The 1:+2 / 1:-2 resonance

On the linear level elliptic equilibria of Hamiltonian systems are mere superpositions of harmonic oscillators. Non-linear terms can produce instability, if the ratio of frequencies is rational and the Hamiltonian is indefinite. In this paper we study the frequency ratio 1/2 and its unfolding. In particular we show that for the indefinite case (1:-2) the frequency ratio map in a neighbourhood of the origin has a critical point, i.e. the twist condition is violated for one torus on every energy surface near the energy of the equilibrium. In contrast, we show that the frequency map itself is non-degenerate (i.e. the Kolmogorov non-degeneracy condition holds) for every torus in a neighbourhood of the equilibrium point. As a byproduct we are able to obtain another proof of fractional monodromy in the 1:-2 resonance.

💡 Research Summary

The paper investigates the dynamics of Hamiltonian systems near an elliptic equilibrium when the linearized frequencies are in a 1:2 ratio, focusing on the two distinct cases distinguished by the sign of the quadratic part of the Hamiltonian: the definite case (1:2) and the indefinite case (1:-2). In the linear approximation an elliptic equilibrium is simply a superposition of independent harmonic oscillators, but once nonlinear terms are added the situation changes dramatically if the frequency ratio is rational. In particular, when the quadratic form is indefinite, resonant interactions can generate instability.

The authors employ normal‑form theory to bring the Hamiltonian into a simplified expression that isolates the resonant terms up to a prescribed order. Introducing action variables (I_1, I_2) they define the frequency map (\omega(I)=\bigl(\partial H/\partial I_1,\partial H/\partial I_2\bigr)). The Jacobian matrix (D\omega) encodes the “twist” of the invariant tori: the twist condition (non‑vanishing determinant of (D\omega)) is a standard hypothesis in the classical KAM theorem. By explicit computation of (D\omega) for the 1:-2 resonance the authors show that there is always a single point ((I_1^\ast,I_2^\ast)) on each nearby energy surface where (\det D\omega=0). Consequently, on every energy level close to the equilibrium there exists exactly one torus that violates the twist condition. This torus is “non‑twist” and therefore lies outside the scope of the usual (twist‑based) KAM theorem.

Despite this local failure of the twist condition, the global frequency map remains non‑degenerate in the Kolmogorov sense: the Hessian of the Hamiltonian with respect to the actions never loses rank, and the gradient of (\omega) never vanishes identically. Hence the Kolmogorov non‑degeneracy condition, which underlies the more general KAM theorem, holds for every torus in a neighbourhood of the equilibrium. In practical terms, while a single non‑twist torus appears on each energy surface, the surrounding family of invariant tori is still robust under small perturbations, and a large measure of quasi‑periodic motions survives.

A secondary but noteworthy contribution is a new proof of fractional monodromy in the 1:-2 resonance. Traditional explanations of fractional monodromy rely on the topology of the energy‑momentum fibration, showing that transporting a basis of cycles around a loop encircling the singular value yields a transformation by a rational matrix rather than an integer one. The authors reinterpret this phenomenon through the lens of the frequency map: the critical point of (\omega) creates a branch‑cut–like structure in action space, and when one follows a closed loop around the critical value the action variables undergo a rational shift. This rational shift manifests as fractional monodromy, providing a clear geometric picture that ties the monodromy directly to the twist‑failure point.

In summary, the paper establishes two complementary results for the indefinite 1:-2 resonance: (1) the twist condition is violated on exactly one torus per nearby energy surface, and (2) the Kolmogorov non‑degeneracy condition holds everywhere in a neighbourhood of the equilibrium. The coexistence of these properties clarifies why KAM theory can still be applied in a weakened form, and it deepens our understanding of resonant Hamiltonian dynamics, especially in contexts such as molecular vibrations, celestial mechanics, and other physical systems where 1:2 resonances with mixed sign quadratic forms naturally arise.