Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Faraday isolators



It bothers me that Faraday isolators don't obey optical reversibility. The
fact that looking through one I could see you but you not see me is
disquieting. In particular, I don't quite see how to save Kirchhoff's law
(absorptivity = emissivity) with these gizmos.

The arrangement I have in mind is simple. A black sample is inserted into a
cavity having perfectly reflective walls. Then I make a small window in the
walls into which I place a Faraday isolator. Now what prevents the sample
from radiating away energy continuously and cooling down to 0 K? The basic
idea is so simple I'm sure some of you have thought of it before.

To dispense with a couple of objections off the bat:

(1) a Faraday isolator only works at one wavelength - okay, put an ideal
filter in front of it (i.e., one which perfectly transmits a narrow
passband and perfectly reflects everything else)

(2) a Faraday isolator only transmits one linear polarization - okay, how
about the following arrangement:
F
S / F

where S is the sample, / is an ideal 45-degree cube polarizing beamsplitter
which perfectly transmits say the p-polarization and perfectly reflects the
s-polarization, and F are two isolators consisting of a magnetic rotator
and an output linear polarizer (which perfectly reflects the rejected
polarization from the outside and perfectly transmits the other
polarization from both sides) oriented at 45 degrees (to pass the rotated
light in the usual way). Radiation from the outside which passes through
either F gets sent out through the open bottom channel of the cube. Notice
that I have studiously avoided absorption in all components so there is
nothing to heat up the sample.

Carl Mungan <cmungan@uwf.edu> (who only reads the digest)