On the limit cycle of an inflationary universe (*)

Transcription

1 IL NUOVO CIMENTO VOL. 112 B, N. 6 Giugno 1997 On the limit cycle of an inflationary universe (*) L. SALASNICH (**) Dipartimento di Matematica Pura ed Applicata, Università di Padova Via Marzolo 8, I Padova, Italy INFN, Sezione di Padova - Via Marzolo 8, I Padova, Italy (ricevuto il 25 Settembre 1996; approvato il 26 Novembre 1996) Summary. We study the dynamics of a scalar inflaton field with a symmetric double-well potential and prove rigorously the existence of a limit cycle in its phase space. By using analytical and numerical arguments we show that the limit cycle is stable and give an analytical formula for its period. PACS Cq Particle-theory and field-theory models of the early Universe (including cosmic pancakes, cosmic strings, chaotic phenomena, inflationary universe, etc.). PACS Lm Nonlinear or nonlocal theories and models. The nonlinear and chaotic behaviour of classical field theories is currently subject of intensive research [1-3] and, in this respect, it is of great interest to investigate the existence and properties of limit cycles, which are inherently nonlinear phenomena [4, 5]. In a previous paper [6] we studied the stability of a scalar inflaton field with a symmetric double-well self-energy. We showed that the value of the inflaton field in the vacuum is a bifurcation parameter which changes the phase space structure and that for some functional solutions of the Hubble constant the system goes to a limit cycle, i.e. to a periodic orbit. In this paper we analyze the properties of this limit cycle by using analytical and numerical arguments. We show that the limit cycle is unique and stable and give an analytical formula for its period. To solve the three major cosmological problems, i.e. the flatness problem, the homogeneity problem, and the formation of structure problem, it is generally postulated that the universe, at a very early stage after the big bang, exhibited a short period of exponential expansion, the so-called inflationary phase [7-10]. All the inflationary models assume the existence of a scalar field f, the so-called inflation field, (*) The author of this paper has agreed to not receive the proofs for correction. (**) salasnichhmath.unipd.it 873

2 874 L. SALASNICH with Lagrangian [9, 11] (1) L4 1 2 m f m f2v(f), where the potential V(f) depends on the type of inflation model considered. Here we choose a real field but also complex scalar can be used [9]. The scalar field, if minimally coupled to gravity, satisfies the equation (2) pf4f O 13u ȧ a v ḟ2 1 a 2 2 f42 V f, where p is the covariant d Alembertian operator and a is the cosmological scale factor. The parameter G is the gravitational constant (G4M p 22 with ˇ4c41 and M p 41.2Q GeV the Plank mass) and H4 ȧoa is the Hubble constant, which in general is a function of time (Hubble function). We suppose that in the universe there is only the inflaton field, so the Hubble function H is related to the energy density of the field by (3) H 2 1 u k a 4 ȧ 2 v 1 y k 2 a a 4 8pG ḟ ( f) 2 2 1V(f)z. Immediately after the onset of inflation, the cosmological scale factor a grows exponentially [9]. Thus the term 2 foa 2 is generally believed to be negligible and, if the inflaton field is sufficiently uniform (i.e. ḟ 2, ( f) 2 bv(f)), we end up with a classical nonlinear scalar field theory in one dimension: (4) f O 13H(f) ḟ1 V f 40, where the Hubble function H satisfies the equation (5) H 2 4 8pG V(f). 3 The potential V(f) depends on the type of inflation model considered and we choose a symmetric double-well potential (6) V(f) 4 l 4 (f 2 2v 2 ) 2, where 6v are the values of the inflaton field in the vacuum, i.e. the points of minimal energy of the system. One of the main difficulties in constructing models with potentials suitable for inflation is that these potentials must be flat enough to allow a sufficiently long period of inflation [8, 9]. In this respect our model is very schematic but it can be seen as a toy model for classical nonlinear dynamics with the attractive feature that it emerges from inflationary cosmology. Obviously, the complete study of the dynamics of the inflaton may be addressed only by a complete quantum field theory approach able to predict not only the behaviour of the classical value of the inflaton field but also the associated

3 ON THE LIMIT CYCLE OF AN INFLATIONARY UNIVERSE 875 quantum fluctuations. However, the quantization of the inflation scenario is still an open problem [10] (an interesting stochastic approach can be found in [12]). In paper [6] we showed that the inflaton field value in the vacuum v is a bifurcation parameter. If v40 in the phase space there is only one stable fixed point (f40, ḟ40), which is an attractor. Instead for vc0 there are three fixed points: (f40, ḟ40), which is unstable, and (f46v, ḟ40), which are stable. The Hubble function is determined by solving eq. (5). There are four possible continuous solutions: (7) H(f) 46gNf 2 2v 2 N, Fig. 1. The Hubble function vs. time (top) and the phase space trajectory of the inflaton field (bottom); for H(f) 4gNf 2 2v 2 N with g41o2, l43 and v41. Initial conditions: f40 and ḟ 41O2.

4 876 L. SALASNICH but also (8) H(f) 46g(f 2 2v 2 ), where g4k2pglo3 is the friction parameter. The choice of the solution is crucial for the dynamical evolution of the system. By using the Bendixon criterion [13] (discussed in detail in [6]) we obtain that if H(f) 4gNf 2 2v 2 N then the Hubble function does not change sign and we do not find periodic orbits. The 4th-order Runge-Kutta numerical integration [14] of the equations of motion shows that for v c 0 the inflation field approaches one of its two stable fixed-point attractors, and that the Hubble function goes to zero with an oscillatory behaviour (see fig. 1). Instead, if we choose H(f) 4g(f 2 2v 2 ), then the Hubble Fig. 2. The Hubble function vs. time (top) and the phase space trajectory of the inflaton field (bottom); for H(f) 4g(f 2 2v 2 ) with g41o2, l43 and v41. Initial conditions: f40 and ḟ 44.

5 ON THE LIMIT CYCLE OF AN INFLATIONARY UNIVERSE 877 function can change sign and the Bendixon criterion admits for vc0 the existence of a limit cycle. In fact, the numerical calculations plotted in fig. 2 show that a limit cycle exists and that the Hubble function oscillates forever. Now we want analyze in detail the properties of this limit cycle. The equation of motion of the inflaton field with H(f) 4g(f 2 2v 2 ) reads (9) f O 13g(f 2 2v 2 ) ḟ1lf(f 2 2v 2 ) 40. This equation can be written as d yḟ13g f (10) (u 2 2v 2 )duz1lf(f 2 2v 2 ) 40, dt 0 and if we put f (11) F(f) 43 (u 2 2v 2 )du4f(f 2 23v 2 ), G(f)4f(f 2 2v 2 ), 0 and also v4 ḟ1gf(f), we obtain the system (12). / ḟ 4v2gF(f), v. 42lG(f). For systems of this kind the Lienard theorem [15, 16] states that there is a unique and stable limit cycle if the following conditions are satisfied: F(f) is an odd function and F(f) 40 only at f40 and f46a; F(f)E0 for 0 EfEa, F(f) D0 and is increasing for fda; G(f) is an odd function and fg(f) D0 for all fda. It is easy to check that the functions F(f) and G(f) defined by (11) satisfy all the conditions of the Lienard theorem with a4v. The cubic force G(f) tends to reduce any displacement for large NfN, whereas the damping F(f) is negative at small NfN and positive at large NfN. Since small oscillations are pumped up and large oscillations are damped down, it is not surprising that the system tends to settle into a self-sustained oscillation of some intermediate amplitude. Figures 2 and 3 show that both internal and external initial conditions generate trajectories which approach the limit cycle, so we have also a numerical evidence of the stability of the limit cycle. Let us consider a typical trajectory of the Lienard system (12). After the scaling c4lv we obtain (13). F(f)l ḟ 4lkc2 g l /, ċ 42G(f). The cubic nullcline c4 (gol)f(f) is the key to understand the motion [5]. Suppose that lc1 and the initial condition is far from the cubic nullcline, then (13) implies NḟNAO(l)c1; hence the velocity is enormous in the horizontal direction and tiny in the vertical direction, so trajectories move practically horizontally. If the initial condition is above the nullcline then ḟ D 0, thus the trajectory moves sideways toward

6 878 L. SALASNICH Fig. 3. The Hubble function vs. time (top) and the phase space trajectory of the inflaton field (bottom); for H(f) 4g(f 2 2v 2 ) with g41o2, l43 and v41. Initial conditions: f421o2 and ḟ 40. the nullcline. However, once the trajectory gets so close that cc (log) F(f), then the trajectory crosses the nullcline vertically and moves slowly along the backside of the branch until it reaches the knee and can jump sideways again. The period T of the limit cycle is essentially the time required to travel along the two slow branches, since the time spent in the jumps is negligible for large l. By symmetry, the time spent on each branch is the same so we have (14) t B TC2 dt, t A

7 ON THE LIMIT CYCLE OF AN INFLATIONARY UNIVERSE 879 where A and B are the initial and final points on the positive slow branch. To derive an expression for dt we note that on the slow branches with a good approximation cc (gol) F(f) and thus (15) dc dt C g l F 8(f) df dt 43 g l (f 2 2v 2 ) df dt. Since from (13) dcodt42f(f 2 2v 2 ), we obtain (16) dtc23 g l df f, on the slow branches. The slow positive branch begins at f A 42gvOl and ends at f B 4 gvol, hence (17) Because g4k2pglo3 we have t B TC2dtC26 g f Bdf l f C6 g l ln 2. f A t A (18) TC2 ln2o 6pG l. Note that the period is v-independent. In summary, we have proved the existence and stability of a limit cycle in the phase space of a scalar inflaton field f with a symmetric double-well potential V(f) and a friction term in the equation of motion proportional to V(f). Then we have obtained an analytical estimation of the period of the limit cycle. *** The author is greatly indebted to V. R. MANFREDI and M. ROBNIK for many enlightening discussions. R E F E R E N C E S [1] KAWABE T. and OHTA S., Phys. Lett. B, 334 (1994) 127; KAWABE T., Phys. Lett. B, 343 (1995) 225. [2] SALASNICH L., Phys. Rev. D., 52 (1995) 6189; GRAFFI S., MANFREDI V. R. and SALASNICH L., Mod. Phys. Lett. B, 7 (1995) 747. [3] SEGAR J. and SRIRAM M. S., Phys. Rev. D, 53 (1996) [4] NAYFEH A. H. and BALACHANDRAN B., Applied Nonlinear Dynamics (J. Wiley, New York) [5] FARKAS M., Periodic Motion (Springer, Berlin) [6] SALASNICH L., Mod. Phys. Lett. A, 10 (1995) [7] GUTH A. H., Phys. Rev. D, 23 (1981) 347. [8] LINDE A. D., Phys. Lett. B, 108 (1982) 389; 129 (1983) 177. [9] LINDE A. D., Particle Physics and Inflationary Cosmology (Harwood Academic Publishers, London) 1988.

8 880 L. SALASNICH [10] BRANDENBERGER R. H., in SUSSP Proceedings: Physics of the Early Universe, edited by J. A. PEACOCK, A. F. HEAVENS and A. T. DAVES (Institute of Physics Publishing, Bristol) [11] ITZYKSON C. and ZUBER J. B., Quantum Field Theory (McGraw-Hill, New York) [12] BECK C., Nonlinearity, 8 (1995) 423. [13] BENDIXSON I., Acta Math., 24 (1901) 1. [14] Subroutine D02BAF, The NAG Fortran Library, Mark 14, Oxford: NAG Ltd. and USA: NAG Inc. (1990). [15] LIENARD A., Rev. Gen. Electr., 23 (1928) 901. [16] JORDAN D. W. and SMITH P., Nonlinear Ordinary Differential Equations (Oxford University Press, Oxford) 1987.