-
In a single-field inflationary model, the Lagrangian density can be written as
$ {\mathcal{L}}(g_{\alpha\beta}, \psi)=\frac{1}{2}g^{\alpha\beta}\nabla_{\alpha}\psi\nabla_{\beta}\psi+V(\psi), $
(1) where ψ is the inflaton field, and the potential function is represented by
$ V(\psi) $ .The full action with the unit
$ \hbar=c=1 $ is given by$ S=\frac{1}{16\pi G} \int {\rm d}^{4}x \sqrt{-g} {\mathcal{R}} - \int {\rm d}^{4}x \sqrt{-g}\mathcal{L} , $
(2) where the first term represents the Einstein–Hilbert action.
If we consider the universe to be isotropic and homogeneous, the evolution of the inflaton field
$ \psi(t) $ and the scale factor$ a(t) $ can be expressed by the Friedmann equation:$ 3H^2=k\left[\frac{1}{2}\dot{\psi}^2+V(\psi)\right], $
(3) $ \begin{array}{*{20}{l}} \ddot{\psi}+3H\dot{\psi}=-V'(\psi), \end{array} $
(4) where the Hubble parameter
$ H=\dfrac{\dot{a}}{a} $ , and$ k=8\pi G $ . The dot represents differentiation with respect to t, whereas prime indicates differentiation with respect to ψ. When the slow-roll parameters are small, we note a period of accelerated expansion, i.e., inflation, and the corresponding slow-roll parameters can be written as [40]$ \epsilon(\psi)=\frac{m_p^2}{2}\left[\frac{V'(\psi)}{V(\psi)}\right]^2, $
(5) $ \eta(\psi)=m_p^2\left[\frac{V''(\psi)}{V(\psi)}\right]. $
(6) Here,
$ {m_p}^2=\dfrac{1}{G} $ , and$ m_p $ denotes the Planck mass. As the slow-roll parameters are small, the time-derivatives of ψ in Eqs. (3) and (4) can be justifiably ignored, and the potential term is much greater than the kinetic term, i.e.,$ \dfrac{1}{2}\dot{\psi}^2 \ll V(\psi) $ . Thus, Eq. (3) becomes$ 3H^2\approx k V(\psi) $ . Therefore, when$ \psi \approx $ constant, the energy density corresponding to the scalar field has a constant value, which results in a de sitter solution.In addition to exponential acceleration, non-exponential accelerated expansion can also perform work [41–43]. One may consider, for instance, a power law expansion, and the corresponding scale factor is given by
$ a(t)\sim t^q $ , where the power law index q is greater than unity ($ q>1 $ ). In the conventional PLI, the canonical scalar field has an exponential potential of the form$V(\psi)=V_0{\rm e}^{-\sqrt{\frac{2}{q}}(\frac{\psi}{M_p})}$ , with$ V_0 $ as a constant [44–48]. With the exponential potential, the slow-roll parameters from Eqs. (5) and (6) can be written as$ \epsilon=\frac{1}{q}, $
(7) $ \eta=\frac{2}{q}. $
(8) Within the slow-roll paradigm,
$ \epsilon, \eta\; \ll 1 $ , and hence,$ q\gg 1 $ . For the de Sitter inflationary paradigm, the exponential expansion terminates when the slow-roll approximation is no longer applicable. In the PLI, q is a constant parameter, and hence, the termination of inflation is not clear. Thus, an exit mechanism should be incorporated into the entire inflationary scenario such that with a progress in the expansion, the inflationary phase matches the standard hot Big Bang model, i.e., the radiation or the matter dominated region.Any given inflationary model is usually characterized by the scalar spectral index,
$ n_s $ , and the tensor-to-scalar ratio, r [49, 50]. These two parameters are related to the slow-roll parameters ($ \epsilon, \eta $ ) as$ \begin{array}{*{20}{l}} n_s-1 = 2\eta-6\epsilon, \end{array} $
(9) $ \begin{array}{*{20}{l}} r = 16\epsilon. \end{array} $
(10) For the PLI, we have
$ n_s-1 = -\frac{2}{q}, $
(11) $ r = \frac{16}{q}. $
(12) The Planck data [24], combined with the Wilkinson microwave anisotropy probe (WMAP) result [23], require that the value of the scalar spectral index (
