-
In this section, we briefly discuss the main elements corresponding to the present model, deducing the gravitational field equations in a cosmological context. In the subsequent discussion, we will introduce a model characterized by the following action:
$ \begin{equation} S=S_m+\int {\rm d}^4x \sqrt{-\tilde{g}} \big[f(R)+g(\phi) \big], \end{equation} $
(1) where a new generic function is added to the
$ f(R) $ part [32]. This function is further based on specific contractions between the Einstein tensor and energy–momentum tensor [80]:$ \begin{equation} \phi=G_{\mu \nu} T^{\mu \nu}. \end{equation} $
(2) This action can be regarded as a special case and complements the approach presented in Refs. [51, 69, 81]. Here, the energy–momentum tensor is defined in the usual manner as follows:
$ \begin{equation} T_{\mu \nu}=-\frac{2}{\sqrt{-\tilde{g}}}\frac{\delta(\sqrt{-\tilde{g}}L_m)}{\delta \tilde{g}^{\mu \nu}}, \end{equation} $
(3) where
$ L_m $ denotes the Lagrangian for the matter sector. Before proceeding to the derivation of the Einstein field equations, we introduce new elements for consideration. The trace of the energy–momentum tensor is defined as:$ T=T^{\mu \nu}\tilde{g}_{\mu \nu} $ . The variation of the energy–momentum tensor with respect to the inverse metric is equal to the following relation [80]:$ \begin{equation} \frac{\delta T_{\alpha \beta}}{\delta \tilde{g}^{\mu \nu}}=\frac{\delta \tilde{g}_{\alpha \beta}}{\delta \tilde{g}^{\mu \nu}}L_m+\frac{1}{2}\tilde{g}_{\alpha \beta}L_m \tilde{g}_{\mu \nu}-\frac{1}{2}\tilde{g}_{\alpha\beta}T_{\mu\nu}-2\frac{\partial^2 L_m}{\partial \tilde{g}^{\mu\nu} \partial \tilde{g}^{\alpha\beta}}. \end{equation} $
(4) Next, in our computations, we consider the following contraction [80]:
$ \begin{aligned}[b] \Sigma_{\mu\nu}=&G^{\alpha\beta}\frac{\delta T_{\alpha\beta}}{\delta \tilde{g}^{\mu\nu}}=-G_{\mu\nu}L_m+\frac{1}{2}G^{\alpha\beta}\tilde{g}_{\alpha\beta}(\tilde{g}_{\mu\nu}L_m-T_{\mu\nu}) \\&-2 G^{\alpha\beta}\frac{\delta^2 L_m}{\delta \tilde{g}^{\mu\nu} \delta \tilde{g}^{\alpha\beta}}. \end{aligned} $
(5) For the
$ f(R) $ part, the variation in the corresponding action with respect to the inverse metric leads to the associated energy momentum tensor [32],$ \begin{equation} T_{\mu\nu}^{f(R)}=\tilde{g}_{\mu\nu}f(R)-2 R_{\mu\nu}f_R+2\nabla_{\mu}\nabla_{\nu}f_R-2 \tilde{g}_{\mu\nu} \Box f_R, \end{equation} $
(6) where we introduced the derivative with respect to the scalar curvature,
$ f_R=\dfrac{\partial f(R)}{\partial R} $ . In the case of$ g(\phi) $ component, we obtained the following energy–momentum tensor [80]:$ \begin{aligned}[b] T_{\mu\nu}^{g(\phi)}=&\tilde{g}_{\mu\nu}g(\phi)+g_{,\phi}T R_{\mu\nu}-2 g_{,\phi} G_{\nu}^{\beta}T_{\mu\beta}-2 g_{,\phi} G_{\mu}^{\alpha}T_{\nu\alpha} \\&-g_{,\phi}R T_{\mu\nu}-\Box (g_{,\phi} T_{\mu\nu})+\nabla_{\alpha}\nabla_{\mu}(g_{,\phi}T_{\nu}^{\alpha})+\nabla_{\alpha}\nabla_{\nu}(g_{,\phi}T_{\mu}^{\alpha}) \\&-\tilde{g_{\mu\nu}}\nabla_{\alpha}\nabla_{\beta}(g_{,\phi} T^{\alpha\beta})+\tilde{g_{\mu\nu}}\Box(g_{,\phi}T)\\&-\nabla_{\mu}\nabla_{\nu}(g_{,\phi}T)-2g_{,\phi}\Sigma_{\mu\nu}, \end{aligned} $
(7) where
$ g_{,\phi} $ denotes the derivative with respect to the ϕ invariant, i.e.$ \begin{equation} g_{,\phi}=\frac{d g(\phi)}{d \phi}. \end{equation} $
(8) If we further apply the principle of least action, we obtain the final Einstein–like equation [32] as follows:
$ \begin{equation} T_{\mu\nu}^{f(R)}+T_{\mu\nu}^{g(\phi)}+T_{\mu\nu}^{m}=0. \end{equation} $
(9) This leads to the conservation relation as follows:
$ \begin{equation} \nabla^{\mu}\big[T_{\mu\nu}^{f(R)}+T_{\mu\nu}^{g(\phi)}+T_{\mu\nu}^{m}\big]=0. \end{equation} $
(10) This is also known as the continuity equation.
