Roy Lewallen wrote:
Jim Lux wrote:

There are also complications about temperature when referring to
solids, liquids, and gases. The "temperature" of even a weekly
ionized plasma is quite high (e.g. 11000 K per eV), but that more to
do with the velocity of the ions and the mean free path. There's not
much mass there, so the "heat" is small. That is, even though the
ionosphere is "hot" in a temperature sense, it's not very "hot" in a
sensible transfer of heat sense.


BTW, I think the sunburn is not from thermal absorption, but from
photons with enough energy to make the reaction go. The total energy
in the UV of sunlight is MUCH lower than the total energy in the
visible range. The power spectrum of sunlight is pretty close to the
spectral sensitivity of your eyes (which evolved that way to match, I
would think).

A good graph of sunlight power density vs wavelength can be found at
http://en.wikipedia.org/wiki/File:Solar_Spectrum.png. Comparing areas of
various graph sections shows that the UV part of the spectrum contains
maybe 1/5 the amount of energy as the visible part -- plenty enough to
embrittle plastics and fabrics and sunburn skin. But the infrared energy
-- invisible to our eyes -- looks to be at least equal to the visible
energy.



The plastics degradation is definitely an "athermal" effect (because
adding carbon black to the plastic inhibits it, but doesn't change the
absorbed power very much.

But..
note that the scale is in wavelength and the energy is "per nm" (because
that's how spectrophotometers work). the photons have less energy at
lower wavelength. (or, you could plot it in frequency, and then look at
the watts/Hz to integrate)

If you look at power spectral density (e.g. watts/hz) it actually peaks
up around 1000 nm (near IR). The Wien displacement law says that 5250K
peaks up at about 550 nm, but the power spectral density at 550nm (545
THz) is about 2/3 that at the peak.
By the time you get to 350nm (857 THz), the energy per hz is about 10%
of what it is at the peak (at 950nm)

Running a quick numerical integration... (multiplying the power spectral
density every 50 nm by the frequency range).. I get 0.166 for all
wavelengths shorter than 320nm, 2.09 for 320-670, and 3.6 for 670-4000 nm
(there's a missing integration constant, so the numbers have some scale
factor, but the relative amounts should match..)
for the band around 400nm, I get .26 and for the band at 550 about 0.34
and for around 650 about .32... Yes, it peaks at 550 nm as expected.