Lorentzian manifolds and scalar curvature invariants
aa r X i v : . [ g r- q c ] M a r Lorentzian manifolds and scalar curvature invariants
Alan Coley , Sigbjørn Hervik , Nicos Pelavas Department of Mathematics and Statistics,Dalhousie University, Halifax, Nova Scotia,Canada B3H 3J5 Faculty of Science and Technology,University of Stavanger,N-4036 Stavanger, Norway aac, [email protected], [email protected] (Dated: September 1, 2018)We discuss (arbitrary-dimensional) Lorentzian manifolds and the scalar polynomial curvature in-variants constructed from the Riemann tensor and its covariant derivatives. Recently, we have shownthat in four dimensions a Lorentzian spacetime metric is either I -non-degenerate, and hence locallycharacterized by its scalar polynomial curvature invariants, or is a degenerate Kundt spacetime.We present a number of results that generalize these results to higher dimensions and discuss theirconsequences and potential physical applications. I. INTRODUCTION
We address the question of when a Lorentzian manifold (in arbitrary dimensions) can be uniquely characterized(locally) by its scalar polynomial curvature invariants, which are scalars obtained by contraction from a polynomialin the Riemann tensor and its covariant derivatives. This question is not only of mathematical interest, but is also offundamental physical import. We begin by introducing some necessary mathematical terminology and machinery.For a spacetime ( M , g ), a (one-parameter) metric deformation , ˆ g τ , τ ∈ [0 , ǫ ), is a family of smooth metrics on M such that ˆ g τ is continuous in τ , ˆ g = g , and ˆ g τ for τ >
0, is not diffeomorphic to g . We define the set of all scalarpolynomial curvature invariants on ( M , g ) by I ≡ {
R, R µν R µν , R µν R µπ R πν , . . . , C µναβ C µναβ , . . . , R µναβ ; γ R µναβ ; γ , . . . , R µναβ ; γδ R µναβ ; γδ , . . . } . If there does not exist a metric deformation of g having the same set of invariants as g , then we will call the set ofinvariants non-degenerate , and the spacetime metric g will be called I -non-degenerate [1]. Therefore, for a metricwhich is I -non-degenerate the invariants locally characterize the spacetime completely.The Kundt class of spacetimes in n dimensions is defined by those metrics admitting a null vector that is geodesic,expansion-free, shear-free and twist-free. A Kundt metric can be written in the canonical form ( i = 3 , ...n ) [2, 3]d s = 2d u (cid:2) d v + H ( v, u, x k )d u + W i ( v, u, x k )d x i (cid:3) + h ij ( u, x k )d x i d x j . (1)A degenerate Kundt spacetime is a spacetime in which there exists a kinematic null frame (in which the appropriateRicci rotation coefficients L ij are zero [2]) such that all of the positive boost weight (b.w.) terms of the Riemanntensor and all of its covariant derivatives ∇ ( k ) ( Riem ) are zero (in this common frame) [4]. That is, a degenerateKundt spacetime is an aligned algebraically special Riemann type II and aligned algebraically special ∇ ( k ) ( Riem ) type II Kundt spacetime. In terms of the metric (1), written in the canonical (kinematic) frame, the condition that theRiemann tensor is aligned and of algebraically special type II implies that W i,vv = 0, and the condition that ∇ ( Riem )is aligned and of algebraically special type II implies that H ,vvv = 0 (whence it follows that all of ∇ ( k ) ( Riem ) ( k > II ) [4]. We note that the important constant curvature invariant ( CSI )spacetimes [3] and vanishing scalar invariant (
V SI ) spacetimes [5] are degenerate Kundt spacetimes.Let us briefly review the results of [1], particularly those that have generalizations to higher dimensions, in whichthe class of four-dimensional (4D) Lorentzian manifolds that can be completely characterized by the scalar polynomialcurvature invariants constructed from the Riemann tensor and its covariant derivatives was determined. The importantresult that a spacetime metric is either I -non-degenerate or the metric is a degenerate Kundt metric was proven. Thistheorem was proven on a case-by-case basis, depending on the algebraic type, using a b.w. decomposition and, mostimportantly, by determining an appropriate set of projection operators from the Riemann tensor and its covariantderivatives. We recall that the 4D Lorentzian manifolds are characterized algebraically by their Petrov and Segre [6]types or, alternatively, in terms of their Ricci, Weyl (and Riemann) types [2, 7]. A. Coley, S. Hervik and N. PelavasIt is useful to state a