Next, we consider the following cosmological context associated to the FLRW model described by the metric:
$ \begin{equation} {\rm d}s^2=-{\rm d}t^2+a(t)^2\delta_{ij}{\rm d}x^i{\rm d}x^j, i,j=1,2,3. \end{equation} $
(11) In this instance, we consider a universal scale factor a that is dependent on cosmic time. Subsequently, we define the Hubble parameter in the conventional manner,
$ H=\dfrac{\dot{a}}{a} $ , where the dot denotes differentiation with respect to cosmic time. The energy–momentum tensor for the barotropic matter fluid is as follows:$ \begin{equation} T_{\mu\nu}=(\rho_m+p_m)u_{\mu}u_{\nu}+p_m g_{\mu\nu}, \end{equation} $
(12) where
$ \rho_m $ denotes the density and$ p_m $ denotes the pressure, connected via a barotropic equation of state of the form:$ p_m=w_m\rho_m $ . In the expression of the energy–momentum tensor, we embedded the 4-velocity$ u_{\mu}=\delta_\mu^0 $ . Furthermore, for the computations, we consider$ L_m=p_m $ . Subsequently, we will disregard the pressure in the matter sector, considering a theoretical framework that corresponds to a non-relativistic fluid without pressure. Within this cosmological setting, we derive the subsequent modified Friedmann relations [80]:$ \begin{equation} f(R)-6 f_R(\dot{H}+H^2)+6 H \dot{f_R}=\rho_m-g(\phi)-6 \rho_m g_{,\phi} \dot{H}, \end{equation} $
(13) $ \begin{equation} f(R)-2 f_R(\dot{H}+3 H^2)+2 \ddot{f_R}+4 H \dot{f_R}=-p_{\phi}, \end{equation} $
(14) $ \begin{aligned}[b] p_{\phi}=&g(\phi)-2 g_{,\phi}(\rho_m(3 H^2+\dot{H})+H \dot{\rho_m}) \\ &-6 H^2 \rho_m \big[2 \rho_m \dot{H}+H \dot{\rho_m} \big] g_{,\phi\phi}. \end{aligned} $
(15) Then, we can define the effective (total) equation of state as:
$ \begin{equation} w_{tot}=-1-\frac{2}{3}\frac{\dot{H}}{H^2}. \end{equation} $
(16) For the FLRW model without pressure, the matter–geometry invariant acquires the following value,
$ \phi=3 H^2 \rho_m $ , while the scalar curvature is equal to$ R=6(\dot{H}+2 H^2) $ . As expected, the resulting field equations reduce to the fundamental Einstein equations if$ g(\phi)=0 $ and$f(R)={R}/{2}$ . Similarly, the model describes$ f(R) $ theories of gravitation [32] in the absence of the interplay between the matter energy–momentum tensor and Einstein tensor. Finally, we note that at the linear level$ (g(\phi)=g_0\phi) $ , the equations reduce to the relations presented in Ref. [80]. As previously stated, the action presented in this study offers a generalization for the analysis presented in Ref. [80], extending the field equations in a generic manner. -
In this section, we discuss the physical features in the case where the coupling functions are as follows:
$ \begin{equation} f(R)=f_0 R^n, \end{equation} $
(17) and
$ \begin{equation} g(\phi)=g_0 \phi^m, \end{equation} $
(18) where
$ f_0, g_0, n, m $ are constant parameters. To study the cosmological model, we introduce the following dimensionless variables:$ \begin{equation} x=\dfrac{\dot{f_R}}{f_R H}, \end{equation} $
(19) $ \begin{equation} z=\frac{R}{6 H^2}, \end{equation} $
(20) $ \begin{equation} s=\frac{\rho_m}{6 f_R H^2}, \end{equation} $
(21) $ \begin{equation} u=\frac{g(\phi)}{6 f_R H^2}. \end{equation} $
