Re: How do optical logic gates work/how does an optical transistor work?

The basic approach in an optical logic gate, or optical switch, is to have
the intensity of one light beam input control how well another light beam 
is either reflected or transmitted.

The way this is done is rather complicated, so let me try to simplify it a 
bit before giving a real example.  You probably have noticed that metals 
like silver and aluminum reflect most colors of light - in fact, a 
household mirror is simply a glass plate coated with a layer of silver.

Metals can be thought of having a large number of free electrons, available 
to move throughout the material. Effectively, there is no clear gap between 
electrons that are effectively tied up by a nucleus (what we call the 
"valence band"), and electrons that have enough energy to wander freely 
(what we call the "conduction band"). 

In contrast, non-metals that appear to have a color have a clear gap 
between the valence and conduction band. To move an electron from the 
valence to the conduction band, it must absorb a certain amount of energy. 
That energy can come from a photon of light with just the right color. For 
example, an object may have the gap such that red light photons can be 
absorbed to move electrons into the conduction band - when you beam white 
light at this object, it may reflect most of the colors except for red, 
which is absorbed - so, the white light, made of red, orange, yellow, 
green, blue, and violet light, is reflected without the red, and the 
reflected light looks greenish-blue.

So, if we can set up an object such that the gap between the valence and 
the conduction band CHANGES when we hit it with a light beam of a certain 
intensity, we can change how well it will reflect another light beam.

One way to do this is to build a special kind of diode, called a PIN diode,  
where the bottom reflects light as a mirror, the middle has "Multiple 
Quantum Wells", a series of alternating layers of two semiconductor 
materials, such as gallium arsenide (GaAs) and aluminum gallium arsenide 
(AlGaAs), and the top is made a semiconductor that transmits light well. 
This is complicated further by requiring that the bottom "mirror" is n-type 
(some atoms of a different element are introduced or doped into the 
semiconductor so that there are now lots of free electrons), the middle 
(with the multiple quantum wells) is undoped or intrinisic (no atoms of 
different elements are introduced), and the top layer is p-type (some atoms 
of still another element are introduced so that there are now very,very few 
free electrons, but lots of vacancies or locations where electrons can 
enter the valence band). This is called a PIN diode because it is made of 
P-type (top), I-ntrinsic (middle) and N-type (bottom) layers - P / I / N.

If this multiple quantum well PIN diode is reverse biased (a negative 
voltage is applied to the P-type top compared to the N-type bottom), then 
the  multiple quantum wells serve as a place where photons of light of 
certain colors from one input beam can be absorbed - but when the photons 
of light are absorbed, electrons in the multiple quantum wells are moved 
from the valence to the conduction band - which changes how well the 
multiple quantum wells absorb various colors of light. 

So, the first input beam changes how well the other beam of light is 
reflected vs absorbed.

I hope this explanation helps; here are also some related websites:

http://www.gtri.gatech.edu/res-news/SWITCH.html
http://drip.colorado.edu/~blair/research/solitons/switching.html
http://www.ece.ucsb.edu/MOST/4.html