Small Oscillation

ide
Sk Jahiruddin*

gu *Assistant Professor
Sister Nibedita Govt. College, Kolkata
ics
Author was the topper of IIT Bombay M.Sc Physics 2009-2011 batch
He ranked 007 in IIT JAM 2009 and 008 (JRF) in CSIR NET June 2011

He has been teaching CSIR NET aspirants since 2012

ys
Ph

1
©Sk Jahiruddin, 2020 Small Oscillation

Contents

1 Stability and equilibrium 3

ide
1.1 Bounded and Unbounded motion . . . . . . . 4

1.2 Integration of equation of motion in 1D . . . . 5

2.1
gu
2 Small oscillation

Small oscillation in more than one dimension . 10

8
ics
3 General formulation of Small Oscillation 13

3.1 Matrix formulation of small oscillation problem 17

ys

3.2 Rigorous solution of Linear triatomic molecule 25

3.3 Making the non quadratic potential quadratic 32

Ph

[email protected] 2 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

1 Stability and equilibrium

The result F = −dU/dx is useful not only for computing the

ide
force but also for visualizing the stability of a system from the
potential energy plot.

Suppose there is a force on the particle is F = −dU/dx,

and the system is in equilibrium where there is no force i.e.

gu
dU/dx = 0. If this occurs at a minimum of U it is a stable
equilibrium whereas if it is at a maximum of U, the equilibrium
is unstable. Say, dU/dx = 0 occurs at some point x0 . To test
ics
for stability we must determine whether U has a minimum or
a maximum at x0 . One needs to examine d2 U/dx2 at x0 .
ys
Ph

[email protected] 3 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

d2 U d2 U
>0 → Stable; <0 → Unstable;
dx2 dx2

ide
d2 U
= 0 → Neutral
dx2
If the second derivative is positive, the equilibrium is stable;
if it is negative, the system is unstable. If d2 U/dx2 = 0, one
must look at higher derivatives. If all derivatives vanish so

gu
that U is constant in a region about x0 , the system is said to
be in a condition of neutral.
ics
1.1 Bounded and Unbounded motion

In Mechanics we generally divides the motion in two broad

ys

categories. The Bounded system, where energy is negative

and the Unbounded system where the energy is positive.
Bounded system may be periodic or non periodic. In this
Ph

section we will discuss the periodic bounded system their os-

cillation and frequencies.

since kinetic energy K = E − U can never be negative, the

motion of the system is constrained to regions where U < E.

[email protected] 4 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

The points where E = V (K.E = 0) are called turning

points. The following figure is an illustration of bounded mo-
tion of one potential with two different energy and turning
points.

ide
gu
ics
ys

1.2 Integration of equation of motion in 1D

 2 r
1 dx dx 2
E= m + V (x) =⇒ = [E − V (x)]
Ph

2 dt dt m
(1.1)

Hence
dx
dt = q (1.2)
2
m [E − V (x)]]
So the time period as a function of energy is twice the time

[email protected] 5 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

the particle take to go from x1 to x2

r Z x2 (E)
m dx
T (E) = 2 p (1.3)
2 x1 (E) [E − V (x)]

ide
Example 1.1: Example: Consider the potential
shown in the figure. Find the period of oscillation

gu
ics
ys

Time period is twice the time taken to reach from starting

point to end point
Z b Z 2b
dx dx
Ph

T =2 q +2 q
0 2E b 2(E−U0 )
m m
r r
m m
= 2b + 2b
3U U
 0 r 0
1 m
= 2b 1 + √
3 U0

[email protected] 6 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

Example 1.2: Determine the period of oscillation, as a

function of the energy, when a particle of mass m moves
in fields for which the potential energy is

ide
(a) U = A|x|n , (b) U = −U0 / cosh2 αx, −U0 < E < 0, (

(a) is already done in the action angle variable. Here we

will get same ans by doing similar integration.

gu p
T = 2 (2m)

