A summary of the properties

of van der Waals fluids

S. LE VENT, Department of Chemistry, University of Manchester
Institute of Science and Technology, Manchester, M60 1QD, UK.
〈[email protected]〉

Received 29th July 1999

Revised 24th February 2000

The well-known van der Waals equation of state is described in standard textbooks and
undergraduate courses in physical chemistry, and in chemical and mechanical engineering.
Such descriptions generally outline the origin of terms extra to those found in the perfect gas
equation and give typical values of the additional parameters involved. Generally, the
various implications for fluids satisfying the equation are not covered to any extent, such
coverage being scattered throughout more advanced chemical and engineering literature.
The purpose of the present paper is to collect together some of the more important of these.
Considered here are the various forms of the equation, vapour pressure implications, the
compression factor, thermodynamic aspects, Joule and Joule–Kelvin coefficients and
associated inversion phenomena.

Key words: equations of state, fluids, gases, thermodynamics

SYMBOLISM

Generally, quantity names and their symbols conform with those given in Quantities, Units
and Symbols of Physical Chemistry, 2nd edn., International Union of Pure and Applied
Chemistry (IUPAC), Blackwell, Oxford, 1993.

Am molar Helmholtz energy

E excess molar Helmholtz energy
Am
a, b, A, B equation of state parameters (usually van der Waals equation)
Cpm, Cvm isobaric and isochoric molar heat capacities
Gm molar Gibbs energy
E excess molar Gibbs energy
Gm
Hm molar enthalpy
E excess molar enthalpy
Hm
∆H(vap) molar enthalpy of vaporization
n amount of substance (or chemical amount)
p pressure
Pc critical pressure
Pe reduced vapour pressure
Pr reduced pressure

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
258 S. Le Vent

RbTi
q=
2a
Qi ith virial coefficient
R gas constant
Sm molar entropy
E excess molar entropy
Sm
∆So(vap) standard molar entropy of vaporization at chosen standard pressure po
T thermodynamic (or absolute) temperature
Tb Boyle temperature
Tbr reduced Boyle temperature
Tc critical temperature
Ti inversion temperature
Tr reduced temperature
Tir reduced inversion temperature
Um molar internal energy
E excess molar internal energy
Um
Vc critical (molar) volume
Vm molar volume
Vr reduced (molar) volume
Vrl, Vrg reduced volumes of liquid and gas phases at equilibrium
w acentric factor
z compression (formerly compressibility) factor
zc critical compression factor
α isobaric expansion coefficient
µ chemical potential
µl, µg chemical potentials of pure liquid and gas phases
µo standard chemical potential at chosen standard pressure po
µJ Joule coefficient
µJK Joule–Kelvin (or Joule–Thomson) coefficient
φ fugacity coefficient
φ1, φ2 fugacity coefficients at pressures p1 and p2

INTRODUCTION
The well-known extension of the perfect gas equation by Johannes Diderick van der Waals*
in 1873 [4]
 a 
 p + V 2  (Vm − b) = RT (1)
 m

(p = pressure, Vm = molar volume, T = thermodynamic temperature, R = gas constant) forms

an approximate basis for predicting various properties of fluids, i.e. liquids and gases. The
origins of the a and b parameters, each characteristic of a particular fluid (single- or multi-
component), are treated in standard textbooks of physical chemistry, chemical and mechani-
cal engineering, and in articles to a variety of degrees of sophistication, e.g. for two of the
more sophisticated treatments [5, 6]. Broadly speaking, a is associated with the existence of

* Photographs of van der Waals may be found in references [1–3].

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 259

intermolecular attractive forces and b with the idea of a finite molecular size or, in alterna-
tive terms, with the existence of strong, short-range, intermolecular repulsive forces.
Importantly, neither a nor b are dependent on T. Currently, there exist many other equations
of state with far better general predictive power than equation (1). Some of these are essen-
tially modifications of the van der Waals equation, still having two parameters,
commonly designated a and b, but, depending upon the detailed nature of the equation, with
a (but not b) having a dimension appropriate to the equation and distinct from the van der
Waals a, e.g. the Berthelot [7], the Redlich–Kwong [8], the Redlich–Kwong–Soave [9], and
the Peng–Robinson [10] equations. The last two of these are commonly used by chemical
engineers, who need more accurate predictions than those afforded by van der Waals
equation, particularly when dealing with particular classes of compounds, e.g. pure and
mixed hydrocarbons. Unlike the other equations already mentioned, their a factors are
temperature-dependent, the nature of the dependencies (for particular classes of compounds)
being deduced from vapour pressure data. The Dieterici equation [11, 12] is slightly different
in form to equation (1) but still has only two temperature-independent parameters. For some
purposes, it has better predictive features, e.g. for Joule–Kelvin inversion. Other equations of
state have more than two temperature-independent parameters; the Beattie–Bridgeman [13]
and the Benedict–Webb–Rubin [14] equations with five and eight parameters, respectively,
are but two examples of many. The so-called virial equations [15] have parameters which are
all temperature-dependent. They express pVm兾RT as truncated infinite power series of Vm−1
or p, with unity as the first term; depending on the predictive accuracy required, truncation
can be made at any term. In principle, such equations are capable of near perfect prediction,
but nevertheless have some limitations, not least in the limited knowledge of the parameters
at various temperatures [16], but also (in the author’s experience) a general inability to
produce satisfactory Joule–Kelvin inversion curves. Various texts, e.g. [17–19], present
general overviews of these various equation of state.
Despite the existence of the better equations, van der Waals equation still plays an
important part in the teaching of engineering thermodynamics and general physical chemis-
try. In particular, it gives good semi-quantitative explanations of observed consequences of
intermolecular forces, i.e. of deviations from perfect gas behaviour. Although these are
generally well documented, they tend to be scattered throughout the literature. The purpose
of this paper is to collect them together. In some cases, considerations will be appropriate
only for pure substances, but generally this will not be so.

OTHER FORMS OF EQUATION (1)

There are several alternative forms of equation (1). Four trivial, well established, modifica-
tions are as follows.

(a) Expand the equation as a cubic in Vm

 RT  a ab
Vm3 − Vm2  b +  + Vm − =0 (2)
 p  p p

(b) Isolate p on the left side of the equation

RT a
p= − (3)
Vm − b Vm2

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
260 S. Le Vent

(c) Multiply equation (3) by Vm兾RT and replace pVm兾RT by z, the compression (formerly
compressibility) factor

Vm a
z= − (4)
Vm − b RTVm

Binomial expansion of 1兾(Vm – b) allows comparison with virial equations in the

form
∞
∑ Vmii
Q
z = 1+
i =1

a
Q1 = b − ; Qi (i > 1) = b i
RT
(these being improperly temperature-independent).
(d) Replace molar volume, Vm in equations (1)–(4) by volume/amount of substance, V兾n,
with or without subsequent multiplication of the equation by n. Incidentally, it is a
frequent coupled misconception that equation (1) applies only to the special case of
n = 1 mol and that molar volume is the volume of 1 mole; both these points are quite
wrong.
A lesser known modification (better known in analogous application to the Redlich–Kwong–
Soave and Peng–Robinson equations) is obtained by introducing two new parameters:

pa pb
A= , B=
R2 T 2 RT

together with z into equation (1). This leads to a cubic in z, viz.,

z 3 − z 2 ( B + 1) + zA − AB = 0

Finally, there are the well-known reduced forms of equations (1) and (4). A single van der
Waals p versus Vm isotherm has a point of inflexion, for which the values of p, Vm and T are
the critical pressure, (molar) volume and temperature, pc, Vc and Tc for the particular mate-
rial (pure or mixed). Application of the inflexion point requirements,
(∂p ∂Vm ) T = (∂ 2 p ∂ 2 Vm ) T = 0 gives the well-known results: Vc = 3b, pc = a 27b 2 ,
Tc = 8a 27 Rb , zc (the critical compression factor) = pc Vc RTc = 3 8. This value of zc is in
fact much greater than common experimental values, often about 0.29 [20]; in this respect,
the Dieterici equation fares better with a predicted zc of 2 e 2 ≈ 0.271. Experimental values
of a and b can be obtained from pairs of critical constants (pc, Vc and Tc) but, as stressed by
Guggenheim [21], because zc inevitably deviates from 3兾8, it will depend upon which pair is
chosen; optimum a and b are better chosen by least squares fit of equation (1) to experimen-
tal p, Vm, T data. Reduced pressure is defined as pr = p兾pc, with reduced (molar) volume (Vr)
and reduced temperature (Tr) defined analogously. If, in equations (1) and (4), p is replaced
by pc pr = apr 27b 2 and likewise for Vm and T, we obtain the well-known reduced forms
of the van der Waals equation
 3 
 pr + V 2  (3Vr − 1) = 8Tr (5)
 r 

or

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 261

8Tr 3
pr = − (6)
3Vr − 1 Vr2
or
 1 8T  3 1
Vr3 − Vr2  + r  + Vr − =0 (7)
 3 3 pr  pr pr

3 pr Vr 1 9
z= = 1+ − (8)
8Tr 3Vr − 1 8Tr Vr
The important feature of these equations is their freedom from a and b parameters (and R).
This is a general feature of any equation of state with two temperature-independent
parameters, an indicator of the ‘Principle of Corresponding States’, also propounded by van
der Waals and essentially stating that all materials should have the same relationships be-
tween the three reduced quantities or between z and any two reduced quantities. The rela-
tionships are not specifically equations (5)–(8) but are based upon experiment and generally
presented as graphs or tables (rather than equations) (e.g. [22]). (Essentially arising from the
inadequacy of such relationships for more refined purposes, an additional parameter is com-
monly introduced as a modifier of ‘corresponding states’ graphs and table entries. Two
examples of this are zc and the acentric factor, w = –log10 (reduced vapour pressure at 0.7
Tr) – 1, these being specific to the particular material.)

VAPOUR PRESSURE PREDICTIONS FOR PURE FLUIDS

Equations (2) and (7) are soluble analytically for Vm or Vr (as appropriate) by Cardano’s
method [23] and (less tediously) numerically (for given T (or Tr) and p (or pr)) by such
techniques as the Newton–Raphson method. Alternatively, for graphical purposes, one can
simply calculate, for particular T (or Tr), values of p (or pr) for a range of V (or Vr). The well-
known results are shown in Figs 1(a) and (b), here for pr as a function of both Vr and log10Vr
for selected Tr; the second of these is generally regarded as the more satisfactory display.
The pressure is infinite for Vm = b, corresponding to Vr = 1兾3. Note that, for values of Tr
below a particular value, the pressure can improperly be zero or negative at finite values of
Vr. The particular value of Tr is 27兾32, readily deduced from the solution of the Vr quadratic
obtained by putting pr = 0 in equation (5) and then seeing where the two roots are equal; the
corresponding value for Vr is 2兾3. Other well known major qualitative failures of the van der
Waals equation are that, at specific temperatures lower than the critical value (Tr < 1), (a)
there is a range of volume where increased pressure leads to increased volume, and (b) there
is no recognition of equilibrium liquid-gas coexistence where, for a pure fluid, pressure (the
vapour pressure) is invariant over a range of volume from that of liquid on its own to that of
the gas on its own. This is not the case for fluid mixtures, where gas and liquid compositions
are generally different and depend upon relative volumes of the two phases.
For a pure fluid, constant pressure lines can be superimposed upon pr versus Vr (or
log Vr) plots by means of the so-called Maxwell construction [24]. For a particular isother-
mal, these lines intersect the equation-of-state curve (van der Waals or otherwise) at three
values of Vr. The lowest and highest of the molar volumes corresponding to these intersec-
tions, Vml and Vmg, correspond respectively to the extremes of liquid and gas only. The
chemical potentials corresponding to these single-phase points, µl and µg, must be equal
because of the equilibrium requirement. Since (∂µ ∂p) T = Vm , integration along the isother-
mal curve then gives

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
262 S. Le Vent

(a)

(b)

Fig. 1. Reduced isothermal plots of reduced pressure (pr) against (a) reduced volume (Vr)
and (b) common logarithm of Vr(log(Vr)). Each curve is labelled with the value of the
reduced temperature, Tr.

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 263

V = Vg
µg − µl = ∫V = V
l
Vm dp = 0

so
Vr = Vrg
∫V = V
r rl
Vr dpr = 0

Integration by parts then makes this requirement correspond to

Vrg Vrg
∫Vrl
pr dVr − pe (Vrg − Vrl ) = ∫V rl
( pr − pe ) dVr = 0 (9)

where pe is the reduced vapour pressure, and Vrl and Vrg are the reduced volumes corre-
sponding to Vl and Vg. Equation (9) corresponds to the pe horizontal cutting off equal areas,
above and below that horizontal, in Fig. 1(a). Values of pe have been obtained for various Tr
using the following MATHCAD procedure:
(i) Using the graphs of Fig. 1(a) as a rough guide, guess pe;
(ii) Calculate, by numerical solution of equation (6) with pr = pe (requiring initial guess-
ing for iteration), values of Vrl and Vrg—care has to be taken to avoid the intermediate
intersection point;
(iii) Evaluate the second integral of equation (9) (with pr given by equation (6)) and
observe its proximity to zero;
(iv) Progressively adjust the guessed pe until the integral is essentially zero—increasing pe
will reduce the value of the integral and vice-versa.

The time required for numerically evaluating integrals increased dramatically with decreas-
ing Tr—below a value of 0.45, there were problems of non-convergence. Values of pe at
various Tr are shown in Table 1.

Table 1. Calculated reduced vapour pressures

(pe) at various reduced temperatures (Tr) for
pure van der Waals fluids

Tr pe Tr pe

1 1 0.65 0.136
0.9 0.647 0.6 0.087
0.85 0.505 0.55 0.052
0.8 0.383 0.5 0.028
0.75 0.283 0.45 0.013
0.7 0.201

A test for proximity to Clapeyron–Clausius behaviour (which strictly assumes that the
gas phase is perfect) shows near linearity of ln(pe) with 1兾Tr—Fig. 2. The slope and inter-
cept of the least squares graph are –3.53 and 3.47, which would be equated, on a Clapeyron–
Clausius basis, to –∆H(vap)兾RTc and ∆So(vap)兾R – ln (pc兾po), where ∆H(vap) and ∆So(vap)
are the molar enthalpy and standard molar entropy of vaporization, the latter corresponding
to a standard pressure of po. This would imply that

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
264 S. Le Vent

Fig. 2. Dependence of natural logarithm of calculated reduced vapour pressure (ln(pe))

against reciprocal of reduced temperature (1兾Tr) for a pure van der Waals liquid.

