Poincare group


Riemannian geometry

Differential geometry

synthetic differential geometry


from point-set topology to differentiable manifolds

geometry of physics: coordinate systems, smooth spaces, manifolds, smooth homotopy types, supergeometry



smooth space


The magic algebraic facts




tangent cohesion

differential cohesion

graded differential cohesion

id id fermionic bosonic bosonic Rh rheonomic reduced infinitesimal infinitesimal & étale cohesive ʃ discrete discrete continuous * \array{ && id &\dashv& id \\ && \vee && \vee \\ &\stackrel{fermionic}{}& \rightrightarrows &\dashv& \rightsquigarrow & \stackrel{bosonic}{} \\ && \bot && \bot \\ &\stackrel{bosonic}{} & \rightsquigarrow &\dashv& Rh & \stackrel{rheonomic}{} \\ && \vee && \vee \\ &\stackrel{reduced}{} & \Re &\dashv& \Im & \stackrel{infinitesimal}{} \\ && \bot && \bot \\ &\stackrel{infinitesimal}{}& \Im &\dashv& \& & \stackrel{\text{étale}}{} \\ && \vee && \vee \\ &\stackrel{cohesive}{}& ʃ &\dashv& \flat & \stackrel{discrete}{} \\ && \bot && \bot \\ &\stackrel{discrete}{}& \flat &\dashv& \sharp & \stackrel{continuous}{} \\ && \vee && \vee \\ && \emptyset &\dashv& \ast }


Lie theory, ∞-Lie theory

differential equations, variational calculus

Chern-Weil theory, ∞-Chern-Weil theory

Cartan geometry (super, higher)



The Poincaré group ISO(d,1)ISO(d,1) is the isometry group of Minkowski spacetime of dimension d+1d+1. The classical case is ISO(3,1)ISO(3, 1), the group of affine transformations on 4\mathbb{R}^4 which preserve the Minkowski "metric", i.e., the group of maps f: 4 4f: \mathbb{R}^4 \to \mathbb{R}^4 of the form f(x)=Mx+bf(\vec{x}) = M \vec{x} + \vec{b} such that

Q(f(x)f(y))=Q(xy)Q(f(\vec{x}) - f(\vec{y})) = Q(\vec{x} - \vec{y})

where the quadratic form Q(t,x,y,z)=t 2x 2y 2z 2Q(t, x, y, z) = t^2 - x^2 - y^2 - z^2 is often called the Minkowski “norm”. The group elements are multiplied by composing maps.

The Poincaré group GG may also be described as a semidirect product

( 3)O(1,3)(\mathbb{R} \oplus \mathbb{R}^3) \rtimes O(1, 3)

where O(1,3)O(1, 3), the Lorentz group, consists of all linear transformations L: 4 4L: \mathbb{R}^4 \to \mathbb{R}^4 that preserve the Minkowski inner product of signature (1,3)(1, 3).

This is a linear algebraic group (e.g. arXiv:0905.1217, p. 6-7).


Basic structure of the Lorentz group

The Lorentz group is a 6-dimensional Lie group. It has four connected components; the connected component of the identity is called the special Lorentz group and is denoted SO +(1,3)SO^+(1, 3).

  • The superscript of SO +(1,3)SO^+(1, 3) indicates that the group elements take the forward light cone, i.e., the set of vectors
    {v=(t,x,y,z):Q(v)>0,t>0},\{v = (t, x, y, z): Q(v) \gt 0, t \gt 0\},

    to itself, and the SS indicates of course group elements which have determinant 1 as 4×44 \times 4 matrices. It easily follows that elements of SO +(1,3)SO^+(1, 3) preserve orientation of the spatial component 3\mathbb{R}^3.

Passage between the four components of the full Lorentz group is effected by composing with a time-reversal operator

(t,x,y,z)(t,x,y,z)(t, x, y, z) \mapsto (-t, x, y, z)

and with a spatial inversion (or parity-reversal) operator

(t,x,y,z)(t,x,y,z)(t, x, y, z) \mapsto (t, -x, -y, -z)

The special Lorentz group (also called the proper, orthochronous Lorentz group: “orthochronous” here means the forward light cone is mapped to itself, and “proper” means orientation-preserving) may be analyzed further. The subgroup of SO +(1,3)SO^+(1, 3) that fixes the unit time-like vector

u t:=(1,0,0,0)u_t := (1, 0, 0, 0)

may be identified with the group of rotations SO(3)SO(3), since the restriction of the Minkowski norm to the spatial component 3\mathbb{R}^3 is minus the usual Euclidean norm, x 2y 2z 2-x^2 - y^2 - z^2. This subgroup is of course 3-dimensional.

A general element gSO +(1,3)g \in SO^+(1, 3) may be decomposed uniquely in the form

g=β vρg = \beta_v \circ \rho

where ρ\rho is a rotation and β v\beta_v is a boost in the direction vv (vS 2v \in S^2 a unit spatial vector), mapping

u tcosh(β)u t+sinh(β)vu_t \mapsto \cosh(\beta)u_t + \sinh(\beta)v
vsinh(β)u t+cosh(β)vv \mapsto \sinh(\beta)u_t + \cosh(\beta)v

for some parameter β\beta (called the rapidity), and acting as the identity on the spatial plane orthogonal to vv. Thus a boost is described by a pair (v,β)(v, \beta), involving 3 parameters. (Warning: boosts do not compose to form a subgroup.) A boost can be thought of as a relativistic coordinate change from a “laboratory” frame of reference to the frame of reference of an observer moving inertially in the direction vv with speed tanh(β)\tanh(\beta) (relative to the speed of light c=1c = 1), as measured in the laboratory frame.

Universal spin covering

We discuss aspects of the Poincaré spinor group.

The universal cover of SO +(1,3)SO^+(1, 3) is a double cover (the spin double cover)

SL 2()SO +(1,3)SL_2(\mathbb{C}) \to SO^+(1, 3)

constructed as follows: to each x=(x 0,x 1,x 2,x 3)x = (x_0, x_1, x_2, x_3) one associates a Hermitian matrix

X=(x 0+x 1 x 2+ix 3 x 2ix 3 x 0x 1)=x 0(1 0 0 1)+x 1(1 0 0 1)+x 2(0 1 1 0)+x 3(0 i i 0)X = \left( \array{ x_0 + x_1 & x_2 + i x_3 \\ x_2 - i x_3 & x_0 - x_1 } \right) = x_0 \left( \array{ 1 & 0 \\ 0 & 1 } \right) + x_1 \left( \array{ 1 & 0 \\ 0 & -1 } \right) + x_2 \left( \array{0 & 1 \\ 1 & 0 } \right) + x_3 \left( \array{ 0 & i \\ -i & 0 } \right)

whose determinant is the Minkowski norm of xx. We thus identify 4\mathbb{R}^4 with the space HH of Hermitian matrices, and define an action of SL 2()SL_2(\mathbb{C}) on HH:

AX=AXA *,ASL 2(),XHA \cdot X = A X A^*, A \in SL_2(\mathbb{C}), X \in H

Observe that AXA *A X A^* belongs to HH. Also, det(AX)=det(X)det(A \cdot X) = det(X) since AA has determinant 1, so the action preserves the Minkowski norm. Therefore the action

SL 2()GL( 4)SL_2(\mathbb{C}) \to GL(\mathbb{R}^4)

factors through the inclusion O(1,3)GL( 4)O(1, 3) \hookrightarrow GL(\mathbb{R}^4). Furthermore, since SL 2()SL_2(\mathbb{C}) is connected, the action SL 2()O(1,3)SL_2(\mathbb{C}) \to O(1, 3) factors through the connected component SO +(1,3)SO^+(1, 3) of O(1,3)O(1, 3). It is not hard to check that the kernel of the action is {I,I}\{I, -I\}; therefore the map

SL 2()SO +(1,3)SL_2(\mathbb{C}) \to SO^+(1, 3)

is an open homomorphism between connected Lie groups of the same dimension, and is therefore surjective. In this way, we have produced an explicit identification

SL 2()/{±I}SO +(1,3)SL_2(\mathbb{C})/\{\pm I\} \cong SO^+(1, 3)

which exhibits SL 2()SL_2(\mathbb{C}) as a double cover of SO +(1,3)SO^+(1, 3); another way to say it is that SO +(1,3)SO^+(1, 3) is identified with the group PSL 2()PSL_2(\mathbb{C}) of complex Moebius transformations. Finally, there is morphism of covering spaces

SU(2) SL 2() π π SO(3) SO +(1,3)\array{ SU(2) & \hookrightarrow & SL_2(\mathbb{C}) \\ \pi \downarrow & & \downarrow \pi \\ SO(3) & \hookrightarrow & SO^+(1, 3) }

Here the inclusions are homotopy equivalences and the left projection is a universal covering map (as SU(2)S 3SU(2) \cong S^3 is simply connected), therefore SL 2()SL_2(\mathbb{C}) is also simply connected and the projection on the right is a universal covering map. This is the spin double cover; it is crucial for getting a correct mathematical description of fermions in particle theory.

With regard to the inclusion maps above being homotopy equivalences, we remark in passing that the homogeneous space

SO +(1,3)/SO(3)SO^+(1, 3)/SO(3)

is identified with the space of boost maps; concretely, each coset of SO +(1,3)/SO(3)SO^+(1, 3)/SO(3) is of the form β vSO(3)\beta_v SO(3) for a unique boost map β v\beta_v. Topologically, the space of boost maps is 3\mathbb{R}^3 or a 33-ball, hence contractible, and from the long exact homotopy sequence applied to the fibration

SO(3)SO +(1,3)SO +(1,3)/SO(3)SO(3) \to SO^+(1, 3) \to SO^+(1, 3)/SO(3)

we deduce that the inclusion i:SO(3)SO +(1,3)i: SO(3) \to SO^+(1, 3) is a homotopy equivalence.

Similarly the lift to the double covers SU(2)SL 2()SU(2) \to SL_2(\mathbb{C}) is a homotopy equivalence. Since SU(2)S 3SU(2) \cong S^3 is identified with the space of unit quaternions, SU(2)SU(2) and SL 2()SL_2(\mathbb{C}) are simply connected and hence the respective universal covering spaces of SO(3)SO(3) and SO +(1,3)SO^+(1, 3).

  • Remark: Geometrically the space of boosts carries hyperbolic structure as well, in other words the space of boosts carries a structure of hyperbolic 3-space H 3H^3.

Lie algebra presentations

As for any Lie group, there are various mechanisms for describing the Lie algebras of the Lorentz group and of the Poincaré group: by left-invariant vector fields, or by studying “infinitesimal generators” of 1-parameter subgroups, etc. We begin with the Lorentz group. See also Poincaré Lie algebra.

In the vector field picture, one often chooses a basis of the Lorentz algebra? consisting of six elements: the first three

M 12=x yy xM 23=y zz yM 13=z xx zM_{12} = x\partial_y - y\partial_x \qquad M_{23} = y\partial_z - z\partial_y \qquad M_{13} = z\partial_x - x\partial_z

describe rotational flows (around the zz-, xx-, and yy-axes respectively), and the last three

M 01=x t+t xM 02=y t+t yM 03=z t+t zM_{01} = x\partial_t + t\partial_x \qquad M_{02} = y\partial_t + t\partial_y \qquad M_{03} = z\partial_t + t\partial_z

describe hyperbolic or boost flows, (where the boosts are in the directions of the xx-, yy-, and zz-axes, respectively). One may easily compute the commutators by hand and reproduce the standard gobbledygook formula given in physics texts:

[M μν,M ρσ]=η μρM νση μσM νρη νρM μσ+η νσM μρ[M_{\mu\nu}, M_{\rho\sigma}] = \eta_{\mu\rho} M_{\nu\sigma} - \eta_{\mu\sigma} M_{\nu\rho} - \eta_{\nu\rho} M_{\mu\sigma} + \eta_{\nu\sigma} M_{\mu\rho}

where lower-case Greek letters range over 0,1,2,30, 1, 2, 3 and η\eta is the 4×44 \times 4 matrix representing the Minkowski quadratic form.

(I could be off by a sign here. It depends on whether η\eta has one plus and three minuses, or one minus and three pluses.)

By integrating the vector fields, we obtain in each of these cases flows or 1-parameter subgroups, e.g.,

sexp(sM 12)=(1 0 0 0 0 cos(s) sin(s) 0 0 sin(s) cos(s) 0 0 0 0 1)s \mapsto \exp(s M_{12}) = \left( \array{ 1 & 0 & 0 & 0 \\ 0 & \cos(s) & -\sin(s) & 0 \\ 0 & \sin(s) & \cos(s) & 0 \\ 0 & 0 & 0 & 1 } \right)


sexp(sM 01)=(cosh(s) sinh(s) 0 0 sinh(s) cosh(s) 0 0 0 0 1 0 0 0 0 1)s \mapsto \exp(s M_{01}) = \left( \array{ \cosh(s) & \sinh(s) & 0 & 0 \\ \sinh(s) & \cosh(s) & 0 & 0 \\ 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 1 } \right)

Differentiating these maps at s=0s = 0 gives matrix representations for the Lie algebra elements M μνM_{\mu\nu}.

For the 10-dimensional Poincaré algebra, we need to give, in addition to infinitesimal generators for rotations and boosts, four more elements which generate translations. In the vector field picture, these Lie algebra elements are represented by

P 0= t,P 1= x,P 2= y,P 3= zP_0 = \partial_t, P_1 = \partial_x, P_2 = \partial_y, P_3 = \partial_z

and these of course commute; the brackets with the M μνM_{\mu\nu} are

[M μν,P ρ]=η μρP νη νρP μ[M_{\mu\nu}, P_\rho] = \eta_{\mu\rho} P_\nu - \eta_{\nu\rho} P_\mu

Integration of the vector fields P ρP_\rho leads to the expected translations, e.g.,

exp(sP 0)(t,x,y,z)=(t+s,x,y,z)\exp(s P_0)(t, x, y, z) = (t + s, x, y, z)

The Poincare group in physics

The Poincaré group is basic to relativistic physics, since the fundamental principle of relativity is that physical laws are required to be invariant with respect to the action of the Poincaré group on spacetime. (There is some fine print here: in some cases, e.g., where physical laws are not invariant under space or time inversions, one must restrict to the action of the group 4SO +(1,3)\mathbb{R}^4 \rtimes SO^+(1, 3). If one is dealing with fermions, one considers invariance with respect to an action of the universal cover 4SU 2()\mathbb{R}^4 \rtimes SU_2(\mathbb{C}).)

Such physical laws may be classical or quantum, according to the description of physical states and observables in the theory. For example, in classical mechanics, pure states correspond to points in a symplectic phase space such as the cotangent bundle of Minkowski 4-space; in quantum mechanics, pure states are described by unit vectors in a suitable Hilbert space such as L 2( 4)L^2(\mathbb{R}^4). In either case, a relativistic theory will involve an action or representation of the Poincaré group, together with a structure which governs the dynamics of the theory, e.g., a Hamiltonian.

In the quantum case, a fundamental relativistic condition is that probability amplitudes ψ|ϕ\langle \psi|\phi \rangle, obtained by pairing an initial state ϕ\phi with a final state ψ\psi, are invariant under the action of the Poincaré group. This condition says

ψ|ϕ=gψ|gϕ\langle \psi|\phi \rangle = \langle g \cdot \psi|g \cdot \phi \rangle

for every gg in the Poincaré group. Thus the representation of the Poincaré group on Hilbert space is required to be unitary. Due to the noncompactness of the Poincaré group, unitary representations on finite-dimensional Hilbert spaces are scarce; one must really pass to unitary representations on infinite-dimensional Hilbert spaces to get anything interesting.

In particular, an elementary particle in quantum physics is sometimes defined to be an irreducible unitary representation of the Poincaré group on L 2( 4)L^2(\mathbb{R}^4).

Unitary representations

groupsymboluniversal coversymbolhigher coversymbol
orthogonal groupO(n)\mathrm{O}(n)Pin groupPin(n)Pin(n)Tring groupTring(n)Tring(n)
special orthogonal groupSO(n)SO(n)Spin groupSpin(n)Spin(n)String groupString(n)String(n)
Lorentz groupO(n,1)\mathrm{O}(n,1)\,Spin(n,1)Spin(n,1)\,\,
anti de Sitter groupO(n,2)\mathrm{O}(n,2)\,Spin(n,2)Spin(n,2)\,\,
conformal groupO(n+1,t+1)\mathrm{O}(n+1,t+1)\,
Narain groupO(n,n)O(n,n)
Poincaré groupISO(n,1)ISO(n,1)Poincaré spin groupISO^(n,1)\widehat {ISO}(n,1)\,\,
super Poincaré groupsISO(n,1)sISO(n,1)\,\,\,\,
superconformal group

Revised on August 19, 2016 03:48:32 by Urs Schreiber (