(22) Furthermore, we use the next non–independent variables:
$ \begin{equation} \varsigma=\frac{\ddot{R}}{H^4}, \end{equation} $
(23) $ \begin{equation} \Delta=\frac{\dot{\rho_m}}{f(R) H}. \end{equation} $
(24) Considering the above definitions, we can rewrite the Friedmann constraint Eq. (13) in the following manner:
$ \begin{equation} z\left(\frac{1}{n}-1\right)+1+x=s-u-2 u m z +4 u m, \end{equation} $
(25) by reducing the dimension of the autonomous system with one unit. Hence, we remain with the following autonomous system
$ \{z,s,u\} $ . Considering the transformation to the e-folding number$N={\rm log}(a)$ , we can approximate the dynamics of the cosmological system at the linear level as follows:$ \begin{equation} \frac{{\rm d}z}{{\rm d}N}=z \left(\frac{x}{n-1}-2 z+4\right), \end{equation} $
(26) $ \begin{equation} \frac{{\rm d}s}{{\rm d}u}=\frac{\Delta z}{n}+s (-x)-2 s z+4 s, \end{equation} $
(27) $ \begin{equation} \frac{{\rm d}u}{{\rm d}N}=u \left(\frac{\Delta m z}{n s}+2 (m-1) (z-2)-x\right). \end{equation} $
(28) Then, the acceleration Eq. (14) can be expressed in the following manner:
$ \begin{equation} \Delta = \frac{s \left(6 n z \left((2 m (4 m-5)+3) (n-1) u+(n-2) x^2+2 (n-1) x-n+1\right)-6 (n-1) z^2 (2 m (2 m-1) n u+n-3)+n (n-1)^2 \varsigma\right)}{12 m^2 (n-1) u z^2}, \end{equation} $ (29) By obtaining a direct relation between Δ and ς, an additional relation between these two non–independent variables is obtained via direct differentiation of the Friedmann constraint Eq. (13), i.e.,
$ \begin{aligned}[b] & \frac{n (6 z ((n-2) x^2-n x+2 (n-1) (z-2)+x)+(n-1)^2 \varsigma)}{6 (n-1) z^2} \\=&\Delta +m u \Bigg(-\frac{2 n (z-2) (2 m (z-2)-2 z+1)}{z} \\&+\frac{\Delta (-2 m (z-2)-1)}{s}-\frac{2 n x}{n-1}\Bigg). \end{aligned} $
(30) In this approach, we can close the autonomous system of equations by obtaining specific relations for the non–independent variables ς and Δ. The final form of the autonomous system is as follows:
$ \begin{equation} \frac{{\rm d}z}{{\rm d}N}=-\frac{z \left(n u (2 m z-4 m+1)+2 n^2 z-4 n^2-n s-3 n z+5 n+z\right)}{(n-1) n}, \end{equation} $
(31) $ \begin{equation} \frac{{\rm d}s}{{\rm d}u}=s\left(2 m u z-4 m u+\frac{z}{n}-s+u-3 z+5+ \frac{A}{B}\right) , \end{equation} $
(32) $ \begin{aligned}[b] A=&n s (2 m u z-3 n+3)-2 m u (z^2 (2 m n u-(n-1) \\&\times(2 (m-1) n+1))+n z (6 m n-4 m u-6 m\\&-7 n+u+8)-(4 m-3) (n-1) n), \end{aligned} $
(33) $ \begin{equation} B=(n-1) n (m u (-2 m z+2 m-1)+s), \end{equation} $
(34) $ \begin{aligned}[b] \frac{{\rm d}u}{{\rm d}N}=&u\Bigg(2 m u z-4 m u+2 (m-1) (z-2)\\&+\frac{z}{n}-s+u-z+1+\frac{C}{D}\Bigg), \end{aligned} $
(35) $ \begin{aligned}[b] C=&m (n s (2 m u z-3 n+3)-2 m u (z^2 (2 m n u-(n-1)\\&\times (2 (m-1) n+1))+n z (6 m n-4 m u-6 m\\&-7 n+u+8)-(4 m-3) (n-1) n)), \end{aligned} $
(36) $ \begin{equation} D=(n-1) n (m u (-2 m z+2 m-1)+s). \end{equation} $
(37) The critical points for the autonomous system are obtained by setting the r.h.s. of Eqs. (26)–(28) to zero. In this case, we obtained the following critical points, associated to different cosmological solutions, which are attached to various epochs in the evolution of the Universe.
The first critical point investigated in our analysis is located in the phase space structure at the following coordinates:
$ \begin{equation} \mathcal{Q}_1=\big[z=0, s=2, u=0 \big], \end{equation} $
(38) By describing a radiation era
$ (w_{tot}=\dfrac{1}{3}) $ , where the dynamics in mainly influenced by the matter component via the s variable. The corresponding eigenvalues are as follows:$ \begin{equation} \Bigg[-2,3-7 m,\frac{4 n-3}{n-1} \Bigg]. \end{equation} $
(39) It can be observed that this epoch can be either stable or saddle, depending on the coupling parameters n and m.
The second critical point can be observed at the following coordinates:
$ \begin{equation} \mathcal{Q}_2=\Bigg[z=2-\frac{3}{2 n}, s=\frac{(13-8 n) n-3}{2 n^2}, u=0 \Bigg], \end{equation} $
(40) which is influenced by the value of the curvature coupling without any influence from the matter–geometry interplay. The total equation of state
$(w_{\rm tot}=\dfrac{1}{n}-1)$ can mimic an accelerated expansion era by fine–tuning the value of the curvature coupling parameter n. From a dynamical perpective, we obtained the following eigenvalues:$ \begin{aligned}[b] & \Bigg[3-\frac{3 m (n+1)}{n}, \\ &\frac{\pm \sqrt{256 n^4-864 n^3+1025 n^2-498 n+81}-3 n+3}{4 (n-1) n} \Bigg]. \end{aligned} $
(41) From a physical perspective, the value of the effective matter density parameter should be real and positive,
$ s>0 $ . We showed in Fig. 1, a specific region, where$ \mathcal{Q}_2 $ denotes a stable node, where all of the eigenvalues exhibit negative real parts, and the effective matter density parameter is restricted to$ [0, 1] $ interval. In Fig. 2, we observe that the value of the total equation of state depends on the scalar curvature coupling parameter. It should be noted that the main physical features for the critical points$ \mathcal{Q}_1 $ and$ \mathcal{Q}_2 $ are not influenced by the matter–geometry invariant. These solutions have been identified in different prior studies, being specific to the$ f(R) $ component in the action [21]. For these solutions, the matter–geometry invariant affects the physical features through the dynamics, influencing the values of the corresponding eigenvalues.Figure 1. (color online) Specific region, where the critical point
$ \mathcal{Q}_2 $ , is stable and physically viable, with$ s \in [0, 1] $ .The third critical point represents a geometrical dark energy solution with the dynamics driven mainly by the matter–geometry coupling invariant,
$ \mathcal{Q}_3=\Bigg[z=0, s=0, u=\frac{8 m-5}{8 m^2-6 m+1} \Bigg]. $
(42) The solution corresponds to a radiation epoch
$ (w_{\rm tot}=\dfrac{1}{3}) $ with the following eigenvalues:$ \Bigg[\frac{6-10 m}{1-2 m},\frac{5-8 m}{2 m-1},\frac{m (8 n-2)-4 n}{(2 m-1) (n-1)} \Bigg]. $
(43) By performing appropriate fine–tuning, we can obtain viable restrictions for the constant parameters such that this solution exhibits a saddle dynamic. For example, if we consider the following restriction,
$ \begin{equation} m<\frac{1}{2}\lor m>\frac{5}{8}, \end{equation} $
(44) then we can obtain a saddle behavior associated to the radiation era.
Next, solution
$ \mathcal{Q}_4 $ can be determined in the phase space structure at the following coordinates:$ \begin{aligned}[b] \mathcal{Q}_4=&\Bigg[z=0, s=-\frac{2 m (5 m-3) (9 m-5)}{(m-1) (m (18 m-13)+3)}, \\ u=&\frac{27 m-9}{m (18 m-13)+3}+\frac{2}{m-1} \Bigg], \end{aligned} $
(45) with a similar behavior as
$ \mathcal{Q}_3 (w_{\rm tot}=\dfrac{1}{3}) $ . In this case, the effective matter density parameter is affected by the matter–geometry interplay, without any influence by the curvature coupling. We obtained the following specific eigenvalues:$ \Bigg[-\frac{2 \left(35 m^2-36 m+9\right)}{14 m^2-23 m+9},\frac{5-9 m}{m-1},\frac{4 (m n+m-n)}{(m-1) (n-1)} \Bigg]. $
(46) Similar to the previous case, appropriate fine–tuning can identify specific regions where the behavior corresponds to a saddle dynamics, compatible to the known history of the Universe. In Fig. 3, we plot the variation in the matter density parameter for different values of m, the auxiliary variable associated with the interplay between matter and geometry.
For the critical point
$ \mathcal{Q}_5 $ , we have the following coordinates:$ \mathcal{Q}_5=\Bigg[z=2, s=-\frac{2 m (n-2)}{(2 m-1) n},u=-\frac{2-n}{n-2 m n} \Bigg], $
(47) which is a particular solution corresponding to a de–Sitter epoch, where the dynamics corresponds to a cosmological constant,
$ (w_{\rm tot}=-1) $ . The equations related to the eigenvalues are too complex to present in this manuscript. As an alternative, we show in Fig. 4 a particular region of interest, where the cosmological solution is stable and physically viable, considering the following restriction:$ s \in [0, 1] $ . The numerical evolution toward$ \mathcal{Q}_5 $ solution is presented in Fig. 5, showing the compatibility between the analytical analysis and numerical approach$ (m=1, n=1.7) $ .Figure 4. (color online) Region where the cosmological solution
$ \mathcal{Q}_5 $ is physically viable and stable$ s \in [0, 1] $ .Figure 5. (color online) Evolution in the phase space toward the de–Sitter solution
$ \mathcal{Q}_5 $ for different initial conditions in the attractor basin.The last class of critical points can be determined at the following coordinates:
$ \begin{aligned}[b] \mathcal{Q}_6^{\pm}=&\Bigg[z=\frac{\pm\sqrt{4 m^2+4 (1-4 m) m n+(2 m n+n)^2}+m (6 n-2)-n}{4 m n}, s=0, \\u=&\frac{-28 m^2 n^2\mp\left(2 m (n-2)+2 n^2-3 n+1\right) \sqrt{4 m^2 ((n-4) n+1)+4 m n (n+1)+n^2}+\mathcal{R}}{8 m^2 n (3 m-n-1)} \Bigg], \end{aligned} $ (48) with
$ \mathcal{R}=40 m^2 n-4 m^2+4 m n^3-8 m n+2 m+2 n^3-3 n^2+n $ . Given that the two solutions are very similar, we discuss only the$ \mathcal{Q}_6^{+} $ case. As noted, the matter–geometry coupling and scalar curvature component influence the location in the phase space structure and corresponding dynamical features. This epoch can be considered as a curvature–matter–geometry solution, where the geometrical–matter components completely dominate in terms of effective density parameters. Based on this point, we obtained the following total equation of state:$ \begin{equation} w_{tot}=\frac{-\sqrt{4 m^2+4 (1-4 m) m n+(2 m n+n)^2}-4 m n+2 m+n}{6 m n}, \end{equation} $
(49) with influences from the scalar curvature coupling and matter–geometry invariant, respectively. By fine–tuning these parameters, we can obtain different classes of cosmological eras. For example, if we set
$ m=1 $ and$ n=\dfrac{2}{3} $ , then we can obtain a matter-dominated epoch, whereas for$ m=2 $ and$ n=\dfrac{1}{3} $ , we obtain a radiation behavior. In Fig. 6, we plot the variation in the total equation of state for different values of m and n parameters. We can observe that by fine–tuning we can obtain different values, corresponding to de–Sitter, quintessence, and phantom regimes. The final analytical expressions for the corresponding eigenvalues are very cumbersome and are not displayed in the manuscript. Instead, in Fig. 7 , we focus on displaying a non–exclusive possible interval, where the cosmological solution is represented by a saddle critical point. -
Next, we study the phase space structure for the case where the geometrical coupling function extends the fundamental Einsten–Hilbert action,
$ f(R,\phi)=\dfrac{R}{2}+g_0 \phi^{\alpha} $ , considering a power law representation for the matter–geometry invariant. In this case, we should introduce the following dimensionless variables:$ s=\frac{\rho_m}{3 H^2}, $
(50) $ x=\frac{g(\phi)}{3 H^2}, $
(51) $ y=\frac{\dot{H}}{H^2}. $
(52) In terms of dimensionless variables, the Friedmann constraint in Eq. (13) becomes:
$ \begin{equation} -1+s-x(2 \alpha y+1)=0, \end{equation} $
(53) which reduces the dimension of the phase space by one unit. Hence, the final independent dimensionless variables are:
$ \{s,x\} $ . Then, the associated dynamical equations have the following form:$ \begin{equation} \frac{{\rm d}s}{{\rm d}N}=-2 s y + \mathcal{L}, \end{equation} $
(54) $ \begin{equation} \frac{{\rm d}x}{{\rm d}N}=\frac{\alpha x \mathcal{L}}{s}+2 \alpha x y-2 x y, \end{equation} $
(55) where the additional non–independent variable is defined as:
$ \mathcal{L}=\dfrac{\dot{\rho_m}}{3 H^3} $ . The specific form of the acceleration equation (14) can be used to extract the non–independent variable,$ \begin{equation} \mathcal{L}=-\frac{s ((2 \alpha -1) x (2 \alpha y+3)-2 y-3)}{2 \alpha ^2 x}. \end{equation} $
(56) Hence, we obtain the final form of the autonomous system:
$ \begin{equation} \frac{{\rm d}s}{{\rm d}N}=\frac{s \left(x \left(4 \alpha ^2+2 \alpha +(1-4 \alpha ) \alpha s-1\right)+s-2 (\alpha -1) \alpha x^2-1\right)}{2 \alpha ^3 x^2}, \end{equation} $
(57) $ \begin{equation} \frac{{\rm d}x}{{\rm d}N}=\frac{-x (-4 \alpha +\alpha s+1)+s+2 \alpha (2-3 \alpha ) x^2-1}{2 \alpha ^2 x}. \end{equation} $
(58) As in the previous case, we identified the following critical points by analyzing the r.h.s. of the autonomous system.
The first critical point is located at the following coordinates:
$ \begin{equation} \mathcal{U}_1=\Bigg[s=\frac{2 \alpha }{2 \alpha -1}, x=\frac{1}{2 \alpha -1} \Bigg], \end{equation} $
(59) representing a de–Sitter epoch
$(w_{\rm tot}=-1)$ with the following eigenvalues:$ \begin{equation} \Bigg[\frac{2 \alpha (4-7 \alpha )\pm \sqrt{4 (\alpha -1) \alpha ((\alpha -7) \alpha +4)+1}-1}{4 \alpha ^2} \Bigg]. \end{equation} $
(60) For this point, we illustrate the value of s variable in Fig. 8, specifically for cases where α is negative. The s variable denotes the matter density parameter and, from a physical standpoint, should be positive. It is evident that this critical point corresponds to an era where the geometric coupling function is as a cosmological constant. When viewing it dynamically and in instances where α is negative, the point acts as an attractor with negative eigenvalues.
The second critical point represents an epoch located in the phase space at:
$ \begin{equation} \mathcal{U}_2^{\pm}=\Bigg[s=0, x=\frac{\pm \sqrt{-8 \alpha ^2+8 \alpha +1}-4 \alpha +1}{8 \alpha -12 \alpha ^2} \Bigg], \end{equation} $
(61) an epoch which can describe the accelerated expansion with the effective equation of state:
$ \begin{equation} w_{\rm tot}=\frac{-2 \alpha \pm \sqrt{1-8 (\alpha -1) \alpha }+1}{6 \alpha }, \end{equation} $
(62) which is sensitive to the value of the geometrical coupling parameter α. We note that the values of the matter density parameter is zero, a feature where the geometrical–matter coupling plays a fundamental role in the effective dynamics at the background level. Subsequently, we briefly describe
$ \mathcal{U}_2^{+} $ solution. The dynamics in a specific region for various values of the coupling parameter α is represented in Fig. 9. As shown, this critical point can explain various stages in the history of our Universe, the super–acceleration, radiation$ (\alpha =\dfrac{2}{3}) $ , and matter domination$ (\alpha =1) $ via fine–tuning. The accelerated expansion ($w_{\rm tot} < -\dfrac{1}{3}$ ) is obtained in the following interval:$ (\dfrac{1}{4} \left(2-\sqrt{6}\right)\leq \alpha <0) $ . Based on a dynamical perspective for the$ \mathcal{U}_2^{+} $ solution, we obtain the following eigenvalues:$ \begin{aligned}\\[-10pt] \Bigg[ \frac{2 (3 \alpha - 2) \left(8 \alpha ^2 + 4 \left( \sqrt{-8 \alpha ^2 + 8 \alpha + 1} - 2\right) \alpha - \sqrt{-8 \alpha ^2 + 8 \alpha + 1} - 1\right)}{\alpha \left(\sqrt{-8 \alpha ^2+8 \alpha +1}-4 \alpha +1\right)^2}, \frac{2 (2 \alpha - 1) \left(24 \alpha ^3 - 6 \left( \sqrt{-8 \alpha ^2 + 8 \alpha + 1} + 3\right) \alpha ^2 + \sqrt{-8 \alpha ^2 + 8 \alpha + 1} + 4 \alpha + 1\right)}{\alpha ^2 \left(\sqrt{-8 \alpha ^2+8 \alpha +1}-4 \alpha +1\right)^2} \Bigg].\end{aligned} $ (63) Considering the analytical analysis of the previous eigenvalues,
$ \mathcal{U}_2^{+} $ solution is a stable node (with real and negative eigenvalues) in the following interval:$ \frac{1}{4} \left(2-\sqrt{6}\right)<\alpha <0 $ . Finally, we observe that the latter solution can explain a variety of cosmological eras, with a high sensitivity to the values of the geometrical–matter coupling embedded into the α coefficient. For the$ \mathcal{U}_2^{-} $ solution, we have a similar behavior. It should be noted that$ \mathcal{U}_2^{-} $ cannot explain matter and radiation epochs, while the accelerated expansion phenomenon is favored. -
In this section, we consider the following subclass of functions defined in the following manner:
$ \begin{equation} f(R,\phi)=\frac{R}{2}+g_0 {\rm e}^{\alpha \phi}, \end{equation} $
(64) where
$ f_0 $ and α denote constant parameters. As noted, the action is based on the fundamental Einstein–Hilbert action, extended with an exponential representation for the matter–geometry component. For this specific model, we introduce the following auxiliary variables:$ \begin{equation} s=\frac{\rho_m}{3 H^2}, \end{equation} $
(65) $ \begin{equation} x=\frac{g(\phi)}{3 H^2}, \end{equation} $
(66) $ \begin{equation} y=H^2 \dot{H}, \end{equation} $
(67) $ \begin{equation} z=\frac{1}{H^2}. \end{equation} $
(68) Moreover, we add the following non–independent variable:
$ \begin{equation} \mathcal{L}=H \dot{\rho_m}, \end{equation} $
(69) This can be used to form the autonomous system. In these variables, the Friedmann constraint Eq. (13) reduces to the following relation:
$ \begin{equation} 1+x+s(-1+18 \alpha x y)=0. \end{equation} $
(70) Then, the autonomous system becomes:
$ \begin{equation} \frac{{\rm d}s}{{\rm d}N}=\frac{z^2 \mathcal{L}}{3}-2 s y z^2, \end{equation} $
(71) $ \begin{equation} \frac{{\rm d}x}{{\rm d}N}=18 \alpha s x y-2 x y z^2+3 \alpha x \mathcal{L}, \end{equation} $
(72) $ \begin{equation} \frac{{\rm d}z}{{\rm d}N}=-2 y z^3, \end{equation} $
(73) while the acceleration Eq. (14) is equal to:
$ \begin{aligned}[b] &\frac{3 x \left(108 \alpha ^2 s^2 y+6 \alpha s \left(y z^2+3 \alpha \mathcal{L}+3\right)+z^2 (2 \alpha \mathcal{L}-1)\right)}{z^3}\\=&2 y z+\frac{3}{z} \end{aligned} $
(74) Given the specific form of the acceleration Eq. (14), we can extract the
$ \mathcal{L} $ variable, remaining with a three-dimensional dynamical system$ \{s,x,z\} $ . As in the previous cases, we determine the associated critical points by analyzing the right hand side of the autonomous system of differential equations. For the last physical model, where the coupling function is represented by an exponential, we identified only one class of cosmological solutions located in the phase space structure at the following coordinates:$ \begin{equation} \Xi_1=\Big[s=\frac{18 \alpha +z^2}{18 \alpha }, x=\frac{z^2}{18 \alpha } \Big]. \end{equation} $
(75) The critical line corresponds to a de–Sitter solution
$(w_{\rm tot}=-1)$ , describing an epoch where the dynamics induced by the interplay between geometry and matter mimics a cosmological constant behavior. In Fig. 10, we plot a specific region in the$ \{z, \alpha \} $ space, where the critical line$ \Xi_1 $ is viable considering the existence conditions.Figure 10. (color online) Specific region where the de–Sitter critical point
$ \Xi_1 $ is viable, with the matter density parameter in the$ [0, 1] $ interval.At this cosmological solution, we obtained the following eigenvalues,
$ \begin{equation} \Bigg[0,-\frac{2268 \alpha ^3+11 \alpha z^4+360 \alpha ^2 z^2 \pm \sqrt{\alpha ^2 \left(104976 \alpha ^4-23 z^8-720 \alpha z^6-7128 \alpha ^2 z^4-46656 \alpha ^3 z^2\right)}}{648 \alpha ^3+6 \alpha z^4+144 \alpha ^2 z^2} \Bigg], \end{equation} $ (76) describing a non–hyperbolic solution, which can saddle in some specific intervals. For example, if consider the following restrictions,
$ \begin{equation} z\in \mathbb{R}\land z\neq 0\land -\frac{z^2}{6}<\alpha <-\frac{z^2}{18}, \end{equation} $
(77) then we obtain a de–Siter epoch with a saddle behavior, where the dynamics corresponds to a cosmological constant added in the Einstein field equations.
Physical aspects of ${ \boldsymbol{f}(\boldsymbol R,\boldsymbol G_{\boldsymbol\mu \boldsymbol\nu}\boldsymbol T^{\boldsymbol\mu \boldsymbol\nu})} $ modified gravity theories
- Received Date: 2023-04-07
- Available Online: 2023-10-15
Abstract: The paper extends basic Einstein–Hilbert action by incorporating an invariant derived from a specific contraction between the Einstein tensor and energy momentum tensor. This represents a non–minimal coupling between the space–time geometry and matter fields. The fundamental Einstein–Hilbert action is extended by considering a generic function