1 IIT Delhi - CML 100:2 – Waves

u(x,t
Classical waves )

Figure 2.1: A vibrating string

A linear partial differential equation

𝜕 2 𝑢(𝑥, 𝑡) 1 𝜕 2 𝑢(𝑥, 𝑡)
= 2
𝜕𝑥 2 𝑣 𝜕𝑡 2

Boundary conditions

𝑢(0, 𝑡) = 0 and 𝑢(𝑙, 𝑡) = 0 at all times

Separation of variables: A technique used when the two variables are independent

𝑢(𝑥, 𝑡) = 𝑋(𝑥)𝑇(𝑡)

which gives

1 𝑑2 𝑋(𝑥) 1 𝑑2 𝑇(𝑡)
=
𝑋(𝑥) 𝑑𝑥 2 𝑣 2 𝑇(𝑡) 𝑑𝑡 2

Since LHS is only dependent on the position and RHS on time they must be equal to a constant, 𝐾

𝑑2 𝑋(𝑥) 𝑑2 𝑇(𝑡)
− 𝐾𝑋(𝑥) = 0 𝑎𝑛𝑑 − 𝐾𝑣 2 𝑇(𝑡) = 0
𝑑𝑥 2 𝑑𝑡 2

These are linear differential equations with constant coefficients.

Solutions 𝐾 = 0, 𝐾 < 0, 𝐾 > 0

For 𝐾 = 0, the solution are trivial – no use

For 𝐾 > 0, its a subset of the solution for 𝐾 < 0

For 𝐾 < 0

𝑑2 𝑦
Lets rewrite the equation in this form 𝑑𝑥 2 + 𝑘 2 𝑦 = 0 where 𝐾 = −𝛽 2

We need a solution that when differentiated twice gives back the same function. Lets try 𝑦 = 𝑒 𝛼𝑥

This gives, (𝛼 2 + 𝛽 2 ) 𝑦(𝑥) = 0

i.e. 𝛼 = ± 𝑖 𝛽

The general solution is then

2 IIT Delhi - CML 100:2 – Waves

𝑦(𝑥) = 𝑐1 𝑒 𝑖𝛽𝑥 + 𝑐2 𝑒 −𝑖𝛽𝑥

Using Euler’s formula,

𝑒 ±𝑖𝜃 = 𝑐𝑜𝑠𝜃 ± 𝑖 sin 𝜃

we get

𝑦(𝑥) = 𝐴 cos 𝛽𝑥 + 𝐵 sin 𝛽𝑥

See for yourself that the case of 𝐾 > 0 is a special case of the solutions for 𝐾 < 0.

Boundary conditions: See the two equations for 𝑋(𝑥)𝑎𝑛𝑑 𝑇(𝑡)

𝑋(0) = 0 implies 𝐴 = 0

𝑋(𝑙) = 0 implies 𝑋(𝑙) = 𝐵 sin 𝛽𝑙 = 0 𝐵 = 0 is trivial. So, sin 𝛽𝑙 = 0 gives

𝛽𝑙 = 𝑛𝜋 𝑛 = 1,2,3 …

𝑛 = 0 is not a solution because the wave does not exist.

𝑛𝜋𝑥
𝑋(𝑥) = 𝐵 sin
𝑙

Also, for 𝑇(𝑡)

𝑑2 𝑇(𝑡)
+ 𝛽 2 𝑣 2 𝑇(𝑡) = 0
𝑑𝑡 2

The general solution (remember 𝛽 = 𝑛𝜋/𝑙) is

𝑇(𝑡) = 𝐷 cos 𝜔𝑛 𝑡 + 𝐸 sin 𝜔𝑛 𝑡

So the amplitude of the wave 𝑢 is given by (it now depends on 𝑛)

𝑛𝜋𝑥
𝑢𝑛 (𝑥, 𝑡) = 𝑋(𝑥)𝑇(𝑡) = (𝐵𝑛 sin ) (𝐷𝑛 cos 𝜔𝑛 𝑡 + 𝐸𝑛 sin 𝜔𝑛 𝑡)
𝑙
𝑛𝜋𝑥
= (𝐹𝑛 cos 𝜔𝑛 𝑡 + 𝐺𝑛 sin 𝜔𝑛 𝑡) sin
𝑙

Superposition

As each 𝑢𝑛 (𝑥, 𝑡) is a solution to the linear differential equation, so is any sum of the 𝑢𝑛 (𝑥, 𝑡) ′𝑠.

