THERMODYNAMIC POTENTIALS AND PROPERTIES

Thermodynamic potentials and properties .................................................................................................... 1

Thermodynamic potentials. System equilibrium ...................................................................................... 2
Equations of state and thermodynamic coefficients ................................................................................. 4
System stability ......................................................................................................................................... 5
General dependence of state functions...................................................................................................... 5
The material data needed by Thermodynamics ............................................................................................ 8
Thermodynamic models of substances ..................................................................................................... 8
Perfect liquid model .............................................................................................................................. 9
Incompressible liquid model ................................................................................................................. 9
Dilatable liquid model ........................................................................................................................... 9
Perfect gas model ................................................................................................................................ 10
Ideal gas model ................................................................................................................................... 10
Van der Waals model .......................................................................................................................... 10
Van der Waals reduced model ............................................................................................................ 11
Redlich-Kwong model ........................................................................................................................ 11
Virial expansion model ....................................................................................................................... 12
Corresponding states model (introduction) ......................................................................................... 12
Properties of a pure substance and their graphical presentation. Diagrams............................................ 13
Simple pure substances ....................................................................................................................... 13
Phase diagrams for a pure substance................................................................................................... 14
Other thermodynamic diagrams .......................................................................................................... 15
Corresponding states model (analysis) ............................................................................................... 16
Boyle’s temperature. Inversion temperature. Acentric factor ............................................................. 18
Equations of state with three or more parameters ............................................................................... 19
Type of problems .................................................................................................................................... 20

THERMODYNAMIC POTENTIALS AND PROPERTIES

We deal here with two thermodynamic subjects that in most other presentations are separately covered:
'Properties of pure substances', usually covered at the very beginning of Thermodynamic courses, and
Potentials or 'General relations', usually covered at the very end of Thermodynamic courses. In our view,
it is better to deal at the same time with the functional structure of Thermodynamic (which functions are
needed and how others relate to them) and the actual models and particular values for those functions (the
material properties of selected substances). If not, one might end without realising the minimum set of
properties which define the thermodynamic behaviour of a substance, or the many important relations
amongst them, as the relation between vaporisation enthalpy and vapour pressure data, the relation
between the speed of sound and density data, etc.

It may be argued that the combined presentation of Potentials and Properties is too wide, but one may do
a short combined presentation (even shorter than here done) and leave most details for discussion under
other headings to follow: Phase changes, Real gases, Mixtures, etc.

Thermodynamic potentials and properties 1

Briefly, the aim here is to place on a common field the two isolated models of perfect gases and perfect
liquids that we have freely used, showing how one can model any possible behaviours of a substance at
any state, primarily for pure substances, but opening the door to any kind of mixture.

THERMODYNAMIC POTENTIALS. SYSTEM EQUILIBRIUM

An isolated system is at equilibrium when its entropy is a maximum, but what about a non-isolated
system?

We have introduced many state variables, like E (or U), V, S, T, p, i, and , others being not state
variables but path integrals, like W, Q, Emdf, Sgen and I. These state variables (many of them only defined
at equilibrium states), and many others we intend to introduce to ease our thermodynamic analysis, are
not all independent. In fact, the number of intensive independent variables at equilibrium, what is known
as the variance of the system, is just 2 for simple compressible systems (without change in composition),
and 2+C1 for a single-phase multicomponent system with C different chemical species, as deduced in
Chapter 2: Entropy, where we got T, p and i, (1+1+C) as intensive variables defining the equilibrium,
and we got the Gibbs-Duhem equation as their only constraint.

Thermodynamic potentials are single functions of some variables (call them their eigenvariables) from
which all other equilibrium variables can be deduced. The genuine thermodynamic potential is entropy,
from which T, p and i were derived in (2.4-5). We call eigenvariables the variables that were kept
constant during the evolution towards equilibrium; in the case of entropy they were, U, V and ni, thus, the
genuine thermodynamic potential is S(U,V,ni), and its total differential form dS=(1/T)dU+(p/T)dV-
(i/T)dni. From the many thermodynamic potentials one may define, we only consider the four energetic
potentials:

Internal energy: U(S,V,ni), dU=TdS-pdV+idni (4.1)

Enthalpy: H(S,p,ni)U+pV, dH=TdS+Vdp+idni (4.2)
Helmholtz function: A(T,V,ni) UTS, dA=SdT-pdV+idni (4.3)
Gibbs function: G(T,p,ni) U+pVTS=HTS=ini, dG=SdT+Vdp+idni (4.4)

where Euler equation (2.13) has been used in (4.4). Any of the above thermodynamic potentials, as a
function of its eigenvariables, contain all the thermodynamic information of the equilibrium states, and
may be physically interpreted in a similar manner as entropy: entropy was a function that got extremal
(maximum) as a system with constant U, V and ni (the eigenvariables) evolved towards equilibrium;
similarly, internal energy is a function that gets extremal (minimum in this case, Fig. 4.1) as a system
with constant S, V and ni (its eigenvariables) evolved towards equilibrium; similarly, enthalpy is a
function that gets extremal (minimum) as a system with constant S, p and ni (its eigenvariables) evolved
towards equilibrium; and so on.

Figure 4.1 helps to visualise the change from the entropy potential (that is maximum at equilibrium) to
the internal energy potential (that is minimum at equilibrium).

Thermodynamic potentials and properties 2

Fig. 4.1. Entropy variation with internal energy, U, and another unspecified non-equilibrium parameter, 
(e.g. a gradient of temperature). Notice that from a non-equilibrium state P, the equilibrium state
reached is different if the evolution is at U=constant (maximum entropy) or at S=constant
(minimum energy).

From the Mathematics point-of-view, the different potentials are just Legendre transformations as used in
other disciplines, as the change from the Lagrangian to the Hamiltonian function in Mechanics. In fact,
the existence of potential functions is not genuine of Thermodynamics but a general rule in all branches
of Physics, where one (the observer) tries to establish simple laws of Nature as: "this variable does not
change with time' or 'this variable takes a extremum value between two instants in time".

The interest in using these new thermodynamic functions lies in the following facts:
 Enthalpy, H, named by Kamerlingh Onnes in late 19th century as 'heat function', directly measures
the heat exchanged by a closed system in an isobaric process (the most common case in all
practical heat transfer cases), and directly measures heat-and-work exchange in flowing systems
(the most important systems in engineering). The basic idea to keep in mind is that enthalpy-
change, H, is the sum of energy-change, U, plus the work of expansion, pV.
 And A and G (traditionally named 'free energy' and 'free enthalpy', respectively), directly measure
the exergy of a process at T-and-V constant, and at T-and-p constant, respectively, the latter of key
interest for processes in the presence of an environment like the Earth atmosphere that can be
assumed to keep constant T=T0 and p=p0.

An isolated system evolves trying to maximise its entropy, at constant energy, but practical systems
interact with their surroundings (perfect isolation is a limit idea), and tend to minimise the
thermodynamic potential appropriate to their interaction with the environment. Thence:
 Isolated systems evolve such that dS/dt>0 (with dU/dt=0).
 Non-dissipative rigid systems evolve such that dU/dt<0 (with dS/dt=0).
 Real practical systems in an environment at constant T and p evolve such that dG/dt<0.

This is, for instance, why a liquid system (e.g. water in a closed evacuated flask) reaches an equilibrium
without all changing to a liquid state (that has less energy) or all changing to gas (that has more entropy);
instead, a two-phase liquid-vapour equilibrium is established such that both effects balance:
Gliq,vap=Hliq,vapTSliq,vap=0 (it will be shown as an Exercise in Chapter 9 that this corresponds to
T=373 K for water at 100 kPa). Notice also a general rule that can be deduced from dG=dHTdS<0: as
temperature increase, processes that increase entropy (mixing, gasification, decomposition, dissociation,

Thermodynamic potentials and properties 3

ionisation) are favoured against processes that decrease energy (that is also why water droplets and
oxygen molecules do not separate from lighter nitrogen molecules in the air).

EQUATIONS OF STATE AND THERMODYNAMIC COEFFICIENTS

Any of the thermodynamic potentials (e.g. G=G(T,p,ni)) holds all the thermodynamic information about a
system at equilibrium. If we call equations of state (EOS) to its partial derivatives (G/T, G/p,
G/ni), the set of all of them plus an irrelevant integration constant (because only increments of the
potential function were defined) also hold all the information. We will also introduce some functions
directly related to the second partial derivatives of the thermodynamic potentials, and call them
thermodynamic coefficients because they do not change much with the state; similarly, the set of all
thermodynamic coefficients plus two integration constants (one of them irrelevant) also holds all the
information about a system at equilibrium. Taking Gibbs potential because its eigenvariables are the more
natural, one gets from (4.4)

G /T p , ni
  S , and S=S(T,p,ni) is called energetic equation of state (4.5)
 G /  p T ,n  V , and V=V(T,p,ni) is called thermal equation of state (4.6)
i

 G /  ni T , p  i , and the i=i(T,p,ni) are called chemical equations of state (4.7)

But not all the equations of state are independent since they come from a single potential function and
thus must verify that crossed second derivatives match, a mathematical property of analytical functions
known as Schwarz relations in Mathematics and as Maxwell relations in Thermodynamics. That means
e.g. that:

 2G  2G  ( S ) V
   (4.8)
 T  p  p T  p T ,n  T i
p , ni

and this is the reason why for systems of constant composition there is only need of one equation of state,
chosen as the V=V(T,p) and simply named 'the equation of state' (but notice that it has not all the
information; it remains S/T).

From the second partial derivatives of the potentials we define the following so called thermodynamic
coefficients:

 Specific (by unit of mass) or molar (by unit of amount of substance) isobaric thermal capacity:

s
cp  T (4.9)
T p , ni

 Isobaric thermal expansion (dilation):

1 V
 (4.10)
V T p , ni

 Isothermal compressibility (similarly, an isentropic compressibility is also introduced):

Thermodynamic potentials and properties 4

 1  V
 (4.11)
V  p T ,ni

Exercise 1. Compatibility of equations of state

SYSTEM STABILITY
A system will evolve if some internal restrain is released (initial non-equilibrium) or if some constrain at
the frontier is imposed (boundary non-equilibrium). We know that for an isolated system, an equilibrium
state will be finally reached, although it might take longer than our life span (we call metastable
equilibrium when the system appears to be stable for finite times under small enough disturbances).
Stable systems return to their initial equilibrium state after a small perturbation is applied. Unstable
systems move away from their initial equilibrium state to a distant equilibrium state after a small
perturbation is applied, although metastable states may require a sizeable perturbation to become
unstable. In Thermodynamics, when a system gets unstable (as when heating water a lot), a new phase
appears (e.g. vapour), or new chemical compounds form, with different properties than the initial system
(different densities, different compositions, etc.), i.e. a different equilibrium is reached.

Besides indicating when a system becomes unstable and phase transitions appear, the analysis of the
stability teaches some general conditions for the thermodynamic functions. At a stable equilibrium, d2S<0
implies that T>0, p>0 and >0, as well as cp>0 (thermal stability), >0 (mechanical stability) and
i/nj>0 (chemical stability). Notice that no restriction holds on dilation, , which, although usually
positive, may be negative as in the case of pure water between 0 ºC and 4 ºC (there is some restriction,
really, because from the generalised Mayer equation, cp-cv=2vT/ (see below) and the stability
conditions one may concludes that 2<cp/(vT)). A more extensive analysis of System stability can be
found aside.

From now on till the study of Mixtures (Chapter 7) we only consider constant-composition systems, i.e.
evolutions that keep ni=constant, what applies to pure chemical substances, and to closed mixtures
without phase change, and, of course, without chemical reactions in any case.

GENERAL DEPENDENCE OF STATE FUNCTIONS

Many variables have been defined to characterise equilibrium states, and still some others are currently
used, as the heat capacity at constant volume:

s
cv  T (4.12)
T v

and the ratio of heat capacities:

cp
 (4.13)
cv

Thermodynamic potentials and properties 5

also known as isentropic exponent because of its appearance in (1.13) (and sometimes also isentropic
coefficient although it can be distinguished from a more general definition of the latter as k≡∂lnp/∂lnv|s,
coinciding with the heat-capacity ratio in the ideal gas model).

Also a variable directly related to the isentropic compressibility can be defined that later will be shown to
correspond to the speed of propagation of small perturbations (the speed of sound), c:

p 1
c  (4.14)
 s  s

Finally, a variable called compressibility factor (not to be confused with the compressibility coefficient) is
introduced as:

pv
Z (4.15)
RT

where R is the gas constant. But as said before, there are only two intensive state variables independent,
and all the others can be expressed as functions of these two. The natural choice for the two independent
variables is T and p, and the general dependence of the others, based on the definition of the coefficients
and on Maxwell relations) is as follows:

dv   vdT   vdp (4.16)

c c v
ds  p dT   vdp  p dT  dp 
T T T p
(4.17)
c  c c
 v dT  dv  v dp  p dv
T  T  Tv
 v 
dh  c p dT  (1   T )vdp  c p dT   v  T  dp 
 T p 
 (4.18)
 v  1  T c  c
  cv   dT  dv   v  v  dp  p dv
       v
 T   p 
du  cv dT    1 pdv  cv dT   T  p  dv 
 p   T v 
(4.19)
cv c 
  c p   vp  dT   T   p  vdp  dp   p  p  dv
  v 
 1  1
dZ  Z     dT  Z     dp (4.20)
 T  p

after which the following relations can be obtained:

h
cp  (4.21)
T p
u
cv  (4.22)
T v

Thermodynamic potentials and properties 6

  2 vT
c p  cv  (4.23)

c p  2v    cv 2 p
 T  vT   2   ; T (4.24)
p T 2  T v T 2 v
T p  p T

 
 (4.25)
T p pT
cp  c2
   (4.26)
cv  s cT2

where the isothermal speed of sound, cT, is also shown for historical reasons (it is no longer used after
experiments show that the actual speed of sound coincides with the isentropic model proposed by Laplace
and not with the isothermal model proposed by Newton).

Notice that (4.24) gives the general dependence of cp(p) and cv(v) on the EOS f(p,v,T)=0 (cp(T) and cv(T)
are unrelated to the EOS). For instance, if the EOS is of the form v=Tf(p), like for ideal gases, v=RT/p,
then d2v/dT2|p=0 and cp cannot depend on p, whereas if the EOS is of the form p=Tf(v), like for ideal
gases, p=RT/v (and for the van der Waal EOS too), then d2p/dT2|v=0 and cv cannot depend on v. Another
way to find the dependence of energy functions on non-thermal variables may be obtained from (4.18-
22): dh=cpdT+(vTv/T|p)dp and du=cvdT+(Tp/T|vp)dv.

Another variable of interest in isenthalpic expansions is the Joule-Kelvin coefficient, named for James
Prescott Joule and William Thomson (later 1st Baron Kelvin) who established the effect in 1852
elaborating earlier work by Joule on gas expansion at constant internal energy). The Joule-Kelvin or
Joule-Thomson coefficient, JK or JT, is defined as:

T (1   T )v 1 h
JK    (4.27)
p h cp c p p T

Usually, the Joule-Kelvin coefficient is positive and a gas cools a little when throttled (isenthalpic
expansion), what may be used for refrigeration (e.g. at room temperature, JK=2.2∙10-6 K/Pa for air,
JK=4.4∙10-6 K/Pa for CH4, JK=11∙10-6 K/Pa for CO2, JK=28∙10-6 K/Pa for NH3, JK=0.34∙10-6 K/Pa for
H2), but above a certain temperature, called inversion temperature (which depends on pressure) the
coefficient becomes negative, and a decrease in pressure causes an increase in temperature. The J-K
cooling may be important; e.g. if nitrogen from a bottle at 20 MPa and 300 K is let to flow along a
thermally-insulated porous plug to 100 kPa, the escaping gas is at 270 K (below 0 ºC). Be aware that the
inversion temperature just mentioned relates to T/p|h=0, which is unrelated to the temperature inversion
defined in meteorology, dT/dz=0 (i.e. the condition in which the temperature of the atmosphere increases
with altitude instead of the normal decreasing.

Besides the above state variables, density (=1/v), exergy (that is a combination of system and
environment variables as in (3.6)), mechanical energy (that is independent of the internal equilibrium
state), total energy (that is the sum of internal energy and mechanical energy), some other mechanical
variables (as the velocity or the height of the system), and a few physicochemical state variables to be
Thermodynamic potentials and properties 7
introduced later, Thermodynamics makes use of a few path integrals (not state functions): work, heat,
entropy generation and irreversibility. And the omnipresent universal gas constant Ru=8.314 J/(molK),
that enters because of the arbitrary choice of Avogadro number (the number of entities taken as unit of
amount of substance). Notice, by the way, that as no distinction is made on the writing of intensive
variables (e.g. v is used for both specific volume and molar volume), so no distinction is made on the
writing between the universal gas constant (the molar value given above) and the gas constant for a given
substance, RRu/M, where M is the molar mass of the substance.

Exercise 2. Generalised Mayer relation

Exercise 3. Sound speed in air and water

THE MATERIAL DATA NEEDED BY THERMODYNAMICS

THERMODYNAMIC MODELS OF SUBSTANCES
We concentrate here on models for pure substances, leaving to Chapter 7: Mixtures, the modelling of
mixtures (in terms of pure substance models).

It was said at the beginning of the Chapter that any potential function holds all the material properties of a
system at equilibrium, but it is unusual to find the thermodynamic data in such a form, except for the most
accurate reference data, nowadays compiled as a multiparameter empirical fitting of Helmholtz potential
A(T,). Most usually, the data are presented in two functions (T,p) and cp(T,p0), or in non-dimensional
form as Z(T,p) and cp(T,p0)/R, and always for a restricted region of validity in the T-p field. For the
condensed states of a substance p0=100 kPa is usually taken, but for the gaseous states (and condensed
states based on it), it is preferred to base the data on p00 since in this limit =p/(RT) and cp/p=0.

Before stating any particular model, an important idea to keep in mind is that "there isn't any model
covering all real details", a truth not only for thermodynamic data, but for any human-conceivable model.
Empirical models can be very accurate on the small domain covered by the experimental data used to fit
the parameters in the model, but they are usually nonsense when extrapolated. Theoretical models, based
on molecular models and statistical mechanics, may be not so accurate, but they have fewer substance-
dependent parameters, these parameters have physical meaning, and extrapolation may be good enough.
Sometimes, however, this dichotomy is arbitrary, since there are semi-empirical equations, which
combine features of theoretical and empirical equations (e.g. Antoine's equation versus Clausius-
Clapeyron vapour-pressure equation, Redlich-Kwong versus van der Waals equation of state, or the virial
expansion).

It is also customary to present the material data in a graphical form because the human brain works very
well with graphics (but computers deal much better crunching numbers than recognising graphical
patterns). The two functions, (T,p) and cp(T,p0), can be studied separately because their modelling is
different, but we have preferred here to deal with complete models at once.

Brief, thermodynamic computations today are done manually when using simple models, or automatically
for more accurate predictions using long lists of coefficients and some kind of polynomial or other type of
function fitting of experimental data, and graphics may be used to plan or check computations.

Thermodynamic potentials and properties 8

Amongst the most simple data models, we have the following ones (the recommendation is, as for any
model, to use the simplest model producing acceptable results):

Perfect liquid model

The perfect liquid model, PLM (perfect here means cp(T,p0)=c=constant), is also applicable to solids, and
is:

PLM: (T,p)==constant, and cp(T,p0)= c=constant (4.28)

A table of liquids and solids with their  and c values is presented aside. No subindex is given to the
thermal capacity here because with this model cp and cv coincide (and =0 and =0); this approximation
may be judged by the case of water at 25 ºC and 100 kPa: cp=4185 J(kg·K) and cv=4140 J(kg·K), just 1%.

Incompressible liquid model

The incompressible liquid model, ILM (not much used; incompressible here means =constant; it also
applies to solids), is:

ILM: (T,p)= =constant, and any cp(T,p0), usually c(T,p0)=aiTi with ai constant(4.29)

As an example, for liquid water at 100 kPa, c is nearly 4220 J/(kg∙K) at both 0 ºC and 100 ºC, with a
minimum of 4180 J/(kg∙K) at around 35 ºC. For liquid water at its vapour pressure, c=4500 J/(kg∙K) at
200 ºC, 5600 J/(kg∙K) at 300 ºC, growing to infinity at the critical point (374 ºC). Notice that it may be
inconsistent to account for T-changes on thermal capacity and not on density.

Dilatable liquid model

The dilatable liquid model, DLM (used in dilatometric and thermoelastic problems), also applicable to
solids, is:

DLM: (T,p)=0[1- (T-.T0)], with 0 and  constant, and any cp(T,p0) (4.30)

A list of thermal expansion coefficients is presented aside. Notice that linear expansion coefficients,
lin≡(1/L)L/T|p=vol/3, are commonly tabulated and used in connection with solid materials (because
only their main dimension is of interest), whereas volumetric expansion coefficients (4.10) are always
used for fluids. In spite of the fact that for Materials Engineering the basic thermal data are thermal
expansion and thermal conductivity, only trivial problems of dilatometry are usually tackled in
Thermodynamics, leaving real thermoelastic problems (including thermal shock and thermal crack) to be
analysed in Elasticity and Materials Science. When designing structures, mechanisms, and even coatings,
it is important to account for expansion joints (if not, they will be established by themselves!). Notice, for
instance, that reinforced concrete works well because steel-rods and concrete have similar expansion
coefficients, =12·10-6 K-1; otherwise tensions and cracks would develop when the temperature changed.
Notice also that expansion may be due to other than thermal phenomena, as to moisture (e.g. moist wood
expands, but moist ropes contract).

Thermodynamic potentials and properties 9

As an example, consider what might happen if a steel reservoir were totally filled with gasoline at a
certain temperature (e.g. 20 ºC), after some heating or cooling (e.g. by 10 ºC). The thermal expansion of
steel is linear=12∙10-6 1/K whereas for gasoline it is about vol=900∙10-6 1/K, what shows that, as in most
cases, liquid reservoirs can be considered rigid relative to the thermal expansion of the liquid.

On heating, two problems may arise, depending on if the reservoir is open or closed. If open, the
expanding liquid would overflow in the amount V=VT, where V is the volume of the reservoir, with a
risk of an open fire. If closed, a pressure rise of p=VT / would result, probably breaking the
reservoir.

On cooling, two problems may arise, depending on if the reservoir is open or closed. If open, the
contracting liquid would ingest an amount of ambient air V=VT, with a risk of an internal explosion
in the event of a spark. If closed, the contracting liquid would produce a void up to the vapour pressure of
the liquid; in spite of the high vapour pressure of gasolines (they are mixtures with composition varied
from winter to summer) would surely collapse the a thin-wall reservoir.

Exercise 4. Effect of thermal expansion on a pendulum

Perfect gas model

The perfect gas model, PGM, is the model most used in Thermodynamics. Perfect here means
cp(T,p0)=cp=constant, besides being an ideal gas. This model is:

p
PGM:  (T , p)  , (or more usually pv=RT), and cp(T,p0)= cp=constant (4.31)
RT

A data table of gases is presented aside, giving their molar mass M to compute RRu/M, their thermal
capacity cp, and some other properties.

Ideal gas model

The ideal gas model, IGM, usually refers just to the equation of state, but we said we only consider full
models (equation of state and thermal capacity). This model can be based on the dynamics of point-like
particles non-interacting among themselves, except for collisions. Notice that here 'ideal' means =p/(RT),
or in the more usual form pv=RT, whereas in Fluid Mechanics 'ideal' usually means inviscid. The
complete model is:

p
IGM:  (T , p)  , and any cp(T,p0), usually cp(T,p0)=aiTi with ai constant(4.32)
RT

A table of gases with their molar thermal capacity as a function of T, for very small pressures, p0, is
presented separately under Thermal data. Notice that, although cp(T,p0) is monotonically increasing, a
minimum appears at finite pressure values and low temperature; e.g. for dry air at 100 kPa, cp=1040
J/(kg·K) at 100 K, cp=1007 J/(kg·K) at 300 K, and cp=1141 J/(kg·K) at 1000 K.

Van der Waals model

The van der Waals model, VWM, is:

Thermodynamic potentials and properties 10

  a
VWM:  p  2   v  b   RT , with a and b constant, and any cp(T,p0) function (4.33)
 v 

This model is the simplest one that can be used for the whole fluid range (gas and liquid regions),
although it is more of academic interest than for real computations (its accuracy is not good). For this and
most other models, no simple explicit expression for the density exists; this is a cubic equation of state in
v (there have been many other cubic equation of state proposed, as Redlich-Kwong's, below). Other forms
of (4.33) are Z=v/(vb)a/(vRT) and v3(b+RT/p)v2+(a/p)vab/p=0.

Van der Waals proposed that equation of state in 1875 trying to enhance the ideal gas model, pV=mRT, to
incorporate two effects: the finite volume of the particles, and their attractive force. For the former effect,
he proposed to replace the whole volume of the reservoir by the free volume available to the particles,
once their own volume subtracted, i.e. passing from pV=mRT to p(VNd3/6)=mRT, with N being the
number of particles and d their diameter, assuming rigid spheres; notice that by experimentally measuring
b for a substance, with (4.33), one can work-out the size of the molecules by b=NAd3/6, NA=N/n being
Avogadro's number). As for the second effect, he proposed to replace the actual pressure, p, in pv=RT, by
the ideal pressure the particles would impose on the walls (which would be higher in absence of their
mutual attraction), p+pattr, with pattr being proportional to the density of particles, squared to account for
the surface effect (i.e. a spherical shell of particles pulling from a central particle, pattr1/v2, what finally
yields (4.33). Notice that by experimentally measuring a for a substance, with (4.33), one can work-out
the characteristic energy of interaction, 0, for the important Lennard-Jones potential:
u=40[(r0/r)12(r0/r)6], with r0=d/2 being the molecular radius, r the intermolecular distance, and u the
interaction energy between two molecules.

Van der Waals reduced model

The van-der-Waals reduced model, VWRM, is the simplest universal analytical model, i.e. a model
independent from the substance, whose properties only enter as scaling factors. This is a two-parameter
equation of state, obtained by analytically solving for the constants a and b in (4.33) in terms of the p-T-
values of the substance at its critical point (TCR and pCR, to be explained below), to yield (see Problem 5):

 3  1 8
VWRM:  pR  2   vR    TR , and any cp(T,p0) function (4.34)
 vR   3 3

Any other bi-parametric equations of state could be used in reduced variables in a similar way, as for
instance Redlich-Kwong's, but the numeric results would be different to the 3 and 8/3 in (4.34). In any
case, all these reduced analytical universal models are inaccurate and only of academic interest.

Exercise 5. Van der Waals equation in reduced form

Redlich-Kwong model
The Redlich-Kwong model, RKM, proposed in 1949, is reputed to be one of the most accurate among all
two-constant equations of state, also called van-der-Waals' derived equations (VWM extensions);
nevertheless, deviations of saturated liquid molar volume for a wide variety of substances lay typical
within (2020)% of experimental values (for the whole temperature range). This model is a VWM
Thermodynamic potentials and properties 11
correction to the attractive term in (4.33), a/v2, and, as most attraction-term extensions, is cubic in v.
(extensions to the repulsive-term in (4.33), vb, yield non-cubic equations). The Redlich-Kwong model
is:

 a 
RKM:  p    v  b   RT , with a and b constant, and any cp(T,p0) function(4.35)
 v  v  b  T 


Virial expansion model

The virial expansion model, VEM (from Lat. vis, force), proposed by Kamerlingh-Onnes in 1912, gives
an asymptotic series expansion to the thermal equation of state; i.e., we know that in the limit of very low
pressure (above absolute zero temperature) all substances become ideal gases, pv=RT (or Z=1), so it is
natural to expect that the real behaviour can be approach by an expansion on pressure (or in the inverse of
molar volume):

either Z  1  Bp  Cp 2  ...

VEM:  B' C' , with B=B(T), C=C(T)..., and any cp(T,p0) function(4.36)
 or Z  1    ...
 v v2

The simplest virial equation only retains the second term (B(T) or B'(T), B'(T)B(T)), what yields a
quadratic equation in v (not a cubic one, as for VWM, RKM and many others). It happens that the second
virial coefficient is negative for low temperatures and positive for high temperatures, showing the
dominance of intermolecular attraction over repulsion at low temperatures, and vice versa; e.g. when van
der Waals equation is expanded in v, one gets Z=1+(b-a/(RT))/v+b2/v2+...

Corresponding states model (introduction)

The molar volume of a fluid, v, at a given state, T-p, depends a lot on the particular fluid at hand, but it is
found in practice that if all values are scaled with values at the critical point of that substance, results are
quite similar, i.e. v/vCR=f(T/TCR,p/pCR) is only slightly dependent on the substance, what may be used as
an approximate prediction of p-v-T values based solely on critical point properties, and, of course, the
particular equation of state used (the function f). Any equation of state could be used to build such a
universal p-v-T plot of 'corresponding states’, as the VWRM presented above, but the most widely used
corresponding states model, CSM, and the only one deserving that name here, is to be a graphical plot
drawn by graphically fitting experimental data from several substances on a Z-pR diagram. Because of the
practical difficulty to differentiate and integrate graphically, auxiliary graphics are presented to help
compute enthalpy and entropy changes, as explained in the analysis below.

The CSM is universal except for the scaling in terms of TCR and pCR (i.e., applicable to any substance
without requiring special data about it, other than TCR and pCR), it is simple (not requiring heavy
computations but a look-and-see graph) and the most accurate of the models just presented (typical
uncertainty around 5% except near the critical point), justifying a detailed description here, after the
properties of a pure substance are first presented.

The corresponding states model can also be applied to the liquid state, but the uncertainty can be large.
2/7
Some useful correlations for reduced liquid densities, R=/cr, at saturation, are R=1/Zcr(1TR) , and
Thermodynamic potentials and properties 12
R=1+(3/4)(1TR)+(7/4)(1TR)1/3, the latter due to Guggenheim (1940s), who also proposed
R,V=1+(3/4)(1TR)(7/4)(1TR)1/3 for the reduced vapour density at saturation (for TR>0.5).

A key point in evaluation the goodness of a thermodynamic model for a fluid is its ability to yield
accurate vapour pressure data, what is best understood by using phase-change diagrams shown below.

PROPERTIES OF A PURE SUBSTANCE AND THEIR GRAPHICAL PRESENTATION. DIAGRAMS

Simple pure substances
Thermodynamic problems usually focus on simple chemically-pure substances and their mixtures, mostly
fluids: water, air, alcohols, short-chain hydrocarbons, and the like. A substance is a chemically identified
pure or mixed matter (solid or fluid): water, wine, oil, blood, DNA. A large share of substances of
practical interest are molecules composed of just hydrogen, H, and C, N, O and F atoms (see the periodic
table); including the lower row, Si, P, S and Cl one covers practically all fluids of interest.

Substances with small molecules show well-defined phase-transition points (e.g. 273 K and 373 K for
solid-liquid and liquid-gas transitions of pure water at 100 kPa without kinetic effects), whereas
substances with large molecules (macromolecules usually refers to M>>1 kg/mol, i.e. with more than 100
atoms per molecule) do not show so sharp phase-transition points, and are only stable in condensed form
because they chemically decompose on heating, before boiling.

Materials are solid bodies with intrinsic properties (apart of the shape) that render them useful, mainly for
structures, but also for electronics, optics, biomedicine, etc. They are usually classified in terms of
physical behaviour as: ceramics, metals, polymers and composites; although other times they are
classified in terms of their application (structural materials, electronic materials, and so on).

Except for air and water that are ubiquitous, most other working substances are procured from a local
supplier, what may be a nuisance for 'special-container substances' like gases (there are some additional
notes on natural and commercial gases aside).

Of course, the 'pure substance' is just an idealisation: there are always impurities in any real substance,
but a 90%-pure substance (10% impurities) may be good enough for some purposes, a 99%-pure
substance is usually named 'commercially pure', and a 99.99%-pure is usually referred as 'laboratory
pure', although it depends a lot on the application (e.g. in the electronics industry, sometimes 6 nines are
required, i.e. 99.9999% purity, less than 1 ppm of impurities). Besides, if there is no segregation within a
mixture during a thermodynamic process, the mixture may be treated as a pure substance, as usually made
with air (until segregation of one of its components, water vapour, is studied in the Humid Air chapter.

Properties of matter (as a system of N elementary particles of type say a) may be of three general types:
constitutive, if only depend on a; colligative, if only depend on N; and additives, if they depend on the
product of N and an attribute of a (e.g. mass, momentum, energy, entropy).

Simple chemically-pure substances, like water, usually have simple molecules, i.e. of a few atoms, and
their thermal behaviour is as follows (think on a cylinder-piston system to avoid contact with ambient

Thermodynamic potentials and properties 13

air). At low enough temperatures (near T=0 K) they are all crystalline solids (except the rarity of quantum
helium). When the solid is heated enough, it always melts at a precise temperature known as melting
temperature, Tm (or freezing temperature, Tf), which depends very little on pressure. When the liquid is
heated enough, it boils at a precise temperature known as boiling (or condensing) temperature, Tb. When
the gas is heated enough, it decomposes in atoms and ions and Chemical Thermodynamics is needed to
analyse these reactions. These solid-liquid-gas phase transitions apply only to few-atom molecular
substances and crystalline networks (metallic and ceramic), and only for quasi-static processes, but do not
apply to many-atom molecular substances like ceramic or plastic macromolecules that yield amorphous
solid phases, and do not show well-defined melting temperatures but glass-transition temperatures, where
thermal expansion has a jump (in crystalline solids it is density that shows a jump).

Phase diagrams for a pure substance

Phase transformations can be represented in a p-T diagram (called phase diagram for a pure substance), as
shown in Fig. 4.2 for the case of pure water at equilibrium (in a metastable state, distilled water at 100
kPa may be kept as undercooled liquid well below 0 ºC, and as superheated liquid above 100 ºC).
Although water is perhaps the simple substance showing more complex behaviour, all simple-fluids have
nearly the same layout, particularly if only the region close to the liquid-vapour equilibrium is of interest,
with a critical point CR, a triple point TR and the normal boiling point BP in between.

Fig. 4.2. Phase diagram for H2O to scale. Notice how different they may look in linear scale, and in
logarithmic scale for pressure.

A first feature of p-T diagrams is a quasi-vertical line separating the solid and liquid phases, showing the
fact that the melting point scarcely depends on pressure (except at huge pressures, of interest only in
geology and astrophysics, up to now, where all substances tend to the solid state, usually a different
allotrope of the normal solid, see Fig. 4.2b). Although the melting line is practically vertical in all cases,
its inclination is usually exaggerated in sketches to show that most substances have a positive slope, but
water and a few rare others have negative slopes, as later explained.

A second feature is that the boiling temperature depends a lot on applied pressure and two limits appear in
the vapour pressure curve, known as triple point (TR) and critical point (CR) respectively (Fig. 4.2). For
low enough pressures, the solid directly sublimates to gas without passing through the liquid phase, and
for high enough pressures the liquid does not boil but expands and expands without phase change.

Thermodynamic potentials and properties 14

Because a smooth process (without phase changes) can join a liquid state to a gas state, and vice versa,
around the critical point, it is a matter of consensus to distinguish between a liquid state and a gas state;
usually, a fluid state is named ‘liquid’ if it is below its critical pressure and to the left of the vapour-
pressure curve in Fig. 4.2; a fluid state is named ‘vapour’ if it is below its vapour-pressure curve and to
the left of its critical temperature; a fluid state is named ‘gas’ if it is below its critical pressure and to the
right of its critical temperature; finally, a fluid state is named ‘supercritical fluid’ if it is above its critical
pressure (usually in the range p/pCR=1..2 and T/TCR=1.0..1.1). The term 'vapour pressure curve' usually
refers to the liquid-vapour-equilibrium (LVE) vapour pressure, but sometimes the solid-vapour-
equilibrium curve is included.

The liquid state is not abundant in the universe, with the remarkable exception of water on Earth, but
there can be realised very-low-temperature liquids (cryogenic fluids, with quantum helium being the only
substance remaining liquid as T→0 for p<2.5 MPa), and very-high-temperature liquids (liquid metals; at
room pressure, hafnium is a liquid from 2500 K to 4900 K).

Notice that the melting line cannot end in a critical solid-liquid point but splits at additional triple points
(solid-solid-liquid), because there cannot be a smooth transition from a crystalline symmetry to the liquid
randomness. Triple points of pure substances of low molar mass are fixed points, i.e. can only occur at a
precise temperature and pressure, and they are key points in thermometry. The approximation of triple-
point data by the melting temperature and its corresponding vapour pressure is usually good enough for
most purposes (e.g. the triple point of water is TTR273.16 K (by definition of the temperature unit) and
pTR6111 Pa (by experimentation), whereas the melting point is 273.15 K and its corresponding vapour
pressure using Antoine equation, to be explained in Chapter 6: Phase change, is pv(273.15 K)=620 Pa; a
very good approximation.

A final feature is that the axes (T=0, p=0) in the phase diagram are unattainable limits (nowadays some
10-6 K and some 10-12 Pa have been reached, but 0 K would imply an infinite series of cooling steps and 0
Pa would imply lack of any mass at sight). As this singular point is approached, the the slope of the
sublimation curve tends to zero in the p-T diagram, and to infinity in a lnp-T diagram

Other thermodynamic diagrams

Thermodynamic diagrams are graphical plots of, either the substance properties (e.g. the p-h diagram for
water), or some particular processes applied to the working substance (as used in most thermodynamic
exercises, e.g. Ex. 2.3).

Besides the p-T or phase diagram, several other diagrams are in current use to present pure-substance data
or to sketch particular processes of a pure substance, as shown in Fig. 4.3. The p-v diagram is mainly used
to sketch mechanical processes like cylinder-piston evolutions. The p-h diagram is most used to present
data for substances, and in refrigeration problems. The T-s diagram is the most used in teaching and
sketches of general process. The h-s diagram is the traditionally used in steam problems. Notice that the
triple point in the p-T diagram becomes a triple line in the other diagrams, due to the coexistence of three
phases with different properties, whereas the critical point remains a point because the two coexisting
phases have the same properties. It must be noticed also that, to present a complete set of data for a

Thermodynamic potentials and properties 15

substance, one needs one potential, or two equations of state, so that it is not enough to have the family of
isochors (lines of constant volume) in a p-T diagram, neither the family of isotherms in the p-h diagram,
nor the family of isobars in a T-s diagram; but the family of isentropics in a p-h diagram suffice, as well
as the family of isobars in a h-s diagram, because the potential h(s,p) is made available in both cases.
Exercise 6. The h-s or Mollier diagram

Fig. 4.3. Main thermodynamic diagrams used to plot pure-substance data or to sketch particular
processes.

More accurate than the graphical presentation of data (although less handy) is a set of tables with
numerical values, usually a double entry table for the gas phase and a single entry one for the saturated
states.

Notice that some facts on the behaviour of pure substances may be misleading when mixtures are
considered. Some examples follow: One may believe that water cannot be solid or gas at room conditions;
one says “at 15 ºC and 100 kPa, water is a liquid” when one should say “at 15 ºC and 100 kPa, pure water
is a liquid”, but at the same conditions water is a gas when dissolved in air. One may believe that air
cannot be in the liquid state at room conditions, in spite that all fish and aquatic plants breathe it (gases
and solids dissolved in liquids are in the liquid state). One may believe that the critical point marks the
maximum temperature for a liquid, without realising that the critical point for mixtures departs a lot from
the critical points of pure substances. One may believe that the maximum density for a substance is when
as a pure solid (or pure liquid in some exceptional cases like water or silicon), without ever thinking on
the possibility that a mixture with another substance might enclose more molecules per unit volume than
the pure condensed phase (e.g. one litre of metal hydride may hold more hydrogen than if filled with
liquid hydrogen).

Corresponding states model (analysis)

As said before, graphical representations of simple pure-substance data, on the same diagram, have all the
same topology (the one shown in Fig. 4.3), suggesting that if conveniently scaled with the triple, or better
the critical point data (because this is the key point in the whole fluid phase, liquid and vapour), they
might match. Of course real substances do not match perfectly when plotted in reduced variables, but the
accuracy may be acceptable and the study worth enough (it would perfectly match if the molecular model
had only two parameters, as in the Lennard-Jones model referred above, or for any other two-parameters
model as VWM or RKM).

The most widely used corresponding states diagram (see above how they can be generated) comes back to
Hougen and Watson in 1900, who chose an overlay fitting with ZCR=0.27 (see following point), and is
sketched in Fig. 4.4 and plotted aside under Thermal data:

Thermodynamic potentials and properties 16

 Fig. 4.4. Sketch of the compressibility factor, Z, and the compressibility corrections hcc and scc , in
terms of reduced pressure and temperature, TR≡T/TCR and pR≡p/pCR, for the corresponding
states model (CSM).

The value of the compressibility factor at the critical point, ZCR=pCRvCR/(RTCR) (not to be confused with
the compressibility coefficient, , which tends to infinity at the critical point), most clearly shows the
difference between reduced models and experiments. Experimental data show that ZCR(H2O)=0.23,
ZCR(NH3, alcohols, ketones...)=0.24..0.26, ZCR(hydrocarbons)=0.26..0.28, and ZCR(permanent
gases)=0.28..0.30, whereas for the analytical models ZCR(PGM, IGM)=1, ZCR(vWRM)=0.38,
ZCR(RKRM)=0.32, and the choice of ZCR(CSM)=0.27. Instead of ZCR, other reduced parameters may be
chosen to compare models with experiments; e.g. the reduced boiling temperature, TbR≡T/bTCR, the
reduced Boyle’s temperature, TBR≡T/BTCR (see following point), the reduced inversion temperature,
TIR≡T/ITCR, the acentric factor (see following point), or the complete vapour-pressure curve (in reduced
variables); a simple approximation to the latter is Guggenheim's rule, lnpR=A(11/TR), with A6, as
explained in Chapter 6: Phase change. For many substances, the reduced boiling-point temperature is
TbR=0.6 (e.g. 0.58 for H2O, 0.59 for NH3, 0.64 for C4H10, 0.71 for CO2, 0.66 for CH3OH, 0.61 for N2,
0.62 for C3H8). The reduced triple-point temperature, however, has large scatter: TTPR=0.42 for H2O
(pTPR=0.0045), TTPR=0.71 for CO2 (pTPR=0.014), TTPR=0.48 for NH3 (pTPR=0.0088), TTPR=0.23 for C3H8
(pTPR=0.023), etc. If critical-point data for a substance is not available, a crude approximation in terms of
the normal boiling point is: TCR1.6Tb, CR0.37b, and pCR=ZCRCRRTCR.

Enthalpy and entropy corrections to the ideal gas state are performed with the help of the auxiliary graphs
h-pR and s-pR. Other functions are computed from enthalpy and entropy; e.g. u=h(pv). The
explanation of how the enthalpy and entropy correction can be computed from the compressibility factor
is very illustrative of how to use the data available in any model (CSM, VWM, RKM...). To begin with,
only increments of energy and entropy were defined operationally (Eqs. 1.3 and 2.6); thus the problem is
how to compute h and s from the given data v(T,p) and cp(T,p0). The general expressions to do that
are (4.17) and (4.18), and the integration can be performed along any desirable path since h and s are state
functions and thus path-independent. On view of the available data, it is advantageous to follow the three-
step path shown in Fig. 4.5 since this is the only one that only requires the values of cp at very low
pressure, thus (4.17) and (4.18) develop to:

Thermodynamic potentials and properties 17

 Fig. 4.5. This is the best integrating path from 1 to 2, on view of the data at p0.

p 0 T2 p2
h2  h1   (1  T1 )v dp   c p (T , p  0)dT   (1  T2 )v dp (4.36)
p1 T1 p 0

p 0 T2 c p (T , p  0) p2
s2  s1   ( v) dp   dT   ( v) dp (4.37)
p1 T1 T p 0

and in reduced variables, with v=ZRT/p from (4.15) and v=v/T|p from 4.16):

h2  h1 pR  0  Z dpR T2 c p (T , p  0) pR 2 Z dpR
 TR21   dTR  TR22  (4.38)
RTCR pR1  TR pR
pR T1 R pR  0  T
R pR
pR

pR  0  Z  dp T2 c (T , p  0)
  dp
s2  s1  Z  TR  Z  R (4.39)
pR 2
  Z  TR1  R  p R dTR  
R p   TR  pR T RTR p  0  2
 TR2  pR
 pR 
R 1 R
 pR 
1

what may be viewed as an ideal gas contribution (that may be further approximated with the perfect gas
model), plus some compressibility corrections in the way::

h2  h1   h2id  h1id   Dh2cc  Dh1cc (4.40)

s2  s1   s2id  s1id   Ds2cc  Ds1cc (4.41)

where:

Δhid T2 c p (TR , p  0) Δhcc Z dpR


pR
dTR ,  TR2  (4.42)
RTCR T1 R RTCR pR  0  T pR
R pR

T2 c p (TR , p  0)
  dp
Δsid Δs cc  Z  1  TR  Z
pR2
 R (4.43)
pR
 dTR  ln , 
R T1 RTR pR1 R p R  0   TR  pR
 pR 

Boyle’s temperature. Inversion temperature. Acentric factor

Boyle's temperature, TB, is defined as the isotherm that goes out from p=0 with zero slope in the Z-pR
diagram, i.e. it is the temperature at which real gas behaviour departs the least from the ideal gas model as
pressure increases (the closest to the horizontal line Z=1 in the Z-pR diagram); some experimental values
for Boyle temperature are: TB(N2)=327 K (TBRTB/TCR=2.60), TB(O2)=405 K (TBR=2.61), TB(Ar)=410 K
(TBR=2.72), TB(CO2)=713 K (TBR=2.35), TB(H2)=110 K (TBR=3.3), TB(CH4)=510 K (TBR=2.67), to be
compared with TBR(VWRM)=27/8=3.4, and TBR(CSM)=2.5. The fact that normal temperature on Earth
(around 300 K) is close to air Boyle’s temperature (around 350 K) makes the ideal gas model applicable
to air at relatively large pressure (which is not the case for most other gases).

Thermodynamic potentials and properties 18

The inversion temperature, TI, is the temperature separating the region of heating from the region of
cooling, when adiabatically throttling a gas, and corresponds to JK=0 in (4.27); some experimental
values for the inversion temperature are: TI(N2)=620 K (TIR=4.9), TI(O2)=765 K (TIR=4.9), TI(He)=45 K
(TIR=8.5), TI(Ne)=270 K (TIR=6.1), TI(H2)=200 K (TIR=6.1), TI(CO2)=1250 K (TIR=4.1), TI(CH4)=1010 K
(TIR=5.3), to be compared with TIR(VWRM)=27/4=6.8, and TIR(CSM)5.

The acentric factor, , introduced by Pitzer in 1955 as a measure of the non-sphericity of a gas molecule.
He realise that monoatomic gases (really all noble gases except helium) at TR=0.7 (a reduced temperature
he chose for comparisons), have a reduced vapour pressure, pvR, close to pvR=0.1 (i.e. log10pvR=1).
Besides, he realised that other gases departing the same amount from that reference point, have similar
departures in other reduced variables. He then proposed to define the acentric factor as
  1  log10 pvR T 0.7 , and use it as a third parameter in the enhanced corresponding state model
R

described below. Values of Pitzer's acentric factors can be found aside, where it can be seen that neon,
argon and xenon have 0.

Furthermore, the acentric factor may be used to have a reduced vapour pressure curve, more accurate than
Guggenheim's approximation, in the form lnpR=5.4(1+)(11/TR).

Equations of state with three or more parameters

The corresponding states method developed above (CSM) gives all the volumetric information of a fluid,
v(T,p), in terms of only two parameters for each fluid, TCR and pCR. Several enhanced corresponding
states methods have been developed, introducing additional parameters of the substance. The first choice
might seem to be a three-parameter model based on TCR, pCR and vCR (or ZCR=pCRvCR/(RTCR), but the
experimental measure of vCR has a large uncertainty (the position of a maximum), and it is preferable to
use the tern TCR, pCR and  (the acentric factor, defined just above), basically because two sets of graphics
(or tabulations) are enough, i.e. Z(pR,TR,)=Z0(pR,TR)+Z1(pR,TR) at least for ||<0.2, where Z0(pR,TR) is
a similar graph to the Z(pR,TR)-graph used in the simple corresponding state model (CSM), which was
prepared to best fit hydrocarbon gases with ZCR=0.27, but now fitting the noble-gas data (other than He,
as said), which have ZCR=0.29 (and pvR|TR=0.7=0.1, whereas in our Z(pR,TR)-graph pvR|TR=0.7<0.1;
pvR|TR=0.70.08).

As a substitute of this three-parameters extended corresponding-states-method, one may use the following
truncated virial expansion model in reduced variables, based on the term TCR, pCR and, that we name
acentric virial model, AVM), and which gives accurate predictions for most real gases (except highly
polar molecules like water) in the range 0.6<TR<2:

 0.083 0.422   0.139 0.172  

Z  1  Bp  1    2.6      5.2   pR , and any cp(T,p0) function (4.44)
  T R TR   T R TR  

Exercise 7. Corresponding states corrections

Exercise 8. Corrections based on an analytical equation of state

Thermodynamic potentials and properties 19

TYPE OF PROBLEMS
The housekeeping problems of how thermodynamic variables are dependent on each other, constitutes the
bulk of the problems in this Chapter that is the most formal and mathematical of all. The type of problems
may be quite varied:
1. Deduce one particular equation from others (many, according to the many equations presented).
2. Find the redundancy in some thermodynamic data given.
3. Deduce the shape of different curve families in the different diagrams.
4. Extract indirect thermodynamic functions from usual diagrams.
5. Compare the accuracy (and burden) of different thermodynamic models.
6. Use the corresponding states model to quickly elucidate the goodness of more simple models.

Back to Index

Thermodynamic potentials and properties 20