What have we been discussing?
As you can see, Aurele has extended for the next 2 weeks the discussion regarding optical antennas INCLUDING some related or less related issues which pop up! So perhaps we can get further on Wednesday, perhaps even to section 4 (applications) but also to see if there is further interest in addressing the qualitative discussion in 3.8, and 3.9 (nonlinearity in metals!). We’ll just have to skip past the final sentence of section 3.7 which I still cannot believe and is such an amaxing finding that it should have required its own paper 😉
However last time we did get into a serious discussion of the complex refractive index of metals which goes beyond the question of antennas but is obviously of great importance to anyone who uses metals at optical frequencies (such as Aluminumized mirrors) and wants to understand how they work — quantitatively. In particular there was the issue of epsilon_infinity (assymptotic value at high frequencies well past the plasma frequency): should it be unity, or something else? And if it IS something >1, then does that not affect the plasma frequency?
I had referred to a table in Feynman table 32-3 (in my old edition) taken from Kittel which tabulates the predicted and experimental plasma frequencies of alkalii metals (from an old experiment by Robert Wood, I believe, who realized that metals would become transparent in the far UV). Now I have done a calculation based on the conductivity, valence, and density of Lithium (all well known) but assuming epsilon_infinity = 1 (as Feynman does) and have plotted the result here:
  http://www.strw.leidenuniv.nl/~meisner/LithiumIndexTheory.jpg
This is very consistent with Feynman/Kittel but differs significantly from this reference (which shouldn’t be taken as ABSOLUTELY true, but is the best I can find):
  http://refractiveindex.info/?group=METALS&material=Lithium
Note that he plots Re{n} at the top and Im{n} down below vs. microns, as I do. Comments?
