Mathematics for Physics 3: Dynamics and Dierential Equations 87

19 Lecture 19: Central forces and the orbit equation

In the previous lecture we studied the motion under the influence of a conservative central force. We first saw that
the spherical symmetry of the potential gives rise to the conservation of the angular momentum vector L = r p.
The fact that the direction of L is fixed implies that the motion is confined to a plane. This directed us to using plane
polar coordinates r and to describe the motion. Furthermore, we observed that the conservation of the magnitude
of L implies that
r2 h = |L|/m (314)
is a constant of motion, thus reducing the three dimensional problem to one dimension. The remaining equation is
the radial one,
h2 Fr (r)
r 3 = . (315)
r m
This is a non-linear second order dierential equation.

Circular orbits
The simplest example of an orbit in a central potential is that of a circular orbit where the distance from the origin
is constant, r = a. One immediately deduces then that r = 0 and r = 0 so the orbit equation (323) yields

r4 Fr (r) a3 Fr (a)
r = 2
with r = a = h= . (316)
h m m
The initial conditions clearly need to be such that r2 = h with h given by (316). This can be achieved if all of the
following conditions hold at time t = 0:
The particle is at r(0) = a
It has no radial velocity: r(0)
=0

Its velocity in the direction (tangent to the orbit) v = a is such that the angular momentum mr2 is mh
with h given by (316), namely
h aFr (a)
v = a = =
a m
This is equivalent to the requirement that the force will be
mv 2
Fr (a) =
a
which is precisely the value of the required centripetal force for a circular motion, as we learnt in Lecture 4.

Stability of circular orbits

Let us now examine the stability of a circular orbit. Clearly this depends on the details of the force Fr (r). In some
cases a nearly circular orbit will remain stable, and in others a small perturbation may increase in an unbounded
manner.
To this end assume that r(t) = a + (t) where is a small perturbation, (0) a. Let us expand the general
radial orbit equation assuming that is small, and verify if the solution is consistent with this assumption. If it is,
we have verified stability. Substituting r = a + into (315) we get:

h2 Fr (a + )
3
= . (317)
(a + ) m
Expanding in /a we get:
h2 Fr (a) dFr (a)
1 3 = + . (318)
a3 a m dr m

3 F (a)
Now since for the circular orbit we have, according to equation (316), h = a m r
, eq. (318) reduces to:

3h2 1 dFr (a)
+ = 0. (319)
a4 m dr

If
3h2 1 dFr (a)
>0 (320)
a4 m dr
Mathematics for Physics 3: Dynamics and Dierential Equations 88

then we define
3h2 1 dFr (a)
2
a4 m dr
and deduce that (319) describes harmonic motion, (t) = A cos(t + ), which is bounded by its amplitude A. It
thus follows that if the condition (320) is met, a small perturbation around a circular orbit will remain small. On
the other hand, if the condition (320) is violated the perturbation may grow indefinitely.
It is useful to translate our stability condition for the circular orbit to a condition that does not involve h. This
can be done using the relation established above, equation (316), h2 = a3 Fr (a)/m, yielding:
1 dFr (a) 3
< (321)
Fr (a) dr a
where we used the fact that Fr (a) < 0. As expected, the stability condition depends only on the force law.

Derivation of the orbit equation for r()

An alternative to solving the radial equation of motion (315) directly is to solve for r(). To this end we need to
determine the radial velocity r and acceleration r in term of r() and its derivatives. Let us denote derivatives with
respect to by a prime, e.g. dr()/d = r (). We then have:
dr d r ()
r = = r () = h 2
d dt r
and
d r () d d r () h2 r ()r2 2r(r ())2 h2 (r ())2
r = h 2 =h = 2 = 4 r () 2
dt r dt d r2 r r4 r r
where we repeatedly used (314) to substitute for .
Substituting the result for r into the radial equation of motion (315) we get:

h2 (r ())2 h2 Fr (r)
4
r () 2 3 = . (322)
r r r m
or, equivalently,
(r ())2 r4 Fr (r)
r () 2 r() = . (323)
r h2 m
This is a non-linear second order dierential equation, which allows direct determination of the orbit r().
It is called the orbit equation.

Determination of the force law from the orbit equation

A simple application of (323) is the determination of the force law given a particular orbit.
Consider the following hyperbolic trajectory:
l
r() = (324)
1 + cos()
with > 1 (note that this trajectory corresponds to the near branch of a hyperbola, where the far branch would be
described by a similar equation with the 1 in the denominator replaced by -1). Because r > 0 this must correspond
to 1 + cos() > 0, namely < < where = arccos(1/). The picture is then as follows: the particle starts
far away along one ray, say = , it then bends such that its trajectory is tangent at = 0, where its distance
l
to the centre of force is minimal, rmin = 1+ , and then proceeds to increase its distance (in a symmetric way to the
approach). An example is shown in figure 15.
Let us now compute the force law based on this trajectory using (323). We have:

h2 m (r ())2
Fr (r) = 4 r () 2 r() . (325)
r r
Using now (324) we get:

1
r() = l (326)
1 + cos()

sin()
r () = l (327)
(1 + cos())2

cos()(1 + cos())2 + 2(1 + cos())2 sin2 () cos() + 2 cos2 () + 22 sin2 ()
r () = l = l (328)
(1 + cos())4 (1 + cos())3
Mathematics for Physics 3: Dynamics and Dierential Equations 89

Figure 15: Hyperbolic trajectories with eccentricity = 2 in polar coordinates. The origin (from which r is measured)
is the centre of the force. Here both the near branch r() = l/(1+ cos()) and the far branch r() = l/(1+ cos())
are shown, alongside the two asymptotic lines.

Substituting these into (325) we obtain:

h2 m (r ())2 h2 m r2 h2 m/l
Fr (r) = 4 r () 2 r() = 4 = .
r r r l r2

where we skipped straightforward algebra in simplifying the square brackets. The result is that the corresponding
force is an attractive square law force.