Perturbed \((2n - 1)\)-dimensional Kepler problem and the nilpotent adjoint orbits of \(\operatorname{U}(n, n)\) (Q2217964)

Consider the space \[\mathcal{N}:=\{(\mathcal{Z},z)\in\mathfrak{gl}(2n,\mathbb{C})\times \mathrm{Gr}(n,\mathbb{C}^{2n})\mid \mathrm{Im}(\mathcal{Z})\subset z \subset \mathrm{Ker}(\mathcal{Z})\},\] which is a bundle of certain nilpotent matrices over the Grassmannian \(\mathrm{Gr}(n,\mathbb{C}^{2n})\) of \(n\)-dimensional complex vector spaces in \(\mathbb{C}^{2n}\). Denote by \(\operatorname{U}(n,n)\) the subgroup of \(\mathrm{GL}(2n,\mathbb{C})\) made of matrices that preserve the scalar product \[\langle v,w\rangle = v^+ \phi w,\quad v,w\in \mathbb{C}^{2n}.\] Here, \(\phi\in\mathfrak{gl}(2n,\mathbb{C})\) is a Hermitian matrix of signature \((+,\ldots,+,-,\ldots,-)\) and the superscript \(+\) denotes the Hermitian transpose. The author notes that there exists an involution on \(\mathcal{N}\) whose fixed point set, denoted \(\mathcal{N}_0\), is isomorphic to \(T^*\operatorname{U}(n)\). Using this isomorphism, one can push onto \(\mathcal{N}_0\) a natural action of \(\operatorname{U}(n,n)\) on \(T^*\operatorname{U}(n)\), see Proposition 3.2. Furthermore, the action of \(\operatorname{U}(n,n)\) preserves a symplectic form on \(T^*\operatorname{U}(n)\), and it admits a momentum map into the Lie-Poisson space \(\mathfrak{u}(n,n)^*\). This momentum map can be used to classify \(\mathrm{Ad}(\operatorname{U}(n,n))\)-orbits inside \(\operatorname{pr}_1(\mathcal{N}_0)\subset \mathfrak{u}(n,n)\), where we use the projection \(\operatorname{pr}_1:\mathcal{N}\to \mathfrak{gl}(2n,\mathbb{C})\) onto the first component, see Proposition 3.3. The remainder of this paper focuses on several models for the \(\mathrm{Ad}(\operatorname{U}(n,n))\)-orbit \(\mathcal{N}_{0,1}\subset pr_1(\mathcal{N}_0)\) which consists of the rank 1 nilpotent elements of \(\mathfrak{u}(n,n)\). Section 4 deals with the study and comparison of various Hamiltonian systems defined on theses models of \(\mathcal{N}_{0,1}\), which are all equivalent. In Section 5, it is explained how one of these systems leads to the (usual) 3-dimensional Kepler system when \(n=2\). This motivates the terminology of \textit{\((2n-1)\)-dimensional Kepler Problem} used in the title for the equivalent forms of the \(\operatorname{U}(n,n)\)-Hamiltonian system considered previously. Finally, a generalisation of this system is studied in Section 6 based on [the author and \textit{E. Wawreniuk}, J. Phys. A, Math. Theor. 49, No. 26, Article ID 265202, 33 p. (2016; Zbl 1353.81077)]. It is proved that this generalisation can be integrated by quadratures.