1) I think the primary reason for the high pressure
is to achieve a short mean free path (rather than
to just increase the number of traps). Corona
occurs when the E-field times a mean free path
is comparable to a molecular ionization potential.
Some guy named Geiger had something to say about
this.
2) I heretofore never thought of a gas as having
"traps" but I think BC has correctly described
the physics. The "trapped" charges diffuse
around a little more than they would in an
amorphous solid, but compared to the free motion
the charges would have in a vacuum or in a
perfect crystal, they are pretty much trapped.
Diffusion is reeeally slow compared to free motion.
3) In my previous note I used the word "corona
point" when I should have said "field emission
point". Same geometry, slightly different physics.
Hugh: I tried to answer you off-list but the fascist
anti-spam filter at your ISP won't accept mail from
me... which is ironic, in that it has essentially
"profiled" me as an abuser, even though my mailer
is (and always has been) 100% standards-compliant and
abuse-free. Evidently it is relying too heavily on
"secondary characteristics".
====
Speaking of spam: For the last 2 weeks I've been using
the latest beta version of Mozilla as my mail reader.
It comes with a trainable Bayesian spam detector. It
has reduced the amount of spam I have to deal with by
an order of magnitude, from 100 per day to about 10 per
day. The ones that get through tend to be very short
and/or have nearly every word misspelled, presumably
intentionally in order to circumvent this sort of
checker. As far as I can tell there have been no
false positives (falsely trashed legit messages) out
of several thousand decisions.
I'm sure its performance will deteriorate as the
spammers evolve to meet the challenge. But there's
room for the detector to get better, too. Right now
it just collects statistics word-by-word; it could
use word-pairs or higher-level linguistic analysis.
And it could get smarter about misspellings and
nonsense words.