1: \documentclass[preprint,nofootinbib]{revtex4}

2: 

3: \usepackage{graphicx}

4: 

5: \begin{document}

6: 

7: \title{From de Sitter to de Sitter

8: \\{\small A non-singular inflationary universe driven by vacuum}}

9: 

10: \author{Saulo Carneiro\footnote{saulo@fis.ufba.br,

11: saulocarneiro@yahoo.com}}

12: 

13: \affiliation{Instituto de F\'{\i}sica, Universidade Federal da

14: Bahia, 40210-340, Salvador, BA, Brazil \\ International Centre for

15: Theoretical Physics, Trieste, Italy\footnote{Associate member}}

16: 

17: \begin{abstract}

18: A semi-classical analysis of vacuum energy in the expanding

19: spacetime suggests that the cosmological term decays with time, with

20: a concomitant matter production. For early times we find, in Planck

21: units, $\Lambda \approx H^4$, where $H$ is the Hubble parameter. The

22: corresponding cosmological solution has no initial singularity,

23: existing since an infinite past. During an infinitely long period we

24: have a quasi-de Sitter, inflationary universe, with $H \approx 1$.

25: However, at a given time, the expansion undertakes a phase

26: transition, with $H$ and $\Lambda$ decreasing to nearly zero in a

27: few Planck times, producing a huge amount of radiation. On the other

28: hand, the late-time scenario is similar to the standard model, with

29: the radiation phase followed by a dust era, which tends

30: asymptotically to a de Sitter universe, with vacuum dominating

31: again.

32: \end{abstract}

33: 

34: \maketitle

35: 

36: \begin{flushright}

37: {\small {\it ``Matter has to find a way to avoid the annihilation of

38: its volume"}\\ George Lema\^{\i}tre

39: 

40: \vspace{0.25cm}

41: 

42: {\it ``however, [...] `the beginning of the world' really

43: constitutes a beginning"}\\ Albert Einstein}

44: \end{flushright}

45: 

46: ${ }$

47: 

48: The cosmological constant problem has been a theme of theoretical

49: discussion for decades, and has turned into a central point of

50: modern cosmology since recent observations suggested the existence

51: of a negative-pressure component in the cosmic energy content

52: \cite{Weinberg}.

53: 

54: The problem arises when we try to associate such a component with

55: the vacuum energy density predicted by quantum field theories. In

56: the case, for example, of a massless scalar field, the energy

57: density associated to its quantum fluctuations is given by

58: \begin{equation} \label{Lambdanua}

59: \Lambda_0 \approx \int_0^{\infty} \omega^3 d\omega.

60: \end{equation}

61: 

62: This divergent integral can be regularized by imposing a superior

63: cutoff $m$, leading to $\Lambda_0 \approx m^4$. This may also be

64: performed by introducing a bosonic distribution function in

65: (\ref{Lambdanua}),

66: \begin{equation}\label{Lambdareg}

67: \Lambda_0 \approx \int_0^{\infty}

68: \frac{\omega^3\,d\omega}{e^{\omega/m}-1}\approx m^4.

69: \end{equation}

70: The regularization procedure is thus equivalent to assume a thermal

71: distribution of vacuum fluctuation modes, at a characteristic

72: temperature $m$.

73: 

74: A natural choice for $m$ is the energy scale of QCD condensation,

75: the latest cosmological vacuum transition we know, since vacuum

76: fluctuations above this cutoff - which has the order of the pion

77: mass - would generate quark de-confinement. Unfortunately, even with

78: this value (many orders of magnitude below the Planck mass, also

79: usually taken as a natural cutoff), the obtained vacuum density is

80: around $40$ orders of magnitude higher than the presently observed

81: cosmological constant. That is the problem.

82: 

83: We should observe, however, that the above reasoning is based on

84: QFT in flat spacetime. In this case, the energy-momentum tensor

85: appearing in Einstein's equations must be zero, and, therefore,

86: (\ref{Lambdareg}) should be exactly canceled by a bare

87: cosmological constant. Such a cancelation should occur for any

88: vacuum contribution derived in flat spacetime.

89: 

90: Now, what would happen if we could calculate the vacuum density in

91: the expanding background? The regularized result would depend on the

92: curvature, and, after subtracting $\Lambda_0$, we should obtain a

93: renormalized, time-dependent cosmological term $\Lambda$, decaying

94: from high initial values to smaller ones, as the universe expands

95: \cite{Ozer,Schutzhold,GRF2005}. This renormalization is similar to

96: what happens in the Casimir effect, where the important thing is not

97: the vacuum density itself, but the difference between its values

98: inside a bounded region and in unbounded space.

99: 

100: The variation in the vacuum density leads, on the other hand, to

101: matter production, in order to preserve the conservation of total

102: energy implied by Einstein's equations. Indeed, in the realm of a

103: spatially homogeneous and isotropic spacetime, the Bianchi

104: identities lead to the conservation equation

105: \begin{equation} \label{continuidade}

106: \dot{\rho}_T + 3H(\rho_T+p_T)=0.

107: \end{equation}

108: Here, $H=\dot{a}/a$ is the Hubble parameter, while $\rho_T$ and

109: $p_T$ are, respectively, the total energy and pressure of cosmic

110: fluid. By introducing the matter density and pressure, and writing

111: $\rho_T = \rho_m + \Lambda$ and $p_T = p_m - \Lambda$, we

112: have\footnote{Since the vacuum has the same symmetry as spacetime,

113: its energy-momentum tensor has the form

114: $T_{\Lambda}^{\mu\nu}=\Lambda g^{\mu\nu}$, where $g^{\mu\nu}$ is the

115: metric tensor, and $\Lambda$ is a scalar function of coordinates (in

116: the FLRW spacetime, just a function of time). Therefore, it has the

117: same structure as for a perfect fluid in co-moving observers, with

118: $p_{\Lambda} = - \Lambda$.}

119: \begin{equation} \label{continuidade2}

120: \dot{\rho}_m+3H(\rho_m+p_m)=-\dot{\Lambda}.

121: \end{equation}

122: This shows that matter is not conserved - the decaying vacuum acting

123: as a source of entropy.

124: 

125: But how to evaluate the vacuum contribution in the expanding

126: spacetime? A possible answer is suggested by a semi-classical

127: analysis of the equation of motion of a minimally coupled massless

128: scalar field, $D^{\mu}D_{\mu}\phi=0$, where $D$ denotes the

129: covariant derivative. In a FLRW spacetime, it assumes the form

130: \begin{equation} \label{KG}

131: 3H\frac{\partial\phi}{\partial t}+\frac{\partial^2\phi}{\partial

132: t^2}-\nabla^2\phi=0.

133: \end{equation}

134: 

135: Since the space is isotropic, let us consider a plane wave solution

136: propagating in the radial direction. As for any plane wave, the

137: wavelength is supposed very small compared to the cosmological

138: scale. Therefore, $H$ changes very slowly compared to the wave

139: function, and the solution has the form

140: \begin{equation} \label{phi}

141: \phi \approx \phi_0\, e^{-br} e^{-i(\omega t-kr)},

142: \end{equation}

143: with

144: \begin{equation} \label{k}

145: k = \frac{\sqrt{2}\omega}{2}\left[1+\sqrt{1+\left(

146: \frac{3H}{\omega}\right)^2}\right]^{1/2},

147: \end{equation}

148: \begin{equation} \label{b}

149: b =

150: \frac{3\sqrt{2}H}{2}\left[1+\sqrt{1+\left(\frac{3H}{\omega}\right)^2}\right]^{-1/2}.

151: \end{equation}

152: As one can see, the wave amplitude decreases with $r$, with a depth

153: length equal to $b^{-1}$.

154: 

155: The energy-momentum tensor of this scalar field is

156: \begin{equation} \label{T}

157: T^{\nu}_{\mu}=\partial_{\mu}\phi\,

158: \partial^{\nu}\phi^{\dagger}-\frac{1}{2}\,\partial_{\sigma}\phi\,

159: \partial^{\sigma}\phi^{\dagger}\, \delta^{\nu}_{\mu}.

160: \end{equation}

161: Taking its time component, and using (\ref{phi}), we derive

162: \begin{equation} \label{rho}

163: \rho \approx \left(\omega^2+b^2\right)\phi\phi^{\dagger}.

164: \end{equation}

165: One can interpret this result by saying that the scalar particle

166: performs, superposed to the mode of frequency $\omega$, a thermal

167: motion of temperature $b$.

168: 

169: On this basis, we may evaluate the vacuum fluctuations by doing the

170: shift $m \rightarrow m+b$ in (\ref{Lambdareg}). The dominant

171: contributions to the integral will be given by modes with $\omega

172: \approx m + b$. For $b << m$, the dominance occurs for $\omega

173: \approx m$, which implies, through (\ref{b}), that $b \approx H$.

174: For $b \sim m$ or $b>>m$, the dominance occurs for $\omega\approx

175: b$, leading again, through (\ref{b}), to $b \approx H$. In this

176: case, however, the plane-wave approximation (\ref{phi})-(\ref{b})

177: cannot be used anymore, and the identity between $b$ and $H$ will be

178: taken as an {\it ad hoc} assumption.

179: 

180: Then, after subtracting $\Lambda_0$, we obtain

181: \begin{equation}\label{Lambda}

182: \Lambda \approx (m+H)^4-m^4.

183: \end{equation}

184: This has a similar structure as in the Casimir effect. Actually, for

185: $H>>m$ we have the same cutoff-independent result, $\Lambda \approx

186: H^4$, with $H^{-1}$ playing the role of a distance between Casimir

187: plates.

188: 

189: Therefore, in the limit of very early times, the cosmological term

190: scales as $\Lambda \approx H^4$, while for later times ($H<<m$) it

191: scales as $\Lambda \approx m^3 H$ (we should, however, be careful

192: with this last conclusion, as discussed below). Let us investigate

193: the corresponding cosmological scenarios. For simplicity, we will

194: only consider the spatially flat case.

195: 

196: Leading the Friedmann equation $\rho_T = 3H^2$ into the

197: conservation equation (\ref{continuidade}), using for matter the

198: equation of state of radiation, $p_m = \rho_m/3$, and taking for

199: the vacuum our early-time result $\Lambda = 3H^4$ (the constant

200: factor is not important, being taken three for convenience), we

201: obtain the evolution equation

202: \begin{equation}\label{evolucao}

203: \dot{H}+2H^2-2H^4=0.

204: \end{equation}

205: Apart an integration constant which determines the origin of time,

206: its solution is

207: \begin{equation}\label{H}

208: 2t=\frac{1}{H}-\tanh^{-1}H.

209: \end{equation}

210: 

211: The evolution of $H$ is plotted in Figure 1. As one sees, this

212: universe has no initial singularity, existing since an infinite

213: past, when $H$ approaches asymptotically the Planck value $H=1$.

214: During an infinitely long period we have a quasi-de Sitter,

215: inflationary expansion, with $H\approx1$. But at a given time

216: (chosen around $t=0$) we have a huge phase transition, with a

217: characteristic time scale of a few Planck times, during which $H$

218: (and so $\Lambda$) falls to nearly zero.

219: 

220: The transition can also be understood in terms of the energy

221: content. The energy density of radiation is $\rho_m = \rho_T -

222: \Lambda$, and its relative energy density is $\Omega_m = 1 - H^2$.

223: Therefore, the transition leads from an empty, vacuum-dominated

224: universe to a radiation-dominated phase, with $\Omega_m$ approaching

225: $1$ asymptotically (see Figure 2, where we also plot the relative

226: energy density of vacuum, $\Omega_{\Lambda}=H^2$). This behavior can

227: also be described with help of the deceleration factor we obtain

228: from (\ref{H}). It is $q = 1 - 2H^2$, and suddenly changes from

229: $-1$, in the quasi-de Sitter phase, to $1$, in the radiation one

230: (Figure 3).

231: 

232: Let us now consider the limit of late times, for which

233: $\Lambda=\sigma H$, with $\sigma\approx m^3$. We have shown

234: elsewhere \cite{Borges} that, in the radiation phase, $a\propto

235: t^{1/2}$, with $\rho_m = 3/(4t^2)$, as in the standard model. On the

236: other hand, in the dust phase we have

237: \begin{equation}\label{a}

238: a = C\left( e^{\sigma t/2}-1\right)^{2/3},

239: \end{equation}

240: where $C$ is an integration constant.

241: 

242: For early times ($\sigma t<<1$), we have $a\propto t^{2/3}$, as in

243: the Einstein-de Sitter model. This decelerated phase follows until

244: very recently, when vacuum begins dominating again and the expansion

245: reenters in an accelerated phase, which, as one can see from

246: (\ref{a}), tends asymptotically to a de Sitter universe. We have

247: analyzed the redshift-distance relation for supernovas Ia in this

248: model, obtaining a fit of observational data as good as in the

249: $\Lambda$CDM model. The obtained present values of $H$ and

250: $\Omega_m$ and the universe age are also in good accordance with

251: other observations \cite{Borges}. Finally, the analysis of evolution

252: of density perturbations until the present time shows no important

253: difference compared to the $\Lambda$CDM model.

254: 

255: In the de Sitter limit, the Hubble parameter is given, as we know,

256: by $H=\sqrt{\Lambda/3}$. Therefore, using $\Lambda \approx m^3 H$,

257: one derives the results $H\approx m^3$ and $\Lambda\approx m^6$. The

258: former is an expression of the famous Eddington-Dirac large number

259: coincidence, provided we take $m$ of the order of the pion mass,

260: i.e., the order of the energy scale of the QCD chiral transition, as

261: initially supposed. The last relation, on the other hand, was

262: suggested by Zel'dovich four decades ago, on the basis of different

263: arguments.

264: 

265: Nevertheless, we should be careful before concluding that the

266: present universe evolves as described above. Our approach is based

267: on a macroscopic, semi-classical reasoning, and we still do not have

268: a microscopic description of vacuum decay. At late times the decay

269: probably depends on the mass of the produced particles, and so we

270: have no guarantee that vacuum is still decaying. If it stopped

271: decaying at some earlier time, we just have, after the primordial

272: transition, a $\Lambda$CDM universe.

273: 

274: To conclude, some words about the entropy of this universe. If the

275: vacuum fluctuations are thermally distributed, as suggested by

276: (\ref{Lambdareg}), the number of states inside a volume $V$ may be

277: estimated as

278: \begin{equation}\label{N}

279: N \approx V \int_0^{\infty}

280: \frac{\omega^2\,d\omega}{e^{\omega/(m+H)}-1}.

281: \end{equation}

282: 

283: For late times we have $H<<m$, and $N\approx Vm^3$. In the de Sitter

284: limit, taking $V$ as the Hubble volume, and $m^3\approx H$, we

285: obtain $N\approx H^{-2}\approx 10^{120}$. That is, in the final de

286: Sitter phase, the entropy inside the Hubble sphere is equal to the

287: area of its surface, which is an expression of the holographic

288: conjecture \cite{Bousso,GRF2003}.

289: 

290: On the other hand, during the primeval quasi-de Sitter phase, we

291: have $H>>m$, and (\ref{N}) leads to $N\approx VH^3$. Now, by taking

292: the Hubble volume we obtain $N\approx 1$, which is, again, equal to

293: the area of the Hubble surface. In this way, one may conclude that

294: the primordial phase transition leads a universe of very low entropy

295: into a state of very high entropy. The thermodynamic and time arrows

296: coincide.

297: 

298: ${ }$

299: 

300: I am indebt to R. Abramo, A. Saa, I. Shapiro and C. Pigozzo for

301: useful discussions and assistance.

302: 

303: \newpage

304: 

305: \begin{thebibliography}{}

306: 

307: \bibitem{Weinberg} S. Weinberg, Rev. Mod. Phys. {\bf 61}, 1 (1989);

308: V. Sahni and A. Starobinsky, Int. J. Mod. Phys. D {\bf 9}, 373

309: (2000); P. J. E. Peebles and B. Ratra, Rev. Mod. Phys. {\bf 75},

310: 559 (2003).

311: 

312: \bibitem{Ozer} M. Ozer and O. Taha, Phys. Lett. A {\bf171}, 363 (1986);

313: Nucl. Phys. B {\bf{287}}, 776 (1987); O. Bertolami, Nuovo Cimento

314: {\bf 93}, 36 (1986); K. Freese {\it et al}, Nucl. Phys. B {\bf

315: 287}, 797 (1987); J. C. Carvalho, J. A. S. Lima and I. Waga, Phys.

316: Rev. D {\bf 46}, 2404 (1992).

317: 

318: \bibitem{Schutzhold} R. Sch\"utzhold, Phys. Rev. Lett. {\bf 89}, 081302

319: (2002); R. Horvat, Phys. Rev. D {\bf 70}, 087301 (2004); I.

320: Shapiro, J. Sol\`a and H. Stefancic, JCAP {\bf 0501}, 012 (2005);

321: F. Bauer, Class. Quant. Grav. {\bf 22}, 3533 (2005); J. S. Alcaniz

322: and J. A. S. Lima, Phys. Rev. D {\bf 72}, 063516 (2005); R.

323: Aldrovandi, J. P. Beltr\'an Almeida and J. G. Pereira, Grav. \&

324: Cosmol. {\bf 11}, 277 (2005).

325: 

326: \bibitem{GRF2005} S. Carneiro, Int. J. Mod. Phys. D {\bf 14}, 2201 (2005);

327: S. Carneiro and J. A. S. Lima, Int. J. Mod. Phys. A {\bf 20}, 2465

328: (2005); S. Carneiro and A. E. Montenegro Jr., Braz. J. Phys. {\bf

329: 35}, 1052 (2005).

330: 

331: \bibitem{Borges} H. A. Borges and S. Carneiro, Gen. Rel. Grav. {\bf 37},

332: 1385 (2005); S. Carneiro, C. Pigozzo, H. A. Borges and J. S.

333: Alcaniz, astro-ph/0605607.

334: 

335: \bibitem{Bousso} R. Bousso, Rev. Mod. Phys. {\bf 74}, 825 (2002).

336: 

337: \bibitem{GRF2003} G. A. Mena Marug\'{a}n and S. Carneiro, Phys. Rev. D {\bf 65},

338: 087303 (2002); S. Carneiro, Int. J. Mod. Phys. D {\bf 12}, 1669

339: (2003).

340: 

341: \end{thebibliography}

342: 

343: \vspace{1.5cm}

344: 

345: \begin{figure}[hb]

346: \begin{center}

347: \includegraphics[height=6cm,width=10cm]{fig1.eps}

348: \end{center}

349: \caption{The Hubble parameter as a function of time (in Planck

350: units)}

351: \end{figure}

352: 

353: 

354: \begin{figure}[h]

355: \begin{center}

356: \includegraphics[height=6cm,width=10cm]{fig2.eps}

357: \end{center}

358: \caption{The relative energy densities of radiation and vacuum as

359: functions of time}

360: \end{figure}

361: 

362: 

363: \begin{figure}[hb]

364: \begin{center}

365: \includegraphics[height=6cm,width=10cm]{fig3.eps}

366: \end{center}

367: \caption{The deceleration parameter as a function of time}

368: \end{figure}

369: 

370: \end{document}

371: