The Split-Algebras and Non-compact Hopf Maps

arXiv:0905.2792v1 [math-ph] 18 May 2009

Kazuki Hasebe Department of General Education, Kagawa National College of Technology, Takuma-cho, Mitoyo-city, Kagawa 769-1192, Japan email: [email protected] Abstract We develop a non-compact version of the Hopf maps based on the split-algebras. The split-algebras consist of three species: split-complex numbers, split-quaternions, and split-octonions. They correspond to three non-compact Hopf maps that represent topological mapping between higher dimensional hyperboloids in different dimensions with hyperboloid-bundles. We realize such non-compact Hopf maps in two ways: one is to utilize the split-imaginary unit, and the other is to utilize the ordinary imaginary unit. The connections of the hyperboloid-bundles in the non-compact Hopf maps are evaluated, and they are naturally regarded as non-compact gauge fields of monopoles.

1

Introduction

In the sequel papers in 1931 and 1935 [1, 2], Heinz Hopf introduced the notion of topological maps from spheres to spheres in different dimensions:

7

S 15

S −→

S3 −→ S8

S1

−→ S4

S2

(1st) (2nd) (3rd)

Such maps are entitled as the Hopf maps in honor of his name. There are no other fibrations between spheres with sphere-bundles according to Adams’ theorem [3]. The importance of the Hopf maps is now widely appreciated in fibre-bundle theory of mathematics, and in applications of topology to physics [4, 5]. In particular, the three Hopf maps have direct relevance to monopoles; the 1st Hopf map is the underlying mathematics of Dirac’s U (1) monopole [6], the 2nd Hopf map is that of Yang’s SU (2) monopole [7], and the 3rd Hopf map is that of the SO(8) monopole [8]. Interestingly, the Hopf maps are deeply related to the normed division algebras; i.e. complex numbers C, quaternions H, and octonions O [9]. In the language of the division algebras, the Hopf maps can be rewritten as

SH1 SO1

−→

SC1 −→ OP 1

S1

CP 1

R −→ HP 1

1

(1st) (2nd) (3rd)

It would be worthwhile to mention the relevant groups of the Hopf maps. The isometries of the sphere-base manifolds, S 2 , S 4 and S 8 , are respectively given by SO(3) ≃ SU (2)/Z2 ≃ U Sp(2)/Z2 , SO(5) ≃ U Sp(4)/Z2 , SO(9),

(1)

where U Sp(2n) denotes the compact symplectic group. Similarly, the holonomy groups of such sphere-base manifolds are respectively given by SO(2) ≃ U (1) ≃ S 1 ,

SO(4) ≃ SU (2) × SU (2) ≃ S 3 × S 3 ,

SO(8) ≃ S 7 × S 7 × G2 .

(2)

It should be noticed that the holonomy groups and the sphere-bundles are closely related: the sphere-bundles S 1 , S 3 , and S 7 appear on RHS in the expression of the holonomy groups (2). The 1st and 2nd Hopf maps are special in the sense that their sphere-bundles have counterparts of the (special) unitary group manifolds, i.e. U (1) ≃ S 1 , SU (2) ≃ S 3 . (There does not exist a corresponding group manifold with S 7 .) The automorphism groups of the division algebras are respectively given by Aut(C) = Z2 ,

Aut(H) = SO(3),

Aut(O) = G2 .

(3)

Few years after the discovery of quaternions by William R. Hamilton [10], James Cockle introduced the notion of the split-algebras [11, 12]. Similar to the original division algebras, there exist three species of split-algebras; splitcomplex numbers C′ , split-quaternions H′ , and split-octonions O′ . The splitalgebras have similar properties to the original division algebras except for their split signatures. Based on the split-algebras, we introduce the notion of noncompact Hopf maps as

HH1 ′ HO1 ′

−→

HC1′ −→ O′ P 1

H1

R −→ ′ 1 HP

C′ P 1

(1st) (2nd) (3rd)

where H 1 represents “one-dimensional” hyperboloid (hyperbola) in the space of the corresponding split-algebras. Defining higher dimensional hyperboloids (ultra-hyperboloids) H p,q 1 : p X i=1

xi xi −

q+1 X j=1

y j y j = −1,

the non-compact Hopf maps are rewritten as 1 In

particular, S n = H 0,n , dS n = H 1,n−1 , and AdS n = H n−1,1 .

2

(4)

4,3

H 8,7

H −→

H 2,1 −→ H 4,4

H 1,0

−→ H 2,2

H 1,1

(1st) (2nd) (3rd)

The relevant groups of the non-compact Hopf maps are summarized as follows. The isometries of the hyperboloid-base manifolds, H 1,1 , H 2,2 and H 4,4 , are respectively SO(2, 1) ≃ SU (1, 1)/Z2 ≃ Sp(2, R)/Z2 , SO(3, 2) ≃ Sp(4, R)/Z2 ,

SO(5, 4),

(5)

where Sp(2n, R) represents the real symplectic group. The holonomy groups of the hyperboloid-base manifolds are respectively SO(1, 1) ≃ U(1) ≃ H 1,0 ,

SO(2, 2) ≃ SU (1, 1) × SU (1, 1) ≃ H 2,1 × H 2,1 ,

SO(4, 4) ≃ H 4,3 × H 4,3 × G2(2) .

(6)

(U(1) will be defined in Subsec.3.1.) On RHS in the expression of the noncompact holonomy groups (6), one may notice that the hyperboloid-bundles H 1,0 , H 2,1 , and H 4,3 appear. The 1st and 2nd non-compact Hopf maps are special in the sense that their hyperboloid-bundles have the counterparts of the non-compact (special) unitary group manifolds, i.e. U(1) ≃ H 1,0 , SU (1, 1) ≃ H 2,1 . (There does not exist a corresponding group manifold with H 4,3 .) The automorphism groups of the split-algebras are respectively given by Aut(C′ ) = Z2 ,

Aut(H′ ) = SO(2, 1),

Aut(O′ ) = G2(2) .

(7)

Apparently, there are close analogies between the relevant groups of the original and the non-compact Hopf maps. In this paper, we explore explicit realization of the non-compact Hopf maps. We introduce two ways of the realization, one of which (referred as “Realization I” in the context) is performed by using the split-imaginary unit j: j 2 = 1, and the other (referred as “Realization II”) by the ordinary imaginary unit i: i2 = −1. The generators of non-compact groups are represented by finitedimensional hermitian matrices, with use of the split-imaginary unit. Taking advantage of this property, in Realization I, we demonstrate the realization of the non-compact Hopf maps quite analogously to that of the original Hopf maps 2 . In Realization II, we adopt the ordinary imaginary unit, and realize the non-compact Hopf maps. There, finite-dimensional non-hermitian (but skew hermitian) matrices are utilized to express the non-compact generators. The connections of fibres and their curvatures are also evaluated, and they are naturally interpreted as SO(1, 1), SU (1, 1) and SO(4, 4) monopole gauge fields. 2 The

realizations of the original Hopf maps can be referred to Refs.[1, 2, 13, 14]

3

The organizations are as follows. In Sec.2, as a preliminary, we introduce 0th non-compact Hopf map. In Sec.3, Sec.4 and Sec.5, we develop the noncompact version of the 1st, 2nd and 3rd Hopf maps, respectively. The SO(2, 1), SO(3, 2) Dirac, and SO(5, 4) Majorana spinors play crucial roles in constructing the non-compact Hopf maps. Sec.6 is devoted to summary and discussions. In Appendix A, a non-compact version of the supersymmetric (1st) Hopf map is provided.

2

0th Non-compact Hopf Map

As a warm-up, we first introduce the non-compact version of the 0th Hopf map. The original 0th Hopf map is given by Z

2 S1, S 1 −→

(8)

where 1D sphere S 1 is defined as usual: x1 2 + x2 2 = 1.

(9)

The map (8) is realized by identifying “opposite points” on the 1D sphere: (x1 , x2 ) ∼ −(x1 , x2 ).

(10)

We introduce the non-compact version of the 0th Hopf map as Z

2 H 1,0 , H 1,0 −→

(11)

where H 1,0 is one-dimensional hyperboloid, i.e. hyperbola defined by x1 2 − x2 2 = −1,

(12)

and the hyperbola consists of two branches; left and right in (x2 , x1 )-space. The non-compact 0th Hopf map is realized by identifying the “opposite points” (with respect to the origin) between the right and left branches of the hyperbola as Eq.(10). Then, H 1,0 on RHS of the non-compact map (11) represents onebranch of the hyperboloid.

3

1st Non-compact Hopf Map

The 1st non-compact Hopf map is introduced as H 1,0

H 2,1 −→ H 1,1 , or

U (1)

AdS 3 −→ AdS 2 .

(13) (14)

Here, AdS n denotes the n-dimensional anti de Sitter space, and the base manifold AdS 2 ≃ H 1,1 is a two-dimensional one-leaf hyperboloid. 4

3.1

Split-Complex Numbers

The split-imaginary unit j is introduced so as to satisfy j 2 = 1,

j ∗ = −j,

(15)

where ∗ denotes complex conjugation. With two real numbers x and y, the spilt-complex number is defined as z = x + jy,

(16)

z ∗ = x − jy.

(17)

z ∗ z = zz ∗ = x2 − y 2 .

(18)

z1∗ z1 − z2∗ z2 = −1,

(19)

and its complex conjugation is

Then, The “normalized” split-complex number, satisfying z ∗ z = zz ∗ = −1, represents H 1,0 . The hyperbola in the 2D split-complex space HC1′ is defined as

and HC1′ represents H 2,1 . Utilizing the split-imaginary unit j instead of i, we define the non-compact U(1) group whose group element is given by ejϑ = cosh ϑ + j sinh ϑ, which has one-to-one correspondence to the SO(1, 1) group element   cosh ϑ sinh ϑ g= . sinh ϑ cosh ϑ

(20)

(21)

Therefore, U(1) is isomorphic to the SO(1, 1) group U(1) ≃ SO(1, 1).

(22)

With real x and y, the SO(1, 1) group element is generally represented  numbers  x y as g = , satisfying the constraint detg = x2 − y 2 = 1. The constraint is y x the definition of H 1,0 . Then, U(1) ≃ H 1,0 .

3.2

(23)

Realization I

We introduce the spilt-Pauli matrices as     0 1 0 −j 1 2 σ = , σ = , 1 0 j 0 5

  1 0 σ = . 0 −1 3

(24)

The split-imaginary unit j appears in σ 2 , and the split Pauli matrices satisfy the relation {σ i , σ j } = 2ηij , (25) where ηij = diag(+1, −1, +1). The split-Pauli matrices can also be regarded as the gamma matrices in (2+1)D. The split-Pauli matrices are hermitian and satisfy the SU (1, 1) algebra [σ i , σ j ] = 2jǫij k σ k ,

(26)

with ǫ123 = 1. It is noted that, in general, generators of non-compact groups cannot be expressed by finite-dimensional hermitian matrices, but here, because of the special property of the split-imaginary unit j (15), the SU (1, 1) generators are represented by hermitian and finite-dimensional matrices. Since σ 2 satisfies − (σ i )∗ = σ 2 σ i (σ 2 )−1 ,

(27)

σ 2 is the charge conjugation matrix of SU (1, 1). σ 2 has the following properties (σ 2 )−1 = −σ 2 = (σ 2 )t = (σ 2 )∗ ,

(28)

and satisfies the consistency condition of charge conjugation matrix (σ 2 )∗ σ 2 = σ 2 (σ 2 )∗ = 12 .

(29)

(12 denotes 2 × 2 unit matrix.) With use of the split Pauli matrices (24), the 1st non-compact Hopf map is expressed as φ → xi = φ† σ i φ. (30) Here, φ denotes a SU (1, 1) Dirac spinor (the non-compact 1st Hopf spinor)   u φ= , (31) v which satisfies the normalization condition φ† φ = u∗ u + v ∗ v = uR 2 − uI 2 + vR 2 − vI 2 = 1,

(32)

where u = uR + juI , v = vR + jvI . (The lower indices R and I represent real and imaginary parts, respectively.) Thus, φ denotes coordinates on H 2,1 . xi given by (30) are explicitly x1 = u∗ v + v ∗ u,

x2 = −ju∗ v + jv ∗ u,

x3 = u∗ u − v ∗ v,

(33)

and satisfy X

i,j=1,2,3

ηij xi xj = (x1 )2 − (x2 )2 + (x3 )2 = (φ† φ)2 = 1.

6

(34)

Then, xi denote coordinates on H 1,1 . Inverting the map (30), φ is represented as   1 1 + x3 jϑ φ= p (35) 1 2 e . 2(1 + x3 ) x − jx

The U(1) phase ejϑ corresponding to the H 1,0 -fibre is cancelled in the map (30). The connections of H 1,0 -fibre are induced as A = dxi Ai = −jφ† dφ,

(36)

1 ǫij3 xj . 2(1 + x3 )

(37)

with Ai = The curvatures are also derived as

1 i dx ∧ dxj Fij , 2

F = dA = or

3.3

(38)

1 Fij = ∂i Aj − ∂j Ai = − ǫijk xk . 2

(39)

Realization II

In the case of the ordinary imaginary unit, the non-compact fibre U(1) ≃ H 1,0 is replaced by the ordinary compact fibre U (1) ≃ H 0,1 , and then the 1st noncompact Hopf map takes the different form from (13): H 0,1

H 2,1 −→ H 2,0 , or

(40)

U(1)

2

AdS 3 −→ Euclid.AdS 2 .

(41)

2

Euclid.AdS is a two-leaf hyperboloid, while AdS in (14) is a connected (oneleaf) hyperboloid. The SU (1, 1)(≃ SO(2, 1)) generators τ i are realized as τ 1 = iσ 1 ,

τ 2 = iσ 2 ,

where σ i denote the ordinary Pauli matrices     0 1 0 −i σ1 = , σ2 = , 1 0 i 0

τ 3 = σ3 ,

σ3 =



(42)

 1 0 . 0 −1

(43)

τ i are non-hermitian but skew hermitian: (τ i )† = −τi .

(44)

In detail, τ 1 and τ 2 are anti-hermitian, while τ 3 is hermitian. τ i satisfy the anticommutation relation {τ i , τ j } = −2η ij , (45) 7

with ηij = diag(+, +, −), and their commutation relations are given by [τ i , τ j ] = 2iǫijk τk .

(46)

The charge conjugation matrix σ 1 satisfying − (τ i )∗ = σ 1 τ i (σ 1 )−1 ,

(47)

(σ 1 )−1 = (σ 1 )t = (σ 1 )∗ = (σ 1 )† .

(48)

has the properties By a unitary transformation, σ 1 can be diagonalized to give σ 3 . Multiplying σ 3 to τ i , hermitian matrices are constructed as σ 3 τ i = (−σ 2 , σ 1 , 1).

(49)

A normalized SO(2, 1) Dirac spinor φ = (u, v)t (1st non-compact Hopf spinor) is introduced as 2 − vI2 = 1, φ† σ 3 φ = u∗ u − v ∗ v = u2R + u2I − vR

(50)

and φ represents coordinates on H 2,1 . We now realize the 1st non-compact Hopf map (40) as φ → xi = φ† σ 3 τ i φ, (51) or i

x1 = iu∗ v − iv ∗ u,

x2 = u∗ v + v ∗ u,

x3 = u∗ u + v ∗ v.

(52)

2,0

x are coordinates on H as found X ηij xi xj = x2 + y 2 − z 2 = −(φ† σ 3 φ)2 = −1.

(53)

i,j=1,2,3

Inverting (51), the 1st non-compact Hopf spinor is expressed as   1 1 + x3 iϑ φ= p 2 1 e , 2(1 + x3 ) x − ix

(54)

where U (1) phase factor corresponds to the H 0,1 -fibre. The associated fibreconnections are given by A = −iφ† σ 3 dφ = dxi Ai , where Ai =

1 ǫij3 xj . 2(1 + x3 )

(55)

(56)

Similarly, the curvatures are derived as 1 Fij = ∂i Aj − ∂j Ai = − ǫijk xk . 2 8

(57)

4

2nd Non-compact Hopf Map

In this section, based on the split-quaternions, we explore the realization of the 2nd non-compact Hopf map: H 2,1

H 4,3 −→ H 2,2 .

