Chapter 1

Introduction
to Mechanics
Section 1

Newton’s Laws

NEWTON’S THREE LAWS: dp

=F
dt

For a point particle: p = mv = m r·

First law (Law of inertia)[C]: It is always possible to find a
coordinate system with respect to which isolated bodies move For a rigid body: linear momentum of the CM.
  uniformly. This coordinate system is called an inertial reference
d 2r
f r a m e .
 =F
dt 2

Third law (C): If a given body A exerts a force on another body B,

The usual statement “A body remains in its state of rest or in
then body B also exerts a force on body A with equal magnitude
uniform linear motion as long as no external force acts on it to
but opposite in direction.
change” is not suitable. In an accelerating frame, a force-free
body will appear to accelerate. Not always valid. If the Sun moved by a distance, then the action
and reaction with Earth will not take on the new values until the
Inertial reference frame: In which the law of inertial is valid. Any
signal reaches the Earth. This is because the Newtonian
reference frames moving with a constant velocity wrt stationary
dynamics assumes instantaneous propagation of signal (action
frame is a inertial reference frame.
at a distance).
Examples: free space, a freely falling elevator.
Main points: Second order differential equation.
Second law (C): The change in the linear momentum of a body
Other assumptions:
is proportional to the external force acting on the body and is in
the direction of the force.

2
Absolute space: Motion takes pace in the background of
absolute space. The space is same for all observers.
Consequence: The length of a rod is the same for all observers.
Absolute Time: Evolution of physical system in “time”.
moves uniformly with equal rate for all observers.
Consequence: Time interval between two events is the same
for all observers.
Mass conservation
Infinite speed
(1) Solve for the position and velocity of a damped linear oscillator
whose mass m=1, the viscous coefficient is γ. The initial
condition is [x(0),v(0)].
(2) Do the same for a forced undamped oscillator with initial
Time
condition is [x(0),v(0)]. Solve for both cases: ω0 ≠ ωf and ω0 = ωf
(3) A spherical ball of mass m falls under gravity in a viscous
fluid. Find the position and velocity of the ball as a function of
time. Assume that the mass starts at rest from a height h
above the ground. Apply the above solution to a raindrop
whose radius is 1 mm. Assume the dynamic viscosity of air to
be 10−5 kg/(m s).

Examples:

Oscillator: Equation

m x·· + 2γ x· + ω02 x = F0 cos ωf t

(1) Linear frictionless oscillator with the initial condition is

v(0)
[x(0),v(0)]: x(t) = x(0)cos ωt + sin ωt
ω

(2) Solve x·· = x: x(t) = x(0)cosh t + v(0)sinh t

Exercises:

3
Section 2

Energy

∮
Work energy theorem: Conservative force fields: Fields for which F ⋅ dr = 0 for any
The work done on a particle by a force F as it moves from point
Q

∫P
A to point B: closed path. For such fields, F ⋅ dr is only function of the end
B

∫A
WAB = F ⋅ dr points, but remain independent of paths. Therefore, we choose
P as the reference point and define potential energy U(r) as

which can be rewritten as r

∫P
U(r) = − F(r′) ⋅ dr′.

[2 ]
B B
dv dr 1 2
∫A dt dt
[1] WAB = m ⋅ dt = mv = TB − TA
A Therefore,

B B A
where TA, TB are the kinetic energy of the particle at the points A
∫A ∫P ∫P
[2] F ⋅ dr = F ⋅ dr − F ⋅ dr = − UB + UA
and B respectively.

The above statement is work energy theorem: The difference in Eqs. [1,2] yield
the kinetic energy of a particle between two points is equal to the
net work done on the particle by the external forces during the TA + UA = TB + UB,

transit.
The sum of kinetic and potential energies is the total energy, and
Conservative force fields and Potentials: the above statement is the conservation of total energy of a
mechanical systems that are conservative.

4
Example: these charges. (a) Find the position(s) of equilibrium point(s).
(b)  Compute the potential near the equilibrium point. (c) What
(1) Spring force F = − k x. Therefore
x
is the nature of the equilibrium point? Describe the motion of
1 1
∫x
U(x) = − (−k x)d x = k x 2 − k x02 the test particle near this point. (d) Does your analysis depend
0
2 2
on the sign of the test charge?

Hence, the potential energy is U(x) = k x 2 /2 apart from a constant, (3) What is the potential of an electric dipole? Compute the
and the total energy electric field induced by it. Sketch the potential.

1 ·2 1 2 (4) Philip Morse modelled the potential energy of a diatomic

E= mx + kx
2 2
molecule using a potential energy function:

(2) For gravitational force F = − GMm /r 2 r̂ , U(r) = De{1 − exp[−a(r − re)]}2 where r is the distance between
the atoms of the dipolar molecule re is the equilibrium bon

( rB rA )
B B
1 1
∫A ∫A
U(B) − U(A) = − F ⋅ dr = − Fr dr = − GMm − distance and De is the depth of the potential well. (a) Sketch
the potential. (b) Find the minima of the potential. (c) Expand
the potential near its minimum and compute the frequency of
By choosing rA = ∞, we deduce
oscillation.
GMm
U(r) = −
r

Exercises:
(1) A charged particle +Q is fixed at the origin. A test charged
particle +q of mass m is fired head-on towards the charged
particle +Q with velocity v∞ from ∞. Compute the shortest
distance of approach between the particles.

(2) Four immovable charges of magnitude +Q are placed at the

vertices of a square. A test charge +q is moving in the field of

5
Section 3

System of Particles and Conservation Laws

fa = fa,ext + fa,int = fa,ext + fa,b

1. Linear Momentum ∑

b

Centre of mass: where fa,b is the force on the a-th particle due to the b-th particle.

∑ mara fa = fa,ext + fa,b

∑ ∑ ∑∑
Ma CM =
RCM =
∑ ma a a a b

a: Particle label From Newton’s third law:

CM reference frames: Frame in which RCM = 0. fa,b + fb,a = 0

Linear momentum of a system: Therefore,

P= mava = ma(VCM + v′a) = M VCM = PCM dP

∑ ∑
f = Fext.
∑ a,ext
=
dt a
Equation of motion:
If Fext = 0, then the total linear momentum of the system P is a
· d VCM
P=M mav· a = f
∑ ∑ a
= constant. This is the statement of the conservation of linear
dt a
momentum.

fa is the net force on the a-th particle. Example: Two masses (m) connected by a spring. Example Ch
13

6
 If the torque, N, on the particle is zero, then the angular
momentum L of the particle is a constant.

This law is called conservation of angular momentum.

3. Angular Momentum of a System of

2. Angular Momentum of a Single Particle
Particles

y
The Angular momentum of a system of particles is a sum of

P .
m
V angular momentum of individual particles. The angular
momentum of such system about a reference point is
r

.
O
L=
∑
a
ra × pa,

where ra denotes the position vector of the ath particle from the
x
reference point, and pa its linear momentum. Using the CM
We define the angular momentum of the particle about a coordinates, r = R CM + r′a and v = VCM + v′a:
reference point O, where OP = r, as
L= ra × pa
∑
a
L = r × p,
ma(R CM + r′a) × (VCM + v′a)
∑
=
where p = m r· is the linear momentum of the particle measured in a
the reference frame. The above formulas is valid in any reference
= R CM × PCM + r′a × p′a
∑

frame, inertial or noninertial. Now.
Therefore,
dL dp dr dp
=r× + ×p=r× =r×F=N
dt dt dt dt L = R CM × PCM + LCM = Lorbit + LCM,

where
where N is the torque on the particle about the reference point O.
7
 LCM = r′a × p′a
∑

is the angular momentum of the body about the CM, The time derivative of LP is

d LP ··
ma(ra − RP ) × (··ra − RP ).
∑
=
dt
4. Equation of Motion Using Angular a

Momentum Using ma··ra = fa is the net force acting on the ath particle, we
obtain
We assume that x y coordinate system of the figure is an inertial
d LP ··
(ra − RP ) × (fa − maRP ).
∑
reference frame, and the reference point O belongs to it. The =
dt a
angular momentum of the body about O is
··
(ra − RP ) × (−maRP ) can interpreted as the torque on
∑
LO = mara × pa. The term
∑

a
a
··
the rigid body due to the pseudo force −maRP. The force fa is a
We also compute angular momentum LP about another point P sum of external and internal forces, i.e.,
shown in the following figure:
fa = fa,ext + fa,b
∑

·
LP = mar′a × p′a = ma(ra − RP ) × (r· a − RP ),
∑ ∑
b
a a

where fa,ext is the external force on the ath particle, while fa,b is the
internal force on the ath particle due to the bth particle as shown
y
in Fig. 15.4. Therefore,

d LP ·· 1
(ra − RP ) × (fa,ext − maRP ) + (r′ × f + r′b × fb,a)
∑ ∑ ∑ a a,b
ra
r'a
=
dt a
2 a b
P

RP
O x

8
where r′a = ra − RP. According to Newton's third law, fa,b = − fb,a. d LP ··
(ra − RP ) × fa,ext − M(R CM − RP ) × RP. or
∑
=
dt
Therefore, a

d LP ··
d LP ·· 1 = Next − M(R CM − RP ) × RP,
(ra − RP ) × (fa,ext − maRP ) + (r′ − r′b) × fa,b dt
∑ ∑∑ a
[3.1] =
dt a
2 a b

where Next is the external torque acting on the rigid body about
the point P, and the second term of the above equation is the
··
torque about the point P induced by the pseudo force −M R CM.
Note that the second term vanishes for one of the following
cases:

(1) RP = R CM, i.e., CM is the reference point P. This result is due

to the fact that the torque induced by a pseudo force wrt the CM
is zero.
··
(2) RP = 0, or the reference point P is not accelerating. Naturally,
the pseudo force is zero for this case.
Figure 15.4 Internal forces between particles a and b.
··
(3) (R CM − RP ) | | RP, a case rarely used in solving problems on
rigid bodies.
Now we apply the strong form of Newton's third law, according to
which the action and reaction forces are not only equal and
opposite, but they are also along the line joining the two particles. Hence, using the CM or any non-accelerating point as a
Under this condition, the total internal torque (the second term of reference, we obtain
Eq. [3.1] vanishes. Therefore,
dL
= Next . (15.3.10)
dt

Next = 0, then L is a constant. That is,

9
If the net torque on a system of particles is zero, then the total
angular momentum of the system is conserved.
Rigid body
Solving For Motion Of A Rigid Body
The equation of motion for the centre of mass of the rigid body in
an inertial frame is
d PCM
= Fext (1)
dt
where PCM is the linear momentum of the CM, and Fext is the
external force acting on the rigid body.
The equation of motion for the angles of the rigid body is
dL
= Next, (2)
dt
where L is the angular momentum of the rigid body about a
reference point, and Next is the external torque acting on the rigid
body about the same reference point.
It is important to note that computation of L and Next requires a
reference point. The reference point can be either the centre of
mass or a non-accelerating point of the rigid body.
Rotation About a Single Axis
Example 1: A cylinder rolls down an incline without slipping.
Describe the motion of the cylinder.
Solution: A cylinder is rolling down the inclined plane without
slipping. The forces acting on the rigid body are gravity and
friction. The frictional force opposes the tendency to slip, hence
the frictional force is along − ̂x.
For rolling without slipping, it is convenient to use the clockwise
rotation as positive. We denote the velocity of the CM by v, and
the angular velocity of the cylinder along − ẑ by Ω. Note that Ω is
the same for all the points on the cylinder. Let a be the
acceleration of the cylinder, and α its angular acceleration. We
take the CM of the cylinder to be the reference point. We can do
so even though the CM is accelerating because the pseudo force
passes through the CM, and it causes no additional torque.
Therefore, the equation of motion of the cylinder is
ma = mg sin θ − f, (1)
Iα = f R. (2)
10
The constraint that the cylinder rolls down without slipping yields Ia k
f = = mg sin θ . (6)
R2 1+k
v = ωR (3)
where k = I /(m R 2). For the cylinder I = MR 2 /2, hence k = 1/2 and
By taking a time derivative of Eq. (3) we obtain a = (2g/3)sin θ.

a = αR (4) 1 2
XCM = at , (7)
2

1 2
ϕ= αt , (8)
2
z
A block slides with an acceleration a = g sin θ.

Note that f ≤ μN, where μ is the coefficient of friction, and N is the

normal force. If the frictional force cannot support rolling, then
the ball will slip and roll, and we cannot apply the condition
a = αR.

For rolling without slipping, the frictional force does not do any

∫
work ( f . dr = 0) because the contact point does not slip (dr = 0).

Since the frictional force does not dissipate any energy, we can
Now we have three equations (1,2,4), and three unknowns a , α, also solve the above problem by applying conservation of total
and f. Solving these equations yields energy:
g sin θ 1 1
a= (5) mgx sin θ = 2
mVCM + IΩ2, (9)
1+k 2 2

and

11
with VCM = ΩR. The LHS is the loss in potential energy, and the a = g(sin θ − μ cos θ)
RHS is the gain in KE. The time derivative of the above equation
α = k(g/R)cos θ.
yields the same acceleration as Eq. (5).

In the reference frame x yz, acceleration of any point on the
cylinder is aP = a − αr = αR − αr, where r is the distance of the Example 2: The cylinder of Example 16.1 is pushed up the
point P from the axis of the cylinder. Clearly the bottom-most incline in such a way that the cylinder rolls up the incline without
point (the contact point) of the cylinder has aP = 0. As a slipping. Describe the motion of the cylinder.
consequence, we can choose the bottom-most point of the
Solution: When the cylinder is rolled up, the frictional force is still
cylinder as the reference point since it is a non-accelerating point
upward in order to oppose the tendency of the cylinder to slip
of the rigid body. The equation for motion for the cylinder about
down the incline. Hence the equations of motion are exactly the
the contact point is
same as Eq. (1-4) of the above example, As a result, the cylinder
I′α = (mg sin θ)R (10) has the same acceleration as Eq. (5); the cylinder slows down
while ascending the incline. Note that the motion for this case is
2
with I′ = I + MR . Hence
the time reversed motion of Exercise 16.1. It is interesting to note
g sin θ that the time-reversal symmetry is respected in this problem even
α= , (11)
R(1 + k) in the presence of frictional force. This is because the rolling
frictional force does not dissipate energy, hence it does not break
g sin θ
a= , (12) the time-reversal symmetry.
1+k

which is consistent with Eq. (5).

We remark that the non-accelerating point of the cylinder at

different times are different. Exercises
(1) A hoop of mass M and radius R is lying horizontally on a
When the ball slips, f = μg cos θ. Hence
smooth surface. A bug of mass m moves slowly on the hoop.

12
Compute the trajectories of the bug and the centre of mass of the (6) A charged ball of mass m, radius R and charge q is resting on
hoop. a horizontal slab. We apply an electric field E on the ball. The
direction of the electric field is parallel to the slab. Describe the
(2) Two particles rotate around each other in a circular orbit under
motion of the ball.

the influence of gravitational pull of each other. The time period of
orbit is T. At a given time, the particles are stopped suddenly, and
then let go towards each other. After what interval of time, will the
particles meet each other?

(3)The spring mass–system shown in Fig. (a) is pulled with force F.

Compute the position of both the blocks as a function of time.

(4) A spring–mass system shown in Fig. (b) is resting on a

horizontal surface. What force should be applied to the upper
plate so that the lower one gets lifted after the pressure is
removed.

(5) Two masses m1 and m2 are hanging on the two sides of a

pulley that has moment of inertia I. Assume the string to be
massless and inextensible. (a) Compute the acceleration of the
masses. (b) Compute the tension of the string.

13
Chapter 2

Constraints,
Lagrange’s
equations
Section 1

Constraints

The position of the particle or system follows certain rules due to 1. Principle of Virtual work
constraints:

Holonomic constraint: f (r1 . r2, . . . rn, t) = 0 System under equilibrium:

Constraints that are not expressible as the above are called The total force on each particle fa = 0
nonholonomic.
Virtual displacement δra : Arbitrary infinitesimal change in the
Examples: position of the a-th particle keeping the constraints. This is called
virtual displacement.
Rigid body: ra,b = constant
Therefore, the sum of virtual work is zero:
Rolling without slipping: VCM = ωRCM
fa ⋅ δra = 0
∑

particle moving on a circle. a

Generalized coordinates: qi with i = 3N − C, where C is the

Note that fa = fa,ext + fa,int.
number of constraints.
We choose δra such that
ra = ra(qi)
fa,int ⋅ δra = 0,
∑

Note that qi, q· i are independent variables. a

15
then The internal forces between the molecules of the plank does not
do any work under displacement δθ.
fa,ext ⋅ δra = 0.
∑

a The normal forces do no work.

This is the principle of virtual work. Work done by the frictional force:

Example: A plank resting agains at a wall. The bottom surface is W1 = fd x = f lδ(sin θ) = f l cos θδθ
frictional with the friction force = f.
Work done by mg:

W2 = mgδy = mg(l /2)δ(cos θ) = − mg(l /2)sin θδθ

y
N2 Using principle of virtual work: W1 + W2 = 0.

2f
Therefore tan θ = .
mg

θ 2. D’Alembert’s Principle
N1
y For dynamics
mg

fa = p· a
x f x
Hence

(fa − p· a) ⋅ δra = 0.
∑

a
Virtual displacement: δθ.

16
Again choose δra such that the virtual work done by the internal
dt [ ∂q· j ] ∂qj
d ∂T ∂T
m v· a,i δra,i =
∑ ∑
[1.2] − δqj
forces is zero. Hence
a,i j

(fa,ext − p· a) ⋅ δra = 0
∑

a where

Now some algebra: 1 2

∑ 2 a,i
T= mv

∂U ∂ra,i ∂U
∑ ∑ ∂ra,ext,i ∂qj ∑ ∂qj j
[1.1] fa,ext,i δra,i = − δqj = − δq is the kinetic energy of the system. The displacement δqj is
a,i a,i, j j
arbitrary. Therefore, using Eqs. [1.1, 1.2] we obtain
∂ra,i, j
mav· a,i δra,i = mav· a,i
dt [ ∂qj ] ∂qj
d ∂T ∂T ∂U
∑ ∑
δqj
∂qj − =
a,i a,i, j ∂qj

∂ra,i d ∂ra,i Typically, ∂U/∂q· j = 0. Then

dt [ ∂qj ]
d
∑ ∑
= ma va,i − mava,i δqj
j
dt a,i
∂qj

dt [ ∂qj ] ∂qj
d ∂L ∂L
− =0
dra,i ∂ra,i ∂ra,i
Note: r·a,i = q· j +
∑ ∂qj
=
dt ∂t
j where L=T-U is the Lagrangian of the system.

∂r·a,i ∂ra,i Advantages of the Lagrangian formalism

Hence, = ·
∂qj ∂ qj
No need to worry about constraint forces, simpler

d ∂ra,i ∂ 2ra,i ∂ 2ra,i ∂r·a,i

dt [ ∂qj ] ∑
= · Analytical, For example, Mécanique analytique by Lagrange does

∂q ∂ · qk + ∂q ∂t = ∂q
q
k j k j j not have a single figure.

substitution of which in the above yields

17
Examples: ·· 3
θ = sin θ
2
(1) a free particle
(3) Construct Lagrangian for a cylinder rolling down an incline.
(2) a particle in 2D in radial polar coordinate (r, ϕ)

1 ·2 ·
L= m(r + r 2ϕ2) − U(r)
2 Exercises:

Hence the equations of motion are (1) A particle is sliding on a uniformly rotating wire. Write down

· the Lagrangian of the particle. Derive its equation of motion.

m ··r = mr ϕ2 − ∂rU
(2) Verify D’Alembert’s principle for a block of mass M sliding
d ·
m (r 2ϕ) = − ∂ϕU or down a wedge with an angle of θ.

dt

·· · 1
m(r ϕ + 2r· ϕ) = − ∂ϕU
r

(2) Consider the plank discussed before. Let us assume the

ground surface to be frictionless.

Generalized coordinate = θ

1 ·2 ·2 1 1 · 1 ·
The KE = T = m(x + y ) + ml 2θ 2 = ml 2θ 2
2 2 12 6

1
The potential energy U = mgy = mgl sin θ
2

The Lagrangian L = T − U

The equation of motion yields

18
Chapter 3

Principle of
Least Action
Section 1

Variational Calculus

Function of functions We will derive an equation for the required function q(t) that
extremizes the action. We will compute action for another
· t)
L = L(q, q,
function
L is a function of q(t), which itself is a function of t.
q(t, α) = q(t,0) + αη
Objective:
where αη is the deviation from the required function. Here α is a
Extremize action · t). The change in action due to the above is
number and η(q, q,

∫t [ ∂q ]
t2 t2 t2
∂L ∂L
∫t ∫t
S= · t)dt
L(q, q, δS = · t)]dt =
[δL(q, q, αη + · α η· + HOT dt
1 1
∂q
1

with the ends fixed at (t1, q1) and (t2, q2). where HOT stands for the higher order terms. For extremization,
we take the limit α → 0 (ignore HOT). An integration by parts

q2 yields

∫t [ ∂q· ] [ ∂q· ] ∫t [ dt ( ∂q· )]

t2 t2t
∂L · ∂L 2
d ∂L
α η dt = αη − αηdt
q1 1 t 1 1

The variation of q at the ends must vanish, that is η=0 at the

t1 t2
ends. Hence, the boundary term vanishes. Therefore,

20
 ∫t [ ∂q dt ( ∂q· )]
t2 Therefore,
∂L d ∂L
δS = − αηdt

dt ( ∂q· dt ) ∂q 2
1
d ∂ df ∂ 2f · ∂ 2f
= q+
∂q∂t
Since η is arbitrary,
and
dt ( ∂q· ) ∂q
d ∂L ∂L
=

∂q ( dt ) ∂q 2
∂ df ∂ 2f · ∂ 2f
= q+
∂q∂t
Note: The following Lagrangian
Hence the additional terms cancel each other. Q.E.D.
L′(q, q, · t) + d f (q, t)
· t) = L(q, q,
dt
NOTE: On many occasions, the dependent variable is x rather
yields the same equation of motion. than time. On those cases, we replace q· by q′.

t2 t2 t2
df For Multi Variables
∫t ∫t ∫t dt
Proof (a): S′ = · t)dt =
L′(q, q, · t)dt +
L(q, q, dt
1 1 1

= S + f (q2, t2) − f (q1, t1) Here the generalized variables are qi’s. Hence

Hence, δS′ = δS. QED L = L(qi, q·i, t)

Proof (b): For this case, qi(t, α) = qi(t,0) + αηi. Hence Eq. () becomes

d ∂f · ∂f
f = q+
dt ∂q ∂t t2 t2

∑ { ∂qi ∂qi }
∂L ∂L
∫t ∫t
δS = [δL(qi, q·i, t)]dt = αηi + · α η·i + HOT dt
Hence 1 1 i

∂q· ( dt ) ∂q
∂ df ∂f
=

21
 t2 Examples:
∂qi dt ( ∂q·i )
∂L d ∂L
∫t ∑
δS = − αηi dt
1 i (1) Minimize the distance between two points in 3D:

(dx) (dx)
2 2 2
Since it is valid for arbitrary ηi, we obtain dy dz
∫1
D= 1+ + dx

dt ( ∂q· i ) ∂qi
d ∂L ∂L
=
x: independent variable

y,z: Dependent variables

Beltrami Identity

(dx) (dx)
2 2
dy dz
L= 1+ +
If L is not an explicit function of time t, then

∂L ∂L ∂L
q·i · = const
∑ ∂qi
L− Since = = 0,
i
∂y ∂z

∂L ∂L
Proof: = C1 and = C2
∂y′ ∂z′

dt ( ∂qi )
dL d ∂L
dt ∑
LHS = − qi · Therefore,
i

y ′2 2 z ′2
∂L ∂L · ∂L ∂L d ∂L = C1 and = C22
qi + · q··i − q··i · − q· i 1+y +z
′2 ′2 1+y +z
′2 ′2
∂t ∑
= +
i
∂qi ∂qi ∂qi dt ∂q· i
Hence, y ′2 + z ′2 = constant. Therefore, y′ and z′ are constants.
=0 Hence, the particle moves on a straight line.

Here we have used the equation of motion. (2) Minimize the time of descent between two points in a
gravitational field:

22
 Cx
y′ = .
1 − Cx

A substitution of Cx = sin2 θ yields y′ = tan θ. Therefore,

A y
dx sin 2θ dy 1 − cos 2θ
= and = ,
dθ C dθ C

whose parametric solution with initial condition (x=0, y=0) is

B
x = A(1 − cos ϕ) and y = A(ϕ − sin ϕ)

where ϕ = 2θ. The above is an equation of cycloid.

x
y
B
ds B 1 + y ′2
∫A v ∫A
T= = dx
2gx

Hence, the Lagrangian is

x

1 + y ′2
L=
x

Since ∂L /∂y = 0, ∂L /∂y′ = C, a constant, which yields Note: We chose the vertical axis as x, so that L is independent of
y. It helps simplify the solution. If we interchanged the axes, the
y ′2 time will be
= C.
x(1 + y ′2)

dsB B 1 + y ′2
∫A v ∫A
Hence T= = d x.
2gy

23
Hence the Lagrangian will be equation is of higher order. Also, this method is not convenient
when we have more constraints and variables.
′2
1+y
L=
y Lagrange proposed an alternative. He suggests that the variables
x, y and a new variable λ be made independent. Idea is to look for
Using Beltrami identity, we obtain a contour of f (x, y) that is tangent to the g(x, y) = c curve. See
figure below.
∂L 1
L − q· · = =C
∂q y(1 + y ′2)

Therefore,

C−y
y′ = ,
y

whose solution is same as before except the change of axis.

From Wikipedia
Extremization under constraints
The intersection point is the desired extremum point. Here
Detour to Lagrange multiplier
∇f (x, y) = − λ ∇g(x, y) and g(x, y) = c.
We illustrate using an example. Suppose we want to Extremize
f (x, y) under the constraint that g(x, y) = c. The constraint would The above equations can be derived by extremizing
make f (x, y) a function of single variable (say x) that can be
maximized using the standard method. However solving for x F(x, y, λ) = f (x, y) + λ[g(x, y) − c]

from the constraint equation could be tricky if the constraint

wrt x, y, λ that yields

24
 fx = λgx We return to the original point, but with a constant perimeter. That
is, x(t1) = x(t2) = x0 and y(t1) = y(t2) = y0. We want to know a
fy = λgy
function that yields maximum area.

g(x, y) = c 1 t2 ·
∫
S= ·
(x y − y x)dt
2 t1
Example:

(1) Find minimum of the function x 2 + y 2 under the constraint that under the constraint that
y − x − 1 = 0. t2

∫t
I= (x· 2 + y· 2)dt
Solution: We minimize the function 1

F = (x 2 + y 2) + λ(y − x − 1) Hence we extremize

By taking derivatives wrt x, y, λ we obtain 1 · · +λ

L= (x y − y x) x· 2 + y· 2
2
2x − λ = 0; 2y + λ = 0; y = x + 1
which yields
whose solution is y = 1/2,x = − 1/2,λ = − 1/2.
λ x·
dt ( 2 x· 2 + y· 2 )
1· d 1
y− − y+ =0
Application to variation calculus 2

Extremize functions under constraints variationally.

λy·
dt ( 2 x· 2 + y· 2 )
1 d 1
− x· − x+ =0
We illustrate using an example. 2

(1) Parametric curve

Using x· = d x /dt, y· = dy/dt we obtain
x = x(t), y = y(t)
λy· λ x·
x+ = C1, y − = C2
x· 2 + y· 2 x· 2 + y· 2

25
which yields (4) Consider the functional

b
(x − C1)2 + (y − C2)2 = λ 2
∫a
S[y] = (P(x)y ′2 + Q(x)y 2)d x
which is an equation of a circle. The parameter λ is determined
by the perimeter of the circle. Find the extrema of the above subject to the condition that

∫a
w(x)y 2 d x = 1
Exercises:

(1) A rope of linear density γ and length l is hanging by two The resulting equation is called Sturm-Liouville problem. Relate
supports that are located horizontally 2a apart. Assuming this equation to the Schrodinger’s equation.
equilibrium position for the rope, compute its equation.

(2) On a sphere, the great arc is defined as the curves that

minimizes the distance travelled between the given two points.
Compute the equation of a great arc.

(3) Analyze the variational problems corresponding to the

following functionals. In each case take y(0) = 0 and y(1) = 1.

∫0
(a) y ′2 d x

∫0
(b) y y′d x

∫0
(c) x y y′d x

26
27
Section 2

Variation Principle in Mechanics

Hamilton’s Principle:
dt ( ∂q· i ) ∂qi
d ∂L ∂L
=
Every mechanical system is characterized by a Lagrangian,
L(q , q· , t) or L(q, q,
i i
· t) in brief, and the motion of he system is such

that the action Law of Inertia

t2
Homogeneity of space: Every mechanical system and its

∫t
S= · t)dt
L(q, q, properties are invariant under a space translation.
1

Homogeneity of time: Every mechanical system and its

computed between the starting point (q1, t1) and (q2, t2) is
minimum.
Here qi represent the generalized coordinates, chosen in such a
way so as to eliminate the constraints. The number of
generalized coordinates is 3N − C, where N is the number of
particles, and C is the number of constraints. For a rigid body
we need 6 generalized coordinates: 3 for the CM, and 3 for the
Euler angles.
Using the derivation of the Section 3.1, we deduce the equation
of motion of the system:
properties are invariant under a time translation.
Isotropy of space: Every mechanical system and its properties
are invariant under a space rotation.
Inertial frame: A reference frame can always be chosen in which
space is homogeneous and isotropic, and time is homogeneous.
Law of inertia holds in such a frame.
Due to homogeneity of space and time, L can only depend
on the velocity of the particle v. Due to the isotropy of space, it
can depend only on a function of v 2:

L = L(v 2) (1)

28
Since L is independent of r, Therefore the additional term is

∂L /∂v = constant, hence v = constant. d

mv ⋅ ϵ = [mr ⋅ ϵ].
dt
This is the law of inertia.
Thus we deduce that the Lagrangian of the free particle is
Galilean invariance: Reference frames moving with constant
1 2
velocities relative to one another are equivalent. We cannot L= mv ,
2
differentiate one from another.
where m is the mass of the particle.
Lagrangian of a free particle
1 2 m (Δr)2
Action S = mv (t2 − t1) = .
2 2 (t2 − t1)
In a reference frame moving with a velocity -ε with respect to the
reference to that of Eq. (1), the velocity of the particle would be For action to be minimum m > 0.

v’ = v + ε. Lagrangian of a system of particles

Therefore, the Lagrangian of the particle in the new frame is
Additive property for noninteracting systems implies that for N
2 2 2
L′ = L[(v + ϵ) ] = L[v + 2v ⋅ ϵ + ϵ ]
free particles,
Assuming ε to be small, 1
m v2
∑ ∑2 a a
L= La =
∂L a
L′(v ′2) = L(v 2) + 2 2v ⋅ ϵ
∂v
Interaction among these particles is introduced using a potential
The additional term can be represented as d f (r, t)/dt only when function U:

∂L 1 1
= m = const. m v 2 − U(r1, r2, . . . , rn)
∑2 a a
∂v 2 2 L=
a

29
Homogeneity of space requires that U is dependent on the μ x·· = − k x
distance between the particles | ri − rj | , but not on their absolute
whose solution is
positions. The equation of motion of the particle is
x(t) = c1 cos ωt + c2 sin ωt
∂U
m v· a = −
∂ra
with ω = k /μ. Clearly, the later coordinates are more
convenient.
This is the Newton’s law.

Sometimes, we consider a system under the influence of an

Example: Two masses m1 and m2 coupled by a spring of a spring
external potential.
constant k. The Lagrangian of the system would be
1 2
1 1 1 L= mv − U(r, t)
L = m1x·1 2 + m2 x·2 2 − k(x2 − x1)2 2
2 2 2

Example: The Lagrangian of a forced oscillator is

that yields the equations of motions as
1 ·2 1 2
m1x··1 = − k(x1 − x2) L= m x − k x + F0 sin ωf t
2 2
m1x··2 = k(x1 − x2).
An important note: Interaction speed either infinite or a finite
We could also write L in terms of the CM and relative coordinate constant for all the reference frame.

x = x2 − x1 as
Cyclic coordinates and conservation
1 · 1 1 If a variable q does not appear explicitly in the Lagrangian (but q·
L = M X 2CM + μ x· 2 − k x 2
2 2 2 appears), then the equation of motion yields

where μ is the reduced mass. The above equation yield ∂L

= constant in time
∂q·
·
XCM = constant and
Hence it is conserved. Here q is called a cyclic variable.

30
 ·· ·· · ·
Example: Ml12ϕ1 + m2l1l2 cos(ϕ1 − ϕ2)ϕ2 − m2l1l2 sin(ϕ1 − ϕ2)ϕ1 ϕ2

1. In the spring mass system, L is independent of XCM, hence · ·

= − Mgl1 sin ϕ1 − m2l1l2 sin(ϕ1 − ϕ2)ϕ1 ϕ2
∂L ·
· = m XCM = constant, which is the linear momentum of the ·· ·· · ·
∂XCM m2l22ϕ2 + m2l1l2 cos(ϕ1 − ϕ2)ϕ1 + m2l1l2 sin(ϕ1 − ϕ2)ϕ1 ϕ2

CM. · ·
= − m2 gl2 sin ϕ2 + m2l1l2 sin(ϕ1 − ϕ2)ϕ1 ϕ2
2. Double pendulum: figure
3. Cone problem: constraint: z = ρ cot α
1 ·2
T1 = m1l1ϕ1 1 ·
2 L= m[ ρ· 2 + z· 2 + ρ 2ϕ2] − mgz
2
1
T2 = m2(x· 22 + y· 22) 1 ·
2 = m[ ρ· 2cosec2α + ρ 2ϕ2] − mgρ cot α
2
where x2 = l1 cos ϕ1 + l2 cos ϕ2 and y2 = l1 sin ϕ1 + l2 sin ϕ2. Therefore,
Since L is not an explicit function of ϕ, ϕ is a cyclic variable.
1 · · · · Hence
T2 = m2[l1ϕ21 + l2ϕ22 + 2l1l2 cos(ϕ1 − ϕ2)ϕ1 ϕ2]
2
∂L 2 ·
· = mρ ϕ = Lz = const
U = − m1gl1 cos ϕ1 − m2 g(l1 cos ϕ1 + l2 cos ϕ2) ∂ϕ

Hence The equation of motion for ρ is

·· 2 ·
m ρcosec α = mρ ϕ2 − mg cot α or
1 · 1 · · ·
L= Ml1ϕ21 + m2l2ϕ22 + m2l1l2 cos(ϕ1 − ϕ2)ϕ1 ϕ2] − Mgl1 cos ϕ1 − m2 gl2 cos ϕ2
2 2
·· 2
Lz2
m ρcosec α = − mg cot α
mρ 3
The equations of motion for the bobs are
We will solve this equation later in Chapter 5.

31
 c. A double pendulum consisting of two bars of lengths l1 and
l2, and masses m1 and m2.
Constrained Mechanical Systems
Examples: d. A vertical spring-mass system (mass m and spring constant
k) whose base is being vibrated vertically with a frequency of
(1) Simple pendulum: We use the constraint that r = l.
ωf.
The Lagrangian of the pendulum is
e. A pendulum whose length is l, and whose bob’s mass is m.
1 2 ·2 1 ·2 Make the base of the bob oscillate (i) horizontally as A cos ωf t
L= mr θ + m r + mgr cos θ + λ(r − l )
2 2
and (ii) vertically as A cos ωf t.
The equations of motion of the system are
2. Suppose a system depends on the generalized coordinate q
·· · q,
·· t). Show that
m θ = − g sin θ and its first two derivatives, that is L = L(q, q,
· the equation of motion for such system is
m ··r = λ − mg sin θ + mr θ 2

dt 2 ( ∂q·· ) dt ( ∂q· ) ∂q
d 2 ∂L d ∂L ∂L
r =l − + =0

Note that λ is the constraint force, which is tension at present.

1 1
Apply this result to the Lagrangian L = − mqq·· − kq 2.
2 2
Exercises:

1. Compute the Lagrangian for the following systems. After that

derive the equations of motion and then solve them.

a. Two pendulums of equal lengths (l) and masses m1 and m2

that are coupled together by a spring of a spring constant k.

b. A bar pendulum.
32
Chapter 4

Symmetries
and
Conservation
Laws
xx
Section 1

Hamiltonian
∂L
Legendre transform pμ =
∂q· μ
A system is described by a function f with x, y as independent
variables.
Legendre’s transform to go from q· μ to pμ as a independent
f = f (x, y) variable. The Hamiltonian

∂f ∂f pμq· μ − L,
∑
df = dx + dy = ud x + vdy H=
∂x ∂y
Hence
Now we want to describe the system in terms of x, v.
∂L ∂L ∂L
( pμδq· μ + q· μδpμ − μ δq μ − · μ δq· μ) −
∑
dH = δt
Since d(v y) = vdy + ydv, ∂q ∂q ∂t

d(v y − f ) = − ud x + ydv ∂L
( pμδq· μ + q· μδpμ − p· μδq μ − pμδq· μ) −
∑
dH = δt
∂t
The new function is g = g(x, v) = v y − f.
∂L
∂g ∂g dH = q· μδpμ − p· μδq μ − δt
= − u; =y ∂t
∂x ∂v
Hence
Hamiltonian
∂H · ∂H ∂H ∂L
Lagrangian L = L(qμ, q· μ, t) q· μ = ; pμ = − μ ; =−
∂pμ ∂q ∂t ∂t

34
 dL ∂L ∂L ∂L ∂L
Note: = μ q· μ + · μ q··μ + δt = p· μq· μ + pμq··μ +
dt ∂q ∂q ∂t ∂t

dH ∂L
=−
dt ∂t

Homogeneity of time implies that for an isolated system, the

potential function is not an explicit function of time. For a pair
charge particles, U depends only the distance between the
charged particles. Hence, for an isolated system, ∂L /∂t = 0.
Hence

dH
= 0,
dt

or the total energy of the system is a constant. This is the

statement of conservation of energy, which is derived using the
homogeneity property of the space.

Exercise

1. Argue that the energy is conserved for spring-mass system and

planetary system.

2. An oscillator of mass m and spring constant k is forced with

F0 cos(ωf t). Write down the Lagrangian for the same. Is the
energy conserved for this system?

35
Section 2

More Conservation laws

∂L d ∂L d ∂L
Conservation of Linear Momentum
∑ ∂ra ∑ dt ∂va
= =
∑ ∂va
= 0.
dt
Homogeneity of space: If each particle of an isolated system is a a a

shifted by a constant distance ε, then the system’s Lagrangian

∂L
Therefore, the total linear momentum P =
∑ ∂va
should not change. Again think of a collection of charged of the system is
a
particle. Interaction potential is function only of the distances
between the particles. Here conserved. It is derived from the homogeneity of space.

∂L On many occasions, a system may be symmetric about some

∑ ∂ra
δL = ⋅ϵ =0
a
translations, but not arbitrary translations. For example for a line
charge aligned along the z axis, the system is invariant under any
Since ε is arbitrary, ∂L
∑ ∂va,z
translation along z. Therefore only Pz = is conserved for
a
∂L
∑ ∂ra
= 0.
such system.
a

If the Lagrangian is not an explicit function of qi, then the

Since
generalized force Fi = ∂L /∂qi = 0, and the generalized momenta
∂L d ∂L
= , ∂L
∂ra dt ∂va pi =
∂q· i
we obtain
is conserved.

36
Conservation of Angular Momentum is a constant. This is related to the isotropy of space.
Isotropy of space: The system or Lagrangian of an isolated
If the system is invariant under rotation about an axis, the angular
system is invariant under the rotation of the whole system by an
momentum about the axis is conserved.
arbitrary angle.
The above three symmetries (homogeneity and isotropy of space,
Let us rotate the system by angle δφ about an axis. The vector
and homogeneity in time) have never been broken. So far, we
δφ is a vector aligned along the axis, and its magnitude is δφ.
have not observed any violation of conservation laws of energy,
Under this operation, each particle is shifted by
linear momentum, and angular momentum.
δra = δϕ × ra and δva = δϕ × va.
Robust conservation

Therefore, the change in the Lagrangian under the above Example:

operation is

∂L ∂L Galilean invariance:
∑ ∂ra
δL = ⋅ δra + ⋅ δva = 0 or
a
∂v a

Vr is the relative velocity between the two inertial frames. For a

p· a ⋅ δϕ × ra + pa ⋅ δϕ × va = 0
∑
δL = set of particles,
a

1 1
mav2a and L′ = ma(va + Vr )2
∑2 ∑2
L=
Therefore,
a a

d
ra × pa = 0.
dt ∑
δϕ ⋅ Hence
a

Since δφ is arbitrary, the total angular momentum of the system
∑ [2 a a ] [ ∑ a a] r
1
about the origin δL = L′ − L = δm (v + Vr )2 = m v ⋅ V = PCM ⋅ Vr
a a
d
L= ra × pa
dt ∑

a
for small Vr.
37
We do not set δL zero like in earlier cases because it will lead to α2
L′ = 2 T − α kU = α k L
PCM = 0, a special case. We set β

dF d The above equation implies that

δL = = [RCM ⋅ Vr ]
dt dt
β = α 1−k/2.
Comparing the above we obtain
For oscillator, k = 2. Hence β = const, therefore, the time period of
d RCM
PCM = , the oscillator is independent of the amplitudes.
dt
For the gravitational potential k = − 1, hence β = α 3/2, which is the
which is an identity.
Kepler’s third law.

Detour to nondimensionalization
Self similarity
The equation of motion remains unchanged if


L → L′ = cL Exercise:
1. Construct conserved quantities for two masses coupled by a
where c is a constant. We use this fact to derive scaling
spring.
transformation, which is stretching a system by a factor α:
2. A particle moves in the field for the following system.
ra → r′a = α ra
Construct the conserved quantities of the charged particle:

Suppose the potential of the function is a homogeneous function Infinite charged plate, infinite line charge, two charges of equal

of the coordinates, then magnitude, infinite half-plane charge, charged helix, charged
torus.
U(r′a) = U(α ra) = α kU(ra)

Now to keep L → L′ = cL, we transform t → t′ = βt. Hence

38
3. Construct conserved quantities for the system of Exercise 1 of
Section 3.2 as well as a system of two masses m1 and m2
coupled by a spring of a spring constant k.

4. What are the conserved quantities for a particle moving on the

inner side of the cone?

39
Section 3

Noether’s theorem

We can derive conservation laws for continuous transformation. ∂L′ ∂L′ dt′ ∂L′ dq ′μ ∂L′ d q· ′μ
= + + ·μ
A general transformation is ∂ϵ ∂t′ dϵ ∂q μ dϵ ∂q dϵ

t′ = t + ϵτ ∂L′ ∂L′ μ ∂L′ d q· ′μ

= τ + μ ζ + ·μ
∂t′ ∂q ∂q dϵ
q ′μ = q μ + ϵζ μ
dq ′μ dq μ + ϵdζ μ
The new action is Using = ,
dt′ dt + ϵdτ

∫ [ dt ]
dt′
∫ ∫
L′dt′ − L dt = L′ − L dt d dq ′μ ·
we obtain = ζμ − q· ′μτ.
·
dϵ dt′

For minimum,
dϵ ( dt′ ) dϵ
d dt′ d · = τ·
Also, = (1 + ϵ τ)

dϵ [ dt ]
d dt′ dF
L′ −L =
ϵ=0
dt
Therefore

or ∂L · · = dF or
L τ· + τ + p· μζ μ + pμ(ζμ − q· ′μτ)
∂t dt

dϵ [ dt ]ϵ=0 [ dϵ ]ϵ=0
d dt′ dL′ dF
L + =
dt ∂H · dF
−Hτ· − τ + p· μζ μ + pμζμ = or
∂t dt
Note:
pμζ μ − Hτ − F = const.

40
This is the Noether’s theorem. Therefore, the conserved quantity with F = 0 is

Examples: pζ − H = constant

(1): By choosing τ = 1,ζ μ = 0,F = 0, we obtain H = constant. ∂L

where p = · = m x· exp(bt /m).
∂x
(2): By choosing τ = 0,ζ μ = 1,F = 0, we obtain pμ = constant.

( 2m )
bx
Hence pζ = m x· − exp(bt /m) = − bx x· exp(bt /m).
(3) Choose τ = 0,F = 0, x′ = x − ϵy, y′ = ϵx + y that yields Mz =
constant.

(2 )
1 ·2 1 2
Also H = p x· − L = m x + k x exp(bt /m)
(4) Galilean invariance: τ = 0,F = RCM ⋅ Vr, ζ μ = Vrμt. It yields a 2

conserved quantity as
Therefore the conserved quantity is
PCM ⋅ Vr t − RCM ⋅ Vr = const
(2 )
1 ·2 1 2 1 ·
m x + k x + bx x exp(bt /m)
2 2
which is identically zero since RCM = PCM t.

(5) Damped linear oscillator:

(2 )
1 ·2 1 2
L= m x − k x exp(bt /m)
2

The above Lagrangian is invariant under the transform t′ = t + ϵτ

and x′ = x + ϵζ with τ = 1 and ζ = − bx /2m. Note that

· ·
( 2m )
·x′ = x + ϵ ζ = x· 1 − ϵ b
1 + ϵ τ·

Substitution of the above implies the invariance of L under the

above transformation.
41
Section 4

Discrete Symmetries

Time Reversal Symmetry:

U(−t) = U(t),

hence the Lagrangian is symmetric under time reversal. (b) (a)

We cannot distinguish between the forward and time-reversed Can’t figure out who is the real gun man!
dynamics.
(b) (a)

Parity or Mirror Symmetry

r → − r = [(x, y, z) → (x, − y, z)] + [(x, − y, z) → (−x, − y, − z)]
g g
Parity = mirror (about xz plane) + rotation about the y axis by π.

symmetry of mirror reflection.

Can’t figure out the real projectile motion

42
Proper and pseudo vectors A cross product of two pseudo vectors is a proper vectors.
When a mirror is placed as an xz plane, the symmetry of mir- Examples of pseudo vectors
ror reflection is mathematically expressed as
(1) Angular momentum L = r × p since the linear momentum
z′ = z, y′ = − y, x′ = x; t′ = t p = mv is a proper vector. Also, spin angular momentum of a
Hence, body.

v′x = vx, v′y = − vy, v′z = vz (2) Angular velocity Ω, which is defined using v = Ω × r.

(3) Magnetic field B since

a′x = ax, a′y = − ay, a′z = az
μ0 Id l × r
4π ∫ r 3
A vector that transforms as above is called proper vector or B=
real vector.

A cross product of two proper vectors transforms very differ-

ently under mirror transformation. Let us consider two proper vec-
tors A and B. A cross product of these two vectors is

C=A×B
= (Ay Bz − Az By) ̂x + (Az Bx − Ax Bz ) ŷ + (Ax By − Ay Bx ) ẑ B B

Since A and B are proper vectors, the rules of mirror reflec- F F

tions of yield

C′x = − Cx, C′y = Cy, C′z = − Cz

V (b) (a) V
A vector that transforms like C is called pseudo vector or axial
vector. Thus, Can’t figure out real experiment!

Cross product of two proper vectors is a pseudo vector. Does the mirror symmetry hold for all experiment?

Two corollaries of the above statements are Ans: No! In Beta decay

A cross product of a proper vector and a pseudo vector is a

proper vector.

43
 (b) (a) Let us understand the phenomena from the fundamental forces
and fundamental particle point of view. Find a schematic diagram
e e
v v of neutriono.

S S

Wu and coworkers discovered that that the electrons (beta)

are emitted preferentially opposite to the direction of the spin of
the Cobalt-60 nucleus (see Fig. (a)). Note that the spin vector per-
pendicular to the mirror does not change sign under mirror reflec- Left-handed Right-handed
tion, but the velocity vector does. Hence in the mirror image, the neutrino neutrino
electrons would move preferentially in the direction of the spin,
as shown in Fig. (b). Thus the law that the beta particles have a
preferential velocity opposite to their spin is violated in the mirror If the nature respects the mirror symmetry, then both left-handed
reflection. This is how mirror symmetry is violated in a beta decay and right-handed neutrinos should exist. In experiments however
experiment. we do not find right-handed neutrino. Contrast this observations
Mathematically, the violation of the mirror symmetry is due to with electrons that come in both positive and negative spin
the appearance of nonzero value of the pseudo scalar Q = S ⋅ v in flavors. The aforementioned observation of neutrino violates P-
the Lagrangian. Here S is the spin of the Cobalt-60 nucleus, and symmetry.
v is the velocity of the electron. Such quantities vanish in experi-
ments respecting mirror symmetry (e.g., gravitational and electro- Charge Conjugation
magnetic interactions). The potential of weak nuclear force how-
Particle → Antiparticle
ever is a combination of a proper scalar and a pseudo scalar.
This kind of potential was first proposed by Marshak and Sudar- For left-handed neutrinos, we expect the charge conjugation
shan, and Feynman and Gell-mann in 1957. to yield a left-handed antineutrino as shown in the following fig-
ure.
44
 Therefore, for weak interactions, a combination of the two, the
mirror and charge conjugation is respected. This is called CP sym-
metry. Note that matter and antimatter physics differ due to the C
violation.

However CP is also violated in K meson decay. There are two

paths in its decay process: electron antineutrino and the other an
electron neutrino. The amplitudes of decay differs for both the
Left-handed Left-handed processes. Hence CP is violated. A careful observation however
indicate that CPT, where T is the time reversal symmetry, is re-
neutrino antineutrino
spected in this experiment.

However we do not find left-handed antineutrionos. Hence the

charge conjugation symmetry is also violated for neutrinos, in fact Exercises
for weak interactions. 1. Which of the following forces would violate mirror symmetry?

However we observe that right-handed antineutrino exist, as (a) F=q v×B (b) F=q v×E
shown below.
(c) F=q v×B+mg (d) F=E+B

(e) F=E×B

In the above, E is the electric field, B is the magnetic fields, and g

is the acceleration due to gravity.

Left-handed
Right-handed
neutrino
antineutrino

45
Section 5

Relativistic Lagrangian

Relativistic Dynamics ∂L mv
p0 = =
∂v 1 − v 2 /c 2
A particle follows a trajectory for which the elapsed proper time
is minimum. This action is Lorentz invariant. Therefore the energy or the Hamiltonian of the free particle is

∫
S = − α dτ mc 2
H0 = p ⋅ v − L =
1 − v 2 /c 2

∫
= −α 1 − v 2 /c 2 dt
For a charged particle in an electromagnetic field

q d xμ
∫ ∫
In the nonrelativistic limit, 2 μ
S = (−mc )dτ + Aμd x = (−mc )dτ + Aμ 2
dt
c dt
1 v2
∫[ 2 c2 ]
S =−α 1− dt. Hence the Lagrangian of the particle is

L = L0 − q(ϕ − A ⋅ v/c)
Hence, when we choose α = mc 2, we obtain the usual
Lagrangian as L = mv 2 /2. Hence, the relativistic Lagrangian is The generalized momentum is

∂L
L0 = − mc 2 1 − v 2 /c 2 p= = p0 + qA/c = γ mv + qA/c
∂v
The linear momentum is
Hence γ 2 = 1 + (p − qA/c)2 /m 2c 2.

Hence the energy of the system is

46
 H = p ⋅ v − L = γ mc 2 + qϕ = m 2c 4 + (pc − qA)2 + qϕ

In the nonrelativistic limit,

1 2
L= mv − q(ϕ − A ⋅ v/c)
2

(p − qA/c)2
H= + qϕ
2m

The equation of motion is

d
(mvi + qAi /c) = − ∂i ϕ + (∂i Aj )vj
dt

d Ai ∂A ∂A
Using = i + i vj, we obtain
dt ∂t ∂xj

∂Aj
[ ∂t ]
d
(mvi) = − q ∂i ϕ + + (v × B)i = q(E + v × B)i
dt

which is the Lorentz equation.

47
Chapter 5

Solution of
Equation of
Motion
Section 1

One-dimensional Motion
1 2 ·2
L= ml ϕ − mgl(1 − cos ϕ)
2
Generic features
where φ is the displacement in the counterclockwise direction.
The equation of motion
U(x)
··
ϕ = − g sin ϕ,

E which is a nonlinear equation. Difficult to solve.

Note: Easy solution for small φ though.

a b c x
It is easier to solve the energy equation which is first order in
time:

1 ·2
For the given energy E, l ϕ = g(cos ϕ − cos ϕ0)
2

x<a and b<x<c: Not allowed since V > E.

We compute φ(t) using
a<x<b: oscillatory motion ϕ
dϕ′
∫0
t=
x>c: Asymptotically the particle will go to infinity. (2g/l )(cos ϕ′ − cos ϕ0)

l ϕ dϕ′
g ∫0 2
Pendulum =
(sin2(ϕ0 /2) − sin2(ϕ′ /2))

49
 (2) Repeat the exercise for (a) U(x) = − U0 /cosh2 α x where
−U0 < E < 0, (b) U = U0 tan2(α x), and (c) U(x) = A | x |n .
We make a change of variable

sin(ϕ/2) (3) A particle experiences a force F(x) = 1/x 3 − 1/x 2. Compute the
=x potential for the particle. Identify the equilibrium point as well
sin(ϕ0 /2)
as the fixed points. Sketch the phase space plots.
that yields
(4)
l x d x′ l
g ∫0
t= = K(x, k)
(1 − x ′2)(1 − k 2 x ′2) g

where k = sin(ϕ0 /2) and K(x, k) is Jacobi's elliptic integral of the

first kind.

We can compute the time period of the pendulum using

l 1 d x′ l
g ∫0
T =4 =4 K(x = 1,k)
(1 − x ′2)(1 − k 2 x ′2) g

For small ϕ0, k → 0. Hence

l
T = 2π
g

Exercises:

(1) Compute the time period of a particle of mass m which is

moving in a potential field U = x 4 /4 with an energy E.

50
51
Section 2

Two-dimensional motion

The inversion of the above equations yields

(d3 d4) (y) ( sin(ωt) )

d1 d2 cos(ωt)
A (y) =
1. Two-dimensional Isotropic Oscillator x x
=
The potential of an isotropic oscillator is

1 2 2 Therefore
V(x, y) = ω (x + y 2)
2
(x y) A T A (y) = 1 or
x

and hence the Lagrangian is

1 1
L = m(x· 2 + y· 2) + ω 2(x 2 + y 2)
2 2
and equations of motion are
m x·· = − ω 2 x; m ··y = − ω 2 y,
or
F = − mω 2r
The solution of the above equations are
x(t) = c1 cos(ωt) + c2 sin(ωt) = A cos(ωt − ϕA)
y(t) = c3 cos(ωt) + c4 sin(ωt) = B cos(ωt − ϕB)
α x 2 + βy 2 + γx y = 1 with
alpha = (d12 + d32); β = (d22 + d42); γ = 2(d1d2 + d3d4)
which is an equation of an ellipse since 4αβ − γ 2 = det(A T A) > 0 .
2. Motion under a central potential
We solve for the motion of a particle under the influence of
central potential: U(r) = U(r). Hence, the force F(r) = − ∇U(r) is
radial. Therefore, the angular momentum L about the centre of
the force is conserved. By definition L = r × p, hence L ⋅ r = 0.
Therefore, the motion of the particle will be in the plane
perpendicular to L. Therefore, we use two-dimensional radial-
polar coordinate system:

52
 ·
r = r r̂ + r ϕ ϕ̂

In the next section, we consider the motion of planet under the The potential for a pair of charges with same sign is repulsive:
influence of a star. α α
U(r) = = . (2)
| r1 − r2 | r
3. Motion under −α/r potential
The above potential is applicable to In the following discussion, we will solve for the motion of a
particle under an attractive potential. In particular, we will solve
1. star-plant system
for the motion of a planet. We will verify the three laws of Kepler.
2. Motion of a classical electron in an atom For the solar planets:

3. Scattering between two masses, and two charged particles Now we solve the equation of motion for star-planet system
shown in the above figure in the CM reference frame. Hence,
Note that the forces between two masses, and that between two
oppositely charged particles is attractive, i.e., m2 m1
r1 = r; r2 = − r, (3)
m1 + m2 m1 + m2
α α
U(r) = − = − , (1)
| r1 − r 2 | r
where r = r1 − r2. The kinetic energy of the system is

where α>0, and r1, r2 are the positions of the planet and star 1 1 1
T= m1v12 + m2v22 = μ r· 2,
respectively as shown in the following figure. 2 2 2

where μ = m1m2 /(m1 + m2) is the reduced mass. We denote | r· | by

v. Clearly r = r1 − r2 is a convenient variable to solve the above
problem.

The Lagrangian of the system is

1 2
L= μv − U(r).
2

53
Since ∂L /∂t = 0, the total energy equation of the system is

1 2
μv + U(r) = E
2

is conserved. Since the motion is in two dimensions,

1 ·2 · α
μ(r + r 2ϕ2) − = E
2 r
1 ·2
μr + Ueff (r) = E(4)
2

with
α L2 Figure 1: Plot of effective potential Ueff (r) vs. r
Ueff (r) = − + .
r 2μr 2

Equation (1) is an equation of a particle that is moving in an effec-

tive potential Ueff (r). Please note that the particle is moving in two The potential Ueff (r) has a minimum
dimensions, but we are able to map the equation of motion to a
α 2m
one-dimensional problem. A plot of Ueff (r) vs. r is shown in the fol- U0 = − , (5)
2Lz2
lowing figure.

at

Lz2
r0 = . (6)
mα

where | L | = Lz is the magnitude of the angular momentum. Note

that U(r) depends only on the angular momentum L, and not on E.

The effective potential follows the following properties:

1. E = Umin: r = r0, hence the trajectory is a circle of radius r0.

54
2. Umin < E < 0: E = Ueff yields two values of r where r· = 0, Lz2 α
E0 = = mv02 = .
which are rmin and rmax. Since m r· 2 /2 > 0, we obtain mr02 r0
rmin < r < rmax. BOUNDED SYSTEM.
Naturally, the time scale is
3. E ≥ 0: E = Ueff yields only one point where r· = 0. The planet
r0 Lz3
approaches to rmin, after which it gets scattered to ∞. t0 = = .
v0 mα 2
The energy equation is
Using the above set of scales, we make the following change of

2 ( dt )
2
1 dr αLz2 variables
μ + − = E, (7)
2μr 2 r
r = r0r̃; t = t0t̃, v = v0 ṽ, E = E0 Ẽ . Lz = L0 L̃z .
When we non-dimensionalise the equation discussed in the ear-
lier section using appropriate length- and time scales, they ap-
pear somewhat simpler. For non-dimensionalization, we use the Since Lz = L0, L̃z = 1, which yields
radius of the circular orbit as the length scale. For the circular or- dϕ
bit, L̃z = r̃ 2 = 1.
d t̃
v02 α
m = 2 . Here the time derivative is in terms of new non-dimensional time
r0 r0 t̃. When we substitute the above in the energy equation, we ob-
tain
Using Lz = mv0r0 for a circle, we obtain the radius of the circular

2 ( d t̃ )
2
orbit in terms of Lz as 1 d r̃ 1 1
+ 2 − = Ẽ . (8)
2r̃ r̃
Lz2
r0 = .
mα A change of variable ũ = 1/ r̃ provides

For a circular orbit, the velocity of the planet is d r̃ 1 d ũ 1 d ũ dϕ d ũ

=− 2 =− 2 =− .
dt ũ dt ũ dϕ dt dϕ
Lz
v0 = ,
mr0 Substitution of the above in Eq. (8) yields

while its energy is

55
2 ( dϕ )
2
1 d ũ ũ2 Solutions 1,2 are bound orbits, and 3,4 are unbound orbits. Plane-
+ − ũ = Ẽ. (9)
2 tary orbits are either of type 1 or type 2. Thus, Kepler's first law
that planetary orbits are ellipses is verified. Parabolic and hyper-
The solution of the above equation is bolic trajectories can be observed for the unbounded orbits.

1
ũ = =1+ 1 + 2Ẽ cos ϕ . (10)
r̃

Here we choose + sign that corresponds to the counterclockwise

rotation of the planet, as against the - sign which would lead to a
clockwise rotation. In addition, we choose C = π /2, which implies
that the nearest point of approach of the planet to the star is at
ϕ = 0.

Here one focus of the curve is at the origin (position of the star).

Definition: Figure 2: Keplerian orbits: circle, ellipse, parabola, and hyperbola

Latus Rectum 2p = 2r0 = 2 In dimensional form Eq. (10) is written as

Eccentricity e = 1 + 2Ẽ. r0 2ELz2

=1+ 1+ cos ϕ.
r mα 2
1 2
Clearly Ẽ = (e − 1).
2 2Lz2
Latus Rectum 2p = 2r0 =
There are four possible solutions: mα

1. e = 0 (Ẽ = − 1/2). 2ELz2

Eccentricity e = 1+ .
mα 2
2. 0 < e < 1 (Ẽ < 0).

3. e = 1 (Ẽ = 0). CLASSIFICATION OF KEPLERIAN ORBITS

Circle (e = 0)
4. e > 1 (Ẽ > 0).

56
A circle is a conic section with e = 0. Hence 1
Ẽ = − .
2ã
1
r̃ = 1 and Ẽ = −
2 Let us shift the origin of the coordinate system to the centre of
the ellipse by an amount ã − r̃min = ãe. The coordinates in the
In dimensional form
shifted system as (x′, y′) are
Lz2 cos ϕ e + cos ϕ
r = p = r0 = . x̃′ = x̃ + ãe = + ãe = ã ,
mα 1 + e cos ϕ 1 + e cos ϕ
and
sin ϕ 1 − e 2 sin ϕ
ỹ′ = ỹ = r̃ sin ϕ = = b̃ .
α mα 2 1 + e cos ϕ 1 + e cos ϕ
E=− =− ,
2r0 2Lz2
A simple manipulation shows that
This orbit corresponds to the minimum of the effective potential
x̃′2 ỹ′2
Ueff (r). + = 1,
ã 2 b̃2
Ellipse (0 < e < 1)
Since e ≥ 0, we conclude that Ẽ ≥ − 1/2. Note that the circular or-
When e < 1, the trajectory is an ellipse (Figs. 2). Hence bit of unit radius (p = 1) has total energy Ẽ = − 1/2.

1
r̃max =
1−e In non-dimensional units, the time period of the elliptical orbit is

1 1 T̃ = 2π ã b̃ = 2π(1 − e 2)−3/2 = 2π ã3/2 (11)

ã = (r̃min + r̃max) = .
2 1 − e2
We test the above for the solar planets and find a very good agree-
The minor radius is obtained by minimising y = r̃ sin ϕ, which ment.
yields

1
b̃ = .
1 − e2

57
 the corresponding points are called aphelion and perihelion re-
spectively, while for the geo satellites (including the Moon), the re-
spective points are {\em apogee} and {\em perigee} respectively.

Parabola (e = 1)

A parabla is a conic section with e = 1, and it corresponds to

E = 0. Hence

1
r̃ = , (23)
1 + cos ϕ

and it is plotted in Fig. 3. In the Cartesian coordinate the above

Fig. 3: Time period T of the planets in the solar system as a equation is
function of a = aP /aE, where aP, aE are the major radii of the planet
2px = p 2 − y 2,
and the Earth. For Earth a = 1.
which is an equation of a parabola.
In dimensional form
Hyperbola (e > 1)
1 1 α
a= r = r =
1 − e 2 0 2 | Ẽ | 0 | E | Hyperbolas are conic sections with e > 1, and it corresponds to
E > 0. Equation (13) shows that r → ∞ for cos ϕ0 = − 1/e. The
r0 Lz a
b= = = Lz range of allowed ϕ is (−ϕ0, ϕ0). By several algebraic manipulations
1 − e2 2m | E | 2αm
we obtain

T = t0r0−3/22πa 3/2 = 2πa 3/2 m /α or x ′2 y ′2

− = 1, (25)
a2 b2
m
T = t02π(1 − e 2)−3/2 = πα which is the equation of a hyperbola (see Fig.9.13).
2 | E |3
In Fig. 3 we plot all the for kinds of conic curves for a given p = 1.
Notation: The farthest point of the ellipse is called apsis, while the The circular orbit (e = 0) has radius p. As e increases,rmin de-
closest point apsides. For the orbits of solar planets and comets, creases and the point of closest approach inches towards the ori-
58
gin. The elliptical, parabolic, and hyperbolic orbits shown in the Laplace-Runge-Lenz vector
figure are for e = 0.5,1 and 3 respectively. It is also important to
note that all the orbits in the figure have same angular momentum
since L = mαp, but different energies. For Kepler's problem, the vector N = v × L − α r/r is a constant.
This vector is called the Runge--Lenz vector.
Fig. 9.14 exhibits the trajectories of solar planets, with most of
them closely resembling circles. Mercury has eccentricity of Proof:
0.206, which is maximum among all the solar planets. Mars has · ̂
v × L = Lz(r· r̂ + r ϕ ϕ) × ẑ
the second highest eccentricity of 0.093.
d r̃ d ϕ̃ ̂
= v0 Lz( r̂ + r̃ ϕ) × ẑ
d t̃ d t̃

d ϕ̃ d r̃ ̂
=α (− ϕ + r̃ r̂ )
d t̃ dϕ

1
= α(−e sin ϕ ϕ̂ + r̂ )
r̃

Therefore,

N = v × L − α r/r = − αe ̂x.

Figure 3 Trajectories of solar planets: (a) Mercury, Venus, Earth,
and Mars in increasing order; (b) Jupiter, Saturn, Uranus, and
Naputune; The planets Mercury, Mars, Earth, and Mars appear to-
gether near the origin. We normalise the distance in the units of
Earth-Sun distance.
Keplerian Orbit as a function of time
Landau
Hence the direction of N is along the apogee, that is from the foci
of the ellipse to the point with rmax. The above result also shows
that N
Laplace-Runge-Lenz vector is an additional constant for Kepler’s
problem. Thus, for the 2D motion, we have three conserved quan-
tities: E, Lz, N. There is only one free variable here, which is ϕ, the
azimuthal angle. Hence the motion is very constrained.
SOLVING THE FORCE EQUATION OF A PLANET

59
We can also derive the solution of the planetary orbit from the which yields
force equation that in non-dimensionalised form is
c1 = 2Ẽ + 1.

( d t̃ )
2 2
d r̃ dϕ 1
− r̃ = − , Therefore,
d t̃ 2 r̃ 2

A change of variable r̃ = − 1/ ũ yields ũ = 1 + 2Ẽ + 1 cos ϕ,

which is same as those derived using the energy method.

d t̃ ( dϕ )
d 2r̃ d d ũ d 2ũ dϕ 2
2 d ũ
= − =− = − ũ .
d t̃ 2 dϕ 2 d t̃ dϕ 2

A substitution of the above in Eq. (30) with an application of

dϕ 4. Tips to solve problems on Keplerian orbits
L̃z = r̃ 2 = 1 yields Here we assume that the mass and the parameter α is fixed.
d t̃

d 2ũ 1. A nondimensional planetary orbit is uniquely determined by its

+ ũ = 1,
dϕ 2 eccentricity e and latus rectum p.

whose solution is 2. The length scale r0 could be determined by various dimen-

sional quantities.
ũ = 1 + c1 cos(ϕ + c2) .
3. e and p are independent of mass of the planet, while the en-
By choosing rmin at ϕ = 0, we obtain c2 = 0. Hence ergy and angular momentum are proportional to the planetary
mass.
ũ = 1 + c1 cos ϕ .
In the following, we will solve three problems using the above
The constant c_1 is determined using the energy equation. At ideas.
rmin,max, du /dϕ = 0, substitution of which in Eq. (9.6.10) yields
Example 9.4 Halley's comet is in an elliptical orbit around the
1 2 Sun whose mass is 2 × 1030 kg. The eccentricity of the comet's or-
ũ − ũmax,min = Ẽ,
2 max,min bit is 0.967, and the period is 76 years. Using these data, deter-
mine the closest and the farthest distance of Halley's comet from
ũmax,min = 1 ± 2Ẽ + 1,

60
the Sun. Determine the speed of the comet when it is closest to Note that E is proportional to the mass of the comet. At the clos-
the Sun? est point to the Sun,

Solution The problem statement provides the eccentricity e and GMs m 1 2

E=− + mvmax.
the time period. Thus rmin 2

T = 2πa 3/2 m /α = 2πa 3/2 / GMs , (1)
where Ms = 2 × 1030 kg is the mass of the Sun, and
G = 6.67 × 10−11m3 /s2 kg is Newton's gravitational constant, we
compute the semi-major radius a of the comet as
1/3
GMsT 2
( 4π 2 )
a= = 2.69 × 1012 meters . (2)
Hence we obtain the speed at rmin as vmax = 54.3 km/s, which is
approximately 1.8 times the average orbital speed of the Earth.
Example 9.5 For a Earth's satellite, rmin and rmax are 10000 km
and 6000 km respectively. The mass of the satellite is 2000 kg.
Compute the eccentricity, energy, angular momentum, and mini-
mum and maximum speeds of the satellite.
Solution:

Note that a is independent of the mass of the comet. Using a, we rmax − rmin 4 1
The eccentricity of the satellite e = = =
obtain rmax + rmin 16 4

rmin = a(1 − e) = 8.9 × 1010m ∼ 0.59 AU, The semi-major radius of the satellitea = (rmax + rmin) /2 = 8000 km

rmax = a(1 + e) = 5.3 × 1012m ∼ 3.53 AU. The latus rectum of the elliptical orbit p = r0 = a(1 − e 2) = 7500 km.

where AU, called the astronomical unit, is the distance between Also, α = GMm = gRE2 m = 7.2 × 1017 Joule m
the Sun and the Earth, and it is equal to 1.5 × 1012 meters. We can
compute the latus rectum of the ellipse using Therefore, the angular momentum =
L = mr0α = 1.039 × 1014 kg m2 /sec
p = r0 = rmin(1 + e) = a(1 − e 2)
α
The energy of the comet is computed using Eq. (9.5.26) as Energy = E = − = 4.5 × 1010Joules
2a
α GMs m GMs m
E = Ẽ = (e 2 − 1) =− = − m × 2.48 × 107J . Since Lz = m rmin Vmax, we obtain
r0 r0 2a

61
 1.03 × 1014 GMm
Vmax = = 8.5km /s The energy at point A of the circular orbit = −
2 × 103 × 6 × 106 4Re

1.03 × 1014 GMm

Similarly Vmin = = 5km /s The energy at point B of the circular orbit = −
2 × 103 × 107 8Re

For the elliptical orbit

Example 9.6 A space vehicle of mass 2000 kg is orbiting around a = (rmin + rmax )/2 = 3Re
the Earth in a circular orbit with the radius of orbit as 2Re. We wish
4−2 1
to transfer the vehicle to a circular orbit of radius 4Re (see Fig. e= =
4+2 3
9.15). One of the schemes to transfer the vehicle is to use a semi-

( 3)
elliptical orbit as shown in the figure. What velocity changes are 1 8
p = (1 + e)rmin = 1+ 2Re = Re
required at the points of intersection, A and B? What is the 3
change in energy of the system in the two configurations?
α GMm GMm
E=− =− =−
2a 2 × 3Re 6Re

1 2 GMm
E= mv − ,
2 r

the velocity of the vehicle at A for the elliptical orbit

m( 2Re )
2 GMm 2GM
VAe = E+ =
3Re
Figure 4 Example 9.6: A space vehicle is transferred from an in-
ner orbit to an outer orbit.
and at B
Solution:

m( 4Re )
2 GMm GM
VBe = E+ =
Total energy of a circular orbit of radius r is −GMm /(2r), where M 6Re
is the mass of the planet and m is the mass of the satellite. There-
fore,
62
For the circular orbit, v = GM/r, where r is the radius of the cir- convenient to work with a non dimensionalized form of the energy
cular orbit. Therefore, equation. For non dimensionalization we use r0 = Lz /(mω) as

GM the length scale, 1/ω as the time scale, and L ω as the energy
VAc = scale, which yields
2Re
t′ E′
r = r0r′; t = ; E= .
GM ω Lω
VBc =
4Re
The energy equation in terms of the non dimensionalized vari-
ables is
Hence,

2 ( d t̃ )
2
1 d r̃ 1 1
+ 2 + r̃ 2 = Ẽ .
Re ( ) ( 2)
GM 2 1 2 1
ΔVA = − = gRe − 2r̃ 2
3 2 3
The above equation resembles an equation of a particle executing

(2 6)
1 1 a one-dimensional motion under the effective potential
ΔVB = gRe −
1 1 2
Ueff = + r̃ ,
2r̃ 2 2
In the next section, we will work out the trajectories of a two-
dimensional oscillator. 54. Two-dimensional oscillator in radial- which is plotted in Fig. 9.4. The effective potential has a mini-
polar coordinate mum value of 1 at r̃ = 1. At the minimum value, the orbit of the
particle is a circle with unit radius. For Ẽ > 1,rmin ≤ r ≤ rmax as indi-
cated in Fig. 9.4. We will solve for the orbits for this regime.
The equation of a two-dimensional oscillator in radial-polar coordi-
nate is

2 ( dt )
2
1 dr Lz2 1 2 2
m + + mω r = E, (1)
2mr 2 2

where Lz is the angular momentum of the oscillator, and m and ω

are the mass and frequency, respectively, of the oscillator. It is

63
 ũ2 = Ẽ + Ẽ 2 − 1 sin 2ϕ

or

(x̃ 2 + ỹ2)Ẽ + 2 x̃ ỹ Ẽ 2 − 1 = 1

A change of variable x̃ = x̃′ + ỹ′ and ỹ = x̃′ − ỹ′ leads to

x̃′2(Ẽ + Ẽ 2 − 1) + ỹ′2(Ẽ − Ẽ 2 − 1) = 1,

which is an equation of an ellipse whose semi-major and semi-

minor radii are (Ẽ − Ẽ 2 − 1)−1/2 and (Ẽ + Ẽ 2 − 1)−1/2 respec-
tively. Note that Ẽ = 1 yields a circular trajectory of unit radius.
Figure 5: For an isotropic oscillator, plot of effective potential
Ueft = 1/(2r 2) + r 2 /2 vs. r. Note that it is much easier to solve this problem in Cartesian
coordinate.
A change of variable ũ = 1/ r̃ yields
5. Virial Theorem
( dϕ )
2
d ũ 1
+ ũ2 + 2 = 2Ẽ. We define the time average of a quantity Q as
ũ
1 T
T ∫0
which is integrated using ⟨Q⟩ = Q(t)dt,

ũd ũ
∫ ∫
= dϕ.
where T is the time period of the planetary orbit. Let us take
(Ẽ 2 − 1) − (ũ2 − Ẽ )2
Q = p ⋅ v, where p is the linear momentum of the planet. Using
the product rule
The solution of the above equation is
d
2
ũ − Ẽ p⋅v= (p ⋅ r) − ma ⋅ r,
sin −1
= 2(ϕ + C ) dt
Ẽ 2 −1
where a = − α r/r 3 is the acceleration of the planet. Temporal aver-
where C is a constant, which is chosen to be zero. Therefore aging of each of the terms of the above equation yields
⟨p ⋅ v⟩ = 2 ⟨T ⟩,

64
 −m ⟨a ⋅ r⟩ = ⟨α /r⟩ = − ⟨U ⟩, and

⟨ dt ⟩ T→∞ T ∫ dt
d 1 d 1
dt (p ⋅ r) = lim [p ⋅ r] = 0
t=T
(p ⋅ r) = lim
T→∞ T t=0

because (p ⋅ r) is finite for a bounded orbit.Therefore,

2 ⟨T ⟩ = − ⟨U ⟩ .

This is called the virial theorem . Clearly

⟨E⟩ = ⟨T ⟩ + ⟨U ⟩ = − ⟨T ⟩ = ⟨U ⟩/2.

The above procedure is quite general, and it is used to derive the

relationship between kinetic and potential energy for other forms
of central potentials.

6. General Central Force Problem Figure 6 Effective potential for (a) 0 < n < 2 with α = 1, (b) n = 2
A general attractive central force potential of the power law form
with α = 1 (solid line) and α = 0.1 (dashed line), (c) n > 2 with
is
α = 1, (d) m > 0 with β = 1.
α
U(r) = − , or U(r) = βr m, The behaviour of these systems are
r n

with α, n, β, m > 0. The effective potentials are 1. 0 < n < 2: Ueff has a similar behaviour as that for −1/r (Kepler-
ian) potential. Therefore, we obtain bounded orbits for E < 0
Lz2 and unbounded orbits for E > 0.
Ueff = U(r) +
2mr 2
2. The orbits for n > 2 are inward spiral towards the centre.
and they have been plotted in Fig. 9.5.
3. n = 2 with Ueff = α′ /r 2: we obtain unbounded and bounded or-
bits depending on the value of α′ and E.

4. βr m: all the orbits are bounded.

65
Also exact solutions exists only for n = 1 and m = 2, and they If we demand that U(r = ∞) = 0, then clearly, E = 0. The force is
have been solved in this chapter. It has been shown rigorously −dU(r)/dr ∝ r −3, and its direction is radially inward.
that the orbits are closed only for these two cases.
Example 2: Consider a potential
Example 1: An object of unit mass orbits in a central potential
α
U(r). Its orbit is given by r = a exp(−bϕ) with b > 0. What are the U(r) = − ,
r2
constants of motion? Find the potential.
where α > 0. Solve for the trajectories of a particle moving under
Solution At ϕ = 0, r = a. As ϕ increases, r decreases. The veloc- this potential.
ity of the particle is
· Solution The energy equation for a particle of mass m moving un-
v = r· r̂ + r ϕ ϕ̂ der the above potential is
· ·
= a exp(−bϕ)(−b ϕ) r̂ + r ϕ ϕ̂

2 ( dt ) ( 2m
2
Lz2
) r2
1 dr 1
· m + −α = E,
̂
= r ϕ(−b r̂ + ϕ).

The constants of motion are angular momentum L and energy E: where the effective potential is
·
Lz = r 2ϕ Lz2
( 2m ) r2
1
Ueff = −α .
·
= a 2 exp(−2bϕ)ϕ,

1 ·
· Using r = 1/u and Lz = mr 2ϕ, the energy equation is transformed
E = (r· 2 + r 2ϕ2) + U(r)
2 to
1 2 ·2 2
( dϕ ) (
2

Lz )
= r ϕ (b + 1) + U(r) . du 2mα 2mE
2 + 1− 2 u2 = 2
Lz
Lz2
= (b 2 + 1) + U(r) (1)
2r 2 or

Since E is a constant, U(r) = − αr −2 with du

=± γ − βu 2 ,
dϕ
α = Lz2(b 2 + 1)/2. (2)
66
where γ = 2mE /L 2 and β = (1 − 2mα /Lz2). In terms of β, the effec- 5. β = 0: For this case, the coefficient of u 2 of Eq. (2) is zero,
tive potential is hence E > 0 or γ > 0. Therefore, the solution of Eq. (3) is
u = γϕ or rϕ = γ. The particle trajectory for this case

2m ( L ) r2
Lz2 2mα 1 Lz2 β
Ueff = 1− 2 = × . similar to Fig. 7(b), except that the convergence to the origin
2m r 2
is slower for this case compared to item 2.
PS: In this problem we cannot construct a length scale purely
from Lz, m and α, like in −α /r potential. Here, the length scale is
α /E. From the above equation, it is appropriate to choose β/γ
as the length scale.

The solutions of the equation for different cases are

1 γ
1. γ > 0,β > 0: = cos( βϕ). Here r(ϕ) increases with ϕ,
r β
as illustrated in Fig. 7(a). Here β = 0.01 and γ = 2.

1 γ
2. γ > 0,β < 0: = sinh( | β | ϕ). Here the particle falls to
r |β|
the centre of the potential, as illustrated in Fig. 7(b). Here
β = − 0.1 and γ = 2.