=2
r
Z

2m
·
0
(E/A)1/n

 1/n Z 1
E dy
p
dx
(E − Axn )
ics
p
E A 0 (1 − y n )

By the substitution y n = u the integral is reduced to a beta

function, which can be expressed in terms of gamma functions:
ys

r  1/n
2 2πm E Γ(1/n)
T = ·
Γ 12 + 1/n


n E A
Ph

Do (b) and (c) similarly. The ans are

p
(b) T = (π/α) (2m/|E|),
p
(c) T = (π/α) [2m/ (E + U0 )]

[email protected] 7 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

2 Small oscillation
Nearly every bound system oscillates like a harmonic oscillator
if it is slightly perturbed from its equilibrium position.

ide
gu
ics
Look at the figure. Expand U (r) about r0 , the position of
the potential minimum.
2 
 
dU  1 d U
ys

U (r) = U (r0 ) + (r − r0 ) + (r − r0 )2 2  + · · ·
dr r0 2
 dr r0
However, since U is a minimum at r0 , (dU/dr)|r0 = 0. Fur-
Ph

thermore, for sufficiently small displacements. we can neglect

the terms beyond the third in the power series. In this case,
kx2
U (x) = constant +
2
This is the potential energy of a harmonic oscillator, where
d2 U 

k= ; the effective spring constant
dr2  r0

[email protected] 8 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

Frequency of small oscillation is

∂U
mẍ = F (x) = − = −kx, ẍ + ω 2 x = 0,
∂x

ide
r r 
k 2
1 d U 
ω= =
m m dr2 

r0

Example 2.3: The time period of a particle of mass m

(a) π
sgu
, undergoing small oscillations around x = 0 , in the
potential V
mL2
V0
= V x
s


0 cosh L , is

(b) 2π
mL2
2V0
[NET Dec 2018]
ics
s s
mL2 2mL2
(c) 2π (d) 2π
V0 V0
ys

dV
Solution: First make dx = 0 to find the equlibrium point.

x V0  x/L −x/L
 dV V0  x/L −x/L

V0 cosh( ) = e +e → = e −e
L 2 dx 2L
Ph

So
dV
= 0 → ex/L − e−x/L = 0 → x = 0
dx
Now the double derivative
d2 V V0  x/L −x/L
 1
= e + e = V;
dx2 2L2 L2

[email protected] 9 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

d2 V 

V0
=⇒ =
dx2 x=0 L2
So the frequency of small oscillation is

ide
s  r
1 d2 V  V0
ω= =
m dx2 x=0 mL2

Hence the time period

2.1
gu T = 2π/ω = 2π
s
mL2
V0

Small oscillation in more than one di-

ics
mension
You have already know the theory in small oscillation in 2
and 3 dimension, I’ll not repeat that. You can proceed in
ys

both way. Either calculate the equation of motion and solve

directly or just calculate the V and T matrix and need to find
the frequencies of small oscillation by getting
Ph

|V − ω 2 T = 0|

we will proceed with interesting examples.

As the first and most simple system of vibrating mass

points, we consider the free vibration of two mass points, fixed

[email protected] 10 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

to two walls by springs of equal spring constant, as is shown

in the figure

ide
gu
ics
The two mass points shall have equal masses. The dis-
placements from the rest positions are denoted by x1 and x2 ,
respectively. We consider only vibrations along the line con-
necting the mass points.
ys

When displacing the mass 1 from the rest position, there

acts the force −kx1 by the spring fixed to the wall, and the
force +k (x2 − x1 ) by the spring connecting the two mass
Ph

points. Thus, the mass 1 obeys the equation of motion

mẍ1 = −kx1 + k (x2 − x1 )

For the second mass we get

mẍ2 = −kx2 − k (x2 − x1 )

[email protected] 11 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

We first determine the possible frequencies of common vibra-

tion of the two particles. The frequencies that are equal for all
particles are called eigenfrequencies. The related vibrational
states are called eigen- or normal vibrations. These defini-

ide
tions are generalized for a N -particle system. We take the
solution
x1 = A1 cos ωt, x2 = A2 cos ωt
i.e., both particles shall vibrate with the same frequency ω .

gu
The specific type of the solution, be it a sine or cosine function
or a superposition of both, is not essential. We would always
get the same condition for the frequenty, as can be seen from
ics
the following calculation.
Put the solution in the equation of motion

A1 −mω 2 + 2k − A2 k = 0


(2.1)
ys

2


−A1 k + A2 −mω + 2k = 0

The system of equations has nontrivial solutions for the am-

plitudes only if the determinant of coefficients D vanishes:
Ph

 
 −mω 2 + 2k −k 
2
2
D= = −mω + 2k − k2 = 0
 
2
−k −mω + 2k

 

We thus obtain an equation for determining the frequencies:

4 k 2 k2
ω −4 ω +3 2 =0
m m
[email protected] 12 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

The positive solutions of the equation are the frequencies

r r
3k k
ω1 = and ω2 =
m m

ide
These frequencies are called eigenfrequencies of the system;
the corresponding vibrations are called eigenvibrations or nor-
mal vibrations. To get an idea about the type of the normal vi-
brations, we insert the eigenfrequency into the equation (2.1)

gu
For the amplitudes, we find

A1 = −A2 for ω1 =
r
3k
m
ics
and r
k
A1 = A2 for ω2 =
m
ys

3 General formulation of Small Os-

cillation
Ph

Consider a system with {qi } as the generalized coordinates.

Since the system is conservative, V (q1 , q2 , . . . , qN ) . Lagrange
defined equilibrium as a configuration in which all generalized
forces vanish, i.e. ∂V
∂qi = 0.

∂ 2V
> 0 Signifies Stable equilibrium
∂qi ∂qj
[email protected] 13 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

and
∂ 2V
< 0 Signifies unstable equilibrium
∂qi ∂qj
Without loss of generality, let us shift the equilibrium posi-

ide
tion to the origin (q1 = q2 = . . . , = qN = 0). If the system is
disturbed to a configuration {qi } , we can write,
X  ∂V
V (q1 , q2 , . . .) = V (0, 0, . . .) +
∂qi

+ gu
1 X ∂ 2V
 

2 i,j ∂qi ∂qj 0

qi qj

+ Higher order terms

i
ics
where the partial derivatives are evaluated at the posi-
tion of equilibrium and all higher order terms” which involve
third order or higher corrections are neglected. If the poten-
ys

tial energy is measured from its minimum value, we choose

V (0, 0, ...) = 0. Along with
 
Ph

∂V
=0
∂qi 0

The leading term in the change in potential energy is then

1 X ∂ 2V
 
>0
2 i,j ∂qi ∂qj 0

[email protected] 14 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

for stable equilibrium. Let us write

 2 
∂ V
Vij =
∂qi ∂qj 0

ide
so that X1
V = Vij qi qj with Vij = Vji
i,j
2
Now, the kinetic energy in general a quadratic in generalised

gu
velocities, and can be written as

T =
1X
2 i,j
mij q̇i q̇j
ics
The coefficients mij are in general functions of the coordinates
qk , but may be expanded in Taylor series about the equilib-
rium position
ys

X  ∂mij 
mij (q1 , q2 , . . .) = mij (0, 0, . . .) + qk + . . .
∂qk 0
k
 
Ph

∂mij
It turns out that the quantities and the higher order
∂qk
0
derivatives are negligibly small so that the coefficients mij s
can be essentially treated as constants having the same values
as they would have in the equilibrium position. Thus around
the equilibrium position,

mij = Tij

[email protected] 15 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

Then the Lagrangian has the following structure:

1X
L=T −V = (Tij q̇i q̇j − Vij qi qj )
2 i,j

ide
Now if you solve the Lagrange’s equation of motion
 
d ∂L ∂L
− =0
dt ∂ q̇k ∂qk

gu
for the given Lagrangian you will get (you need not to do the
general case by hand, you will do in the examples)
X
Vij − ω 2 Tij Ai = 0


ics
i

The equation is a homogeneous equation in Ai s and the con-

dition for existence of the solution is
ys

det Vij − ω 2 Tij = 0

This is equation is called the secular equation. We are only

Ph

interested in real roots of the above equation. ωk0 s determined

from this equation are known as characteristic frequencies or
eigenfrequencies.

[email protected] 16 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

3.1 Matrix formulation of small oscillation

problem
Using symmetry properties of V and T (symmetric matrix)

ide
we can write the secular equation as
X
(Vki − λTki ) Ai = 0
i

gu
where λ = ω 2 . Define a column vector

A=

A1
 A 
 2 

ics
 ··· 


AN

The matrices V and T are given by

ys

   
V11 V12 . . . V1N T11 T12 . . . T1N
 V
 21 V22 . . . V2N   T21 T22 . . . T2N
  
V =  and T = 


 ... ... ... ...   ... ... ... ... 
Ph

VN 1 VN 2 . . . VN N TN 1 TN 2 . . . TN N

So the matrix equation can be written as a matrix equation

V A = ω2T A (3.1)

Note that this equation is not in the form of an eigenvalue

equation as V A = λA but V A = λT A. If T is invertible,(in

[email protected] 17 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

general it is) one can get an eigenvalue equation T −1 V A =

ω 2 IA.
since we have N homogeneous equations, we have N modes,
i.e. N solutions for ω 2 . Let us denote the k -th mode frequency

ide
by ωk2 = λk . Let the vector A corresponding to this mode be
written as  
Ak1
 A 
 k2 
Ak = 

We have
gu  ... 
AkN

V Ak = λk T Ak

ics
Taking conjugate of this equation and changing the index k
to i, we get
Ãi V = λi Ãi T
ys

where we have used à to denote the transpose of the matrix

A.. From V Ak = λk T Ak we get by multiplying with Ãk
Ph

Ãk V Ak
λk =
Ãk T Ak

We have V Ak = λk T Ak . Multiply by Ãi in LHS. You get

Ãi V Ak = λk Ãi T Ak

[email protected] 18 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

We also have Ãi V = λi Ãi T . Multiply by Ak in RHS. You

get
Ãi V Ak = λi Ãi T Ak

ide
From the last two equations

(λk − λi ) Ãi T Ak = 0

Thus, if the eigenvalues are non-degenerate, i.e. if λi 6= λk ,

gu
we get the orthogonality condition

Ãi T Ak = 0
ics
Note that this is different from the orthogonality condition
on eigen vectors for a regular eigenvalue equation. Since (3.1)
does not uniquely determine A, we define normalization con-
dition as
ys

Ãi T Ai = I (3.2)
The normalization of the eigenvectors are not necessary un-
less specifically stated in the problem. We can get the idea
Ph

of normal coordinates without normalization as we see in the

examples. Fundamentally speaking we have a set of coordi-
nate η. We are trying to achieve another set of coordinates
ζ, which are related by

η = Aζ =⇒ ζ = A−1 η

[email protected] 19 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

In the form of ζ the equations look like

ζ¨i + ω 2 ζ = 0

ide
The set of ζ coordinates are called normal coordinates and
these vibrations in the coordinate of ζ are called normal mode.
Example 3.4: Consider two masses m1 = 2m and m2 =
m connected by three springs, as shown in the follow-

gu
ing figure. Find the frequency of oscillation and normal
modes.
ics
Solution: The Lagrangian of the system is
ys

1 1 1 1 1
L = m1 ẋ21 + m2 ẋ22 − k1 x21 − k2 (x2 − x1 )2 − k3 x22
2 2 2 2 2
Ph

Define T and V matrix

! !
∂ 2T m1 0 ∂ 2V 5k −k
T = = ; V = =
∂ ẋi ∂ ẋj 0 m2 ∂xi ∂xj −k 3k

Non trivial solution exist if

det Vik − ω 2 Tik = 0

[email protected] 20 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

which gives
 
 5k − 2mω 2 −k 
=0
 
−k 3k − mω 2

 

ide
The secular equation has the solution ω 2 = 7k/2m and 2k/m.
The normal modes are now found by substituting the values
of ω’s in to the secular equation and evaluation of eigenvectors

(V − ω 2 T )A = 0

Writing A =gu A1
A2
!
, we get, for ω 2 = 7k
2m
ics
! ! ! !
7k 2m 0 A1 5k −k A1
=
2m 0 m A2 −k 3k A2
For this normal mode we have A2 = −2A1 so that we get
unnormalized
ys

!
1
A=
−2
2k
For ω 2 = m, a similar calculation gives
Ph

!
1
A=
1
So we get the unnormalised transformation matrix
!
1 1
A=
−2 1

[email protected] 21 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

We have taken the coordinates as x1 and x2 . The normal

coordinates
η = Aζ =⇒ ζ = A−1 η
! ! !

ide
ζ1 1 −1 x1
=
ζ2 2 1 x2
q q
If we write ω1 = 2m and ω2 = 2k
7k
m , the general solution is

gu
ζ1 = a1 eiω1 t + b1 e−iω1 t ; ζ2 = a2 eiω2 t + b2 e−iω2 t

where a1 , a2 , b1 , b2 are the constants which to be determined

ics
by the initial conditions. The solution can also be written by
in terms of old (x1 , x2 ) coordinates
! ! !
x1 1 1
= c1 cos ω2 t + c2 sin ω2 t
ys

x2 1 1
! !
1 1
+ c3 cos ω1 t + c4 sin ω1 t
−2 −2
Ph

where c1 , c2 , c3 , c4 are the constants which to be determined

by the initial conditions x1 (0), x2 (0), x˙1 (0), x˙2 (0).

[email protected] 22 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

Example 3.5: Three masses m each, initially located

equidistant from one another on a horizontal circle of
radius R. They are connected in pairs by three springs of

ide
force constant k each and of unstretched length 2πR/3.
The spring threads the circular tract so that the mass
is constrained to move on the circle. Find the normal
modes.

gu
ics
ys

Solution: Let θ1 , θ2 and θ3 be the angular displacements

of the three masses from their equilibrium positions. The
Ph

Lagrangian is given by L = T − V, where,

1 2

2 2 2

T = mR θ̇1 + θ̇2 + θ̇3
2
and
1 2h 2 2 2
i
V = kR (θ1 − θ2 ) + (θ2 − θ3 ) + (θ3 − θ1 )
2
[email protected] 23 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

T and V matrices are

   
2 −1 −1 1 0 0
V = kR2  −1 2 −1  T = mR2  0 1 0 
   

ide
−1 −1 2 0 0 1
You can write the secular equation as before and easily show
q
3k
that the normal modes are given by ω = 0 and ω = m
(doubly degenerate).

gu
For ω = 0, the secular equation is


2k −k −k


A1
R2  −k 2k −k   A2  = 0


ics
−k −k 2k A3
which gives A1 = A2 = A3 . The normalized normal coordinate
is given by from AT T A = I. We get
 
ys

1
1
√ 1
 
3mR2
1
Ph

For ω 2 = 3k/m, we have

  
−k −k −k A1
 −k −k −k   A2  = 0
  

−k −k −k A3
which gives A1 + A2 + A3 = 0. The eigenvectors cannot be
uniquely fixed. However, if we use the orthogonality with

[email protected] 24 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

the zero mode, we can take these two choices (2, −1, 1) and
(0, 1, −1) which are orthogonal to the first vector and them-
selves orthogonal. Normalization gives

ide
   
2 0
1 1
√ −1  ; √ 1 
   
 
6mR2 2mR2
−1 −1

3.2
molecule gu
Rigorous solution of Linear triatomic

We now consider the model of linear triatomic molecule. In

ics
the equilibrium configuration of the molecule, two atoms of
mass m are symmetrically located on each side of an atom of
mass M as shown in the figure.
ys
Ph

All three atoms are on one straight line, the equilibrium

distances apart are denoted by b . For simplicity, we shall
first consider only vibrations along the line of the molecule,
and the actual complicated interatomic potential will be ap-
proximated by two springs of force constant k joining the three

[email protected] 25 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

atoms. The positions of three atoms are denoted by x1 , x2 ,

and x3 In these coordinates, the potential energy is

k k
(x2 − x1 − b)2 + (x3 − x2 − b)2
V =

ide
2 2
We now introduce coordinates relative to the equilibrium
positions:

where gu ηi = xi − x0i

x02 − x01 = b = x03 − x02

ics
The potential energy is now
k k
V = (η2 − η1 )2 + (η3 − η2 )2
2 2
or
ys

k 2
η1 + 2η22 + η32 − 2η1 η2 − 2η2 η3


V =
2
Hence the V matrix is
Ph

 
k −k 0
V =  −k 2k −k 
 

0 −k k

The kinetic energy is

m 2
M
T = η̇1 + η̇32 + η̇22
2 2
[email protected] 26 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

So the T matrix is
 
m 0 0
T= 0 M 0 
 

ide
0 0 m
The secular equation is
 
 k − ω2m −k 0 
 
2
 
V − ω T =  −k 2k − ω 2 M −k =0
 

gu 
 0 −k k − ω2m

Evaluate the determinant. You will get the equation



ics
ω 2 k − ω 2 m k(M + 2m) − ω 2 M m = 0

The solutions are

r s  
k k 2m
ω1 = 0. ω2 = , ω3 = 1+
ys

m m M

Now insert the values of the frequencies into the equation of

motion
Ph

V − ω2T A = 0




mω 2 − k a1 + ka2 +0a3 =0


⇒ ka1 + M ω 2 − 2k a2 +ka3 = 0
 


0a1 +ka2 + mω 2 − k a3 = 0
(3.3)

[email protected] 27 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

one by one
Insertion of ω = ω1 = 0 into (3.3) yields a1 = a2 = a3 . The
eigenfrequency ω1 = 0 does not correspond to a vibrational
motion, but represents only a uni- form translation of the

ide
entire molecule: • → ◦ → •
Inserting ω = ω2 = (k/m)1/2 into (3.3) yields a1 = −a3 , a2 =
0; i.e., the central atom is at rest, while the outer atoms vi-
brate against each other: ← • ◦ • →

gu
Inserting ω = ω3 = {k/m(1 + 2m/M )}1/2 into (3.3) yields
a1 = a3 , a2 = −(2m/M )a1 , i.e., the two outer atoms vibrate
in phase, while the central atom vibrates with opposite phase
ics
and with another amplitude: • →← ◦• →

So without normalization the normal coordinates are

ζ1 = η1 + η2 + η3
ys

ζ2 = η1 − η3 (3.4)
2m
ζ3 = η1 − η2 + η3
M
Ph

With the equations

...
 
k k 2k
ζ1 + 0ζ1 = 0; ζ¨2 = ζ2 = 0; ζ¨3 + + ζ3 = 0
m m M
(3.5)
Now we will evaluate the normalized eigenvectors seriously.

ˆ For ω = 0, the equation V A − ω 2 T A = 0 can be written

[email protected] 28 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

as   
k −k 0 A1
 −k 2k −k   A2  = 0
  
0 −k k A3

ide
which gives A1 = A2 = A3 . Taking each of them to be
equal to 1 (in some unit) we get
 
1
1  
A=√ 1

gu N
1
where N is to be fixed by normalization. Normalization
condition gives
ics
  
m 0 0 1
1  
1 1 1  0 M 0  1  = 1
  
N
0 0 m 1
ys

i.e. M + 2m = N. Thus the normalized vector is

 
1
1
Ph

√ 1
 
2m + M
1

ˆ Consider ω 2 = k
m.
For this case, we have
  
0 −k 0 A1
 −k k −k   A2  = 0
  
0 −k 0 A3

[email protected] 29 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

which gives A2 = 0 and A1 = A3 . Using the orthogonality

with the mode ω1 , we take this mode to be
 
1

ide
1 
√  0 

N
−1

Normalization condition
  

N gu
1   m 0 0 1
1 0 −1  0 M 0   0  = 1


0 0 m


−1

ics
fixes N = 2m and the displacement vector is
 
1
1 
√  0 

2m
−1
ys

which shows the central mass to be at rest while the end

masses displace by the same amount in the opposite di-
Ph

rection.
ˆ For ω 2 = k
m + 2k
M, the equation is
  
−2 km
M −k 0 A 1
kM
 −k − m −k   A2  = 0
  

0 −k −2 km
M A3

[email protected] 30 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

which gives A1 = A3 and A2 = − 2m

M A1 . The normalization
can be done as in the previous cases. The displacement
vector is given by

ide
 
r 1
M
−2m/M
 
2mM + 4m2
 
1
We can now get the new set of coordinates. As we have started

gu
with the set of coordinates η and found the normalized eigen-
vectors then the relation between the old set of coordinate
and new set of coordinate is
ζ = A−1 η
ics
η = Aζ =⇒
The A matrix is
     
1 1 r 1
1   1  M
A = √  1 √ 0   −2m/M 
   

2mM + 4m2
2m + M 2m
ys

1 −1 1
 q 
√ 1 √1 M
 2m+M 2m 2mM +4m2 
q
 1 2m
=  √2m+M 0 − M 2 +2mM


Ph

 q 
1 1 M
√
2m+M
− √2m 2mM +4m2

We get the new coordinate set after some tedious calculation

which you need not to do, just see the results.
 √m √ √ 
√ m
√ M √

−1
 2m+M
q 2m+M 2m+M
q 
1
A = 
 2 0 − 12  
M 2m M
√
4m+2M
− √2m+M √
4m+2M

[email protected] 31 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

Verify that AA−1 = I Hence

1 √ √ √ 
ζ1 = √ mη1 + M η2 + mη3
2m + M

ide
r
1
ζ2 = (η1 − η3 )
2 "r #
1 M √
ζ3 = √ (η1 + η3 ) − 2mη2
2m + M 2

gu
The equation of motions are in normalized coordinates

ζ¨1 + ω12 ζ1 = 0; ζ¨2 + ω22 ζ2 = 0; ζ¨3 + ω32 ζ3 = 0

ics
3.3 Making the non quadratic potential quadra

Example 3.6: Consider a particle of mass m moving in

two dimensions in a potential
ys

1 1 1
V (x, y) = − kx2 + λ0 x2 y 2 + λ1 x4 , k, λ0 , λ1 > 0
2 2 4
Ph

(a) At what point (x0 , y0 ) is the particle in stable equi-

librium?
(b) Give the Lagrangian appropriate for small oscilla-
tions about this equilibrium position.
(c) What are the normal frequencies of vibration in (b)?

[email protected] 32 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

Solution: (a) A point where

∂V /∂x = 0, ∂V /∂y = 0, ∂ 2 V /∂x2 > 0, ∂ 2 V /∂y 2 > 0

ide
and
∂ 2V 2 ∂ 2V ∂ 2V
2
(dx) + 2 dxdy + 2
(dy)2 > 0
∂x ∂x∂y ∂y
is a point of stable equilibrium.
p  For the
 p given potential
 we
find two such points, k/λ1 , 0 and − k/λ1 , 0

gu
(b) V is a minimum at a point of stable equilibrium (x0 , y0 ) .
At a neighboring point (x, y), we have, to second order of the
small qualities x − x0 , y − y0 ,
ics
V (x, y) = V (x0 , y0 )
"
∂ 2V
  2 
1 2 ∂ V
+ (x − x 0 ) + 2 (x − x0 ) (y − y0 )
2 ∂x2 x0 ,y0 ∂x∂y x0 ,y0
#
ys

 2 
∂ V
+ 2
(y − y0 )2
∂y x0 ,y0
+ ···
Ph

 q
k
Evaluate the derivatives for the equilibrium point . λ1 , 0
You will get upto second order approximation of x − x0 and
y − y0
 r !2 
2 2
k 1 k kλ0 y 
V (x, y) = − + 2k x − +
4λ1 2 λ1 λ1

[email protected] 33 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

q 
k
Shift the coordinate system to the new origin λ1 , 0 .
r
k
x0 = x − , y0 = y
λ

ide
and take the new origin as the reference level for potential
energy. Then
 
0 0 1 02 λ0 02
V (x , y ) = k x + y

gu
And the Lagrangian is
1
2

1
λ1


λ0
L = T − V = m ẋ02 + ẏ 02 − k x02 + y 02



ics
2 2 λ1
Similarly for the other point of equilibrium, we set
r
k
x00 = x + , y 00 = y
ys

λ
x00 , y 00 replacing x0 , y 0
(c) The secular equation
Ph

 
 2k − mω 2 0 
V − ω 2 T  = 0
 
→ =0
 
kλ0
 0 λ1 − mω 2 

The solutions are

r r
2k kλ0
ω1 = , ω2 =
m mλ1
[email protected] 34 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

These are the normal angular frequencies for small oscillations

of the system, about both of the points of equilibrium.
Example 3.7: Two mass points of equal mass m are

ide
connected to each other and to fixed points by three
equal springs of force constant k, as shown in the dia-
gram. The equilibrium length of each sprint is a . Each
mass point has charge +q, and they repel each other ac-
cording to Coulomb law. Set up the secular equation for

gu
the eigenfrequencies.
ics
Solution: This problem also has two generalized coordi-
ys

nates x1 and x2 , the coordinates of the two masses starting

from left. x10 = a and x20 = 2a are the initial positions of the
masses. Let η1 and η2 be the displacements from the equi-
librium positions. (η1 = x1 − x10 , η2 = x2 − x20 ) So kinetic
Ph

energy operator will be

1 1
T = mη̇12 + mη̇22
2 2

[email protected] 35 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

The T matrix will be

!
m 0
T =⇒
0 m

ide
As far as potential energy is concerned, besides the contri-
bution from the three springs, we will have one extra contri-
bution due to the Coulomb repulsion between the two charges
1 1 1 q2
V = kη12 + k (η2 − η1 )2 + kη22 +

=
gu
2
1
2
2
2 2
2


2kη1 − 2kη1 η2 + 2kη2 +
4π0 |x1 − x2 |
q2
4π0 |η2 − η1 − a|
ics
Because |η2 − η1 | /a  1, we obtain by expanding the
last term in the powers of |η2 − η1 | /a, and retaining up to
quadratic terms
−1
q2

1 |η2 − η1 |
ys

2 2


V = 2kη1 − 2kη1 η2 + 2kη2 + 1−
2 4π0 a a
1 2 2

q2 q 2 |η2 − η1 |
= 2kη1 − 2kη1 η2 + 2kη2 + +
2 4π0 a 4π0 a2
Ph

q 2 (η2 − η1 )2
+ + ···
4π0 a3
Noting that constant terms and the terms which are linear in
displacement coordinates do not contribute to the potential
energy matrix, we have
2q 2 2q 2 2q 2
      
1
V = 2k + η12 −2 k+ η1 η2 + 2k + η22
2 4π0 a3 4π0 a3 4π0 a3

[email protected] 36 Physicsguide
©Sk Jahiruddin, 2020 Small Oscillation

Hence the V matrix is

    
q2 2q 2
2 k + a3 − k + a3

ide
V =  2q 2
 
q2
 
− k + a3 2 k + a3

So we note that the presence of charges on the two masses

have effectively modified the force constants of the springs.

gu
The problem can be easily solved by setting up the secular


equation det V − ω 2 T = 0
ics
ys
Ph

[email protected] 37 Physicsguide