Cayley-Klein Lie bialgebras: Noncommutative spaces, Drinfel'd doubles and kinematical applications

The Cayley-Klein (CK) formalism is applied to the real algebra ${so}(5)$ by making use of four graded contraction parameters describing in a unified setting 81 Lie algebras, which cover the (anti-)de Sitter, Poincar\'e, Newtonian and Carrollian algebras. Starting with the Drinfel'd-Jimbo real Lie bialgebra for ${so}(5)$ together with its Drinfel'd double structure, we obtain the corresponding CK bialgebra and the CK $r$-matrix coming from a Drinfel'd double. As a novelty, we construct the (first-order) noncommutative CK spaces of points, lines, 2-planes and 3-hyperplanes, studying their structural properties. By requiring to deal with real structures, it comes out that there exist 63 specific real Lie bialgebras together with their sets of four noncommutative spaces. Furthermore, we find 14 classical $r$-matrices coming from Drinfel'd doubles, obtaining new results for the de Sitter ${so}(4,1)$ and anti-de Sitter ${so}(3,2)$ and for some of their contractions. These geometric results are exhaustively applied onto the (3+1)D kinematical algebras, not only considering the usual (3+1)D spacetime but also the 6D space of lines. We establish different assignations between the geometrical CK generators and the kinematical ones which convey physical identifications for the CK contraction parameters in terms of the cosmological constant/curvature $\Lambda$ and speed of light $c$. We finally obtain four classes of kinematical $r$-matrices together with their noncommutative spacetimes and spaces of lines, comprising all $\kappa$-deformations as particular cases.


Introduction
The notion of Cayley-Klein (CK) Lie algebras along with their corresponding Lie groups and symmetric homogeneous spaces date back to early studies of projective metrics. In particular, CK Lie groups appear in a natural way within the context of the consideration by Klein that most geometries are, in fact, subgeometries of projective geometry and also in relation to Cayley's theory of projective metrics [1][2][3]. However, the complete classification of CK geometries, understood as geometries endowed with a projective metric, was not given by Klein himself. Meanwhile the two-dimensional (2D) ones were studied under the name of "quadratic geometries" by Poincaré, who followed a modern group theoretical approach. The classification of CK geometries for arbitrary dimension N was finally achieved by Sommerville in 1909 [4], where he showed that there exist exactly 3 N different CK geometries in dimension N , each of them corresponding to a different choice of the kind of measure of distance between points, lines, 2planes,. . . , (N − 1)-hyperplanes, which can be either of elliptic, parabolic or hyperbolic type [3]. Then CK groups are just the motion groups of the CK geometries acting as groups of isometries of the symmetric homogeneous CK spaces. In dimension N , such CK groups are semisimple pseudo-orthogonal groups SO(p, q) (p + q = N + 1) and some of their contractions such as, for instance, the inhomogeneous ISO(p ′ , q ′ ) (p ′ + q ′ = N ).
In order to set up the main ideas and the formalism that we shall follow along the whole paper, let us consider the well known nine 2D CK geometries [3,[5][6][7][8][9][10][11][12]. These emerge as the different possibilities for considering the measure of distance between two points and the measure of an angle between two lines, being each of them either of elliptic, parabolic or hyperbolic type. The CK groups are the simple real Lie groups SO(3) and SO (2,1), the non-simple inhomogeneous Euclidean ISO (2) and Poincaré ISO (1,1), and the twice inhomogeneous Galilean IISO(1) (in this notation ISO(1) ≡ R). The 2D CK geometries are constructed through the coset spaces of the above 3D Lie groups with a precise 1D isotropy subgroup. Early, the usual procedure for describing these geometries made use of hypercomplex numbers with two hypercomplex units ι 1 and ι 2 [2,3,5,6]. We recall that a hypercomplex number is defined by z := x + ι y, where (x, y) are two real coordinates and ι is a hypercomplex unit such that ι 2 ∈ {−1, +1, 0}. Hence there are three possible kind of hypercomplex numbers according to the specific unit ι: (1) If ι 2 = −1, then ι is an elliptical unit providing the usual complex numbers; (2) If ι 2 = +1, ι is a hyperbolic unit yielding the so-called split complex, double or Clifford numbers; and (3) if ι 2 = 0, ι is a parabolic unit leading to the dual or Study numbers. Alternatively, 2D CK geometries can also be studied in terms of two real graded contraction parameters ω 1 and ω 2 , which can take positive, negative or zero values [7][8][9][10][11][12]. By taking into account the above two approaches, we display the specific 2D CK geometries in Table 1, where they are named in their original geometric form [3], as well as in their physical (or kinematical) terminology (second and third rows).
Consequently, the family of 2D CK geometries contains nine homogeneous spaces of constant curvature: the three classical Riemannian spaces in the first row of Table 1; the three (Newtonian) spaces with a degenerate metric in the second row; and the three pseudo-Riemannian or Lorentzian spaces in the third row.

Measure of distance
Measure Elliptic Parabolic Hyperbolic of angle contraction [13]. The CK algebra has a single quadratic Casimir given by The 2D CK geometry is then defined as the following coset space between the CK Lie group SO ω 1 ,ω 2 (3) with Lie algebra so ω 1 ,ω 2 (3) (1) and the isotropy subgroup of a point H, spanned by J 12 [9]:  (2), which is of constant (Gaussian) curvature equal to ω 1 and with metric signature given by diag(+1, ω 2 ) (see Table 1).
Spacetimes of constant curvature in (1+1) dimensions arise as particular cases of the CK geometry (3) through different assignations between the geometrical generators J ab and the kinematical ones, which require appropriate relations between the CK parameters (ω 1 , ω 2 ) and physical quantities. Explicitly, let {P 0 , P 1 , K} be the generators of time translations, space translations and boost transformations. Under the particular identification where Λ is the cosmological constant and c is the speed of light, we find that the CK algebra (1) adopts the form [K, P 0 ] = P 1 , [K, P 1 ] = 1 c 2 P 0 , [P 0 , P 1 ] = −ΛK.
Thus, under the relations (4), the CK group SO ω 1 ,ω 2 (3) with Lie algebra (5) becomes a kinematical group acting as the group of isometries of six relevant (1+1)D spacetime models [14] of constant curvature equal to −Λ, which are all of them contained in (3). These are the three Lorentzian spacetimes with metric signature diag(+1, −c −2 ), mentioned in the third row of Table 1, along with their non-relativistic limit c → ∞ leading to the three Newtonian spacetimes with degenerate metric diag(+1, 0) in the second row of Table 1 (NH means Newton-Hooke).
It should be noted that all of the kinematical algebras and spacetimes can also be described within the CK framework except for the static algebra [14]; at this dimension the latter is just the abelian algebra.
In principle, the very same results can be obtained by making use of the formalism in terms of hypercomplex numbers [2,3,5,6] since, roughly speaking, one finds the relations ι 2 ∼ −ω. Nevertheless, besides to simply deal with real numbers instead of hypercomplex ones, the main differences between both approaches clearly appear in the pure contracted case corresponding to consider the parabolic unit ι 2 = 0 and to set ω = 0. In particular, the contraction of exponentials of a Lie generator J could give rise to different results; for instance: where x is a real number. We stress that this kind of exponentials often appear in quantum groups and, moreover, terms depending on some √ ω m will be omnipresent throughout the paper (within Lie bialgebra structures), so that both procedures could be no longer equivalent in a quantum deformation framework. Thus we shall make use of the graded contraction approach with real parameters ω m in such manner that a smooth and well-defined ω m → 0 limit of all the expressions that we shall present here will be always feasible.
In addition, we stress that the same CK group SO(2, 1) appears three times in Table 1, and two of their CK spaces are "similar" (see [15,16] for a very detailed description of these three geometries in terms of hypercomplex numbers). The structure of the three CK geometries involved, hyperbolic and the (anti-)de Sitter ones, can be better understood by considering not only the usual CK space (3) shown in Table 1 but also their 2D homogeneous spaces of lines, as it was performed in [8,12], and likewise for the Euclidean and Poincaré groups which appear twice in Table 1. In arbitrary dimension N , the CK algebra depends on N real graded contraction parameters ω m (m = 1, . . . , N ) and is denoted by so ω 1 ,...,ω N (N + 1). This family comprises 3 N semisimple and non-semsimple real Lie algebras (being some of them isomorphic) which share common geometric and algebraic properties. The signs of the parameters ω m = 0 determine a specific real form so(p, q) and when at least one ω m vanishes the CK algebra becomes a non-semisimple one.
In algebraic terms, a CK algebra can be defined as a graded contracted Lie algebra from so(N +1) [17] which keeps the same number of algebraically independent Casimir invariants as in the semisimple case, regardless of the values of ω m [18]. This definition implies that all of the 3 N particular CK algebras have the same rank (even for the most contracted case with all ω m = 0), so that they are also known as quasisimple orthogonal algebras [18,19]. From this viewpoint they can be seen as the "closest" contracted algebras to the semisimple ones. In this respect, we remark that the CK contraction sequence ensures to always obtain a non-trivial quadratic Casimir (like (2)) which, in turn, means that there always exists a non-trivial metric on the N D CK geometry, although degenerate in many cases. This fact explains the absence of the static algebra in the CK family. Obviously, if one goes beyond the CK Lie algebra contraction sequence, then one can get the static algebra, and finally arrive at the abelian algebra [20].
In this paper we shall focus on the physically relevant dimension N = 4 thus covering the (3+1)D spacetimes of constant curvature. Hence the CK algebra and group will depend on a set of four real graded contraction parameters ω = (ω 1 , ω 2 , ω 3 , ω 4 ) so comprising 3 4 = 81 specific Lie algebras. For each CK algebra/group we shall consider four types of symmetric homogeneous spaces: the usual 4D CK space of points (spacetimes) (6) along with the 6D space of lines, 6D space of 2-planes and 4D space of 3-hyperplanes, being all of them of constant curvature and equal to ω 1 , . . . , ω 4 , respectively. Therefore in this paper by a CK geometry it will be understood the set of these four homogeneous spaces associated with a given Lie algebra in the CK family and not only the usual space of points (6), which is the one commonly considered in the literature.
In this geometrical setting, we initially study CK Lie bialgebras and their associated noncommutative spaces in order to further develop their physical applications. Thus we, firstly, review the basics on quantum groups that will be used along the paper in Section 2. And, secondly, we present a two-fold work in Sections 3-6 with two main but interrelated parts, whose structure, objectives and results are as follows.
1. Starting with the Drinfel'd-Jimbo Lie bialgebra for so (5) and also considering its Drinfel'd double structure in Section 3, we obtain the corresponding CK bialgebra along with the classical CK r-matrix coming from a Drinfel'd double in Section 4. As a novelty, we construct, by means of quantum duality, the first-order (in the quantum coordinates) noncommutative CK spaces of points, lines, 2-planes and 3-hyperplanes. We analyse their properties and always require to deal with real structures. It comes out that, finally, there are 63 specific real Lie bialgebras together with their sets of four (first-order) noncommutative spaces, which are summarized in Tables 2 and 3, respectively. Additionally, we find 14 classical r-matrices coming from Drinfel'd double real structures: there are four cases (I)-(IV) for the simple algebras and 10 more cases for their contractions. In this way, we obtain new results for de Sitter so(4, 1) (case (II)) and anti-de Sitter so(3, 2) (case (IV)) Drinfel'd doubles and for some of their contractions, which are displayed in Table 4. 2. The above geometric results are exhaustively applied onto the (3+1)D kinematical algebras [14], not only considering the usual (3+1)D spacetime but also the 6D space of lines; the classical picture for each kinematical algebra is presented in Section 5 and outlined in Table 5. In Section 6, we establish different assignations between the geometrical CK generators and the kinematical ones which convey appropriate physical identifications for the CK contraction parameters ω in terms of the cosmological constant/curvature Λ and speed of light c. In this process we obtain four classes of kinematical r-matrices and, for some algebras, also r-matrices coming from Drinfel'd doubles. These classes are called A, B, C and D, matching, in this order, with the above cases (I)-(IV). The resulting kinematical bialgebras are given in Table 6, while their corresponding first-order noncommutative spacetimes and spaces of lines are shown in Table 7. We stress that the class C covers the kappa-deformations.
We would like to mention that although in this work we do not construct the complete quantum kinematical algebras and their associated full noncommutative spacetimes and spaces of lines, we comment on related known results and open problems in Sections 6.5 and 6.6, respectively. To finish with, several conclusions and a more exhaustive list of open problems, also concerning the geometric CK setting, are drawn in Section 7.

Fundamentals on Quantum Groups
In this Section we review the basic background on quantum groups necessary for the paper along with updates and physical motivation related to the main results here presented. We shall focus on quantum deformations of Lie algebras (with a Hopf algebra structure) along with their connection with Lie bialgebras, Poisson-Lie groups, Poisson homogeneous spaces, noncommutative spaces and Drinfel'd doubles. More details on these topics can be found in [21][22][23][24][25][26].

Lie Bialgebras and Quantum Algebras
Let us consider an nD Lie algebra g = span{X 1 , . . . , X n } with commutation relations given by The Lie algebra g is endowed with a Lie bialgebra structure (g, δ) [27] if there exists a map δ : g → g ∧ g called the cocommutator verifying two conditions: (ii) The dual map δ * : g * ⊗ g * → g * is a Lie bracket on the dual Lie algebra g * of g.
Therefore any cocommutator δ can be written in a skew-symmetric form as in such a manner that the antisymmetric factors f jk i turn out to be the structure constants of the dual Lie algebra g * = span{x 1 , . . . ,x n }: The duality between the generators of g and g * is determined by a canonical pairing given by the bilinear form And the cocycle condition (8) leads to the following compatibility equations among the structure constants c k ij (7) and f jk i (10): For some Lie bialgebras the 1-cocycle δ is coboundary [27], that is, it can be obtained from an element r ∈ g ⊗ g in the form The element r is the so-called classical r-matrix which can always be written in a skew-symmetric form and must be a solution of the modified classical Yang-Baxter equation such that If the Schouten bracket (15) does not vanish for an r-matrix written in the skew-symmetric form (13), then the Lie algebra g is endowed with a quasitriangular (or standard) Lie bialgebra structure (g, δ(r)). The vanishing of the Schouten bracket corresponds to the classical Yang-Baxter equation and (g, δ(r)) is called triangular (or nonstandard) Lie bialgebra.
We remark that the deformation parameter, that we shall denote z throughout the paper and such that q = e z , is already contained within δ (9) and r (13) in the factors f jk i and r ij as a global multiplicative constant. This, in turn, means that the non-deformed or "classical" limit z → 0 (i.e. q → 1) leads to a trivial coproduct δ = 0 with all f jk i ≡ 0 and classical r-matrix r = 0 with all r ij ≡ 0, being g * an abelian Lie algebra thus with commutative generatorsx i (10).
A quantum algebra U z (g) is a Hopf algebra deformation of the universal enveloping algebra U (g) of g constructed as formal power series C[[z]] in a deformation indeterminate parameter z and coefficients in U (g), that is, U z (g) = U (g)⊗ C[[z]]. The Hopf algebra structure of U z (g) is determined by the coproduct ∆ z , counit ǫ and antipode γ mappings [24][25][26]. In particular, the coproduct ∆ z : U z (g) → U z (g) ⊗ U z (g) must be an algebra homomorphism and fulfil the coassociativity condition where Id is the identity map, giving rise to a coalgebra structure (U z (g), ∆ z ). Once ∆ z is obtained, the remaining maps, ǫ and γ, can directly be deduced from the Hopf algebra axioms providing the complete Hopf algebra structure. Hence hereafter we shall only focus on the coalgebra structure of U z (g) assuming the existence of the corresponding counit and antipode.
The remarkable point is that any quantum algebra U z (g) is determined at the first-order in z by a Lie bialgebra (g, δ). Explicitly, if we write the coproduct ∆ z as a formal power series in z, the cocommutator δ (9) is just the skew-symmetric part of the first-order term ∆ 1 in z, namely where σ is the flip operator σ(X i ⊗ X j ) = X j ⊗ X i and ∆ 0 is called the primitive (non-deformed) coproduct. Therefore each Lie bialgebra (g, δ) determines a quantum deformation (U z (g), ∆ z ) and the equivalence classes (under automorphisms) of Lie bialgebra structures on g will provide all its possible quantum algebras.
We recall that for semisimple Lie algebras all their Lie bialgebra structures are coboundaries, so that all their possible quantum deformations are determined by classical r-matrices. The paradigmatic type of them is provided by the so-called Drinfel'd-Jimbo deformations [28][29][30]; the corresponding Drinfel'd-Jimbo r-matrix for the compact real form so(5) [31] will be our starting point for the detailed study of the CK Lie bialgebras which will be performed in Sections 3 and 4, respectively. However, even for semisimple Lie algebras the determination of all the Lie bialgebra structures through classical r-matrices is a cumbersome task and, in fact, there are only classifications for the Lorentz algebra so(3, 1) [32] and for the related real forms so(4) and so(2, 2) [33,34]; from a kinematical viewpoint these classifications for so (3,1) and so (2,2) correspond to (2+1)D (anti-)de Sitter r-matrices [35]. Therefore, in the (3+1)D case, which is the one that we shall consider throughout this paper, there are no such classifications for the simple algebras so(p, q) with p + q = 5, although we remark that there are some partial results. In particular, it was shown in [36] that there only exist two two-parametric classical r-matrices for the (anti-)de Sitter algebras so(4, 1) and so(3, 2) keeping primitive (undeformed) the time translation generator and a single rotation generator. And the classification of their r-matrices which preserve a Lorentz so(3, 1) sub-bialgebra has been, very recently, obtained in [37] starting with the former full (2+1)D classification [32,33].
Concerning other kinematical algebras, we also recall that the obtention of Lie bialgebras mainly cover low-dimensional cases such as the (1+1)D Galilei algebra (isomorphic to the Heisenberg-Weyl algebra h 3 ) [44][45][46][47], the 2D Euclidean algebra iso(2) [48], the (1+1)D centrally extended Galilei algebra [49][50][51] and the (1+1)D centrally extended Poincaré algebra in the light-cone basis (isomorphic to the oscillator algebra h 4 ) [52,53]. With the exception of the latter, all of them have both coboundary and non-coboundary Lie bialgebra structures. In general, for solvable and nilpotent Lie algebras many of their Lie bialgebra structures are noncoboundaries; in this respect, see [54,55] and references therein for the classifcation of 3D Lie bialgebras. Finally, we point out that, very recently, the classification of 4D indecomposable coboundary Lie bialgebras has been carried out in [56], which shows how the difficulties of this task grow when the dimensions of the Lie bialgebras increase.

Quantum Groups and Noncommutative Spaces
Let us consider a quantum algebra (U z (g), ∆ z ) with underlying Lie bialgebra (g, δ) and let G be the Lie group with Lie algebra g. A quantum group (G z , ∆ Gz ) is a noncommutative algebra of functions on G defined as the dual Hopf algebra to the quantum algebra (U z (g), ∆ z ). Explicitly, let m Gz and m z be the noncommutative products in G z and U z (g), respectively. The duality between the Hopf algebras (G z , m Gz , ∆ Gz ) and (U z (g), m z , ∆ z ) is established by means of a canonical pairing , : where X, Y ∈ U z (g), f, g ∈ G z , and f ⊗ g, X ⊗ Y = f, X g, Y .
The duality relation (18) implies that the noncommutative product m Gz in the quantum group G z is defined by the coproduct ∆ z in the quantum algebra U z (g) and, conversely, the expression (19) implies that the coproduct ∆ Gz in G z is given by the noncommutative product m z in U z (g). By taking into account that the first-order term in z of the coproduct ∆ z is defined by the cocommutator δ (17), a straightforward consequence of the above Hopf algebra duality is that the commutation relations for the quantum group G z at the first-order in the quantum (noncommutative) coordinatesx i are given by the dual map δ * of δ (9), that is, with fundamental Lie brackets (10).
Furthermore, each quantum group (G z , ∆ Gz ) can be associated with a Poisson-Lie group (G, Π), with Poisson structure Π, and the latter with a unique Lie bialgebra structure (g, δ). In particular, it is well known [27] that Poisson-Lie structures on a connected and simply connected Lie group G are in one-to-one correspondence with Lie bialgebra structures. Hence, quantum groups are quantizations of Poisson-Lie groups, that is, quantizations of the Poisson-Hopf algebras of multiplicative Poisson structures on Lie groups [24,25,30]. In the case of coboundary Lie bialgebras (g, δ(r)), coming from a skew-symmetric classical r-matrix (13), the Poisson structure Π of the Poisson-Lie group is given by the so-called Sklyanin bracket [24,30] {f, g} = n i,j=1 where ∇ L i and ∇ R i are left-and right-invariant vector fields on G. Next, a Poisson homogeneous space of a Poisson-Lie group (G, Π) is a Poisson manifold (M, π) endowed with a transitive group action ⊲ : G × M → M which is a Poisson map with respect to the Poisson structure on the manifold M and the product Π ⊕ π of the Poisson structures on G and M . In this paper we shall consider that the manifold M is an ℓD homogeneous space of a Lie group G (the motion group of M ) with isotropy subgroup H whose Lie algebras are g and h, respectively. Moreover, throughout this paper we will be interested in pointed Poisson homogeneous spaces, i.e. Poisson homogeneous spaces in which the origin is fixed, and we will not study how the Poisson structure is modified when this origin is changed. The Lie algebra g, understood as a vector space, can be written as the sum of two vector subspaces The generators of h leave a point of M invariant, which is taken as the origin O of the space, so they play the role of rotations around O, while the ℓ generators belonging to t move O along ℓ basic directions, behaving as translations on M . The group parameters (u 1 , . . . , u ℓ ) of the generators of t lead to ℓ coordinates of M and they span the annihilator h ⊥ of the vector subspace h in the dual Lie algebra g * [57].
In principle, the Lie algebra g of G may admit several coboundary Lie bialgebra structures (g, δ(r)), that is, different classical r-matrices. Then a particular Poisson homogeneous space (M, π) can be constructed by endowing the motion group G with the Poisson-Lie structure Π (20) for a given classical r-matrix, and the homogeneous space M (21) with a Poisson bracket π that has to be compatible with the group action ⊲ : G × M → M . Therefore, according to the possible classical r-matrices of g it follows that the Lie group G may be endowed with several Poisson-Lie structures Π (20), each of them leading to a different Poisson homogeneous space [58].
A distinguished type of Poisson homogeneous spaces are those in which the Poisson bracket π is obtained as the canonical projection on M with coordinates (u 1 , . . . , u ℓ ) of the Poisson-Lie bracket Π [57,59] (see also [60,61]). In terms of the underlying Lie bialgebra, this requirement corresponds to imposing the so-called coisotropy condition for the cocommutator δ with respect to the isotropy subalgebra h of H given by [57,59] A particular and very restrictive case of the above condition is that the subalgebra h be a sub-Lie bialgebra, which implies that the Poisson homogeneous space is constructed through an isotropy subgroup H which is a Poisson subgroup (H, Π) of (G, Π). Furthermore, since the quantum group (G z , ∆ Gz ) is the quantization of the Poisson-Lie group (G, Π), the quantization of a coisotropic Poisson homogeneous space (M, π) fulfilling (22) provides a quantum homogeneous space M z with quantum coordinates (û 1 , . . . ,û ℓ ), onto which the quantum group G z co-acts covariantly [62]. The coisotropy condition (22) ensures that the commutation relations that define M z at the first-order in all the quantum coordinates close a Lie subalgebra which is just the annihilator h ⊥ of h on the dual Lie algebra g * and such relations determine a Lie subalgebra of g * (10). Generically, M z is called a noncommutative space.
Probably, the best known and most studied example of noncommutative spaces is the socalled κ-Minkowski spacetime coming from the κ-Poincaré algebra [63][64][65][66][67][68][69] where κ is the quantum deformation parameter which is proportional to the Planck mass and here related to z as κ = z −1 . The quantum algebra U z (g) and quantum group G z correspond to the κ-Poincaré algebra and κ-Poincaré group. In this case, the underlying homogeneous space (21) is the flat (3+1)D Minkowski spacetime constructed as the coset space of the Poincaré group G = ISO (3,1) with the Lorentz isotropy subgroup H = SO(3, 1): Thus the dimension is ℓ = 4 and the coordinates (u 1 , . . . , u 4 ) are identified with the time and spatial ones (x 0 , x i ) (i = 1, 2, 3). The κ-Poincaré classical r-matrix [67] provides a quasitriangular quantum deformation of the Poincaré algebra [64][65][66] such that the Lorentz subalgebra h = so(3, 1) fulfils the coisotropy condition (22) thus giving rise to the subalgebra h ⊥ whose generators are the quantum coordinates (x 0 ,x i ) dual to the time and space translation generators. The complete quantization of h ⊥ is just the κ-Minkowski spacetime M 3+1 κ which is defined by the commutation relations given by [67]: which are covariant under the κ-Poincaré quantum group [69]. We remark that M 3+1 κ is a linear algebra which coincides exactly with the one obtained through the Sklyanin bracket (20) of the underlying classical r-matrix [67] which provides the (linear) Poisson homogeneous spacetime. Therefore no higher-order terms in the classical and quantum coordinates arise. By contrast, when the κ-deformation is applied to a curved manifold instead of (24), higher-order terms in the coordinates appear in the Poisson homogeneous spacetime, so that the corresponding quantization is not straightfoward at all, as the recent constructions of the κ-noncommutative (anti-)de Sitter [70], Newtonian and Carrollian [71] spacetimes explicitly show.
In Section 6.3 we shall recover the κ-Minkowski space M 3+1 κ (25) and the κ-Poincaré Lie bialgebra as a particular case of "time-like" deformations within the CK family of Lie bialgebras, a fact that is already well known [31,102]. However, as we shall show in Section 4.3, what is a striking point is that a formally similar structure to M 3+1 κ arises as the first-order noncommutative CK space of points which is shared by 63 CK bialgebras. Moreover the complete (in all orders in the quantum coordinates) noncommutative CK space of points is kept linear and shared by 27 CK bialgebras. Consequently, a linear noncommutative space similar to (25) is somewhat "ubiquitous" which, in turn, suggests that additional "structures" should be taken into account. In fact, this is one of the main aims of this paper and as a novel result we shall explicitly show in Section 4.3 that the consideration of other noncommutative spaces beyond the space of points (kinematically, spacetimes) associated with a given quantum algebra (namely, noncommutative spaces of lines, 2-planes and 3-hyperplanes) does allow one to distinguish mathematical and physical properties between two quantum algebras with the same underlying noncommutative space of points.

Drinfel'd Double Structures
Let us assume that the dimension of the Lie algebra g is even n = 2d. In this case, g is a Lie algebra of a Drinfel'd double Lie group [30] if there exists a basis {Y 1 , . . . , Y d , y 1 , . . . , y d } of g such that the commutation relations (7) can be written as Hence g can be split into two Lie subalgebras g 1 = span{Y 1 , . . . , Y d }, g 2 = span{y 1 , . . . , y d } with structure constants C γ αβ and F αβ γ , respectively. Both subalgebras are dual to each other, g * 2 = g 1 , by means of the duality defined with respect to the nondegenerate symmetric bilinear form , : g × g → R given by which is "associative" or invariant in the sense that The triple (g 1 , g 2 = g * 1 , g) is called a Manin triple and the Drinfel'd double Lie group is the unique connected and simply connected Lie group G with Lie algebra g. Therefore the Lie algebra g, verifying (26), is the double Lie algebra of g 1 and of its dual algebra g * 1 = g 2 . By construction, each Drinfel'd double structure for g has a canonical classical r-matrix which is a solution of the classical Yang-Baxter equation (16). Moreover, the universal enveloping algebra U (g) of g has always a quadratic Casimir element given by which is directly related to the bilinear form (27). The tensorized form of C reads which is ad-invariant under the action of g, that is The element Ω leads to a skew-symmetric classical r-matrix for the Drinfel'd double Lie algebra g from the canonical one (28) in the form: which is a solution of the modified classical Yang-Baxter equation (14) (its Schouten bracket does not vanish now), so that r D defines a quasitriangular or standard quantum deformation of g with a coboundary Lie bialgebra (g, δ D (r D )) determined by a cocommutator δ D through the relation (12).
Concerning the Drinfel'd-Jimbo quantum deformations of semisimple Lie algebras [28][29][30], it is known that they are closely related to quantum deformations of Drinfel'd doubles, that is, quantum doubles, in such a manner that they are "almost" but not strictly speaking quantum doubles [24]. Nevertheless, it is remarkable that proper Drinfel'd double structures for the four Cartan series of semisimple Lie algebras on C have been obtained in [103,104] by enlarging the Lie algebras with an appropriate number of central extensions.
From a physical viewpoint, it is worth stressing that Drinfel'd double structures are naturally related to (2+1)D gravity, which is a quite different theory from the full (3+1)D one [105,106]. In particular, (2+1)D gravity is a topological theory which admits a description as a Chern-Simons theory with gauge group given by the group of isometries of the corresponding spacetime of constant curvature [107,108]. The phase space structure of (2+1)D gravity is related to the moduli space of flat connections on a Riemann surface, the symmetries of which are given by certain Poisson-Lie groups [109,110] such that the Poisson structure on this space admits a description in terms of coboundary Lie bialgebras associated with the gauge group. Hence quantum group symmetries arise as the quantum counterparts of the (semiclassical) Poisson-Lie symmetries of the classical theory. The essential fact in the (2+1)D gravity framework is that the relevant quantum group symmetries are those coming from some classical r-matrices corresponding to Drinfel'd double structures [111][112][113][114][115][116][117], which ensures that the Fock-Rosly condition [110] is fulfilled. The symmetric component of such admissible classical r-matrices, which is just the element Ω (29) when these are written in the symmetric form (28), must be dual to the Ad-invariant symmetric bilinear form in the Chern-Simons action. As a consequence, the κ-Poincaré and κ-(anti-)de Sitter symmetries are not compatible [111,112] with the Chern-Simons formulation of (2+1)D gravity. Furthermore, the Chern-Simons approach to non-relativistic (2+1)D quantum gravity has also been developed in [118,119] by making use of a two-fold central extension of the Galilei [120] and Newton-Hooke algebras, and their full quantum deformation has been obtained in [121]. Additionally, Drinfel'd doubles also play a prominent role in state sum or spin foam models for (2+1)D gravity as shown in [122,123] in the context of the Turaev-Viro model and invariant.
We recall that the classifications of non-isomorphic 4D and 6D real Drinfel'd double structures were carried out in [124] and [125], respectively, while their Hopf algebra quantizations were constructed in [126]. From these results and also from [54], there were obtained the classifications of Drinfel'd double structures for the (2+1)D (anti-)de Sitter algebras in [115], (2+1)D Poincaré algebra and centrally extended (1+1)D Poincaré algebra in [127], and 3D Euclidean algebra in [128]. By contrast, results concerning Drinfel'd double structures in the (3+1)D case are very scarce, only covering the real so(5) and anti-de Sitter so(3, 2) algebras given in [129]. In this respect, we advance that in Section 4.4 we shall obtain two new classical r-matrices coming from Drinfel'd doubles: one for the de Sitter so(4, 1) and another for the anti-de Sitter so(3, 2).
3 The Drinfel'd-Jimbo Lie Bialgebra for so (5) Let us consider the real orthogonal Lie algebra so(5) with generators {J ab } (a < b; a, b = 0, 1, . . . , 4) fulfilling the Lie brackets and such that those commutators involving four different indices are equal to zero. The universal enveloping algebra of the Lie algebra so(5) is endowed with two (second-and fourth-order) Casimir operators [18,130]. The quadratic one, coming from the Killing-Cartan form, is given by A fine grading group Z ⊗4 2 of so(5) is spanned by the four commuting involutive automorphisms Θ (m) (m = 1, . . . , 4) of (31) defined by [17,20]: Each involution Θ (m) provides a Cartan decomposition of so(5) in invariant and anti-invariant subspaces denoted h (m) and t (m) respectively: These subspaces verify that where h (m) is a Lie subalgebra such that while the vector subspace t (m) is not a subalgebra.
A faithful matrix representation of so(5), ρ : so(5) → End(R 5 ), is given by where e ab is the 5 × 5 matrix with a single non-zero entry 1 at row a and column b (a, b = 0, 1, . . . , 4), fulfilling the orthogonal matrix condition where ρ(J ab ) T is the transpose matrix of ρ(J ab ).

Symmetric Homogeneous Spaces
According to each automorphism Θ (m) (33) and its associated Cartan decomposition (34) we construct four symmetric homogeneous spaces, of the type (21), as the coset spaces [131][132][133] where the four generators in the rectangle span the subspace t (4) . The coset space is given by which is interpreted as the symmetric homogeneous space of 3-hyperplanes. The four generators in t (4) play the role of translations on S (4) , while H (4) is the isotropy group of a 3-hyperplane.
Some remarks are in order. Firstly, the four spaces (39) are of positive constant curvature in the sense that their sectional curvature K is equal to +1 and they are endowed with a Riemmanian metric (so with positive definite signature). Secondly, the 4D spaces of points and 3-hyperplanes are of rank 1, that is, there is a single invariant under the action of SO(5) for a pair of points (the ordinary distance) or 3-hyperplanes. The 6D spaces of lines and 2-planes are of rank 2 [132,133] so that there are two independent invariants under the action of SO (5) for a pair of lines (an angle and the distance between two lines) or 2-planes (see [134] for the Euclidean case). And thirdly, there is a relevant automorphism of so(5) defined by leaving J 04 and J 13 invariant (up to the minus sign). Thus the map D interchanges the spaces of points and 3-hyperplanes, and the spaces of lines and 2-planes: Note that D 2 = Id. The map D will be called polarity, since for so(3) reduces to the well known duality in projective geometry interchanging the 2D space of points with the 2D space of lines (see [8] and references therein) which, only at this dimension, are both of rank 1. Note that this map is sometimes called ordinary duality [8], although in this paper we will always call it polarity in order to avoid confusion with the completely unrelated notion of quantum duality.

Lie Bialgebra
Let us consider the Drinfel'd-Jimbo quantum deformation of the real compact form g = so(5) [28][29][30] which in the basis (31) is generated by the following classical r-matrix [31] Recall that z is the quantum deformation parameter (such that q = e z ) and, hereafter, it will be assumed that z is an indeterminate real parameter. We remark that r 04,13 is a solution of the modified classical Yang-Baxter equation (14) so that this underlies a quasitriangular Hopf algebra structure. Therefore the corresponding cocommutator is coboundary [27], δ = δ(r), so that this is obtained from (47) through the relation (12), yielding The resulting real Lie bialgebra (so(5), δ(r 04,13 )) is so determined by the commutation rules (31) and cocommutator (48). The indices in r 04,13 (47) indicate the primitive generators J 04 and J 13 . The first primitive generator J 04 is the "main" one in the sense that, once the CK scheme of contractions be introduced and further applied to kinematical algebras, it will provide dimensions of the deformation parameter z since the product zJ 04 must be dimensionless [7,135]; this fact will be studied in detail in Section 6. The last term of r 04,13 is a classical r-matrix r 13 = zJ 23 ∧J 12 giving rise to the Drinfel'd-Jimbo Lie bialgebra (so (3), δ(r 13 )), with so(3) = span{J 12 , J 13 , J 23 } and primitive generator J 13 , which is a sub-Lie bialgebra of (so(5), δ(r 04,13 )); thus J 13 plays the role of a "secondary" primitive generator in r 04,13 [31].
Now we analyse how to implement the Z ⊗4 2 -grading of so(5) into (so (5), δ(r 04,13 )). This requires to generalize the action of the automorphisms Θ (m) : so(5) → so(5) (33) onto the cocommutator δ : so(5) → so(5) ⊗ so(5) and also to consider a possible action on the quantum deformation parameter [7,135]. Recall that δ (48) is the skew-symmetric part of the first-order term in z (17) of the full coproduct ∆ of the real quantum algebra U z (so (5) such that, as mentioned above, z is an indeterminate parameter. Since z is linked to the "main" primitive generator J 04 through the product zJ 04 , both elements must be transformed in the same way. By taking into account that J 04 → −J 04 under the four maps Θ (m) , then z → −z as well. Hence we define four z-maps (m = 1, . . . , 4) where Θ (m) is given in (33). Notice that the second relation in (49) can directly be applied to the r-matrix (47), consistently with the relation (12), and extended to higher-order tensor product spaces. When the z-maps (49) are applied either to δ (48) or to the r-matrix (47) one finds that only Θ (2) z and Θ (3) z remain as involutive automorphisms of (so(5), δ(r 04,13 )), meanwhile Θ (1) z and Θ (4) z are no longer involutions. This is a consequence of the presence of the term r 13 = zJ 23 ∧ J 12 in r 04,13 which does not appear neither in the Drinfel'd-Jimbo quantum deformation of so(3) [7] nor in so(4) [135]; for these latter deformations the whole initial Z ⊗2 2 -and Z ⊗3 2 -grading is kept, respectively, but for (so(5), δ(r 04,13 )) there only remains a Z ⊗2 2 -grading spanned by Θ Likewise, the polarity (44) can be implemented into (so(5), δ(r 04,13 )) by also considering an action on the deformation parameter defined by [7,135] with D given in (44), so that D 2 z = Id. It can be checked that the r-matrix (47) remains invariant under this "z-polarity" (and, obviously, the cocommutator (48) as well). Note that both primitive generators J 04 and J 13 are kept unchanged under D (up to the minus sign) as shown in (45).
From (48) it is straightforward to prove that this deformation fulfils the coisotropy condition (22) [57,59] for the four isotropy subalgebras h (m) (36): Thus each of them would provide a Poisson homogeneous space. Notice, however, that none of them leads to a Poisson subgroup since the condition (23) is not satisfied:

Dual Lie Algebra and Noncommutative Spaces
According to Section 2.1, we denote byx ab (a < b; a, b = 0, 1, . . . , 4) the generators in g * = so(5) * dual to J ab in g = so(5) with canonical pairing defined by (11): In contrast to the commutation rules of so(5) (31), the Lie brackets of so(5) * involving four different indices (54) are no longer equal to zero.
Next we express the dual Lie algebra so(5) * as the sum of two vector spaces where h are, in this order, the annihilators in so(5) * of the vector subspaces h (m) and t (m) of so(5) given in (34) and verifying (35). From the results presented in Section 2.2, it turns out that each h (m) ⊥ leads to a linear noncommutative space which is the first-order in the quantum coordinates of the full noncommutative space associated with the homogeneous space (39). We shall denote such a first-order noncommutative space by S Following [60,61] we analyse the relations h for each m. Such structures do depend on the chosen quantum deformation (here determined by (47)) and they are directly deduced from (53) and (54): 2. Noncommutative space of lines S 3. Noncommutative space of 2-planes S 4. Noncommutative space of 3-hyperplanes with S Consequenlty, the four first-order noncommutative spaces S (m) z close on a Lie subalgebra of ⊥ , as it should be, since this is a direct consequence of the coisotropy condition (51) [57,[59][60][61]. Furthermore, the noncommutative spaces of lines S The pairing (52) allows us to define the following maps in so(5) * from (33) by considering again an action on z: Similarly to (49), the action of θ (m) z can be extended to the tensor product space so(5) * ⊗ so(5) * , although we shall not make use of it in this paper. It can be checked that only θ (2) z and θ such that the map d is dual to D (44) through the pairing (52) and d z is an automorphism of so(5) * which, as expected, interchanges the noncommutative spaces of points and 3-hyperplanes, and the noncommutative spaces of lines and 2-planes: to be compared with (46).

