Vous êtes sur la page 1sur 17

PHY 252, Spring 2006

Lecture Notes on Thermodynamics


(based on Equilibrium Thermodynamics, 3rd ed., C. J. Adkins, C.U.P, 1983)
T. J. Newman
Lecture 1: Introduction to Thermodynamics, and Discussion of the Zeroth Law
• The subject of thermodynamics was developed in the nineteenth century, spurred
by the engineering demands of steam power in the industrial revolution.
• Remarkably, the laws of thermodynamics are still regarded as universally correct
– 150 years after their discovery. All modern theories in physics are consistent
with thermodynamics.
• One reason for the “longevity” of these laws is the fact that they are stated for
macroscopic systems. Thermodynamics has nothing to say about the microscopic
details of a system. Hence thermodynamics was untouched by the quantum
revolution. The laws of thermodynamics are expressed in terms of macroscopic
thermodynamic variables.
• Another remarkable feature of thermodynamics is its predictive power. From
three, rather vague laws, one is able to deduce a large number of powerful
statements about the relationships between various thermodynamic quantities.
• Let us begin with some definitions:
o a thermodynamic system is some particular region (of the universe)
separated from its surroundings by a boundary;
ƒ e.g. for a gas, the system would be the gas itself, and the boundary
could be a cylinder in which the gas is contained.
o thermodynamics is concerned with the response of the system to external
influences:
ƒ such influences may be macroscopic, in which case the energy
supplied to (or removed from) the system is termed work;
ƒ or, the influences may be microscopic (meaning that the influence
is transmitted via molecular motion with no macroscopic
quantification), in which case the energy supplied to (or removed
from) the system is termed heat.
o as examples of work and heat, consider a gas contained in a cylinder, at
one end of which is a frictionless piston. If the piston is pushed into the
cylinder, the gas is compressed, and energy is supplied to the gas as work.
Alternatively, if the piston is held static, but heated by a candle flame,
energy will be supplied to the gas via the piston in the form of heat.
o special types of boundaries (or walls) may be used to limit the external
influences. Adiabatic walls are such that no heat may be supplied to (or
removed from) the system. Examples of such boundaries are the silvered
walls of a thermos flask, between which is a vacuum – thus disallowing
transmission of heat via conduction or radiation. Walls which allow
transmission of heat are termed diathermal. Two systems connected via
diathermal walls are said to be in thermal contact. A system shielded
from both heat and work is said to be isolated.

1
• The state of a thermodynamic system is described by thermodynamic variables.
Such variables are either extensive or intensive. Extensive variables scale with
the mass of the system, e.g. volume, internal energy, length etc. Intensive
variables are essentially local in character, e.g. pressure, tension, etc. Intensive
variables (with the nature of a force) and extensive variables (with the nature of a
spatial displacement) appear together as conjugate pairs, and their product has
the dimension of energy; e.g. pressure and volume, tension and length.
• If a system is perturbed by heat or work it will typically be displaced from its
thermodynamic state. If a system, having been perturbed, is left unperturbed for a
sufficiently long time, it is generally found that the system will reach a state in
which no further change in its thermodynamic variables is measurable – this is
termed a state of thermodynamic equilibrium. It is important to remember, that
even though a system may be in a state of thermodynamic equilibrium, it is still
highly dynamic at a microscopic level; e,g, a macroscopic volume of gas may be
in thermodynamic equilibrium, such that no measurable change is observed in its
pressure or volume, but at the microscopic level, the molecules of the gas will be
moving and colliding with tremendous dynamic activity.
• An equilibrium state of the system can be described by particular values of the
thermodynamic variables, e.g. pressure and volume for a gas. A fundamental class
of processes (meaning changes of the state of the system) is reversible processes.
In a reversible process, each infinitesimal change to the state of the system takes
the system from one equilibrium state to another. Thus, each infinitesimal change
can be exactly reversed. Reversible processes must be performed slowly (i.e.,
quasistatically), and require the absence of memory in the system (i.e., no
hysteresis). The essential point is that a reversible process is such that at any time
the state of the system can be described fully by the macroscopic thermodynamic
variables. If a process is irreversible, density and energy gradients are initiated in
the system, and then the state of the system requires (spatio-temporal) information
at intermediate (“mesoscopic”) scales for its description.
• Since reversible processes take the system through a continuous set of equilibrium
states, such processes can be represented as a line on a diagram in the space of
thermodynamic variables. E.g., for a gas, a reversible process may be represented
by a line on a graph of pressure versus volume.
• It is natural to inquire as to how many thermodynamic variables are required to
uniquely define the thermodynamic state of the system (in equilibrium). This
number is referred to as the number of degrees of freedom of the system. This
number can be determined by listing all possible thermodynamic variables of the
system, and all possible relationships between them. The number of degrees of
freedom is then the number of variables minus the number of constraints.
• E.g., for a gas, we might list five variables: pressure (p), volume (V), temperature
(θ), density, mass. For a gas of constant mass, we have two immediate
constraints: i) mass=constant, and ii) density=mass/volume. So, the number of
degrees of freedom appears to be three (p,V,θ). In fact, each thermodynamic
system has an equation of state, which relates temperature to the other
thermodynamic variables. So, this constitutes a third constraint, and we therefore
have two degrees of freedom for our (simple) gas, conveniently listed as (p,V).

2
• This number could well be higher for more complicated systems, such as a gas
composed of ions, for which electromagnetic properties would be required to
define the state of the system.
• Having defined several terms and discussed some fundamental concepts of
thermodynamics, let us discuss the Zeroth Law of Thermodynamics. It is
known as such, as it was formulated years after the first and second laws were
well established.
• The zeroth law: “If systems A and B are each separately in thermal equilibrium
with a third system C, then A and B must also be in thermal equilibrium with each
other.”
• The zeroth law can be used to infer the existence of a quantity, the equality of
which for two different systems, defines the systems to be in thermal equilibrium
– this quantity is temperature.
• We give here a formal argument for the existence of temperature:
o we consider three systems, each of which, for the sake of argument, is a
gas, whose state is defined by pressure and volume;
o we fix the state of system 3 as (p3,V3);
o if system 1 is now brought into thermal equilibrium with system 3 then we
are not at liberty to change p1 and V1 independently – the fact of
equilibrium between systems 1 and 3 places an additional constraint on the
state of system 1. Therefore, for a given pressure in system 1, the volume
of system 1 must be uniquely determined. In other words, there must be a
functional relationship between the thermodynamic variables of systems 1
and 3;
o we write this relationship as F13(p1,V1,p3,V3)=0;
o similarly, for system 2 to be in thermal equilibrium with system 3 we have
a relationship between the thermodynamic variables of these two systems,
which we write as F23(p2,V2,p3,V3)=0;
o we can formally solve these two equations for p3 yielding
ƒ p3= f13(p1,V1,V3) and p3= f23(p2,V2,V3), which implies
ƒ f13(p1,V1,V3) = f23(p2,V2,V3) – equation (1);
o now, we invoke the zeroth law, which states that systems 1 and 2 must
also be in thermal equilibrium, implying that the thermodynamic variables
of systems 1 and 2 must be functionally related;
o thus, we write F12(p1,V1,p2,V2)=0 – equation (2);
o equation (2) states the functional relationship between the thermodynamic
variables of systems 1 and 2 with no need for explicit mention of the
volume of system 3 – thus, the variable V3 in equation (1) must cancel out,
leaving equation (1) in the form
ƒ Φ1(p1,V1)= Φ2(p2,V2) – equation (3);
o equation (3) is an expression of thermal equilibrium between systems 1
and 2 and shows that there exists a function Φ, of p and V, for each
system, yielding a numerical value; the equality of this numerical value for
systems 1 and 2 establishes the condition of thermal equilibrium;
o this numerical value θ is termed the empirical temperature, and the
relationship θ=Φ(p,V) is the equation of state for the gas.

3
• As we shall see in later lectures, a fundamental temperature emerges from the
theory of thermodynamics, and this is termed thermodynamic temperature, and
commonly given the symbol T.
• for an ideal gas (defined later), the equation of state takes a particularly simple
form:
o pV=RT for one mole of gas,
o where R=8.32 JK-1mol-1 is termed the gas constant;
• it turns out that at low pressures, all gases tend to the limit of the ideal gas;
• this provides an experimental means to measure the thermodynamic temperature;
namely, to measure the ratio of pV/R for a gas in the limit of low pressure;
• T is measured in units of Kelvins – one unit of Kelvin (symbol K) is equal to one
degree of Celsius (oC);
• the Kelvin scale is fixed by defining 273.16K to be the temperature at the triple
point of water (i.e., the temperature at which water coexists with ice and water
vapor);
• this, at first glance, odd number, establishes the Kelvin scale as a centigrade scale;
i.e. that there are 100 units of Kelvin between the freezing and boiling points of
water at one atmosphere of pressure;
• absolute zero occurs at 0K, which equals -273.15oC;
• gas thermometers, calibrated at low pressure, are used to establish rigorous
temperature measurements for international standards;
• for practical purposes, other types of thermometers are used, which use a physical
property which varies approximately linearly over a particular range of
temperatures; e.g.,
o mercury thermometer (range -39oC – 350oC)
o platinum resistance thermometer (range -200oC – 1200oC)
o radiation pyrometer (range > 1000oC).

4
Lecture 2: The First Law of Thermodynamics

• The First Law of Thermodynamics serves to place heat on an equal footing with
work as a form of energy which can be transmitted to and from a thermodynamic
system.
• The “common” form of the first law is “energy is conserved if heat is taken into
account.”
• A more rigorous form of the first law is: “If the state of a thermally isolated
system is changed by the performance of work, the amount of work required
depends only on the initial and final states of the system.”
• We can envision many different ways in which work (e.g. mechanical and
electrical) can be transmitted to a system, thereby changing the system from state
1 to state 2. The first law states that for a thermally isolated system, the amount of
work will always be the same, no matter what manner the work is performed, and
the consequent intermediate states the systems passes through.
• The immediate consequence of the first law is that a given change of state of a
system is exactly associated with a definite amount of energy – that energy
transmitted to the system as work. Thus, the total energy of the system must be a
function of the thermodynamic state of the system. This “energy of the system” is
called the internal energy of the system, and denoted by U.
• Denoting the work done on the system by W, the first law may be written as
o ∆U=W (for an adiabatic process).
• If work is applied to a system which is not thermally isolated, we write more
generally: ∆U=W+Q, where Q is the heat delivered to the system.
• In this sense, heat is defined as that form of energy responsible for changing the
internal energy of a system which cannot be accounted for in terms of measurable
work.
• Since U is a function of state, ∆U is uniquely defined given we know the initial
and final states of the system. The first law then tells us that the sum (W+Q) is
also uniquely defined – however, the heat Q or the work W are not independently
defined by knowing the initial and final states of the systems. Heat and work are
not functions of state – i.e. it is not meaningful to speak of the amount of heat or
amount of work contained in a system in a given thermodynamic state.
• We can write the first law in differential form (i.e. for infinitesimal changes in
state): dU = dW + dQ ;
o the symbol d indicates a small change in a quantity which is not a
function of state.
• Note, if a system is changed from state 1 to state 2 along a particular reversible
path, then the change in work, and hence the change in heat, are uniquely defined.

5
• Now, consider a set of systems in thermal contact with one another, but otherwise
isolated. The total amount of energy of the set is conserved, since as a composite
system, no heat or work is exchanged with the surroundings. However, heat
energy can be exchanged between the individual systems.
• So for the total system ∆U=0.
• Labeling each individual system by i we have
o ∆U = ∑ ∆U i = ∑ (Qi + Wi ) = 0 ;
i i

o but no work is performed and so Wi=0, yielding ∑Q


i
i = 0.

• So, for this set of systems, “heat is conserved.” This fact is the basis for the
“method of mixtures” in calorimetry. The amount of heat transferred to a
control system (measured by means of a change in temperature) can be used to
infer the heat transferred from a second system.
• Let us state our convention for the signs of work and heat. If W is positive this is
work performed on the system, and if Q is positive this is heat supplied to the
system. Thus, system 1 is “hotter” than system 2 if heat flows from 1 to 2 in the
approach to thermal equilibrium.
• In our discussion of the first law in differential form we write small changes in the
amounts of work and heat as dW and dQ , as these quantities are not “exact
differentials”, i.e. they are not small changes in functions of state. We shall see
that it is possible to rewrite these quantities in terms of exact differentials, so long
as the change from one state of the system to another occurs reversibly. We shall
examine this now for small changes in work done, but the discussion for heat
requires new concepts which emerge from the second law.
• If work is performed on a system such that the system changes from one state to
another reversibly, then we can relate the work done to the action of a generalized
force, and its conjugate generalized displacement: dW =Fdx.
• Let us consider the case of a gas in a cylinder, with force applied via a frictionless
piston (of area A) at one end. If the gas is in equilibrium at pressure p, then the
force applied by the piston is F=pA. We now imagine compressing the gas by
moving the piston inward by an amount dx. Then the work done is
o dW = Fdx = pAdx = -pdV , where dV is the (negative) volume change of
the gas.
• This expression is very useful for calculating the work done on a gas in changing
it reversibly from state 1 (defined by thermodynamic variables p1,V1) to state 2
(defined by thermodynamic variables p2,V2):
V2
o we have W = − ∫ p (V )dV
V1

• In order to evaluate this integral for a particular case, we need to know the volume
dependence of the pressure. As an example, consider an isothermal (i.e. “same
temperature”) process for an ideal gas.
• Then p(V)=RT/V, and consequently,
o W=RTln(V1/V2).

6
• An important measure of the thermodynamic nature of a system is the heat
capacity – this indicates the amount of heat absorbed by the system for a given
increase in temperature, with, importantly, some given constraint on the
thermodynamic variables.
• Common heat capacities are CV and Cp – the heat capacities at constant volume
and constant pressure respectively.
• For a gas we have dU = dW + dQ and dW = -pdV, which together give
o dQ = dU + pdV
• Taking volume and temperature as our thermodynamic variables we have
∂U ∂U ∂U ⎛ ∂U ⎞
dU = dT + dV , and so dQ = dT + ⎜ + p ⎟ dV , therefore our
∂T V ∂V T ∂T V ⎝ ∂V T ⎠
expression for the heat capacity at constant volume becomes
dQ ∂U
o CV = =
dT V ∂T V
• By taking pressure and temperature as the thermodynamic variables, we leave it
as an exercise to the reader to show that
dQ ∂U ∂V
o Cp = = +p
dT p ∂T p ∂T p
• From the two expressions above, assuming that the derivatives of the internal
energy are similar in size, it appears that generally CV < Cp since for most
substances the volume increases on increasing temperature at constant pressure,
∂V
and so > 0.
∂T p
o an exception is water at atmospheric pressure between 0oC and 4oC, in
which temperature range water expands on cooling.
• We will return to an explicit expression relating these two heat capacities for an
ideal gas, but first we shall define an ideal gas experimentally, and connect this
with some mathematical relationships.
• An ideal gas can be empirically defined as a gas which satisfies two experimental
laws:
o Boyle’s law: for isothermal processes in a gas, the product pV is constant;
o Joule’s law: the internal energy of a gas is a function of temperature only.
• It is possible to use these two laws to derive the equation of state for an ideal gas,
but this derivation requires results which follow from the second law.
• For now, we content ourselves with probing further the physical content of
Joule’s law.
o Joule’s experiment involved the free expansion of a gas – in this
experiment no external work is performed on the gas, and no heat is
exchanged with the surroundings;
o then, from the first law, dU=0.

7
• Experimentally, Joule found that under free expansion, the gas did not change
∂T
temperature, i.e. = 0.
∂V U
• Note, the constraint in this partial derivative is that the internal energy is held
fixed, since, as discussed above, dU=0.
• Using the reciprocity theorem of calculus, we have
∂T ∂T ∂U 1 ∂U
o =− =− , where in the last step we have used the
∂V U ∂U V ∂V T CV ∂V T
form of the heat capacity at constant volume derived above.
∂T ∂U
o Now, CV is a positive constant, and so = 0 implies = 0 ; in
∂V U ∂V T
other words, the internal energy does not depend on volume. Since we
have been using volume and temperature as our two independent
thermodynamic variables, this implies that internal energy is a function of
temperature only – i.e., for an ideal gas U=U(T).
• Returning to our expressions for the heat capacities above, we see that for an ideal
∂U ∂U ∂V
gas we have = , and so C p = CV + p .
∂T V ∂T p ∂T p
• As mentioned before, the equation of state for one mole of an ideal gas is pV=RT,
and so we easily find C p = CV + R . This expression is satisfied rather well for a
wide range of gases at low to moderate pressures.
• As a final topic in this lecture, let us briefly discuss enthalpy, otherwise known
by the misleading name of “heat content.” Enthalpy H is defined as
o H=U+pV, and is consequently a function of state.
• From this definition we see that dH=dU+pdV+Vdp. Using the first law, we have
dU = dQ − pdV , and so dH = dQ + Vdp .
• Thus, for thermodynamic processes at constant pressure (i.e. isobaric changes),
we have dH = dQ p . It follows from this, that the heat capacity at constant
∂H
pressure has a simple form in terms of the enthalpy: C p = .
∂T p
• Note, the change of enthalpy under isobaric change is equal to the heat absorbed
(hence the alternative name for enthalpy, heat content). Enthalpy is commonly
used to describe thermodynamic processes at constant pressure, such as chemical
reactions. It is also useful for the thermodynamic description of fluid flow with
heat transfer (i.e. generalizations of Bernoulli’s equation to account for heat).

8
Lecture 3: The Second Law of Thermodynamics
• The first law establishes heat as a form of energy and restates the conservation of
energy when work and heat are taken into account. However, the first law says
nothing about the direction of thermodynamic processes, which, as we have
discussed, is a fundamental element of the subject. When two systems at different
temperatures are placed in thermal contact, heat will be exchanged until the
composite system is in thermal equilibrium. Energy would be conserved by a
process in which heat is exchanged from a cooler to a hotter system, but this is
never observed (without the additional input of work). Thus, thermal processes in
nature have a particular directionality, and it is the function of the second law to
describe this.
• A second function of the second law is to describe the process by which useful
work may be produced from heat. As we shall see, there is a natural limit to such
processes, which can be quantified in terms of the efficiency of an engine
designed to accomplish this.
• A heat engine is a “thermodynamic machine” which operates by passing through
a set of thermodynamic processes which constitutes a cycle. After each cycle, a
certain amount of heat has been converted into work, and the heat engine returns
to its initial state. As we shall see, it is impossible for a heat engine to convert a
given amount of heat completely into work – there will always be a certain
amount of heat discarded. Thus, we can speak of the efficiency of a heat engine
in terms of the ratio of heat absorbed by the engine, to work produced:
work out W
efficiency, η = =
heat in Q1
• Referring to the expression above, Q1 is the heat absorbed by the engine in one
cycle, Q2 is the heat discarded, and W is the work performed. Applying the first
law, and remembering that the engine returns to its initial state after one cycle, we
have ∆U = 0 = Q1 − Q2 −W , and so W = Q1 – Q2. Therefore, we can rewrite the
efficiency as η = 1 – Q1/Q2.
• We now turn to a particular cyclic process which plays a fundamental role in
thermodynamics – this is known as the Carnot cycle. It consists of four
processes. It is convenient to describe the thermodynamic system undergoing the
cycle as the “working substance,” (w.s.). Each step of the cycle is reversible.
o i) w.s. expands isothermally at temperature θ1, absorbing heat Q1;
o ii) w.s. expands adiabatically, the temperature changing from θ1 toθ2;
o iii) w.s. is compressed isothermally at temperature θ2 rejecting heat Q2;
o iv) w.s. is compressed adiabatically from θ2 toθ1 (its initial state).
• It is a good exercise to draw the isotherms and adiabatic processes of the Carnot
cycle for a gas on a p-V diagram. The work done by the Carnot engine over one
cycle can be calculated as the area in the p-V plane enclosed by the cycle.
• Each of the four processes is reversible, and so a Carnot engine can be run in
reverse: absorbing heat Q2 at the lower temperature θ2 and delivering an amount
of heat Q1 at the higher temperature θ1. This requires that work be performed on
the engine. This type of Carnot cycle is essentially a heat pump.

9
• We now turn to the formal statement of the second law. In fact, there are two
statements, which can be shown to be equivalent:
o Kelvin statement: “No cyclic process is possible whose sole result is the
complete conversion of heat into work.”
o Clausius statement: “No cyclic process is possible whose sole result is
the transfer of heat from a cooler body to a hotter body.”
• The Kelvin statement places the emphasis on denying perfect efficiency of a heat
engine, while the Clausius statement places the emphasis on the natural
directionality of heat flow from hotter to cooler – i.e., the macroscopic
irreversibility of nature.
• To show the relatedness of these two different statements of the second law, let us
show that falsifying the Clausius statement is equivalent to falsifying the Kelvin
statement.
o We consider a heat engine which delivers heat to a thermal reservoir at
temperature θ1, having absorbed the heat from a thermal reservoir at the
lower temperature θ2. No work is performed on this engine, and thus the
engine violates the Clausius statement.
o We now couple this idealized engine to a “real” heat engine (not
necessarily a Carnot engine), which satisfies the second law. We will
adjust the rate of the second engine, such that in one cycle of the
composite engine, the same amount of heat (say, Q2) is delivered to the
cooler reservoir by the real engine as is absorbed by the idealized engine.
Then, in one cycle the composite engine absorbs an amount of heat Q1-Q2
from the hotter reservoir and converts this completely into work. Thus, the
composite engine violates the Kelvin statement.
• By a similar construction it is straightforward to show that a heat engine which
violates the Kelvin statement, when coupled to a real heat engine, leads to a
composite engine which violates the Clausius statement. Thus, these two
statements are complementary expressions of the second law.
• It is possible to use similar arguments to those above to prove that hotness and
temperature are directly related. We will not pursue this here, but we refer the
reader to Adkins, §4.4 for more details.
• We will now proceed with some arguments which lead to the definition of
thermodynamic temperature.
• First we shall prove Carnot’s theorem: “No heat engine operating between two
given reservoirs can be more efficient than a Carnot engine operating between
these same two reservoirs.”
• To prove this we make a composite engine from a Carnot engine and a
hypothetical engine which is more efficient. We then show that the composite
engine violates the Clausius statement of the second law.
o We denote the Carnot engine by C and the more efficient hypothetical
engine by H.
o We choose to drive C in reverse, such that it absorbs heat QC2 from the
reservoir at θ2 and delivers heat QC1 to the reservoir at θ1. H is run in the
forward direction, absorbing heat QH1 from the reservoir at θ1 and
delivering heat QH2 to the reservoir at θ2, producing work WH.

10
o We adjust the cycle of C such that it uses precisely the work WC=WH from
H in order to pump heat.
o Now, we have assumed that ηH>ηC, which, from the definition of
efficiency means that WH/QH1 > WC/QC1. Since WC=WH we have QC1>QH1.
o So the net amount of heat delivered toθ1 (which from the first law is the
same as the net amount of heat absorbed from θ2, since no net work is
done) is positive, and so the composite heat engine violates the Clausius
statement of the second law.
o Thus we have shown that ηC ≥ ηother.
• Consider replacing the hypothetical engine H by a reversible engine R (which is
not necessarily a Carnot engine). Then, the above proof amounts to ηC ≥ ηR. But,
given both C and R are reversible, we can run both engines in reverse, and by an
identical argument show that ηR ≥ ηC. These statements can only be reconciled if
ηC = ηR.
• Thus, we have the result that “All reversible heat engines working between the
same reservoirs are equally efficient.” The Carnot engine is special in that it is a
reversible engine with a single cycle working between two heat reservoirs.
• From the above result of the universal efficiency of reversible heat engines, it
must be the case that ηR depends only on the temperatures of the reservoirs
between which it works. Since η = 1 – Q2/Q1, we can write this previous
statement mathematically as Q1/Q2 = f(θ1,θ2) where f is a universal function of the
two temperatures. We stress this is true only for reversible engines.
• Let us now couple two Carnot engines together in the following manner:
o C1 absorbs heat Q1 from a reservoir at temperature θ1, and delivers heat Q2
to a cooler reservoir at temperature θ2. Likewise, C2 absorbs heat Q2 from
the reservoir at temperature θ2, and delivers heat Q3 to a still cooler
reservoir atθ3. Then we have Q1/Q2=f(θ1,θ2) and Q2/Q3=f(θ2,θ3).
o Now, the composite engine has the net effect of absorbing Q1 from θ1, and
delivering Q3 toθ2. Thus, Q1/Q3=f(θ1,θ3).
o This implies that f(θ1,θ3) = f(θ1,θ2) f(θ2,θ3).
o This equation can only be correct if f factorizes as f(θ1,θ2)=T1(θ1)/T2(θ2).
• Thus, we have Q1/Q2 = T1/T2, where T is defined as the thermodynamic
temperature. The unit of T is defined to be the Kelvin as discussed in Lecture 1.
• So, to conclude, the thermodynamic temperature T is defined so that
o “The ratio of the thermodynamic temperatures of two heat reservoirs is
equal to the ratio of the amount of heat exchanged between those
reservoirs by a reversible heat engine working between them.”
• Thus, for the efficiency of a reversible heat engine, we have η = 1 – T2/T1.
o The efficiency of a reversible heat engine increases as the temperature
difference between the two reservoirs at which it works increases.
o A final note: we have stated (but, not yet proved) that the equation of state
of an ideal gas is pV = RT , where T in this equation is the thermodynamic
temperature. This is why thermodynamic temperature is calibrated using
low pressure gas thermometry.

11
Lecture 4: Entropy

• In the previous lecture we proved a number of results related to the efficiency of


heat engines. In particular we proved i) Carnot’s theorem ηR ≥ ηother, where ηR is
the efficiency of a reversible engine, and ii) ηR=1 – T2/T1, for a reversible engine
working between two reservoirs at (thermodynamic) temperatures T1 and T2.
• Consider a reversible engine R and an arbitrary engine A working between two
thermal reservoirs. Let the amounts of heat absorbed and discarded be Q1 and Q2
for A, and Q1r and Q2r for R. We have ηR=1 – Q2r/Q1r = 1 – T2/T1, and η=1 –
Q2/Q1. Carnot’s theorem then implies that Q2/Q1≥ T2/T1, and so Q1/T1≤ Q2/T2.
• Our notation so far in our discussion of heat engines has been that Q2 is the
(positive) amount of heat discarded from the system at the temperature T2. Let us
make a more general definition that Qi is the heat absorbed by the system (i.e.
heat engine) from a reservoir at temperature Ti. Then, Qi is to be taken as negative
if it is discarded from the engine rather than absorbed. The advantage of this
notation is that we rewrite the inequality derived above in the simple form:
2

∑ Q / T ≤ 0 . Our aim in the next few paragraphs is to generalize this result to a


i =1
i i

completely general cyclic process occurring in a system (denoted by G), which


may contain an arbitrary number of heat transfer processes at different
temperatures, any number of which may be irreversible.
• Now, in a given infinitesimal part of the cycle (which we denote by i), a particular
amount of heat dQi will enter or leave G. We will quantify this heat transfer by
feeding it through a small Carnot engine C which operates between G and a
thermal reservoir held fixed at temperature T0.
• In one cycle of C (which occurs during the ith part of the cycle of G), an amount
of heat dQi' is absorbed by C from the reservoir, and an amount of heat dQi is
delivered from C to G at temperature Ti. If, during this small part of the cycle of
system G, G is in thermal equilibrium, then C and G are in thermal equilibrium at
Ti during the heat transfer from C to G. If, on the other had, G is not in thermal
equilibrium, we cannot define a temperature for G, and we then simply regard Ti
to be the temperature of C during which the delivery of heat from C to G occurs.
• Since C is reversible, we have dQi' /T0= dQi /Ti, and so dQi' =T0 dQi /Ti. Thus, over
the entire cycle of G, the amount of heat absorbed from the thermal reservoir is
∑ dQi' = T0 ∑ dQi / Ti .
i i

• If we now regard the general system G and the Carnot engine C as one composite
engine G′, the total heat absorbed from the thermal reservoir is equal to the total
heat delivered to G′. In one cycle of G′ the change of internal energy of the
composite engine is zero, and so, from the first law, we have that the amount of
work performed by G′ is simply WG′= T0 ∑ dQi / Ti . But, Kelvin’s statement of the
i
second law states that no cyclic process is possible in which the sole result is the
complete conversion of heat into work. Therefore, we must have that WG′≤0.

12
• Since T0>0, this immediately implies that ∑ dQ / T ≤0.
i
i i

• Taking the limit of each infinitesimal element of heat exchange to zero, we can
rewrite this fundamental result as v∫ dQ / T ≤0 for any cyclic process.
• Now, let us take this argument one more step forward. If the cycle in G is actually
reversible, then we could run both G and C in reverse, and in each infinitesimal
step the heat exchange between C and G would be exactly reversed. We would
then have v∫ dQ / T ≥0 for any reversible cyclic process.
• This relationship can only be compatible with our previous result if, for a
reversible process, v∫ dQ / T =0.
• To summarize, we have proven Clausius’ Theorem: “For any closed cycle,
v∫ dQ / T ≤0, where the equality necessarily holds for a reversible cycle.” It is
very important to remember that T is the temperature at which heat is supplied to
the system. In thermal equilibrium, T is also the temperature of the system itself.
• Now, finally, we are ready to introduce the fundamental quantity underlying the
second law; namely, entropy, which we denote by the symbol S. We define, for
an infinitesimal reversible change: the infinitesimal entropy change dS to be
dS = dQrev / T , where the “rev” subscript emphasizes reversible heat transfer.
• From this definition, we have for the entropy change from state 1 to state 2:
2
S 2 − S1 = ∫ dQrev / T . We can show that entropy, so defined, is a function of state.
1

• Consider a reversible cycle, and pick two arbitrary states on the cycle A and B.
o Denote the path (i.e. continuous set of states along the cycle) connecting
A to B by C, and the path connecting B to A by D. Then Clausius’
theorem states that v∫ dQ / T =0, and so, from the definition of entropy,
ACBDA

we have v∫
ACBDA
dS =0.

o But, v∫
ACBDA
dS = ∫
ACB
dS + ∫
BDA
dS , and so ∫
ACB
dS = − ∫
BDA
dS = ∫
ADB
dS .

o This last result shows that the entropy difference SB – SA is independent of


the intermediate states connecting A and B – and thus, entropy is a
function of state.
• Let us now consider the change in entropy for irreversible processes.
o We consider an irreversible change connecting states A and B. We choose
an arbitrary reversible path R connecting B to A, which then defines a
cycle.
o From Clausius’ theorem we have v∫ dQ / T ≤0.
B B
o Then ∫
A,irrev
dQ / T + ∫
BRA
dQ / T ≤0, which implies ∫
A,irrev
dQ / T ≤ − ∫
BRA
dQ / T .

13
B
o We can rewrite this last result as ∫
A,irrev
dQ / T ≤ ∫
ARB
dQ / T .

o Then remembering that ∫


ARB
dQ / T =
ARB
∫ dS = S B − S A , we have from the
B
above inequality: S B − S A ≥ ∫
A,irrev
dQ / T .

o Shrinking the states B and A to be infinitesimally close, this result can be


rewritten in the simple form dS ≥ dQ /T, where equality holds if the
change is reversible. This is the fundamental expression of the second
law of thermodynamics.
• Note, that in a thermally isolated system dQ =0, and so dS ≥ 0. This is the law of
increase of entropy: “the entropy of an isolated system cannot decrease.” Thus,
the equilibrium state of an isolated system is that state of maximal entropy.
This result, coupled with Boltzmann’s identification of entropy with the number
of microstates, forms the basis of equilibrium statistical mechanics. We will
discuss this briefly in future lectures.
• The fact that entropy cannot decrease (and thus the system evolves inexorably
toward thermal equilibrium) in an isolated system gives macroscopic systems a
directionality in time. Although the microscopic equations of motion are valid
with time running either forward or backward, only the forward direction is
permissible at the macroscale.
• We now turn to a more practical aspect of entropy.
o Recall our differential expression of the first law: dU = dQ + dW , which
is true for reversible and irreversible processes.
o We showed in lecture 2, that for reversible processes, we can write
dW = − pdV (considering the system to be a fluid).
o Likewise, we have just showed, that for reversible processes, we can write
dS = dQ / T , or equivalently: dQ = TdS .
o Thus, for reversible processes, we are able to write the first law in the
form: dU = TdS − pdV .
o But, and here is the key point, each term in the above equation is a
function of state, and so we can integrate this equation between any two
states of the system along any reversible path we choose, and always get
the same result. Thus, we have dU = TdS - pdV is true for any change, be
it reversible or irreversible. The point is that we can only identify the first
term on the right-hand-side with heat transfer (and, likewise, the second
term on the right-hand-side with work done) if the change is reversible.
However, the sum of these two terms is always equal to the change in the
internal energy of the system.
• It is noteworthy that the first law, when stated formally, introduces a new function
of state – the internal energy U. Likewise, as we have seen, the second law, after a
series of careful arguments, introduces another function of state – the entropy S.

14
Lecture 5: Miscellaneous Results in Thermodynamics
• From the first and second laws of thermodynamics we have derived the
fundamental equation dU = TdS − pdV .
• It proves useful to define related “energy functions,” otherwise known as
thermodynamic potentials:
o enthalpy: H = U + pV
o Helmholtz free energy: F = U – TS
o Gibbs free energy: G = U – TS + pV
• The significance of these potentials is more easily appreciated if we study their
differential form:
o dH = TdS + Vdp
o dF = –SdT – pdV
o dG = –SdT + Vdp
• Depending on the thermodynamic constraints on the system, it is convenient to
use a particular potential to describe the system. E.g., if the system is held at
constant temperature and constant volume, then these constraints are easily
incorporated by using the Helmholtz free energy – in this case, the equilibrium
condition for the system is simply dF=0. Similarly, the Gibbs free energy is often
used to describe equilibrium between phases (since they share the same pressure
and temperature). The Helmholtz free energy also has a fundamental significance
in statistical mechanics, as it is very simply related to the key quantity in this
subject; namely, the partition function.
o We will use the thermodynamic potentials here to derive some very useful
thermodynamic relations, known as the Maxwell relations. Consider the
differential form for the Helmholtz free energy: dF = –SdT – pdV. From
this equation we have expressions for the entropy and pressure, viz,
∂F ∂F
S =− and p = −
∂T V ∂V T
o Now, differentiating S with respect to V, and p with respect to T gives us
two equal expressions for the second derivative of F with respect to V and
T. We therefore have the Maxwell relation:
∂S ∂p
=
∂V T ∂T V
o We choose to rewrite this Maxwell relation by inverting the derivatives:
∂T ∂V
=
∂p V ∂S T
o The same procedure on the other three energy functions (U, H, and G)
yields three more Maxwell relations:
∂T ∂V ∂T ∂p ∂T ∂p
= , =− , =−
∂p s ∂S P ∂V S ∂S V ∂V p ∂S T
o These relations are extremely useful in thermodynamic calculations. For
instance, one can replace a derivative of entropy with a derivative of a
“simpler” thermodynamic variable such as temperature.

15
• Let us now reconsider the constant pressure and constant volume heat capacities:
dQ p dQ V
Cp = , CV = . Since dQ =TdS it is straightforward to show that
dT dT
∂S ∂S
Cp = T , CV = T .
∂T p ∂T V
• Now, we know that our simple fluid system is described by two degrees of
freedom. In other words, we can write any function of state as a function of two
thermodynamic variables. So, considering S as a function of T and V we
∂S ∂S
have dS = dT + dV . Likewise, considering V as a function of T and p
∂T V ∂V T
(i.e. formally writing the equation of state of the system as V=V(p,T)), we have
∂V ∂V
dV = dT + dp . Substituting this expression into the expression for dS,
∂T p ∂p T
∂S ∂S ∂V ∂S ∂V
we easily find dS = dT + dT + dp . From here, and the
∂T V ∂V T ∂T p ∂V T ∂p T
definitions of the heat capacities in terms of derivatives of entropy, as given
∂S ∂V
above, we find C p = CV + T .
∂V T ∂T p
• We can simplify this expression for the heat capacities by invoking a Maxwell
relation to replace the derivative of entropy with respect to volume. Thus, we find
2
∂p ∂V ∂p ⎛ ∂V ⎞
C p = CV + T = CV − T ⎜⎜ ⎟⎟ , where in the last step we have
∂T ∂T
V p ∂V T ⎝ ∂T p ⎠

used the reciprocity theorem of calculus.


• To relate this apparently abstract equation to physically measurable quantities, we
define two simple thermodynamic quantities: the isobaric cubic expansivity:
1 ∂V 1 ∂V
βp = , and the isothermal compressibility: κ T = − . Then, from
V ∂T p V ∂p T
the final relation for the heat capacities derived above, we have the result:
C p − CV = VT β p 2 / κT
• This is a fundamental and universal result, relating heat capacities to expansivity
and compressibility.
• Another universal result can be derived for the ratio of the heat capacities. We
1 ∂V
introduce the adiabatic compressibility κ S = − . Writing the ratio of the
V ∂p S
isothermal and adiabatic compressibilities, using the reciprocity theorem in both
numerator and denominater, rearranging terms, and then invoking the chain rule,
we find the beautiful result: κ T / κ S = C p / CV .
• This ratio of heat capacities is commonly denoted by the symbol γ.

16
• Let us end this lecture by returning to the ideal gas. In a previous lecture, we
defined the ideal gas by the two empirical laws: pV=constant (Boyle’s law), and
U=U(T) (i.e., internal energy is a function solely of temperature – Joule’s law).
Let us use these laws to derive the equation of state for the ideal gas.
• We have dU = TdS − pdV . Writing dS in terms of dT and dV, we can easily
∂U ∂S
show that =T − p . But, from Joule’s law, the left-hand-side is zero,
∂V T ∂V T
∂S
and so we have p = T . We can replace the derivative of entropy using a
∂V T
∂p
Maxwell relation, giving us the simple equation p = T . This may be
∂T V
immediately integrated to give ln p = ln T + f (V ) , where f(V) is an arbitrary
function of volume. We rearrange this equation in the form pf (V ) = T , where
f is an arbitrary function of V related to f. But from Boyle’s law, at constant
temperature, pV is constant, and so we must have f (V ) ∝ V . Thus we have the
equation of state of the ideal gas: pV = nRT , where n is the number of moles of
gas, and we have defined the constant of proportionality to be the molar gas
constant R = 8.12 JK-1.
• Finally we derive the equation for an adiabatic process for an ideal gas. We return
to our previous result for the ratio of heat capacities. We have
∂V
Cpκ ∂p T ∂V ∂V
γ= = T = . We can therefore write =γ .
CV κ S ∂V
∂p S
∂p T
∂p S
• Now, for an isothermal change we have Boyle’s law pV=constant, and so
∂V V
Vdp + pdV = 0 . This can be rearranged to give = − . Using this with our
∂p T p
expression above relating isothermal and adiabatic compressibilities, we have
∂V V
=− . This can be immediately integrated to give ln p = −γ ln V + const.
∂p S γp
(for an isentropic process). Rearranging, we have the familiar result pV γ = const.

• Here ends the supplemental notes on thermodynamics. There are many more
topics in thermodynamics which we now have the background to discuss, but we
have run out of time. Most interesting of all is the topic of changes of phase. We
will cover this in a less rigorous manner in the classroom.

17

Vous aimerez peut-être aussi