---
title: Elliptic Hill Four-Body Problem
url: https://www.emergentmind.com/topics/elliptic-hill-four-body-problem
type: topic
---

# Elliptic Hill Four-Body Problem

The Elliptic Hill Four-Body Problem (EH4BP) is a Hill approximation of the planar elliptic restricted four-body problem in a neighborhood of the smallest primary \(m_3\). In the formulation currently available in the literature, it is expressed as a small perturbation of the circular Hill four-body problem, with the eccentricity \(e\) of the primaries’ orbits as the small parameter [2507.22416]. A persistent point of clarification in the subject is that much of the Hill four-body literature concerns circular autonomous models, spatial extensions, or related time-periodic Hill-type systems, rather than a genuine elliptic Hill four-body problem in this strict sense [1412.3775].

## 1. Definition, scope, and terminological boundaries

A recurrent misconception is to identify any Hill-type four-body model with an elliptic one. The literature is explicit that several influential papers do **not** study an elliptic Hill four-body problem. The McGehee regularization study with oblate bodies analyzes a planar, circular, autonomous Hill approximation of the restricted four-body problem in a co-rotating frame, built on a fixed triangular relative equilibrium with oblateness, and states that it is “not an elliptic one” [2209.13625]. The foundational Hill approximation of Burgos-García and Gidea is likewise a circular autonomous base model derived from the equilateral restricted four-body problem, not a genuinely elliptic Hill four-body formulation [1412.3775]. The contact-geometry work on the spatial Hill four-body problem also treats the circular equilateral autonomous setting and distinguishes it from an elliptic generalization [2407.06927].

Within that landscape, the EH4BP appears as a specific local model: the Hill limit of the **planar elliptic restricted four-body problem** near the smallest primary, formulated as a nonautonomous periodic perturbation of the circular Hill four-body problem [2507.22416]. This gives the topic a dual character. On one side, it belongs to Hill theory through the \(m_3^{1/3}\)-type local blow-up. On the other, it belongs to elliptic restricted-body dynamics through its explicit periodic dependence on the true anomaly and the eccentricity parameter \(e\).

This delimitation matters structurally. Circular Hill four-body models are autonomous and possess a Jacobi-like first integral; the elliptic model studied so far is a \(2.5\)-degree-of-freedom Hamiltonian system with explicit time dependence, so the circular energy becomes a drifting quantity rather than an exact invariant [2507.22416].

## 2. Derivation from the planar elliptic restricted four-body problem

The starting point is the planar elliptic restricted four-body problem (ER4BP): a massless particle moves under the gravitational attraction of three primaries \(m_1>m_2>m_3\), where the primaries form an equilateral central configuration and move on elliptic orbits around their common center of mass [2507.22416]. In inertial Cartesian coordinates \((X,Y)\), with complex notation \(Z=X+iY\), the equations are
\[
\frac{d^2Z}{dt^2}=-G\sum_{i=1}^{3}\frac{m_i(Z-Z_i)}{\rho_i^3}, \qquad \rho_i=|Z-Z_i|.
\]

The primaries are taken along a homographic Lagrange solution \(q_i(t)=\phi(t)a_i\), with \(\phi(t)=re^{if}\). The radial variable \(r\) and true anomaly \(f\) satisfy the Kepler equations, and
\[
r=\frac{c^2/\lambda}{1+e\cos f}.
\]
Using pulsating coordinates
\[
Z=re^{if}z,
\]
and taking \(f\) as the new time variable, one obtains
\[
\frac{d^2z}{df^2}+2i\frac{dz}{df} = \frac{1}{1+e\cos f} \left( z-\sum_{i=1}^3 \frac{\mu_i(z-z_i)}{|z-z_i|^3} \right), \qquad \mu_i=\frac{m_i}{M}.
\]
Writing \(z=x+iy\), this becomes
\[
\ddot x-2\dot y=\Omega_x,\qquad \ddot y+2\dot x=\Omega_y,
\]
with
\[
\Omega(x,y;m_1,m_2,m_3) = \frac{1}{1+e\cos f} \left( \frac12(x^2+y^2)+\sum_{i=1}^3 \frac{m_i}{r_i} \right).
\]

After the canonical change
\[
\dot x=p_x+y,\qquad \dot y=p_y-x,
\]
the ER4BP Hamiltonian becomes
\[
H= \frac12(p_x^2+p_y^2)+yp_x-xp_y+\frac12(x^2+y^2) -\frac{1}{1+e\cos f} \left( \frac12(x^2+y^2)+\sum_{i=1}^3 \frac{m_i}{r_i} \right).
\]

For small eccentricity,
\[
\frac{1}{1+e\cos f}=1-e\cos f+\mathcal O(e^2),
\]
so
\[
H_{ER4BP}=H_{CR4BP} +e\cos(f)\left(\frac12(x^2+y^2)+\sum_{i=1}^3\frac{m_i}{r_i}\right) +\mathcal O(e^2).
\]

The Hill approximation is then taken near the smallest primary \(m_3\): the coordinates are shifted to \(m_3\), and the conformally symplectic scaling
\[
(x,y,p_x,p_y)\mapsto m_3^{1/3}(x,y,p_x,p_y)
\]
is applied. The limit \(m_3\to 0\) yields the Hamiltonian
\[
H_{EH4BP}=H_{CH4BP} +e\cos(f)\left(\frac12(x^2+y^2)+U(x,y)\right) +\mathcal O(e^2),
\]
where
\[
H_{CH4BP} = \frac12(p_x^2+p_y^2)+yp_x-xp_y-U(x,y),
\]
and
\[
U(x,y)= -\frac18 x^2+\frac{3\sqrt3}{4}(1-2\mu)xy+\frac58 y^2+\frac{1}{\sqrt{x^2+y^2}}, \qquad m_1=1-\mu,\quad m_2=\mu.
\]

After a planar rotation that diagonalizes the quadratic part, the circular Hamiltonian is written
\[
H_{CH4BP} = \frac12(p_x^2+p_y^2)+yp_x-xp_y-U_{rot}(x,y),
\]
with
\[
U_{rot}(x,y)=-ax^2-by^2+\frac{1}{\sqrt{x^2+y^2}},
\]
\[
\lambda_1=\frac32(1-d),\qquad \lambda_2=\frac32(1+d),\qquad d=\sqrt{1-3\mu+3\mu^2},
\]
and
\[
a=\frac12(1-\lambda_2),\qquad b=\frac12(1-\lambda_1).
\]
The corresponding effective potential is
\[
\Omega_{eff}(x,y)=\frac12(\lambda_2x^2+\lambda_1y^2)+\frac{1}{\sqrt{x^2+y^2}}.
\]

In this form, the EH4BP is a small time-periodic perturbation of the circular Hill four-body problem, with \(t=f\) and perturbation
\[
G(x,y,t)=\cos(t)H_1(x,y), \qquad H_1(x,y)=\frac12(x^2+y^2)+U_{rot}(x,y).
\]
That perturbative representation is the basis of the diffusion theory developed in the subject [2507.22416].

## 3. Circular limit and the invariant structures inherited from it

The circular Hill four-body problem is the unperturbed backbone of the elliptic theory. Its original derivation showed that the limiting Hamiltonian inherits dynamical features from both the restricted three-body problem and the restricted four-body problem, and that it reduces to the classical Hill three-body problem when \(\mu=0\) [1412.3775]. The planar circular model was then developed further through numerical continuation of periodic-orbit families, including direct and retrograde tertiary-centered orbits, Lyapunov families, and short- and long-period families around the additional equilibria \(L_3\) and \(L_4\) [1508.00875].

In the rotated circular Hamiltonian, the equilibria are
\[
L_1=\left(\frac{1}{\sqrt[3]{\lambda_2},0\right),\quad
L_2=\left(-\frac{1}{\sqrt[3]{\lambda_2},0\right),\quad
L_3=\left(0,\frac{1}{\sqrt[3]{\lambda_1}\right),\quad
L_4=\left(0,-\frac{1}{\sqrt[3]{\lambda_1}\right).
\]
For the parameter value
\[
\mu=0.00095,
\]
used in the EH4BP diffusion study, the model is interpreted in relation to the Sun–Jupiter system [2507.22416]. In the circular problem, \(L_1\) and \(L_2\) are the center-saddle equilibria that organize the transport mechanism, while \(L_3\) and \(L_4\) are the center-center equilibria relevant to other periodic-orbit families [1412.3775].

The EH4BP diffusion construction uses a narrow energy interval above the first bottleneck energy. In the circular problem,
\[
h_{L_1}=H_0(L_1)=H_0(L_2)=-2.16286.
\]
The numerical study takes
\[
[h_\alpha,h_\beta]=[-2.10446079,\,-2.07715457],
\]
a regime in which the Hill region has an inner domain around the tertiary connected to an outer domain through the necks near \(L_1\) and \(L_2\) [2507.22416].

Around each center-saddle point \(L_i\), \(i=1,2\), there is a Lyapunov family
\[
\lambda_i(h),\qquad h\in[h_\alpha,h_\beta],
\]
and the union
\[
\Lambda_0^i=\bigcup_{h\in[h_\alpha,h_\beta]}\lambda_i(h)
\]
is a \(2\)-dimensional normally hyperbolic invariant manifold with boundary. Each \(\Lambda_0^i\) carries symplectic action-angle coordinates \((I,\theta)\) on an annulus
\[
\mathbb A=\{(I,\theta)\mid I\in[I_\alpha,I_\beta],\ \theta\in\mathbb T^1\},
\]
with inner dynamics
\[
R^t(I,\theta)=(I,\theta+t\omega(I)),\qquad \omega(I)=\frac{1}{T(I)}.
\]
The diffusion mechanism requires that the stable and unstable manifolds of these NHIMs have both homoclinic and heteroclinic transverse intersections,
\[
W^u(\Lambda_0^i)\pitchfork W^s(\Lambda_0^i),\qquad
W^u(\Lambda_0^i)\pitchfork W^s(\Lambda_0^j),\ i\neq j,
\]
which are verified numerically in the energy range under consideration [2507.22416].

## 4. Arnold diffusion in the small-eccentricity regime

The defining result for the EH4BP is a geometric mechanism for Arnold diffusion in the unperturbed energy \(H_0\). The perturbation is written
\[
H_e(x,y,p_x,p_y,t)=H_0(x,y,p_x,p_y)+e\,G(x,y,t)+\mathcal O(e^2),
\]
with \(H_0=H_{CH4BP}\). Since \(H_e\) is time-periodic, \(H_0\) is no longer conserved, and its drift becomes the quantity of interest [2507.22416].

The key objects are the scattering maps associated with transverse homoclinic and heteroclinic channels. For a homoclinic channel \(\Gamma\),
\[
\sigma=\Omega^+_{\mid \Gamma}\circ (\Omega^-_{\mid \Gamma})^{-1},
\]
where \(\Omega^\pm\) are the wave maps from \(W^{s,u}(\Lambda)\) to \(\Lambda\). In the unperturbed CH4BP, the scattering map in action-angle coordinates has the form
\[
s_0(I,\theta)=(I,\theta+\Delta(I)),
\]
so it preserves the action exactly and acts only by an angle shift.

Under the elliptic perturbation, the scattering map becomes
\[
s_e(I,\theta)= s_0(I,\theta)+e\,J\nabla S\circ s_0(I,\theta)+\mathcal O(e^2),
\]
with
\[
J=\begin{pmatrix}0&-1\\1&0\end{pmatrix}.
\]
Consequently,
\[
I' - I = -e\frac{\partial S}{\partial \theta}(I,\theta+\Delta(I)) +\mathcal O(e^2).
\]
The generating function \(S\) is given by Melnikov-type improper integrals along the homoclinic or heteroclinic orbit. Thus the sign of \(-\partial_\theta S\) determines whether one application of the scattering map raises or lowers the action, and therefore raises or lowers the circular energy \(H_0\) [2507.22416].

Two diffusion arguments are provided. The first uses a single scattering map and Birkhoff’s Ergodic Theorem. If the average of
\[
-\frac{\partial S}{\partial\theta}(I,\theta+\Delta(I))
\]
over the annulus is nonzero, then there exist pseudo-orbits for which repeated scattering produces an \(O(1)\) increase in action after \(O(1/e)\) iterates. The second uses two scattering maps and, at each step, chooses one that increases the action:
\[
-\frac{\partial S^j}{\partial\theta}\circ \sigma_0^j(I,\theta)>c.
\]
This produces a pseudo-orbit of the iterated function system with order-one drift independent of \(e\) [2507.22416].

A shadowing lemma then converts such pseudo-orbits into true trajectories, subject to a recurrence hypothesis for the inner map or, alternatively, through a dichotomy argument if trajectories leave every neighborhood of the pseudo-orbit. The resulting theorem states that there exist \(e_0>0\), \(C>0\) such that for every \(0<e<e_0\), there is an orbit and a time \(T>0\) with
\[
|H_0(\Phi_e^T(z))-H_0(z)|>C.
\]
The theorem yields finite \(O(1)\) drift in the circular energy. It does **not** prove arbitrarily large energy growth [2507.22416].

The numerical verification is carried out for
\[
\mu=0.00095.
\]
For the homoclinic mechanism, the averaged positivity condition is verified numerically, with sample lower bounds leading to
\[
>0.015\cdot 4.50=0.0675.
\]
For the heteroclinic mechanism, the corresponding bound is
\[
>0.015\cdot 6.18=0.0927.
\]
The two-scattering-map construction yields stronger pointwise positivity bounds, and the paper remarks that the heteroclinic channels are numerically more efficient than the homoclinic ones [2507.22416].

## 5. Adjacent elliptic and time-periodic four-body models

The EH4BP sits inside a broader family of nonautonomous four-body models, but those models are not interchangeable.

A particularly close relative is the Hill Restricted 4-Body Problem (HR4BP) for the Sun–Earth–Moon system. It is a coherent \(\pi\)-periodic Hill-type model, not an elliptic Hill four-body problem in the strict sense, yet it exhibits the same structural consequences of periodic forcing: the Earth–Moon triangular equilibria are replaced by periodic orbits, notably the dynamical equivalent of \(L_4\), and nearby invariant objects become \(2\)-dimensional invariant tori rather than the invariant circles of the autonomous CR3BP [2402.18081]. The same model has also been used to compute resonant periodic-orbit families by Melnikov screening and pseudo-arclength continuation, with the resonance condition
\[
bT_g=aT^*,
\]
which is the natural closure condition in a forced Hill-type system [2402.19181]. These results do not define the EH4BP, but they show how time-periodic forcing reorganizes Hill-type dynamics through resonant periodic orbits, invariant tori, and manifold-mediated transport.

Another nearby body of work studies elliptic relative equilibria of restricted or full four-body problems without taking a Hill limit. In the restricted Lagrangian-triangle case, the linearized Poincaré map splits into Keplerian, elliptic Lagrangian, and essential parts, and the stability of the essential block depends on mass-curvature parameters \(\alpha,\beta\) and eccentricity \(e\) through \(\omega\)-Maslov index theory [1907.13475]. For planar four-body elliptic relative equilibria, symplectic reduction separates the Kepler two-body linearized subsystem from an \(8\)-dimensional essential subsystem with coefficients depending periodically on
\[
r(\theta)=\frac{p}{1+e\cos\theta},
\]
which is methodologically close to the anomaly-based reduction used in elliptic Hill settings [2104.11132]. This suggests that symplectic reduction, Floquet analysis, and Maslov-index methods are natural tools for further EH4BP stability theory, although that transfer remains an inference rather than a theorem.

A more global elliptic restricted four-body model also exists in the form of a bi-elliptic restricted problem with three equal-mass primaries on an elliptic Lagrangian homographic solution, written in a rotating-pulsating frame with true anomaly \(f\) as the independent variable [1506.06632]. That model yields a vertical equation
\[
\ddot z+\omega z=0,
\]
for quasi-planar motion, but it does not perform a Hill scaling and therefore is not an EH4BP.

## 6. Circular, spatial, and oblate foundations

The circular Hill four-body theory remains indispensable because the EH4BP is built as its small-eccentricity perturbation. The original Hill approximation derived the limiting Hamiltonian by translating to the small primary, scaling by \(m_3^{1/3}\), Taylor-expanding the distant-body potentials, and taking \(m_3\to 0\), so that the nearby Kepler singularity survives while the distant primaries remain only through a quadratic tidal field [1412.3775]. The planar periodic-orbit study showed that this circular model contains families \(f\), \(g\), \(a\), \(H_a\), \(H_b\), and short- and long-period families around \(L_3\) and \(L_4\), together with horizontal and vertical stability changes absent from the classical Hill three-body problem [1508.00875]. The spatial circular model further revealed that a second distant disturbing body changes the stability of familiar Hill families, alters their bifurcation structure, and creates new spatial periodic-orbit families absent in the classical spatial Hill problem [2112.00135].

The spatial circular problem also has a geometric regularity property: for every \(\mu\in[0,\frac12]\) and every energy below the first critical value
\[
H(L_{1/2})=-\frac32\sqrt[3]{\lambda_2},
\]
the bounded regularized energy surface is of contact type, with
\[
\widetilde{\Sigma}_c^b \cong (S^*S^3,\xi_{st}),
\qquad
\widetilde{\Sigma}_c^b|_{\mathrm{Fix}(\sigma)}\cong (S^*S^2,\xi_{st})
\]
in the spatial and planar cases respectively [2407.06927]. Those statements are proved in the autonomous circular setting and rely on fixed energy hypersurfaces; they do not transfer verbatim to the EH4BP.

Oblateness introduces another circular extension. The oblate-tertiary Hill four-body problem replaces the Newtonian singularity near the smallest primary by
\[
-\frac1r-\frac{c}{r^3}
\]
in the planar oblate case and creates six equilibria, including two \(z\)-axis equilibria that are absent when \(c=0\) [1812.10852]. The McGehee regularization analysis then shows that, for the planar oblate circular model, collision with the tertiary can be regularized by a blow-up adapted to quasi-homogeneous singularities, and that the collision manifold changes qualitatively when the tertiary oblateness coefficient crosses zero: for oblate tertiary there is a collision torus, at \(c_3=0\) there is a double saddle-node bifurcation, and for prolate tertiary there are no collisions [2209.13625]. These are precise autonomous statements. Their methodological value for the EH4BP lies in the blow-up, collision-manifold geometry, and quasi-homogeneous scaling, not in any established elliptic theorem.

Taken together, these circular, spatial, contact-geometric, and oblate results define the current state of the field. The EH4BP is no longer merely hypothetical, because a concrete planar small-eccentricity model and a diffusion mechanism have now been formulated [2507.22416]. At the same time, most of the detailed structural results in the surrounding literature—exact Jacobi geometry, contact-type energy levels, Levi-Civita or McGehee regularization, branch and block regularization, and the global taxonomy of periodic-orbit families—remain established primarily for circular autonomous Hill four-body problems [2209.13625]. This suggests that the EH4BP should presently be regarded as an emerging nonautonomous extension of a much more mature circular theory.

Source: https://www.emergentmind.com/topics/elliptic-hill-four-body-problem