Einstein manifolds with torsion and nonmetricity

Dietmar Silke Klemm1,2 , Lucrezia Ravera2

1
Dipartimento di Fisica, Università di Milano,
Via Celoria 16, 20133 Milano, Italy.
2
INFN, Sezione di Milano,
arXiv:1811.11458v2 [gr-qc] 18 Feb 2020

Via Celoria 16, I-20133 Milano, Italy.

[email protected], [email protected]

Abstract

Manifolds endowed with torsion and nonmetricity are interesting both from the physical
and the mathematical points of view. In this paper, we generalize some results presented in
the literature. We study Einstein manifolds (i.e., manifolds whose symmetrized Ricci tensor is
proportional to the metric) in d dimensions with nonvanishing torsion that has both a trace
and a traceless part, and analyze invariance under extended conformal transformations of the
corresponding field equations. Then, we compare our results to the case of Einstein manifolds
with zero torsion and nonvanishing nonmetricity, where the latter is given in terms of the Weyl
vector (Einstein-Weyl spaces). We find that the trace part of the torsion can alternatively be
interpreted as the trace part of the nonmetricity. The analysis is subsequently extended to
Einstein spaces with both torsion and nonmetricity, where we also discuss the general setting
in which the nonmetricity tensor has both a trace and a traceless part. Moreover, we consider
and investigate actions involving scalar curvatures obtained from torsionful or nonmetric con-
nections, analyzing their relations with other gravitational theories that appeared previously in
the literature. In particular, we show that the Einstein-Cartan action and the scale invariant
gravity (also known as conformal gravity) action describe the same dynamics. Then, we consider
the Einstein-Hilbert action coupled to a three-form field strength and show that its equations
of motion imply that the manifold is Einstein with totally antisymmetric torsion.
1 Introduction
In the 19th century, the branches of mathematics and physics experienced an extraordinary
progress with the emergence of non-Euclidean geometry. In particular, the development of Rie-
mannian geometry led to many important results, among which is the rigorous mathematical for-
mulation of Einstein’s general relativity.
In spite of the success and predictive power of general relativity, there are still some open
problems and questions, whose understanding and solution may need the formulation of a new
theoretical framework as well as generalizations and extensions of Riemannian geometry. One
possible way of generalizing Riemannian geometry consists in allowing for nonvanishing torsion and
nonmetricity (metric affine gravity) [1] (see also [2–8] and the recent work [9]). There are several
physical (and mathematical) reasons which motivate the introduction of torsion or nonmetricity in
the context of gravitational theories (see [1] for details). For instance, nonmetricity is a measure
for the violation of local Lorentz invariance, which has been attracting some interest recently.
Furthermore, nonmetricity and torsion find applications in the theory of defects in crystals, where,
in particular, nonmetricity describes the density of point defects, while torsion is interpreted as
density in line defects [10]. Moreover, as shown in [11], incorporating torsion and nonmetricity may
allow to explore new physics associated with defects in a hypothetical spacetime microstructure.
Recently, in [12–14] the authors discussed the propagation of matter fields in theories with torsion
and nonmetricity. Further applications include quantum gravity [15] and cosmology [16–18].
Moreover, torsion is related to the translation group and to the energy-momentum tensor of
matter, while nonmetricity is related to the group GL(4, R)/SO(3, 1) (in four dimensions) and to
the hypermomentum current (see Refs. [5] and [6,8], where, in particular in the latter, equations of
motion in metric affine manifolds were studied); the trace of the nonmetricity (the Weyl vector) is
related to the scale group and to the dilation (or scale) current. In particular, in matter the shear
and dilation currents couple to nonmetricity, and they are its sources. It is to the dilation current
that the Weyl vector is coupled.
Historically, a remarkable generalization of Riemannian geometry was first proposed in 1918 by
Weyl (cf. e.g. [19–22] for an introduction), who introduced an additional symmetry in an attempt of
unifying electromagnetism with gravity geometrically [23, 24]. In Weyl’s theory, both the direction
and the length of vectors are allowed to vary under parallel transport. However, Weyl’s attempt to
identify the trace part of the connection associated with stretching and contraction with the vector
potential of electromagnetism failed, due to observational inconsistencies (see e.g. [25] for details).
Subsequently, there were many attempts to adjust the theory. Finally, following [26], Weyl showed
that a satisfactory theory of electromagnetism can be achieved if the scale factor is replaced by a
complex phase. This was the origin of what is now well-known as the U(1) gauge theory1 .
The trace part of the connection introduced by Weyl is known as the Weyl vector. When it is
given by the gradient of a function, there exists a scale transformation (dilatation) that sets the
vector to zero. In this case, Weyl geometry is said to be integrable (parallel transported vectors
along closed paths return with unaltered lengths) and there exists a subclass of global gauges in
which the geometry is Riemannian.
Although Weyl’s theory of electromagnetism fails, there has been a renewed interest in it [28,29].
Indeed, there are motivations for seeking a deeper understanding of general relativity formulated
1
See [27] and references therein for interesting details on ‘the dawning of gauge theory’.

1
within the framework of Weyl geometry (and especially of integrable Weyl geometry), in particular
concerning scale invariant general relativity and higher symmetry approaches to gravity involving
conformal invariance [25]. Always in Weyl’s perspective, conformal (higher curvature) gravity
theories were constructed and studied in detail in [30–32]. Furthermore, in [33] an observational
constraint to the non-integrability of lengths in the original Weyl theory was placed for the first
time.
A Weyl manifold is a conformal manifold equipped with a torsionless but nonmetric connection,
called Weyl connection, preserving the conformal structure. Then, it is said to be Einstein-Weyl if
the symmetric, trace-free part of the Ricci tensor of this connection vanishes (and the symmetric
part of the Ricci tensor of the Weyl connection is proportional to the metric). Thus, Einstein-Weyl
manifolds represent the analog of Einstein spaces in Weyl geometry and are less trivial than the
latter, which have necessarily constant curvature in three dimensions.
Einstein-Weyl spaces were studied in [34–46], and they are also relevant in the context of (fake)
supersymmetric supergravity solutions [47–52]. Einstein-Weyl geometry is particularly rich in three
dimensions [34, 35], where it has an equivalent formulation in twistor theory [53], which provides a
tool for constructing selfdual four-dimensional geometries. Selfdual conformal four-manifolds play
a central role in low-dimensional differential geometry, and a key tool in this context is provided by
the so-called Jones-Tod correspondence [54], in which the reduction of the self-duality equation by
a conformal vector field is given by the Einstein-Weyl equation together with the linear equation for
an abelian monopole (in other words, the Jones-Tod correspondence is a correspondence between
a self-dual space with symmetry and an Einstein-Weyl space with a monopole). Einstein-Weyl
structures are also related to certain integrable systems, like the SU(∞) Toda field equations [55]
or the dispersionless Kadomtsev-Petviashvili equation [56].
On the other hand, as already mentioned, another generalization of Riemannian geometry is
given by the introduction of a nonvanishing torsion, which is the case for the Einstein-Cartan
theory [57–61], where the geometrical structure of the manifold is modified by allowing for an anti-
symmetric part of the affine connection (see also [62] for a recent review on torsional constructions
and metric affine gauge theories). Cartan suggested that spacetime torsion is related to the intrinsic
angular momentum, before the concept of spin was introduced. Cartan’s theory was then reinter-
preted as a theory of gravitation with spin and torsion [63–65]. Subsequently, the introduction of a
non-vanishing torsion has been widely analyzed in general relativity and in the setting of telepar-
allel gravities [66–72], as well as in other contexts. In particular, in [73, 74] the torsion tensor was
related to the Kalb-Ramond field [75]. Furthermore, the relation between torsion and conformal
symmetry was studied by several authors, and it turned out that torsion plays an important role in
conformal invariance of the action and behaves like an effective gauge field [76,77]. Subsequently it
was shown that in the nonminimally coupled metric-scalar-torsion theory, for some special choice
of the action, torsion acts as a compensating field and the full theory is conformally equivalent
to general relativity at a classical level [78, 79]. More recently, in [80] the metric-torsional confor-
mal curvature of four-dimensional spacetime was constructed, and in [81] different types of torsion
were investigated, together with their effect on the dynamics and conformal properties of fields.
Conformal invariance was also analyzed in generalizations of Einstein-Cartan spaces including non-
metricity [82–85], and in [86] an exhaustive classification of metric affine theories according to their
scale symmetries was presented (see also [87]). Finally, in a cosmological context, it was proposed
in [88, 89] that a nonvanishing torsion can serve as an origin for dark energy. Let us also mention,
here, that a generic theory (without matter) involving terms quadratic in torsion and nonmetricity

2
will be classically equivalent at low energy to Einstein’s theory, as discussed in [90] and references
therein. From a mathematical point of view, Einstein manifolds with skew-symmetric torsion (i.e.,
totally antisymmetric torsion) were analyzed in [91, 92].
Motivated by the fact that nonmetric and torsionful connections are interesting both from the
physical and the mathematical point of view, in this paper we generalize some results presented
previously in the literature. In particular, we study Einstein manifolds in d dimensions with
nonvanishing torsion that has both a trace and a traceless part, and we analyze invariance under
extended conformal transformations (see Refs. [78, 82], where these transformations are defined for
metric affine spaces) in this context. Then, we compare our results to the case of Einstein spaces
with zero torsion and nonvanishing nonmetricity, where the latter is given in terms of the Weyl
vector. We find that the trace part of the torsion can alternatively be interpreted as the trace part
of the nonmetricity. Subsequently, we extend our analysis to the case of Einstein manifolds with
both torsion and nonmetricity (Einstein-Cartan-Weyl spaces), where we allow for both a trace and
a traceless part of the nonmetricity tensor. Finally, we construct and investigate actions involving
scalar curvatures obtained from torsionful or nonmetric connections, and analyze their relations
with other gravitational theories known in the literature. In particular, we consider the Einstein-
Cartan action and discuss its relationship with scale invariant gravity (also known as conformal
gravity, which is invariant under Weyl transformations) [93–102], showing that they describe the
same dynamics. Then, we study the Einstein-Hilbert action coupled to a three-form Hµνρ and
shew that its equations of motion imply that the manifold is Einstein with skew-symmetric torsion.
Furthermore, it turns out that the equations of motion of Einstein gravity coupled to a three-form
may also be retrieved from a constrained action that contains the scalar curvature of a connection
with torsion. Let us specify that in this work we will focus on the vacuum, without considering
matter.
The remainder of this paper is organized as follows: In section 2, we consider Einstein spaces
with torsion that has both a trace and a traceless part. In particular, we find the field equations
satisfied by an Einstein-Cartan space. Then, the invariance under extended conformal (Weyl)
transformations of the latter is studied and the results are compared to the case of Einstein-Weyl
manifolds, which have nonvanishing nonmetricity but zero torsion. In section 3, we extend the anal-
ysis to Einstein-Cartan-Weyl manifolds, and add thereby also a traceless part to the nonmetricity
tensor. In section 4, the Weyl invariant Einstein-Cartan action is studied and shown to be equiva-
lent to scale invariant gravity (i.e., conformal gravity), which involves the presence of a scalar field
φ. Subsequently, in section 5 we consider the Einstein-Hilbert action coupled to a three-form, and
show that the resulting field equations imply that the space is Einstein with torsion, where the
latter is proportional to Hµνρ . We conclude our work with some comments and possible future
developments. In the appendix we collect some technical details.

2 Einstein manifolds with torsion

We first consider a d-dimensional Einstein manifold with metric gµν and nonvanishing torsion
(i.e., a so-called Einstein-Cartan manifold)2 . The connection Γλµν can be decomposed as

Γλµν = Γ̃λµν + N λµν , (2.1)

2
Our convention for the metric signature is (−, +, +, · · · , +).

3
where Γ̃λµν are the connection coefficients of the Levi-Civita connection (i.e., the Christoffel sym-
bols) and N λµν is called the distortion. Here, the latter can be written as3

1
Nλµν = (Tνλµ − Tλνµ − Tµνλ ) , (2.2)
2
where T λµν = eaλ T aµν is the torsion4 , antisymmetric in the last two indices,

T λµν = Γλµν − Γλνµ . (2.3)

Let us also introduce the contorsion (or contortion), antisymmetric in the first two indices,
1
Kνλµ = (Tνλµ − Tλνµ − Tµνλ ) . (2.4)
2
Observe that the distortion (2.2) can then be written as

Nλµν = Kνλµ . (2.5)

In [91,92], Einstein manifolds with skew-symmetric torsion were analyzed. Below, we shall consider
a general decomposition of the torsion tensor, which can be decomposed in a traceless and a trace
part as
1  λ 
T λµν = T̆ λµν + δ ν Tµ − δλµ Tν . (2.6)
d−1
In particular, we have T̆ νµν = 0 and Tµ ≡ T νµν . Notice that 2N λ[µν] = T λµν . The distortion (2.5)
becomes then
1 1  1
Nλµν = (Tνλµ − Tλνµ − Tµνλ ) = Kνλµ = T̆νλµ − T̆λνµ − T̆µνλ + (gµν Tλ − gµλ Tν ) , (2.7)
2 2 d−1
and thus (2.1) reads
1 λ  1  
Γλµν = Γ̃λµν + T̆ν µ − T̆ λνµ − T̆µνλ + gµν T λ − δµλ Tν . (2.8)
2 d−1

The explicit expression for the Riemann tensor R̃λρµν = ∂µ Γλνρ − ∂ν Γλµρ + Γλµσ Γσνρ − Γλνσ Γσµρ of
the Einstein-Cartan connection Γλµν is given in the appendix (see eq. (A.1)). There as well as in
the following, ∇ denotes the covariant derivative of the Levi-Civita connection. The corresponding
Ricci tensor Rρν = Rµρµν is given by (A.2). In particular, one gets

d−2 1 1 2−d µ 1
R[ρν] = ∇[ν Tρ] − T̆µσ[ν T̆ρ]µσ − T µ T̆[ρν]µ + T T̆µνρ + ∇µ T̆ µνρ . (2.9)
d−1 2 d−1 2(d − 1) 2
3
As we will see in sec. 3, in the case of torsionful, nonmetric connections the distortion is generally defined as
Nλµν = 21 (Tνλµ − Tλνµ − Tµνλ ) + 12 (Qλµν + Qλνµ − Qµλν ), where Qλµν is the nonmetricity tensor (we will introduce
and define it later). In the present section we first restrict ourselves to the case of vanishing nonmetricity, namely we
consider a metric, torsionful connection. The nonmetric torsion-free case (where Nλµν = 21 (Qλµν + Qλνµ − Qµλν ))
will be discussed at the end of the current section when we will explore Einstein-Weyl spaces.
4 λ
ea denotes the inverse vielbein and early latin indices a, b, . . . refer to the tangent space. The torsion 2-form is
defined as T a = dea + ω ab ∧ eb .

4
Note that if we set the traceless part of the torsion to zero, T̆ λµν = 0, we are left with
d−2 d−2 d−2
R[ρν] = ∇ T = ∂ T ≡ Fνρ , (2.10)
d − 1 [ν ρ] d − 1 [ν ρ] 2(d − 1)
where
Fνρ ≡ ∂ν Tρ − ∂ρ Tν . (2.11)
In general, one has thus
d−2 1 1 2−d µ 1
R[ρν] = Fνρ − T̆µσ[ν T̆ρ]µσ − T µ T̆[ρν]µ + T T̆µνρ + ∇µ T̆ µνρ . (2.12)
2(d − 1) 2 d−1 2(d − 1) 2
One can also construct another Ricci tensor by contracting the second and the third index of
the Riemann tensor. However, the Ricci tensor obtained in this way coincides with (A.2), since
Rλρµν = −Rρλµν is still valid (while it fails to be for nonmetric connections).
The Ricci scalar reads
(d − 2)(1 − d) 1 1
R = g ρν Rρν = R̃ + 2
Tµ T µ + 2∇µ T µ + T̆µνρ T̆ µνρ − T̆νρµ T̆ µνρ . (2.13)
(d − 1) 4 2
Let us now define an Einstein space with torsion by
R(ρν) = λgρν (2.14)
for some function λ. Using (A.2), this becomes
1 h i
R̃ρν + gρν ∇µ T µ + (d − 2)∇(ν Tρ) + (d − 3)T µ T̆(ρν)µ
d−1
1
+ [(2 − d)gρν Tµ T µ + (d − 2)Tν Tρ ] (2.15)
(d − 1)2
1 1
+ T̆ρ µσ T̆νµσ − T̆µσ(ρ T̆ν) µσ − ∇µ T̆(ρν)µ = λgρν ,
4 2
whose trace yields
d−2 1 1
R̃ + 2∇µ T µ − Tµ T µ + T̆ µρσ T̆µρσ − T̆ µρσ T̆ρσµ = λd, (2.16)
d−1 4 2
and thus  
1 µ d−2 µ 1 µρσ 1 µρσ
λ= R̃ + 2∇µ T − Tµ T + T̆ T̆µρσ − T̆ T̆ρσµ . (2.17)
d d−1 4 2
Hence, in terms of Riemannian data, (2.14) becomes
1 h i 1
R̃ρν + (d − 2)∇(ν Tρ) + (d − 3)T µ T̆(ρν)µ + [(d − 2)Tν Tρ ]
d−1 (d − 1)2
1 1
+ T̆ρ µσ T̆νµσ − T̆µσ(ρ T̆ν) µσ − ∇µ T̆(ρν)µ (2.18)
4  2 
1 d−2 µ d−2 µ 1 µτ σ 1 µτ σ
= gρν R̃ + ∇µ T + Tµ T + T̆ T̆µτ σ − T̆ T̆τ σµ ,
d d−1 (d − 1)2 4 2
which is a set of nonlinear partial differential equations characterizing an Einstein manifold with
torsion, henceforth termed Einstein-Cartan space.

5
2.1 Extended conformal invariance in Einstein-Cartan manifolds
We will now show that (2.14) is invariant under extended conformal transformations discussed
in [78]. Thus, let us consider the extended conformal (Weyl) transformations
′
gµν 7→ gµν = e2ω gµν ,
(2.19)
T λµν 7→ T ′λ λ λ λ
µν = T µν + δ ν ∂µ ω − δ µ ∂ν ω,

where ω = ω(x) is an arbitrary scalar field. Therefore, we have

Tµ 7→ Tµ′ = Tµ + (d − 1)∂µ ω, T̆ λµν 7→ T̆ ′λ λ

µν = T̆ µν . (2.20)

Moreover, (2.19) leads to the following transformation for the connection:

Γρµν 7→ Γ′ρµν = Γρµν + δρν ∂µ ω, (2.21)

which is called, specifically, a special projective transformation of the connection (see, for instance,
Refs. [86, 87]), also known as λ transformation. Let us observe that, actually, the combination of
the conformal metric transformation in (2.19) plus the special projective transformation (2.21) of
the affine connection is called a frame rescaling (see Refs. [86,87], where frame rescalings have been
considered in metric affine spaces, also including Einstein-Cartan ones).
For the Riemann tensor, the Ricci tensor and the scalar curvature, we get respectively

Rσρµν 7→ R′σρµν = Rσρµν ,

′
Rρν 7→ Rρν = Rρν , (2.22)
R 7→ R′ = e−2ω R.

Now, (2.14) implies R = λd, so that (2.14) is equivalent to

1
R(ρν) = Rgρν , (2.23)
d
which is obviously invariant under extended conformal transformations given by (2.19).

2.2 Comparison with Einstein-Weyl spaces

A Weyl structure on a manifold Σ consists of a conformal structure [g] = {f g|f : Σ → R+ },
ˆ fulfilling
and a torsion-free connection ∇
ˆ ν gλµ = 2Θν gλµ ,
∇ (2.24)

for some one-form Θ on Σ (the Weyl vector). The condition (2.24) is invariant under the transfor-
mation
′
gµν 7→ gµν = e2ω gµν , Θµ 7→ Θ′µ = Θµ + ∂µ ω. (2.25)
One can then define the nonmetricity tensor, which reads
ˆ ν gλµ = −2Θν gλµ .
Qµνλ = −∇ (2.26)

6
In this case the distortion is given by
1
N λµν = (Qλµν + Qλνµ − Qµλν ) = −δλν Θµ − δλµ Θν + Θλ gµν . (2.27)
2
ˆ is
A Weyl structure is said to be Einstein-Weyl [20] if the symmetrized Ricci tensor Wρν of ∇
proportional to some metric g ∈ [g],
1
W(ρν) = gρν W, (2.28)
d
where W is the scalar curvature of the Weyl connection ∇.ˆ It is given by5

W = R̃ + (d − 2)(1 − d)Θµ Θµ + 2(d − 1)∇µ Θµ . (2.29)

The condition (2.28) can be rewritten in terms of Riemannian data as

1 h i
R̃ρν + (d − 2)Θρ Θν + (d − 2)∇(ν Θρ) = gρν R̃ + (d − 2)∇µ Θµ + (d − 2)Θµ Θµ . (2.30)
d
The scope of this subsection is to compare the field equations for Einstein manifolds with torsion,
(2.18), with the Einstein-Weyl equations (2.30). To this end, let us define
Tµ
Aµ ≡ , (2.31)
d−1
such that, under the first transformation in (2.20), we have

Aµ 7→ A′µ = Aµ + ∂µ ω. (2.32)

Using (2.31) in (2.18), one gets

1 1
R̃ρν + (d − 2)Aρ Aν + (d − 2)∇(ν Aρ) + (d − 3)Aµ T̆(ρν)µ + T̆(ρ µσ T̆ν)µσ − T̆µσ(ρ T̆ν) µσ − ∇µ T̆(ρν)µ
 4  2
1 1 1
= gρν R̃ + (d − 2)∇µ Aµ + (d − 2)Aµ Aµ + T̆ µτ σ T̆µτ σ − T̆ µτ σ T̆τ σµ .
d 4 2
(2.33)

Thus, for T̆ λµν = 0, (2.33) exactly coincides with (2.30) if we identify Aµ with Θµ , i.e., Tµ →
(d − 1)Θµ . This is actually not surprising, since for T̆ λµν = 0 the torsion two-form is given by

1
T aµν = eaν Tµ − eaµ Tν = eaν Aµ − eaµ Aν = − (ea ∧ A)µν T a = A ∧ ea .


⇒ (2.34)
d−1
Then, the first Cartan structure equation gives

dea + ω a b ∧ eb = A ∧ ea ⇒ dea + (ω a b − δa b A) ∧ eb = 0. (2.35)

We can then define a new connection ω̂ ab as

ω̂ ab = ω ab − η ab A, (2.36)
5
See also the results of sec. 3 in the case of zero torsion.

7
which is torsion-free
dea + ω̂ ab ∧ eb = 0, (2.37)
but nonmetric, since ω̂ (ab)
6= 0. The trace part of the torsion can thus always be shuffled into a
Weyl vector and vice-versa. In the latter case, a Weyl structure gets translated into a conformal
structure [g] together with a torsionful connection D which is compatible with [g],
Dµ gνλ = 0. (2.38)
The torsion of D has only a trace part Tµ , and (2.38) is invariant under the transformation (2.19),
(2.20).
Finally, note that a duality between torsion and nonmetricity has also been discussed in [103]
in a slightly different context.

3 Einstein manifolds with torsion and nonmetricity

Let us now consider Einstein spaces with both torsion and nonmetricity (we will call these
Einstein-Cartan-Weyl manifolds), and study the Weyl invariance of the corresponding field equa-
tions.
With respect to section 2, we will in addition allow for a nonmetricity tensor of the form (2.26),
where ∇ ˆ has also torsion. We are thus considering only the trace part of the nonmetricity. The
consequences of adding a traceless part will be analyzed at the end of this section. The connection
Γ̂λµν of the Einstein-Cartan-Weyl manifold is given by

Γ̂λµν = Γ̃λµν + N λµν , (3.1)

where the Γ̃λµν are the Christoffel symbols, and the distortion N λµν reads
1 1
Nλµν = (Tνλµ − Tλνµ − Tµνλ ) + (Qλµν + Qλνµ − Qµλν ) , (3.2)
2 2
that is, in the present context,
1  1
Nλµν = T̆νλµ − T̆λνµ − T̆µνλ + (gµν Tλ − gµλ Tν ) + Θλ gµν − Θµ gλν − Θν gλµ . (3.3)
2 d−1
The Ricci tensor of ∇,ˆ that is R̂ρν = R̂µρµν , is given in the appendix (see eq. (A.3)). Note that
one can also construct another Ricci tensor Rρν = R̂µµρν (commonly referred to as the homothetic
curvature), since for nonmetric connections the Riemann tensor is not necessarily antisymmetric
in the first two indices. In our case we have
Rρν = d (∇ν Θρ − ∇ρ Θν ) , (3.4)
and thus the Ricci scalar associated with the homothetic curvature is identically zero. On the other
hand, the nonvanishing Ricci scalar is given by
R̂ = gρν R̂ρν
(d − 2)(1 − d) 1 1
= R̃ + Tµ T µ + 2∇µ T µ + T̆µνρ T̆ µνρ − T̆νρµ T̆ µνρ (3.5)
(d − 1)2 4 2
µ µ µ
+ (d − 2)(1 − d)Θµ Θ + 2(d − 1)∇µ Θ + 2(2 − d)Θ Tµ .

8
Observe that, if we define
Ťµ ≡ Tµ + (d − 1)Θµ , (3.6)
the Ricci scalar (3.5) becomes
(d − 2)(1 − d) 1 1
R̂ = R̃ + 2
Ťµ Ť µ + 2∇µ Ť µ + T̆µνρ T̆ µνρ − T̆νρµ T̆ µνρ , (3.7)
(d − 1) 4 2
which corresponds to the Ricci scalar of a metric connection with torsion (cf. eq. (2.13)), whose
trace part is given by Ťµ .
We define an Einstein-Cartan-Weyl space by
R̂(ρν) = λgρν (3.8)
for some function λ. Using (A.3), this can be rewritten in the equivalent form
1 h i 1
R̃ρν + (d − 2)∇(ν Tρ) + (d − 3)T µ T̆(ρν)µ + [(d − 2)Tν Tρ ]
d−1 (d − 1)2
1 1
+ T̆ρ µσ T̆νµσ − T̆µσ(ρ T̆ν) µσ − ∇µ T̆(ρν)µ
4 2
2(d − 2)
+ (d − 2)Θν Θρ + (d − 2)∇(ν Θρ) + Θ(ν Tρ) + (d − 3)Θµ T̆(νρ)µ
" d − 1 (3.9)
1 d−2 d−2 1 1
= gρν R̃ + ∇µ T µ + Tµ T µ + T̆ µτ σ T̆µτ σ − T̆ µτ σ T̆τ σµ
d d−1 (d − 1)2 4 2
#
2(d − 2)
+ (d − 2)∇µ Θµ + (d − 2)Θµ Θµ + Θµ Tµ ,
d−1

which is a system of nonlinear partial differential equations characterizing an Einstein-Cartan-Weyl

manifold.

3.1 Extended conformal invariance of the Einstein-Cartan-Weyl equations

Let us now discuss the extended conformal invariance of (3.8). In an affine manifold such as an
Einstein-Cartan-Weyl one, the most general extended conformal (Weyl) transformations involving
an arbitrary scalar field ω = ω(x) which leave the curvature tensor invariant are given by (see [82])
gµν 7→ e2ω gµν ,
T λµν 7→ T λµν + 2(1 − ξ)δλ[ν ∂µ] ω, (3.10)
Qλµν 7→ Qλµν − 2ξ∂µ ωδλν ,

where ξ denotes an arbitrary parameter that we are free to include [82, 85]6 . In particular, for the
one-forms Θ and T and for T̆ λµν we find

Θµ 7→ Θµ + ξ∂µ ω,
Tµ 7→ Tµ + (1 − ξ)(d − 1)∂µ ω, (3.11)
T̆ λµν 7→ T̆ λµν ,
6 ˆ µ gνρ = 2Θµ gνρ transforms covariantly.
Note that (3.10) implies that ∇

9
and the connection Γ̂ transforms according to

Γ̂ρµν 7→ Γ̂ρµν + (1 − ξ)δρν ∂µ ω. (3.12)

This ensures the invariance of the curvature tensor due to its special projective invariance (see, for
instance, Refs. [86, 87]).
Thus, for the Riemann tensor, the Ricci tensor and the scalar curvature one obtains respectively

R̂σρµν 7→ R̂σρµν , R̂ρν 7→ R̂ρν , R̂ 7→ e−2ω R̂. (3.13)

Eq. (3.8) implies R̂ = λd, so that (3.8) is equivalent to

1
R̂(ρν) = R̂gρν , (3.14)
d
which is clearly invariant under the extended conformal transformations written above.
Let us finally make some comments on two particular cases, namely ξ = 1 and ξ = 0.

• For ξ = 1 one has

Tµ 7→ Tµ , Θµ 7→ Θµ + ∂µ ω. (3.15)
Observe that (3.15) corresponds to the transformation (2.32), for Aµ = Θµ , discussed in
section 2 in the context of a Weyl structure (that is with nonmetricity and zero torsion).
Moreover, note that this is the only case in which the connection is also invariant, Γ̂ρµν 7→ Γ̂ρµν .
In fact, setting ξ = 1 into (3.10) and (3.11) leads to a conformal transformation of the metric in
an affine space, namely a transformation under which the metric tensor picks up a conformal
factor e2ω while the affine connection is left unchanged (see Refs. [86, 87]).

• For ξ = 0 we get the extended conformal transformation discussed in [78] in the context of
a torsion theory which leads to a special projective transformation for the connection. In
particular, in this case we have

Tµ 7→ Tµ + (d − 1)∂µ ω, Θµ 7→ Θµ , (3.16)

which reproduces exactly the transformation in (2.20) for Tµ discussed in section 2 for mani-
folds with torsion and vanishing nonmetricity, together with

Γ̂ρµν 7→ Γ̂ρµν + δρν ∂µ ω, (3.17)

which is a special projective transformation (3.17) for the connection. On the other hand,
let us observe that the combination of the conformal metric transformation in (3.10) plus the
special projective transformation (3.17) is called, according to [86, 87], a frame rescaling.

We can conclude that there are two unique transformations which single out torsion or nonmetricity.
This is in agreement with [82]. Note that the same results could have been obtained by considering
(3.7), together with the definition (3.6), that is by reabsorbing the nonmetricity and exploiting the
transformations of sec. 2 for an Einstein-Cartan manifold with torsion and vanishing nonmetricity.

10
3.2 Adding a traceless part to the nonmetricity tensor
In the following we extend the above analysis to include a traceless part of the nonmetricity as
well. Interestingly, in the case where the latter is totally symmetric, it can be viewed as representing
a massless spin-3 field [104, 105].
Thus, we decompose
Qλµν = −2Θµ gνλ + Q̆λµν , (3.18)
where Q̆νµν = 0. Using (2.6) and (3.18) in (3.2), the distortion becomes

1  1
Nλµν = T̆νλµ − T̆λνµ − T̆µνλ + (gµν Tλ − gµλ Tν )
2 d−1
1 
+ Θλ gµν − Θµ gλν − Θν gλµ + Q̆λµν + Q̆λνµ − Q̆µλν
2 (3.19)
1
= K̆νλµ + (gµν Tλ − gµλ Tν ) + Θλ gµν − Θµ gλν − Θν gλµ + M̆λµν
d−1
= Kνλµ + Mλµν ,

where we defined the so-called disformation (also known as deflection tensor)

1
Mλµν = (Qλµν + Qλνµ − Qµλν )
2
1 
(3.20)
= Θλ gµν − Θµ gλν − Θν gλµ + Q̆λµν + Q̆λνµ − Q̆µλν
2
= Θλ gµν − Θµ gλν − Θν gλµ + M̆λµν ,

which is symmetric in the last two indices. K̆νλµ and M̆νλµ are respectively the traceless part of
Kνλµ and Mνλµ ,

1  1 
K̆νλµ = T̆νλµ − T̆λνµ − T̆µνλ , M̆νλµ = Q̆λµν + Q̆λνµ − Q̆µλν . (3.21)
2 2
From (3.1) one obtains for the connection
1 λ  1  
Γ̂λµν = Γ̃λµν + T̆ν µ − T̆ λνµ − T̆µνλ + gµν T λ − δµλ Tν
2 d−1 (3.22)
1  
+ Θλ gµν − Θµ δλν − Θν δλµ + Q̆λµν + Q̆λνµ − Q̆µλν .
2
ˆ is given in the appendix (see (A.4)), and
The explicit expression for the Ricci tensor R̂ρν of ∇
it contains extra contributions from the traceless tensor Q̆λµν . The homothetic curvature is still
given by (3.4), while the Ricci scalar is

(d − 2)(1 − d)
R̂ = R̃ + Tµ T µ + 2∇µ T µ
(d − 1)2
+ (d − 2)(1 − d)Θµ Θµ + 2(d − 1)∇µ Θµ + 2(2 − d)Θµ Tµ (3.23)
1  1 
+ T̆µνρ T̆ µνρ − 2T̆µνρ Q̆µνρ + Q̆µνρ Q̆µνρ − T̆νρµ T̆ µνρ + T̆µνρ Q̆µνρ + Q̆νρµ Q̆µνρ .
4 2

11
Observe that, by defining

Ťµ ≡ Tµ + (d − 1)Θµ , Ťµνρ ≡ T̆µνρ − Q̆µνρ , (3.24)

where Ť νµν = 0, and using the fact that the symmetries of T̆µνρ and Q̆µνρ imply

T̆νρµ Q̆µνρ = 0, T̆ µνρ Q̆νρµ = T̆ µνρ Q̆µρν , (3.25)

one can shew that the Ricci scalar (3.23) can be written as

(d − 2)(1 − d) 1 1
R̂ = R̃ + 2
Ťµ Ť µ + 2∇µ Ť µ + Ťµνρ Ť µνρ − Ťνρµ Ť µνρ , (3.26)
(d − 1) 4 2

which corresponds to the Ricci scalar of a metric connection with nonvanishing torsion, whose trace
and traceless parts are now respectively given by Ťµ and Ťµνρ . This is analogous to the case in
which one does not include a traceless contribution for the nonmetricity, cf. eq. (3.7).
As before, we define an Einstein-Cartan-Weyl space by eq. (3.8), which becomes in the present
context
1 h µ
i 1
R̃ρν + (d − 2)∇(ν Tρ) + (d − 3)T T̆(ρν)µ + [(d − 2)Tν Tρ ]
d−1 (d − 1)2
1 1
+ T̆ρ µσ T̆νµσ − T̆µσ(ρ T̆ν) µσ − ∇µ T̆(ρν)µ
4 2
2(d − 2)
+ (d − 2)Θν Θρ + (d − 2)∇(ν Θρ) + Θ(ν Tρ) + (d − 3)Θµ T̆(νρ)µ
d−1
2−d µ d−4 d−4
+ T Q̆µ(νρ) + Tµ Q̆ρ µν − (d − 2)Θµ Q̆µ(νρ) + Θµ Q̆ρ µν
d−1 2(d − 1) 2
1 1 1 1
− Q̆µρσ Q̆µν σ + 2∇µ Q̆µ(νρ) − ∇µ Q̆ν µρ + T̆µ(ρ σ Q̆µν)σ + T̆ µ(ρσ Q̆ν)σµ − T̆(ρµσ Q̆ν)µσ
4 " 2 2 2
1 d−2 d−2 2(d − 2) µ
= gρν R̃ + ∇µ T µ + 2
Tµ T µ + (d − 2)∇µ Θµ + (d − 2)Θµ Θµ + Θ Tµ
d d−1 (d − 1) d−1
#
1  µτ σ µτ σ
 1
µτ σ µτ σ

µτ σ
+ T̆ T̆µτ σ + Q̆ Q̆µτ σ − T̆ T̆τ σµ + Q̆ Q̆τ σµ − T̆ Q̆µτ σ ,
4 2
(3.27)

which represents a system of nonlinear partial differential equations characterizing an Einstein-

Cartan-Weyl manifold with the most general form of torsion and nonmetricity.
Finally, we can consider the transformations (3.10). In particular, we have

Q̆λµν 7→ Q̆λµν . (3.28)

For the curvature tensors one still has the transformation laws given in (3.13), so that the Einstein-
Cartan-Weyl equations (3.8) are again invariant under extended conformal transformations for
arbitrary parameter ξ.

12
4 Einstein-Cartan action and scale invariant gravity
Let us consider the action
√
Z
4
 
S= dd x −gφ2 R − κφ d−2 , (4.1)

where R is the Ricci scalar (2.13) of a torsionful but metric connection, φ denotes a scalar field,
and κ is a constant. Along the same lines of [85], (4.1) can be rewritten as

d √
 
d−2 1 1
Z
4
2 µ µ µνρ µνρ
S = d x −gφ R̃ − Tµ T + 2∇µ T + T̆µνρ T̆ − T̆νρµ T̆ − κφ d−2 , (4.2)
d−1 4 2

with R̃ the scalar curvature of the Levi-Civita connection. One easily shows that (4.2) is invariant
under
2−d
gµν 7→ e2ω gµν , φ→7 e 2 ω φ, Tµ 7→ Tµ + (d − 1)∂µ ω, T̆ λµν 7→ T̆ λµν . (4.3)
Using the traceless part of the contorsion defined in (3.21), the action (4.2) becomes

d √
 
d−2
Z
4
2 µ µ µνρ
S = d x −gφ R̃ − Tµ T + 2∇µ T − K̆νρµ K̆ − κφ d−2 , (4.4)
d−1

and its variation w.r.t. Tµ and K̆νρµ yields respectively

2(d − 1) ∇µ φ
Tµ = − , K̆µ[νρ] = 0. (4.5)
d−2 φ
Notice that Tµ can be eliminated by an extended conformal transformation and is thus pure gauge.
Using the definition (3.21) and the fact that the traceless part of the torsion is antisymmetric in the
last two indices, we get T̆µνρ = 2K̆µ[νρ] = 0, and therefore also K̆µνρ = 0, in agreement with [79,85].
Varying the action (4.4) w.r.t. gµν and φ leads to
 
2 1 2d
φ R̃µν − gµν R̃ + ∇µ φ∇ν φ − 2φ∇ν ∇µ φ
2 d−2
2 1 2d
+ 2gµν φ∇ρ ∇ρ φ − gµν ∇ρ φ∇ρ φ + gµν κφ d−2 = 0, (4.6a)
d−2 2
4(d − 1) d d+2
φR̃ − ∇ρ ∇ρ φ − κφ d−2 = 0, (4.6b)
d−2 d−2

where we have used the expression for Tν in (4.5) as well as K̆µνρ = 0. Observe that the trace of
(4.6a) implies (4.6b), which can be understood as a consequence of φ being pure gauge.
Let us now consider the action

d √
 
4(d − 1)
Z
2d
2 µ
S = d x −g φ R̃ + ∇µ φ∇ φ − κφ d−2 , (4.7)
d−2

which is called scale invariant (also known as conformal gravity). It turns out that the equations
of motion following from (4.7) are precisely (4.6a) and (4.6b) obtained from (4.4) after having
used the expressions for the torsion. The actions (4.1) and (4.7) describe thus the same dynamics.

13
Notice also that, plugging Tµ (cf. (4.5)) and K̆µνρ = 0 into (4.4), one gets, up to a surface term7 ,
the conformal gravity action (4.7) (see also [85]).
One can also show that the action (4.1) implies that the spacetime is Einstein with torsion,
which is a completely new result. To see this, observe that eq. (4.6a) can be rewritten as

∇ρ ∇ρ φ 2d ∇ρ φ∇ρ φ d d−2
 
2d ∇µ φ∇ν φ ∇ν ∇µ φ 1 d 4
R̃µν + −2 = gµν R̃ − 2d + − κφ . (4.8)
d−2 φ2 φ d 2 φ d−2 φ2 2

Using also (4.6b), this can be cast into the form

∇ρ ∇ρ φ 2d ∇ρ φ∇ρ φ
 
2d ∇µ φ∇ν φ ∇µ ∇ν φ 1
R̃µν + −2 = gµν R̃ − 2 + . (4.9)
d−2 φ2 φ d φ d−2 φ2

On the other hand, consider the system (2.18) characterizing an Einstein-Cartan manifold, and
use the result (4.5) for the trace part of the torsion as well as T̆µνρ = 0. Then (2.18) boils down
precisely to (4.9).
Let us also observe that, as already mentioned in [85], conformal (Weyl) invariance allows to
2−d √
rescale φ 7→ e d ω φ. One can use this freedom to gauge fix φ = 1/(4 πG), where G is Newton’s
constant. Then the action (4.7) becomes

1 √ 
Z 
S= dd x −g R̃ − 2Λ , (4.10)
16πG

where we chose κ = 2Λ(16πG)2/(d−2) . The Einstein-Hilbert action with cosmological constant can
thus be viewed as a gauge fixed version of the action (4.7).
Finally, let us recall that the trace part of the torsion can also be interpreted as the trace part
of the nonmetricity (cf. sec. 2.2). If we set the traceless part of the torsion to zero, this leads to
the action
√
Z
4
 
S = dd x −gφ2 W − κφ d−2 , (4.11)

which is invariant under

2−d
gµν 7→ e2ω gµν , φ 7→ e 2
ω
φ, Θµ 7→ Θµ + ∂µ ω. (4.12)

The variation of (4.11) w.r.t. Θµ yields

2 ∇µ φ
Θµ = − . (4.13)
d−2 φ

Again, one can easily show that the actions (4.11) and (4.7) describe the same dynamics. (4.11)
implies that the spacetime is Einstein-Weyl, where the Weyl vector is given by (4.13), and is
thus pure gauge. Notice in this context that there is no known action principle that leads to the
Einstein-Weyl equations with non-exact Weyl vector.
√ h i
7
dd x −g − 4(d−1) ∇µ (φ∇µ φ) .
R
The surface term is d−2

14
5 Einstein-Hilbert action coupled to a 3-form as Einstein-Cartan
gravity
The Einstein-Hilbert action coupled to a 3-form field strength reads

d √
 
1
Z
µνρ
S1 = d x −g R̃ − Hµνρ H , (5.1)
12
where Hµνρ is given in terms of a gauge potential Bµν ,

Hµνρ = ∂µ Bνρ + ∂ν Bρµ + ∂ρ Bµν , Bµν = −Bνµ . (5.2)

The variation of (5.1) w.r.t. Bµν leads to

∇µ Hµνρ = 0, (5.3)

while varying gρν gives

1 1 1
R̃ρν − gρν R̃ + gρν Hµτ σ H µτ σ − Hρ µσ Hνµσ = 0. (5.4)
2 24 4
On the other hand, consider the system (2.18) satisfied by an Einstein manifold with torsion.
Assume that Tµ = 0 and take T̆µνρ to be completely antisymmetric. Then (2.18) boils down to
 
1 µσ 1 1 µτ σ
R̃ρν − T̆µσν T̆ρ = gρν R̃ − T̆ T̆µτ σ . (5.5)
4 d 4
We would like to compare this with (5.4). To this end, take the trace of (5.4), which leads to
d−6
R̃ = H 2, H 2 ≡ Hµτ σ H µτ σ . (5.6)
12(d − 2)
Now subtract its trace part from (5.4) to obtain
1 1 1
R̃ρν − gρν R̃ − Hρ µσ Hνµσ + gρν H 2 = 0, (5.7)
d 4 4d
which coincides precisely with (5.5) if we identify Hµνρ = T̆µνρ . The equations of motion following
from (5.1) can thus be interpreted as implying that the spacetime is Einstein with skew-symmetric
torsion Hµνρ satisfying (5.3). Notice however that the equations (5.4) are more restrictive than
(5.5), since they contain in addition the trace part (5.6), while (5.5) is traceless. This is somehow
reminiscent of hyper Cauchy-Riemann (hyper-CR, or Gauduchon-Tod) spaces [106], where on top
of the (trace-free) Einstein-Weyl equations there is a constraint on the scalar curvature.
Quite remarkably, the equations (5.3), (5.4) can also be retrieved from the constrained action

d √
  
1
Z
µνρ
S2 = d x −g R + λ T̆µνρ − √ (∂µ Bνρ + ∂ν Bρµ + ∂ρ Bµν )
3
√

d − 2 1 1
Z
= dd x −g R̃ − Tµ T µ + 2∇µ T µ + T̆µνρ T̆ µνρ − T̆νρµ T̆ µνρ
d−1 4 2
 
µνρ 1
+λ T̆µνρ − √ (∂µ Bνρ + ∂ν Bρµ + ∂ρ Bµν ) , (5.8)
3

15
where R denotes the scalar curvature of a torsionful but metric connection (cf. (2.13)), λµνρ is a
Lagrange multiplier, and Bµν is antisymmetric. The variation of (5.8) w.r.t. Tµ , Bµν , λµνρ , T̆µνρ
and g µν gives respectively
Tµ = 0, ∇µ λ[µνρ] = 0, (5.9)
1
T̆µνρ = √ (∂µ Bνρ + ∂ν Bρµ + ∂ρ Bµν ), (5.10)
3
1  νρµ 
λµνρ = T̆ + T̆ ρµν − T̆ µνρ , (5.11)
2
1 1 3
R̃µν − gµν R̃ + gµν T̆τ ρσ T̆ τ ρσ − T̆µ τ ρ T̆ντ ρ = 0, (5.12)
2 8 4
where we already used Tµ = 0 in (5.12). (5.10) implies that the traceless part of the torsion is
completely antisymmetric, and thus (5.11) reduces to
1
λµνρ = T̆ µνρ . (5.13)
2
Plugging this into the last eq. of (5.9) leads to

∇µ T̆ µνρ = 0. (5.14)

Finally, using (5.10) in (5.14) and (5.12), one gets precisely (5.3) and (5.4). The actions S1 and S2
describe therefore the same dynamics.

6 Discussion
Motivated by the interest in connections with torsion and nonmetricity both from the physical
and the mathematical point of view, we first generalized here some results that appeared previously
in the literature. In particular, we considered Einstein spaces with nonvanishing torsion that has
both a trace and a traceless part (Einstein-Cartan manifolds), and showed that the resulting field
equations are invariant under extended conformal transformations. We then compared our results
to Einstein manifolds with zero torsion but nonvanishing nonmetricity, where the latter is given in
terms of the Weyl vector Θµ (Einstein-Weyl spaces). We saw that, if the traceless part of the torsion
is set to zero, then the system of partial differential equations characterizing Einstein-Cartan spaces
exactly coincides with the Einstein-Weyl equations if the torsion trace Tµ is replaced by (d − 1)Θµ .
Subsequently, we extended our analysis to the case of Einstein manifolds with both torsion and
nonmetricity (Einstein-Cartan-Weyl spaces), allowing for both a trace and a traceless part of the
nonmetricity tensor.
Moreover, we considered actions involving scalar curvatures obtained from torsionful or non-
metric connections, and investigated their relations with other gravitational theories, obtaining
completely new results in this context. In particular, we analyzed a conformally (Weyl) invariant
action with torsion and its relation with scale invariant gravity, which involves a scalar φ, and
found that they reproduce the same dynamics. Furthermore, we have shown that the action (4.1)
implies that the spacetime is Einstein with torsion. Then, the Einstein-Hilbert action coupled to
a three-form field strength Hµνρ was considered, and it was shown that its equations of motion
imply that the manifold is Einstein with skew-symmetric torsion. Furthermore, it turned out that

16
the equations of motion of Einstein gravity coupled to a three-form may also be retrieved from a
constrained action that contains the scalar curvature of a connection with torsion. Let us stress
that in this paper we concentrated on the vacuum, without considering the presence of matter.
Among the solutions to Einstein’s field equations, Einstein spaces are of particular relevance
in physics, think for instance of the Kerr-(A)dS solution or of string compactifications on e.g.
Sasaki-Einstein manifolds. Since nature could accommodate for torsion and nonmetricity, it seems
reasonable to generalize the concept of Einstein spaces to torsionful and nonmetric connections.
The manifolds analyzed in this paper may also have applications in the classification and physical
study of (fake) supersymmetric supergravity solutions in the same way as Einstein-Weyl manifolds
provide the base space for fake supersymmetric solutions in de Sitter supergravity [47–52]. Under
the physical point of view, this analysis is particularly relevant in higher dimensions, since, in
d > 4, it is highly nontrivial to determine whether a given near-horizon geometry can be extended
to a full black hole solution (due to the fact that the strong uniqueness theorems that hold in four
dimensions [107–112] break down and there exist different black holes with the same asymptotic
charges and different black hole solutions with the same near-horizon geometry). Progress in
classifying near-horizon geometries can help to face this problem, as it was proven in [51], where
the authors, after having showed that a class of solutions of minimal supergravity in five dimensions
is given by lifts of three-dimensional Einstein-Weyl structures of hyper-CR type, considered the
task of reconstructing all supersymmetric solutions from such near-horizon geometry, demonstrating
that the moduli space of infinitesimal supersymmetric transverse deformations of the near-horizon
data is finite-dimensional if the spatial section of the horizon is compact.
Always in this context, a new result has recently been obtained in [113], where it has been shown
that the horizon geometry for supersymmetric black hole solutions of minimal five-dimensional
gauged supergravity is that of a particular Einstein-Cartan-Weyl structure in three dimensions,
involving the trace and traceless part of both torsion and nonmetricity, and obeying some precise
constraint; in the limit of zero cosmological constant, the set of nonlinear partial differential equa-
tions characterizing this Einstein-Cartan-Weyl structure reduces to that of a hyper-CR Einstein-
Weyl structure in the Gauduchon gauge, which was shown in [51] to be the horizon geometry in
the ungauged BPS (Bogomol’nyi-Prasad-Sommefield) case.
The analysis of this paper might also be extended in other directions. In particular, it would be
interesting to generalize the construction of [88] concerning the Chern-Simons formulation of three-
dimensional gravity involving torsion and nonmetricity, and the recent results presented in [114] in
the context of double field theory. One could also investigate possible generalizations of [104, 105].
On the other hand, a future development of our work may consist in possible generalizations
of the Jones-Tod correspondence [54] between selfdual conformal four-manifolds with a conformal
vector field and abelian monopoles on Einstein-Weyl spaces in three dimensions. Especially one
could ask whether Einstein-Cartan-Weyl manifolds can arise in a similar way by symmetry reduction
from higher dimensions.
Finally, a further direction for future research would be a geometrical investigation of the results
on unconventional supersymmetry presented recently in [115], where torsion plays a fundamental
role, under the perspective developed here.

17
A Riemann and Ricci tensors
The Riemann tensor of the Einstein-Cartan connection Γλµν introduced in section 2 reads

Rλρµν = ∂µ Γλνρ − ∂ν Γλµρ + Γλµσ Γσνρ − Γλνσ Γσµρ

1 h i
= R̃λρµν + gρν ∇µ T λ − gρµ ∇ν T λ + δµλ ∇ν Tρ − δνλ ∇µ Tρ
 d−1  
1 1 λ σ  1
λ σ

+ δ T −T̆σµρ + T̆µρσ − T̆ρµσ + δµ T −T̆σνρ + T̆νρσ + T̆ρνσ
d−1 2 ν 2
1 h λ i
+ T T̆ρµν − Tρ T̆ λµν
d − 1( )
1 1 σh    i
+ T gνρ T̆σ λ µ + T̆ λµσ + T̆µ λ σ − gµρ T̆σ λ ν + T̆ λνσ + T̆ν λ σ
d−1 2
1 h
λ λ

λ λ

σ

λ λ
i
+ gνρ Tµ T − gµρ Tν T + g ρµ δν − gρν δµ Tσ T + T ρ δ µ Tν − δν Tµ
(d − 1)2
1h λ σ       i
+ T̆ ν T̆µρσ + T̆µρσ T̆σλ ν + T̆νλ σ − T̆σµρ T̆ σλν + T̆ λν σ + T̆νλσ + T̆ λν σ + T̆νλσ T̆ρµσ
4
1h λ    i
+ T̆σ ν T̆ρµσ − T̆ρ λµ T̆νρσ + T̆ρνσ + T̆σνρ T̆ σλµ + T̆ λ µ σ + T̆µ λσ
4
1 h λ σ  i
− T̆ µ + T̆µ λσ T̆νρσ + T̆ρνσ
4
1h i
+ ∇µ T̆ λνρ + ∇µ T̆ν λ ρ + ∇µ T̆ρ λν − ∇ν T̆ λµρ − ∇ν T̆µ λ ρ − ∇ν T̆ρ λ µ ,
2
(A.1)

where R̃λρµν and ∇ denote respectively the Riemann tensor and the covariant derivative of the
Levi-Civita connection. The first line of (A.1) follows from the definition [Dµ , Dν ] ωρ + T σµν Dσ ωρ =
−Rλρµν ωλ , where D denotes the connection with coefficients Γ. The corresponding Ricci tensor is
given by
1 1
Rρν = Rµρµν = R̃ρν + [gρν ∇µ T µ + (d − 2)∇ν Tρ ] + [(2 − d)gνρ Tµ T µ + (d − 2)Tν Tρ ]
d−1 (d − 1)2
( )
1 1 µh   i
+ T (2 − d) T̆µνρ − T̆νρµ + (d − 4)T̆ρνµ
d−1 2
1 1 
+ T̆νµσ T̆ρµσ + T̆µνσ T̆ρµσ + ∇µ T̆ µνρ − ∇µ T̆νρ µ − ∇µ T̆ρν µ .
4 2
(A.2)

On the other hand, the Ricci tensor of the Einstein-Cartan-Weyl connection Γ̂λµν introduced in

18
section 3 is
1 1
R̂ρν = R̂µρµν = R̃ρν + [gρν ∇µ T µ + (d − 2)∇ν Tρ ] + [(2 − d)gνρ Tµ T µ + (d − 2)Tν Tρ ]
d−1 (d − 1)2
( )
1 1 µh   i
+ T (2 − d) T̆µνρ − T̆νρµ + (d − 4)T̆ρνµ
d−1 2
1 1 
+ T̆νµσ T̆ρµσ + T̆µνσ T̆ρµσ + ∇µ T̆ µνρ − ∇µ T̆νρ µ − ∇µ T̆ρν µ
4 2
+ (d − 2)Θν Θρ + (2 − d)Θµ Θµ gνρ + gνρ ∇µ Θµ + (d − 1)∇ν Θρ − ∇ρ Θν
1
+ [(d − 2)Θρ Tν + (d − 2)Θν Tρ + 2(2 − d)gνρ Θµ Tµ ]
d−1
d−2 µ  d−4
+ Θ T̆µνρ − T̆νρµ + Θµ T̆ρνµ ,
2 2
(A.3)

where ∇ denotes again the Levi-Civita connection.

ˆ
Finally, adding a traceless part to the nonmetricity tensor, we have that the Ricci tensor of ∇
reads, explicitly,
1 1
R̂ρν = R̂µρµν = R̃ρν + [gρν ∇µ T µ + (d − 2)∇ν Tρ ] + [(2 − d)gνρ Tµ T µ + (d − 2)Tν Tρ ]
d−1 (d − 1)2
( )
1 1 µh   i
+ T (2 − d) T̆µνρ − T̆νρµ + (d − 4)T̆ρνµ
d−1 2
1 1 
+ T̆νµσ T̆ρµσ + T̆µνσ T̆ρµσ + ∇µ T̆ µνρ − ∇µ T̆νρ µ − ∇µ T̆ρν µ
4 2
+ (d − 2)Θν Θρ + (2 − d)Θµ Θµ gνρ + gνρ ∇µ Θµ + (d − 1)∇ν Θρ − ∇ρ Θν
1
+ [(d − 2)Θρ Tν + (d − 2)Θν Tρ + 2(2 − d)gνρ Θµ Tµ ]
d−1
d−2 µ  d−4
+ Θ T̆µνρ − T̆νρµ + Θµ T̆ρνµ
2 ( 2 )
1 1 µh   i
+ T (2 − d) Q̆µνρ + Q̆νρµ + (d − 4)Q̆ρµν
d−1 2
d−2 µ  d−4
− Θ Q̆µνρ + Q̆νρµ + Θµ Q̆ρµν
2 2
1 1  µσ 
− Q̆µρσ Q̆µν σ + Q̆ν Q̆ρµσ − Q̆ν µσ Q̆ρσµ − ∇µ Q̆ν µρ + ∇µ Q̆νρ µ + ∇µ Q̆ρν µ
4 2
1 µ σ 1 µσ 1 1  
+ T̆ ρ Q̆µνσ − T̆ρ Q̆νµσ − T̆ν µσ Q̆ρµσ + T̆ µν σ Q̆ρσµ − Q̆ρµσ ,
2 2 2 2
(A.4)

which, indeed, now contains extra contributions from the traceless tensor Q̆λµν .

19
B Acknowledgements
L. R. acknowledges illuminating discussions with L. Andrianopoli, B. L. Cerchiai, M. Trigiante
and J. Zanelli. D. K. is supported partly by INFN.

References
[1] F. W. Hehl, J. D. McCrea, E. W. Mielke and Y. Ne’eman, “Metric affine gauge theory of gravity:
Field equations, Noether identities, world spinors, and breaking of dilation invariance,” Phys.
Rept. 258 (1995) 1 [gr-qc/9402012].

[2] E. Schrodinger, “Space-time Structure,” Cambridge 1954.

[3] A. Einstein, “Meaning of Relativity,” Princeton 1955.

[4] J.A. Schouten, “Ricci Calculus,” Springer, Berlin 1954.

[5] M. Blagojevic and F. Hehl (eds.), “Gauge Theories of Gravitation,” Imperial College Press,
London 2013.

[6] Y. Ne’eman and F. W. Hehl, “Test matter in a space-time with nonmetricity,” Class. Quant.
Grav. 14 (1997) A251 [gr-qc/9604047].

[7] R. F. Sobreiro and V. J. Vasquez Otoya, “Aspects of nonmetricity in gravity theories,” Braz.
J. Phys. 40 (2010) 370 [arXiv:0711.0020 [hep-th]].

[8] D. Puetzfeld and Y. N. Obukhov, “Equations of motion in metric-affine gravity: A covariant

unified framework,” Phys. Rev. D 90 (2014) no.8, 084034 [arXiv:1408.5669 [gr-qc]].

[9] L. Järv, M. Rünkla, M. Saal and O. Vilson, “Nonmetricity formulation of general relativity and
its scalar-tensor extension,” Phys. Rev. D 97 (2018) no.12, 124025 [arXiv:1802.00492 [gr-qc]].

[10] E. Kröner, “The differential geometry of elementary point and line defects in Bravais crystals,”
Intern. J. Theor. Phys. 29 (1990) 1219.

[11] A. Delhom-Latorre, G. J. Olmo and M. Ronco, “Observable traces of nonmetricity: New

constraints on metric-affine gravity,” Phys. Lett. B 780 (2018) 294 [arXiv:1709.04249 [hep-th]].

[12] P. Luz and V. Vitagliano, “Raychaudhuri equation in spacetimes with torsion,” Phys. Rev. D
96 (2017) no.2, 024021 [arXiv:1709.07261 [gr-qc]].

[13] S. Speziale, “Raychaudhuri and optical equations for null geodesic congruences with torsion,”
Phys. Rev. D 98 (2018) no.8, 084029 [arXiv:1808.00952 [gr-qc]].

[14] D. Iosifidis, C. G. Tsagas and A. C. Petkou, “Raychaudhuri equation in spacetimes with torsion
and nonmetricity,” Phys. Rev. D 98 (2018) no.10, 104037 [arXiv:1809.04992 [gr-qc]].

[15] C. Pagani and R. Percacci, “Quantum gravity with torsion and nonmetricity,” Class. Quant.
Grav. 32 (2015) no.19, 195019 [arXiv:1506.02882 [gr-qc]].

20
[16] J. Stelmach, “Nonmetricity driven inflation,” Class. Quant. Grav. 8 (1991) 897.

[17] E. A. Poberii, “Nonmetricity driven inflation,” Helv. Phys. Acta 67 (1994) 745.

[18] J. Beltrán Jiménez and T. S. Koivisto, “Modified gravity with vector distortion and cosmo-
logical applications,” Universe 3 (2017) no.2, 47.

[19] R. Adler, M. Bazin, M. Schiffer, “Introduction to general relativity,” Mc.Graw-Hill 1965.

[20] D. M. J. Calderbank and H. Pedersen, “Einstein-Weyl geometry,” in Essays on Einstein Man-

ifolds (eds. C. R. LeBrun and M. Wang), Surveys in Differential Geometry, vol. V, International
Press.

[21] G. B. Folland, “Weyl manifolds,” J. Diff. Geom. 4 (1970) 145.

[22] C. Romero, J. B. Fonseca-Neto and M. L. Pucheu, “General relativity and Weyl geometry,”
Class. Quant. Grav. 29 (2012) 155015 [arXiv:1201.1469 [gr-qc]].

[23] H. Weyl, “Gravitation and electricity,” Sitzungsber. Preuss. Akad. Wiss. Berlin (Math. Phys.)
1918 (1918) 465.

[24] H. Weyl, “A new extension of relativity theory,” Annalen Phys. 364 (1919) 101 [Surveys High
Energ. Phys. 5 (1986) 237] [Annalen Phys. 364 (1919) 101].

[25] J. T. Wheeler, “Weyl geometry,” Gen. Rel. Grav. 50 (2018) no.7, 80 [arXiv:1801.03178 [gr-qc]].

[26] F. London, “Quantum mechanical interpretation of the Weyl theory,” Z. Phys. 42 (1927) 375
[Surveys High Energ. Phys. 5 (1986) 253].

[27] L. O’Raifeartaigh, “The Dawning of Gauge Theory” (Princeton University Press, Princeton,
NJ, 1997), p. 249.

[28] E. Scholz, “The unexpected resurgence of Weyl geometry in late 20-th century physics,” Ein-
stein Stud. 14 (2018) 261 [arXiv:1703.03187 [math.HO]].

[29] C. Barceló, R. Carballo-Rubio and L. J. Garay, “Weyl relativity: A novel approach to Weyl’s
ideas,” JCAP 1706 (2017) no.06, 014 [arXiv:1703.06355 [gr-qc]].

[30] M. R. Tanhayi, S. Dengiz and B. Tekin, “Weyl-Invariant Higher Curvature Gravity Theories
in n Dimensions,” Phys. Rev. D 85 (2012) 064016 [arXiv:1201.5068 [hep-th]].

[31] S. Dengiz, E. Kilicarslan and B. Tekin, “Weyl-gauging of Topologically Massive Gravity,”

Phys. Rev. D 86 (2012) 104014 [arXiv:1209.1251 [hep-th]].

[32] S. Dengiz, “A noncompact Weyl-Einstein-Yang-Mills model: A semiclassical quantum gravity,”

Annals Phys. 383 (2017) 560 [arXiv:1609.02475 [hep-th]].

[33] I. P. Lobo and C. Romero, “Experimental constraints on the second clock effect,” Phys. Lett.
B 783, 306 (2018) [arXiv:1807.07188 [gr-qc]].

[34] K. P. Tod, “Compact 3-dimensional Einstein-Weyl structures,” J. London Math. Soc. s2-45
(1992), 341.

21
[35] H. Pedersen and K. P. Tod, “Einstein metrics and hyperbolic monopoles,” Class. Quantum
Grav. 8 (1991), 751.

[36] H. Pedersen and K. P. Tod, “Three-dimensional Einstein-Weyl geometry,” Adv. Math. 97

(1993), 74.

[37] H. Pedersen and A. Swann, “Riemannian submersions, four-manifolds and Einstein-Weyl ge-
ometry,” Proc. London Math. Soc. s3-66 (1993), 381.

[38] H. Pedersen and A. F. Swann, “Einstein-Weyl geometry, the Bach tensor and conformal scalar
curvature,” J. reine angew. Math. 441 (1993), 99.

[39] H. Pedersen, Y. S. Poon and A. F. Swann, “The Einstein-Weyl equations in complex and
quaternionic geometry”, Differential Geometry and its Applications 3 (1993), 309.

[40] H. Pedersen, Y. S. Poon and A. F. Swann, “The Hitchin-Thorpe inequality for Einstein-Weyl
manifolds”, Bull. London Math. Soc. 26 (1994), 191.

[41] H. Pedersen, Y. S. Poon and A. F. Swann, “Einstein-Weyl deformations and submanifolds,”

International J. Math. 7 (1996), 705.

[42] A. B. Madsen, H. Pedersen, Y. S. Poon and A. F. Swann, “Compact Einstein-Weyl manifolds

with large symmetry group,” Duke Math. J. 88 (1997), 407. Corrigendum: Duke Math. J. 100
(1999), 167.

[43] D. M. J. Calderbank and P. Tod, “Einstein metrics, hypercomplex structures and the Toda
field equation,” Differ. Geom. Appl. 14 (2001) 199 [math/9911121 [math-dg]].

[44] D. M. J. Calderbank, “Selfdual Einstein metrics and conformal submersions,” math/0001041

[math-dg].

[45] D. M. J. Calderbank and M. Dunajski, “Scalar-flat Lorentzian Einstein-Weyl spaces,” Class.

Quant. Grav. 18 (2001) L77 [math/0104208 [math-dg]].

[46] D. M. J. Calderbank and H. Pedersen, “Selfdual spaces with complex structures, Einstein-Weyl
geometry and geodesics,” Ann. Inst. Fourier (Grenoble) 50 (2000), 921 [math.DG/9911117].

[47] P. Meessen and A. Palomo-Lozano, “Cosmological solutions from fake N = 2 EYM supergrav-
ity,” JHEP 0905 (2009) 042 [arXiv:0902.4814 [hep-th]].

[48] J. B. Gutowski and W. A. Sabra, “Solutions of minimal four-dimensional de Sitter supergrav-

ity,” Class. Quant. Grav. 27 (2010) 235017 [arXiv:0903.0179 [hep-th]].

[49] J. Grover, J. B. Gutowski, C. A. R. Herdeiro, P. Meessen, A. Palomo-Lozano and W. A. Sabra,

“Gauduchon-Tod structures, Sim holonomy and de Sitter supergravity,” JHEP 0907 (2009) 069
[arXiv:0905.3047 [hep-th]].

[50] J. B. Gutowski, A. Palomo-Lozano and W. A. Sabra, “Einstein-Weyl structures and de Sitter

supergravity,” Class. Quant. Grav. 29 (2012) 105006 [arXiv:1109.5257 [hep-th]].

22
[51] M. Dunajski, J. Gutowski and W. Sabra, “Einstein-Weyl spaces and near-horizon geometry,”
Class. Quant. Grav. 34 (2017) no.4, 045009 [arXiv:1610.08953 [hep-th]].

[52] M. Randall, “Three-dimensional near-horizon metrics that are Einstein-Weyl,”

arXiv:1704.06433 [math.DG].

[53] N. J. Hitchin, “Complex manifolds and Einstein equations, twistor geometry and nonlinear
systems” (eds H. D. Doebner and T. D. Palev), Lecture Notes in Math. 970, Springer, Berlin,
1982, 79.

[54] P. Jones and K. Tod, “Minitwistor spaces and Einstein-Weyl spaces,” Class. Quant. Grav. 2
(1985) no.4, 565.

[55] R. S. Ward, “Einstein-Weyl spaces and SU(∞) Toda fields,” Class. Quant. Grav. 7, L95 (1990).

[56] M. Dunajski, L. J. Mason and P. Tod, “Einstein-Weyl geometry, the dKP equation and twistor
theory,” J. Geom. Phys. 37 (2001) 63 [math/0004031 [math-dg]].

[57] É. Cartan, “Sur une généralisation de la notion de courbure de Riemann et les espaces à
torsion,” C. R. Acad. Sci. Paris 174 (1922), 593.

[58] É. Cartan, “Sur les variétés à connexion affine et la théorie de la relativité généralisée
(premiÚre partie),” Ann. Éc. Norm. Sup 40 (1923), 325.

[59] É. Cartan, “Sur les variétés à connexion affine et la théorie de la relativité généralisée (première
partie) (Suite),” Ann. Éc. Norm. Sup 41 (1924), 1.

[60] É. Cartan, “Sur les variétés à connexion affine et la théorie de la relativité généralisée II,”
Ann. Éc. Norm. Sup 42 (1925), 17.

[61] É. Cartan, “Sur les variétés à connexion affine et la théorie de la relativité généralisée,” (Gau-
thier Villars, 1955). See also “On manifolds with an affine connection and the theory of general
relativity”, transl. A. Magnon and A. Ashtekar (Bibliopolis, Napoli, 1986).

[62] Y. F. Cai, S. Capozziello, M. De Laurentis and E. N. Saridakis, “f(T) teleparallel gravity and
cosmology,” Rept. Prog. Phys. 79 (2016) no.10, 106901 [arXiv:1511.07586 [gr-qc]].

[63] D. W. Sciama, “Recent developments in general relativity”, Festschrift for Infeld, Pergamon,
Oxford (1962), 415.

[64] D. W. Sciama, “The Physical structure of general relativity,” Rev. Mod. Phys. 36 (1964) 463
Erratum: [Rev. Mod. Phys. 36 (1964) 1103].

[65] T. W. B. Kibble, “Lorentz invariance and the gravitational field,” J. Math. Phys. 2 (1961)
212.

[66] F. W. Hehl, P. Von Der Heyde, G. D. Kerlick and J. M. Nester, “General relativity with spin
and torsion: Foundations and prospects,” Rev. Mod. Phys. 48 (1976) 393.

[67] I. L. Shapiro, “Torsion: Theory and possible observables,” hep-th/9811072.

23
[68] I. L. Shapiro, “Physical aspects of the space-time torsion,” Phys. Rept. 357 (2002) 113 [hep-
th/0103093].

[69] S. Capozziello, G. Lambiase and C. Stornaiolo, “Geometric classification of the torsion tensor
in space-time,” Annalen Phys. 10 (2001) 713 [gr-qc/0101038].

[70] T. Watanabe and M. J. Hayashi, “General relativity with torsion,” gr-qc/0409029.

[71] S. Jensen, “General relativity with torsion: Extending Wald’s chapter on curvature,” 2005.

[72] A. Golovnev, “Introduction to teleparallel gravities,” arXiv:1801.06929 [gr-qc].

[73] J. Scherk and J. H. Schwarz, “Dual models and the geometry of spacetime,” Phys. Lett. 52B
(1974) 347.

[74] A. Saa, “Strings in background fields and Einstein-Cartan theory of gravity,” Class. Quant.
Grav. 12 (1995) L85 [hep-th/9307095].

[75] M. Kalb and P. Ramond, “Classical direct interstring action,” Phys. Rev. D 9 (1974) 2273.

[76] Y. N. Obukhov, “Conformal invariance and spacetime torsion,” Phys. Lett. A 90 (1982) 13.

[77] J. W. Maluf, “Conformal invariance and torsion in general relativity,” Gen. Rel. Grav. 19
(1987) 57.

[78] G. German, “Brans-Dicke type models with torsion,” Phys. Rev. D 32 (1985) 3307.

[79] J. A. Helayel-Neto, A. Penna-Firme and I. L. Shapiro, “Conformal symmetry, anomaly and

effective action for metric-scalar gravity with torsion,” Phys. Lett. B 479 (2000) 411 [gr-
qc/9907081].

[80] L. Fabbri, “Metric-torsional conformal gravity,” Phys. Lett. B 707 (2012) 415 [arXiv:1101.1761
[gr-qc]].

[81] S. Chakrabarty and A. Lahiri, “Different types of torsion and their effect on the dynamics of
fields,” Eur. Phys. J. Plus 133 (2018) no.6, 242.

[82] L. L. Smalley, “Brans-Dicke type models with nonmetricity,” Phys. Rev. D 33 (1986) 3590.

[83] T. Dereli and R. W. Tucker, “A broken gauge approach to gravitational mass and charge,”
JHEP 0203 (2002) 041 [hep-th/0112123].

[84] W. Drechsler and D. Hartley, “The role of the internal metric in generalized Kaluza-Klein
theories,” J. Math. Phys. 35 (1994) 3571.

[85] T. Y. Moon, J. Lee and P. Oh, “Conformal invariance in Einstein-Cartan-Weyl space,” Mod.
Phys. Lett. A 25 (2010) 3129 [arXiv:0912.0432 [gr-qc]].

[86] D. Iosifidis and T. Koivisto, “Scale transformations in metric-affine geometry,”

arXiv:1810.12276 [gr-qc].

24
[87] D. Iosifidis, “Metric-Affine Gravity and Cosmology/Aspects of Torsion and non-Metricity in
Gravity Theories,” arXiv:1902.09643 [gr-qc].

[88] S. L. Cacciatori, M. M. Caldarelli, A. Giacomini, D. Klemm and D. S. Mansi, “Chern-Simons

formulation of three-dimensional gravity with torsion and nonmetricity,” J. Geom. Phys. 56
(2006) 2523 [hep-th/0507200].

[89] A. N. Ivanov and M. Wellenzohn, “Einstein-Cartan gravity with torsion field serving as an
origin for the cosmological constant or dark energy density,” Astrophys. J. 829 (2016) no.1, 47
[arXiv:1607.01128 [gr-qc]].

[90] R. Percacci, “Gravity from a particle physicists’ perspective,” PoS ISFTG (2009) 011
[arXiv:0910.5167 [hep-th]].

[91] A. C. Ferreira, “Einstein four-manifolds with skew torsion,” J. Geom. Phys. 61 (2011) 2341.

[92] I. Agricola, A. C. Ferreira, “Einstein manifolds with skew torsion,” The Quarterly Journal of
Mathematics 65 (2014), Issue 3, 717 [arXiv:1209.5886 [math.DG]].

[93] P. A. M. Dirac, “New basis for cosmology,” Proc. Roy. Soc. Lond. A 165 (1938) 199.

[94] P. A. M. Dirac, “Long range forces and broken symmetries,” Proc. Roy. Soc. Lond. A 333
(1973) 403.

[95] P. D. Mannheim, “Conformal cosmology with no cosmological constant,” Gen. Rel. Grav. 22
(1990) 289.

[96] P. K. Aluri, P. Jain and N. K. Singh, “Dark energy and dark matter in general relativity with
local scale invariance,” Mod. Phys. Lett. A 24 (2009) 1583 [arXiv:0810.4421 [hep-ph]].

[97] L. P. Chimento, A. S. Jakubi and D. Pavon, “An exact model of conformal quintessence,”
Phys. Rev. D 69 (2004) 083511 [astro-ph/0401016].

[98] C. Wetterich, “Conformal fixed point, cosmological constant and quintessence,” Phys. Rev.
Lett. 90 (2003) 231302 [hep-th/0210156].

[99] E. Gunzig, A. Saa, L. Brenig, V. Faraoni, T. M. Rocha Filho and A. Figueiredo, “Su-
perinflation, quintessence, and nonsingular cosmologies,” Phys. Rev. D 63 (2001) 067301 [gr-
qc/0012085].

[100] V. Faraoni, “Superquintessence,” Int. J. Mod. Phys. D 11 (2002) 471 [astro-ph/0110067].

[101] B. Kelleher, “Scale invariant gravity: Spacetime recovered,” Class. Quant. Grav. 21 (2004)
483 [gr-qc/0307091].

[102] B. Kelleher, “Scale invariance in gravity and implications for the cosmological constant,”
Class. Quant. Grav. 21 (2004) 2623 [gr-qc/0310109].

[103] D. Iosifidis, A. C. Petkou and C. G. Tsagas, “Torsion/non-metricity duality in f(R) gravity,”

Gen. Relativ. Gravit. 1055 51, 66 (2019) arXiv:1810.06602 [gr-qc].

25
[104] N. Boulanger and I. Kirsch, “A Higgs mechanism for gravity. Part II. Higher spin connec-
tions,” Phys. Rev. D 73 (2006) 124023 [hep-th/0602225].

[105] P. Baekler, N. Boulanger and F. W. Hehl, “Linear connections with propagating spin-3 field
in gravity,” Phys. Rev. D 74 (2006) 125009 [hep-th/0608122].

[106] P. Gauduchon and K. P. Tod, “Hyper-Hermitian metrics with symmetry,” J. Geom. Phys.
25 (1998) 291.

[107] S. W. Hawking, “Black holes in general relativity,” Commun. Math. Phys. 25 (1972) 152.

[108] W. Israel, “Event horizons in static vacuum space-times,” Phys. Rev. 164 (1967) 1776.

[109] B. Carter, “Axisymmetric black hole has only two degrees of freedom,” Phys. Rev. Lett. 26
(1971) 331.

[110] D. C. Robinson, “Uniqueness of the Kerr black hole,” Phys. Rev. Lett. 34 (1975) 905.

[111] W. Israel, “Event horizons in static electrovac space-times,” Commun. Math. Phys. 8 (1968)
245.

[112] P. O. Mazur, “Proof of uniqueness of the Kerr-Newman black hole solution,” J. Phys. A 15
(1982) 3173.

[113] D. S. Klemm and L. Ravera, “Supersymmetric near-horizon geometry and Einstein-Cartan-

Weyl spaces,” Phys. Lett. B 793 (2019) 265 [arXiv:1904.03681 [hep-th]].

[114] V. A. Penas, “Deformed Weitzenböck connections and double field theory,” Fortsch. Phys.
67, 1800077 (2019) arXiv:1807.01144 [hep-th].

[115] L. Andrianopoli, B. L. Cerchiai, R. D’Auria and M. Trigiante, “Unconventional supersymme-

try at the boundary of AdS4 supergravity,” JHEP 1804 (2018) 007 [arXiv:1801.08081 [hep-th]].

26