Definition

Let be a finite-dimensional Lie group with g\mathfrak g, and let λg\lambda\in\mathfrak g^*. The coadjoint orbit through λ\lambda is the

Oλ=Gλ={Adgλ:gG}g\mathcal O_\lambda=G\cdot\lambda =\{\operatorname{Ad}_g^*\lambda:g\in G\}\subseteq\mathfrak g^*

for the . If Gλ={gG:Adgλ=λ}G_\lambda=\{g\in G:\operatorname{Ad}_g^*\lambda=\lambda\} is the , then Oλ\mathcal O_\lambda carries the canonical immersed-manifold structure for which G/GλOλG/G_\lambda\to\mathcal O_\lambda is a GG-equivariant . Its dimension is dimGdimGλ\dim G-\dim G_\lambda.

Tangent space and homogeneous structure

At μOλ\mu\in\mathcal O_\lambda, the is

TμOλ={adXμ:Xg}.T_\mu\mathcal O_\lambda =\{\operatorname{ad}_X^*\mu:X\in\mathfrak g\}.

Its kernel presentation identifies this tangent space with g/gμ\mathfrak g/\mathfrak g_\mu, where gμ\mathfrak g_\mu is the Lie algebra of the stabilizer. Thus each coadjoint orbit is a .

Kirillov–Kostant–Souriau form

Every coadjoint orbit has a canonical invariant symplectic form. With Xμ#=ddt0Adexp(tX)μX^\#_\mu=\frac{d}{dt}|_{0}\operatorname{Ad}_{\exp(tX)}^*\mu, one common convention is

ωμ(Xμ#,Yμ#)=μ,[X,Y].\omega_\mu(X^\#_\mu,Y^\#_\mu)=\langle\mu,[X,Y]\rangle.

This is well defined, closed, and nondegenerate, so Oλ\mathcal O_\lambda is a . This structure underlies the described in Kirillov, “The Method of Orbits”.

Examples and sign conventions

For an every coadjoint orbit is a single point. After identifying so(3)R3\mathfrak{so}(3)^*\cong\mathbb R^3, the nonzero coadjoint orbits of SO(3)\mathrm{SO}(3) are spheres. Authors who define or with the opposite sign write the negative of the displayed symplectic form; the underlying orbit is unchanged.

References
  1. A. A. Kirillov, Elements of the Theory of Representations, Springer, 1976. DOI record. Relevant: “The Method of Orbits.”
  2. Ana Cannas da Silva, Lectures on Symplectic Geometry, Springer, 2008. DOI record. Relevant: Hamiltonian group actions and coadjoint orbits.