4.1

(58)

Split-Quaternions

The split-quaternions, 1, q1 , q2 , q3 , are introduced so as to satisfy the following algebras: q1 2 = −q2 2 = q3 2 = q1 q2 q3 = 1, qi qj = −qj qi ,

(59)

with i, j = 1, 2, 3. As in the case of quaternions, the commutativity does not hold in the algebra of the split-quaternions. Other relations, such as q1 q2 = −q3 , q2 q3 = q1 , q3 q1 = −q2 ,

(60)

are derived by (59). The conjugation of the split-quaternions is given by 1∗ = 1,

qi ∗ = −qi .

(61)

A split-quaternion is generally represented as h = r0 1 + r1 q1 + r2 q2 + r3 q3 ,

(62)

h∗ = r0 1 − r1 q1 − r2 q2 − r3 q3 .

(63)

h∗ h = hh∗ = r0 2 − r1 2 + r2 2 − r3 2 .

(64)

and its conjugation is

Then, One may confirm the split-quaternions indeed possess the split-signatures. The normalized split-quaternion is defined as h∗ h = hh∗ = −1, which represents H 2,1 . Similarly, the hyperbola in the 2D split-quaternion space HH1 ′ is defined as h∗1 h1 − h∗2 h2 = −1, (65)

which is equal to H 4,3 . It is also important to notice that the split-quaternions are realized by the use of the split-Pauli matrices : qi = jσ i ,

(66)

where j in front of σ i are needed to satisfy Eq.(61) with the definition (qi )∗ = (jσ i )† . 9

4.2

Realization I

The SO(3, 2) anticommutation relations are given by {γa , γb } = −2ηab ,

(67)

with ηab = diag(+1, −1, +1, −1, −1). Such gamma matrices γ a = (γ i , γ 4 , γ 5 ) (i = 1, 2, 3) can be represented as       0 jσ i 0 12 12 0 i 4 5 γ = , γ = . (68) , γ = −jσ i 0 12 0 0 −12 It is noted that the “split-quaternions” appear in the off-diagonal elements of γ a , and γ a are hermitian: γ 1 , γ 3 are pure imaginary antisymmetric matrices, and γ 2 , γ 4 and γ 5 are real symmetric matrices. Then, the SO(3, 2) generators constructed by the formula σ ab = −j 14 [γ a , γ b ] are also represented by hermitian and finite dimensional matrices. Just as in the SO(2, 1) case, with use of j, SO(3, 2) generators are represented by hermitian and finite-dimensional matrices. The charge conjugation matrix of SO(3, 2) is constructed by multiplying the purely imaginary gamma matrices:  2    σ 0 ǫ 0 b = jγ 1 γ 3 = − = j , (69) 0 σ2 0 ǫ where ǫ = ǫαβ =



0 −1

 1 . 0

(70)

b indeed satisfies ∗ bγa b−1 = γa∗ , bσab b−1 = −σab ,

(71)

b−1 = −b = bt = b∗ .

(72)

and has the properties The consistency condition of charge conjugation matrix is satisfied as b∗ b = bb∗ (= −b2 ) = 1.

(73)

The charge conjugation of SO(3, 2) four-component spinor ψ = (ψ1 , ψ2 , ψ3 , ψ4 )t is constructed as ψc = −b−1 ψ ∗ = bψ ∗ , (74) and its magnitude is given by ψc† ψc = −ψ † ψ.

(75)

We now realize the 2nd non-compact Hopf map as ψ → xa = ψ † γ a ψ.

10

(76)

Here, ψ is a normalized SO(3, 2) spinor satisfying the condition ψ† ψ =

4 X

ψµ∗ ψµ =

µ=1

4 X

µ 2 (ψR ) −

µ=1

4 X

(ψIµ )2 = 1,

(77)

µ=1

µ where ψ µ = ψR + jψIµ . The condition (77) is the definition of H 4,3 , and then ψ denotes coordinates on H 4,3 . Meanwhile, xa given by (76) are explicitly

x1 = −j(ψ4∗ ψ1 + ψ3∗ ψ2 − ψ2∗ ψ3 − ψ1∗ ψ4 ),

x2 = −ψ4∗ ψ1 + ψ3∗ ψ2 + ψ2∗ ψ3 − ψ1∗ ψ4 ,

x3 = −j(ψ3∗ ψ1 − ψ4∗ ψ2 − ψ1∗ ψ3 + ψ2∗ ψ4 ),

x4 = ψ3∗ ψ1 + ψ4∗ ψ2 + ψ2∗ ψ3 + ψ2∗ ψ4 , x5 = ψ1∗ ψ1 + ψ2∗ ψ2 − ψ3∗ ψ3 − ψ4∗ ψ4 ,

(78)

and satisfy the relation X ηab xa xb = (x1 )2 − (x2 )2 + (x3 )2 − (x4 )2 − (x5 )2 = −(ψ † ψ)2 = −1. a,b=1,2,3,4,5

(79) Thus, xa represent coordinates on H 2,2 . Inverting the 2nd non-compact Hopf map (76), ψ can be represented as   1 (1 + x5 )φ , (80) ψ=p 4 i 2(1 + x5 ) (x + jx σi )φ

where φ is a normalized 2-component spinor (32) representing the AdS 3 -fibre. A straightforward calculation shows that the degrees of φ is cancelled in the mapping (76). Utilizing the explicit form of (80), the AdS 3 -fibre connections are evaluated as A = −jψ † dψ = dxa φ† Aa φ, (81) where Am = −

1 ηmni xn σ i , 2(1 + x5 )

A5 = 0.

(82)

Here, m, n = 1, 2, 3, 4, and the split-’t Hooft symbol ηmni is made of the splitmetric ηmn = diag(+1, −1, +1, −1) as ηmni = ǫmni − ηmi ηn4 + ηm4 ηni .

(83)

Since the gauge group is SU (1, 1), Aa may be physically interpreted as the SU (1, 1) monopole gauge fields. The curvatures F = dA + jA ∧ A = 11

1 a dx ∧ dxb Fab 2

(84)

or Fab = ∂a Ab − ∂b Aa + j[Aa , Ab ]

(85)

are computed as 1 Fmn = xm An − xn Am + ηmni σ i , 2 5 Fm5 = −F5m = −(1 + x )Am .

4.3

(86)

Realization II

With use of the ordinary imaginary unit, the (3 + 2)D gamma matrices γ a are realized as γ i = τ i ⊗ σ2 , γ 4 = 1 ⊗ σ1 , γ 5 = γ 1 γ 2 γ 3 γ 4 = 1 ⊗ σ3 ,

(87)

     −iτ i 0 1 1 0 4 5 , γ = , γ = , 0 1 0 0 −1

(88)

or i

γ =



0 iτ i

where i = 1, 2, 3, and τ i are the SU (1, 1) generators (42). γ a are skew hermitian: (γ a )† = −γa .

(89)

In detail, γ 1 and γ 2 are anti-hermitian matrices while γ 3 , γ 4 and γ 5 are hermitian matrices. Their anticommutation relations are given by {γ a , γ b } = −2η ab ,

(90)

with η ab = ηab = diag(+, +, −, −, −), and their commutators yield the SO(3, 2) generators 3 1 (91) σ ab = −i [γ a , γ b ], 4 or     1 ijk τk 0 1 −τ i 0 i4 ij σ = ǫ , σ = , 0 τi 0 τk 2 2     1 0 τi i 0 1 σ i5 = , σ 45 = . (92) 2 τi 0 2 −1 0 σ ab are non-hermitian, and their hermitian conjugations are (σ ab )† = σab . The SO(3, 2) charge conjugation matrix r is given by  1  σ 0 2 3 r = −γ γ = , 0 σ1 3

(93)

(94)

The SO(3, 2) algebra of σab is given by [σab , σcd ] = −i(ηac σbd −ηad σbc +ηbd σac −ηbc σad ).

12

where σ 1 is the SO(2, 1) charge conjugation matrix. Indeed, r satisfies the relations r† γ a r = (γ a )∗ , r† σ ab r = −(σ ab )∗ , (95)

and has the properties:

r† = rt = r−1 = r.

(96)

Diangonalizing r, we obtain k=



σ3 0

 0 , σ3

(97)

and k satisfies k † = k t = k −1 = k.

(98)

a

Multiplying k to γ , hermitian matrices are constructed as k a = kγ a ,

(99)

or     0 iσ 2 0 −iσ 1 0 2 3 , k = , k = i12 −iσ 2 0 iσ 1 0    3  0 σ3 σ 0 5 k4 = , k = . σ3 0 0 −σ 3

k1 =



 −i12 , 0 (100)

kσ ab are hermitian as well. A normalized SO(3, 2) Dirac spinor ψ (the 2nd non-compact Hopf spinor) is introduced so as to satisfy ψ † kψ = ψ1∗ ψ1 − ψ2∗ ψ2 + ψ3∗ ψ3 − ψ4∗ ψ4 = 1,

(101)

and, thus, ψ denotes coordinates on H 4,3 . With such a normalized ψ, the 2nd non-compact Hopf map is realized as ψ → xa = ψ † k a ψ.

(102)

Since k a are hermitian, xa given by (102) x1 = ψ1∗ ψ4 − ψ2∗ ψ3 + ψ4∗ ψ1 − ψ3∗ ψ2 ,

x2 = −i(ψ1∗ ψ4 + ψ2∗ ψ3 − ψ4∗ ψ1 − ψ3∗ ψ2 ),

x3 = −i(ψ1∗ ψ3 + ψ2∗ ψ4 − ψ3∗ ψ1 − ψ4∗ ψ2 ),

x4 = ψ1∗ ψ3 − ψ2∗ ψ4 + ψ3∗ ψ1 − ψ4∗ ψ2 ,

x5 = ψ1∗ ψ1 − ψ2∗ ψ2 − ψ3∗ ψ3 + ψ4∗ ψ4 ,

(103)

are real, and they satisfy the relation X ηab xa xb = (x1 )2 + (x2 )2 − (x3 )2 − (x4 )2 − (x5 )2 a,b=1,2,3,4,5

= −(ψ † kψ)2 = −1. 13

(104)

Thus, xa are regarded as coordinates on H 2,2 . Inverting the 2nd non-compact Hopf map (102), the non-compact 2nd Hopf spinor is expressed as   1 (1 + x5 )φ ψ= p , (105) 4 i 2(1 + x5 ) (x − ix τi )φ

where φ = (u, v)t is the 1st non-compact Hopf spinor satisfying the condition (50), and represents the H 2,1 -fibre. The associated fibre-connections are introduced as A = −iψ † kdψ = dxa φ† Aa φ (106) where Am = −ηmni

xn τ i, 2(1 + x5 )

A5 = 0,

(107)

with the split-’t Hooft symbol given by ηmni = ǫmni4 + ηmi ηn4 − ηni ηm4 .

(108)

The curvatures are evaluated by the formula Fab = ∂a Ab − ∂b Aa + i[Aa , Ab ],

(109)

or 1 Fmn = xm An − xn Am + ηmni τ i , 2 Fm5 = −F5m = (1 + x5 )Am .

5

(110)

3rd Non-compact Hopf Map

Here, we explore realization of the 3rd non-compact Hopf map H 4,3

H 8,7 −→ H 4,4 .

(111)

In the constructions of the 1st and 2nd non-compact Hopf maps, SO(2, 1) and SO(3, 2) Dirac spinors played a crucial role, while in the 3rd non-compact Hopf map, SO(5, 4) Majorana spinor does the role

5.1

Split-Octonions

The split-octonions, 1, e1 , e2 , · · · , e7 , are introduced so as to satisfy the following relations: {eI , eJ } = −2ηIJ , [eI , eJ ] = 2fIJK eK , 14

(112)

1 e1 e2 e3 e4 e5 e6 e7

1 1 e1 e2 e3 e4 e5 e6 e7

e1 e1 −1 −e3 e2 e5 −e4 e7 −e6

e2 e2 e3 −1 −e1 e6 −e7 −e4 e5

e3 e3 −e2 e1 −1 e7 e6 −e5 −e4

e4 e4 -e5 −e6 −e7 1 −e1 −e2 −e3

e5 e5 e4 e7 −e6 e1 1 −e3 e2

e6 e6 −e7 e4 e5 e2 e3 1 −e1

e7 e7 e6 −e5 e4 e3 −e2 e1 1

Table 1: The structure constants of split-octonions can be read from the table. For instance, f132 = f145 = f167 = f246 = f527 = f347 = f356 = −1. Other structure constants can be derived from cyclic permutations of the indices. where ηIJ = diag(+1, +1, +1, −1, −1, −1, −1), and fIJK (I, J, K = 1, 2, · · · , 7) denote an antisymmetric tensor known as structure constants of the split-octonions [See Table 1]. As in the case of octonions, the split-octonions do not respect associativity as well as commutativity. The complex conjugation of split-octonion is defined as 1∗ = 1, (eI )∗ = −eI . (113) Generally, split-octonion is expressed as o = r0 1 +

7 X

rI eI ,

(114)

I=1

and its conjugation is o∗ = r0 1 − Then, o∗ o = oo∗ = r0 2 +

7 X

I,J=1

ηIJ rI rJ =

X

rI eI .

X

(rI )2 −

I=0,1,2,3

(115)

X

(rI )2 ,

(116)

I=4,5,6,7

and the normalized split-octonion, satisfying o∗ o = −1, represents H 4,3 . Similarly, the hyperbola HO1 ′ in “2D” split-octonion space, o∗1 o1 − o∗2 o2 = −1,

(117)

represents H 8,7 .

5.2

Realization I

The SO(5, 4) gamma matrices, ΓA (A = 1, 2, · · · , 9), which satisfy {ΓA , ΓB } = 2η AB , 15

(118)

(η AB = ηAB = diag(+, −, −, +, +, −, −, +, +)) are introduced as       0 jγ I 18 0 0 18 I 9 8 Γ = , , Γ = , Γ = 18 0 −jγ I 0 0 −18

(119)

with I = 1, 2, · · · , 7. ΓA are hermitian, and γ I (off-diagonal block of ΓI ) are SO(4, 3) gamma matrices, satisfying {γ I , γ J } = 2η IJ , with ηIJ = diag(−1, +1, +1, −1, −1, +1, +1). γ I are explicitly     0 0 0 −σ 1 0 0 0 σ2  0 0 σ1 0  0 −σ 2 0  ,  , γ2 =  0 γ1 =  2  0  0 −σ σ1 0 0  0 0 −σ 1 0 0 0 σ2 0 0 0     3 0 0 0 12 0 0 0 −σ   0 0 σ3 0  0 12 0   , γ4 = j  0 , γ3 =   0  0 σ3 0 0  −12 0 0  −σ 3 0 0 0 −12 0 0 0     0 0 12 0 0 0 12 0    0 0 0 −12   , γ 6 =  0 0 0 12  , γ5 = j  12 0 0 0  −12 0 0 0  0 12 0 0 0 12 0 0   12 0 0 0  0 1 0 0  2 , γ7 =  (120)  0 0 −12 0  0 0 0 −12 where σ i (i = 1, 2, 3) are the split-Pauli matrices (24). γ 1 , γ 4 , γ 5 are purely imaginary matrices, and their product yields a charge conjugation matrix of SO(4, 3),   0 −b 1 4 5 d = −jγ γ γ = j , (121) b 0 where b is the SO(3, 2) charge conjugation matrix (69). Indeed, d satisfies the relations ∗ − γI ∗ = d · γI · d−1 , − σIJ = d · σIJ · d−1 , (122) with SO(5, 4) generators σIJ = −j 41 [γI , γJ ]. d has the following properties d−1 = d = dt = d∗ ,

(123)

and satisfies the consistency condition d∗ d = dd∗ (= d2 ) = 18 .

(124)

Meanwhile, the purely imaginary matrices of SO(5, 4) are Γ2 , Γ3 , Γ6 , Γ7 , and their product yields the charge conjugation matrix of SO(5, 4)   d 0 2 3 6 7 B=Γ Γ Γ Γ =− , (125) 0 d 16

which satisfies Γ∗A = BΓA B −1 ,

− Σ∗AB = BΣAB B −1 .

(126)

The properties of B are given by B −1 = B = B t = B ∗ ,

(127)

and the consistency condition is satisfied also: B ∗ B = BB ∗ (= B 2 ) = 116 .

(128)

With (16-component) SO(5, 4) spinor Ψ, the Majorana condition is imposed as Ψ = −B −1 Ψ∗ ,

(129)

and the SO(5, 4) Majorana spinor carries 16 real degrees of freedom. From (129), one may notice that the SO(5, 4) Majorana spinor is expressed as   U jUc   Ψ= (130)  V . jVc Here, (U, jUc )t and (V, jVc )t are SO(4, 3) 8-component Majorana spinor. Each of U and V denotes SO(3, 2) 4-component Dirac spinor, and Uc and Vc are their charge conjugations respectively determined by Eq.(74). Then, in total, Ψ carries 16 real degrees of freedom that come from U and V . With the above preparation, we realize the 3rd non-compact Hopf map as Ψ → xA = Ψ† ΓA Ψ.

