Advanced Celestial Dynamics....

Aditya Bhoj

Contents
1 Introduction 2

2 Hamiltonian Framework of Celestial Dynamics 2

2.1 Canonical Coordinates and Action–Angle Variables . . . . . . . . . . . . . . 2
2.2 KAM and Nekhoroshev Theorems . . . . . . . . . . . . . . . . . . . . . . . . 2
2.3 Example 1: Width of the 2:1 Jovian Resonance . . . . . . . . . . . . . . . . 3

3 Post–Newtonian Celestial Mechanics 3

3.1 Einstein–Infeld–Hoffmann Equations . . . . . . . . . . . . . . . . . . . . . . 3
3.2 Example 2: 2PN Perihelion Advance of Mercury . . . . . . . . . . . . . . . 3

4 Chaotic Diffusion and Long–Term Stability 3

4.1 Angular–Momentum Deficit (AMD) and Hill Criteria . . . . . . . . . . . . . 3

5 Stellar Structure Theory 4

5.1 Hydrostatic Equilibrium and Virial Theorem . . . . . . . . . . . . . . . . . . 4
5.2 Polytropic Lane–Emden Equation . . . . . . . . . . . . . . . . . . . . . . . . 4
5.3 Chandrasekhar Limit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
5.4 Convective Overshoot and Mixing–Length Calibration . . . . . . . . . . . . . 4

6 Advanced Stellar Evolution 4

6.1 Nuclear Energy–Generation Chains . . . . . . . . . . . . . . . . . . . . . . . 4
6.2 Late–Stage Evolution and Core Collapse . . . . . . . . . . . . . . . . . . . . 5
6.3 Neutron–Star Equation of State and g–Modes . . . . . . . . . . . . . . . . . 5

7 Problem Set 5

1
1 Introduction
The dynamical behaviour of celestial bodies and the internal physics of stars represent two
extreme—but intimately linked—realms of gravitational physics. Modern research demands
fluency not only with classical Hamiltonian mechanics, but also with general–relativistic cor-
rections, nonlinear dynamics, and quantum–statistical matter under supra–nuclear density.
This project provides a terse yet thorough tour of very advanced topics spanning:
• Kolmogorov–Arnold–Moser (KAM) stability and Nekhoroshev theory.
• Mean–motion and secular resonances, chaotic diffusion, and AMD/Hill criteria.
• Post–Newtonian (PN) celestial mechanics to 2PN order with applications to perihelion
precession and gravitational–wave back–reaction.
• Stellar structure from polytropic Lane–Emden solutions up to MHD–stabilised neu-
tron–star equations of state.
• Energy–generation chains (pp, CNO), convective overshoot, mixing–length calibration,
and g–mode seismology.
Each major theory is accompanied by carefully chosen, fully–solved problems that force
the reader to apply the mathematics at professional level.

2 Hamiltonian Framework of Celestial Dynamics

2.1 Canonical Coordinates and Action–Angle Variables
Consider an isolated, hierarchical N –body system. After Jacobi reduction to eliminate linear
momentum, the Hamiltonian reads
N −1  2 
X pi Gm0 mi
H= − + ε R(ri , pi ), (1)
i=1
2µi ri
where ε ≡ max(mi /m0 ) and R is the disturbing function. Passing to Delaunay variables
(L, G, H; ℓ, g, h):
p
L = µ Gm0 a, ℓ = M,
√
G = L 1 − e2 , g = ω,
H = G cos i, h = Ω,
renders the Keplerian part integrable. The perturbation drives slow evolution of (G, H).

2.2 KAM and Nekhoroshev Theorems

For ε ≪ 1 the iso–energetic non–degeneracy and Diophantine condition √ on fundamental
frequencies ω(L) ensure survival of invariant tori of measure 1 − O( ε) [1]. Nekhoroshev
−a
b


estimates bound action drift ∆I ≲ ε over times tesc ∼ exp ε with a, b > 0, explaining
9
Solar–System stability > 10 yr despite micro–chaos.

2
2.3 Example 1: Width of the 2:1 Jovian Resonance
Problem. Using the pendulum approximation for a first–order p:(p−1) resonance, derive
the semi–major–axis half–width ∆a of the exterior 2:1 resonance with Jupiter (mJ = 9.5 ×
10−4 M⊙ ) at low e.

Solution. The resonant Hamiltonian in Poincaré variables reduces to

β
H = − (δn)2 − 2αe cos ϕ, ϕ = 2λ − λJ − ϖ, (2)
2
(1)
with coefficients (see [2]) β = 3na/2, α = na2 mJ Gb1 (α∗ )/4, α∗ = (a/aJ ). Equating libration
energy at separatrix yields
r r
p ∆a 8 αe mJ (1)
∆n = 4 αe/β, ∴ = =8 e |b1 |1/2 . (3)
a 3 βn 3
Setting e = 0.05 gives ∆a ≈ 0.017 au.

3 Post–Newtonian Celestial Mechanics

3.1 Einstein–Infeld–Hoffmann Equations
Retaining terms through O(c−2 ) leads to the 1PN Lagrangian
X mi v 2  3vi2 vi2 + vj2 − 3vi ·vj
 X  
i Gmi mj G(mi + mj )
L1PN = 1+ 2 + 1+ 2
− 2
. (4)
i
2 8c i<j
2r ij 2c r ij c

Variation yields the celebrated EIH accelerations.

3.2 Example 2: 2PN Perihelion Advance of Mercury

Using Gauss’ perturbative equations with EIH and 2PN terms [3], show that for Mercury
ϖ̇2PN ∈ [−18, −4] µas cty−1 , two orders below current ephemeris precision, but measurable
via BepiColombo tracking.

Solution Sketch. Insert Eqs. (30)–(35) of [3] with a = 0.387 au, e = 0.206, f0 variable.
Averaging over mean anomaly gives quoted interval.

4 Chaotic Diffusion and Long–Term Stability

4.1 Angular–Momentum Deficit (AMD) and Hill Criteria
P √ p
Define system AMD C = mi GM⊙ ai (1 − 1 − e2i cos ii ). For two–planet systems Hill
stability requires C < Cc (α, µ) [4, 5]. Recent work [6] unifies the criterion with AMD.
Numerical experiments confirm abrupt transition to collisional outcomes once C > Cc .

3
5 Stellar Structure Theory
5.1 Hydrostatic Equilibrium and Virial Theorem
Writing  
1 d 2 dP
r = −4πGρ(r), (5)
r2 dr dr
combining with the equation of state closes the system. The scalar virial theorem yields
Egrav + 2Eth = 0 for polytropes.

5.2 Polytropic Lane–Emden Equation

1/n−1
For P = Kρ1+1/n set ρ = ρc θn and r = aξ, a2 = (n+1)Kρ 4πG
c
to obtain
 
1 d 2 dθ
ξ + θn = 0. (6)
ξ 2 dξ dξ

Analytic solutions exist for n = 0, 1, 5. Numerical integration for n = 3 models the Sun’s
envelope.

5.3 Chandrasekhar Limit

√
Degenerate Fermi pressure P ∝ x(2x2 −3) x2 + 1+3 sinh−1 x. In the ultra–relativistic limit
P = Kρ4/3 yields polytrope n = 3; the mass becomes independent of central density:

−2
MCh = 1.456 µe /2 M⊙ . (7)

5.4 Convective Overshoot and Mixing–Length Calibration

√ 2
Mixing–length theory (MLT) dictates convective flux Fc = ρcP gδHP αMLT (∇ − ∇e )3/2 with
free parameter αMLT . 3D RHD simulations constrain αMLT (Teff , log g) to ±0.1 [9]. Overshoot
mixing obeys diffusion with Dov = βvc Hp exp(−2z/f Hp ) [7].

6 Advanced Stellar Evolution

6.1 Nuclear Energy–Generation Chains
pp–chain dominates for Tc < 1.5 × 107 K with rate
2 −2/3 −1/3

εpp = ε0 XH ρT6 exp −3.380T6 . (8)

CNO cycle takes over above T ∼ 1.7 × 107 K with εCNO ∝ T 17 [8]. Updated S–factors [11]
raise the rate 15% at T = 25 MK, impacting turn–off ages by ≈150 Myr.

4
6.2 Late–Stage Evolution and Core Collapse
Silicon burning proceeds in quasi–nuclear–statistical equilibrium, halting at 56 Ni when pho-
todisintegration matches α–captures.CollapseensuesonceMcore → 1.46M⊙ and electrons be-
come ultra–relativistic.

6.3 Neutron–Star Equation of State and g–Modes

Modern chiral EFT matched to perturbative QCD yields P (ε) bands compatible with GW170817
tidal deformabilities. Buoyancy due to composition gradients excites core g–modes with
fg ≳ 600 Hz only if a quark–hadron mixed phase exists [10].

7 Problem Set
Problem 3 — Lane–Emden Scaling
Show that for an n = 1 polytrope the central–to–mean density ratio is π 2 /3. Deduce the
radius of a fully convective 0.2M⊙ star with K = 1.0 × 1013 cgs.

Solution. For n = 1, analytic solution θ = sin ξ/ξ vanishes at ξ1 = π. Mean density

Rξ
ρ̄ = 3M/4πR3 = ρc ξ33 0 1 θn ξ 2 dξ = ρc 3 sin ξ1 − ξ1 cos ξ1 )/ξ13 = ρc 3/π 2 . Hence ρc /ρ̄ = π 2 /3.
1
Radius R = aξ1 with a as above; inserting numbers gives R ≈ 0.24R⊙ .

Problem 4 — AMD Stability Threshold

Consider two Earth–mass planets at a1 = 1 au, a2 = 1.6 au. Assuming co–planar orbits, find
the maximum common eccentricity emax compatible with Hill/AMD stability.

Solution. With µ = m/M⊙ = 3 × 10−6 and α = a1 /a2 = 0.625, the critical AMD [4]
√ 2 3/2
1/2 1/2 ( α + 1) − (1 + µ) (1 + αµ)1/2
Cc = µa1 a2 . (9)
(1 + µ)1/2
√ √ √ √
Using e1 = e2 = e the system AMD is C = 2µ GM⊙ ( a1 + a2 )(1− 1 − e2 ). Numerically
solving C = Cc gives emax ≈ 0.18.

Conclusion
We have surveyed a spectrum of cutting–edge topics in celestial dynamics and stellar physics,
weaving together rigorous mathematics, relativistic corrections, and micro–physics of dense
matter. The problems supplied are designed to stretch intuition and build true research–level
competence.

5
References
[1] V. I. Arnold, Russ. Math. Surveys 18, 9 (1963).

[2] C. D. Murray and S. F. Dermott, Solar System Dynamics (Cambridge Univ. Press,
1999).

[3] L. Iorio, arXiv:2212.07323 (2022).

[4] J. Laskar and A. Petit, A&A 605, A72 (2017).

[5] A. Petit et al., A&A 617, A93 (2018).

[6] S. Fromenteau et al., A&A 620, A101 (2018).

[7] Q. S. Zhang, ApJS 205, 18 (2013).

[8] C. Bertulani, Introduction to Nuclear Astrophysics (Wiley, 2013).

[9] A. Jermyn et al., ApJ 948, L16 (2023).

[10] R. Essick et al., Phys. Rev. C 104, 065804 (2021).

[11] G. Imbriani et al., Phys. Rev. C 111, 025805 (2025).