Orbital Mechanics

These notes provide an alternative and elegant derivation of Kepler’s three laws for the
motion of two bodies resulting from their gravitational force on each other.

Orbit Equation and Kepler I

Consider the equation of motion of one of the particles (say, the one with mass m) with
respect to the other (with mass M ), i.e. the relative motion of m with respect to M :

r
r = −µ , (1)
r3
with µ given by
µ = G(M + m). (2)
Let h be the specific angular momentum (i.e. the angular momentum per unit mass) of m,

h = r × ṙ. (3)

The × sign indicates the cross product. Taking the derivative of h with respect to time, t,
we can write
d
(r × ṙ) = ṙ × ṙ + r × r̈
dt
= 0+0
= 0 (4)

The first term of the right hand side is zero for obvious reasons; the second term is zero
because of Eqn. 1: the vectors r and r̈ are antiparallel. We conclude that h is a constant
vector, and its magnitude, h, is constant as well. The vector h is perpendicular to both r
and the velocity ṙ, hence to the plane defined by these two vectors. This plane is the orbital
plane.
Let us now carry out the cross product of r̈, given by Eqn. 1, and h, and make use of the
vector algebra identity
A × (B × C) = (A · C)B − (A · B)C (5)
to write
µ  2

r̈ × h = − (r · ṙ)r − r ṙ . (6)
r3
 –2–

The r · ṙ in this equation can be replaced by rṙ since

r · r = r2 ,

and after taking the time derivative of both sides,

d d 2
(r · r) = (r ),
dt dt
2r · ṙ = 2rṙ,
r · ṙ = rṙ. (7)

Substituting Eqn. 7 into Eqn. 6 gives

µ  2

r̈ × h = − 3 rṙr − r ṙ ,
r !
ṙ ṙ
= µ − r ,
r r2
 
d r
= µ . (8)
dt r

Integrating the latter equation and considering h is constant we get

 
r
ṙ × h = µ +e , (9)
r
where the vector e is an integration constant called the Laplace-Runge-Lenz (LRL) vector.
It is clear that the vectors e and h are perpendicular to one another. Hence, e must be a
vector in the orbital plane!
Finally, let us make the scalar product of both sides of Eqn. 9 with the vector r. We have
 
r·r
r · (ṙ × h) = µ +e·r
r
= µ (r + re cos θ) , (10)

where θ is the angle between the vectors r and e. Applying the vector algebra identity

A · (B × C) = C · (A × B) (11)

to the left-hand side of Eqn. 10 we get

r · (ṙ × h) = h · (r × ṙ),
= h · h,
= h2 . (12)
 –3–

Replacing the left hand side of Eqn. 10 by Eqn. 12 then gives

h2 = µ (r + re cos θ) ,

or
h2 /µ
r= . (13)
1 + e cos θ

In analytical geometry, the general equation of an ellipse in polar coordinates, r and θ, with
one of the ellipse’s foci as the origin of the coordinate frame (see Figure 3.6 and Eqution
3.42 in the Ryden-Peterson textbook), is

a(1 − e2 )
r= , (14)
1 + e cos θ
The distance r is the magnitude of the position vector r, which makes an angle θ with the
reference axis along the line of apsides. This angle is called the true anomaly. The quantity
a is called the semimajor axis and is half the length of the largest diameter of the ellipse,
called the major axis. The two foci are located on the major axis and are equidistant from
the center of the ellipse. That distance is equal to ae. For e = 0, r = a, and the curve is
a circle with radius a. The two foci of a circle coincide with the center of the circle. Note
that the periapse distance, rp , and apoapse distance, ra , are obtained by entering f = 0 and
f = π, respectively, in Eqn. 14. Doing so we get

rp = a(1 − e), (15)

and
ra = a(1 + e). (16)

Comparing Eqns. 13 and 14, we conclude that the orbit of m around M is a conic section,
with a semi major axis a and eccentricity e related to h and µ via the equation

h2
= a(1 − e2 ),
µ
or p
h= µa(1 − e2 ) . (17)
The magnitude e of the LRL vector e is the eccentricity of the conic section. For 0 ≤ e < 1,
the conic section is an ellipse. In that case, the curve is closed and the mass m describes a
closed orbit around the attracting mass M , located at one of the foci of the ellipse. What
value of the angle θ makes r a minimum? The answer is, of course, that value of θ that makes
1 + e cos θ a maximum, which is when cos θ = +1, or θ = 0, i.e. when r is parallel to e. Thus,
 –4–

the LRL vector e is a vector that points from the point of central attraction to the point of
closest approach, the periapse point. The opposite point on the ellipse, when θ = π, is called
the apoapse point.1
For e = 1, r → ∞ as θ −→ ±π, which describes a parabola. For e > 1, the orbit is a
hyperbola. In this case r −→ ∞ along asymptotes defined by values of θ = θ∞ < π and given
by e cos θ∞ = −1. For e ≥ 1, Equation 13 holds unchanged, and parabolic or hyperbolic orbits
do occur in nature. For example, non-periodic comets describe hyperbolic orbits around the
Sun; they approach the Sun, swing by once, and then move away along their hyperbolic
path, to never come back. For parabolas and hyperbolas, however, the geometric description
of Eqn. 14 takes on a slightly different form. From here on,we restrict ourselves to elliptical
orbits unless specifically stated otherwise.

Kepler II

Let us now consider a right-handed, Cartesian coordinate frame with origin O at the center
of mass of the M, m system, and with the x, y-plane coinciding with the orbital plane. We
also consider a system of polar coordinates (r, θ), with origin at M , and a system of two
orthogonal, corotating unit vectors r̂ and θ̂ with cartesian coordinates (cos θ, sin θ, 0) and
(− sin θ, cos θ, 0), respectively. I refer to Section 3.1.1 in Ryden & Peterson for specifics and
figures. The velocity v = ṙ can be written as

v = vR r̂ + vT θ̂, (18)

with vR and vT the radial and tangential components of v, respectively. It is shown on page
64 of Ryden & Peterson that
vR = ṙ, (19)
and
vT = rθ̇. (20)

1
Both these labels can be modified to include the name of the attracting body. For example, for motion
around the Sun, we refer to the point of closest approach as the perihelion point. For the Moon and other
satellites of the Earth we call this point the perigee. In a binary star system, the point of closest approach of
one star as it orbits the other is called the periastron point. Similarly, the point of maximum distance would
be called the aphelion, apogee, apastron.
 –5–

In terms of the corotating r̂, θ̂, k̂ frame, the specific angular momentum vector h can be
written as
h = r × ṙ 
 r̂ θ̂ k̂ 
 
=  r 0 0 
 
 
 ṙ rθ̇ 0 
= 0r̂ + 0θ̂ + r2 θ̇k̂
= r2 θ̇k̂. (21)
Since h is a constant vector, r2 θ̇ is constant. Consider the position vector r sweeping an area
dA as the mass m moves in its orbit from the position at time t to the position at time t + dt.
dA can be considered to be the area of an infinitesimal triangle with sides r and rdθ, so we
can write
1
dA = r2 dθ, (22)
2
or
1 2
dA = r θ̇dt
2
1
= hdt. (23)
2
Integrating this from time t1 to time t2 , when m is at position 1 and 2, respectively, the area
A swept by the position vector is
1
A = h(t2 − t1 ). (24)
2
Hence, equal ∆t’s give equal A’s, which is Kepler’s second law, the law of areas: the position
vector sweeps out equal areas in equal intervals of time.

Kepler III

In Eqn. 24, let ∆t = t2 −t1 be the time for one complete revolution of m around M . This time
interval is called the period of the orbital motion. Let P be this period. The corresponding
area A swept by the position vector must then be the area of the entire ellipse, given by the
equation
A = πab, (25)
with a the semimajor axis and b the semiminor axis of the ellipse. The latter is related to
the former via the eccentricity e: √
b = a 1 − e2 . (26)
 –6–

For ∆t = P , Eqn. 24 then becomes

√ 1
πa2 1 − e2 = hP, (27)
2
and with Eqn. 17,
√ 1p
πa2 1 − e2 = µa(1 − e2 )P. (28)
2
√
The 1 − e2 drops out, and after squaring both sides and rearranging variables we get
Newton’s form of Kepler’s third law:
a3
µ = G(M + m) = 4π 2 . (29)
P2

The Vis Viva Equation

The following is an alternative derivation of Leibniz’ vis viva equation, the important Equa-
tion 3.67 in Ryden & Peterson.
The magnitude v of the velocity v of m with respect to M can be written as
v 2 = vR 2 + vT 2 , (30)
or, using Eqns. 19 and 20, as
v 2 = ṙ2 + r2 θ̇2 . (31)
In here, the ṙ can be obtained from differentiating Eqn. 14, which leads to
ṙ = a(1 − e2 ) sin θ θ̇ (1 + e cos θ)−2
p
r2 µa(1 − e2 )
= e sin θ
a(1 − e2 ) r2
r
µ
= e sin θ. (32)
a(1 − e2 )
The θ̇ comes from Eqn. 21:
h
θ̇ = 2
rp
µa(1 − e2 )
= (33)
r2
The vis viva equation then follows by substituting Eqns. 32 and 33 into 31 and carrying out
some algebra:
 
2 2 1
v =µ − . (34)
r a
 –7–

This is a very important equation. It tells us that, for given masses M and m, the orbital
speed only depends on the distance r between the two bodies and the orbit’s semi major
axis.
Applying vis viva at the periapse point, with r given by Eqn. 15, yields the orbital speed at
periapse passage, r r
µ 1+e
vp = , (35)
a 1−e
which corresponds to the maximum value v can have. Similarly, vis viva and Eqn. 16 give
the orbital speed at apoapse, r r
µ 1−e
va = , (36)
a 1+e
which is the minimum orbital speed of m. Note that the product
µ
vp va = (37)
a
is independent of the eccentricity e.

Energy

The total mechanical energy of the system of two bodies (M, m) is the sum of the kinetic
energy of M , the kinetic energy of m, and the gravitational potential energy of the (M, m)
system. Choose a coordinate frame with origin at the center of mass (CM) of the system.
Vectors r1 and r2 are the position vectors of M and m, respectively. Clearly, the relative
position of m with respect to M is
r = r2 − r1 . (38)
Because the position vector of the CM is the zero vector (CM is at O), and using the definition
of the CM, we have
M r1 + mr2 = 0, (39)
or
M r1 = −mr2 . (40)
Combining the latter with 38 gives
m
r1 = − r, (41)
M +m
and after taking the time derivative,
m
v1 = − v. (42)
M +m
 –8–

Likewise we obtain
M
v2 = + v. (43)
M +m
The total mechanical energy E then becomes:
1 1 Mm
E = M v1 2 + mv2 2 − G
2 2 r
 2  2
1 m 2 1 M Mm
= M v + m v2 − G
2 M +m 2 M +m r
 
1 Mm Mm
= v2 − G . (44)
2 M +m r
The quantity in parentheses, M m/(M + m), has the dimension of mass and is called the
reduced mass of the system. Substituting v 2 by the vis viva Eqn. 34, and after some algebra,
we get
1 Mm
E=− G , (45)
2 a
or, given µ = G(M + m),  
1 Mm µ
E=− . (46)
2 M +m a
Note that we have used the viv viva equation for eliptical motion, so Eqn. 45 gives the
total energy of the system for elliptical orbits. Given the masses M and m, E is uniquely
determined by the semimajor axis a of the orbit. Also, E is negative, indicating a bound
system.
For parabolic orbits, both the vis viva (Eqn. 34) and energy (Eqn. 45) equations are still
valid, provided we set a = ∞. We then have
Eparab = 0, (47)
and the vis viva equation becomes
2 2G(M + m)
vpara = . (48)
r
p
A particle m moving in a parabolic orbit with this parabolic speed v = 2G(M + m)/r will
make it to infinity, i.e. will “escape” the gravitational pull of M . This speed is referred to as
the escape speed, vesc . If m << M , the escape speed reduces to
r
2GM
vesc = , (49)
r
which is Eqn. 3.62 in Ryden & Peterson.
For hyperbolic orbits, the total energy is positive and given by
1 Mm
E= G . (50)
2 a
 –9–

Virial Theorem

The −GM m/a part of Eqn. 45 represents the “mean potential energy” of the system, with
the mean taken over one orbital cycle. Thus for a two-body orbit we find the total energy to
be equal to half the time-averaged potential energy. This is the so-called virial theorem for
a gravitationally bound system of many particles. The theorem can be expressed as
1
< E >= < U >, (51)
2
or, since
< E >=< K > + < U >, (52)
as
1
< K >= − < U > . (53)
2
A rigorous derivation of the virial theorem is in Ryden & Peterson Section 3.4.
The virial theorem is very useful in astronomy in the study of large stellar system such as star
clusters and galaxies, It also plays an important role during the star formation process. When
part of a nebula (a cloud of interstellar gas and dust) collapses gravitationally, its (negative)
potential energy decreases (the inter-particle distance decreases, hence the absolute value of
the PE increases, but since PE is a negative number, the PE decreases). According to the
virial theorem, half of the lost PE goes into KE of the particles, i.e. the internal energy (and
temperature) of the collapsing blob increases. What happens to the other half? That other
half is carried out of the blob of collapsing gas by photons, i.e. it gets radiated away by light.
So, during star formation, before the onset of nuclear fusion, stars in the process of forming
already shine.

Interplanetary Travel

We now have all the ingredients for launching probes and get them from one orbit to another
around the same planetary body, or from one planet to another: see the idea of Hohmann
transfer orbits in Ryden & Peterson, Section 3.3, and our discussion in class.