Celestial Mechanics

K. Yavuz Ekşi

Department of Physics
İstanbul Technical University

October 16, 2025

 Outline
1 Introduction

2 Kepler’s Laws

3 Conic sections
Geometry of Ellipse
Geometry of Conic Sections
4 Motion in Polar Coordinates

5 Conservation of Angular Momentum

Law of Equal Areas
Law of Equal Areas and Central Forces
Motion is in a Plane
6 Motion is a Conic Section
Derivation of Kepler’s Third Law
7 Two-body problem

8 Stability of Circular Orbits

 Outline
1 Introduction

2 Kepler’s Laws

3 Conic sections
Geometry of Ellipse
Geometry of Conic Sections
4 Motion in Polar Coordinates

5 Conservation of Angular Momentum

Law of Equal Areas
Law of Equal Areas and Central Forces
Motion is in a Plane
6 Motion is a Conic Section
Derivation of Kepler’s Third Law
7 Two-body problem

8 Stability of Circular Orbits

 Outline
1 Introduction

2 Kepler’s Laws

3 Conic sections
Geometry of Ellipse
Geometry of Conic Sections
4 Motion in Polar Coordinates

5 Conservation of Angular Momentum

Law of Equal Areas
Law of Equal Areas and Central Forces
Motion is in a Plane
6 Motion is a Conic Section
Derivation of Kepler’s Third Law
7 Two-body problem

8 Stability of Circular Orbits

 Outline
1 Introduction

2 Kepler’s Laws

3 Conic sections
Geometry of Ellipse
Geometry of Conic Sections
4 Motion in Polar Coordinates

5 Conservation of Angular Momentum

Law of Equal Areas
Law of Equal Areas and Central Forces
Motion is in a Plane
6 Motion is a Conic Section
Derivation of Kepler’s Third Law
7 Two-body problem

8 Stability of Circular Orbits

 Outline
1 Introduction

2 Kepler’s Laws

3 Conic sections
Geometry of Ellipse
Geometry of Conic Sections
4 Motion in Polar Coordinates

5 Conservation of Angular Momentum

Law of Equal Areas
Law of Equal Areas and Central Forces
Motion is in a Plane
6 Motion is a Conic Section
Derivation of Kepler’s Third Law
7 Two-body problem

8 Stability of Circular Orbits

 Outline
1 Introduction

2 Kepler’s Laws

3 Conic sections
Geometry of Ellipse
Geometry of Conic Sections
4 Motion in Polar Coordinates

5 Conservation of Angular Momentum

Law of Equal Areas
Law of Equal Areas and Central Forces
Motion is in a Plane
6 Motion is a Conic Section
Derivation of Kepler’s Third Law
7 Two-body problem

8 Stability of Circular Orbits

 Outline
1 Introduction

2 Kepler’s Laws

3 Conic sections
Geometry of Ellipse
Geometry of Conic Sections
4 Motion in Polar Coordinates

5 Conservation of Angular Momentum

Law of Equal Areas
Law of Equal Areas and Central Forces
Motion is in a Plane
6 Motion is a Conic Section
Derivation of Kepler’s Third Law
7 Two-body problem

8 Stability of Circular Orbits

 Outline
1 Introduction

2 Kepler’s Laws

3 Conic sections
Geometry of Ellipse
Geometry of Conic Sections
4 Motion in Polar Coordinates

5 Conservation of Angular Momentum

Law of Equal Areas
Law of Equal Areas and Central Forces
Motion is in a Plane
6 Motion is a Conic Section
Derivation of Kepler’s Third Law
7 Two-body problem

8 Stability of Circular Orbits

 Introduction

• The roots of modern science are found in the study of the motion of
planets.
• The biggest challenge was the retrograde motion of planets
• Within the framework of the geocentric model of Aristoteles this was
understood by epicycles.
• Copernicus: rediscovered the heliocentric model of Aristarchus.
• Galileo: observations challenged the geocentric idea.
• Tycho Brahe: accurate observations of positions of the planets.
• Kepler: 3 empirical “laws” summarizing the observational data.
• Newton: presented the laws of motion that explains Kepler’s “laws”
and more.
• The industry revolution relied on the Newtonian mechanics.
 Kepler’s First Law

A planet orbits the Sun in an ellipse, with the Sun at one focus of
the ellipse
Astronomica Nova (1609)
 Kepler’s Second Law

A line connecting a planet to the Sun sweeps equal areas in equal

time intervals
Astronomica Nova (1609)
 Kepler’s Third Law

• Square of the orbital period is

proportional to the cube of the
distance.

P2 = a3

where P is the sidereal orbital period

of the planet, measured in years, a is
the average distance of the planet
from the Sun in AU.
The power-law relation a = P2/3
Harmonica Mundi (1619) (The
appears linear in a log-log plot.
Harmony of the World)
• The sidereal period is the time that it
takes the object to make one full
orbit around the Sun, relative to the
stars.
 Remarks on Kepler’s “Laws”

• Not really laws but “empirical relations”.

• Not valid for all celestial objects.
• Empirical relations is the first step of investigating the nature of a
system.
• These relations are to be explained by fundamental laws. (Newton’s
laws of motion and Universal Law of Gravitation in this case.)
 Another example of empirical relation: HR
diagram

This is to be explained by stellar structure and evolution theory.

 Terminology

• The distance between

perihelion and aphelion
is called the major axis.
• minor axis
• The length of major axis
is 2a
• The distance a is called
semi-major axis.
• The distance b is called
the semi-minor axis.
 Geometry of Ellipse
Definition

• Not just any oblate circle.

• Defined as the set of points the sum of whose distances r + r′ from two
foci is constant.
r + r′ = constant.

• If the two foci coincide then r′ = r, we obtain a circle. Thus a circle is

a special ellipse.
 Drawing an Ellipse

• An ellipse can be
constructed by tying a
string to two pins and
drawing with the pencil
stretching the string taut.
• Each pin constitutes one
focus of the ellipse.

r + r′ = constant.
 Terminology

• The Sun is at the principal focus F.

• The other focus, F ′ , is empty.
• The orbit is described by polar
coordinates r and θ
• r is the distance measured from the
principal focus
• true anomaly θ is measured
counterclockwise from the major axis of
the ellipse beginning with the direction
toward perihelion.
 Terminology

• The point where a planet

is closest to the Sun is
called the perihelion (In
Turkish: Günberi).
• The point where the
planet is most distant to
the Sun is called the
aphelion (In Turkish:
Günöte).
 Terminology

For a sattelite orbiting around

the Earth
• perigee is the closest
position
• apogee is the most
distant position.
 Terminology

More generally these would

be called
• periapsis
• apoapsis
 Fixing the Constant

• Recall r + r′ = constant.
• What is this constant?
• At θ = 0 one obtains

r + r′ = 2a (1)

• If the two foci coincide

then r′ = r and we obtain
r = a which is the
equation describing a
circle.
 Eccentricity

The eccentricity of the ellipse

is defined as the distance
between the two foci, 2q
divided by major axis, 2a.

q
ε≡ (2)
a
 Eccentricity

• At an orbital point on the semi-minor

axis r = r′
• r + r′ = 2a requires r = r′ = a.
• Pythagoras’ theorem: q2 + b2 = r2
√
• As r = a we obtain q = a2 − b2
• Using this in ε ≡ q
a

√ r
a2 − b2 b2
ε= = 1− (3)
a a2
 Ellipse
in polar coordinates

• Pythagorean theorem
r′2 = r2 sin2 θ + (2aε + r cos θ)2
• r + r′ = 2a
• eliminating r′
• The general equation for an ellipse in
polar coordinates.

a(1 − ε2 )
r= , 0≤ε<1 (4)
1 + ε cos θ
 Relation between perihelion and aphelion
distance
2
a(1−ε )
• From r = 1+ε cos θ
• θ = 0 gives perihelion
rp = a(1 − ε)

• θ = π gives the aphelion

ra = a(1 + ε)

• Thus
ra 1+ε
=
rp 1−ε
 Exercise: Halley’s Comet

• Halley’s comet follows a very eccentric orbit with ε = 0.9673.

• At the perihelion the comet is 0.59 AU from the Sun.
• What is the comet’s the aphelion?
• Answer: ra = 35 AU i.e. outside the orbit of Neptune.
 Exercise: Halley’s Comet

• Halley’s comet follows a very eccentric orbit with ε = 0.9673.

• At the perihelion the comet is 0.59 AU from the Sun.
• What is the comet’s the aphelion?
• Answer: ra = 35 AU i.e. outside the orbit of Neptune.
 Exercise: Halley’s Comet

• Halley’s comet follows a very eccentric orbit with ε = 0.9673.

• At the perihelion the comet is 0.59 AU from the Sun.
• What is the comet’s the aphelion?
• Answer: ra = 35 AU i.e. outside the orbit of Neptune.
 Exercise: Halley’s Comet

• Halley’s comet follows a very eccentric orbit with ε = 0.9673.

• At the perihelion the comet is 0.59 AU from the Sun.
• What is the comet’s the aphelion?
• Answer: ra = 35 AU i.e. outside the orbit of Neptune.
 Exercise: The orbit of Mars

The semimajor axis of Mars’s orbit is 1.5237 AU. The planet’s orbital
eccentricity is ε = 0.0934.
• Determine (a) the perihelion distance (Answer: rp = 1.3814 AU)
• (b) its aphelion distance (Answer: ra = 1.6660 AU)
• (c) the variation in Mars’ orbital distance from the Sun between
perihelion and aphelion (Answer: %19)
 Exercise: The orbit of Mars

The semimajor axis of Mars’s orbit is 1.5237 AU. The planet’s orbital
eccentricity is ε = 0.0934.
• Determine (a) the perihelion distance (Answer: rp = 1.3814 AU)
• (b) its aphelion distance (Answer: ra = 1.6660 AU)
• (c) the variation in Mars’ orbital distance from the Sun between
perihelion and aphelion (Answer: %19)
 Exercise: The orbit of Mars

The semimajor axis of Mars’s orbit is 1.5237 AU. The planet’s orbital
eccentricity is ε = 0.0934.
• Determine (a) the perihelion distance (Answer: rp = 1.3814 AU)
• (b) its aphelion distance (Answer: ra = 1.6660 AU)
• (c) the variation in Mars’ orbital distance from the Sun between
perihelion and aphelion (Answer: %19)
 Exercise: The orbit of Mars

The semimajor axis of Mars’s orbit is 1.5237 AU. The planet’s orbital
eccentricity is ε = 0.0934.
• Determine (a) the perihelion distance (Answer: rp = 1.3814 AU)
• (b) its aphelion distance (Answer: ra = 1.6660 AU)
• (c) the variation in Mars’ orbital distance from the Sun between
perihelion and aphelion (Answer: %19)
 Geometry of Conic Sections

• An ellipse is one of a class of

curves known as conic sections
• The other conic sections are
parabola and hyperbola.
• All these figures are found by
passing a plane through a cone.
 Geometry of Conic Sections

• The general equation for a conic section in polar coordinates is

k
r= (5)
1 + ε cos θ

• Comparing this with that of ellipse

a(1 − ε2 )
r=
1 + ε cos θ

• we obtain
k = a(1 − ε2 ) for ellipse. (6)
 Conic Sections and Orbits
Not only Ellipse!

It is rewarding to study the ellipse in the context of conic sections because in

understanding planetary motion using Newton’s laws of motion one finds
that parabolic or hyperbolic orbits are also possible, depending on the energy
of the system.
 Ellipse in Cartesian coordinates

• It is possible to show that in Cartesian

coordinates an ellipse is described by

x2 y2
+ =1 (7)
a2 b2

• Exercise: Show that this is correct. Hint:

Start with the following relations
q q
r = (x − q)2 + y2 r′ = (x + q)2 + y2
(8)
Use these in the defining equation of ellipse
(r + r′ = 2a) and take squares and also use
b2 = a2 − q2 .
 Conic section in Cartesian coordinate
system
In the Cartesian coordinate system, the graph of a quadratic equation in two
variables is always a conic section, and all conic sections arise in this way.
The most general equation will be of the form

Ax2 + Bxy + Cy2 + Dx + Ey + F = 0 (9)

where any of A, B and C are constants.

• if B2 − 4AC < 0, the equation represents an ellipse; As a special case,
if A = C and B = 0, the equation represents a circle;
• if B2 − 4AC = 0, the equation represents a parabola;
• if B2 − 4AC > 0, the equation represents a hyperbola.
Exercise: Show that Eqn for ellipse

x2 y2
+ =1
a2 b2
indeed satisfies B2 − 4AC < 0, the condition for an ellipse. Hint: What are
A, B and C in this case?
 Conic sections as sections of a cone

• The equation for conic section can be rearranged by taking the linear
part to the other side, yielding

Ax2 + Bxy + Cy2 = −(Dx + Ey + F) (10)

• In this form, a conic section is realized exactly as the intersection of the

graph of the quadratic form z = Ax2 + Bxy + Cy2 and the plane
z = −(Dx + Ey + F).
• Ellipses and hyperbolae can be realized by a horizontal plane
(D = E = 0), while parabolae require that the plane be slanted.
• Degenerate conics correspond to degenerate intersections, such as
taking slices such as z = −1 of a positive-definite form.
 Conic sections and guidelines

• As a conic section, the parabola is

produced by the intersection of a conical
surface with a plane parallel to the side
of the cone.
• A parabola is the geometric locus of all
points on a plane that have equal
distance from a fixed guideline and fixed
focal point.
 Conic sections and guidelines

• As a conic section, the parabola is

produced by the intersection of a conical
surface with a plane parallel to the side
of the cone.
• A parabola is the geometric locus of all
points on a plane that have equal
distance from a fixed guideline and fixed
focal point.
 Conic sections and guidelines
If the distance between the guideline and focus
is 2p then
d = 2p − r cos θ (11)
and as d = r one can solve r from
r = 2p − r cos θ to obtain
2p
r= (12)
1 + cos θ
where p is the distance of closest approach to
It is possible to show that in
the parabola’s focus (a parabola has only one
Cartesian coordinates the
focus) at θ = 0. Comparing this with
parabola above can be
r = 1+ε kcos θ we see that
expressed as y2 = −4px.
ε = 1, k = 2p for parabola. (13) Exercise: Show this. Hint:
Notep that
r = (p + x)2 + y2 and
r = d = p − x.
 Hyperbola

• The hyperbola is the geometric locus of all

points of a plane whose distances from two
fixed points on the plane (the focal points)
F and F ′ have a constant distance,
r − r′ = constant.
• The constant can be fixed by noting that

r − r′ = 2a (14)

when θ = π.
 Hyperbola
• The eccentricity of the hyperbola is defined as the ratio ε = c/a.
• Note that ε > 1 for a hyperbola.
• one can obtain by cosinus theorem that
q
r′ = r2 + (2c)2 + 4rc cos θ (15)

and by using this in the defining Eqn (14) and that c = εa

a(ε2 − 1)
r= (16)
1 + ε cos θ
which is the hyperbol equation in polar coordinates. Comparison with
Eqn. (5) implies

k = a(1 − ε2 ) for hyperbola. (17)

 Hyperbola in Cartesian coordinates

It is possible to show that in Cartesian coordinates an hyperbola is described

by
x2 y2
− =1 (18)
a2 b2
where b2 = c2 − a2 and ±b/a is the slope of the asyptotes of the hyperbola.
 Eccentricity of conic sections

Orbits can be any of the conic sections

depending on the eccentricity:
• An orbit with ε = 0 is a circle.
• An orbit with 0 < ε < 1 is an ellipse.
• An orbit with ε = 1 is a parabola.
• An orbit with ε > 1 is a hyperbola.

We will later see that eccentricity is determined by the total energy of the
system of two objects in question.
 Conic section and guidelines

• In general all conic section figures can be defined in the way we defined
parabola.
• A conic section then is defined as the locus of points the ratio of whose
distance to a fixed point F (the focus) and a line L (the directrix) is
fixed.
• The corresponding conic section consists of all points whose distance to
F equals ε times their distance to L.
• For 0 < ε < 1 we obtain an ellipse, for ε = 1 a parabola, and for ε > 1
a hyperbola.
 Exercise

The average distance of all points on the ellipse from either focus is equal to
the semi-major axis length a.
To prove this:
• Consider an arbitrary point on the ellipse,
P(x, y)
• and its reflection point accross the y-axis,
P′ (−x, y)
• The distance from point P to the focus in the
+x-axis is r
• By symmetry, the distance from the
complementary point P′ to the focus on the
+x axis is r′ , where r′ is the distance from
point P to the focus on the −x-axis.
 Exercise

The average distance of all points on the ellipse from either focus is equal to
the semi-major axis length a.
To prove this:
• The average distance of the two points from
the focus on the +x-axis is then
r + r′ 2a
⟨r⟩ = = =a (19)
2 2

• Since this relation holdsfor all (P, P′ ) pairs,

regardless of the choice of P, it is true that
the average distance ⟨r⟩ from the focus over
the entire ellipse is a.
 Motion in Polar Coordinates

• We later will find the motion of the

system is restricted to a plane.
• This means that we can use the polar
coordinates for describing the motion.
• The unit vectors in polar coordinates are
related to the unit vectors in Cartesian
coordinates as
r̂ = cos θi + sin θj
(20)
θ̂ = − sin θi + cos θj
 Motion in Polar Coordinates

• Ex: Show that

dr̂/dθ = θ̂
(21)
dθ̂/dθ = −r̂

d() d() dθ
• Thus, the chain rule dt = dθ dt implies

dr̂/dt = θ̇θ̂
(22)
dθ̂/dt = −θ̇r̂
 Kinematics
• The position of a particle in polar coordinates with unit vectors is given
as
r = rr̂ (23)

• Thus the velocity v = ṙ is

v = ṙr̂ + rθ̇θ̂ (24)

• And the acceleration a = v̇ is

a = (r̈ − rθ̇2 )r̂ + (2ṙθ̇ + rθ̈)θ̂ (25)

which can also be written as

1d 2
a = (r̈ − rθ̇2 )r̂ + (r θ̇)θ̂ (26)
r dt
 Kinematics

• The specific angular momentum h = |r × v| of a particle is defined as

its angular momentum divided by its mass.

h = r2 θ̇ (27)

• Using this one obtains

h
v = ṙr̂ + θ̂ (28)
r
and
h2
 
1 dh
a= r̈ − 3 r̂ + θ̂ (29)
r r dt

It can immediately be seen from the above equation that if h is constant in

time the acceleration has only a radial component.
 Law of Equal Areas

• As the Figure shows |a × b| = ab sin φ

gives the area of the parallelogram spanned
by a and b.
• As a triangle defined by these vectors is just
the half of this area 21 |a × b| = 12 ab sin φ
gives the area of the triangle defined by two
the vectors a and b.
• We can “vectorize” the area by defining a
The product of two
unit vector n̂ perpendicular to its surface: vectors give the area of the
A = An̂. Thus parallelogram spanned by
1 them.
A= a × b. (30)
2
gives a vector whose magnitude is equal to
the area and direction is perpendicular to the
surface.
 Law of Equal Areas

• As the Figure shows |a × b| = ab sin φ

gives the area of the parallelogram spanned
by a and b.
• As a triangle defined by these vectors is just
the half of this area 21 |a × b| = 12 ab sin φ
gives the area of the triangle defined by two
the vectors a and b.
• We can “vectorize” the area by defining a
The product of two
unit vector n̂ perpendicular to its surface: vectors give the area of the
A = An̂. Thus parallelogram spanned by
1 them.
A= a × b. (30)
2
gives a vector whose magnitude is equal to
the area and direction is perpendicular to the
surface.
 Law of Equal Areas

• As the Figure shows |a × b| = ab sin φ

gives the area of the parallelogram spanned
by a and b.
• As a triangle defined by these vectors is just
the half of this area 12 |a × b| = 12 ab sin φ
gives the area of the triangle defined by two
the vectors a and b.
• We can “vectorize” the area by defining a
The product of two
unit vector n̂ perpendicular to its surface: vectors give the area of the
A = An̂. Thus parallelogram spanned by
1 them.
A= a × b. (30)
2
gives a vector whose magnitude is equal to
the area and direction is perpendicular to the
surface.
 Law of Equal Areas

• A = 12 a × b can be used to write

1
r × dr = dAn̂ (31)
2
where n̂ is a unit vector perpendicular to
the plane of the triangle.
• By dividing both sides with dt we find
The triangle formed by r and
1
r × v = Ȧn̂ (32) dr.
2
 Kepler’s 2nd “law” (Law of equal areas)

• Now, Kepler’s second law, with the

present notation, states that Ȧ is a
constant.
• When r is small v must be large to keep
rate Ȧ the same. Multiplying the
equation above with mass 2m we find

L = 2mȦn̂ (33)

where L = r × p = mr × v is the
angular momentum.
• It can immediately be seen that Kepler’s
second law is a manifestation of the
conservation of angular momentum.
The opposite is also true: Conservation of angular momentum always
implies that Ȧ is constant, i.e. that Kepler’s second “law” is valid.
 Law of Equal Areas and Central Forces
• We will show that Kepler’s second law
implies a central force.
• As Ȧ is constant it’s derivative vanishes.
• Then by taking the time derivative of both
sides of the above equation

d
(r × v) = 0 (34)
dt

• This can be expanded as

dr dv
× v + r× =0 (35)
dt dt

• The first term vanishes because v × v = 0.

Thus r × a = 0 is obtained.
 Law of Equal Areas and Central Forces
• We will show that Kepler’s second law
implies a central force.
• As Ȧ is constant it’s derivative vanishes.
• Then by taking the time derivative of both
sides of the above equation

d
(r × v) = 0 (34)
dt

• This can be expanded as

dr dv
× v + r× =0 (35)
dt dt

• The first term vanishes because v × v = 0.

Thus r × a = 0 is obtained.
 Law of Equal Areas and Central Forces
• r × a = 0 requires that the θ component of
the acceleration in Eqn.(29)

h2
 
1 dh
a = r̈ − 3 r̂ + θ̂
r r dt

vanishes:
1 dh
0= (36)
r dt
which implies that h is a constant. Thus we
rewrite Eqn.(27)

h = r2 θ̇ = constant (37)

• Actually, by noting that dr = rdθ we find

1 2 1 2
2 r dθ = dA and that 2 r θ̇ = Ȧ. Hence
h = 2Ȧ.
 Law of Equal Areas and Central Forces
• r × a = 0 requires that the θ component of
the acceleration in Eqn.(29)

h2
 
1 dh
a = r̈ − 3 r̂ + θ̂
r r dt

vanishes:
1 dh
0= (36)
r dt
which implies that h is a constant. Thus we
rewrite Eqn.(27)

h = r2 θ̇ = constant (37)

• Actually, by noting that dr = rdθ we find

1 2 1 2
2 r dθ = dA and that 2 r θ̇ = Ȧ. Hence
h = 2Ȧ.
 Central force

• As F is proportional to a we find

r×F=0 (38)

• This implies that r ∥ F i.e. that the force

and the radius vectors are collinear. The
force acts along the radius vector.
• Such a force is called a central force.
 Central force

• Thus the second law of Kepler requires a

central force field.
• The opposite is also true: A central force
field requires Kepler’s second law to be
valid.
• If the force is a central force, F = F(r)r̂,
and F = ma
 Example:
• At perihelon and aphelion ṙ = 0 and so
v = (h/r)θ̂. Thus

h h
vp = , va = (39)
rp ra

• As h is a constant
vp ra
= (40)
va rp

• Note that this is simply the conservation

of angular momentum: mra va = mrp vp .
Using rp = a(1 − ε) and ra = a(1 + ε)

vp 1+ε
= (41)
va 1−ε
 Exercise:

The semi-major axis of Mars’s orbit is 1.5237 AU and the planet’s orbital
eccentricity is 0.0934. Find the ratio of it’s velocities vp /va
(Answer: 1.6660/1.3814 = 1.206)
 Motion is in a Plane

• For central forces the torque vanishes:

N = r × F = r × F(r)r̂ = 0. (42)

• From Newtonian laws of dynamics we

know that the change in angular
momentum, L, is equal to torque
N = dL/dt
• Note also that the plane in
• As N = 0 for a central force we find that
which motion takes place is
L = constant perpendicular to the (constant)
• In order that L remains constant as a angular momentum vector L.
vector r and v should always remain in
the same plane.
 Discussion

• Note that we did not need the force to be

an inverse square law but only that it is a
central force, or equivalently that
angular momentum is conserved, for
explaining the second law of Kepler. Constancy of the angular
• This, however, is tricky since not all momentum vector implies that
central forces would lead to a closed the motion is confined to a
orbit: Bertrand’s theorem plane.
 Central force

• In the case of a central force depending only on the distance between

the objects
F = F(r)r̂ (43)

• The 2nd law of motion, F = ma, gives

h2
F(r) = m(r̈ − ) (44)
r3
and
dh
=0 (45)
dt
which implies that the specific angular momentum h = r2 θ̇ is a
constant.
 Newton’s Universal law of gravity

• We now assume that the central force is the universal law of gravity by
Newton
GM⊙ m
F=− r̂ (46)
r2
where M⊙ is the mass of the Sun and minus sign means that the force is
directed towards the center
• Thus, the 2nd law of motion, F = ma, gives

h2
 
GM⊙ m
− = m r̈ − 3 (47)
r2 r

which is a nonlinear ODE!

 Solution of the ODE
• In order to solve this equation we will make a change of variables
1
u≡ (48)
r

• Instead of finding r(t) we will find r(θ) since this makes it possible to
compare the solution with the conic section equation in polar
coordinates. We first find u(θ).
 2

• Thus − GMr2⊙ m = m r̈ − hr3 becomes

2
 
2 u 2 2d
−GM⊙ mu = m −h u − h2 u3
dθ2
and cancelling mu2 on both sides and dividing by h2 we get
GM⊙ d2 u
= +u
h2 dθ2
which is a linear equation!
2
• But how did we obtain r̈ = −h2 u2 dθ
d u
2?
 Solution of the ODE
• Let us first obtain ṙ: By chain-rule

dr dr du dθ
=
dt du dθ dt

• By using θ̇ = h/r2 = hu2 and dr/du = −1/u2

du
ṙ = −h (49)
dθ

• Now
dṙ dṙ dθ
r̈ = =
dt dθ dt

• By using θ̇ = h/r2 = hu2 and dṙ/dθ = −hd2 u/dθ2 from Eqn.(49)

d2 u
r̈ = −h2 u2 (50)
dθ2
 SHO in g

• The equation

GM⊙ d2 u
= +u
h2 dθ2
is similar to the equation one would get in
studying SHO in gravity.
• Fk = −kx thus

mg − kx = mẍ
 Solution of the ODE
• Using Eqns.(48) and (50) in Eqn.(47)

d2 u
−GM⊙ u2 = −h2 u2 − h2 u3 (51)
dθ2

• By arranging we obtain

d2 u GM⊙
+u= (52)
dθ2 h2

• Defining
GM⊙
z≡u− (53)
h2

• we obtain
d2 z
+z=0 (54)
dθ2
which is the simple harmonic oscillator (SHO) equation.
• The SHO equation d2 z
dθ 2
+ z = 0 has the solution

z = A cos θ + B sin θ (55)

or equivalently
z = A cos (θ + θ0 ) (56)

• θ0 is a constant angle depending on the choice of the coordinate frame.

• Thus referring to Eqn.(53)

GM⊙
u = A cos(θ + θ0 ) + (57)
h2

• and finally to Eqn.(48)

1
r= GM⊙
(58)
A cos (θ + θ0 ) + h2
• By multiplying the denominator and nominator with h2 /GM⊙ we can
rewrite this as
h2 /GM⊙
r= Ah2
(59)
1 + GM⊙ cos(θ + θ0 )

• This very much looks like the conic section equation

k
r(θ) =
1 + ε cos θ

• By comparing the two equations, we identify

h2
k= (60)
GM⊙

but it is not yet possible set θ0 and ε because we do not know the sign
of A.
• Now we will find A in terms of the energy E.
1 2 GM⊙ m
E ≡K+U = mv − (61)
2 r
where v2 = ṙ2 + r2 θ̇2 (recall v = ṙr̂ + rθ̇θ̂).
• Using θ̇ = h/r2 and ṙ = dθdr h
r2
as obtained before
• We thus obtain
"  #
2
1 h2 dr 2 GM⊙ m
E= m 4 +r − (62)
2 r dθ r
GM⊙
• By taking the derivative of u = A cos(θ + θ0 ) + h2
du
= −A sin(θ + θ0 ) (63)
dθ
and noting that dr
dθ = dr du
du dθ = − u12 dθ
du du
= −r2 dθ
dr
= r2 A sin(θ + θ0 ) (64)
dθ
and using this in (62) we obtain
1 2 2 2  GM⊙ m
E= mh A sin (θ + θ0 ) + r−2 − (65)
2 r
• Let us now eliminate r from
1 2 2 2  GM⊙ m
E= mh A sin (θ + θ0 ) + r−2 −
2 r
by using
1 GM⊙
= + A cos(θ + θ0 ) (66)
r h2
we find
"  2 #
1 2 2 2 GM⊙
E = mh A sin (θ + θ0 ) + + A cos(θ + θ0 )
2 h2
 
GM⊙
− GM⊙ m + A cos(θ + θ 0 )
h2
• The last equation
"  2 #
1 2 2 2 GM⊙
E = mh A sin (θ + θ0 ) + + A cos(θ + θ0 )
2 h2
 
GM⊙
− GM⊙ m + A cos(θ + θ0 )
h2

• Expanding and simplifying gives

(GM⊙ )2
 
1 2 2
E= m h A − (67)
2 h2

• or rather
1/2
2Eh2

GM⊙
A=− 2 1+ 2 2 (68)
h G M⊙ m
• Using this in the conic equation (59)

h2 /GM⊙
r=  1/2 (69)
2Eh2
1− 1+ 2 m
G2 M⊙
cos (θ + θ0 )

• If we now choose θ0 = π so that cos(θ + θ0 ) = − cos θ

h2 /GM⊙
r=  1/2 (70)
2
1 + 1 + G22Eh
2
M m
cos θ
⊙

• This can, now, be compared with the conic equation to find

1/2
2Eh2

ε= 1+ 2 m
(71)
G2 M ⊙
• Hence the eccentricity is determined by the energy and the angular
momentum of the body l = mh.
• Note that only E can make the term 2Eh2
2 m
G2 M⊙
negative.
• If E < 0 then 0 < ε < 1 and this gives an ellipse. The orbit is circular if
the energy has its minimum value

G2 M⊙
2
m
Emin = − (72)
2h2
obtained by requiring that the square root in Eqn.(71) is real.
• If E is positive ε > 1 and one obtains a hyperbola.
• If E = 0 then ε = 1 and we obtain a parabola.
• Note that from Eqn.(71) one can write the total energy as

G2 M ⊙
2
m 2
E= (ε − 1) (73)
2h2
or more simply E = Emin (1 − ε2 ).
 2
• Recall also that we have found (see Eqn.(60)) k = GM h
for a conic
⊙
section
• and that (see Eqn.(6)) k = a(1 − ε2 ) for an ellipse giving us

h2
1 − ε2 = (74)
aGM⊙

• And using this we obtain

GM⊙ m
E=− (75)
2a
 Exercise:

Show that  
2 1
v2 = GM − (76)
r a
Hint: Use Eqns.(75) and (61).
• If the total energy of the objects is negative (potential energy dominates
the kinetic energy) the system is bound and the orbits are elliptical.
• If the total energy is positive (kinetic energy of the system exceeds the
gravitational potential energy) the two objects are not bound and they
follow a hyperbolic orbit. After one near approach they recede from
each other indefinitely.
• The orbit will be parabolic in the critical case that the total energy is
zero i.e. kinetic energy is equal to the potential energy.
• Observations have shown that some comets approaching the Sun from
very large distances have orbits that are practically parabolic, although
they may be slightly elliptic or slightly hyperbolic.
• One of the big advances brought about by Newton’s theory was the
realization that planetary and cometary orbits can be understood in
terms of one and the same theory of gravitation. Prior to Newton
nobody had a clue about such a connection.
 Derivation of Kepler’s Third Law

• We now derive the Kepler’s third law

P2 = constant a3

from Newton’s laws.

• Note that we have written Kepler’s second law as

h
Ȧ = (77)
2
where h is a constant.
• By integrating this over the period P we obtain

hP
A= (78)
2
where A is the area of the ellipse which is equal to A = πab
 Derivation of Kepler’s Third Law
• We now take the square of both sides to obtain

4A2 4π 2 a2 b2
P2 = = (79)
h2 h2

• Recall from the section on ellipse that b2 = a2 (1 − ε2 ) (see Eqn.(3)).

• Recalling our result in Eqn.(74)

h2
1 − ε2 = . (80)
aGM⊙

• We now use this in Eqn.(79)

4π 2 3
P2 = a (81)
GM⊙

which shows that the constant Kepler’s third law P2 = constanta3 is

4π 2
GM⊙ .
 Derivation of Kepler’s Third Law

• Note that if one uses the law to study the orbits of moons moving
around planets one has to replace the mass of the Sun M⊙ with the
mass of the given planet.
• Newton’s theory of gravitation explains not only the law itself but also
the origin of the proportionality factor and thus how the law can be
applied to other systems.
 Two-body problem

• In the Kepler problem, the Sun was stationary at the origin of the
coordinate system and it was only the planets that were moving.
• Although this is a good approximation given that the Sun possesses
more than 99% of the mass of the Solar System, the approximation
breaks down for binary star systems where the masses of the stars are
comparable.
• According to Newtons third law, action-reaction, both objects are
subject to a force with the same magnitude and both should accelerate.
• This is called the two-body problem.
 Two-body problem

• The two-body problem can be reduced to

the single body (Kepler) problem by
defining the reduced mass
m1 m2
mr ≡ (82)
m1 + m2
where m1 and m2 are the masses of the
Two bodies with mass m1 and
objects, and center of mass (CoM)
m2 , at positions r1 and r2 , and
m1 r1 + m2 r2 their center of mass at R.
Rcm ≡ , (83)
m1 + m2
where r1 and r2 are the positions of the
objects
 Two-body problem

• The force that body 2 applies onto body

1 is
Gm1 m2
F21 = − êr (84)
|r1 − r2 |2
where
r1 − r2
êr = . (85)
|r1 − r2 |
Two bodies with mass m1 and
• We now define m2 , at positions r1 and r2 , and
their center of mass at R.
r ≡ r1 − r2 (86)
 Two-body problem

• This allows us to write the equation of

motion as
Gm1 m2
m1 r¨1 = − êr , (87)
r2
where êr = r/r.
• The force that body 1 applies onto body
Two bodies with mass m1 and
2 is F12 and acccording to Newton’s m2 , at positions r1 and r2 , and
third law of motion their center of mass at R.
F12 = −F21 (88)

