Geometric optics on phase space.

Publication date 4 November 2004

aberration coma distorsion Hamiltonian optics optical phase space paraxial optics reflection refraction screen coordinates

Mathematics Subject Classification ID

Applications of Lie groups to the sciences; explicit representations (22E70) Applications of differential geometry to physics (53Z05) Research exposition (monographs, survey articles) pertaining to optics and electromagnetic theory (78-02) Geometric optics (78A05) Finite-dimensional groups and algebras motivated by physics and their representations (81R05)

Review text

This book is devoted to the Hamiltonian (geometric) model of the light. At the very beginning the optical phase space is identified with a four-dimensional manifold which has the topology of the cotangent bundle \(T^*S^2\) of the two-sphere \(S^2\). By the classical Darboux theorem any symplectic (phase space) manifold is locally symplectomorphic to a standard one for the respective dimensions and in the case under consideration this is \({\mathbb R}^4\) with its canonical symplectic structure. From the purely mathematical point of view one of the main problems here is to find the orbits the groups of which inside their dual Lie algebra spaces coincide with \(T^*S^2\). In these terms the answer is given in the paper by \textit{N. Mukunda, R. Simon} and \textit{E. Sudarshan} [On the orbits in the Lie algebras of some (pseudo) orthogonal groups, Indian J. Pure Appl. Math. 19, 91--124 (1988; Zbl 0643.22008)]. A much more important issue is, however, to find the explicit action of the generators on these orbits as the real power of the Hamiltonian mechanics is hidden in the canonical transformations which mix the so called canonical coordinates and preserve the equations of motion. Depending on the explicit form of the Hamiltonian the infinite dimensional group of symplectomorphisms reduces in physically important cases to a finite-dimensional Lie group. Specially, for the case of paraxial optics the corresponding Lie group is Sp\((4,{\mathbb R})\). By conjugation of its linear canonical action on \({\mathbb R}^4\) via direct and inverse diffeomorphisms between the optical (curved) phase space and the flat space \({\mathbb R}^4\) one gets a (local!) nonlinear canonical action of the group Sp\((4,{\mathbb R})\) on the phase space. This picture is made clear within the first ninety pages of the book. The rest of it is devoted to profound exploitation of this rich group structure for the purposes of optics. Formally, it is structured in four Parts and fifteen Chapters supplemented by three Appendices. Out of \(165\) References in the book more than \(40\) belong to the author which makes clear the fact that this is an original monograph on the subject. It covers a lot of the mathematical formalism of Lie algebras and Lie groups like Cartan's classification of the semisimple Lie algebras and introduces from scratch such fundamental notions like rank, root diagrams, weights, etc. Special attention is also paid to the Weyl unitary trick, contractions and the exponential mapping, all of which are carefully explained and exemplified in lower dimensions. The generalizations to higher dimensional cases are also presented but everything is done keeping an eye on its relevance to the subject of the book. As an example, the accidental isomorphisms between some low-dimensional Lie algebra are described in detail but the respective meaning and consequences for optics are pointed out for all dimensions. The same refers to the Gauss and Iwasawa decomposition of various matrices. In this manner one can be easily convinced that the maximal compact subgroup is associated with the fractional Fourier transformation, that the abelian subgroup plays the role of pure magnifier and that the nilpotent subgroups represent thin lenses. In optics an imager is defined as an optical system the output of which is a sharp image of the object. Its group-theoretical analog is the solvable subgroup. Much of the roots of these notions are encoded in the plane paraxial optics with the phase space \({\mathbb R}^2\) and the respective symplectic group Sp\((2,{\mathbb R})\) which is treated in great detail as well. One has to notice that on the Lie algebra level \(\mathfrak{sp}(2,{\mathbb R}) \simeq \mathfrak{sl}(2,{\mathbb R}) \simeq \mathfrak{su}(1,1)\simeq \mathfrak{so}(2,1)\) while the corresponding Lie groups of the first three Lie algebras are covering spaces for the last one. This fact has also optical implications as their parametrizations provide in a quite natural way different optical transformations. Let us mention as well that some years ago the reviewer [\textit{I. Mladenov}, Saxon-Hutner theorem for periodic multilayers, C. R. Acad. Bulg. Sci. 40, No. 10, 35--38 (1987; Zbl 0646.34053)] considered the one-dimensional propagation of plane polarized monochromatic light for which Maxwell's system of equations reduces to a linear second order ordinary differential equation. The latter can be viewed as a one-dimensional Schrödinger equation for which the corresponding transfer matrix belongs just to one of the isomorphic groups SL\((2,{\mathbb R}) \simeq \text{SU}(1,1)\) (depending on the chosen basis in the space of wave functions). This work was mentioned here as an example of the general trend towards establishing a close relationship between optics and quantum mechanics which is really profound and thus opens the possibility to transfer ideas and results in both directions. The author had included such an example as well - the so called Maxwell fish-eye problem which exhibits the same hidden symmetries SO(4) for the elliptic and SO\((3,1)\) for the hyperbolic cases as the Kepler problem in classical mechanics. All the above refers to the paraxial regime but beyond it and ideal optical systems lie the real optics and the so called metaxial regime. The group-theoretical treatment in this situation relies on generators acting in the phase space which are polynomials truncated to a common degree that reflects the aberration order. Computations become really hard and one needs to use some computer algebra system like \texttt{Reduce} or as the author had done (cf. Reference [161] in the book), to create its own in order to compose the previously computed by hand aberrating systems with free parameters. We will not dwell anymore on these developments despite that these calculations are enormously important for electronic microscopes, accelerators design, etc. However, just for the reader's appreciation we will mention an alternative approach based this time on the Heisenberg Lie group (let us stress that the three-dimensional Heisenberg group is a closed subgroup of the symplectic group Sp\((4,{\mathbb R})\) and that this is proved in the book under review). Details can be found in the book by \textit{W. Schempp} [Magnetic resonance imaging: mathematical foundations and applications (Wiley, New York) (1998; Zbl 0930.92015)]. To end this review we remark that the exposition is quite clear and pedagogical and can be recommended for both a lecture course and a self study because each Part/Chapter opens with its own Introduction. The text is supplemented by a representative list of important papers and books on the subject and an Index which is exhaustive and very well organized. Last but not least - typesetting and the overall design of the book and the fine graphics included inside are quite nice.

Recommendations

Cited in

(44)

Describes a project that uses

Uses Software

REDUCE

This page was built for publication: Geometric optics on phase space.

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4825140)