Difference between revisions of "Problems of the Hot Universe (the Big Bang Model)"

At adiabatic expansion $aT \approx const$ and thus $a_{Pl} T_{Pl} \approx a_0 T_0 $. Then it follows that at Planck temperature $T_{Pl} \sim 10^{32} K$ the region which the presently observed Universe ($ a_0 \sim 10^{28} \mbox{\it cm}$, $T_0 \sim 1K $) originated from had the size of order of $10^{ - 4} \mbox{\it cm}$, which is $29$ orders of magnitude greater than the Planck length. What is the reason of such discrepancy?

The maximum size of the region where the causality relations can be established is the particle horizon: $$ L_p (t) = a(t)\int_0^t {\frac[[:Template:Dt']][[:Template:A(t')]]}. $$ At the radiation-dominated era $a(t) \sim t^{1/2} $ and $L_p (t) = 2t$. For $t \to 0$ the particle horizon $L_p (t)$ shrinks faster than the scale factor $a(t)$. Therefore at any moment of time in the past one can find regions of size $L_p,$ which are causally disconnected. It contradicts to well-established (with accuracy $10^{ - 5} $) isotropy of CMB. This contradiction is the essence of the horizon problem.

It is well known that the total energy density in the Universe is very close to the critical one: $\Omega \approx 1$. This condition is satisfied at least with $2\%$ accuracy. The problem of flatness of the Universe lies in the fact that the density of early Universe had to be incredibly close to the critical one in order to provide such correspondence at present. To convince oneself in this fact use the first Friedman equation in the form $$ \Omega - 1 = \frac{k}[[:Template:\dot a^2]]. $$ For matter-dominated Universe $a \sim t^{2/3}$, and $\dot a \sim t^{ - 1/3} $, then $a\dot a^2 = const$, therefore \begin{equation} \label{prob_of_mod_2_sol_1} \Omega - 1 \sim ka \sim kt^{2/3}. \end{equation} Under assumption that the matter-domination era endured from the time of equality between densities of matter and radiation $t_{eq} \approx 50\,000$ years to present time $t_0 \approx 1.4 \times 10^{10}$ years, then it follows from (\ref{prob_of_mod_2_sol_1}) that \begin{equation} \label{prob_of_mod_2_sol_2} \frac{{\Omega \left( {t_{eq} } \right) - 1}}{{\Omega \left( {t_0 } \right) - 1}} = \frac{{a\left( {t_{eq} } \right)}}{{a\left( {t_0 } \right)}} = \left( {\frac{{t_{eq} }}[[:Template:T 0]]} \right)^{2/3} = \left( {\frac[[:Template:50\,000]]{{1.4 \times 10^{10} }}} \right)^{2/3} \approx 2 \times 10^{ - 4}. \end{equation} If at present $\Omega \left( {t_0 } \right) - 1$ is less than $\Delta$, then it follows from (\ref{prob_of_mod_2_sol_2}) that at $t = t_{eq} $ after the Big Bang $\Omega - 1$ was less than $2\Delta \times 10^{ - 4}$. Let us proceed even further back in time. Assume that at time $t_{Pl} < t < t_{eq} $ the Universe was dominated by radiation. Then $a \sim t^{1/2} ,\,\,\dot a \sim t^{ - 1/2} $ and therefore $$ \Omega - 1 \sim ka^2 \sim kt. $$ Then $$ \begin{array}{l} \displaystyle\frac{{\Omega \left( {t_{Pl} } \right) - 1}}{{\Omega \left( {t_0 } \right) - 1}} = \left( {\frac{{a\left( {t_{Pl} } \right)}}{{a\left( {t_{eq} } \right)}}} \right)^2 \frac{{a\left( {t_{eq} } \right)}}{{a\left( {t_0 } \right)}} = \frac{{t_{Pl} }}{{t_{eq} }}\left( {\frac{{t_{mr} }}[[:Template:T 0]]} \right)^{2/3} \approx 10^{ -60}, \\ \\ \displaystyle\Omega \left( {t_{Pl} } \right) - 1 \approx \Delta \times 10^{ - 60}. \\ \end{array} $$ It means that if the present value of relative density $\Omega $ is close to unity with accuracy of few per cent, than at Planck time it could not differ from unity more than on $10^{ - 60} $. It should be stressed that the flatness problem does not imply a contradiction between theory and observations, it rather concerns about the reason why Nature chose such a strange initial condition.

Entropy density in present Universe by order of magnitude equals to photon number density $\left( n_{\gamma }\sim 400\ \mbox{ \it cm}^{-3} \right)$ and therefore \[S\sim n_{\gamma }R_{0}^{3}\sim 10^{88}.\] Let us now calculate the entropy in early Universe. As specific entropy is $s\sim n_{\gamma }\sim T^3,$ then inside the horizon \[S\sim R_{H}^{3}{{T}^{3}}\sim {{H}^{-3}}{{T}^{3}}.\] As was shown above, in the early Universe $H\sim T^2/M_{Pl}$ and therefore \[S\sim\left( \frac{M_{Pl}}{T} \right)^{3}.\] Thus at Planck epoch inside the horizon one has $T\sim {{M}_{Pl}}$ and \[S\sim 1.\] It means that early Universe consisted of ${{10}^{88}}$ independent, causally disconnected regions! The Big Bang model lacks a mechanism which could transform such a Universe into the presently observed one (with isotropy on level of $10^{-5}$).

It was shown in Chapter 6 that current value of entropy of the Universe equals to $S \sim 10^{88} $. If the total entropy is conserved during the expansion of the Universe (recall that entropy conservation is embedded into the Friedman equations: $\dot \rho + 3H\left( {\rho + p} \right)= 0$ $ \to \;dE + pdV = 0$), then the Universe had to have such huge value of entropy already at the very moment of birth.

The wavelength $\lambda _p $ corresponding to a given perturbation grows as any other length scale according to the law $\lambda _p \propto a$. The Hubble radius is $$R_H = H^{ - 1} = \left( {\frac[[:Template:\dot a]] {a}} \right)^{ - 1}. $$ If $a(t) \propto t^q $, then $R_H \propto a^{1/q} $. Therefore $$ \left( {\frac[[:Template:\lambda p]] [[:Template:R H]]} \right) \propto a^{(q-1)/ q}. $$ For radiation $(q = 1/2)$, as well as for matter $(q = 2/3)$, one has $q < 1$, thus we come to the conclusion that the primary fluctuations had to be correlated on scales considerably larger than the Hubble radius. Therefore any mechanism of primary inhomogeneities generation in the Big Bang model comes to contradiction with the causality principle. Indeed, if the inhomogeneities generation would occur according to the causal mechanism then the corresponding length scales must evidently lie within the Hubble sphere, i.e. $\lambda _p < R_H $. However for example for $q = 1/2$ (as for radiation) one obtains $$ {\frac[[:Template:\lambda p]] [[:Template:R H]]} \propto a^{ - 1} $$ and the above cited condition is deliberately violated for small $a$.

The required condition can be presented in the form $$ - \frac{d} [[:Template:Dt]]\left( {\frac[[:Template:\lambda p]] [[:Template:R H]]} \right) < 0. $$ Taking into account that $\lambda _p \propto a$ and $$R_H = H^{ - 1} = \frac{a} [[:Template:\dot a]],$$ one can see that this condition is equivalent to the requirement $$-\ddot{a}< 0,$$ i.e. expansion of early Universe had to be accelerated. In other words the early Universe had to pass the accelerated (inflationary) expansion phase if generation of primary perturbations (fluctuations) corresponds to any physical mechanism.

At present our Universe is homogeneous and isotropic at scales of order of $ct_0 $. Initial size of the inhomogeneity is $$ l_i \sim ct_0 \frac{a_i}{a_0 }. $$ Let us compare it with the corresponding causality scale $l_{caus} \sim ct_i $: $$ \frac{l_i }{l_{caus} } \sim \frac{t_0 a_i }{t_i a_0 }. $$ Assuming that the scale factor grows as a power law function of time $\dot a \sim a/t$ one obtains $$ \frac{l_i }{l_{caus} } \sim \frac{\dot a_i }{\dot a_0 }. $$ For a Universe dominated by gravity $\dot a_i > \dot a_0 $. Therefore the scale of inhomogeneity will always remain greater than that of causality.

Use the relation obtained in the previous problem $$ \frac{l_i }{l_{caus} } \sim \frac{t_0 a_i }{t_i a_0 }. $$ Replace $t_i \to t_{Pl} $ and take into account that $aT \sim const$ and $$ \frac{a_{Pl} }{a_0 } \sim \frac{T_0 }{T_{Pl} } \sim 10^{ - 32} $$ to obtain $$ \frac{l_i }{l_{caus} } \sim \frac{10^{17}} {10^{-43}} 10^{ - 32} \sim 10^{28}. $$ Therefore in order to reproduce the presently observed homogeneity of CMB the condition $\delta \rho /\rho \sim 10^{ - 5} $ must be satisfied in $10^{84}$ causally disconnected regions at Planck time.

Consider spherically symmetric distribution of matter. As total energy is conserved then $$ \begin{array}{l} \displaystyle E^{tot} = E_i^K + E_i^P = E_0^K + E_0^P,\\ \displaystyle E_i^K = E_0^K \left( {\frac{\dot a_i }{\dot a_0 }} \right)^2,\\ \displaystyle \frac{E^{tot} }{E_i^K} = \frac{E_i^K + E_i^P }{E_i^K } = \frac{E_0^K + E_0^P }{E_0^K }\left( {\frac{\dot a_0 }{\dot a_i }} \right)^2.\\ \end{array} $$ As $E_0^K \sim \left| {E_0^P } \right|$ and $$ \left( {\frac{\dot a_0 }{\dot a_i }} \right)^2 \le 10^{ - 28} $$ (see the previous problem), then $$ \frac{E^{tot} }{E_i^K } \le 10^{ - 56}. $$ The obtained inequality means that velocities of initial Hubble flow should be tuned so that the huge negative potential energy of matter could compensate the huge positive kinetic energy. A tiny deviation of order $10^{ - 54} \% $ in initial velocity distribution would dramatically change history of the Universe.