number of partial results of when a spacetime metric is I -non-degenerate (which is how thetheorem in 4D was actually proven), which will be exploited in obtaining and stating results in higher dimensions.First, it was proven that if a 4D spacetime metric is locally of Ricci type I or Weyl type I (i.e., algebraically general)the metric is I -non-degenerate [1]. This indicates that, in general, a spacetime metric is I -non-degenerate and themetric is locally determined by its curvature invariants. For the algebraically special cases the Riemann tensor itselfdoes not provide sufficient information to determine all of the required projection operators, and it is necessary toalso consider the covariant derivatives. In terms of the b.w. decomposition, for an algebraically special metric (whichhas a Riemann tensor with zero positive b.w. components) which is not Kundt, by taking covariant derivatives of theRiemann tensor positive boost weight components are found and a set of higher derivative projection operators aredetermined. Therefore, if the 4D spacetime metric is algebraically special, but ∇ R , ∇ (2) R , ∇ (3) R , or ∇ (4) R is of type I or more general, the metric is I -non-degenerate. The remaining metrics which do not acquire positive boost weight components when taking covariant derivativeshave a very special curvature structure; indeed, they are degenerate Kundt metrics [4]. This implies that metricsnot determined by their curvature invariants must be of degenerate Kundt form; i.e., degenerate Kundt metrics arenot I -non-degenerate . This exceptional property of the the degenerate Kundt metrics essentially follows from thefact that they do not define a unique timelike curvature operator. The degenerate Kundt spacetimes are classifiedalgebraically by the Riemann tensor and its covariant derivatives in the aligned kinematic frame [4].We note parenthetically that these results are of importance to the equivalence problem of characterizing Lorentzianspacetimes (in terms of their Cartan invariants) [6]. Clearly, by knowing which spacetimes can be characterized bytheir scalar curvature invariants alone, the computations of the invariants (i.e., simple scalar invariants) is much morestraightforward and can be done algorithmically. On the other hand, the Cartan equivalence method also contains,at least in principle, the conditions under which the classification is complete. II. HIGHER DIMENSIONS
Let us now consider higher dimensions. The results are proven on a case-by-case basis in terms of algebraic types;hence we need to utilize the higher dimensional algebraic classification of tensors [2, 7]. We note that recently anumber of exact higher dimensional algebraically special spacetimes have been studied [8]. In particular, similarresults to those that occur in 4D (discussed above) occur in the algebraically general cases. For Ricci type I or Gthe proof is essentially identical to the 4D case. We know the Segre types in higher dimensions and for Ricci type Ithey are of the form { , . . . } , { , (11)1 . . . } , { , (111) . . . } etc., or { z ¯ z . . . } , { z ¯ z (11) . . . } , etc. We note that thestabilizer of a Ricci type I tensor is always contained in the compact group O ( n − O ( n − Theorem II.1.
If a spacetime metric is of Ricci type I or G, the metric is I -non-degenerate. We can provide a compelling argument that in higher dimensions a general Weyl type I or G spacetime is I -non-degenerate. A bivector formalism in higher dimensions is needed to characterize (invariantly) the curvature operatorsof the Weyl tensor [9]. In general, the Weyl tensor can be decomposed using the eigenspace projection operators as: C = N + X A λ A ⊥ A , where N is nilpotent. In the case of the Weyl tensor, each of the projection operators, ⊥ A , are of type D with respectto a certain frame. Now, if all of the projection operators are of type D with respect to the same frame, then theWeyl tensor is of type II or simpler. Therefore, for the Weyl tensor to be of type I/G, these frames cannot be aligned.Each of these Weyl projection operators can be used to construct a projection operator of the tangent space of sometype { (1 , .. .. } . Since these are not aligned, and the fact that Weyl tensor is of type I/G (and not simpler), bysuccessive projections we can isolate a timelike direction and construct a timelike projection operator. It then followsthat a Weyl type I or G spacetime is I -non-degenerate . Therefore, we have that: Conjecture II.2.
