Arxiv:2011.05677V2

arXiv:2011.05677v2 [quant-ph] 22 May 2021 pnetnldSuee tt naBohFour-hyperboloid Bloch a on State Squeezed Spin-entangled n[3,teato tlzdanncmat2dHp a n B a and map Hopf 2nd non-compact a utilized author the [23], In an an h previous the h obtained the the nage pnpiso l nee pn.W lrf basic clarify We spins. integer all of spin-pairs entangled mhsso hi itntost h original the corre to spin distinctions their entanglement, on spin emphasis of entropy Neumann von as such SO SO h lc hyperboloid Bloch The SO (2 ainlIsiueo ehooy ediClee Ayashi, College, Sendai Technology, of Institute National (4 (4 , , )etnino h qezdsae.I h rsn ae,we paper, present the In states. squeezed the of extension 3) , )vrino qezdvcu ae nteohrBohfour-h Bloch other the on based vacuum squeezed of version 1) )suee aumi nepee sasproiino nin an of superposition a as interpreted is vacuum squeezed 1) SO SO (2 (4 , , )suee aum nve fteShigrsfruaino formulation Schwinger’s the of view In vacuum. squeezed 3) )suee aumi atclrfu-oesuee state squeezed four-mode particular a is vacuum squeezed 1) H 2 nele h unu emtyo h original the of geometry quantum the underlies [email protected] auiHasebe Kazuki a 5 2021 25, May SO Abstract (2 , )case. 1) rpriso the of properties ain n netit eain with relations uncertainty and lations ohfour-hyperboloid loch edi 8-18 Japan 989-3128, Sendai, ute useteie oderive to idea the pursue further yperboloid nt ubro maximally of number ﬁnite SO SO (4 , nua momentum, angular f (2 )suee vacuum, squeezed 1) o ut iia to similar quite not , H )suee states. squeezed 1) 4 eso that show We . H 2 , 2 oexplore to Contents

1 Introduction 2

2 Bloch hyperboloid and squeezed states 3 2.1 Blochsphereandspin-coherentstate ...... 4 2.1.1 Schwinger operator realization ...... 5 2.2 Bloch hyperboloid and pseudo-spin-coherent state ...... 7 2.2.1 SU(1, 1) algebra and the Schwinger operator realization ...... 7 2.2.2 SU(1, 1) coset representative and the squeeze operator ...... 9 2.3 Two-mode SO(2, 1) squeeze vacuum and entangled spin-coherent states ...... 10 2.3.1 SO(2, 1) two-mode squeeze operator and the interaction Hamiltonian ...... 10 2.3.2 Two two-mode squeezed vacua as entangled spin-coherent states ...... 10

3 Bloch four-hyperboloid and SO(4, 1) squeeze operator 11 3.1 SO(4, 1)algebraandBlochfour-hyperboloid ...... 11 3.1.1 H4 and the SO(4, 1)algebra ...... 11 3.1.2 Bloch four-hyperboloid and SO(4, 1)cosetrepresentative ...... 13 3.2 SO(4, 1)squeezeoperator ...... 15 3.2.1 SO(4, 1) Schwinger operator realization and SO(4, 2)operators ...... 15 3.2.2 Subgroup decomposition of the SO(4, 1)squeezeoperator ...... 16 3.2.3 SO(4, 1) squeeze operator and the interaction Hamiltonian ...... 17

4 SO(4, 1) squeezed vacuum and its basic properties 19 4.1 SO(4, 1)squeezedvacuum...... 19 4.1.1 Spin interpretation ...... 19 4.1.2 Spinentanglement ...... 20 4.2 DimensionalReduction...... 21 4.3 Statisticalproperties ...... 23 4.3.1 Statistics about spins ...... 23 4.3.2 vonNeumannentropy ...... 25 4.4 SO(4, 1)uncertaintyrelations...... 26

5 Summary 27

A Basic properties of the SO(2, 1) two-modesqueezedvacuum 28 A.1 One-mode SO(2, 1) squeezed vacuum (Majoranarepresentation) ...... 28 A.2 Two-mode SO(2, 1)squeezedvacuum(Dirac representation)...... 29

B Expectation values of the SO(4, 1) and SO(2, 3) squeezed vacua 30 B.1 SO(4, 1)squeezedvacuum...... 30 B.2 SO(2, 3) Sp(4; R)squeezedvacua...... 32 ≃ B.2.1 Two-mode SO(2, 3) squeezed vacuum (Majoranarepresentation) ...... 32 B.2.2 Four-mode SO(2, 3)squeezedvacuum(Diracrepresentation) ...... 33

1 1 Introduction

Squeezed state [1, 2, 3, 4, 5, 6] is a particularly important quantum state in quantum optical information and practical applications (see [8, 9, 10] as nice reviews.) In quantum optical information, two-mode squeezed light realizes an entangled state of photons and it has been employed in state-of-the-art experiments of quantum mechanics (see [7] for instance and references therein). Also in practical application to quantum telecommunication, the squeezed state has attracted a lot of attention due to its intrinsic anisotropic property of quantum noise. The uncertainty region of squeezed state is not isotropic unlike the coherent state (laser) and an unfavorable quantum noise is reduced in a speciﬁc direction. Up to now, many theoretical extensions of the squeezed state have been proposed [11, 12, 13, 14, 15, 16, 17, 18, 19, 20]. In an attempt of extending the formalism of the squeezed state, mathematical aesthetics and logically natural extension should be appreciated. One fascinating way to formulate a theory of physics may resort to geometry. Needless to say, most successful theories of physics in this approach may be Einstein’s general relativity of Riemann geometry and Yang-Mills theory of ﬁbre-bundle geometry. Also in quantum information theory, there is a reasonable fact that we can believe that geometric approach plays a key role, since qubit, the fundamental object in quantum information, is described by the geometry of Bloch sphere associated with the Hopf map [21, 22], and geometric structure seems to be inherent in the formulation. As Bloch sphere provides a geometric description of qubit, Bloch hyperboloid plays a similar role in description of squeezed state [see [23] and references therein]. Mathematically, Bloch sphere realizes the basemanifold of the 1st Hopf map, and similarly Bloch hyperboloid denotes the basemanifold of the non- compact 1st Hopf map. For 2D hyperboloids, we have two kinds of hyperboloids, two-leaf hyperboloid and one-leaf hyperboloid, which are related to the two kinds of the non-compact 1st Hopf maps:

H2 H2,1/S1, H1,1 H2,1/H1. (1) ≃ ≃ In both cases, the total bundle space is H2,1. The base-manifolds are H2 and H1,1 with compact ﬁbre S1 and non-compact ﬁbre H1 respectively. The geometric origin of the original squeezed state is accounted for by the left of (1) with the Bloch hyperboloid H2. Interestingly, the 1st non-compact Hopf maps have their higher dimensional cousins, i.e. the 2nd and 3rd non-compact Hopf maps [24, 25]. This observation led the author to propose an extension of squeezed state based on the geometry of higher dimensional hyperboloids [23].1 As a higher dimensional analogue of (1), we have the 2nd non-compact Hopf maps:

H4 H4,3/S3, H2,2 H4,3/H2,1. (6) ≃ ≃ As in (1), the total space in both cases of (6) is given by H4,3, and H4 and H2,2 respectively realize the base-manifolds with compact ﬁbre S3 and non-compact ﬁbre H2,1. We utilized H2,2 (the right of (6))

1The non-compact Hopf maps have already been applied to several subjects, such as twistorial quantum Hall eﬀect [26], non-Hermitian topological insulators [27, 28], and cosmological models of matrix model [29, 30, 31]. Higher dimensional hyperboloid Hp,q is deﬁned as p p+q+1 xixi xixi = 1. (2) − − Xi=1 i=Xp+1 As a coset space, Hp,q is represented as Hp,q SO(p,q + 1)/SO(p,q), (3) ≃ and the topology is Hp,q Rp Sq. (4) ≃ ⊗ In low dimensions, Hp,q realizes Anti-de Sitter, de Sitter and Euclidean anti-de Sitter spaces:

H2,0 = EAdS2, H1,1 = dS2 = AdS2, H4,0 = EAdS4, H3,1 = AdS4, H1,3 = dS4. (5)

2 to construct an Sp(4; R) SO(2, 3) extension of squeezed state [23]. In the present paper, we propose ≃ an SO(4, 1) version of squeezed state based on the H4 geometry (the right of (6)). It turns out that the SO(4, 1) squeezed vacuum attains a more natural generalization of the original squeezed vacuum than the previous SO(2, 3) case in several respects. The SO(4, 1) squeezed vacuum accommodates an interesting spin structure related to the hierarchy of the non-compact 2nd Hopf map:

1 S3 S S2 −→ H4,3 H4 −→ H4,3 can be regarded as a fibre-bundle of base-manifold H4 with S3-fibre which itself denotes the fibre- bundle space of 1st Hopf map of the Bloch sphere. This implies the existence of inherent spin geometry in the present SO(4, 1) formulation. Before proceeding to detail discussions, we mention differences between the previous SO(2, 3) formulation [23] and the present SO(4, 1) formulation in a group theory point of view. For the original SO(2, 1) squeezed state, the associated two-hyperboloid is given by H2 SO(2, 1)/SO(2) SU(1, 1)/U(1) Sp(2; R)/U(1) ≃ ≃ ≃ U(1; H′)/U(1). (7) ≃ The symmetry group SO(2, 1) allows Majorana representation as well as the Dirac representation, which leads to one-mode and two-mode squeeze operators. The holonomy group is a compact group SO(2) U(1). ≃ Meanwhile in the previous work [23], we utilized a Bloch four-hyperboloid (right of (6)): H2,2 SO(2, 3)/SO(2, 2) Sp(4; R)/(SU(1, 1) SU(1, 1))) ≃ ≃ ⊗ U(2; H′)/(U(1; H′) U(1; H′)). (8) ≃ ⊗ The symmetry group SO(2, 3) also accommodates Majorana representation as well as the Dirac representation, and the corresponding two-mode and four-mode SO(2, 3) squeezed states are constructed. A crucial difference to the original case is the holonomy group SO(2, 2) SU(1, 1) SU(1, 1), which is non-compact ≃ ⊗ unlike SO(2). In the present work, we will adopt the other Bloch four-hyperboloid (left of (6)): H4 SO(4, 1)/SO(4) USp(2, 2)/(SU(2) SU(2))) ≃ ≃ ⊗ U(1, 1; H)/(U(1; H) U(1; H)). (9) ≃ ⊗ The symmetry group SO(4, 1) does not accommodate Majorana representation, and so the two-mode realization of SO(4, 1) squeezed state is not possible, and hence the SO(4, 1) squeezed operator can be realized only by a four-mode Dirac representation. An analogous point to the original case is the holonomy group, which is given by a compact group SO(4) SU(2) SU(2) like SO(2). ≃ ⊗ This paper is organized as follows. In Sec.2, we review spin-coherent state and squeezed state with emphasis on their relations to the geometry of Bloch sphere and hyperboloid. We also discusses interesting relationship between spin-coherent states and squeezed states. In Sec.3, the SO(4, 1) squeeze operator is introduced based on the geometry of the Bloch four-hyperboloid H4. In Sec.4, we propose a physical interpretation of the SO(4, 1) squeezed vacuum in spin geometry and explore its basic properties, such as von Neumann entropy, spin correlations and uncertainty relations. Sec.5 is devoted to summary and discussions.

2 Bloch hyperboloid and squeezed states

We begin with a review of the Bloch sphere and Bloch hyperboloid and their associated coherent states. The Schwinger operator formulation is not only useful to represent spin state with arbitrary spin but also

3 closely related to the squeeze operator.

