1: \expandafter\ifx\csname qd\endcsname\relax
2: \documentstyle[12pt,twoside]{article}\input{cbmacro}\fi
3: \expandafter\ifx\csname ok\endcsname\relax
4: \def\re{{\rm Re}} \def\im{{\rm Im}}
5: {\catcode`@=11 \@addtoreset{equation}{section}\@addtoreset{figure}{section}}
6: \def\theequation{\thesection .\arabic{equation}}
7: \def\thefigure{\thesection .\arabic{figure}}
8: \addtolength{\topmargin}{-48pt}
9: \addtolength{\textheight}{90pt}
10: \addtolength{\evensidemargin}{-31pt}
11: \addtolength{\oddsidemargin}{-12pt}
12: \addtolength{\textwidth}{30pt}
13: \addtolength{\footskip}{21pt}
14: \setlength{\parindent}{0pt}\frenchspacing
15: \begin{document}
16: \fi
17: %\addtocontents{toc}{\protect\vspace*{\tocsp}}
18: \section{Summary and conclusions}
19: We have linearized the semiclassical system of Schr\"odinger equation and
20: Einstein equations (\ref{EinSchro}) around the de~Sitter -- Bunch-Davies solution
21: in order to investigate the stability of this solution against small
22: fluctuations of the quantum state and small, spatially homogeneous and
23: isotropic perturbations of the k=0 FRW de~Sitter metric. This linearization
24: in the sense of a stability analysis is the only approximation appearing
25: in the present work.\\
26: The condition of finite energy density for the initial quantum state has been
27: discussed and the Schr\"odinger equation was completely solved. The expectation
28: value of the energy momentum tensor has been calculated as a functional of the
29: metric perturbation. The momentum integrations have been carried out
30: analytically. A procedure for the isolation of divergencies was developed.
31: These were removed by a renormalization through the subtraction of
32: De~Witt-Schwinger terms.\\
33: The linearized semiclassical Einstein equations became a linear
34: integro-differential equation for the metric perturbation. After finding out
35: the initial data which can be specified the general solution was obtained in
36: terms of its Laplace transform. This Laplace transform has been analyzed
37: analytically as well as numerically using the necessary high numerical
38: precision. The general solution contains only two potential instabilities:
39: a constant mode which corresponds just to a constant spatial rescaling of the
40: underlying de~Sitter spacetime i.e. a pure gauge, and an instability on the
41: Planck time scale which is outside of the scope of a semiclassical theory.\\
42: Thus we have shown that de~Sitter spacetime and Bunch-Davies vacuum are stable
43: within our semiclassical theory!\\
44: The same result was obtained in reference \cite{IsaRog} for the special case of
45: minimal coup\-ling $\xi=0$. Since there the momentum integrals are not evaluated
46: explicitly, complicated estimates were necessary in order to achieve this and
47: there is no numerical analysis.\\
48: Our conclusion above is in contradiction with some claims existing in the
49: literature. In particular in ref. \cite{Mottola} an instability on the Hubble
50: time scale $H_0^{-1}$ was found. They use indeed a different coordinate system
51: (k=+1 FRW), but the coordinate lines $t=\hbox{const.}$, on which the spatially
52: homogeneous initial data of the perturbation are specified, tend to coincide
53: within the k=0 and k=+1 FRW parametrizations for late times $t\to\infty$.
54: Therefore the answer to the stability issue should be the same.
55: Some criticism on appendix A of ref. \cite{Mottola} is already contained in
56: ref. \cite{IsaRog}. Furthermore the diverging perturbation $\s(\eta)$ of the
57: conformal factor found in ref. \cite[section 5]{Mottola} is very near to a
58: pure gauge mode. It can be transformed into our $I(t)$ remaining
59: small for all times and contains no physical divergence.\\
60: Finally a few possibilities for a continuation of this work should be
61: mentioned: Fermi fields could be included, but presumably they wouldn't alter
62: the stability argumentation. The investigation of the minimally coupled
63: massless case would be interesting, because the renormalized two point function
64: in de~Sitter spacetime (see refs. \cite{Allen, FoVi, AlFo})
65: as well as our result (\ref{drhoI}) are containing a
66: logarithmic infrared divergence (the linear infrared divergencies of the
67: $\psi\,$- and $G\,$-functions cancel). At last one could try to access the
68: region of quantum gravity via the Wheeler-De~Witt equation.\\[10mm]
69: I would like to thank W. Buchm\"uller for many helpful discussions in the
70: course of this work and D. Litim for many valuable comments on the
71: manuscript.
72:
73: \expandafter\ifx\csname ok\endcsname\relax
74: \end{document}\fi
75: