PHYS 312, Worksheet for Lecture 3

1. Solve the problem f 00 (θ) = −f (θ), f (0) = 1, f 0 (0) = 0 using an exponential ansatz
f (θ) = ekθ .

Substitute the ansatz into the equation to obtain the auxiliary equation k 2 = −1. The
solutions to this equation are k = ±i. The general solution is therefore f (θ) = c1 eiθ + c2 e−iθ .
Substitute initial conditions: f (0) = c1 + c2 = 1 and f 0 (0) = ic1 − ic2 = 0. The second
iθ −iθ
equation gives c1 = c2 and then the first gives c1 = c2 = 21 . Therefore f (θ) = e +e2
.

2. Find the most general solution to the wave equation

∂2 ∂2
c2 u − u=0
∂x2 ∂t2
for u(x, t) on the region 0 ≤ x ≤ L and −∞ < t < ∞, with boundary conditions u(0, t) =
u(L, t) = 0. Hint: The most general solution is simply a sum of all possible solutions of the
form f (x)g(t).
As a first step, make a plan.

The plan is 1. separate variables, 2. solve each separate equation, 3. put together solu-
tions to the PDF by multiplying together solutions from step 2, 4. add all possible solutions
with arbitrary coefficiens together.

Solution: Separating the equation using u = f (x)g(t), we obtain

f 00 (x) = Af (x) and g 00 (t) = c2 Ag(t) .

In the x direction, there is a boundary condition which implies that f (0) = f (L) = 0. We
are going to have to check all three cases! √ √
− Ax Ax
Case 1: A > 0. The general solution √
is f (x)
√
= c 1 e + c 2 e . Boundary conditions:
− AL AL
f (0) = c1 + c2 = 0 and f (L) = c1 e + c2 e = 0. The first equation gives c2 = −c1 ,
substitute into the second: √ √
c1 e− AL − c1 e AL = 0
 √ √ 
c1 e− AL − e AL = 0

Either c1 = 0, which implies that f (x) = 0, which is a trivial solution we are not interested
in, or √ √
e− AL − e AL = 0
√ √
e AL
= e− AL

√
=1 e2 AL

√
Since A and√L are real and positive, we can take a logarithm of both sides to solve, which
implies that 2 AL = ln 1 = 0. That can’t be right, so there is no nonzero solution for this
case!
Case 2: A = 0. The general solution is f (x) = c1 + c2 x, whose graph is a straight line.
To go through zero at both x = 0 and x = L, the line must be given by f (x) = 0. Again,
that is a trivial solution we are not interested in.
Case 3: A < 0. It’s easier to p do case 3 inp the trigonometric function approach. The
general solution is f (x) = c1 sin |A|x + c2 cos |A|x. Then,

f (0) = 0 ⇒ c1 sin 0 + c2 cos 0 = c2 = 0

and p p  nπ 2
f (L) = 0 ⇒ c1 sin |A|L = 0 ⇒ |A|L = nπ ⇒ A = −
L
The final answer in case 3 is then f (x) = a sin πnx/L for any integer n. Moreover, negative
integers lead to the same answer as positive ones: a sin π(−n)x/L = −a sin πnx/L and n = 0
leads to f (x) = 0, so n can actually be taken to be a positive integer.

nπ 2
fn = an sin πnx


L
, An = − L
n = 1, 2, 3, . . .

You should be familiar with this formula as the standing wave!

Digression: We can also get there using the √ complex √ exponentials. It is instructive, so let’s go
−i |A|x i |A|x
over it. The general solution is f (x) = c1 e + c2 e . Repeating the calculation for case
3, we arrive at √
e2i |A|L = 1

We cannot now take a logarithm to solve the equation (unless you understand that the logarithm
is a multi-valued function in the complex plane). Let’s rewrite in terms of real and imaginary
components:  p   p 
cos 2 |A|L + i sin 2 |A|L = 1

Thereal and imaginary parts of the above equation must each be equal. The real part implies that
p p
cos 2 |A|L = 1, so 2 |A|L = 2πn for n an integer. With this condition, the imaginary part is
 p 
automaticaly satisfied, sin 2 |A|L = sin 2πn = 0. Solving for A:
p
2 |A|L = 2πn
 πn 2
A=
L
Substituting it all back in,
√ √
f (x) = c1 e−i |A|x + c2 ei |A|x = c1 e−iπnx/L − c1 eiπnx/L = 2ic1 sin πnx/L

For a real solution, c1 = a/(2i) where a is a real constant, at which point we obtain the answer we
had before. If we don’t require a real solution, c1 can be anything.
That was a rather messy way to do it.
 Since we know the separation constant, the g equation is a snap,
 nπc 2
g 00 (t) = c2 An g(t) = − g(t)
L
with solutions    
nπct nπct
gn (t) = bn,1 cos + bn,2 sin
L L
These are written in the obviously real form, but complex exponentials are ok as well (see
below).
Putting things together, we have solutions of the form
    
nπct nπct πnx
un = bn,1 cos + bn,2 sin sin
L L L

The general solution is a sum! There are many ways to write it,
∞     
X nπct nπct πnx
u(x, t) = bn,1 cos + bn,2 sin sin
n=1
L L L
∞ 
X nπct nπct
 πnx
u(x, t) = an,+ ei L + an,− e−i L sin
n=1
L
∞   
X nπc(t − Tn πnx
u(x, t) = cn cos sin
n=1
L L
an,± , bn,1 , bn,2 , cn and Tn are various arbitrary coefficients. These can be determined if
initial conditions are known.