Drinfel'd Double Structure
In [129] it was shown that the real Lie algebra so(5) has a classical r-matrix coming from a Drinfel'd double structure for the classical complex Lie algebra c 2 . We now review the main results according to the notation introduced in Section 2.3.
Let us consider the complex Lie algebra c 2 in a Chevalley basis with generators {h l , e ±l } (l = 1, 2) fulfilling the Lie brackets given by Then the 10 generators {h l , e ±m } with l = 1, 2 and m = 1, . . . , 4 span the Lie algebra c 2 in the Cartan-Weyl basis. As a shorthand notation we denote e m ≡ e +m and f m ≡ e −m so that the full commutation rules of c 2 read To unveil the Drinfel'd double structure for c 2 , we consider the linear combination of the two generators h 1 and h 2 belonging to the Cartan subalgebra given by [129]: Finally, the identification allows us to express the commutation relations (64) with the new Cartan generators (65) in the required form (26), thus obtaining a Drinfel'd double structure for c 2 with two 5D subalgebras g 1 = span{Y a ≡ e a } and g 2 = span{y a ≡ f a } (a = 0, . . . , 4), which are dual to each other by means of the canonical pairing (27).
• Lie subalgebra g 1 = span{e 0 , . . . , e 4 }: • Lie subalgebra g 2 = g * 1 = span{f 0 , . . . , f 4 }: From these results the real Lie algebra so(5) ∼ c 2 is obtained in the basis with generators {J ab } obeying the commutation rules (31) through the following change of basis [129]: , whose inverse reads The canonical pairing (27) now reads Then the canonical classical r-matrix (28) turns out to be is the ad-invariant element (29) corresponding to the tensorized expression of the Casimir C (32). Hence the skew-symmetric form for r can is obtained by substracting Ω, as in (30). We introduce explicitly the quantum deformation parameter z multiplying this result by 2iz as r D = 2iz(r can − Ω), obtaining the real r-matrix which in terms of the Drinfel'd-Jimbo classical r-matrix (47) considered for so (5) reads Hence a Reshetikhin twist with the commuting primitive generators must be added to (47) in order to obtain a classical r-matrix coming from a Drinfel'd double structure. Recall that it is possible to consider a generalized two-parametric r-matrix [136] showing the effects of the twist with quantum deformation parameter ϑ on the former deformation determined by r 04,13 and properly recovering the Drinfel'd double r-matrix whenever ϑ = z. We remark that bothr D and r z,ϑ are quasitriangular classical r-matrices (like r 04,13 ), so that they are solutions of the modified classical Yang-Baxter equation (14), while the twist itself determines a triangular r-matrix with vanishing Schouten bracket (16). In this sense,r D and r z,ϑ can be regarded as "hybrid" classical r-matrices [114,137].
4 The Drinfel'd-Jimbo Lie Bialgebra for the Cayley-Klein Algebra so ω (5) The Z ⊗4 2 -grading of so(5) generated by the four automorphisms Θ (m) (33) enables one to obtain a particular set of contracted real Lie algebras [17,20] through the graded contraction formalism [138,139]. These are the so-called orthogonal Cayley-Klein (CK) algebras or quasisimple orthogonal algebras [18,19,31,133] (see [140] for their description in terms of hypercomplex units). We collectively denote them by so ω (5) as this family of contracted algebras depends explicitly on four real graded contraction parameters ω = (ω 1 , ω 2 , ω 3 , ω 4 ). Alternatively, each contraction parameter ω m (m = 1, 2, 3, 4) can be introduced in the initial commutation rules of so(5) (31) by means of the following mapping provided by the involution Θ (m) (33): The composition of the four (commuting) mappings gives [31] Φ (J ab where the contraction parameter with two indices ω ab is defined by Next we apply the map (68) with all the factors √ ω ab = 0 onto the commutation rules of so (5) (31) obtaining the Lie brackets corresponding to the CK family so ω (5) which are given by without sum over repeated indices and with all the remaining brackets being equal to zero. This is just the same result coming from a particular solution of the Z ⊗4 2 -graded contraction equations for so(5) [17] (see [20] for the general solution). Explicitly, the non-vanishing commutation relations of so ω (5) read (71) We remark that although the factor √ ω ab = 0 in the map (68) can be an imaginary number, enabling to change the real form of the algebra, the resulting commutation relations (70) of so ω (5) only comprise real Lie algebras. Moreover, the zero value for ω ab is consistently allowed in (70), which is equivalent to apply an Inönü-Wigner contraction [13,31], leading to a more abelian (contracted) Lie algebra. Consequently, each graded contraction parameter ω m can take a positive, negative or zero value in (70), and when ω m = 0, it can be reduced to ±1 through a scaling of the Lie generators. Hence so ω (5) contains 3 4 = 81 specific real Lie algebras, being some of them isomorphic.
Moreover, the CK algebra so ω (5) (70) is always endowed with two non-trivial Casimirs regardless of the values of ω. One of them is the quadratic Casimir coming from the Killing-Cartan form which is given by [18] to be compare with (32). Observe that in the most contracted case, with all ω m = 0, C = J 2 04 . The second Casimir is a fourth-order one that can be found explicitly in [18] and this is related to the Pauli-Lubanski operator. Also in the most contracted case the fourth-order Casimir does not vanish. In this respect, we recall that the CK Lie algebras are the only graded contracted algebras from so(N + 1) [17,18] that preserve the rank of the semisimple algebra, understood as the number of algebraically independent Casimirs, which at this dimension is equal to two.
It is worth stressing that the polarity D (44) also remains as an automorphism of the whole family of CK algebras in such a manner that this map interchanges isomorphic Lie algebras within the family in the form so interchanging the contraction parameters ω 1 ↔ ω 4 and ω 2 ↔ ω 3 . Consequently, the CK algebras with ω 4 = 0 (76) are related, through D, to those with ω 1 = 0 (75) and so they are isomorphic; those with ω 4 = ω 3 = 0 are twice-inhomogeneous algebras and isomorphic to the ones with ω 1 = ω 2 = 0 (77); those with ω 3 = 0 are also Newton-Hooke type algebras isomorphic to (80); and the (single) flag algebra (81) remains unchanged under D.
We also recall that all the CK algebras in so ω (5) (even the flag algebra) have the same number of functionally independent Casimirs [18]. At this dimension, there are two (secondand fourth-order) Casimir invariants, exactly equal to the rank of the simple algebra so(5); for this reason they are also called quasisimple orthogonal algebras.

Symmetric Homogeneous Cayley-Klein Spaces
Since, by construction, the Z ⊗4 2 -grading is preserved for the CK algebra so ω (5), the same Cartan decompositions (34) in invariant h so But now from (70), we can express the relations (35) by taking into account the contraction parameter ω m : This, in turn, means that again for any value of ω m , h (m) ω is always a Lie subalgebra but the subspace t (m) becomes an abelian subalgebra when ω m = 0.
Next, as in Section 3.1 we construct the homogeneous CK spaces as the coset spaces [131][132][133] where SO ω (5) is the CK Lie group with Lie algebra so ω (5) and H (m) ω is the isotropy subgroup of SO ω (5) with Lie algebra h (m) ω . We recall that, usually, a CK geometry (6) is identified with the space of points S (1) ω , without taking into account other spaces. Along this paper, a CK geometry will be understood as the full set of the four homogeneous spaces (85).
The four spaces S 1. 4D CK space of points: 2. 6D CK space of lines: 3. 6D CK space of 2-planes: 4. 4D CK space of 3-hyperplanes: We stress that, strictly speaking, only the rank-one spaces S ω and S (4) ω are of constant curvature in the sense that all their sectional curvatures are equal to ω 1 and ω 4 , respectively. However, the rank-two spaces S ω are not, in general, of constant curvature in the above sense, but they are as close to constant curvature as a rank-two space would allow [132]. In particular, the sectional curvature K of the space of lines S (2) ω along any 2-plane direction spanned by any two tangent vectors (J 0i , J 0j ), (J 1i , J 1j ) and (J 0i , J 1i ) (i, j = 2, 3, 4) is constant and equal to ω 2 , but the remaining sectional curvatures could be different but proportional to ω 2 . When ω 2 = 0, S ω is a proper flat space with K = 0. And similarly for S ω .
By taking into account the above comments, we can say, roughly speaking, that the coefficients ω = (ω 1 , ω 2 , ω 3 , ω 4 ) that label the CK family so ω (5) are just the constant curvatures of the four aforementioned spaces. Therefore two isomorphic algebras in the family so ω (5) lead to two different sets of four homogeneous spaces through their corresponding Lie groups, and such sets of spaces are those which determine each specific CK geometry amongst the 81 ones. In this respect, we remark that the polarity D (44), that relates isomorphic CK algebras in the form (82), also interchanges the homogeneous CK spaces as in (46): For instance, the 4D Euclidean algebra iso (4)  It is worth remarking that it is possible to construct other 4D and 6D symmetric homogeneous spaces from the CK group SO ω (5) which, depending on each particular CK geometry, could be different from the four above ones (85). In particular, any composition of the automorphisms Θ (m) (33), which form a basis for the Z ⊗4 2 -grading, gives rise to another automorphism which provides another Cartan decomposition, like (83), and from it the corresponding coset space can be constructed. For instance, the composition Θ (1) Θ (4) leads to the 6D symmetric homogeneous space (to be compared with (41)), which can be interpreted as another 6D CK space of lines. This fact can clearly be appreciated in the Lorentzian spacetimes where there exist time-like and spacelike lines. In the single case of so (5) with ω = (+, +, +, +) all of such possible CK spaces are equivalent to the four spaces (39). Furthermore, it is also possible to obtain generalizations of the polarity D (44) relating such other homogeneous spaces belonging to different CK geometries with the same (isomorphic) CK algebra. In the 2D case shown in Table 1, the prefix "Co-" in the name of some CK geometries reminds the action of D that here interchanges ω 1 ↔ ω 2 , so keeping the three geometries in the diagonal unchanged. The full description of all the 2D CK spaces and the generalizations of the polarity D can be found in [7,12].

