algorithm with adaptive step size was used to

1013: solve for $f$ and $g$, with numerical errors well

1014: within the $1\%$ level.

1015: 

1016: The results of this analysis for $l=1$ and for the low (weak) and

1017: intermediate (GUT) inflation scales discussed below Eq.~\ref{scaleAB}

1018: are shown in Fig.~\ref{fg}. The solid lines show contours of constant (log)

1019: amplification factor $f$ (left) and $g$ (right) versus the bubble

1020: starting radius $z_0$ and mass parameter $Q$, with bubble mass

1021: increasing toward the top. The shaded regions indicate regions which

1022: we have disallowed as bubble starting radii because the bubble would

1023: not be classically buildable for $r<r_{\rm BH}$ (marked as $Q > Q_{\rm

1024:   BH}$), or the bubble is in the forbidden region $Q < V(z)$ of the

1025: effective 1D equation of motion Eq.~\ref{juncteom}, or the bubble

1026: would be too small to be treated classically. We choose the latter

1027: radius as fifty times the Compton wavelength $z_{\rm compton}$ of a

1028: piece of the bubble wall~\footnote{The mass of a piece of wall of

1029:   scale $s$ is $M \simeq s^2 \sigma$, where $\sigma$ is the wall

1030:   surface energy density; the Compton wavelength is then found by

1031:   setting $M=s^{-1}$, yielding $s = z_{\rm compton} \simeq

1032:   \sigma^{-1/3}$.}. The choice of fifty Compton wavelengths is rather

1033: arbitrary; the effect of a larger bound would be to exclude more of

1034: the parameter space in Fig.~\ref{fg}.  This (unshaded) parameter space

1035: includes all classical initial conditions which could be set up by the

1036: observer in region I of the SdS conformal diagram.

1037: 

1038: \begin{figure*}

1039: \includegraphics[width=8.6cm]{f_l1_gut.eps}

1040: \includegraphics[width=8.6cm]{g_l1_gut.eps}

1041: \includegraphics[width=8.6cm]{f_l1_weak.eps}

1042: \includegraphics[width=8.6cm]{g_l1_weak.eps}

1043: \caption{(Color online) Contour plot of $Log_{10}[f(l=1,z_{0},Q,T_{\rm max})]$ (left) and  $Log_{10}[g(l=1,z_{0},Q,T_{\rm max})]$ (right) for $M_I=10^{14}$ GeV (top) and $M_I=100$ GeV (bottom). \label{fg}}

1044: \end{figure*}

1045: 

1046: It can be seen in Fig.~\ref{fg} that the growth of the perturbations

1047: is in general larger for higher-mass bubbles (smaller $|Q|$, larger

1048: $-\log_{10}(-Q)$). The lower the inflation scale, the closer to zero

1049: the peak in the potential function becomes, and the smaller $|Q|$

1050: (higher mass) bubbles are allowed, so at low inflation scales $f$ and

1051: $g$ can be very large. Growth for the Planck-scale inflation bubbles

1052: is very small, with $f$ of order 10 and $g$ of order 1, and is not

1053: plotted.

1054: 

1055: The enhanced growth at small $|Q|$ is due to the suppression of the

1056: term in Eq.~\ref{sdsdimeom} proportional to $l(l+1)(-Q)$, which always

1057: acts to stabilize the perturbations. Another consequence of this

1058: suppression is that the range in $l$ over which solutions are unstable

1059: depends on $Q$; as a general rule of thumb, approximately a few times

1060: $(-Q)^{-1/2}$ $l$-modes are unstable (note that this is unlike the

1061: case true vacuum bubbles, for which only the $l=0,1$ modes are

1062: unstable).  An example of the $f$ function for $Q=-10^{-4}$ with the

1063: intermediate (GUT) inflation scale parameters is shown in

1064: Fig.~\ref{usj}. The $f$ functions for very large $l$ modes are stable

1065: and approach sinusoids with amplitudes less than one (see the inset of

1066: Fig.~\ref{usj}), meaning that the perturbations are never larger than

1067: their initial size.

1068: 

1069: \begin{figure}

1070: \includegraphics[width=8.6cm]{unstable.eps}

1071: \caption{$f(l,z_{0}=.5,Q=-10^{-4},T)$ for various $l$. The inset

1072:   shows the oscillatory behavior of $f$ for large $l$. \label{usj}}

1073: \end{figure}

1074: 

1075: \subsection{Initial Conditions and Evolution to the Turning Point}

1076: 

1077: Having fully characterized the growth of the perturbations, we now

1078: require an estimate for their initial values when the bubble is

1079: formed.  There is no reason to expect that a region of false vacuum

1080: will fluctuate into existence with anything near spherical symmetry,

1081: nor is it likely to have thin walls (there is no instanton or other

1082: mechanism to enforce these symmetries). Since low-$l$ (relative to

1083: $(-Q)^{-1/2}$) modes are unstable, an initially aspherical bubble will

1084: only become more aspherical; this is in marked contrast to true vacuum

1085: bubbles, which both start spherical, and tend to become more spherical

1086: as they expand.

1087: 

1088: Suppose, however, that we consider the best-case scenario in which a

1089: bubble {\em is}, by chance or design, spherically symmetric.  It will

1090: nevertheless inevitably be dressed with zero-point quantum

1091: fluctuations of the perturbation field. We may then check whether

1092: these fluctuations alone, considered as initial values for the

1093: perturbations of a bubble starting with a given $Q$ and $z_{0}$,

1094: suffice to make the bubble nonlinearly aspherical by turnaround.

1095: 

1096: We assume that the ensemble average of the quantum fluctuations at the

1097: time of nucleation is zero; but the ensemble average of the square of

1098: the field (the space-like two-point function $\langle

1099: \Phi(\theta,\phi) \Phi(\tilde{\theta},\tilde{\phi})\rangle \equiv

1100: \langle \Phi \tilde{\Phi} \rangle$) will not generally vanish.  We can

1101: write the mode functions (Eq.~\ref{phimodes}) in terms of it as:

1102: \begin{eqnarray}\label{philm2}

1103: \langle \Phi_{\rm lm}^{2} \rangle & = & \\

1104: && \int d\Omega d\tilde{\Omega} \nonumber

1105:  \langle \Phi \tilde{\Phi} \rangle  Y_{\rm lm}(\theta,\phi) Y_{\rm lm}^{*}(\tilde{\theta},\tilde{\phi}).

1106: \end{eqnarray}

1107: By spherical symmetry, the two-point function can be written as a function of the angular separation $\Psi$ between $(\theta,\phi)$ and

1108: $(\tilde{\theta},\tilde{\phi})$, and decomposed into Legendre polynomials:

1109: \begin{equation}

1110: \langle \Phi \tilde{\Phi} \rangle = \sum_{l} C_{l} P_{l}(\cos \Psi). 

1111: \end{equation}

1112: Using the addition theorem for spherical harmonics, we can write this as

1113: \begin{equation}

1114: \langle  \Phi \tilde{\Phi} \rangle  = \sum_{\rm l',m'} \frac{4 \pi}{2l'+1} C_{l

1115: '} Y_{\rm l'm'}^{*}(\theta,\phi) Y_{\rm l'm'}(\tilde{\theta},\tilde{\phi}).

1116: \end{equation}

1117: Substituting this into Eq.~\ref{philm2} and using the orthogonality of the

1118: spherical harmonics yields the relation:

1119: \begin{equation}\label{initcond}

1120: \langle \Phi_{\rm lm}^2\rangle  = \frac{4 \pi C_{l}}{2l+1}.

1121: \end{equation}

1122: 

1123: Given some space-like two point function at the time the bubble is

1124: nucleated, we can obtain the $C_{l}$ from

1125: \begin{equation}\label{Cl}

1126: C_{l} = \frac{2l+1}{4 \pi} \int_{-1}^{1} d(\cos \Psi) \langle \Phi \tilde{\Phi}\rangle  P_{l}(\cos \Psi)

1127: \end{equation}

1128: and therefore set the typical initial amplitudes of the mode functions

1129: as the r.m.s. value $\langle \Phi_{\rm lm}^2\rangle ^{1/2}$ from

1130: Eq.~\ref{initcond}.  The velocity field can be decomposed into

1131: spherical harmonics just as $\Phi$ was, and the analysis performed

1132: above carries over exactly. The typical initial size of the velocity

1133: mode functions is then given by

1134: \begin{equation}\label{initcondg}

1135: \langle \Phi_{\rm lm}'^2\rangle  = \frac{4 \pi A_{l}}{2l+1}.

1136: \end{equation}

1137: with 

1138: \begin{equation}\label{Al}

1139: A_{l} = \frac{2l+1}{4 \pi} \int_{-1}^{1} d(\cos \Psi) \langle \Phi' \tilde{\Phi}'\rangle  P_{l}(\cos \Psi).

1140: \end{equation}

1141: 

1142: The initial amplitudes in Eq.~\ref{initcond} and Eq.~\ref{initcondg}

1143: can now be evolved to the turning point, and the mode functions

1144: re-summed. The ensemble average of the r.m.s. fluctuations in $\Phi$

1145: at any time at a given point will then be:

1146: \begin{eqnarray}

1147: \label{Phiavg}

1148: \langle {\Phi(T)}^2\rangle  &=&  \sum_{l} \frac{2l+1}{4 \pi} \langle {\Phi_{\rm lm}(T)}^2 \rangle \\ \nonumber

1149: &=&  \sum_{l} \left[ C_{l}^{1/2} f(l,z_0,q,T) + A_{l}^{1/2} g(l,z_0,q,T) \right]^2, 

1150: \end{eqnarray}

1151: which can be evaluated at $T=T_{\rm max}$.

1152: 

1153: A full model of the two-point functions $\langle\Phi\tilde\Phi\rangle$

1154: and $\langle\Phi'\tilde\Phi'\rangle$ would involve quantizing the mode

1155: functions on the curved spacetime of the bubble wall worldsheet, which

1156: has a metric depending on $z(T)$.  Further, to treat large

1157: fluctuations, we would need to include non-linear terms in the

1158: equation of motion. The exact two-point function is therefore a rather

1159: formidable object to compute. As a simplified model, we will employ

1160: the two-point functions of a massless scalar field in flat spacetime,

1161: and replace the spatial distance $r$ with the distance along the

1162: bubble wall $r_0\Psi$. This massless scalar corresponds to the

1163: perturbations on a flat wall separating domains of equal energy

1164: density in Minkowski space \cite{GVQ91}. Corrections to this picture

1165: in the presence of curvature should be small over small regions of the

1166: bubble wall. We are also neglecting the large difference in energy

1167: densities across the bubble wall, which will give the field a

1168: (negative) mass to first order. The apparent divergence of the

1169: correlator due to this negative mass will be rendered finite by the

1170: non-linear terms which must be introduced to discuss large

1171: fluctuations. In light of all these difficulties, and several more

1172: approximations we will make below, this should be considered as a

1173: first, rough estimate of the amplitude of the quantum fluctuations on

1174: the bubble at the time of nucleation.

1175: 

1176: The space-like two-point function in Minkowski space at large separations is given by

1177: \begin{equation}\label{philarger}

1178: \langle \phi({\bf x}) \phi({\bf y}) \rangle =\frac{\sigma^{-1}}{4 \pi r}, 

1179: \end{equation}

1180: where $r \equiv |\bf{x}-\bf{y}|$. As in the work of Garriga and Vilenkin \cite{GVQ91}, we introduce a smeared field operator to obtain a well-defined answer at close separations

1181: \begin{equation}

1182: \phi_{s} \equiv \frac{1}{\pi s^2} \int_{|\bf{y} - \bf{x}|< s} d^2 y \phi(\bf{y}), 

1183: \end{equation}

1184: where $s$ is a smearing length, chosen to be the Compton wavelength

1185: $s=\sigma^{-1/3} $ of

1186: a piece of wall, which is a physically reasonable lower bound on the size of a measurable region. The smeared correlator will then be given by

1187: \begin{eqnarray}\label{smeared2p}

1188: \langle \phi_{s}({\bf x}) \phi_{s}({\bf y}) \rangle = \frac{1}{\pi^2 s^4} && \int_{\rm |{\bf z}-{\bf x}| < s} d^2 z \\ \nonumber && \int_{\rm |{\bf q}-{\bf y}| < s} d^2 q \langle \phi({\bf z}) \phi({\bf q}) \rangle,

1189: \end{eqnarray}

1190: which evaluates at ${\bf x} = {\bf y}$ to

1191: \begin{equation}\label{phismallr}

1192: \langle \phi_{s}^{2} \rangle = (\frac{1}{2} + {\cal G}) \sigma^{-2/3},

1193: \end{equation}

1194: where ${\cal G}=.916...$ is Catalin's constant. We have been unable to

1195: integrate Eq.~\ref{smeared2p} to obtain the exact form of the smeared

1196: two-point function for all $r$. However, it must smoothly interpolate

1197: between the value at $r=0$ in Eq.~\ref{phismallr} to the functional

1198: form at $r \gg s$ given by Eq.~\ref{philarger}. We therefore employ

1199: the following `toy model' smeared field correlator

1200: \begin{eqnarray}\label{phiconnect}

1201: \langle \phi({\bf x}) \phi({\bf y})  \rangle  &=& \frac{\sigma^{-1}}{4\pi s} \left[ (2\pi + 4\pi {\cal G} -1)e^{-r^2/2s^2} \right. \\ \nonumber && \left. + \frac{1}{r/s + 1} \right],

1202: \end{eqnarray}

1203: which has the correct asymptotics. With $r=r_0\Psi$, the dimensionless

1204: form of the toy correlator is given by:

1205: \begin{eqnarray}

1206: \label{eq-2pt}

1207: \langle  \Phi \tilde{\Phi} \rangle  = \frac{(-Q)}{(24 \pi)^2}  && \left[ (2\pi + 4\pi {\cal G} -1)e^{-(R^2/2) \Psi^2} \right. \\ \nonumber && \left. + \frac{1}{R \Psi + 1} \right],

1208: \end{eqnarray} 

1209: where $R\equiv r_0/s$.

1210: 

1211: The velocity-velocity correlator can be calculated using the

1212: Hamiltonian approach. For spacelike separations, this is given by

1213: \begin{eqnarray}

1214: \langle \phi'( {\bf x}) \phi'( {\bf y}) \rangle &=& \sigma^{-1} \int \frac{|p| d^2 p}{2(2 \pi)^2} \exp[i {\bf p} ({\bf x}-{\bf y} )] \\ \nonumber &=& \sigma^{-1} \int_{0}^{\Lambda} \frac{dp}{4 \pi} p^2 J_{0}(pr)

1215: \end{eqnarray}

1216: where we have introduced a hard momentum cut-off $\Lambda$ to obtain a

1217: finite answer. This cutoff will correspond to the inverse smearing

1218: length, with higher momentum scales higher accounted for in the

1219: smeared operator. The integral can be evaluated in terms of

1220: generalized hypergeometric functions as

1221: \begin{equation}\label{phi'larger}

1222: \langle \phi'({\bf x}) \phi'({\bf y}) \rangle = \frac{\Lambda^3 \sigma^{-1}}{12 \pi}  \ _{1}F_{2} \left( \frac{3}{2};1,\frac{5}{2};-\frac{\Lambda^2 r^2 }{4} \right).

1223: \end{equation} 

1224: 

1225: At close separations, we construct a smeared operator $\phi_{s}'$. The expectation value of this operator at zero separation is

1226: \begin{equation}\label{phi'smallr}

1227: \langle \phi_{s}'^{2} \rangle = \frac{\log(64) + 2\gamma -4}{2 \pi^2},

1228: \end{equation}

1229: where $\gamma=.577...$ is the Euler-Mascheroni constant. Smoothly

1230: connecting the small $r$ (Eq.~\ref{phi'smallr}) and large $r$

1231: (Eq.~\ref{phi'larger}) behavior as in Eq.~\ref{phiconnect}, the

1232: $\Phi'$ (defined similarly to $z'$) correlator on the bubble at the time

1233: of nucleation is:

1234: \begin{eqnarray}\label{phi'2pt}

1235: \langle  \Phi' \tilde{\Phi}' \rangle &=& \frac{(-Q)}{12 \pi} \left[ _{1}F_{2} \left( \frac{3}{2};1,\frac{5}{2};-\frac{R^2 \Psi^2}{4} \right) \right. \\ \nonumber & & \left. + \left(\frac{12 \gamma -6 \log(64) -24}{\pi} -1 \right)e^{-(R^2/2) \Psi^2} \right]

1236: \end{eqnarray}

1237: 

1238: The integrals Eq.~\ref{Cl} and Eq.~\ref{Al} for the correlators

1239: Eq.~\ref{eq-2pt} and Eq.~\ref{phi'2pt} must be evaluated

1240: numerically. Calculation of the coefficients for every $l$ and $R$ is

1241: unfortunately unfeasible because of the highly oscillatory behavior of

1242: the integrands and sheer number of mode functions that must be

1243: considered.  However, we have been able to deduce sufficiently good

1244: approximate fits for $C_l$ and $A_l$ as a function of both $l$ and

1245: $R$. In both cases, the power is dominated by a peak at $l\simeq R$.

1246: 

1247: The $C_l$ are nicely fit by the function

1248: \begin{eqnarray}

1249:  C_l &=& \frac{(-Q)}{(24 \pi)^2} \frac{83}{100} \frac{\sqrt{2}l + 1

1250:  }{\sqrt{2} R +1} \\ \nonumber && \exp \left\{\frac{-1}{4R^2-2} \left[

1251:    (\sqrt{2} l + 1 )^2 -(\sqrt{2} R + 1)^2 \right] \right\}.

1252: \end{eqnarray}

1253: The proposed fit for the $A_l$ consists of two power laws matched at

1254: the $R=l$ peak. For $l \leq R$, the best fit is $A_l=(-Q) l^{5/4}/( 10

1255: R^{9/4})$ and for $l>R$, the fit is $A_l=3 (-Q) R^{1.8} / (200 l^{2.6}

1256: )$. Because these power law indices are slightly uncertain, we

1257: only count the modes with $l \leq R$ in the $A_l$. This is

1258: conservative, and also justified because these modes will not

1259: contribute significantly to the sum in Eq.~\ref{Phiavg}.

1260: 

1261: With these initial conditions, we can now evolve each mode function

1262: using Eq.~\ref{levolve} and then re-sum in Eq.~\ref{Phiavg} to find

1263: the average size of the fluctuations at the turning point. We have

1264: calculated $f$ and $g$ up to the $l$ corresponding to the last

1265: unstable mode of the smallest $|Q|$, for all three inflation scales

1266: ($M_I=100$ GeV, $M_I=10^{14}$ GeV, and $M_I=10^{17}$ GeV). The results

1267: for weak- and GUT-scale inflation are shown in Fig.~\ref{fg}, where

1268: the dotted line indicates the boundary of the region over which the

1269: perturbations become non-linear (non-linear to the left of the

1270: line). It can be seen that in this model, even just quantum

1271: perturbations of the bubble wall grow nonlinear in bubbles that start

1272: at radii less than about one-tenth of the turnaround radius; this

1273: grossly violates the assumption of spherical symmetry used in

1274: tunneling calculations. On the other hand, none of the parameter space

1275: in the high inflation scale case went non-linear, and at all scales

1276: there is {\em always} a region of initial bubble radii near the

1277: turnaround radius, for which nonlinearity never occurs.

1278: 

1279: \subsection{Thick Walls and Radiation}\label{TWRAD}

1280: Just as we have no reason to expect a fluctuated region to be

1281: spherically symmetric, we have no reason to assume that it will have

1282: thin walls. An analysis of thick-walled true vacuum bubbles was

1283: undertaken in Ref.~\cite{GVQ91,VV91}, where it was found that the

1284: instabilities found in thin-walled case are still present in the form

1285: of deformations normal to the bubble profile. In the case of small

1286: false vacuum bubbles, there is no obvious consideration (such as a

1287: corresponding instanton) to supply the profile of the bubble wall, and

1288: so we can merely conjecture by precedent that the instability would be

1289: retained in the thick-walled case as well.

1290: 

1291: Another consideration, applying to bubbles smaller than the false

1292: vacuum horizon size, is whether inflation is spoiled by non-vacuum

1293: contributions to the energy density. The perturbations on the bubble

1294: wall translate into gravitational waves~\cite{II97,II99}, and since

1295: the bound bubble solutions remain relatively close to their

1296: gravitational radius and become distorted over many different length

1297: scales on a relatively short time scale (see the quasi-exponential

1298: growth in Fig.~\ref{usj}), they will be emitters of copious

1299: gravitational radiation. Another problem arises if the kinetic and

1300: gradient energy of the field becomes appreciable in the bubble

1301: interior, either from intrusion of the wall (for example, imagine a

1302: bubble being pinched in half by some non-linear perturbation), or from

1303: particle production or other scalar modes propagating in from the

1304: wall. If the emission of energy into the interior of the bubble from

1305: any combination of these modes makes a significant contribution to the

1306: equation of state, then inflation will not occur.

1307: 

1308: \section{Summary \& Discussion} \label{conclusions}

1309: 

1310: In this paper we have examined the feasibility of producing

1311: inflationary universes from a spacetime with a small cosmological

1312: constant. Reviewing the solutions allowed by the junction conditions,

1313: there are several distinct possibilities if one allows for violations

1314: of the null energy condition. The first is the Farhi, Guth \& Guven

1315: (FGG) mechanism, in which (referring to Fig.~\ref{confdiags1} and Fig.~\ref{confdiags2}) the IV-I-II/III (solution 1) or IV-III-II/III (solution 2) solution is fluctuated

1316: in the expanding phase and then tunnels to an unbound solution, as

1317: shown in Fig.~\ref{tunnelst}. To the outside observer in region I of

1318: the SdS conformal diagram it appears that the bubble has disappeared

1319: behind the horizon, but on the other side of the wormhole, both an

1320: inflationary universe and a non-inflating universe (an asymptotically

1321: true-vacuum de Sitter region) have come into existence.  In one case,

1322: shown in solution 5 of Fig.~\ref{confdiags2}, we can imagine an observer in

1323: region III, then take the Schwarzschild mass $m\rightarrow 0$ limit

1324: (thus removing regions I, II and IV entirely) and interpret the spacetime as that of the analytically continued Coleman-De Luccia instanton. Since the very same spacetime takes part in both the FGG mechanism and the Coleman-De Luccia false vacuum instanton, there may be some way to smoothly interpolate between these two processes;

1325: we intend to explore this idea further in future work.

1326: 

1327: In addition to the FGG and regular instanton mechanisms, there are two

1328: more possibilities. First, an inflationary universe might be directly

1329: produced by some null energy condition (NEC) violating fluctuation

1330: into one of the unbound or monotonic solutions shown in Solutions 5-13 of

1331: Fig.~\ref{confdiags2}. To the outside observer in region I of the SdS

1332: conformal diagram, these solutions would be indistinguishable from the

1333: fluctuation of a black hole, but they would secretly entail the

1334: creation of everything on the other side of the wormhole, as in the

1335: FGG tunneling. A second method of direct production is the fluctuation

1336: (which does not require an NEC violation) of the IV-I-II'/III-II

1337: solution (solution 3 and 4). These solutions only exist in the case where the

1338: interior and exterior cosmological constants are comparable, and so

1339: would not correspond to the nucleation of inflation from a universe

1340: like ours, but they might be of interest in understanding transitions

1341: between nearly degenerate vacua.

1342: 

1343: Examining these classical solutions to first order, we have shown that

1344: an instability to aspherical perturbations exists in those solutions

1345: which possess a turning point. This includes the bound (solutions 1 and 2)

1346: and unbound solutions (solutions 5-9). In the latter case there is

1347: no clear way to set an initial radius or initial perturbation

1348: amplitude, so we can say only that collapsing bubbles are violently

1349: unstable. The bound solutions are amenable to quantitative

1350: investigation, and we have focused on the growth of perturbations in

1351: the expanding phase that precedes tunneling in the FGG mechanism. For

1352: bound expanding bubbles formed by the fluctuations of a scalar field

1353: in de Sitter space, there is no instanton to enforce spherical

1354: symmetry, so we would expect initial aspherical perturbations to be

1355: relatively large. Since there is no detailed model for the fluctuating

1356: scalar field to see {\em how} large, we have instead calculated an

1357: estimate of the minimal deviations from spherical symmetry in light of

1358: quantum fluctuations, and present this as an extremely rare, best case

1359: scenario for spherical symmetry. These minimal fluctuations were then

1360: evolved to the turning point of the bound solutions, which is the

1361: point in the FGG mechanism where there is a chance for tunneling to an

1362: inflationary universe to occur. Of the three representative energy

1363: scales for inflation (false vacuum energy densities) we have studied,

1364: the evolved minimal perturbations on a Plank scale bubble remain small

1365: over most of the allowed parameter space, while the perturbations on

1366: GUT and weak scale bubbles can grow nonlinear if they start at a

1367: sufficiently small ($\alt 10\%$) fraction of the turnaround radius.

1368: Thus even in the best-case scenario some bubbles become nonlinear, but

1369: on the other hand there will always in principle be some that do not.

1370: 

1371: The instability introduces complications into the use of the FGG

1372: mechanism as a means of baby universe production. The existing

1373: calculations of the tunneling rate rely heavily on the assumption of

1374: spherical symmetry. It is unclear how to perform a similar calculation

1375: for a (possible non-linearly) perturbed bubble, as the number of

1376: degrees of freedom has drastically increased and the assumption of a

1377: minisuperspace of spherically symmetric metrics is no longer

1378: good. Further, the bubble interior will become filled with scalar

1379: gradient and kinetic energy and gravity waves, possibly upsetting the

1380: interior sufficiently to prevent inflation. One might argue that in an

1381: eternal universe there is plenty of time to wait around for a

1382: fluctuation which is sufficiently spherical. However, to fully

1383: understand the importance of the FGG mechanism, one must both have a

1384: model of the scalar field fluctuations, which would predict the

1385: distribution of bubble shapes and masses, and also a model for

1386: tunneling in the presence of asphericities. In addition, one must

1387: understand the competition among the various processes which result in

1388: the formation of inflating false vacuum regions, and thus show that

1389: the FGG mechanism is not a relative rarity.  These problems represent

1390: significant calculations in a theory of quantum gravity which we do

1391: not yet possess, but progress in these areas would undoubtedly improve

1392: our understanding of the initial conditions for inflation.

1393: 

1394: \begin{acknowledgments}

1395: The authors wish to thank A. Albrecht, T. Banks, P.J. Fox, S. Gratton, and T. Mai for their assistance in the development of this work.

1396: \end{acknowledgments}

1397: 

1398: \suppressfloats

1399: 

1400: \begin{thebibliography}{}

1401: 

1402: 

1403: \bibitem{GF87}

1404: E.~Farhi and A.~H.~Guth,

1405: %``An Obstacle To Creating A Universe In The Laboratory,''

1406: Phys.\ Lett.\ B {\bf 183}, 149 (1987).

1407: %%CITATION = PHLTA,B183,149;%%

1408: \bibitem{FGG90}

1409: E.~Farhi, A.~H.~Guth and J.~Guven,

1410: %``Is It Possible To Create A Universe In The Laboratory By Quantum

1411: %Tunneling?,''

1412: Nucl.\ Phys.\ B {\bf 339}, 417 (1990).

1413: \bibitem{FMP89}

1414: W.~Fischler, D.~Morgan and J.~Polchinski,

1415: %``Quantum Nucleation Of False Vacuum Bubbles,''

1416: Phys.\ Rev.\ D {\bf 41}, 2638 (1990); W.~Fischler, D.~Morgan and J.~Polchinski,

1417: %``Quantization Of False Vacuum Bubbles: A Hamiltonian Treatment Of

1418: %Gravitational Tunneling,''

1419: Phys.\ Rev.\ D {\bf 42}, 4042 (1990).

1420: \bibitem{BGG87}

1421: S.~K.~Blau, E.~I.~Guendelman and A.~H.~Guth,

1422:   %``The Dynamics Of False Vacuum Bubbles,''

1423:   Phys.\ Rev.\ D {\bf 35}, 1747 (1987).

1424: \bibitem{DKS02}

1425: L.~Dyson, M.~Kleban and L.~Susskind,

1426: %``Disturbing implications of a cosmological constant,''

1427: JHEP {\bf 0210}, 011 (2002)

1428: [arXiv:hep-th/0208013].

1429: \bibitem{AS04}

1430: A.~Albrecht and L.~Sorbo,

1431: %``Can the universe afford inflation?,''

1432: Phys.\ Rev.\ D {\bf 70}, 063528 (2004)

1433: [arXiv:hep-th/0405270].

1434: \bibitem{Carroll:2004pn}

1435: S.~M.~Carroll and J.~Chen,

1436: %``Spontaneous inflation and the origin of the arrow of time,''

1437: arXiv:hep-th/0410270.

1438: \bibitem{Hollands:2002yb}

1439: S.~Hollands and R.~M.~Wald,

1440: %``An alternative to inflation,''

1441: Gen.\ Rel.\ Grav.\  {\bf 34}, 2043 (2002)

1442: [arXiv:gr-qc/0205058].

1443: \bibitem{penrose} R. Penrose, in ``Fourteenth Texas Symposium on

1444: Relativistic Astrophysics, Dallas, Tex., 1988'' (New York: Annales of

1445: the New York Academy of Sciences, 1989)

1446: \bibitem{LW87}

1447: K.~M.~Lee and E.~J.~Weinberg,

1448: %``Decay Of The True Vacuum In Curved Space-Time,''

1449: Phys.\ Rev.\ D {\bf 36}, 1088 (1987).

1450: \bibitem{G94}

1451: R.~Basu, A.~H.~Guth and A.~Vilenkin,

1452: %``Quantum creation of topological defects during inflation,''

1453: Phys.\ Rev.\ D {\bf 44}, 340 (1991);

1454: J.~Garriga,

1455: %``Nucleation rates in flat and curved space,''

1456: Phys.\ Rev.\ D {\bf 49}, 6327 (1994)

1457: [arXiv:hep-ph/9308280].

1458: \bibitem{Banks:2002nm}

1459: T.~Banks,

1460: %``Heretics of the false vacuum: Gravitational effects on and of vacuum decay.

1461: %II,''

1462: arXiv:hep-th/0211160.

1463: \bibitem{Bousso:2004tv}

1464: R.~Bousso,

1465: %``Cosmology and the S-matrix,''

1466: Phys.\ Rev.\ D {\bf 71}, 064024 (2005)

1467: [arXiv:hep-th/0412197].

1468: \bibitem{DV05}

1469: S.~Dutta and T.~Vachaspati,

1470: %``Islands in the Lambda-sea,''

1471: Phys.\ Rev.\ D {\bf 71}, 083507 (2005)

1472: [arXiv:astro-ph/0501396].

1473: \bibitem{VT99}

1474: T.~Vachaspati and M.~Trodden,

1475: %``Causality and cosmic inflation,''

1476: Phys.\ Rev.\ D {\bf 61}, 023502 (2000)

1477: [arXiv:gr-qc/9811037].

1478: \bibitem{BKT87}

1479: V.~A.~Berezin, V.~A.~Kuzmin and I.~I.~Tkachev,

1480: %``Thin Wall Vacuum Domains Evolution,''

1481: Phys.\ Lett.\ B {\bf 120}, 91 (1983);

1482: V.~A.~Berezin, V.~A.~Kuzmin and I.~I.~Tkachev,

1483: %``Dynamics Of Bubbles In General Relativity,''

1484: Phys.\ Rev.\ D {\bf 36}, 2919 (1987).

1485: \bibitem{S86}

1486: H.~Sato, Prog. Theor. Phys. {\bf 76}, 1250 (1986).

1487: \bibitem{APS89}

1488: A.~Aurilia, M.~Palmer and E.~Spallucci,

1489: %``Evolution Of Bubbles In A Vacuum,''

1490: Phys.\ Rev.\ D {\bf 40}, 2511 (1989).

1491: \bibitem{LR77}

1492: K.~Lake and R.~C.~Roeder,

1493: %``Effects Of A Nonvanishing Cosmological Constant On The Spherically

1494: %Symmetric Vacuum Manifold,''

1495: Phys.\ Rev.\ D {\bf 15}, 3513 (1977).

1496: \bibitem{CP04}

1497: T.~R.~Choudhury and T.~Padmanabhan,

1498: %``Concept of temperature in multi-horizon spacetimes: Analysis of

1499: %Schwarzschild-de Sitter metric,''

1500: arXiv:gr-qc/0404091.

1501: \bibitem{C77}

1502: S.~R.~Coleman,

1503: %``The Fate Of The False Vacuum. 1. Semiclassical Theory,''

1504: Phys.\ Rev.\ D {\bf 15}, 2929 (1977)

1505: [Erratum-ibid.\ D {\bf 16}, 1248 (1977)].

1506: \bibitem{CC77}

1507: C.~G.~.~Callan and S.~R.~Coleman,

1508: %``The Fate Of The False Vacuum. 2. First Quantum Corrections,''

1509: Phys.\ Rev.\ D {\bf 16}, 1762 (1977).

1510: \bibitem{CD80}

1511: S.~R.~Coleman and F.~De Luccia,

1512: %``Gravitational Effects On And Of Vacuum Decay,''

1513: Phys.\ Rev.\ D {\bf 21}, 3305 (1980).

1514: \bibitem{P65}

1515: R.~Penrose,

1516: %``Gravitational Collapse And Space-Time Singularities,''

1517: Phys.\ Rev.\ Lett.\  {\bf 14}, 57 (1965).

1518: \bibitem{L91}

1519: A.~D.~Linde,

1520: %``Hard art of the universe creation (stochastic approach to tunneling and

1521: %baby universe formation),''

1522: Nucl.\ Phys.\ B {\bf 372}, 421 (1992)

1523: [arXiv:hep-th/9110037].

1524: \bibitem{GM04}

1525: J.~Garriga and A.~Megevand,

1526: %``Decay of de Sitter vacua by thermal activation,''

1527: Int.\ J.\ Theor.\ Phys.\  {\bf 43}, 883 (2004)

1528: [arXiv:hep-th/0404097].

1529: \bibitem{GHTW04}

1530: A.~Gomberoff, M.~Henneaux, C.~Teitelboim and F.~Wilczek,

1531: %``Thermal decay of the cosmological constant into black holes,''

1532: Phys.\ Rev.\ D {\bf 69}, 083520 (2004)

1533: [arXiv:hep-th/0311011].

1534: \bibitem{AFW91}

1535: F.~C.~Adams, K.~Freese and L.~M.~Widrow,

1536: %``Evolution Of Nonspherical Bubbles,''

1537: Phys.\ Rev.\ D {\bf 41}, 347 (1990).

1538: \bibitem{GV91}

1539: J.~Garriga and A.~Vilenkin,

1540: %``Perturbations on domain walls and strings: A Covariant theory,''

1541: Phys.\ Rev.\ D {\bf 44}, 1007 (1991).

1542: \bibitem{G93}

1543: J.~Guven,

1544: %``Covariant perturbations of domain walls in curved space-time,''

1545: Phys.\ Rev.\ D {\bf 48}, 4604 (1993)

1546: [arXiv:gr-qc/9304032].

1547: \bibitem{GVQ91}

1548: J.~Garriga and A.~Vilenkin,

1549: %``Quantum fluctuations on domain walls, strings and vacuum bubbles,''

1550: Phys.\ Rev.\ D {\bf 45}, 3469 (1992).

1551: \bibitem{VV91}

1552: T.~Vachaspati and A.~Vilenkin,

1553: %``Quantum state of a nucleating bubble,''

1554: Phys.\ Rev.\ D {\bf 43}, 3846 (1991).

1555: \bibitem{II97}

1556: A.~Ishibashi and H.~Ishihara,

1557: %``Bubble wall perturbations coupled with gravitational waves,''

1558: Phys.\ Rev.\ D {\bf 56}, 3446 (1997)

1559: [arXiv:gr-qc/9704058].

1560: \bibitem{II99}

1561: A.~Ishibashi and H.~Ishihara,

1562: %``Equation of motion for a domain wall coupled to gravitational field,''

1563: Phys.\ Rev.\ D {\bf 60}, 124016 (1999)

1564: [arXiv:gr-qc/9802036].

1565: 

1566: \end{thebibliography}

1567: 

1568: \end{document}

1569: