Research paperClassical multiseparable Hamiltonian systems, superintegrability and Haantjes geometry
Introduction
The prominence of integrable models in many areas of pure and applied mathematics and theoretical physics has motivated, in the last decades, a resurgence of interest in the algebraic and geometric structures underlying the notion of integrability. The study of the geometry of Hamiltonian integrable systems has a long history, dating back to the classical works by Liouville, Jacobi, Stäckel, Eisenhart, Arnold, etc. The approaches proposed in the literature are intimately related with the problem of the determination of suitable coordinate systems guaranteeing the additive separation of the Hamilton-Jacobi (HJ) equation.
Recently, many new ideas coming from differential and algebraic geometry, topology and tensor analysis, have contributed to the formulation of important approaches such as the theory of bi-Hamiltonian systems [23], the Lenard-Nijenhuis geometry [30] and the theory of Dubrovin-Frobenius manifolds [10]. These theoretical developments shed new light on the multiple connections among integrability, topological field theories, singularity theory, co-isotropic deformations of associative algebras, etc. Besides, several integrable models, both classical and quantum ones, have recently been discovered, in particular in the domain of superintegrability.
Superintegrable systems are a special class of integrable systems which possess surprisingly rich algebraic and geometric properties [32], [33], [38], [42], [52], [53]. Essentially, they possess more independent integrals of motion than degrees of freedom. If a system with degrees of freedom admits functionally independent integrals of motion (the maximal number allowed), it is said to be maximally superintegrable; if it admits integrals, then is minimally superintegrable.
Among the most famous examples of superintegrable models we mention the classical harmonic oscillator, the Kepler potential, the Calogero-Moser potential, the Smorodinsky-Winternitz systems and the Euler top. The presence of “hidden” symmetries, expressed by integrals which are second (or higher) degree polynomials in the momenta, usually allows us to determine the dynamical behaviour of superintegrable models. In the maximal case, the bounded orbits are closed and periodic [38]. As is well known, the phase space topology is also very rich: it can be described in terms of a symplectic bifoliation, determined by the standard Liouville-Arnold invariant fibration [1] of Lagrangian tori, and complemented by a co-isotropic polar foliation [14], [38].
Although the present work focuses on classical systems, we point out that quantum superintegrable systems also possess many relevant properties. Exact quantum solvability of a Hamiltonian system, related to the existence of suitable Lie algebras of raising and lowering operators, could be regarded as the quantum analogue of the classical notion of maximal superintegrability [52].
A fundamental class of integrable models is the separable one: these models are characterized by the fact that one can find at least a system of canonical coordinates in which the corresponding Hamilton-Jacobi equation takes the additively separated form
The problem of finding separating variables for integrable Hamiltonian systems has been extensively investigated. In 1904, Levi-Civita proposed a test which permits to establish whether a given Hamiltonian separates in an assigned coordinate system [22]. Another important result, due to Benenti [3], states that a family of Hamiltonian functions are separable in a set of canonical coordinates if and only if they are in separable involution, i.e. they satisfy the relationswhere no summation over is understood. However, such theorem as well as the Levi–Civita test are not constructive, and do not allow us to determine a complete integral of the Hamilton–Jacobi equation.
A constructive approach for the determination of separating variables was given by Sklyanin [47] within the framework of Lax systems. The Hamiltonian functions are separable in a set of canonical coordinates if there exist suitable equations, called the Jacobi-Sklyanin separation equations for , having the formThese equations allow us to construct a solution of the HJ equation. In fact, by solving Eq. (3)with respect to , we get the additively separated form (1).
Nevertheless, the three above-mentioned criteria of separability are not intrinsic: in order to be applied, they require the explicit knowledge of the local chart . To overcome such a drawback, in the last decades the modern theory of separation of variables (SoV) has been conceived in the context of symplectic and Poisson geometry; in particular, the bi-Hamiltonian theory has offered a fundamental geometric insight into the theory of integrable systems [13], [30].
The main purpose of the present work is to establish a novel relationship between the theory of separable Hamiltonian systems and the geometry of an important class of tensor fields, the Haantjes tensors, introduced in [19] as a relevant, natural generalization of the notion of Nijenhuis tensors [40], [41]. The class of Nijenhuis tensors plays a significant role in differential geometry and the theory of almost-complex structures, due to the celebrated Newlander-Nirenberg theorem [39].
Our approach is based on the notion of manifolds, introduced in [51] by analogy with the theory of manifolds [13], [30] for finite-dimensional Hamiltonian systems (see also [15], [27] and [28] for a treatment of integrable hierarchies of PDEs). Essentially, an manifold is a symplectic manifold endowed with an algebra of (1,1) tensor fields with vanishing Haantjes torsion, which are compatible with the symplectic structure. Under the hypotheses of the Liouville-Haantjes (LH) theorem proved in [51], a non-degenerate Hamiltonian system is completely integrable in the Liouville-Arnold sense if and only if it admits a structure.
In our context, Haantjes chains represent in the Haantjes framework the generalization of the notion of Lenard-Magri chain [24] and of generalized Lenard chain [12], [26] defined previously for quasi-bi-Hamiltonian systems [36], [37]. By means of these structures, one obtains a complete description of the integrals of the motion of a system in terms of the associated Haantjes operators.
The problem of SoV can also be recast and studied in our approach. Precisely, as stated in Theorem 2 below, if an integrable system admits a semisimple structure, one can derive a set of coordinates, that we shall call the Darboux-Haantjes (DH) coordinates, representing separation coordinates for the Hamilton-Jacobi equation associated with the system. In these coordinates, the symplectic form takes a Darboux form, and the operators of the Haantjes algebra take all simultaneously a diagonal form. As we will show in Theorem 3, and in the examples of Sections 8 and 10, multiseparable systems possess different Haantjes structures associated in a nontrivial way with their separation coordinates.
In the study of separable Hamiltonian systems, the theory of manifolds has proved to be a powerful tool. Since any semisimple Haantjes algebra admits a Nijenhuis generator, we deduce that the coordinates in the Haantjes scenario are also Darboux-Nijenhuis (DN) coordinates of the theory.
In the semisimple case, we shall prove that there is a one-to-one correspondence between a given manifold and an equivalence class of structures (see Section 3.7). Interestingly enough, given a semisimple manifold, one can find Nijenhuis generators which fulfill both the algebraic and the differential compatibility conditions, required by the theory.
However, from a general, theoretical point of view, the two theories are not equivalent. Indeed, in the non-semisimple case, there are integrable systems, as the Post-Winternitz system (discussed in [51]), which admit a non-Abelian structure endowed with three Haantjes generators. In other words, there exist manifolds with several generators. In these cases, an alternative description in terms of a standard structure is not available, since in the approach the existence of a unique Nijenhuis tensor is assumed.
This article is organized into two parts. In the first one, including Sections 2 and 3, for the sake of self-consistency we briefly summarize the notions necessary for the study of multiseparable systems. Precisely, in Section 2, the basic definitions concerning Nijenhuis and Haantjes tensors are proposed; in Section 3, the Haantjes geometry is reviewed. In particular, the notions of Haantjes algebras, manifolds and Darboux-Haantjes coordinates are revised. In the second part, starting from Section 4, we shall propose the original results of our work. In Section 4, we propose the main theorem concerning the existence of Haantjes structures for separable systems. In Section 5, as a direct application of the theory previously developed, we solve the problem of SoV for a family of Drach-Holt type systems, that were previously considered to be non-separable. Interestingly enough, the new separating variables we found are defined in the full phase space. In Section 6, this theorem is extended to the case of multiseparable (and superintegrable) models. We propose in Section 7 a novel geometric construction: a lift of operators from the configuration space of dimension two to , which generalizes the standard Yano lift [58]. By means of our procedure, a Haantjes operator can be lifted into another Haantjes operator (unlike the Yano lift, which only preserves Nijenhuis operators, but not the Haantjes ones). Section 8 is devoted to the study of the Haantjes structures for the Smorodinsky-Winternitz systems in the plane, whereas Section 9 deals with the study of the anisotropic oscillator. In Section 10, the manifolds associated with certain important multiseparable systems in three dimensions are determined. Future research perspectives are discussed in the final Section 11.
Section snippets
Nijenhuis and Haantjes operators
In this Section, we review some basic algebraic results concerning the theory of Nijenhuis and Haantjes tensors. For a more complete treatment, see the original papers [19], [40] and the related ones [17], [41].
Integrability
In order to formulate our approach to separability, we shall review the relationship between Haantjes geometry and integrability. A more detailed treatment as well as the proofs of the statements reviewed here are available in Refs.[50], [51].
We start recalling that a reference frame is a set of vector fields satisfying the following property: given an open set , the frame represents a basis of the tangent space . Given two frames and , assume that
Main theorem
The next theorem represents our main result concerning the existence of separation variables in the theory of manifolds. Theorem 2 Jacobi-Haantjes Let be an Abelian semisimple manifold of class and be a set of functions belonging to a Haantjes chain generated by a function via a basis of operators . Then, each set of DH coordinates provides us with separation variables for the Hamilton–Jacobi equation associated with each function . Conversely, if is a symplectic
Separation of variables for a Drach-Holt type system
The approach proposed in this paper offers an effective procedure to construct algorithmically separating variables admitted by Hamiltonian integrable systems. As a paradigmatic example, we shall study the case of a system showing an irrational dependence on its coordinates. Precisely, we shall consider a three-parametric deformation of the Holt potential, that has been introduced in [8]:It is integrable in the manifold , with a
Multiseparable systems and Haantjes geometry
A particularly interesting instance of the previous theory is represented by the case of multiseparable systems. They are Hamiltonian systems that can be separated in more than one coordinate system in their phase space. Fundamental physical examples of multiseparable systems are the -dimensional harmonic oscillator and the Kepler system. Another important class is represented by the four Smorodinsky-Winternitz systems, which are the only systems in the Euclidean plane admitting orthogonal
Construction of Haantjes operators in : A novel geometric lift
The problem of constructing Haantjes operators on an -dimensional manifold is, in general, a hard one, as it entails solving a system of nonlinear PDE of first order in the unknown components of the operators we wish to determine.
However, when , being the configuration space of a given mechanical system, one can plan to simplify the problem of constructing a manifold by means of a geometric procedure which allows us to lift a Haantjes operator to a
Haantjes structures for multiseparable systems in
The Smorodinsky-Winternitz (SW) systems are a family of superintegrable systems defined in the Euclidean plane , which were introduced first as quantum-mechanical systems in [16], [31], [59] and later studied from a group theoretical point of view in [46], [52]. They are all multiseparable in and admit three independent integrals of motion, expressed in terms of second-degree polynomials in the momenta. Also, they are separable in at least two different orthogonal coordinate systems in
Anisotropic oscillator with Rosochatius terms
We shall determine now the structures of an important physical model: the two-dimensional anisotropic oscillator with Rosochatius terms [45]. This system in the general case with degrees of freedom has been studied in [43], [44], [60], where its maximal superintegrability was established. In particular, in [43] the higher-order (missing) integral was determined by means of a geometric approach based on the Marsden-Weinstein reduction procedure. The anisotropic oscillator on curved spaces
structures for multiseparable systems in
We shall study in detail three relevant examples of multiseparable systems in the Euclidean space , in the context of Haantjes geometry. One of them is maximally superintegrable, the other two are minimally superintegrable ones [11]. For a different treatment, in the framework of the Killing-Stäckel theory, see [4], [9]. In the following analysis, , , denote the components of the angular momentum in the Cartesian frame.
Future perspectives
The problem of finding separating variables for integrable Hamiltonian systems is certainly among the most relevant ones of Classical Mechanics. The theory of manifolds offers a twofold contribution to this fundamental problem, being of both conceptual and applicative nature.
From a conceptual point of view, we have shown that the existence of semisimple structures ensures that of separating variables. Precisely, as stated in Theorem 2, if an integrable system admits a semisimple Abelian
CRediT authorship contribution statement
Daniel Reyes Nozaleda: Conceptualization, Writing – original draft. Piergiulio Tempesta: Conceptualization, Writing – original draft. Giorgio Tondo: Conceptualization, Writing – original draft.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgement
The research of D. R. N. has been supported by the Severo Ochoa Programme for Centres of Excellence in R&D (CEX2019-000904-S), Ministerio de Ciencia, Innovación y Universidades, Spain. The research of P. T. has been supported by the research project PGC2018-094898-B-I00, Ministerio de Ciencia, Innovación y Universidades, Spain, and by the Severo Ochoa Programme for Centres of Excellence in R&D (CEX2019-000904-S), Ministerio de Ciencia, Innovación y Universidades, Spain.
The research of G. T. has
References (60)
- et al.
The anisotropic oscillator on curved spaces: a new exactly solvable model
Ann Phys
(2016) - et al.
Integrable Hamiltonian hierarchies
Lect Not Phys
(2008) Sulla integrazione dell’equazione di Hamilton–Jacobi per separazione di variabili
Math Ann
(1904)A simple model of the integrable Hamiltonian equation
J Math Phys
(1978)- et al.
A non separable quantum superintegrable system in 2D real euclidean space
J Phys A
(2011) Separation of variables: new trends
Prog Theor Phys Suppl
(1995)- et al.
Generalized lenard chains, separation and superintegrability
Phys Rev E
(2012) - et al.
Mathematical aspects of classical and celestial mechanics
(1997) Separability structures on Riemannian manifolds
Lect. Notes Math.
(1980)Orthogonal separable dynamical systems
Math Publ, Silesian Univ Opava
(1993)
Necessary conditions for existence of non-degenerate Hamiltonian structures
Commun Math Phys
General algebraic identities for the Nijenhuis and Haantjes torsions
Izvestya Mathematics
Two degrees of freedom quasi–bi–Hamiltonian systems
J Phys A
Higher-order superintegrability of a holt related potential
J Phys A
Separation of variables and Killing tensors in the euclidean three-space
Geometry of 2D topological field theories
Superintegrability in classical mechanics
Phys Rev A
Bi–Hamiltonian systems and separation of variables: an example from the Boussinesq hierarchy
Theor Math Phys
Separation of variables for bi-Hamiltonian systems
Math Phys Anal Geom
Superintegrable hamiltonian systems: geometry and perturbations
Acta Appl Math
Differential-geometric approach to the integrability of hydrodynamics chains: the Haantjes tensor
Mat Ann
On higher symmetries in quantum mechanics
Phys Lett
Theory of vector-valued differential forms. part i
Indag Mathematicae
On -forming sets of eigenvectors
Indag Mathematicae
Bi-Hamiltonian structures and Stäckel separability
J Geom Phys
Computational linear and commutative algebra
A geometrical approach to nonlinear solvable equations
Geometry and soliton equations
Atti Accad. Sci. Torino Suppl. physics
Lenard chains for classical integrable systems
Theoret and Math Phys
Recursion operators and Frobenius manifolds
SIGMA
Cited by (10)
Generalized Nijenhuis Torsions and Block-Diagonalization of Operator Fields
2023, Journal of Nonlinear SciencePolarization of generalized Nijenhuis torsions
2023, Contemporary Mathematics