---
title: Extended Brown Hamiltonian Dynamics
url: https://www.emergentmind.com/topics/extended-brown-hamiltonian
type: topic
---

# Extended Brown Hamiltonian Dynamics

Searching arXiv for the specified papers and closely related work on the extended Brown Hamiltonian and Brown corrections.
The extended Brown Hamiltonian is a secular, integrable Hamiltonian model for weakly hierarchical three-body systems in the test-particle limit, developed to describe modified von Zeipel–Lidov–Kozai (ZLK) oscillations when classical double-averaged quadrupole theory and the classical Brown correction are no longer sufficiently accurate. In the formulation introduced by Lei and Grishin, the model retains the quadrupole-order disturbing function but incorporates nonlinear short-period effects accumulated from both the outer and inner orbits, yielding a closed-form Hamiltonian and a closed-form transformation between mean and osculating elements [2505.13780]. Subsequent work analyzed the model’s phase-space structure, the associated modified Lidov integral \(C_{\rm ZLK}\), and applications to irregular satellites of the giant planets, where the model provides an efficient criterion for identifying ZLK libration in the weak-hierarchy regime [2507.11184; 2605.12864].

## 1. Conceptual setting and scope

The extended Brown Hamiltonian is formulated for a hierarchical triple in the test-particle limit, with an inner body of negligible mass orbiting a central mass \(m_0\) and an outer perturber of mass \(m_p\) moving on a fixed Kepler orbit [2505.13780]. The relevant dynamical context is a low-hierarchy or weak-hierarchy triple, where nonlinear perturbations associated with periodic oscillations of both the inner and outer binaries accumulate secularly and materially affect long-term dynamics [2505.13780; 2507.11184].

The motivating systems explicitly include high-altitude lunar satellites influenced by the Earth, planetary satellites perturbed by the Sun, and stellar binaries affected by a supermassive black hole [2505.13780]. In such systems, the classical double-averaged quadrupole Hamiltonian captures the leading ZLK mechanism, while Brown’s 1936 correction accounts for first-order nonlinear short-period effects of the outer orbit; however, these contributions alone can become insufficient when the hierarchy is mild and the inner-orbital period is also comparable to the ZLK timescale [2505.13780].

In this framework, the extended Brown Hamiltonian is not a higher-multipole expansion of the disturbing function. Rather, it is obtained by truncating the disturbing function at quadrupole order while retaining all nonlinearities up to \(\mathcal O\!\left((n_p/n)^2\right)\) produced by successive von Zeipel averaging [2505.13780]. A plausible implication is that its improvement over classical secular theory arises from a more complete treatment of short-period accumulation rather than from adding octupole or higher spatial harmonics.

## 2. Hamiltonian structure and closed-form terms

The basic canonical variables are Delaunay variables. In one normalization, these are written as
\[
L=\sqrt{\mu a},\quad G=L\sqrt{1-e^2},\quad H=G\cos i,\quad (l,g,h)=(M,\omega,\Omega),
\]
with \(\mu=Gm_0\) [2505.13780]. In the normalized formulation used in the analytical study, one works with
\[
g=\omega,\quad G=\sqrt{1-e^2},\quad h=\Omega,\quad H=G\cos i,
\]
where \(\{g,G\}\) and \(\{h,H\}\) are canonically conjugate pairs and \(H\) is an exact integral because \(h\) is cyclic [2507.11184].

After double averaging and retention of nonlinear corrections through second order in the mean-motion ratio, the secular Hamiltonian can be written, up to an additive or overall constant factor, as
\[
{\cal H}_{\rm ext}(e,i,\omega)
=
-\,{\cal C}_0
\Bigl[
{\cal F}_{20}
+\varepsilon_{21}{\cal F}_{21}
+\varepsilon_{22}{\cal F}_{22}
\Bigr],
\]
where
\[
{\cal C}_0=\tfrac{3}{8}\,\frac{Gm_p}{a_p}\bigl(\tfrac{a}{a_p}\bigr)^2(1-e_p^2)^{-3/2}
\]
and the three dimensionless components are given in closed form [2505.13780].

The classical quadrupole term is
\[
{\cal F}_{20}(e,i,\omega)
=
\tfrac16\,(2+3e^2)(3\cos^2i-1)
+\tfrac52\,e^2\sin^2i\cos2\omega,
\]
and the classical Brown correction is
\[
{\cal F}_{21}(e,i,\omega)
=
\tfrac{3}{16}\sqrt{1-e^2}\cos i
\Bigl[
2\sin^2i
+e^2(33+17\cos^2i+15\sin^2i\cos2\omega)
\Bigr].
\]
The new extended contribution is
\[
\begin{aligned}
{\cal F}_{22}(e,i,\omega)
&=
\tfrac{3}{512}\Bigl\{
227\,e^2(8+e^2)
-\cos^4i\,(56-472e^2+701e^4) \\
&\quad
-2\cos^2i\Bigl(\tfrac{376}{3}-360e^2-305e^4\Bigr)
-95e^4\sin^4i\cos4\omega \\
&\quad
+4e^2\sin^2i\Bigl[186+\tfrac{109}{3}e^2+5(18-37e^2)\cos^2i\Bigr]\cos2\omega
\Bigr\}.
\end{aligned}
\]
These formulae are stated to be closed form with respect to the eccentricities of the inner and outer orbits, and the generating function used for the mean–osculating transformation is likewise closed form and converges for any \(e<1\), unlike power-series expansions [2505.13780].

A second, algebraically equivalent representation emphasizes the constant
\[
j_z=\frac{H}{L}=\sqrt{1-e^2}\cos i=\cos i_*,
\]
where \(i_*\) is the mutual inclination at zero eccentricity, and introduces
\[
e_z=e\sin i\sin\omega
=
e\sqrt{1-j_z^2/(1-e^2)}\sin\omega.
\]
In that representation,
\[
{\cal F}(e,\omega;j_z)
=
{\cal F}_{20}
+\varepsilon_{21}{\cal F}_{21}
+\varepsilon_{22}{\cal F}_{22},
\]
with
\[
{\cal F}_{20}=2e^2-5e_z^2+j_z^2-\tfrac13,
\]
\[
{\cal F}_{21}=\frac{3}{8}j_z\Bigl(1-j_z^2+24e^2-15e_z^2\Bigr),
\]
\[
{\cal F}_{22}
=
\frac{1}{64}
\Bigl\{
8e^2(13e^2+22e_z^2+4j_z^2+120)
-94j_z^2
-3\bigl[95e_z^4+6e_z^2(15j_z^2+31)+7j_z^4\bigr]
\Bigr\},
\]
which is the form used in the applications to irregular satellites [2605.12864].

## 3. Ordering, derivation, and approximation regime

The derivation assumes a negligible-mass inner body, a fixed outer Kepler perturber, and a hierarchical semimajor-axis ratio \(a_1/a_2\ll1\), while truncating the disturbing function at quadrupole order \(\propto (a_1/a_2)^2\) [2505.13780]. A bookkeeping parameter \(\epsilon\sim(n_p/n)\) is introduced, and the Hamiltonian is expanded to second order in \(\epsilon\) [2505.13780].

Two small parameters control the corrections:
\[
\varepsilon_{21}\sim\frac{n_p}{n}\,\frac{m_p/(m_0+m_p)}{(1-e_p^2)^{3/2}},
\qquad
\varepsilon_{22}\sim\bigl(\tfrac{n_p}{n}\bigr)^2\,\frac{m_p/(m_0+m_p)}{(1-e_p^2)^3},
\]
with \(\varepsilon_{21}\) associated with the classical Brown correction and \(\varepsilon_{22}\) with the new extended term [2505.13780]. In the irregular-satellite scaling used later, these parameters are expressed as
\[
\varepsilon_{21}\sim \frac{1}{\sqrt3}\alpha_H^{3/2},
\qquad
\varepsilon_{22}\sim \tfrac13\alpha_H^3,
\]
where \(\alpha_H=a/r_H\) [2605.12864].

The derivation proceeds by two successive von Zeipel canonical transformations [2505.13780]. In Step I, one removes the inner fast angle \(l\) via a generating function \(S_1\), producing the single-averaged Hamiltonian and then removing its short-period remainder. In Step II, one removes the outer fast angle \(l_p\) via \(S_1^*\), producing the double-averaged Hamiltonian \({\cal F}_{20}\) and then removing its short-period remainder, which generates \({\cal F}_{21}\). The second-order terms in the final averaging produce two secular corrections: the classical Brown term \({\cal F}_{21}\), from accumulated outer short-period effects, and the new extended term \({\cal F}_{22}\), from accumulated inner short-period effects [2505.13780].

The final secular problem is integrable, with a single degree of freedom in \((g,G)\) and with \(L\) and \(H\) conserved [2505.13780]. This is central to the later analytical development of fixed points, separatrices, maximum eccentricities, and critical inclinations [2507.11184].

The reported validity regime is “mildly to weakly hierarchical triples,” with truncation at \(\mathcal O(\alpha_H^3)\) stated to be valid up to \(\alpha_H\sim0.5\) [2605.12864]. This suggests that the model is intended precisely for systems that are not asymptotically deep in the hierarchy, but are still hierarchical enough for the quadrupole truncation to remain meaningful.

## 4. Mean–osculating transformation and secular equations

A distinctive feature of the framework is that it provides not only the secular Hamiltonian but also a closed-form canonical generating function for transforming between osculating elements \((a,e,i,\Omega,\omega,M)\) and secular or mean elements \((a^*,e^*,i^*,\Omega^*,\omega^*,M^*)\) [2505.13780]. To first order in \((\varepsilon_{21},\varepsilon_{22})\), the generating function is written as
\[
S=S_1(a^*,e^*,i^*,\omega^*,\Omega^*,M^*)+S_1^*(a^*,e^*,i^*,\omega^*,\Omega^*,M_p^*),
\]
where \(S_1\) removes inner-period terms and \(S_1^*\) removes outer-period terms [2505.13780].

The resulting change of variables is given in Lagrange-planetary-equations form, evaluated at the mean elements:
\[
\delta a=\frac{2}{na}\frac{\partial S}{\partial M},
\qquad
\delta e=\frac{\eta}{na^2e}\Bigl(\eta\,\partial_M S-\partial_\omega S\Bigr),
\]
\[
\delta i=\frac{1}{na^2\eta\sin i}\Bigl(\cos i\,\partial_\omega S-\partial_\Omega S\Bigr),
\qquad
\delta\Omega=\frac{1}{na^2\eta\sin i}\partial_i S,
\]
\[
\delta\omega=\frac{\eta}{na^2}\Bigl(\tfrac1e\partial_e S-\tfrac{\cot i}{\eta^2}\partial_i S\Bigr),
\qquad
\delta M=-\frac{2}{na}\partial_a S-\frac{\eta^2}{na^2e}\partial_e S,
\]
with \(\eta=\sqrt{1-e^2}\) and all partial derivatives taken at the mean elements [2505.13780].

The secular equations of motion under \({\cal H}_{\rm ext}=-{\cal C}_0{\cal F}\) are
\[
\frac{de}{dt}=-\frac{\eta}{na^2e}\frac{\partial{\cal F}}{\partial\omega},
\qquad
\frac{d\omega}{dt}=\frac{\eta}{na^2}\Bigl(\tfrac1e\partial_e{\cal F}-\tfrac{\cot i}{\eta^2}\partial_i{\cal F}\Bigr),
\qquad
\frac{d\Omega}{dt}=\frac{1}{na^2\eta\sin i}\partial_i{\cal F},
\]
and may be decomposed as
\[
\frac{dX}{dt}
=
\bigl(dX/dt\bigr)_{20}
+\varepsilon_{21}\bigl(dX/dt\bigr)_{21}
+\varepsilon_{22}\bigl(dX/dt\bigr)_{22},
\]
with each contribution available in closed form [2505.13780].

The paper states that these expressions may be inserted directly into a 1D ODE integrator or used to find analytical Jacobian-elliptic solutions [2505.13780]. This indicates that the model is not only conceptually integrable but also practically deployable as a reduced dynamical system.

## 5. Modified Lidov integral and phase-space organization

A central analytical development in later papers is the modified Lidov integral \(C_{\rm ZLK}\), introduced to characterize the modified ZLK dynamics under the extended Hamiltonian [2507.11184; 2605.12864]. In the classical quadrupole problem, the invariant is
\[
c_2=e^2-\tfrac52 e_z^2,
\]
with
\[
e_z=e\sin i\sin\omega,
\]
and equivalently
\[
c_2=\tfrac12\bigl({\cal F}_{20}-H^2+\tfrac13\bigr)
\]
in the normalized notation [2507.11184].

In the extended model,
\[
C_{\rm ZLK}=C_{20}+\varepsilon_{21}C_{21}+\varepsilon_{22}C_{22},
\]
with
\[
C_{20}=e^2-\tfrac52 e_z^2,
\]
\[
C_{21}=\tfrac92 H\Bigl(e^2-\tfrac58 e_z^2\Bigr),
\]
\[
C_{22}
=
\frac{1}{64}
\Bigl\{
8e^2\bigl(\tfrac{13}{2}e^2+11e_z^2+2H^2+60\bigr)
-\tfrac32 e_z^2\bigl[95e_z^2+6(15H^2+31)\bigr]
\Bigr\},
\]
and one checks directly that \(\{{\cal F},C_{\rm ZLK}\}=0\) and \(\{H,C_{\rm ZLK}\}=0\), so \(C_{\rm ZLK}\) is a second integral of the 1-degree-of-freedom system [2507.11184].

In the alternative \(j_z\)-based representation, the same quantity satisfies
\[
{\cal F}
=
2C_{\rm ZLK}
+\Bigl(j_z^2-\tfrac13\Bigr)
+\frac{3}{8}\varepsilon_{21}j_z(1-j_z^2)
-\frac{1}{64}\varepsilon_{22}j_z^2(94+21j_z^2),
\]
which implies \(\dot C_{\rm ZLK}=0\) exactly [2605.12864]. The later applications paper explicitly calls \(C_{\rm ZLK}\) a “practical diagnostic index” and a “decisive parameter” for identifying the ZLK resonance [2605.12864].

The separatrix between libration and circulation is characterized by
\[
C_{\rm ZLK}=0,
\]
and, in the \((e,\omega)\) phase portrait, the level curve \(C_{\rm ZLK}=0\) passes through the unstable fixed point at \(e=0,\ \omega=\pi/2\) [2507.11184; 2605.12864]. Trajectories with \(C_{\rm ZLK}>0\) correspond to full circulation of \(\omega\) from \(0\) to \(2\pi\), while trajectories with \(C_{\rm ZLK}<0\) are confined to libration around \(\omega=\pi/2\) or \(3\pi/2\) [2605.12864]. Accordingly, the analytic criterion
\[
C_{\rm ZLK}<0
\]
is used to identify ZLK libration in the weak-hierarchy regime [2605.12864].

A recurrent theme is that the Brown correction breaks the prograde–retrograde symmetry of classical quadrupole ZLK dynamics. Under the classical model, \({\cal F}_{20}(H)\) is even in \(H\), implying invariance under \(i\leftrightarrow \pi-i\). The second-order correction \({\cal F}_{22}\) is also even in \(H\), but the first-order Brown term \({\cal F}_{21}\propto H\) is odd in \(H\), so the full Hamiltonian is asymmetric under \(H\leftrightarrow -H\) [2507.11184]. The same asymmetry appears in the Lidov integrals, where only \(C_{21}\propto H\) is odd in \(H\) [2507.11184]. This is the formal reason that ZLK properties in prograde and retrograde regimes are no longer symmetric.

## 6. Analytical results for modified ZLK dynamics

The analytical study in Paper II develops perturbative expressions for the principal geometric and dynamical features of modified ZLK oscillations under the extended Brown Hamiltonian [2507.11184]. The four highlighted characteristics are the location of the ZLK center, the maximum eccentricity on the separatrix, the critical inclination for the onset of libration, and the separatrix boundary itself.

For the stable fixed point, stationarity with \(\partial{\cal F}/\partial\omega=0\) implies \(2\omega=\pi\), and \(\partial{\cal F}/\partial e=0\) leads to a quartic in
\[
x\equiv G_{\rm fix}^2\equiv 1-e_{\rm fix}^2.
\]
The perturbative solution is written as
\[
x=x_0+\varepsilon_{22}x_1+\varepsilon_{22}^2x_2+\varepsilon_{22}^3x_3,
\]
with
\[
x_0=\sqrt{\frac{5(8+9\varepsilon_{21}H)}{3(8-9\varepsilon_{21}H)}}\,|H|,
\]
and the corresponding fixed-point eccentricity and inclination given by
\[
e_{\rm fix}=\sqrt{1-x},
\qquad
i_{\rm fix}=\arccos\bigl(H/\sqrt{x}\bigr)
\]
[2507.11184].

For the separatrix maximum eccentricity, one defines
\[
z\equiv 1-e_{\max}^2,
\]
and imposes
\[
{\cal F}(e=0,\omega=0)={\cal F}(e=e_{\max},\omega=\pi),
\]
or equivalently \(C_{\rm ZLK}=0\), obtaining a quartic in \(z\) [2507.11184]. The perturbation solution is
\[
z=z_0+\varepsilon_{22}z_1+\varepsilon_{22}^2z_2+\varepsilon_{22}^3z_3,
\qquad
z_0=\frac{5(8+9\varepsilon_{21}H)}{3(8-9\varepsilon_{21}H)}H^2,
\]
from which
\[
e_{\max}=\sqrt{1-z},
\qquad
i_{\max}=\arccos\bigl(H/\sqrt{z}\bigr)
\]
follow [2507.11184].

For the critical inclination, the onset of libration is determined by the small-eccentricity saddle conditions
\[
\frac{\partial^2{\cal F}}{\partial e^2}\Big|_{e=0}=0,
\qquad
\frac{\partial^2{\cal F}}{\partial\omega^2}\Big|_{e=0}=0,
\]
leading to a quartic in
\[
w\equiv \cos i_{\rm crit}.
\]
The perturbative solution is
\[
w=w_0+\varepsilon_{22}w_1+\varepsilon_{22}^2w_2+\varepsilon_{22}^3w_3,
\]
with zeroth-order term
\[
\begin{aligned}
w_0
=
\pm\sqrt{
\frac{3}{5}\Bigl(1+\tfrac{729}{640}\varepsilon_{21}^2+\tfrac{649539}{163840}\varepsilon_{21}^4+\cdots\Bigr)
-\frac{27}{40}\varepsilon_{21}\Bigl(1+\tfrac{729}{320}\varepsilon_{21}^2+\cdots\Bigr)
},
\end{aligned}
\]
where \(+\) corresponds to prograde and \(-\) to retrograde motion, and
\[
i_{\rm crit}=\arccos w
\]
[2507.11184].

The perturbation scheme itself is developed in two variants. “Approach I” treats both \(\varepsilon_{21}\) and \(\varepsilon_{22}\) as small and expands in a double series. “Approach II” absorbs \(\varepsilon_{21}\) exactly by starting from \({\cal F}_{20}+\varepsilon_{21}{\cal F}_{21}\) and then treats only \(\varepsilon_{22}\) as small [2507.11184]. The reported quoted results are those of Approach II, expanded to third order in \(\varepsilon_{22}\), and the paper states that analytical predictions show excellent agreement with numerical results [2507.11184].

## 7. Applications to irregular satellites and empirical performance

The principal numerical tests reported for Paper I concern four high-altitude moons of Jupiter: Pasiphae, Kore, Callirrhoe, and Philophrosyne [2505.13780]. For these systems, the reported parameters are approximately \(\varepsilon_{21}\approx 0.17\) and \(\varepsilon_{22}\approx 0.03\) [2505.13780]. The paper compares four models for Pasiphae and Kore: the classical \({\cal F}_{20}\) only, \({\cal F}_{20}+\varepsilon_{21}{\cal F}_{21}\), the full extended model \({\cal F}_{20}+\varepsilon_{21}{\cal F}_{21}+\varepsilon_{22}{\cal F}_{22}\), and direct \(N\)-body integration [2505.13780]. It reports that only the full extended model captures both the period and amplitude well, with errors \(\lesssim1\%\), whereas the truncated models have \(>10\%\) errors and can even predict the wrong libration-versus-circulation behavior; similar agreement is reported for all four satellites over thousands of years [2505.13780].

Paper II extends this numerical validation to extensive \(N\)-body integrations over grids in \((i_*,e,\alpha_H)\), up to \(\sim 300\) secular ZLK periods [2507.11184]. The analytic center, boundary, and critical-inclination curves are stated to lie on top of numerically identified libration islands with typical relative errors
\[
\Delta e_{\rm fix},\ \Delta e_{\max}\lesssim 10^{-4}
\]
for Approach II, and
\[
\Delta i_{\rm crit}\lesssim 0.1^\circ\text{--}0.2^\circ,
\]
even into the weak-hierarchy regime \(\alpha_H\lesssim0.5\) [2507.11184].

Paper III applies the modified Lidov integral \(C_{\rm ZLK}\) to the known population of irregular satellites of the giant planets [2605.12864]. Using mean orbital elements computed from direct \(N\)-body averaging and then forming \(\alpha_H\), \(j_z\), \(\varepsilon_{21}\), \(\varepsilon_{22}\), and \(C_{\rm ZLK}\), the study reports that exactly 27 objects satisfy \(C_{\rm ZLK}<0\) [2605.12864]. Direct \(N\)-body simulations confirm 26 of these as sustained librators, with the sole exception of S/2019 S1, whose discrepancy is attributed to its proximity to the separatrix, \(C_{\rm ZLK}\approx0\) [2605.12864]. The same work states that the red contour \(C_{\rm ZLK}=0\) marks the separatrix between circulating and librating phases in individual \((e,2\omega)\) portraits, and that the extended Brown trajectories “hug” the osculating curves almost perfectly [2605.12864].

These applications clarify both the utility and the limitations of the framework. The utility lies in providing a high-accuracy secular model and a simple analytical resonance indicator in regimes where classical averaging is quantitatively inadequate. The limitation is embedded in the assumptions: test-particle dynamics, a fixed outer Kepler perturber, quadrupole truncation of the disturbing function, and neglect of terms beyond \(\mathcal O[(a_1/a_2)^2(n_p/n)^2]\) or \(\mathcal O(\alpha_H^3)\) [2505.13780; 2605.12864]. A common misconception would be to treat the model as a generic three-body theory valid without these restrictions; the published formulation is explicitly tailored to weakly hierarchical, quadrupole-order, test-particle secular dynamics.

Source: https://www.emergentmind.com/topics/extended-brown-hamiltonian