adiabaticity and reversibility

[Leigh Palmer <palmer@SFU.CA>]

> > we used the word 'adiabatic' to describe a PROCESS not a 'wall'.
>
> Right!
>
> A good illustration of this is ye olde cylinder of gas with piston.  If you
> pull out the piston too quickly, the process would be better described
> using the sudden approximation than the adiabatic approximation.  At the
> other extreme, if you pull out the piston out too slowly, you need to worry
> about thermal conduction and all sorts of other non-adiabatic
> processes.  But if you do it baby-bear-just-right, the adiabatic
> approximation may be a very useful description.
>
> Also note that nothing is *exactly* adiabatic -- but it is still
> appropriate to use terms like "adiabatic process" to describe ideal
> processes, and to describe real-world processes that are adiabatic to a
> good approximation.

This is a misuse of the term "adiabatic", though it is a common one.
Pulling the piston out "too quickly" should read "at a speed comparable
to the speed of sound", and that corresponds to an irreversible process
which can certainly also be adiabatic. If the piston is withdrawn (or
even advanced) with a speed much less than the speed of sound the
process is said to be *reversible*, and ideality is approached as the
speed approaches zero. "Adiabatic" refers to the ideal of nontransfer
of heat. "Adiabatic demagnetization", for example, refers to the
removal of a magnetic field after the salt pill has been *thermally
isolated* from the previous refrigeration stage in a low temperature
experiment. It does not refer to the manner in which the magnetic field
is reduced in time, though that is a commonly misconceived idea. Free
expansion of a gas is another example of an adiabatic process which is
irreversible, and the "adiabatic approximation" applied to this case
is that it is carried out sufficiently quickly that exchange of heat
with the walls of the container during the process can be neglected.

This reminds me of one of my favorite demonstrations. How many of you
do the Clement-Desormes experiment to measure the ratio of specific
heats in a gas? I always do this as a lecture demonstration when I
teach thermodynamics. It really works very well when one learns to
perform the expansion in a manner which justifies the adiabatic
approximation.

Leigh