If the Weyl type is I or G, then the spacetime is I -non-degenerate. The first step in providing a rigorous proof to this result is to investigate the curvature operators in higher dimensionsand to classify these for the various algebraic types [9]. Indeed, with the aid of these operators is it then hoped thatall of the results obtained in 4D can also be shown to be true in higher dimensions. In addition, such a formalismmay lead to simpler proofs of some of the results outlined below.orentzian manifolds and scalar curvature invariants 3Finally, the result that a spacetime is I -non-degenerate follows for any curvature operator (not just those constructedfrom the Ricci or Weyl tensors) that is of general algebraic type I or G: Corollary II.3.
If any curvature operator is of general algebraic type I or G, then the spacetime is I -non-degenerate. We have thus far presented some general results; that is, in general (defined algebraically) a spacetime is I -non-degenerate. It is also possible to present a very special result; namely, that a spacetime which is degenerate Kundt isnot I -non-degenerate. By construction, all degenerate Kundt spacetimes with the same boost weight zero terms butwith different negative b.w. terms will have precisely the same set of scalar curvature invariants since no negativeboost-weight terms can appear in any scalar polynomial invariant in a degenerate Kundt spacetime (see theorem II.7below). Therefore, we have that: Theorem II.4.
A spacetime which is degenerate Kundt is not I -non-degenerate. In higher dimensions the intermediate cases are much harder to deal with (than in the 4D case). Let us first presenta partial result. In the analysis in 4D it was determined for which Segre types for the Ricci tensor the spacetime is I -non-degenerate (similar results were obtained for the Weyl tensor). In each case, it was found that the Ricci tensor,considered as a curvature operator, admits a timelike eigendirection. Therefore, if a spacetime is not I -non-degenerate,its Ricci tensor must be of a particular Segre type (corresponding to the non-existence of a unique timelike direction).Therefore, consider the following algebraic types for the Ricci tensor (or any other (0 ,
2) curvature operator writtenin ‘Segre form’):1. { ... } , { ... } , { ... } ,..., { ... ) } ,2. { (21)11 ... } , { (21)(11)1 ... } , { (21)(111) ... } ,..., { (21)(111 ... ) } ,3. { (211)11 ... } , { (211)(11)1 ... } , { (211)(111) ... } ,..., { (211)(111 ... ) } ,4. { ... } , { ... } , { ... } ,..., { ... ) } , ...It follows that if the Ricci tensor (or any (0 ,
2) curvature operator – and similar results are true in terms of theWeyl tensor in bivector form) is not of one of these types (for example, of type { , ... } ), then the spacetime is I -non-degenerate. It is plausible that if the Ricci tensor and all other curvature operators are all of one of these types,then the spacetime is degenerate Kundt and not I -non-degenerate.In addition, in higher dimensions suppose there exists a frame in which all of the positive b.w. terms of the Riemanntensor and all of its covariant derivatives ∇ ( k ) ( Riem ) are zero (in this frame) (i.e., the spacetime is of ‘type II to allorders’), it is plausible that the resulting spacetime is degenerate Kundt (i.e., the appropriate Ricci rotation coefficients L ij are zero [2]). This is true in 4D (as a result of the theorems of [1]). It is likely true in higher dimensions, butthis might be difficult to prove in general. In particular, we would like to prove that the degenerate Kundt metricsare the only metrics not determined by their curvature invariants (i.e., not I -non-degenerate) in any dimension. It ishoped that higher dimensional generalizations of all of the 4D Theorems can be proven with the aid of Weyl operatorsin higher dimensions [9]. However, we do note that many of the results in [1, 4, 10] can be generalized to higherdimensions.Let us present a number of partial results. First, two higher dimensional results were proven in [4]. Let K n denotethe subclass of Kundt metrics such that there exists a kinematic frame in which the Riemann tensor up to andincluding its n th covariant derivative have vanishing positive b.w. components. Theorem II.5.
