# Reflection at the Quantum Level

1. Nov 1, 2005

### -Job-

If i throw a tennis ball against a wall in a 45 degree angle, it reflects with a 45 degree angle. A beam of light against a mirror also reflects in a 45 degree angle.
If i throw a tennis ball against some irregular surface in a 45 degree angle then the ball, depending on the shape of the surface, will possibly not reflect in a 45 degree angle.
If my ball is a photon and my surface is the mirror, at the quantum level the mirror, composed of its atoms and molecules is likely an irregular surface, hence i would expect that a beam of light against a mirror would reflect in an angle depending on where it hit the mirror.
Is a reflected photon absorved and then emitted or is it just never absorved at all? I'm a little confused.

2. Nov 1, 2005

### CarlB

The reflection of a photon by a mirror is simple when described as a part of the "wave" nature of light. You're looking at it from the "particle" point of view and that is a lot more complicated.

To do the light reflection as a particle, you have to sum up the contributions of all the places the light could possibly have been absorbed at, and then, using a "propagator" figure out where the light could have gone from that spot. In the end, you'll get the same answer, but it will take a lot more trouble.

Feynman has a good discussion of this in his lay oriented book "QED, the strange theory of matter and light" (\$11 at Amazon):

https://www.amazon.com/exec/obidos/tg/detail/-/0691024170/qid=1130893372

Carl

3. Nov 1, 2005

### ZapperZ

Staff Emeritus
Notice one very important thing - a typical mirror is made of a metal.

What this implies is that within the visible range, the conduction electrons play the most significant role in this. The transition from one band to the next upon absorption of a photon involves a conservation of the transverse momentum (separated by the reciprocal lattice vector). And because these are electrons, their response to the photons E-field causes a retransmission that lags in phase by pi/2.

Zz.

4. Nov 1, 2005

### -Job-

Excellent, that was the only explanation that seemed to fit, thanks.