1: \chapter[A ``homogeneous'' LTB model]{The backreaction effect through a ``homogeneous'' dust LTB model}
2: \label{pool}
3:
4: In this chapter we will try to build an inhomogeneous universe model searching for a back-reaction effect of sub-Hubble perturbations.
5: In particular we will aim at finding the average dynamics effects described in Sect.~\ref{avese}.
6:
7: We will make use of the dust LTB model, described in Appendix~\ref{ltbm}.
8: It has been shown that the LTB solution can be
9: used to fit the observed $d_{L}(z)$ without the need of dark energy (for
10: example in \cite{Alnes:2005rw}). To achieve this result, however, it is necessary
11: to place the observer at the center of a rather big underdensity.
12: This is in contrast with the experimental verifications of the cosmological principle.\\
13: To overcome this fine-tuning problem we will build a model that does not single out the center as a preferred place.
14: To this end, we paid attention to choose appropriate densities and curvature profiles.
15:
16: To be clear, the density profile we have in mind is:
17: %
18: \begin{equation} \label{seni}
19: \rho=\sum_{i}c_{i}\cos k_{i}r+\tilde{\rho}
20: \end{equation}
21: %
22: In fig.~\ref{densi} there is a sketch of that density profile.
23: %
24: %%%%%%%%%%%%%%%%%%%%
25: \begin{figure}[htb]
26: \begin{center}
27: \includegraphics[width=15cm]{densi.eps}
28: \caption{\small \slshape An example of the density profile (\ref{seni}) we are going to use in the Lema\^{i}tre-Tolman-Bondi model.}
29: \label{densi}
30: \end{center}
31: \end{figure}
32: %%%%%%%%%%%%%%%%%%%%%
33: %
34: This shape is intermediate between a cubic lattice without preferred positions and a standard LTB model in which the center is singled out by its different density and/or spatial curvature. It is a sort of exact ``spherical lattice'': even though the inhomogeneities of the density are centered on the center of symmetry, an observer moving around such a universe should not note sizeable differences between one place and the other.
35:
36:
37:
38: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
39: \section{Flat case} \label{flafla}
40: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
41:
42: We will now consider a flat LTB model. This analysis will be useful when dealing with more elaborate models with $E \neq 0$.
43:
44: \subsection{Our model}
45: %%%%%%%%%%%%%%%%%%%%%%
46:
47: The density we will use has only inhomogeneities of one size:
48: %
49: \begin{equation} \label{densia}
50: \rho(r, \bar{t})=\rho_{M}+\rho_{M}\cos k \, r
51: \end{equation}
52: %
53: and does not have the weak singularity mentioned in \cite{Vanderveld:2006rb}.
54: This density profile is clearly nothing more than a toy-model: the present-day universe is far more inhomogeneous. However, it will have the main features in order to look for a back-reaction effect.\\
55: In (\ref{densia}) $k=2 \pi /\lambda$ and we will choose $\lambda=0.05$ which, with the units of Table \ref{units}, corresponds to $420$ Mpc.
56: The voids will be roughly 20 times smaller than $r_{BB}$, the comoving distance traveled by a photon since the big bang.
57:
58: As said in Appendix~\ref{ltbm}, we fixed the gauge freedom by choosing $\bar{t}(r)=\bar{t}$ and $Y(r, \bar{t})=r$.
59:
60: We will show how its density evolves in the next section when we will study the curved case. In the flat case the evolution is indeed reversed: structures are not forming, but spreading out. We will show better this in the Chapter~\ref{s-c}.
61:
62: \subsection{Back-reaction effects}
63: %%%%%%%%%%%%%%%%%%%%%%
64:
65: We built this model in search of the back-reaction effects described in Sect.~\ref{avese}: we are interested in the quantities on the right hand side of eq. (\ref{cq1}), which in the units of Table~\ref{units} is:
66: %
67: \begin{equation}
68: H^{2}_{D} = {4 \over 9} \langle \rho\rangle_D -{1 \over 6} {\langle {\cal R} \rangle}_D -{1 \over 6} Q_D
69: \end{equation}
70: %
71: We have given the analytical expression of their constituents for a LTB model in Appendix~\ref{ltbm}.
72:
73: In the flat case one finds out immediately from (\ref{riccis}) that $\langle {\cal R} \rangle_D=0$ and with a short calculation that $Q_{D}=0$, thus there is no backreaction at all i.e. the expansion rate of the LTB model and the one of the FRW model are the same.
74: It exactly holds:
75: %
76: \begin{equation}
77: q_{D}=1/2
78: \end{equation}
79: %
80: These results do not depend on the choice of the free functions of the LTB model.
81: This is not trivial: in the flat case the local equations are invariant under the averaging process: this is the cause of the absence of back-reaction.
82:
83: However is not possible to conclude that there is no backreaction in the curved case.
84: In the flat case we have, from eq. (\ref{scala}), $a_{D}=Y$ and this can be seen, alternatively, as the reason why we found no back-reaction: the equation for $Y$ are already averaged as pointed out in Appendix~\ref{ltbm} and therefore there can not be any different average dynamics.
85: In the curved case, however, $a_{D} \neq Y$. The expansion is now:
86: %
87: \begin{equation}
88: \langle \theta \rangle_D = {1 \over V_{D}} \int_D \theta \, dV=
89: \frac{\int_{0}^{R}dr \; \theta \; Y^{2}Y' / W}
90: {\int_{0}^{R} dr \; Y^{2}Y' / W}
91: = {{\dot V}_D\over V_D} = 3 {{\dot a}_D \over a_D} = 3 H_{D}
92: \end{equation}
93: %
94: where $R$ marks the outer shell of the spherical domain $V_{D}$. It is not anymore true that $H_{D}=\dot{Y}/Y$.
95:
96:
97:
98:
99: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
100: \section{Curved case}
101: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
102: Introducing the curvature enriches the model with respect to the flat case. We need to introduce the curvature for two reasons: the first is to have a more natural evolution and the second is to have a non trivial domain dynamics.
103:
104: \subsection{Our model}
105: %%%%%%%%%%%%%%%%%%%%%%
106:
107: Except for the non-zero curvature, we will use the same model studied in the flat case.
108: As explained in Appendix~\ref{ltbm}, the LTB curvature $E$ is connected to the FRW curvature $k$ by $k=-2E/r^{2}$.
109: (It is not normalized to unity.)
110: If $E \sim \textrm{const}$, away from the observer there will be no clash with the CMB measurement of $k=0$.
111: However, the center will be singled out in contrast with our idea of ``homogeneous'' model.
112: To avoid this, when we will introduce the curvature, we will demand to have $E=0$ on the average.
113:
114: We could think to use an oscillatory curvature, similarly to the behavior of the density. However there is too much freedom in the choice of $E$: it is, together with $\rho$ and $\bar{t}$, a free function of the model.\\
115: $E$, however, is fixed by the initial conditions: requesting $E=0$ forced the velocity of the shells to compensate the inhomogeneities to have a flat space.
116: Thus we can think that a more natural situation will happen if such a compensation does not take place.
117: To have a realistic evolution, we demand therefore that there are no initial peculiar
118: velocities at time $\bar{t}$, that is, to have an initial expansion $H$
119: independent of $r$. From Eq.\ (\ref{cucu}) this implies:
120: %
121: \begin{equation}
122: \label{Er}
123: E(r)=\frac{1}{2} H_{FRW}^2(\bar{t}) \, r^{2}-\frac{1}{6\pi}\frac{M(r)}{r}
124: \end{equation}
125: %
126: The graph of $E(r)$ chosen in this way is shown in Fig.\ \ref{E}: as you can see we have exactly
127: an oscillating behavior without the arbitrariness of requesting it. As seen from
128: the figure, the curvature $E(r)$ is small compared with unity. Indeed, in many
129: formulae $W=(1+ 2 E)^{1/2}\simeq 1+E$ appears, therefore one should compare $E$
130: with $1$. In spite of its smallness, the curvature will play a crucial role to
131: allow a realistic evolution of structures, as we will see in the next section.
132:
133: %%%%%%%%%%%%%%%%%%%%
134: \begin{figure}[htb]
135: \begin{center}
136: \includegraphics[width=16.2 cm]{curvatureh.eps}
137: \caption{\small \slshape Curvature $E(r)$ and $k(r)$ necessary for the initial conditions of no
138: peculiar velocities.}
139: \label{E}
140: \end{center}
141: \end{figure}
142: %%%%%%%%%%%%%%%%%%%%%
143:
144: \subsection{The dynamics}
145: %%%%%%%%%%%%%%%%%%%%%%
146:
147:
148: In Fig.\ \ref{densifi2} we show the evolution of $\rho(r,t)$ for three times:
149: $t=\bar{t}=-0.8$ (the Big Bang is at $t_{BB}=-1$), $t = -0.4$, and $t=0$
150: (corresponding to today).
151: As you can see, overdense regions start contracting and become thin shells (mimicking structures), while underdense
152: regions become larger (mimicking voids), and eventually occupy most of the
153: volume.
154:
155: The densities are normalized with respect to the average density $\langle \rho\rangle_D \simeq \rho_{FRW}(t)$ where the latter is the density of the FRW model which is at the initial time $\rho_{FRW}(\bar{t})=\langle \rho(r,\bar{t})\rangle_D$.
156: The equalities are not exact because the oscillations in the density smooth out not exactly within a finite volume: this is due to the growing of the surface area with the respect to the radius.
157:
158: Moreover, remember that $r$ is only a label for the shell whose Euclidean position at time
159: $t$ is $Y(r,t)$. In Fig.\ \ref{densifi2} we have normalized $Y(r,t)$ using $r_{FRW}=Y(r,t)/a(t)$ where $a(t)$ is again the scale factor of the FRW model just introduced.
160:
161: %%%%%%%%%%%%%%%%%%%%
162: \begin{figure}[htb]
163: \begin{center}
164: \includegraphics[width=16 cm]{densifi2.eps}
165: \caption{\small \slshape Behavior of the
166: density profiles $\rho(r,t)$ with respect to $r_{FRW}=Y(r,t)/a(t)$, at times $t=\bar{t}=-0.8$ (blue),
167: $t=-0.4$ (green) and $t=t_0=0$ (red).
168: The values of $\rho_{FRW}(t)$ are $1,\ 2.8,$ and $25$, for $t=0,\
169: -0.4,\ -0.8$, respectively. }
170: \label{densifi2}
171: \end{center}
172: \end{figure}
173: %%%%%%%%%%%%%%%%%%%%%
174:
175:
176: \clearpage
177: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
178: \subsection{Back-reaction effects}
179: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
180:
181:
182: We will now look again at the quantities on the right hand side of eq. (\ref{cq1}), in particular, $Q_{D}$ and $\langle {\cal R} \rangle_D$.
183:
184: We have found that $Q_{D}$ is negligible while this is not the case for $\langle {\cal R} \rangle_D$.
185: The reason is simple: the only difference with respect to the flat model in which both are zero is the curvature. While $E$ appears explicitly in $\langle {\cal R} \rangle_D$, in $Q_{D}$ it has only an indirect effect coming from the fact that now $a_{D} \neq Y$.
186:
187: From the integrability condition (\ref{condi}), we can therefore deduce that, being $Q_{D}$ and $\langle {\cal R} \rangle_D$ decoupled, the latter will evolve like a standard curvature term: ${\langle {\cal R} \rangle}_D \sim 1/a_{D}^{2}$.
188: In other words, we will have $w_{BR}\simeq -1/3$.
189: We have verified this numerically: in Fig.~\ref{ricci} we plotted the time evolution of $\langle {\cal R} \rangle_D$.
190:
191: %%%%%%%%%%%%%%%%%%%%
192: \begin{figure}[htb]
193: \begin{center}
194: \includegraphics[width=13 cm]{ricci.eps}
195: \caption{\small \slshape Time evolution for $\langle {\cal R} \rangle_D$. It evolves as a FRW curvature term: ${\langle {\cal R} \rangle}_D \sim 1/a_{D}^{2}$.}
196: \label{ricci}
197: \end{center}
198: \end{figure}
199: %%%%%%%%%%%%%%%%%%%%%
200:
201: \clearpage
202: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
203: \section{Discussion}
204: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%
205:
206: We have applied the averaging scheme of Sect.~\ref{avese} to a toy model that features the growing of structures.
207: We looked quantitatively at the importance of the back-reaction sources $Q_{D}$ and $\langle {\cal R} \rangle_D$.
208: While we found a (very small) contribution from the average curvature, no appreciable effects were coming from $Q_{D}$.
209:
210: We trace this to the exact absence of back-reaction for the flat LTB model. The ultimate reason is that the equation automatically performs a ``euclidean'' average with respect to the local quantity $Y$: averaging something twice will not, of course, give sizeable effects. We believe that this is caused by spherical symmetry (Birkhoff theorem). \\
211: In the curved case, the curvature enters only indirectly in $Q_{D}$ and therefore we still do not have interesting domain effects.
212:
213: From this study we learn that LTB models with observers at the center are not interesting to study the formalism developed by Buchert.
214: This could be, however, due to a poor use of them more than to their deficiency with respect to back-reaction studies in general.
215: Having already discussed the possibility that the questions put by Buchert's formalism are not the observationally-meaningful questions,
216: we will try to make a better use of the LTB model.
217: Having seen that spherical symmetry suppresses back-reaction effects, we will try to go beyond this limitation building a Swiss-cheese (toy) model for the universe in the next Chapter.
218:
219:
220:
221:
222: We will conclude acknowledging that recently and independently the reference \cite{Biswas:2006ub} developed a similar density profile, using, however, a different curvature: we have used an oscillating-around-zero curvature which does not single out the center of symmetry.
223:
224:
225:
226:
227:
228:
229:
230:
231:
232: