94 DOC.
81 MAY 1915
Example:
solution
(volume
V) semipermeable
membrane
solvent
(volume
V[0])
If there
is
no
diffusion at absolute
zero,
which
is
what
I
believe,
then there
can
be
no
adiabatic
enlargement
of
the
solvent’s volume at
T
=
0[2]
(naturally
no
more
than
an
isothermal
one).
A
shifting
of
the
semipermeable plane
would not
result
in
any
diffusion
(no
matter how
slowly
it
is
carried
out).
In
the
case
where
the
envisaged process
is
a
chemical
reaction,
no more
can
I
see
the
existence in
principle
of
a
“zero
point
reaction.”
It
is
my conviction, therefore,
that
nothing
further
can
be achieved here from
a
purely
thermodynamic
approach.
On
the other
hand,
there
are even
strong grounds
for Nernst’s
theorem
not
being
valid for mixed
processes.
Consider
that
with
the
previously envisaged
diffusion
process
the
dissolved
substance and
the
solvent differ
only slightly
from
each
other
(dissolving
of
one
isotope
metal
type
into the
other)
and
observe
the
following
reversible
process
(A-B-C-D):
T1
T
B C
V1
A
V
D
V2
V
is
the
volume
throughout
which, by
virtue
of
the
semipermeable membrane,
the
dissolved
substance
is
distributed.
T1
is
a
temperature
high enough
that the
law of osmotic
pressure unquestionably applies.
In
the
special
case
examined
here,
it
ought
hardly
be relevant to
the
specific
heat
per
molecule whether
the
neighboring
molecules
are
molecules of
the
same
sort
or
of
the
isotopic
sort.
That
is
why
at
constant
V
the heat
capacity is
independent
very
close to V.
Under this
scarcely
circumventible
assumption,
if
the
system’s entropy
is
referred to
as
S,
the relation
will
be
SB-SA
=
SC-SD