(i) ∆H(vap)=29.4 J mol–1 Tc (10)

and

(ii) for po = 101.325 kPa, i.e. 1 atm,

∆S o (vap)  pc 
= 28.9 + 8.31 ln   (11)
J K −1 mol −1  101.325 kPa 

neither of which is supported experimentally. This point is illustrated in Table 2 by data for
the first ten n-alkanes [25].
If we now regard the two-phase constant pressure lines (commonly known as evaporation
condensation lines) as a necessary adjunct to the original van der Waals isothermals, the
curve joining the outer intersections of the lines and the corresponding curves will enclose
the region of liquid-gas equilibrium coexistence. It is known that, experimentally, this region
can be entered to a small extent while maintaining the original single phase, but this does not
correspond to an equilibrium situation. The region enclosed by the curve, joining van der
Waals maxima to the critical point and continuing as the curve joining van der Waals
minima, contains isothermals which show an increase of volume with increasing pressure;
since this situation is clearly impossible, the region is designated as the ‘unstable’ region
[24] (to a single phase). The portion of the two-phase region outside this is designated the
‘metastable’ region [24].

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 265

Table 2. Comparison of observed and calculated molar enthalpies and

entropies of vaporization for n-alkanes

Alkane Tc兾K pc兾MPa Tb兾K ∆H(vap)兾 ∆H(vap)兾 ∆So(vap)兾 ∆So(vap)兾

carbon kJ mol–1 kJ mol–1 J K–1 mol–1 J K–1 mol–1
number (experiment) (calculated) (estimated) (calculated)

1 190.6 4.61 111.7 8.2 5.6 73.4 60.6

2 305.5 4.90 185.2 14.7 9.0 79.4 61.1
3 369.8 4.25 231.1 19.0 10.9 82.2 59.9
4 425.1 3.78 272.7 22.4 12.5 82.2 59.0
5 469.8 3.37 309.2 25.8 13.8 83.5 58.0
6 507.7 3.03 341.9 28.9 14.9 84.5 57.1
7 540.3 2.74 371.6 31.8 15.9 85.6 56.3
8 568.8 2.49 398.9 34.4 16.7 86.2 55.5
9 594.7 2.29 424.0 36.9 17.5 87.0 54.8
10 617.8 2.11 447.3 38.8 18.2 86.8 54.1

∆H(vap) and ∆So(vap) values are for the normal boiling point Tb and a pressure of 101.325 kPa;
∆So(vap) = ∆H(vap)兾Tb.
Calculated values of ∆H(vap) and ∆So(vap) are based upon equations (10) and (11).

THE COMPRESSION FACTOR AND THE BOYLE TEMPERATURE

Van der Waals reduced z-isotherms for a variety of Tr are shown in Figs 3(a), (b) and (c); (c)
is merely an enlargement of the lower part of (b). They have the same general shapes as
‘corresponding states’ curves [26]. For Tr less than 1, the van der Waals equation would
predict three z values for a limited region below pr = 1; the two infinite gradient points on
each of these curves correspond to the maximum (higher pr) and minimum of corresponding
curves in Figs 1(a) and (b). For pure fluids, one should, of course, replace the reverse-S-
shaped curves by vertical lines at corresponding reduced vapour pressures (as given in the
previous section); these lines are not included in Figs 3(b) and (c). The upper and lower
intersections of the curves and lines are respectively the z values for gas and liquid phases,
when these are in equilibrium. At values of Tr less than or equal to 27兾32, the van der Waals
curves pass through the point (pr = 0, z = 0); Figs 3(b) and (c) include the unrealistic nega-
tive pr region for the sole purpose of showing curvature continuity. The origin point corre-
sponds to the unrealistic (pr = 0, finite Vr) points of the isotherms of Figs 1(a) and (b). As
with ‘Corresponding State’ curves, curves for Tr greater than about 2 cross within a small
region, here around (pr = 8.5, z = 1.3); as Tr → ∞, the curves approach the limiting curve
z = 1 + 1兾(3Vr – 1).
The Boyle temperature is defined as the temperature at which (∂z ∂p) T = (∂z ∂pr ) Tr = 0
at p = pr = 0. (For van der Waals (z versus P or pr) curves below Tr = 27兾32, one excludes
the two values of the differential coefficient at the unrealistic point (pr = 0, z = 0); in other
words, one must impose the additional limit of z = 1 for the derivative.) Above and below
the Boyle temperature, the limiting values of the differential coefficients are positive and
negative, respectively. Completely equivalent to these limiting values being zero are that
(∂z ∂Vm−1 ) T = (∂z ∂Vr−1 ) Tr = 0 at Vm = Vr = ∞. Partial differentiation of equation (8) and
application of the limiting condition then gives the reduced Boyle temperature
Tbr ( = Tb Tc ) = 27 8 ; multiplying this by Tc then gives the well established result that the
Boyle temperature Tb = a Rb . At the Boyle temperature, equation (8) becomes

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
266 S. Le Vent

(a)

(b)

Fig. 3(a) and (b).

