# Sturm-Liouville theory applied to solve Helmholtz equation

1. Oct 27, 2012

### JordanGo

1. The problem statement, all variables and given/known data

Find the eigenfunctions of the Helmholtz equation:
$\frac{d^2y}{dx^2}+k^2y = 0$

with boundary conditions:
$y(0)=0$
$y'(L)=0$

2. Relevant equations

General Solution:
$y = Asin(kx) + Bcos(kx)$

3. The attempt at a solution

I found that at y(0) that B=0 and that at y'(L) that Akcos(kL)=0.
What do I do with these results?

2. Oct 27, 2012

### LCKurtz

Since you already have B=0, you don't want to have A = 0 to avoid a trivial solution. So you need cos(kL) = 0. What values of kL will do that? That will determine the values of k for which the equation has non-trivial solutions. You may also need to check your original equation when k = 0 to see if that gives an eigenvalue too.

3. Oct 27, 2012

### JordanGo

Ok, so to avoid trivial solutions, we can say that:
$k=\frac{pi}{2L} , k=\frac{3pi}{2L}$
Within the range of [0,2*pi].

But how is knowing a value of k going to help find the eigenfunctions?
If the values of k are the eigenvalues, how do I use them to find the eigenfunctions?

4. Oct 27, 2012

### HallsofIvy

Staff Emeritus
You have already said that B= 0. In order that the solution not be trivial, you must have $k= \pi/(2L)$ or $k= 3\pi/(2L)$. That means that the non-trivial solutions are $Asin(\pi x/(2L))$ and $Asin(3\pi x/(2L))$. Those are the eigenfunctions.

5. Oct 27, 2012

### LCKurtz

And, of course, there is no reason to limit the $kL$ to $[0,2\pi]$. In fact, you best not do that if you want all the eigenvalues and eigenfunctions. And don't forget to check the system when $k=0$.

6. Oct 28, 2012

### JordanGo

Thanks a lot for the help! You guys are great!

One more thing, when you say to check the equation when k=0, you mean use the general equation (knowing A=0) and say that y=B is a solution?

7. Oct 28, 2012

### LCKurtz

I mean that $y'' + k^2 y = 0$ becomes $y''=0$ with your B.C.s $y(0)=0,\, y'(L)=0$. You don't want to forget to check that because sometimes you get a nontrivial solution which isn't included in the sine-cosine case.

8. Oct 29, 2012

### JordanGo

so when k is zero, we know that y'(L)=0 and y''(x)=0, thus y=C , which is another constant?

9. Oct 29, 2012

### LCKurtz

What is the general solution to y''(x)=0? Make it satisfy y(0)=0 and y'(L)=0.

10. Oct 29, 2012

### JordanGo

Is this Laplace's equation? We haven't seen how to solve it yet... Can you help me? I've searched the internet and it shows plenty of ways to solve it in different coordinate systems, not really sure what to do with it...

11. Oct 29, 2012

### LCKurtz

Are you talking about y''(x)=0? It is just a simple calculus problem. Integrate it twice.

12. Oct 29, 2012

### JordanGo

ok, so the solution to y''=0, y(0)=0 and y'(L)=0 is that either y=0 or y=C (some constant)

13. Oct 29, 2012

### LCKurtz

No comment until you show your work.

14. Oct 29, 2012

### JordanGo

When k=0, the general solution is: y(x)=Ax+B.
First boundary condition says y(0)=0=B, thus B=0, and y'(L)=0=A, thus A=0.
So k=0 is not an eigenvalue.

15. Oct 29, 2012

### LCKurtz

Good. So $y = C \ne 0$ is not an option.