# How does Quantum Tunneling work?

1. May 3, 2006

### simon009988

When something tunnels, is it just because of the fact that the wavefunction just decided to randomly pick a place that would be disallowed by classical physics? or is there more to it?

2. May 3, 2006

### Pengwuino

In a lamens sense (because that's all i know :rofl: :rofl: :rofl: ), a particle can tunnel through a potential barrier because there is always an uncertainty of it's energy and position. Theres a calculable probability that the particle will tunnel through the barrier because we can't be precisely certain what the particles energy will be. If the potential barrier was higher then the particle's energy + uncertainty, it becomes an infinite potential well and can't tunnel.

3. May 3, 2006

### ZapperZ

Staff Emeritus
Er... no, this is not correct. If you look at the simple 1D tunneling treatment, there's nothing here that signifies "uncertainty", because you can do this for a single particle. Furthermore, in balistic/elastic tunneling, the energy of the particle is preserved as it leaves the barrier. So its energy is well-defined enough, especially when we're talking about a free particle.

A quantum particle can tunnels because it isn't described in the classical sense. It is described via the QM wavefunction in which, when you apply the boundary conditions at the two edges and inside the barrier, you will find that the wavefunction can "leak through" the barrier and continues to the other side. This gives a finite and non-zero probability of the existence of the particle across the barrier.

Zz.

4. May 3, 2006

### Flatland

so there's no such thing as an impenetratable barrier?

Last edited: May 4, 2006
5. May 4, 2006

### ZapperZ

Staff Emeritus
Make it infinitely high and infinitely wide.

Why is this relevant to this thread?

Zz.

6. May 4, 2006

huh???????

7. May 4, 2006

### koantum

In the real world (governed by quantum mechanics) you donâ€™t ask how something works. You calculate the probabilities of possible measurement outcomes on the basis of actual measurement outcomes. If something has a probability greater than zero, it can happen. That's all there is to it.

8. May 5, 2006

### DaveC426913

Tunnelling is a poor description of the phenomenon; it does not actually go through the barrier, nor does it acquire enough energy to go over it.

Simply put, we do not question how it will come to be on the other side; we simply acknowledge that it did.

Q: But how?
A: By behaving according to the rules of QM.

When we look at a wave function of a particle (which will describe where we find it when we go looking for it), we see that part of the wave function continues outside the barrier. Thus, in a small fraction of times when we look for the particle, we will simply find it outside the barrier.

A key concept here, is that we cannot be "continually" looking at the particle. Measurement is an effectively instantaneous event. Best we can do is record where it is when we do look.

9. May 5, 2006

### ZapperZ

Staff Emeritus
But you also need to be careful here in saying something like this. I can show you evidence that yes, it DOES go through the barrier. How? Let's do electron tunneling as an example. I can imbed the insulating layer between two conductors with magnetic impurities. Such impurity does not change the electrostatic potential barrier, but can still interact with the electrons IF the electrons actually had to pass through the insulating barrier. And they do! Tunneling measurements, especially in superconductor-insulator-normal metal junctions are remarkably different when the barrier is free of magnetic impurities versus those that do have them. There are more inelastic scattering when there are more magnetic impurities in the barrier.

So no, the tunneling phenomenon isn't just "electron on one side - and electron appearing on the other side, and one cannot say anything about what is going on in the middle". We can, and we have evidence that the tunneling process does require the particle to traverse through the barrier itself.

Zz.

10. May 14, 2006

### NoTime

Interesting. I'm not surprised at the result, but at the fact it's been done.

Has anybody tried layering the doping of the barrier layer?
Say top, middle or bottom 1/3.
Would that kind of structuring be feasible with current technology?

Also is there any sort of transit time associated with this?