Section 2D. State Functions and Exact Differentials
Section 2D. State Functions and Exact Differentials
We now need to have a mathematical aside to introduce some useful information about functions
of more than one variable (multivariable calculus).
( ∂x ) ( ∂y )
∂f ∂f
df = dx + dy
y x
( ∂x )
∂f
where, for example, the partial derivative involves taking the derivative of the function
y
with respect to x, while taking y to be a constant.
Example: Write the different df for f (x, y) = A x 3 + Bx 2 y + Cx y 2 + D y 2 + E.
( ∂x )
∂f
= 3A x 2 + 2Bx y + C y 2 = g(x, y)
y
( ∂y )
∂f
= Bx 2 + 2Cx y + 2D y = h(x, y)
x
( ∂x ) ( ∂y )
∂f ∂f
d f = g(x, y)d x + h(x, y)d y, and g(x, y) = and h(x, y) = ,
y x
is an exact differential if and only if, the following condition is satisfied:
( ∂y ) ( ∂x ) ( ∂y∂x ) ( ∂x∂y )
∂g ∂h ∂2 f ∂2 f
= i.e., =
x y
The satisfaction of this criterion implies that the order of application of differentiation is not
important for functions with exact differential. An exact differential is an infinitesimal quantity
that, when integrated, gives a result independent of the path. State functions have exact
differentials. Path functions do not have exact differentials, and thus upon integration gives a
result that depends on the path.
25
( ∂y )
∂g
= 0 + 2Bx + 2C y
x
( ∂x )
∂h
= 2Bx + 2C y + 0
y
These are equivalent, so df for the example function is an exact differential.
Thermodynamic state functions depend on more than one state variable, e.g., U(T,P), U(p,V), or
U(T,V), etc... Suppose we take U = U(T,V). The change in internal energy U when temperature T
and volume V are changing is:
( ∂T ) ( ∂V )
∂U ∂U
dU = dT + dV = CV dT + πT dV
V T
The differential dU depends on the constant volume heat capacity, and we have introduced a new
quantity: the internal pressure πT. Note that for a perfect gas πT = 0, as the internal energy
depends only in the temperature, and, hence, there will be no change in the internal energy for
changing volume at constant temperature.
( ∂T ) ( ∂p )
∂V ∂V
dV = dT + dT
p T
( p2 )
nR −n RT
= dT + dp
p
To show that dV is an exact differential, we must show
( ∂p ( ∂T )p) ( ∂T ( ∂p )T)
∂ ∂V ∂ ∂V
=
T p
From the results above, we have
∂( p ) ( )
−nRT
nR ∂
p2 −n R −n R
= ⟹ =
∂p ∂T p2 p2
T
p
Therefore, we conclude that indeed dV is an exact differential.
What about inexact differentials? Let us consider as an example the heat for the reversible
expansion of a perfect gas due to heating.
26
( ∂T )
∂U n RT n RT
d q = dU + pdV = dT + dV = CV dT + dV
V
V V
For a perfect gas, the constant volume heat capacity does not depend on volume, i.e,
( ∂V )
∂CV
=0
T
However,
∂( V )
nRT
nR
=
∂T V
V
These are clearly not equal, so dq is an inexact differential, and therefore, is a path function.
However, what if we consider a change in entropy dS (we will discuss much more on entropy in
Section 3: The 2nd and 3rd Laws), where
dq CV nR
dS = = dT + dV
T T V
Now, we have
( )
∂(CV /T )
=0
∂V T
and
∂( V )
nR
=0
∂T
V
Hence, dS is an exact differential and S is a state function.
In order to more clearly see how the value of an integral between specific end points (states of
the system) depends on the pathway, let’s consider the heat for a very specific change of state for
1 mole of a monoatomic perfect gas:
T1,V1 T2,V2
n RT 3 RT
For this process, d q = nCV,m dT + dV = RdT + dV where we have used
V 2 V
3
CV,m = R for a monoatomic gas and the fact we are considering 1 mole.
2
Among the many possible pathways to get from State 1 to State 2, two are the following
Constant T expansion Constant V heating
Path 1: T1,V1 T1,V2 T2,V2
( V1 ) 2
T2V2 V2 T2
RT1 3 V2 3
∫T V ∫V ∫T 2
Path 1: q1 = dq = dV + RdT = RT1 ln + R(T2 − T1)
1 1 1
V 1
( V1 )
T2V2 T2 V2
3 RT2 3 V2
∫T V ∫T ∫V V
Path 2: q2 = dq = RdT + dV = R(T2 − T1) + RT2 ln
1 1 1
2 1
2
These are clearly not equal, q1 ≠ q2, and therefore the value depends on the path.