Re: IONS

[Doug Craigen <dcc@cyberspc.mb.ca>]

> Put a light bulb in a ceramic socket.
>
> Turn on the lamp.
>
> Break the glass envelope with a hammer.
>
> Apply a propane torch to the part of the glass supporting the internal
> wires until it glows.
>
> Remove the flame, and wait to see if it cools.

I haven't been following PHYS-L for a few days, so pardon me if I am off
track here.  I gather that the question is why does glass become
conductive when heated.  The fact is that in general semiconductors and
insulators become more conductive as heated.  This is easily understood
in terms of a band structure diagram.


///////   Conduction Band
-------

--------   fermi level


-------
///////   Valence Band


For semiconductors and insulators the fermi level lies in a more or less
empty region between a full valence band and an empty conduction band. 
It is very often "pinned" in position  (doesn't change much with
temperature) by a number of defects - e.g. "dangling bonds".  The defect
states are spatially localized so they give rise to very little
conductivity (a little does occur by charge carriers "hopping" from one
to another - this is a thermally activated process).

If the gap between the fermi level and either of the bands is very large
compared to kT then we are dealing with Boltzmann statistics for the
probability of finding electrons in the conduction band or holes in the
valence band.  You integrate the probability of occupation over the
state density for the various energy levels in each band to find out how
many electrons and holes there are - the result will be an exponentially
activated number of each.  E.g. n_electrons is proportional to
exp(-E/kT).
In the case of a transparent material such as glass, E will likely be in
the order of an electron volt or more (the total band gap must be more
than 3 eV for transparency, and the fermi level will usually be
somewhere near the middle).  So, the conductivity will rise extremely
quickly with temperature - many orders of magnitude if you go up a few
hundred degrees.

**** anyways, there's your basic parameters that need to be
characterized for any given insulator to determine how fast its carrier
density rises with temperature

*************

There is one other question about this demonstration, how do you know
its the glass that is conducting the electricity?  Light bulb glass is
commonly tampered with in various ways.  For example, over a decade ago
I recall reading about a proposal to coat the inside with SnO2 to act as
a heat mirror to increase the efficiency of the bulb.  If your glass
bulb has a surface semiconductor coating like this present, it will
dominate the electrical properties of your test.  One can also imagine
various unintentional ways that coatings may be deposited onto the
internal parts of a light bulb.


/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\

Doug Craigen          "Technology with purpose"
                       http://www.dctech.com