From: thad@hammerhead.com (Thaddeus J. Beier)
Date: Tue, 20 Feb 1996 14:51:17 +0800
To: cypherpunks@toad.com
Subject: Optical repeaters
Message-ID: <199602200501.VAA10510@hammerhead.com>
MIME-Version: 1.0
Content-Type: text/plain



You should be able to use optical repeaters in a quantum cryptography situation.
The way lasers (these optical repeaters are lasers) amplify is that a photon
can stimulate the emission of an identical (in all respects) photon from an
excited atom as it goes by.  So you get two photons, with exactly the same
direction, frequency, polarization, and so on; without measuring those states.
This is how lasers work.

But, and this is the interesting part; you cannot use this feature to tap
the line.  Measuring either one of these photons would disturb the other,
destroying the state that it has.

It has been proven over and over that you cannot measure the polarization of
a photon in two axes at the same time, the measurment at one axis destroys
that information.  So you think, "Fine, I'll just stimulate the emission of
an identical photon, and measure one horizontally and one vertically. 
I'm only measuring each photon once, but since I know that the two photons
are identical, I can deduce the polarization in both axes of the one photon
from these two measurements."  Makes sense.  Doesn't work.  There was a
fabulous article in Scientific American, I think August 1978 that described
almost exactly this experiment.  You can think about it a number of different
ways, but the upshot is that you cannot defeat quantum cryptography this way,
the uncertainty principle will not let you.

thad
-- Thaddeus Beier                     thad@hammerhead.com
   Technology Development                   408) 286-3376
   Hammerhead Productions        http://www.got.net/~thad