1:
2: \documentclass[prd,twocolumn,superscriptaddress,showpacs,nofootinbib,preprintnumbers]{revtex4}
3:
4:
5:
6:
7:
8:
9: \usepackage{amsmath}
10: \usepackage{amsfonts}
11: \usepackage{graphicx}
12: \usepackage{dcolumn}
13: %\usepackage{hyperref}
14:
15: \newcommand{\op}{\`}
16: \def\be{\begin{equation}}
17: \def\ee{\end{equation}}
18: \def\ba{\begin{eqnarray}}
19: \def\ea{\end{eqnarray}}
20: \def\bs{\begin{subequations}}
21: \def\es{\end{subequations}}
22: \renewcommand{\S}{{\text{\tiny $\phi$}}}
23: \newcommand{\T}{{\text{\tiny $T$}}}
24: \newcommand{\sH}{{\text{\tiny $\phi H$}}}
25: \newcommand{\sV}{{\text{\tiny $\phi V$}}}
26: \newcommand{\tH}{{\text{\tiny $TH$}}}
27: \newcommand{\tV}{{\text{\tiny $TV$}}}
28: \renewcommand{\H}{{\text{\tiny $H$}}}
29: \newcommand{\V}{{\text{\tiny $V$}}}
30: \newcommand{\eff}{{\text{\tiny eff}}}
31: \newcommand{\vp}{\varphi}
32: \newcommand{\p}{{\partial}}
33:
34: \usepackage{color}
35: \def\cob{\color{blue}}
36: \def\cor{\color{red}}
37:
38:
39: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
40:
41:
42:
43:
44: \begin{document}
45:
46: \title{Chaotic dynamics in preheating after inflation}
47:
48: \author{Yoshida Jin}
49: \email{jin@gravity.phys.waseda.ac.jp}
50: \affiliation{Department of Physics, Waseda University, Okubo 3-4-1,
51: Shinjuku, Tokyo 169-8555, Japan}
52: \author{Shinji Tsujikawa}
53: \email{shinji@nat.gunma-ct.ac.jp}
54: \affiliation{Department of Physics, Gunma National College of
55: Technology, Gunma 371-8530, Japan}
56: \date{\today}
57:
58: \begin{abstract}
59:
60: We study chaotic dynamics in preheating after inflation in which
61: an inflaton $\phi$ is coupled to another scalar field $\chi$
62: through an interaction $(1/2)g^2\phi^2\chi^2$. We first estimate the
63: size of the quasi-homogeneous field $\chi$ at the beginning of
64: reheating for large-field inflaton potentials $V(\phi)=V_0\phi^n$
65: by evaluating the amplitude of the $\chi$ fluctuations on scales
66: larger than the Hubble radius at the end of inflation.
67: Parametric excitations of the
68: field $\chi$ during preheating can give rise to chaos between two
69: dynamical scalar fields. For the quartic potential ($n=4$,
70: $V_0=\lambda/4$) chaos actually occurs for
71: $g^2/\lambda <{\cal O}(10)$ in a linear regime
72: before which the backreaction of created
73: particles becomes important. This analysis is supported by several
74: different criteria for the existence of chaos. For the quadratic
75: potential ($n=2$) the signature of chaos is not found by the time
76: at which the backreaction begins to work, similar to the case of
77: the quartic potential with $g^2/\lambda \gg 1$.
78: \end{abstract}
79: \pacs{98.80.Cq}
80:
81: \maketitle
82:
83:
84: %%%%%%%%%%%%%%%%%
85: \section{Introduction}
86: %%%%%%%%%%%%%%%%%
87:
88: Reheating after inflation is an extremely important stage
89: to generate elementary particles present in current universe.
90: In the original version of the reheating
91: scenario which is now called {\it old reheating}, the decay of
92: an inflaton field is characterized by a perturbative Born
93: process \cite{oldre}.
94: However this process is not efficient for the success of the
95: GUT-scale baryogenesis scenario.
96: Later it was found that the existence of a nonperturbative
97: stage--dubbed {\it preheating}-- can lead to an explosive
98: particle production prior to the Born
99: decay \cite{TB,KLS1,Boya}.
100:
101: During preheating scalar particles $\chi$ coupled to the inflaton
102: $\phi$ are efficiently generated by parametric resonance through
103: an interaction $(1/2)g^2\phi^2\chi^2$.
104: The existence of the preheating stage provides several interesting
105: possibilities such as the GUT-scale baryogenesis \cite{GUT},
106: nonthermal phase transition \cite{nonther}, the enhancement of
107: metric perturbations \cite{mpre} and
108: the formation of primordial black holes \cite{PBH}.
109: In the chaotic inflationary scenario characterized by the potential
110: $V(\phi)=V_0\phi^n$, the field perturbations $\delta \chi$ obey
111: the Mathieu equation (for $n=2$) or the Lame equation (for $n=4$),
112: which determines the structure of resonance at the linear regime.
113: When the backreaction of created particles begins to violate
114: the coherent oscillation of $\phi$, this tends to work to
115: suppress exponential growth of the field fluctuations.
116: The system enters a fully nonlinear stage after which the mode-mode
117: coupling (rescattering) between perturbations is
118: crucially important \cite{KT,KLS2}.
119:
120: Typically the contribution of the dynamical background field
121: $\chi$ is neglected in standard analysis of particle creations in
122: preheating. This may be justified for a quadratic inflaton
123: potential, since large-scale $\chi$ modes are exponentially
124: suppressed during inflation for the coupling $g$ required for
125: preheating \cite{Iva}. However the situation is different
126: for a quartic inflaton potential with the coupling $g$
127: of order $g^2/\lambda={\cal O}(1)$ \cite{Bruce,Fabio}.
128: In this case the quasi-homogeneous field $\chi$ can play an
129: important role for the dynamics of preheating.
130: In fact Podolsky
131: and Starobinsky \cite{Podolsky} pointed out that chaos may occur
132: for the self-coupling potential $V(\phi)=(1/4) \lambda \phi^4$
133: when the coupling $g^2/\lambda$ is not too much larger than of
134: order unity. Since it is not obvious whether chaos actually occurs
135: or not in this model only by analytic estimations, we shall
136: perform detailed numerical investigation with/without the
137: backreaction effect of created particles.
138: We shall estimate the size of the field $\chi$ at
139: the beginning of reheating by evaluating the amplitude of
140: the $\chi$ fluctuations for the modes larger than the
141: Hubble radius.
142: To judge the existence of chaos in preheating
143: we will adopt several different methods-- such as the
144: Toda-Brumer test \cite{Toda,Brumer}, Lyapunov exponents \cite{Lya}
145: and a fractal map \cite{Barrow}.
146:
147: It was already found that chaos appears for hybrid-type inflation
148: models \cite{EM,Bellido,Bastero,Zibin}
149: (see also Refs.~\cite{Cornish,BT,Joras}).
150: Hybrid inflation is a rather special model in a sense that the
151: symmetry breaking
152: field automatically grows by tachyonic instability even if it is
153: suppressed
154: during inflation. The necessary condition for chaos is that there
155: exist at least two
156: dynamical fields and neither of them is too much smaller than another
157: field.
158: The hybrid model satisfies this condition, since two fields can have
159: frequencies
160: which are the same order after symmetry breaking.
161: The presence of mixing terms between two fields
162: leads to a new instability of perturbations in addition to
163: tachyonic/resonance
164: instabilities \cite{Bastero}.
165: This new type of instability is clearly associated with the presence
166: of chaos.
167:
168: In this work we shall investigate the existence of chaos for
169: large-field potentials $V(\phi)=V_0\phi^n$ with an interaction
170: $(1/2)g^2\phi^2\chi^2$.
171: We estimate the variance of large-scale modes in $\chi$ at
172: the end of inflation, which is relevant to the initial condition of
173: the
174: quasi-homogeneous field $\chi$ for preheating.
175: The field $\chi$ is amplified by parametric
176: resonance, which can give rise to chaos after $\chi$ grows to
177: satisfy the Toda-Brumer condition.
178: We shall numerically solve background equations
179: together with perturbed equations for both quartic ($n=4$) and
180: quadratic ($n=2$) potentials.
181: Our main interest is to find the signature of chaos and the parameter
182: range of
183: the coupling $g$ in which chaos can be seen before the backreaction
184: effect of
185: created particles becomes important.
186: Since chaos can alter the standard picture of preheating by
187: parametric resonance,
188: it is of interest to clarify the situation in which chaos appears.
189:
190: Recent observations suggest that the quartic potential
191: $V(\phi)=(1/4)\lambda \phi^4$ is under an observational pressure,
192: while the quadratic potential $V(\phi)=(1/2)m^2\phi^2$ is allowed
193: \cite{obser}. This depends on the number of $e$-folds $N$ before
194: the end of inflation at which observable perturbations are
195: generated. In the case of quartic potential this corresponds to $N
196: \sim 64$ by assuming instant transitions between several
197: cosmological epochs \cite{efolds}. The likelihood analysis
198: including WMAP and SDSS datasets shows that the quartic potential
199: is marginally allowed by using $N=64$ \cite{Tegmark}. Therefore it
200: is premature to rule out this model completely from current
201: observations. The quartic potential corresponds to a system in
202: which the background equations can be reduced to the ones in
203: Minkowski spacetime by introducing conformal variables. This has
204: an advantage for the investigation of chaotic dynamics during
205: preheating. As we see later, the quartic potential exhibits a
206: stronger chaos compared to the one for the quadratic potential.
207:
208:
209:
210: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
211: \section{The field variance for long wavelength modes}
212: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
213:
214: In this section we shall estimate the field variance
215: for long wavelength modes of the field
216: $\chi$ coupled to the inflaton $\phi$ through an
217: interaction $(1/2)g^2\phi^2\chi^2$.
218: The effective potential in our system is
219: %
220: \begin{eqnarray}
221: V(\phi, \chi)=V_0\phi^n+\frac12 g^2\phi^2\chi^2\,.
222: \label{effpo}
223: \end{eqnarray}
224: %
225: We are mainly interested in two inflaton potentials:
226: (i) the quadratic one ($n=2$) and (ii) the quartic one ($n=4$).
227: In this work we do not implement nonminimal
228: couplings \cite{nonmini}
229: between the field $\chi$ and the scalar curvature $R$.
230:
231: In a flat Friedmann-Lemaitre-Robertson-Walker (FLRW) background
232: with a scale factor $a$, the background equations
233: for the system (\ref{effpo}) are
234: %
235: \begin{eqnarray}
236: \label{back2}
237: & &\ddot{\phi}+3H\dot{\phi}+V_{\phi}=0\,,\\
238: \label{back3}
239: & &\ddot{\chi}+3H\dot{\chi}+V_{\chi}=0\,, \\
240: \label{back}
241: & & H^2 \equiv \left(\frac{\dot{a}}{a}\right)^2
242: =\frac{\kappa^2}{3}
243: \left[\frac12 \dot{\phi}^2+ \frac12 \dot{\chi}^2+
244: V(\phi,\chi)\right]\,,
245: \end{eqnarray}
246: %
247: where $V_\phi=nV_0\phi^{n-1}+g^2\chi^2\phi$,
248: $V_\chi=g^2\phi^2\chi$ and
249: $\kappa^2 = 8\pi/m_p^2$ with $m_p$
250: being the Planck mass.
