Department of Physics Indian Institute of Technology Madras

Physics-1 PH1010 — Tutorial 8 Solutions

Q1

(a) Given U (r) = − rCn . Therefre the Uef f is

C L2
Uef f = − + (1)
rn 2mr2

Radius of the circular orbit is given by the condition

′
Uef f (r0 ) = 0
nC L2
=⇒ − =0
r0n+1 mr03
  1
n m C n−2
=⇒ r0 = (2)
L2
nmC
=⇒ r0n−2 = (3)
L2
′′ (r ) > 0.
(b) For circular orbit to be stable against the radial perturbations, we need Uef f 0
Therefore,
′′ n(n + 1)C 3L2
Ueff =− +
rn+2 r04
 02 
1 3L n(n + 1)C
= 4 + (4)
r0 m r0n−2

1
 Substituting the value of r0n−2 from Eq.(3) into the Eq.(4) we get

′′ L2
Ueff = (2 − n). (5)
mr04
′′ > 0 only if n < 2. For integer n, the only possible value is n = 1.
Therefore, Ueff
′′ (r ) →
Note that for n = 0, the particle moves in a straight line, r0 → ∞ (Eq.(2)), and Uef f 0
0.
(c) Frequency of small (radial) oscillations is :
s
′′
Ueff(r 0)
ωr = (6)
m
L mC 2
= = . (7)
mr02 L3
L
The angular frequency of orbital motion in the circular path is ω0 = mr02
. Comparing ωr
in Eq.(6) with ω0 , we see ωr = ωo .
This implies that the perturbed orbit is ” closed ”. This makes sense, since all orbits for
the 1/r potential are indeed closed paths, like ellipses in this particular case

Q2

L2
• For a parabolic orbit ϵ = 1, therefore rmin = 2mα

• For the given trajectory, we require

L2
R= . (8)
2mα
Where R is radius of the planet, L is angular momentum, α = G × M × m, M is mass of
the planet and m is mass of the asteroid. Therefore ,
L2
R= (9)
2GM m2

2
 • Let ω be the angular velocity at the closest approach, then

L = mR2 ω (10)

• Therefore, using Eq.(9) and Eq.(10) we have ω 2 = 2GM

R3
= 8πG
3 ρm . Where ρm = M
4πR3 /3
is
the mass density of the planet.

Q3

(a) We can assume with no loss of generality that L ⃗ = L ẑ, θ = π/2 and θ̇ = 0. Thus, the
particle is moving in the xy-plane and we can replace r̂ by ρ̂. We need to show that
⃗
dA/dt = 0 in order to show that it is a constant of motion. Before that, notice that
⃗ ⃗ ⃗ is a vector lying in the plane of the orbit.
A · L ≡ 0 implying that A
It suffices to show the vanishing of the following time derivative to prove the constancy
⃗
of A.  
d h ⃗
i dr̂ ⃗ d⃗ p dr̂ ⃗ k
mk r̂ + L × p⃗ = mk + L × = mk − L × r̂ , (11)
dt dt dt dt r2
where we have used Newton’s Second Law to obtain the last term. Now, in the xy-plane,
dr̂ ∂ r̂ ∂ r̂ L
= φ̇ + θ̇ = φ̂φ̇ = φ̂ .
dt ∂φ ∂θ mr2
Substituting the above result in Eq. (11), we obtain (on specialising to the xy-plane)
d h ⃗ × p⃗ = kL [φ̂ − (ẑ × ρ̂)] = 0
i
mk r̂ + L
dt r2
⃗ is a constant of motion.
Thus, A
Geometric Interpretation: Since A ⃗ is a constant of motion, its value can be evaluated
at any instant of time. We choose the instant when the particle is closest to the center
of attraction. At this instant, ṙ = 0 and r = rmin . This implies that
L 2
p⃗ = mrmin φ̇ = φ̂ ⇒ ⃗ × p⃗ = − L r̂ .
L
rmin rmin

One can thus see that A ⃗ ∝ r̂, that is, it is a constant vector whose direction is along the
vector joining the center of attraction to the point of closest approach. The constancy
of the Laplace-Runge-Lenz vector implies that the orbit does not precess. For instance,
suppose we add a term c/r2 (c is a “small positive constant”) to the gravitational potential,
then A⃗ is no longer a constant vector and the orbit will precess, that is, the point of closest
approach will slowly move.

(b) The time-average of a quantity O is related to the angular-average of another quantity.

1 T 1 T dt
Z Z
2πm 2
⟨O⟩t ≡ dt O = dφ O = ⟨r O⟩φ ,
T 0 T 0 dφ LT

where we have used L = mr2 φ̇ to obtain the last equality above. For an elliptical orbit,
one has
r0
= (1 + ϵ cos φ) ,
r

3
 where ϵ2 = 2Ẽ + 1 = 2(E/U0 ) + 1. Using the expression for r0 , one can see that the above
relation between the time and angular averages can be rewritten as
(1 − ϵ2 )3/2 2
⟨O⟩t = ⟨r O⟩φ .
r02

We need to evaluate the angular and time average of r. Let us first evaluate the angular-
average: 1 Z 2π
1 dφ r0
⟨r⟩φ = r0 =p .
2π 0 (1 + ϵ cos φ) (1 − ϵ2 )
The time average of r is obtained by converting into an angular-average as shown above:
Z 2π
(1 − ϵ2 )3/2 3 2 3/2 1 dφ 1 + (ϵ2 /2)
⟨r⟩t = ⟨r ⟩φ = r0 (1 − ϵ ) = r0 .
r02 2π 0 (1 + ϵ cos φ)3 1 − ϵ2
Note: The two averages ⟨r⟩t and ⟨r⟩φ are not equal except when ϵ = 0 which is the case
when the ellipse becomes a circle.
(c) Since E is a constant of motion one has ⟨E⟩t = ⟨E⟩φ = E. Let T and U denote the kinetic
and potential energy of the particle. One has

⟨T ⟩t + ⟨U ⟩t = E ⇒ ⟨T ⟩t = E − ⟨U ⟩t .

It thus sufficient to compute ⟨U ⟩t . One obtains

(1 − ϵ2 )3/2
 
1
⟨U ⟩t = −k = −k ⟨r⟩φ = −U0 (1 − ϵ2 ) = 2E
r t r02
(Note: E = U0 (1 − ϵ2 )/2.)
Thus, one has ⟨T ⟩t = E − ⟨U ⟩t = −E. Thus, the ratio of the time average of T and U is
−1/2. This is called virial theorem.

Q4

• (a) A circular orbit exists at the value of r for which the derivative of the effective potential
(which is the negative of the effective force) is zero. This is simply the statement that the
right-hand side of the equation
L2
mr̈ = − V ′ (r) (12)
mr3
is zero. , so that r̈ = 0.
Since V ′ (r) = βkrk−1 , Eq. 12 gives
1/(k+2)
L2 L2

− βkrk−1 = 0 =⇒ r0 = . (13)
mr3 mβk

If k is negative, then β must also be negative if there is to be a real solution for r0 .

Z 2π
1 1 dφ
Let In ≡ . One can show that I1 = 1/(1 − ϵ2 )1/2 , I2 = 1/(1 − ϵ2 )3/2 , I3 = (1 +
2π 0 (1 + ϵ cos φ)n
(ϵ2 /2))/(1 − ϵ2 )5/2 , . . .

4
• (b) The long method of finding the frequency is to set r(t) ≡ r0 + ϵ(t), where ϵ represents
the small deviation from the circular orbit, and to then plug this expression for r into Eq.
12. The result (after making some approximations) is a harmonic-oscillator equation of
the form ϵ̈ = −ωr2 ϵ.
Alternate method: By introducing the effective potential, we have reduced the problem
to a one dimensional problem in the variable r. We know that the frequency of small
oscillations is, r
V ′′ (x0 )
ω= ,
m
where x0 is the equilibrium point. That is, to find the frequency of small oscillations, we
just need to calculate the second derivative of the potential. For the problem at hand,
we must use the effective potential, because that is what determines the motion of the
variable r. We therefore have r
′′ (r )
Veff 0
ωr = . (14)
m
If you work through the r ≡ r0 + ϵ method described above, you will find that you are
basically calculating the second derivative of Veff , but in a rather cumbersome way.
Using the form of the effective potential, we have

3L2 3L2
 
′′ 1
Veff (r0 ) = 4 + βk(k − 1)r0k−2 = 4 k+2
+ βk(k − 1)r0 .
mr0 r0 m

Using the r0 from Eq. 13, this simplifies to

r
′′ (r )
√
′′ L2 (k + 2) Veff 0 L k+2
Veff (r0 ) = =⇒ ωr = = . (15)
mr04 m mr02

We could get rid of the r0 here by using Eq. 13, but this form of ωr will be more useful
in part (c).
Note that we must have k > −2 for ωr to be real. If k < −2, then Veff ′′ (r ) < 0, which
0
means that we have a local maximum of Veff , instead of a local minimum. In other words,
the circular orbit is unstable. Small perturbations grow, instead of oscillating around zero.

• (c) Since L = mr02 θ̇ for the circular orbit, we have ωθ ≡ θ̇ = L/ mr02 . Combining this

with Eq. 15, we find

ωr √
= k + 2.
ωθ
A few values of k that yield rational values for this ratio are (the plots of the orbits are
shown in Figure):

5
 – k = −1 =⇒ ωr /ωθ = 1 : This is the gravitational potential. The variable r makes
one oscillation for each complete revolution of the (nearly) circular orbit.
– k = 2 =⇒ ωr /ωθ = 2 : This is the spring potential. The variable r makes two
oscillations for each complete revolution.
– k = 7 =⇒ ωr /ωθ = 3 : The variable r makes three oscillations for each complete
revolution.
– k = −7/4 =⇒ ωr /ωθ = 1/2 : The variable r makes half of an oscillation for each
complete revolution. So we need to have two revolutions to get back to the same
value of r.
There is an infinite number of k values that yield closed orbits. But note that this
statement applies only to orbits that are nearly circular. Also, the ”closed” nature of the
orbits is only approximate, because it is based on Eq. 14 which is an approximate result
based on small oscillations. The only k values that lead to exactly closed orbits for any
initial conditions are k = −1 (gravity) and k = 2 (spring), and in both cases the orbits
are ellipses. This result is known as Bertrand’s theorem.

Q5

• (a) Note that the satellite transfers from one orbit to another by firing its rockets in the
tangential direction, forward or backward, when it is at the perigee of its initial orbit. By
our choice of axis, we can choose that this occurs at ϕ = 0 (and with the initial phase
factor δ set to be zero). Also, as the rockets are fired in a tangential direction, the velocity
just after firing is still in the same direction, which is perpendicular to the radius from
Earth to the satellite. Therefore, the position at which the rockets are fired is also the
perigee for the final orbit (with again the initial phase factor δ chosen to be zero).
At perigee, since there is no radial component to the velocity, the angular momentum is
L = mrv. The radius of the satellite’s orbit does not change at the instant when the
rockets are fired, only its velocity changes, eventually leading to a different orbit than its
original one.

6
 Recall that, the orbital equation is given by

1 GM m2
= (1 + ϵ cos θ),
r(θ) L2

where, evidently, M denotes the mass of the Earth, while m represents the mass of the
satellite.
At θ = 0, from the initial circular and the intermediate elliptical orbits, we have

L22 L21
R1 = = ,
2GM m2 (1 + ϵ) 2GM m2

where ϵ is the eccentricity of the intermediate elliptical orbit. The above result implies
that
L22 = L21 (1 + ϵ)
and as we have set L2 = λL1 , we obtain that

λ2 = (1 + ϵ).

Similarly, at θ = π, from the intermediate elliptical and the final circular orbits, we have

L22 L23
2R1 = =
2GM m2 (1 − ϵ) 2GM m2

so that
L22 = L23 (1 − ϵ)

Since L3 = λ′ L2 , we obtain that

2
λ′ = (1 − ϵ)−1 .

Now, note that, in the case of the elliptical orbit,

rmax = 2R1 = a(1 + ϵ), rmin = R1 = a(1 − ϵ),

and, hence,
1+ϵ
= 2,
1−ϵ
so that ϵ = 1/3. Therefore, λ = 4/3 and λ′ = 3/2.
p p

• (b) Note that the energy of the orbit is related to the eccentricity ϵ as follows:

2
GM m2 m
1 − ϵ2


E=−
2L2
In the case of circular orbits of radius R, since there is no radial velocity involved, the
velocity remains the same throughout the orbit. Thus,
1 α
E = mv 2 −
2 R

7
 and
GM m
v2 = .
R
Therefore, in the case of the initial and final circular orbits in our we have
GM GM
v12 = , v32 =
R1 2R1
which implies that
v1
v3 = √
2
In other words, the velocity of the satellite actually decreases, as it goes from the initial
circular orbit to the final circular orbit. However, the loss in kinetic energy is suitably
compensated by an increase in the potential energy (due to the increased radius) leading
to an increase in the total energy of the satellite. Note that E3 = −E1 /2.

Q6

dr
r = keαθ =⇒ = kαeαθ = αr
dθ
By substituting in the orbital equation:
 2
1 dr 2mE 1 2mV (r)
= − 2−
r2 dθ L 2 r L2
L2 L2 1 dr 2
 
V (r) = E − −
2mr2 2m r2 dθ
L2
=E− (1 + α2 )
2mr2
dV (r) L2
F (r) = − = − 3 (1 + α2 )
dr mr
Now, we are focusing on computing θ(t) and r(t):

L dθ L
θ̇ = 2
=⇒ =
mr dt mk e2αθ
2

e2αθ
mk 2 e2αθ dθ = Ldt =⇒ mk 2 = Lt + C1
2α
where, C1 is arbitary integration constant. That gives how θ changes with time t.
 
1 2αL
θ(t) = ln t + C1
2α mk 2

With that, we can find how the radial coordinate change with time t.
r
αθ(t) 2αL
r(t) = ke = t + k 2 C1
m