(131)

Here, Ψ is a normalized SO(5, 4) Majorana spinor (3rd non-compact Hopf spinor) satisfying Ψ† Ψ = U † U − Uc† Uc + V † V − Vc† Vc = 1.

(132)

Since Uc† Uc = −U † U and Vc† Vc = −V † V , (132) can be rewritten as Ψ† Ψ = 2(U † U + V † V ) =2

4 X

(URµ URµ + VRµ VRµ − UIµ UIµ − VIµ VIµ )

µ=1

= 1,

(133)

17

and thus, Ψ denotes coordinates on H 8,7 . xA given by (131) are explicitly x1 = −2(U1 V3 + U2 V4 − U3 V1 − U4 V2 ) + (c.c.),

x2 = 2j(−U1 V3 + U2 V4 + U3 V1 − U4 V2 ) + (c.c.),

x3 = 2j(U1 V4 + U2 V3 − U2 V2 − U4 V1 ) + (c.c.),

x4 = −2(U1 V4 − U2 V3 + U3 V2 − U4 V1 ) + (c.c.),

x5 = −2(U1 V2 − U2 V1 − U3 V4 + U4 V3 ) + (c.c.),

x6 = 2j(−U1 V2 + U2 V1 − U3 V4 + U4 V3 ) + (c.c.),

x7 = 2j(U1∗ V1 + U2∗ V2 + U3∗ V3 + U4∗ V4 ) + (c.c.), x8 = 2(U1∗ V1 + U2∗ V2 + U3∗ V3 + U4∗ V4 ) + (c.c.),

x9 = 2(U1∗ U1 + U2∗ U2 + U3∗ U3 + U4∗ U4 ) − (U → V ), and satisfy X

(134)

ηAB xA xB

A,B=1,2,··· ,9

= (x1 )2 − (x2 )2 − (x3 )2 + (x4 )2 + (x5 )2 − (x6 )2 − (x7 )2 + (x8 )2 + (x9 )2

= (Ψ† Ψ)2 = 1.

(135)

Thus, xA denote coordinates on H 4,4 . Inverting (131), the non-compact 3rd Hopf spinor is represented as   1 (1 + x9 )Φ Ψ= p , 8 I 2(1 + x9 ) (x + jxI γ )Φ

(136)

where γ I are SO(4, 3) gamma matrices (120). From the Majorana condition of Ψ (129), the 8-component spinor Φ in (136) should satisfy Φ = d · Φ∗ ,

(137)

with d given by (121). Thus, Φ denotes a SO(4, 3) Majorana spinor that possesses 8 real degrees of freedom, and Φ can be represented as   1 ψ , (138) Φ= √ 2 jψc where ψ is a SO(3, 2) Dirac spinor, and ψc is its charge conjugation (74). Further, from (133), Φ or ψ should satisfy the condition 1 † (ψ ψ − ψc† ψc ) 2 4 4 X X = ψ† ψ = ψR ψR − ψI ψI

Φ† Φ =

µ=1

= 1.

µ=1

(139) 18

Thus, ψ denotes the 2nd non-compact Hopf spinor, and, at the same time, the fibre of the present 3rd Hopf map. We have confirmed the hierarchical structure of the Hopf maps, and the non-compact 3rd Hopf spinor is represented as     ψ 9   (1 + x ) jψc 1   .  Ψ= p (140)   9 ψ 2 (1 + x ) (x8 + jx γ I ) I jψc One might anticipate that the connections of H 4,3 -fibre would be induced by the formula A = −jΨ† dΨ. (141) Indeed, expressing A as A = dxA Φ† AA Φ,

(142)

we obtain AM = −

1 σMN xN , 1 + x9

A9 = 0.

(143)

Here, M, N = 1, 2, · · · , 8, and σMN are SO(4, 4) “W eyl+” generators given by 1 σIJ = −j [γI , γJ ], 4 1 σI8 = −σ8I = − γI . 2

(144)

However, due to the special properties of the Majorana spinor, the connections (142) vanish. Indeed, Φ† σAB Φ = Φt (d · σAB )Φ = 0, since d · σAB are antisymmetric matrices. To derive the connections of H 4,3 -fibre , we may use   (1 + x9 )18 ˜ = p 1 Ψ , (145) 8 I 2(1 + x9 ) (x 18 + jxI γ )

and the connections (143) are derived by the formula ˜ † dΨ. ˜ A = dxA AA = −j Ψ

(146)

Since the gauge group is SO(4, 4), AA may be interpreted as the SO(4, 4) monopole gauge fields. The corresponding SO(4, 4) curvatures FAB = ∂A AB − ∂B AA − j[AA , AB ]

(147)

are evaluated as FMN = −xM AN + xN AM + σMN , FM9 = −F9M = −(1 + x9 )AM . 19

(148)

5.3

Realization II

As a preliminary, we first construct Majorana representation of the SO(5, 4) group based on the split-octonion structure constants. With e0 = 1, the splitoctonion algebras (112) are expressed as eI eJ = −ηIJ e0 + fIJK eK ,

(149)

eA eB = fABC eC ,

(150)

or where A, B, C = 0, 1, · · · , 7. (The split-octonion structure constants, fABC , can be explicitly obtained from Table 1.) With use of fABC , SO(4, 3) gamma matrices λI (I = 1, 2, · · · , 7) are derived as (λI )AB = −fIAB ,

(151)

or λ1

λ3

λ5

λ7

  −σ 2 0 0 0 0 −σ 3 2  0   −σ 0 0  σ3 0 = i , λ2 =  2  0   0 σ 0 0 0 0 0 0 σ2 0 0    1 0 0 0 −σ 0 0  0  σ1 0 0 0 0  4    , λ = = 0 0 0 σ1  −12 0 0 0 −σ 1 0 0 −12    0 0 0 0 −σ 2 0 2   0 0 0 0 0 σ , λ6 =  = i  0  σ2 12 0 0 0  −12 0 0 −σ 2 0 0   0 0 0 −σ 2 2  0 0 −σ 0  . = i  0 σ2 0 0  σ2 0 0 0 

 0 0  , σ3  0  −12 0 0 −12  , 0 0  0 0  0 −12 12 0  , 0 0  0 0 0 0 0 −σ 3

(152)

Indeed, λI satisfy IJ

{λI , λJ } = −2η IJ

(153) I

with ηIJ = η = diag(+1, +1, +1, −1, −1, −1, −1). λ are all real, and skew anti-hermitian matrices: (λI )t = −λI . (154) In detail, λ1,2,3 are real antisymmetric matrices, and λ4,5,6,7 are real symmetric matrices. With λ0 ≡ 18 , λ0 and λI (I = 1, 2, · · · , 7) represent the SO(4, 4) “Weyl +” gamma matrices. With use of λI , the SO(5, 4) gamma matrices ΓA are constructed as ΓI = iλ8−I ⊗ σ 2 , Γ8 = 18 ⊗ σ 1 , Γ9 = 18 ⊗ σ 3 , 20

(155)

or ΓI =



0 −λ8−I

λ8−I 0



, Γ8 =



0 18

  18 18 , Γ9 = 0 0

 0 . −18

(156)

The “octonion structure constants” appear in the off-diagonal elements of ΓA . ΓA satisfy [ΓA , ΓB ] = 2ηAB , (157) with ηAB = diag(−1, −1, −1, −1, +1, +1, +1, +1, +1). ΓA are real skew hermitian matrices: (ΓA )t = ΓA . (158) Γ1,2,3,4 are real antisymmetric (anti-hermitian) matrices, and Γ5,6,7,8,9 are real symmetric (hermitian) matrices. The SO(5, 4) generators are given by 1 ΣAB = −i [ΓA , ΓB ], 4

(159)

or     i λI σIJ 0 0 , ΣI8 = − , 0 σIJ 2 0 −λI     1 0 −18 i 0 λI , Σ89 = −i , = 0 2 λI 0 2 18

ΣIJ = ΣI9

(160)

where σIJ are the SO(3, 4) generators 1 σIJ = i [λI , λJ ]. 4

(161)

Since ΓA are real matrices, the corresponding SO(5, 4) generators (159) are purely imaginary. Thus, we have confirmed that the present representation is Majorana, and the charge conjugation matrix is 116 . Though all of ΓA are not hermitian, the (real symmetric) hermitian matrices can be made as K A = KΓA , (162) where

 3 Σ 0

 0 , Σ3

(163)

 14 Σ = 0

 0 . −14

(164)

K= and

3

Explicitly, K A are given by    0 Σ3 λ8−I 0 8 KI = , K = −Σ3 λ8−I 0 Σ3  3  Σ 0 K9 = , 0 −Σ3 21

 Σ3 , 0 (165)

where Σ3 λI are the following real antisymmetric matrices:     2 0 −σ 3 0 0 σ 0 0 0  3  0 σ2 0 0 0 0  0 ,  , Σ3 λ2 = σ Σ3 λ1 = −i  0 0 0 0 −σ 3  0 σ2 0  0 0 σ3 0 0 0 0 σ2     0 0 12 0 0 −σ 1 0 0  σ 1 0 0 12  0 0 0  ,  , Σ3 λ4 =  0 Σ3 λ3 =  −12 0 0 0 0 0 0 −σ 1  0 −12 0 0 0 0 σ1 0     0 0 σ2 0 0 0 0 12 0  0 0 −σ 2  0 −12 0   , Σ3 λ6 =  0 , Σ3 λ5 = i  2 σ  0 0 0 0  12 0 0 0 −σ 2 0 0 −12 0 0 0   2 0 0 0 σ 2  0 0 σ 0 . (166) Σ3 λ7 = i  2 0 σ 0 0 2 σ 0 0 0 In the Majorana representation, the SO(5, 4) Majorana spinor is simply given by the (16-component) real spinor Ψ∗ = Ψ.

(167)

We now introduce the explicit realization of the 3rd non-compact Hopf map as Ψ → xA = Ψt K A Ψ,

(168)

where Ψ is a normalized SO(5, 4) Majorana spinor (the non-compact 3rd Hopf spinor) satisfying Ψt KΨ =

4 X

µ=1

Ψµ 2 +

12 X

µ=9

Ψµ 2 −

8 X

µ=5

Ψµ 2 −

16 X

Ψµ 2 = 1,

(169)

µ=13

and Ψ represents coordinates on H 8,7 . With two SO(4, 4) Majorana-Weyl spinors, Ψ+ and Ψ− (each of which represents 8-component real spinor), Ψ is represented as  + Ψ Ψ= , (170) Ψ− and xA given by (168) are expressed as t

xI = 2(Ψ+ ) Σ3 λ8−I Ψ− , t

x8 = 2(Ψ+ ) Σ3 Ψ− , t

t

x9 = (Ψ+ ) Σ3 Ψ+ − (Ψ− ) Σ3 Ψ− , 22

(171)

or − + − + − + − + − x1 = Ψ+ 1 Ψ8 − Ψ2 Ψ7 + Ψ3 Ψ6 − Ψ4 Ψ5 − (Ψ ↔ Ψ ),

− + − + − + − + − x2 = Ψ+ 1 Ψ7 + Ψ2 Ψ8 − Ψ3 Ψ5 − Ψ4 Ψ6 − (Ψ ↔ Ψ ),

− + − + − + − + − x3 = Ψ+ 1 Ψ6 − Ψ2 Ψ5 − Ψ3 Ψ8 + Ψ4 Ψ7 − (Ψ ↔ Ψ ),

− + − + − + − + − x4 = Ψ+ 1 Ψ5 + Ψ2 Ψ6 + Ψ3 Ψ7 + Ψ4 Ψ8 − (Ψ ↔ Ψ ),

− + − + − + − + − x5 = −Ψ+ 1 Ψ4 − Ψ2 Ψ3 − Ψ5 Ψ8 − Ψ6 Ψ7 − (Ψ ↔ Ψ ),

− + − + − + − + − x6 = −Ψ+ 1 Ψ3 + Ψ2 Ψ4 − Ψ5 Ψ7 + Ψ6 Ψ8 − (Ψ ↔ Ψ ),

− + − + − + − + − x7 = −Ψ+ 1 Ψ2 − Ψ3 Ψ4 − Ψ5 Ψ6 − Ψ7 Ψ8 − (Ψ ↔ Ψ ),

− + − + − + − x8 = Ψ+ 1 Ψ1 + Ψ2 Ψ2 + Ψ3 Ψ3 + Ψ4 Ψ4

− + − + − + − − Ψ+ 5 Ψ5 − Ψ6 Ψ6 − Ψ7 Ψ7 − Ψ8 Ψ8 ,

2

2

2

2

2

2

2

2

+ + + + + + + x9 = (Ψ+ 1 + Ψ2 + Ψ3 + Ψ4 ) − (Ψ5 + Ψ6 + Ψ7 + Ψ8 ) 2

2

2

2

2

2

2

2

− − − − − − − − (Ψ− 1 + Ψ2 + Ψ3 + Ψ4 ) + (Ψ5 + Ψ6 + Ψ7 + Ψ8 ).

(172)

A straightforward calculation shows that xA satisfy X ηAB xA xB A,B=1,2,··· ,9

= −(x1 )2 − (x2 )2 − (x3 )2 − (x4 )2 + (x5 )2 + (x6 )2 + (x7 )2 + (x8 )2 + (x9 )2 =(

4 X

2 (Ψ+ µ) −

t

2

µ=1

8 X

2 (Ψ+ µ) +

µ=5

4 X

µ=1

2 (Ψ− µ) −

8 X

2 2 (Ψ− µ) )

µ=5

= (Ψ KΨ) = 1,

(173)

and xA are regarded as coordinates on H 4,4 . Inverting (168), Ψ can be represented as   1 (1 + x9 )Φ Ψ= p , (174) 8 I 2(1 + x9 ) (x − λI x )Φ where Φ is SO(4, 3) real 8-component spinor that satisfies the constraint t

3

ΦΣ Φ=

4 X

µ=1

2

(Φµ ) −

8 X

(Φµ )2 = 1.

(175)

µ=5

Φ denotes the H 4,3 -fibre. As discussed below (144), naively anticipated connections A = −iΨt KdΨ vanish due to the Majorana property of Ψ. Hence, defining   1 (1 + x9 )18 ˜ (176) Ψ= p 8 i , 2(1 + x9 ) x 18 − λi x we evaluate the connections of H 4,3 -fibre as

˜ t KdΨ ˜ = dxA Σ3 AA , A = −iΨ 23

(177)

and obtain 1 σMN xN , 1 + x9 A9 = 0.

AM =

(178)

Here, σMN are 1 σIJ = −i [λI , λJ ], 4

1 σI8 = −σ8I = i λI . 2

(179)

(Σ3 σIJ and Σ3 σI8 are pure imaginary antisymmetric hermitian matrices.) Similarly, the corresponding curvatures FAB = ∂A AB − ∂B AA − i[AA , AB ]

(180)

are computed as FMN = −xM AN + xN AM − σMN , FM9 = −F9M = (1 + x9 )AM .

6

(181)

Summary and Discussion

Based on the split-algebras, we developed the non-compact version of the Hopf maps. With use of the split-imaginary unit in place of the ordinary imaginary unit, the ultra-hyperboloids with split-signatures naturally appeared, and the non-compact Hopf maps between ultra-hyperboloids were introduced as well. With the explicit realization of the non-compact Hopf maps, the associated non-compact fibre structures were also evaluated. As briefly mentioned in Introduction, the Hopf fibrations correspond to the fibre configurations of the monopoles with minimum charge, and the stabilities of such U (1), SU (2), and SO(8) monopole with arbitrary charges are topologically guaranteed by the homotopy theorems π1 (U (1)) ≃ Z,

π3 (SU (2)) ≃ Z,

π7 (SO(8)) ≃ Z ⊕ Z.

(182)

Similarly, the three non-compact Hopf maps may respectively correspond to the fibre configurations of SO(1, 1), SU (1, 1), and SO(4, 4) monopoles with minimum charge. Like the original monopoles, such non-compact monopoles are expected to carry the discrete topological charges. However, the author does not know rigorous discussions about homotopy between non-compact spaces that would account for the topological stability of non-compact monopoles. The homotopy theorems for non-compact monopoles may have to be exploited in a future research. As also mentioned in Introduction, there are wide applications of the original Hopf maps to physics. In particular, there are intriguing relations reported between the division algebras and the N = 1 SUSY Yang-Mills theory and 24

Real numbers Complex numbers Quaternions Octonions

Division algebras SO(2, 1) ≃ SL(2, R) SO(3, 1) ≃ SL(2, C) SO(5, 1) ≃ SL(2, H) SO(9, 1) ≃ SL(2, O)

Split algebras SO(2, 1) ≃ SL(2, R) SO(2, 2) ≃ SL(2, C′ ) SO(3, 3) ≃ SL(2, H′ ) SO(5, 5) ≃ SL(2, O′ )

Table 2: Lorentz groups and split-Lorentz groups Green-Schwarz superstring formalism in special dimensions [18, 19]. In the language of group theory, the 1st Hopf map is related to the SL(2, C) group which is the (3+1)D Lorentz group, and the octonion group related to the 3rd Hopf map corresponds to the Lorentz group in (9+1)D [20, 21]: the critical dimension of the superstring theory. In the case of the non-compact Hopf versions, one may infer similar correspondences (see Table 2), and the split-algebras naturally bring the notion of space-time with split-signatures (split-space-time) [22, 23]. It is stimulating, and may be worthwhile, to speculate applications of the splitalgebras to physics. Indeed, recently, analogies of quantum Hall effect and the twistor theory were explored based on the 2nd non-compact Hopf map [24]. Though direct relations to the split-algebras have not been discussed yet, splitspace-time groups play a crucial role in the “doubled geometry” formulation of string theory [25]. Since split-space-times generally have more than one-time dimension, one may be anxious about the negative-norm problem. However, at the same time, it may also be probable that dualities between space-times with different signature metrices secure the theories from the fatal negative norm problems [26, 27]. Apparently, this also needs more detail investigations. Several other generalizations of the Hopf maps have also been reported. For instance, the supersymmetric version of the 1st Hopf map was explored in Refs.[15, 16]. (We discuss its non-compact version in Appendix.) Though yet to be found, it might even exist supersymmetric versions of the 2nd and 3rd Hopf maps 4 . The θ-deformed 2nd Hopf map has also been explored in Ref.[29]. Such generalization may also be applicable to the present non-compact Hopf maps.

Acknowledgement The author would like to thank Professor Mikio Nakahara for useful conversations. This work was partially supported by Sumitomo foundation. 4 An

attempt to this direction, one may see Ref.[28].

25

Appendix A

Non-compact Supersymmetric Hopf Map

The supersymmetric (1st) Hopf map is introduced by replacing the original bosonic manifolds with their supersymmetric counterparts [15, 16]: S1

S 3|2 −→ S 2|2 .

(183)

The body of the base-manifold S 2|2 is S 2 , and the isometry group of S 2|2 is U OSp(1|2) whose bosonic subgroup is SU (2). Meanwhile, the non-compact version of the supersymmetric Hopf map is given by H 1,0

H 2,1|2 −→ H 1,1|2 , or

U (1)

AdS 3|2 −→ AdS 2|2 .

(184) (185)

The body of the base-manifold H 1,1|2 is H 1,1 , and the isometry group of AdS 2|2 is OSp(1|2) whose bosonic subgroup is SU (1, 1).

A.1

Realization I

The OSp(1|2) algebra is given by [li , lj ] = jǫijk lk , 1 [li , lα ] = (σ i )β α lβ , 2 1 α β {l , l } = (ǫσ i )αβ li , 2

(186)

where ǫ123 = 1, and ηij = diag(+1, −1, +1). With the split-Pauli matrices σ i , and τ 1 = (1, 0)t , τ 2 = (0, 1)t , the OSp(1|2) generators are represented as     1 σi 0 1 0 τα i α l = , l = , (187) 2 0 0 2 −(ǫτ α )t 0 or    0 1 0 0 −j 1 1 l1 = 1 0 0 , l2 = j 0 2 2 0 0 0 0 0    0 0 1 0 0 1 1 lθ1 = 0 0 0 , lθ2 =  0 0 2 2 0 1 0 −1 0

26

  0 1 0 1 0 , l3 = 0 −1 2 0 0 0  0 1 . 0

 0 0 , 0 (188)

The OSp(1|2) spinor takes the form of   u χ = v  , η

(189)

where the first two components of ψ, u and v, are Grassmann even, while the last component, η, is Grassmann odd. With li and lα , the non-compact supersymmetric Hopf map (184) is realized as χ → xi = 2χ‡ li χ, θα = 2χ‡ lα χ. (190) Here, χ is a normalized OSp(1|2) spinor (the supersymmetric non-compact Hopf spinor) that satisfies χ‡ χ = u∗ u + v ∗ v − η ∗ η = (uR )2 + (vR )2 − (uI )2 − (vI )2 − η ∗ η = 1,

(191)

where the superadjoint ‡ is defined as χ‡ = (u∗ , v ∗ , −η ∗ ).

(192)

5

Thus, χ is regarded as coordinates on AdS 3|2 . Meanwhile, xi and θα given by (190) are x1 = u∗ v + v ∗ u, x2 = −ju∗ v + jv ∗ u, x3 = u∗ u − v ∗ v, θ1 = u∗ η − η ∗ v, θ2 = v ∗ η + η ∗ u,

(193)

and satisfy ηij xi xj + ǫαβ θα θβ = 1.

(194)

Thus, xi and θα denote coordinates on AdS 2|2 . Inverting the map (190), χ can be expressed as   1 (1 + x5 )(1 − 4(1+x 5 ) θǫθ) 1 1 (x1 − jx2 )(1 −  jϑ χ= p (195) 4(1+x5 ) θǫθ) e . 2(1 + x5 ) 3 1 2 1 2 (1 + x )θ + (x − jx )θ The U(1)-fibre connections are induced as A = dxi Ai + dθα Aα = −jχ‡ dχ,

(196)

with   2 + x3 1 j 1 + ǫ x θǫθ , ij3 2(1 + x3 ) 2(1 + x3 ) 1 Aα = j(xi σ i ǫθ)α . 2

Ai =

(197)

5 ∗ represents the pseudo-conjugation, (η ∗ )∗ = −η, (η η )∗ = η ∗ η ∗ . In particular, (θ 1 )∗ = 1 2 1 2 θ 2 , (θ 2 )∗ = −θ 1 .

27

Ai and Aα are a supermultiplet of the OSp(1|2) group, and they may be interpreted as (non-compact) gauge fields of supermonopole. The corresponding curvatures are given by F = dA =

1 i 1 dx ∧ dxj Fij + dxi ∧ dθα Fiα − dθα ∧ dθβ Fαβ , 2 2

(198)

or Fij = ∂i Aj − ∂j Ai ,

Fiα = ∂i Aα − ∂α Ai ,

Fαβ = ∂α Aβ + ∂β Aα ,

(199)

which are explicitly Fij = −

  1 3 j ǫ x θǫθ , 1 + ij3 2(1 + x3 ) 2

1 j(σ j ǫθ)α (ηij − 3xi xj ), 2   3 = jxi (σi ǫ)αβ 1 + θǫθ . 2

Fiα = Fαβ

A.2

(200)

Realization II

The non-compact supersymmetric Hopf map (184) cannot be expressed by using the ordinary imaginary unit, and the alternative non-compact supersymmetric Hopf map is introduced as [17] H 0,1

H 2,1|2 −→ H 2,0|2 , or

U(1)

AdS 3|2 −→ Eucl.AdS 2|2 .

(201)

(202)

The OSp(1|2) algebra is given by [li , lj ] = iǫij k lk , 1 [li , lα ] = (τ i )β α lβ , 2 1 {lα , lβ } = (ǫt τi )αβ li , 2

(203)

where τ i are the SU (1, 1) generators (42), and ηij = diag(+1, +1, −1). The OSp(1|2) generators are     1 1 τi 0 0 τα α i , l = . (204) l = 2 0 0 2 −(ǫτ α )t 0 Their complex representation is given by ˜li = −(li )∗ and ˜lα = ǫαβ lβ , and ˜li and ˜lα are related to li and lα by the relation ˜ li = R† li R, ˜lα = R† lα R. 28

(205)

R is the charge conjugation matrix 

0 1  R= 1 0 0 0

 0 0 , 1

(206)

which has the following properties R† = Rt = R−1 = R.