251:
252: We define the number of $e$-folds before the end of inflation, as
253: %
254: \begin{eqnarray}
255: N \equiv {\rm ln}\,(a_f/a(t))\,,
256: \end{eqnarray}
257: %
258: where $a_f$ is the scale factor at the end of inflation.
259: Employing the slow-roll approximation, $|\ddot{\phi}| \ll
260: |3H\dot{\phi}|$
261: and $\dot{\phi}^2 \ll V(\phi)$, we easily find that
262: ${\rm d}\phi/{\rm d}N=n/(\kappa^2\phi)$.
263: Here we neglect the contribution coming from the $\chi$-dependent
264: terms.
265: Integrating this equation, we obtain
266: %
267: \begin{eqnarray}
268: \phi^2-\phi_f^2=\frac{2n}{\kappa^2}N\,.
269: \end{eqnarray}
270: %
271: The field value at the end of inflation ($\phi_f$) is determined
272: by setting the slow-roll parameter, $\epsilon_{s} \equiv (1/2\kappa^2)
273: (V_\phi/V)^2$, is unity.
274: This gives $\phi_f/m_p=n/(4\sqrt{\pi})$,
275: thereby leading to
276: %
277: \begin{eqnarray}
278: \label{phislow}
279: \phi^2=\frac{n}{4\pi} \left(N+\frac{n}{4}\right)
280: m_p^2\,.
281: \end{eqnarray}
282: %
283:
284: Let us consider a perturbation $\delta \chi$ in the field
285: $\chi$. Then the momentum-space first-order perturbed
286: equation is given by
287: %
288: \begin{eqnarray}
289: \label{dchikeq}
290: \delta \ddot{\chi}_{k}+3H\delta \dot{\chi}_{k}+
291: \left(\frac{k^2}{a^2}+g^2\phi^2 \right)\delta \chi_{k}=0\,,
292: \end{eqnarray}
293: %
294: where $k$ is a comoving wavenumber.
295: Note that we neglected the backreaction of
296: gravitational perturbations.
297: When the effective mass of the field $\chi$ is larger than
298: of order the Hubble rate, i.e., $g^2\phi^2>(3H/2)^2$,
299: the evolution of super-Hubble perturbations is characterized by
300: underdamped oscillations with the dependence \cite{Iva}
301: %
302: \begin{eqnarray}
303: \label{chi1}
304: \delta \chi_{k} \propto a^{-3/2}\,.
305: \end{eqnarray}
306: %
307: Hence large-scale perturbations are exponentially suppressed
308: during inflation.
309: Meanwhile when $g^2\phi^2<(3H/2)^2$ super-Hubble
310: perturbations evolve as
311: %
312: \begin{eqnarray}
313: \label{chi2}
314: \delta \chi_{k} \propto \exp \left[ -\left(
315: \frac32 H-\sqrt{\frac94 H^2-g^2\phi^2}
316: \right){\rm d}t \right]\,.
317: \end{eqnarray}
318: %
319: This shows that in the massless limit ($g^2\phi^2 \ll H^2$)
320: the amplitude of $\delta \chi_{k}$ decreases very slowly.
321: In what follows we shall consider the quadratic and quartic
322: models separately.
323:
324: %
325: \subsection{Quadratic model}
326: %
327:
328: For the inflation potential $V(\phi)=\frac12m^2\phi^2$,
329: the coupling $g$ is required to be greater than of order
330: $10^{-4}$ in order for preheating to occur \cite{KLS2}.
331: In this case the resonance parameter, $q \equiv g^2\phi^2/(4m^2)$,
332: is much larger than 1 at the beginning of preheating \cite{KLS2}.
333: Since $H \sim m$ at the end of inflation,
334: the existence of the preheating stage demands the condition
335: $g^2\phi^2 \gg H^2$ during inflation.
336: Hence the evolution of large-scale $\chi$ fluctuations
337: is characterized by Eq.~(\ref{chi1}).
338:
339: Let us consider the modes which are outside the Hubble radius at
340: the end of inflation ($0<k<k_f=a_{f}H_{f})$.
341: Since these modes are effectively massive with slowly changing mass,
342: they can be treated as an adiabatic state characterized by
343: $\delta \chi_{k}=1/(a^{3/2}\sqrt{2\omega_{k}})$ where
344: $\omega_k^2 \simeq g^2\phi^2$.
345: Hence the amplitude $|\delta \chi_{k}|^2$ at the end of
346: inflation is estimated as
347: %
348: \begin{eqnarray}
349: \label{chiend}
350: |\delta \chi_{k} (t_{f})|^2=\frac{1}{2a_{f}^3g\phi_f}\,.
351: \end{eqnarray}
352: %
353:
354: Then we can obtain the variance of the fluctuation
355: $\delta \chi_{k}$ for $k<k_f$:
356: %
357: \begin{eqnarray}
358: \langle \delta \chi^2_{k} (t_{f}) \rangle_{k<k_{f}}
359: &=&
360: \frac{1}{2\pi^2} \int_{0}^{k_f}
361: k^2 |\delta \chi_{k} (t_{f})|^2 {\rm d} k \nonumber \\
362: &=& \frac{\phi_{f}^2}{9\pi g}
363: \sqrt{\frac{4\pi}{3}}
364: \left(\frac{m}{m_{p}}\right)^3\nonumber \\
365: \label{dchik}
366: &=& \frac{1}{9\pi \sqrt{12\pi} g}
367: \left(\frac{m}{m_{p}}\right)^3 m_{p}^2\,,
368: \end{eqnarray}
369: %
370: where we used $\phi_{f}=m_{p}/(2\sqrt{\pi})$
371: in addition to the slow-roll approximation
372: $H^2 \simeq 4\pi m^2\phi^2/3m_p^2$.
373:
374: For the quadratic potential ($n=2$) the inflaton mass is
375: constrained to be $m \simeq 10^{-6}m_p$ from
376: the COBE normalization \cite{LL}.
377: Then Eq.~(\ref{dchik}) shows that the variance is
378: the function of $g$ only.
379: In Fig.~\ref{n2} we plot $\sqrt{\langle \delta \chi^2_{k}
380: (t_{f}) \rangle_{k<k_f}}$ as a function of $g$.
381: It can be regarded as a quasi-homogeneous mode
382: at the beginning of preheating.
383: This at least measures the minimum amplitude
384: of the homogeneous $\chi$ field.
385:
386: %%%%%%%%%%
387: \begin{figure}
388: \begin{center}
389: \includegraphics[height=3.2in,width=3.4in]{fig1.eps}
390: \caption{The amplitude of the variance
391: $\langle \delta \chi^2_{k}
392: (t_{f}) \rangle_{k<k_f}^{1/2}$ at the end of inflation
393: in terms of the function of the coupling $g$
394: for the quadratic potential ($n=2$).
395: The variance gets smaller for larger $g$.
396: }
397: \label{n2}
398: \end{center}
399: \end{figure}
400: %%%%%%%%%%
401:
402:
403: \subsection{Quartic model}
404:
405: For the quartic potential $V(\phi)=(1/4)\lambda\phi^4$,
406: it is known that preheating occurs even when the coupling
407: $g$ is in the range $g^2\phi^2 \lesssim H^2$ \cite{Kaiser,GKLS}.
408: When $g^2/\lambda={\cal O}(1)$, for example, the background
409: dynamics transits from the ``massless regime''
410: [$g^2\phi^2<(3H/2)^2$] to the ``massive regime''
411: [$g^2\phi^2>(3H/2)^2$] during inflation \cite{Bruce,Fabio,Zibin2}.
412: The critical number of $e$-folds, $N_c$, in which the
413: the evolution of the perturbation $\delta \chi_{k}$ transits
414: from Eq.~(\ref{chi2}) to Eq.~(\ref{chi1}) is determined by the
415: condition
416: $9H_c^2/4=g^2\phi_{c}^2$, which gives \cite{Zibin2}
417: %
418: \begin{eqnarray}
419: N_{c}=\frac23 \frac{g^2}{\lambda}
420: =\ln \left(\frac{a_{f}}{a_{c}}\right)
421: \simeq \ln \left(\frac{k_{f}}{k_{c}}\right)\,.
422: \label{NC}
423: \end{eqnarray}
424: %
425: We note that we used the slow-roll condition
426: %
427: \begin{eqnarray}
428: \label{Quarback}
429: H^2 \simeq \frac{2\pi \lambda \phi^4}{3m_{p}^2},\quad
430: \phi^2 \simeq \frac{N}{\pi}m_p^2\,.
431: \end{eqnarray}
432: %
433:
434: The modes which are inside the Hubble radius at transition
435: time $t=t_{c}$
436: (but larger than the Hubble radius at the end of inflation)
437: evolves as effective massive fields for $t>t_{c}$.
438: Then for $k_{c}<k<k_{f}$ the amplitude
439: $|\delta \chi_{k}|^2$ at the end of
440: inflation is given by Eq.~(\ref{chiend}),
441: which gives the variance
442: %
443: \begin{eqnarray}
444: \langle \delta \chi^2_{k} (t_{f})
445: \rangle_{k_{c}<k<k_{f}}
446: &=&
447: \frac{1}{2\pi^2} \int_{k_{c}}^{k_f}
448: k^2 |\delta \chi_{k} (t_{f})|^2 {\rm d} k \nonumber \\
449: \label{delchikd}
450: &=& \frac{\lambda}{18\pi^3}
451: \sqrt{\frac{2\lambda}{3g^2}}
452: (1-e^{-2g^2/\lambda})m_p^2\,, \nonumber \\
453: \end{eqnarray}
454: %
455: where we used the relation $k_{c}=k_{f}\exp (-2g^2/3\lambda)$
456: coming from Eq.~(\ref{NC}).
457:
458: For $k<k_{c}$ the modes exit the Hubble radius before
459: the transition time $t=t_c$.
460: At the Hubble radius crossing ($t=t_k$) the amplitude of the
461: perturbation
462: $\delta \chi_{k}$ is given by
463: %
464: \begin{eqnarray}
465: |\delta \chi_k(t_{k})|^2=\frac{H^2(t_{k})}{2k^3}\,,
466: \end{eqnarray}
467: %
468: which comes from the quantization of a standard massless
469: scalar field \cite{LL}.
470: Since $\delta\chi_k$ evolve as Eq.~(\ref{chi2}) for
471: $t_{k}<t<t_{c}$, we find that the amplitude of the
472: perturbations at $t=t_c$ is given by \cite{Zibin2}
473: %
474: \begin{eqnarray}
475: |\delta \chi_{k} (t_{f})|^2=
476: \frac{H^2(t_{k})}{2k^3}e^{-3F(N_{k})}\,,
477: \end{eqnarray}
478: %
479: where
480: %
481: \begin{eqnarray}
482: F(N_k)&=&N_{k}-N_{c}-\sqrt{N_{k}(N_{k}-N_{c})}
483: \nonumber \\
484: & &+N_{c} \ln \left( \frac{\sqrt{N_{k}}+
485: \sqrt{N_{k}-N_{c}}}{\sqrt{N_{c}}}\right)\,.
486: \end{eqnarray}
487: %
488: Here $N_{k}$ is the number of $e$-folds at $t=t_{k}$.
489: The perturbations evolve as Eq.~(\ref{chi2})
490: for $t_{c}<t<t_{f}$.
491: Using Eq.~(\ref{NC}) we obtain
492: the following amplitude at the end of inflation:
493: %
494: \begin{eqnarray}
495: |\delta \chi^2_{k} (t_{f})|=
496: \frac{H^2(t_{k})}{2k^3}e^{-3F(N_{k})-2g^2/\lambda}\,.
497: \end{eqnarray}
498: %
499: Then the variance of perturbations $\delta \chi_{k}$
500: for the modes $k<k_{c}$ is given by
501: %
502: \begin{eqnarray}
503: \langle \delta \chi_{k}^2 (t_{f}) \rangle_{k<k_{c}}
504: &=&
505: \frac{1}{2\pi^2} \int_{k_i}^{k_c}
506: |\delta \chi_{k} (t_{f})|^2 k^3 {\rm d}(\ln k) \nonumber \\
507: &=& \frac{1}{4\pi^2}
508: \int_{N_{c}}^{N_{i}} {\rm d}N_k H^2(t_{k})
509: e^{-3F(N_{k})-2g^2/\lambda}\,, \nonumber \\
510: &=& \frac{\lambda m_p^2}{6\pi^3}
511: \int_{N_{c}}^{N_{i}} {\rm d}N_k
512: N_k^2 e^{-3F(N_{k})-2g^2/\lambda}\,,
513: \end{eqnarray}
514: %
515: where we used Eq.~(\ref{Quarback}) and
516: ${\rm d}(\ln k) \simeq {\rm d}(\ln a)=-{\rm d}N_k$.
517: Note that $k_{i}$ is the minimum wavenumber relevant for
518: the maximum scale of cosmological perturbations.
519: In order to obtain
520: $\langle \delta \chi^2_{k} (t_{f}) \rangle_{k<k_{c}}$,
521: it is necessary to solve the following differential equation:
522: %
523: \begin{eqnarray}
524: \frac{{\rm d}}{{\rm d}N_{k}}
525: \langle \delta \chi^2_{k} (t_{f}) \rangle_{k<k_{c}}
526: =\frac{\lambda m_p^2}{6\pi^3}
527: N_k^2 e^{-3F(N_{k})-2g^2/\lambda},
528: \end{eqnarray}
529: %
530: which should be integrated from $N_{c}=2g^2/3\lambda$
531: to $N_{i}$. Note that $N_{i}$ roughly corresponds to the
532: total number of $e$-folds during inflation.
533: At least we require the condition $N_{i}>60$.
534: We shall choose $N_{i}=60, 100, 1000$
535: in order to see the sensitivity for the change of this number.
536:
537: Finally the variance for the modes $k<k_{f}$ is given by the
538: following sum:
539: %
540: \begin{eqnarray}
541: \langle \delta \chi^2_{k} (t_{f}) \rangle_{k<k_f}
542: =\langle \delta \chi^2_{k} (t_{f}) \rangle_{k_{c}<k<k_{f}}
543: +\langle \delta \chi^2_{k} (t_{f}) \rangle_{k<k_c}\,.
544: \end{eqnarray}
545: %
546: When $g^2/\lambda \lesssim {\cal O}(1)$ one has
547: $N_{c} \lesssim {\cal O}(1)$. This shows that most of
548: the contribution to the variance
549: $\langle \delta \chi^2_{k} (t_{f}) \rangle_{k<k_f}$
550: comes from the modes $k<k_{c}$.
551: Meanwhile when $g^2/\lambda \gg 1$, i.e., $N_{c} \gg 1$,
552: the modes $k_{c}<k<k_{f}$ dominate the total variance.
553: In Fig.~\ref{n4} we plot $\sqrt{\langle \delta \chi^2_{k} (t_{f})
554: \rangle_{k<k_f}}$ as a function of $g^2/\lambda$
555: for $N_{i}=60, 100, 1000$ and $\lambda=10^{-13}$.
556: We find that the variance does not depend on the
557: values $N_{i}$ for $g^2/\lambda \gtrsim 3$, which reflects the fact that
558: Eq.~(\ref{delchikd}) is independent of $N_{i}$.
559: The difference appears for $g^2/\lambda \lesssim 3$,
560: because of the fact that
561: $\langle \delta \chi^2_{k} (t_{f}) \rangle_{k<k_c}$
562: is dependent on the values $N_{i}$.
563: We shall use the values $\langle \delta \chi^2_{k} (t_{f})
564: \rangle_{k<k_f}$
565: obtained for $N_{i}=60$ as a minimum initial condition
566: of the quasi-homogeneous $\chi$ field
567: at the beginning of preheating.
568:
569: %%%%%%%%%%
570: \begin{figure}
571: \begin{center}
572: \includegraphics[height=3.2in,width=3.4in]{fig2.eps}
573: \caption{The amplitude of the variance
574: $\langle \delta \chi^2_{k}
575: (t_{f}) \rangle_{k<k_f}^{1/2}$ at the end of inflation
576: in terms of the function of the coupling $g$
577: for the quartic potential ($n=4$).
578: The variance is dominated by the modes $k_{c}<k<k_{f}$
579: for $g^2/\lambda \gg 1$, whereas the dominant contribution
580: comes from the modes $k<k_{c}$ for
581: $g^2/\lambda<{\cal O}(1)$.
582: }
583: \label{n4}
584: \end{center}
585: \end{figure}
586: %%%%%%%%%%
587:
588: We note that Podolsky and Starobinsky \cite{Podolsky}
589: estimated the size of the quasi-homogeneous $\chi$ field
590: for $g^2/\lambda={\cal O}(1)$
591: by using a Fokker Planck equation \cite{Sta}.
592: This equation is valid when the mass of the
593: field $\chi$ is smaller than of order the Hubble
594: rate \cite{NNS,Hosoya89,Salopek},
595: which corresponds to $g^2/\lambda \lesssim {\cal O}(1)$.
596: We checked that our estimation of the variance shows good agreement
597: with the one based on the Fokker Planck approach for
598: $g^2/\lambda \lesssim {\cal O}(1)$.
599: In the parameter regime $g^2/\lambda \gg 1$
600: the Fokker Planck equation is no longer valid.
601: Hence we need to use our estimation given above
602: in order to know the size of the quasi-homogeneous $\chi$ field
603: at the end of inflation.
604:
605: %%%%%%%%%%%%%%%%%%%%%%%%
606: \section{Basic properties of preheating and chaos}
607: %%%%%%%%%%%%%%%%%%%%%%%%
608:
609: %%%%%%%%%%%%%%%%%%%%%%%%
610: \subsection{Preheating and the role of
611: the quasi-homogeneous field $\chi$}
612: %%%%%%%%%%%%%%%%%%%%%%%%
613:
614: In the presence of the coupling $(1/2)g^2\phi^2\chi^2$, the
615: coherent oscillation of the inflaton leads to the excitation of
616: the field $\chi$ during preheating through the resonance term
617: $g^2\phi^2 \chi$ in Eq.\,(\ref{back3}). For the quadratic potential
618: ($n=2$) parametric resonance is efficient when the condition,
619: $q= g^2\phi^2/(4m^2) \gg 1$, is satisfied. To be
620: more precise the field $\chi$ grows for $g \gtrsim 3.0 \times
621: 10^{-4}$ by overcoming the friction due to cosmic expansion
622: \cite{KLS2}. In this model the growth of $\chi$ ends when the
623: system enters a narrow resonance regime ($q \lesssim 1$) or the
624: backreaction effect of created particles breaks the coherent
625: oscillation of the field $\phi$. For the quartic potential ($n=4$)
626: resonance bands exist for the parameter space characterized by
627: $n(2n-1)<g^2/\lambda<n(2n+1)$, where $n=1, 2, 3,...$
628: \cite{Kaiser,GKLS}. The center of the band, $g^2/\lambda=2n^2$,
629: corresponds to the largest Floquet index ($\mu_{\rm max} \simeq
630: 0.28$).
631:
632: If the field $\chi$ is strongly suppressed during inflation,
633: this does not contribute to the background dynamics
634: even if it is amplified during
635: reheating\footnote{The amplification of $\chi$ is limited by
636: the backreaction effect of created particles.}.
637: As we see in Figs.~\ref{n2} and \ref{n4}, it is expected
638: that $\chi$ does not dynamically
639: become important during preheating for larger $g$.
640: However, for the quartic potential, parametric resonance
641: takes place even for $g^2/\lambda={\cal O}(1)$,
642: in which case the quasi-homogeneous $\chi$ field
643: does not suffer from strong
644: suppression during inflation.
645: Then if the quasi-homogeneous field $\chi$ is amplified during
646: preheating,
647: this can contribute to the background dynamics.
648:
649: In our model the equations for field perturbations in Fourier space
650: may be written as \cite{Podolsky}
651: %
652: \begin{eqnarray}
653: \label{dphi}
654: & &\delta \ddot{\phi}_k+3H\delta\dot{\phi}_k+
655: \left[\frac{k^2}{a^2}+n(n-1)V_0\phi^{n-2}+g^2\chi^2\right]
656: \delta \phi_k \nonumber \\
657: & & =-2g^2\phi\chi \delta \chi_k \,, \\
658: & &\delta \ddot{\chi}_k+3H\delta\dot{\chi}_k+
659: \left(\frac{k^2}{a^2}+g^2\phi^2\right)\delta \chi_k
660: \nonumber \\
661: & & =-2g^2\phi\chi \delta \phi_k\,,
662: \label{dchi}
663: \end{eqnarray}
664: %
665: where $k$ is a comoving wavenumber.
666: This corresponds to the equations in which
667: the contributions from metric perturbations
668: are dropped (see Ref.~\cite{mpre}).
669: Since metric perturbations are enhanced only when field
670: fluctuations grow sufficiently, it is a good approximation to
671: neglect them except for a nonlinear stage of preheating.
672:
673: The terms on the r.h.s. of Eqs.~(\ref{dphi}) and (\ref{dchi}) are
674: not usually taken into account in standard analysis of preheating
675: \cite{TB,KLS1,KLS2}, since the field $\chi$ was supposed to be
676: dynamically unimportant. However this can give considerable
677: contributions to the dynamics of field perturbations provided that
678: the quasi-homogeneous field $\chi$ is not strongly suppressed
679: during inflation.
680: In fact these terms lead to a mixing between the perturbations of
681: two fields, whose behavior is absent in standard analysis of
682: preheating unless rescattering effects are taken into account at a
683: nonlinear stage.
684:
685: We start integrating background and perturbation
686: equations from the end of inflation.
687: We adopt the Bunch-Davies vacuum state for the initial
688: condition of perturbations for the modes inside the
689: Hubble radius.
690: The total variances of the fields $\varphi=\phi, \chi$
691: integrated in terms of $k$ are
692: %
693: \begin{eqnarray}
694: \langle \delta \varphi^2 \rangle=
695: \frac{1}{2\pi^2} \int k^2
696: |\delta \varphi_k|^2 {\rm d}k\,.
697: \end{eqnarray}
698: %
699: We implement the variances $\langle \delta \phi^2 \rangle$
700: and $\langle \delta \chi^2 \rangle$
701: for both background and perturbation equations
702: as a Hartree approximation \cite{KLS2}.
703: We note that this is for estimating
704: the time at which backreaction effects become important.
705: After the system enters a fully nonlinear stage,
706: one can not trust the analysis using the Hartree approximation.
707: Our interest is to find a signature of chaos
708: before the backreaction sets in.
709:
710:
711: \subsection{The condition for chaos}
712:
713: In this subsection we review several conditions for the
714: existence of chaos and apply them to our effective
715: potential (\ref{effpo}).
716: Let us consider the first-order differential equations
717: %
718: \begin{eqnarray}
719: \label{dyna}
720: \dot{x}_i=F_i(x_j)\,,
721: \end{eqnarray}
722: %
723: and their linearized equations,
724: %
725: \begin{eqnarray}
726: \delta \dot{x}_i
727: =\frac{\partial F_i}
728: {\partial x_j}\, \delta x_j\,,
729: \label{eq:linearized}
730: \end{eqnarray}
731: %
732: where $\delta x_i$ is the perturbation vector connecting
733: two nearby trajectories
734: and $\partial F_i/\partial x_j$ is the Jacobian matrix
735: of $F_i(x_j)$.
736:
737: The scalar-field equations (\ref{back2}) and (\ref{back3}) are
738: expressed
739: by the form (\ref{dyna}) by setting $x_1=\phi$, $x_2=\dot{\phi}$,
740: $x_3=\chi$ and $x_4=\dot{\chi}$.
741: Then one can evaluate the Jacobian matrix
742: $\partial F_i/\partial x_j$ and its eigenvalues $\mu$
743: for a general system characterized by an effective potential
744: $V=V(\phi, \chi)$.
745: Note that we do not account for the linearized
746: equation for $H$, since metric perturbations are neglected
747: in our analysis.
748: The eigenvalues of the matrix
749: $\partial F_i/\partial x_j$ are given by
750: %
751: \begin{eqnarray}
752: \label{mu}
753: \mu=\frac12 \left[-3H \pm \sqrt{9H^2+4\gamma}
754: \right]\,,
755: \end{eqnarray}
756: %
757: where
758: %
759: \begin{eqnarray}
760: \label{gam}
761: \gamma&=&\frac12 \biggl[ -(V_{\phi\phi}+V_{\chi\chi})
762: \nonumber \\
763: & & \pm
764: \sqrt{(V_{\phi\phi}+V_{\chi \chi})^2-4(V_{\phi \phi}
765: V_{\chi \chi}-V_{\phi \chi}^2)}\biggr]\,.
766: \end{eqnarray}
767: %
768:
769: The necessary condition for the existence of chaos is that one of the
770: eigenvalues is at least positive. In an expanding background ($H>0$)
771: this corresponds to the condition $\gamma>0$ from Eq.~(\ref{mu}).
772: Since we are now considering a situation in which both effective
773: masses of
774: $\phi$ and $\chi$ are positive ($V_{\phi\phi}>0$, $V_{\chi\chi}>0$),
775: $\gamma$ can take a positive value when
776: %
777: \begin{eqnarray}
778: V_{\phi \phi} V_{\chi \chi}-V_{\phi \chi}^2<0\,.
779: \end{eqnarray}
780: %
781: This is so-called the Toda-Brumer test \cite{Toda,Brumer}
782: that is used to judge the existence of chaos.
783: For our effective potential (\ref{effpo})
784: this translates into the condition
785: %
786: \begin{eqnarray}
787: \label{chicon}
788: \chi^2>\frac{n(n-1)V_0}{3g^2}\phi^{n-2}\,.
789: \end{eqnarray}
790: %
791:
792: When $n=2$ and $V_0=m^2/2$, this corresponds to
793: %
794: \begin{eqnarray}
795: \label{chin=2}
796: \chi>\frac{m}{\sqrt{3}g}\,,
797: \end{eqnarray}
798: %
799: whereas for $n=4$ and $V_0=\lambda/4$ we get
800: %
801: \begin{eqnarray}
802: \chi>\sqrt{\frac{\lambda}{g^2}}\,\phi\,.
803: \label{chin=4}
804: \end{eqnarray}
805: %
806: For the quadratic potential the initial value of $\chi$ is
807: $10^{-4}\sqrt{g}$ times smaller
808: than the value which leads to chaotic instability,
809: see Eqs.~(\ref{dchik}) and (\ref{chin=2}).
810: Therefore the system is expected to enter a chaotic phase after
811: the field is amplified more than $10^4/\sqrt{g}$ times.
812: For the quartic potential the condition for chaos is not so severe
813: compared to the quadratic potential, since the term on the
814: r.h.s. of Eq.~(\ref{chin=4}) decreases with time.
815:
816: It is worth commenting on the difference about the instabilities of
817: chaos and parametric resonance.
818: Although the field $\chi$ exhibits an exponential growth by
819: resonance, this is different from the chaotic instability
820: in which the evolution of the system is very sensitive to slight
821: change of initial conditions.
822: In fact none of the
823: eigenvalues of the Jacobi matrix is positive in the regime
824: where the condition (\ref{chicon}) is not satisfied.
825: This means that chaos is absent in the region
826: $\chi^2<n(n-1)V_0/(3g^2)\phi^{n-2}$ even if the
827: field shows an exponential growth by parametric resonance.
828:
829: Since the Toda-Brumer test is not a sufficient condition for the
830: existence of chaos, we shall use other criteria as well such as
831: Lyapunov exponents \cite{Lya}.
832: The Lyapunov exponents measure the logarithm of
833: of the expansion of a small volume in a $N$-dimensional
834: phase space. In this case the system possesses $N$
835: Lyapunov exponents. Chaos is accompanied by an
836: increase of the size of the $N$-volume at least in one direction and
837: a maximal Lyapunov exponent $h$ characterizes
838: the signature of chaos.
839: When $h$ approaches a positive constant asymptotically,
840: this shows the existence of chaos since the initial displacement of
841: the $N$-volume grows exponentially.
842: If $h$ approaches to 0 asymptotically, this means the
843: absence of chaos since the orbits are periodic or quasi-periodic.
844: Strictly speaking Lyapunov exponents are defined in the
845: limit $t \to \infty$.
846: Nevertheless one can check the existence of chaos by
847: investigating the behavior of the system for sufficiently
848: large values of $t$.
849: In fact, as we see later, the maximal Lyapunov exponent
850: begins to grow toward a constant value when chaos appears.
851: Although the backreaction effect of created particles can alter the
852: background dynamics, it is possible to see the signature of chaos
853: before the backreaction sets in.
854:
855: In addition to the above two criteria, there exists another criterion
856: for the existence of chaos-- which is so called
857: a fractal map \cite{Barrow}.
858: This strategy is useful because of gauge independence which
859: comes from using a topological character.
860: In the next section we shall also use this criterion
861: to confirm the presence of chaos in addition to other methods.
862:
863:
864: %%%%%%%%%%%%%%%%%%%%%%%%%%
865: \section{Chaotic dynamics for the quartic potential}
866: %%%%%%%%%%%%%%%%%%%%%%%%%%
867:
868: For the quartic potential ($n=4$) the system is effectively
869: reduced to a Hamiltonian system in Minkowski spacetime
870: by introducing conformal variables: $\tilde{\phi}\equiv a\phi$
871: $\tilde{\chi}\equiv a\chi$ and $\eta \equiv \int a^{-1} dt$.
872: Using the fact that $a \propto \eta$ and $a'/a,\ a''/a \to 0$
873: during reheating in this model, the background equations
874: (\ref{back2}), (\ref{back3}) and (\ref{back}) can be written as
875: %
876: \begin{eqnarray}
877: & & \tilde{\phi}''+\lambda\tilde{\phi}^3
878: +g^2\tilde{\chi}^2\tilde{\phi}=0\,, \\
879: & & \tilde{\chi}''+g^2\tilde{\phi}^2\tilde{\chi}=0, \\
880: a'{}^2 &=& \frac{8\pi}{3m_p^2}\left(
881: \frac{1}{2} \tilde{\phi}'{}^2
882: +\frac{1}{2} \tilde{\chi}'{}^2
883: +\frac{\lambda}{4} \tilde{\phi}^4
884: +\frac{g^2}{2}\tilde{\phi}^2\tilde{\chi}^2\right) \nonumber \\
885: &\equiv& \frac{8\pi}{3m_p^2}E={\rm const}\,,
886: \end{eqnarray}
887: %
888: where a prime denotes the derivative with respect to conformal
889: time $\eta$. The trajectories of the fields are bounded by
890: %
891: \begin{eqnarray}
892: \label{bound}
893: \frac{\lambda}{4} \tilde{\phi}^4
894: +\frac{g^2}{2}\tilde{\phi}^2\tilde{\chi}^2
895: \le E\,.
896: \end{eqnarray}
897: %
898: This boundary is plotted as a dotted curve in Fig.~\ref{pspace}.
899:
900: %%%%%%%%%%
901: \begin{figure}
902: \begin{center}
903: \includegraphics[height=3.3in,width=3.3in]{fig3.eps}
904: \caption{A phase-space trajectory in the ($\phi$, $\chi$) plane for
905: $g^2/\lambda=2$. The dotted curve corresponds to the boundary
906: given in Eq.~(\ref{bound}).
907: }
908: \label{pspace}
909: \end{center}
910: \end{figure}
911: %%%%%%%%%%
912:
913: {}From the Toda-Brumer test (\ref{chin=4}),
914: we can expect that chaos occurs when $\chi$ becomes
915: comparable to $\phi$ for $g^2/\lambda={\cal O}(1)$.
916: When $1<g^2/\lambda<3$, corresponding to the first
917: resonance band \cite{Kaiser,GKLS},
918: the variance of the field $\chi$ is larger than
919: $\chi_{f} \sim 10^{-8} m_p$ right after the end
920: of inflation from Fig.~\ref{n4}.
921: We require the parametric excitation of $\chi$
922: to give rise to chaos, since
923: $\chi$ is much smaller than $\phi$ at the beginning of reheating.
924: Therefore the coupling $g$ needs to lie inside of
925: the resonance band for the existence of chaos.
926: In Fig.~\ref{pspace} we show an example of
927: the background trajectory for $g^2/\lambda=2$.
928: We find that two fields evolve chaotically in the phase-space of
929: the ($\phi, \chi$) plane by choosing several different initial
930: conditions.
931:
932: %%%%%%%%%%
933: \begin{figure}
934: \begin{center}
935: \includegraphics[height=3.0in,width=3.3in]{fig4.eps}
936: \caption{Evolution of $\tilde{\chi}$, $\langle \delta \tilde{\chi}^2
937: \rangle^{1/2}$ and $\langle \delta \tilde{\phi}^2\rangle^{1/2}$
938: (normalized by $m_p$) as a function of $x \equiv
939: \sqrt{\lambda} \phi_I \eta$ for $g^2/\lambda=2$
940: when the backreaction effect of created particles is
941: neglected. The horizontal line shows the border of the Toda-Brumer
942: test.
943: The perturbations in $\phi$ and $\chi$ exhibit instabilities
944: associated with chaos once the field $\chi$
945: is sufficiently amplified.
946: }
947: \label{chidelchi}
948: \end{center}
949: \end{figure}
950: %%%%%%%%%%
951:
952: When $g^2/\lambda=2$ the Toda-Brumer test gives the condition
953: $\tilde{\chi}/m_p \gtrsim 0.5$ for the existence of chaos, which
954: corresponds to the time $x \equiv \sqrt{\lambda}\phi_I\eta \gtrsim 45$
955: (see Fig.~\ref{chidelchi}).
956: We find that the fluctuations $\langle \delta \tilde{\chi}^2 \rangle$
957: and $\langle \delta \tilde{\phi}^2 \rangle$ exhibit rapid increase
958: with a similar growth rate after the field $\chi$ satisfies
959: the Toda-Brumer test.
960: It is expected that this is associated with the presence of chaos
961: rather than parametric excitation of the $\chi$ fluctuation.
962: The quasi-homogeneous field $\chi$ is amplified by parametric
963: resonance
964: for $x \lesssim 45$ but stops growing after that.
965: This comes from the fact that the resonance does not occur
966: once the homogeneous oscillation of $\phi$ is broken by the growth
967: of $\chi$. The $g^2\phi^2$ term on the
968: l.h.s. of Eq.~(\ref{dchi}) also becomes ineffective at this stage, but
969: the presence of the mixing term on the r.h.s. leads to
970: a new type of instability associated with chaos.
971: The rapid growth of $\langle \delta \tilde{\phi}^2 \rangle$
972: seen in Fig.~\ref{chidelchi} also comes from
973: the mixing term on the r.h.s. of Eq.~(\ref{dphi})
974: rather than from the parametric excitation of sub-Hubble modes with
975: $3/2<k^2/(\lambda \phi_I^2)<\sqrt{3}$ \cite{Kaiser,GKLS}.
976:
977: %%%%%%%%%%
978: \begin{figure}
979: \begin{center}
980: \includegraphics[height=3.3in,width=3.5in]{fig5.eps}
981: \caption{Evolution of the maximal Lyapunov exponent $h$
982: for $g^2/\lambda=2, 8, 18$. We do not implement the backreaction
983: effect of created particles. We find that the exponent $h$
984: approaches a constant value in these cases.
985: }
986: \label{lyapunov}
987: \end{center}
988: \end{figure}
989: %%%%%%%%%%
990:
991: In order to check the existence of chaos, we plot
992: the evolution of the maximal Lyapunov exponent $h$
993: for several different values of $g$ in Fig.~\ref{lyapunov}.
994: The exponent decreases at the initial stage
995: even though $\chi$ is amplified by parametric resonance.
996: However $h$ begins to grow after $\chi$
997: satisfies the Toda-Brumer test (\ref{chin=4}).
998: The system eventually approaches a phase with
999: a positive constant $h$, which shows the existence of chaos.
1000: The growth of the maximal Lyapunov exponent
1001: is regarded as a signature of chaos,
1002: since this behavior can not be seen in the absence of chaos.
1003: We checked that $h$ continues to decrease and converges
1004: toward 0 in power if the mixing terms do not exist on the r.h.s. of
1005: Eqs.~(\ref{dphi}) and (\ref{dchi}).
1006:
1007: In Fig.~\ref{fractal} we show a fractal map for $g^2/\lambda=2$
1008: with slight change of initial conditions in terms of $\tilde{\chi}_I$
1009: and $\tilde{\dot{\chi}}_I$. We set exit pockets when the field $\chi$
1010: becomes larger than $|\tilde{\chi}|=1.2$.
1011: When an orbit reaches a pocket we assign colors to many initial
1012: conditions
1013: as in the following way;
1014: white if an orbit falls down to an upper pocket ($\tilde{\chi}>1.2$),
1015: black if it falls down to a lower pocket ($\tilde{\chi}<-1.2$).
1016: Figure \ref{fractal} is the result of the above manipulation,
1017: which shows that the map of initial conditions is fractal.
1018: This means that orbits are sensitive to initial data,
1019: thereby showing the existence of chaos.
1020:
1021: %%%%%%%%%%
1022: \begin{figure}
1023: \begin{center}
1024: \includegraphics[height=6.0in,width=3.2in]{fig6.eps}
1025: \caption{A fractal map for $g^2/\lambda=2$.
1026: We show the map of initial conditions $\chi_I$ and
1027: $\dot{\chi}_I$ with exit pockets characterized by
1028: $|\tilde{\chi}| \ge 1.2$. The white color corresponds to
1029: orbits which give $\tilde{\chi} \ge 1.2$, whereas
1030: the black one to orbits which give $\tilde{\chi} \le -1.2$.
1031: The upper panel corresponds to the change of initial conditions
1032: by 0.1\%. The lower panel is an extended figure,
1033: in which initial conditions change by 0.005\%.
1034: These figures exhibit fractal structures, which result
1035: from sensitivity to initial conditions.
1036: }
1037: \label{fractal}
1038: \end{center}
1039: \end{figure}
1040: %%%%%%%%%%
1041:
1042: The above discussion neglects the backreaction effect of
1043: created particles. If we account for it as
1044: a Hartree approximation, Eqs.~(\ref{back2}), (\ref{back3}),
1045: (\ref{back}), (\ref{dphi}) and (\ref{dchi}) are modified by replacing
1046: the $\phi^2$, $\chi^2$ and $\phi^3$ terms for
1047: $\phi^2+\langle \delta \phi^2 \rangle$, $\chi^2+\langle \delta \chi^2
1048: \rangle$
1049: and $\phi^3+3\phi \langle \delta \phi^2 \rangle$, respectively
1050: \cite{KLS2}.
1051: By Eq.~(\ref{dphi}) the backreaction becomes important when
1052: %
1053: \begin{eqnarray}
1054: \label{backcon}
1055: \sqrt{\langle \delta \tilde{\chi}^2 \rangle} \gtrsim
1056: \sqrt{\frac{3\lambda}{g^2}}\,\tilde{\phi}\,.
1057: \end{eqnarray}
1058: %
1059: This is similar to the necessary condition for chaos for
1060: the background field $\chi$, see Eq.~(\ref{chin=4}).
1061: When $g^2/\lambda={\cal O}(1)$, the Toda-Brumer test
1062: is satisfied before the condition (\ref{backcon}) is fulfilled.
1063: As seen in Fig.~\ref{chidelchi} both $\tilde{\chi}^2$ and
1064: $\langle \delta \tilde{\chi}^2 \rangle$
1065: have similar amplitudes initially, but the growth of sub-Hubble
1066: fluctuations
1067: occurs later than that of $\tilde{\chi}^2$.
1068: At the time when the Toda-Brumer test (\ref{chin=4})
1069: is satisfied ($x \simeq 45$), $\langle \delta \tilde{\chi}^2 \rangle$
1070: is much smaller than $\tilde{\chi}^2$ for $g^2/\lambda=2$.
1071: The backreaction effect becomes important around $x=60$
1072: in this case.
1073:
1074: By implementing the backreaction as a Hartree approximation,
1075: we find that this typically tends to work to suppress the growth
1076: of field fluctuations.
1077: As illustrated in Fig.~\ref{chidelchiback} the fluctuations do not
1078: exhibit rapid
1079: growth after the backreaction begins to work ($x \gtrsim 60$).
1080: Nevertheless we need to caution that linear perturbation theory
1081: is no longer valid at this stage. For completeness it is required
1082: to account for the mode-mode coupling (rescattering)
1083: between the fluctuations \cite{KT}. In fact we found a
1084: numerical instability for $x \gtrsim 80$ in the simulation of
1085: Fig.~\ref{chidelchiback}, which signals the limitation of
1086: the Hartree approximation.
1087: It is of interest to see the effect of chaos at this fully nonlinear
1088: stage,
1089: but this is a non-trivial problem because of the complex nature
1090: of reheating. Note that the chaotic period ends at some time
1091: to complete reheating.
1092: It is difficult to judge when chaos ends in our system,
1093: since the mechanism for the decay of $\phi$ and $\chi$
1094: after preheating is not completely known.
1095:
1096: %%%%%%%%%%
1097: \begin{figure}
1098: \begin{center}
1099: \includegraphics[height=3.0in,width=3.2in]{fig7.eps}
1100: \caption{Evolution of $\tilde{\chi}$, $\langle \delta
1101: \tilde{\chi}^2\rangle^{1/2}$ and $\langle \delta
1102: \tilde{\phi}^2\rangle^{1/2}$
1103: (normalized by $m_p$) for $g^2/\lambda=2$
1104: when the backreaction effect is taken into account.
1105: The horizontal line corresponds to the Toda-Brumer border.
1106: }
1107: \label{chidelchiback}
1108: \end{center}
1109: \end{figure}
1110: %%%%%%%%%%
1111:
1112:
1113: When $g^2/\lambda={\cal O}(1)$ one can find out the existence of chaos
1114: during a short period before the backreaction begins to work.
1115: As we already mentioned, the criterion for chaos is
1116: given by Eq.~(\ref{chin=4}), whereas the criterion for the
1117: backreaction corresponds to Eq.~(\ref{backcon}).
1118: The initial value of the quasi-homogeneous field $\chi$
1119: gets smaller for larger $g^2/\lambda$, as illustrated in
1120: Fig.~\ref{n4}.
1121: This means that the condition (\ref{backcon}) tends to be satisfied
1122: prior to the time at which the quasi-homogeneous field $\chi$
1123: grows to satisfy the Toda-Brumer test (\ref{chin=4}).
1124: In Fig.~\ref{g5000} we plot the evolution of the system for
1125: $g^2/\lambda=5000$ with the backreaction effect of
1126: created particles.
1127: In this case the backreaction becomes important around $x=50$
1128: before the quasi-homogeneous field $\chi$ increases sufficiently
1129: to satisfy the Toda-Brumer test.
1130: Therefore, when $g^2/\lambda \gg 1$, we do not find a signature
1131: of chaos before the perturbations reach a nonlinear regime.
1132:
1133: %%%%%%%%%%
1134: \begin{figure}
1135: \begin{center}
1136: \includegraphics[height=3.0in,width=3.2in]{fig8.eps}
1137: \caption{As in Fig.~\ref{chidelchiback} for $g^2/\lambda=5000$.
1138: }
1139: \label{g5000}
1140: \end{center}
1141: \end{figure}
1142: %%%%%%%%%%
1143:
1144: For the coupling $g$ that belongs to the resonance bands
1145: $n(2n-1)<g^2/\lambda<n(2n+1)$, we find that
1146: chaos occurs for
1147: %
1148: \begin{eqnarray}
1149: g^2/\lambda < {\cal O}(10)\,,
1150: \end{eqnarray}
1151: %
1152: before the backreaction begins to work.