In the higher-dimensional Kundt class, K implies K n for all n ≥ . Theorem II.6.
If for a spacetime, ( M , g ) , the Riemann tensor and all of its covariant derivatives ∇ ( k ) ( Riem ) aresimultaneously of type D (in the same frame), then the spacetime is degenerate Kundt . Note that the degenerate Kundt spacetimes can be written in the form (1), where H ( v, u, x k ) = v H (2) ( u, x k ) + vH (1) ( u, x k ) + H (0) ( u, x k ) , W i ( v, u, x k ) = vW (1) i ( u, x k ) + W (0) i ( u, x k ) . Let us next present a new result:
Theorem II.7.
For a degenerate Kundt spacetime the boost weight 0 components of all curvature tensors are identicalto the corresponding Kundt spacetime where H (1) ( u, x k ) = H (0) ( u, x k ) = W (0) ( u, x k ) = 0 . Consequently, theircurvature invariants will also be identical. A. Coley, S. Hervik and N. Pelavas
Proof.
Let us introduce the null-frame (Kundt frame): ℓ = d u, n = d v + ( H (2) v + H (1) v + H (0) )d u + ( W (1) i v + W (0) )d x i , m i = e ij ( u, x k )d x j , δ ij m i m j = g ij ( u, x k )d x i d x j (2)where the functions H (2) , H (1) , H (0) , W (1) i and W (0) do not depend on v . By calculating the Riemann tensor withrespect to this null-frame, we get that the Riemann tensor, R , has the following b.w. decomposition: R = ( R ) + ( R ) − + ( R ) − , (3)where the components of ( R ) do not depend on v and involve only the functions H (2) and W (1) i and the transversemetric g ij ( u, x k ).Consider now the n th-covariant derivatives, symbolically written ∇ ( n ) R . By using the Kundt frame, a covariantderivative of an arbitrary covariant tensor T can be written symbolically: ∇ T = ∂T − X Γ ∗ T, (4)where ∂ are partial derivatives with respect to the frame, and Γ are the connection coefficients. In the Kundt framethe connection coefficents of positive b.w. are all zero; consequently, the piece P Γ ∗ T cannot raise the b.w. Regardingthe partial derivatives, ℓ µ ∂ µ ≡ ∂ v and n raises and lowers the b.w., respectively. The partial derivatives with respectto m i are of b.w. 0. Therefore, if T is of boost-order 0, then the b.w. +1 and 0 components of ∇ T will be: • b.w. +1: ∂ v ( T ) • b.w. 0: (Γ) ( T ) , m i ( T ) , ∂ v ( T ) − First, from theorem II.5, we note that since these are K they are K n , implying ( ∇ ( n ) R ) b> = 0; i.e., all positive b.w.components vanish. The highest b.w. terms are thus the b.w. 0 components. We thus need to consider the b.w. 0components in more detail.Consider first ∇ R . We note that the terms (Γ) ( R ) , m i ( R ) all give the desired result. However, we need toinvestigate ∂ v ( R ) − in more detail. This comes from the covariant derivative ℓ µ ∇ µ ( R ) − ≡ ∇ + ( R ) − . Now, weobserve that the Bianchi identity R µν ( αβ ; δ ) = 0 enables us to rewrite the troublesome derivatives ∇ + ( R ) − in termsof derivatives with respect to n and m i : R − ijk ;+ = − R − ik +; j − R − i + j ; k , R − + − i ;+ = − R − + i +; − − R − ++ − ; i . Consequently, the components of ∇ + ( R ) − can be written in terms of other well-behaving b.w. 0 terms.Let us now show that we can always write the b.w. 0 terms ∇ + ( ∇ ( n ) R ) − , in terms of well-behaving b.w. 0components of lower or equal order. We will show this by induction; therefore, assume it is true for ∇ + ( ∇ ( n ) R ) − , n ≥
1. These components have the form R µναβ ; δ ...δ n where, by the induction assumption, δ n = {− , i } . We need tocheck the components of ∇ + ( ∇ ( n ) R ) − : R µναβ ; δ ...δ n + .By the generalised Ricci identity we have that: R µναβ ; δ ...δ n − δ n + = R µναβ ; δ ...δ n − + δ n + X h R ∗ ∇ ( n − R i µναβδ ...δ n − δ n + We notice that all of the terms on the right-hand side are well-behaved; therefore, the left-hand side is well-behavedalso. Morever, we also see that all of the b.w. 0 terms of the form ∇ + ( ∇ ( n ) R ) − can be written in terms of well-behaved components of lower or equal order. Therefore, by induction, components of ( ∇ ( n ) R ) are well-behaved forall n ≥
