Starting
with the first law
dU
= TdS - PdV,
we
can express the enthalpy as
dH
= dU + PdV + VdP = TdS + VdP,
the
Helmholtz free energy as
dA
= dU - TdS - SdT = -SdT - PdV,
and
the Gibb's free energy as
dG =
dH - TdS - SdT = -SdT + VdP.
These
four relations are exact differentials.
dU
= TdS - PdV
dH
= TdS + VdP
dA
= -SdT - PdV
dG
= -SdT + VdP
For
each of these we can derive an expression based on the equality of the second
cross derivative. These relations, also
known as Maxwell relations are fundamentally useful for thermodynamics.

Let
us begin the with the internal energy.
We can write the internal energy as a total derivative
Note
that this implies that

Now
take the second derivative (i.e. the cross derivative)

And
since the second cross derivatives are equal (for an exact differential)

This
symmetry leads us to state that the natural variables of the internal energy
are S and V. You might ask why I do not
consider V and T. After all we can
write

And
we can see immediately that

So

But
then we are left with the second term.
It is important to note that

But
rather

The
difference between these two partial derivatives is in the variable that is
held fixed. How do we find the value of
(¶U/¶V)T?
We
will show a Maxwell relation can be used to find an expression for (¶U/¶V)T. However, it is clear that the variables V
and T are going to lead to a more complicated expression for U than the
variables V and S did.
Let
us consider the Maxwell relations for A.
![]()
This
implies that

Now
take the second cross derivative of each of these.

Because
of the equality of the second cross derivatives we have

This
is perhaps the most useful of the Maxwell relations. Let us use it first to find (¶U/¶V)T.
Start
with A = U - TS. Take the derivative
with respect to V at constant temperature.

Use
the Maxwell relation for A above to
substitute for (¶S/¶V)T.
Note
that (¶A/¶V)T = -P so we have


While
this would be inconvenient to use in the first law, it is a useful equation and
is sometimes known as a thermodynamic equation of state. For example, for an ideal gas P = nRT/V and
therefore (¶P/¶T)V = nR/V. If we
substitute into the thermodynamic equation of state we find

This
is not a surprise. We have shown that
for an ideal gas U = 3/2nRT and therefore the internal energy depends only on
the temperature and not on the volume.
McQuarrie
and Simon show that the same is true for a hard sphere gas P(V-nb) = nRT (page
891), but not for van der Waal's or other equations of state that include
"a" parmeters that depend on the intermolecular potential energy.
We now use the Maxwell relation for A
to derive a general relationship between Cp and Cv. Recall that we have stated that for an ideal
gas Cp - Cv = nR.
Let us further examine the relationship.
Even
though V and T are not the natural variables for U we can start with


Likewise,
we can express U in terms of the variables P and T

We can find a natural
relationship between these two using the differential expression for the
volume.
![]()
Rearranging
we have

We
see that

Using
the fact that U = H - PV we have

And noting that

we obtain

The
Maxwell relation allows us to write this as

Noting
that

We
can substitute in for (¶P/¶T)V to obtain

We
define the isothermal compressibility

And
the coefficient of thermal expansion
![]()
Using
these definition the difference in the heat capacities becomes
![]()
Although
we have focused on the Maxwell relations for U and A, the same procedure leads
to Maxwell relations for H and G. These
are described in Chapter 22.
The Maxwell relations and associated
machinery are NOT something you memorize.
Everything can be derived from dU = TdS – PdV which is a combined
statement of the First and Second Laws.
Since
you can get dH, dG, and dA from this expression, you can obtain the Maxwell
relations from the expression of each of these as a total derivative.
|
State
function |
Total
derivative |
Maxwell
relation |
|
dU |
TdS – PdV |
(¶T/¶V)S = -(¶P/¶S)V |
|
dH |
TdS + VdP
|
(¶T/¶P)S = (¶V/¶S)P |
|
dA |
–SdT
– PdV |
(¶S/¶V)T = (¶P/¶T)V |
|
dG |
–SdT
+ VdP |
-(¶S/¶P)T = (¶V/¶T)P |

The
standard state for a gas at any temperature is the hypothetical ideal gas at
one bar. To calculate corrections to
the entropy due to deviations from ideality we begin with the Maxwell relation
for dG.
This
expression allows us to calculate the pressure dependence of the entropy.

Suppose
we wish to determine the deviation from ideality at 1 bar of pressure (i.e. at
the standard state). At some very low
pressure all gases behave ideally so we consider an ideal gas at pressure Pid.

The
quantity (¶V/¶T)P can be determined from the equation of state for the
gas. For a hypothetical ideal gas this
same expression is

Adding
these last two equations together gives the difference from ideality at 1 bar
of pressure

Now
we need only find an equation of state that we can use to determine (¶V/¶T)P. We can use the virial equation of state for
example
![]()
From
the virial expansion we have

Which
gives a particularly simple expression for the deviation from the ideal entropy
![]()
The
temperature and pressure dependencies of the Gibb's free energy

We
shown that dG = -SdT + VdP. This
differential can be used to determine both the pressure and temperature
dependence of the free energy. At
constant temperature SdT = 0 and dG = VdP.
The integrated form of this equation is
For
one mole of an ideal gas we have

The
same formula can be obtained from
![]()
We
can use the above expression to indicate the free energy at some pressure P
relative to the pressure of the standard state P = 1 bar.
![]()
G0(T)
is the standard molar Gibb’s free energy for a gas. As discussed above the standard molar Gibb’s free energy is the
free energy of one mole of the gas at 1 bar of pressure. The Gibb’s free energy increases
logarithmically with pressure. This is
entirely an entropic effect.
If
we are dealing with a liquid or a solid the molar volume is more or less a
constant as a function of pressure.
Actually, it depends on the isothermal compressibility, k = -1/V(¶V/¶dP)T, but since k
is typically a number of the order 10-4 atm-1 for liquids
and 10-6 atm-1
for solids. For our purposes we
can treat the volume as a constant and we obtain
![]()
The
Gibb’s free energy also depends on temperature. The easiest way to determine the temperature dependence is to
begin with G = H – TS and divide through by T.
![]()
We
differentiate this expression with respect to T keeping P fixed.

The
last two terms cancel since (¶S/¶T)P = CP/T.

Which
can be applied to any process in which case it is written as

You
might think that we could calculate the temperature dependence of the free
energy directly from DG = DH – TDS and, in fact, we can. We can
calculate the enthalpy referenced to the enthalpy at absolute zero and the
absolute entropy. However, as discussed
on page 903 and 904 of McQuarrie and Simon, this is a cumbersome procedure and
involves some duplication of effort since DfusS = DfusH/Tfus and DvapS = DvapH/Tvap so the
calculations done for entropies and enthalpies of phase transitions will
cancel. As we have already seen this is
consistent with the fact DfusG = 0 and DvapG = 0. A plot of G(T) vs. T has a constant
negative slope with discontinuities at the phase transitions. The reason for the negative slope (see
Figure 22.7) is that

Of
course, the entropies of solid, liquid, and vapor are different giving rise to
the changes in slope at the phase transitions.