Accordingly, the equation of motion of

the second body is
Gm1 m2
m2 r¨2 = êr , (89)
r2
 Two-body problem

• It is possible to show that the

acceleration of the CoM vanishes:
m1 r̈1 + m2 r̈2
R̈cm =
m1 + m2
F21 + F12 (90)
=
m1 + m2
=0

• This is merely a statement of the

conservation of momentum.
 Two-body problem

• We can now express the position vectors

of the two particles in terms of r and
Rcm :
m2
r1 = Rcm + r,
m1 + m2
m1 (91)
r2 = Rcm − r
m1 + m2

• Taking the time derivatives of these 2

times and multiplying with m1 and m2
respectively,
m1 m2
m1 r̈1 = r̈,
m1 + m2
m1 m2 (92)
m2 r̈2 = − r̈
m1 + m2
 Two-body problem

• Using the definition of reduced mass,

Eqn.(82),

m1 r̈1 = mr r̈, m2 r̈2 = −mr r̈ (93)

• Referring to the equations of motion,

Eqns.(87) & (89),
Gm1 m2 Gm1 m2
− êr = mr r̈, êr = −mr r̈
r2 r2
(94)
which are identical.
 Two-body problem

• Thus the two equatons of motion have

been reduced to a single equation

G(m1 + m2 )
− êr = r̈ (95)
r2
where we refered to the definition of
reduced mass, Eqn.(82).
• We have already solved this equation
and showed that it gives conic sections.
The only change is M → m1 + m2 , i.e.
that the two masses appear on equal
footing in the equations.
 Two-body problem

• Because of this symmetry of the system

4π 2
P2 = a3 (96)
G(M⊙ + m)

which reduces to the earlier result if

M⊙ ≫ m.
• This actually is the case in the Solar
system because the Sun is about 1000
times more massive than Jupiter, the
most massive planet in orbit.
 Stability of circular orbits
• Let us consider
h2
F(r) = m(r̈ − ) (97)
r3

• Here h = r2 θ̇ is a constant.
• If the central force is the universal law of gravity by Newton

GM⊙ m
F=− r̂ (98)
r2
where M⊙ is the mass of the Sun
• Then
h2
 
GM⊙ m
− = m r̈ − (99)
r2 r3
or rather
GM⊙ h2
− 2
= r̈ − 3 (100)
r r
 Stability of circular orbits

• Assume the planet is on a circular orbit

with radius r0

ṙ = r̈ = 0 (101)

• Then − GM ⊙ h2
r2
= r̈ − r3
implies

GM⊙ h2
= (102)
r02 r03

• Is this orbit stable?

 Stability of circular orbits

• Let us give a radial perturbation to the

orbit
r = r0 + r1 (103)
where r1 ≪ r0
• If the orbit is stable then r1 = r1 (t)
should remain small and not grow
indefinitely.

The reference orbit r0 and a

small perturbation r1 on it.
 Stability of circular orbits

• Using r = r0 + r1 in

GM⊙ h2
− = r̈ −
r2 r3
and noting that r̈0 = 0

GM⊙ h2
− 2
= r̈1 − 3
(104)
(r0 + r1 ) (r0 + r1 )

• This can be written as

−2 −3
h2
 
GM⊙ r1 r1
− 2 1+ = r̈1 − 3 1 +
r0 r0 r0 r0
(105)
 Stability of circular orbits

• Let us define x ≡ r1 /r0 and note that

r1 ≪ r0 implies x ≪ 1. Hence
(1 + x)n ≃ 1 + nx + O(x2 ).
• Keeping to a first order, we can write
−2 −3
h2
 
GM⊙ r1 r1
− 2 1+ = r̈1 − 3 1 +
r0 r0 r0 r0
as
GM⊙ h2
− (1 − 2x) = r0 ẍ − (1 − 3x)
r02 r03
(106)
 Stability of circular orbits

2
• Using GM2⊙ = h3 , the above equation can
r0 r0
be arranged as
GM⊙
ẍ + x=0 (107)
r03

• Note that this is of the form ẍ + ω 2 x = 0

where
GM⊙
ω2 ≡ (108)
r03

• This means x(t) = A cos ωt + B sin ωt

i.e. the perturbations are sinusoidal.
 Stability of circular orbits

• Thus a circular orbit in an inverse square

gravitational field is stable to radial
perturbations.
• Note that the period of the oscillations
P = 2πω is equal to

2π 3/2
Ppert = √ r0 (109)
GM
which is equal to the Keplerian period.
• This matching can also be seen by
finding that θ̇ = ω.
 Exercise:

Consider a gravitational field with the particular force field

K
F(r) = − (110)
rn
and show that stable orbits exist only for n < 3.
 Exercise:

Show that for

K K′
F(r) = − − (111)
r2 r4
where both K and K ′ positive, a circular orbit with radius r0 is stable for
r02 K > K ′ .
 Exercise:

Consider a gravitational potential

K
V(r) = − e−r/a (112)
r
where K and a are positive. Find F(r) by F = − dV
dr and show that a circular
orbit is stable if √
r0 1+ 5
< ≃ 1.62 (113)
a 2
 Exercise:

Show that the stability condition for general F(r) reads

3 F ′ (r0 )
+ >0 (114)
r0 F(r0 )

(Hint: Consider the Taylor expansion of F near r0 .)