# Fixed points and infrared completion of quantum gravity

###### Abstract

The phase diagram of four-dimensional Einstein-Hilbert gravity is studied using Wilson’s renormalization group. Smooth trajectories connecting the ultraviolet fixed point at short distances with attractive infrared fixed points at long distances are derived from the non-perturbative graviton propagator. Implications for the asymptotic safety conjecture and further results are discussed.

Introduction. — An important challenge in theoretical physics relates to the quantum nature of gravity and the incorporation of the gravitational force into the successful Standard Model of Particle Physics. A promising avenue is provided through Weinberg’s asymptotic safety conjecture, according to which a quantum theory of metric gravity may very well exist on a fundamental level Weinberg (1979). The scenario stipulates the existence of an ultraviolet (UV) fixed point which renders the theory finite even beyond the Planck scale, in a manner similar to the well-known weakly-coupled UV fixed point of quantum chromodynamics (QCD). In gravity, and unlike QCD, the fixed point is expected to be interacting and its investigation requires non-perturbative methods. Substantial support in favor of a gravitational UV fixed point has been accumulated in recent years based on continuum studies, see e.g. Niedermaier and Reuter (2006); Litim (2008); Codello et al. (2009); Litim (2011); Reuter and Saueressig (2012) and references therein, lattice simulations Hamber (2009); Ambjorn et al. (2012), and holography Litim et al. (2012).

A method of choice in the study of gravity in the continuum is given by Wilson’s renormalization group, based on the infinitesimal integrating-out of momentum degrees of freedom. It permits a systematic analysis even of strongly correlated and strongly coupled theories and offers the prospect for a deeper understanding of gravity in its extremes of shortest and largest distances. Furthermore, insights achieved for other strongly correlated systems such as critical scalar theories and QCD, reviewed in Litim and Zappala (2011); Pawlowski (2011), can now be exploited for gravity.

A useful laboratory for quantum gravity is given by the Einstein-Hilbert approximation, which retains Newton’s coupling and the cosmological constant as the relevant couplings. By now, the study of gravitational fixed points within the Einstein-Hilbert theory and the phase diagram has lead to a consistent picture with a non-trivial short-distance fixed point in Newton’s coupling and the cosmological constant, see Niedermaier and Reuter (2006); Codello et al. (2009); Litim (2011); Reuter and Saueressig (2012) and references therein. Infrared fixed points corresponding to a vanishing or negative cosmological constant Donkin and Pawlowski (2012); Litim and Satz (2012) have been detected as well. For positive cosmological constants, the IR behavior is plagued by additional divergences, though the conjectured existence of an IR fixed point Bonanno and Reuter (2002) has received some attention recently Donkin and Pawlowski (2012); Litim and Satz (2012); Nagy et al. (2012); Rechenberger and Saueressig (2012); Contreras and Litim (2012).

In this Letter, we derive the fixed points and the phase diagram of gravity from the RG flow for the non-perturbative graviton propagator. Our derivation of the RG flow differs from previous studies in two main aspects: firstly, we evaluate the flow on flat backgrounds, which permits a better control on how the propagating degrees of freedom drive the RG flow for the relevant couplings. It also provides a consistency check for earlier studies based on background field methods Niedermaier and Reuter (2006); Litim (2008); Codello et al. (2009); Litim (2011); Reuter and Saueressig (2012), bi-metric formulations, Manrique et al. (2011), and geometric flows Donkin and Pawlowski (2012). Secondly, we adopt a new bi-local projection technique to identify the scale-dependence of Newton’s coupling and the cosmological constant, offering an improved resolution of the relevant fluctuations in the UV and IR limits of the theory. As a result, we obtain the first global phase diagram with a stable ultraviolet fixed point which connects smoothly with an attractive infrared fixed point at positive cosmological constant.

Gravitational renormalization group. — The present work is done within the functional RG approach which is set up below. It is then utilized for the derivation of the RG flow for Newton’s coupling and the cosmological constant . Under the RG momentum flow, the classical couplings and turn into scale-dependent couplings, whose values depend on the RG scale parameter . As a function of the latter, the couplings interpolate between the short and long distance regimes of the theory. It is convenient to introduce the scale dependent, dimensionless couplings and , where denotes the graviton wave-function renormalization. In terms of these, the corresponding RG functions have the form

(1) |

with denoting the logarithmic RG ‘time’, and an arbitrary reference scale. We also introduced the anomalous dimension of the graviton and the ‘anomalous dimension’ of the cosmological constant . From (1) one can see that the fixed point condition is satisfied if the canonical running due to the mass dimension of and is exactly counter-balanced by the running induced by quantum fluctuations. In case of an UV fixed point this scaling is approached in the limit , thus leading to divergence-free couplings at arbitrarily small distances. In order to deduce expressions for and we consider the following gauge-fixed effective action,

(2) |

In (2) is the Ricci scalar, and the second and third terms are the gauge fixing term and the ghost action. We use the harmonic gauge condition where is the covariant derivative with respect to the background connection. The background metric is necessary to construct the gauge fixing and allows for a split with a fluctuating graviton field . The background metric comes with its own auxiliary background diffeomorphisms. The effective action is invariant under combined transformations for the background and the fluctuating field Litim and Pawlowski (1999, 1998); Freire et al. (2000); Litim and Pawlowski (2002); Folkerts et al. (2012). The Faddeev-Popov operator reads explicitly . The scale-dependence of is irrelevant in Landau-DeWitt gauge, Litim and Pawlowski (1998); Folkerts et al. (2012); Donkin and Pawlowski (2012) which we adopt throughout.

We now turn to the scale-dependence of the effective action in (2). It is governed by Wetterich (1993),

(3) |

where . The Wilsonian regulator ensures finiteness of the equation, is the full propagator and . The trace over the operator product implies a sum over momenta, all internal indices and fields with a minus sign for the ghost fields. We choose a regulator of the form Litim (2001, 2004); Pawlowski (2007)

(4) | ||||

which allows for a largely analytical access. The RG flow (3), (4) together with (2) enables us to compute (1).

Propagator flow and flat backgrounds. — The scale dependence of the effective action in (2) is completely governed by that of . It is only left to specify our projection procedure for the flow : Firstly we use flat backgrounds . This allows us to distinguish on the right hand side of the flow between propagating modes and the non-dynamical background Folkerts et al. (2012). This distinction is crucial in Yang–Mills theory to obtain a confining potential of the order-parameter Braun et al. (2010). Moreover, the present setting allows us to study the dependence of on external momenta . Following this route, we obtain the flow equation by functional differentiation of (2). It has the diagrammatic representation given in Figure 1. All vertices and propagators are fully dressed. The flow of the inverse propagator contains the three- and the four-point function on the RHS which follow by functional differentiation of (2). The extensive algebra was carried out with the help of Form and xTensor Vermaseren (2000); Martingarcia et al. (2007). Note also that the ghosts couple only linearly to the graviton in our approximation.

Einstein-Hilbert gravity. — Next we turn to the definition of the -functions (1) within the Einstein-Hilbert approximation using an appropriate, bi-local, projection in momentum space.

We recall that the relevant information is stored in the dynamical spin two degrees of freedom of the graviton, which is obtained by projecting the flow onto the transverse-traceless (TT) subspace of the symmetric rank-four tensors. Applying this TT projection to the LHS of the flow equation leads to

(5) |

From an analysis of the momentum structure one infers a quadratic divergence of each diagram in Figure 1, which we have checked explicitly. Hence, one would expect a similar quadratic divergence for large momenta of the right hand side of the TT-flow, denoted with . For large momenta the flow can be calculated analytically and we find a cancellation of all divergent terms, leading to the exact asymptotic behavior

(6) |

This general behavior of is displayed in Figure 2 for and .

