1: \documentclass[prd,preprint,amssymb,amsmath,showpacs]{revtex4}
2: \usepackage{graphicx}
3:
4: \setlength{\voffset}{-1.7cm}
5:
6: \begin{document}
7: %============================================%
8: %<<<<<<<<<<<<<< TITLE PAGE >>>>>>>>>>>>>>>>>>%
9: %============================================%
10: %
11: %=========================================%
12: % preprint numbers and other informations %
13: %=========================================%
14: {\baselineskip0pt
15: \leftline{\large\baselineskip16pt\sl\vbox to0pt{\hbox{\it Department of Physics}
16: \hbox{\it Osaka City University}\vss}}
17: \rightline{\large\baselineskip16pt\rm\vbox to20pt{\hbox{OCU-PHYS-189}
18: \hbox{AP-GR-5}
19: % \hbox{\today}
20: \vss}}
21: }
22: \vskip1.5cm
23: %=========================================%
24: % Title
25: \title{Volume Expansion of Swiss-Cheese Universe}
26: %
27: % Authors
28: \author{Hiroshi Kozaki}
29: \email{furusaki@sci.osaka-cu.ac.jp}
30: %
31: \author{Ken-ichi Nakao}
32: \email{knakao@sci.osaka-cu.ac.jp}
33: %
34: % Affiliation
35: \affiliation{%
36: Department~of~Physics,~Graduate~School~of~Science,~Osaka~City~University,
37: Osaka 558-8585, Japan%
38: }
39: %
40: % Abstract
41: \begin{abstract}
42: In order to investigate the effect of inhomogeneities on
43: the volume expansion of the universe, we study
44: modified Swiss-Cheese universe model.
45: Since this model is an exact solution of Einstein equations,
46: we can get an insight into non-linear dynamics of
47: inhomogeneous universe from it. We find that inhomogeneities make the
48: volume expansion slower than that of the
49: background Einstein-de Sitter universe when those can be regarded as
50: small fluctuations in the background universe.
51: This result is consistent with the previous studies based on the
52: second order perturbation analysis.
53: On the other hand, if the inhomogeneities can not be treated as
54: small perturbations,
55: the volume expansion of the universe depends on the type of
56: fluctuations. Although the volume expansion rate approaches
57: to the background value
58: asymptotically, the volume itself can be finally arbitrarily smaller
59: than the background one and can be larger than that of the background
60: but there is an upper bound on it.
61: \end{abstract}
62: %
63: \pacs{04.25.Nx,04.30.Db,04.20.Dw}
64: \maketitle
65:
66: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
67: \section{INTRODUCTION}
68: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
69:
70: The standard Big Bang scenario is based on an assumption of the
71: homogeneous and isotropic distribution of matter and
72: radiation. This assumption then leads to the Robertson-Walker
73: spacetime geometry and the Friedmann-Lema\^{\i}tre (FL)
74: universe model through the Einstein equations.
75: This model has succeeded in explaining various important observational
76: facts: Hubble's expansion law, the content of light elements and
77: the isotropic cosmic microwave background radiation
78: (CMBR)\cite{ref:Weinberg}.
79:
80: The CMBR conversely gives a strong observational basis for the
81: assumption of homogeneity and isotropy of our universe by its highly
82: isotropic distribution together with the Copernican principle;
83: we know that our universe was highly
84: isotropic and homogeneous at least on the last scattering
85: surface where CMBR comes from\cite{ref:Smoot}. Hence, in the early stage of
86: our universe, the linear perturbation analysis in the
87: FL universe model is a powerful tool
88: to investigate the dynamical evolution of our universe\cite{ref:KS84}.
89:
90: In order to perform the perturbation analysis,
91: we need an appropriate background universe model,
92: i.e., information of the Hubble parameter, the density
93: parameter and further the equation of state of the matter,
94: radiation and so on, in the real universe.
95: To fix the background universe, we use the observational data
96: in the neighborhood of our galaxy. Especially the Hubble parameter
97: is determined from the data about the distance-redshift relation
98: within $100h^{-1}$Mpc except for the type
99: Ia supernova\cite{{Schmidt:1998ys},{Perlmutter:1999np}}.
100: However, the universe in a region within $100h^{-1}$Mpc is
101: highly inhomogeneous and hence there are non-trivial prescriptions
102: to identify the present inhomogeneous universe to the
103: homogeneous and isotropic FL universe model.
104: If those procedure are not appropriate, we might miss
105: finding the correct background universe.
106:
107: It is often stated that the spatially averaged observational data
108: in the vicinity of our galaxy are recognized as those of the
109: background FL universe model.
110: The Hubble parameter determined by the observed
111: distance-redshift relation in our universe
112: is regarded as the expansion rate of the volume
113: of the region co-moving to matter.
114: There are several researches for the effects of inhomogeneities
115: on the volume expansion of the
116: universe\cite{ref:futamase89,ref:futamase97,ref:Tomita,ref:Russ,
117: ref:Mukhanov,ref:Abramo,ref:Nambu,ref:Nambu-2,ref:Nambu-3,ref:Geshnizjani}.
118: Especially, Nambu applied the renormalization
119: group method to the second order
120: cosmological perturbation theory and claimed that the expansion of
121: the dust filled universe is decelerated
122: by the inhomogeneities\cite{ref:Nambu}.
123:
124: In the real universe, this back reaction effect might be very
125: small\cite{ref:Russ}.
126: However in order to get deeper
127: insight into the dynamics of the inhomogeneous universe,
128: we will consider the situation in which the
129: back reaction of the inhomogeneities seems to be effective.
130: For this purpose, we consider the Swiss-Cheese
131: universe model and investigate the volume expansion in it.
132: The original Swiss-Cheese universe model is
133: constructed by choosing non-overlapping spherical regions in
134: the background homogeneous and isotropic dust filled
135: universe and then replacing
136: these regions by the Schwarzschild space-time whose mass parameter
137: is identical with the ``gravitational''
138: mass of the dust fluid in the removed
139: region. On the other hand,
140: in this article, we consider a modified version;
141: we first remove spherical regions from the homogeneous and isotropic
142: dust filled universe and then fill these regions with
143: spherically symmetric but inhomogeneous dust balls.
144: A spherically symmetric inhomogeneous dust ball is described
145: by the Lema\^{\i}tre-Tolman-Bondi (LTB) solution
146: which is an exact solution of the Einstein equations, and hence
147: by this procedure, we obtain an exact solution of Einstein equations,
148: which represents an inhomogeneous universe.
149: Using this solution, we can study non-linear effects
150: of inhomogeneities on the volume expansion of the universe
151: without use of perturbation analysis.
152:
153: In the LTB solution, shell crossing singularities are generic.
154: Since the LTB solution is no longer valid
155: after the occurrence of the shell crossing,
156: we need to change the treatment if it occurs.
157: As a crude approximation to describe
158: the dynamics after the shell crossing, we adopt a model
159: in which the shell crossing
160: region is replaced by a spherical dust shell.
161:
162: This article is organized as follows.
163: In section \ref{sec:SC}, we explain how to construct
164: modified Swiss-Cheese universe models which are studied in this article.
165: We investigate the volume expansion rate in the case of small
166: perturbations in section \ref{sec:perturb}
167: and in the highly inhomogeneous case in \ref{sec:non-linear}.
168: In section \ref{sec:non-linear},
169: an alternative model is also constructed which describes the universe after
170: the shell crossing and investigate the dynamics of this model.
171: Finally, section \ref{sec:summary} is devoted to summary and
172: discussion.
173:
174: We use the units in which $c=G=1$ throughout the paper.
175:
176:
177: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
178: \section{MODIFIED SWISS CHEESE UNIVERSE MODEL} \label{sec:SC}
179: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
180:
181: In this section, we give a prescription to construct
182: MSC universe model.
183: First we consider a Einstein-de Sitter (EdS) universe and then remove
184: spherical regions from it; these removed regions should not overlap
185: with each other. Next these regions filled with inhomogeneous dust
186: balls with the same radii and the same gravitational mass
187: as those of the removed homogeneous dust balls.
188: In this MSC universe model, each inhomogeneous region
189: is described by the Lema\^{\i}tre-Tolman-Bondi (LTB) solution
190: which is an exact solution of Einstein equations.
191:
192: LTB solution describes the dust filled spherically symmetric spacetime.
193: Adopting synchronous and co-moving coordinate system,
194: the line element is written as
195: \begin{align}
196: ds^{2}=&-dt^{2}+\gamma_{ij}dx^{i}dx^{j} \notag \\
197: =&-dt^{2}+\frac{Y'{}^{2}(t,\chi)}{1-\chi^{2}k(\chi)}d\chi^{2}
198: +Y^{2}(t,\chi)(d\theta^{2}+\sin^{2}\theta d\varphi^{2}),
199: \label{eq:line-element}
200: \end{align}
201: where the prime ${}'$ denotes the differentiation
202: with respect to the radial coordinate $\chi$.
203: In this coordinate system, components
204: of 4-velocity $u^{a}$ of a dust fluid element are
205: \begin{equation}
206: u^{a}=(1,~0,~0,~0).
207: \end{equation}
208: The stress-energy tensor $T_{ab}$ is then given by
209: \begin{equation}
210: T_{ab}=\rho(t,\chi) \delta^{0}_{a} \delta^{0}_{b},
211: \end{equation}
212: where $\rho(t,\chi)$ is the rest mass density of the dust.
213:
214: Einstein equations lead to the equations for the areal radius $Y(t,\chi)$
215: and the rest mass density $\rho(t,\chi)$ of the dust;
216: \begin{align}
217: \dot{Y}^{2} =& -\chi^{2}k(\chi) + \frac{2M(\chi)}{Y},\label{eq:einstein} \\
218: \rho =& \frac{M'(\chi)}{4\pi Y'Y^{2}},\label{eq:density}
219: \end{align}
220: where $k(\chi)$ and $M(\chi)$ are arbitrary functions
221: and the dot $\dot{}$ denotes the differentiation
222: with respect to $t$.
223:
224: We set $M(\chi)$ as
225: \begin{equation}
226: M(\chi)=\frac{4\pi\rho_{0}}{3}\chi^{3}, \label{eq:mass-form}
227: \end{equation}
228: where $\rho_{0}$ is a non-negative arbitrary constant.
229: The above choice of $M(\chi)$ does not loose any generality.
230: Eqs.~(\ref{eq:line-element})-(\ref{eq:density}) are
231: invariant for the rescaling of the radial coordinate $\chi$,
232: \begin{equation}
233: \chi \rightarrow \tilde{\chi}=\tilde{\chi}(\chi).
234: \end{equation}
235: Considering this property,
236: we can choose above form of $M(\chi)$ as long as $\rho Y'>0$.
237:
238:
239: The solutions of eq.~(\ref{eq:einstein}) are given as follows:
240:
241: \noindent
242: In the region where $k(\chi)>0$,
243: \begin{align}
244: Y=&\frac{4\pi\rho_{0}}{3k}(1-\cos\eta)\chi, \label{eq:k>0}\\
245: t-t_{0}(\chi)=&\frac{4\pi\rho_{0}}{3k^{3/2}}(\eta-\sin\eta);
246: \label{eq:t-solution}
247: \end{align}
248: in the region where $k(\chi)=0$,
249: \begin{equation}
250: Y=\Bigl[ 6\pi\rho_{0} \bigl\{ t-t_{0}(\chi) \bigr\}^{2} \Bigr]^{1/3}\chi;
251: \label{eq:k=0}
252: \end{equation}
253: in the region where $k(\chi)<0$,
254: \begin{align}
255: Y=&\frac{4\pi\rho_{0}}{3|k|}(\cosh\eta-1)\chi, \label{eq:k<0}\\
256: t-t_{0}(\chi)=&\frac{4\pi\rho_{0}}{3|k|^{3/2}}(\sinh\eta-\eta),%\label{}
257: \end{align}
258: where $t_{0}(\chi)$ is an arbitrary function.
259: Note that $t_{0}(\chi)$ is the time when a shell focusing
260: singularity appears, where `shell focusing singularity' means
261: $Y=0$ for $\chi>0$ and $Y'=0$ at $\chi=0$.
262: In this article, we consider a region of $t>t_{0}$,
263: and hence the time $t=t_{0}$ corresponds to
264: the Big Bang singularity. Here, we focus on a case
265: of $t_{0}=0$, i.e., simultaneous Big Bang.
266: For simplicity, we consider the simplest version of MSC universe
267: models shown in fig.~\ref{fig:MSC};
268: there is only one inhomogeneous spherical region at the center
269: in each identical cubic region $\Omega$.
270: We focus on only one cubic co-moving
271: region $\Omega$ (fig.~\ref{fig:MSC-2}).
272: \begin{figure}
273: \begin{center}
274: \includegraphics{MSC.eps}
275: \end{center}
276: \caption{The MSC universe model.
277: Each shaded region represents the inhomogeneity and is described by
278: LTB solution.
279: }
280: \label{fig:MSC}
281: \end{figure}
282: %
283: \begin{figure}
284: \begin{center}
285: \includegraphics{MSC-2.eps}
286: \end{center}
287: \caption{One cubic region $\Omega$
288: of the MSC universe model.
289: $\ell$ and $\chi_{\rm sc}$ are the co-moving scales of this
290: cubic region and the inhomogeneous region respectively.}
291: \label{fig:MSC-2}
292: \end{figure}
293: In this article, we consider a following model.
294: Assuming $0<\chi_{1}<\chi_{2}<\chi_{3}<\chi_{\rm sc}$,
295: \begin{equation}
296: k(\chi)=
297: \begin{cases}
298: %
299: k_{0} & \text{for~~~} 0\leq\chi<\chi_{1}\\
300: %
301: \dfrac{k_{0}}{2\chi^{2}}
302: \left\{
303: \dfrac{(\chi^{2}-\chi_{2}^{2})^{2}}{\chi_{1}^{2}-\chi_{2}^{2}}
304: +\chi_{1}^{2}+\chi_{2}^{2}
305: \right\} & \text{for~~~} \chi_{1}\leq\chi<\chi_{2} \rule{0pt}{20pt} \\
306: %
307: \dfrac{k_{0}}{2\chi^{2}}
308: \left(\chi_{1}^{2}+\chi_{2}^{2}\right)
309: & \text{for~~~} \chi_{2}\leq\chi<\chi_{3} \rule{0pt}{26pt}\\
310: %
311: \dfrac{k_{0}}{2\chi^{2}}\left(\chi_{1}^{2}+\chi_{2}^{2}\right)
312: \left\{
313: \left(\dfrac{\chi^{2}-\chi_{3}^{2}}
314: {\chi_{\rm sc}^{2}-\chi_{3}^{2}}
315: \right)^{2}-1
316: \right\}^{2}
317: & \text{for~~~} \chi_{3}\leq\chi<\chi_{\rm sc} \rule{0pt}{26pt}
318: \end{cases},
319: \label{eq:region-2-3}
320: \end{equation}
321: where $k_{0}$ is constant. In order to
322: guarantee $1-\chi^{2}k>0$, the following inequality should hold,
323: \begin{equation}
324: \kappa:=\dfrac{k_{0}}{2}(\chi_{1}^{2}+\chi_{2}^{2})<1.
325: \label{eq:kappa-def}
326: \end{equation}
327:
328: We consider two cases; one is the case
329: of $k_{0}>0$ and the other is the case of $k_{0}<0$.
330: Then we investigate the volume expansion rate
331: of a cubic co-moving region $\Omega$.
332: The volume $V$ is defined by
333: \begin{equation}
334: V(t):=\int_{\Omega}\sqrt{\gamma}d^{3}x, \label{eq:volume-def}
335: \end{equation}
336: where $\gamma$ is the determinant of the spatial metric $\gamma_{ij}$.
337: The volume expansion rate is defined as ${\dot V}/V$.
338:
339: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
340: \section{The Case of Small Fluctuations} \label{sec:perturb}
341: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
342:
343: In a region where
344: \[
345: 0< \frac{9t|k|^{3/2}}{2\pi\rho_{0}}\ll 1
346: \]
347: is satisfied, the areal radius
348: $Y(t,\chi)$ is written in the form of power series as
349: \begin{equation}
350: Y(t,\chi)=a(t)\chi
351: \left(
352: 1-\frac{1}{20}\epsilon-\frac{3}{2800}\epsilon^{2}
353: \right)+O(\epsilon^{3}) , \label{eq:perturbation}
354: \end{equation}
355: where
356: \begin{equation}
357: \epsilon(t,\chi):=\left( \frac{9t}{2\pi\rho_{0}} \right)^{2/3}k
358: =\left( \frac{12t}{M(\chi)} \right)^{2/3}\chi^{2}k,
359: \end{equation}
360: and
361: \begin{equation}
362: a(t):=(6\pi\rho_{0}t^{2})^{1/3}. \label{eq:s-factor}
363: \end{equation}
364: Further, we consider the case of $|k|\chi^{2}\ll1$, which has been studied
365: by Nambu by the second order perturbation analysis.
366: The components of the metric tensor are written as
367: \begin{align}
368: g_{\chi\chi}
369: &= a^{2}
370: \Bigg[
371: 1
372: +\left(\dfrac{M}{12t}\right)^{2/3}\epsilon
373: -\frac{1}{10}\frac{d}{d\chi}(\chi\epsilon)
374: +O(\epsilon^{2})
375: \Bigg], \label{eq:metric-perturbation-1}\\
376: g_{\theta\theta}
377: &= a^{2}\chi^{2}
378: \Bigg[
379: 1-\frac{1}{10}\epsilon+O(\epsilon^{2})
380: \Bigg], \label{eq:metric-perturbation-2}\\
381: g_{\varphi\varphi}
382: &= a^{2}\chi^{2}\sin^{2}\theta
383: \Bigg[
384: 1 -\frac{1}{10}\epsilon+O(\epsilon^{2})
385: \Bigg]. \label{eq:metric-perturbation-3}
386: \end{align}
387: From the above equations, it is easily seen that
388: in the limit $\epsilon\rightarrow 0$ with $t$ fixed,
389: the metric tensor becomes that of EdS. Since
390: the outside region is EdS universe,
391: $\epsilon$ should vanish at the boundary $\chi=\chi_{\rm sc}$
392: by the continuity of the metric tensor.
393:
394: To compare our result with the study by Nambu\cite{ref:Nambu},
395: we impose conditions that spatial averages of
396: the Cartesian components of the metric tensor and of the density agree
397: with those of the \textit{background} EdS universe up to the first
398: order of $\epsilon$.
399:
400: The spatial average of a quantity $F$ is defined as follows:
401: \begin{equation}
402: \langle F \rangle := \left(\int_{\Omega}d^{3}x\right)^{-1}
403: \int_{\Omega} F d^{3}x.
404: \end{equation}
405:
406: We consider a Cartesian coordinate system $(x,y,z)$ which
407: is related to the spherical polar coordinate system
408: $(\chi,\theta,\varphi)$ in the ordinary manner as
409: \begin{align}
410: x&=\chi\sin\theta\cos\varphi, \notag \\
411: y&=\chi\sin\theta\sin\varphi, \notag \\
412: z&=\chi\cos\theta. \nonumber
413: \end{align}
414: Then the spatial averages of Cartesian components of the spatial
415: metric $\gamma_{ij}$ are obtained as
416: \begin{align}
417: \langle \gamma_{ij} \rangle
418: &=a^{2}\left[
419: 1+\dfrac{4\pi}{3\ell^{3}}
420: \int_{0}^{\chi_{\rm sc}}
421: \left\{
422: \left( \dfrac{M}{12t} \right)^{2/3}\epsilon\chi^{2}
423: -\dfrac{1}{10}\dfrac{d}{d\chi}(\chi^{3}\epsilon)
424: \right\}
425: d\chi
426: \right]\delta_{ij}
427: +O(\epsilon^{2}) \notag \\
428: &=(6\pi\rho_{0}t^{2})^{2/3}
429: \left(
430: 1+\dfrac{4\pi}{3\ell^{3}}\int_{0}^{\chi_{\rm sc}}k\chi^{4}d\chi
431: \right)\delta_{ij}+O(\epsilon^{2}),
432: \end{align}
433: where we have used $\epsilon=0$ at $\chi=\chi_{\rm sc}$.
434: Therefore we should define the scale factor $a_{\textsc{b}}(t)$
435: of the background EdS universe as
436: \begin{equation}
437: a_{\textsc{b}}(t)
438: :=a(t)\left(
439: 1+\dfrac{4\pi}{3\ell^{3}}\int_{0}^{\chi_{\rm sc}}k\chi^{4}d\chi
440: \right)^{1/2}, \label{eq:linear-s-factor}
441: \end{equation}
442: The above equation means that although the outside homogeneous region
443: has the same geometry of EdS universe, it does not agree with
444: {\it the background} EdS universe as long as
445: \begin{equation}
446: \int_{0}^{\chi_{\mbox{\scriptsize{sc}}}}k\chi^{4}d\chi \neq 0.
447: \label{eq:condition-k}
448: \end{equation}
449:
450: The rest mass density $\rho$ of the dust is written
451: in the form of a power series with respect to $\epsilon$ as
452: \begin{equation}
453: \rho=\dfrac{1}{6\pi t^{2}}\left\{1+\dfrac{1}{20\chi^{2}}
454: \dfrac{d}{d\chi}(\chi^{3}\epsilon)\right\}+O(\epsilon^{2}).
455: \end{equation}
456: The spatial average of $\rho$ is obtained as
457: \begin{equation}
458: \langle\rho\rangle =\dfrac{1}{6\pi t^{2}}+O(\epsilon^{2}).
459: \end{equation}
460: Hence the background energy density $\rho_{\textsc{b}}$ is defined by
461: \begin{equation}
462: \rho_{\textsc{b}}(t):=\dfrac{1}{6\pi t^{2}}. \label{eq:linear-density}
463: \end{equation}
464: Here note that the background Hubble equation
465: \begin{equation}
466: \left(\dfrac{{\dot a}_{\textsc{b}}}{a_{\textsc{b}}}\right)^{2}
467: =\dfrac{8\pi}{3}\rho_{\textsc{b}}, \label{eq:Hubble-eq}
468: \end{equation}
469: holds.
470:
471: By eq.~(\ref{eq:perturbation}), the 3-dimensional volume element
472: $\sqrt{\gamma}$ is written as
473: \begin{multline}
474: \sqrt{\gamma}
475: = a^{3}\chi^{2}\sin\theta
476: \biggl[
477: 1
478: -\frac{1}{20\chi^{2}}\frac{d}{d\chi}(\chi^{3}\epsilon)
479: +\frac{1}{2}\chi^{2}k(\chi)
480: +\frac{1}{700\chi^{2}}\frac{d}{d\chi}(\chi^{3}\epsilon^{2}) \\
481: -\frac{1}{40}k(\chi)\frac{d}{d\chi}(\chi^{3}\epsilon)
482: +\frac{3}{8}\chi^{4}k^{2}(\chi)
483: \biggr]+O(\epsilon^{3}).
484: \end{multline}
485: Using the above equation, the volume $V$ defined by
486: eq.~(\ref{eq:volume-def}) is obtained as
487: \begin{equation}
488: V(t)= a_{\textsc{b}}^{3}(t)
489: \left(
490: \ell^{3}+V_{1}+V_{2}t^{2/3}
491: \right)+O(\epsilon^{3}), \label{eq:linear-volume}
492: \end{equation}
493: where
494: \begin{align}
495: V_{1}&:= \dfrac{3\pi}{2}
496: \int_{0}^{\chi_{\rm sc}}
497: \chi^{6}k^{2}
498: d\chi
499: -\dfrac{2\pi^{2}}{3\ell^{6}}
500: \left(
501: \int_{0}^{\chi_{\rm sc}}
502: k\chi^{4}
503: d\chi
504: \right)^{2}, \\
505: V_{2}&:=-\dfrac{\pi}{20}
506: \left( \frac{9}{2\pi\rho_{0}} \right)^{2/3}
507: \int_{0}^{\chi_{\rm sc}} k^{2}\chi^{4} d\chi <0.
508: \end{align}
509: In eq.~(\ref{eq:linear-volume}), $a_{\textsc{b}}^{3}\ell^{3}$
510: is the 3-dimensional volume measured by background EdS
511: geometry and extra terms come form inhomogeneities.
512: These terms do not include first order perturbations,
513: but come from second order perturbations.
514:
515: The volume expansion rate is given by
516: \begin{equation}
517: \frac{\dot{V}}{V}
518: = 3\frac{\dot{a}_{\textsc{b}}}{a_{\textsc{b}}}
519: +\frac{2V_{2}}{3\ell^{3}}t^{-1/3}+O(\epsilon^{3}).
520: \end{equation}
521: The first term of the R.H.S. in the above equation
522: corresponds to the background part.
523: On the other hand, the second term implies
524: that the back-reaction of inhomogeneities decelerates
525: the volume expansion.
526: It is worthwhile to note that this result does not depend on
527: the detailed functional form of $k(\chi)$.
528:
529: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
530: \section{The Case of Non-Linear Fluctuations}\label{sec:non-linear}
531: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
532:
533: In this section, we study the cases where $|k|\chi^{2}$ is
534: not necessarily much smaller than unity.
535: As in the case treated in the previous section, in
536: order to specify an effect due to inhomogeneity,
537: we need a background homogeneous
538: cubic region to be compared with an inhomogeneous one.
539: However the background homogeneous universe
540: introduced in the previous section
541: is not appropriate for the non-linear case; for example,
542: too small $k$ makes the background scale factor $a_{\textsc{b}}$
543: defined in eq.~(\ref{eq:linear-s-factor}) negative.
544: In order to introduce an appropriate background,
545: we consider the rest mass $M_{\rm R}$ defined by
546: \begin{align}
547: M_{\rm R}
548: &:= \int_{\Omega} \rho u^{0}\sqrt{-g} d^{3}x \notag \\
549: &= 4\pi
550: \int_{0}^{\chi_{\rm sc}}
551: \dfrac{\rho Y'Y^{2}}{\sqrt{1-\chi^{2}k}}
552: d\chi
553: +\rho_{0}
554: \left(
555: \ell^{3}-\dfrac{4\pi}{3}\chi_{\rm sc}^{3}
556: \right) \notag \\
557: &= \rho_{0}\ell^{3}
558: \left\{
559: 1+\dfrac{4\pi}{\ell^{3}}
560: \int_{0}^{\chi_{\rm sc}}
561: \left(\dfrac{1}{\sqrt{1-\chi^{2}k}}-1\right)\chi^{2}
562: d\chi
563: \right\}, \label{eq:M-def}
564: \end{align}
565: where $g$ is the determinant of the metric tensor of spacetime.
566: $M_{\rm R}$ is a conserved quantity by virtue of the continuity
567: of the rest mass density, $\partial_{a}(\rho u^{a}\sqrt{-g})=0$, where
568: $\partial_{a}$ is a partial derivative.
569: We introduce a background as a cubic region with the same rest
570: mass as the corresponding inhomogeneous cubic region.
571: Note that in general, the rest mass of the dust $M_{\rm R}$ in a cubic
572: co-moving region disagrees with that of the original EdS universe.
573: Hence a cubic region of the original EdS universe is not background.
574:
575: The rest mass density $\rho_{\textsc{b}}$ and
576: scale factor $a_{\textsc{b}}$ of the background
577: are introduced in the following manner,
578: \begin{equation}
579: \rho_{\textsc{b}}a_{\textsc{b}}^{3}\ell^{3}:=M_{\rm R}. \label{eq:relation}
580: \end{equation}
581: Even if we specify $M_{\rm R}$, $\rho_{\textsc{b}}$ and $a_{\textsc{b}}$
582: are not fixed completely; we need one more condition.
583: Here we impose a condition in which the volume $V(t)$
584: approaches to the volume
585: $a_{\textsc{b}}^{3}(t)\ell^{3}$ of the background
586: cubic region for $t\rightarrow0$.
587: In the limit of $t\rightarrow0$, the volume $V$ behaves as
588: \begin{align}
589: V&= a^{3}\ell^{3}
590: +4\pi
591: \int_{0}^{\chi_{\rm sc}}
592: \left(
593: \dfrac{Y'Y^{2}}{\sqrt{1-\chi^{2}k}}-a^{3}\chi^{2}
594: \right)
595: d\chi \notag \\
596: &\longrightarrow
597: a^{3}\ell^{3}
598: \left\{
599: 1
600: +\dfrac{4\pi}{\ell^{3}}
601: \int_{0}^{\chi_{\rm sc}}
602: \left( \dfrac{1}{\sqrt{1-\chi^{2}k}}-1 \right)\chi^{2}
603: d\chi
604: \right\}
605: \end{align}
606: Hence the background scale factor $a_{\textsc{b}}$ is given by
607: \begin{equation}
608: a_{\textsc{b}}(t):=a\left\{1+\dfrac{4\pi}{\ell^{3}}
609: \int_{0}^{\chi_{\rm sc}}\left(\dfrac{1}{\sqrt{1-\chi^{2}k}}-1\right)
610: \chi^{2}d\chi\right\}^{1/3}. \label{eq:a-def}
611: \end{equation}
612: Here note that in the case of $\chi^{2}|k|\ll1$,
613: the above definition of $a_{\textsc{b}}$ agrees with
614: eq.~(\ref{eq:linear-s-factor}) up to the first order of $\chi^{2}k$.
615: From eqs.~(\ref{eq:M-def}), (\ref{eq:relation}) and (\ref{eq:a-def}),
616: we find that the background rest mass density $\rho_{\textsc{b}}$ is
617: completely the same as eq.~(\ref{eq:linear-density}).
618: We can easily see that the background Hubble equation
619: (\ref{eq:Hubble-eq}) also holds.
620:
621:
622: From eq.~(\ref{eq:density}), we can easily see that if $Y'$ vanishes,
623: the rest mass density $\rho$ becomes infinite and hence a singularity
624: forms there. This is called shell crossing singularity.
625: Hellaby and Lake showed that
626: a necessary and sufficient condition for the appearance of a shell
627: crossing singularity is \cite{ref:Hellaby-Lake}
628: \begin{align}
629: k'&> 0~~~~~{\rm for~the~region}~~k>0, \\
630: \left( \chi^{2}k \right)'&> 0~~~~~{\rm for~the~region}~~k\leq0.
631: \label{eq:condition}
632: \end{align}
633: In the case of $k_{0}>0$, the first condition does not hold
634: and hence a shell crossing singularity does not appear. On the other
635: hand, in the case of $k_{0}<0$,
636: shell crossing singularities $Y'=0$ necessarily
637: appear since $(\chi^{2}k)'>0$ in the region
638: $\chi_{3}\leq\chi<\chi_{\rm sc}$.
639:
640: \subsection{The Case of $k_{0}>0$}
641:
642: To estimate the volume $V$, we rewrite it in the form,
643: \begin{equation}
644: V=a^{3}\left(\ell^{3}-\dfrac{4\pi}{3}\chi_{\rm sc}^{3}\right)
645: +4\pi\int_{0}^{Y_{\rm sc}}\dfrac{Y^{2}dY}{\sqrt{1-\chi^{2}k}},
646: \label{eq:volume-2}
647: \end{equation}
648: where $Y_{\rm sc}:=a(t)\chi_{\rm sc}$.
649: From eq.~(\ref{eq:k>0}), we can see that $Y$ vanishes
650: at $\eta=2\pi$ by gravitational collapse, and hence by
651: substituting $\eta=2\pi$ into eq.~(\ref{eq:t-solution}), the singularity
652: formation time $t=t_{\rm sg}(\chi)$ is obtained as
653: \begin{equation}
654: t=t_{\rm sg}(\chi)=\dfrac{8\pi^{2}\rho_{0}}{3k^{3/2}}.
655: \end{equation}
656: Denoting the inverse function of $t_{\rm sg}(\chi)$ by
657: $\chi_{\rm sg}(t)$, the region of
658: $0\leq \chi \leq\chi_{\rm sg}(t)<\chi_{\rm sc}$
659: has already collapsed at time $t$ larger than
660: $8\pi^{2}\rho_{0}/3k_{0}^{3/2}$.
661: $\chi_{\rm sg}(t)$ approaches to $\chi_{\rm sc}$ for $t\rightarrow
662: \infty$ asymptotically. Here note that in the
663: integrand of eq.~(\ref{eq:volume-2}),
664: $\chi=\chi_{\rm sg}(t)$ at $Y=0$
665: and $\chi=\chi_{\rm sc}$ at $Y=Y_{\rm sc}$. Since $\chi^{2}k(\chi)$ is
666: decreasing function with respect to $\chi$ in the region
667: $\chi_{3}<\chi<\chi_{\rm sc}$
668: and vanishes just at $\chi=\chi_{\rm sc}$, we find
669: that $0 < \chi^{2}k(\chi)\leq \chi_{\rm sg}^{2}k(\chi_{\rm sg})$ holds
670: in the integrand at sufficiently large $t$.
671: This means that $\chi^{2}k(\chi)$ also approaches to zero asymptotically,
672: since $\chi_{\rm sg}^{2}k(\chi_{\rm sg})\rightarrow
673: \chi_{\rm sc}^{2}k(\chi_{\rm sc})=0$ for $t\rightarrow\infty$, and thus
674: \begin{equation}
675: V\longrightarrow a^{3}\left(\ell^{3}-\dfrac{4\pi}{3}\chi_{\rm sc}^{3}\right)
676: +4\pi\int_{0}^{Y_{\rm sc}}Y^{2}dY=a^{3}\ell^{3}. \label{eq:volume-3}
677: \end{equation}
678: This equation means that
679: the volume expansion rate approaches to the background value
680: asymptotically, i.e.,
681: \begin{equation}
682: \dfrac{\dot{V}}{V}\longrightarrow
683: 3\dfrac{\dot{a}}{a}=3\dfrac{\dot{a}_{\textsc{b}}}{a_{\textsc{b}}}
684: \text{~~~for~} t\rightarrow\infty.
685: \end{equation}
686: However the volume itself may approach to much different value from
687: the background one $a_{\textsc{b}}^{3}\ell^{3}$.
688: $\sqrt{1-\chi^{2}k}$ can be made arbitrarily smaller than unity
689: in the region $\chi_{2}\leq\chi<\chi_{3}$; in the limit of
690: $\kappa\rightarrow1$, $\sqrt{1-\chi^{2}k}\rightarrow 0$ in this region
691: (see eqs.~(\ref{eq:region-2-3}) and
692: (\ref{eq:kappa-def})).
693: Thus, if we set $\kappa$ to be very close to unity,
694: we obtain
695: \begin{equation}
696: a_{\textsc{b}} \sim a\left(\dfrac{4\pi}{\ell^{3}}
697: \int_{0}^{\chi_{\rm sc}}
698: \dfrac{\chi^{2}d\chi}{\sqrt{1-\chi^{2}k}}\right)^{1/3}
699: \gg a.%\label{}
700: \end{equation}
701: In this case, the volume $V$ approaches to the value
702: much smaller than
703: the background one, asymptotically.
704: The volume expansion is also much different
705: from that of the background in the intermediate stage
706: (see fig.~\ref{fig:positive}).
707: \begin{figure}[ht]
708: \begin{center}
709: \includegraphics[keepaspectratio]{volume.eps}
710: \caption{ \label{fig:positive}%
711: Volume expansion with positive $k(\chi)$.
712: The dotted line is the temporal evolution of the co-moving volume
713: measured by the background EdS geometry
714: and the dashed line is that measured by the outer EdS geometry.
715: The time is set to unity,
716: when the scale of the spherical inhomogeneous region $a(t)\chi_{\rm sc}$
717: agrees with that of the horizon scale $a(t)/\dot{a}(t)$.
718: The scale factor at $t=1$ is also set to unity
719: and the co-moving scale $\ell$ is set to $4\chi_{\rm sc}$.
720: Hence the co-moving volume measured by the outer EdS geometry at $t=1$
721: is $\ell^3(=6^3=216)$.}
722: \end{center}
723: \end{figure}
724: \subsection{The Case of $k_{0}<0$}
725:
726: As mentioned in the above, shell
727: crossing singularities $Y'=0$ necessarily appear in this model.
728: Before it, the volume expansion is shown in fig.~\ref{fig:ratio.before}.
729: For $k(\chi)<0$, background scale factor $a_{\textsc{b}}$ cannot differ
730: from original one very much.
731: We plot the ratio of the volume $V(t)$ to
732: $a^{3}\ell^{3}$ and $a_{\textsc{b}}^{3}\ell^{3}$.
733: We find that the inhomogeneities decelerate the volume expansion
734: before shell crossing singularity appears.
735:
736:
737: \begin{figure}[ht]
738: \begin{center}
739: \includegraphics[keepaspectratio]{ratio-1.eps}
740: \caption{\label{fig:ratio.before}%
741: Temporal evolution of the ratios before shell crossing
742: singularity appears.
743: The time is set to unity,
744: when the scale of the spherical inhomogeneous region $a(t)\chi_{\rm sc}$
745: agrees with that of the horizon scale $a(t)/\dot{a}(t)$.
746: Dashed line corresponds to the shell crossing time.
747: $V_{\textsc{msc}}$ is the volume of the co-moving cubic region
748: $\Omega$ in modified Swiss Cheese universe.
749: $V_{\textsc{b}}$ is the volume of $\Omega$ measured by the background
750: EdS geometry, i.e., $V_{\textsc{b}}=a_{\textsc{b}}^{3}\ell^{3}$.
751: $V_{\textsc{out}}$ is the volume of $\Omega$ measured by the outer
752: EdS geometry, i.e., $V_{\textsc{out}}=a^{3}\ell^{3}$.}
753: \end{center}
754: \end{figure}
755:
756: Here, we investigate the volume expansion rate
757: after the appearance of this shell crossing singularity.
758: The structure formed by the shell crossing
759: depends on what is approximated by the dust matter.
760: In case that the dust matter is extremely cold fluid,
761: a shock wave will form after the shell crossing.
762: If the dust matter consists of collisionless particles,
763: a spherical wall will form. When the width of the shock wave or
764: of the wall is much smaller than the radius of it, we can treat the shock
765: front or the wall as a timelike singular hypersurface,
766: where `timelike' means that the unit normal vector $n^{a}$
767: to the hypersurface is spacelike, i.e., $n^{a}n_{a}=1$.
768:
769: A timelike singular hypersurface is characterized
770: by its surface-stress-energy tensor defined by
771: \begin{equation}
772: S_{ab}:=\lim_{\varepsilon\rightarrow0}\int_{-\varepsilon}^{\varepsilon}
773: T_{cd}h_{a}^{c}h_{b}^{d}dx
774: \end{equation}
775: where $x$ is a Gaussian coordinate ($x=0$ on the hypersurface) in the
776: direction of the normal vector $n^{a}$,
777: and $h_{a}^{c}:=\delta_{a}^{c}-n_{a}n^{c}$ is a projection operator.
778: A timelike singular hypersurface with a surface-stress-energy tensor
779: of the form
780: \begin{equation}
781: S_{ab}=\sigma v_{a}v_{b},
782: \end{equation}
783: is called a world sheet generated by a trajectory of a dust shell,
784: where $\sigma$ is the surface-energy density and $v_{a}$ is
785: the 4-velocity of an infinitesimal surface element of the
786: dust shell. Hereafter we focus on this case.
787:
788: In order to get an insight into the dynamics after shell
789: crossing, we study the volume expansion of a cubic region
790: with a spherically symmetric dust shell.
791: In the case of $k_{0}<0$, $k(\chi)$ is negative in the
792: LTB region; $(\chi^{2}k)'\leq0$ in the inner region $0\leq\chi\leq\chi_{3}$,
793: while $(\chi^{2}k)'>0$ in the outer region
794: $\chi_{3}<\chi<\chi_{\rm sc}$.
795: In accordance with eq.~(\ref{eq:condition}), shell crossing
796: necessarily occur in the outer region and we assume that
797: this region collapses into a dust shell. Hence we focus on a situation
798: in which $(\chi^{2}k)'<0$ holds inside the dust shell.
799: The model is constructed by enclosing the interior LTB
800: region by a spherically symmetric timelike singular
801: hypersurface (see fig.~\ref{fig:configuration}).
802: \begin{figure}
803: \begin{center}
804: \includegraphics{DS.eps}
805: \caption{\label{fig:configuration} \\
806: The schematic configuration of the shell. }
807: \end{center}
808: \end{figure}
809:
810: The analysis by using the dust shell model works only before
811: the dust shell reaches the boundary of the cubic co-moving
812: region $\Omega$. When the dust shell reaches the boundary, it collides with
813: other dust shells centered in surrounding cubic regions.
814: In this article, we do not consider the dynamics
815: after collisions of the dust shells and assume that
816: there is enough time before collisions of dust shells.
817:
818: The interior region of the dust shell is described by the LTB solution
819: with a line element,
820: \begin{equation}
821: ds_{-}^{2}
822: = -dt_{-}^2
823: +\dfrac{{Y'}^{2}(t_{-},\chi_{-})}{1-\chi_{-}^{2}k(\chi_{-})}d\chi_{-}^{2}
824: +Y^{2}(t_{-},\chi_{-})
825: \left(
826: d\theta^{2}+\sin^{2}\theta d\varphi^{2}
827: \right),
828: \end{equation}
829: where the prime ${}'$ means the derivative with respect to $\chi_{-}$.
830: The equation for the areal radius $Y$ is given
831: by the same equation as eq.~(\ref{eq:einstein}), i.e.,
832: \begin{equation}
833: \dot{Y}^{2}=-\chi_{-}^{2}k(\chi_{-})+\dfrac{2M(\chi_{-})}{Y},
834: \label{eq:density-LTB}
835: \end{equation}
836: where the dot $\dot{}$ means the derivative with respect to $t_{-}$.
837: Here we focus on the late time behavior of the volume expansion.
838: In this case, since $Y$ is monotonically increasing with respect to
839: $t_{-}$, the ``gravitational potential'' $2M(\chi_{-})/Y$
840: becomes much smaller than $-\chi_{-}^{2}k(\chi_{-})$
841: and hence we ignore this potential term in eq.~(\ref{eq:density-LTB}).
842: This approximation corresponds to an assumption in which the interior region
843: of the dust shell is described by Minkowski geometry.
844: Then the solution for $Y$ is easily obtained as
845: \begin{equation}
846: Y(t_{-},\chi_{-})=\left\{ -\chi_{-}^{2}k(\chi_{-}) \right\}^{1/2}t_{-}.
847: \label{eq:Y-solution}
848: \end{equation}
849:
850: The line element of the outer region is
851: written as
852: \begin{equation}
853: ds_{+}^{2}
854: = -dt_{+}^{2}
855: +a^{2}(t_{+})
856: \left\{
857: d\chi_{+}^{2}
858: +\chi_{+}^{2}
859: \left(
860: d\theta^{2}+\sin^{2}\theta d\varphi^{2}
861: \right)
862: \right\}, \label{eq:FRW}
863: \end{equation}
864: and Einstein equations reduce to the Hubble equation as
865: \begin{equation}
866: (\dot{a}\chi_{+})^{2}=\frac{M_{+}(\chi_{+})}{a\chi_{+}},
867: \label{eq:density-FRW}
868: \end{equation}
869: where $M_{+}(\chi_{+})$ is related to the rest mass density $\rho_{+}$
870: of the outer region as
871: \begin{equation}
872: M_{+}(\chi_{+})
873: = \frac{4\pi}{3}\rho_{+}a^{3}\chi^{3}_{+}.
874: \label{eq:mass}
875: \end{equation}
876: Note that by virtue of the spherical symmetry, angular coordinates,
877: $\theta$ and $\varphi$, are common for both interior and exterior
878: regions of the dust shell.
879:
880: A physically and geometrically clear prescription to treat a
881: timelike singular
882: hypersurface has been presented by Israel\cite{ref:Israel}.
883: In his prescription, junction conditions on metric tensor across
884: the singular hypersurface lead to equations
885: to determine the singular hypersurface itself, i.e.,
886: the equation of motion of a dust shell in our case.
887: Using his prescription, dynamics of a vacuum void
888: surrounded by a spherical dust shell in expanding universe has been
889: analyzed by Maeda and Sato\cite{ref:Maeda}. We can use their results
890: since the situation considered here is completely the same as theirs.
891:
892:
893: By virtue of the spherical symmetry of the system considered here,
894: the trajectory of the dust shell is given by
895: \begin{equation}
896: t_{-}=t_{\rm s-}(t_{+}),~~~
897: \chi_{\pm}=\chi_{\rm s\pm}(t_{+}),~~~\theta={\rm constant}~~~{\rm and}~~~
898: \varphi={\rm constant},
899: \end{equation}
900: where the time coordinate $t_{+}$ in the exterior region
901: has been adopted as an independent temporal variable.
902: The areal radius $R$ of the dust shell is then given by
903: \begin{equation}
904: R(t_{+}):=a\left(t_{+}\right)
905: \chi_{\rm s+}=Y_{-}\left(t_{\rm s-},\chi_{\rm s-}\right)
906: \end{equation}
907:
908: Maeda and Sato derived a differential
909: equation for the areal radius $R$ of the dust shell as\cite{ref:Maeda}
910: \begin{equation}
911: \dfrac{d^{2}R}{dt_{+}^{2}}
912: = \frac{1}{2R}
913: \left\{-\left(1+VV_{H}+2V^{2}+V_{H}^{2}\right)
914: +(1-4V^{2})(1+2VV_{H}+V_{H}^{2})^{1/2}\right\},\label{eq:R-EOM}
915: \end{equation}
916: where
917: \begin{align}
918: H&:= \dfrac{\dot{a}(t_{+})}{a(t_{+})}, \\
919: V_{H}&:= HR, \\
920: V&:=\dfrac{dR}{dt_{+}}-V_{H}.
921: \end{align}
922: Using a solution of eq.~(\ref{eq:R-EOM}), the
923: coordinate radius $\chi_{\rm s+}$ of the dust shell
924: in the outer region is given by
925: \begin{equation}
926: \chi_{\rm s+}=\dfrac{R}{a(t_{+})}.
927: \end{equation}
928: Equations for $t_{\rm s-}$ and $\chi_{\rm s-}$ on the dust shell are
929: given by
930: \begin{align}
931: \dfrac{dt_{\rm s-}}{dt_{+}}
932: = &\left\{1-\chi_{\rm s-}^{2}k(\chi_{\rm s-})\right\}^{1/2}
933: \left\{
934: 1-a^{2}(t_{+})\left(\dfrac{d\chi_{\rm s+}}{dt_{+}}\right)^{2}
935: +\left(\dfrac{dR}{dt_{+}}\right)^{2}
936: \right\}^{1/2} \notag \\
937: &-\left\{-\chi_{\rm s-}^{2}k(\chi_{\rm s-})\right\}^{1/2}
938: \dfrac{dR}{dt_{+}}, \label{eq:t-EOM}\\
939: \dfrac{d\chi_{\rm s-}}{dt_{+}}
940: = &\dfrac{1}{Y_{-}'(t_{\rm s-},\chi_{\rm s-})}
941: \left(
942: \dfrac{dR}{dt_{+}}
943: -{\dot Y}_{-}(t_{\rm s-},\chi_{\rm s-})\dfrac{dt_{\rm s-}}{dt_{+}}
944: \right),
945: \end{align}
946: where the dot $\dot{ }$ and the prime $'$ are derivatives
947: with respect to $t_{-}$ and $\chi_{-}$, respectively.
948: The volume $V_{\rm in}$ inside the dust shell is written as
949: \begin{align}
950: V_{\rm in}
951: &= 4\pi
952: \int_{0}^{\chi_{\rm s-}}
953: \dfrac{Y'Y^{2}}{\sqrt{1-\chi^{2}k(\chi)}}
954: d\chi
955: =4\pi
956: \int^{R}_{0}
957: \dfrac{Y^{2}}{\sqrt{1+Y^{2}/t_{\rm s-}^{2}}}
958: dY \notag \\
959: &= 2\pi t_{\rm s-}^{3}
960: \left[
961: \dfrac{R}{t_{\rm s-}}\sqrt{1+\left(\dfrac{R}{t_{\rm s-}}\right)^{2}}
962: -\ln
963: \left\{
964: \dfrac{R}{t_{\rm s-}}
965: +\sqrt{ 1+\left( \dfrac{R}{t_{\rm s-}} \right)^{2} }
966: \right\}
967: \right], \label{eq:V-in}
968: \end{align}
969: where we have used eq.~(\ref{eq:Y-solution}) in the second equality
970: to estimate $\chi^{2}k(\chi)$ in the integrand.
971: We consider a normalized volume ${\tilde V}_{\rm in}$ by
972: the volume of the removed homogeneous dust ball
973: in the original EdS universe,
974: \begin{equation}
975: {\tilde V}_{\rm in}:=\dfrac{V_{\rm in}}{4\pi R^{3}/3}=
976: \dfrac{3}{2}\left(\dfrac{t_{\rm s-}}{R}\right)^{3}\left[\dfrac{R}{t_{\rm s-}}
977: \sqrt{1+\left(\dfrac{R}{t_{\rm s-}}\right)^{2}}
978: -\ln\left\{\dfrac{R}{t_{\rm s-}}+\sqrt{1+\left(\dfrac{R}{t_{\rm s-}}\right)^{2}}
979: \right\}\right]. \label{eq:V-tilde}
980: \end{equation}
981: ${\tilde V}_{\rm in}$ is
982: monotonically decreasing function of $R/t_{\rm s-}$;
983: it approaches to unity
984: in the limit of $R/t_{\rm s-}\rightarrow0$, while
985: it vanishes in the limit $R/t_{\rm s-}\rightarrow\infty$
986: (see fig.~\ref{fig:V_in}).
987: In order to see temporal behavior of
988: ${\tilde V}_{\rm in}$, we need to solve eqs.~(\ref{eq:R-EOM}) and
989: (\ref{eq:t-EOM}).
990: \begin{figure}[ht]
991: \begin{center}
992: \includegraphics{V_in.eps}
993: \caption{\label{fig:V_in}Behavior of $\tilde{V}_{\textrm{in}}$ as
994: a function of $R/t_{\textrm{s}-}$. }
995: \end{center}
996: \end{figure}
997:
998: For sufficiently large $t_{+}$, the areal radius $R$
999: of the dust shell becomes much smaller than
1000: the cosmological horizon scale $H^{-1}$.
1001: In this case, the the motion of the
1002: dust shell is well described by the Newtonian approximation.
1003: Maeda and Sato showed that for sufficiently large $t_{+}$,
1004: $R$ behaves as\cite{ref:Maeda}
1005: \begin{equation}
1006: R(t_{+})\propto t_{+}{}^{(15+\sqrt{17})/24}\sim t_{+}{}^{0.797}
1007: \sim t_{\rm s-}{}^{0.797}.
1008: \end{equation}
1009: Hence we find that $R/t_{\rm s-}\propto t_{\rm s-}{}^{-0.203}\rightarrow0$
1010: for $t_{\rm s-}\rightarrow \infty$. Using this result and
1011: eq.~(\ref{eq:V-tilde}), we find
1012: ${\tilde V}_{\rm in}\longrightarrow 1$ for
1013: $t_{\rm s-}\rightarrow\infty$, and hence for $t_{+}\rightarrow\infty$,
1014: \begin{equation}
1015: V\longrightarrow a^{3}(t_{+})\ell^{3}.
1016: \end{equation}
1017: This equation means that the volume expansion rate approaches to
1018: the background value asymptotically, i.e.,
1019: \begin{equation}
1020: \dfrac{\dot{V}}{V}\longrightarrow
1021: 3\dfrac{\dot{a}}{a}=3\dfrac{\dot{a}_{\textsc{b}}}{a_{\textsc{b}}}
1022: \text{~~for~}t\rightarrow \infty.
1023: \end{equation}
1024: The effect of inhomogeneities on the volume expansion rate
1025: vanishes after the dust shell
1026: becomes much smaller than the horizon scale $H^{-1}$
1027: (see fig.~\ref{fig:ratio.after}).
1028: However, as in the case of $k_{0}>0$,
1029: the volume itself is different from the background
1030: value $a_{\textsc{b}}^{3}\ell^{3}$. By eq.~(\ref{eq:a-def}), we
1031: find that the asymptotic value of $V$ is larger than
1032: the background value $a_{\textsc{b}}^{3}\ell^{3}$.
1033: \begin{figure}[ht]
1034: \begin{center}
1035: \includegraphics[keepaspectratio]{ratio-2.eps}
1036: \caption{ \label{fig:ratio.after}%
1037: Temporal evolution of the ratios after shell crossing
1038: singularity appears. %
1039: The time is set to unity,
1040: when the scale of spherical inhomogeneous region $a(t)\chi_{\rm sc}$
1041: agrees with that of the horizon scale $a(t)/\dot{a}(t)$.
1042: $V_{\textsc{msc}}$ is the volume of the co-moving cubic region
1043: $\Omega$ in modified Swiss Cheese universe.
1044: $V_{\textsc{b}}$ is the volume of $\Omega$ measured by the background
1045: EdS geometry, i.e., $V_{\textsc{b}}=a_{\textsc{b}}^{3}\ell^{3}$.
1046: $V_{\textsc{out}}$ is the volume of $\Omega$ measured by the outer
1047: EdS geometry, i.e., $V_{\textsc{out}}=a^{3}\ell^{3}$.}
1048: \end{center}
1049: \end{figure}
1050: However it should be noted that there is an upper limit on the
1051: asymptotic value of $V$.
1052: Since $k$ can be arbitrarily small in this model,
1053: $\sqrt{1-\chi^{2}k}$ can be made arbitrarily larger than unity
1054: except at $\chi=0$ and $\chi=\chi_{\rm sc}$. Hence we obtain
1055: \begin{equation}
1056: a_{\textsc{b}}(t_{+})
1057: >a(t_{+})\left(1-\dfrac{4\pi}{3\ell^{3}}\chi_{\rm sc}^{3}\right)^{1/3}
1058: >a(t_{+})\left(1-\dfrac{\pi}{6}\right)^{1/3},
1059: \end{equation}
1060: where the last inequality is obtained by setting
1061: $\chi_{\rm sc}=\ell/2$. Using this inequality, we obtain
1062: \begin{equation}
1063: a_{\textsc{b}}^{3}(t_{+})\ell^{3}
1064: <a^{3}(t_{+})\ell^{3} < \left(1-\dfrac{\pi}{6}\right)^{-1}
1065: a_{\textsc{b}}^{3}(t_{+})\ell^{3}
1066: \sim 2.10\times a_{\textsc{b}}^{3}(t_{+})\ell^{3}.
1067: \end{equation}
1068: In contrast with the case of $k_{0}>0$, the volume itself
1069: can not be so different from the background value.
1070:
1071:
1072: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1073: \section{SUMMARY AND DISCUSSION}
1074: \label{sec:summary}
1075: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1076:
1077: We have investigated an effect of inhomogeneities on the volume expansion
1078: in modified Swiss-Cheese universe model.
1079: We considered two cases; the inhomogeneities collapse into
1080: black holes ($k_{0}>0$), while inhomogeneities
1081: expands faster than the background volume expansion ($k_{0}<0$).
1082: When inhomogeneities can be treated as perturbations of
1083: Einstein-de Sitter universe, the volume expansion is decelerated
1084: due to the second order contribution of the perturbations
1085: in both models. This result agrees with Nambu's second order
1086: perturbation analysis.
1087:
1088: Although the choice of background homogeneous universe is
1089: straightforward in the case of $|k|\chi^{2}\ll 1$,
1090: it is not in the case of non-linear situation.
1091: We introduced the background homogeneous universe
1092: in order to satisfy the conditions;
1093: the cubic region in the background homogeneous universe
1094: with the same rest mass have the same evolution of the volume
1095: as that of corresponding region in the modified Swiss-Cheese universe
1096: in the limit of $t\rightarrow 0$.
1097:
1098:
1099: In the case of non-linear fluctuation with $k_{0}>0$, we find that
1100: the volume expansion rate approaches to that
1101: of background universe asymptotically.
1102: Since the modified Swiss-Cheese model is an exact solution
1103: of Einstein equation,
1104: we can obtain the precise behavior of the volume expansion.
1105: We set it so that for $t\rightarrow 0$,
1106: the temporal evolution agrees with background one.
1107: Then we found that for $t\rightarrow\infty$,
1108: the temporal evolution agrees with that of outer EdS universe.
1109: We can see these asymptotic behaviors analytically
1110: but we have to rely only on the numerical method to obtain the behavior
1111: of the intermediate stages.
1112: From the result of the numerical calculation (fig.~\ref{fig:positive}),
1113: we found that volume expansion is decelerated by the inhomogeneities.
1114: This behavior coincides with a result obtained by Nambu.
1115: We note that his result is based on the perturbation theory
1116: but our result is not.
1117: In our highly non-linear example, the volume expansion rate becomes
1118: negative at the intermediate stage.
1119: This result may be the characteristics of the effect of
1120: non-linear fluctuations which can not be treated by the perturbation
1121: method.
1122:
1123: In the case of $k_{0}<0$, the shell crossing singularity appears
1124: in the inhomogeneous region.
1125: In this article, we assumed that a spherical dust shell forms
1126: after the shell crossing singularity appears.
1127: We find that the inhomogeneities decelerate the volume expansion
1128: before it appears (fig.~\ref{fig:ratio.before}).
1129: This is consistent with the previous results by Nambu but the
1130: inhomogeneities cannot be treated as the perturbation of homogeneous
1131: universe.
1132: After a spherical dust shell forms, the volume expansion rate approaches
1133: to that of the background universe asymptotically.
1134: Our dust shell universe model (fig.~\ref{fig:configuration})
1135: is a crude approximation.
1136: Therefore the behavior of the volume expansion obtained
1137: (fig.~\ref{fig:ratio.after}),
1138: especially at the early stage after the shell crossing,
1139: contains the influences of this crudeness.
1140: But the asymptotic behavior ($t\rightarrow \infty$) might be
1141: free from this approximation.
1142:
1143: We fixed the form of $k(\chi)$.
1144: Temporal evolution of the volume depends on the form of $k(\chi)$
1145: but the asymptotic behavior may be independent of the form of $k(\chi)$;
1146: \begin{align*}
1147: V&\longrightarrow a^{3}\ell^{3} \text{~~for~~}
1148: t\rightarrow \infty.
1149: \end{align*}
1150:
1151:
1152:
1153: %%+++++++++++++++++++++++++++++++++++++++++++++
1154: %
1155: \vskip0.5cm
1156: {\large{\bf Acknowledgements}}
1157:
1158: We are grateful to colleagues in Department of Physics,
1159: Osaka City University for helpful discussions.
1160:
1161: \begin{thebibliography}{99}
1162: %
1163: \bibitem{ref:Weinberg} S.~Weinberg, {\it Gravitation and Cosmology}
1164: (Wiley, New York, 1973).
1165: %
1166: \bibitem{ref:Smoot} G.~F.~Smoot {\it et al}., Astrophys.~J.~Lett.
1167: {\bf 396}, L1 (1992).
1168: %
1169: \bibitem{ref:KS84} H.~Kodama and M.~Sasaki,
1170: Prog.~Theor.~Phys.~Suppl. {\bf 78}, 1 (1984)
1171: %
1172: \bibitem{Schmidt:1998ys} B.~P.~Schmidt {\it et al.},
1173: Astrophys.\ J.\ {\bf 507}, 46 (1998).
1174: %
1175: \bibitem{Perlmutter:1999np}
1176: S.~Perlmutter {\it et al.}, Astrophys.\ J.\ {\bf 517}, 565 (1999).
1177: %
1178: \bibitem{ref:futamase89}
1179: T.~Futamase, Mon. Not. R. astr. Soc. {\bf 237}, 187 (1989).
1180: %
1181: \bibitem{ref:futamase97}
1182: T.~Futamase, Phys. Rev. {\bf D53}, 681 (1997).
1183: %
1184: \bibitem{ref:Tomita}
1185: K.~Tomita, Prog. Theor. Phys. {\bf 37}, 831 (1967).
1186: %
1187: \bibitem{ref:Russ}
1188: H.~Russ, M.H.~Soffle, M.~Kasai and G.~B{\" o}rner, Phys. Rev.
1189: {\bf D56}, 2044 (1997).
1190: %
1191: \bibitem{ref:Mukhanov}
1192: V.~F.~Mukhanov, L.~R.~W.~Abramo and R.~H.~Brandenberger,
1193: Phys. Rev. Lett. {\bf 78}, 1624 (1997).
1194: %
1195: \bibitem{ref:Abramo}
1196: L.~R.~W.~Abramo and R.~H.~Brandenberger,
1197: Phys. Rev. {\bf D56}, 3248 (1997).
1198: %
1199: \bibitem{ref:Nambu}
1200: Y.~Nambu, Phys.~Rev. {\bf D62}, 104010 (2000).
1201: %
1202: \bibitem{ref:Nambu-2}
1203: Y.~Nambu, Phys.~Rev. {\bf D63}, 044013 (2001).
1204: %
1205: \bibitem{ref:Nambu-3}
1206: Y.~Nambu, Phys.~Rev. {\bf D65}, 104013 (2002).
1207: %
1208: \bibitem{ref:Geshnizjani}
1209: G.~Geshnizjani and R.~H.~Brandenberger,
1210: arXiv:gr-qc/0204074.
1211: %
1212: \bibitem{ref:Hellaby-Lake}
1213: C.~Hellaby and K.~Lake, Astrophys.~J.~{\bf 290}, 381(1985),
1214: %
1215: \bibitem{ref:Israel}
1216: W. Israel, Nuovo Cimento {\bf 44B} (1966), 1; \\
1217: -----, Nuovo Cimento {\bf 48B} (1967), 463; \\
1218: -----, Phys. Rev. {\bf 153} (1967), 1388.
1219: %
1220: \bibitem{ref:Maeda}
1221: K.~Maeda and H.~Sato, Prog.~ Theor.~Phys.~{\bf 70}, 772 (1983).\\
1222: K.~Maeda and H.~Sato, Prog.~ Theor.~Phys.~{\bf 70}, 1276 (1983).
1223: \end{thebibliography}
1224:
1225: \end{document}
1226: