60 2 The First Law

When a system is heated, the total energy transferred as Self-test 2D.1 Calculate the values of q, w, and ΔU for an
heat is the sum of all individual contributions at each point of irreversible isothermal expansion of a perfect gas against a
the path: constant non-zero external pressure.
Answer: q = pexΔV, w = −pexΔV, ΔU = 0
f
q= ∫ dq  (2D.2)
i,path

Notice the differences between this equation and eqn 2D.1.

First, the result of integration is q and not Δq, because q is 2D.2 Changes in internal energy
not a state function and the energy supplied as heat cannot be
expressed as qf − qi. Secondly, the path of integration must be Consider a closed system of constant composition (the only
specified because q depends on the path selected (e.g. an adia- type of system considered in the rest of this Topic). The inter-
batic path has q = 0, whereas a non-adiabatic path between the nal energy U can be regarded as a function of V, T, and p, but,
same two states would have q ≠ 0). This path dependence is ex- because there is an equation of state that relates these quan-
pressed by saying that dq is an ‘inexact differential’. In general, tities (Topic 1A), choosing the values of two of the variables
an inexact differential is an infinitesimal quantity that, when fixes the value of the third. Therefore, it is possible to write U
integrated, gives a result that depends on the path between the in terms of just two independent variables: V and T, p and T,
initial and final states. Often dq is written đq to emphasize or p and V. Expressing U as a function of volume and tempera-
that it is inexact and requires the specification of a path. ture turns out to result in the simplest expressions.
The work done on a system to change it from one state to
another depends on the path taken between the two speci-
fied states. For example, in general the work is different if the
(a) General considerations
change takes place adiabatically and non-adiabatically. It fol- Because the internal energy is a function of the volume and
lows that dw is an inexact differential. It is often written đw. the temperature, when these two quantities change, the inter-
nal energy changes by

Example 2D.1 Calculating work, heat, and change in  ∂U   ∂U  General expression

dU =   dV +  dT  (2D.3)
 ∂T  V
for a change in U
internal energy  ∂V T with T and V

Consider a perfect gas inside a cylinder fitted with a piston. The interpretation of this equation is that, in a closed system of
Let the initial state be T,Vi and the final state be T,Vf. The
constant composition, any infinitesimal change in the internal
change of state can be brought about in many ways, of which
energy is proportional to the infinitesimal changes of volume
the two simplest are the following:
and temperature, the coefficients of proportionality being the
• Path 1, in which there is free expansion against zero two partial derivatives (Fig. 2D.2).
external pressure;
• Path 2, in which there is reversible, isothermal expansion.
Calculate w, q, and ΔU for each process.
Internal energy, U

Collect your thoughts To find a starting point for a calculation

in thermodynamics, it is often a good idea to go back to first
U
principles and to look for a way of expressing the quantity to
be calculated in terms of other quantities that are easier to
calculate. It is argued in Topic 2B that the internal energy of
a perfect gas depends only on the temperature and is inde-
pendent of the volume those molecules occupy, so for any
isothermal change, ΔU = 0. Also, ΔU = q + w in general. To
U+ ( ∂U
∂V )
T
dV + (
∂T )
∂U
V
dT

solve the problem you need to combine the two expressions, dT Temperature, T
selecting the appropriate expression for the work done from Volume, V dV
the discussion in Topic 2A.
The solution Because ΔU = 0 for both paths and ΔU = q + w,
in each case q = −w. The work of free expansion is zero (eqn Figure 2D.2 An overall change in U, which is denoted dU, arises
2A.7 of Topic 2A, w = 0); so in Path 1, w = 0 and therefore q = when both V and T are allowed to change. If second-order
0 too. For Path 2, the work is given by eqn 2A.9 of Topic 2A (w infinitesimals are ignored, the overall change is the sum of
= −nRT ln(Vf/Vi)) and consequently q = nRT ln(Vf/Vi). changes for each variable separately.
 2D State functions and exact differentials 61

