You are on page 1of 2

KALMAN KNIZHNIK - DERIVATION OF PLANETARY ORBIT EQUATION

We begin the derivation by defining our variables. Let one planet with mass m1 be located at
r1 and the second planet with mass m2 be located at r2 . Let us call r r1 r2 . Furthermore
let us introduce the reduced mass = (m1 m2 )/(m1 + m2 ) and the position of the center-of-mass
R = (m1 r1 + m2 r2 )/(m1 + m2 ). Now, the kinetic energy the system is
1
1
T = m1 r1 2 + m2 r2 2
2
2

(1)

we can rewrite it in terms of the reduced mass and center of mass coordinate. It is easy to show
(for instance using Griffiths QM problem 5.1) that
r1 = R +

m2
r,
m1 + m2

r2 = R

m1
r
m1 + m2

(2)

Plugging this into equation 1, then, the kinetic energy is


1 2 1 2
T = MR
+ r
2
2

(3)

where M = m1 + m2 , and the Lagrangian L = T U of the system is


1 2 1 2 Gm1 m2
L = MR
+ r +
2
2
r

(4)

where I have used the known Newtonian formula for the potential energy between two masses
U =

Gm1 m2
r

(5)

Now, the following procedure is made much easier by going into the center-of-mass reference frame,
= 0. Thus the Lagrangian is
so that R
1
Gm1 m2
L = r2 +
2
r

(6)

We would like to get rid of the vector r , and to do that we need to pick a coordinate system. It

turns out to be most natural, and perhaps obvious, to use polar coordinates, where r = r
r + r ,
so the Lagrangian can be written
1
Gm1 m2
L = (r 2 + r2 2 ) +
2
r

(7)

and from the angular Euler-Lagrange equation we obtain:


L
= r2 = const = l

l
= 2
r

(8)

while the radial Euler-Lagrange equation yields:


Gm1 m2
r2

(9)

l2
Gm1 m2

r3
r2

(10)

r = r 2
inserting from equation 8 into equation 9 gives:

r=

This is the differential equation of motion of a planet, and solving it would give us the form of its
orbit. To solve it, we make the substitution r u1 , where u = u(). Since u is a function of
only, the time derivatives are:
dr
d(1/u)
1 du
1 du
du l
=
= 2
= 2 =
dt
dt
u dt
u d
d

(11)

where in the last equality I have used the formula for in equation 8. Then
l d du
l d d du
l2 u2 d2 u
d2 r
=

dt2
dt d
d dt d
2 d2

(12)

So we can finally rewrite the equation of motion (10) as

l2 u2 d2 u
l2 u3
=
Gm1 m2 u2
d2

d2 u
Gm1 m2
= u +
2
d
l2

(13)

We are almost there. Now let w() = u() Gm1 m2 /l2 . Then
d2 w
= w
d2

(14)

which has the well known solution w() = Acos( + ), where both A and are constants. We can
always choose = 0 by a convenient choice of , so the solution is w() = Acos(), or, going back
to u: u() = Acos() + Gm1 m2 /l2 , and finally going back to the relative position r:
r() =

1
Acos() +

Gm1 m2
l2

(15)

and transforming to the standard notation for a conic section:


r() =

l2
Gm1 m2

1 + cos()

(16)

where I have called  = Al2 /Gm1 m2 . This is the famous equation of planetary motion.
The theory of conic sections tells us that this is the equation (in polar coordinates) of an ellipse
with eccentricity . If we take an ellipse with semi major axis a and semi minor axis b, then the
eccentricity is defined as
b2
(17)
=1 2
a
from which it is evident that a circular orbit is the special case of  = 0, a parabolic orbit is the
special case where  = 1, and it turns out that a hyperbolic orbit has  > 1. We can also define the
perihelion, where = 0
rmin =

l2
Gm1 m2

1+

(18)

and the aphelion, where = :


rmax =

l2
Gm1 m2

1
as the positions of closest and farthest approach, respectively.

(19)