A similar cancellation has been observed in the context of the Yang–Mills–gravity system where it leads to a vanishing gravitational one-loop contribution to the running of the gauge coupling Ebert et al. (2008); Folkerts et al. (2012). To further exploit the structure of the RG flow for (see Fig. 2), we note that it displays a dip as a function of external momenta at about , with a curvature opposite to the one found at . Since fluctuations are integrated out at momenta of the order of the RG scale , a consistent choice for the evaluation of is given by the symmetric point , which we adopt to identify the RG flow for Newton’s coupling. It also follows from (5) that the cosmological constant appears as a graviton mass defined at vanishing momenta. Therefore, the RG flow for the vacuum energy is deduced from at . Consequently, we are lead to (1) via a bi-local projection in momentum space,

(7) | ||||

The virtue of our set-up is that the running of couplings is sensitive to the global (momentum) behavior of the theory as encoded in its two-point function. In addition, the flow of the cosmological constant can be obtained completely analytically. Writing , our RG flow (1) takes the final form

(8) | ||||

The last terms originate from the ghosts, and the functions and are known numerically and plotted in the relevant regime in Figure 3.

Fixed points and phase diagram. —

no ghosts | 1.95 | 0.11 | 0.21 | ||

with ghosts | 2.03 | 0.22 | 0.45 |

The phase diagram of the RG flows (1) with (8) displays four connected fixed points , , and in the physical regime, see Figure 4. The fixed point at denotes the asymptotically safe UV fixed point whose coordinates and scaling exponents are given in Tab. 1. It is characterised by two UV relevant real eigenvalues, which turn into a complex conjugate pair in the absence of ghost field fluctuations. A complex conjugate pair of eigenvalues, as found in many previous studies, can be lifted via additional interactions e.g. the inclusion of higher derivative interactions. For our set-up, we conclude that the degeneracy in the UV scaling of the Ricci scalar and the vacuum energy is already lifted by the ghost sector. The fixed point at is the well-known repulsive Gaußian IR fixed point. It corresponds to classical gravity in the IR with a vanishing cosmological constant. The fixed point at is IR attractive in both couplings and governs infrared gravity with a negative vacuum energy. In addition we find a non-trivial IR fixed point at

(9) |

that is connected to the UV fixed point via smooth RG trajectories. It is a distinctive property of the present approach that it admits an infrared completion of quantum gravity for positive cosmological constant. Naively, one would expect classical RG-scaling in the IR fixed point regime Bonanno and Reuter (2002), for signatures thereof see Litim (2004); Contreras and Litim (2012). Here, we find an asymptotic IR behaviour with non-classical scaling exponents and . This IR behavior satisfies the exact scaling relation within the given numerical accuracy. It implies an asymptotic weakening of gravity, see also Donkin and Pawlowski (2012). A more detailed analysis will be presented in Christiansen et al. (2012). In our set-up, the non-classical scaling of the propagator is a consequence of strong IR effects. It will thus be interesting to analyse the impact of this scaling on the observable background Newton constant and background cosmological constant, see Donkin and Pawlowski (2012).

It is important to stress that physical short-distance initial conditions in the vicinity of the UV fixed point leads to trajectories in the regions Ia and Ib. In region Ia all trajectories end into the new IR fixed point with large anomalous dimensions for the graviton propagator; in region Ib the trajectories end in the IR fixed point with classical scaling for the graviton propagator. These two regions are separated by the IR instable separatrix leading to the Gaußian FP at . In turn, all other trajectories terminate in singularities and have no classical regime. Hence, the strong-gravity region is shielded by the separatrices () and () from regions which admit an extended semi-classical regime. The fact that a strong-gravity regime is shielded dynamically by the RG flow is also present in previous global UV-IR studies Donkin and Pawlowski (2012); Litim and Satz (2012), and thus appears to be a generic feature of asymptotically safe gravity.

Summary. — We have studied the phase diagram of quantum gravity with a novel projection technique based on the graviton two-point function. Our study also disentangles the role of fluctuation and background fields, and confirms that gravity becomes asymptotically safe. A further new result are global RG trajectories connecting the short distance fixed point with weak-gravity long distance fixed points. Interestingly, the strong gravity regime is dynamically shielded.

Acknowledgements — This work is supported by the Helmholtz Alliance HA216/EMMI, and by the Science and Technology Facilities Council (STFC) under grant number ST/J000477/1.

## References

- Weinberg (1979) S. Weinberg, General Relativity: An Einstein centenary survey, Eds. Hawking, S.W., Israel, W; Cambridge University Press pp. 790–831 (1979).
- Niedermaier and Reuter (2006) M. Niedermaier and M. Reuter, Living Rev.Rel. 9, 5 (2006).
- Litim (2008) D. F. Litim (2008), eprint 0810.3675.
- Codello et al. (2009) A. Codello, R. Percacci, and C. Rahmede, Annals Phys. 324, 414 (2009), eprint 0805.2909.
- Litim (2011) D. F. Litim, Phil.Trans.Roy.Soc.Lond. A369, 2759 (2011), eprint 1102.4624.
- Reuter and Saueressig (2012) M. Reuter and F. Saueressig, New J.Phys. 14, 055022 (2012), eprint 1202.2274.
- Hamber (2009) H. W. Hamber, Gen.Rel.Grav. 41, 817 (2009), eprint 0901.0964.
- Ambjorn et al. (2012) J. Ambjorn, A. Goerlich, J. Jurkiewicz, and R. Loll (2012), eprint 1203.3591.
- Litim et al. (2012) D. F. Litim, R. Percacci, and L. Rachwal, Phys.Lett. B710, 472 (2012), eprint 1109.3062.
- Litim and Zappala (2011) D. F. Litim and D. Zappala, Phys. Rev. D83, 085009 (2011), eprint 1009.1948.
- Pawlowski (2011) J. M. Pawlowski, AIP Conf.Proc. 1343, 75 (2011), eprint 1012.5075.
- Donkin and Pawlowski (2012) I. Donkin and J. M. Pawlowski (2012), eprint 1203.4207.
- Litim and Satz (2012) D. Litim and A. Satz (2012), eprint 1205.4218.
- Bonanno and Reuter (2002) A. Bonanno and M. Reuter, Phys.Lett. B527, 9 (2002), eprint astro-ph/0106468.
- Nagy et al. (2012) S. Nagy, J. Krizsan, and K. Sailer, JHEP 1207, 102 (2012), eprint 1203.6564.
- Rechenberger and Saueressig (2012) S. Rechenberger and F. Saueressig, Phys.Rev. D86, 024018 (2012), eprint 1206.0657.
- Contreras and Litim (2012) C. Contreras and D. F. Litim (2012), eprint in preparation.
- Manrique et al. (2011) E. Manrique, M. Reuter, and F. Saueressig, Annals Phys. 326, 463 (2011), eprint 1006.0099.
- Litim and Pawlowski (1999) D. F. Litim and J. M. Pawlowski, World Sci. pp. 168–185 (1999), eprint hep-th/9901063.
- Litim and Pawlowski (1998) D. F. Litim and J. M. Pawlowski, Phys. Lett. B435, 181 (1998), eprint hep-th/9802064.
- Freire et al. (2000) F. Freire, D. F. Litim, and J. M. Pawlowski, Phys. Lett. B495, 256 (2000), eprint hep-th/0009110.
- Litim and Pawlowski (2002) D. F. Litim and J. M. Pawlowski, JHEP 0209, 049 (2002), eprint hep-th/0203005.
- Folkerts et al. (2012) S. Folkerts, D. F. Litim, and J. M. Pawlowski, Phys.Lett. B709, 234 (2012), eprint 1101.5552.
- Wetterich (1993) C. Wetterich, Phys.Lett. B301, 90 (1993).
- Litim (2001) D. F. Litim, Phys. Rev. D64, 105007 (2001), eprint hep-th/0103195.
- Litim (2004) D. F. Litim, Phys. Rev. Lett. 92, 201301 (2004), eprint hep-th/0312114.
- Pawlowski (2007) J. M. Pawlowski, Annals Phys. 322, 2831 (2007), eprint hep-th/0512261.
- Braun et al. (2010) J. Braun, H. Gies, and J. M. Pawlowski, Phys.Lett. B684, 262 (2010), eprint 0708.2413.
- Vermaseren (2000) J. A. M. Vermaseren (2000), eprint arXiv:math-ph/0010025.
- Martingarcia et al. (2007) J. Martingarcia, R. Portugal, and L. Manssur, Computer Physics Communications 177, 640 (2007), eprint 0704.1756.
- Ebert et al. (2008) D. Ebert, J. Plefka, and A. Rodigast, Phys. Lett. B660, 579 (2008), eprint 0710.1002.
- Christiansen et al. (2012) N. Christiansen, J. M. Pawlowski, and A. Rodigast, in preparation (2012).