2.1 Bloch sphere and spin-coherent state The geometry of the Bloch sphere [21] stems from the 1st Hopf map:2

1 S2 S S2, (10) −→ 3 and the Hopf map is explicitly realized as the map from the 2-component normalized spinor (ψ†ψ = 1 : S ) 3 2 to a normalized three component vector ( i=1 xixi = 1 : S ) [32]: P ψ1 ψ = x = ψ†σ ψ. (11) ψ → i i 2

Here, σi (i =1, 2, 3) are the Pauli matrices:

0 1 0 i 1 0 σx = , σy = − , σz = . (12) 1 0 i 0 0 1 − When we parameterize x S2 by using the usual polar coordinates i ∈

x1 = sin θ cos φ, x2 = sin θ sin φ, x3 = cos θ, (13)

ψ can be expressed as θ 1 i 2 φ 1 cos( 2 ) e− i χ ψ(φ,θ,χ)= 1 e− 2 . (14) θ i 2 φ sin( 2 ) e !

i 1 χ 1 Here, e− 2 denotes the overall U(1) phase that corresponds to the S ﬁbre in (10). Importantly, ψ has a physical meaning as spin-coherent state [33, 34, 35] that is aligned to the direction of (θ, φ) on the S2 (see the left of Fig.1): (x(θ, φ) σ)ψ = ψ. (15) · Applying the Euler angle rotation [36] to the spin state aligned to z direction, we can readily reproduce (14) as

(1/2) 1 θ i 1 (χ+φ) 1 θ i 1 (χ φ) 0 ψ(φ,θ,χ) D (φ,θ,χ) = cos( ) e− 2 + sin( ) e− 2 − . (16) ≡ 0 2 0 2 1 The rotation matrix in (16) is known as the Wigner’s (spin 1/2) D-matrix of the SU(2) group :

θ i 1 (χ+φ) θ i 1 (χ φ) φ θ χ 2 2 (1/2) i σz i σy i σz cos( 2 ) e− sin( 2 ) e − D (φ,θ,χ)= e− 2 e− 2 e− 2 = 1 − 1 , (17) θ i 2 (χ φ) θ i 2 (χ+φ) sin( 2 ) e− − cos( 2 ) e ! where the range of the parameters are

0 φ< 2π, 0 θ < π, 0 χ< 4π. (18) ≤ ≤ ≤ For later convenience, we also introduce coset representative of the SU(2) group.3 As S2 is expressed as a coset S2 S3/S1 SU(2)/U(1), (19) ≃ ≃ 2See [22] for the roles of the 2nd Hopf map in the context of entanglement of qubits. 3Coset representative (see [38] for instance) is also referred to as the non-linear realization [39, 40, 41].

4 Figure 1: Bloch sphere and Bloch hyperboloid for the spin and pseudo-spin-coherent states.

the corresponding coset representative is obtained as

θ θ iφ θ −iφ + iφ − iθ P ymσm3 (e σ e σ ) cos 2 sin 2 e− e m=1,2 = e− 2 − = − , (20) sin θ eiφ cos θ 2 2 where y1 = cos φ, y2 = sin φ, (21) and 1 + 0 1 0 0 σ = ǫ σ , σ = , σ− = . (22) ij 2 ijk k 0 0 1 0 (20) can be factorized as

φ θ φ iθ P ymσm3 i σz i σy i σz (1/2) Euler decomposition : e m=1,2 = e− 2 e− 2 e 2 = D (φ, θ, φ) (23) − and θ −iφ + θ θ iφ − iθ P ymσm3 tan e σ ln(cos )σz tan e σ Gauss decomposition : e m=1,2 = e− 2 e− 2 e 2 . (24) Notice that the Euler decomposition (23) is realized as a special case of the D-matrix and the physical meaning of the coset representative is transparent in this decomposition.

2.1.1 Schwinger operator realization We construct the spin-coherent state with arbitrary spin magnitude using the Schwinger operator formalism [37, 36]. With the Schwinger boson operator

a ψˆ (25) ≡ b subject to [a,a†] = [b,b†]=1, [a,b] = [a,b†]=0, (26) the spin operators are expressed as 1 Sˆ = ψˆ† σ ψ.ˆ (27) i 2 i In detail, 1 1 1 Sˆ = (a†b + b†a), Sˆ = i (a†b b†a), Sˆ = (a†a b†b), (28) x 2 y − 2 − z 2 −

5 and so 2 Sˆ = Sˆ(Sˆ + 1) (29) with 1 1 Sˆ = (ˆn +ˆn )= (a†a + b†b). (30) 2 a b 2 The spin states S,S that satisfy | zi 2 Sˆ S,S = S(S + 1) S,S , Sˆ S,S = S S,S (31) | zi | zi z| zi z| zi are constructed as 1 na nb S,Sz na,nb = (a†) (b†) 0, 0 , (32) | i ≡ | i √na!nb! | i

where the relations between the boson numbers (na,nb) and spin values (S,Sz) are depicted in Fig.2, and hereafter we utilize n ,n and S,S interchangeably. | a bi | zi

Figure 2: The SU(2) irreducible representation with ﬁxed S is speciﬁed by each of the oblique blue lines of na + nb =2S .

We introduce D-operator as iφSˆz iθSˆy iχSˆz Dˆ(φ,θ,χ) e− e− e− (33) ≡ whose matrix representation is the D-matrix:

(S) S,S Dˆ(φ,θ,χ) S′,S′ = D (φ,θ,χ) ′ δ ′ , (34) h z| | zi Sz ,Sz · S,S i.e., (S) iφSz iθSy iχSz D (φ,θ,χ)= e− e− e− . (35)

In (35), Si=x,y,z denote the SU(2) matrices with spin magnitude S. Under the SU(2) transformation of Dˆ(φ,θ,χ), (25) transforms as an SU(2) spinor

(1/2) Dˆ(φ,θ,χ)† ψˆ Dˆ(φ,θ,χ)= D (φ,θ,χ) ψ.ˆ (36)

In a similar manner to (16), we can readily construct the spin-coherent state with arbitrary spin magnitude:

iSzχ S,S (φ,θ,χ) = Dˆ(φ,θ,χ) S,S = e− Dˆ(φ, θ, 0) S,S , (37) | z i | zi | zi which satisﬁes

2 1 Sˆ S,S (φ,θ,χ) = (n + n )(n + n + 2) S,S (φ,θ,χ) , | z i 4 a b a b | z i 1 (x(θ, φ) Sˆ) S,S (φ,θ,χ) = (n n ) S,S (φ,θ,χ) . (38) · | z i 2 a − b | z i

6 With (34), (37) can be concisely expressed as a linear combination of the spin states with expansion coeﬃcients of the components of D-matrix:4

S (S) S,S (φ,θ,χ) = D (φ,θ,χ) ′ S,S′ , (39) | z i Sz ,Sz | zi S′ = S zX− which is a ket state expression of (16) generalizing the spin 1/2 to arbitrary spin magnitude. In the Schwinger boson notation, (39) is represented as

2S 1 (S= (na+nb)) 2 1 ′ ′ 1 S,Sz(φ,θ,χ) = D (φ,θ,χ) (n n ), (na n ) na′ ,nb′ . (40) | i 2 a− b 2 − b | i n′ ,n′ =0 aXb 2.2 Bloch hyperboloid and pseudo-spin-coherent state Here, we extend the above discussions to the case of hyperboloid. The SU(1, 1) version of D-operator realizes the squeeze operator in its special form.

2.2.1 SU(1, 1) algebra and the Schwinger operator realization The non-compact Hopf map is given by

1 H2,1 S H2, (41) −→ and, in the same sense of the Bloch sphere, we refer to the base-manifold H2 as the Bloch hyperboloid. 2,1 H is a group manifold of the SU(1, 1) whose invariant matrix is σz. The SU(1, 1) matrix generators can be realized as the following non-Hermitian matrices 1 1 1 1 τ i = i σ ,i σ , σ , (42) 2 { 2 x 2 y 2 z} which satisfy τ i, τ j =2ηij , [τ i, τ j ]= 2iǫijkτ , (43) { } − k with η = diag( , , +), ǫ123 =1. (44) ij − − 2 2 2,1 With two-component spinor subject to ψ†σ ψ = ψ ψ = 1 : H , the non-compact Hopf map is z | 1| − | 2| explicitly realized as i i ψ x = ψ†κ ψ, (45) → where κi are Hermitian matrices made of τ i:

κi = σ τ i = σ , σ , 1 . (46) z {− y x 2} xi (45) satisfy i j i 1 1 2 2 3 3 2 η x x = x x = x x x x + x x = (ψ†κψ) =1, (47) ij i − − and denote the coordinates on H2. When xi are parameterized as

x1 = sinh ρ sin φ, x2 = sinh ρ cos φ, x3 = cosh ρ, (48)

4 (S) Interestingly, D-matrices themselves have a physical meaning; D (φ,θ,χ) ′ denote 2S + 1 degenerate eigenstates of Sz ,Sz (S Sz )th Landau level in the monopole background with magnetic charge Sz (see [44] for instance). −| |

7 ψ can be realized as ρ i φ cosh( ) e 2 χ 2 i 2 ψ(φ,ρ,χ)= φ e , (49) ρ i 2 sinh( 2 ) e− ! which also acts as the pseudo-spin-coherent state [34, 42, 43] (see the right of Fig.1):

i (x (ρ, φ) τi)ψ = ψ. (50)

ψ can be derived by acting the SU(1, 1) rotation to the pseudo-spin state aligned to z-direction:

1 ψ(φ,ρ,χ)= D(1/2)(φ,ρ,χ) , (51) 0 where D(1/2)(φ,ρ,χ) is the SU(1, 1) version of the D-matrix5

ρ i 1 (χ+φ) ρ i 1 (χ φ) φ z ρ x χ z cosh( ) e 2 sinh( ) e 2 (1/2) i 2 τ i 2 τ i 2 τ 2 2 − − D (φ,ρ,χ)= e e− e = 1 1 . (53) ρ i 2 (χ φ) ρ i 2 (χ+φ) sinh( 2 ) e − cosh( 2 ) e− ! The ranges of the parameters are

ρ = [0, ), φ = [0, 2π), χ = [0, 4π). (54) ∞ With the Schwinger boson operators, we can construct the SU(1, 1) operators as

i i Tˆ = ψˆ†κ ψ.ˆ (55)

Here, ψˆ is deﬁned as a ψˆ , (56) ≡ b † whose components satisfy ˆ ˆ ˆ ˆ [ψα, ψβ† ] = (σz)αβ, [ψα, ψβ]=0. (57) The SU(1, 1) Hermitian operators are represented as [45, 46, 47]

x 1 y 1 z 1 1 Tˆ = i (a†b† ab), Tˆ = (a†b† + ab), Tˆ = (a†a + b†b)+ , (58) 2 − 2 2 2 and the ladder operators are + Tˆ = a†b†, Tˆ− = ab. (59) The SO(2, 1) Casimir is derived as

i x 2 y 2 z 2 1 C = Tˆ Tˆ = (Tˆ ) (Tˆ ) + (Tˆ ) = (ψˆ†σ ψˆ) (ψˆ†σ ψˆ + 2), (60) i − − 4 z z where ψˆ†σ ψˆ = a†a bb† = a†a b†b 1. (61) z − − − The SU(1, 1) D-operator is given by

iφTˆz iρTˆx iχTˆz Dˆ(φ,ρ,χ)= e e− e . (62)

Under the SU(1, 1) transformation of Dˆ(φ,ρ,χ), ψˆ transforms as an SU(1, 1) spinor

(1/2) Dˆ(φ,ρ,χ)† ψˆ Dˆ(φ,ρ,χ)= D (φ,ρ,χ) ψ.ˆ (63) 5One may explicitly verify that (53) satisﬁes the conditions of the SU(1, 1) group: (1/2) † (1/2) (1/2) D (φ,ρ,χ) σz D (φ,ρ,χ)= σz, det(D (φ,ρ,χ)) = 1. (52)

8 2.2.2 SU(1, 1) coset representative and the squeeze operator In the parameterization y (φ) = sin φ, y (φ) = cos φ S1, (64) 1 2 ∈ an SU(1, 1) coset representative for H2 SU(1, 1)/U(1) is obtained as ≃ ρ ρ iφ ρ m n ρ iφ + ρ −iφ − cosh( ) sinh( ) e i 2 ǫmny (φ) τ 2 e t + 2 e t 2 2 M(ρ, φ)= e = e− = ρ iφ − ρ , (65) sinh( ) e− cosh( ) − 2 2 with ǫ = ǫ = 1 and 12 − 21

+ 1 x 1 y 0 1 1 x 1 y 0 0 t = i τ + τ = , t− = i τ + τ = . (66) − 2 2 0 0 2 2 1 0 − (65) can be factorized as

i φ τ z i ρ τ x i φ τ z (1/2) Euler decomposition : M(ρ, φ)= e 2 e 2 e− 2 = D (φ, ρ, φ), (67) − − and

tanh( ρ ) eiφ t+ ln(cosh ρ )τ z tanh( ρ ) e−iφ t− Gauss decomposition : M(ρ, φ)= e− 2 e− 2 e 2 ρ m 1 ρ ρ m 1 tanh y ( τm iτm3) log(cosh ) τz tanh y ( τm+iτm3) = e− 2 2 − e− 2 e 2 2 , (68)

where 1 1 τ ij = i [τ i, τ j ]= ǫijkτ . (69) − 4 −2 k ρ m In the Gauss decomposition (68), the Poincar´eDisc coordinates, tanh 2 y (φ), appear on the shoulder of the exponent. The operator that corresponds to the coset representative (65) is constructed as

ρ m n i ǫmny (φ) Tˆ ρ iφ ρ iφ Sˆ(ρ, φ)= e 2 = exp e a†b† + e− ab . (70) − 2 2 Notice that (70) is nothing but the two-mode squeeze operator in quantum optics [2, 3, 4]. Thus, the SU(1, 1) D-operator realizes the squeeze operator in its special form. Since the SU(1, 1) D-operator sig- niﬁes hyperbolic rotation, the squeeze operator can be interpreted as a hyperbolic rotation operator. As exempliﬁed in the above, (70) is factorized as

iφTˆz iρTˆx iφTˆz Euler decomposition : Sˆ(ρ, φ)= Dˆ(φ, ρ, φ)= e e e− − − φ ρ † † φ i (ˆna+ˆn ) (a b +ab) i (ˆna+ˆn ) = e 2 b e− 2 e− 2 b , (71)

and

tanh ρ eiφ Tˆ+ 2 ln(cosh ρ )Tˆz tanh ρ e−iφ Tˆ− Gauss decomposition : Sˆ(ρ, φ)= e− 2 e− 2 e 2

nˆa+ˆnb 1 tanh( ρ )eiφ a†b† 1 tanh( ρ )e−iφ ab = e− 2 e 2 . (72) cosh ρ cosh( ρ ) 2 2 The Euler decomposition (71) demonstrates the physical meaning of the squeeze operation, i.e., the SU(1, 1) hyperbolic rotation. Meanwhile, the Gauss decomposition is very useful in deriving the number state expansion of squeezed state [47], as (72) is realized as an almost normal ordered form. The phase φ can be absorbed in the phase redeﬁnition of the Schwinger boson operators6

iα iβ a e− a, b e− b, (73) → → 6The commutation relations (26) are immune to the phase redeﬁnition of each Schwinger operator (73).

9 the squeeze operator is transformed as

ρ iφ ρ iφ ρ ρ Sˆ(ρ, φ) = exp e a†b† + e− ab S(ρ) = exp a†b† + ab , (74) − 2 2 → − 2 2 where 1 1 α = φ + γ, β = φ γ. (75) 2 2 − (The degree of freedom of γ still remains.)

2.3 Two-mode SO(2, 1) squeeze vacuum and entangled spin-coherent states 2.3.1 SO(2, 1) two-mode squeeze operator and the interaction Hamiltonian With the Gauss decomposition (72), the two-mode squeezed vacuum7 is readily obtained as

n 1 ∞ ρ sq(ρ, φ) = Sˆ(ρ, φ) 0 0 = tanh( ) einφ n n , (76) | i | i ⊗ | i cosh ρ − 2 | i ⊗ | i 2 n=0 X where 1 n 1 n n n = a† 0 b† 0 . (77) | i ⊗ | i √n! | i⊗ √n! | i sq(ρ, φ) signiﬁes an entangled state of the number states. Removing the right U(1) factor of (71), we | i introduce the Schwinger-type squeeze operator as [23]

ˆz ˆx ˆ(ρ, φ)= eiφT eiρT . (78) S The diﬀerence between the squeezed vacuum of the Dirac-type and that of the Schwinger-type is just the phase factor: i φ sq(ρ, φ) = Sˆ(ρ, φ) 0 0 = e− 2 ˆ(ρ, φ) 0 0 , (79) | i | i ⊗ | i S | i ⊗ | i and so the Dirac-type and Schwinger-type squeezed vacua are physically identical. In the interaction picture, the two-mode Hamiltonian realizes as the time-evolution operator of the squeeze operator (70), which is int ρ iφ ρ iφ Hˆ (ρ, φ)= i e a†b† + i e− ab. (80) − 2 2

2.3.2 Two two-mode squeezed vacua as entangled spin-coherent states For later convenience, we point out an interesting correspondence between the SU(2) spin-coherent states and the SU(1, 1) squeezed states. Let us consider a direct product of two independent squeezed vacua with same squeeze parameter ρ. We also assume that a squeezed vacuum is realized in the Schwinger boson space of a and c, and another in the Schwinger boson space of b and d:8

1 ρ iφ † † ρ iφ′ † † 2 e a c 2 e b d sq(ρ, φ) a,c sq(ρ, φ′) b,d = 2 ρ e− e− 0, 0, 0, 0 | i ⊗ | i cosh 2 | i

1 ∞ ρ ′ = ( tanh( ))n+mei(nφ+mφ ) n,m,n,m , (81) 2 ρ − 2 | i cosh 2 n,m=0 X 7See Appendix A about basic properties of the two-mode squeezed vacuum. 8(81) is a four-mode analogue of the two-mode SU(1, 1) SU(1, 1) state of [48]. ⊗

10 where n,m,n,m n m n m . Recall that the Schwinger interpretation of spin states, in which | i ≡ | ia| ib| ic| id n,m is interpreted as the spin state S,S , and the sum of all particles numbers is replaced with the sum | i | zi of all possible spin degrees of freedom (see Fig.2):

S . (82) → n,m=0,1,2,3, S=0,1/2,1,3/2, Sz = S X ··· X ··· X− With this replacement, (81) can be rewritten as

S ∞ ′ ′ 1 ρ 2S iS(φ+φ ) iSz(φ φ ) sq(ρ, φ) a,c sq(ρ, φ′) b,d = ( tanh( )) e e − S,Sz a,b S,Sz c,d, | i ⊗| i cosh2( ρ ) − 2 | i ⊗| i 2 S=0,1/2,1, Sz= S X ··· X− (83) which denotes a linear combination of direct products of two identical spin states with all possible spins. In each spin sector, we have a maximally entangled spin state:

S ′ iSz(φ φ ) e − S,S S,S . (84) | zia,b ⊗ | zic,d Sz= S X− Though in (83) the left and right-hand sides are mathematically identical thorough the Schwinger operator formalism, they describe totally diﬀerent physics: The left-hand side denotes a separable state of two squeezed states, while the right-hand side stands for a maximally entangled spin state.

3 Bloch four-hyperboloid and SO(4, 1) squeeze operator

In this section, we exploit the four-hyperboloid geometry (Fig.3) and the SO(4, 1) coset representative. The geometric structures of H2 and H4 are respectively given by

H2 R2 S1 R , H4 R4 S3 R . (85) ≃ ≃ ⊗ + ≃ ≃ ⊗ + H4 can be regarded as a natural generalization of H2 whose latitude is expanded from S1 to S3. Such an S3 structure brings an internal spin structure to the SO(4, 1) formulation. Adopting a set of four Schwinger operators as an SO(4, 1) representation, we derive an SO(4, 1) squeeze operator and discuss its internal spin structure.

3.1 SO(4, 1) algebra and Bloch four-hyperboloid 3.1.1 H4 and the SO(4, 1) algebra The four-hyperboloid H4 is given by

5 4 η xaxb = xmxm + x5x5 = 1 (86) ab − m=1 a,bX=1 X with ηab = η = diag( , , , , +). (87) ab − − − − As a coset, H4 is represented as H4 SO(4, 1)/SO(4). (88) ≃

11 Figure 3: 2D hyperboloid H2 (left), 4D hyperboloid H4 (middle) and its S3-latitude (right)

The SO(4, 1) gamma matrices γa (a =1, 2, 3, 4, 5) are introduced so as to satisfy

γa,γb =2ηab, (89) { } and the SO(4, 1) generators Σab are constructed as 1 Σab = i [γa,γb], (90) − 4 which satisfy [Σab, Σcd]= iηacΣbd iηadΣbc + iηbdΣac iηbcΣad. (91) − − Hereafter, we take the SO(4, 1) gamma matrices and the generators as

0 q¯m 1 0 γm = − , γ5 = 2 , (92) qm 0 0 1 − 2 with qm being quaternions, qm = iσ , 1 , q¯m = iσ , 1 , (93) {− i 2} { i 2} and then9 1 ηi σ 0 1 0q ¯m Σmn = mn i , Σm5 = i . (97) −2 0η ¯i σ − 2 qm 0 mn i 9The independent generators of Σab are

m 1 σi 0 1 0 0 1 0q ¯ , , i m . (94) − 2 0 0 − 2 0 σi − 2 q 0 Among Σab (97), Σmn are Hermitian matrices, while Σm5 are anti-Hermitian matrices. iΣab (97) are recognized as the U(1, 1; H) generators, since their components are quaternions

m ab 1 qi 0 1 0 0 1 0q ¯ iΣ = , , m , (95) { 2 0 0 2 0 qi 2 q 0 } and they satisfy 1 0 1 0 (iΣab)† = iΣab. (96) 0 1 − 0 1 − −

12 i i Here, ηmn andη ¯mn are the ’t Hooft symbols:

ηi = ǫmni4 + δmiδn4 δm4δni, η¯i = ǫmni4 δmiδn4 + δm4δni, (98) mn − mn − with ǫ1234 = 1. Notice that γa and Σab amount to ﬁfteen matrices that satisfy the SO(4, 2) SU(2, 2) ≃ algebra. With the SU(2, 2) invariant matrix k = γ5, (99) we can “hermitianize” the SU(2, 2) generators as

ka kγa, kab = kΣab, (100) ≡ or 0 q¯m 0 iσ 0 1 1 0 km = − = − i , − 2 , k5 = 2 . (101) qm 0 iσ 0 1 0 0 1 − i − 2 2 Also note that neither the Hermitian matrices ka satisfy (89) nor kab satisfy the SO(4, 1) algebra (91).

3.1.2 Bloch four-hyperboloid and SO(4, 1) coset representative The 2nd non-compact (hybrid) Hopf map is [25]

3 H4,3 S H4, (102) −→ which is explicitly given by a a Ψ x =Ψ†k Ψ. (103) → t Here Ψ = (Ψ1 Ψ2 Ψ3 Ψ4) denotes a four-component spinor subject to

2 2 2 2 4,3 Ψ† k Ψ= Ψ + Ψ Ψ Ψ = 1 : H , (104) | 1| | 2| − | 3| − | 4| and xa (103) automatically satisfy

4 a b m m 5 5 2 4 η x x = x x + x x = (Ψ†kΨ) = 1 : H . (105) ab − m=1 X We can realize Ψ as 1 (1 + x5) 1 Ψ= 2 ψ (106) 5 xmq 2(1 + x ) m t with two-component spinor ψ = (ψ1 ψ2) psubject to

3 ψ†ψ = 1 : S . (107)

With polar angle coordinate representation10

x1 = sinh ρ sin χ sin θ cos φ, x2 = sinh ρ sin χ sin θ sin φ, x3 = sinh ρ sin χ cos θ, x4 = sinh ρ cos χ, x5 = cosh ρ, (109)

10The ranges of the parameters are

0 ρ< , 0 χ, θ π, 0 φ < 2π. (108) ≤ ∞ ≤ ≤ ≤

13 (106) is represented as cosh( ρ ) 1 Ψ= 2 2 ψ. (110) sinh( ρ ) ymq 2 m m 4 m m 3 4 Here, y ( m=1 y y = 1) denote coordinates on the normalized S -“circle” at latitude ρ on H (see Fig.3): P 1 ym = xm = sin χ sin θ cos φ, sin χ sin θ sin φ, sin χ cos θ, cos χ , (111) sinh ρ { } and so

4 4 iφ 4 m m m (cos χ i sin χ cos θ) i sin χ sin θ e− m y q = y q = − − = ( y q¯ )†. (112) m − i sin χ sin θ eiφ (cos χ + i sin χ cos θ) m m=1 m=1 m=1 X X − X Notice that Ψ acts as the SO(4, 1) pseudo-spin-coherent state on H4:

a (x γa)Ψ=Ψ, (113)

which suggests ψ Ψ= M(ρ,χ,θ,φ) , (114) 0 where M(ρ,χ,θ,φ) denotes an SO(4, 1) rotation11

ρ ρ 4 m cosh( ) 12 sinh( ) y q¯m iρ P4 ymΣ M(ρ,χ,θ,φ)= 2 2 m=1 = e m=1 m5 , (116) sinh( ρ ) 4 ymq cosh( ρ ) 1 2 m=1 m P 2 2 ! with P 1 0q ¯m Σ = Σm5 = i . (117) m5 − 2 qm 0 The last expression of (116) implies M(ρ,χ,θ,φ) is a coset representative of SO(4, 1) group, since Σm5 act as the broken generators associated with the symmetry breaking, SO(4, 1) SO(4). Generalizing the → SO(5) coset matrix for S4 [53], the Euler decomposition of (116) is given by

iρΣ45 M(ρ,χ,θ,φ)= H(χ,θ,φ)† e− H(χ,θ,φ). (118) · · Here, H(χ,θ,φ) denotes an SU(2) matrix of the form

iχΣ34 iθΣ13 iφΣ12 HL(χ,θ,φ) 0 H(χ,θ,φ)= e e e− = , (119) · · 0 H (χ,θ,φ) R with

θ i 1 (χ φ) θ i 1 (χ+φ) χ θ φ cos( ) e− 2 − sin( ) e− 2 i 2 σ3 i 2 σ2 i 2 σ3 2 2 HL(χ,θ,φ)= HR( χ,θ,φ)= e− e e = θ i 1 (χ+φ) θ i 1 (χ φ) . (120) − · · sin( ) e 2 cos( ) e 2 − − 2 2 ! 11In the polar coordinate representation, (116) is expressed as M(ρ,χ,θ,φ)= ρ ρ ρ −iφ cosh( 2 ) 0 sinh( 2 )(cos χ + i sin χ cos θ) i sinh( 2 )sin χ sin θ e ρ − ρ − ρ  0 cosh( ) i sinh( ) sin χ sin θ eiφ sinh( )(cos χ i sin χ cos θ) 2 − 2 − 2 − . sinh( ρ )(cos χ i sin χ cos θ) i sinh( ρ )sin χ sin θ e−iφ cosh( ρ ) 0 − 2 − 2 2   i sinh( ρ )sin χ sin θ eiφ sinh( ρ )(cos χ + i sin χ cos θ)) 0 cosh( ρ )   2 − 2 2  (115)

14 Obviously (118) is a natural generalization of the SU(1, 1) case (67) with replacement of the U(1) factor i φ τ z e− 2 with the SU(2) factor H(χ,θ,φ) (119). The Gauss decomposition is also given by

ρ 0 ymq¯ ρ 1 0 ρ 0 0 M(ρ,χ,θ,φ) = exp tanh( ) m exp ln(cosh( )) 2 exp tanh( ) 2 0 0 · − 2 0 1 · 2 ymq 0 − 2 m ρ 1 ρ ρ 1 = exp tanh( ) ym ( γ iΣ ) exp ln(cosh( )) γ exp tanh( ) ym ( γ + iΣ ) , − 2 · 2 m − m5 · − 2 5 · 2 · 2 m m5 (121) which realizes a natural generalization of (68), again.

3.2 SO(4, 1) squeeze operator 3.2.1 SO(4, 1) Schwinger operator realization and SO(4, 2) operators As the SO(4, 1) group does not accommodate the Majorana representation,12 two-mode SO(4, 1) squeeze operator does not exist unlike the SO(2, 3) case [23]. We then consider a set of four Schwinger operators in correspondence with the Dirac representation:

a b Ψ=ˆ   (126) c†   d†   whose components satisfy ˆ ˆ ˆ ˆ [Ψα, Ψβ† ]= kαβ , [Ψα, Ψβ]=0. (127) Sandwiching (100) with (126), we introduce the following Hermitian operators:

a a ab ab Xˆ = Ψˆ †k Ψˆ , Xˆ = Ψˆ †k Ψˆ . (128)

Xˆ a are explicitly

1 2 Xˆ = i(a†d† + b†c† ad bc), Xˆ = (a†d† b†c† + ad bc), − − − − − − 3 4 Xˆ = i(a†c† b†d† ac + bd), Xˆ = (a†c† + b†d† + ac + bd), − − − − 5 Xˆ = a†a + b†b + c†c + d†d +2, (129)

12The SO(4, 1) gamma matrices (92) and generators (94) satisfy

(γa)∗ = CγaC−1, (Σab)∗ = CΣabC−1, (122) − where C denotes the charge conjugation matrix 0 1 0 0  10 0 0 C = − . (123) 0 0 0 1    0 0 1 0  −  With C (123) the Majorana condition is imposed as

∗ ΨM = CΨM, (124) and so (Ψ∗ )∗ = C∗CΨ = Ψ . (125) M M − M ∗ ∗ Under the usual deﬁnition of the complex conjugation (Ψ ) = Ψ, (125) implies ΨM = 0, which means that the Majorana spinor does not exist in the SO(4, 1) group.

15 Xˆ ab consist of the SO(4) particle conserving operators Xˆ mn

12 1 13 1 Xˆ = (a†a b†b c†c + d†d), Xˆ = i ( a†b + b†a c†d + d†c), −2 − − 2 − − 14 1 23 1 Xˆ = (a†b + b†a + c†d + d†c), Xˆ = (a†b + b†a c†d d†c), −2 2 − − 24 1 34 1 Xˆ = i (a†b b†a c†d + d†c), Xˆ = (a†a b†b + c†c d†d), (130) − 2 − − −2 − − and the remaining particle number non-conserving operators Xˆ m5

15 1 25 1 Xˆ = (a†d† + b†c† + ad + bc), Xˆ = i (a†d† b†c† ad + bc), 2 − 2 − − 35 1 45 1 Xˆ = (a†c† b†d† + ac bd), Xˆ = i (a†c† + b†d† ac bd). (131) 2 − − − 2 − − Notice that the Hermitian matrices kab (100) do not satisfy the SO(4, 1) algebra, but the corresponding Hermitian operators Xˆ ab do satisfy the SO(4, 1) algebra. SO(4, 1) invariant operators are constructed as

a 1 ab Xˆ Xˆ = Xˆ Xˆ = (Ψˆ †kΨ)(ˆ Ψˆ †kΨ+4)ˆ , (132) a 2 ab Xa

1 na nb nc nd na,nb,nc,nd = (a†) (b†) (c†) (d†) 0, 0, 0, 0 . (137) | i √na!nb!nc!nd! | i

Obviously, the particle numbers, na, nb, nc and nd, uniquely specify the eigenstates.

3.2.2 Subgroup decomposition of the SO(4, 1) squeeze operator From a view of the SO(4, 2) group theory, we delve into internal spin and pseudo-spin structures. The SO(4, 2) group accommodates two important subgroups, the SO(4) group for spins and the SO(2, 2) group for pseudo-spins:

SO(4) SU(2) SU(2) SO(4, 1) SO(4, 2), (138a) ≃ L ⊗ R ⊂ ⊂ SO(2, 2) SU(1, 1) SU(1, 1) SO(2, 3) SO(4, 2). (138b) ≃ T ⊗ B ⊂ ⊂ 13For oscillator realization of the SU(2, 2) group and its representation, one may consult Refs.[54, 55, 56, 57, 58, 59, 60].

16 For SO(4) SU(2) SU(2) , the corresponding left and right SU(2) spin operators are constructed as ≃ L ⊗ R 1 a † 1 a 1 c † 1 c Lˆ = ηi Xˆ mn = σ , Rˆ = η¯i Xˆ mn = σ , (139) i 4 mn b 2 i b i 4 mn d 2 i d which satisfy 3 3 LîLî = Lˆ(Lˆ + 1), RîRî = Rˆ(Rˆ +1) (140) i=1 i=1 X X with 1 1 Lˆ = (ˆn +ˆn ), Rˆ = (ˆn +ˆn ). (141) 2 a b 2 c d The SU(2, 2) eigenstate (137) can be specified by the spin group quantum numbers, 1 1 1 1 Lˆ = (ˆn +ˆn ), Rˆ = (ˆn +ˆn ), Lˆ = (ˆn nˆ ), Rˆ = (ˆn nˆ ). (142) 2 a b 2 c d z 2 a − b z 2 c − d Meanwhile for SO(2, 2) SU(1, 1) SU(1, 1) , the corresponding top and bottom SU(1, 1) pseudo-spin ≃ T ⊗ B operators are introduced as

a † 1 a b † 1 b Tˆi = κi , Bˆi = κi , (143) c 2 c d 2 d † † † † which satisfy 3 3 i j i j ηij Tˆ Tˆ = Tˆ(Tˆ + 1), ηij Bˆ Bˆ = Bˆ(Bˆ + 1), (144) i,j=1 i=1 X X with η = diag( , , +) and ij − − 1 1 Tˆ = (ˆn nˆ 1), Bˆ = (ˆn nˆ 1). (145) 2 a − c − 2 b − d − In detail,

x 1 1 45 1 3 y 1 1 35 1 4 Tˆ = i (a†c† ac)= X X , Tˆ = (a†c† + ac)= Xˆ Xˆ , 2 − 2 − 4 2 −2 − 4 z 1 1 1 34 1 5 x 1 1 45 1 3 Tˆ = (a†a + c†c)+ = Xˆ + Xˆ , Bˆ = i (b†d† bd)= X + X , 2 2 2 4 2 − 2 4 y 1 1 35 1 4 z 1 1 1 34 1 5 Bˆ = (b†d† + bd)= Xˆ Xˆ , Bˆ = (b†b + d†d)+ = Xˆ + Xˆ . (146) 2 2 − 4 2 2 −2 4 The SU(2, 2) eigenstate (137) can be speciﬁed by the pseudo-spin quantum numbers of Tˆ, Bˆ, Tˆz and Bˆz.

3.2.3 SO(4, 1) squeeze operator and the interaction Hamiltonian The existence of spin and pseudo-spin groups in the SO(4, 1) group implies that the SO(4, 1) formalism is large enough to incorporate both spin and squeezed state structures. Replacing the non-Hermitian matrices in (116) with the corresponding Hermitian operators, we introduce the SO(4, 1) squeeze operator:

4 m 4 m m5 iρ P y (χ,θ,φ)Xˆm5 iρ P y (χ,θ,φ)Xˆ Sˆ(ρ,χ,θ,φ)= e m=1 = e− m=1 . (147) The interaction Hamiltonian is then given by

4 Hˆ int(ρ,χ,θ,φ)= ρ ymXˆ − m5 m=1 X ρ iφ iφ = sin χ (cos θ(a†c† b†d†) + sin θ(e− a†d† + e b†c†)) i cos χ (a†c† + b†d†) + h.c. (148) 2 − − 

17 The SO(4, 1) squeeze operator is a unitary operator that satisﬁes

1 Sˆ(ρ,χ,θ,φ)† = Sˆ(ρ,χ,θ,φ)− = Sˆ( ρ,χ,θ,φ). (149) − In (148), we can easily see that the phase φ can be absorbed in the phase redeﬁnition of the Schwinger operators: iα iβ iα iβ a e− a, b e b, c e c, d e− d, (150) → → → → where α + β = φ. (151) To grasp physical meaning of the SO(4, 1) squeeze operator (147), we rewrite the SO(4, 1) squeeze operator in the language of the SU(2) spin and the SU(1, 1) pseudo-spin groups. For this purpose, we apply the Euler decomposition (118) to the squeeze operator (147): 14

Sˆ(ρ,χ,θ,φ)= Hˆ (χ,θ,φ)† Sˆ(ρ) Hˆ (χ,θ,φ), (154) · · where iχXˆ 34 iθXˆ 13 iφXˆ 12 Hˆ (χ,θ,φ) e e e− = Dˆ (χ, θ, φ) Dˆ (χ, θ, φ) (155) ≡ · · L − − · R − and iρXˆ 45 Sˆ(ρ) e− = Sˆ (ρ) Sˆ (ρ). (156) ≡ T · B Here,

iχLˆz iθLˆ y iφLˆz iχRˆz iθRˆy iφRˆz iρTˆx iρBˆx Dˆ (χ,θ,φ) e− e− e− , Dˆ (χ,θ,φ) e− e− e− , Sˆ (ρ) e , Sˆ (ρ) e . L ≡ R ≡ T ≡ B ≡ (157) Consequently, the SO(4, 1) squeeze operator can be expressed as15

Sˆ(ρ,χ,θ,φ)= Dˆ (φ, θ, χ) Dˆ ( φ, θ, χ) Sˆ (ρ) Sˆ (ρ) Dˆ (χ, θ, φ) Dˆ (χ, θ, φ), (158) L − · R − − · T · B · L − − · R − Removing the right SU(2) factor of (154), we introduce the Schwinger-type SO(4, 1) squeeze operator:16

ˆ(ρ,χ,θ,φ)= Hˆ (χ,θ,φ)† Sˆ(ρ)= Dˆ (φ, θ, χ) Dˆ ( φ, θ, χ) Sˆ (ρ) Sˆ (ρ). (160) S · L − · R − − · T · B 14One may of course utilize the Gauss decomposition of Sˆ(ρ,χ,θ,φ) (121) in deriving the number state expansion of the SO(4, 1) squeezed vacuum:

ρ m 1 ρ ρ m 1 Sˆ(ρ,χ,θ,φ) = exp tanh( ) y ( Xˆm iXˆm5) exp log(cosh( ) Xˆ5) exp tanh( ) y ( Xˆm + iXˆm5) , (152) − 2 2 − · − 2 · 2 2 where

m 1 −iφ † † iφ † † † † † † y ( Xˆm iXˆm5)= i sin χ sin θe a d + i sin χ sin θe b c + (cos χ + i sin χ cos θ)a c + (cos χ i sin χ cos θ)b d , 2 − −

Xˆ5 =n ˆa +n ˆb +n ˆc +n ˆd + 2. (153) In the main text, we however adopted the Euler decomposition as it manifests a clear physical meaning of the SO(4, 1) squeezed vacuum. 15We used Dˆ(χ,θ,φ)† = Dˆ( φ, θ, χ) in the derivation of (158). − − − 16The corresponding squeeze matrix of the Schwinger-type is given by

45 (ρ,χ,θ,φ)= H(χ,θ,φ)† e−iρΣ M · 1 1 1 1 ρ θ i 2 (χ−φ) ρ θ −i 2 (χ+φ) ρ θ i 2 (χ−φ) ρ θ −i 2 (χ+φ) cosh 2 cos 2 e cosh 2 sin 2 e sinh 2 cos 2 e sinh 2 sin 2 e 1 − 1 − 1 1  ρ θ i 2 (χ+φ) ρ θ −i 2 (χ−φ) ρ θ i 2 (χ+φ) ρ θ −i 2 (χ−φ) cosh 2 sin 2 e cosh 2 cos 2 e sinh 2 sin 2 e sinh 2 cos 2 e = 1 1 − 1 − 1 .  ρ θ −i 2 (χ+φ) ρ θ i 2 (χ−φ) ρ θ −i 2 (χ+φ) ρ θ i 2 (χ−φ)   sinh 2 cos 2 e sinh 2 sin 2 e cosh 2 cos 2 e cosh 2 sin 2 e  − ρ − 1 − ρ 1 ρ − 1 − − ρ 1   sinh sin θ e i 2 (χ φ) sinh cos θ ei 2 (χ+φ) cosh sin θ e i 2 (χ φ) cosh cos θ ei 2 (χ+φ)  − 2 2 − 2 2 2 2 2 2  (159)

18 4 SO(4, 1) squeezed vacuum and its basic properties

Armed with the spin and pseudo-spin based expression of the SO(4, 1) squeeze operator, we now derive a mathematical expression of the SO(4, 1) squeezed vacuum and discuss its physical properties.

4.1 SO(4, 1) squeezed vacuum 4.1.1 Spin interpretation We simply construct the SO(4, 1) squeezed vacuum by applying the SO(4, 1) squeeze operator to the vacuum: Sq(ρ,χ,θ,φ) = Sˆ(ρ,χ,θ,φ) 0, 0, 0, 0 . (161) | i | i The squeezed vacuum is identical for the Dirac-type Sˆ and the Schwinger-type ˆ: S Sq(ρ,χ,θ,φ) = Dˆ (φ, θ, χ) Dˆ ( φ, θ, χ) Sˆ (ρ) Sˆ (ρ) 0, 0, 0, 0 | i L − · R − − · T · B | i = ˆ(ρ,χ,θ,φ) 0, 0, 0, 0 . (162) S | i This is a property similar to the original SO(2, 1) case, in which the squeezed vacua in the Dirac-type and the Schwinger-type are physically equivalent (79). (In the case of the Sp(4; R) SO(2, 3) squeezed ≃ states [23], the Dirac-type and Schwinger-type squeezed vacua are physically diﬀerent.) Thus, the SO(4, 1) squeezed vacuum inherits the uniqueness of the squeezed vacuum. Expanding the two squeeze operators, SˆT and SˆB, in terms of the number operators, we can express the squeezed vacuum (162) as

n+m 1 ∞ ρ Sq(ρ,χ,θ,φ) = tanh( ) Dˆ (φ, θ, χ)Dˆ ( φ, θ, χ) n,m n,m . (163) | i 2 ρ − 2 L − R − − | i ⊗ | i cosh ( 2 ) n,m=0 X With the spin interpretation (32), we can rewrite (163) as

Here, we changed the summation from n and m to S and Sz according to (82). Thus, the SO(4, 1) four- mode squeezed vacuum can be understood as a linear combination of two-body spin-coherent states. In (164) the first spin direction is specified by (θ, φ) while the second spin direction is specified by (θ, φ), − and then, by tuning θ and φ, we can manipulate relative configurations of two spins.17 Extracting the U(1) phase factor of the spin state iχSz S,S (φ,θ,χ) = e− S,S (θ, φ) , (165) | z i | z i we can express (164) as

2S S 1 ρ 2iχSz Sq(ρ,χ,θ,φ) = tanh( ) e S,Sz(θ, φ) S,Sz(θ, φ) , (166) | i cosh2( ρ ) − 2 | i ⊗ | − i 2 S=0,1/2,1,3/2, Sz= S X ··· X− 17In the four-mode interpretation, due to the independent phase choice of the four-mode operators, the angle φ can be “gauged” away (151), but in the spin interpretation, only the overall phase of a and b and that of c and d are allowed to change, since the angular momentum operators (139) are just immune to overall phase redeﬁnition of these operators. Thus in the spin interpretation, φ cannot be gauged away and is indeed related the relative direction of two spins.

19 where S,S (θ, φ) Dˆ(φ, θ, 0) S,S . (167) | z i≡ | i One may ﬁnd that (166) naturally generalizes the number-state expansion of the original squeezed vacuum (76) [Fig.4].

Figure 4: The original SO(2, 1) squeezed vacuum is an entangled state of inﬁnite (separable) pairs of number-states. The SO(4, 1) squeezed vacuum is an entangled state of inﬁnite (maximally entangled) spin-pairs.

4.1.2 Spin entanglement We can represent (164) as

1 ρ 2S Sq (ρ,χ,θ,φ) = √2S +1 tanh( ) S S (χ,θ,φ) , (168) | i cosh2( ρ ) − 2 | ⊗ ii 2 S=0,1/2,1,3/2, X ··· where S S (χ,θ,φ) denotes a (normalized) spin-pair state: | ⊗ i 1 S S S (χ,θ,φ) S,Sz(φ, θ, χ) S,Sz( φ, θ, χ) | ⊗ ii ≡ √2S +1 | − i ⊗ | − − i Sz= S X− S S S 1 (S) (S) = D (φ, θ, χ)L ,S D ( φ,θ,χ)R ,S S,Lz S, Rz √2S +1 − z z − z z | i ⊗ | i Sz= S Lz= S Rz= S X− X− X− S S 1 (S) (S) = (D (φ, θ, χ) D ( χ, θ, φ))L ,R S,Lz S, Rz . √2S +1 − − − − z z | i ⊗ | i Lz= S Rz= S X− X− (169) In the last equation, we used the property of the D-matrix: D(S)(χ,θ,φ)t = D(S)(φ, θ,χ). (170) −

20 (168) exhibits a duplicate entanglement which we refer to as the hybrid entanglement: 1. An inﬁnite number of spin-pairs are entangled with the squeeze parameter ρ. This manifests a natural generalization of the entanglement of the original SO(2, 1) squeezed vacuum.

2. In each spin-sector, two spins are maximally entangled. This is a unique feature in the SO(4, 1) case. In the original SO(2, 1) squeezed vacuum, the corresponding state is simply a product state, n n . | i⊗| i The spin part (169) can be expressed as a superposition of the integer total spin states only:

I S S =0 1 2 2S. (171) ≡ ⊗ ⊕ ⊕ ⊕···⊕ This is a generalization of the fact that even total number states n n only appear in the expansion | i ⊗ | i j,jz of the two-mode SO(2, 1) squeezed vacuum. Using the Clebsch-Gordan coeﬃcient Cj1 ,m1;j2,m2 , we can construct total integer spin states

S S I,Iz I, Iz = C S,Lz S, Rz , (172) | ii S,Lz;S,Rz | i ⊗ | i Lz= S Rz= S X− X− and (169) is expressed as

2S I 1 I,Iz S S (χ,θ,φ) = C (χ,θ,φ) I, Iz (173) | ⊗ ii √2S +1 S | ii I=0 Iz = I X X− where S S I,Iz I,Iz (S) (S) C (χ,θ,φ) C (D (φ, θ, χ) D ( χ, θ, φ))L ,R . (174) S ≡ S,Lz;S,Rz − − − − z z Lz= S Rz= S X− X− Consequently, the SO(4, 1) squeezed vacuum (168) is actually expressed as a linear combination of an inﬁnite number of spin-pairs with total integer spins:

I Sq (ρ,χ,θ,φ) = CI,Iz (ρ,χ,θ,φ) I, I , (175) | i | zii I=0,1,2, Iz = I X ··· X− with 1 ρ CI,Iz (ρ,χ,θ,φ) ( tanh( ))2S CI,Iz (χ,θ,φ). (176) ≡ cosh2( ρ ) − 2 S 2 S= I , I +1, I +2, 2 2X2 ··· ρ Up to O(tanh( 2 )), (175) is represented as 1 ρ Sq (ρ,χ,θ,φ) = 0, 0 tanh( ) (cos χ + i sin χ cos θ) 1, 1 + √2i sin χ sin θ cos φ 1, 0 | i cosh2( ρ ) | ii− 2 | ii | ii 2 ρ + (cos χ i sin χ cos θ) 1, 1 + √2 sin χ sin θ sin φ 0, 0 + O(tanh2( )) . − | − ii | ii 2 (177)

4.2 Dimensional Reduction Here, we discuss reductions of the squeezed vacuum associated with lower dimensional geometries of the Bloch four-hyperboloid. First, we consider a 0D reduction: We focus on the the lowest point of the upper leaf of H4, i.e., (x1, x2, x3, x4, x5) = (0, 0, 0, 0, 1), which corresponds to no squeezing ρ = 0, and so the squeeze operator (147) becomes trivial: Sˆ =1. (178)

21 The squeezed vacuum (166) is reduced to the trivial vacuum:

Sq = 0, 0, 0, 0 . (179) | i | i Next, let us consider a 2D reduction, H4 H2, which is realized at θ = 0: → (x1, x2, x3, x4, x5)=(0, 0, sin χ sinh ρ, cos χ sinh ρ, cosh ρ). (180)

The interaction Hamiltonian (148) is reduced to

int ρ iχ ρ iχ int int Hˆ (ρ, χ, 0, φ)= i a†c† e i b†d† e− + h.c. = Hˆ (ρ,χ)+ Hˆ (ρ, χ) − 2 − 2 T B − = Hˆ int(ρ, χ, 0, 0), (181) and then the squeezed vacuum (166) becomes

2S S 1 ρ 2iχSz Sq = sq(ρ,χ) T sq(ρ, χ) B = tanh( ) e S,Sz S,Sz , | i | i ⊗ | − i cosh2( ρ ) − 2 | i ⊗ | i 2 S=0,1/2,1, Sz= S X ··· X− (182) which is obvious from (83). Lastly, let us consider 1D reductions. For χ = 0, we have (x1, x2, x3, x4, x5)= (0, 0, 0, sinh ρ, cosh ρ) and the squeezed state is reduced to

S 1 ρ 2S 1 Sq = √2S + 1( tanh( )) S,Sz S,Sz . (183) | i cosh2( ρ ) − 2 · √2S +1 | i ⊗ | i 2 S=0,1/2,1, Sz= S X ··· X− π 1 2 3 4 5 Similarly for χ = θ = φ = 2 , i.e. (x , x , x , x , x )=(0, sinh ρ, 0, 0, cosh ρ), we have

int π π π ρ ρ int int Hˆ (ρ, , , )= i (a†d† ad)+ i (b†c† bc)= Hˆ (ρ, 0)+ Hˆ (ρ, π), (184) 2 2 2 − 2 − 2 − (a,d) (b,c) and

Sq = sq(ρ, 0) sq(ρ, π) | i | ia,d ⊗ | ib,c S ∞ 1 ρ 2S 1 S Sz = √2S + 1 ( tanh( )) ( 1) − S,Sz S, Sz . (185) cosh2( ρ ) − 2 · √2S +1 − | i ⊗ | − i 2 S=0 Sz= S X X− Each spin sector of (185) realizes a maximally entangled spin-singlet state

S 1 S Sz ( 1) − S,Sz S, Sz , (186) √2S +1 − | i ⊗ | − i Sz= S X− which is utilized in the context of violations of Bell inequality for arbitrary spins [49, 50, 51, 52].

22 4.3 Statistical properties 4.3.1 Statistics about spins The probability function to ﬁnd a state S,L S, R in the SO(4, 1) squeezed vacuum is18 | zi ⊗ | zi P (ρ,χ,θ,φ) ( S,L S, R ) Sq.(ρ,χ,θ,φ) 2 S,Lz,Rz ≡ | h z| ⊗ h z| | i| 1 ρ = tanh4S( ) (D(S)(φ, θ, χ) D(S)( χ, θ, φ)) 2. (189) 4 ρ 2 Lz,Rz cosh ( 2 ) | − − − − | We also introduce the following probability function to ﬁnd the two spin-coherent state with L = R = S:

S S 2S +1 4S ρ PS(ρ) PS,L ,R (ρ,χ,θ,φ)= tanh ( ). (190) ≡ z z cosh4( ρ ) 2 Lz= S Rz= S 2 X− X− Meanwhile, the probability function for the original SO(2, 1) squeezed vacuum is (see Appendix A) 1 ρ P (ρ)= tanh2n( ), (191) n 2 ρ 2 cosh ( 2 ) which monotonically decreases as the squeeze parameter increases. (190) exhibits non-trivial behavior with respect to S, and its maximum value is attained generally at S = 0 [Fig.5]. This is because of the internal 6 spin degrees of freedom that appears as the numerator in (190).

Figure 5: Behaviors of the probability functions for the SO(2, 1) (blue) and SO(4, 1) (red) for ρ = 2 and ρ = 3. The maximum value of PS(ρ = 2, 3) is attained at 2S = 1, 4, while the maximum value of P (ρ =2, 3) is attained at n = 0. For ρ >> 1, the maximum value is attained at 2S 1 eρ. n ≃ 4 It is straightforward to derive ρ Sˆ = S P (ρ,χ,θ,φ) = sinh2( ), h i S 2 S=0,1/2,1, X ··· 3 ρ 1 ρ Sˆ2 = S2 P (ρ,χ,θ,φ)= sinh4( )+ sinh2( ), (192) h i S 2 2 2 2 S=0,1/2,1, X ··· 18Since product of two D-matrices should be an SU(2) group element, it satisﬁes S (D(S)(φ,θ, χ) D(S)( χ, θ, φ)) 2 = 2S + 1. (187) | − − − − Lz ,Rz | Lz ,RXz=−S 4S ρ 4 ρ From the formula, S=0,1/2,1,···(2S + 1) tanh ( 2 ) = cosh ( 2 ), one may see that the summation of (189) indeed becomes unity: P S

PS,Lz,Rz (ρ,χ,θ,φ) = 1. (188) S=0,1/X2,1,3/2,··· LzRXz=−S

23 and the variation of spin-magnitude is 1 ∆Sˆ2 = Sˆ2 Sˆ 2 = sinh2 ρ. (193) h i h i − h i 8 The ratio of the spin ﬂuctuation ∆Sˆ2 1 ρ h i = coth( ) (194) q Sˆ √2 2 h i monotonically decreases from to 1 as ρ increases from 0 to (left of Fig.6). The expectation values ∞ √2 ∞ of the left and right spins are (see Appendix B.1) ρ 3 ρ 1 ρ Lˆ = Rˆ = sinh2( )= Sˆ , Lˆ2 = Rˆ2 = sinh4( )+ sinh2( )= Sˆ2 . (195) h i h i 2 h i h i h i 2 2 2 2 h i

ρ 2S In the expansion of the squeezed vacuum (166), the expansion coeﬃcient in front of the spin-pair, tanh( 2 ) , will become dominant for higher S as ρ increases, and so higher spin states will become signiﬁcant. The correlation between the left and right spins is obtained as 3 ρ 1 ρ Lˆ Rˆ = sinh4( )+ sinh2( )= Sˆ2 . (196) h · i 2 2 2 2 h i (196) can be derived by using the expectation values of the number operators . Consequently, ∆L2 = h i Lˆ2 Lˆ 2, ∆R2 = Rˆ2 Rˆ 2 and cov(L, R)= Lˆ Rˆ Lˆ Rˆ , are given by h i − h i h i h i − h i h · i − h ih i 1 ∆L2 = ∆R2 = cov(L, R)= sinh2 ρ = ∆Sˆ2 . (197) h i h i 8 h i

The linear correlation coeﬃcient, J SO(4,1)(L, R)= cov(L,R) , takes the maximum value √ ∆L2 ∆R2 h i h i J SO(4,1)(L, R)=1, (198) which should be compared with that for the SO(2, 1) two-mode squeezed vacuum (233b):

SO(2,1) J (na,nb)=1. (199)

As the original SO(2, 1) squeezed vacuum exhibits maximal correlation, the SO(4, 1) squeezed vacuum has the maximal correlation between L and R spins. This result is reasonable since, without the spin degrees of freedom, the SO(4, 1) squeeze vacuum is almost equivalent to the original SO(2, 1) squeezed vacuum (Fig.4) and two-mode correlations are analogous in the SO(4, 1) and SO(2, 1) states. For z-components of left and right spins, a bit of calculation shows ’(see Appendix B.1) 1 L = R =0, ∆L 2 = ∆R 2 = sinh2 ρ, h zi h zi h z i h z i 8 1 cov(L , R )= L R L R = sinh2 ρ (1 2 sin2 χ sin2 θ), (200) z z h z · zi − h zih zi 8 − and then SO(4,1) cov(Lz, Rz) 2 2 J (Lz, Rz)= =1 2 sin χ sin θ. (201) ∆L 2 ∆R 2 − h z i h z i π q SO(4,1) At χ = θ = 2 , the linear correlation (201) takes maximum negative correlation J (Lz, Rz) = 1, SO(4,1) − while either at χ = 0, π or θ = 0, π, (201) takes the maximum positive correlation J (Lz, Rz) =+1 (Right of Fig.6). There is no correlation for χ and θ that satisfy sin2 χ sin2 θ =1/2.

24 Figure 6: Left: Behaviors of the spin magnitude (blue), spin ﬂuctuation (red) and normalized spin ﬂuctu- ation (magenta) and with respect to the squeeze parameter. Right: The angular coordinate dependence of SO(4,1) J (Lz, Rz). (The blue plane denotes J = 0 for comparison.)

4.3.2 von Neumann entropy With the usual deﬁnition of the density operator,ρ ˆ Sq (ρ,χ,θ,φ) Sq.(ρ,χ,θ,φ) , we construct the ≡ | ih | reduced density operator

S ρˆ(Red) S, R ( φ, θ, χ) ρˆ S, R ( φ, θ, χ) ≡ h z − − | | z − − i S=0,1/2,1, Rz= S X ··· X− S 1 4S ρ = tanh ( ) S,Lz(φ, θ, χ) S,Lz(φ, θ, χ) , (202) cosh4( ρ ) 2 | − ih − | 2 S=0,1/2,1, Lz= S X ··· X− and so the von Neumann entropy is given by

4 ρ 1 ∞ ρ cosh ( ) SSO(4,1) = tr(ˆρ(Red) logρ ˆ(Red))= (2S + 1) tanh4S( ) log 2 . (203) vN − cosh4( ρ ) 2 tanh4S( ρ ) 2 S=0,1/2,1.3/2, 2 X ··· The coeﬃcient (2S +1) on the last right-hand side of (203) comes from the SU(2) spin degrees of freedom. Meanwhile, the von Neumann entropy of the original squeezed vacuum (235) is given by

2 ρ 1 ∞ ρ cosh ( ) SSO(2,1) = tanh2n( ) log 2 . (204) vN cosh2( ρ ) 2 tanh2n( ρ ) 2 n=0,1,2, 2 X ··· The von Neumann entropy of the SO(4, 1) (203) grows rapidly than that of the SO(2, 1) (204), because of the internal spin degrees of freedom (Fig.7).

25 Figure 7: The von Neumann entropies of the SO(2, 1) (blue) (204) and SO(4, 1) (red) (203) squeezed vacua.

4.4 SO(4, 1) uncertainty relations From the SO(4, 1) Schwinger operators, we introduce non-commutative 4D coordinates:19 1 1 X1 = (a + a† + c + c†), X2 = i (a a† + c c†), 2√2 − 2√2 − − 1 1 X3 = (b + b† + d + d†), X4 = i (b b† + d d†), (208) 2√2 − 2√2 − − which satisfy the Heisenberg-Weyl algebra: 1 [X ,X ] = [X ,X ]= i , [X ,X ] = [X ,X ] = [X ,X ] = [X ,X ]=0. (209) 1 2 3 4 2 1 3 1 4 2 3 2 4 From the expectation values for the squeezed vacuum (see Appendix B), we have

X = X = X = X =0, (210) h 1i h 2i h 3i h 4i and the variants (∆X )2 (X )2 X 2 = (X )2 (m =1, 2, 3, 4) (211) h m i ≡ h m i − h mi h m i as 1 (∆X )2 = (∆X )2 = (cosh ρ sinh ρ cos χ), h 1 i h 3 i 4 − 1 (∆X )2 = (∆X )2 = (cosh ρ + sinh ρ cos χ). (212) h 2 i h 4 i 4 Consequently, 1 1 1 (∆X )2 (∆X )2 = (∆X )2 (∆X )2 = + sinh2 ρ sin2 χ . (213) h 1 ih 2 i h 3 ih 4 i 16 16 ≥ 16 19One can adopt other 4D coordinates:

1 † † 1 † † X1 = (a + a + b + b ), X2 = i (a a + b b ), 2√2 − 2√2 − − 1 † † 1 † † X3 = (c + c + d + d ), X4 = i (c c + d d ), (205) 2√2 − 2√2 − − which satisfy same algebra as (209). However, their expectation values for the SO(4, 1) squeezed vacuum are

2 2 2 2 1 X1 = X2 = X3 = X4 = 0, (∆X1) = (∆X2) = (∆X3) = (∆X4) = cosh ρ, (206) h i h i h i h i h i h i h i h i 4 and so 2 2 2 2 1 2 1 (∆X1) (∆X2) = (∆X3) (∆X4) = cosh ρ . (207) h i · h i h i · h i 16 ≥ 16 Therefore, for (205), only the trivial vacuum (ρ = 0) saturates the uncertainty bound.

26 As the squeeze parameter ρ increases, the uncertainty (213) monotonically increases. The inequality is saturated at χ =0, π:

2 1 ρ 2 1 ρ 2 1 ρ 2 1 ρ (∆X ) = e∓ , (∆X ) = e± , (∆X ) = e∓ , (∆X ) = e± . (214) h 1 i 4 h 2 i 4 h 3 i 4 h 4 i 4 To remove the χ-dependence in (212), let us apply the following coordinate rotation (same as in the original case (225)):

X cos χ sin χ X X cos χ sin χ X 1 2 − 2 1 , 3 2 2 3 , (215) X → sin χ cos χ X X → sin χ cos χ X 2 2 2 2 4 − 2 2 4 but we end up with 1 (∆X )2 = (∆X )2 = (cosh ρ sinh ρ(cos2 χ + sin2 χ cos θ)), h 1 i h 3 i 4 − 1 (∆X )2 = (∆X )2 = (cosh ρ + sinh ρ(cos2 χ + sin2 χ cos θ)). (216) h 2 i h 4 i 4 As mentioned around (151), we can remove φ-dependence by the redeﬁnition of the Schwinger operators but not χ, so the angle dependence in (212) is not generally removal unlike the original SO(2, 1) case (see Appendix A). However, for the special case θ = 0, χ-dependence in (216) can be removed:

2 2 1 ρ 2 2 1 ρ (∆X ) = (∆X ) e− , (∆X ) = (∆X ) e . (217) h 1 i h 3 i → 4 h 2 i h 4 i → 4 Recall that for θ = 0 the SO(4, 1) is simply reduced to the direct product of two SO(2, 1) squeezed states with opposite angles, χ and χ (182), and in each SO(2, 1) squeezed vacuum, the minimum uncertainty is − realized (Fig.8).

Figure 8: The uncertainty regions for θ = 0.

5 Summary

In this work, we proposed a generalized framework of the squeezed state that includes the spin degrees of freedom based on the Bloch four-hyperboloid (Table 1). While the obtained SO(4, 1) squeezed vacuum is a four-mode squeezed state in the photon picture, it can be interpreted as an entangled spin state with the Schwinger’s angular momentum operator formalism. The four parameters of the SO(4, 1) squeezed vacuum have a clear geometric origin in the Bloch four-hyperboloid: The squeeze parameter corresponds to the radial coordinate and the three internal spin parameters correspond to the three angular coordinates on

27 SO(2, 1) squeezed vacuum SO(4, 1) squeezed vacuum

Symmetry group SO(2, 1) = Sp(2; R)= U(1; H′) SO(4, 1) = U(1, 1; H) Representation Majorana/Dirac Dirac Topological map 1st non-compact Hopf map 2nd non-compact (hybrid) Hopf map Quantum manifold Bloch two-hyperboloid H2 Bloch four-hyperboloid H4 Degrees of freedom One/two-mode photons Four-mode photons or two-mode spins

Table 1: Comparison between the SO(2, 1) squeezed vacuum and the SO(4, 1) squeezed vacuum.

S3-latitude. As the original SO(2, 1) squeezed vacuum is expressed by a superposition of identical number two-mode photons, the SO(4, 1) squeezed vacuum is realized as a superposition of identical magnitude spin- pairs. The SO(4, 1) squeezed vacuum exhibits a hybrid entanglement: One is about the infinite set of the spin-pairs and the other is about two spins in each spin-pair. The linear correlation coefficient between the left and right spins realizes the maximum correlation. The statistical properties of the SO(4, 1) squeezed vacuum are qualitatively different to the original SO(2, 1) squeezed vacuum in several respects, such as the spin magnitude dependence of the probability function and the angular coordinate dependence of the linear correlation between the left and right third-spin-components. The uncertainty relation was also generalized in 4D with coordinate dependence of the Bloch four-hyperboloid. The Bloch four-hyperboloid has provided a formalism that accommodates higher spins and squeezed states on the same footing. Since non-compact geometry can incorporate compact geometry as its internal structure, theory based on non-compact geometry generally provides a more comprehensive framework than that on compact geometry. While the importance of compact geometry in quantum information has already been appreciated [61], non-compact or indefinite signature geometry has less been exploited so far. We hope that significance of non-compact geometry in quantum information will be further unveiled in future developments.

Acknowledgements

This work was supported by JSPS KAKENHI Grant Number 16K05334 and 16K05138.

A Basic properties of the SO(2, 1) two-mode squeezed vacuum

This section is mainly based on Sec.7.7 of Gerry and Knight [7].

A.1 One-mode SO(2, 1) squeezed vacuum (Majorana representation) For the one-mode SO(2, 1) squeezed vacuum

2 ρ eiφa† + ρ e−iφa2 sq(ρ, φ) = e− 4 4 0 , (218) | i | i we have ρ ρ 1 nˆ = sinh2( ), nˆ2 = sinh4( )+ sinh2 ρ. (219) h i 2 h i 2 2 The variant of the number is 1 ∆ˆn2 = sinh2 ρ, (220) h i 2 and then 2 ∆n ρ h i = √2 coth( ) √2. (221) q n 2 ≥ h ai

28 The non-commutative coordinates which satisfy 1 [X ,X ]= i (222) 1 2 2 are deﬁned as 1 1 X = (a + a†), X = i (a a†). (223) 1 2 2 − 2 − The relevant expectation values are 1 1 X = X =0, (∆X )2 = (cosh ρ sinh ρ cos φ), (∆X )2 = (cosh ρ + sinh ρ cos φ). (224) h 1i h 2i h 1 i 4 − h 2 i 4 An appropriate choice of the rotation

X cos φ sin φ X 1 2 − 2 1 , (225) X → sin φ cos φ X 2 2 2 2 can remove the angle dependence of the variants:

2 1 ρ 2 1 ρ (∆X ) = e− , (∆X ) = e . (226) h 1 i 4 h 2 i 4

A.2 Two-mode SO(2, 1) squeezed vacuum (Dirac representation) For the SO(2, 1) two-mode squeezed vacuum

ρ eiφa†b†+ ρ e−iφab sq(ρ, φ) = e− 2 2 0 , (227) | i | i the probability function is given by 1 ρ P (ρ)= n ,n sq(ρ, φ) 2 = δ tanh2na ( ), (228) na,nb a b na,nb 2 ρ 2 |h | i| · cosh ( 2 ) which satisﬁes ∞ ∞ Pna,nb (ρ)= Pn(ρ)=1. (229) n ,n =0 n=0 aXb X Here, P denotes the probability to ﬁnd n n in the squeezed vacuum: n | i ⊗ | i 1 ρ P (ρ) P (ρ)= tanh2n( ). (230) n n,n 2 ρ 2 ≡ cosh ( 2 ) The number averages are evaluated as

∞ ρ nˆ = nˆ = n P (ρ, φ) = sinh2( ), (231a) h ai h bi · n 2 n=0 X ∞ ρ 1 nˆ2 = nˆ2 = nˆ nˆ = n2 P (ρ, φ) = sinh4( )+ sinh2 ρ, (231b) h ai h b i h a bi · n 2 4 n=0 X and the variants and covariants are 1 ∆n 2 = ∆n 2 = cov(n ,n )= sinh2 ρ. (232) h a i h b i a b 4

29 The ratio of the number ﬂuctuation and the linear correlation are respectively given by

2 ∆na ρ h i = coth( ) 1, (233a) q n 2 ≥ h ai cov(na,nb) J(na,nb)= =1. (233b) ∆n 2 ∆n 2 h a i h b i q The ratio (233a) monotonically decreases to unity, as the squeeze parameter increases. The statistics of the number is a super-Poissonian statistics. With the density operatorρ ˆ sq(ρ, φ) sq(ρ, φ) , the reduced density operator is deﬁned as ≡ | ih | 1 ρ ρˆ(Red) tr ρˆ = tanh2n( ) n n , (234) ≡ R 2 ρ 2 | iLh |L cosh ( 2 ) n=0,1,2, X ··· and the associated von Neumann entropy is

2 ρ 1 ∞ ρ cosh ( ) S = tr(ˆρ(Red) logρ ˆ(Red))= tanh2n( ) log 2 . (235) vN − 2 ρ 2 2n ρ cosh ( 2 ) n=0 tanh ( 2 ) X Expectation values for the non-commutative coordinates 1 1 X1 = (a + a† + b + b†), X2 = i (a a† + b b†), (236) 2√2 − 2√2 − − are given by 1 1 X = X =0, (∆X )2 = (cosh ρ sinh ρ cos φ), (∆X )2 = (cosh ρ + sinh ρ cos φ), (237) h 1i h 2i h 1 i 4 − h 2 i 4 which are identical to the one-mode case (224).

B Expectation values of the SO(4, 1) and SO(2, 3) squeezed vacua

Even without knowing a simple number state expansion of squeezed vacuum, we can readily evaluate squeezed vacuum expectation values by the covariance of the Schwinger operator:

Sˆ† Ψˆ Sˆ = MΨˆ , (238)

which implies the following useful formula

O(Ψ)ˆ = Sq O(Ψ)ˆ Sq = 0 Sˆ†O(Ψ)ˆ Sˆ 0 = 0 O(Sˆ†Ψˆ Sˆ) 0 = 0 O(MΨ)ˆ 0 . (239) h i h | | i h | | i h | | i h | | i B.1 SO(4, 1) squeezed vacuum For the SO(4, 1) squeezed vacuum (161), we can derive

2 2 0 Sˆ†a Sˆ 0 = 0 Sˆ†c Sˆ 0 = 0 Sˆ†ac†Sˆ 0 =0, h | | i h | | i h | | i 2 2 2 ρ 2 2 2 ρ 0 Sˆ†aa†Sˆ 0 = M + M = cosh ( ), 0 Sˆ†cc†Sˆ 0 = M + M = cosh ( ), h | | i | 11| | 12| 2 h | | i | 33| | 34| 2 2 2 2 ρ 2 2 2 ρ 0 Sˆ†a†aSˆ 0 = M + M = sinh ( ), 0 Sˆ†c†cSˆ 0 = M + M = sinh ( ), h | | i | 13| | 14| 2 h | | i | 31| | 32| 2 1 0 Sˆ†acSˆ 0 = M M ∗ + M M ∗ = sinh ρ(cos χ + i sin χ cos θ), (240) h | | i 11 31 12 32 −2

30 where M(ρ,χ,θ,φ) is (115). With these results, the expectation value of

2 1 2 1 2 2 2 2 (X ) = (a+a† +c+c†) = (a +aa† +a†a+a† +c +cc† +c†c+c† +2ac+2ac† +2a†c+2a†c†), (241) 1 8 8 is evaluated as 1 (X )2 = (cosh ρ sinh ρ cos χ). (242) h 1 i 4 − In a similar manner, we have ρ nˆ = nˆ = nˆ = nˆ = sinh2( ), h ai h bi h ci h di 2 ρ ρ nˆ2 = nˆ2 = nˆ2 = nˆ2 = 2 sinh4( ) + sinh2( ), h ai h b i h c i h di 2 2 1 ∆n 2 = ∆n 2 = ∆n 2 = ∆n 2 = sinh2 ρ, (243) h a i h b i h c i h d i 4 which are exactly equal to (231) and (232). Furthermore with ρ nˆ nˆ = nˆ nˆ = sinh4( ), h a bi h c di 2 1 ρ nˆ nˆ = nˆ nˆ = sinh2 ρ (cos2 χ + sin2 χ cos2 θ) + sinh4( ), h a ci h b di 4 2 1 ρ nˆ nˆ = nˆ nˆ = sinh2 ρ sin2 χ sin2 θ + sinh4( ). (244) h a di h b ci 4 2 the covariances, cov(n ,n )= nˆ nˆ nˆ nˆ , are derived as a b h a bi − h aih bi

cov(na,nb) = cov(nc,nd)=0, 1 cov(n ,n ) = cov(n ,n )= sinh2 ρ (1 sin2 χ sin2 θ), a c b d 4 − 1 cov(n ,n ) = cov(n ,n )= sinh2 ρ sin2 χ sin2 θ, (245) a d b c 4 1 The linear correlation coeﬃcients, J(na,nb)= 2 2 cov(na,nb), are √ ∆na ∆nb h ih i

J(na,nb)= J(nc,nd)=0, J(n ,n )= J(n ,n )=1 sin2 χ sin2 θ, a c b d − 2 2 J(na,nd)= J(nb,nc) = sin χ sin θ. (246) The linear correlation coeﬃcients (246) depend on the angular coordinates, while for the original squeezed vacuum they are constant (233b). The linear correlation coeﬃcients (246) become the maximum value 1 π at χ =0, π or θ =0, π, while they vanish either at χ = θ = 2 . The expectation values for the left and right spins are evaluated as 1 1 ρ Lˆ = Rˆ = nˆ + nˆ = sinh2( ), Lˆ = Rˆ =0, (247) h i h i 2h ai 2h bi 2 h zi h zi and 1 3 ρ 1 ρ Lˆ2 = Rˆ2 = (ˆn +ˆn )2 = sinh4( )+ sinh2( ), h i h i 4h a b i 2 2 2 2 1 1 ρ Lˆ Rˆ = (ˆn +ˆn )(ˆn +ˆn ) = sinh2 ρ + sinh4( ), h · i 4h a b c d i 8 2 1 1 Lˆ2 = Rˆ2 = (ˆn nˆ )2 = sinh2 ρ, h zi h zi 4h a − b i 8 1 1 Lˆ Rˆ = (ˆn nˆ )(ˆn nˆ ) = sinh2 ρ (1 2 sin2 χ sin2 θ). (248) h z · zi 4h a − b c − d i 8 −

31 Consequently, 1 1 ∆L2 = Lˆ2 Lˆ 2 = ( ∆n2 + 2cov(n ,n )+ ∆n2 )= sinh2 ρ = ∆R2 , h i h i − h i 4 h ai a b h b i 8 h i 1 ∆L2 = Lˆ2 Lˆ 2 = sinh2 ρ = ∆R2 , (249) h zi h zi − h zi 8 h zi and 1 cov(L, R)= Lˆ Rˆ Lˆ Rˆ = (cov(n ,n ) + cov(n ,n ) + cov(n ,n ) + cov(n ,n )) h · i − h ih i 4 a c a d b c b d 1 = sinh2 ρ, 8 1 cov(L , R )= Lˆ Rˆ Lˆ Rˆ = (cov(n ,n ) cov(n ,n ) cov(n ,n ) + cov(n ,n )) z z h z · zi − h zih zi 4 a c − a d − b c b d 1 = sinh2 ρ (1 2 sin2 χ sin2 θ). (250) 8 −

B.2 SO(2, 3) Sp(4; R) squeezed vacua ≃ For comparison, we also derive expectation values for the SO(2, 3) squeezed vacua [23]. In the SO(2, 3) case, we have two-mode and four-mode squeezed states, since SO(2, 3) allows Majorana representation and Dirac representation. The SO(2, 3) squeezed vacua are also physically distinct for the Dirac-type and the Schwinger-type. In this section, we utilize the following parameterization of the coordinates on H2,2:

x1 = sin θ cos χ sinh ρ, x2 = sin θ sin χ sinh ρ, x3 = cos φ sin θ cosh ρ, x4 = sin φ sin θ cosh ρ, x5 = cos θ, (251) which satisfy (x1)2 + (x2)2 (x3)2 (x4)2 (x5)2 = 1. (252) − − − −

B.2.1 Two-mode SO(2, 3) squeezed vacuum (Majorana representation) • Dirac-type

θ nˆ = nˆ = sinh2 ρ sin2( ), h ai h bi 2 θ θ ∆n 2 = ∆n 2 = sinh2 ρ sin2( ) 1 + cosh(2ρ) sin2( ) , h a i h b i 2 2 1 cov(n ,n )= sinh2 ρ sin2 θ, a b 4 1 J(na,nb)= 2 θ . (253) 4(1 + cosh(2ρ) sin ( 2 )) Similar to the SO(4, 1) case (see Appendix B.1), the expectation values for the Dirac-type SO(2, 3) squeezed vacuum have angle dependence but in a diﬀerent manner.

• Schwinger-type

32 ρ nˆ = nˆ = sinh2( ), h ai h bi 2 1 ∆n 2 = ∆n 2 = ∆n 2 = sinh2 ρ, h a i h b i h b i 2 cov(na,nb)=0,

J(na,nb)=0. (254)

The Schwinger-type SO(2, 3) squeezed vacuum is just given by a direct product of two SO(2, 1) squeezed vacuum states [23]. Therefore, (254) is just given by two copies of (219) and (220), and there is no correlation between a and b modes.

B.2.2 Four-mode SO(2, 3) squeezed vacuum (Dirac representation) • Dirac-type

θ nˆ = nˆ = nˆ = nˆ = sinh2 ρ sin2( ), h ai h bi h ci h di 2 θ θ ∆n 2 = ∆n 2 = ∆n 2 = ∆n 2 = sinh2 ρ sin2( )(1 + sin2 ρ sin2( )), h a i h b i h c i h d i 2 2 1 θ cov(n ,n ) = cov(n ,n )= sinh2(2ρ) sin4( ), cov(n ,n ) = cov(n ,n )=0, a b c d 4 2 a c b d 1 cov(n ,n ) = cov(n ,n )= sinh2 ρ sin2 θ, a d b c 4 1 θ J(n ,n )= J(n ,n )= cosh2 ρ sin2( ), J(n ,n )= J(n ,n )=0, a b c d 2 2 θ 2 a c b d 1 + sinh ρ sin ( 2 ) 1 θ J(n ,n )= J(n ,n )= cos2( ). (255) a d b c 2 2 θ 2 1 + sinh ρ sin ( 2 ) As in the two-mode case (253), (255) exhibits non-trivial angle dependence but in a diﬀerent manner. As ρ increases, J(na,nb) and J(nc,nd) monotonically increase while J(na,nd) and J(nb,nc) monotonically decrease.

• Schwinger-type

ρ nˆ = nˆ = nˆ = nˆ = sinh2( ), h ai h bi h ci h di 2 1 ∆n 2 = ∆n 2 = ∆n 2 = ∆n 2 = sinh2 ρ, h a i h b i h c i h d i 4 cov(na,nb) = cov(nc,nd)=1, cov(na,nc) = cov(na,nd) = cov(nb,nc) = cov(nb,nd)=0,

J(na,nb)= J(nc,nd)=1, J(na,nc)= J(na,nb)= J(nb,nc)= J(nb,nd)=0. (256)

This is a straightforward generalization of the two-mode case (254).

References

[1] Horace P. Yuen, “Two-photon coherent states of the radiation ﬁeld”, Phys. Rev. A 13 (1976) 2226.

33 [2] James N. Hollenhorst, “Quantum limits on resonant-mass gravitational-radiation detectors”, Phys,Rev. D 19 (1979) 1669.

[3] Carlton M. Caves, “Quantum-mechanical noise in an interferometer”, Phys. Rev. D 23 (1981) 1693- 1708.

[4] D. F. Walls, “Squeezed states of light”, Nature 306 (1983) 141-146.

[5] Bonny L. Schumaker and Carlton M. Caves, “New formalism for two-photon quantum optics. I. Quadrature phases and squeezed states”, Phys. Rev. A 31 (1985) 3068.

[6] Bonny L. Schumaker and Carlton M. Caves, “New formalism for two-photon quantum optics. II. Mathematical foundation and compact notation”, Phys. Rev. A 31 (1985) 3093.

[7] Christopher C. Gerry, Peter L. Knight, “Introductory Quantum Optics”, Cambridge University Press (2004).

[8] P.D. Drummond, Z. Ficek, “Quantum Squeezing”, Springer (2003).

[9] Hans-A. Bachor, Timothy C. Ralph, “A Guide to Experiments in Quantum Optics”, Wiley-VCH (2019).

[10] Roman Schnabel, “Squeezed states of light and their applications in laser interferometers”, Phys. Rep. 684 (2017) 1-51.

[11] G. J. Milburn, “Multimode minimum uncertainty squeezed states”, J. Phys. A: Math. Gen. 17 (1984) 737-745.

[12] R. F. Bishop, A. Vourdas, “General two-mode squeezed states”, Z. Phys. B-Cond.Matt. 71 (1988) 527-529.

[13] Xin Ma, William Rhodes, “Multimode squeeze operators and squeezed states”, Phys. Rev. A 41 (1990) 4625.

[14] D. Han, Y. S. Kim, Marilyn E. Noz, Leehwa Yeh, “Symmetries of two-mode squeezed states”, J. Math. Phys. 34 (1993) 5493.

[15] Arvind, B. Dutta, N. Mukunda, R. Simon, “Two-mode quantum systems: invariant classiﬁcation of squeezing transformations and squeezed states”, Phys.Rev. A 52 (1995) 1609-1620.

[16] Arvind, B. Dutta, N. Mukunda, R. Simon, “The real symplectic groups in quantum mechanics and optics”, J. Phys. 45 (1995) 471-497.

[17] Emi Yukawa, Kae Nemoto, “Classiﬁcation of spin and multipolar squeezing”, J. Phys. A: Math. Theor. 49 (2016) 255301.

[18] K. Svozil, “Squeezed Fermion states”, Phys. Rev. Lett. 65 (1990) 3341.

[19] H. A. Schmitt and A. Mufti, “Squeezing via superpositions of even and odd Sp(2,R) coherent states”, Phys.Rev. A 44 (1991) 5988.

[20] H. A. Schmitt, “Osp(4/2, R) supersymmetry and the one- and two-photon dressed Jaynes-Cummings hamiltonian”, Optics Communications 95 (1993) 265-268.

[21] Felix Bloch, “Nuclear Induction”, Phys. Rev. 70 (1946) 460.

34 [22] Rémy Mosseri, Rossen Dandoloff, “Geometry of entangled states, Bloch spheres and Hopf fibrations”, J. Phys. A: Math. Gen. 34 (2001) 10243; quant-ph/0108137.

[23] Kazuki Hasebe, “ Sp(4; R) Squeezing for Bloch Four-Hyperboloid via The Non-Compact Hopf Map”, J.Phys.A: Math.Theor. 53 (2020) 055303; arXiv:1904.12259.

[24] Kazuki Hasebe, “The Split-Algebras and Non-compact Hopf Maps”, J. Math. Phys.51 (2009) 053524; arXiv:0905.2792.

[25] Kazuki Hasebe, “Non-Compact Hopf Maps and Fuzzy Ultra-Hyperboloids”, Nucl.Phys. B 865 (2012) 148-199; arXiv:1207.1968.

[26] Kazuki Hasebe, “Split-Quaternionic Hopf Map, Quantum Hall Eﬀect, and Twistor Theory”, Phys.Rev.D81 (2010) 041702; arXiv:0902.2523.

[27] Kenta Esaki, Masatoshi Sato, Kazuki Hasebe, Mahito Kohmoto, “Edge states and topological phases in non-Hermitian systems”, Phys. Rev. B 84 (2011) 205128; arXiv:1107.2079.

[28] Masatoshi Sato, Kazuki Hasebe, Kenta Esaki, Mahito Kohmoto , “Time-Reversal Symmetry in Non- Hermitian Systems”, Prog. Theor. Phys. 127 (2012) 937-974; arXiv:1106.1806.

[29] Harold C. Steinacker, “Cosmological space-times with resolved Big Bang in Yang-Mills matrix models”, JHEP 02 (2018) 033; arXiv:1709.10480.

4 [30] Marcus Sperling, Harold C. Steinacker, “The fuzzy 4-hyperboloid Hn and higher-spin in Yang-Mills matrix models”, Nucl.Phys. B 941 (2019) 680; arXiv:1806.05907.

[31] A. Stern, Chuang Xu, “Signature change in matrix model solutions”, Phys.Rev. D 98 (2018) 086015; arXiv:1808.07963.

[32] See, for instance, Kazuki Hasebe, “Hopf Maps, Lowest Landau Level, and Fuzzy Spheres”, SIGMA 6 (2010) 071; arXiv:1009.1192.

[33] F. A. Arecchi, Eric Courtens, Robert Gilmore, Harry Thomas, “Atomic coherent states in quantum optics’, Phys. Rev. A 6 (1972) 2211.

[34] A. M. Perelomov, “Coherent States for Arbitrary Lie Group”, Commun. Math. Phys. 26 (1972) 222- 236.

[35] J. M. Radcliﬀe “Some properties of coherent spin states”, J. Phys. A 4 (1971) 313.

[36] See for instance, J. J. Sakurai, “Modern Quantum Mechanics”, Addison Wesley (1993).

[37] J. Schwinger, “Quantum Theory of Angular Momentum”, edited by L. C. Biedenharn and H. Van Dam, Academic Press (1965) p. 229.

[38] Robert Gilmore, “Lie Groups, Physics, and Geometry’, Cambridge University Press (2008).

[39] S. Coleman, J. Wess, B. Zumino, “Structure of Phenomenological Lagrangians. I”, Phys. Rev. 177 (1969) 2239-2246.

[40] C. G. Callan, Jr., S. Coleman, J. Wess, B. Zumino, “Structure of Phenomenological Lagrangians. II”, Phys. Rev. 177 (1969) 2247-2250.

[41] Abdus Salam, J. Strathdee, “On Kaluza-Klein Theory”, Ann. Phys. 141 (1982) 316-352.

35 [42] Robert Gilmore, Jian-Min Yuan, “Group theoretical approach to semiclassical dynamics: Single mode case”, J. Chem. Phys. 86 (1987) 130.

[43] Monique Combescure, Didier Robert, “Coherent States and Applications in Mathematical Physics”, Springer (2012).

[44] Kazuki Hasebe, “Relativistic Landau Models and Generation of Fuzzy Spheres”, Int.J.Mod.Phys.A 31 (2016) 1650117; arXiv:1511.04681.

[45] K. W´odkiewicz, J.H. Eberly, “Coherent states, squeezed ﬂuctuations, and the SU(2) and SU(1,1) groups in quantum-optics applications”, J.Opt.Soc.Am. B 2 (1985) 458.

[46] Bernard Yurke, Samuel L. McCall, and John R. Klauder, “SU(2) and SU(1,1) interferometers”, Phys.Rev. A 33 (1986) 4033.

[47] Christopher C. Gerry, “Application of SU(1,1) coherent states to the interaction of squeezed light in an anharmonic oscillator”, Phys. Rev. A 35 (1987) 2146.

[48] Christopher C. Gerry and Adil Benmoussa, “Two-mode coherent states for SU(1,1) SU(1,1)”, Phys. ⊗ Rev. A 62 (2000) 033812.

[49] Anupam Garg and N. D. Mermin, “Bell Inequalities with a Range of Violation that Does Not Diminish as the Spin Becomes Arbitrarily Large”, Phys. Rev. Lett. A 49 (1982) 1294.

[50] Asher Peres, “Finite violation of a Bell inequality for arbitrarily large spin”, Phys. Rev. A 46 (1992) 4413.

[51] Christopher C. Gerry and Jaroslav Albert, “Finite violations of a Bell inequality for high spin: An optical realization”, Phys. Rev. A 72 (2005) 043822.

[52] Christopher C. Gerry, Adil Benmoussa, Edwin E. Hach III, Jaroslav Albert, “Maximal violations of a Bell inequality by entangled spin-coherent states”, Phys. Rev. A 79 (2009) 022111.

[53] Kazuki Hasebe, “SO(5) Landau models and nested matrix geometry”, Nucl.Phys. B 956 (2020) 115012; arXiv:2002.05010.

[54] A. O. Barut, A. B¨ohm, “Reduction of a Class of O(4, 2) Representations with Respect to SO(4, 1) and SO(3, 2)”, Jour. Math. Phys. 11 (1970) 2938.

[55] M. Gunaydin, C. Saclioglu, “Bosonic construction of the Lie algebras of some non-compact groups appearing in supergravity theories and their oscillator-like unitary representations”, Phys. Lett. B 108 (1982) 180-186.

[56] M. Gunaydin, C. Saclioglu, “Oscillator-like unitary representations of non-compact groups with a jordan structure and the non-compact groups of supergravity”, Commun. Math. Phys. 87 (1982) 158- 179.

[57] I. Bars, M. Gunaydin, “Unitary representations of non-compact supergroups”, Commun. Math. Phys. 91 (1983) 31-51.

[58] M. Gunaydin, D. Minic, M. Zagermann, “4D doubleton conformal theories, CPT and IIB strings on AdS S5”, Nucl. Phys. B 534 (1998) 96-120. 5 ×

36 [59] Sudarshan Fernando, Murat Gunaydin, “Minimal unitary representation of SU(2,2) and its deforma- tions as massless conformal ﬁelds and their supersymmetric extensions”, J. Math. Phys. 51 (2010) 082301; arXiv:0908.3624.

[60] Karan Govil, Murat Gunaydin, “Deformed Twistors and Higher Spin Conformal (Super-)Algebras in Six Dimensions”, JHEP 03 (2015) 026; arXiv:1401.6930.

[61] See, for instance, Ingemar Bengtsson, Karol Zyczkowski, “Geometry of Quantum States’, Cambridge University Press (2006).