1: %vyzor preambuly v LaTeX-u:

2: \NeedsTeXFormat{LaTeX2e}

3: \documentclass[12pt,a4paper]{article}

4: \usepackage{t1enc}

5: %\usepackage[slovak]{babel} %<--slovenska podpora

6: %\usepackage[latin2]{inputenc} %<--tiez koli pisaniu po slovenky

7: \usepackage[dvips,final]{graphicx} %<--ked chces vkladat obrazky

8: \usepackage{tabularx} %<--ked chces robit nieco s tabulkami

9: \usepackage{amsfonts}

10: \usepackage{amssymb}

11: \usepackage{amsmath}

12: \usepackage{epsfig}

13: \usepackage{epsf}

14: \usepackage{psfrag}

15: 

16: 

17: 

18: \def\du{\unskip\smash{\lower 1.4ex \hbox{\char34}}\kern-.2ex}

19: \def\hu{\kern-.2ex\hbox{\char92}}

20: \def\uv#1{ \du #1\hu}

21: 

22: \newcommand{\bdis}{\begin{displaymath}}

23: \newcommand{\edis}{\end{displaymath}}

24: \newcommand{\be}{\begin{equation}}

25: \newcommand{\ee}{\end{equation}}

26: \newcommand{\mbb}{\mathbb}

27: \newcommand{\mcal}{\mathcal}

28: \newcommand{\pd}{\partial}

29: \newcommand{\noi}{\noindent}

30: 

31: \newcommand{\la}{\lambda}

32: \newcommand{\La}{\Lambda}

33: \newcommand{\ep}{\epsilon}

34: \newcommand{\om}{\omega}

35: \newcommand{\Om}{\Omega}

36: 

37: \newcommand{\Ran}{\mbox{Ran}}

38: \newcommand{\Ker}{\mbox{Ker}}

39: %\newcommand{\dim}{\mbox{dim}}

40: 

41: \newtheorem{thm}{}[section]

42: \newtheorem{pr}[thm]{}

43: \newtheorem{pr*}[thm]{*}

44: \newtheorem{definition}{Defin\' icia}

45: %\theoremstyle{citing}

46: 

47: 

48: % nastavenia strany:

49: \voffset=-20mm %horny okraj

50: \oddsidemargin=-0.5cm %pravy okraj

51: \textwidth=180mm

52: \textheight=245mm \pagestyle{plain}

53: 

54: \begin{document}

55: \baselineskip=7mm

56: \newpage

57: 

58: \title{False vacuum decay with gravity in a critical case}

59: \author{Michal Demetrian \\

60: Comenius University \\

61: Mlynska Dolina M 105 Bratislava 4 \\

62: 842 25 Slovak Republic}

63: \maketitle

64: 

65: \abstract{The vacuum decay in a de Sitter universe is

66: studied within semiclassical approximation for the class of effective inflaton

67: potentials whose curvature at the top is close to a critical value.

68: By comparing the actions of the Hawking - Moss instanton and the Coleman - de Luccia instanton(s) the mode of

69: vacuum decay is determined. The case when the fourth derivative of the effective potential at its top is less than

70: a critical value is discussed. }

71: 

72: \section{Introduction}

73: 

74: The idea of vacuum decay in a de Sitter universe (the transition of the inflaton field from false vacuum with positive

75: energy density to true vacuum with (almost) zero energy density caused by the quantum mechanical instability of

76: the false vacuum)

77: was developed by Coleman and de Luccia in \cite{cdl} and plays

78: an important role in the cosmological inflationary scenario. It is considered as a mechanism of transition

79: to a Friedman universe in old inflation \cite{guth} (in this scenario, the rapidly growing

80: bubbles of almost true vacuum are created in the sea of false vacuum; collisions of such bubbles were expected to

81: produce a Friedman universe, however it was established afterwards that there is not time enough for that)

82: and emerges also in the scenario of open inflation,\cite{bgt}, \cite{lm} (in this case, only single bubble is created and filled by the

83: configuration of inflaton that can evolve classically to the true vacuum; reheating is provided by decay

84: of inflaton particles into other particles via parametric resonance

85: \cite{lindekofman} at the stage when classical inflaton field oscillates around

86: the true vacuum).

87: 

88: We consider single scalar

89: field $\Phi$ with self-interaction given by the nonnegative function $V(\Phi)$ - effective potential - that has two

90: nondegenerate minima, one of them strictly positive (false vacuum) and second one equal to zero (true vacuum). These

91: vacua are supposed to be separated by a finite potential barrier. Let $V$  reach its local maximum $V_M$ at $\Phi_M$.

92: Furthermore, let us denote by $H(\Phi)$ the Hubble parameter corresponding to the energy-density $V(\Phi)$:

93: $H(\Phi)=\sqrt{8\pi V(\Phi)/3}$, especially $H_M=\sqrt{8\pi V_M/3}$. In order to study quantum transition of the

94: inflaton, in fact, one does not need to have the potential described above, namely the potential may have no

95: local minima (vacua), since the existence of potential barrier is sufficient for this purpose.

96: Supposing $O(4)$ symmetry supplemented by the

97: regularity we get the following (Euclidean) equations of motion and boundary conditions for Euclidean version of

98: the scale parameter $a$ and the inflaton $\Phi$

99: \begin{eqnarray} \label{ee}

100: & &

101: a''=-C\left(\Phi'^2+V\right)a, \quad \Phi''+3\frac{a'}{a}\Phi'=V'_{\Phi} \\

102: & &

103: \label{eel2}

104: a(0)=0, \ a'(0)=1, \Phi'(0)=\Phi'\left(\tau_f\right)=0,

105: \end{eqnarray}

106: where the constant $C$ equals $8\pi/3$ and $\tau_f>0$ is defined by the equation $a(\tau_f)=0$.

107: For any suitable potential there exists a trivial solution of the above problem that reads

108: \be \label{hm_instanton}

109: \Phi_{HM}=\Phi_M,\ a_{HM}=H_M^{-1}\sin\left(H_M\tau\right), \quad \mbox{with} \quad \tau_f=\frac{\pi}{H_M}

110: \ee

111: and is called the Hawking-Moss instanton \cite{hm}. This instanton mediates

112: the vacuum decay in such a way that

113: the inflaton jumps up to the top of the barrier in the horizon-size domain and afterwards the inflaton leaves

114: (by quantum or thermal fluctuations) the

115: unstable equilibrium and evolves classically to the true vacuum. There are also trivial solutions

116: corresponding to inflaton lying in stable equilibria in the true and false vacuum, respectively.

117: However, under some additional conditions, the

118: problem (\ref{ee}-\ref{eel2})

119: has also nontrivial solutions (with variable $\Phi$) called Coleman - de Luccia (CdL) instantons

120: (or bounces) \cite{cdl}. Following the ideas of paper \cite{vj2} (see also \cite{tanaka} and recently

121: \cite{weinberg}) CdL instantons can be characterized by how many times the inflaton crosses the top of the barrier. We

122: talk about the CdL instanton of the $l$th order if the inflaton crosses the top $l$-times.

123: \\

124: The boundary conditions (\ref{eel2}) provide the action of a CdL instanton (that follows from the general

125: Einstein-Hilbert action) to be finite. The action is a very important quantity for an instanton since it determines

126: the probability of the vacuum decay per unit space-time volume in the form $\exp(-S)$ .

127: This quantity

128: can be transformed, according to \cite{vj2}, into the following simple form

129: \be \label{act0}

130: S=2\pi^2\int_0^{\tau_f}\left[\left(\frac{1}{2}\Phi'^2+V\right)a^2-\frac{1}{C}

131: \left(aa'^2+1\right)\right]a{\rm d}\tau=-\frac{4\pi^2}{3C}\int_0^{\tau_f}a{\rm d}\tau .

132: \ee

133: It is easy to find that the action of the Hawking - Moss instanton is given by

134: \be \label{hmaction}

135: S_{HM}=-\frac{\pi}{H_M^2} .

136: \ee

137: %Being motivated by the previous computation, it seems to be useful to introduce dimensionless

138: %quantity $\sigma$ by the equation

139: %\be \label{sigmadef}

140: %S=-\frac{4\pi^2}{3CH_M^2}\sigma .

141: %\ee

142: %..................................... !!!!!!!!!!!! ...............................

143: 

144: 

145: 

146: 

147: \section{Near-to-limit CdL instanton of the first order and its action}

148: 

149: As it was sketched in \cite{js} and finally proved in \cite{vj1} the CdL instanton necessarily exists for

150: potentials with $V_M''/H_M^2<-4$ and may exist if $V''/H^2<-4$ for some value of $\Phi$ in the potential barrier. If the

151: fraction $V_M''/H_M^2$ approaches one of the values $-l(l+3),\ l=1,2,3 \dots$, the near-to-limit CdL instanton (that

152: approaches necessarily existing Hawking-Moss instanton) may exist. If $V_M''/H_M^2<-l(l+3)$ then the instanton

153: of $l$th order necessarily exists. Our task is to compute the difference between the action of the near-to-limit

154: CdL instanton of the first order and the action of the related Hawking - Moss instanton. This task has been considered in

155: \cite{vj2}, but our treatment will be different.

156: By making use of the re-scaled Euclidean time

157: $x=H_M\tau$ and the shifted inflaton field $y=y(x)=\Phi(\tau(x))-\Phi_M$ we rewrite equations (\ref{ee}) in to the form

158: \be \label{ee1}

159: a''=-C\left(y'^2+\frac{V}{H^2_M}\right)a, \quad y''+3\frac{a'}{a}y'=\frac{V'_y}{H_M^2}.

160: \ee

161: Since now the prime denotes differentiation with respect to $x$. Introducing the expansion of the relevant

162: quantities, including the dimensionless Euclidean action

163: \bdis

164: \sigma=-\frac{3CH_M^2}{4\pi^2}S , \qquad \left(\sigma_{HM}=2\right)

165: \edis

166: into the series in the inflaton amplitude $k$

167: \be \label{k-exp}

168: y(x)=\sum k^nu_n(x), \ -\frac{V_M''}{H_M^2}=4+\sum k^n\Delta_n, \ a(x)=CH_M^{-1}\sum k^nv_n(x), \

169: \sigma=2+\sum k^nw_n

170: \ee

171: and expanding the potential into the powers of $y$

172: \bdis

173: V=V_M+\frac{1}{2}V_M''y^2+\frac{1}{6}V_M'''y^3+\frac{1}{24}V_M''''y^4+ \dots

174: \edis

175: we replace the nonlinear equations (\ref{ee1}) by the infinite system of linear equations

176: \begin{eqnarray} \label{linsys}

177: & &

178: u_n''(x)+3\frac{\cos(x)}{\sin(x)}u_n'(x)+4u_n(x)=\mcal{U}_n(x), \nonumber \\

179: & &

180: v_n''(x)+v_n(x)=\mcal{V}_n(x)\sin(x)

181: \end{eqnarray}

182: in which the functions $\mcal{U}_n$ and $\mcal{V}_n$ can be computed order by order if we know

183: the functions $u_{n-1},u_{n-2},\dots ,u_0$. Functions $u_n$ and $v_n$ are defined on the interval

184: $[0,x_f^{(n)}]$, where $x_f^{(n)}$ is defined as the point in which the scale factor $a$ computed up to the $n$th order

185: in $k$ vanishes. The value of the functions $v_n$ and $v_n'$ must vanish at $x=0$ (this

186: follows from (\ref{eel2})) and $u_n$ must be regular.

187: We know that

188: \bdis

189: u_0=0, \ v_0=\sin\left(x\right) \quad \mbox{and} \quad u_1=\cos(x) .

190: \edis

191: Furthermore, $V_M'=0$ implies that $v_1$ vanishes. The next nonzero term in $a$ is given by $v_2$ that must obey

192: \bdis

193: v_2''+v_2=-\frac{1}{4}\left[\sin(x)-3\sin(3x)\right].

194: \edis

195: The solution is

196: \be \label{v2}

197: v_2(x)=\frac{1}{4}\left[\frac{5}{8}\sin(x)+\frac{1}{2}x\cos(x)-\frac{3}{8}\sin(3x)\right] .

198: \ee

199: Solving equation $v_0(x)+k^2v_2(x)=0$ with the accuracy up to the order $k^2$ and

200: supposing the solution is close $x=\pi$, we find that the shifted right end-point is given by

201: \bdis

202: x_f^{(2)}=\pi-\frac{1}{8}C\pi k^2\equiv \pi+\delta^{(2)} .

203: \edis

204: Knowing $v_2$ we can compute the contribution of the order $k^2$

205: to the difference between the actions of CdL and Hawking - Moss instanton that is defined by eqs. (\ref{act0}) and

206: (\ref{k-exp}). The result is

207: \bdis

208: w_2=\int_0^{\pi}v_2(x){\rm d}x=0.

209: \edis

210: This means that we cannot distinguish between the action of

211: a near-to-limit CdL instanton and the related Hawking - Moss instanton in the second order of inflaton amplitude and

212: we must continue our computations. Equation for $u_2$ reads

213: \bdis

214: u_2''+3\frac{\cos(x)}{\sin(x)}u_2'+4u_2=\frac{1}{2}\frac{V_M'''}{H_M^2}\cos^2(x)

215: \edis

216: and its regular solution is

217: \be \label{u2}

218: u_2(x)=\frac{1}{24}\frac{V_M'''}{H_M^2}\left[1-2\cos^2(x)\right] .

219: \ee

220: Now, we can derive equation for $v_3$ and its solution

221: \be \label{v3}

222: v_3''+v_3=\frac{V_M'''}{48H_M^2}\left[2\sin(2x)-5\sin(4x)\right] \ \Rightarrow

223: v_3=-\frac{V_M'''}{72H_M^2}\left[\sin(2x)-\frac{1}{2}\sin(4x)\right] .

224: \ee

225: There is no shift of the right end point $x_f$ since $v_3(\pi)=0$,

226: and we can compute the $k^3$-contribution to the action as follows

227: \bdis

228: w_3=\int_0^\pi v_3(x){\rm d}x=0.

229: \edis

230: This result forces us to continue up to the fourth order in $k$. The shift of the right end point $x_f$

231: with respect to $\pi$ (which

232: is actually of the order $k^2$)

233: must be taken under consideration in the equation for $u_3$. Introducing new a independent variable

234: \bdis

235: X=\frac{\pi x}{\pi+\delta^{(2)}}=x\left(1+\frac{1}{8}Ck^2+o\left(k^2\right)\right)\equiv K x

236: \edis

237: we obtain supplementary formulas

238: \bdis

239: v_0\left( K^{-1}X\right)=\sin\left( X\right)-\frac{Ck^2}{8}X\cos\left( X\right),\

240: v_2\left( K^{-1}X\right)=v_2\left( X\right),

241: \edis

242: \bdis

243: v_0(x)+k^2v_2(x)=\sin\left( X\right)\left[1-\frac{Ck^2}{8}X\frac{\cos\left( X\right)}{\sin\left( X\right)}+

244: \frac{1}{4}Ck^2\left(1-\frac{3}{2}\cos^2\left( X\right)\right)\right] \equiv

245: \sin\left( X\right)+e\left( X\right)

246: \edis

247: and

248: \bdis

249: \frac{v_0'(x)+k^2v_2'(x)}{v_0(x)+k^2v_2(x)}=K\frac{\cos\left( X\right)+\frac{{\rm d}e\left( X\right)}{{\rm d}X}}

250: {\sin\left( X\right)+e\left( X\right)}=K

251: \left[\frac{\cos\left( X\right)}{\sin\left( X\right)}+\frac{3}{4}Ck^2\cos\left( X\right)\sin\left( X\right)\right]

252: \edis

253: which provides us with the equation for $u_3$ of the form

254: \begin{eqnarray*}

255: & &

256: \frac{{\rm d}^2u_3\left( X\right)}{{\rm d}X^2}+3\frac{\cos\left( X\right)}{\sin\left( X\right)}

257: \frac{{\rm d}u_3\left( X\right)}{{\rm d}X}-

258: \frac{V_M''}{H_M^2}u_3\left( X\right)=\\

259: & &

260: \left[ \frac{1}{K^2}\left( 4+\frac{V_M''}{H_M^2}\right)+\frac{13}{4}C+

261: \frac{1}{24}\left(\frac{V_M'''}{H_M^2}\right)^2\right]\cos\left( X\right)+

262: \left[\frac{1}{6}\frac{V_M''''}{H_M^2}-\frac{9}{4}C-

263: \frac{1}{12}\left(\frac{V_M'''}{H_M^2}\right)^2\right]\cos^3\left( X\right) \equiv \\

264: & &

265: A\cos\left( X\right)+B\cos^3\left( X\right).

266: \end{eqnarray*}

267: The only regular ($\frac{{\rm d}u_3(0)}{{\rm d}X}=\frac{{\rm d}u_3(\pi)}{{\rm d}X}=0$)

268: solution to this equation is given by the formula

269: \bdis

270: u_3\left( X\right)=\beta\cos^3\left( X\right)

271: \edis

272: with the constant $\beta$ to be determined from the system of linear equation

273: \bdis

274: 6\beta=A,\ -14\beta=B\ \Rightarrow \

275: \beta=-\frac{1}{14}

276: \left\{\frac{1}{6}\frac{V_M''''}{H_M^2}-\frac{9}{4}C-\frac{1}{12}

277: \left(\frac{V_M'''}{H_M^2}\right)^2\right\} .

278: \edis

279: However, the fixation of $\beta$ is only a supplementary consequence of previous system of linear equations,

280: since their main purpose is to determine the value of $k^2$ as a function of $A$ and $B$.

281: Namely, we obtain from them the following

282: "quantization rule" for $k^2$ as a function of $4+V_M''/H_M^2$

283: \be \label{k2quant}

284: k^2=-\frac{4+\frac{V_M''}{H_M^2}}

285: {\frac{2}{7}\left[8C+\frac{1}{48}\left(\frac{V_M'''}{H_M^2}\right)^2+\frac{1}{4}\frac{V_M''''}{H_M^2}\right]} .

286: \ee

287: If the denominator of the fraction on the right hand side is positive then we have, for small

288: negative numerator a near-to-limit CdL instanton of the first order whose inflaton amplitude

289: is given by (\ref{k2quant}). We will return to the case when the denominator is negative later.

290: Now, let us concentrate

291: on computation of the action of this near-to-limit CdL instanton. Performing some

292: tedious algebra one derives equation for $v_4$ of the form

293: \bdis

294: v_4''+v_4=\left[\aleph_0+\aleph_2\cos^2(x)+\aleph_4\cos^4(x)\right]\sin(x) ,

295: \edis

296: where we have introduced the parameters

297: \begin{eqnarray*}

298: \aleph_0 & = &

299: -\frac{15}{16}C+\frac{1}{288}\left(\frac{V_M'''}{H_M^2}\right)^2 , \\

300: \aleph_2 & = &

301: \frac{159}{224}C-\frac{2}{21}\left(\frac{V_M'''}{H_M^2}\right)^2+\frac{3}{28}\frac{V_M''''}{H_M^2} \\

302: \aleph_4 & = &

303: \frac{27}{56}C+\frac{1}{7}\left(\frac{V_M'''}{H_M^2}\right)^2-\frac{9}{56}\frac{V_M''''}{H_M^2} .

304: \end{eqnarray*}

305: The solution is

306: \begin{eqnarray} \label{v4}

307: & &

308: v_4=\frac{1}{192}\left\{ -12\left(8\aleph_0+2\aleph_2+\aleph_4\right)x\cos(x)+ \right. \nonumber \\

309: & &

310: \left.

311: \sin(x)\left[ 96\aleph_0+36\aleph_2+23\aleph_4-2\left(6\aleph_2+5\aleph_4\right)\cos(2x)-2\aleph_4

312: \cos(4x)\right]\right\}

313: \end{eqnarray}

314: and with the help of it we finish the computations with a nonzero contribution to the action of a

315: surprisingly simple form

316: \be \label{w4}

317: \Delta S^{(4)}\equiv -\frac{4\pi^2k^4w_4}{3CH_M^2}=\frac{2\pi^2}{15}\frac{k^2}{H_M^2}

318: \left( 4+\frac{V_M''}{H_M^2}\right) .

319: \ee

320: Formula (\ref{w4}) tells us that a near-to-limit CdL instanton of the first order has, in the case

321: $V''_M/H_M^2<-4$, less action than the related Hawking - Moss instanton and therefore, if no other instantons exist,

322: it is the instanton governing the false vacuum decay. \\

323: Let us demonstrate the power of the formulas (\ref{k2quant}) and (\ref{w4}) on a concrete example. We will consider

324: the often mentioned quartic potential

325: \be \label{quarticpot}

326: V(\Phi)=\frac{1}{2}\Phi^2-\frac{1}{3}\delta\Phi^3+\frac{1}{4}\lambda\Phi^4 ,

327: \ee

328: where $\delta$ and $\lambda$ are supposed to be positive. The non-negativeness of the potential and the existence of

329: the false vacuum (the true vacuum is located at $\Phi=0$) require that the $\delta$ parameter belongs to the interval

330: $(\delta_m,\delta_M)$, where

331: \bdis

332: \delta_m=2\sqrt{\lambda},\ \delta_M=3\sqrt{\frac{\lambda}{2}}\approx 1.06 \delta_m .

333: \edis

334: The positions of the false vacuum ($\Phi_{fv}$) and of the top of the barrier are given by

335: \bdis

336: \Phi_{fv}=\frac{\delta}{2\lambda}+\sqrt{\frac{\delta^2}{4\lambda^2}-\frac{1}{\lambda}}=

337: \frac{\sqrt{1-Z^2}}{(1-Z)\sqrt{\lambda}}, \

338: \Phi_M=\frac{\delta}{2\lambda}-\sqrt{\frac{\delta^2}{4\lambda^2}-\frac{1}{\lambda}}=

339: \frac{\sqrt{1-Z^2}}{(1+Z)\sqrt{\lambda}} ,

340: \edis

341: where

342: \bdis

343: Z=\sqrt{1-\frac{4\lambda}{\delta^2}}, \quad Z\in\left[ 0,\frac{1}{3}\right] .

344: \edis

345: The potential (\ref{quarticpot}) in the $(Z,\lambda)$ parametrization has the form

346: \bdis

347: V=\frac{1}{2}\Phi^2-\frac{2}{3}\frac{\sqrt{\lambda}}{\sqrt{1-Z^2}}\Phi^3+\frac{1}{4}\lambda\Phi^4 .

348: \edis

349: We are interested in the quantities

350: \bdis

351: H_M^2\equiv \frac{8\pi}{3}V_M=\frac{2\pi}{9\lambda}\frac{(1-Z)(1+3Z)}{(1+Z)^2}, \quad

352: V_M''=-\frac{2Z}{1+Z} .

353: \edis

354: From these expressions it follows that the effective curvature of the potential at its top is given by

355: \be \label{efcurvq}

356: \frac{V_M''}{H_M^2}=-\frac{9\lambda}{\pi}\frac{Z(1+Z)}{(1-Z)(1+3Z)}

357: \ee

358: and is a monotonically increasing function of $Z$.

359: If we denote by $\lambda_S$ the value of $\lambda$ at which (at given $Z$) $V_M''/H_M^2=-4$, then

360: \bdis

361: \lambda_S=\frac{4\pi}{9}\frac{(1-Z)(1+3Z)}{Z(1+Z)} .

362: \edis

363: It will be useful to express the fractions $V'''_M/H_M^2$ and $V_M''''/H_M^2$ entering the formulae

364: (\ref{k2quant}) and (\ref{w4}) in terms of the parameters of the quartic potential. After some algebra one finds out

365: that

366: \bdis

367: \frac{V_M'''}{H_M^2}=-\frac{9\lambda^{3/2}}{2\pi}\frac{\left(1+Z\right)^2}{(1-Z)\sqrt{1-Z^2}}, \quad \mbox{and} \quad

368: \frac{V_M''''}{H_M^2}=\frac{27\lambda^2}{\pi}\frac{\left( 1+Z\right)^2}{(1-Z)(1+3Z)}

369: \edis

370: and by using the relation (\ref{efcurvq}) one gets the dependence of the fractions in question on the effective

371: curvature $V_M''/H_M^2$ of the potential at its top.

372: \be \label{frfo}

373: \frac{V_M'''}{H_M^2}=\frac{\sqrt{\pi}}{6}\left(\frac{1+Z}{1-Z}\right)^{1/2}\left(\frac{1}{Z}+3\right)^{3/2}

374: \left(-\frac{V_M''}{H_M^2}\right)^{3/2}, \

375: \frac{V_M''''}{H_M^2}=\frac{\pi}{3}\frac{(1-Z)(1+3Z)}{Z^2}\left(\frac{V_M''}{H_M^2}\right)^2 .

376: \ee

377: Now, we are ready to compare the predictions of the formulas (\ref{k2quant}) and (\ref{w4}) with the

378: numerical solutions of exact instanton equations (\ref{ee1}). In order to perform the numerical analysis of the

379: instanton equations we will fix the parameter $Z$ of the quartic potential to have the value

380: corresponding to central point

381: between the case when the false vacuum energy density $V_{fv}$ is negligible in comparison to

382: $V_M$ (this situation corresponds to the thin-wall approximation considered in \cite{cdl} and

383: is analyzed in part numerically in \cite{samuel}),

384: and the case when the false vacuum disappears. Since the energy density in false vacuum is given by

385: \bdis

386: V_{fv}=\frac{1}{12\lambda}\frac{(1+Z)(1-3Z)}{(1-Z)^2} ,

387: \edis

388: the fraction $V_{fv}/V_M$, that depends on $Z$ only, equals $1/2$ if

389: \bdis

390: \left(\frac{1+Z}{1-Z}\right)^3\frac{1-3Z}{1+3Z}=\frac{1}{2} \quad \mbox{with} \quad Z\in[0,1/3].

391: \edis

392: This equation determines $Z$ as

393: \be \label{Z_value}

394: Z\approx 0.278\ .

395: \ee

396: (For the values of $Z$ for which $V_{fv}/V_M\lessapprox 1/2$

397: the function $V_{fv}/V_M$ can be replaced with a good accuracy

398: by $-12Z+12/3$; this expression would give $Z=21/72\approx 0.292$.) We have solved numerically

399: the exact instanton equations with this choice of the parameter $Z$ and compared these results with the approximative

400: formulae (\ref{k2quant}) and (\ref{w4}), see figure \ref{f1}.

401: 

402: \begin{figure}[h]

403: \centering

404: %\psfrag{a}

405: %{$-\frac{V_M''}{H_M^2}$}

406: %\psfrag{b}{$\Phi_M,\ \Phi_{left}$}

407: %\includegraphics[]{phi_oc.eps}

408: \psfrag{a}

409: {$-\frac{V_M''}{H_M^2}$}

410: \psfrag{b}{$k_{num},\ k_{theor}$}

411: \includegraphics[]{k_oc.eps}

412: \psfrag{a}

413: {$-\frac{V_M''}{H_M^2}$}

414: \psfrag{b}{$H_M^2w_{4\ num},\ H_M^2w_{4\ theor}$}

415: \includegraphics[]{action_oc.eps}

416: \caption{The prediction of the analytical formula (\ref{k2quant}) for the instanton width in $\Phi$ direction

417: (lower, doted, line)

418: is compared with the numerical computations of this quantity in the left graph.

419: The

420: range of $-V_M''/H_M^2$ is taken $[4,10]$; at the value $4$ the CdL instanton of the first order appears and

421: at the value approximately $10$ the second order CdL instantons appear, \cite{vj1}, \cite{socdlja}.

422: Finally,

423: the right graph shows the theoretical dependence of the first-order CdL instanton action according to (\ref{w4})

424: (lower, doted, line) together with numerically obtained values of this quantity.}

425: \label{f1}

426: \end{figure}

427: 

428: \section{CdL instanton(s) of the first order in the case with subcritical value of the fourth derivative of the

429: effective potential at its top}

430: 

431: Let us consider an effective potential which has,

432: for a suitable choice of parameters, such a shape that the denominator in the formula (\ref{k2quant})

433: is negative and at the same time it is possible to change continuously the sign of the nominator. If the sign of the term

434: $-4-V_M''/H_M^2$ is positive, then there must be at least one CdL instanton of the first order,

435: as discussed previously. But what

436: happens when we pass through zero to negative values of $-4-V_M''/H_M^2$, keeping

437: $\left[8C+\frac{1}{48}\left(\frac{V_M'''}{H_M^2}\right)^2+\frac{1}{4}\frac{V_M''''}{H_M^2}\right]$ a negative

438: constant? The formula (\ref{k2quant}) ensures that we have the near-to-limit CdL instantons in the

439: region with $-V_M/H_M^2$ (a little bit) less than $4$. By the continuity argument, this set of instantons

440: must be lined-up to the "overcritical" instantons existing for $-V_M''/H_M^2>4$.

441: In order to investigate the structure of the

442: instanton solutions it is helpful to use the method of representation of a

443: CdL instanton proposed by Tanaka in \cite{tanaka}. Let us consider the two dimensional "phase" plane

444: $(\Pi,\Phi)$, where $\Phi$ stands for some value (to be determined later) of the inflaton and $\Pi$ stands for some

445: value of the conjugated momentum $2\pi^2a^3\Phi'$. For a given $V$ we can start the evolution,

446: using to the Euclidean equations of motion (\ref{ee}), (\ref{eel2}) for $a$ and $\Phi$, with

447: any initial value $\Phi_i$ of the inflaton and we will come, in some finite Euclidean time $\bar{\tau}(\Phi_i)$,

448: to the point at which

449: $a$ reaches its maximum. Let $\Phi_i^+$ be an arbitrary value of $\Phi$ located to

450: the right of $\Phi_M$. Taking this $\Phi_i^+$

451: as the initial value for the system (\ref{ee}), (\ref{eel2}) we obtain the point

452: $(\bar{\Pi}^+,\bar{\Phi}^+)\equiv (\Pi(\bar{\tau}),\Phi(\bar{\tau}))$, and

453: varying $\Phi_i^+$ we can draw the curve

454: \bdis

455: \mcal{C}^+=\left\{ (\bar{\Pi}^+,\bar{\Phi}^+),\ \Phi_i^+>\Phi_M\right\} .

456: \edis

457: Analogically, varying the initial value of the inflaton $\Phi_i^-$ located to the left of $\Phi_M$ we construct the curve

458: $\mcal{C}^-$. Intersections of the curves $\mcal{C}^+$ and $\mcal{C}^-$ correspond to the CdL instantons.

459: (The curve $\mcal{C}^+$ does not intersect itself and the same holds for $\mcal{C}^-$).

460: 

461: %\begin{figure}[h]

462: %\centering

463: %\psfrag{a}

464: %{$-\frac{V_M''}{H_M^2}$}

465: %\psfrag{b}{$\Phi$}

466: %\psfrag{c}{$(-V''_M/H_M^2)_{min}$}

467: %\includegraphics[]{schema.eps}

468: %\caption{The schematic picture of the instanton solutions structure around $-V''_M/H_M^2=4$ (the origin of the

469: %coordinate system, horizontal axis represents the values of $-V_M''/H_M^2$ and

470: %the vertical axis stands for the initial values of $\Phi$ of CdL instanton).

471: %For $-V''_M/H_M^2>4$ there is at least one CdL instanton of the first order. In some range

472: %$(-V''_M/H_M^2)_{min}<-V''_M/H_M^2<4$ there are at least two CdL instantons of the first order.}

473: %\label{schema}

474: %\end{figure}

475: 

476: The existence of two CdL instantons of the

477: first order for given $-V_M''/H_M^2$ opens the question which instanton governs the vacuum decay. Let us

478: investigate a concrete realization of the kind of vacuum decay described above. In the appendix it is shown that

479: for a class of generalizations of the quartic potential we cannot obtain negative value of the fourth derivative of the

480: potential at $\Phi_M$ if we require that the potential contains both the false and true vacuum. Let us

481: relax these requirements and consider the potential

482: \be \label{ipol}

483: V(\Phi)=\frac{3}{8\pi}-\frac{1}{2}g_2\Phi^2-\frac{1}{24}g_4\Phi^4 \ ,

484: \ee

485: where $g_2,g_4$ are positive constants. The top of the potential is at $\Phi_M=0$ and

486: \begin{align*}

487: &V_M''=-g_2\ ,&

488: &H_M^2=1\ ,&

489: &-\frac{V_M''}{H_M^2}=g_2\ .&

490: \end{align*}

491: The choice

492: \bdis

493: g_4=300

494: \edis

495: ensures that

496: \bdis

497: 8C+\frac{1}{48}\left(\frac{V_M'''}{H_M^2}\right)^2+\frac{1}{4}\frac{V_M''''}{H_M^2}=

498: 8C-\frac{1}{4}g_4

499: \edis

500: is negative. We have performed numerical analysis of the structure of CdL instanton solutions in this theory for values

501: of $-V_M''/H_M^2$ close to $4$ from both sides. The structure of the instanton solution is fully characterized by the

502: Tanaka's curves $\mcal{C}$. Since the potential (\ref{ipol}) is an even function of $\Phi$ we need only one of the curves

503: $\mcal{C}^+$ and $\mcal{C}^-$ ($\mcal{C}^-$ is the mirror image of $\mcal{C}^+$ with respect to vertical axis in the

504: $(\bar{\Pi},\bar{\Phi})$ plane).

505: The instanton solutions are determined by the points at which $\mcal{C}^+$

506: (or $\mcal{C}^-$) crosses the vertical axis. The Tanaka's curves $\mcal{C}^+$ with $-V_M''/H_M^2$ close to $4$

507: and $g_4=300$ for the theory (\ref{ipol}) are shown on the graphs in figure \ref{tcfig}. These graphs tell us that

508: for $V_M''/H_M^2$ close above $4$ there is only one (no near-to-limit) CdL instanton, for $-V_M''/H_M^2=4$ lying

509: between approximately $3.966$ and $4$ there are two CdL instantons, and for $-V_M''/H_M^2$ less than approximately

510: $3.966$ there are no CdL instantons. Finally, the structure of the instanton solutions in the theory (\ref{ipol}) and

511: the $-V_M''/H_M^2$-dependence of the instanton action are

512: shown in figure \ref{instrfig}.

513: 

514: 

515: \begin{figure}

516: \centering

517: \includegraphics[]{tc396.eps}

518: \includegraphics[]{tc398.eps}

519: \includegraphics[]{tc40.eps}

520: \includegraphics[]{tc41.eps}

521: \caption{The Tanaka's curves $\mcal{C}^+$ (i.e. $\bar{\Pi}^+$ versus $\bar{\Phi}^+$) in the theory (\ref{ipol})

522: with $g_4=300$. The

523: graphs are plotted, from the left to the right and from top to bottom,

524: for the values of $-V_M''/H_M^2=3.96,\ 3.98,\ 4.00$ and $4.10$ respectively.}

525: \label{tcfig}

526: \end{figure}

527: 

528: 

529: \begin{figure}

530: \centering

531: \includegraphics[]{inststr.eps}

532: \includegraphics[]{actionstr.eps}

533: \caption{Left graph: the initial value of the CdL instanton solution versus $-V_M''/H_M^2$ in the theory (\ref{ipol}) with

534: $g_4=300$.

535: There are two

536: bifurcation points in the parameter $-V_M''/H_M^2$ for solutions of

537: the instanton equations. At $-V_M''/H_M^2=4$ the number of

538: (nontrivial, i.e. no-HM-instanton) solutions changes from $1$ to $2$, and at a value approximately $3.966$ the

539: number of CdL instantons changes from $0$ to $2$. Right graph: the difference $\Delta S$ between the

540: action of CdL and HM intanton (the action of the HM instanton is normalized to $2$).

541: Lower curve describes the subcritical near-to-limit

542: CdL instanton and upper curve corresponds to the no-near-to-limit CdL instanton. These curves merge in the

543: point $-V_M''/H_M^2\approx 3.966$ mentioned above. For $-V_M''/H_M^2\gtrapprox 3.975$ the no-near-to-limit instanton governs

544: the vacuum decay, for the values below this the vacuum decay is governed by the HM instanton.

545: }

546: \label{instrfig}

547: \end{figure}

548: 

549: \section{Conclusion}

550: 

551: The false vacuum decay in a de Sitter universe has been investigated for near-to-critical values of the curvature

552: of the effective potential. An approximate formula for the Euclidean action of the near-to-limit CdL instanton has

553: been found by expanding the inflaton and the metric into to powers of the inflaton in a different way than in our

554: previous work \cite{vj2}. We have focused on the case when the fourth derivative

555: of the effective potential at its top has a subcritical value and $-V_M''/H_M^2$ is running from both sides

556: around its critical value $4$.

557: We conclude that there is a range of the parameter $-V_M''/H_M^2$ less than $4$ for which at least

558: two CdL instantons exist. One of them

559: is the near-to-limit instanton that can be described by the approximate formulas, together with its action,

560: derived in the first part of the paper.

561: The second instanton must exist because of the necessity to disconnect the energy curve from the potential when the

562: starting point of the curve moves towards the true vacuum \cite{vj1}.

563: The near-to-limit instanton in this case mediates the vaccum decay with

564: a less probability than the related HM instanton. However, the

565: vacuum decay is not governed by the HM instanton in this case but by the no-near-to-limit CdL instanton. On the

566: other hand, we have shown on a concrete example that for sufficiently small values of $-V_M''/H_M^2$ the HM

567: instanton has the least action from the three instantons in question and is to be considered as the instanton

568: governing the vacuum decay.

569: 

570: 

571: \noi

572: {\bf Acknowledgment}: I would like to thank Vlado Balek for fruitful discussions. This work was supported by the

573: Slovak grant VEGA 1/0250/03.

574: 

575: 

576: 

577: \appendix

578: %\setcounter{equation}{A}

579: 

580: \section{Non-negativeness of $V_M''''$ for a class of generalizations of quartic potential}

581: 

582: We will consider the following class of potentials

583: \be \label{polypot}

584: V_n=\frac{1}{2n}\Phi^{2n}-\frac{1}{2n+1}\delta\Phi^{2n+1}+\frac{1}{2n+2}\lambda\Phi^{2n+2}

585: \ee

586: with $n$ an arbitrary integer and $\lambda$ and $\delta$ are positive parameters. Requirement of non-negativeness of

587: $V_n$ and existence of both vacua leads to the restriction on $\delta$ at given $\lambda$

588: \bdis

589: \delta_m^{(n)}=2\sqrt{\lambda}<

590: \delta<\frac{2+\frac{1}{n}}{\sqrt{1+\frac{1}{n}}}\sqrt{\lambda}=\delta_M^{(n)} .

591: \edis

592: The width $\delta_M^{(n)}-\delta_m^{(n)}$ of the allowed interval of $\delta$ tends to zero as $n$ grows to infinity.

593: We want to answer the question whether it is possible to choose the parameters $\delta$ and $\lambda$ in such a way that

594: we obtain a negative denominator in the fraction on the right hand side of (\ref{k2quant}). Negativeness of

595: $V_M''''$ is obviously necessary for this. The top of the barrier of $V_n$ is reached at  $\Phi_M$ for which we have

596: \bdis

597: \delta\Phi_M=\frac{2}{1+Z}, \quad \mbox{with} \quad

598: Z=\sqrt{1-\frac{4\lambda}{\delta^2}} .

599: \edis

600: The range of the parameter $Z$ follows from the range of $\delta$, namely

601: \be \label{Zrestr}

602: Z\in\left[ 0,\frac{1}{2n+1}\right] .

603: \ee

604: The direct computation of $V''''_{n\ M}$ with the crucial help of the parametrization $(\lambda,Z)$ gives the result

605: \be \label{V_Mingeneralzed}

606: V''''_{n M}=\frac{6}{\lambda^{n-2}}\frac{(1-Z)^{n-2}}{(1+Z)^{n-2}}

607: \left[ -4n^2+4n-1+\frac{2n(2n-1)}{1+Z}\right] .

608: \ee

609: For any given $n$ there are obviously values of $Z$ (not fulfilling (\ref{Zrestr})) for which

610: $V''''_{nM}$ has both positive and negative values. The restriction (\ref{Zrestr}) changes

611: the situation. At $Z=0$ the expression (\ref{V_Mingeneralzed}) is positive. The zero point of (\ref{V_Mingeneralzed}) is

612: given by

613: \bdis

614: Z_0=\frac{(2n-1)2n}{4n^2+1-4n}-1\longrightarrow^{n\to\infty} 0 ,

615: \edis

616: and we easily find that $Z_0$ never belongs to the required range of $Z$ because

617: \bdis

618: Z_0-\frac{1}{2n+1}=\frac{4n-2}{(4n^2+1-4n)(2n+1)}>0 .

619: \edis

620: Therefore $V''''_{n\ M}$ cannot be negative.

621: 

622: 

623: 

624: \begin{thebibliography}{20}

625: \bibitem{cdl}

626: S. Coleman and F. de Luccia, Phys. Rev. D {\bf 21}, 3305 (1980).

627: \bibitem{guth}

628: A.H. Guth, Phys. Rev. {\bf D 23}, 347 (1981).

629: \bibitem{bgt}

630: M. Bucher, A.S. Goldhaber and N. Turok, Phys. Rev. D {\bf 52}, 3314 (1995), hep-ph/9411206.

631: \bibitem{lm}

632: A.D. Linde and A. Mezhlumian, Phys. Rev. D {\bf 52}, 6789 (1995), astro-ph/9506017.

633: \bibitem{lindekofman}

634: L. Kofman, A.D. Linde and A. Starobinsky, Phys. Rev. D {\bf 56}, 3258 (1997), hep-ph/9704452;

635: L. Kofman, A. Linde and A. Starobinsky,

636: Phys. Rev. Lett. 73, 3195 (1994), hep-th/9405187.

637: \bibitem{hm}

638: S.W. Hawking and I.G. Moss, Phys. Lett. {\bf 101 B}, 35 (1982).

639: \bibitem{js}

640: L.G. Jensen and P.J. Steinhardt, Nucl. Phys. {\bf B237}, 176 (1984).

641: \bibitem{vj1}

642: V. Balek and M. Demetrian, Phys. Rev. D {\bf 69}, 063518 (2004), gr-qc/0311040.

643: \bibitem{vj2}

644: V. Balek and M. Demetrian, Phys. Rev. D {\bf 71}, 023512 (2005), gr-qc/0409001.

645: \bibitem{socdlja} 

646: M. Demetrian, gr-qc/0408012.

647: \bibitem{samuel}

648: D.A. Samuel and W.A. Hiscock, Phys. Rev. D {\bf 44}, 3052 (1991).

649: \bibitem{tanaka}

650: T. Tanaka, Nucl. Phys. B {\bf 556}, 373 (1999), gr-qc/9901082.

651: \bibitem{weinberg}

652: J.C. Hackworth and E.J. Weinberg, Phys Rev. D {\bf 71}, 044014 (2005), hep-th/0410142.

653: \end{thebibliography}

654: 

655: 

656: 

657: \end{document}

658: