You are on page 1of 83

THERMODYNAMIC EVALUATION OF GREEN ENERGY

TECHNOLOGIES

COURSE READER

May 2016

Contents
1 Carbon Dioxide Separation and Entropy
1.1

1.2

The Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.1.1

The Thermodynamic Problem . . . . . . . . . . . . . . . . . .

1.1.2

The Ideal Gas Law . . . . . . . . . . . . . . . . . . . . . . . .

First Law of Thermodynamics . . . . . . . . . . . . . . . . . . . . . .

1.2.1

Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

10

1.2.2

Heat . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

11

1.2.3

Internal Energy and the First Law . . . . . . . . . . . . . . .

12

1.2.4

Heating an Object . . . . . . . . . . . . . . . . . . . . . . . .

13

1.2.5

Reversible and Irreversible Processes . . . . . . . . . . . . . .

15

1.2.6

Defining the Second Law . . . . . . . . . . . . . . . . . . . . .

17

1.3

Entropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

18

1.4

Minimum Work for Separating Carbon Dioxide . . . . . . . . . . . .

20

1.4.1

25

Minimum Work for Gas Separation . . . . . . . . . . . . . . .

2 Fuel Cells and Chemical Equilibria


2.1

27

Introduction to Fuel Cells . . . . . . . . . . . . . . . . . . . . . . . .

27

2.1.1

Thermodynamic Problem . . . . . . . . . . . . . . . . . . . .

29

Chemical Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

30

2.2.1

The Chemical Potential . . . . . . . . . . . . . . . . . . . . .

32

2.3

The Gibbs Free Energy . . . . . . . . . . . . . . . . . . . . . . . . . .

32

2.4

Chemical Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . .

37

2.5

Chemical Equilibrium in a Fuel Cell . . . . . . . . . . . . . . . . . . .

39

2.2

iv

2.6

Voltage of a Fuel Cell . . . . . . . . . . . . . . . . . . . . . . . . . . .

41

2.6.1

A Note on Units . . . . . . . . . . . . . . . . . . . . . . . . .

43

2.6.2

Dependence of Voltage on Partial Pressure . . . . . . . . . . .

44

2.6.3

Voltages with Different Fuels . . . . . . . . . . . . . . . . . . .

45

2.7

Law of Mass Action . . . . . . . . . . . . . . . . . . . . . . . . . . . .

47

2.8

Fuel Cell vs. Heat Engine . . . . . . . . . . . . . . . . . . . . . . . .

48

3 Solar Thermal, Phase Transitions, Heat Engines

50

3.1

The Three Phases of Water . . . . . . . . . . . . . . . . . . . . . . .

52

3.2

Phase Transition in Water . . . . . . . . . . . . . . . . . . . . . . . .

55

3.3

The Clausius-Clapeyron Relation . . . . . . . . . . . . . . . . . . . .

58

3.4

Evaporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

60

3.5

The Carnot Engine . . . . . . . . . . . . . . . . . . . . . . . . . . . .

63

3.5.1

Isothermal Expansion of Gas . . . . . . . . . . . . . . . . . . .

63

3.5.2

The Carnot Cycle . . . . . . . . . . . . . . . . . . . . . . . . .

66

The Carnot Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . .

68

3.6.1

Stacking a Carnot Heat Engine and a Heat Pump . . . . . . .

69

3.6.2

Case Studies of a Carnot Engine and a Heat Pump . . . . . .

71

3.6.3

Carnot Engine as the Most Efficient Heat Engine . . . . . . .

73

3.6.4

for a Heat Pump . . . . . . . . . . . . . . . . . . . . . . . .

73

Efficiency of a Solar Thermal Power Plant . . . . . . . . . . . . . . .

74

3.6

3.7

List of Tables
2.1

Chemical species and reactions in a fuel cell . . . . . . . . . . . . . .

40

3.1

Table describing the four parts of the Carnot cycle given in Fig. 3.10.

67

vi

List of Figures
1.1

Correlation between fossil fuel combustion and atmospheric concentration of CO2 . Adapted and used with permission from M. L. Machala,
Carbon-neutral Synthetic Fuels, large.stanford.edu/PHYS240. . . . . .

1.2

Correlation between atmospheric concentration of CO2 , CH4 and T.


Courtesy of the Intergovernmental Panel on Climate Change, 2001;
http://www.ipcc.ch, N. Oreskes, Science 306, 1686, 2004; D. A. Stainforth et al, Nature 433, 403, 2005. . . . . . . . . . . . . . . . . . . . .

1.3

Minimum work required to separate one mole of CO2 at difference concentrations of CO2 . From Wilcox, J. Carbon Capture, 2012, Springer
Publishing. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.4

The initial state of mixed gases and the final state with CO2 separation
and other gases. From Wilcox, J. Carbon Capture, 2012, Springer
Publishing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.5

Representation of differential volume dV . . . . . . . . . . . . . . . .

1.6

Schematic of the difference between heat and work

1.7

Representation of how the internal energy of a closed system may change. 12

1.8

Depending on the arrangement of the molecules, the entropyor disorder

. . . . . . . . . .

11

can differ. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

13

A partitioned box containing carbon dioxide and vacuum. . . . . . . .

15

1.10 A gas lattice with J sites and N molecules. . . . . . . . . . . . . . . .

19

1.11 The processes of mixing and demixing gases. . . . . . . . . . . . . . .

20

1.9

1.12 A barrier divides moles of CO2 with 1 moles of other gas. Both
gases have the same pressure. . . . . . . . . . . . . . . . . . . . . . .

vii

22

1.13 The entropy of mixing as a function of the CO2 fraction. . . . . . . .

24

1.14 The specific entropy of mixing as a function of the CO2 fraction. . . .

24

1.15 The partial molar entropy as a function of the CO2 fraction. . . . . .

25

1.16 The minimum work required to separate a mol of CO2 as a function


of the CO2 fraction. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1

Hydrogen and oxygen gas separated by a membrane that is permeable


to protons but not electrons . . . . . . . . . . . . . . . . . . . . . . .

2.2

28

A fuel cell with a completed circuit, in the form of a light bulb. The
net reaction is 2H2 + O2 2H2 O. . . . . . . . . . . . . . . . . . . . .

2.3

26

29

Hydrogen and oxygen gas at the same temperature and pressure, separated by a barrier. All walls, including the internal purple one, are
impermeable to heat, work and matter. . . . . . . . . . . . . . . . . .

2.4

30

Using the slope and intercept, we can map the monotonically-decreasing


function U to a value of G for every value of U and V . It is important
for U to be monotonically decreasing such that there is only one value
of V for every value of U . . . . . . . . . . . . . . . . . . . . . . . . .

34

2.5

An illustration of a system in a heat and work reservoir. . . . . . . .

36

2.6

The voltage of a fuel cell as a function of the pressure. The intercept


is given by 0rxn and can be changed by modifying the chemistry of the
reaction. Increasing the reactant concentration increases the voltage,
while increasing the product concentration decreases it. . . . . . . . .

2.7

45

The efficiency of a fuel cell as compared to the Carnot efficiency of a


heat engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

48

3.1

Schematic of a solar thermal power plant. This module will focus on


51

3.2

the thermodynamics of phase transitions and heat engines. . . . . . .


as a function of the temperature
The specific Gibbs free energy (G)
at a certain
(T ). The stable phase of water is the one with the lowest G
and S.
. . . .
T . Here, we neglect the temperature dependence of H

53

3.3

Schematic of Gibbs free energy before and after boiling. . . . . . . . .

55

3.4

Gibbs free energy of water vapor and liquid water during boiling. . .

57

viii

3.5

Sketch of the Clausius-Clapeyron relation given by equation 3.13. The


boiling temperature rises with higher pressure. . . . . . . . . . . . . .

3.6

60

Schematic of boiling from vacuum (top) and evaporation from a dry


environment (bottom) with liquid water at room temperature. In both
cases, the liquid water evaporates/boils into water vapor (left) initially,
but the rate of evaporation (blue arrows) and condensation (red arrows) eventually equilibrate after waiting a long time, leaving a partial
pressure of water equal to the water vapor pressure. . . . . . . . . . .

3.7

62

A piston in a heat reservoir isothermally expanding. Here the wall of


the piston is impermeable to matter but allows heat transfer from the
reservoir. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.8

Pressure, volume, temperature, and entropy during isothermal expansion of a gas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.9

64
65

A piston that isothermally expands and compresses between two pressures. No net work is being done in a cycle. . . . . . . . . . . . . . .

66

3.10 The T vs S diagram of the Carnot engine. . . . . . . . . . . . . . . .

67

3.11 The Carnot efficiency, defined in equation 3.15, as a function of the


hot reservoir temperature. . . . . . . . . . . . . . . . . . . . . . . . .

68

3.12 A heat pump and a heat engine stacked. The work from the heat
engine is being used to power the heat pump. Here, QH,1 and QC,2
are positive because heat flows out, whereas QC,1 and QH,2 are
negative as heat flows in. . . . . . . . . . . . . . . . . . . . . . . . . .

69

3.13 The maximum efficiency of a solar thermal power plant at different


temperatures and light concentration (C). This expression accounts
for both efficiency of sunlight to heat, after account for re-radiation,
and the efficiency of the heat to work through a Carnot engine. . . .

ix

75

Chapter 1
Carbon Dioxide Separation and
Entropy
1.1

The Problem

Figure 1.1: Correlation between fossil fuel combustion and atmospheric concentration
of CO2 . Adapted and used with permission from M. L. Machala, Carbon-neutral
Synthetic Fuels, large.stanford.edu/PHYS240.

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

Industrialized and developing nations emit tens of millions of tons of carbon dioxide (CO2 ) into the atmosphere every year. Natural processessuch as CO2 uptake
by biomass through photosynthesis and absorption by oceanscannot recapture and
sequester CO2 at the rate humans are producing it, particularly through the combustion of fossil fuels. This leads to an increase in CO2 concentration in the atmosphere
shown in Fig. 1.1.

Figure 1.2: Correlation between atmospheric concentration of CO2 , CH4 and T. Courtesy of the Intergovernmental Panel on Climate Change, 2001; http://www.ipcc.ch,
N. Oreskes, Science 306, 1686, 2004; D. A. Stainforth et al, Nature 433, 403, 2005.
It has been shown that atmospheric greenhouse gas concentration track average global
temperature. Figure 1.2 shows the correlation over 0.4 million years between concentrations of methane (CH4 ) and CO2 in relation to temperature, T . However, atmospheric CO2 levels never surpassed 325 parts per million by volume (ppmv) during
these hundreds of millennia, until the modern day. In 2015, CO2 concentrations levels
exceed 400 ppmv.
As anthropogenic climate change and its adverse effects have been linked to a green
house gas emissions (largely CO2 ), many have proposed capturing and sequestering
CO2 from the atmosphere as well as from industrialized processes (e.g. from flue

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

gas) before it escapes into the atmosphere. But which process is more reasonable
both economically and from an energetic perspective? Figure 1.3 shows the minimum

Figure 1.3: Minimum work required to separate one mole of CO2 at difference concentrations of CO2 . From Wilcox, J. Carbon Capture, 2012, Springer Publishing.
amount of energy it takes to extract 1 mol of CO2 as a function of CO2 concentration in
the original gas. Notice the stark difference between the energy needed to remove one
mole of CO2 at atmospheric concentrations versus industrial flue gas concentrations.
During this chapter you we learn how to construct this graph and will be able to
evaluate the thermodynamic viability of one capture process versus another.

1.1.1

The Thermodynamic Problem

As we will show in this module, principles of thermodynamics sets the absolute minimum amount of energy required to capture and separate CO2 . Lets begin by precisely
defining the problem. Initially, CO2 is mixed with other gases. In the atmosphere,
this is mostly N2 and O2 , with CO2 concentration 0.04 vol%. At the smokestack of
a coal combustion power plant, CO2 is mixed with H2 O and N2 , and has a concentration typically in the range of a few tens of percent. The process of capture and
separation will separate CO2 into a pure stream from the other gases.

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

Figure 1.4: The initial state of mixed gases and the final state with CO2 separation
and other gases. From Wilcox, J. Carbon Capture, 2012, Springer Publishing
Lets more precisely define the problem using initial and final state. Because thermodynamics does not describe the time-dependent behavior of a system, we will describe
the composition of the gas mixtures rather than the flow rates. Specifically, we can
use a closed volume of gas with appropriate internal partitioning to describe the initial and final states of the system (Fig. 1.4).
The initial state is a single volume of gas consisting of CO2 mixed with other gases
(which we are not specifying). The final state is the same volume but divided into
two partitions with a wall that does not allow the gas to go back and forth: the first
chamber contains pure CO2 , while the second chamber contains the other gases.
Our goal is to calculate the minimum amount of energy needed to carry out the
separation process per mole of CO2 . This represents the absolute best case scenario.
As we will show later, a process requiring energy less than this minimum value is
thermodynamically impossible and is a perpetual motion machine.

1.1.2

The Ideal Gas Law

Before discussing such topics as the First and Second Law of Thermodynamics in the
context of CO2 capture, we must first define an important relation between variables
that we can control in a gaseous systemnamely pressure, volume, number of gas

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

molecules, and temperature. An equation relating these variables is known as the


Ideal Gas Law. Recall that pressure (P) is defined as force applied per unit area:
P=

Force
Area

(1.1)

and that momentum is defined as:


Momentum = Force Time

(1.2)

Combining Equations 1.1 and 1.2 gives another relation for pressure
P =

Momentum
Force
=
Area
Time Area

(1.3)

Consider a gas molecule moving in a box of volume V. This molecule bounces around
inside the box hitting the walls periodically. As the molecule bounces it imparts
outward momentum to the walls of the box. If we add trillions of gas molecules to
the box, the momentum imparted the walls increases per equation 1.3. This outward
momentum is what we experience as pressure. But how does this relate to temperature and number of gas molecules?
First, we make three assumptions. The molecules:
1. collide with the wall elastically, meaning no energy is lost
2. are very small
3. do not interact with one another
These assumptions define an ideal gas.
The rate at which molecules collide with the wall depends on the volume of the box
and the velocity of the gas molecules. For a given period of time, only a fraction of
the gas molecules in the box will strike the wall, transfer momentum, and contribute

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

to the pressure. Therefore, after Eq. 1.3, pressure is given by:


(Average momentum of the gas molecules)

(1.4)

(Fraction of gas molecules striking the wall per second per area)

Figure 1.5: Representation of differential volume dV


We can determine the differential volume dV of gas that will strike the wall of area
dA in one second (Fig. 1.5). To simplify the geometry, take dA to lie on the y-z plane.
The size of this prismatic differential volume, dV , is given by the product of dA and
the length, l, the latter of which is the velocity of the molecule in the x-direction, vx ,
times one second. The fraction of gas in this differential volume is given by:
dV
dA x (length)
dA x (vx x 1s)
=
=
V
V
V

dV
V

.
(1.5)

Thus, the fraction of gas that strikes dA per second is


v 
dV
x
=
dA
V xs
V

(1.6)

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

Since we assumed that the collision of a gas molecule with the wall is elastic, the
kinetic energy of the molecule is conserved and the initial velocity of the molecule
vx will rebound with vx . By conservation of momentum and per Equation 1.2 the
momentum imparted to the wall is 2mvx .
Substituting this result and Eq. 1.6 into Eq. 1.4, we obtain:
P =

X 1
i

!
2mvx,i

summed over the number of molecules. The

 v 
x,i

V
1
2

2
X mvx,i
i

(1.7)

accounts for the probability of half

of the molecules striking one wall, dA and the other half striking the opposite wall,
which does not exist in reality.
Now we consider the internal energy (U ) of the gases in the box defined as the
translational kinetic energy of the total number of gas molecules, N .
1X
mvi2
2 i
1X
2
2
2
=
m(vx,i
+ vy,i
+ vz,i
)
2 i

U=

(1.8)

If we sum over all of the molecules in the box, N , and consider the average kinetic
energy of all of the gas molecules, < KE >trans = 21 m < v >2 , then we can rewrite
Equation 1.8 as
U = N < KE >trans

(1.9)

Notice similarities between Eqns. 1.7 and 1.8. Lets begin by manipulating Equation
1.7which included the momentum of the gas molecules striking the wall. Because we
only consider the x-direction, we multiply the righthand side of Eq. 1.7 by a factor

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

of 31 assuming the molecules move in all directions randomly.


P =

2
X mi vx,i

1
=
3

2 X mi vi2
2 i V

2
=
3

1 X mi vi2
2 i V

Substituting Eqns. 1.8 and 1.9 yields


P =

2
X mi vx,i
i

2
=
3

N < KE >trans
V
 
2 U
=
3 V

Lets do a sanity check to confirm that

U
V


(1.10)

has the same unit as pressure

U
Energy
=
V
Volume
Force Length
=
Volume
Force
=
Area

(1.11)

Rearranging the equation for pressure gives


2
PV = U
3

(1.12)

Its helpful to rescale U to a convenient value with which we are familiar. Lets define
temperature as being linearly proportional to < KE >trans :
U T

(1.13)

Moreover, we select a prefactor for T to scale with the boiling point and freezing
point of water, such that
H2 O
2O
Tboiling
TfHreezing
= 100

(1.14)

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

and results in
3
N < KE >trans = U = N kB T
2

(1.15)

This is where the Boltzmann constant (kB ) originated. If we substitute Eq. 1.15 into
1.12 we obtain
P V = N kB T

(1.16)

The Boltzmann factor is related to another variable to which you may be familiarthe
universal gas constant, R, by a factor of Avagadros number, NA , or one mole.
R = NA kB

(1.17)

Substituting Eq. 1.17 into 1.16 gives


PV =

N
RT = nRT
NA

(1.18)

where n is the number of moles of gas molecules. Thus, the Ideal Gas Law can be
written as
P V = nRT

(1.19)

We have shown how pressure, volume, number of molecules, and temperature are
related for an ideal gas and that pressure exerted on a container is due to the thermal
motion of gas molecules.

1.2

First Law of Thermodynamics

There are two ways that energy can be transferred into and out of a thermodynamic
system: work and heat.

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

1.2.1

10

Work

Work is the transfer of energy through directed motion. In our convention, we


define positive work as work done by the system on the surrounding (e.g. a piston
the systemexpanding against atmospherethe surrounding), and negative work is
done by the surrounding on the system (a piston being compressed). Figure 1.6
shows a schematic of work. Recall from mechanics that work W done is given by
W = F x

(1.20)

where F is the force and x the displacement. Here the operator refers to the
difference between the initial and final state. In differential form, this is written as
dW = F dx
= P Adx

(1.21)

= P dV
Here, P is the pressure (which is defined as

F
,
A

or force per unit area), A the area, and

V the volume. This expression states that the differential work equals the pressure
multiplied by the differential change in volume.
For example, consider doing work on a piston that is held at constant temperature
(e.g. by contacting it to a thermal reservoir), Fig. 1.6. From the Ideal Gas Law, we
know that P V = nRT = constant. Substituting this relation for P (V ), we can solve
for the work done by the system when the piston expands from V1 to V2
Z
W =

dW
Z

V2

nRT
dV
V
V1
 
V2
= nRT ln
V1
=

This is known as isothermal expansion.

(1.22)

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

1.2.2

11

Heat

Figure 1.6: Schematic of the difference between heat and work


Heating, on the other hand, is the transfer of energy through random motion. An
example is the random motion of gas molecules. The sign for heat is positive for heat
added into the system.
The difference between heat and work is that for work, the atoms are all traveling
in the same direction. For heat, on the other hand, the molecules are traveling
in random directions. Both heat and work added to a system involve a transfer of
kinetic energy, but they are in different directions. Figure 1.6 shows a schematic of the

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

12

difference between heat and work. For work, gas molecules can all move in a specific
directionthe piston moving upwardand energy in the form of work is transferred for
this movement. When heat is added to the system, the random velocities of the gas
molecules increases.

1.2.3

Internal Energy and the First Law

Change in the internal energy (dU) = Heat In (dQ) - Work Done (dW)
dU = dQ dW

(1.23)

Figure 1.7: Representation of how the internal energy of a closed system may change.
In this class we will use the following definitions to classify systems: open, closed,
and isolated. An open system allows the passage of heat, work and mass into and out
of the system. A closed system is impermeable to mass but still allows the transfer
of heat and work. An isolated system allows nothing to pass in and out. Consider
the isolated system shown in Fig. 1.7 where there is no energy flow into or out of the
system. There is a closed system contained within this larger, isolated system. The

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

13

first law of thermodynamics states that


dU = (dQ1 dQ2 ) (dW1 dW2 )

(1.24)

In other words, the total energy of an isolated system is conserved. The net heat
and work coming into and out of the inner box must be zero. Since the universe is a
isolated system, the total energy of the universe is conserved.

1.2.4

Heating an Object

To heat up a gas is to increase the random motion of the gas molecules. There are
two components to heating:
1. Increasing randomness or disorder of the gas molecules
2. Increasing the average velocity (< KE >) of the gas molecules
Lets consider four indistinguishable molecules sitting in a 4x4 lattice, shown in Fig.
1.8, with each lattice position identical to the other (i.e. no edge or corner effects).

Figure 1.8: Depending on the arrangement of the molecules, the entropyor disorder
can differ.

(A): the four molecules are clustered in groups of 4


(B): they are clustered in groups of 2
(C): there is no cluster (only groups of 1)

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

14

For simplicity, assume that the clusters cannot rotate.


The measure of disorder is related to the number of unique ways to arrange the
molecules in the lattice. The greater the number of unique arrangements, the greater
the disorder. We count the number of possible arrangements in Fig. 1.8.
(A): 9 Ways
(B): 12 12 = 144 ways
(C) 16 16 16 16 = 65536 ways
Note that we are overcounting here by allowing molecules to occupy the same lattice
position (because they are identical). We are ignoring the few cases in (B) where the
configuration is in clusters of 4, as well as for (C).
Now, suppose that the average kinetic energy < KE > for (A), (B), and (C) are all
equal, meaning the molecules have the same average velocity. Which of the following
transformations requires a greater amount of heat?
(A) (B), or
(A) (C)
Recall that heating is the increase of random motion. Since the average speed of
the three systems are the same, the change in the randomness also determines the
extent of heating. Thus, (A) (C) requires a greater amount of heating.
From this, we make the following definition:
Heating = < KE > Change in Disorder
In differential form, heating is represented by dQ, average kinetic energy is represented by T (at constant temperature) and changes in disorder are represented by a
change in entropy (dS). Here, S is the entropy which is a measure of disorder. We

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

15

will give a more quantitative definition later.


In the following equation we substitute variables in for < KE >, disorder and heating.
dQ = T dS

1.2.5

(1.25)

Reversible and Irreversible Processes

Figure 1.9: A partitioned box containing carbon dioxide and vacuum.


As we will show through an example, Eq. 1.25 only holds under very specific conditions. Consider a partitioned box (Fig. 1.9) with walls impermeable to energy a
matter, where the left partition contains a gas (e.g. CO2 ) and the right partition
contains no gas (i.e. vacuum). Now, lets break the wall. The gas from the left
partition expands to fill the entire box. This process is known as the free expansion
of gas and results in the increase of disorder in the systemthe number of potential
gas lattice sites has increased. Therefore, the entropy of the system increases and
S = Sf inal Sinit > 0. In contrast to isothermal compression or expansion (see
1.2.1), no work is done as we assume that breaking the wall requires negligible work,
like popping a balloon. So we have:
W = 0

(1.26)

Again, refers to the difference between the final and initial states. Moreover,

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

16

because the box is an isolated system, the internal energy remains constant:
U = 0

(1.27)

Combining these two equalities with the First Law of Thermodynamics, we obtain:
Q = U + W = 0

(1.28)

This result contradicts with Eq. 1.25 and our earlier statement that the entropy of
the system increases when the wall is broken. It turns out Eq. 1.25 only applies to
a reversible process, that is, a process that can be reverted back to the initial state
without doing additional work. In this free expansion of gas thought problem, it is not
a reversible process because no work is needed to expand the gas, yet work is required
to compress the gas back to the initial state. There are a few more equivalent ways
to define an irreversible process:
Occurs in one direction
Occurs spontaneously without work being done on the system
A reversible process, at any given point during the process, must not be far from
equilibrium (where equilibrium reflects a state of balance, with no net flow of matter
or energy). In other words, a reversible process occurs sufficiently slowly such that
the system is essentially in equilibrium at all time. This is also known as a quasistatic
process.
Note that a reversible process must be quasistatic, but not all quasistatic processes are
reversible. As an example, consider the system in Fig. 1.9. Now as a thought problem, add a special valve on the partition wall which lets one molecule flow through at
a time. The system is near equilibrium (i.e. static) at all times, but the gas expands
spontaneously and irreversibly as work must be done to remove the gas molecules.
Returning to Eq. 1.25, it only applies to reversible processes. Irreversible processes,

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

17

such as the free expansion of gas, cause the entropy to increase more than the heat
added to the system. Therefore, we modify Eq. 1.25:
dQ T dS

(1.29)

with dQrev = T dS

(1.30)

where the subscript rev indicates a reversible process.

1.2.6

Defining the Second Law

There are three ways to state the Second Law of Thermodynamics


An isolated system always tends to a configuration that maximizes disorder
(entropy)
The disorder of an isolated system (such as the universe) increases due to occurrence of spontaneous and irreversible processes
dQ/T dS, where dQrev /T = dS is the reversible limit
Whereas the First Law specifies the conservation of energy, the Second Law specifies
the direction of the process. For example, the Second Law says that gas expands when
a barrier to vacuum is broken. The First Law does not state the order of events, only
that the internal energy of the system is the same whether or not it is confined by
the barrier.
How do reversibility and the Second Law connect to carbon dioxide separation?
For a closed system, the First Law tells us that the internal energy is fixed, thus
dU = dQ dW = 0
dQ = dW

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

18

The Second Law states that the disorder of a closed system stays constant or increases
but never decreases.
dQ T dS

(1.31)

Combining the two, we obtain the final expression for a closed system, where the
differential work is less than the temperature multiplied by the differential entropy.
dW T dS

(1.32)

In other words, to decrease the entropy of the system, such as separating a gas mixture
containing carbon dioxide, oxygen, and nitrogen to just carbon dioxide and oxygen
and nitrogen, work must be done on the system (dW < 0). Specifically, the minimum
work needed on the system is given by
dWmin = T dS = dQrev

(1.33)

Equation 1.33 is the work required for a reversible process. Any irreversibility would
increases the amount of work needed to decrease entropy.

1.3

Entropy

To evaluate the minimum work to separate and capture carbon dioxide, we need to
understand entropy, which is the measure of disorder of a system.
Consider the lattice gas in Fig. 1.10, where there are J sites and N molecules.
Assuming that each lattice position can be occupied by at least one molecule, there
are J ways to place the first molecule and J 1 ways to place the second molecule,
etc. The number of ways to place N molecules in J sites is given by
=

J!
(J N )!

(1.34)

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

19

Figure 1.10: A gas lattice with J sites and N molecules.


where the exclamation mark ! represents the factorial. If the molecules are indistinguishable, then we overcounted by N !. Accounting for this, Eq. 1.34 becomes:
=

J!
(J N )!N !

(1.35)

Thus, is the number of ways to configure indistinguishable gas molecules in the lattice. For gas, the size of each equivalent cell is very small, which can be obtained
from quantum mechanics.
Boltzmann postulated that the entropy is related to the number of configurations of
the molecules in the lattice.
S = kB ln

(1.36)

Here, kB is the Boltzmann constant, which is 1.381023 JK 1 . For indistinguishable


molecules, it is given by Eq. 1.35. Combining Eqns. 1.35 and 1.36 yields the following
expression for entropy

S = kB ln

J!
(J N )!N !


(1.37)

Using the Stirling approximation, ln N ! N ln N N , which holds for large values

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

20

of N
S = kB (J ln J J (J N ) ln(J N ) + J N N ln N + N )
= kB (J ln J (J N ) ln(J N ) N ln N )


N
J
N ln
= kB J ln
J N
J N

(1.38)

If J  N , the first term in the expression approaches kB Jln(1) = 0. This approximation is valid when the number of gas molecules is much less than the number of
lattice sites (i.e. low density), which is typically true for a gas except under very high
pressures. Thus, in this limit:
S = N kB ln

N
J

(1.39)

This expression describes the entropy of a gas in a lattice.

1.4

Minimum Work for Separating Carbon Dioxide

As described in the problem statement, carbon dioxide separation can be depicted as


follows:

Figure 1.11: The processes of mixing and demixing gases.


We want to calculate the minimum work needed to carry out the demixing (separation) process. For an isolated system, the First Law states that dU = dQ dW = 0.
The Second Law states that dQ T dS. Combining this with the First Law yields

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

21

dW T dS. We call the lower bound the minimum work, Wmin .


During gas demixing, the system goes from the final state to the initial state shown
in Fig. 1.11, whereas during mixing, the system goes from the initial state to the final
state. Because demixing and mixing involves two identical end states, the associated
entropy changes are precisely opposite of one another:
Sdemix = Smix

(1.40)

Conceptually, it is easier to think about mixing, so we write:


Wmin = T dSmix

(1.41)

The minimum work is the work required to carry out the process reversibly. To
evaluate the minimum work, we define the system depicted in 1.11 with the following
parameters:
= volume fraction of CO2
ntot = total moles of gas
Vtot = total volume of gas
Conceptually, what happens to the gas when they mix, as shown in Fig. 1.11?
1. The number of moles of the each gas component does not change, assuming no
exchange of matter with the outside world and no chemical reactions.
2. While the number of moles does not change, the volume occupied by each gas
changes.
Recall from Eq. 1.39, S = N kB ln NJ , where N is the number of molecules and J is
the number of gas lattice sites. Remember that we can convert N kB = nR, where n
is the number of moles, N is the total number of molecules, and R is the gas constant.

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

22

Since N is proportional to n and J is proportional to V , we can rewrite Eq. 1.39 as:


S = nR ln

n
+ constant
V

(1.42)

where the constant represents the proportionality constant contained in the log. Thus,
as the volume expands at constant n and T , the entropy increases. This makes sense
because increasing the volume increases the number of lattice sites, decreases the
fraction of sites occupied, and thus, increases the disorder of the gas. Applying the
ideal gas law, P V = nRT , we obtain:
S = nR ln P + constant

(1.43)

Fig. 1.12 depicts several partitioned boxes, each with a different volume fraction of
CO2 , defined as or

VCO2
,
Vtot

where VCO2 is the volume of CO2 and Vtot is the total volume

of the all gas. Both chambers have the same pressure. Let us consider what happens
qualitatively after we break the wall as a function of , in each scenario.

Figure 1.12: A barrier divides moles of CO2 with 1 moles of other gas. Both
gases have the same pressure.

1. Scenario A: CO2 expands and its volume doubles, and so does the other gas
2. Scenario B: CO2 expands by a factor of 4, where the other gas expands by 33%.
3. Scenario C: CO2 expands by a factor of 10, where the other gas expands by
11%.

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

23

Thus, we see that the extent of volume expansion for each gas depends on , the
volume fraction of CO2 . Let us look at the entropy of mixing:
Smix = Sf inal Sinit

(1.44)

where refers to the difference before and after breaking the wall. We consider the
entropy of each component:
CO2
other
Sinit = Sinit
+ Sinit

 

ntot
(1 )ntot
= ntot R ln
+ constant + (1 )ntot R ln
+ constant
Vtot
(1 )Vtot
ntot
= ntot R ln
+ constant
Vtot

where ntot is the total number of moles of gas molecules. After mixing by breaking
the wall
other
2
Sf inal = SfCO
inal + Sf inal
 


(1 )ntot
ntot
+ constant + (1 )ntot R ln
+ constant
= ntot R ln
Vtot
Vtot


ntot
= (ntot R ln ) + ((1 )ntot R ln(1 )) + ntot R ln
+ constant
Vtot

It can be shown that the constant term is equal for both Sinit and Sf inal . Thus, we
can write the entropy of mixing as
Smix = Sf inal Sinit = ntot R ( ln + (1 ) ln(1 ))

(1.45)

Does this make sense? Lets consider the boundary cases. When = 0 or = 1,
Smix = 0. In other words, if there is no wall to begin with, there is no change in
the entropy of the system.
The first term represents the entropy gained by expanding CO2 . The second term
represents the entropy gained by the other gas. To understand the behavior of the

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

24

entropy of mixing at intermediate values of , lets plot Smix against nCO2 (Fig.
1.13).

Figure 1.13: The entropy of mixing as a function of the CO2 fraction.

1. At =

nCO2
ntot

= 0 or 1, Smix = 0 because the system contains either CO2 or

the other gas.


2. At =

nCO2
ntot

= 0.5, both gases expands by the same amount and maximizes

Smix .
The entropy of mixing, as plotted, is not normalized to the size of the system. This
makes comparisons across multiple situations difficult. For example, we are interested
in calculating the minimum work to separate CO2 per mole of CO2 .

Figure 1.14: The specific entropy of mixing as a function of the CO2 fraction.

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

Therefore, we also consider S mix =

Smix
.
nCO2

25

This is known as the specific entropy,

plotted in Fig. 1.14.


S mix =

Smix
1
= R ( ln + (1 ) ln(1 ))
nCO2

(1.46)

Figure 1.15: The partial molar entropy as a function of the CO2 fraction.
As the fraction of CO2 increases, the magnitude of the specific entropy of mixing (and
demixing) tends to zero. Finally, sometimes its useful to calculate the incremental
change in Smix per incremental addition of CO2 . This is known as the partial molar
entropy and is given by the derivative of Smix with respect to nCO2 , or

Smix
,
nCO2

and

is plotted in Fig. 1.15.

1.4.1

Minimum Work for Gas Separation

To wrap up, we now have enough information to determine the minimum work required to separate carbon dioxide as a function of the number of moles of gas ntot
and the fraction of carbon dioxide .
The minimum work to separate a mole of gas into CO2 and other gases is given
by:
Wmin = T Sdemix = T Smix
= ntot RT ( ln + (1 ) ln(1 ))

(1.47)

CHAPTER 1. CARBON DIOXIDE SEPARATION AND ENTROPY

26

The minimum work for a single mole of carbon dioxide is given by


Wmin =

1
RT ( ln + (1 ) ln(1 ))

(1.48)

Here, the unit is J mol1 CO2 . This equation shows that it takes more work to capture a mole of CO2 from a source with low CO2 concentrations than from a source
with a high CO2 concentration. Thus, capturing a mole of CO2 from flue gas would
require less work than capturing a mole of CO2 from the atmosphere.
Sometimes, its also convenient to express in terms of partial pressures instead of volume fractions, where pCO2 = ptot and pother = (1 )Ptot . Note that pCO2 + pother =
Ptot , where Ptot is the total pressure.
The minimum work for a mol of CO2 is given by

Wmin = RT

pCO2 pother pother


ln
+
ln
Ptot
pCO2
Ptot


(1.49)

This is plotted in Fig. 1.16.

Figure 1.16: The minimum work required to separate a mol of CO2 as a function of
the CO2 fraction.

Chapter 2
Fuel Cells and Chemical Equilibria
2.1

Introduction to Fuel Cells

Modern society requires a steady supply of electricity. There are many ways to
generate electricity, including hydroelectric dams, wind turbines, and solar cells. The
dominant source of electricity arises from burning of fossil fuels like coal or natural
gas, and then using a heat engine to convert the heat of combustion to electricity.
As we will learn in module 3, such heat engines are often fairly inefficient (around 30
% heat to electrical), and cannot be scaled down: for example, you cannot create a
small and efficient natural gas heat engine to power your computer.
An alternative way to convert fuel to electricity is a fuel cell. Here, fuel such as
hydrogen or methane (natural gas) reacts with oxygen to form water and/or carbon
dioxide, and this process directly generates electricity. This process can be much more
energy efficient than combustion, often nearing 70 %. The increased efficiency not
only reduces fuel consumption but also minimizes the release of harmful greenhouse
gases. Additionally, fuel cells are scalable, and can be used to power a device as small
as an integrated circuit to as large as a city.
To understand fuel cells, consider what happens if we mix two reacting gases,
such as hydrogen and oxygen, in a chamber. The fuel will combust, releasing heat.
However, this thermal energy cannot be converted to work unless it is combined with
a heat engine. Next, consider what happens if we separate the hydrogen and oxygen
27

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

28

gas into separate chambers. The wall dividing the two is replaced by a membrane
that can only conduct protons, but not electrons (Fig. 2.1).

Figure 2.1: Hydrogen and oxygen gas separated by a membrane that is permeable to
protons but not electrons
In Fig. 2.1, a small amount of oxygen gas molecules will dissociate into oxygen ions
and travel across the membrane to react with hydrogen. However, once that happens,
an electrostatic force will develop between the two chambers, which will attract the
protons back into the hydrogen chamber, where they recombine with the electrons to
form hydrogen. As a result, no net reaction will take place and no work is generated.
To create an operational fuel cell, the electron must also be able to conduct from
the hydrogen to the oxygen chamber, enabling a complete, charge-balanced reaction.
Thus, creating an electrical pathway from the hydrogen to the oxygen chamber is
necessary to operate a fuel cell, as shown in Fig. 2.2
In the fuel cell, the oxygen gas in the anode receives electrons through an external
circuit and dissociates into oxygen ions through the half reaction O2 + 4e 2O2 .
The oxygen ions conduct through the membrane, and migrate to the anode and react
with hydrogen to form water through the half reaction 2O2 + 2H2 2H2 O + 4e .
Because the hydrogen wants to react with the oxygen to form water (in this module,
you will learn why), this driving force gives rise to a voltage between the cathode and
the anode. Electrical work is generated when current flows.

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

29

Figure 2.2: A fuel cell with a completed circuit, in the form of a light bulb. The net
reaction is 2H2 + O2 2H2 O.

2.1.1

Thermodynamic Problem

In this module, we aim to understand the maximum electrical work that can be
derived from a fuel cell. Electrical work equals voltage multiplied by the amount
of charge passed. The charge passed is proportional to the amount of gas reacted:
for example, reacting one oxygen molecule results in a transfer of 4 electrons. The
current is defined as the charge passed per unit of time, so increasing the current
results in an increased fuel consumption.
The voltage, on the other hand, depends strongly on the thermodynamics of the
system. In particular, the voltage under zero current condition, also known as the
open-circuit voltage, is governed purely by thermodynamic conditions. Thermodynamic variables like temperature, pressure, and the composition of the gas all directly
affect the voltage of a fuel cell. To understand how much electrical work can be produced by a fuel cell, we have to be able to calculate the voltage of a fuel cell, and
understand what factors affect the voltage.

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

2.2

30

Chemical Work

Figure 2.3: Hydrogen and oxygen gas at the same temperature and pressure, separated by a barrier. All walls, including the internal purple one, are impermeable to
heat, work and matter.
In module 1, we learned that mechanical work is defined as P dV . However, this
is not the only form of work possible. Let us consider this thought experiment,
shown schematically in Fig. 2.3: An isolated enclosure of gas is separated by a
barrier impermeable to matter (purple). One chamber contains oxygen, and the other
contains hydrogen. Both sides are at the same temperature and pressure. What
happens when the barrier is broken? Intuitively, we expect the following chemical
reaction to take place
2H2 + O2 2H2 O

(2.1)

Specifically
1. The number of hydrogen and oxygen atoms remain unchanged, consistent with
the enclosure being impermeable to matter.
2. 2 mol of hydrogen and 1 mol of oxygen react to form 2 mol of water. In other
word, the chemical bonds rearrange while conserving the number of constituent
atoms, but not the number of molecules.
3. Because the reaction is exothermic, we expect the temperature to increase.

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

31

Since the enclosure is isolated, no heat flows in and no external mechanical work
is being done on the system. Thus, dQ = dW = 0, and the First Law gives dU = 0.
Since the internal energy is constant, temperature should not change. This contradicts
point 3. However, since the First Law (dU = dQ dW ) must be obeyed, we conclude
that our existing definition of work (dW = P dV ) is incomplete. In other words, we
are missing contributions to dW .
To be more precise, we are missing chemical work, which is the transfer of potential
energy contained in the chemical bonds of the molecules to kinetic energy of a system,
just as mechanical work (P dV ) can be converted to internal energy that increases the
temperature, like by a compressing piston. Here, we define chemical work as
Chemical Work = hdn T sdn

(2.2)

where h is the molar enthalpy, or the bond energy, T the temperature, s the molar
entropy, and dn is the change in the number of moles. The first term in equation
2.2 relates to the change in the bond energy of the molecules (enthalpy). The second
term relates to the change in the degree of randomness in the molecules (entropy).
Equation 2.2 considers a system with a single chemical species, such as water. For
a system composed of many species, such as hydrogen, oxygen, and water, we can
sum the chemical work for the ith component of the system:
Chemical Work =

X
(hi dni T si dni )

(2.3)

By incorporating chemical work as well as mechanical work, we can write an


expanded first law of thermodynamics
dU = dQ dW

dUrev = T dS P dV +

X
(hi T si )dni

(2.4)

(2.5)

Equation 2.5 summarizes this section. In the expression for the internal energy, the
first term relates to the heat added to or removed from the system. The second term

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

32

relates to the mechanical work done on or by the system. The third term is the
chemical energy transferred into the system via the rearrangement of bonds and/or
the exchange of mass.

2.2.1

The Chemical Potential

Now that we have defined chemical work, we introduce a new variable, the chemical
potential of species i, denoted by the variable i :
i = hi T si

(2.6)

Like T and P , the chemical potential is independent of the quantity of the material.
Substituting equation 2.6 into equation 2.5 yields
dUrev = T dS P dV +

i dni

(2.7)

2.3

The Gibbs Free Energy

The internal energy is given by equation 2.7. As written, the expression has three
independent variables: (S, V, n). Consider the special case that S and V are held
constant in a system, such that dS = 0 and dV = 0. Based on equation 2.7, this
indicates that the chemical work done on the system equals the change in the internal
energy. We can write the expression below:

dU =

i dni |S,V

(2.8)

The subscripted S and V after | indicate that those variables are held constant. This
equation can also be rearranged to provide an explicit definition of i :

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA


i =

U
ni

33


(2.9)
S,V,nj6=i

The subscripted S, V , and j 6= i again indicate that these variables are kept constant.
Thus, chemical potential of species i is the change in the internal energy per mole
of ith species added, while holding entropy, volume and all other material amount
constant.
While keeping S and V constant appears to simplify the equation for internal energy, it is very difficult to measure S and even more difficult to control it. As a result,
they are rarely held constant in a real system. Furthermore, in most chemical plants
and devices (including fuel cells), the pressure, not volume, is controlled. Therefore,
we want to define a new thermodynamic potential which are functions of T and P ,
like written below.
d() = ()1 dT + ()2 dP +

i dni

(2.10)

In other words, we want to transform the internal energy U (S, V, ni ) into another
function G(P, T, ni ). For now, consider G as an arbitrary function; we will return
later to its physical meaning. For simplicity, we first show how to mathematically
transform U (V ) G(P ), and then generalize this transformation from U (S, V, ni )
G(P, T, ni ).
STEP 1
Since U is a function of S, V, ni , we can expand dU into the partial derivatives.
This expression is the multivariable form of the Taylor Series, taken to the first order
approximation.

dUrev =

U
S


dS +

V,ni

U
V


dV +
S,ni

X  U 
i

ni

dni

(2.11)

V,S,nj6=i


Comparing this equation to 2.7, in this equation, we see that U
= T and
S V,ni

U
= P where the subscript rev has been omitted from U for clarity. Here,
V S,ni

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

34

S, V, ni are the independent variables that we control, whereas U, T, P are the dependent variables.
STEP 2
We consider the special case where dS = 0 and dni = 0. The Taylor series
expansion for U (equation 2.7) gives

dU =

U
V


dV = P dV

(2.12)

S,ni

Fig. 2.4 plots U as an arbitrary function of V . The local derivative of the curve is
equal to P ; since the pressure must be positive, U must monotonically decrease as
a function of V . Suppose we pick an arbitrary point (V0 , U0 ). We draw a tangent
line at the point (V0 , U0 ) and label the y-intercept of this line as G0 , which equals
U0 P V0 . More generally, we can map every value of U to a value of G via the
equation U = G P V , where G is the intercept and P is slope of the dashed line
in Fig. 2.4.

Figure 2.4: Using the slope and intercept, we can map the monotonically-decreasing
function U to a value of G for every value of U and V . It is important for U to be
monotonically decreasing such that there is only one value of V for every value of U .
Combining U = G P V , equation 2.12, and the product rule yields:

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

35

dG = dU + d(P V )
dG = P dV + P dV + V dP
dG = V dP
Now, we have created a thermodynamic function G whose independent variable
is P rather than V . In this expression, if the pressure is kept constant, there is no
change to G.
STEP 3
Applying the same transformation we did in STEP 2 to V to the variable S, we
can convert U (S, V, ni ) to G(T, P, ni ). First, the full expression for G is given:

Grev = U + P V T S

(2.13)

It follows that:
dGrev = dU + d(P V ) d(T S)
X
= T dS P dV +
i dni + P dV + V dP SdT + T dS
i

= V dP SdT +

(2.14)

i dni

This completes the transformation from U (S, V, ni ) to G(T, P, ni ). We have created


a thermodynamic function that is dependent on T and P rather than S and V . Now,
by holding T and P constant, as is commonly done during chemical reactions, G is
only dependent on dni , the change in the amount of species in the systems:

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

dGrev =

X
i

i =

i dni |T,P

G
ni

36

(2.15)


(2.16)
T,P,nj6=i

The variable G is called the Gibbs free energy, and is the greatest amount of work
that can be extracted from a system at constant T and P . Moreover, equation 2.16
tells us that the chemical potential of species i is the partial molar Gibbs free energy
at constant T , P , and nj6=i .

Figure 2.5: An illustration of a system in a heat and work reservoir.

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

37

To illustrate this, consider a system in contact with an infinite heat and mechanical
work reservoirs, schematically shown in Fig. 2.5. We label the system with the
subscript sys and the reservoir with the subscript res. Thus, both T and P of the
system and the reservoir remain constant. If the heat and work is transferred and done
quasi-statically, then the maximum work that can withdrawn from the combination
of the system and the reservoir is given by:
Total Work Withdrawn = dUsys T dSres + P dVres
The first term relate to the change in the internal energy of the system. However,
the change in internal energy due to heat transfer (second term) or work done on the
reservoir (third term) cannot be withdrawn.
Since dVres = dVsys , dSres = dSsys , and T and P are constant, we can write

Total Work Withdrawn = dUsys + T dSsys P dVsys


= d(Usys T Ssys + P Vsys )
= dGsys
Thus, dGsys is the maximum work extracted from a system at constant P and T
when it is in contact with external reservoirs. In a fuel cell, P and T can be controlled
and kept constant, so dG yields the maximum work.

2.4

Chemical Equilibrium

The Gibbs free energy allows us to define equilibrium for a system at constant T and
P . These three conditions must be satisfied by a system when it is in equilibrium.
1. There is no net flow of energy or matter within the system
2. There is no net exchange of energy or matter with the outside world
3. There is no unbalanced driving forces or potentials

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

38

Note the term net in no net flow and no net exchange. Equilibrium does not
mean there is no flow and no exchange of matter or energy, only that the exchange
is balanced in the forward and reverse directions. More formally, G is minimized
against perturbations in the composition of the system (n1 , n2 ...) at constant T and
P when the system is at equilibrium:


G
ni


= i = 0 for all i

(2.17)

T,P,nj 6=i

In other words, an infinitesimally small modification in the composition of the system


yields no change in the Gibbs free energy. This is analogous to how the derivative of
the gravitational potential energy of a ball in a well is 0.
We can also apply this formalism to chemical reactions. Consider the following
reaction at constant T and P :
aA + bB cC + dD

(2.18)

where the lowercase letters are the stoichiometric coefficients. An example is 2H2 +
O2 2H2 O, where the stoichiometric coefficients are 2, 1, and 2. Now, consider
what happens to dG when the reaction in equation 2.18 occurs, where dG is equal to
the change in free energy associated with each species, or dGi :

dG = dGA + dGB + dGC + dGD

(2.19)

= A dnA + B dnB + C dnC + D dnD


As written, dG has four degrees of freedom, namely the change in the number of mol of
each species (dnA , dnB , dnC , dnD ). In actuality, however, the dni are not independent
from one another. The reason is that the reaction must proceed while obeying the
stoichiometry given in equation 2.18 in order to conserve mass. For example, for
every a mol of species A consumed, b mol of species B must also be consumed, while
c mol of species C and d mol of species D must be produced.
Thus, for the reaction shown in equation 2.18, we can substitute dnA = adx,

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

39

dnB = bdx, dnC = cdx, and dnD = ddx, where dx is the extent of the reaction.
Combining the above substitutions for dnA , dnB , dnC , dnD with equation 2.19 yields

dG = aA dx bB dx + cC dx + dD dx
= (aA bB + cC + dD )dx


G
x

(2.20)


= aA bB + cc + dD
T,P

Now, there is only one degree of freedom, x, or the extent of reaction. At equilibrium,
the Gibbs free energy is minimized with regard to the change in the composition,
at a constant temperature and pressure. The more general statement of chemical
equilibrium is given below.


G
x


=
T,P

vi i = 0

(2.21)

where, vi is the stoichiometric coefficient of species i. It is negative for reactants and


positive for products.

2.5

Chemical Equilibrium in a Fuel Cell

We want to determine the voltage of the fuel cell (shown schematically in Fig. 2.2) at
open circuit, when the current is infinitesimally small. The reason is that any flowing
current would cause a change in voltage due to resistances in the cell (on the zeroth
order, you can think of this as Ohms Law, V = IR). The first step is to consider
the components of the fuel cell where equilibrium is attained at open circuit, so that
we can apply equation 2.21 to these components/species. To attain open circuit, we
would remove the lightbulb and replace it with an infinitely resistive open circuit:
Table 2.1 lists the different species and reactions in a fuel cell. There, only the first,
fourth, and fifth reactions can reach equilibrium because all the participating species
can access each other. The other reactions cannot equilibrate because the species are
physically blocked from each other. The electrons are blocked by the broken circuit,

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

Species
O2
e
H2 O
H2 /e /O2 /H2 0
O2 /O2 /e
H2 /H2 O/O2

Location
Membrane
Anode/Cathode
Anode
Anode
Cathode
Anode/Cathode

Can Equilibrate
Yes
No
No
Yes
Yes
No

40

Reaction
2
OA
OC2
2
2H2 + 2OA
2H2 O + 4e
A
2
O2 + 4e

2O
C
C
2H2 + O2 2H2 O

Table 2.1: Chemical species and reactions in a fuel cell


and the gases are blocked by the membrane. To calculate the open-circuit voltage of
the fuel cell, we recognize that the following reactions must reach equilibrium at open
circuit:
2
1. 2H2 + 2OA
2H2 O + 4e
A
2
2. O2 + 4e
C 2OC
2
3. OA
OC2

The subscripts denote whether the reaction is happening at the cathode (C) or the
anode (A). Note that, strictly speaking, number 3 is not a chemical reaction, but a
transport process. Applying the equilibrium condition (equation 2.21), we obtain
1. 2H2 + 2O2 = 2H2 O + 4e
A

2. O2 + 4e = 2O2
C

3. O2 = O2
C

Combining these equations and rearranging, we obtain the chemical potential difference between electrons in the cathode and the anode as a function of the chemical
potential of the gas species.
e e =
C

1
(2H2 O O2 2H2 )
4

(2.22)

The right hand side of equation 2.22 gives the difference in the chemical potential
between the product (H2 O) and reactants (H2 and O2 ), multiplied by the appropriate

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

41

stoichiometric coefficients. The left hand side gives the difference in the electron
chemical potential between the anode and the cathode. Equation 2.22 tells us that the
difference in the chemical potential of electrons in the anode and cathode is precisely
balanced by the difference in chemical potential of the reactants and products. The
factor of 4 results from 4 electrons transferred for every molecule of oxygen consumed.
Note for advanced readers: we left out the electrostatic potential to simplify this
problem.

2.6

Voltage of a Fuel Cell

At the end of the last section, equation 2.22 shows that the difference in the chemical
potential of the reactants and products create a difference in the chemical potential
of electrons between the anode and the cathode. We can use a voltmeter to measure
the open-circuit voltage, which is related to the difference in the electron chemical
potential between the two electrodes:
V =


1 
e e
C
A
F

where F is Faradays constant, which equals 96,500 Coloumbs of charge per mol of
electron, and is given in units of J/mol. The negative sign results from electrons
being negatively charged. Thus, a higher electric potential results in a lower energy.
Combining the definition of the voltage with equation 2.22, we arrive at the following
equation for the open-circuit voltage of a fuel cell
V =


1 
1
e e =
(2H2 O O2 2H2 )
C
A
F
4F

(2.23)

Here, the voltage (also known as the Nernst potential) is a direct measure of the
change in chemical potential of the species (H2 , O2 , H2 O) upon reaction. In other
words, the fuel cell develops a precise voltage difference between the anode and cathode to balance out the chemical potential difference between the gas in the two electrodes (H2 and O2 /H2 O), so that equilibrium is achieved locally in the anode and
the cathode and electrolyte as in Table. 2.1.

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

42

Now that we have obtained the relationship between the fuel cell voltage and the
gas chemical potentials, we relate the chemical potential and the voltage to the partial
pressure of each gas specie. Recall from equation 2.6 that the chemical potential is
related to the partial molar enthalpy (h) and the partial molar entropy (s) of the
species:
i = hi T si
Recall from module 1 that the partial molar entropy of a gas is related to the concentration, or the partial pressure (pi ). In module 1, we learned that the entropy (capital
S) of a gas is given by:
Si = ni R ln pi + constant

(2.24)

where R is the gas constant and


 ni is the number of mol of a gas. It can be shown
i
that the partial molar entropy, S
, is given by
ni
si = R ln pi

(2.25)

Strictly speaking, the argument of a logarithm should be unitless, so we modify the


above expression:
si = s0i R ln

pi
pref

(2.26)

Here, pref is the standard pressure such that si = s0i when pi = pref . In the homework
and exams, we typically choose 1 atm or 1 bar as pref . Combining equation 2.26 with
equation 2.6, we can write the chemical potential of species i as:
i = hi T s0i + RT ln

pi
pref

(2.27)

Here, s0i not only accounts for the configurational entropy at standard state, but also
all other sources of entropy, such as the number of ways a molecule can rotate. Now,
we combine the hi and T s0i terms into a single term that we refer to as 0i . This

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

43

allows us to lump all terms that are independent of the partial pressure pi :
i = 0i + RT ln

pi
pref

(2.28)

This equation divides the chemical potential into two contributions. The first term,
0i , is the standard chemical potential of species i. This is the chemical potential at a
defined temperature T and partial pressure pref . When the partial pressure of species
i changes, i changes in accordance with the second term in equation 2.28.
By combining this expression for the chemical potential (equation 2.28) with the
expression for the voltage of the fuel cell (equation 2.23), we can write the partialpressure-dependent open-circuit voltage:

1
V =
4F




(pH2 /pref )2 pO2 /pref
0
0
0
O2 + 2H2 2H2 O + RT ln
(pH2 O /pref )2

(2.29)

The first three terms give the standard chemical potential of the reaction, whereas
the last term describes the configurational entropy of the gas molecules. Equation 2.29
states that the voltage of a fuel cell is not only dependent on the types of reactants
and products, but also on the concentration of each species. We combine the first
three terms into the standard reaction potential 0rxn = (0O2 + 20H2 20H2 O ),
where indicate the difference between the initial and final state of the reaction. The
standard reaction potential is the change in the Gibbs free energy of the system per
mol of reactants reacted when the reactants and products are have a partial pressure
equal to pref .

2.6.1

A Note on Units

You may also see the voltage written in terms of a single molecule, rather than a mol
of molecules as in equation 2.29. With the proper conversion factors, the following
equation is identical to the one in equation 2.29:
1
V =
4e

0O2

20H2

20H2 O


+ kB T ln

(pH2 /pref )2 pO2 /pref


(pH2 O /pref )2



CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

44

Here, i is the chemical potential of a single molecule of a species, given in units of


electron-volts (eV), rather than the chemical potential of a mol of the species, which
would have units of J/mol. An electron-volt is the potential energy of an energy
across a voltage of 1 V. e in this equation is the charge of a single electron, which
equals 1.6 1019 Coulombs. All the constants here differ from ones in equation 2.29
by Avogadros number, NA . For example, F = NA e, R = NA kB , and the i
in equation 2.29 (which has units of J/mol) equals Avogadros number multiplied by
the i in this equation (which has units of eV ).
For this course, we will teach in terms of F , R, and i in units of J/mol, as
in equation 2.29. These units are more typical of chemistry. More physics-oriented
courses will use e, kB , i in terms of eV. You are welcome to use those units if you
feel more comfortable. However, make sure you do not mix up the units in a single
equation. If you mix up units, you may get unphysical quantities on the order of
1020 V or 1020 V. If you do find yourself getting those values, check that you are
using consistent units.

2.6.2

Dependence of Voltage on Partial Pressure

To analyze the fuel cell voltage further, we sketch voltage as a function of the partial
pressure of each species in Fig. 2.6. Here, one components partial pressure is varied
at a time. When Pi = Pref , the fuel cell voltage is given only by the standard reaction
potential. The voltage increases when the partial pressure of oxygen or hydrogen is
increased. In contrast, the voltage decreases when the partial pressure of water is
increased. To rationalize the origins, we revisit the net reaction of hydrogen and
oxygen to form water (2H2 + O2 2H2 O).
When this reaction equilibrates, there will be significantly more water than hydrogen and oxygen. In other words, we say that the equilibrium of this reaction lies far to
the right (i.e., favoring the product). We can use this knowledge of equilibrium to explain why the open-circuit voltage of a fuel cell modulates with the partial pressures.
Let us imagine a fuel cell with a lot of water, and very little hydrogen and oxygen.
In such a fuel cell, there will be almost no driving force toward creating more water.

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

45

Figure 2.6: The voltage of a fuel cell as a function of the pressure. The intercept is
given by 0rxn and can be changed by modifying the chemistry of the reaction. Increasing the reactant concentration increases the voltage, while increasing the product
concentration decreases it.
Accordingly, this fuel cell generates a small voltage because there is no driving force
for the hydrogen and oxygen to react and form water. Alternatively, if the partial
pressures of hydrogen and oxygen is very high, then there will be a significant driving
force for creating water, giving a high fuel cell voltage, which is consistent with the
voltage plot based on equation 2.29.

2.6.3

Voltages with Different Fuels

The voltage of a fuel cell depends strongly on the type of fuel and oxidant. So far,
we have considered hydrogen and oxygen as reactants and water as the product. For
this reaction, we define the reaction potential 0rxn , and equals -475,000 J/mol of
0

O2 . Thus, 4Frxn = 1.23V at 300 K.


What happens if we change the mobile ion from oxygen ion to proton? In such a
fuel cell, the reactions become:

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

46

Cathode : O2 + 4H + + 4e 2H2 O
Anode : 2H2 4H + + 4e
Total : O2 + 2H2 2H2 O
Because the total reaction and the number of electrons transferred are the same,
the open-circuit voltage is the same and equals the expression in 2.29. This illustrates
an important concept in thermodynamics: thermodynamic variables like open-circuit
voltage and Gibbs free energy is independent of the path taken. Here, regardless of
whether the proton or the oxygen ion is the mobile carrier, the change in the Gibbs
free energy per mol to react hydrogen and oxygen into water is always equal. Thus,
the open-circuit voltage of a fuel cell based on proton conduction equals one based
on oxygen ion conduction.
Finally, let us consider changing the fuel from hydrogen to methane. If oxygen
ions are mobile, this undergoes the following reactions:

Cathode : 2O2 + 8e 4O2


Anode : CH4 + 4O2 CO2 + 2H2 O + 8e
Sum : 2O2 + CH4 2H2 O + CO2
0

Here, 0rxn = -763,000 J/mol at 473 K, and 8Frxn = 1.10V . The voltage
differs from the hydrogen fuel cell because the change in bond energy upon reaction
of methane and oxygen into water and carbon dioxide is significantly different from
the change in bond energy from the reaction of hydrogen and oxygen. Thus, while
changing the mobile ion cannot change the open-circuit voltage, changing the type of
reactants can.

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

2.7

47

Law of Mass Action

The concepts in this module can be used to derive the Law of Mass Action, which is
a very useful relationship. You can find versions of this law throughout many topics
in the physical, chemical, and biological sciences. It governs relationships as diverse
as the number of free electrons in a semiconductor, the yield of ammonia produced in
a chemical reaction plant, the kinetics of a digestion enzyme, and the rate in which
an infectious disease spreads in a population.
In this module, we present the Law of Mass Action in terms of chemical equilibrium. Consider the general reaction aA + bB cC + dD, and recall the conditions
for equilibrium at a constant T and P from equation 2.21, reproduced below:
aA bB + cC + dD = 0

(2.30)

i
Substituting equation 2.28 (i = 0i + RT ln Ppref
) and rearranging, we obtain the

following equilibrium condition:

a0A

b0B

c0C

d0D

= RT

ln

peq
A
pref

a


ln

peq
B
pref

b


+ ln

peq
C
pref

c


+ ln

(2.31)
where the superscript eq represents the equilibrium state. The left hand side of this
equation is 0rxn , the standard reaction potential. Rearranging yields:
peq
C
pref

c 

peq
D
pref

d

peq
A
pref

a 

peq
B
pref

B



K = exp

0rxn

RT

=

(2.32)

Here, K, is referred to as the equilibrium constant. Equation 2.32 is the Law of


Mass Action, which relates the reactant and product partial pressure to the standard
chemical potential of the reaction and T . Going back to equation 2.31, we see that
the equilibrium occurs when the disorder of the system (entropy) perfectly balances
the change in the bond energy (enthalpy).
A more general relationship for the Law of Mass Action is shown here: for a

peq
D
pref

d )

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

48

chemical reaction with the stoichiometric coefficients written as vi :


 P
 Y  eq vi
pi
i v i i
exp
=
RT
pref
i

(2.33)

where a negative vi indicates a reactant and a positive vi indicates a product, and


indicates the product of all the terms enclosed.

2.8

Fuel Cell vs. Heat Engine

At the beginning of the module, we stated that a fuel cell can be much more efficient
than a heat engine. Now, we will close this module by defining and calculating the
efficiency of a fuel cell and a heat engine.

Figure 2.7: The efficiency of a fuel cell as compared to the Carnot efficiency of a heat
engine
We define the efficiency as the maximum possible electrical work harvested divided
by the change in the bond energy, or the reaction enthalpy (rxn ). It can be shown
that the reaction enthalpy is the amount of heat released if we directly burn the fuel.
At constant pressure, this is known as the thermal efficiency therm . Using oxygen
and hydrogen as reactants in the fuel cell, Fig. 2.7 shows that the efficiency is high

CHAPTER 2. FUEL CELLS AND CHEMICAL EQUILIBRIA

49

at low temperatures and drops with increased temperature. The reason that the fuel
cell efficiency drops is because the reaction 2H2 + O2 2H2 O decreases in entropy
upon the creation of water. Lowering the entropy means that heat is released, which is
discarded to the environment rather than contribute to the open-circuit voltage of the
fuel cell. At high temperatures, T ds plays a much larger contribution to the chemical
potential (equation 2.6), and therefore reduces 0rxn and the open-circuit voltage.
However, there are other fuels where the voltage does not fall with temperature, as
you will see in your homework.
Figure 2.7 paints a rather pessimistic picture: the thermal efficiency of a hydrogen/oxygen fuel cell is a a little more than 90% at room temperature, and drops to
below 80% at 1000K. only a little more than 80 % at room temperature. In other
words, the electrical energy generated by such a fuel cell is less than 80 % of heat
released from burning the fuel directly. Why, then, should we use a fuel cell? It turns
out that the heat generated by combusting hydrogen with oxygen cannot be converted
to electrical work at 100 % efficiency. As you will learn in module 3, conversion of
thermal energy (heat) into work (electrical, mechanical, etc) is limited by the Carnot
efficiency, which is plotted in red in 2.7. Comparing the fuel cell thermal efficiency
to the Carnot efficiency, we see that fuel cells have a clear efficiency advantage at
temperatures below 1,000 K.

Chapter 3
Solar Thermal, Phase Transitions,
Heat Engines
In module 2, we described fuel cells as one way to generate electricity from fuel. While
fuel cells are very efficient at converting energy from chemical bonds into electric work,
they require a steady source of fuel to operate, such as hydrogen. Because it readily
reacts with atmospheric oxygen, hydrogen gas is very rare on this planet. As a result,
hydrogen is typically created through methane reformation to meet demand, but
carbon dioxide is released in the process. Alternatively, some fuel cells directly use
methane but also release carbon dioxide that contributes to global warming.
One clean and abundant source of energy is the sun. Forty minutes of solar
irradiation (energy of sunlight reaching the earths surface) contains as much energy
as all of humanity ueses in an entire year. To put it another way, we can meet the
energy demands of the world by converting just 10% of the solar energy that shines
on the state of Arizona into electricity.
One way to harness solar power is to use solar cells. In these devices, solar energy is
converted directly to electricity through a semiconductor like silicon. While solar cell
technology has improved greatly in the last decade, they have several disadvantages:
firstly, these devices can only convert some of the sunlight to electricity (typically
15-20%); the rest is not absorbed or released as heat. Secondly, silicon processing
is expensive and energy-intensive. Third, solar cells cannot produce any electricity
50

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 51

when the sun does not shine, such as at night, while there is always demand for
electricity. In the absence of energy storage, solar cells can only produce about 20% of
daily electrical demand until another source is required to supplement its production
deficiencies.

Figure 3.1: Schematic of a solar thermal power plant. This module will focus on the
thermodynamics of phase transitions and heat engines.
An alternative method to create electricity from the sun is using a solar thermal
plant, shown in Fig. 3.1. Here, sunlight is concentrated by mirrors and used to heat
water or another working fluid. The water boils and passes through a heat engine.
In the heat engine, the difference in the temperature between the water vapor (or
steam) and the liquid water is converted to mechanical work using a turbine. This
mechanical work can be converted to electricity in a generator using the principles of
electromagnetic induction. Solar thermal electricity generation overcomes the three
problems of solar cells: solar thermal plants use most of the energy from the sun
and can attain higher efficiencies, the mirrors in a solar thermal plant are generally
less expensive than silicon, and such plants can store energy in the form of heat to

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 52

continue to produce energy in the absence of sunlight.


As we see in the schematic in Fig. 3.1, there are two key components to a solar
thermal plant. First, the water needs to transform from the liquid phase to the
gas phase during heating. Second, the water vapor runs a heat engine to generate
mechanical work. Thus, this module is divided into two parts: phase transitions (i.e.,
boiling water) and heat engines.
To simplify, we will describe the two processes separately and use the Carnot
engine as a model of a heat engine. Combining the two processes in a single thermodynamic cycle, as in a Rankine Cycle used in many real solar thermal plants, is
beyond the scope of this class. You may wish to take a more advanced thermodynamics course in mechanical engineering if you are more interested in heat engines.

3.1

The Three Phases of Water

Water is highly abundant on earth and is the most important material for sustaining
life. Over 70% of the earths surface is covered by water, and the human body is more
than 60% water. Due to its abundance and lack of toxicity, it is the working fluid of
choice in most heat engines.
Water exists in three phases: solid ice, liquid water, and gaseous water vapor.
Figure 3.2 sketches the specific Gibbs free energy of water as a function of temperature
=H
T S,
so the slope of G
vs T gives
for all three phases of water. Recall that G
For simplicity, we assume that H
and S do not
S while the y-intercept gives H.
depend on temperature.
First, we will discuss the specific enthalpy, or the y-intercept of Fig. 3.2. Here,
the specific enthalpy of a molecule such as liquid water is the bond energy relative to
a defined state. In module 2, we chose the elemental form of O and H as the reference
state; however, because there is only one molecular species here, we can choose the
enthalpy of water vapor at 0 K as the reference. Negative enthalpy implies stronger
intermolecular bonds, such as the hydrogen bonding between liquid water molecules.
does not matter (we can choose any
However, we note that the specific value of H
arbitrary point as our reference); choosing a different reference would move all the

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 53

as a function of the temperature (T ).


Figure 3.2: The specific Gibbs free energy (G)
at a certain T . Here, we
The stable phase of water is the one with the lowest G
and S.

neglect the temperature dependence of H


between, however, must be
curves in figure 3.2 up or down. The difference in the H
specifically defined.
The three phases of water have significantly different specific enthalpies of formations, which represent the strength of the intermolecular bonds (bonds between
molecules), relative to the reference state (which we choose as water vapor at 0 K).
Solid bonds are the strongest (therefore giving the most negative specific enthalpy).
This explains why melting ice to water absorbs heat from the environment because
the specific enthalpy of ice is more negative than that of liquid water. We note here
that the enthalpy refers to the energy to form bonds, not the change in the average
kinetic energy of the molecules associated with heating or cooling. Solid ice and liquid
water at 0 C in fact have the same kinetic energy, since it is only dependent on the
temperature.

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 54

Next, the slopes equal S of each phase and indicate that each phase has a
different value of specific entropy. This makes sense because a gas has more disorder
than a liquid, which in turn has more disorder than a solid. In a crystalline solid like
ice, the position of each water molecule is fixed in a lattice. In liquid water, the liquid
molecules flow but are still tightly bound. Gas molecules have the most disorder and
travel freely in all three dimensions. This is why the gas phase has the most negative
slope, while the solid has the least negative slope. Note all specific entropy here is
positive because it is defined by the number of configurations in a system, and not
referenced to a particular state. However, if we choose a reference point, the results
are identical because only the relative entropy difference, not the absolute number,
between different phases affect the equilibrium behavior.
Recall from module 2 that a system in equilibrium minimizes the Gibbs free energy.
At T > Tm and T < Tb ,
In Fig. 3.2, we observe that, at T < Tm , ice has the lowest G.
At T > Tb , water vapor has the lowest G.
Thus, under
liquid water has the lowest G.
equilibrium conditions, water is a solid below a temperature of Tm , a liquid between
Tm and Tb , and a gas above Tb .
of ice and water is equal (G
s = G
l ). Thus, ice and water
When T = Tm , the G
coexist and are equally stable, and Tm is known as the melting point. If we observe
that both ice and water are thermodynamically stable (e.g., both are present in
equilibrium in an isolated glass of water), then we know that the temperature is
equal to the melting point. Similarly, at T = Tb , liquid water and water vapor are
l = G
g , and
equally stable, in a temperature known as the boiling point, because G
the subscripts l and g refers to the liquid and gas phase, respectively.
Summarizing the paragraphs above, water molecules adopt the phase with the
lowest specific Gibbs free energy. The temperature relationships are written below:
At T < Tm , solid ice is the most stable phase
At T = Tm , solid ice and liquid water are equally stable
At Tm < T < Tb , liquid water is the most stable
At T = Tb , liquid water and gaseous water vapor are equally stable

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 55

At T > Tb , water vapor is the most stable phase


The list above only describes the thermodynamically stable, equilibrium phase.
Liquid water can exist at a temperature below Tm . However, such a phase is not in
equilibrium and is known as supercooled water. Likely, it has not formed ice due to
kinetic limitations, i.e. there is not yet a nucleation point for ice crystals to form.

3.2

Phase Transition in Water

The previous section showed that two phases of water can coexist simultaneously at
the melting and boiling temperatures. If heat is continuously added to or removed
from the water (such as in a water kettle), one phase is continuously transformed to
another. In this section, we will discuss the thermodynamics of phase transition, such
as when 1 mol of liquid water boils to form 1 mol of water vapor.
We begin with a schematic of the total Gibbs free energy during the boiling of
water at the boiling temperature in a system where mass cannot be exchanged with
the outside world.

Figure 3.3: Schematic of Gibbs free energy before and after boiling.
We consider the Gibbs free energy of each phase before boiling has started (such
as at 99.9 C) and after boiling has finished (such as at 100.1 C) in Fig. 3.3. We

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 56

start with 1 mol of liquid water and end with 1 mol of water vapor. When 0% of the
water has boiled, all of the water is in liquid form, so the Gibbs free energy of the
gas phase is 0. Similarly, when 100 % of water boils, all of the water is in the gas
phase, and the Gibbs free energy of the liquid phase is 0 because there is no more
liquid water. Since the water is transformed at the boiling temperature T = Tb , the
total Gibbs free energy at 0% water vapor and 100 % water vapor is equal.
Here, we consider what happens between 0% and 100% liquid water. Recall from
module 2 that the variation dG is given by
dG = SdT + V dP +

i dni

(3.1)

For boiling at constant T and P , equation 3.1 simplifies to:

dG =

i dni

(3.2)

Here, i indexes the phases just as it indexes the different species in module 2. During boiling, both liquid and gaseous phases of water are present. To determine the
relationships between of g , dng , l , and dnl , we use a chemical reaction to represent
the equilibrium between the two phases of water, we write:
H2 O(l) H2 O(g)
As a result, we write that the total variation in dG is given by a combination of
dGl and dGg , which are the variation in the liquid and vapor phases
dG = dGg + dGl

(3.3)

where dGg = g dng and dGl = l dnl . Addditionally, since liquid water and gaseous
water vapor are in equilibrium at the boiling temperature, we apply the chemical
equilibrium criterion from module 2 (equation 2.21) to obtain:
l = g

(3.4)

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 57

Therefore, the chemical potential of the two phases must be equal when they coexist
during a phase transition. Moreover, for a closed system (fixed amount of H and O
atoms), conservation of mass requires:
dnl = dng

(3.5)

By substituting equations 3.4 and 3.5 into equation 3.3, we compute dGl and dGg :

dGg = g dng

(3.6)

dGl = l dnl = g dng


These equations tell us that the Gibbs free energy of a phase is proportional to the
molar quantity of that phase. Using this relationship, Fig. 3.4 completes the plot
in Fig. 3.3. As shown in Fig. 3.6, dGg = dGl during the phase transformation at
constant T = Tb and constant P .

Figure 3.4: Gibbs free energy of water vapor and liquid water during boiling.

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 58

3.3

The Clausius-Clapeyron Relation

By popular convention, Tm for water is 0 C and Tb is 100 C. However, upon closer


inspection this is not always true. A pressure cooker can keep liquid water at 120 C
without boiling. Similarly, if you live in a high-altitude city like Denver, the boiling
temperature of water drops below 100 C, which is why cooking recipes often offer
different instructions for high-altitude environments. Similarly, if you put a glass of
liquid water in a vacuum, you will see it start boiling at room temperature.
In this section, you will learn how Tb changes with the pressure. First, we write
down the the variation of G with T , P and n for both the liquid and gas phase:

dGl = Sl dTb + Vl dP + ul dnl

(3.7)

dGg = Sg dTb + Vg dP + ug dng


As P changes, Tb changes while maintaining the two phase coexistance. In other
words, the specific Gibbs free energy must remain equal regardless of the values of T
and P , i.e., Gl = Gg . Mathematically, this means that
dGl = dGg .

(3.8)

If this condition is violated, either the liquid or the gas phase will have a lower G,
make it the more stable phase. As written, the system has three degrees of freedom
(Tb , P , n). However, we are only interested in the relationship between Tb and P .
By recognizing that at a given (and constant) Tb and P , the total Gibbs free energy
does not change with the composition, we are free to choose values of n and hold
it constant. Thus, for mathematical convenience, we hold both nl and ng constant.
This is equivalent to saying that we are looking at the pure phases, i.e., fully liquid at
the onset of the phase transformation, and fully gas at the completion of the phase
transformation. Applying to this Eqs. 3.7 and 3.8 we obtain:
Sl dTb + Vl dP = Sg dTb + Vg dP

(3.9)

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 59

Here, dTb and dP are the change in the boiling temperature and pressure. Rearranging
this expression yields
(Vg Vl )dP = (Sg Sl )dTb
Rearranging again yields the differential equation
Vg Vl
dTb
=
dP
Sg Sl

(3.10)

To simplify this equation, recall that for a reversible process the heat flow Qrev =
T S. In other words, the change in entropy (S) in the reversible limit equals the
heat flow (Qrev ) divided by the temperature (T ). The amount of heat exchanged
during boiling is something that can be measured and tabulated. It is known as Lb ,
the latent heat of boiling equalling Qrev /n, where n is the number of moles of the
material. It can also be written as

Lb =

Tb (Sg Sl )
n

(3.11)

where S = Sg Sl . Substituting equation 3.11 into equation 3.10 yields:



dTb
Tb
Vg Vl
=
dP
Lb

(3.12)

where Vg and Vl are the specific volumes, or the volume per mol. This is known as the
Clausius-Clapeyron relation, and describes how the boiling temperature of a liquid
changes with the pressure. Lb , Vg , and Vl are all dependent on the temperature.
In the case of water, the molar volume of liquid water ( 0.018 L/mol) is much
less then the molar volume of water vapor ( 30 L/mol according to ideal gas law at
373 K at 1 atm). As a result, the Vl term is negligible, and we can write a simpler
Clausius-Clapeyron relation:
dTb
Tb
Vg
dP
Lb

(3.13)

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 60

Figure 3.5 shows this relationship and how the boiling temperature rises with the
pressure. At the red line, liquid and gas are equally favorable and coexist. Above the
red line, liquid phase is stable; below the red line, gas phase is stable.

Figure 3.5: Sketch of the Clausius-Clapeyron relation given by equation 3.13. The
boiling temperature rises with higher pressure.

3.4

Evaporation

In the last section, we learned that water converts from a liquid to gas if we raise the
temperature above the boiling temperature. This is known as boiling, and a typical
indication of boiling is the formation of bubbles within the liquid. However, if we
leave out a small amount of liquid water for a long time at room temperature in
an open environment, the liquid water will eventually phase transform into gaseous
water vapor. However, the temperature of the water is significantly below the boiling
temperature, whereas Fig. 3.2 indicates that only the liquid phase is stable between
the melting and boiling temperatures. In this section, we will discuss why the water
evaporates into a gas even below the boiling point.
First, consider the case of a jar of water inside a closed chamber. We pull a
vacuum on the chamber so that there exists liquid water and vacuum above it, as in

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 61

the top row of Fig. 3.6. We then remove the pump, and seal the container but keep
it at a constant temperature T , room temperature. If we let the system equilibrate,
we can imagine that the following steps would happen:
1. If the vacuum is perfect, the boiling temperature of water nears absolute 0 K
according to the Clausius-Clapeyron equation 3.12. Even without a perfect
vacuum, the boiling temperature is below room temperature. As a result, the
water would start to boil and some of the liquid water transforms to water
vapor.
2. At some point, enough water will boil to raise pressure of the chamber to the the
equilibrium pressure Peq , which is defined by the Clausius-Clapeyron equation
at temperature T .
3. Now, liquid water and water vapor are in equilibrium, as in the top right panel
of Fig. 3.6. Since the chemical potential of water in the liquid and gas phases
are equal, this means that the net evaporation rate equals the net condensation
rate
Next, we change the situation by replacing the vacuum with 1 bar of dry nitrogen
at room temperature. The temperature of water is below the boiling point (100 C),
the pressure is 1 bar, so no boiling is expected. However, water at the surface would
still evaporate, but condensation is impossible because the environment has no water
vapor.
After waiting a long time, the chamber contains a mixture of water vapor and
nitrogen, shown in the bottom panels of Fig. 3.6, and the liquid water level has
decreased. In other words, some of the liquid water evaporated into water vapor.
The vapor pressure of water, pvapor , is the equilibrium partial pressure in which the
rate of evaporation equals the rate of condensation. This means the chemical potential
of water vapor equals the chemical potential of liquid water.
At a defined temperature, pvapor is given by the Clausius-Clapeyron equation.
To understand why, recall the conditions from which we derived that equation. We

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 62

Figure 3.6: Schematic of boiling from vacuum (top) and evaporation from a dry environment (bottom) with liquid water at room temperature. In both cases, the liquid
water evaporates/boils into water vapor (left) initially, but the rate of evaporation
(blue arrows) and condensation (red arrows) eventually equilibrate after waiting a
long time, leaving a partial pressure of water equal to the water vapor pressure.
stated that, at equilibrium, l = g , which also holds true here. Thus, using the same
derivation, we can state that
L
dpvapor


dT
T Vg

(3.14)

Here, L is the heat of evaporation, a positive number because it takes heat to break
the intermolecular bonds of liquid water, and equals the latent heat of boiling. This
is why sweating cools the body down because evaporation absorbs heat from the
environment. On a humid day, when the partial pressure of water in the atmosphere
approaches the vapor pressure, it is very difficult for our body to expel heat through
evaporation. This explains why a humid hot day can feel hotter than a dry day at
the same temperature.
What if we were to constantly replenish the air with dry nitrogen by opening the
chamber? In that case, water would continually evaporate because the water vapor

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 63

would be swept away. After waiting a long time, all of the water would evaporate at
room temperature.
You may notice that the west coast of the US has much drier air than the east
coast, even near a large reservoir of water (the Pacific Ocean). The reason is that
the water on the west coast is much cooler than on land, a result of ocean currents.
Therefore, the water vapor in the air over land is approximately in equilibrium with
liquid water in the oceans. However, because the water in the ocean is cooler than
the water on land, the partial pressure of water vapor over land is less than the vapor
pressure at the land temperature. Conversely, near the Atlantic Ocean, the ocean and
land temperatures are often comparable, and the humidity can near 100 %, which is
when the partial pressure of water equals the vapor pressure.
Finally, we can arrive at a stricter definition of boiling. The boiling temperature
is the temperature in which the vapor pressure exceeds the total pressure of the
environment. At 100 C, pvapor of water equals 1 bar, and would induce water to boil.

3.5

The Carnot Engine

So far in this module, we have discussed the thermodynamics of phase transitions,


namely the boiling of water, which is an essential step to convert sunlight into a
thermal-energy-carrying fluid (steam). Additionally, a solar thermal plant requires a
device that converts thermal energy (heat) to mechanical energy (work), and subsequently to electricity. A heat engine is a device that converts heat to work. In the
second part of this module, you will learn about the Carnot engine, which is the most
efficient heat engine possible.

3.5.1

Isothermal Expansion of Gas

To describe the Carnot engine, we first consider a piston in contact with a heat
reservoir (Fig. 3.7). The walls of the piston are impermeable to matter, but transfer
heat. The temperature of the piston equals the temperature of the reservoir, but the
pressure in the piston is higher than the environment. As a result, the piston expands

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 64

Figure 3.7: A piston in a heat reservoir isothermally expanding. Here the wall of the
piston is impermeable to matter but allows heat transfer from the reservoir.
and does work on the environment, but its temperature is constant. This is known
as isothermal expansion. We assume the process proceeds quasistatically.
To model this behavior, we use the ideal gas law P V = nRT . n and T are constant
in this isothermal expansion, so P V = constant. Figure 3.8 plots the isothermal
expansion as a function of the P and V as well as the T and S.
On the left panel of Fig. 3.8, we plot the P and V of the system, given by P V =
constant, as the system expands from (P1 , V1 ) to (P2 , V2 ). Recall that the mechanical
work is given by dW = P dV . As a result, the expanding gas is doing work, which
can be harvested through the motion of the piston. At a constant temperature,
dU = dQ dW = 0, so the heat that flows into the piston from the environment
must equal the work done by the piston. In the reversible limit, dQrev = T dS, so
the entropy of the piston must increase. This is consistent with what is learned in
module 1, where a lower pressure indicates more disorder and higher entropy.
Here, if the piston continues to expand, we have a system that converts heat to
work at 100 % efficiency. Unfortunately, in order for us to carry out this conversion

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 65

Figure 3.8: Pressure, volume, temperature, and entropy during isothermal expansion
of a gas
indefinitely, the piston has to be infinitely large (as the piston expands indefinitely).
What we want is a piston that can convert heat to work, but also be able to restore
itself back to its original state in terms of T , P , and V .
To restore the piston to the original state, we can compress the piston back to its
original temperature and pressure. Here, work is done quasistatically on the piston
to increase its pressure. Figure 3.9 plots the P-V and the T-S diagram of this cycle.
In the P-V digram, the piston expands from (P1 , V1 ) to (P2 , V2 ), and then back to
(P1 , V1 ) at a constant temperature. However, the positive work done by the piston
during expansion equals the negative work done on the piston during compression,
and no net work is being generated. Graphically, integrating P dV in Fig. 3.9 from
(P1 , V1 ) to (P2 , V2 ) and back to (P1 , V1 ) yields no net area under the curve. Similarly,
the heat absorbed during expansion from (T, S1 ) to (T, S2 ) equals the heat released
during the compression back to (T, S1 ). Again, this is shown graphically in Figure
3.9 (recall integrating the T-S curve gives heat transferred into the system). Thus,
there is no net conversion of heat to work.

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 66

Figure 3.9: A piston that isothermally expands and compresses between two pressures.
No net work is being done in a cycle.

3.5.2

The Carnot Cycle

To be able to harness work from the piston, we want the work done by the piston
during expansion to be greater than the work done on the piston during compression. In the reversible limit, this means that the heat absorbed during expansion is
greater than the heat released during compression. We will use the transfer of heat
to understand this process. Since dQrev = T dS, in order to absorb more heat than is
released (while keeping S equal at the start and end of the cycle), the piston need to
absorb heat at a higher temperature TH and release it at a colder temperature TC . If
the heating and cooling can be done at constant entropy (i.e., no heat flow in or out
of the piston), a complete cycle is obtained as plotted in the right panel of Fig. 3.10.
The left panel plots the P-V diagram of the same process. Steps 2 3 and 4
1 are known as adiabatic expansion and compression (i.e., no flow of heat), and the
P-V curve can be calculated using some math starting with the formulas dU = dW ,
dW = P dV , dU = 3/2nRdT , and the ideal gas law.
Table 3.1 describes all the processes in the Carnot engine of Fig. 3.10.
We want to calculate the heat-to-work efficiency, defined as =

Work Done
.
Heat In

We

can integrate the P-V diagram (left panel of Fig. 3.10) through dW = P dV , but that
will involve a bit of math. Instead, we invoke the First Law of Thermodynamics,
dU = dQ dW , to calculate the heat and the work. For a complete Carnot cycle

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 67

Figure 3.10: The T vs S diagram of the Carnot engine.


Description
Step
1 2 Isothermal Expansion
Adiabatic Expansion
23
3 4 Isothermal Compression
3 4 Adiabatic Compression

Temperature
TH
TH TC
TC
TC TH

Entropy
S1 S2
S2
S2 S1
S1

Heat Qrev
TH (S2 S1 )
0
TC (S1 S2 )
0

Table 3.1: Table describing the four parts of the Carnot cycle given in Fig. 3.10.
shown in Fig. 3.10, a cycle begins and ends at the same temperature, so Ucycle = 0.
Thus, we know that Qcycle = Wcycle . Here, Qcycle is the sum of all the terms in
the last column of table 3.1 and equals
Qcycle = Wcycle = (TH TC )(S2 S1 )
Thus, the thermal energy to work efficiency can be calculated, where Heat In
describes the amount of heat added to the piston during expansion.

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 68

Work Done
Heat In
(TH TC )(S2 S1 )
=
TH (S2 S1 )
TH TC
=
TH
TC
=1
TH

carnot =

(3.15)

This equation gives the Carnot efficiency and describes the maximum efficiency of a
heat engine operating between TH and TC , in units of Kelvin. It is only a function
of the temperature and independent of all other variables. The Carnot efficiency is
plotted in Fig. 3.11 using of 300 K as the temperature of the cold reservoir. Here,
the Carnot efficiency increases when the temperature of the hot reservoir is higher.

Figure 3.11: The Carnot efficiency, defined in equation 3.15, as a function of the hot
reservoir temperature.

3.6

The Carnot Theorem

The Carnot theorem states that the Carnot engine is the most efficient heat engine.
Anything more efficient is a perpetual motion machine. In this section, we shall prove
the Carnot theorem.

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 69

3.6.1

Stacking a Carnot Heat Engine and a Heat Pump

A heat pump is a heat engine operating in reverse, where work is done to move heat
from a lower temperature to a higher temperature. Consider the device shown in
figure 3.12, which is a heat engine is coupled to to a heat pump. Both the engine
and the pump are thermally connected to the same hot and cold reservoir. The heat
engine generates work by transferring heat from the hot reservoir to the cold reservoir
through the cycle shown in Fig. 3.10. The heat pump uses the heat engines work to
pump heat from the cold to the hot reservoir by running the cycle in figure 3.10 in
reverse. We use subscript 1 to denote the heat engine, and subscript 2 to denote
the heat pump.

Figure 3.12: A heat pump and a heat engine stacked. The work from the heat engine
is being used to power the heat pump. Here, QH,1 and QC,2 are positive because
heat flows out, whereas QC,1 and QH,2 are negative as heat flows in.
Let us first consider the heat and work from the heat engines perspective. From

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 70

the First Law of Thermodynamics, the net heat absorbed equals the net work done
by the engine, as the internal energy of the heat engine does not change in a cycle.
The net heat equals QH,1 plus QC,1 . Note that QC,1 is negative because heat is
rejected to the cold reservoir:
W1 = QH,1 + QC,1

(3.16)

The efficiency of the heat engine is defined as the ratio between the work done and
the heat absorbed from the hot reservoir. This can be written as:
1 =

W1
QH,1

(3.17)

Its value is given by equation 3.15. By combining equation 3.16 and 3.17, we can
solve for the heat (QH,1 and QC,1 ) as a function of the work

1
W1
1


1
= 1
W1
1

QH,1 =

(3.18)

QC,1

(3.19)

We can write similar expressions for the heat pump. Because work is being done on
the heat pump, W2 is negative, and equals W1 . Similarly, QH2 is negative
while QC,2 becomes positive.

W2 = W1
1
QH,2 = W2 = 1/2 W1
2

(3.20)

QC,2 = (1 1/2 )W2 = (1 1/2 )W1

(3.22)

(3.21)

Combining the equations for the hot reservoir (equations 3.18 and 3.21) and cold
reservoir (equations 3.19 and 3.22), we derive the net flow of heat out of the hot
reservoir to the heat engine and pump and the net heat from the cold reservoir to the

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 71

heat engine and pump:

3.6.2

QH = QH,1 + QH,2 = (1/1 1/2 )W1

(3.23)

QC = QC,1 + QC,2 = (1/2 1/1 )W1

(3.24)

Case Studies of a Carnot Engine and a Heat Pump

Now that the problem is set up and all of the relationships are derived, we employ
several case studies where we vary the efficiencies of the heat engine and the heat
pump in Fig. 3.12. Through these case studies, we shall show that the Carnot engine
is the most efficient heat engine possible. Anything more efficient is a perpetual
motion machine that can spontaneously reduces its entropy by pumping heat from a
cold reservoir to a hot reservoir without external work being done on the system, a
phenomenon that violates the Second Law of Thermodynamics.
Case 1: 1 = 2 = carnot
In this first case, assume that both the engine and heat pump operate at the Carnot
efficiency, carnot , given in equation 3.15. From equations 3.23 and 3.24, the heat flow
out of the two engines can be written as:

{QH = (1/carnot 1/carnot )W1 } = 0

(3.25)

{QC = (1/carnot 1/carnot )W1 } = 0

(3.26)

Hence, when the efficiency of the engine and the pump equals the Carnot efficiency,
there is no net flow of heat from the hot or the cold reservoir. This makes sense
because the heat discharged by the heat engine to the cold reservoir is precisely the
same amount that is taken up by the heat pump. In this case, the entropy of the
entire system does not change, which is idealized but consistent with the Second Law
of Thermodynamics.

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 72

Case 2: 1 < carnot , 2 = carnot


Now we put a realistic heat engine with 1 < carnot . We continue to take the efficiency
of the heat pump at the Carnot efficiency. Now, the heat flows can be written as

{QH = (1/1 1/carnot )W1 } > 0

(3.27)

{QC = (1/carnot 1/1 )W1 } < 0

(3.28)

This implies that heat flows out of the hot reservoir into the cold reservoir. This
increases the entropy and is consistent with the Second Law of Thermodynamics.
This is expected, as using a nonideal engine results in an increase in entropy.
Case 3: 1 > carnot , 2 = carnot
Next, we consider the case of engine 1 exceeding the Carnot efficiency. Because
1 > carnot , then 1/1 < 1/carnot

{QH = (1/1 1/carnot )W1 } < 0

(3.29)

{QC = (1/carnot 1/1 )W1 } > 0

(3.30)

Here, heat flows from the cold reservoir to the hot reservoir, and results in a
spontaneous decrease in entropy without work being done on the system. Accordingly,
this system violates the Second Law of Thermodynamics, and constitutes a perpetual
motion machine.
You may notice that, strictly speaking, Case three violates the Second Law of
Thermodynamics whenever 1 > 2 , for any heat pump with an efficiency 2 . In
other words, to be consistent with the Second Law, 1 2 , for any possible heat
pump that we can put in 2 . Since we know that there exists a heat pump with an
efficiency of carnot , then 1 < carnot .
We arrive at the Carnot Theorem: any heat engine with an efficiency

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 73

greater than the Carnot efficiency violates the Second Law of Thermodynamics and is a perpetual motion machine.

3.6.3

Carnot Engine as the Most Efficient Heat Engine

While the Carnot theorem is derived and first considered in the case of the heating
and expansion of gases, it is universally applicable to processes that extract work
from a temperature gradient, without necessarily needing pistons or gas expansions.
For example, we can think of thermoelectrics as heat engine that directly converts a
temperature gradient into electric work. Solar cells can also be thought of as a heat
engine, where the temperature of the sun and earth are the different heat reservoirs.
In any heat engine, the Carnot theorem defines the maximum efficiency; any efficiency
exceeding it is a perpetual motion machine.

3.6.4

for a Heat Pump

Finally, we consider what the Carnot theorem says about the of a heat pump. A
heat pump moves heat from a cold reservoir to a hot reservoir when work is being
done of the heat pump. One example is an air conditioner, which removes heat
from a cold room and expels it outside the building. During winter, it may also be
used to pump heat from the environment into a building. In this case, we have to
reconsider what the efficiency of a heat pump means. Previously, for a heat engine,
we defined efficiency through equation 3.17, rearranged in equation 3.31. A higher
efficiency means we can derive more work with less heat.
QH = W

(3.31)

In a heat pump, however, we want to pump more heat using less work. Hence,
we want to be able to minimize as defined in equation 3.31. A high efficiency
() is a bad heat pump; for example, = 100% implies that all the work is directly
converted to heat, without pumping any heat from the outside. Similarily, an
near 0 implies that we can pump a lot of heat from the cold reservoir to the hot

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 74

reservoir without performing any work, which is a violation of the Second Law of
Thermodynamics.
To summarize, a good heat engine would have a high because it derives the most
work from heat. A good heat pump would have a low because it requires the least
work to pump a given amount of heat. This is why the heat pump community avoids
using the word efficiency and use the term cost of performance, defined as 1/,
as their figure of merit.
The Carnot theorem tells us that the minimum of a heat pump is the Carnot
efficiency, which is attained by running the Carnot cycle in Fig. 3.10 backwards. This
is the minimum work you need to pump a given amount of heat from a cold to a hot
reservoir. The subsection below shows why we cannot have a heat pump with an
less than carnot . We can see this using the following case study, which reused the
device in figure 3.12
Case 4: 1 = carnot , 2 < carnot
Inserting these two relations into equations 3.24 and 3.23 yields:

{QH = (1/carnot 1/2 )W1 } < 0

(3.32)

{QC = (1/2 1/carnot )W1 } > 0

(3.33)

This is again a perpetual motion machine, where heat is pumped from a cold
reservoir to a hot reservoir without external work. Hence, the Carnot theorem not
only tells us the maximum efficiency of a heat engine, but also the minimum , or
the maximum cost of performance, of a heat pump.

3.7

Efficiency of a Solar Thermal Power Plant

Previously, equation 3.15 and Fig. 3.11 states that the efficiency of the Carnot engine
increaes at higher TH for a given TC . However, such an expression does not address
thermodynamic loss mechanisms. One loss mechanism that influences the working

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 75

temperature is the re-radiation of electromagnetic waves by the hot reservoir. All


matter with a temperature above absolute zero emits energy as electromagnetic radiation. This explains the red light a toaster emits when it is hot. The power of this
radiation, known as blackbody radiation, can be expressed by the Stefan-Boltzmann
relation:
P = AT 4

(3.34)

where A is the surface area of the body, is a constant equal to 5.67108 W K 4 m2 ,


and T is the temperature of the emitting object. This expression states that the
power of blackbody radiation scales with the temperature to the fourth. Equation
3.34 indicates that very high temperatures for the TH reservoir increases the loss due
to re-radiation.

Figure 3.13: The maximum efficiency of a solar thermal power plant at different
temperatures and light concentration (C). This expression accounts for both efficiency
of sunlight to heat, after account for re-radiation, and the efficiency of the heat to
work through a Carnot engine.
The full sunlight to heat efficiency can be writen as the product of the sun to heat
and the heat to work efficiencies:
solar to work = solar to heat heat to work
At low temperatures, solar to heat is high due to low re-radiation but heat to work is
low in accordance with the Carnot efficiency. At high temperatures, the opposite is

CHAPTER 3. SOLAR THERMAL, PHASE TRANSITIONS, HEAT ENGINES 76

true. Thus, there exists a limit on the optimal temperature and maximum efficiency
of a solar thermal plant. However, one approach around this is to use concentrated
sunlight, such that the area for solar irradiation is not the same as the area for
re-radiation. In other words, when using concentrated sunlight, we can reduce the
effective area in equation 3.34 while keeping the total solar energy constant. As a
result, concentrating solar power yields an increased maximum power and increased
efficiency. Figure 3.13 shows the maximum efficiency of a solar thermal power plant
as a function of TH and concentration ratios (C, where 1 C is the unconcentrated
sunlight) accounting for both the Carnot efficiency and the reradiation of the hot
reservoir. Increased light concentration increases the maximum temperature and the
efficiency of solar thermal plants. At infinite concentrations, the area for re-radiation
is negligible, and the overall efficiency approaches the Carnot efficiency.

You might also like