1153: We note that this result is obtained by using the value
1154: $\sqrt{\langle \delta \chi^2_{k} (t_{f}) \rangle_{k<k_f}}$ derived
1155: in Sec.~II as the initial condition of $\chi$ at the beginning of
1156: preheating.
1157: When $g^2/\lambda > {\cal O}(10)$ the field fluctuation
1158: $\langle \delta \tilde{\chi}^2 \rangle$ satisfies the condition
1159: (\ref{backcon}) prior to the time at which the necessary condition
1160: for chaos is fulfilled.
1161: Provided that $g^2/\lambda \gg 1$, the standard
1162: Floquet theory of parametric resonance \cite{Kaiser,GKLS} is valid
1163: at the linear level, since the effect of the quasi-homogeneous field
1164: $\chi$
1165: is not important relative to its perturbations on sub-Hubble scales.
1166:
1167:
1168: %%%%%%%%%%%%%%%%%%%%%%%%%%
1169: \section{Quadratic potential}
1170: %%%%%%%%%%%%%%%%%%%%%%%%%%
1171:
1172: For the quadratic potential the system can not be
1173: reduced to the analysis in Minkowski spacetime
1174: by introducing conformal quantities.
1175: Therefore the analysis in the quadratic potential is more complicated
1176: than
1177: in the case of the quartic potential
1178: in an expanding background.
1179:
1180: We start our analysis by studying the two-field dynamics in a
1181: frictionless background.
1182: In this case the fields oscillate coherently without
1183: an adiabatic damping due to cosmic expansion.
1184: The system has the field equations
1185: corresponding to $H=0$ in Eqs.~(\ref{back2}) and (\ref{back3})
1186: together with the constraint
1187: %
1188: \begin{eqnarray}
1189: \label{energy}
1190: E=\frac12 \dot{\phi}^2+\frac12 \dot{\chi}^2+
1191: V(\phi, \chi)\,,
1192: \end{eqnarray}
1193: %
1194: where $E$ is conserved.
1195: Unlike the case of an expanding background, the field $\chi$ is
1196: enhanced only when the system is inside of resonance bands
1197: from the beginning of preheating.
1198: This comes from the fact that the field $\chi$ does not shift to other
1199: stability/instability bands in the absence of cosmic
1200: expansion \cite{KLS2}. We wish to study the existence of chaos
1201: for the coupling $g$ that leads to parametric excitation of
1202: $\chi$ in an expanding background ($g \gtrsim 3.0 \times 10^{-4}$).
1203: First we carry out the analysis in a conserved Hamiltonian system
1204: given above and then proceed to the case in which the expansion
1205: of universe is taken into account.
1206:
1207: As illustrated in Fig.~\ref{n2}, the quasi-homogeneous $\chi$
1208: field at the end of inflation is estimated to be
1209: $\chi_f \lesssim 10^{-9}$-$10^{-8}m_{p}$ for the coupling
1210: $g \gtrsim 3.0 \times 10^{-4}$, which is smaller than that
1211: in the quartic potential with $g^2/\lambda={\cal O}(1)$.
1212: Hence the field $\chi$ for the quadratic potential
1213: is more strongly suppressed during inflation
1214: for the values of $g$ relevant to efficient preheating.
1215:
1216: In Fig.~\ref{qua1} we plot the evolution of the background field
1217: $\chi$
1218: together with $\sqrt{\langle \delta \chi ^2\rangle}$ and
1219: $\sqrt{\langle \delta \phi^2\rangle}$ for the coupling $g=3.0\times
1220: 10^{-4}$.
1221: Note that we implement the backreaction of sub-Hubble
1222: field fluctuations as a Hartree approximation.
1223: In this case the Toda-Brumer test gives the condition
1224: $\chi/m_p>1.9\times 10^{-3}$ by Eq.~(\ref{chin=2}).
1225: For the quadratic potential the backreaction begins to work for
1226: $\langle \delta \chi^2 \rangle \gtrsim m^2/g^2$,
1227: which is basically a similar condition to the Toda-Brumer test
1228: for the background field $\chi$.
1229: As shown in Fig.~\ref{qua1} the quasi-homogeneous field $\chi$
1230: does not satisfy the Toda-Brumer test, since the backreaction
1231: becomes important before $\chi$ grows sufficiently.
1232:
1233: %%%%%%%%%%
1234: \begin{figure}
1235: \begin{center}
1236: \includegraphics[height=3.0in,width=3.2in]{fig9.eps}
1237: \caption{Evolution of $\chi$, $\langle \delta \chi ^2\rangle^{1/2}$
1238: and
1239: $\langle \delta \phi^2\rangle^{1/2}$ (normalized by $m_p$)
1240: for $g=3.0\times 10^{-4}$
1241: without the friction due to cosmic expansion.
1242: The horizontal line represents the border of the Toda-Brumer test.
1243: We implement the backreaction effect of created particles
1244: as a Hartree approximation.
1245: }
1246: \label{qua1}
1247: \end{center}
1248: \end{figure}
1249: %%%%%%%%%%
1250:
1251: The quasi-homogeneous $\chi$ field at the end of inflation gets
1252: smaller for larger $g$, see Eq.~(\ref{dchik}).
1253: Figure \ref{qua2} shows the evolution of the system
1254: for $g=3.0 \times 10^{-3}$, in which case the initial variance of the
1255: field $\chi$ is suppressed relative to the case $g=3.0 \times
1256: 10^{-4}$.
1257: Although the condition for chaos using
1258: the Toda-Brumer test (\ref{chin=2}) gets milder for larger $g$,
1259: this property is compensated by
1260: the suppression of the quasi-homogeneous field $\chi$
1261: at the beginning of preheating.
1262: Therefore it is difficult to satisfy the necessary condition for chaos
1263: before the backreaction begins to work.
1264: We carried out numerical simulations for other values of
1265: $g$ and found that the signature of chaos is not seen
1266: in the frictionless system as long as the backreaction effect
1267: is taken into account.
1268:
1269: %%%%%%%%%%
1270: \begin{figure}
1271: \begin{center}
1272: \includegraphics[height=3.0in,width=3.2in]{fig10.eps}
1273: \caption{As in Fig.~\ref{qua1} with $g=3.0 \times 10^{-3}$.
1274: }
1275: \label{qua2}
1276: \end{center}
1277: \end{figure}
1278: %%%%%%%%%%
1279:
1280: If we implement the effect of cosmic expansion, the energy of the
1281: system given by Eq.~(\ref{energy}) decreases.
1282: In fact the time-derivative of $E$ is given by
1283: %
1284: \begin{eqnarray}
1285: \frac{{\rm d}E}{{\rm d}t}=-3H(\dot{\phi}^2+\dot{\chi}^2)
1286: \simeq -3HE\,,
1287: \end{eqnarray}
1288: %
1289: where we used the approximation $E \simeq \dot{\phi}^2+\dot{\chi}^2$.
1290: Then the energy lost during one oscillation of
1291: the inflaton ($\Delta t \simeq 1/m$) is estimated as
1292: %
1293: \begin{eqnarray}
1294: \frac{\Delta E}{E} \simeq -3\frac{H}{m} \simeq
1295: -\sqrt{24\pi} \frac{\phi}{m_p}\,.
1296: \end{eqnarray}
1297: %
1298: Therefore the energy loss is large at the beginning of
1299: preheating ($|\Delta E/E| \gtrsim 0.1$),
1300: but it becomes smaller and smaller with time.
1301: The analysis in the frictionless system can be used in an expanding
1302: background when the condition, $|\Delta E/E| \ll 1$, is satisfied.
1303:
1304: In addition to the energy loss, we need to
1305: caution that the structure of
1306: resonance changes in the presence of cosmic expansion.
1307: The field $\chi$ passes many instability/stability bands, which is
1308: generally called {\it stochastic resonance} \cite{KLS2}.
1309: Parametric resonance ends when the resonance parameter,
1310: $q=g^2\phi^2/(4m^2)$, drops down to less than of order unity,
1311: whose property is different from the analysis in Minkowski spacetime.
1312:
1313: In spite of above complexities, it is possible to check whether
1314: the signature of chaos is seen or not in a linear perturbation regime.
1315: Note that the Toda-Brumer condition is valid in an expanding
1316: background. We run our numerical code including
1317: the backreaction of sub-Hubble field fluctuations for the coupling
1318: $g$ relevant to efficient preheating ($g \gtrsim 3.0 \times 10^{-4}$)
1319: and find that the background $\chi$ stops growing by the backreaction
1320: effect before the Toda-Brumer condition is satisfied.
1321: This property is similar to the case discussed in the frictionless
1322: background. Therefore chaos does not appear for the quadratic
1323: potential at least in a regime before the backreaction sets in.
1324:
1325:
1326: %%%%%%%%%%%%%%%%%%%%%%%%%%
1327: \section{Conclusions}
1328: %%%%%%%%%%%%%%%%%%%%%%%%%%
1329:
1330: In this paper we discussed chaotic dynamics in two-field preheating
1331: with monomial inflaton potentials $V(\phi)=V_0\phi^n$.
1332: A scalar field $\chi$ coupled to inflaton with an interaction
1333: $(1/2)g^2\phi^2\chi^2$ is amplified by parametric resonance
1334: when the inflaton oscillates coherently.
1335: As long as the background field $\chi$ is not
1336: strongly damped in an inflationary epoch, $\chi$ can grow to
1337: the same order as $\phi$, thereby giving rise to a possibility
1338: of chaos during preheating.
1339:
1340: First we estimated the amplitude of the quasi-homogeneous
1341: $\chi$ field at the beginning of reheating by considering
1342: the variance of the $\chi$ fluctuations on scales larger than
1343: the Hubble radius at the end of inflation.
1344: The variance $\sqrt{\langle \delta \chi^2_{k}
1345: (t_{f}) \rangle_{k<k_f}}$ is plotted as a function of the coupling
1346: $g$ in Figs.~\ref{n2} and \ref{n4}.
1347: For the quartic inflaton potential with $g^2/\lambda={\cal O}(1)$
1348: there exist large-scale perturbation modes ($k<k_{c}$)
1349: which are not strongly suppressed during inflation.
1350: Then this gives large contribution to the total variance of
1351: $\chi$ relative to the case $g^2/\lambda \gg 1$ at the end of
1352: inflation.
1353:
1354: Typically the background field $\chi$ is assumed to be
1355: negligible in standard analysis of particle creations in
1356: preheating, but its presence leads to a mixing between two fields.
1357: This gives rise to a chaotic instability in addition to the
1358: enhancement of perturbations by parametric resonance. We note that
1359: standard Floquet theory using Mathieu or Lame equation ceases to
1360: be valid when the chaos is present. This new channel of
1361: instability can alter the maximum size of field fluctuations if
1362: two dynamical fields do not decay for a long time.
1363:
1364: In order to study whether chaos really appears or not, we solved the
1365: background equations (\ref{back2}), (\ref{back3}) and (\ref{back})
1366: together with perturbed equations (\ref{dphi}) and (\ref{dchi}).
1367: For a quartic potential ($n=4$ and $V_0=\lambda/4$)
1368: parametric resonance can occur for the coupling $g^2/\lambda$ of order
1369: unity. In this case the quasi-homogeneous field $\chi$ satisfies the
1370: Toda-Brumer test (\ref{chin=4})
1371: before the backreaction effect of created particles
1372: becomes important. Since this is only the necessary condition for
1373: chaos, we also
1374: evaluated Lyapunov exponents for $g^2/\lambda={\cal O}(1)$ in order
1375: to confirm the
1376: presence of chaos. We find that the maximal Lyapunov exponent begins
1377: to increase
1378: toward a positive constant value after the Toda-Brumer condition is
1379: satisfied,
1380: which shows the existence of chaos.
1381: Our analysis using a Fractal map also implies that the system exhibits
1382: chaotic behavior when $g^2/\lambda$ is not too much larger than unity.
1383: For larger $g^2/\lambda$, the backreaction of field fluctuations on
1384: sub-Hubble
1385: scales works earlier than the time at which the Toda-Brumer test is
1386: satisfied for the background field $\chi$. We find signatures of chaos
1387: for $g^2/\lambda <{\cal O} (10)$ at the linear regime before the
1388: backreaction
1389: begins to work. For $g^2/\lambda \gg 1$ the quasi-homogeneous field
1390: $\chi$ gets smaller
1391: at the beginning of preheating, in which case perturbations
1392: enter a nonlinear region before chaos can be seen.
1393:
1394: The system with a quadratic ($n=2$) potential can not
1395: be effectively reduced to the analysis in Minkowski spacetime
1396: unlike the quartic potential.
1397: We first analyze the preheating dynamics in a frictionless
1398: background and then proceed to the case in which the
1399: expansion of universe is taken into account.
1400: In this model the background field $\chi$
1401: does not grow sufficiently to satisfy the Toda-Brumer test
1402: in the presence of the backreaction effect of created particles.
1403: We find that this result holds both in Minkowski and expanding
1404: backgrounds. Therefore chaos can not be observed at least
1405: at the linear stage of preheating before the backreaction sets in.
1406:
1407: When chaos is present, this means that the second field $\chi$
1408: gives a non-negligible contribution for the background dynamics.
1409: This generally gives a strong correlation between adiabatic and
1410: isocurvature metric perturbations \cite{corre}, which can lead to
1411: the amplification of curvature perturbations for the self-coupling
1412: inflation model \cite{zetagrow}. This reduces the tensor to scalar
1413: ratio $r$, which is favorable from the observational point of
1414: view. Of course the ratio $r$ is not sufficient to judge whether
1415: the model is rescued or not, since large contribution of
1416: isocurvature perturbations modifies the CMB power spectrum. It is
1417: still interesting that the appearance of chaos has a possibility
1418: to reduce the ratio $r$.
1419:
1420: In this work we did not study the nonlinear dynamics of the system
1421: during which the mode-mode coupling between perturbations plays an
1422: important role. We expect that the presence of chaos persists even
1423: in such a nonlinear stage provided that two interacting fields
1424: $\phi$ and $\chi$ are dynamically important. The chaos would
1425: finally disappear after the energy densities of scalar fields are
1426: converted to that of radiation. It is of interest to extend our
1427: analysis to such a regime including Born decays of scalar fields
1428: for a complete understanding of chaos in reheating.
1429:
1430:
1431: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1432: \section*{ACKNOWLEDGMENTS}
1433: It is a pleasure to thank Kei-ichi Maeda and Alexei Starobinsky for
1434: useful discussions.
1435: This work was partially supported by a Grant for The 21st Century COE
1436: Program
1437: (Holistic Research and Education Center for Physics Self-organization
1438: Systems)
1439: at Waseda University.
1440: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1441:
1442:
1443:
1444:
1445: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1446:
1447:
1448:
1449:
1450:
1451: \begin{thebibliography}{40}
1452:
1453: \bibitem{oldre}
1454: A.~Dolgov and A.~Linde,
1455: %''Baryon aymmetry in the inflationary universe,''
1456: Phys. Lett. {\bf 116B}, 329 (1982);
1457: L.~Abbott, E.~Farhi and M.~Wise,
1458: %''Particle production in the new inflationary cosmology,''
1459: Phys. Lett. {\bf 117B}, 29 (1982).
1460:
1461: \bibitem{TB}
1462: J.~H.~Traschen and R.~H.~Brandenberger,
1463: Phys. Rev. D {\bf 42}, 2491 (1990);
1464: Y.~Shtanov, J.~H.~Trashen and R.~H.~Brandenberger,
1465: Phys. Rev. D {\bf 51}, 5438 (1995).
1466:
1467: \bibitem{KLS1}
1468: L.~Kofman, A.~D.~Linde and A.~A.~Starobinsky,
1469: %``Reheating after inflation,''
1470: Phys.\ Rev.\ Lett.\ {\bf 73}, 3195 (1994)
1471: [arXiv:hep-th/9405187].
1472:
1473: \bibitem{Boya}
1474: D.~Boyanovsky, H.~J.~de Vega, R.~Holman, D.~S.~Lee and A.~Singh,
1475: %``Dissipation via particle production in scalar field theories,''
1476: Phys.\ Rev.\ D {\bf 51}, 4419 (1995)
1477: [arXiv:hep-ph/9408214].
1478:
1479: \bibitem{GUT}
1480: E.~W.~Kolb, A.~D.~Linde and A.~Riotto,
1481: %``GUT baryogenesis after preheating,''
1482: Phys.\ Rev.\ Lett.\ {\bf 77}, 4290 (1996)
1483: [arXiv:hep-ph/9606260];
1484: E.~W.~Kolb, A.~Riotto and I.~I.~Tkachev,
1485: %``GUT baryogenesis after preheating: Numerical study
1486: %of the production and decay of X-bosons,''
1487: Phys.\ Lett.\ B {\bf 423}, 348 (1998)
1488: [arXiv:hep-ph/9801306].
1489:
1490: \bibitem{nonther}
1491: L.~Kofman, A.~D.~Linde and A.~A.~Starobinsky,
1492: %``Non-Thermal Phase Transitions After Inflation,''
1493: Phys.\ Rev.\ Lett.\ {\bf 76}, 1011 (1996)
1494: [arXiv:hep-th/9510119];
1495: I.~I.~Tkachev,
1496: %``Phase Transitions at Preheating,''
1497: Phys.\ Lett.\ B {\bf 376}, 35 (1996)
1498: [arXiv:hep-th/9510146];
1499: S.~Khlebnikov, L.~Kofman, A.~D.~Linde and I.~Tkachev,
1500: %``First-order nonthermal phase transition after preheating,''
1501: Phys.\ Rev.\ Lett.\ {\bf 81}, 2012 (1998)
1502: [arXiv:hep-ph/9804425].
1503:
1504: \bibitem{mpre}
1505: H.~Kodama and T.~Hamazaki,
1506: %``Evolution of cosmological perturbations in a stage
1507: %dominated by an oscillatory scalar field,''
1508: Prog.\ Theor.\ Phys.\ {\bf 96}, 949 (1996)
1509: [arXiv:gr-qc/9608022];
1510: A.~Taruya and Y.~Nambu,
1511: %``Cosmological perturbation with two scalar fields
1512: %in reheating after inflation,''
1513: Phys.\ Lett.\ B {\bf 428}, 37 (1998)
1514: [arXiv:gr-qc/9709035];
1515: B.~A.~Bassett, D.~I.~Kaiser and R.~Maartens,
1516: %``General relativistic preheating after inflation,''
1517: Phys.\ Lett.\ B {\bf 455}, 84 (1999)
1518: [arXiv:hep-ph/9808404];
1519: M.~Parry and R.~Easther,
1520: %``Preheating and the Einstein field equations,''
1521: Phys.\ Rev.\ D {\bf 59}, 061301 (1999)
1522: [arXiv:hep-ph/9809574];
1523: F.~Finelli and R.~H.~Brandenberger,
1524: %``Parametric amplification of gravitational fluctuations
1525: % during reheating,''
1526: Phys.\ Rev.\ Lett.\ {\bf 82}, 1362 (1999)
1527: [arXiv:hep-ph/9809490];
1528: B.~A.~Bassett, F.~Tamburini, D.~I.~Kaiser and R.~Maartens,
1529: %``Metric preheating and limitations of linearized gravity. II,''
1530: Nucl.\ Phys.\ B {\bf 561}, 188 (1999)
1531: [arXiv:hep-ph/9901319];
1532: S.~Tsujikawa and B.~A.~Bassett,
1533: %``When can preheating affect the CMB?,''
1534: Phys.\ Lett.\ B {\bf 536}, 9 (2002)
1535: [arXiv:astro-ph/0204031].
1536:
1537: \bibitem{PBH}
1538: A.~M.~Green and K.~A.~Malik,
1539: Phys.\ Rev.\ D {\bf 64}, 021301 (2001)
1540: [arXiv:hep-ph/0008113];
1541: B.~A.~Bassett and S.~Tsujikawa,
1542: Phys.\ Rev.\ D {\bf 63}, 123503 (2001)
1543: [arXiv:hep-ph/0008328];
1544: F.~Finelli and S.~Khlebnikov,
1545: %``Large metric perturbations from rescattering,''
1546: Phys.\ Lett.\ B {\bf 504}, 309 (2001)
1547: [arXiv:hep-ph/0009093];
1548: F.~Finelli and S.~Khlebnikov,
1549: %``Metric perturbations at reheating:
1550: %The use of spherical symmetry,''
1551: Phys.\ Rev.\ D {\bf 65}, 043505 (2002)
1552: [arXiv:hep-ph/0107143];
1553: T.~Suyama, T.~Tanaka, B.~Bassett and H.~Kudoh,
1554: %``Are black holes over-produced during preheating?,''
1555: Phys.\ Rev.\ D {\bf 71}, 063507 (2005)
1556: [arXiv:hep-ph/0410247].
1557:
1558: \bibitem{KT}
1559: S.~Y.~Khlebnikov and I.~I.~Tkachev,
1560: %``Classical decay of inflaton,''
1561: Phys.\ Rev.\ Lett.\ {\bf 77}, 219 (1996)
1562: [arXiv:hep-ph/9603378];
1563: S.~Y.~Khlebnikov and I.~I.~Tkachev,
1564: %``Resonant decay of Bose condensates,''
1565: Phys.\ Rev.\ Lett.\ {\bf 79}, 1607 (1997)
1566: [arXiv:hep-ph/9610477].
1567:
1568: \bibitem{KLS2}
1569: L.~Kofman, A.~D.~Linde and A.~A.~Starobinsky,
1570: %``Towards the theory of reheating after inflation,''
1571: Phys.\ Rev.\ D {\bf 56}, 3258 (1997)
1572: [arXiv:hep-ph/9704452].
1573:
1574: \bibitem{Iva}
1575: P.~Ivanov,
1576: %``On generation of metric perturbations during preheating,''
1577: Phys.\ Rev.\ D {\bf 61}, 023505 (2000)
1578: [arXiv:astro-ph/9906415];
1579: K.~Jedamzik and G.~Sigl,
1580: %``On metric preheating,''
1581: Phys.\ Rev.\ D {\bf 61}, 023519 (2000)
1582: [arXiv:hep-ph/9906287];
1583: S.~Tsujikawa,
1584: %``Preheating with extra dimensions,''
1585: JHEP {\bf 0007}, 024 (2000)
1586: [arXiv:hep-ph/0005105];
1587: B.~A.~Bassett, M.~Peloso, L.~Sorbo and S.~Tsujikawa,
1588: %``Fermion production from preheating-amplified
1589: % metric perturbations,''
1590: Nucl.\ Phys.\ B {\bf 622}, 393 (2002)
1591: [arXiv:hep-ph/0109176].
1592:
1593: \bibitem{Bruce}
1594: B.~A.~Bassett and F.~Viniegra,
1595: %``Massless metric preheating,''
1596: Phys.\ Rev.\ D {\bf 62}, 043507 (2000)
1597: [arXiv:hep-ph/9909353].
1598:
1599: \bibitem{Fabio}
1600: F.~Finelli and R.~H.~Brandenberger,
1601: %``Parametric amplification of metric fluctuations
1602: %during reheating in two field models,''
1603: Phys.\ Rev.\ D {\bf 62}, 083502 (2000).
1604:
1605: \bibitem{Podolsky}
1606: D.~I.~Podolsky and A.~A.~Starobinsky,
1607: %``Chaotic reheating,''
1608: Grav.\ Cosmol.\ Suppl.\ {\bf 8N1}, 13 (2002)
1609: [arXiv:astro-ph/0204327].
1610:
1611: \bibitem{Toda}
1612: M.~Toda,
1613: Phys.\ Lett.\ A {\bf 48} (1974) 335.
1614:
1615: \bibitem{Brumer}
1616: P. Brumer,
1617: J.\ Comp.\ Phys. {\bf 14} (1974) 391.
1618:
1619: \bibitem{Lya}
1620: G.~L.~Baker and J.~P.~Gollub,
1621: {\it chaotic dyansmics}: an introduction
1622: (Cambridge University Press, Cambridge, UK, 1996).
1623:
1624: \bibitem{Barrow}
1625: J.~D.~Barrow and J.~Levin,
1626: %``Chaos in the Einstein-Yang-Mills equations,''
1627: Phys.\ Rev.\ Lett.\ {\bf 80}, 656 (1998)
1628: [arXiv:gr-qc/9706065];
1629: J.~D.~Barrow,
1630: %``Chaotic Behavior In General Relativity,''
1631: Phys.\ Rept.\ {\bf 85} (1982) 1.
1632:
1633: \bibitem{EM}
1634: R.~Easther and K.~i.~Maeda,
1635: %``Chaotic dynamics and two-field inflation,''
1636: Class.\ Quant.\ Grav.\ {\bf 16}, 1637 (1999)
1637: [arXiv:gr-qc/9711035].
1638:
1639: \bibitem{Bellido}
1640: J.~Garcia-Bellido and A.~D.~Linde,
1641: %``Preheating in hybrid inflation,''
1642: Phys.\ Rev.\ D {\bf 57}, 6075 (1998)
1643: [arXiv:hep-ph/9711360].
1644:
1645: \bibitem{Bastero}
1646: M.~Bastero-Gil, S.~F.~King and J.~Sanderson,
1647: %``Preheating in supersymmetric hybrid inflation,''
1648: Phys.\ Rev.\ D {\bf 60}, 103517 (1999)
1649: [arXiv:hep-ph/9904315].
1650:
1651: \bibitem{Zibin}
1652: J.~P.~Zibin,
1653: %''Dynamical chaos and the growth of cosmological fluctuations,''
1654: arXiv:hep-ph/0108008.
1655:
1656: \bibitem{BT}
1657: B.~A.~Bassett and F.~Tamburini,
1658: %``Inflationary reheating in grand unified theories,''
1659: Phys.\ Rev.\ Lett.\ {\bf 81}, 2630 (1998)
1660: [arXiv:hep-ph/9804453].
1661:
1662: \bibitem{Cornish}
1663: N.~J.~Cornish and J.~J.~Levin,
1664: Phys. Rev. D {\bf 53}, 3022 (1996).
1665:
1666: \bibitem{Joras}
1667: S.~E.~Joras and V.~H.~Cardenas,
1668: %''Chaos and preheating,''
1669: Phys. Rev. D {\bf 67}, 043501 (2003)
1670: [arXiv:gr-qc/0108088].
1671:
1672: \bibitem{obser}
1673: H.~V.~Peiris {\it et al.},
1674: %``First year Wilkinson Microwave Anisotropy
1675: %Probe (WMAP) observations: Implications for inflation,''
1676: Astrophys.\ J.\ Suppl.\ {\bf 148}, 213 (2003);
1677: V.~Barger, H.~S.~Lee, and D.~Marfatia,
1678: %``WMAP and inflation,''
1679: Phys.\ Lett.\ B{\bf 565}, 33 (2003) [arXiv:hep-ph/0302150];
1680: W.~H.~Kinney, E.~W.~Kolb, A.~Melchiorri and A.~Riotto,
1681: %``WMAPping inflationary physics,''
1682: Phys.\ Rev.\ D {\bf 69}, 103516 (2004)
1683: [arXiv:hep-ph/0305130];
1684: S.~M.~Leach and A.~R.~Liddle, Phys. Rev. D{\bf 68}, 123508 (2003)
1685: [arXiv:astro-ph/0306305];
1686: S.~Tsujikawa and A.~R.~Liddle,
1687: %``Constraints on braneworld inflation from CMB anisotropies,''
1688: JCAP {\bf 0403}, 001 (2004)
1689: [arXiv:astro-ph/0312162];
1690: S.~Tsujikawa and B.~Gumjudpai,
1691: Phys.\ Rev.\ D {\bf 69}, 123523 (2004)
1692: [arXiv:astro-ph/0402185].
1693:
1694: \bibitem{efolds}
1695: A.~R.~Liddle and S.~M.~Leach,
1696: %``How long before the end of inflation were
1697: %observable perturbations produced?,''
1698: Phys.\ Rev.\ D {\bf 68}, 103503 (2003)
1699: [arXiv:astro-ph/0305263].
1700:
1701: \bibitem{Tegmark}
1702: M.~Tegmark {\it et al.} [SDSS Collaboration],
1703: %``Cosmological parameters from SDSS and WMAP,''
1704: Phys.\ Rev.\ D {\bf 69}, 103501 (2004)
1705: [arXiv:astro-ph/0310723].
1706:
1707: \bibitem{nonmini}
1708: B.~A.~Bassett and S.~Liberati,
1709: %``Geometric reheating after inflation,''
1710: Phys.\ Rev.\ D {\bf 58}, 021302 (1998)
1711: [Erratum-ibid.\ D {\bf 60}, 049902 (1999)]
1712: [arXiv:hep-ph/9709417];
1713: S.~Tsujikawa, K.~i.~Maeda and T.~Torii,
1714: %``Resonant particle production with non-minimally
1715: %coupled scalar fields in preheating after inflation,''
1716: Phys.\ Rev.\ D {\bf 60}, 063515 (1999)
1717: [arXiv:hep-ph/9901306];
1718: {\it ibid}, Phys.\ Rev.\ D {\bf 60}, 123505 (1999)
1719: [arXiv:hep-ph/9906501];
1720: {\it ibid}, Phys.\ Rev.\ D {\bf 61}, 103501 (2000)
1721: [arXiv:hep-ph/9910214];
1722: S.~Tsujikawa and H.~Yajima,
1723: Phys.\ Rev.\ D {\bf 62}, 123512 (2000)
1724: [arXiv:hep-ph/0007351].
1725:
1726: \bibitem{LL}
1727: A.~R.~Liddle and D.~H. ~Lyth, {\em Cosmological inflation and
1728: large-scale structure}, Cambridge University Press (2000).
1729:
1730: \bibitem{Kaiser}
1731: D.~I.~Kaiser,
1732: %``Post inflation reheating in an expanding universe,''
1733: Phys.\ Rev.\ D {\bf 53}, 1776 (1996)
1734: [arXiv:astro-ph/9507108].
1735:
1736: \bibitem{GKLS}
1737: P.~B.~Greene, L.~Kofman, A.~D.~Linde and A.~A.~Starobinsky,
1738: %``Structure of resonance in preheating after inflation,''
1739: Phys.\ Rev.\ D {\bf 56}, 6175 (1997)
1740: [arXiv:hep-ph/9705347].
1741:
1742: \bibitem{Zibin2}
1743: J.~P.~Zibin, R.~H.~Brandenberger and D.~Scott,
1744: %``Backreaction and the parametric resonance
1745: %of cosmological fluctuations,''
1746: Phys.\ Rev.\ D {\bf 63}, 043511 (2001)
1747: [arXiv:hep-ph/0007219].
1748:
1749: \bibitem{Sta}
1750: A.~A.~Starobinsky, in Field Theory, Quantum Gravity,
1751: and Strings, proceedings of the Seminar, Meudon and Paris,
1752: France, 1984-1985, edited by H.~T.~de Vega
1753: and N.~Sanchez, Lecture Notes in Physics, Vol. 246
1754: (Springer-Verlag, New York, 1986), p 107.
1755:
1756: \bibitem{NNS}
1757: K.~i.~Nakao, Y.~Nambu and M.~Sasaki,
1758: %``Stochastic Dynamics Of New Inflation,''
1759: Prog.\ Theor.\ Phys.\ {\bf 80}, 1041 (1988).
1760:
1761: \bibitem{Hosoya89}
1762: A.~Hosoya, M.~Morikawa and K.~Nakayama,
1763: %``Stochastic Dynamics Of Scalar Field
1764: %In The Inflationary Universe,''
1765: Int.\ J.\ Mod.\ Phys.\ A {\bf 4}, 2613 (1989).
1766:
1767: \bibitem{Salopek}
1768: D.~S.~Salopek and J.~R.~Bond,
1769: %``Stochastic Inflation And Nonlinear Gravity,''
1770: Phys.\ Rev.\ D {\bf 43}, 1005 (1991).
1771: %%CITATION = PHRVA,D43,1005;%%
1772:
1773:
1774: \bibitem{corre}
1775: D.~Langlois, Phys.\ Rev.\ D {\bf 59}, 123512 (1999)
1776: [arXiv:astro-ph/9906080];
1777: C.~Gordon, D.~Wands, B.~A.~Bassett and R.~Maartens,
1778: Phys.\ Rev.\ D {\bf 63}, 023506 (2001) [arXiv:astro-ph/0009131];
1779: S.~Tsujikawa, D.~Parkinson and B.~A.~Bassett,
1780: Phys.\ Rev.\ D {\bf 67}, 083516 (2003) [arXiv:astro-ph/0210322].
1781:
1782: \bibitem{zetagrow}
1783: J.~P.~Zibin, R.~H.~Brandenberger and D.~Scott,
1784: Phys.\ Rev.\ D {\bf 63}, 043511 (2001) [arXiv:hep-ph/0007219];
1785: S.~Tsujikawa, B.~A.~Bassett and F.~Viniegra,
1786: JHEP {\bf 0008}, 19 (2000) [arXiv:hep-ph/0006354].
1787:
1788: \end{thebibliography}
1789: \end{document}\Documents
1790: