1: \documentstyle[eqsecnum,preprint,aps,epsbox]{revtex}
2:
3: \def\beq{\begin{equation}}
4: \def\eeq{\end{equation}}
5: \def\bx{{\bf x}}
6: \def\bk{{\bf k}}
7: \newcommand{\mpl}{{m_{\rm Pl}}}
8: \newcommand{\etal}{{\it et al.\ }}
9: \renewcommand{\prd}{{\it Phys.\ Rev.\ D~}}
10: \renewcommand{\prl}{{\it Phys.\ Rev.\ Lett.\ }}
11: \newcommand{\plb}{{\it Phys.\ Lett.\ B~}}
12:
13: \begin{document}
14:
15: \draft
16: \title{Generality of Topological Inflation}
17:
18: \author{Nobuyuki Sakai\thanks{Electronic address:
19: sakai@ke-sci.kj.yamagata-u.ac.jp}}
20: \address{Faculty of Education, Yamagata University, Yamagata
21: 990-8560, Japan}
22: \address{Graduate School of Science and Engineering, Yamagata
23: University, Yamagata 990-8560, Japan}
24: \address{Osservatorio Astronomico di Roma, Via Frascati, 00040 Monte
25: Porzio Catone, Italia}
26:
27: \maketitle
28:
29: \begin{abstract}
30: Many authors claimed that a large initial inhomogeneity prevents the onset
31: of inflation and therefore inflation takes place only if the scalar field
32: is homogeneous or appropriately chosen over the horizon scale. We show
33: that those arguments do not apply to topological inflation. The core of a
34: defect starts inflation even if it has much smaller size than the horizon
35: and much larger gradient energy than the potential, as long as the
36: vacuum expectation value is large enough ($\gtrsim0.3\mpl$) and the core is not
37: contracting initially. This is due to stability of false vacuum.
38: \end{abstract}
39:
40: \vskip 1cm
41: \begin{center}
42: PACS number(s): 98.80.Cq, 11.27.+d, 04.20.Dm
43: \end{center}
44:
45: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
46: %2345678901234567890123456789012345678901234567890123456789012345678901234567890
47: \newpage
48: \tighten
49: %\baselineskip = 18pt
50: \section{Introduction}
51:
52: Inflation gives a natural solution of the horizon problem of the big-bang
53: universe. However, if inflation requires homogeneous initial conditions
54: over the super-horizon scale, one cannot say that it is a solution of the
55: horizon problem, though it reduces the problem by many orders of
56: magnitude. To answer this naive question, many people have investigated
57: how initial inhomogeneity affects the onset of inflation
58: \cite{GP,KK,CNN,confGP,VT,Cou,BG}. The most remarkable work was done by
59: Goldwirth and Piran \cite{GP}, who solved the full Einstein equations for
60: a spherically symmetric spacetime. Their results are summarized as
61: follows. New inflation is so sensitive to initial inhomogeneity that it
62: requires homogeneity over a region of several horizon sizes is needed.
63: Chaotic inflation is not so affected by initial inhomogeneity but requires
64: a sufficiently high average value of the scalar field over a region of
65: several horizon sizes. Later many people explored the same problem in
66: different approaches \cite{CNN,confGP,VT,Cou,BG}, which as a whole confirmed
67: the results obtained by Goldwirth and Piran. This suggests failure of the
68: ``ordinary" scenario of inflationary cosmology and therefore some
69: alternatives were discussed \cite{VT,Cou,BG},
70:
71: In this paper we investigate whether topological inflation \cite{LV} also
72: requires initial homogeneity over the horizon size. The simple model of
73: topological inflation is described by the potential,
74: \beq\label{V}
75: V={\lambda\over 4}(\Psi^2-\eta^2)^2.
76: \eeq
77: However, the same type of inflation takes place in many plausible theories
78: such as a nonminimally coupled massive scalar field \cite{SY}, some models of
79: supergravity \cite{sugra}, and some superstring inspired models
80: \cite{string,BAR}.
81: The inflationary condition for the model (\ref{V}) was investigated by
82: numerical analysis \cite{SSTM}: $\eta\gtrsim0.33\mpl$ for non-gauged defects.
83: Although we claimed that the critical value of $\eta$ does not depend on
84: initial conditions \cite{SSTM}, it seems to lack of convincible
85: explanations according to recent arguments \cite{VT,Cou,BG}. We therefore
86: return to the same model and clarify the effect of initial inhomogeneity
87: both analytically and numerically.
88:
89: The plan of this paper is as follows. In Sec II, we discuss analytically the
90: scalar field dynamics under the approximation that the metric is homogeneous.
91: In Sec III, we numerically solve the Einstein equations for a global monopole
92: with a large gradient. Section IV is devoted to a summary and discussion.
93: We use the units of $c=\hbar=1$ throughout the paper.
94:
95:
96: \section{Scalar Field Dynamics in a Homogeneous Spacetime}
97:
98: Let us consider the scalar field dynamics under the approximation that
99: the metric is homogeneous, following Berera and Gordon \cite{BG}.
100: The equation of the scalar field $\Psi(t, \bx)$ is
101: \beq\label{sfeq}
102: \ddot\Psi+3H\dot\Psi-{\nabla^2\Psi\over a^2}+{dV\over d\Psi}=0,
103: \eeq
104: where an overdot denotes time-derivatives, $a(t)$ is the scale factor and
105: $H\equiv\dot a/a$.
106: Although $a(t)$ is not a solution of the exact Einstein equations, it is
107: instructive to write down its evolution equation under the present
108: approximation,
109: \beq\label{aeq}
110: {\ddot a\over a}={8\pi\over 3\mpl^2}(-\dot\Psi^2+V).
111: \eeq
112: This equation tells us that inflation does not take place ($\ddot a<0$) if
113: the kinetic term dominates over the potential. It should be noted that a gradient
114: term does not appear in Eq.(\ref{aeq}). Gradient energy can affect cosmic expansion
115: only by exciting kinetic energy through the field equation (\ref{sfeq}).
116:
117: Because inflation can take place at $\Psi\approx0$, we approximate the
118: potential (\ref{V}) as $V=\lambda\eta^4/4-\lambda\eta^2\Psi^2/2$. By
119: Fourier expansion
120: $\Psi(t,\bx)=\sum\Psi_{\bk}(t)e^{i\bk\cdot\bx}$,
121: Eq.(\ref{sfeq}) becomes
122: \beq\label{keq}
123: \ddot\Psi_{\bk}+3H\ddot\Psi_{\bk}+{k^2\over a^2}\Psi_{\bk}-\lambda\eta^2
124: \Psi_{\bk}=0.
125: \eeq
126: We assume the slow-roll condition for a homogeneous and isotropic
127: spacetime, $|\ddot\Psi|\gg3H|\dot\Psi|$ and $\dot\Psi ^2/2\gg V$ at
128: $\Psi\approx0$, which is equivalent to $\eta\gg\mpl/\sqrt{6\pi}$.
129: (Actually the numerical finding $\eta>0.33\mpl$ \cite{SSTM} is more precise
130: condition.) Under
131: this condition what happens if the $k^2$-term of (\ref{keq}) dominates?
132: Clearly $\Psi_{\bk}$ begins oscillation, which contributes kinetic energy.
133: In the case of other nontopological inflationary models, inflation never
134: happens if kinetic energy dominates until the mean field rolls down to the
135: potential minimum and the slow-roll condition breaks down. On the other
136: hand, in the case of topological inflation, the false vacuum $\Psi=0$ is
137: topologically stable and never decays. As long as the core region
138: $\Psi\approx0$ is not contracting at an initial time ($\dot a\ge 0$),
139: oscillations are eventually damped. Because a local region of $\Psi\approx
140: 0$ always exists, inflation takes place eventually.
141:
142: The above arguments are rough because we have neglected the inhomogeneity of
143: the spacetime. This is why we shall move on to numerical analysis in the next
144: section. Nevertheless, the above arguments will help us to understand our
145: numerical results intuitively.
146:
147: \section{Numerical Analysis of a Monopole with a Large Gradient}
148:
149: Here we take inhomogeneity of spacetime into account by solving the
150: Einstein equations coupled to the scalar field equation numerically. We
151: restrict ourselves to the global monopole system. We believe that the main
152: features of dynamics do not change for domain walls and global strings,
153: according to the previous work \cite{SSTM}. Because we have already known
154: that the inhomogeneity of gauge fields affects the onset of inflation
155: \cite{Sak} and our present concern is pure effects of the scalar field
156: inhomogeneity, we do not consider gauged defects here.
157:
158: We therefore consider the Einstein-Higgs system, which is described by the
159: action,
160: \beq\label{action}
161: {\cal S}=\int d^4x\sqrt{-g}\left[{\mpl^2\over 16\pi}{\cal R}
162: -\frac12(\partial_{\mu}\Psi^a)^2-V(\Psi) \right], ~ ~ (a=1,2,3),
163: \eeq
164: with the potential (\ref{V}).
165: The action (\ref{action}) yields the field equations,
166: \begin{eqnarray} \label{Ein}
167: G_{\mu\nu}&\equiv&{\cal R}_{\mu\nu}-\frac12g_{\mu\nu}{\cal R}
168: ={8\pi\over\mpl}T_{\mu\nu},\\
169: \label{Psi}
170: \Box\Psi^a&=&\frac{\partial V(\Psi)}{\partial\Psi^a},
171: \end{eqnarray}
172: with
173: \begin{equation}
174: T_{\mu\nu}\equiv\nabla_{\mu}\Psi^a\nabla_{\nu}\Psi^a
175: -g_{\mu\nu}\left[\frac12(\nabla\Psi^a)^2+V(\Psi)\right].
176: \end{equation}
177: We assume a spherically symmetric spacetime
178: and adopt the metric,
179: \beq\label{metric}
180: ds^2=-dt^2+A^2(t,r)dr^2+B^2(t,r)r^2(d\theta^2+\sin^2\theta d\varphi^2).
181: \eeq
182: For the scalar field, we adopt the hedgehog ansatz:
183: \beq\label{hg}
184: \Psi^a=\Psi(t,r)(\sin\theta\cos\varphi,\sin\theta\sin\varphi,\cos\theta)
185: \eeq
186:
187: In the following we shall write down the field equations (\ref{Ein}) and
188: (\ref{Psi}) with the assumptions (\ref{metric}) and (\ref{hg}).
189: To begin with, we introduce the extrinsic curvature tensor $K_{ij}$:
190: \begin{equation}
191: K^r_r=-{\dot A\over A},~~~ K^{\theta}_{\theta}(=K^{\varphi}_{\varphi})
192: =-{\dot B\over B},~~~
193: K\equiv K^i_i.
194: \end{equation}
195: Next, following Nakamura {\it et al.}\cite{Nak}, we introduce
196: auxiliary dynamical variables,
197: \begin{equation}
198: a\equiv{A-B\over r^2},~~~ k\equiv{K^{\theta}_{\theta}-K^r_r\over r^2},
199: \end{equation}
200: and use a new space variable, $x\equiv r^2$, instead of $r$. The advantages of
201: those variables were discussed in Ref.\cite{SYM}.
202:
203: We introduce further auxiliary variables,
204: \beq
205: C\equiv-{B'\over B}, ~~~
206: \psi\equiv{\Psi\over r}, ~~~ \varpi\equiv\dot\psi, ~~~ \xi\equiv\psi',
207: \eeq
208: where a prime denotes $\partial/\partial x$.
209: A full set of dynamical variables is
210: $a,~B,~C, ~\psi,~\xi, ~K,~k$ and $\varpi$. The time-derivatives of the first
211: five variables are given by the definitions above:
212: \begin{eqnarray}
213: &&\dot a=-a K^r_r+B k,~~~
214: \dot B=-B K^{\theta}_{\theta}, ~~~ \dot C={K^{\theta}_{\theta}}',
215: \label{dotB} \\ \label{dotpsi}
216: &&\dot\psi=\varpi, ~~~ \dot\xi=\varpi'.
217: \end{eqnarray}
218: Note that the value of ${K^{\theta}_{\theta}}'$ in Eq.(\ref{dotB}) is
219: determined by the momentum constraint (\ref{MC}) below, which gives a more
220: accurate value than a finite difference of $K^{\theta}_{\theta}$.
221:
222: Now we write down the field equations (\ref{Ein}) and (\ref{Psi}) as
223: %%%%%%%%% HC
224: \begin{eqnarray}\label{HC}
225: -G^t_t&\equiv&{K^2-k^2x^2\over 3}
226: +{8x\over A^2}\left(C'-{A'C\over A}-{3C^2\over 2}\right)
227: +{4\over A^2}\left[{A'\over A}+4C+{a(A+B)\over 4B^2}\right] \nonumber\\
228: &=& {8\pi\over\mpl^2}\left[{x\varpi^2\over 2}+{2x\xi\over A^2}(x\xi+\psi)
229: +{\psi^2\over 2A^2}+{\psi^2\over B^2}+V \right],
230: %%%%%%%%% MC
231: \\ \label{MC}
232: {G_{tr}\over 4r}&\equiv& {K^{\theta}_{\theta}}'+k\left({1\over2}-xC\right)
233: ={2\pi\varpi\over\mpl^2}(2x\xi+\psi),\\
234: %%%%%%%%% dot K
235: \label{dotK}
236: -{\cal R}^t_t&\equiv&\dot K-{K^2+2x^2k^2\over 3}
237: ={8\pi\over\mpl^2} (x\varpi^2-V),\\
238: %%%%%%%%% dot k
239: \label{dotk}
240: {{\cal R}^r_r-{\cal R}^{\theta}_{\theta}\over r^2}&\equiv&
241: \dot k-kK+{2\over A^3}\left[2AC'-2A'C+aC+a'-{a^2\over B^2}
242: \left({A\over 2}+B\right)\right] \nonumber\\
243: &=&{8\pi\over\mpl^2 A^2}
244: \left[4\xi(x\xi+\psi)-{\psi^2 a(A+B)\over B^2}\right],\\
245: %%%%%%%%% dot varpi
246: \label{dotvarpi}
247: \dot\varpi&=&K\varpi+{4\over A^2}\left[x\xi'-x\xi\left({A'\over A}+2C\right)
248: +{5\xi\over 2} \right] \nonumber\\
249: &&-{2\psi\over A^2}\left[{A'\over A}+2C+{a(A+B)\over B^2}\right]
250: -\lambda\psi(\Psi^2-\eta^2).
251: \end{eqnarray}
252: Equations (\ref{HC}) and (\ref{MC}) are constraint equations, which fixes
253: initial values of the metric. Equations (\ref{dotK})-(\ref{dotvarpi}) together
254: with (\ref{dotB}) and (\ref{dotpsi}) provides the time evolution of the seven
255: variables.
256:
257: Let us discuss how to discretize
258: a space at each time into a mesh and approximate spatial derivatives.
259: Because our field equations are written with $x$ instead of $r$, the simplest
260: way is to use a regular mesh with separation $\Delta x$ in $x$-space, as we did in
261: Ref.\cite{SYM}.
262: However, this mesh gives lower resolution for a center than for a far region in
263: terms of $r$-space, which is unsuitable for the present case, where the central
264: region has a large gradient. Therefore, we adopt a regular mesh with separation
265: $\Delta r$ in $r$-space,
266: \beq
267: r_1=0,~ r_2=\Delta r,~ ...,~ r_i=(i-1)\Delta r,~ ...,
268: \eeq
269: and approximate $x$-derivatives of any function $F(x)$ by
270: \beq
271: {dF\over dx}(x_i)={1\over 2r_i}{dF\over dr}(x_i)
272: \approx{1\over 2r_i}{F_{i+1}-F_{i-1}\over2\Delta r}
273: ={F_{i+1}-F_{i-1}\over x_{i+1}-x_{i-1}}.
274: \eeq
275:
276: We impose further regularity condition on all dynamical variables at $r=0$,
277: $F(0)$, as follows. Although all evolution equations do not contain any
278: diverging factor at $r=0$ like $1/r$, we do not use them to obtain $F_1=F(0)$
279: at each time. Instead, following Hayley and Choptuik \cite{HC}, we
280: employ a ``quadratic fit",
281: \beq
282: F_1={4F_2-F_3\over 3}.
283: \eeq
284: We find that this simple method is very effective to keep good accuracy in the
285: center.
286:
287: As initial conditions of the scalar field, we suppose
288: \beq
289: \Psi(t=0,r)=\eta\tanh\left({r\over\delta}\right), ~~~
290: \dot\Psi(t=0,r)=0,
291: \eeq
292: where $\delta$ is a free parameter, which controls the initial gradient at
293: $\Psi\approx0$. If the previous arguments applied to topological
294: inflation, $\delta$ would have to be larger than the horizon size
295: $H_{0}^{-1}\equiv(8\pi V(0)/3\mpl^2)^{-1/2}$ for the onset of inflation.
296: Now our question is specified: can inflation start even if $\delta \ll
297: H_{0}^{-1}$? We therefore choose $\delta=0.01H_0^{-1}$ below as an example.
298:
299: For the metric, there are four unknown variables, $B,~a,~K$ and $k$,
300: while there are two constraint equations (\ref{HC}) and (\ref{MC}).
301: Therefore, two of the four variables are arbitrary chosen, and this choice is
302: an important problem. If we choose $K=$constant, for example,
303: large initial inhomogeneity of $\Psi$ causes a black hole or a
304: singularity, as discussed by Chiba \etal \cite{CNN}~~
305: The appearance of black holes or singularities, however, does not mean failure
306: of inflation, because inflation is a local phenomenon.
307: To judge whether inflation occurs or not,
308: we should solve the evolution equations; however, it is difficult to solve them
309: for such spacetimes numerically. For such a technical reason, we set initial
310: conditions without a black hole nor a singularity by adopting the
311: following two ways, under the condition that the initial 3-space is not
312: contracting, $K\le 0$.
313:
314: (A) Following Goldwirth and Piran \cite{GP}, we set a homogeneous and
315: isotropic 3-space by introducing another massless scalar field $\chi$ to
316: compensate inhomogeneous energy density of $\Psi$. We assume
317: $\chi(t=0,r)=0$ and $\dot\chi(t=0,r)>0$. The initial metric functions become
318: $B=1,~a=0,~k=0$ and $K=$constant.
319:
320: (B) Without introducing a massless scalar field, we set $B=1$ and $a=0$ and solve
321: the constraint equations (\ref{HC}) and (\ref{MC}) to obtain $K$ and $k$.
322:
323: We demonstrate solutions with $\lambda=0.1$, $\eta=\mpl$ and
324: $\delta=0.01H_0^{-1}$ in Figs.\ 1 and 2. In both cases we see in (a) and (b) the
325: monopole core expands. As Eq. (\ref{dotK}) suggests, whether the expansion rate is
326: increasing or not is determined by the sign of $\dot\Psi^2-V$ even in an inhomogeneous
327: spacetime. Therefore, once kinetic energy becomes
328: smaller than the potential, inflation starts even though gradient energy is still
329: dominant. This behavior is actually seen in Figs.\ 1(c)-(e) and 2(c)-(d).
330:
331: The evolution time in Fig.\ 2 is relatively short because calculation breaks down
332: soon after the black-hole horizon appear at $t\approx 0.27H_0^{-1}$. However, we
333: believe it demonstrates sufficiently that inflation actually begins.
334:
335: It should be emphasized that topological inflation can take place even if
336: $\rho_{grad}\gg V$, as shown in Figs.\ 1(c) and 2(c). This is in contrast
337: with the previous results \cite{GP}, which showed that new inflation
338: requires a more stringent condition than $\rho_{grad}<V$.
339:
340: \section{Summary and Discussion}
341:
342: We have investigated whether topological inflation requires
343: initial homogeneity over the horizon size. Contrary to the results for
344: nontopological inflationary models, which have been investigated by many
345: people, topological inflation can take place even if the initial false
346: vacuum has one percent of the horizon size and gradient energy of $10^5
347: V(0)$. This is due to topological stability of false vacuum. In this sense
348: topological inflation explains most naturally the onset of inflation in a
349: really chaotic universe.
350:
351: Here we make a comment on the papers by Vachaspati and Trodden \cite{VT}, who
352: considered initial conditions for inflation based on the null Raychaudhuri
353: equation, and concluded that {\it homogeneity on super-horizon scales must be
354: assumed as an initial condition.} They claimed that the size of the initial
355: inflationary patch, $X$, must be greater than the inflationary horizon,
356: $H_{inf}^{-1}$, which must be larger than the (pre-inflationary) background
357: FRW inverse Hubble size at the time inflation starts, $H_{FRW}^{-1}$, that is,
358: \beq
359: X\ge H_{inf}^{-1}\ge H_{FRW}^{-1}.
360: \eeq
361: However, what they really proved is only the second inequality,
362: $H_{inf}^{-1}\ge H_{FRW}^{-1}$, to which we have no objection.
363: On the other, they just {\it assumes} the first inequality, although its
364: validity is more crucial. Actually, what we showed in the present paper is
365: topological inflation can takes place even if $X<H_{inf}^{-1}$.
366:
367: Another note is added.
368: Stability of false vacuum is a key feature of
369: topological inflation, which sometimes leads to the idea that inflation
370: cannot end \cite{BAR}. It is, however, not to worry. Once inflation
371: begins, the evolution of a local region is described by Eq.(\ref{sfeq})
372: without the gradient term. Therefore, just like new inflation or chaotic
373: inflation, $\Psi$ simply rolls down to the potential minimum except for
374: $\Psi=0$, that is, inflation ends in any region except the one point. More
375: detailed arguments are given in Ref.\cite{SNH}.
376:
377: \acknowledgements
378: The author thanks Luca Amendola for discussions and his hospitality at
379: Osservatorio Astronomico di Roma. He also thanks David Coule for critical
380: comments. Numerical computation of this work was carried out at the Yukawa
381: Institute computer facility.
382: This work was supported in part by the exchanging-researcher project
383: between Japan Society for the Promotion of Science and National Research
384: Council of Italy, 2003, and by Ministry of Education, Culture, Sports,
385: Science and Technology, Grant-in-Aid for Young Scientists (B), 2003,
386: No.\ 00267402.
387:
388: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
389: \begin{thebibliography}{99}
390: %\baselineskip = 18pt
391:
392: \bibitem{GP}Goldwirth D S and Piran T 1989 \prd {\bf 40} 3263 \\
393: Goldwirth D S and Piran T 1990 \prl {\bf 64} 2852 \\
394: Goldwirth D S 1991 \prd {\bf 43} 3204 \\
395: Goldwirth D S and Piran T 1993 {\it Phys. Rep. \ }{\bf 214} 223
396: %
397: \bibitem{KK}Kung J H and Brandenberger H 1989 \prd {\bf 40} 2532 \\
398: Kung J H and Brandenberger H 1990 \prd {\bf 42} 1008 \\
399: Kurki-Suonio H Laguna P and Matzner R A 1993 \prd {\bf 48} 3611
400: \bibitem{CNN}Chiba T Nakao K and Nakamura T 1994 \prd {\bf 49} 3886
401: \bibitem{confGP}
402: Calzetta E and Sakellariadou M 1992 \prd {\bf 45} 2802 \\
403: Shinkai H and Maeda K 1994 \prd {\bf 49} 6367 \\
404: Deruelle N and Goldwirth D S 1995 \prd {\bf 51} 1563 \\
405: Nambu Y and Taruya A 1996 {\it Class. Quntum Grav.\ }{\bf 13} 705 \\
406: Iguchi O and Ishihara H 1997 \prd {\bf 56} 3216
407: %
408: \bibitem{VT}Vachaspati T and Trodden M 1999 \prd {\bf 61} 023502 \\
409: Vachaspati T and Trodden M 1999 {\it Mod. Phys. Lett. A} {\bf 14} 1661
410: \bibitem{Cou}Coule D 2000 \prd {\bf 62} 124010
411: \bibitem{BG}Berera A and Gordon C 2001 \prd {\bf 63} 063505
412: %
413: \bibitem{LV}Linde A 1994 \plb {\bf 327} 208 \\
414: Vilenkin A 1994 \prl {\bf72} 3137
415: \bibitem{SY}Sakai N and Yokoyama J 1999 \plb {\bf456} 113
416: \bibitem{sugra}
417: Izawa K -I Kawasaki M and Yanagida T 1999 {\it Prog. Theor. Phys.}
418: {\bf101} 1129\\
419: Kawasaki M Sakai N Yamaguchi M and Yanagida T 2000 \prd {\bf 62} 123507 \\
420: Kawasaki M and Yamaguchi M 2002 \prd {\bf 65} 103518
421: \bibitem{string}Bento M C and Bertolami O 1996 \plb {\bf 384} 98 \\
422: Ellis J Kaloper N Olive K A and Yokoyama J 1999 \prd{\bf 59} 103503 \\
423: Matsuda T 2003 arXiv hep-ph/0302204
424: \bibitem{BAR}Brandenberger H Alexander S Rozali M 2003 arXiv hep-th/0302160
425: %
426: \bibitem{SSTM}Sakai N Shinkai H Tachizawa T and Maeda K 1996 \prd {\bf 53}
427: 655
428: \bibitem{Sak}Sakai N 1996 \prd {\bf 54} 1548
429: \bibitem{Nak}Nakamura T Maeda K Miyama S and Sasaki M 1981
430: {\it Prog. Theor. Phys.} {\bf 63} 1229 \\
431: Nakamura T Oohara K and Kojima Y 1987 {\it Prog. Theor. Phys. Suppl.}
432: {\bf 90} 1
433: \bibitem{SYM}Sakai N Yokoyama J and Maeda K 1999 \prd {\bf 59} 103504
434: \bibitem{HC}Hawley S H and Choptuik M W 2000 \prd {\bf 62} 104024
435: \bibitem{SNH}Sakai N Nakao and K Harada T 2000 \prd {\bf 61} 127302
436:
437: \end{thebibliography}
438:
439: \newpage
440: \small
441: %\baselineskip = 15pt
442:
443: \begin{figure}
444: \begin{center}
445: (a)\psbox[scale=0.40]{f1a.ps}
446: (b)\psbox[scale=0.40]{f1b.ps}
447: (c)\psbox[scale=0.40]{f1c.ps}
448: \end{center}
449: \end{figure}
450:
451: \newpage
452: \begin{figure}
453: \begin{center}
454: (d)\psbox[scale=0.40]{f1d.ps}
455: (e)\psbox[scale=0.40]{f1e.ps}
456: \end{center}
457: \end{figure}
458:
459: \noindent
460: {\bf FIG. 1}. A solution for the case (A). We show the evolution of the
461: inflaton field, the volume element, and gradient energy
462: $\rho_{grad}=(\partial\Psi/\partial r)^2/(2A^2)$ in (a), (b) and (c),
463: respectively.
464: For kinetic energy $\rho_{kin}=(\partial\Psi/\partial t)^2/2$, (d) and (e)
465: report its early and late behaviors, respectively.
466: The abscissa is the arial radius of the spacetime $B(t,r)r$, normalized by the horizon size $H_0^{-1}$.
467:
468: \begin{figure}
469: \begin{center}
470: (a)\psbox[scale=0.40]{f2a.ps}
471: \end{center}
472: \end{figure}
473:
474: \newpage
475: \begin{figure}
476: \begin{center}
477: (b)\psbox[scale=0.40]{f2b.ps}
478: (c)\psbox[scale=0.40]{f2c.ps}
479: (d)\psbox[scale=0.40]{f2d.ps}
480: \end{center}
481: \end{figure}
482:
483: \noindent
484: {\bf FIG. 2}. A solution for the case (B). We show the evolution of the
485: inflaton field, the volume element, gradient energy and kinetic energy in
486: (a), (b), (c) and (d), respectively.
487:
488: \end{document}
489:
490: