Cosmological constant and renormalization of gravity

In arXiv:1601.02203 and arXiv:1702.07063, we have proposed a topological model with a simple Lagrangian density and have tried to solve one of the cosmological constant problems. The Lagrangian density is the BRS exact and therefore the model can be regarded as a topological theory. In this model, the divergence of the vacuum energy coming from the quantum corrections from matters can be absorbed into the redefinition of the scalar field. In this paper, we consider the extension of the model in order to apply the mechanism to other kinds of divergences coming from the quantum correction and consider the cosmology in an extended model.


I. INTRODUCTION
By the recent cosmological observations, we now believe the accelerating expansion of the present universe, whose simplest model may be given by a cosmological term with a small cosmological constant. We also know that the quantum correction coming from the contributions from matters to the vacuum energy, which may be identified with the cosmological constant, terribly diverges and we need the very finely tuned counterterm to cancel the divergence. For the discussion about the small but non-vanishing vaccum energy, see [1] for example. In order to solve the problem of the large quantum corrections to the vacuum energy, the unimodular gravity theories  have been proposed and discussed. Other scenarios like the sequestering mechanism have been also proposed [29][30][31][32][33][34][35]. In [36], we have proposed a new model which could be regarded with a topological field theory and desicussed the cosmology in the model [37]. We should note that by the quantum corrections from the matter, the following terms, besides the cosmological constant, are generated, L qc = αR + βR 2 + γR µν R µν + δR µνρσ R µνρσ . (1) Here the coefficient α diverges quadratically and β, γ, and δ diverge logarithmically without the cut-off scale. If we further include the quantum corrections from the graviton, infinite numbers of diverging quantum corrections appear. In order to solve these problems of the divergences and the renormalizations, we extend the model in [36] and discuss the cosmology given by the extended models. In the next section, we review on the topological model for the cosological constant based on [36]. In Section III, we extend the toplogical model and consider more general divergences. In Section IV, we consider the cosmology given by the extended topological model. The last section is devoted for summary and discussions.

II. TOPOLOGICAL MODEL FOR COSMOLOGICAL CONSTANT PROBLEM
The action of the model in [36] has the following form, Although λ and ϕ are ordinary scalar fields but b and c are fermionic (Grassmann odd) scalar fields and we regard that b is an anti-ghost field and c with a ghost field. 12 In (2), S matter is the action of matter and L gravity is the Lagrangian density of arbitrary gravity. There does not appear any parameter in the action (2) except in the parts of S matter and L gravity .
By separating the gravity Lagrangian density L gravity into the sum of some constant Λ, which may include the large quantum corrections from matter, and other part L (0) gravity , L gravity = L (0) gravity − Λ, we redefine the scalar field λ by λ → λ − Λ. The obtained action has the following form, Note that the action (3) does not include the cosmological constant Λ and therefore the constant Λ never affects the dynamics in the model. This tells that the large quantum corrections from the matters can be tuned to vanish. As shown in [36], there appear the ghosts, which generate the negative norm states in the quantum theory, in the model (2). The existence of the negative norm states makes the so-called Copenhagen interpretation invalid and therefore the model becomes inconsistent. The negative norm states, however, can be eliminated by defining the physical states which are annihilated by the BRS charge [40]. We can find that the action (2) is invariant under the infinite numbers of the BRS transformation, Here ǫ is a fermionic (Grassmann odd) parameter and λ 0 should satisfy the following equation by putting λ = λ 0 , which is obtained by the variation of the action (2) with respect to ϕ. 3 If the physical states are defined as the states invariant under the BRS transformation in (4), the negative norm states can be eliminated by the Kugo-Ojima mechanism in the gauge theory [41,42]. 4 Because Eq. (4) tells that λ − λ 0 is given by the BRS transformation of the anti-ghost b and therefore the vacuum expectation value of λ − λ 0 must vanish in the physical states. Therefore there occurs the spontaneous breakdown of the corresponding BRS symmetry in case that the vacuum expectation value of λ − λ 0 does not vanish. For the broken BRS symmetry, it is impossible to impose the physical state condition. It should be noted, however, there is one and only one unbroken BRS symmetry in the infinite numbers of the BRS symmetries in (4). The point is that Eq. (5) is nothing but the field equation for λ. Because the real world is realized by one and only one solution of (5) for λ, one and only one λ 0 is chosen so that λ = λ 0 and therefore the corresponding BRS symmetry is not broken, which eliminates the negative norm states, which are the ghost states, and the unitarity is guaranteed. Although the quantum fluctuations are prohibited by the BRS symmetry, λ 0 can include the classical fluctuation as long as λ 0 satisfies the classical equation (5).
We can regard the Lagrangian density in the action (2), as the Lagrangian density of a topological field theory [43], where the Lagrangian density is given by the BRS transformation of some quantity. If we consider the model which only include the scalar field ϕ but whose Lagrangian density identically vanishes, which tells that the action is trivially invariant under any transformation of ϕ. We may fix the gauge symmetry by imposing the following gauge condition, By following the paper [44], we find the gauge-fixed Lagrangian is given by the BRS transformation (4) of −b (1 + ∇ µ ∂ µ ϕ) and we obtain Then we find the Lagrangian density (6) is given by the BRS transformation of the quantity −b (1 + ∇ µ ∂ µ ϕ) up to the total derivative terms if λ 0 = 0. The action is not given by the BRS transformation (4) with the non-vanishing λ 0 , which could be a reason why the Lagrangian density (6) gives non-trivial and physically relevant contributions.