1 |γ | Figure 7 Trajectories for U(r) = − α /r 2 potential for the four

3. γ < 0,β < 0: = cosh( | β | ϕ). Here too the trajectory cases discussed discussed in Example 9.9: (a) case 1: β = 0.01,
r β
γ = 2; (b) case 2: β = − 0.1, γ = 2; case 3: β = − 0.1, γ = − 2;
falls to the origin, as illustrated in Fig. 7(c). Here β = − 0.1 case 4: β = − 0.1, γ = 0.
and γ = − 2.

4. γ = 0,β < 0: r = r(0)exp(− | β | ϕ). Here β = − 0.1. The solu-

tion shown in Fig. 7(d) 7. Stability of Keplerian Orbits

67
The force equation for the planet is x(t) = ̂x0 exp(iωt); y(t) = ŷ 0 exp(iωt)
2
··r = L − α = − dUeff ,
Substitution of the above in the equation yields
m 2r 3 mr 2 dr
2 dΩ2
−ω ̂x0 − 2iΩ0ω ŷ 0 = − ̂x0 R
where the effective potential Ueff plotted in Fig. is convex. Hence dR
the orbit will be stable.
−ω 2 ŷ 0 + 2iΩ0ω ̂x0 = 0.
RAYLEIGH CRITERION
Elimination of ̂x0, ŷ 0 in the above yields
Let us analyze the dynamics of the orbit in more detail. For sim-
plicity, we will work out the stability of a circular Keplerian orbit of 2 2 dΩ2
ω (ω − 4Ω20 −R )=0
radius R0, We denote the angular velocity of the planet at a dis- dR
tance R from the centre as ω(R). We employ a reference frame
which is attached at a distance R0 and it rotating with an angular whose solutions are ω = 0 and
velocity of ω(R). Now consider a planet of mass m located at dΩ2 1 d
2
(x,y) and moving with velocity v = (x, · y).
· ω = 4Ω20 +R = 3 (R 4Ω2)
dR R dR
In the undisturbed state, the gravitational force acting on planet For stability ω 2 > 0. The above condition is called the Rayleigh cri-
provides the centripetal force −m[Ω(R0 + x)]2(R0 + x) ̂x to facilitate terion for centrifugal stability.
the circular motion. In the rotating frame, the equation of motion
is For the planetary orbit, Ω2 = CR −3,where C is a constant, hence

m ··r = − m[Ω(R0 + x)]2(R0 + x) ̂x + m[Ω(R0)]2(R0 + x) ̂x − 2mΩ × r· ω 2 = Ω2.

In terms of components Therefore the orbit is stable.

2 MAGNETOROTATION INSTABILITY
dΩ
x·· − 2Ω0 y· = − xR
dR Balbus proposed that the magnetic field acting like a spring can
··y + 2Ω x· = 0 destabilize a Keplerian orbit. He modelled the magnetic force as
0
−Kr, and hence modelled the equation of motion of the planet as
where Ω0 = Ω(R0). To solve the equations, we attempt

68
 2
dΩ massive objects exhibits a chaotic motion. It has been shown rig-
x·· − 2Ω0 y· = − xR − Kx
dR orously that for some set of parameters, the final states of the
two nearby initial states will diverge exponentially. Malhotra, Hol-
··y + 2Ω x· = − Ky
0 man, and Ito (2001) argued however that the time-scale for this
divergence is over 5 to 10 million years, and the escape time-
By following a similar procedure as above, we obtain
scale of the planets is beyond the age of the Sun. These studies
dΩ2 show that even though a two-body problem is exactly solvable,
( dR )
4 2 2
ω − (2K + κ )ω + K K + R =0 three-body and many-body problems are much more complex.
These issues will be discussed when we take up nonlinear dynam-
ics and chaos.
2 dΩ2
where κ = 4Ω20 +R . The solutions of the above equation are
dR
9. Additional interesting results on central
dΩ2 potential orbits
( dR )
2ω = (2K + κ ) ±
2 2 2 2
(2K + κ ) − 4 K + R

1. Bertrand theorem: For central potentials, the closed orbits are

2
For instability we require that ω < 0 (imaginary ω), which is possi- possible only for potentials −α /r and kr 2. Other potentials
ble if could yield bounded orbits, but they will not be closed. See
dΩ2 Goldstein et al. for further discussion.
K+R < 0.
dR
2. Interestingly, potentials −α /r and kr 2 have exact solution also in
For Keplerian orbits, the above condition translates to K < 3Ω20. quantum mechanics.
Thus a weak magnetic field can destabilize a Keplerian orbit.
3. In quantum mechanics, the operators L 2 and Lz commute with
This instability is invoked to explain the accretion process in astro-
Hamiltonian. This is also termed as SO(3) symmetry (special
physics.
orthogonal group or rotation in 3D). Rotation symmetry about
8. Chaos in Three-body Problem the z axis yields degeneracy with m quantum numbers. Thus
A critical question is whether a planet under the influence of the energy could be function of the principle quantum number n
other two masses is stable or not. Poincaré was the first mathe- and l. This conservation law is applicable to wavefunction.
matician to argue that that the a planet under the influence of two
69
 Contrast this with the classical particle dynamics in which all
the three components Lx, L y, Lz are conserved.
4. Laplace-Runge-Lenz vector is an additional constant for
Kepler’s problem. It has a consequence on the degeneracy of
H atom. The symmetry of Kepler’s problem can be mapped to
SO(4). Consequently, different l quantum numbers for a given
n have the same energy. http://www.cds.caltech.edu/
~marsden/wiki/uploads/projects/geomech/
Alemicds205final.pdf
Exercises
1. A particle has constant angular velocity, but its radial velocity
is proportional to its distance from the origin. Derive an equa-
tion for the trajectory of the particle?
2. A bead of mass m slides on a long rod, which is rotating with a
constant angular velocity Ω. The coefficient of friction between
the rod and the bead is μ, a constant. Work out your answers in
a fixed inertial frame whose origin coincides with the fixed end
of the rod.
(a) Write down the equation of motion for the bead in the
radial-polar coordinate system. Is the angular momen-
tum of the particle about the origin conserved?
(b) Solve for the motion of the bead with an initial condition
(r = 0,r· = v0). Show that the system is in equilibrium if
v0 = 0.
· Solve for the
(c) Consider a variation in which μ(r) = kr / r.
motion of the bead.
3. Evaluate the ratio of the mass of the Sun to that of the Earth
given that a terrestrial year is 365.24 terrestrial days, and a lu-
nar month is 27.3 days. Assume that the mean radius of
Earth's orbit is 1.49 × 108 km, and radius of moon's orbit is
3.8 × 105 km, and the orbits are circular.
4. A shell is fired from a Bofors gun with velocity 100 m/s. The an-
gle of the projectile is 45∘ with respect to the horizontal. Com-
70
 pute the trajectory of the shell using (a) g = const, (b) g ∝ − r̂ /r 2
(Kepler's problem). Show that both these results are consistent.
5. Perilion and aphelion (shortest and largest distance from the
Sun) of Pluto's orbit are 29.658 AU and 49.305 AU respec-
tively. The corresponding quantities for the Earth are 0.999860
AU and 1.0000 AU. Compute the orbital time period of Pluto
given that Earth's time period is 365.256 solar days.
6. Compute the ratio of the maximum and minimum speed of the
Earth around the Sun given that the eccentricity of Earth's or-
bit about the Sun is 0.0167. Do the same exercise for Pluto
whose eccentricity is 0.24.
7. A comet orbiting around the Earth has a velocity 10 km/sec at
apogee and 80 km/sec at perigee. Compute the shortest and
largest distances for the comet from Earth's centre.
8. Devise schemes to measure the mass of the moon.
9. Compute mass of the Sun using the orbital properties of the
Earth.
10.A geostationary satellite is one whose position in the sky re-
mains the same for an observer on the Earth. These satellites
are place above Earth's equator. Compute the radius of the
orbit of a geostationary satellite.
11.Consider a general attractive radial force f (r) of the form r β act-
ing on a planet:
(a) Write down the equation of motion for the r coordinate of
the planet.
(b) Identify a stationary orbit, which is circular, for the sys-
tem.
12. An object of unit mass orbits around a central potential U(r).
The orbit of the object is r = 1/(1 + cos ϕ), and its angular momen-
tum is L.
(a) Sketch the trajectory of the object.
(b) Express object's velocity in terms of ϕ and L.
(c) What is the potential on the object?
13. A particle moving in a central potential follows a trajectory
given by rϕ = const. Sketch the trajectory of the particle.
Also, compute the potential of the particle as a function of r.
14. The trajectory of a particle is described by x = a cos(ωt),
y = b sin(ωt). Determine the force that acts on the particle at
every point of the path.
x2 y2
15. The trajectory of a particle is given by a curve 2 + 2 = 1,
a b
and the acceleration of the particle is always along the y-axis.
At t = 0, the particle was at (x = 0,y = b) and had velocity v0.
Determine the force acting on the particle at all time.
16. Estimate the escape velocity on the surface of the Earth.
Also estimate the escape velocity on a asteroid whose radius
is 5 km, and whose density is same as that of the Earth.
71
17. A particle of unit mass is moving with a constant speed v on
a curve y = A sin x. Compute the velocity, acceleration, and
force of the particle at x = 0,π /2,π and 3π /2. Take A = 1.

18. In Section 1 we showed that the trajectory of an isotropic os-

cillator is an ellipse for ϕB − ϕA = π /2.

19. Apply Virial's theorem to the potentials (a) U = − α /r n and (b)

U = cr m with c, m, n > 0. Obtain relationships between the ki-
netic energy, potential energy, and total energy. Of course,
you need to assume bounded orbits for the particle in motion.

20. For Kepler's problem, prove that v × L − α r/r is a constant.

This vector is called the Runge--Lenz vector. What is the direc-
tion of the vector?

21. Numerically solve for particle trajectories when the particle

moves in the potential (a) U(r) = − α /r 3, (b) U(r) = − α /r 3/2, (c)
U(r) = βr 3. Choose appropriate values for α and β.

72
Section 3

Collisions

The effective potential is

Collisions either via contact interactions or fields. Lz2 α

Ueff = +
μr 2 r
First we deal with collision processes via contact forces.
which has only unbound solution.

Rutherford Scattering
We consider scattering between two charged particles, which is
mediated by the electric field between the particles. When the
charged particles have opposite signs, then the solution is same
as the parabolic and elliptic orbits discussed in Sec. XXX.

Here we discuss scattering when the charged particles have the Figure:
same sign. For such systems, the method to solve is same as
that discussed in Sec. XXX, except that the potential energy is In terms of u = 1/r, the equation transforms to

α 1/2

dϕ ( Lz )
U(r) = du 2mE 2mα
r = − 2 u − u2
2 Lz

with α > 0. For such systems, the energy conservation yields

1 ·2 · α
μ(r + r 2ϕ2) + = E
2 r

73
We nondimensionalize the above equation using the scales We sketch the solution in Fig. 1 for a two particle system. The
defined for the Kepler problem. Note however that r = r0 does not centre of the force is at the origin. Contrast the figure with that
correspond to the circle; here r0 just provides a length scale. for Kepler’s problem where the the hyperbola opens its mouth to
the left.

2 ( dϕ )
2 2
1 d ũ ũ
+ + ũ = Ẽ.
2

The solution of the above equation is
y
1
ũ = = 1 + 2Ẽ cos ϕ = e cos ϕ − 1
r̃
r

ELz2
r1

where Ẽ = . Note that e > 1 since E > 0. Hence, the particle x

mα 2 r2

executes a hyperbolic trajectory. The closest distance from the

centre of the force is 1/(e − 1), and r → ∞ when cos ϕ → 1/e.

In dimensional form the solution is

p
= e cos ϕ − 1,
r

where

Lz2
p= = r0,
mα

2ELz2
e= 1+ .
mα 2

74