Cayley-Klein Lie Bialgebra
Let us start with the classical r-matrix r 04,13 (47) for so (5). The Lie bialgebra contraction procedure introduced in [31] shows that it is not only necessary to apply the contraction map Φ (68) to the Lie generators of so(5) in order to obtain a classical r-matrix for the CK algebra so ω (5), but additionally a possible transformation of the quantum deformation parameter z must be considered. We recall that the idea to transform the deformation parameter in contractions of quantum groups was formerly introduced in [153,154]. In our case, the coboundary Lie bialgebra contraction that ensures a well defined limit ω m → 0 (for any m) of both the classical r-matrix and the cocommutator for so(5) is given by the transformation [31] Ψ Then we apply the composition of the maps (68) and (91) to r 04,13 in the form Its Schouten bracket (15) turns out to be It can be checked that r (92) is a solution of the modified classical Yang-Baxter equation (14) for any Lie algebra within the CK family so ω (5) (so for any value of ω = (ω 1 , ω 2 , ω 3 , ω 4 )). The corresponding cocommutator can either be obtained from the Lie bialgebra contraction of (48) or through the relation (12) with (92) giving rise to the CK Lie bialgebra (so ω (5), δ(r)); namely Now some remarks are in order.
• The coisotropy condition (22) is always satisfied by the four CK subalgebras h but none of them fulfils the Poisson subgroup condition for any value of the contraction parameters (even for the flag algebra with all ω m = 0) So far we have obtained, in a unified setting, a family of coboundary Lie bialgebra structures (so ω (5), δ(r)), with quasitriangular classical r-matrix (92) and cocommutator δ (94), which covers 81 particular Lie bialgebras with the aforementioned properties. However, it is worth stressing that for some values of the contraction parameters the CK cocommutator could involve imaginary quantities due to term √ ω 1 ω 4 in the CK r-matrix, although the CK algebras are always real ones. If we require to have a real Lie bialgebra then which excludes the 18 cases with the following values for (ω 1 , ω 2 , ω 3 , ω 4 ): Complex Lie bialgebras: (+, ω 2 , ω 3 , −) and (−, ω 2 , ω 3 , +), ∀ω 2 , ω 3 .
We remark that the commutation relations (98) and (99) define a real dual CK algebra so(5) * under the constraint (97), thus covering the 63 cases given in Table 2. Note also that all the commutators (99) vanish for either ω 1 = 0 or ω 4 = 0, corresponding to the dual algebra of inhomogeneous algebras and their contractions.
Similarly to (55), we express the dual CK algebra so ω (5) * as the sum of two vector spaces  z,ω is the first-order in the quantum coordinates of the complete noncommutative space associated with the homogeneous CK space S (m) ω (85). We display in Table 3 the defining commutation relations for the four noncommutative CK spaces along with the Lie brackets among h    Table 2.
• Noncommutative CK space of points S Now we analyse the structure and properties of such noncommutative CK spaces which do strongly depend on the contraction/curvature parameters. The four noncommutative spaces close on a Lie subalgebra h (m) ⊥,ω , in agreement with the coisotropy condition (96), and the noncommutative spaces of lines and 2-planes are both reductive and symmetric as it was also the case for so(5) * (see (60)). Furthermore, the explicit presence of the curvature parameters allows us to highlight some properties for the contracted noncommutative spaces straightforwardly. In particular, if we set ω 4 = 0 in the noncommutative space of points we find that ⊥,ω , t ⊥,ω , t ⊥,ω .
Likewise, taking ω 1 = 0 in the noncommutative space of 3-hyperplanes we obtain that ⊥,ω . Thus both contracted noncommutative spaces are reductive and symmetric (to be compared with (84)). And a remarkable common property for the four noncommutative spaces is that when ω m = 0, the subspace t (m) ⊥,ω becomes an abelian subalgebra (m = 1, 2, 3, 4): ⊥,ω = 0. Such relations can be applied, for instance, to the inhomogeneous Poincaré and Euclidean algebras with ω 1 = 0 (75) for the noncommutative space of points, to the Newton-Hooke type algebras with ω 2 = 0 (80) for the noncommutative space of lines, to the twice inhomogeneous algebras (Galilei) with ω 1 = ω 2 = 0 (77) for the noncommutative spaces of points and lines, and so on up to reach the flag algebra (81) for the four noncommutative spaces.
Finally, as we advanced at the end of Section 2.2, it is worth stressing that the structure of the first-order noncommutative CK space of points S (1) z,ω , shown in Table 3, is shared by the 63 CK real Lie bialgebras with ω 1 ω 4 ≥ 0 of Table 2 since no ω m appears within S (1) z,ω , and that this is formally similar to the κ-Minkowski spacetime (25). Furthermore, the commutation relations of S (1) z,ω are kept linear under full quantization for the 27 CK bialgebras with parameters (0, ω 2 , ω 3 , ω 4 ), while higher-order terms in the quantum coordinates are expected for the CK bialgebras with ω 1 = 0. Similar properties hold for the first-order noncommutative CK space of 3-hyperplanes S (4) z,ω which remains as a linear full noncommutative space for the 27 CK bialgebras with parameters (ω 1 , ω 2 , ω 3 , 0). Nevertheless, if one looks at the four first-order noncommutative CK spaces in Table 3 altogether, then one finds that the four contraction parameters appear explicitly. Thus it turns out that the set of four noncommutative spaces S (m) z,ω is different for each specific CK bialgebra except for the nine cases with ω 1 = ω 4 = 0, for which all the terms involving any ω m vanish. Consequently, this observation suggests the necessity of constructing other noncommutative spaces beyond the usual noncommutative spacetime for a given quantum deformation. To the best of our knowledge, there are very scarce results in this research direction which concern noncommutative spaces of lines [99,100].
A physical (kinematical) analysis on the noncommutative CK spaces of points and lines will be addressed in Section 6.

Drinfel'd Double Structures for Cayley-Klein Algebras
Let us consider the classical r-matrixr D (66) coming from the Drinfel'd double structure of so (5). We apply the composition of the contraction maps (68) and (91) in the form ∀ ω m = 0, Table 4: Simple Lie algebras with a real r-matrix r D (101) coming from a Drinfel'd double structure according to the sign of the graded contraction parameters ω = (ω 1 , ω 2 , ω 3 , ω 4 ) and bilinear form I ω (74), along with their contractions to non-simple Lie algebras endowed with a real Lie bialgebra (so ω (5), δ D (r D )).

Simple Lie algebras with a Drinfel
obtaining that which is a superposition of the classical CK r-matrix r (92) with a Reshetikhin twist J 13 ∧ J 04 formed by the two commuting primitive generators. We remark that, by construction, r D is a classical r-matrix coming from the Drinfel'd double structure for the simple Lie algebras contained in the CK family so ω (5). Moreover, r D leads to the same Schouten bracket as for r (93) (there are no twist contributions) so that it is a solution of the modified classical Yang-Baxter equation (14).
If we now require r D (101) to define a real Lie bialgebra, (so ω (5), δ D (r D )), we have to impose the restriction corresponding to r (97) together with the new one determined by the twist: which lead to four possible cases shown in Table 4, where we have named them according with their kinematical interpretation that we shall show in Section 6.
The effect of the twist J 13 ∧ J 04 in r D (101) with respect to the CK r-matrix (92) can be highlighted by associating it with a second deformation parameter ϑ in a similar form to (67), that is, such that the r-matrix coming from a Drinfel'd double structure corresponds to the one-parame- The cocommutator δ D , obtained with (12), is just the CK cocommutator δ (94) plus new terms coming from the twist which are denoted δ ϑ . Hence δ D = δ + δ ϑ with δ ϑ given by Consequently, (so ω 2 ,ω 3 (3), δ D ) with generators J 12 , J 13 , J 23 does not remain as a Lie subbialgebra of (so ω (5), δ D (r D )). However, if this is enlarged with the primitive generator J 04 then it provides the Lie sub-bialgebra (so ω 2 ,ω 3 (3) ⊕ so ω 04 (2), δ D ).
The polarity D z (50) and the involutions Θ (2) z and Θ (3) z (49) also hold for the two-parametric deformation determined by (103) provided that ϑ is unchanged. Nevertheless, in the proper Drinfel'd double case with a single deformation parameter z, with the identification (104), the above maps do not remain since z → −z.
As far as the first-order noncommutative spaces associated to (so ω (5), δ D (r D )) is concerned, it is straightforward to prove that the coisotropy condition (22) is only fulfilled for the subalgebras h (1) ω and h (4) ω (see (105)): for any of the 14 Lie algebras displayed in Table 4. Therefore, only the twisted noncommutative CK spaces of points and 3-hyperplanes S (l) z,ϑ,ω (l = 1, 4) can consistently be constructed. In particular, from (105) and applying the quantum duality (10) with pairing (52), we directly obtain the defining commutation relations for the twisted noncommutative CK space of points S (1) z,ϑ,ω : In the same way, S z,ϑ,ω can also be obtained. Clearly S (1) z,ϑ,ω (106) is not isomorphic to S (1) z,ω given in Table 3. Moreover, since ω 2 ω 3 > 0, this factor can be scaled to +1 within the commutators (106) via the scalingsx which shows that S (1) z,ϑ,ω is the common first-order twisted noncommutative CK space of points for the 14 Lie bialgebras shown in Table 4; obviously, higher-order terms in the quantum coordinates may arise for each specific case.
Finally, we stress that it is not ensured at all that a given contracted Drinfel'd double rmatrix r D gives rise to a Drinfel'd double structure for a non-semisimple Lie algebra and, in fact, this problem should be studied case by case. Nevertheless, we can answer negatively to this question for the contracted r-matrices of Table 4. It was established in [127], from the results given in [145], that there does not exist any Drinfel'd double structure for Poincaré, Euclidean and Carroll algebras at this dimension. In contrast, as we commented at the end of Section 2.3, in lower dimensions such structures do exist and the classification of Drinfel'd doubles has recently been performed for the (2+1)D Poincaré [127] and 3D Euclidean algebras [128]. Moreover, to the best of our knowledge, the classification of Drinfel'd doubles for the (anti-)de Sitter algebras has only been carried out in (2+1) dimensions [115]. In (3+1) dimensions there is no such classification for the simple Lie algebras so(p, q) and there has only been constructed the Drinfel'd double structure here considered for so (5), reviewed in Section 3.4, and from it a Drinfel'd double for the anti-de Sitter algebra so(3, 2) [129], that we advance which is just the case (III) in Table 4. Therefore we have obtained two new r-matrices coming from Drinfel'd doubles, one for the de Sitter so(4, 1) and another for the anti-de Sitter so(3, 2) (cases (II) and (IV)), although our results do not convey a complete classification.
The physical (kinematical) interpretation of the CK r-matrices r (92) and r D (101) along with their associated first-order noncommutative spaces will be described in detail in Section 6.

Kinematical Algebras and Homogeneous Spaces
As we have already mentioned in the previous Section, the kinematical algebras introduced in [14] arise as particular cases of graded contractions of so(5) [20,152] so that they appear within the CK family so ω (5) for some specific values of the contraction parameters ω = (ω 1 , ω 2 , ω 3 , ω 4 ) (except for the static algebra which does not belong to the CK family). These kinematical algebras have recently been derived from deformation theory in [145,155]; in this respect, recall that Lie algebra deformations [156] can be regarded as the opposite processes to Lie algebra contractions [13,138,157,158].
In order to deal with kinematical algebras let us introduce a physical basis denoting by P 0 , P = (P 1 , P 2 , P 3 ), K = (K 1 , K 2 , K 3 ) and J = (J 1 , J 2 , J 3 ) the generators of time translations, spatial translations, boosts and rotations, respectively. These ten generators are isometries of a (3+1)D spacetime of constant curvature. The 11 kinematical algebras [14] are contained within a three-parametric Lie algebra, here denoted so Λ,c,λ (5), with commutation relations given by where from now on the indices i, j, k = 1, 2, 3 and sum over repeated indices will be understood.
Recall also that the commutators (107) are a consequence of 3-space isotropy [14] and they are shared by any Lie algebra in so Λ,c,λ (5), while the Lie brackets (108) distinguish the specific kinematical algebra according to the values of the real parameters Λ, c and λ.
The family so Λ,c,λ (5) has two Casimir operators: a quadratic one, coming from the Killing-Cartan form, which is given by and a fourth-order Casimir [18] (with the exception of the static algebra [14] corresponding to set Λ = λ = 0 and c → ∞ in (108), so all of these brackets vanish).

(110)
Each of them provides a type of contraction: the composition PT corresponds to the (flat) spacetime contraction (Λ → 0), the parity P to the speed-space contraction (c → ∞), and the time-reversal T to the speed-time contraction (λ → 0) (see [135] for the (2+1)D kinematical algebras and contractions within the CK family so ω (4) and their Drinfel'd-Jimbo quantum deformation). In other words, the quantities Λ, 1/c 2 and λ behave as graded contraction parameters, each of them corresponding to the Z 2 -grading of so Λ,c,λ (5) determined by PT , P and T , respectively.
From the Lie group SO Λ,c,λ (5) of so Λ,c,λ (5) we construct the (3+1)D spacetime and the 6D space of lines as the coset spaces such that H st and H line are the isotropy subgroups of an event and a line, respectively. Thus these are symmetric homogeneous spaces associated, in this order, with the composition PT and parity P involutions. Similarly to the discussion on the curvature of the CK spaces (86)- (89) in Section 4.1, we remark that the (3+1)D spacetime ST 3+1 is a rank-one homogeneous space such that all their sectional curvatures K are equal and constant. However, the 6D space of lines L 6 is of rank-two and only the sectional curvatures K of any 2-plane direction spanned by any two tangent vectors (P i , P j ), (K i , K j ) and (P i , K i ) (i, j = 1, 2, 3) are equal among themselves and constant, being the remaining ones, (P i , K j ) with i = j, generically non-constant (or zero).
Furthermore, when SO Λ,c,λ (5) is a non-simple Lie group, the metric on either space (111) could be degenerate and, in this case, an invariant foliation arises so that an additional metric defined on each leaf of the foliation is necessary to determine completely the metric structure of the space [132]. Moreover, it is important to take into account that, in principle, the (3+1)D spacetime ST 3+1 does not necessarily coincides with the CK space of points S (1) ω (86) (in most cases it does), and likewise with the 6D space of lines L 6 with respect to the CK space of lines S (2) ω (87). Nevertheless, they can always be identified with another CK space, as for instance (90) for the space of lines. This fact will depend on the kinematical assignation of the geometrical CK generators that we shall study next in Section 6.
In what follows we describe the 11 kinematical algebras although we shall only focus on the homogeneous spaces (111) for nine of them: the Lorentzian, Newtonian and Carrollian cases. Additionally, we shall show how the three classical Riemannian algebras (and their homogenous spaces) can also be recovered from the family so Λ,c,λ (5). These nine kinematical algebras/spaces plus the three Riemannian ones are those which will appear in Section 6 and they are summarized in Table 5.