III. EXTENSION OF TOPOLOGICAL MODEL
The mechanism in the last sectioncan work for the divergences in (1) or more general divergences [37]. In case that we include the divergences in (1), we may generalize the model in (6) as follows, As in the case of the vacuum energy, the divergences are included in the coefficients Λ, α, β, γ, and δ but if we shift the parameters λ (Λ) , λ (α) , λ (β) , λ (γ) , and λ (δ) as follows, we can rewrite the Lagrangian density (9), which tells that we can absorb the divergences into the redefinition of λ (i) , (i = Λ, α, β, γ, δ) and the divergences becomes irrelevant for the dynamics. 5 The Lagrangian density (11) is also invariant under the following BRS transformations where λ (i)0 's satisfy the equation, as in (5). The Lagrangian density (11) is also given by the BRS transformation (12) with λ (i)0 = 0,   = ǫ (L + (total derivative terms)) .
As mentioned, due to the quantum correction from the graviton, the divergences in infinite numbers of quantum corrections appear. Let O i be possible gravitational operators then a further generalization of the Lagrangian density (11) is given by Then all the divergences are absorbed into the redefinition of λ i . The Lagrangian density (15) is invariant under the BRS transformation and given by the the BRS transformation of some quantity and therefore the model can be regarded as a topological field theory, again. 5 In the model which includes the terms given by the square of curvatures in the Lagrangian density, there generally appear the massive scalar mode and/or the massive spin 2 mode. The latter is a ghost and violates the unitarity. These modes do not appear only in the case that the curvature square terms are given by the Gauss-Bonnet combination. The massive scalar mode and the massive spin 2 mode can be regarded as composite modes, which do not appear in the perturbation by the definition of quantum field theory. Therefore we may need to renormalize the curvature square terms to be the Gauss-Bonnet combination. In the Lagrangian density (11), the evolution of the scalar fields λ (i) (i = β, γ, δ) is given by the common equations, ∇ µ ∂µλ (i) = 0 as clear from the Lagrangian density (11). Therefore if we choose the initial condition or boundary condition so that λ (i) 's become the Gauss-Bonnet combination, the combination is preserved in whole space-time and the ghost does not appear.

IV. COSMOLOGY IN EXTENDED MODEL
By the arguments in the last sections, the problems of the divergences in the quantum theory might be solved but there is not any principle to determine the values of the observed cosmological constant and other coupling constants. The values could be determined by the initial conditions or the boundary conditions in the classical theory. Therefore it could be interesting to investigate the cosmology and specify the region of the initial conditions which could be consistent with the evolution of the observed universe. For the model (2), in [37], it has been shown that we need the fine-tuning for the initial conditions although the constraints on the conditions are relaxed a little bit.
For simplicity, we consider the following reduced model, In order to consider the cosmology, we assume b (Λ) = c (Λ) = b (α) = c (α) = 0 because the ghost number should be conserved and superselection rule should hold. We assume that the space-time is given by the FRW universe with flat spacial part, and we assume that all the scalar fields λ (Λ) , ϕ (Λ) , λ (α) , and ϕ (α) only depend on the cosmological time t. Then the variation of the action with respect to the scalar fields and metric gives the following equations.
If λ (Λ) and λ (α) are constant, the second equations in (18) and (19) are satisfied. Then Eq. (20) or (21) gives H is also a constant, Then a solution for the first equations in (18) and (19) is given by Because H is a constant, we find that the de Sitter space-time is a solution of this model. We now consider the stability of the obtained solution describing the de Sitter space-time by considering the perturbation, Then we obtain the following perturbed equations, By using (27), (29), and (30), we obtain By using (30), we delete δH in (26) and (28) and obtain We now define new variables δη (Λ) and δη (α) by and we rewrite (27), (27), (32), and (38) as follows, Furthermore we define Eqs. (37) and (38) give We now write Eqs. (34), (35), (36), (37), and (40) in a matrix form, If there is negative eigenvalue in the matrix A, the solution describing the de Sitter space-time is unstable but as clear from the form of the matrix A, which is triangular, the eigenvalues are given by four 3H 0 's and two 0's. Therefore the solution describing the de Sitter space-time is stable or at least quasi-stable as in the model (3) with L (0) gravity = R 2κ 2 , that is, the case of the Einstein gravity [37]. The stability tells that the solution might describe the acceleratingly expanding universe at present.

V. SUMMARY AND DISCUSSIONS
In summary, in the model (16), the divergences in the cosmological constant and the gravitational constant coming from the quantum corrections may not affect the dynamics but there is no principle to determine the constants, which may correspond to λ (Λ)0 and λ (α)0 in (22). These constants could be determined by the initial conditions or something else and therefore it could be interesting to investigate the cosmology by including the matter as in the model of [37]. In the model (15), however, we need infinite numbers of the initial conditions, which might be unphysical but this problem might give any clue for the quantum gravity. In fact, the second equations in (18) and (19) describe the evolutions of the effective coupling constants λ (Λ) and λ (α) with respect to the scale a(t) as in the renormalization group equations.
Finally we mention on the relation with the Weinberg no-go theorem [49]. In the paper, it was assumed that the system has translational invariance and GL(4) invariance. Then it has been shown that we need the fine-tuning the parameters in order to obtain the vanishing or small cosmological constant. In the paper, the translational invariance was assumed even for the fields and therefore all the fields are constant. As clear from Eq. (24), some of the scalar fields must depend on the time and not constants. Therefore the Weinberg no-go theorem cannot apply for the model in this paper although there might be a problem of the fine-tuning for the initial conditions.