From the responses I've seen trickling back to the list, it appears I may
not have been clear enough, so here are a few additional details.
First, some folks seem to be suggesting that my arrangement is overly ideal
and hence the whole issue is hypothetical. I disagree. Look up the
derivation of the fact that the spectral (as opposed to the total)
emissivity and absorptivity of an object are equal - i.e., wavelength by
wavelength. The simplest derivation postulates an ideal filter which
perfectly transmits one narrow band and perfectly reflects all other
wavelengths. I want to similarly postulate an ideal polarizer which
perfectly transmits one linear polarization and perfectly reflects the
other. Is this such a stretch of imagination? Surely real polarizers exist
which come close to this (say to within 1%)? Even if they don't exist, are
you sure we can squirm out of the problem this way? The diode analogy may
be appropriate: do real diodes exist which come to the necessary ideal or
failing that can we idealize a diode so that it approaches the required
properties without violating any physics? That's really what I want to do
with my polarizers - I'm hoping that the 1% reality factor is merely an
engineering limitation and does not contain any important physics in it.
As for my admittedly somewhat complicated arrangement: I want to be be sure
the light entering the Faraday isolators from the outside does not reflect
back through the rotators repeatedly because after 3 passes it will rotate
to a transmitting orientation! So I purposefully tilt the polarizer on the
sample side so that it is not normal to the sample emission, to avoid
creating an optical cavity in either isolator. This way, after 1 pass the
rejected light originally from the outer world is dumped out. (The reason I
introduced a second isolator was so that the polarization coming from the
sample which is reflected off the tilted polarizer also escapes.) So it
appears no light gets back in, but light is getting out.
Obviously there must be a flaw in my development somewhere if both
Kirchhoff's law and the 2nd law are to stand.
Dr. Carl E. Mungan, Assistant Professor http://www.uwf.edu/~cmungan/
Dept. of Physics, University of West Florida, Pensacola, FL 32514-5751
office: 850-474-2645 (secretary -2267, FAX -3323) email: cmungan@uwf.edu