The cool thing about quantum mechanics is that a man-in-the-middle attach won't work! The reason for this is the Heisenberg Uncertainty Principle, which states

The cool thing about quantum mechanics is that a man-in-the-middle attach won't work! The reason for this is the Heisenberg Uncertainty Principle, which states that there are certain attributes of a particle where we can't know both of them perfectly. The classic example of this is position and energy. The relevant attributes here, though, are angular momentum (spin) in the "left/right" and "up/down" directions. Think of a particle as two arrows, one pointed either left or right, and the other pointed either up or down. The Heisenberg Uncertainty Principle states that if I know with 100% accuracy that a particle is pointed left (or right), I can't know anything about if it's pointed up or down, and vice versa. In fact, if I measure the left/rightness of a particle, the up/downness gets set randomly to be either of them. So quantum communication works more or less like this: We agree that left and up correspond to a 1, and right and down correspond to a zero. So if I want to send a 1, I randomly choose a either a left or an up, and send it to you. Then, on your end, you randomly choose an axis to measure on, and after I've measured, we tell each other (through a public channel) which axis we measured on. If we agree, we use the bit, otherwise, we throw it away. Ok, so now, what if Eve comes in the middle and steals our electron, reads it, and sends it again. Well, she doesn't know what axis to measure on! So if I send it on the up/down axis, Eve measures it on the right/left axis, and you measure it on the up/down axis, you have a 50% chance of getting a wrong answer. So if we take a few bits at random and publish them, we can get a pretty good idea if someone's snooping. Also, Eve can't copy our particle, because if she could, she could read one axis from one particle and the second from the other, which would violate the Uncertainty Principle (this result is known as the No-Cloning theorem).