A most general solution is

∞
𝑛𝜋𝑥
𝑢(𝑥, 𝑡) = ∑(𝐹𝑛 cos 𝜔𝑛 𝑡 + 𝐺𝑛 sin 𝜔𝑛 𝑡) sin
𝑙
𝑛=1

or
3 IIT Delhi - CML 100:2 – Waves

∞
𝑛𝜋𝑥
𝑢(𝑥, 𝑡) = ∑ 𝐴𝑛 cos(𝜔𝑛 𝑡 + 𝜙𝑛 ) sin
𝑙
𝑛=1

where 𝐴 is the amplitude and 𝜙 the phase angle. No matter how the string is plucked its shape will
evolve according to the above equations.

Each 𝑢𝑛 (𝑥, 𝑡) is called a normal mode. The time dependence of each normal mode represents
harmonic motion of frequency 𝜈𝑛 = 𝜔𝑛 /2𝜋 = 𝑣𝑛/2𝑙 (since 𝜔𝑛 = 𝑛𝜋𝑣/𝑙 )

Figure 1.2: The first few modes of vibration

First harmonic = fundamental mode, frequency 𝑣/2𝑙

Second harmonic = first overtone, frequency 𝑣/𝑙

Mid-point of the second harmonic does not change with time. Its fixed at zero. This is a node
which you will also encounter in quantum mechanics. 𝑥 = 0 and 𝑥 = 𝑙 are not nodes – they are
boundary conditions.

These are standing waves.

Add up the first two harmonics, phase shifted by 90°

𝜋𝑥 1 𝜋 2𝜋𝑥
𝑢(𝑥, 𝑡) = cos 𝜔1 𝑡 sin + cos (𝜔2 𝑡 + ) sin
𝑙 2 2 𝑙

Some work: sketch the travelling wave.

4 IIT Delhi - CML 100:2 – Waves

Schrödinger Equation
Lets start with the classical wave equation

𝜕2𝑢 1 𝜕2𝑢
=
𝜕𝑥 2 𝑣 2 𝜕𝑡 2

which as we have seen above can be solved to give

𝑢(𝑥, 𝑡) = 𝜓(𝑥) cos 𝜔𝑡

𝜓(𝑥) is called the spatial amplitude of the wave. This gives,

𝑑2 𝜓 𝜔2
+ 𝜓(𝑥) = 0
𝑑𝑥 2 𝑣2

Since 𝜔 = 2 𝜋𝜈 and 𝜈𝜆 = 𝑣,

𝑑2 𝜓 4𝜋 2
+ 2 𝜓(𝑥) = 0
𝑑𝑥 2 𝜆

Now the total energy

𝑝2
𝐸 = 𝐾𝐸 + 𝑃𝐸 = + 𝑉(𝑥)
2𝑚

and so

𝑝 = {2𝑚[𝐸 − 𝑉(𝑥)]}1/2

Use de Broglie relation 𝜆 = ℎ/𝑝. This is where quantum mechanics comes in

ℎ ℎ
𝜆= = 1
𝑝
{2𝑚[𝐸 − 𝑉(𝑥)]}2

and we get,

𝑑2 𝜓 2𝑚
+ [ 𝐸 − 𝑉(𝑥)] 𝜓(𝑥) = 0
𝑑𝑥 2 ℏ2

or

ℏ2 𝑑2 𝜓
− + 𝑉(𝑥)𝜓(𝑥) = 𝐸 𝜓(𝑥)
2𝑚 𝑑𝑥 2

This is the time-independent Schrödinger equation

In this course we will not worry about the time-dependent Schrödinger equation
5 IIT Delhi - CML 100:2 – Waves

Operators
An operator operates – it does something. For example, an “turn 90° left” is an operator that tells us
to turn left by 90°. Another example could be “walk five paces ahead”.

Mathematical operators tell us to perform a mathematical operation on a function (𝑓(𝑥) to give

another function 𝑔(𝑥).

Some examples of mathematical operators

1
INTEGRATE: ∫0 𝑓(𝑥) = 𝑔(𝑥)

2
SQR: (𝑓(𝑥)) = 𝑔(𝑥)

𝑑
DIFFERENTIATE: 𝑑𝑥 𝑓(𝑥) = 𝑔(𝑥)

In general we can denote an operator using a hat on it,

𝐴̂ 𝑓(𝑥) = 𝑔(𝑥)

Operators and quantum mechanics

In quantum mechanics, we encounter only linear operators. This is one of the postulates of QM
which we will discuss later.

𝐴̂ [𝑐1 𝑓1 (𝑥) + 𝑐2 𝑓2 (𝑥)] = 𝑐1 𝐴̂(𝑓1 (𝑥) + 𝑐2 𝐴̂( 𝑓2 (𝑥)

Here 𝑐1 and 𝑐2 can be complex numbers.

Differentiate and integrate are linear. Squaring is non-linear.

Operators may not commute like numbers, i.e. 𝐴̂ 𝐵̂ 𝑓(𝑥) is not necessarily equal to 𝐵̂𝐴̂ 𝑓(𝑥). As an
example consider the case of a person walking five paces and turning 90°.

Eigenfunctions and Eigenvalues

A function that gets operated and results in the same function apart from a multiplicative factor is
an eigenfunction of the operator

𝐴̂ 𝑓(𝑥) = 𝑎 𝑓(𝑥)

Finding the eigenfunction of the operator and the eigenvalue is called an eigenvalue problem.

The Schrödinger equation can be written as

ℏ2 𝑑2 𝜓
[− + 𝑉(𝑥)] 𝜓(𝑥) = 𝐸 𝜓(𝑥)
2𝑚 𝑑𝑥 2

or

̂ 𝜓(𝑥) = 𝐸 𝜓(𝑥)
𝐻
6 IIT Delhi - CML 100:2 – Waves

where 𝐻̂ is called the Hamiltonian operator and the eigenvalue is the energy. So there is a
correspondence between the operator and a measurable. Such correspondences between operators
and classical-mechanical variables are fundamental to the formalism of QM.
2 2
̂ = − ℏ 𝑑 𝜓2
̂ = 𝑉(𝑥) ∴ 𝐾𝐸
Since the energy is KE + PE, and 𝑃𝐸 2𝑚 𝑑𝑥

𝑝2 𝑑2 𝜓
𝐾𝐸 = ⇒ 𝑝̂ 2 = − ℏ2 2
2𝑚 𝑑𝑥

𝑑2 𝜓 𝑑
∴ 𝑝̂ . 𝑝̂ = − ℏ2 2
𝑜𝑟 𝑝̂ = −𝑖ℏ (the minus sign is needed for the correct direction)
𝑑𝑥 𝑑𝑥

Probability
Discrete Events

An experiment has 𝑛 possible outcomes, each with probability 𝑝𝑗 . We perform the experiment a
large number of times (ideally infinite number of times)

𝑁𝑗
𝑝𝑗 = lim 𝑗 = 1, 2, 3 … . . 𝑛
𝑛→ ∞ 𝑁

0 ≤ 𝑝𝑗 ≤ 1 and Σ𝑝𝑗 = 1 (normalization)

Suppose we get a value 𝑥𝑗 at the 𝑗 𝑡ℎ experiment, then the average is defined as

〈𝑥〉 = Σ 𝑥𝑗 𝑝𝑗 = Σ 𝑥𝑗 𝑝(𝑥𝑗 )

Second momemt
𝑛

〈𝑥 2 〉 = ∑ 𝑥𝑗2 𝑝𝑗
𝑗=1

Second central moment or variance

𝑛
2
𝜎𝑥2 = 〈(𝑥 − 〈𝑥〉)2 〉 = ∑(𝑥𝑗 − 〈𝑥〉) 𝑝𝑗 = 〈𝑥 2 〉 − 〈𝑥〉2
𝑗=1

𝜎𝑥 is called the standard deviation.

Continuous distributions

𝑝𝑟𝑜𝑏 (𝑥, 𝑥 + 𝑑𝑥) = 𝑝(𝑥)𝑑𝑥

𝑏

𝑝𝑟𝑜𝑏 (𝑎 ≤ 𝑥 ≤ 𝑏) = ∫ 𝑝(𝑥)𝑑(𝑥)
𝑎
7 IIT Delhi - CML 100:2 – Waves

Normalization condition
∞

∫ 𝑝(𝑥)𝑑(𝑥) = 1
−∞

Average and standard deviation

∞

〈𝑥〉 = ∫ 𝑥 𝑝(𝑥)𝑑(𝑥)
−∞

〈𝑥 2 〉 = ∫ 𝑥 2 𝑝(𝑥)𝑑(𝑥)
−∞

𝜎𝑥2 = ∫ (𝑥 − 〈𝑥〉)2 𝑝(𝑥)𝑑(𝑥)

−∞

Quantum mechanics and probability

If we restrict the particle in a certain region, then the probability of finding the particle in this region
is one. Outside this region the particle does not exist. Since the intensity of a wave is the square of
the magnitude of the amplitude, mathematically we say 𝜓 ∗ (𝑥)𝜓(𝑥)𝑑𝑥 is the probability that the
particle is located between 𝑥 and 𝑥 + 𝑑𝑥.