0. The theorem is consequently proven.Note that an alternative proof might be given by using ∇ = e ∇ + τ , where e ∇ is the corresponding connection with H (1) ( u, x k ) = H (0) ( u, x k ) = W (0) ( u, x k ) = 0 and τ is a tensor (the remaining piece). The tensor τ will be of boostorder − I -symmetric or Kundt- CSI . We recall that for
CSI metrics, I -non-degeneracy implies that the spacetime is curvature homogeneous to all orders; hence, an important corollaryof the results of [1] is a proof of the CSI -Kundt conjecture in 4D [10], that for a 4D
CSI spacetime then either thespacetime is locally homogeneous or a subclass of the Kundt spacetimes . It is plausible that this result generalizes tohigher dimensions. In the context of string theory, it is of considerable interest to study higher dimensional Lorentzian
CSI spacetimes. In particular, a number of N-dimensional
CSI spacetimes are known to be solutions of supergravitytheory when supported by appropriate bosonic fields [14].First, we have that:orentzian manifolds and scalar curvature invariants 5
Corollary II.8.
A degenerate Kundt spacetime is I -symmetric if and only if the corresponding spacetime with H (1) ( u, x k ) = H (0) ( u, x k ) = W (0) ( u, x k ) = 0 is also I -symmetric. We can also prove the following:
Theorem II.9.
If the spacetime is of type D k , then the components of the curvature tensors are determined by thescalar curvature invariants.Proof. First we note that a type D k spacetime possesses at least one Killing vector (actually, at least three, seeCorollary that follows), namely a boost isotropy. Therefore, let H = SO (1 , d ) be the isotropy group of all thecurvature tensors (which necessarily must be at least of dimension 1; i.e., d ≥ g AB be the projector onto thetangent subspace of the orbits of H . Furthermore, let g IJ be defined so that ( g µν ) = ( g AB ) ⊕ ( g IJ ). This implies thatany curvature tensor can be orthogonally decomposed as a tensor; thus, symbolically ∇ ( k ) R = X φ ⊗R , where φ = ( φ ABC...D ) is a scalar representation of H , and R = ( R IJ...K ) is a tensor over g IJ . Consequently, if1 + d + ˜ d = dimension of spacetime, then ∇ ( k ) R can be considered as an SO ( ˜ d )-tensor. Hence, all scalar curvatureinvariants can be considered as SO ( ˜ d )-invariants. Since this group is compact, the group action separates orbits, andthus the components are determined by the scalar invariants.Note that this means that we can also say that a type D k spacetime is characterized by its invariants (however, in adifferent sense than I -non-degeneracy) [9]. This result also immediately implies that, for example, a spacetime whichis of type D k and CSI is necessarily Kundt and homogeneous [4]. In fact, in general, type D k spacetimes possess atleast three Killing vectors: Corollary II.10.
A spacetime of type D k possesses three Killing vector fields with 2-dimensional timelike orbits.Proof.
We have already pointed out that such a spacetime possesses a boost isotropy. Consider the Lie derivatives £ ℓ I , and £ n I where I is any polynomial curvature invariant. Since this is of type D k , we must have: £ ℓ I = ℓ ( I ) = ℓ µ ∇ µ I = 0(similarly for n ). Therefore, ℓ µ ∇ µ R = 0 for any curvature tensor R and thus there exists a Killing vector ˜ ℓ (similarlyfor n ). Hence, the spacetime possesses three Killing vectors (˜ ℓ , ˜ n , and a boost).Perhaps a more useful consequence is: Corollary II.11.
