We show that the group of linear canonical transformations in a 2N-dimensional phase space is the real symplectic group Sp(2N), and discuss its unitary representation in quantum mechanics when the N coordinates are diagonal. We show that this Sp(2N) group is the well-known dynamical group of the N-dimensional harmonic oscillator. Finally, we study the case of n particles in a q-dimensional oscillator potential, for which N = nq, and discuss the chain of groups Sp(2nq)⊃Sp(2n)×O(q). An application to the calculation of matrix elements is given in a following paper.
We show that the group of linear canonical transformations in a 2N-dimensional phase space is the real symplectic group Sp(2N), and discuss its unitary representation in quantum mechanics when the N coordinates are diagonal. We show that this Sp(2N) group is the well-known dynamical group of the N-dimensional harmonic oscillator. Finally, we study the case of n particles in a q-dimensional oscillator potential, for which N = nq, and discuss the chain of groups Sp(2nq)⊃Sp(2n)×O(q). An application to the calculation of matrix elements is given in a following paper.
Classical mechanics in phase space involves the concepts of Poisson brackets and canonical transformations. In (nonrelativistic) quantum mechanics, based on fundamental Cartesian variables, the kinematics is expressed via the Heisenberg canonical commutation relations (CCR’s). One now has unitary transformations on wave functions in Hilbert space. In both cases, Hamiltonians quadratic in the canonical variables lead to linear equations of motion whose solutions are linear canonical transformations. In all this, the key role is played by the family of real symplectic groups Sp (2 n ,ℝ). They are the ‘noncompact’ versions of the classical compact Lie groups Cl = USp (2 l ) , and their defining representations are in real even dimensional vector spaces. In addition to their significance in the canonical formulation of classical mechanics and its counterpart in quantum mechanics, these groups are important in several specific quantum mechanical problems as well as in classical and quantum optics. We look at some aspects of these groups in this chapter. While we describe the main features, we will not present complete proofs of all statements made, but a motivated reader should be able to supply the details and proceed to make practical uses of these groups. In both classical and quantum mechanics we have in mind only Cartesian fundamental variables. So for n degrees of freedom, or n canonical pairs, the classical phase space is R 2 n , while in quantum mechanics we have the Hilbert space H = L 2(ℝ n ). Since the symplectic groups are somewhat unfamiliar to most students of physics, we study the case n = 1 to begin with, and later go on to general n . The Group Sp (2,ℝ) For one degree of freedom with Cartesian variables, in classical mechanics we have real canonical variables q and p , varying from −∞ to ∞, and the Poisson brackets (PB) {q, p} = 1, {q, q} = {p, p} = 0. (8.1) The phase space is the two dimensional plane ℝ 2 . For the corresponding system in quantum mechanics we have an irreducible pair of hermitian operators q , p obeying the Heisenberg CCR’s In the usual Schrödinger solution to these relations we have a Hilbert space H = L 2 (ℝ) with q , p acting as follows: (Strictly speaking, since q and p are unbounded operators, their domains should be carefully specified; we assume such precautions are implicitly observed.) Now consider linear homogeneous transformations on q , p which in the classical case preserve the PB's (8.1) and in the quantum case the CCR's (8.2).
While groups of point transformations such as rotations, translations and permutations have been extensively used in quantum mechanics, we cannot say the same of the more general groups of canonical transformations. To begin with, for the latter groups we have to determine their unitary representation in appropriate spaces. But once this is achieved canonical transformations can play as important a role in quantum mechanics as point transformations. We illustrate this in relation with the harmonic oscillator in one and many dimensions and for one and many particles. We show how radial matrix elements with respect to harmonic oscillator states can be determined group theoretically both in the Schroedinger and Heisenberg pictures. We discuss the application of canonical transformations to the three body nuclear problem. We briefly outline in the concluding section applications to Coulomb problems, many body systems, accidental degeneracy, and generalized integral transforms.
The conventional quantization of the harmonic oscillator in terms of operators Q and P can be implemented with the help of irreducible unitary representations of the Heisenberg-Weyl group which acts transitively and effectively on the simply connected classical phase space S q,p ≊ 2 . In the description of the harmonic oscillator in terms of angle and action variables ϕ and I the associated phase space S ϕ,I corresponds to the multiply connected punctured plane 2 – {0}, on which the 3-dimensional symplectic group Sp(2, ) acts transitively, leaving the origin invariant. As this group contains the compact subgroup U (1) it has infinitely many covering groups. In the here relevant irreducible unitary representations (positive discrete series) the self-adjoint generator K o of U (l) represents the classical action variable I . It has the possible spectra n + k, n = 0,1,...; k > 0, where k depends on the covering group. This implies different possible spectra for the action variable Hamiltonian ℏωK 0 of the harmonic oscillator. On the other hand, expressing the operators Q and P (non-linearly) in terms of the three generators K 0 etc. of Sp (2, ) leads to the usual framework. Possible physical (experimental) implications and generalizations to higher dimensions are discussed briefly.
In the last few years an extensive literature has developed on linear canonical tranformations and their representation in quantum mechanics. Applications of these results have been made to clustering theory in nuclei, problems of accidental degeneracy, etc. In the present paper we wish to turn our attention to nonlinear canonical transformations. We show that by dealing with appropriate functions fα (α=1,...,2n) of xi, pi (i=1,...,n) rather than with these variables themselves, we can in principle set unambiguously the equations that determine the representation in quantum mechanics of the canonical transformation under study. This result holds when the old and new functions fα have the same spectrum. We discuss specific examples when this last condition is satisfied: nonlinear canonical transformations in the radial variable that were obtained from projection of linear ones in higher-dimensional spaces; canonical transformations that take us from one Hamiltonian to another with the same spectrum, be this one continuous or discrete; canonical transformations that relate two sets of integrals of motion (which include the Hamiltonians) when we are dealing with phase spaces of dimensions higher than 2, etc. We discuss briefly, in the concluding section, the possibility of extending our analysis to canonical tranformations that do not conserve the spectrum of the relevant operators.
We consider pseudounitary quantum systems and discuss various properties of pseudounitary operators. In particular we prove a characterization theorem for block-diagonalizable pseudounitary operators with finite-dimensional diagonal blocks. Furthermore, we show that every pseudounitary matrix is the exponential of i=−1 times a pseudo-Hermitian matrix, and determine the structure of the Lie groups consisting of pseudounitary matrices. In particular, we present a thorough treatment of 2×2 pseudounitary matrices and discuss an example of a quantum system with a 2×2 pseudounitary dynamical group. As other applications of our general results we give a proof of the spectral theorem for symplectic transformations of classical mechanics, demonstrate the coincidence of the symplectic group Sp(2n) with the real subgroup of a matrix group that is isomorphic to the pseudounitary group U(n,n), and elaborate on an approach to second quantization that makes use of the underlying pseudounitary dynamical groups.