Photons coming from atoms, when electron changes levels

Bassalisk

Today I did the usual thinking, watched some courses online etc. and found out(or rather reminded myself) that the EM waves can be produced only when a charge oscillates.(or thats what the professor at Yale said)

But then I remembered LED diodes, and their ability to emit light. Electron goes from higher energy to lower energy state and produces a photon in process. Not until today I gave this a thought.

How come it emits a photon? Photon is a EM wave, it oscillates. So in order to emit a photon, it had to oscillate in the process?
Is this one of those answers "it just does that", "its the way it works" or is there more to it? Because all textbooks are going as deep as: change of energy: emit a photon.

mfb

You do not need a nice oscillation, any classical acceleration will emit EM waves. However, this is a classical view, and cannot be transferred to quantum mechanics easily. The allowed energy levels of the electron have band structures in semiconductors, and the transition between those bands can produce a photon. The momentum of the electron changes, but it is not like an oscillation.

Bassalisk

Hmmm but what does then produce the photon? The change in energy?

I understand that difference in energy levels will produce a that much wave length photon.

But these are wavelengths of nanometers etc, meaning high frequency. How do you get a EM wave, which oscillates, from a simple transition from one level to the other?

Lets go further, what do we mean by one photon? Photon is a particle and a wave. Does it have length? It has energy and wavelength.

But I see EM waves as lets say electron oscillating up and down, and generating oscillating electric field which then generates magnetic field perpendicular to it and the that travels forward into space, as a EM wave.(simple example of electron oscillating up and down)

I am familiar with the fact that ANY acceleration will generate a wave, I learned that from antennas. But if we are talking about wavelengths then we must talk about periodic waves, namely sinusoidal waves. Or at least ones that can be decomposed into F.S.

uart

As mfb said, the classical view doesn't transfer very well to the quantum domain. But the key point is that you don't need oscillations, just rapid acceleration/deceleration, or more generally a change in energy. (Remember that classical oscillations are really just a sinusoidal acceleration).

Note that these recombination energy changes happen over very small distances, typically within atomic dimensions. Remember that xrays, which are very much higher frequency the the LED photons we're discussing here, are produced by electrons simply decelerating as they strike a tungsten anode. There is no oscillation as such, but an electron stops pretty darn quickly when it strikes a tungsten nucleus.

Bassalisk

This is why I have regrets not studying quantum mechanics. I think I understand what you are saying, in a nutshell: "it just works like that". Like there is no explanation for why electric field exists. It just does.

And I took LEDs as an example, because from there i learned about Fermi levels band gaps etc.


I was hoping for an answer like it wiggles in a process of transition etc.

Thank you both for giving me satisfying answer.

mfb

Well, it is better: "It can be calculated".

Every transition has an intrinsic width in the frequency distribution. If you detect the emitted light, you will always see a single photon, but its frequency can vary a bit. In a classical view, this corresponds to deviations from the sinusoidal shape of the wave. However, usually the deviations are extremely small, so don't imagine some fancy shapes here.

Bassalisk

I will get there. I am EE after all. This is just me wanting to know a bit more that I have to.

Thank you for your time.

SK3

Richard Feynman explained (with some input from Frank Vernon and Robert Hellwarth)
A way of visualizing a quantum leap as the flip of a dipole moment. This allows a classical
dipole moment model and the electromagnetic field that it generates and the unit of energy
that it carries to be also thought of as a "photon" in a valid quantum mechanical yet classically able to be visualized process.
see

http://ieeexplore.ieee.org/xpl/login...mber%3D5121723
or
http://adsabs.harvard.edu/abs/1996AmJPh..64.1475K

mikeph

I think this is another case of faith, until you learn more. In this case the spontaneous generation of a photon requires relativistic quantum mechanics.

SK3

Actually not, Feynman used the Schrodinger Equation to demonstrate this. There is a way to do this with the Dirac Equation but I don't know if it was published. The Schrodinger Equation is non-relativistic quantum mechanics. The method is similar to what Rabi used for Nuclear Magnetic Resonance which was a classical non-relativistic model which Julian Schwinger showed was equivalent to the quantum mechanical treatment.
Your comment is in conflict with both Richard Feynman and Julian Schwinger so most people with any knowledge of the subject would tell you you are totally wrong!

SK3

mikeph

Do you have a link that I can access without a password? I'd be interested to read about that.


With my current understanding, spontaneous emission is tricky because the Schrodinger equation leads us to believe that the atom can exist in numerous different stationary states with different energies. So an excited state, being a stationary state, should be stable if left alone. But the atom still transitions to a different stationary state under no apparent external force (certainly no change in the Hamiltonian), and, furthermore, the atom's wavefunction has no mention of any photon! Where does the photon come from?