(207)

R is diagonalized to yield 

 −1 0 0 − κ =  0 1 0 , 0 0 1

(208)

and “hermitian” matrices are constructed by multiplying κ to li and lα as κi = 2κli ,

κα = 2κlα ,

(209)

or  2   1    σ 0 σ 0 1 0 κ1 = − , κ2 = , κ3 = , 0 0 0 0 0 0     0 τ θ1 0 −τ θ2 θ2 κθ 1 = = , κ . −(τ θ2 )t 0 0 (τ θ1 )t

(210)

κi and κα are hermitian in the sense †

κi = κi , κα† = (σ 1 )β α κβ .

(211)

The supersymmetric non-compact Hopf map (201) is realized as χ → xi = χ† κi χ, θα = χ† κα χ, with χ = (u, v, η)t satisfying the normalization condition

(212) 6

χ† κχ = u∗ u − v ∗ v − η ∗ η = 1.

(213)

xi and θα (212) are explicitly x1 = iu∗ v − iv ∗ u, θ1 = u∗ η − η ∗ v,

x2 = u∗ v + v ∗ u, θ2 = −v ∗ η + η ∗ u,

x3 = u∗ u + v ∗ v (214)

6 Here, † and ∗ represent the conventional hermitian conjugation and complex conjugation, respectively; χ† = (u∗ , v∗ , η∗ ), and (η∗ )∗ = η, (η1 η2 )∗ = η2∗ η1∗ . By the “hermiticity” of κi ∗ and κα (211), xi and θ α are “real” in the sense: xi = xi and θ α∗ = (σ1 )β α θ β . Then, θ = (θ 1 , θ 2 )t is a SO(2, 1) Majorana-spinor.

29

and they denote coordinates on Euclidean AdS 2|2 , since ηij xi xj − ǫαβ θα θβ = −(χ† κχ)2 = −1.

(215)

Inverting (212), χ is represented as   1 (1 + x3 )(1 − 4(1+x 3 ) θǫθ) 1 1 (x2 − ix1 )(1 +  · eiϑ , χ= p 4(1+x3 ) θǫθ) 2(1 + x3 ) 3 1 2 1 2 (1 + x )θ + (x − ix )θ

(216)

and the U (1)-fibre connections are derived as A = −iχ† κdχ = dxi Ai + dθα Aα ,

(217)

where   2 + x3 xj 1 1 + θǫθ , Ai = − ǫij 3 2 1 + x3 2(1 + x3 ) 1 Aα = −i xi (θτi ǫ)α . 2

(218)

The corresponding curvatures Fij = ∂i Aj − ∂j Ai , Fiα = ∂i Aα − ∂α Ai ,

Fαβ = ∂α Aβ + ∂β Aα ,

(219)

are also evaluated as 1 3 Fij = − ǫijk xk (1 + θǫθ), 2 2 1 Fiα = −i (θτ j ǫ)α (ηij − 3xi xj ), 2 3 Fαβ = −i(τi ǫ)αβ xi (1 + θǫθ). 2

(220)

References ¨ [1] Heinz Hopf, ”Uber die Abbildungen der dreidimensionalen Sph¨ are auf die Kugelfl¨ ache”, Mathematische Annalen 104 (1931) 637-665. ¨ [2] Heinz Hopf, ”Uber die Abbildungen von Sph¨ aren auf Sph¨ aren niedrigerer Dimension”, Fundamenta Mathematicae 25 (1935) 427-440. [3] J.F. Adams, “On the Non-existence of Elements of Hopf Invariant One”, Annals of Mathematics 72 (1960) 20-104. [4] Mikio Nakahara, “Geometry, Topology and Physics”, (2nd ed.), IOP Publishing, (2003). 30

[5] Chia-Hsiung Tze, Feza G¨ ursey, “On the Role of Division, Jordan and Related Algebras in Particle Physics”, World Scientific, (1996). [6] Paul Dirac, ”Quantised Singularities in the Electromagnetic Field”, Proc. Roy. Soc. (London) A 133 (1931) 60. [7] Chen Ning Yang, “Generalization of Dirac’s monopole to SU2 gauge fields”, J. Math. Phys. 19 (1978) 320. [8] Bernard Grossman, Thomas W. Kephart, James D. Stasheff, “Solutions To Yang-Mills Field Equations In Eight-Dimensions And The Last Hopf Map”, Commun.Math.Phys. 96 (1984) 431, Erratum-ibid.100 (1985) 311. [9] See for instance, John C. Baez, “The Octonions”, Bull. Amer. Math. Soc. 39 (2002), 145-205. Errata in Bull Amer. Math. Soc. 42 (2005), 213; math/0105155. [10] William Rowan Hamilton, “On a new Species of Imaginary Quantities connected with a theory of Quaternions”, Proceedings of the Royal Irish Academy, 2 (1844), pp. 424-434. [11] James Cockle, “On Certain Functions Resembling Quaternions, and on a New Imaginary Algebra”, Phil. Mag. (3) 33 (1848) 435-439. [12] James Cockle, “On a New Imaginary in Algebra”, Phil. Mag. (3) 34 (1849) 37-47. “On Systems of Algebra Involving more than one Imaginary and on Equations of the Fifth Degree”, Phil. Mag. (3) 35 (1849) 434-437. [13] Shou-Cheng Zhang, Jiangping Hu “A Four Dimensional Generalization of the Quantum Hall Effect”, Science 294 (2001) 823; cond-mat/0110572. [14] B.A. Bernevig, J.P. Hu, N. Toumbas, S.C. Zhang “The Eight Dimensional Quantum Hall Effect and the Octonions”, Phys.Rev.Lett. 91 (2003) 236803; cond-mat/0306045. [15] G. Landi, G. Marmo, “Extensions of Lie Superalgebras and Supersymmetric Abelian Gauge Fields”, Phys.Lett.B193 (1987) 61. [16] Kazuki Hasebe, Yusuke Kimura, “Fuzzy Supersphere and Supermonopole”, Nucl.Phys.B709 (2005) 94-114; hep-th/0409230. [17] Kazuki Hasebe, “Hyperbolic Supersymmetric Quantum Hall Effect”, Phys.Rev.D78 (2008)125024, arXiv:0809.4885. [18] Taichiro Kugo, Paul K. Townsend, “Supersymmetry and the Division Algebras”, Nucl.Phys.B221 (1983) 357. [19] J.M. Evans, “Supersymmetric Yang-Mills Theories and Division Algebras”, Nucl.Phys.B298 (1988) 92.

31

[20] Anthony Sudbery, “Division algebras, (pseudo)orthogonal groups and spinors”, J.Phys. A: Math. Gen. 17 (1984) 939-955. [21] K.W. Chung, Anthony Sudbery, “Octonions and the Lorentz and Conformal Groups Of Ten-Dimensional Space-Time”, Phys.Lett.B198 (1987) 161. [22] R. Foot, G.C. Joshi, “Nonstandard Signture of Spacetime, Superstrings, and the Split Composition Algebras”, Lett.Math.Phys. 19 (1990) 65-71. [23] R. Foot, G.C. Joshi, “An Application of The Division Algebras, Jordan Algebras and Split Composition Algebras”, Int.J.Mod.Phys. A 7 (1992) 43954413. [24] Kazuki Hasebe, “Split-Quaternionic Hopf Map, Quantum Hall Effect and Twistor Theory”, arXiv:0902.2523. [25] C M Hull “Doubled Geometry and T-Folds”, JHEP0707:080,2007; hep-th/0605149. [26] C.M. Hull, R.R.Khuri “Branes, Times and Dualities”, Nucl.Phys. B536 (1998) 219-244; hep-th/9808069. [27] Zhanna Kuznetsova, Francesco Toppan “Superalgebras of (split)division algebras and the split octonionic M-theory in (6,5)-signature”, hep-th/0610122. [28] Z. Kuznetsova, A. Nersessian, F. Toppan, V. Yeghikyan . “Second Hopf map and supersymmetric mechanics with Yang monopole”, arXiv:0902.2682. [29] Giovanni Landi, Chiara Pagani, Cesare Reina, “A Hopf bundle over a quantum four-sphere from the symplectic group”, Commun.Math.Phys. 263 (2006) 65-88; math/0407342.

32