For a spacetime of (aligned) type II to all orders, the boost weight 0 components are determined bythe curvature invariants.
This corollary follows simply from the fact that when constructing a complete contraction of an arbitrary tensor oftype II, only the b.w. 0 components will contribute; i.e., if T is of type II, meaning T = ( T ) + ( T ) − + ( T ) − + ... ,then for a complete contraction: Contr[ T ] = Contr[( T ) ] . Consequently, the invariants of T and ( T ) are identical. Thus, since ( T ) is of type D, the above theorem impliesthat its components are determined from the invariants.This further implies a proof of the ‘ CSI F ’ conjecture: Corollary II.12.
A (degenerate) Kundt
CSI spacetime is a spacetime for which there exists a frame with a nullvector ℓ such that all components of the Riemann tensor and its covariants derivatives in this frame have the propertythat (i) all positive boost weight components (with respect to ℓ ) are zero and (ii) all zero boost weight components areconstant. Finally, it should be possible to prove an extension of the type D k result stated in [4]: Conjecture II.13.
If the curvature tensors to all orders are type II (or simpler) and aligned (i.e., ∇ ( n ) R is of typeII), then the spacetime is Kundt. It then follows, under some appropriate assumptions, that degenerate Kundt spacetimes with H (1) = H (0) = W (0) i =0 are of Riemann type D k .In the future we hope to establish all of the higher dimensional generalizations of the results obtained in 4D. Asnoted above, the first step is to investigate the curvature operators in higher dimensions and to classify these for thevarious algebraic types [9]. A. Coley, S. Hervik and N. Pelavas III. DISCUSSION
We have found that the degenerate Kundt metrics are not determined by their curvature invariants (in the sense thatthey are not I -non-degenerate). Degenerate Kundt spacetimes are also special in a number of other ways including, forexample, their holonomy structure, which may lead to novel and fundamental physics. Indeed, in a degenerate Kundtspacetime it is not possible to define a unique timelike curvature operator, and hence a unique timelike direction, anda 1+( n −
1) spacetime splitting of the spacetime is not possible.Supersymmetric solutions of supergravity theories have played an important role in the development of string theory(see, for example, [11]) The existence of Killing spinors accounts for much of the interest in metrics with specialholonomy in mathematical physics. Supersymmetric solutions in M -theory that are not static admit a covariantlyconstant null vector ( CCN V ) [2]. The isotropy subgroup of a null spinor is contained in the isotropy subgroup ofthe null vector, which in arbitrary dimensions is isomorphic to the spin cover of
ISO ( n − CCN V metric is adegenerate Kundt metric (1) with H ,v = 0 and W i,v = 0. This class includes a subset of the Kundt- CSI and the
V SI spacetimes as special cases. The
V SI and
CSI degenerate Kundt spacetimes are of fundamental importance sincethey are solutions of supergravity or superstring theory, when supported by appropriate bosonic fields [14].The classification of holonomy groups in Lorentzian spacetimes is quite different from the Riemannian case. ALorentzian manifold M is either completely reducible , and so M decomposes into irreducible or flat Riemannianmanifolds and a manifold which is an irreducible or a flat Lorentzian manifold or ( R , − dt ), or M is not completelyreducible , which leads to the existence of a degenerate (one-dimensional) holonomy invariant lightlike subspace (theLorentzian manifold decomposes into irreducible or flat Riemannian manifolds and a Lorentzian manifold with inde-composable, but non-irreducible holonomy representation), which gives rise to the recurrent null vector ( RN V ) and
CCN V (Kundt) spacetimes [12, 13]. [A
RN V metric has holonomy Sim( n −
2) and the metric belongs to the classof Kundt metrics (1) with W i = W i ( u, x k ), but is not necessarily a degenerate Kundt metric.]Therefore, the Kundt spacetimes that are of particular physical interest are degenerately reducible, which leads tocomplicated holonomy structure and various degenerate mathematical properties. Indeed, it could be argued that acomplete understanding of string theory is not possible without a comprehensive knowledge of the properties of theKundt spacetimes [13]. For example, as noted above, a degenerate Kundt spacetime is not completely classified by itsset of scalar polynomial curvature invariants (i.e., they have important geometrical information that is not containedin the scalar invariants). All V SI spacetimes and
CSI spacetimes that are not locally homogeneous (including theimportant
CCN V subcase) belong to the degenerate Kundt class [2, 7]. In these spacetimes all of the scalar invariantsare constant or zero. This leads to interesting problems with any physical property that depends essentially on scalarinvariants, and may lead to ambiguities and pathologies in models of quantum gravity or string theory.As an illustration, in many theories of fundamental physics there are geometric classical corrections to general rela-tivity. Different polynomial curvature invariants (constructed from the Riemann tensor and its covariant derivatives)are required to compute different loop-orders of renormalization of the Einstein-Hilbert action. In specific quantummodels such as supergravity there are particular allowed local counterterms [15]. All classical corrections are zero in
V SI spacetimes (and constant in
CSI spacetimes). Indeed, it is possible that a Lorentzian Kundt spacetime doesnot even allow for a low order perturbative expansion.
Acknowledgments
This work was supported by the Natural Sciences and Engineering Research Council of Canada. [1] A. Coley, S. Hervik and N. Pelavas, 2009, Class. Quant. Grav. , 025013 [arXiv:0901.0791].[2] A. Coley, 2008, Class. Quant. Grav. , 033001 [arXiv:0710.1598].[3] A. Coley, S. Hervik and N. Pelavas, 2006, Class. Quant. Grav. , 3053.[4] A. Coley, S. Hervik, G. Papadopoulos and N. Pelavas, 2009, Class. Quant. Grav. , 105016 [arXiv:0901.0394].[5] A. Coley, R. Milson, V. Pravda and A. Pravdova, 2004, Class. Quant. Grav. , 5519 [gr-qc/0410070]; V. Pravda, A.Pravdov´a, A. Coley and R. Milson, 2002, Class. Quant. Grav. , 6213.[6] H. Stephani, D. Kramer, M. A. H. MacCallum, C. A. Hoenselaers, E. Herlt 2003 Exact solutions of Einstein’s field equations,second edition (Cambridge University Press; Cambridge).[7] A. Coley, R. Milson, V. Pravda and A. Pravdova, 2004, Class. Quant. Grav. , L35 [gr-qc/0401008]; R. Milson, A. Coley,V. Pravda and A. Pravdova, 2005, Int. J. Geom. Meth. Mod. Phys. , 41; A. Coley, 2008, Class. Quant. Grav. , 033001. orentzian manifolds and scalar curvature invariants 7 [8] J. Podolsky and M. Ortaggio, 2006, Class. Quantum Grav. , 5785; J. Podolsky and M. Zofka, 2009, Class. QuantumGrav. , 105008; M. Ortaggio, V. Pravda and A. Pravdova, 2009, AIP Conf. Proc. , 195010; M. Godazgar and H. S. Reall, 2009, Class. Quantum Grav. , 165009.[9] A. Coley and S. Hervik, 2010, Class. Quantum Grav. , 015002 [arXiv:0909:1160].[10] A. Coley, S. Hervik and N. Pelavas, 2009, Class. Quant. Grav. , 125011 [arXiv:0904:4877]; A. Coley, S. Hervik and N.Pelavas, 2008, Class. Quantum Grav. , 025008.[11] J. Grover et al. , 2009, JHEP , 069.[12] L. B´erard-Bergery and A. Ikemakhen, 1993, Proc. Sympos. Pure Math. (Amer. Math. Soc., RI), , 27; T. Leistner, 2007,Journal of Differential Geometry, , 423.[13] J. Brannlund, A. Coley and S. Hervik, 2008, Class. Quant. Grav. A24 , 527; Z. Bern et al., 1998, Nucl. Phys. B53062