Y.

Typical examples of irreversible processes

Y.1 The Joule free expansion of a gas
Imagine a gas confined in a partition (of volume V
1
) of an adiabatically insulated container of
total volume V
2
(Figure 1). The other part of the container is initially evacuated. Let T
1
be the initial
temperature of the gas. The following natural question arises: what happens to the gas if it is
allowed to expand suddenly into the
whole volume of the container by
opening the tap? Intuitively, one can
readily say that this phenomenon is
irreversible, since no spontaneous re-
confinement of the gas back into the
original partition can be imagined. The
irreversible character of this expansion
will be proven in the following by means
of the overall framework of the Laws of
Thermodynamics. This phenomenon was
first studied in an attempt to obtain the
caloric equation of state for a perfect gas. The most consistent report was made first by James
Prescott Joule in a paper published in 1845 [1].
In analyzing the phenomenon one may readily observe that the work done by the gas during the
expansion is nil, since there is no opposition from the evacuated compartment of the container.
Therefore, the work channel is closed for the expanding gas, by the very experimental arrangement.
Moreover, the caloric channel is also closed due to the assumed adiabatic enclosure of the system.
Needless to say that the chemical channel is also blocked as no amount of gas is allowed to leak
neither outside the container nor inside it. The First Law of Thermodynamics then states that the
internal energy of the system is constant during the process.
To check the reversibility of the process, assume an infinitesimal version of it, in which the gas
expands from the volume V to V + dV. As there should be a reversible process connecting the initial
and final states of this process, one may write the fundamental equation of Thermodynamics:
T dS dU PdV d v = + . (1)
As the internal energy and the mass of the gas should be the same in both states, dU = 0 and dv = 0,
so that Eq. (1) becomes:

Figure 1: Schematic of the free expansion apparatus.

P
dS dV
T
= . (2)
Therefore, as the process is necessarily an expansion and thus dV > 0, and since P and T are strictly
positive quantities, one may conclude that dS > 0 in the infinitesimal process. On the other hand, the
system being insulated, one has oQ = 0 during the free expansion. Therefore, dS > oQ/T and so, the
process should be irreversible.
Moreover, an important question related to the free expansion of a gas is how its temperature
changes in the final state. To answer this question consider again an infinitesimal free expansion,
for which we again have dU = 0. Writing down the internal energy variation from the caloric
equation of state, one readily obtains:

, , ,
0
V T T V
U U U
dU dT dV d
T V
v v
v
v
c c c | | | | | |
= + + =
| | |
c c c
\ . \ . \ .
. (3)
But, as dv = 0,
,
V
V
U
C
T
v
c | |
=
|
c
\ .
, ( )
, ,
1
V
P V
U P
T P P T
V T
v v
|
c c | | | |
= =
| |
c c
\ . \ .
, Eq. (3) becomes:
( ) 1
V
V
P
dT T dV
C
| = . (4)
As P, C
V
and dV are strictly positive quantities, it turns out that the sign of dT is determined by the
parenthesis of Eq. (4). For an ideal gas, one readily finds:

,
1 1
V
V
P
P T T
v
|
c | |
= =
|
c
\ .
, (5)
so that no temperature variation should be detected. This was actually obtained in the early
experiments on free expansion of gases and led Joule to formulate the thermal equation of state for
an ideal gas as a function of temperature only. However, subsequent experiments revealed that
slight temperature variations can always be detected following the free expansion, so that the ideal
gas model has a limited validity. As discussed in Chapter 8, a more realistic thermal equation of
state for a gas would be given by the Van der Waals formula:
( )
2
2
P a V b RT
V
v
v v
| |
+ =
|
\ .
, (6)
where a and b are the Van der Waals parameters of the specified gas. By taking the partial
derivative with respect to the temperature in both sides of Eq. (6), one readily obtains:

2
2
,
1
V
P
P a
T T V
v
v | | c | |
= +
| |
c
\ .
\ .
, (7)
so that

2 2
2 2
1
1 1
V
T P a a
P V PV
v v
|
| |
= + =
|
\ .
, (8)
which is a strictly negative quantity. Therefore, according to Eq. (4), one may always expect that
real gases get cooled after a free expansion. More exactly, an infinitesimal free expansion of a Van
der Waals gas will produce the following temperature variation:

2
2
V
a
dT dV
C V
v
= . (9)
By partitioning a macroscopic free expansion (from the volume V
1
to V
2
) into infinitesimal such
processes, one may readily compute the corresponding temperature variation of a Van der Waals
gas:
( )
2
exp
1 2
free ansion
V
a V
T
C V V
v A
A = . (10)
Y.2 The Joule-Thomson effect
Another example of irreversible process with gases is the Joule-Thomson effect. This
phenomenon is not only interesting by itself, but has also very important practical implications. It is
named for James Prescott
Joule and William
Thomson, 1st Baron Kelvin
who discovered the effect in
1852 [2, 3] following earlier
work by Joule on free
expansion of gases. The
effect is applied in the Linde
technique as a standard
process in the petrochemical
industry for example, where
the cooling effect is used to
liquefy gases, and also in many cryogenic applications (e.g. for the production of liquid oxygen,
nitrogen and argon).

Figure 2: Schematic of the Joule-Thomson apparatus.
The Joule-Thomson effect refers to the changes in the state of a gas that is expanding under
pressure through a porous plug or a throttle. This kind of gas expansion is actually denoted as
throttling in the literature. The phenomenon takes place in a thermally insulating cylinder (Figure
2), which the plug or throttle divides into two compartments. The gas being initially at the
temperature T
1
in a volume V
1
is pushed with an adiabatic piston, under high constant pressure P
1
,
through the throttle and collected on the other side at a constant pressure P
2
, which is controlled by
another adiabatic piston. The final state of the gas is also characterized by the temperature T
2
and
the volume V
2
. The amount of gas is supposed to remain unchanged. Thus, the gas throttling takes
place with both the chemical and thermal channels closed. The First Law written for the process
takes therefore the form:
U W A = . (11)
As both pressures are kept constant, the total work performed on the system will be:

( ) ( )
1 1 2 2 1 1 2 2
0 0 W P V P V PV P V = = . (12)
Therefore, as AU = U
2
U
1
, Eq. (11) becomes:

1 1 1 2 2 2
U PV U P V + = + , (13)
thus showing that the throttling takes place so that the enthalpy of the gas does not change
(isenthalpic process):
1 2
H H = .
The throttling of a gas is an irreversible process. To prove this, let us consider an infinitesimal
transformation of this kind, where all the state parameters (except of those which are kept constants)
vary by infinitesimal amounts. Consider a reversible process linking the initial and final states. Still
this reversible transformation keeps the enthalpy and the mass constants (since both quantities are
state parameters). Therefore, by using the fundamental equation of Thermodynamics, one may
write:
( )
,
, ,
1 1
H
H H
P V V
dS dU dV d dH dP d dP
T T T T T T T
v
v v

v v
| | | |
= + = =
| |
\ . \ .
. (14)
As the gass pressure decreases following the throttling, dP < 0 and thus Eq. (14) shows that dS >0.
Moreover, as the throttling occurs under adiabatic enclosure, one may observe that oQ = 0.
Therefore, one readily finds that during throttling:

Q
dS
T
o
> (15)
So that the process is irreversible.
To account for the temperature variation during throttling, one may again resort on
infinitesimal isenthalpic processes, for which:

, , ,
0
P T T P
H H H
dH dT dP d
T P
v v
v
v
c c c | | | | | |
= + + =
| | |
c c c
\ . \ . \ .
. (16)
As the mass is constant, one readily obtains:

, , T P
H H
dT dP
P T
v v
(
c c | | | |
=
( | |
c c
\ . \ .

. (17)
The enthalpy derivatives can be obtained explicitly as follows:

( )
, , , ,
,
, , , , ,
T T T T
T
T T T V T
U PV
H H V V
P V P V P
U P V P V
P V T V
V V P T P
v v v v
v
v v v v v
| | c +
c c c c | | | | | | | |
= = =
|
| | | |
c c c c c
\ . \ . \ . \ .
\ .
(
c c c c c | | | | | | | | | |
= + + = +
(
| | | | |
c c c c c
\ . \ . \ . \ . \ .

. (18)
More over, since
, , ,
1
P V T
T P V
V T P
v v v
c c c | | | | | |
=
| | |
c c c
\ . \ . \ .
, it follows that
, , , V T P
P V V
T P T
v v v
c c c | | | | | |
=
| | |
c c c
\ . \ . \ .
, so
that we finally get from Eq. (18):
( )
, ,
1
P
T P
H V
T V V T
P T
v v
o
c c | | | |
= + =
| |
c c
\ . \ .
. (19)
Now, for the other derivative of the enthalpy in Eq. (17) we can most conveniently start from the
definition of the constant-pressure molar heat capacity, C
P
:

( ) ( )
( )
: :
:
:
:
P
P P
P T
P
P P T
P
P P P P
P P
C Q dT dU PdV dT
U U
dT dP PdV dT
T P
U V V
dT P dT P dP dT
T T P
PV
U V U H
dT P dT dT
T T T T T
v o = = + =
( c c | | | |
= + + =
( | |
c c
\ . \ .

( c c c | | | | | |
= + + =
( | | |
c c c
\ . \ . \ .

| | c ( c c c c | | | | | | | |
+ = + =
|
( | | | |
c c c c c
\ . \ . \ . \ .
\ .
(20)
Thus,using Eqs. (17), (19) and (20), the temperature variation in an infinitesimal Joule-Thomson
process reduces to:
( ) 1
P
P
dT T dP
C
u
o = , (21)
where u is the molar volume of the gas. The coefficient of dP in Eq. (21) is called Joule-Thomson
coefficient:
( ) 1
JT P
P
T
C
u
o . (22)
Owing to the fact that dP < 0 in Eq. (21), the gas is cooled following the Joule-Thomson effect
if the coefficient
JT
is positive. For an ideal gas, one readily finds that o
P
= 1/T, so that such a
system cannot be cooled or heated through Joule-Thomson expansion (dT = 0). On the contrary,
real gases obeying the Van der Waals equation of state may either be cooled or heated through such
a throttling process. For any given gas a temperature exists for which the Joule-Thomson coefficient
vanishes. This value of T is called inversion temperature, T
inv
. When T crosses the value of the
inversion temperature the Joule-Thomson coefficient changes sign and the effect of throttling is
reversed. It turns out that the
JT
is a decreasing function of temperature, so that if the T is below
the inversion temperature then the gas is cooling when subject to Joule-Thomson throttling. The
heating effect appears when T is above T
inv
. To see how this works, let us consider again Eq. (6) to
compute o
P
. For convenience, the derivative of the temperature with respect to the volume should
be taken:

( ) ( )
2 2 2
2 3 3
,
1 1
2 2
P
T RT
P a a V b a V b
V R V V R V b V
v
v v v v
v v
v v v
( ( c | |
= + =
| ( (
c
\ .

.
It proves more convenient to use the molar volume u to express the previous derivative:

2 3 2 3
,
1 1 2 1 2 2
1
1
P
T RT a ab RT b a ab
b
V R R
v
v u u u v u u u u
u
(
(
( c | | | |
= + ~ + +
( | | (
c
\ . \ .
(

(23)
where the approximation
1
1
1
b
b
u
u
~ +

has been used. This is a valid assumption since for gases

b u . Using Eq. (23) in the definition of the expansivity o
P
, one readily finds:

2 3
,
1 2 2 2
1 1 1
P P
T RT b a ab b a b
V T
V R RT
v
u
o u u u u u u u
( ( c | | | | | |
= = + + = +
| | | ( (
c
\ . \ . \ .

. (24)
Therefore, by replacing in Eq. (22), the Joule-Thomson coefficient takes the form:

1 2
1 1 1 1
2
1 1
JT
P P
b a b
b a b C C RT
RT
u u

u u u
u u u
(
(
(
| |
(
= ~ +
| (
| | ( \ .

+
|
(
\ .

,
since both b u and a RT u . It may be concluded that

1 2
1
JT
P
a b
b
C RT

u
(
| |
~
| (
\ .

. (25)
By further neglecting b/u with respect to 1 in Eq. (25) the following expression is obtained for the
Joule-Thomson coefficient of a Van der Waals gas:

1 2
JT
P
a
b
C RT

| |
~
|
\ .
. (26)
The corresponding inversion temperature becomes:

2
inv
a
T
Rb
= . (27)
It is now clear that for T < T
inv
the Joule-Thomson coefficient is positive and the gas is cooling.
Helium and hydrogen are two gases whose Joule-Thomson inversion temperatures at one
atmosphere are very low (e.g., about 222 C for helium). Thus, helium and hydrogen will warm
when expanded at constant enthalpy at typical room temperatures. For this reason, simple Linde
cycle liquefiers cannot normally be used to liquefy helium, hydrogen and neon. On the other hand
nitrogen has an inversion temperature of 621 K (348 C) and oxygen has an inversion temperature
of 764 K (491 C): the two most abundant gases in air can be cooled by the Joule-Thomson effect
at typical room temperatures.

Figure 3: PHYWE experimental setup for demonstrating the Joule-Thomson
effect.
Figure 3 shows a laboratory setup for demonstrating the Joule-Thomson effect. The cylinder
where the throttling takes place is made of glass (a sufficiently adiabatic enclosure for the
demonstrative purposes). The gas is injected from a pressurized canister through a long thermal
conductive metal pipe which is coiled on the base cylinder. This component is intended for
thermalizing the gas to the room temperature before throttling. The initial pressure of the gas is read
on a manometer and the pressure after the throttling is the ambient atmospheric one. The
temperature difference is measured by a couple of thermocouples imersed in the two compartments
of the throttling vessel. Typical values for the Joule-Thomson coefficient are of the order of 10
-5

K/Pa at the room temperatures. From Eq. (26) it may be seen that
JT
decreases with the
temperature, so the effect can be obtained more clearly at lower ambient temperatures.

Bibliography
[1] J. P. Joule, Philosophical Magazine, Series 3, XXVI, p. 369 (1845).
[2] W. Thomson and J. P. joule, Proceedings of the Royal Society, vol VII, p. 127 (1853).
[3] J. P. Joule and W. Thomson, Philosophical Transactions, vol. CXLIV, p. 321 (1854).