CHM 305 - Lecture 2 - The Classical Wave Equation

Prof. Marissa Weichman

September 7, 2021

In this lecture we will review and solve the classical wave equation, a partial differential equation
that governs the spatial and temporal behavior of waves. This classical wave equation is going to
provide a conceptual background for Schrödinger’s “quantum wave equation,” which governs the
wavefunctions of quantum mechanical matter, and which we will solve in subsequent lectures. While
classical, today’s lecture will also introduce some math that will be quite useful over the next few
lectures. This material is covered in Chapter 2 of McQuarrie.

1 The Vibrating String

The wavefunction of a classical wave describes how a wave’s amplitude evolves in space and time.
Imagine taking a piece of rope in your hand and shaking it, so waves travel down its length. We
might write down the following wavefunction to describe the propagation of these waves:
  
x t
ψ(x, t) = A sin 2π − (1)
λ T

where x is the distance along the direction of wave propagation, λ is the wavelength, and T = 1/ν
is the period.
We will also use an equivalent expression, which is sometimes more convenient, written as:

ψ(x, t) = A sin (kx − ωt) (2)

where k = 2π/λ is called the wave vector, and ω = 2π/T = 2πν is the angular frequency.
Let’s plot various snapshots of this wave to make some sense of it:

These are snapshots of the spatial structure of ψ(x, t) for special values of t. Looking at these
snapshots collectively, the wave appears to propagate in time towards the positive end of the x
axis. By t = T /2 the wave is 180o out of phase from where it started at t = 0, and by t = T it will
fully return to its state at t = 0.

1
 We can also consider just one point in space, and think about the sinusoidal wave amplitude at
that location over time, as shown above for x = 0.
This is a “traveling wave,” meaning that the node positions move in time. We can solve for the
node positions (e.g. the positions where ψ(x, t) = 0) as a function of time by setting Eqn. 1 to zero:
  
x t
ψ(x, t) =A sin 2π − =0 (3)
λ T
 
x t
∴ 2π − = nπ (4)
λ T
nλ λ
→ x= + ·t (5)
2 T
≡ x0 + v · t (6)

The nodes are therefore moving with a positive velocity along the x axis of:
λ ω
v= = (7)
T k
It’s simple to write down the wavefunction for a wave traveling in the opposite direction, with
a negative velocity, simply by switching the sign of either the k or ω term in Eqn. 2:

ψ(x, t) = A sin (kx + ωt) (8)

It’s also important to note that we will sometimes wavefunctions written in an equivalent
complex exponential form. Euler’s relation tells us that:

eiθ = cos θ + i sin θ (9)

eiθ + e−iθ eiθ − e−iθ

cos θ = , sin θ = (10)
2 2i
So a general expression for a wave traveling to the right can also be written as:

ψ(x, t) = Bei(kx−ωt) + Ce−i(kx−ωt) (11)

This notation will become especially useful in quantum mechanics where wavefunctions are complex.

2
2 The Wave Equation in One Dimension
Classical waves are governed by a partial differential equation that relates the spatial structure to
their temporal structure. For waves traveling through a one-dimensional material (like a piece of
rope), the relevant wave equation is:

∂ 2 ψ(x, t) 1 ∂ 2 ψ(x, t)
= · (12)
∂x2 v2 ∂t2
where v is the speed with which a disturbance or displacement travels down the rope. We make
the assumption that this speed v is the same for all waves, regardless of their frequency.
Eqn. 12 is referred to as the classical non-dispersive wave equation. It is a linear partial differen-
tial equation because ψ(x, t) and its derivatives only appear to the first power. A linear differential
equation has the important property that any linear combinations of its solutions will also be a
solution. For instance, if ψ1 (x, t) and ψ2 (x, t) are solutions, then aψ1 (x, t) + bψ2 (x, t) is also a
solution. This will become crucial!
Let’s check that the general wave from Eqn. 11 satisfies the wave equation:

∂ 2 ψ(x, t) ∂ 2 h i(kx−ωt) −i(kx−ωt)

i
LHS: = Be + Ce (13)
∂x2 ∂x2
=(ik)2 Bei(kx−ωt) + (−ik)2 Ce−i(kx−ωt) (14)

= −k 2 ψ(x, t) ✓ (15)

1 ∂ 2 ψ(x, t) 1 ∂ 2 h i(kx−ωt) −i(kx−ωt)

i
RHS: = Be + Ce (16)
v2 ∂t2 v 2 ∂t2
1 h i
= 2 (−iω)2 Bei(kx−ωt) + (iω)2 Ce−i(kx−ωt) (17)
v
ω2
=− ψ(x, t) = −k 2 ψ(x, t) ✓ (18)
v2
Where in the last step we used v = ω/k from Eqn. 7.

3 Waves with Boundary Conditions

Let’s now imagine that instead of just a piece of rope in our hands, we have a length of rope of
length L pinned down at its ends - like a jump rope, or a guitar string. You probably have some
intuition for what the behavior of this piece of rope will be like, but let’s work through a solution
to the wave equation to get there.
In particular, we want to solve Eqn. 12 for ψ(x, t) with the conditions that

ψ(0, t) = 0 (19)
ψ(L, t) = 0 (20)

for all values of t, as shown below.

3
 We are going to start our solution here by assuming that ψ(x, t) will be a standing wave, not a
traveling wave. The wavefunctions of standing waves are separable, meaning that they factor into
two functions of purely x or t:
ψ(x, t) = X(x) · T (t) (21)
We can therefore think of our standing wave as one spatial envelope, with fixed nodes determined by
locations where X(x) = 0, and whose total intensity oscillates in time according to T (t). Assuming
that our solution will take on this form seems like a major constraint, and that we will only recover
a small subset of possible solutions! However, it will turn out that our standing wave solutions will
form a basis set of so-called normal modes, from which we can patch together any solution we like.
All solutions to this problem can be described in terms of linear combinations of standing waves –
all because our wave equation is linear!
So let’s get started. Plugging Eqn. 21 into Eqn. 12:

∂2 1 ∂2
X(x)T (t) = X(x)T (t) (22)
∂x2 v 2 ∂t2
∂ 2 X(x) 1 ∂ 2 T (t)
= T (t) · = X(x) · (23)
∂x2 v2 ∂t2
1 ∂ 2 X(x) 1 1 ∂ 2 T (t)
→ · = · (24)
X(x) ∂x2 v 2 T (t) ∂t2
The left-hand side of Eqn. 24 is now purely a function of x, while the right-hand side is purely a
function of t. The only way that equality between the two halves can be preserved for all values
of x and t is if both sides are equal to a constant independent of x and t, which we will call
the “separation constant” K. It’s worth pausing for a moment to make sure this makes sense:
both sides must be constants, otherwise, for instance, we could change t to alter the LHS without
touching the RHS and ruin the equality. Therefore:

1 ∂ 2 X(x) ∂2X
· =K → − KX = 0 (25)
X(x) ∂x2 ∂x2

1 1 ∂ 2 T (t) ∂2T
· =K → − Kv 2 T = 0 (26)
v 2 T (t) ∂t2 ∂t2
We now have a much simpler problem to solve: two ordinary differential equations that are func-
tions of just one independent variable. We should also already have a good idea for what their
solutions should look like. In particular, we know that sines and cosines (or equivalently, complex
exponentials) are proportional to their own second derivatives.

4
 Let’s take a stab at solving Eqn. 25 for X(x) first. We will make the educated guess that:

X(x) = A sin(βx) + B cos(βx) (27)

We therefore find:
d2
[A sin(βx) + B cos(βx)] − K [A sin(βx) + B cos(βx)] = 0 (28)
dx2
= −β 2 [A sin(βx) + B cos(βx)] − K [A sin(βx) + B cos(βx)] = 0 (29)

→ −β 2 = K (30)

Since we would like to require that β be real, this implies that K < 0. Otherwise, our guess for
X(x) is a perfectly valid solution to our differential equation in Eqn. 25.
Let’s consider our boundary conditions now, from Eqns. 19 and 20.

X(0) = 0 → A sin(0) + B cos(0) = 0 (31)

A·0+B·1=0 (32)
B=0 (33)

X(L) = 0 → A sin(βL) = 0 (34)

βL = nπ, n = 1, 2, . . . (35)
nπ
β= (36)
L

Putting this all together, we now have

 nπx 
Xn (x) = A sin , n = 1, 2, . . . (37)
L
We can now turn our attention to solving Eqn. 26 for the time-dependent part given by T (t).

d2 T d2 T d2 T  nπv 2
− Kv 2 T = + β 2v2T = + T = 0 (38)
dt2 dt2 dt2 L
where we’ve used our results from Eqns. 30 and 36.
We should now quickly recognize the solutions to differential equations of this form, and we can
just write down the answer:
 nπv   nπv 
Tn (t) = C sin t + D cos t , n = 1, 2, . . . (39)
L L
≡ C sin(ωn t) + D cos(ωn t) (40)

Where we have defined, for simplicity’s sake:

nπv
ωn = (41)
L

5
So finally, our complete time- and position-dependent standing wave solutions are:
ψn (x, t) = Xn (x)Tn (t) (42)
 nπ 
= A sin x [C sin(ωn t) + D cos(ωn t)] (43)
L
 nπ 
= A sin x sin(ωn t + ϕ) (44)
L
where ϕ is an arbitrary phase that captures what the state of the system was at t = 0, since that
was not specified by our boundary conditions.
We can now explore a few implications of our results.
(i) The spatial and temporal structure of standing waves are closely related. Take a look at the
first few standing waves with n = 1, 2, 3, as shown below. The standing wave with index n,
must have n − 1 spatial nodes and oscillate in place with frequency ωn which is proportional
to n. The more spatial oscillations along the length of the rope, the faster it must oscillate.
This might be familiar if you have played any stringed musical instruments!
(ii) There are an infinite series of standing waves. We call the first, node-less wave the “funda-
mental,” while higher-order waves are “harmonics” or “overtones.”
(iii) Standing waves form a basis set to construct arbitrary spatial waveforms. Because our differ-
ential equations are linear, we can take linear combinations, or “superpositions,” of our ψn
solutions to construct arbitrary solutions:
∞
X ∞
X  nπ 
Φ(x, t) = An ψn (x, t) = An sin x sin(ωn t + ϕ) (45)
L
n=1 n=1

It turns out that you can construct any arbitrary function over the window x ⊂ [0, L] by
choosing the appropriate coefficients An . To extend the musical analogy, one can think of
these superposition states as “chords.” The use of an infinite collection of sinusoidal functions
as a basis set to construct other functions may be familiar if you have seen Fourier analysis
in other coursework.
(iv) Classical standing waves are a very good analogy for solutions to the quantum wave equation.
Keep this idea in mind as we start to think about Schrödinger’s equation in the next class.
Quantum objects have states called “eigenfunctions” which behave exactly like standing waves
– they have separable temporal and spatial behavior. And just like standing waves form a
basis set to describe more complex phenomena, quantum eigenstates serve as a basis set to
describe how quantum systems evolve over time.