Internal energy, U
Repulsions
πT dominant, πT < 0

Internal energy, U
U
Perfect gas

Attractions
dominant, πT > 0
Temperature, T
dV
Volume, V
Volume, V

Figure 2D.4 For a perfect gas, the internal energy is independent

Figure 2D.3 The internal pressure, πT, is the slope of U with of the volume (at constant temperature). If attractions are
respect to V with the temperature T held constant. dominant in a real gas, the internal energy increases with volume
because the molecules become farther apart on average. If
repulsions are dominant, the internal energy decreases as the gas
expands.
In many cases partial derivatives have a straightforward
physical interpretation, and thermodynamics gets shapeless
and difficult only when that interpretation is not kept in sight.
The term (∂U/∂T)V occurs in Topic 2A, as the constant-volume 22 atm and the other was evacuated. He then tried to meas-
heat capacity, CV . The other coefficient, (∂U/∂V)T , denoted πT , ure the change in temperature of the water of the bath when a
plays a major role in thermodynamics because it is a measure stopcock was opened and the air expanded into a vacuum. He
of the variation of the internal energy of a substance as its vol- observed no change in temperature.
ume is changed at constant temperature (Fig. 2D.3). Because The thermodynamic implications of the experiment are as
π T has the same dimensions as pressure but arises from the follows. No work was done in the expansion into a vacuum,
interactions between the molecules within the sample, it is so w = 0. No energy entered or left the system (the gas) as heat
called the internal pressure: because the temperature of the bath did not change, so q = 0.
Consequently, within the accuracy of the experiment, ΔU = 0.
 ∂U 
πT =   Internal pressure
Joule concluded that U does not change when a gas expands
 ∂V  T
(2D.4)
[definition]
isothermally and therefore that π T = 0. His experiment, how-
In terms of the notation CV and πT, eqn 2D.3 can now be ever, was crude. The heat capacity of the apparatus was so large
written that the temperature change, which would in fact occur for a
real gas, is simply too small to measure. Joule had extracted
dU = πT dV + CVdT (2D.5)
an essential limiting property of a gas, a property of a perfect
It is shown in Topic 3D that the statement πT = 0 (i.e. the gas, without detecting the small deviations characteristic of
internal energy is independent of the volume occupied by real gases.
the sample) can be taken to be the definition of a perfect
gas, because it implies the equation of state pV ∝ T. In mo-
lecular terms, when there are no interactions between the
molecules, the internal energy is independent of their separa-
tion and hence independent of the volume of the sample and
πT = 0. If the gas is described by the van der Waals equation
Vacuum
with a, the parameter corresponding to attractive interactions,
High pressure
dominant, then an increase in volume increases the average gas
separation of the molecules and therefore raises the internal
energy. In this case, it is expected that πT > 0 (Fig. 2D.4). This
expectation is confirmed in Topic 3D, where it is shown that
πT = na/V 2.
James Joule thought that he could measure π T by observ- Figure 2D.5 A schematic diagram of the apparatus used by Joule
ing the change in temperature of a gas when it is allowed to in an attempt to measure the change in internal energy when
expand into a vacuum. He used two metal vessels immersed a gas expands isothermally. The heat absorbed by the gas is
in a water bath (Fig. 2D.5). One was filled with air at about proportional to the change in temperature of the bath.
62 2 The First Law

(b)Changes in internal energy at tive sign in the definition ensures that the compressibil-
ity is a positive quantity, because an increase of pressure,
constant pressure implying a positive dp, brings about a reduction of volume,
Partial derivatives have many useful properties and some are a negative dV.
reviewed in The chemist’s toolkit 9 of Topic 2A. Skilful use of
them can often turn some unfamiliar quantity into a quantity
Example 2D.2
that can be recognized, interpreted, or measured. Calculating the expansion coefficient of
As an example, to find how the internal energy varies with a gas
temperature when the pressure rather than the volume of the Derive an expression for the expansion coefficient of a perfect
system is kept constant, begin by dividing both sides of eqn gas.
2D.5 by dT. Then impose the condition of constant pressure on
the resulting differentials, so that dU/dT on the left becomes Collect your thoughts The expansion coefficient is defined in
(∂U/∂T)p. At this stage the equation becomes eqn 2D.6. To use this expression, you need to substitute the
expression for V in terms of T obtained from the equation of
 ∂U   ∂V  state for the gas. As implied by the subscript in eqn 2D.6, the
 ∂T  = π T  ∂T  + CV pressure, p, is treated as a constant.
p p
The solution Because pV = nRT, write
As already emphasized, it is usually sensible in thermodynam-
ics to inspect the output of a manipulation to see if it contains pV = nRT pV = nRT

any recognizable physical quantity. The partial derivative on 1  ∂V  1  ∂(nRT / p)  1 nR nR nR 1

α= =  = V × p = pV = nRT = T
the right in this expression is the slope of the plot of volume V  ∂T  p V  ∂T p
against temperature (at constant pressure). This property is
normally tabulated as the expansion coefficient, α , of a sub- The physical interpretation of this result is that the higher the
stance, which is defined as temperature, the less responsive is the volume of a perfect gas
1  ∂V  to a change in temperature.
α= Expansion coefficient
V  ∂T  p  [definition]
(2D.6)
Self-test 2D.2 Derive an expression for the isothermal com-
pressibility of a perfect gas.
and physically is the fractional change in volume that accom- Answer: κT = 1/p
panies a rise in temperature. A large value of α means that the
volume of the sample responds strongly to changes in tem-
perature. Table 2D.1 lists some experimental values of α . For Introduction of the definition of α into the equation for
future reference, it also lists the isothermal compressibility, (∂U/∂T)p gives
κT (kappa), which is defined as
 ∂U 
 ∂T  = απ T V + CV (2D.8)
1  ∂V  Isothermal compressibility
p
κT = −  (2D.7)
V  ∂ p  T [definition] This equation is entirely general (provided the system is
closed and its composition is constant). It expresses the de-
The isothermal compressibility is a measure of the fractional pendence of the internal energy on the temperature at con-
change in volume when the pressure is increased; the nega- stant pressure in terms of CV, which can be measured in one
experiment, in terms of α , which can be measured in an-
other, and in terms of the internal pressure π T. For a perfect
Table 2D.1 Expansion coefficients (α) and isothermal
gas, π T = 0, so then
compressibilities (κT) at 298 K*
 ∂U 
−4
α/(10 K ) −1 −6
κT/(10 bar ) −1  ∂T  = CV (2D.9)
p
Liquids: That is, although the constant-volume heat capacity of a per-
Benzene 12.4 90.9 fect gas is defined as the slope of a plot of internal energy
Water 2.1 49.0 against temperature at constant volume, for a perfect gas CV is
Solids: also the slope of a plot of internal energy against temperature
Diamond 0.030 0.185 at constant pressure.
Lead 0.861 2.18 Equation 2D.9 provides an easy way to derive the relation
between Cp and CV for a perfect gas (they differ, as explained
* More values are given in the Resource section. in Topic 2B, because some of the energy supplied as heat
 2D State functions and exact differentials 63

escapes back into the surroundings as work of expansion when a function of p and T, and the argument in Section 2D.2 for
the volume is not constant). First, write the variation of U can be adapted to find an expression for the
variation of H with temperature at constant volume.
Definition
of C eqn 2D.9

   
p
 
 How is that done? 2D.1 Deriving an expression for the
 ∂H   ∂U 
C p − CV =  − variation of enthalpy with pressure and temperature
 ∂T  p  ∂T  p
Consider a closed system of constant composition. Because H
Then introduce H = U + pV = U + nRT into the first term and is a function of p and T, when these two quantities change by
obtain an infinitesimal amount, the enthalpy changes by
 ∂(U + nRT )   ∂U   ∂H 
dH = 
 ∂H 
dp +  dT
C p − CV =   −  ∂T  = nR(2D.10)  ∂ p 
T  ∂T  p
 ∂T p p

The general result for any substance (the proof makes use of The second partial derivative is Cp. The task at hand is to
the Second Law, which is introduced in Focus 3) is express (∂H/∂p)T in terms of recognizable quantities. If the
enthalpy is constant, then dH = 0 and
α 2TV
C p − CV =
κ T (2D.11)  ∂H 
 ∂ p  dp = − C pdT at constant H
T
This relation reduces to eqn 2D.10 for a perfect gas when α =
1/T and κT =1/p. Because expansion coefficients α of liquids Division of both sides by dp then gives
and solids are small, it is tempting to deduce from eqn 2D.11  ∂H   ∂T 
that for them Cp ≈ CV. But this is not always so, because the  ∂ p  = − C p  ∂ p  = − C p µ
T H
compressibility κT might also be small, so α 2/κT might be large.
That is, although only a little work need be done to push back where the Joule–Thomson coefficient, μ (mu), is defined as
the atmosphere, a great deal of work may have to be done to  ∂T  Joule–Thomson coefficient
µ= (2D.12)
pull atoms apart from one another as the solid expands.  ∂ p  H [definition]

It follows that
Brief illustration 2D.1
(2D.13)
dH = −μCpdp + CpdT
The expansion coefficient and isothermal compressibility of The variation of enthalpy with
water at 25 °C are given in Table 2D.1 as 2.1 × 10−4 K−1 and temperature and pressure
49.0 × 10−6 bar−1 (4.90 × 10−10 Pa−1), respectively. The molar vol-
ume of water at that temperature, Vm = M/ρ (where ρ is the Brief illustration 2D.2
mass density), is 18.1 cm3 mol−1 (1.81 × 10−5 m3 mol−1). Therefore,
from eqn 2D.11, the difference in molar heat capacities (which The Joule–Thomson coefficient for nitrogen at 298 K and
is given by using Vm in place of V) is 1 atm (Table 2D.2) is +0.27 K bar−1. (Note that μ is an intensive
property.) It follows that the change in temperature the gas
(2.1×10−4 K −1 )2 × (298K) × (1.81×10−5 m3 mol −1 ) undergoes when its pressure changes by −10 bar under isen-
C p ,m − CV,m =
4.90 ×10−10 Pa −1 thalpic conditions is
3 −1 −1 −1 −1
     = 0.485Pa m K mol = 0.485JK mol
∆T ≈ µ∆p = + (0.27 K bar −1 ) × (−10 bar) = − 2.7K
−1 −1 −1 −1
For water, Cp,m = 75.3 J K mol , so CV,m = 74.8 J K mol . In some
cases, the two heat capacities differ by as much as 30 per cent.
Table 2D.2 Inversion temperatures (TI), normal freezing (Tf ) and
boiling (Tb) points, and Joule–Thomson coefficients (μ) at 1 atm
and 298 K*
2D.3 Changes in enthalpy
TI/K Tf/K Tb/K μ/(K atm−1)
A similar set of operations can be carried out on the enthalpy, Ar 723 83.8 87.3
H = U + pV. The quantities U, p, and V are all state functions; CO2 1500 194.7 +1.10 +1.11 at 300 K
therefore H is also a state function and dH is an exact differ- He 40 4.2 4.22 −0.062
ential. It turns out that H is a useful thermodynamic function N2 621 63.3 77.4 +0.27
when the pressure can be controlled: a sign of that is the re-
lation ΔH = q p (eqn 2B.2b). Therefore, H can be regarded as * More values are given in the Resource section.