Alain J.

Brizard
Saint Michael's College

Motion in a Central-Force Field

1 Motion in the Center-of-Mass Frame

The Lagrangian for an isolated two-particle system is
m1 m
L = j_r1j 2 + 2 j_r 2j2 ¡ U(r 1 ¡ r2);
2 2
where r1 and r2 represent the positions of the particles of mass m1 and m2 , respectively,
and U(r1 ; r 2) = U(jr1 ¡r 2j) is the potential energy for an isolated two-particle system (see
Figure below).

Let us now de¯ne the position R of the center-of-mass (CM)

m1 r1 + m2 r2
R = ;
m1 + m2
and de¯ne the inter-particle vector r = r1 ¡ r2 , so that the particle positions can be
expressed as
m m
r1 = R + 2 r and r 2 = R ¡ 1 r;
M M
where M = m1 + m2 is the total mass of the two-particle system. The Lagrangian of the
isolated two-particle system thus becomes
M _ 2 ¹ 2
L = _ ¡ U(r);
jRj + jrj
2 2

1
where µ ¶ ¡1
m1 m2 1 1
¹ = = +
m1 + m2 m1 m2
denotes the reduced mass of the two-particle system.
For an isolated system, the CM canonical momentum
@L _
P =
= MR
@R_
is a constant of the motion. The CM reference frame is de¯ned by the condition R = 0,
i.e., we move the origin of our coordinate system to the CM position (the Figure below
shows the case where m1 > m2 ).

In this case, the Lagrangian for an isolated two-particle system in the CM reference
frame is
¹
L(r; r_ ) = j r_ j 2 ¡ U (r);
2
where m2 m
r1 = r and r2 = ¡ 1 r:
M M
Hence, once the Euler-Lagrange equation for r
à !
d @L @L
= ! ¹ Är = ¡ rU(r)
dt @ r_ @r
is solved for r = r(t), the motion of m1 and m2 are determined through r1 (t) = (m2=M ) r(t)
and r2 (t) = ¡ (m1 =M) r(t).

2 Motion in a Central-Force Field

A particle moves under the in°uence of a central-force ¯eld F(r) = F (r) br if the force on
the particle is independent of the angular position of the particle about the center of force

2
and depends only on its distance r from the center of force. Here, the magnitude F (r)
(which is positive for a repulsive force and negative for an attractive force) is de¯ned in
terms of the central potential U (r) as F (r) = ¡ U 0 (r).

2.1 Lagrangian Formalism

The motion of two particles in an isolated system takes place on a two-dimensional plane.
When these particles move in a central-force ¯eld, the Lagrangian is simply
¹³ 2 ´
L = r_ + r 2 µ_2 ¡ U(r); (1)
2
where polar coordinates (r; µ) are most conveniently used. Since the potential U is inde-
pendent of µ, the canonical momentum
@L
pµ = = ¹ r2 µ_ ´ ` (2)
@ µ_
is a constant of motion (here, labeled `). The Euler-Lagrange equation for r, therefore,
becomes the radial force equation
³ ´ `2
¹ Är ¡ r µ_2 = ¹ rÄ ¡ = F (r): (3)
¹ r3
In this description, the planar orbit is parametrized by time, i.e., once r(t) and µ(t) are
obtained, a path r(µ) onto the plane is de¯ned.
Since µ_ does not change sign on its path along the orbit, we may replace r_ and Är with
r0 (µ) and r 00 (µ) as follows. First,
µ ¶0
` r0 ` 1
r_ = µ_ r0 = = ¡ :
¹ r2 ¹ r
Next, using Eq. (2)
2 ` r_ 2 `2 r0
µÄ = ¡ = ¡ ;
¹ r3 ¹ 2 r5
we ¯nd
`2 h i
rÄ = µÄ r0 + µ_ 2 r 00 = ¡ 2 5 2 (r0 )2 ¡ r r00 :
¹ r
Lastly, using the identity
µ ¶ 00 Ã !0
1 r0 1 h 0 2 00
i
= ¡ 2 = 2 (r ) ¡ r r ;
r r r3
we ¯nd an expression for rÄ:
`2 µ 1 ¶00
Är = ¡ ;
¹ 2 r2 r

