Modified Gravity Theories on a Nutshell: Inflation, Bounce and Late-time Evolution

I. INTRODUCTION

• The review surveys modified gravity as a framework for describing inflation, late-time acceleration, and bouncing cosmology, while assessing compatibility with observations and astrophysical constraints.It aims to provide an accessible but technically useful account of metric modified gravity, including unified descriptions of early- and late-time acceleration and selected astrophysical applications.Source: §I · pp. 3–6
• The introduction motivates bouncing cosmology as a nonsingular alternative to Big Bang evolution and notes that modified gravity can realize bounces without some constraints required in general relativity.It discusses cyclic contraction and expansion, perturbation generation, observational signatures, and representative matter-bounce and quintom scenarios.Source: §I · pp. 3–5
• The review organizes its technical treatment around F(R), Gauss-Bonnet, string-inspired, F(T), massive, bigravity, mimetic, and unimodular modified-gravity theories, followed by inflation, dark energy, astrophysical, and bounce analyses.The planned analyses include observational indices, reconstruction methods, unified acceleration histories, compact-object solutions, and dynamical-system stability of reconstructed bounces.Source: §I · pp. 6–7

A. F(R) Gravity

• F(R) gravity replaces the Einstein-Hilbert scalar curvature R with a function F(R), and its metric variation yields modified gravitational field equations.The section formulates the theory using the matter Lagrangian and develops its equations for spatially flat FRW geometry.Source: A. F(R) Gravity · p. 7, §1 · p. 8

1. General properties

• For flat FRW spacetime, F(R) gravity is expressed through the Hubble rate H, scalar curvature R = 12H^2 + 6Ḣ, and an effective equation-of-state description applicable to modified gravity.The formalism also admits algebraic constant-curvature solutions whose vacuum geometries include de Sitter, anti-de Sitter, Schwarzschild-(anti-)de Sitter, and Kerr-(anti-)de Sitter spaces.Source: §1 · p. 8
• F(R) gravity can be recast using effective density and pressure, but this fluid picture must retain higher-derivative curvature contributions; viable dark-energy and unified models are reviewed [290] [291].Source: §1 · p. 9

2. Scalar-tensor description

• The scalar-tensor reformulation introduces an auxiliary field and conformal transformation, but Jordan- and Einstein-frame descriptions can differ in acceleration, singularity type, and effective equation of state [295] [296].The Einstein-frame metric is not the physical matter metric; e^σgµν corresponds to the Jordan-frame metric.Source: §2 · pp. 9, 11
• The chameleon mechanism suppresses fifth-force corrections by making the scalar heavy in high-curvature environments, although an excessively short Compton wavelength invalidates the continuous-matter approximation.The highlighted limitation concerns interatomic vacuum regions, where the Compton length cannot be shorter than atomic separations [297].Source: §2 · p. 10
• Viable models require F′(R) > 0 to exclude antigravity regions, where the graviton becomes a ghost according to the cited field-theory interpretation.Source: §2 · p. 10

3. Viable F(R) gravities

• A viable unified F(R) model must recover Einstein or Minkowski behavior at low curvature, provide quasi-stable inflationary and stable late-time de Sitter solutions, and avoid antigravity and curvature singularities [95] [97] [98] [99] [304] [305].The inflationary de Sitter phase must slowly decay, while the late-time de Sitter solution supplies acceleration.Source: §3 · pp. 11–13
• Adding an R^2 term can eliminate finite-time future singularities while simultaneously generating early-time acceleration; terms behaving as R^n with n > 1 instead permit Big Rip behavior.For f(R) ∼ R^2 no singular solution appears in the analyzed class, whereas 0 < ε < 2 avoids future singularities [323].Source: §3 · pp. 12–13

6. By combining conditions (36) and (41), we find

• F(R) viability requires F′(R) > 0 and F(R) > 0, while stable de Sitter evolution prevents the curvature from reaching an additional unstable de Sitter solution.The reviewed F(R) conditions also include models constrained by large-scale-structure observations.Source: §6 · pp. 13–14
• F(R,T) gravity generally violates conservation of Tµν and requires additional constraints, while non-minimal matter couplings conserve only an effective tensor ˜Tµν.Solvability often depends on imposing further assumptions.Source: §6 · pp. 14–15

1. General properties

• For β < 0, the effective phantom phase is transient because the Einstein term dominates near the would-be Big Rip and removes the singularity.The model can realize weff < −1 even when the matter equation of state satisfies w > −1.Source: §1 · p. 16
• For 0 < β < 1/2, modified Gauss–Bonnet gravity can evolve from deceleration to asymptotic de Sitter acceleration and can unify inflation with late-time acceleration.The f(G) contribution becomes dominant at late times after matter dilution.Source: §1 · p. 16
• Modified Gauss–Bonnet gravity produces no Newton-law correction around de Sitter or flat backgrounds, unlike the significant solar-system corrections found for 1/R gravity.The absence of the correction is attributed to the topological structure of the Gauss–Bonnet invariant.Source: §1 · p. 16
• Introducing auxiliary scalar fields rewrites f(G) gravity as scalar-Einstein-Gauss–Bonnet gravity, whose kinetic extension is described as ghost-free and stable in general.The scalar field is nondynamical before adding a kinetic term, and the ε → 0 limit is smooth.Source: §1 · pp. 16–17

C. String-inspired Gravity

• String-inspired gravity couples dilaton fields to curvature corrections, with coefficients and flux parameters depending on the underlying string theory.The reviewed action leaves the scalar potential and Gauss–Bonnet coupling unspecified, allowing them to arise from non-perturbative string effects.Source: C. String-inspired Gravity · p. 17
• Scalar-Einstein-Gauss–Bonnet models admit de Sitter solutions, can match the current Hubble rate for suitable parameters, and can realize effective phantom behavior unavailable to a canonical scalar without the Gauss–Bonnet term.The parameter h0 can be adjusted through V0 and ξ0, including negative values corresponding to weff < −1.Source: C. String-inspired Gravity · p. 18

D. F(T ) Gravity

• F(T) gravity uses the torsion scalar T, defined through the Weitzenböck connection, instead of curvature R, extending the teleparallel Lagrangian to model inflation and late-time acceleration [146] [363] [366] [367] [368] [369] [370] [371] [372] [373] [374] [375] [376] [377] [378] [379] [380] [381] [382] [383] [384] [385] [386] [387] [388] [389] [390] [391] [392].Source: D. F(T ) Gravity · p. 19
• The standard F(T) formulation is frame dependent because vanishing spin connection restricts local Lorentz invariance; a covariant formulation requires both vierbein and spin connection [361] [393].The Lorentz-violation problem can complicate spherically symmetric solutions and prevents consistent coupling to spinor fields in the standard formulation.Source: D. F(T ) Gravity · p. 19
• For FRW spacetime, the absence of local Lorentz invariance is treated as unimportant, so the review uses the simpler standard formulation with T = −6H2.Source: D. F(T ) Gravity · p. 19
• In vacuum, the consistent F(T) solution is uniquely constrained, whereas the form given by Eq. (90) satisfies the cosmological equations for any H and therefore realizes arbitrary cosmic evolution.The review also identifies F(T) applications to late-time acceleration, astrophysical and metric solutions, inflation, bouncing cosmology, and perturbation evolution [361, 368, 369, 373, 374, 375, 379, 381, 385, 388, 390, 391, 395, 402–405, 406–411].Source: D. F(T ) Gravity · pp. 19–20

E. Massive F(R) Gravity and Massive Bigravity

• The massive F(R) bigravity construction can be obtained by algebraically solving for the auxiliary scalars and substituting them into the action.Because the scalar equations are difficult to solve explicitly, defining the model directly in terms of auxiliary scalars may be easier [424].Source: E. Massive F(R) Gravity and Massive… · pp. 25–26
• A nonlinear massive F(R) theory with a non-dynamical reference metric fµν was shown to be ghost-free using the Hamiltonian formulation in the scalar-tensor frame.The model is proposed in Ref. [425].Source: E. Massive F(R) Gravity and Massive… · p. 26

F. Mimetic F(R) gravity

• Mimetic gravity rewrites the metric using new degrees of freedom, producing a wider solution class that effectively induces dark matter.The mimetic parametrization yields equations beyond those obtained by varying the metric alone.Source: F. Mimetic F(R) gravity · pp. 26–27
• Mimetic F(R) gravity can realize early- and late-time acceleration, inflation compatible with observations, ΛCDM reconstruction, unified acceleration, and cosmological bounces.The review also notes extensions with scalar potentials and Lagrange multipliers.Source: F. Mimetic F(R) gravity · p. 27
• Standard F(R) solutions remain solutions of mimetic F(R) gravity, while nonzero Cφ introduces distinct FRW equations and corresponding solutions.For Cφ = 0 the FRW equations coincide with standard F(R); Cφ ≠ 0 changes them, with the Cφ term behaving as a^-3.Source: F. Mimetic F(R) gravity · p. 27
• A Lagrange-multiplier formulation adds the mimetic constraint and potential, allowing arbitrary cosmological evolution in principle while expecting standard F(R)-like ghost and Newton-law behavior.The claimed ghost-free property is explicitly stated to require verification with a Hamiltonian analysis.Source: F. Mimetic F(R) gravity · p. 28

G. Mimetic f(G) Gravity

• For C = 0, the mimetic f(G) equations reduce to ordinary f(G) gravity, whereas comparison of the resulting expressions identifies differences between the mimetic and non-mimetic theories.The framework also admits Lagrange-multiplier and mimetic-potential generalizations.Source: G. Mimetic f(G) Gravity · pp. 29–30
• Mimetic f(G) gravity reconstructs f(G) from a chosen Hubble history by solving for g(t), converting to g(G), and integrating with respect to the Gauss-Bonnet scalar.The reconstruction can also proceed inversely from a given f(G), potential, or cosmic evolution.Source: G. Mimetic f(G) Gravity · pp. 30–31

H. Unimodular F(R) Gravity

• The non-covariant unimodular F(R) formulation recovers the standard Poisson equation and Newtonian potential after redefining the gravitational constant.The derivation uses a point source, restricted gauge freedom, and the unimodular constraint.Source: H. Unimodular F(R) Gravity · p. 32
• The non-covariant unimodular constraint eliminates the propagating scalar mode, so no correction to Newton’s law occurs.This differs from standard F(R), where the scalar mode produces a nontrivial correction.Source: H. Unimodular F(R) Gravity · p. 32
• The non-covariant unimodular F(R) theory is not fully covariant because its metric determinant is constrained.A covariant formulation introduces a three-form field so that no direct constraint is imposed on the metric.Source: H. Unimodular F(R) Gravity · pp. 32–33
• The covariant formulation introduces an additional field sector whose Hamiltonian has an infinite, unbounded-below eigenvalue, although no quantum transitions between the states occur.The review therefore discusses quantum stability while noting the problematic Hamiltonian spectrum.Source: H. Unimodular F(R) Gravity · p. 33
• In covariant unimodular F(R) gravity, the scalar potential is shifted by a constant multiplier contribution, but the scalar mass and Newton-law corrections are essentially those of standard F(R).The review contrasts this with the non-covariant version, which recovers the standard Newton law.Source: H. Unimodular F(R) Gravity · p. 34

I. Unimodular Mimetic Gravity

• Unimodular mimetic gravity imposes both mimetic and unimodular constraints through Lagrange multipliers, with the FRW time coordinate redefined by dτ = a(t)^3dt.The auxiliary scalar can then be identified with cosmic time under the mimetic constraint.Source: §I · pp. 34–35
• The unimodular-mimetic equations form a reconstruction method: specifying a(t), energy density, and pressure determines the potential and multiplier functions realizing that cosmology.The review identifies Eqs. (216), (213), and (212) as the reconstruction system.Source: §I · pp. 35–36

J. Unimodular Mimetic F(R) Gravity

• Unimodular mimetic F(R) gravity extends the reconstruction with F(R), a mimetic potential, and two multiplier functions, allowing the potential, λ(t), and η(t) to be obtained for an arbitrary scale factor.The construction uses the FRW equations and the general solution of the differential equation for the potential.Source: J. Unimodular Mimetic F(R) Gravity · pp. 36–37

III. INFLATIONARY DYNAMICS IN MODIFIED GRAVITY

• Inflation addresses Big Bang flatness, monopole, entropy, and horizon problems through a period of nearly exponential early-time expansion.The horizon problem is described as the absence of past causal contact between regions that are presently far apart but nearly homogeneous.Source: §III · p. 37
• Inflation occurs when the Hubble radius decreases, with Ṙ_H < 0 equivalent to accelerated expansion, and the review evaluates models through the scalar spectral index and tensor-to-scalar ratio.The discussion assumes vacuum modified gravity for simplicity and does not develop perturbation theory in detail.Source: §III · p. 38

1. Canonical Scalar Field Inflation

• Canonical scalar inflation requires ε1, ε2 ≪ 1, predicts n_s and r from slow-roll parameters, and ends when the first slow-roll index becomes O(1).For Starobinsky or Higgs inflation, the review gives horizon-crossing and leading-N observational expressions.Source: §1 · pp. 39–40
• The Hubble slow-roll expansion can terminate inflation earlier than the lowest-order approximation when higher-order terms become large or singular.For V(ϕ) ∼ ϕ^2, inverse powers of ϕ can cause the expansion to break down before ε reaches O(1).Source: §1 · pp. 40–41

2. Non-Canonical Scalar Field Inflation

• For ω(φ) > 0, a non-canonical scalar can be mapped to a canonical theory only when the defining field redefinition is invertible.If the integral is not invertible, the slow-roll indices must instead be expressed directly through ω(φ) and V(φ).Source: §2 · p. 41
• For ω(φ) < 0, the scalar is phantom, with an unbounded classical energy density and negative-norm states in the quantum theory.The review therefore characterizes ghost-driven phantom inflation as physically unattractive.Source: §2 · p. 42

3. Multi-Scalar Field Inflation

• Multi-scalar inflation generalizes slow-roll dynamics to a curved scalar configuration space with adiabatic and iso-curvature directions defined by the field-space metric.The formalism introduces a turn-rate vector and projects the potential gradient along the adiabatic direction.Source: §3 · p. 42

B. Inflation in f(ϕ, R) Theories of Gravity

• The general f(ϕ,R) action contains F(R), canonical and non-canonical scalar theories, and f(φ)R models as subcases.Its slow-roll equations and scalar-to-tensor ratio are organized through the auxiliary function Qs.Source: B. Inflation in f(ϕ, R) Theories of… · p. 43
• In the canonical-scalar limit, the formalism recovers the standard scalar-to-tensor relation r ≃16ε1.The same framework is used to derive the spectral index under the slow-roll conditions εi ≪ 1.Source: B. Inflation in f(ϕ, R) Theories of… · pp. 43–44

1. Inflation in f(φ)R Theories of Gravity

• For f(φ)R gravity, the slow-roll formalism expresses the spectral index and scalar-to-tensor ratio through generalized slow-roll indices for any analytic coupling function f(φ).A symmetric f(φ) example produces observational-index expressions identical to α-attractor models [527] [528] [529] [530] [531] [532] [533] [534] [535] [536] [537], with β ≪1 there versus β >1 here.Source: §1 · pp. 44–46

2. Inflation from F(R) Theories

• F(R) inflation yields the same leading-order observational indices in the Jordan and Einstein frames, while the R2 term drives graceful exit through an unstable de Sitter point.The final attractor is a de Sitter vacuum whose perturbations grow, producing the exit from inflation.Source: §2 · p. 48

3. Inflation in mimetic-F(R) Theories of Gravity

• Mimetic-F(R) gravity can realize observationally viable inflationary scenarios, although the mimetic R2 model is incompatible while the power-law model can fit data for suitable parameters.The framework uses slow-roll indices and observational constraints from Planck [48] and BICEP2/Keck Array [49].Source: §3 · p. 49
• Mimetic power-law F(R) inflation can produce observationally compatible results, whereas the ordinary power-law model is not compatible with observations.For suitable parameters, compatibility is achieved for the mimetic model, while the ordinary model can make only the spectral index compatible.Source: §3 · pp. 50–51
• Compatibility with current observational data is obtained for N = 50, H0 = 12.04566, M = 2, and n = 1.8, although achieving it requires fine-tuning the parameters.Further details on allowed parameter ranges are given in [142].Source: §3 · pp. 51–52
• A second mimetic power-law example also yields observationally compatible indices for N = 50, c = 0.014, d = −0.1, n = 11, and H0 = 11.The parameter compatibility is again reported alongside a fine-tuning caveat and reference [142].Source: §3 · p. 52
• Mimetic F(R) gravity makes it very difficult to address the graceful-exit problem using the methods described in this chapter.A separate discussion is devoted to this issue later in the chapter.Source: §3 · p. 52

4. Reconstruction of F(R) gravity from Einstein Frame Scalar Potential

• A canonical Einstein-frame potential determines the corresponding Jordan-frame F(R) theory through the canonical transformation and the resulting differential equation.Solving the transformation for the curvature and combining it with the potential reconstruction yields F(R).Source: §4 · p. 52
• For the Starobinsky potential, the reconstruction gives F(R) = R + R2/(8μ2).The review also notes that several attractor potentials asymptotically approach the potential used in the reconstruction.Source: §4 · p. 53

1. Explicit Calculation of the Power Spectrum

• The f(G) reconstruction introduces an auxiliary scalar identified with cosmic time and functions P(t), Q(t), then solves for t(G) to obtain the gravity model realizing a chosen evolution.For intermediate inflation, this procedure is applied at early times and simplified in the small-G limit.Source: §1 · pp. 54–55
• In f(G) gravity, scalar perturbations contain only k2 propagating terms, so the evolution has no k4 scalar modes or associated superluminal propagation.The propagation speed is determined by B1(t).Source: §1 · p. 56
• The reconstructed f(G) realization of intermediate inflation produces a power spectrum that is not scale invariant and is incompatible with observational data.Although ns can agree for α = 13.9 or α = −12, neither value is allowed by the intermediate-inflation scenario.Source: §1 · pp. 57–59

• This conservation question is important because comoving curvature perturbations do not remain conserved in some scenarios, including the matter bounce scenario [61].Source: §2 · p. 59
• The calculation focuses on superhorizon modes satisfying k ≪ a(t)H(t), allowing the perturbation equation’s final term to be neglected.Source: §2 · p. 59
• The function Q(t) and the governing f(G) model determine the post-horizon-crossing evolution through an integral expression.Source: §2 · p. 59
• The f(G) analysis finds that the comoving curvature perturbation is conserved after horizon crossing because Φ(t) remains constant.The integral term governing its evolution eventually decays as cosmic time increases.Source: §2 · p. 59

D. Inflation from F(T ) = T + f(T ) Theories in the Jordan Frame

• For a canonical scalar field, the minimally coupled F(T) theory has Ψ = Φ and only one propagating degree of freedom.Source: D. Inflation from F(T ) = T + f(T )… · p. 60
• D. Inflation from F(T ) = T + f(T ) Theories in the Jordan FrameScalar perturbations are formulated through a differential equation whose structure matches the Einstein-Hilbert case except for the speed of sound parameter.Source: D. Inflation from F(T ) = T + f(T )… · pp. 60–61
• The F(T) perturbation framework uses the gauge-invariant comoving curvature fluctuation ζ and the variable v to calculate the primordial power spectrum and spectral index.The calculation specializes to vacuum F(T) gravity realizing intermediate inflation, with the gravitational function entering implicitly through the sound speed.Source: D. Inflation from F(T ) = T + f(T )… · p. 61
• ns = 0.9644 for n = 0.670576, identical with the 2015 Planck constraint on the spectral index, although the power spectrum is not scale invariant.Source: D. Inflation from F(T ) = T + f(T )… · p. 63

E. Singular Inflation

• E. Singular InflationA Type IV singularity is incorporated into the R2 model by modifying its quasi-de Sitter Hubble rate, requiring α > 1 and additional parameter hierarchies.Source: E. Singular Inflation · p. 63

1. Scenario I

• 1. Scenario IFor 2 < α < 3, the singularity in ǫ4 breaks the slow-roll expansion at higher order and may provide a dynamical mechanism for graceful exit; the resulting ns and r match the R2 model.Source: §1 · p. 64

2. Scenario II

• The parameters ǫ3, ǫ4, and ηH diverge at t = ts, indicating a strong instability at that time and causing inflation to end abruptly at that point.Source: §2 · p. 65

3. Scenario III

• For α > 3 with ts < tf, ǫ3 and ǫ4 remain finite, so inflation ends when ǫ1 becomes order one, with behavior differing from ordinary R2 gravity.The second slow-roll parameter is nonzero in this scenario, and its spectral-index predictions are observationally close to Starobinsky’s model [510].Source: §3 · p. 65

4. The Scenarios with ts = tf

• When ts = tf, divergences in higher slow-roll indices can make the graceful exit more pronounced, whereas α > 3 yields behavior indistinguishable from the ordinary R2 model.The Type IV singular R2 model can also support late-time acceleration, although its early-time singular-Hubble description does not apply at late times [510].Source: §4 · p. 65

F. A Qualitative Study for the Graceful Exit from Inflation Issue

• The graceful-exit formalism uses unstable de Sitter vacua whose growing curvature perturbations trigger the end of inflation, and it tests instability through linearized Hubble perturbations.The method solves for de Sitter solutions, perturbs H(t) = Hd + ∆H(t), and analyzes the resulting differential equation.Source: F. A Qualitative Study for the… · p. 66
• The perturbation analysis is quantitatively general but model dependent; for H0 = 12.04566 and M = 2, ∆H is order one unless c1 and c2 are extremely small, preventing sufficient inflation.The stated example gives ∆H ∼ −0.5 + 0.9c1 + 1.0c2.Source: F. A Qualitative Study for the… · p. 67