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 267

(c)

Fig. 3. Relationship between compression factor (z) and reduced pressured (pr) for
various reduced temperatures (Tr), the values of which label each curve. (c) is an
enlargement of the lower end of (b).

1 1 1
z = 1+ − = 1+
3Vr − 1 3Vr 3Vr (3Vr − 1)
which is equal to 1 at infinite Vr (p = 0) and greater than 1 otherwise. At Vr = 1, this z is 7兾6
and at Vr = 10 it is 871兾870. The latter figure supports the common statement that, at the
Boyle temperature, gases are almost perfect (10Vc being a reasonably typical gas molar
volume: Vm = 940 cm3 mol–1 for CO2). (In terms of virial equations, the second (but not
higher) virial coefficients are zero at the Boyle temperature.)
At each temperature below the Boyle temperature, z is 1 both (i) at p = 0, i.e. at Vr = ∞
and (i) (using equation (8)) at a finite Vr equal to 9 (27 − 8Tr ) = 9 (8(Tbr − Tr )), i.e. to
ab b
Vm = = ;
a − bRT 1 − T
Tb
(when Tr > Trb, Vr is unacceptably negative). By equation (6), the (positive) volumes then
correspond to
8  T 
pr = Tr (27 − 8Tr ) = 8Tr 1 − r 
27  Tbr 
or to
1 RT  RT  T
p = RT  − = 1−
b a  b  Tb 

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
268 S. Le Vent

These are all quadratic relationships and yield similar results. Selecting the first pr versus Tr
relationship, pr is zero at Tr = 0 and at Tr = Tbr (= 27兾8), and pr is a maximum of 27兾4 at
Tr = 12 Tbr ( = 27 16). Below the maximum, there are two Tr values corresponding to each
unit-z pr, i.e. pairs of isothermals cross at z = 1, the two temperatures being equidistant from
1
T .
2 b

SOME ISOTHERMAL THERMODYNAMIC RELATIONSHIPS

The thermodynamic functions considered here are molar internal energy Um, molar enthalpy
Hm, molar entropy Sm, molar Helmholtz energy Am, and molar Gibbs energy Gm. Variation
of these quantities with molar volume will generally be expressed as functions of Vm and T.
Conversion to functions of p and T, via analytical solutions of equation (2), substantially
complicate the relationships. The key primary relationships here are

 ∂Am  RT a
 ∂  = −p = − + (12)
 Vm  T Vm − b Vm2

and
 ∂Sm   ∂p  R
 ∂  =   = (13)
 Vm  T ∂T Vm Vm − b

or in integrated forms
Vm1 − b  1 1 
∆Am = − RT ln − a − 
Vm2 − b  m1
V Vm2 

Vm1 − b
∆Sm = R ln
Vm2 − b
where ∆Xm (X = A or S and subsequently U, G, H or pV) is Xm(T, Vm1) – Xm(T, Vm2). Con-
stant composition is implicit in this and subsequent sections. The relationship for internal
energy change readily follows

 1 1 
∆U m = ∆Am + T∆Sm = − a − 
 Vm1 Vm2 

Relationships for ∆Gm and ∆Hm then follow by corresponding addition of

 V Vm2   1 1 
∆( pVm ) = RT  m1 −  − a − 
 Vm1 − b Vm2 − b   Vm1 Vm2 

to expressions for ∆Am and ∆Um. For pure fluids, Gm is equivalent to chemical potential µ of
the single species.
The various relationships for ∆Xm given here reduce to corresponding perfect gas rela-
tionships when a = b = 0. Furthermore, if these are then subtracted from the corresponding
van der Waals relationships and Vm2 allowed to approach infinity where all fluids become
perfect gases, expressions for excess molar quantities, X m E (= X (van der Waals) –
m
Xm (perfect)), are obtained as follows

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 269

Vm − b a
E
Am = − RT ln −
Vm Vm
Vm − b
E
Sm = R ln
Vm

a
E
Um =−
Vm
 b V − b  2a
E
= RT  − ln m −
Vm  Vm
Gm
 Vm − b

RTb 2a
E
Hm = −
Vm − b Vm
The difference between corresponding molar isobaric and isochoric heat capacities is
generally given by
2
  ∂V    ∂p 
T m   T 
  ∂T  p   ∂Vm  T
C pm − CVm =− =−
 ∂Vm    ∂T  
2
   
 ∂p  T   ∂Vm  
 p

which can then be combined with van der Waals specific relationships
 ∂p  RT 2a
∂  =− + (14)
 Vm  T (Vm − b) 2 Vm2

 ∂T  T 2 a(Vm − b)
∂  = − (15)
 Vm  p Vm − b RVm3

Incidentally,
(a) multiplication of equations (14) and (15) by –Vm and +Vm respectively, followed by
inversion gives the isothermal compressibility and isobaric expansion coefficients;
and
(b) division of corresponding sides of equations (12) and (13) by equation (14) gives
relationships for (∂Am ∂p) T and (∂Sm ∂p) T but still in terms of Vm and T.

The various thermodynamics relationships presented here will not be valid within the
two-phase region.

FUGACITY COEFFICIENT FOR PURE FLUIDS

For any pure fluid, fugacity coefficient φ is defined by

φp
µ = Gm = µ o + RT ln
po

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
270 S. Le Vent

φ is in fact independent of the choice of po. The dependence of chemical potential µ upon
pressure for a pure material is given by (∂µ ∂p) T = Vm . Combination of these relationships
followed by integration then gives

φ p  p1 Vm1
∫p ∫V
1 1
ln  1 1  = Vm dp = z1 − z2 − p dVm (16)
 φ 2 p2  RT 2 RT m2

the last step here using integration by parts. Combination of equations (3) and (16) and
performance of the integration then results in
pb
z1 − 1
φ1 RT − a + a
ln = z1 − z2 − ln
φ2 p
z2 − 2
b RTVm1 RTVm2
RT
Now letting p2 approach zero, whence φ 2 → 1, z2 → 1, a → 0, b → 0, Vm2 → ∞, so
(dropping subscript 1)

ln φ = z − 1 − ln  z −
pb  a
−
 RT  RTVm
Two other forms of this equation are
A
ln φ = z − 1 − ln( z − B) −
z
and, because
pc b 1 a 9
= and =
RTc 8 RTc Vc 8

 p  9
ln φ = z − 1 − ln  z − r  −
 8Tr  8Tr Vr

For the second of these, z can be replaced by one of its equivalents from equation (8).
Isothermal plots of ln φ against pr are shown in Figs 4(a), (b), (c) and (d). Below the
critical temperature, the plots show irregularities associated with the existence of three
values of Vr for each values of pr in the region of the two phase region. Below Tr of 27兾32—
Fig. 4(d)—there are more irregularities associated with the meaningless negative pr region;
these irregularities cannot be shown in Fig. 4(d) because Φ becomes negative at negative pr;
infinite ln Φ at zero pressure corresponds to finite volumes at zero pressure in Figs 1(a) and
(b). The crossing point in each Tr < 1 isotherm corresponds to the reduced vapour pressure
for that temperature. For a pure fluid, the ‘triangular’ section above that point is associated
with the van der Waals deviation from the vapour pressure (pe) within the two phase region.
The sharp ‘corners’ correspond to the maxima and minima in Figs 1(a) and (b); the ‘triangle’
side joining these corners corresponds to the ‘unstable’ region and for that reason has no real
meaning. Ignoring the triangular section, the parts of the residual curve to the low and high
pressure sides of the crossing point are respectively associated with the gas and liquid
phases. The gradient discontinuity at the crossing point becomes greater as Tr is reduced,
with the consequence of a rapid decrease of Φ for the liquid phase as pressure is increased.
An alternative semi-quantitative way of viewing this may be seen in terms of the alternative
general expression for ln Φ, viz.,

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 271

p V 1
ln φ = ∫0  RTm − p  dp
The integrand is generally much smaller in magnitude in the integration range 0 to pe than in
the range pe to some liquid pressure p; in the latter range it will be negative and reasonably
approximately to –1兾p.

JOULE AND JOULE–KELVIN COEFFICIENTS—INVERSION TEMPERATURES

The Joule coefficient, µJ, and the Joule–Kelvin (or Joule–Thomson) coefficient, µJK, are
associated with the temperature change which occurs when a gas is expanded adiabatically
into a vacuum, in the first case freely (constant internal energy) and in the second case
through a throttle (constant enthalpy). The coefficients are defined by µ J = (∂T ∂Vm )U m
and µ JK = (∂T ∂p) H m. Using the well known ‘–1 rule’ (in the form (∂ α兾∂β )γ =
–(∂γ兾∂β)α兾(∂γ兾∂α)β) and then various standard thermodynamic relationships, we obtain

 ∂U m   ∂Am   ∂Sm   ∂p 
∂   ∂  + T  p − T 
 Vm  T  m T
V  ∂Vm  T  ∂T  V
µJ = − =− = m (17)
 ∂U m  CVm CVm
 
 ∂T  V
m

and
 ∂H m   ∂Gm   ∂Sm 
    + T 
 ∂p  T  ∂p  T  ∂p  T
µ JK = − =−
 ∂H m  C pm
 
 ∂T  p

 ∂V   ∂(Vm T ) 
−Vm + T  m   
 ∂T  p  ∂T  p
= = T2 (18)
C pm C pm

Two other equivalent expressions for µJK are

Vm T  ∂ ln z  Vm
  and (Tα − 1)
C pm  ∂T  p C pm

where α is the isobaric expansion coefficient.

Both Joule and Joule–Kelvin coefficients are zero for perfect gases, but for a van der
Waals gas

 ∂p  a
p − T  =− 2
 ∂T  V Vm
m

and (∂Vm兾∂T)p is given by the inverse of equation (15). Consequently, equation (17)
becomes
a
µJ = −
CVm Vm2

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
272 S. Le Vent

(a)

(b)

Fig. 4(a) and (b).

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 273

(c)

(d)

Fig. 4. Dependence of natural logarithm of fugacity coefficient (ln φ) upon reduced

pressure (pr) for various reduced temperatures (Tr), the values of which label each
curve.

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
274 S. Le Vent

so that µJ must always be negative. Equation (18) becomes

T
−Vm +
T 2 a(Vm − b)
−
Vm − b RVm3
µ JK =
C pm
which allows µJK to be either positive or negative. Because Cpm is inevitably positive,
equation (18) indicates that µJK changes sign when (∂Vm ∂T ) p = Vm T or, in terms of
reduced quantities, when (∂Vr ∂T r ) p r = Vr T r . For a particular Vm, the temperature at
which µJK is zero is called the inversion temperature Ti.
Relationships between Ti and either Vm or p, or between corresponding reduced values,
can be determined as follows. Using equation (15),

Ti Ti 2 a(Vm − b)
= −
Vm Vm − b RVm3
which readily converts to

b RbTi
1− = = q (say)
Vm 2a

The various quantities, or combinations of quantities, in the right side of equation (3),
applied at the inversion temperature, are related to q as follows

b bq 2 aq 2
Vm = , Vm − b = , RTi =
1− q 1− q b

so that

2 aq 2 b a(1 − q ) 2 a(1 − q )(3q − 1)

p= − =
bq (1 − q ) b2 b2
which is essentially a relationship between Ti and p.
In terms of reduced quantities

9 6
− + + 3 pr
Tir  ∂Tr  Vr2 Vr3
= =
Vr  ∂Vr  p
(19)
8
r

by differentiation of equation (5). Substitution for pr from equation (6) followed by simplifi-
cation of equation (19) then gives

3(Vr − 1) 2
Tir =
4Vr2
an equation first obtained by Porter [27] in 1906. Rearrangement of this equation gives
1
Vr =
Tir
3−2
3

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 275

and, if this is substituted into equation (6), we obtain directly a relationship between pr and
Tir, viz.,

 T  T 
pr = 9 3 − 2 ir   2 ir − 1
 3  3 

When pr = 0, Tir = 3兾4 or 27兾4; pr reaches a maximum of 9 when Tir = 3. The relationship
between Tir and pr is shown in Fig. 5. If this figure is treated as a phase diagram, i.e. Tir
replaced by Tr, then within and outside the curve, µJK is positive and negative, respectively.
Very close to Tir = 3兾4, the inversion curve intersects the vapour pressure curve (not in-
cluded in the figure); on the low pr side of the intersection, the inversion curve would
represent the liquid phase. For a particular pr below 9, the two inversion temperatures, are
called the higher and lower inversion temperatures.

Fig. 5. Plot of reduced Joule-Kelvin inversion temperature (Tir) against reduced pressure
(pr). Within the curve, the Joule–Kelvin coefficient is positive; outside the curve, the
coefficient is negative.

It is well known that the van der Waals inversion curve does not agree too well with
experiment; the corresponding curve for the Dieterici equation (being thinner and longer) is
a much better predictor in fact [28]. Other, more complicated, equations of states have also
been shown to be reasonable experimental predictors [29]. Nevertheless, the van der Waals
equation does correctly predict two pressure-dependent inversion temperatures, whereas the
truncated virial equation z = 1 + Q1兾Vm predicts only one pressure-dependent inversion tem-
perature, corresponding to dQ1兾dT = Q1兾T; unfortunately, extension to higher order in 1兾Vm
does not appear to give better predictions in the author’s experience, presumably because the
dependence of higher coefficients upon temperature is not generally known with sufficient
accuracy.

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3
276 S. Le Vent

SUMMARY

Despite its obvious deficiencies, the van der Waals equation has many virtues, particularly in
its semi-quantitative predictions of the consequences of deviation from perfect gas behav-
iour. The present paper has aimed to summarize both well known and lesser known conse-
quences of those deviations in van der Waals terms. Naturally, there are aspects which are
not covered, e.g. special relationships which apply near the critical point [30], but the author
has tried to cover the more useful aspects.

REFERENCES
[1] ‘Johannes Diderik van der Waals’, Frontispiece of J. Chem. Educ., 32(3), Suppl. 29 (March 1955).
[2] Kondepudi, D., and Prigogine, I., Modern Thermodynamics. From Heat Engines to Dissipative
Structures, John Wiley, New York, 1988, p. 15.
[3] World Wide Web pages: http://www.library.upenn.edu/etext/smith/w/vanderWaals.html and
http://dbhs.wvusd.k12.ca.us/Gallery/Gallery12.html.
[4] van der Waals, J. D., ‘Over de continuïteit van den gas- en vloeistof-toestand’, Dissertation,
Leiden, 1873. English translation ‘The continuity of the liquid and gaseous states of matter’ in
Physical Memoirs Selected from Foreign Sources Under the Direction of the Physical Society of
London, 1 (No. 3), 333–369 (1890).
[5] Glasstone, S., Theoretical Chemistry, van Nostrand, New York, 1944, pp. 443–446.
[6] Swinbourne, E. S., ‘The van der Waals gas equation. A simple kinetic treatment’, J. Chem. Educ.,
32, 366–369(1955).
[7] Berthelot, D., ‘Determining molecular and atomic weights’, J. Phys., 8, 263–274 (1899).
[8] Redlich, O. and Kwong, J. N. S., ‘On the thermodynamics of solution. V. An equation of state.
Fugacities of gaseous solutions’, Chem. Rev., 44, 233–244, (1949).
[9] Soave, G., ‘Equilibrium constants from a modified Redlich–Kwong equation of state’, Chem. Eng.
Sci., 27, 1197–1203 (1972).
[10] Peng, D.-Y. and Robinson, D. B., ‘A new two-constant equation of state’, Ind. Eng. Chem.,
Fundam., 15, 59–64 (1976).
[11] Dieterici, C., ‘Kinetic theory of liquids’, Ann. Phys., 66, 826–858 (1898).
[12] Dieterici, C., ‘On the critical state’, Ann. Phys., 69, 685–705 (1899).
[13] Beattie, J. A. and Bridgeman, O. C., ‘A new equation of state for fluids. I. Application to gaseous
ethyl ether and carbon dioxide’, J. Amer. Chem. Soc., 49, 1665–1667 (1927).
[14] Benedict, M., Webb, G. B. and Rubin, L. C., ‘An empirical equation for thermodynamic properties
of light hydrocarbons and their mixtures. I. Methane, ethane, propane and butane’, J. Chem. Phys.,
8, 334–345 (1940).
[15] Mason, E. A. and Spurling, T. H., The Virial Equation of State (Topic 10 (The Fluid State),
Volume 2 of the International Encyclopedia of Physical Chemistry and Chemical Physics),
Pergamon, Oxford, 1969.
[16] Dymond, J. H. and Smith, E. B., The Virial Coefficients of Pure Gases and Mixtures: A Critical
Compilation, Clarendon Press, Oxford, 1980.
[17] Perry, P. H. and Green, D. W., (eds), Perry’s Chemical Engineers’ Handbook, 7th edn., McGraw-
Hill, New York, 1997, pp. 4-20 to 4-22.
[18] Smith, J. M., Van Ness, H. C. and Abbot, M. M., Introduction to Chemical Engineering
Thermodynamics, 5th edn., McGraw-Hill, New York, 1996, Chapters 3 and 6.
[19] Sandler, S. I., Chemical and Engineering Thermodynamics, 2nd edn., John Wiley, New York,
1989, Chapter 4.
[20] Guggenheim, E. A., Thermodynamics, 5th edn., North Holland, Amsterdam, 1967, p. 136.
[21] Guggenheim, E. A., Thermodynamics, 4th edn., North Holland, Amsterdam, 1959, p. 165.
[22] Hougen, O. A., Watson, K. M. and Ragatz, R. A., Chemical Process Principles, Part 2.
Thermodynamics, 2nd edn., John Wiley, New York, 1959, pp. 580, 588–591.
[23] Cardano, G. Ars Magna, 1545 (originally) and e.g., Korn, G. A. and Korn, T. M., Mathematical
Handbook for Scientists and Engineers, 2nd edn., McGraw-Hill, New York, 1968, p. 23.
[24] Prigogine, I. and Defay, R., translated Everett, D. H., Chemical Thermodynamics, Longmans,
London, 1954, p. 229–232.

International Journal of Mechanical

DownloadedEngineering Education
from ijj.sagepub.com Vol 29 No 3 STATE UNIV on May 8, 2016
at PENNSYLVANIA
 A summary of the properties of van der Waals fluids 277

[25] Dean, J. A., Lange’s Handbook of Chemistry, 15th edn., McGraw-Hill, New York, 1999, Tables
1.15, 6.2 and 6.5..
[26] Hougen, O. A., Watson, K. M. and Ragatz, R. A., Chemical Process Principles Charts, 3rd edn.,
John Wiley, New York, 1964, Fig. 140a.
[27] Porter, A. W., ‘On the inversion-points for a fluid passing through a porous plug and their use in
testing proposed equations of state’, Phil. Mag., 11, 554–568 (1906).
[28] Moelwyn-Hughes, E. A., Physical Chemistry, Pergamon, London, 1957, p. 586.
[29] Heyes, D. M. and Llaguno, C. T., ‘Computer simulation and equation of state study of the Boyle
and inversion temperature of simple fluids’, Chemical Physics, 168, 61–68 (1992).
[30] Reichl, L. E., A Modern Course in Statistical Physics, 2nd edn., Wiley InterScience, New York,
1998, p. 139–141.

International
Downloaded from ijj.sagepub.com JournalSTATE
at PENNSYLVANIA of Mechanical Engineering
UNIV on May 8, 2016 Education Vol 29 No 3