$ n_s $ ) be in the range of$ n_s\in [0.945, 0.98] $ and the tensor-to-scalar ratio be$ r<0.11 $ (in the model Planck TT+lowP+BAO). The constraint originating from a combination of the Planck [26], BICEP2, and Keck Array data [25, 26] is$ r<0.07 $ . The observational limits on$ n_s $ are equivalent to the limits of the power law index,$ 38\leq q \leq 101 $ , which correspond to$ 0.16 < r < 0.43 $ . This range of r lies beyond the above-mentioned ranges. In Fig. 2, the outermost and middle contours represent the Planck + WP data, and the innermost contour corresponds to the Planck + BICEP2 + Keck Array data. The solid line representing the PLI lies completely outside the outer contour.Thus, the conventional PLI has two drawbacks. The first problem is the graceful exit problem, and the second is the mismatch between observations and theoretical predictions. Hence, one must clarify the Lagrangian density corresponding to the inflaton so that
$ n_s $ and r become consistent with the observational data [43, 51]. -
Conformal transformation is typically used as a mathematical tool to map the equations of motion between physical systems and mathematically equivalent sets of equations, rendering them easier to solve and computationally more convenient to study. In this model, we make two assumptions regarding the gravitational coupling of matter systems. First, different types of matter couple with a particular metric. This implies that all types of fields in different standard models couple in a similar manner to gravity, irrespective of extensive variations in their physical properties. Second, we consider a unique metric describing the background geometry. Thus, gravitational coupling is universal, and the equivalence principle supports this fact with numerous observable results, such as those presented by Damour [52, 53]. This has also been verified empirically several times since the seventeenth century [54, 55]. In this case, the results are extrapolated to the total age of the universe because all the equivalence principle tests have been conducted within a limited time interval (four hundred years since the time of Galileo). However, there exists a possibility that the equivalence principle has been violated at some point during the evolution of the universe. Further, all the equivalence principle tests are restricted to the solar system. It is well known that there exist some screening techniques, based on which the anomalous gravitational coupling of matter can be obscured from experiments, e.g., if a chameleon scalar field interacts with matter [56, 57]), such an interaction cannot be detected empirically. In this case, the local gravity constraints are suppressed in the laboratory as the chameleon field is heavy. Meanwhile, the field can be sufficiently light in the low-density cosmological environment to produce observable effects on a large scale.
In this study, we attempt to consider gravitational coupling to cool off the aforementioned assumptions. We also consider the fact that the two metrics for the matter and gravitational sectors relate to different conformal frames. In this case, we consider a minimally coupled scalar field as depicted by the Lagrangian density (1). As we consider contributions resulting from the various fields of elementary particles, the potential of the inflaton ψ has a large effective mass, which corresponds to a large effective cosmological constant. Further, we consider in the full action that the inflaton part pertains to a different conformal frame [58, 59]. Thus,
$ \bar{g}_{\alpha\beta} ={\rm e}^{-2\xi} g_{\alpha\beta}, $
(13) $ \begin{array}{*{20}{l}} \bar{\psi} = {\rm e}^{\xi} \psi. \end{array} $
(14) From Eq. (13), the inverse metric
$ g^{\alpha\beta} $ and the determinant g = det[$ g_{\alpha\beta} $ ] transform as$ \begin{array}{*{20}{l}} \bar{g}^{\alpha\beta} ={\rm e}^{2\xi} g^{\alpha\beta}, \end{array} $
(15) $ \begin{array}{*{20}{l}} \sqrt{-\bar{g}} = {\rm e}^{-4\xi} \sqrt{-g}. \end{array} $
(16) We consider the conformal transformations as local unit transformations [60–65] with a space-time dependent conversion factor. Usually, ξ depends on the space-time, and it is a smooth, dimensionless function. However, later in the calculation, we will assume that ξ is a function of time only. Hence, the Lagrangian density corresponding to the inflaton can be written as
$ {\mathcal{L}}(\bar{g}_{\alpha\beta}, \bar{\psi})=\frac{1}{2}\bar{g}^{\alpha\beta}\nabla_{\alpha}\bar{\psi}\nabla_{\beta}\bar{\psi}+V(\bar{\psi}). $
(17) Based on the fact that the Lagrangian is invariant under a conformal transformation, the full action from Eq. (2) is given by
$ \begin{eqnarray} S=\frac{1}{2k} \int {\rm d}^{4}x \sqrt{-g} {\mathcal{R}} -\int {\rm d}^{4}x \sqrt{-\bar{g}} {\mathcal{L}}(\bar{g}_{\alpha\beta} \bar{\psi}). \end{eqnarray} $
(18) In terms of
$ g_{\alpha\beta} $ and ψ, the action in Eq. (18) becomes$ \begin{aligned}[b] S =&\frac{1}{2} \int {\rm d}^{4}x \sqrt{-g}\bigg\{\frac{1}{k}{\mathcal{R}}-g^{\alpha\beta} \nabla_{\alpha} \psi \nabla_{\beta}\psi \\& -2\psi g^{\alpha\beta} \nabla_{\alpha} \psi \nabla_{\beta}\xi -\psi^2 g^{\alpha\beta} \nabla_{\alpha} \xi \nabla_{\beta}\xi -V({{\rm e}^{\xi}}\psi){\rm e}^{-4\xi}\bigg\}. \end{aligned} $
(19) The action functional thus obtained depends on two scalar fields, viz., ξ and ψ, which are dynamical in nature with a term belonging to the mixed kinetic type [66, 67]. This type of system has an important application in the formulation of assisted quintessence [68–71], as well as in the amelioration of different dark energy models [67, 72, 73].
We may consider the following:
$ \begin{array}{*{20}{l}} \psi g^{\alpha\beta} \nabla_{\alpha}\psi \nabla_{\beta}\xi = \nabla_{\alpha} (\psi \xi g^{\alpha\beta}\nabla_{\beta}\psi)-\xi g^{\alpha\beta} \nabla_{\alpha} \psi\nabla_{\beta}\psi-\psi \xi \Box\psi. \end{array} $
(20) Using Eq. (20), the action in Eq. (19) can be written as
$ \begin{aligned}[b] S =&\frac{1}{2} \int {\rm d}^{4}x \sqrt{-g} \bigg\{\frac{1}{k}{\mathcal{R}}-(1-2\xi)g^{\alpha\beta} \nabla_{\alpha} \psi \nabla_{\beta}\psi \\& - 2\psi \xi \Box \psi -\psi^2 g^{\alpha\beta} \nabla_{\alpha} \xi \nabla_{\beta}\xi -V({\rm e}^{\xi}\psi){\rm e}^{-4\xi}\bigg\}. \end{aligned} $
(21) It should be noted from Eq. (21) that the scalar field has a dimension of
$ L^{-1} $ or M in a natural unit system. However, the conformal factor is dimensionless. Thus, ξ is a dynamical variable but not a scalar field, and hence, in Eq. (21), we only have one scalar field.When the slow-roll condition is valid, we can write
$ \begin{array}{*{20}{l}} \{(\partial \psi)^2, \Box \psi \} \ll V({\rm e}^{\xi}\psi){\rm e}^{-4\xi}. \end{array} $
(22) Thus, Eq. (21) can be approximated to
$ S=\frac{1}{2} \int {\rm d}^{4}x \sqrt{-g} \left\{\frac{1}{k}{\mathcal{R}} -\psi^2 g^{\alpha\beta}\partial_{\alpha} \xi \partial_{\beta}\xi-V({\rm e}^{\xi}\psi){\rm e}^{-4\xi}\right\}. $
(23) It is well known that during inflation, the potential term is much greater than the kinetic term. Hence, the approximation depicted in Eq. (22) is justified, implying that Eq. (23) is correct. However, we can use Eq. (23) instead of the full action in Eq. (21). It is also interesting to note the existence of an exponential coefficient in the potential term. If ξ increases with time, the coefficient plays the role of a damping factor, and the potential decreases.
-
As an example, we consider a potential of the form [6]
$ V(\bar{\psi})= \nu m_p^4\left(\frac{\bar{\psi}}{m_p}\right)^p,$
(24) where ν and p are constant quantities with
$ \nu\ll 1 $ .The corresponding slow-roll parameters are given by
$ \epsilon=\frac{p^2}{2\gamma^2}\; ,\; \eta=\frac{p(p-1)}{\gamma^2}, $
(25) where
$ \gamma \equiv \psi/m_p $ , and during slow-roll inflation, we can write$ \gamma \gg 1 $ , which will be discussed later in this section.With this, from Eq. (23) and using the monomial potential in Eq. (24), we have
$ \begin{aligned}[b] S=&\frac{1}{2} \int {\rm d}^{4}x \sqrt{-g} \Bigg[\frac{1}{k}{\mathcal{R}} -\psi^2\Big\{g^{\alpha\beta} \partial_{\alpha} \xi \partial_{\beta}\xi\\&+ \nu m_p^{4-p}\psi^{p-2} {\rm d}^{(p-4)\xi}\Big\}\Bigg]. \end{aligned} $
(26) Varying the action with respect to
$ g_{\alpha\beta} $ and ξ produces the required field equations in a spatially flat FRW background, as follows:$ 3 \left(\frac{\dot{a}}{a}\right)^2=\frac{1}{2}k \psi^2\left\{\dot{\xi}^2+\nu m_p^{4-p}\psi^{p-2} {\rm d}^{(p-4)\xi}\right\}. $
(27) $ \ddot{\xi}+3H \dot{\xi}+\frac{1}{2}(p-4)\nu m_p^{4-p}\psi^{p-2} {\rm e}^{(p-4)\xi}=0. $
(28) The corresponding solution can be provided as follows [31]
$ \begin{array}{*{20}{l}} a(t)=a_{0} t^q , \end{array} $
(29) $ \xi(t)=\xi_T - C\ln {\left(\frac{t}{t_T}\right)}, $
(30) where
$\begin{aligned}[b]& q= 4\pi C^2\gamma^2,\quad C=\frac{2}{(p-4)} \; \;\text{and}\; \\& t_T^2=\left[\frac{4(3q-1)}{\nu \gamma^{p-2}(p-4)^2m_p^2 {\rm e}^{(p-4)\xi_T}}\right]. \end{aligned} $
(31) In this case,
$ \xi_T $ is a constant, which is a dimensionless quantity and represents the value of ξ when the inflation terminates. However, following the approach adopted by Kalara et al. [31], one can opt for simplified forms of Eqs. (29)–(31).From Eq. (31), we can assert that during a power law inflationary phase, the universe results in
$ \gamma>\dfrac {|p-4|}{4\sqrt{\pi}} $ . Here, the values of ψ from$ -\infty $ to$ +\infty $ are completely appropriate. If$ \rho_{\psi} $ is the energy density of ψ, and$ \rho_{\psi}< m_p^4 $ , the universe can be described classically. As$ \nu \ll 1 $ , one can constrain the kinetic energy of ψ, viz.,$ (\partial\psi)^2<m_p^4 $ [5, 6].From Eq. (30), it is evident that
${\rm e}^{(p-4)\xi}$ decreases with t. Thus,$\Lambda_{\rm eff}\equiv 4\pi\nu m_p^2\gamma^p {\rm e}^{(p-4)\xi}$ reduces during the inflationary phase such that$\Lambda_{\rm eff}\sim t^{-2}$ . This result matches well with the observational upper limit and with the phenomenological models proposed by Ray et al. [19] and Mukhopadhyay et al. [20]; in these studies, it was concluded that for a flat universe,$ \Lambda\sim t^{-2} $ is true for different models.To discriminate among the different components that may be responsible for the present acceleration of the universe [74, 75], two new geometrical parameters, termed as the statefinder parameters, which depend on the nature of the space-time metric, were introduced by Sahni et al. [76]. These parameters are usually denoted by r and s, but here, we will denote them by
$ r' $ and$ s' $ . The parameters are defined, along with the deceleration parameter$q_{\rm dec}$ , as$ q_{\rm dec}=-1-\frac{\dot{H}}{H^2}, $
(32) and
$ r'=1+\frac{3\dot{H}}{H^2}+\frac{\ddot{H}}{H^3},\; s'=\frac{r'-1}{3\left(q_{\rm dec}-\frac{1}{2}\right)}. $
(33) Using the form of a obtained in Eq. (31), the deceleration and statefinder parameters can be determined to be
$ q_{\rm dec}=-\left(\frac{q+1}{q}\right),\; r'=1+\frac{2-3q}{q^2},\; s= \frac{2}{q}\left(\frac{3q-2}{3q+2}\right). $
(34) We can observe from the above result that despite using a scalar field inflaton to generate the inflationary mechanism, both the statefinder parameters are determined to be constants, which is also the case for the Λ-term. Thus, our description of the inflationary mechanism in terms of
$\Lambda_{\rm eff}$ is further established.Another important point regarding the inflationary model is the exit mechanism, i.e., the mechanism of inflation termination. The exit mechanism has been previously identified as a serious problem in the PLI. In the present study, graceful exit occurs owing to the decay of vacuum density. The inflaton ψ is freezed during the slow-roll paradigm, and the corresponding energy density of ψ is given by
$\rho_{\psi}\equiv\dfrac{1}{2}\dot{\psi}^2+V{\rm e}^{-4\xi}\approx V{\rm e}^{-4\xi}$ . Unlike exponential inflation, the energy density$ \rho_{\psi} $ in the present case does not remain constant but decays during inflation. With the evolution of time,$V{\rm e}^{-4\xi}$ reduces. Thus, at a particular stage, there will exist an instant in time when the kinetic term cannot be neglected in$ \rho_{\psi} $ . At this point in time, the kinetic and potential terms will have the same order of magnitude, i.e., the slow-roll approximation will no longer be valid at this stage, and inflation will terminate.After the inflationary paradigm, the reheating stage initiates, and during this stage, the inflaton begins oscillating near the minimum of its effective potential; consequently, elementary particles are produced. These particles interact with each other and ultimately create a state of thermal equilibrium for the universe at some temperature. During reheating, the inflaton energy is converted into matter and radiation, after which the universe re-enters the hot Big Bang model phase, followed by the dark matter and vacuum phases. In this model, we deal with two dynamical scalar fields; however, the reheating process is actually controlled by the inflaton ψ. At the end of the inflation, the conformal factor tends to have a constant configuration, and the kinetic energy part of the inflaton, which was unimportant during inflation, now becomes important. Indeed, the parts of ψ and ξ change during the phase of reheating, and the model is again reduced to a single-field type. In the present case, the reheating process proceeds in a similar manner as that in the standard Big Bang model.
Next, let us assume that inflation terminates at time
$ t_T $ . From Eq. (30), it is evident that$ \xi \rightarrow \xi_0 $ when$ t \rightarrow t_T $ . This implies that when inflation terminates, ξ assumes a constant value. It is clear from Fig. 1 that practically, even when$ t \ll t_T $ , the variation in ξ is negligible, i.e.,$ t_b \ll t_T $ . Therefore, the action given by Eq. (19) reduces to$ S=\frac{1}{2} \int {\rm d}^{4}x \sqrt{-g}\; \left\{\frac{1}{k}{\mathcal{R}}-g^{\alpha\beta} \nabla_{\alpha} \psi \nabla_{\beta}\psi -V(\psi)\right\}, $
(35) where the conformal factor
${\rm e}^{-4\xi}$ in Eq. (19) becomes a constant factor${\rm e}^{-4\xi_T}$ , and hence, it can be consumed by the potential. Thus, after the termination of inflation, one would expect to observe the following two features: (i) the effective cosmological term$\Lambda_{\rm eff}\sim t^{-2}$ decreases in an identical manner to the energy densities of the radiation and matter dominated phases after the inflationary era in the Big Bang model, and (ii) reheating initiates in a manner very similar to that in the conventional inflationary models.One important factor in any model regarding inflation is the number of e-foldings produced by inflation, which is defined as
$ N\equiv \ln \frac{a_T}{a_b} =\int_{t_b}^{t_T} \frac{{\rm d}a}{a}=\int_{t_b}^{t_T} H {\rm d}t, $
(36) where
$ t_b $ and$ t_T $ denote the times at which inflation begins and terminates, respectively, and$ a_b $ and$ a_T $ are the corresponding scale factors.Substituting the solutions of (29) and (30) in Eq. (27), we can obtain
$ \begin{array}{*{20}{l}} N\sim q\ln\left(\dfrac{t_T}{t_b}\right), \end{array} $
(37) where q can be obtained from Eq. (31).
Next, to overcome the smoothness and flatness problems, we must ensure that
$ N>60 $ . Based on Eqs. (31) and (37) and the fact that$ t_b \ll t_T $ and$ \gamma \gg 1 $ , one can understand that this condition can be achieved easily. In this connection, we would like to mention that with a constant value of ψ as$ \gamma \gg 1 $ , it is possible to easily achieve the slow-roll condition, and we explain the smoothness and flatness problems based on this. Here, ξ is a dynamical variable; however, ξ eventually becomes a constant quantity, as is evident from Fig. 1. Hence, Eq. (23) effectively transforms into Eq. (35), and it implies the end of inflation. Therefore, we can explain the end of inflation through the dynamical behavior of ξ. -
In this section, we compare our model results with recent observational data [23–26].
From Eq. (25), we have
$ n_s-1=-\frac{p(p+2)}{\gamma^2}, $
(38) $ r =\frac{8p^2}{\gamma^2}. $
(39) Equations (38) and (39) are plotted in Fig. 2 for different values of p. Essentially, we do not require an explicit value for γ while plotting the variation in
$ n_s $ and r; this is because on considering the ratio, the factor γ cancels out. Thus, we skip the step to opt for any particular value of the parameter.Figure 2. (color online) Plot of r versus
$ n_s $ for different values of p of the monomial potential (all the coloured lines between$ n_s $ and the extreme right line of the conventional power law inflation). The innermost contour corresponds to the Planck + BICEP2 + Keck Array data, whereas the outermost and middle contours correspond to the Planck + WMAP + BAO data at σ and 2σ confidence limits (CL), respectively.Figure 2 indicates that while the conventional power law inflationary model with exponential potential remains completely outside the region allowed by observational results in the
$ \{r, n_s\} $ space, the present model (Eq. (26)) is in very good agreement with the results reported by Dunkley et al. [23] and Ade et al. [24], and it agrees fairly well with the results reported by Ade et al. [25]) and Akrami et al. [26], along with the highly modified inflationary models. It is worth noting that in the present study, for a minimal coupling scalar field, the power-law potential index must satisfy$ p<4 $ to agree with the observational data reported by Dunkley et al. [23], Ade et al. [24, 25], and Akrami et al. [26]. This is different from the non-minimal coupling scalar field case, for which the power law potential index must satisfy$ p>4 $ [77]. In general, for chaotic inflationary models, the usual convention is the following:$ p \geq 1 $ ; however, in the present model, one can infer that for$ p<1 $ , the latest experimental result can be retrieved particularly well. In this connection, it is worthwhile to mention that the monomial potentials$ V(\psi) = \nu m^4_{Pl} (\psi/m_{Pl})^p $ , as described [see Eq. (24)] by Linde [6], with$ p \geq 2 $ are strongly unfavorable with respect to the$ R^2 $ model. Akrami et al. [26] argued that, for these values, the Bayesian evidence is worse than it was in 2015 owing to the smaller level of tensor modes allowed by BK14 [25]. Notably, models with$ p = 1 $ or$ p = 2/3 $ are more compatible with the data reported by Silverstein and Westphal [78] and McAllister et al. [79, 80]. It is interesting to note that our prescribed value$ p<1 $ agrees well with the second option$ p = 2/3 $ .
Studies on modified power law inflation
- Received Date: 2022-09-11
- Available Online: 2023-03-15
Abstract: In this study, we evaluate power law inflation (PLI) with a monomial potential and obtain a novel exact solution. It is well known that the conventional PLI with an exponential potential is inconsistent with the Planck data. Unlike the standard PLI, the present model does not encounter the graceful exit problem, and the results agree fairly well with recent observations. In our analysis, we calculate the spectral index and the tensor-to-scalar ratio, both of which agree very well with recent observational data and are comparable with those of other modified inflationary models. The employed technique reveals that the large cosmological constant decreases with the expansion of the universe in the case of the PLI. The coupling of the inflaton with gravitation is the primary factor in this technique. The basic assumption here is that the two metric tensors in the gravitational and inflaton parts correspond to different conformal frames, which contradicts with the conventional PLI, where the inflaton is directly coupled with the background metric tensor. This fact has direct applications to different dark energy models and the assisted quintessence theory.