G. Reheating in F(R) Gravity

• In F(R) reheating, curvature oscillations act as matter sources and back-react on the metric, avoiding the need to fine-tune inflaton couplings to Standard Model fields.The reheating phase begins when slow-roll conditions fail and proceeds through an oscillatory curvature regime.Source: G. Reheating in F(R) Gravity · pp. 68–69
• After reheating, the matter sector approaches radiation domination, while the reheating temperature constrains Hi through baryogenesis and monopole-avoidance requirements.The parameter Hi also affects inflationary observables, reheating temperature, and subsequent Friedmann evolution.Source: G. Reheating in F(R) Gravity · pp. 68–70

A. ΛCDM Epoch from F(R) Gravity

• F(R) gravity can realize ΛCDM expansion by reconstructing the function from H2(N), reducing the reconstruction equation to a solvable hypergeometric differential equation.The same realization was obtained in Ref. [572].Source: A. ΛCDM Epoch from F(R) Gravity · p. 71

B. ΛCDM Epoch from Modified Gauss-Bonnet Gravity

• The modified Gauss-Bonnet construction reproduces the ΛCDM universe by reconstructing the auxiliary-field action and choosing functions that yield the desired cosmological solution.The auxiliary scalar can be identified with cosmic time or e-foldings, enabling reconstruction of H = h(N).Source: B. ΛCDM Epoch from Modified… · pp. 72–73

C. Unification of Inflation with Dark Energy Era in F(R) Gravity

• The unified F(R) model generates a quasi-stable inflationary de Sitter phase that gradually becomes unstable as curvature decreases, providing an exit from inflation.The inflationary phase occurs near R = RI and is stable at large curvature but destabilizes during curvature decrease.Source: C. Unification of Inflation with Dark… · pp. 73, 75
• The model reproduces the required small-curvature Einstein-gravity behavior and avoids curvature singularities in the large-curvature limit.Its large-curvature behavior satisfies the condition preventing curvature singularities, while its R → 0 limit satisfies the Einstein-gravity requirement.Source: C. Unification of Inflation with Dark… · pp. 74, 76
• The model avoids antigravity across the considered curvature range and prevents matter instability because U(Rb) is negative.Avoiding antigravity also prevents the associated ghost-like graviton problem described by the viability conditions.Source: C. Unification of Inflation with Dark… · pp. 74, 76
• The Chameleon mechanism makes the scalar sufficiently massive that its Compton length is very small and corrections to Newtonian gravity are practically negligible.A sufficiently large scalar mass suppresses long-range forces and helps maintain compatibility with solar-system observations, although excessive mass may conflict with matter abundance.Source: C. Unification of Inflation with Dark… · pp. 74, 76
• After inflation, reheating can produce radiation and matter domination, followed by late-time acceleration when the curvature approaches R0.During radiation and matter domination, the cosmological evolution can become identical to Einstein gravity before acceleration is recovered.Source: C. Unification of Inflation with Dark… · pp. 75–77
• The unified model can also support a second inflationary era at very high curvature when an R2 term is generated.This double-inflation possibility requires the Universe to begin with R ≫ RI.Source: C. Unification of Inflation with Dark… · p. 77
• The inflationary de Sitter solution is strongly unstable, ensuring a graceful exit from inflation [576].For suitable parameters, the model ultimately approaches R = 0 or may encounter a curvature singularity unless quasi-stability is imposed.Source: C. Unification of Inflation with Dark… · p. 78
• The model yields N ≃107 for α = 5/2, N ≃64 for α = 8/3, and N ≃51 for α = 11/4 [576].These estimates quantify viable inflationary durations for the examined parameter choices.Source: C. Unification of Inflation with Dark… · p. 78
• At late times, the de Sitter vacuum has R = 4Λ, is stable, and becomes the model’s final attractor.The theory also avoids antigravity and reproduces the matter-dominated era when F(R) ≃ R.Source: C. Unification of Inflation with Dark… · p. 79
• The scalar mass can suppress Newton-law corrections, but an excessively large mass may overproduce reheating particles and invalidate the chameleon mechanism.In the solar-system and terrestrial-atmosphere regimes, the inferred masses are extremely large, motivating an upper-limit constraint.Source: C. Unification of Inflation with Dark… · p. 79
• Numerical evolution remains close to ΛCDM at large redshift, with present wDE = −0.994, −0.975, and −0.950 for R0 = 0.6Λ, 0.8Λ, and Λ, respectively [577].As R0 decreases, the model becomes indistinguishable from ΛCDM with wDE = −1; phantom-divide crossing causes no serious phenomenological issue.Source: C. Unification of Inflation with Dark… · pp. 80–81
• Dark-energy density oscillations during matter domination can acquire divergent frequencies, making perturbative analyses inapplicable through nonlinear scalar-curvature oscillations [576].These modes arise because F(R) equations contain fourth-order derivatives, with frequencies controlled by the scalar mass.Source: C. Unification of Inflation with Dark… · p. 81

D. Unification of Inflation with Dark Energy Era in Modified Gauss-Bonnet Gravity

• Choosing H_e ≫ H_l yields early and late constant-Hubble regimes that realize inflation and late-time acceleration in modified Gauss-Bonnet gravity.The construction uses a Hubble rate approaching H_e as N→−∞ and H_l as N→+∞.Source: D. Unification of Inflation with Dark… · p. 81
• The Gauss-Bonnet formalism can achieve unified early- and late-time acceleration without detailing the construction beyond the preceding framework.Source: D. Unification of Inflation with Dark… · p. 81

E. Phantom Dark Energy Era

• F(R) reconstruction introduces auxiliary functions P(φ) and Q(φ) to determine a gravity model realizing a specified cosmological evolution.The method derives P(φ), obtains Q(φ), and then constructs the resulting F(R) theory.Source: E. Phantom Dark Energy Era · p. 82
• The reconstructed vacuum F(R) model crosses the phantom divide: weff > −1 before t_s and weff = −16/15 < −1 afterward, while a Big Rip occurs at t_s.The analysis therefore restricts attention to the era 0 < t < t_s before the singularity.Source: E. Phantom Dark Energy Era · p. 83

1. Dark Energy Oscillations

• The oscillation analysis rewrites the F(R) equations into a redshift-based master differential equation and solves it numerically for collisional and non-collisional matter.The model is modified exponential gravity with Λ = 7.93 ˜m2, ˜γ = 1/1000, ΩM = 0.279, and w = 0.2 or 0.8.Source: §1 · p. 86
• Increasing the matter parameter w makes the dark-energy oscillations more pronounced, with collisional and non-collisional cases compared through ωDE(z).The comparison uses the dark-energy equation-of-state parameter for w = 0.2 and w = 0.8.Source: §1 · pp. 86–87, Figure 2 · p. 87
• For w = 0.8, the Hubble-rate behavior is the same for collisional and non-collisional matter, and the scalar-curvature behavior is likewise reported as the same.The Hubble-rate comparison is shown in Fig. 3; the scalar curvature is discussed alongside it.Source: §1 · p. 87, Figure 3 · p. 87
• The total equation of state never crosses the phantom divide, whereas the dark-energy equation of state oscillates strongly around w = −1.These behaviors are compared for collisional and non-collisional matter across w = 0.2 and w = 0.8.Source: §1 · p. 88, Figure 4 · p. 88

2. Growth Index Evolution

• The growth-index calculation uses subhorizon matter perturbations, with Geff depending explicitly on wavenumber, and focuses on matter domination and later eras.The subhorizon approximation breaks down before matter domination, setting the stated scope of the perturbation analysis.Source: §2 · pp. 88–89
• Matter perturbations differ significantly between collisional and non-collisional cases even when their Hubble-rate evolutions are indistinguishable, making perturbations useful for distinguishing modified-gravity effects.The comparison uses the growth factor at k = 0.1Mpc−1 and w = 0.8.Source: §2 · p. 89

V. ASTROPHYSICAL APPLICATIONS

• The astrophysical applications are motivated by modified gravity’s ability to change the conditions required for compact objects and other solutions relative to Einstein gravity.The section covers neutron and quark stars, black holes, and anti-evaporation within F(R) gravity, while contrasting with massive gravity.Source: §V · p. 90, B. Black holes in F(R) Gravity and… · p. 93

A. Neutron and quark stars from F(R) gravity

• The neutron-star analysis solves modified field equations for static spherical stars using an equation of state, perturbative expansions, and model-dependent F(R) or massive-gravity assumptions.The treatment includes ordinary and quark-core equations of state, with transition densities specified for FPS and SLy models.Source: A. Neutron and quark stars from F(R)… · pp. 90–92
• F(R) gravity can generate stable neutron-star configurations with smaller radii and masses of approximately 2M⊙ for a SLy equation of state.A second stable high-density branch can fit observations better than the corresponding Einstein-gravity model.Source: A. Neutron and quark stars from F(R)… · p. 92
• Specific dRGT massive-gravity models produce more constrained mass-radius relations and smaller maximum masses, preventing compact massive neutron stars under the stated assumptions.The conclusion is limited to standard equations of state and a particular minimal massive-gravity model.Source: A. Neutron and quark stars from F(R)… · p. 93

B. Black holes in F(R) Gravity and anti-evaporation

• F(R) gravity can produce Nariai black-hole anti-evaporation, with horizon growth or oscillatory evaporation depending on perturbation parameters.The Nariai solution is unstable in the analyzed perturbative cases, and oscillation amplitudes can increase.Source: B. Black holes in F(R) Gravity and… · pp. 93–95

C. Wormholes in F(R) Gravity

• F(R) gravity can realize wormhole geometries while ordinary matter satisfies the null energy condition, shifting the required violation into restrictions on the gravitational function.The throat conditions constrain the functional form of F(R), and reconstruction proceeds from a chosen shape function and matter equation of state.Source: C. Wormholes in F(R) Gravity · pp. 95–97
• Wormhole reconstruction specifies the shape function and equation of state, derives R(r) and F_R(r), and integrates to obtain F(R).The review presents an example based on a prescribed relation among energy density and pressures.Source: C. Wormholes in F(R) Gravity · p. 97

A. General Features of Bouncing Cosmologies

• A cosmological bounce contracts to a nonzero minimum scale factor, crosses H(t)=0, and then expands, thereby avoiding an initial singularity.The standard conditions are H<0 before the bounce, H=0 at it, and H>0 afterward.Source: A. General Features of Bouncing… · pp. 97–98
• In the matter-bounce scenario, the Hubble rate changes sign at t=0 while the Hubble radius decreases during contraction and expands after the bounce.Primordial modes are generated near the beginning of contraction and later become relevant as the horizon evolves.Source: A. General Features of Bouncing… · p. 98
• The singular-bounce scenario contains a Type IV singularity and does not yield a scale-invariant near-bounce power spectrum without an additional scenario.Its Hubble radius diverges near the bouncing point, imposing a viability limitation.Source: A. General Features of Bouncing… · p. 99

B. Bounce Cosmology from F(R) Gravity

• Using an e-folding reconstruction, vacuum F(R) gravity can generate the superbounce exactly, while the matter-inclusive solution approaches R+αR^2 at early times and R−R^-1/2+Λ at late times.The early-time form resembles Starobinsky inflation and the late-time form resembles ΛCDM.Source: B. Bounce Cosmology from F(R) Gravity · pp. 99–101

1. Stability in the F(R) reconstructions

• The vacuum and matter-supported F(R) superbounce reconstructions are reported as stable under the linear perturbation analysis.The large-c limit gives J1 = 4c2 + 6 > 0 and J2 = 3c4 + 18c2 −12 > 0 for the stated stability conditions.Source: §1 · p. 102

C. Bounce Cosmology from f(G) Gravity

• The reconstruction method yields two f(G) gravities that realize the superbounce cosmology.The two solutions are subsequently found to be conditionally stable under the adopted linear-stability analysis.Source: C. Bounce Cosmology from f(G) Gravity · p. 103, §1 · p. 104

D. Bounce Cosmology from F(T ) Gravity

• F(T) reconstruction produces an exact gravity realizing the superbounce, including vacuum and matter-fluid solutions.For c = 3(1 + wi), including c = 3 for dust, the superbounce can also be realized by the teleparallel equivalent of General Relativity.Source: D. Bounce Cosmology from F(T ) Gravity · p. 105
• Superbounce cosmologies can be realized in F(R), f(G), and F(T) gravity, with stability analyzed for the corresponding dynamical systems.The review explicitly presents these three modified-gravity realizations and their stability analysis.Source: D. Bounce Cosmology from F(T ) Gravity · p. 105
• Reheating remains insufficiently addressed in modified-gravity bouncing cosmologies and is identified as a topic requiring focused study.The review notes that only limited serious treatment exists and places a comprehensive analysis beyond its scope.Source: D. Bounce Cosmology from F(T ) Gravity · p. 105

VII. CONCLUSION

• The review presents a pedagogical modified-gravity toolbox spanning inflation, dark energy, bouncing cosmology, and multiple theoretical frameworks.It emphasizes self-contained formalism and aims to serve both non-experts and experts while acknowledging that the field is too broad for complete coverage.Source: §VII · p. 105
• Modified gravity can unify inflation and dark energy descriptions, but the review stresses that many theoretical and observational challenges remain.The discussion includes unresolved dark-matter and dark-energy problems, pre-inflationary bounce possibilities, future singularities, and the need for further observations.Source: §VII · pp. 105–106
• Different modified-gravity descriptions of the same pre-inflationary bounce can produce different perturbation spectra.The review contrasts a non-scale-invariant F(R) spectrum with a nearly scale-invariant f(G) spectrum for a singular bounce.Source: §VII · p. 106