3
and the radial force equation (3) becomes

¹ dU (s)
s00 + s = ¡ F (1=s) = ¡ ; (4)
`2 s2 ds
where s(µ) = 1=r(µ) and U(s) = (¹=`2) U(1=s).
Note that the form of the potential can be calculated from the solution s(µ) = 1=r(µ)
as follows. For example, consider the particle trajectory described in terms of the solution
r(µ) = r0 sec(® µ), where r0 and ® are constants, then
³ ´ dU(s)
s00 + s = ¡ ®2 ¡ 1 s = ¡ ;
ds
and thus
1 ³ 2 ´ `2 ³ 2 ´
U (s) = ® ¡ 1 s2 ! U(r) = ® ¡ 1 :
2 2¹ r2
Note also that the function µ(t) is determined from the relation
` ¹ Z µ 2
µ_ = ! t(µ) = r (µ) dµ:
¹r2 (µ) ` 0
Returning to our example, we ¯nd
v
u à !2
¹r02 Z ®µ ¹r 20 u
t1 ®` t
t(µ) = sec2 Á dÁ = tan(®µ) ! r(t) = r0 +
®` 0 ®` ¹r20

and the total energy

®2 `2
E = ;
2¹r20
is determined from the initial conditions r(0) = r0 and r(0)
_ = 0.

2.2 Hamiltonian Formalism

The Hamiltonian for the central-force problem is

p2r `2
H = + + U(r);
2¹ 2 ¹ r2
where pr = ¹ r_ is the radial canonical momentum. Since energy is also conserved, we solve
¹ r_ 2 `2 ¹ r_ 2
E = + + U(r) = + V (r);
2 2 ¹ r2 2
as s
2
r_ = § [ E ¡ V (r) ]; (5)
¹

4
where V (r) is known as the e®ective potential and the sign § depends on initial conditions.
This equation can then be used with Eq. (2) to yield
` ` dr ¡ ds
dµ = 2
dt = 2
= q ; (6)
¹r ¹ r r_ ² ¡ 2 U(s) ¡ s2

where ² = 2¹ E=`2, or q
s 0(µ) = § ² ¡ 2 U (s) ¡ s 2: (7)
We readily check that this equation is a proper solution of the radial force equation (4)
since
s0 [dU=ds + s] dU
s00 = q = ¡ ¡ s
² ¡ 2 U(s) ¡ s2 ds

is indeed identical to Eq. (4). Hence, for a given central-force potential U (r), we can solve
for r(µ) = 1=s(µ) by integrating
Z s
d¾
µ(s) = ¡ q ; (8)
s0 ² ¡ 2 U(¾) ¡ ¾2
where s0 de¯nes µ(s0) = 0, and performing the inversion µ(s) ! s(µ).

2.3 Turning Points

Eq. (7) yields the following energy equation

¹ 2 `2 `2 h 0 2 2
i
E = r_ + + U(r) = (s ) + s + 2 U(s) ;
2 2¹ r 2 2¹
where s0 = ¡ ¹ r=` _ = ¡ pr =pµ . Turning points are those special values of rn (or sn ) (n =
1; 2; :::) for which " #
`2 `2 s 2n
E = U(rn ) + = U (sn ) + ;
2¹ r2n ¹ 2
i.e., r_ (or s0 ) vanishes at these points. If two non-vanishing turning points r2 < r1 < 1
(or 0 < s 1 < s2) exist, the motion is said to be bounded in the interval r2 < r < r1 (or
s1 < s < s2), otherwise the motion is unbounded.

3 Kepler Problem
We now solve the Kepler problem where U(r) = ¡ k=r, where k is a constant, so that
U(s) = ¡ s 0s, where s 0 = ¹k=`2. The turning points for the Kepler problem are solutions
of the quadratic equation
s2 ¡ 2 s0 s ¡ ² = 0;

5
 q
which can be written as s = s 0 § s20 + ²

s1 = s 0 (1 ¡ e) and s2 = s 0 (1 + e);

where q q
e = 1 + ²=s20 = 1 + 2 E`2=¹k2:
We note that motion is bounded when E < 0 (0 < e < 1) and unbounded when E ¸ 0
(e > 1).

3.1 Bounded Keplerian Orbits

We will now look at the bounded case (e < 1). We de¯ne µ(s2 ) = 0, so that for the Kepler
problem, Eq. (8) becomes
Z s
d¾
µ(s) = ¡ q ; (9)
s 0 (1+e) s20 e2 ¡ (¾ ¡ s 0)2

which can easily be integrated by using the identity

µ ¶
¡ dx x
p = d arccos ;
a ¡x
2 2 a

so that Eq. (9) yields µ ¶

s ¡ s0
µ(s) = arccos :
s0 e
This equation can easily be inverted to yield

s(µ) = s 0 (1 + e cos µ): (10)

We can readily check that this solution also satis¯es the radial force equation (4).

3.1.1 Kepler's First Law

The solution for r(µ) is now trivially obtained from s(µ) as

r0
r(µ) = ; (11)
1 + e cos µ
where r0 = 1=s0 denotes the position of the minimum of the e®ective potential
à !
1 `2
V 0 (r0) = 2 k ¡ = 0:
r0 ¹ r0

6
 Eq. (11) generates an ellipse of major radius

r0 k
a = 2
=
1 ¡e 2 jEj

and minor radius v

u
p u `2
b = a 1 ¡ e2 = t
2¹ jEj
and, therefore, yields Kepler's First Law.

3.1.2 Kepler's Second Law

Using Eq. (2), we ¯nd

¹ 2 2¹
dt = r dµ = dA(µ);
` `
R
where dA(µ) = r dr dµ = 12 [r(µ)] 2 dµ denotes an in¯nitesimal area swept by dµ. When
integrated, this relation yields Kepler's Second law
2¹
¢t = ¢A; (12)
`
i.e., equal areas are swept in equal times since ¹ and ` are constants.

3.1.3 Kepler's Third Law

The orbital period T of a bound system is de¯ned as

Z 2¼
dµ ¹ Z 2¼ 2 2¹ 2¼ ¹
T = = r dµ = A = ab
0 µ_ ` 0 ` `

7
where A = ¼ ab denotes the area of an ellipse with major radius a and minor radius b.
Using the expressions for a and b found above, we ¯nd
v v
u u
2¼ ¹ k u `2 u ¹ k2
T = ¢ ¢ t = 2¼ t :
` 2jEj 2¹ jEj (2 jEj)3

If we now substitute the expression for a = k=2jEj and square both sides of this equation,
we obtain Kepler's Third Law
(2¼)2¹ 3
T2 = a : (13)
k
Note that in Newtonian gravitational theory, k=¹ = G (m1 + m2 ); although Kepler's Third
Law states that T 2=a3 is a constant for all planets in the solar system, we ¯nd that this is
only an approximation that holds for m1 À m2.

3.2 Unbounded Keplerian Orbits

We now look at the case where the total energy is positive or zero (i.e., e ¸ 1). Eq. (11)
yields r (1 + e cos µ) = r 0 or
à !2
p er r20
e2 ¡ 1 x ¡ p 2 0 ¡ y2 = :
e ¡1 e2 ¡ 1

For e = 1, the particle orbit is a parabola x = (r02 ¡ y2 )=2r0 , with distance of closest
approach at x(0) = r0=2, while for e > 1, the particle orbit is a hyperbola.

3.3 Laplace-Runge-Lenz Vector

Let us now investigate an additional constant of the motion for the Kepler problem. First,
we consider the time derivative of the vector p £ L, where the linear momentum p and
angular momentum L are
³ ´
p = ¹ r_ rb + rµ_ µb and L = ` zb = ¹r2 µ_ zb:

The time derivative of the linear momentum is p_ = ¡ rU(r) = ¡ U 0 (r) rb while the angular
momentum L = r £ p is itself a constant of the motion so that
d dp
(p £ L) = £ L = ¡ ¹_r ¢ rU r + ¹r ¢ rU r_
dt dt
d
= ¡ (¹ U r) + ¹ (r ¢ rU + U) r_ ;
dt

8
hence the vector A = p £ L + ¹ U(r) r is a constant of the motion if the potential U(r)
satis¯es the condition r ¢ rU(r) + U(r) = 0. For the Kepler problem, with central potential
U(r) = ¡ k=r, the Laplace-Runge-Lenz (LRL) vector
à !
`2
A = p £ L ¡ k¹ rb = ¡ k¹ rb ¡ ` ¹r_ µb
r
is, therefore, a constant of the motion since r ¢ rU = ¡ U.
Since the vector A is constant in both magnitude and direction, we choose its direction
to be along the x-axis and its amplitude is determined at the distance of closest approach
rmin = r0 =(1 + e) and we can easily show that A ¢ rb = A cos µ leads to the Kepler solution
r0
r(µ) = ;
1 + e cos µ
where r0 = `2=k¹ and e = A=k¹.
Note that if the Keplerian orbital motion is perturbed by the introduction of an ad-
ditional potential term ±U(r), we can show that the LRL vector is no longer conserved
(i.e., dA=dt 6= 0) and that the direction of the Keplerian elliptical orbit precesses with a
precession frequency
A dA
!p (µ) = zb ¢ 2 £ ;
A dt
where the unperturbed Kepler solution r(µ) is to be used.

4 Isotropic Simple Harmonic Oscillator

We now investigate the case when the central potential is of the form
k 2 ¹k
U (r) = r ! U (s) = : (14)
2 2`2 s2
The turning points for this problem are expressed as
µ
1 ¡ e ¶4 µ
1 + e ¶4
1 1
1 1
r1 = r0 = and r2 = r0 = ;
1+e s2 1¡e s1
where r0 = (`2=¹k)1=4 = 1=s 0 is the radial position at which the e®ective potential has a
minimum V0 = k r02 and v
u à !2
u kr02
t
e = 1 ¡ :
E
Here, we see that orbits exist and are always bounded for E > V0 (and thus 0 e 1).
Next, using the change of coordinate q = s2 in Eq. (8), we obtain
¡1 Z q dq
µ = q ; (15)
2 q2 (¹=`2 ) [2E q ¡ k] ¡ q2

9
where q2 = (1 + e) ¹E=`2. We now substitute q(') = (1 + e cos ') ¹E=`2 in Eq. (15) to
obtain " Ã !#
1 1 `2 q
µ = arccos ¡1
2 e ¹E
or p
r0 1 ¡ e2
r(µ) = p : (16)
1 + e cos 2µ
This equation describes the ellipse

x2 y2
+ = r02
(1 ¡ e) (1 + e)
p p
of semi-major axis a = r0 1 + e and semi-minor axis b = r0 1 ¡ e. Lastly, we note that
one revolution along the orbit r(µ) corresponds to an angular period of ¼, i.e., r(µ + ¼) =
r(µ), and not 2¼ as found in the Kepler problem. The area of the ellipse A = ¼ ab =
¼ (`2=¹E) while the period is
Z ¼
dµ ¹A ¼`
T (E; `) =
0 _µ = ` = E :

If we introduce the angular frequency ! = ¼=T , then we ¯nd the important relation between
energy and angular momentum E = ` !.

10