Lorentzian Algebras
If we set the parameter λ = 1 and consider c finite, we find that so Λ,c,λ (5) covers the three Lorentzian algebras l Λ of relativistic (3+1)D spacetimes such that the Lie brackets (108) now read where c is the speed of light and Λ is the cosmological constant. Then we obtain the de Sitter (dS) l + = so(4, 1), anti-de Sitter (AdS) l − = so(3, 2) and Poincaré l 0 = iso(3, 1) algebras. The quadratic Casimir (109) for l Λ reads and the fourth-order Casimir, related to the Pauli-Lubanski 4-vector, can be found in [18].

Newtonian Algebras
The non-relativistic limit c → ∞ (or speed-space contraction) of l Λ (112) gives rise to three Newtonian algebras n Λ with Lie brackets where Λ = ±1/τ 2 and τ is again the time universe radius. The second-order Casimir (113) reduces to C = −P 2 + ΛK 2 and the corresponding fourth-order Casimir can be found in [18]. This non-relativistic limit is obtained by setting λ = 1 and c → ∞ in (108) and (109). In this way we find the expanding Newton-Hooke (NH) n + , oscillating NH n − and the Galilei n 0 ≡ iiso(3) algebras, which have the following structure (see (80) and (77), respectively): (3) : The isotropy subgroup H st (111) is now the 3D Euclidean subgroup ISO(3) spanned by rotations and (commuting) Newtonian boosts and the (3+1)D spacetime has the same sectional curvature as in the Lorentizan spacetimes: K = −Λ. The metric is degenerate and corresponds to an "absolute-time", so that there exists an invariant foliation under the action of the Newtonian group N Λ , whose leaves are defined by a constant time, which is determined by a 3D non-degenerate Euclidean spatial metric restricted to each leaf of the foliation [132,147]

Carrollian Algebras
We set λ = 0 and c = 1 in the commutators (108) yielding the three Carrollian algebras c Λ with Lie brackets and second-order Casimir given by Notice that now the parameter Λ has dimensions of length −2 instead of time −2 and the Carrollian boosts have dimensions of speed instead of speed −1 (which were the cases in the Lorentizan and Newtonian algebras).
The algebra c Λ comprises the Para-Euclidean algebra c + ≡ i ′ so(4) (isomorphic to the Euclidean iso(4)), the Para-Poincaré algebra c − ≡ i ′ so(3, 1) (isomorphic to the Poincaré iso(3, 1)) and the proper Carroll algebra c 0 ≡ ii ′ so(3) ≡ i ′ iso(3), which have the following structure [147] (see (76), (78) and (79)): The isotropy subgroup H st (111) is again the 3D Euclidean subgroup ISO(3) spanned by rotations and (commuting) Carrollian boosts, but the (3+1)D spacetime has sectional curvature K = +Λ (instead of K = −Λ as in the Lorentizan and Newtonian spacetimes). The metric is degenerate corresponding to an "absolute-space" and there exists an invariant foliation under the action of the Carrollian group characterized by a 1D time metric restricted to each leaf of the foliation [147]. The isotropy subgroup of a line is H line = R ⊗ SO(3) and the homogeneous space of lines (111) has the same curvature as the spacetimes so equal to +Λ.

The Two Remaining Kinematical Algebras
For the sake of completeness, we also mention that the Para-Galilei algebra [14] arises for λ = 0 and c → ∞, that is, the commutators (108) reduce to for any value of Λ = 0 (apply the map P 0 → ±P 0 /Λ), while the second-order Casimir (109) simply reads C = ΛK 2 . And the static algebra [14] corresponds to the most contracted algebra within the kinematical family for Λ = λ = 0 and c → ∞, with trivial second-order Casimir C = 0. In fact, the static algebra is the only kinematical one which does not appear within the CK family so ω (5) [17], but it can be obtained from the general solution of the grading equations for so(5) [20,152]. Observe that the static algebra is not a quasisimple Lie algebra in the sense that it does not have the same number of Casimir invariants as the simple Lie algebra so(5).
When one compares the commutation relations of the static algebra (115) with those for the Carroll one (114) with Λ = 0, one finds that the Carroll algebra can be regarded as a centrally extended algebra, with non-trivial central extension P 0 , from the static algebra, and there cannot be added any other central extension to the Carroll algebra [14] (see [19] for the central extensions of the CK algebras in any dimension). In this respect, we remark that in [145] the (3+1)D kinematical algebras have been constructed from the static algebra through deformation theory (see [155] for higher dimensions). We also recall that twist deformations for the Para-Galilei, static and Carroll algebras have been obtained in [146].
Neither the Para-Galilei nor the static algebra will appear within the deformations that we shall describe in Section 6 so that they are omitted in Table 5.

Riemannian Algebras
Additionally, but not kinematically, we can set λ = 1 and the speed of light equal to the imaginary unit c = i in (108) finding the commutators with second-order Casimir (109) given by In this way, we obtain so(5) for Λ < 0, iso(4) for Λ = 0 and so(4, 1) for Λ > 0. The generator P 0 now behaves as another space translation, while the generators K are no longer boosts but rotations. The isotropy subgroup H st (111) is SO(4) = K, J , such that we recover the three classical 4D Riemannian spaces of constant sectional curvature K = −Λ: spherical (K > 0), Euclidean (K = 0) and hyperbolic (K < 0) spaces. The isotropy subgroup of a line is H line = SO −Λ (2) ⊗ SO(3) and the corresponding 6D space of lines has positive curvature K = +1 for any value of Λ [132].

Kinematical Lie Bialgebras and Noncommutative Spaces
Our aim now is to interpret in the kinematical framework the Lie bialgebras coming from the classical CK r-matrix r (92) and the Drinfel'd doubles further provided by r D (101) together with the corresponding first-order noncommutative spaces of points and lines displayed in Table 3 and the twisted one (106). With this in mind, we shall perform different identifications between the "geometrical" generators J ab of so ω (5) (71) and the kinematical ones of so Λ,c,λ (5) (107) and (108), which will convey physical correspondences between the contraction/curvature CK parameters ω and Λ, c, λ.
According to [102], the main idea is to start with the main primitive generator J 04 and to identify it either with a spatial translation P i or with the time translation P 0 . Since the product zJ 04 must be dimensionless we shall obtain the so-called [135] "space-like" deformations J 04 ≡ P i , with the deformation parameter z being a fundamental length scale, and the "timelike" deformations J 04 ≡ P 0 , with z being a fundamental time scale. In particular, we shall study firstly three classes of kinematical deformations, called A, B and C, such that their properties are determined by the two primitive (undeformed) generators (J 04 , J 13 ) corresponding to (P 3 , K 2 ), (P 2 , J 2 ) and (P 0 , J 2 ), respectively. We remark that in these three classes the time translation generator P 0 ≡ J 0l for l = 1, 2, 4, and the remaining case P 0 ≡ J 03 would provide results which are equivalent, under certain Lie algebra automorphisms, to those already contained in Table 6: Four classes of real classical r-matrices for the kinematical and Riemannian algebras with commutation relations (107) and (108). For each class, it is displayed the dimensions of the quantum deformation parameter z and the primitive generators (determined by (J 04 , J 13 )), the r-matrix, the specific Lie algebras according to the values of the graded contraction parameters (Λ, c, λ) as in Table 5, the CK parameters ω, and the Drinfel'd double r-matrix r D together with the corresponding case given in Table 4.

#
z Kinematical real r-matrices and Lie algebras the class A, so that we omit it. Therefore, the classes A and B will give rise to space-like deformations, while the class C will lead to time-like ones. Additionally, we shall construct an AdS Lie bialgebra for which z is dimensionless with primitive generators (J 2 , P 0 ), and it will correspond to the new Drinfel'd double of case (IV) in Table 4; we shall call it class D. We point out that such four classes of kinematical deformations will contain the four Drinfel'd doubles for the simple Lie algebras of Table 4.
The main results that will be obtained along Sections 6.1-6.4 concerning the kinematical rmatrices are presented in Table 6. From them, their corresponding first-order noncommutative spacetimes and spaces of lines will be computed being summarized in Table 7. Comments on these results will be performed in Sections 6.5 and 6.6. Table 7: The first-order noncommutative spacetimes ST 3+1 z = x 0 ,x 1 ,x 2 ,x 3 and spaces of lines L 6 z = x 1 ,x 2 ,x 3 ,ξ 1 ,ξ 2 ,ξ 3 for the classes A, B and C of kinematical and Riemannian Lie bialgebras shown in Table 6 with the notation of Table 5; the case D has no associated noncommutative space. First-order twisted noncommutative spacetimes coming from the Drinfel'd double structures and their contractions presented in Table 4 are also written in terms of the twist deformation parameter ϑ such that ϑ ≡ z corresponds to the proper (or contracted) Drinfel'd double.
When we impose, under the identification (117), that the commutation rules of so ω (5) (71) fulfil the common Lie brackets of any kinematical algebra (107) we find that ω 3 = ω 4 = +1. Next the remaining specific kinematical commutation relations (108) imply that ω 1 = −Λ, ω 2 = −1/c 2 and λ = 1. In this way we find a set of six kinematical algebras in so ω (5) with graded contraction parameters and bilinear form I ω (74) given by These are the three Lorentzian (c finite) and the three Newtonian (c = ∞) algebras described in Sections 5.1 and 5.2. Moreover, the three Riemannian algebras of Section 5.5 also appear for c = i, that is, ω 2 = +1. Thus this class A cover nine of the Lie algebras shown in Table 5. Recall that in the Lorentzian cases the sectional curvature of the (3+1)D spacetime ST 3+1 is minus the cosmological constant ω 1 = −Λ, while the 6D space of (time-like) lines L 6 is of negative curvature ω 2 = −1/c 2 . In the Newtonian cases, L 6 is a flat space with ω 2 = 0 (c = ∞). The kinematical automorphisms (110) are related to the CK ones Θ (m) (33) through Now we apply the geometrical-kinematical identification (118) to the CK r-matrix r (92) and to the Drinfel'd double one r D (101) obtaining the following kinematical r-matrices with primitive generators P 3 ≡ J 04 and K 2 ≡ J 13 , so that z has dimensions of a length with dimensionless product zP 3 . Next, the constraint (97) ω 1 ω 4 = −Λ ≥ 0 excludes three cases in order to deal with real bialgebras: dS, expanding NH and hyperbolic algebras, all of them with Λ > 0 (first column in Table 5). And the Drinfel'd double r-matrix r D subjected to the additional condition (102) ω 2 ω 3 = −1/c 2 > 0 only holds for c = i, that is, for the spherical and Euclidean algebras so recovering the cases (I) and (Ia) in Table 4. Thus these results finally comprise six real Lie bialgebras shown in Table 6.
The cocommutator δ for r (120) can then be deduced by applying (12), or by introducing directly the kinematical assignations (118) and (119) into the CK cocommutator (94). It can be checked that the isotropy subalgebras of an event h st and a line h line given by (see (111)) both satisfy the coisotropy condition (22).
Now we proceed to obtain the corresponding first-order noncommutative spacetimes ST 3+1 z and spaces of (time-like) lines L 6 z associated with the homogeneous spaces (111). For this purpose, we introduce the quantum coordinates (x 0 ,x i ,ξ i ,θ i ) dual, in this order, to the generators (P 0 , P i , K i , J i ) (i = 1, 2, 3) via the canonical pairing (11), so with non-zero entries: Hence the first-order noncommutative spaces are defined as the annihilators of the vector subspaces h st and h line (121): Notice that for the Riemannian cases the quantum time coordinatex 0 corresponds to a spatial one, while the noncommutative rapiditiesξ i become quantum angular coordinates. The corresponding defining commutation relations for these noncommutative spaces can be deduced either from the dual of the cocommutator δ for r (120), or by introducing the following identification between the quantum CK coordinatesx ab and the kinematical ones (122) in S (1) z,ω and S (2) z,ω given in Table 3:x 01x02x03x04 x 12x13x14 x 23x24 (which is the dual counterpart of (118)), together with (119). Likewise, the first-order twisted noncommutative spaces of points can be obtained from S (1) z,ϑ,ω (106) by taking into account that it only covers the spherical and Euclidean spaces with c = i (ω 2 ω 3 = +1) so thatx 0 is another quantum spatial coordinate; recall that the proper Drinfel'd double structure corresponds to set ϑ ≡ z (104). All of these noncommutative spaces are explicitly presented in Table 7. c = 1 and ω 1 = Λ. Hence we obtain the three Lorentzian algebras of Section 5.1 within so ω (5) with contraction parameters (ω 1 , ω 2 , ω 3 , ω 4 ) = Λ, −1, −1, +1 , λ = 1, c = 1, (126) and bilinear form I ω (74) given by In terms of Θ (m) (33), the kinematical automorphisms (110) read With the assignations (124) and (126) we find that the space S (1) ω (86) is related to the spacetime ST 3+1 (since PT = Θ (1) ), but the former has curvature K = ω 1 , while the latter has K = −ω 1 = −Λ. The space of (time-like) lines L 6 (111) cannot be identified with S (2) ω (87) (now P = Θ (2) ), but it can be so with the rank-2 CK space associated with the composition of involutions P = Θ (1) Θ (2) Θ (3) (see the comments at the end of Section 4.1).
The r-matrices (92) and (101) turn out to be The primitive generators are P 2 ≡ J 04 and J 2 ≡ −J 13 , and z has dimensions of a length since the product zP 2 is dimensionless; notice that (127) is written in units with c = 1. The constraint ω 1 ω 4 = Λ ≥ 0 excludes the AdS algebra. Moreover, since ω 2 ω 3 = +1, the r-matrix r D is well defined for the dS and Poincaré cases, which correspond to the cases (II) and (IIa) in Table 4, as shown in Table 6. The cocommutator for r (127) can be obtained straightforwardly showing that the coisotropy condition (22) is satisfied for h st (121) in both cases, but only for h line for the Poincaré bialgebra, thus precluding the construction of the noncommutative dS space of (time-like) lines L 6 z (123). We stress that to set c = 1 in (126) implies that the r-matrices (127) are not well defined neither for the non-relativistic algebras with c → ∞, nor for the Riemannian ones with c = i (remind that λ = 1). In fact, the speed of light can be introduced explicitly in r and r D (127) providing the commutation rules (108) by means of the scalings (that preserve the product zP 2 =zP 2 ) yielding the classical r-matrices showing the above exclusions. Note also that it is possible to transform the deformation parameter asz = c 2 z (without preserving zP 2 ) obtaining that which is not real for c = i but both of them have a well defined limit c → ∞ reducing to a Reshetikhin twistr ≡r D =z √ ΛK 3 ∧K 1 .
Next we construct the two first-order noncommutative spacetimes ST 3+1 z and the noncommutative Minkowskian space of (time-like) lines L 6 z (123) by means of the dual of the cocommutator for r (127) or, alternatively, by introducing the kinematical identification x 01x02x03x04 x 12x13x14 x 23x24 dual to (125), together with the contraction parameters (126) in the commutation relations of the dual CK algebra (98) and (99). Similarly, the first-order twisted noncommutative dS and Minkowskian spacetimes can be deduced from S (1) z,ϑ,ω (106) (ω 2 ω 3 = +1). All of these structures are presented in Table 7.
the resulting seven real Lie bialgebras given in Table 6 also appear in Table 4 with the simple algebra AdS corresponding to the case (III). The spherical and Euclidean algebras (cases (I) and (Ia)) are again recovered as in class A, but here for different values for the parameters (Λ, c, λ). Once the cocommutator for r (131) has been computed it can be checked that the coisotropy condition (22) is fulfilled for both subalgebras h st and h line (121) allowing the construction of the two noncommutative spaces (123).
It is worth stressing that this class of time-like deformations cover the so-called kappadeformations such that the deformation parameters z and κ are related through z ∼ 1/κ. Hence the r-matrix for the Poincaré algebra of case (IIIa) in Table 4 underlies the well known κ-Poincaré deformation [64][65][66][67][68], and that for the AdS algebra of case (III) provides the κ-AdS algebra [129,136].
As far as the non-relativistic limit c → ∞ is concerned, we remark that the condition c = 1, in principle, precludes it. To be precise, if we apply the same scalings (128) to the Lie generators keeping z unchanged and so the product zP 0 as well, then c appears explicitly in the commutators (108) and the r-matrices (131) now read which diverge under the limit c → ∞. Nevertheless, we can introduce the scalings (128) but with transformed deformation parameterz = c 2 z (not preserving zP 0 ) finding that [71] r =z( , thus allowing one to apply the limit c → ∞ obtaining that which coincides with the κ-Poincaré r-matrix. The remarkable point is that the corresponding cocommutator is trivial, that is, δ(X) = 0 for all X, so that there is no deformation for the (contracted) Newtonian algebras. In other words, the scheme of contractions for the CK rmatrix (92) and cocommutator δ (94) ensures to always obtain both a non-trivial r-matrix and cocommutator. In this respect, it is worth stressing that it is possible to apply a Lie bialgebra contraction in such a manner that the initial r-matrix diverges but the initial cocommutator gives a non-trivial result [31] (this is called fundamental but non-coboundary Lie bialgebra contraction). This contraction process was applied in [65,159] in order to obtain the κ-Galilei algebra by contracting κ-Poincaré, finding a non-coboundary quantum Galilei algebra which explains its absence in our approach. In fact, such a non-coboundary Lie bialgebra contraction has recently been applied in [71] in order to deduce the κ-Newtonian algebras containing both κ-NH algebras (Λ = 0) together with the above κ-Galilei one (Λ = 0).
The first-order noncommutative spacetimes and spaces of lines for the seven Lie bialgebras contained in this class can be obtained by introducing the identification dual to (129) given bŷ together with the contraction parameters (130) in the commutation rules (98) and (99). In the same way, the first-order twisted noncommutative spacetimes are deduced from S z,ϑ,ω (106) for the seven cases since for all of them ω 2 ω 3 = +1. The explicit expressions for all of these noncommutative spaces can be found in Table 7. 6.4 Class D: Dimensionless Deformation with Primitive Generators (J 2 , P 0 ) As the last class we study how to obtain the AdS deformation of case (IV) in Table 4 in a kinematical basis. With this aim we consider the identification (not considered in [102]) given by Then the Lie brackets (107) gives that ω 1 = ω 2 = ω 3 = ω 4 = −1 and the commutators (108) lead to set Λ = −1, c = 1 and λ = 1. Therefore we obtain a single Lie algebra in this class, AdS ≃ so (3,2), such that (ω 1 , ω 2 , ω 3 , ω 4 ) = (−1, −1, −1, −1 , Λ = −1, c = 1, λ = 1, The kinematical automorphisms (110) read And the CK r-matrices (92) and (101) turn out to be The primitive generators are J 2 ≡ J 04 and P 0 ≡ J 13 , while z is dimensionless like the product zJ 2 (we are working with units with Λ = −1 and c = 1). Note that the constraints ω 1 ω 4 = +1 and ω 2 ω 3 = +1 are automatically satisfied, so that r D is the kinematical expression of the new Drinfel'd double r-matrix of case (IV) in Table 4. By computing the cocommutator for r (133) (or from (94) with the identification (132)), it can be checked that the coisotropy condition (22) is not satisfied for any subalgebra (121), so that there do not exist noncommutative spacetime and space of lines (123) associated with this bialgebra.
It is rather natural to analyse whether there may exist some possible contraction from this AdS deformation, although by following our approach the answer is negative whenever one requires to keep a non-trivial r-matrix and cocommutator. Starting from the AdS commutation relations (108) with Λ = −1, c = 1 and λ = 1, it is possible to introduce explicitly such parameters by means of the scalings (coming from the automorphisms (110)) keeping the dimensionless parameter z. By introducing (134) in r (133) we obtain that so that this diverges under the contractions Λ → 0, c → ∞ and λ → 0. If we transform the deformation parameter asz = − z c 2 Λλ then the above contractions are well defined but only provide twisted r-matrices whose terms are all formed by commuting generators. Tables 6 and 7 highlight the main results so far obtained from a global kinematical viewpoint; the former covers all the information of kinematical bialgebras, while the latter shows their corresponding first-order noncommutative spacetimes and spaces of lines. We now make some observations and also comment on known results as well as on some open problems concerning the kinematical bialgebras and their complete quantum deformation. Similarly, some remarks for the full noncommutative spaces will be addressed in Section 6.6.

Quantum Kinematical Algebras
All the kinematical r-matrices presented in Table 6 underlie quantum kinematical algebras U z (so Λ,c,λ (5)) with real Lie bialgebras (so Λ,c,λ (5), δ(r)) of quasitriangular or standard type as it was described in Section 2.1. If we focus on the Poincaré bialgebra we get three sequence of coboundary contractions: Recall that the class D only contains an isolated AdS bialgebra. From this approach, the nonrelativistic contraction leading to a Galilei bialgebra can only be performed within the space-like deformations belonging to the class A, but none of the three bialgebras in this sequence can be endowed with a (contracted) Drinfel'd double structure. The only possibility to get a spacelike Poincaré bialgebra with associated contracted Drinfel'd double structure is provided by the class B, but now coming from a dS bialgebra, instead of the AdS one of class A. Notice that the difference between the classical r-matrices r D of classes A and B can clearly been appreciated in the kinematical basis. In the class A there appears the term zK 2 ∧ P 3 , which only holds for the Riemannian bialgebras (with c = i and K 2 becoming a rotation generator), while in the class B the Drinfel'd double structure requires to add the term −zJ 2 ∧ P 2 with a proper rotation generator. The sequence for the class B corresponds to perform (II) → (IIa) in Table 4.
The complete Hopf algebra structure for the quantum inhomogeneous kinematical algebras and their further contractions can be found in [102], which belong to the quantum CK family U z (so ω (5)) with arbitrary ω = (0, ω 2 , ω 3 , ω 4 ). Thus such results comprise the quantum deformations of the Poincaré, Galilei and Euclidean bialgebras of class A, the Poincaré bialgebra of class B, along with the κ-Poincaré and κ-Carroll ones of class C; observe that the Euclidean bialgebra of class C is equivalent to that of class A, being just case (Ia) in Table 4. The κ-deformation for the curved Carrollian bialgebras (with Λ = 0) of class C, κ-Para-Euclidean and κ-Para-Poincaré, can be deduced from the results given in [102] by applying the z-polarity D z (50) since this map interchanges the CK bialgebras as in (95), (0, ω 2 , ω 3 , ω 4 ) ↔ (ω 4 , ω 3 , ω 2 , 0), providing their explicit expressions which are given in [71]. Generalized results on twisted (space-and time-like) Poincaré algebras can be found in [87]. For twist deformations of κ-Poincaré and their contractions to κ-Galilei algebras we refer to [86], and for twist deformations of the Carroll algebra see [146]. Table 8: Space-and time-like real classical r-matrices for the (2+1)D kinematical and 3D Riemannian algebras with commutation relations (107) and (108) with indices i, j = 1, 2 and k = 3.

Space-like class
• Lorentzian (c finite): dS so(3, 1) Λ > 0 Poincaré iso(2, 1) Λ = 0 AdS so(2, 2) Λ < 0 • Newtonian (c = ∞): Time-like class C: Nevertheless, quantum deformations for the simple AdS and dS bialgebras have only been achieved for the κ-AdS of class C in [136], as a Poisson-Hopf algebra, showing the hard difficulties of this task. Consequently, the obtention of the quantum algebras for AdS of class A, dS of class B and AdS of class D remain as open problems. In contrast to this (3+1)D case, quantum Drinfel'd-Jimbo deformations for the semisimple (A)dS algebras, so(3, 1) and so(2, 2), are well known and their space-and time-like deformations were formerly obtained in [135] within a CK framework. Later on, the (2+1)D κ-(A)dS algebras were considered in a quantum gravity context in [160], and their twisted deformations, with underlying Drinfel'd double structures [115], were studied in [161].
By taking into account the above comments, it is worth comparing the (2+1)D case with the (3+1)D one with more detail, since they are quite different. In fact, the former is somewhat "special" as it is very well known in quantum gravity. In particular, the six generators {J ab } (a < b; a, b = 0, 1, . . . , 3) span the CK family so ω 1 ,ω 2 ,ω 3 (4) which turns out to be a Lie subalgebra of so ω (5) so with non-vanishing Lie brackets included in (71). It was proven in [31] that the Drinfel'd-Jimbo r-matrix for the family so ω 1 ,ω 2 ,ω 3 (4) simply reads which is ω-independent in contrast to (92). Moreover the r-matrix (136) gives rise to a cocommutator δ through (12) determining a real Lie bialgebra for any value of the three contraction parameters (ω 1 , ω 2 , ω 3 ). The pair of main and secondary primitive generators is (J 03 , J 12 ). Therefore the family of CK bialgebras (so ω 1 ,ω 2 ,ω 3 (4), δ(r)) contains all the 3 3 = 27 possibilities at this dimension [135].
Next we consider the family of kinematical algebras so Λ,c,λ (4) spanned by the six generators {P 0 , P 1 , P 2 , K 1 , K 2 , J 3 } in such a manner that the commutation rules are given by (107) and (108) setting the indices i, j = 1, 2 and fixing k = 3. Starting with the CK r-matrix (136), we look for space-and time-like so Λ,c,λ (4) bialgebras which would be the (2+1)D counterparts of the classes A, B and C in Table 6. Clearly the class B, with commuting primitive generators (P 2 , J 2 ), has no (2+1)D counterpart since at this dimension there does not exist a spatial generator P i commuting with J 3 . Hence we are led to the two classes A and C for so Λ,c,λ (4) such that the former requires to set λ = 1, while the latter obliges to fix c = 1. Each of them covers nine real Lie bialgebras which are displayed in Table 8.
The (2+1)D NH algebras are n + = i 4 (so(1, 1) ⊕ so(2)) and n − = i 4 (so(2) ⊕ so (2)). Obviously, there exists a second class of space-like deformations with primitive generators (P 1 , K 2 ), but this leads to equivalent results already contained within the class A [135]. The time-like class C, corresponding to the kappa-deformation, does not only show the known fact that the (2+1)D κ-Poincaré r-matrix is shared by its curved neighbours (see [161] and references therein), but also that it holds for the three Carrollian and Riemannian deformations. The strong differences between the (2+1)D and (3+1)D deformations become evident when comparing the expressions of Table 8 with those given in Table 6.
Finally, we would like to point out that the CK approach to kinematical deformations may appear to be rather restrictive since this starts with the specific Drinfel'd-Jimbo r-matrix (47) for so (5) and, from it, the CK r-matrix (92) is introduced, which together with r D (101) become the cornerstone of this work. However, we stress that this is not the case provided that one searches for deformations with a fundamental scale determined by the quantum deformation parameter. In particular, the problem of finding time-like classical r-matrices for (3+1)D (A)dS algebras was addressed in [36]. There, it was initially considered the Lorentzian algebras, with commutation relations (107) and (108) (set λ = 1), together with the most generic classical r-matrix depending on 45 deformation parameters. Then it was required to keep underformed the time translation generator P 0 and another commuting generator which was chosen J 3 , that is, δ(P 0 ) = δ(J 3 ) = 0 (recall that P 0 only commutes with generators of rotations). Under these conditions, the solution of the modified classical Yang-Baxter equation (14) gave rise to two two-parametric classical r-matrices. One of them was formed by the superposition of the κ-AdS r-matrix with the twist P 0 ∧ J 3 , which turns out to be just r D of the class C in Table 6 by identifying the two deformation parameters in [36] and interchanging the indices 3 ↔ 2 in the generators (so J 3 ↔ J 2 ) through the appropriate Lie algebra automorphism. Likewise, the second solution can be identified with r D of the class D by identifying again the two deformation parameters and applying the permutation of indices 3 → 2 → 1 → 3 with another algebra automorphism. We remark that no analysis on real Lie bialgebras and Drinfel'd doubles was carried out in [36]. Moreover, although it was claimed that the second solution (r D of class D) was determined by a dimensionful deformation parameter this is not exactly correct if one requires a dimensionless classical r-matrix. In fact, from a dimensionless deformation parameter (like z in class D) it can be introduced a dimensionful one by trivially multiplying it by a global factor.

Noncommutative Spacetimes and Spaces of Lines
The first-order noncommutative spaces for the kinematical bialgebras of Table 6 are shown in Table 7. There are several different situations among the four classes, ranging from the class D, where there is no noncommutative space (and so omitted), to the classes A and C, for which there exist both noncommutative spacetimes and spaces of lines for all the bialgebras. In this sense, the latter classes can be regarded as the prototypes for space-and time-like noncommutative spaces. However, when Drinfel'd double structures are taken into account, the classes B and C become the relevant ones, also providing twisted noncommutative spacetimes.
Although these results do not convey, in general, the full noncommutative spaces, for which all orders in the quantum coordinates must be considered, in some cases they do. Concerning the (3+1)D noncommutative spacetimes, which are those commonly studied in the literature, the first-oder noncommutative spacetimes in Table 7 turns out to be the complete ones for all the cases associated with a flat classical spacetime ST 3+1 (111), so with vanishing sectional curvature K; these are just the four spaces displayed in the middle column of Table 5. Consequently, Table 7 comprises the following full (3+1)D (linear) noncommutative spacetimes: the space-like Minkowskian and Galilean ones together with the 4D Euclidean space in the class A; another space-like Minkowskian spacetime of class B which is equivalent to that of class A under the automorphism corresponding to the permutation of indices 1 → 2 → 3 → 1; and the (time-like) κ-Minkowski and κ-Carroll spacetimes of class C (the Euclidean case is equivalent to that of class A). Additionally, complete twisted noncommutative spacetimes coming from contracted Drinfel'd double structures cover the 4D Euclidean space of class A, the space-like Minkowskian one of class B, and the time-like Minkowskian and Carroll spaces of class C (again the Euclidean space here is equivalent to that of class A). We remark that more general results on twisted space-and time-like Minkowskian noncommutative spacetimes can be found in [87].
To the best of our knowledge, results for (3+1)D noncommutative spacetimes related to curved spacetimes ST 3+1 (111) (so with Λ = 0) only comprise the (nonlinear) κ-AdS space (and its twisted version) [70], as well as the κ-Para-Euclidean and κ-Para-Poincaré [71] of class C. In such noncommutative spaces there appear higher-order terms in the quantum coordinates governed by the cosmological constant/curvature parameter Λ. This fact allows one to distinguish them from the linear (flat) κ-Minkowski and κ-Carroll spaces with Λ = 0, but the latter share the same linear structure. Hence the construction of the space-like noncommutative AdS (class A) and dS (class B) spacetimes remain as open problems, which could be faced by computing their Poisson-Lie structure by means of the Sklyanin bracket (20) and next studying their quantization (similarly to the κ-AdS spacetime [70]).
Noncommutative spaces of lines have scarcely been explored and they have only been constructed for the κ-Minkowski of class C in [100] and in lower dimensions for the 4D (A)dS noncommutative spaces of worldlines in [99]; recall that the three classical Lorentzian spaces of worldlines are of non-zero curvature equal to −1/c 2 , while both NH and Galilean spaces of lines are flat. Although the first-order noncommutative Minkowskian space of lines of class C has vanishing commutators, we stress that the brackets defining its full quantum space are not trivial at all and, in fact, it can be endowed with a symplectic structure everywhere but in the origin. By contrast, observe that the structure of the first-order noncommutative spaces of lines of class A is not trivial (see Table 7). From this viewpoint, noncommutative spaces of lines deserve a deeper study and moreover it would be necessary to construct more noncommutative spaces of lines which when read altogether with their corresponding noncommutative spacetimes could allow for a deeper insight into the structure of each precise quantum deformation.

Conclusions and Outlook
This paper can be seen as a two-fold work with two interlinked parts that we proceed to comment separately.
In the first part of the work (Sections 3 and 4), we have considered the CK formalism for quasiorthogonal Lie algebras and their associated symmetric homogeneous spaces in order to next study their Drinfel'd-Jimbo quantum deformations. The CK approach conveys a built-in scheme of Lie algebra contractions in terms of explicit graded contraction/curvature parameters ω, in such a manner that semisimple together with non-semisimple Lie algebras and their homogeneous spaces can be described in a unified setting, which ranges from the semisimple so(p, q) algebras (providing curved spaces) to the most contracted case in the CK family, the flag algebra (with associated flat spaces). In all the contraction sequence the same number of Casimir invariants (two, in our case) is preserved which, in turn, implies that these CK algebras share many structural properties as we have shown along the paper. As a novelty, we stress that we have not only considered the usual space of points (i.e. spacetimes), but also the symmetric homogeneous CK spaces of lines, 2-planes and 3-hyperplanes. In this global framework, Drinfel'd-Jimbo CK bialgebras have been obtained from the one corresponding to so(5) in a rotational basis, by always requiring the condition of getting a real Lie bialgebra, which finally led to the 63 real Lie bialgebras shown in Table 2. From these results, their dual quantum counterparts have also been deduced giving rise to their corresponding first-order noncommutative spaces of points, lines, etc., for which the coisotropy condition has been imposed, thus ensuring to always obtain a noncommutative space as a subalgebra of the dual Lie bialgebra; the final results have been summarized in Table 3. Furthermore, r-matrices coming from Drinfel'd double structures have been studied in detail as well. In particular, starting with the one corresponding to the real compact form so(5) in the rotational CK basis, three classical r-matrices for the so(p, q) algebras together with ten contracted r-matrices have explicitly been achieved and displayed in Table 4. New results correspond to the dS so(4, 1) algebra of case (II) and the AdS so(3, 2) one of case (IV), along with the contractions from the four classical r-matrices r D for the simple Lie algebras. We remark that such r-matrices, coming from Drinfel'd doubles have provided, in a natural way, first-order twisted noncommutative CK spaces of points and of 3-hyperplanes for the 14 real Lie bialgebras given in Table 4.
Concerning this first part of the paper, there are, at least, two research lines which we plan to face in the future: 1. To construct new dual homogeneous CK spaces with isotropy subalgebras corresponding to the first-order noncommutative CK spaces S (m) ⊥,ω (m = 1, . . . , 4) (100) in a similar form to that followed in [60], but moreover considering their symmetric character according to the z-involutions θ (m) z (61), which in some cases would provide a Z ⊗4 2 -grading in this dual framework. And also to study their mathematical/physical properties.

2.
To perform a similar construction to the one here developed for the Drinfel'd-Jimbo quantum deformations of quasiorthogonal CK algebras for other families of CK algebras [162]. Among them, we remark the quasiunitary CK algebras [163,164] (starting with the su(p, q) algebras) since they are naturally related to the physical quantum space of states for any quantum system [165,166].
In the second part of the work we have focused on the kinematical algebras together with their associated symmetric homogeneous spacetimes and spaces of lines displayed in Table 5. Then we have applied the previous CK approach in order to deduce their corresponding classical r-matrices, r and r D , given in Table 6, thus providing kinematical bialgebras, and from them we have constructed the first-order noncommutative spacetimes and spaces of lines shown in Table 7. A detailed physical discussion on known results and open problems concerning their full quantum algebra deformation and complete quantum spaces has already been carried out in Sections 6.5 and 6.6, respectively. Therefore, to end with, we summarize the main conclusions and open lines of research on this issue: 1. In this paper we have only considered coboundary Lie bialgebra contractions, that is, those Lie bialgebras coming from a contracted classical r-matrix. However, there also exist fundamental Lie bialgebra contractions, under the which the r-matrix diverges but the cocommutator δ is well defined [31] (so ensuring the existence of well defined coproduct ∆ z ). Hence a systematic study of all the possible fundamental but non-coboundary Lie bialgebra contractions starting with the four classical r-matrices for so (3,2) and so(4, 1) in Table 6 is still lacking. These could give rise to new quantum deformations for non-simple kinematical algebras as it was the case for the κ-Newtonian ones already obtained in [71] 2. The quantum algebra deformations for the simple algebra AdS so(3, 2) of the classes A and D, and for dS so(4, 1) of the class B are still unknown. Such structures would be moreover useful in order to obtain the corresponding contracted quantum algebras for the kinematical algebras with Λ = 0 covering the NH, Para-Euclidean and Para-Poincaré algebras. In this contraction process, both coboundary and fundamental non-coboundary Lie bialgebra contractions may be applied.
3. From Table 7 it directly follows that quantum deformations for different kinematical algebras share the same underlying first-order noncommutative spacetime structure. When dealing with the curved cases with Λ = 0, differences among them could arise when higher-orders in the quantum coordinates are taking into account (as it happened for the κ-spacetimes of class C obtained in [70,71]). Nevertheless, the linear noncommutative spacetime structure remains the same for the flat cases with Λ = 0 and this fact holds for the Minkowkian, Galilean and Carroll spaces. Consequently, the construction of the "accompanying" noncommutative spaces of lines may be of interest in order to distinguish them. And, furthermore, new physical consequences could be extracted from such new structures. In this respect, we would like to emphasize that the noncommutative space of worldlines already constructed for the κ-Poincaré algebra in [100] constitutes a prototype example in this direction.
Work on the above lines is currently in progress.