Ind. Eng. Chem. Res.

2002, 41, 953-962 953

Statistical Associating Fluid Theory: A Successful Model for the

Calculation of Thermodynamic and Phase Equilibrium Properties of
Complex Fluid Mixtures
Ioannis G. Economou†
Molecular Modelling of Materials Laboratory, Institute of Physical Chemistry, National Research Center for
Physical Sciences “Demokritos”, GR-15310 Aghia Paraskevi Attikis, Greece

Statistical associating fluid theory (SAFT) is a powerful model for thermodynamic property and
phase equilibrium calculations for fluid mixtures. In this paper, the model development,
modifications, and generalizations proposed over the past decade are reviewed. Emphasis is
given to developments resulting in equations of state applicable to real fluids. In addition,
theoretical models are reviewed. Applications are discussed, and representative calculations
are presented with emphasis on aqueous systems and polymers. Despite its wide acceptance,
SAFT has several limitations that are discussed here.

Introduction multicomponent phase equilibria at low and high pres-

sure.10,11 Significant work was performed over the past
Accurate models for thermodynamic property and decade toward the improvement of SAFT, to become
phase equilibrium predictions of pure compounds and more accurate for different types of fluid systems (such
mixtures over a wide range of temperature and pressure as, for example, polar fluids, copolymers, electrolytes,
are of extreme importance for the optimization of etc.). The purpose of this paper is to provide an overview
existing and the design of new processes and/or materi- on the SAFT developments from the original version to
als in the chemical process industry. Traditional ap- date. The various modifications are discussed in detail
proaches include empirical correlations for the calcula- in terms of their strengths and limitations when applied
tion of single-phase properties and activity coefficient to real fluids.
models or cubic equations of state (EoS) for the calcula- Significant work has been performed on the develop-
tion of phase equilibria.1 These models are usually ment of EoS for model systems (not fully engineering
accurate over a limited range of conditions and types of EoS) based on TPT. Such models are helpful in under-
systems (pure compounds and mixtures). standing specific molecular interactions that control
The development of novel processes at extreme condi- macroscopic properties of real fluids and, in many cases,
tions (such as, for example, processes where one or more are used as the basis for the development of more
of the components is supercritical) and the design of new accurate engineering EoS. These models are also re-
materials (such as, for example, well-defined copoly- viewed. Finally, representative applications are pre-
mers) over the last 2 decades imposed the need for new sented, mainly for aqueous systems and polymer mix-
models. At the same time, significant developments in tures, and model limitations for specific applications are
the area of applied statistical mechanics resulted in a discussed.
number of semiempirical EoS, such as the lattice fluid
theory (LFT),2 the perturbed hard-chain theory,3 and SAFT and Modifications
their modifications. These EoS are more complex than
cubic EoS but significantly more accurate for various In a series of four papers, Wertheim developed a
complex fluids, such as hydrogen bonding fluids, super- statistical thermodynamic theory for fluids with a
critical fluids, and polymers. Furthermore, the tremen- repulsive core and one or more highly directional short-
dous increase of computing power at affordable prices range attractive sites, known as TPT.6-9 In TPT, the
Helmholtz free energy, A, is calculated from a graphical
made these new complex models attractive for process
summation of interactions between different species.
simulation calculations.
Based on the first-order TPT (from here on referred to
A semiempirical EoS that was developed in the late as TPT-1), Chapman, Gubbins, and co-workers devel-
1980s and has gained considerable popularity in both oped an EoS for spherical and chain molecules with one
the academic and industrial communities is the statisti- or more hydrogen-bonding sites.4,12 This model was used
cal associating fluid theory (SAFT). SAFT was developed as the basis for SAFT development. In fact, there are a
by Chapman, Gubbins, Radosz, and co-workers at number of relatively small differences between the
Cornell University and Exxon Research4,5 based on the SAFT model of Chapman et al.5 and the SAFT model of
thermodynamicperturbationtheory(TPT)ofWertheim.6-9 Huang and Radosz10 as will be shown later on. In SAFT,
SAFT was parametrized for a wide range of fluids the Helmholtz free energy is written as the sum of
including organic compounds (paraffins, aromatic hy- contributions due to hard-sphere repulsive interactions,
drocarbons, alcohols, ketones, etc.), light gases, water, due to chain formation through bonding of a number of
and polymers and was shown to correlate accurately hard spheres, and due to association. For the model to
be applicable to real fluids, a dispersion term is added
† E-mail: [email protected]. Tel.: ++30 as a perturbation to the TPT-1 reference fluid. In Figure
1 6503963. Fax: ++30 1 6511766. 1, the physical basis underlying SAFT is shown sche-
10.1021/ie0102201 CCC: $22.00 © 2002 American Chemical Society
Published on Web 10/04/2001
954 Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002

vo ) voo[1 - C exp(-3uo/kT)]3 (Huang and Radosz)

(5)
where uo/k is a dispersion energy parameter per seg-
ment and C ) 0.12 (except for hydrogen which is 0.241).
m, voo, and uo/k are the three nonassociating parameters
for pure fluids which are typically fitted to pure-
component vapor pressure and liquid density data.
For the chain term, the following expression is used
based on Wertheim TPT:

achain 1 - 0.5η
) (1 - m) ln (6)
RT (1 - η)3
For the association, the Helmholtz free energy is
calculated from the expression

[ ]
Figure 1. Schematic representation of the physical basis of the
aassoc M
XA
SAFT model. The reference fluid consists of hard spheres that form
chains (here tetramers) through covalent bonding. Hydrogen RT
) ∑
A)1
ln XA -
2
+ 0.5M (7)
bonding between terminal sites at different chains results in chain
oligomers. The last step is to account for the weak dispersion
forces. where M is the number of association sites per molecule
and XA is the mole fraction of molecules not bonded at
matically. Mathematically, SAFT is expressed by the site A. In eq 7, the sum is taken over all association
following equation for the residual Helmholtz free sites of the molecule. The quantity XA is calculated from
energy per mole: the expression
M
ares(T,F) XA ) (1 + ∑ FXB∆AB)-1 (8)
RT B)1

a(T,F) aideal(T,F) aref(T,F) adisp(T,F) where ∆AB is the association strength evaluated from
) - ) +
RT RT RT RT one of the following expressions:
ahs(T,F) achain(T,F) aassoc(T,F) adisp(T,F) 1 - 0.5η
) + + + (1) ∆AB ) x2vo [exp(AB/kT) - 1]κAB
RT RT RT RT (1 - η)3

(Chapman et al.) (9)

where T and F are the temperature and density of the
system, respectively, and the residual Helmholtz free 1 - 0.5η
energy is with respect to the Helmholtz free energy for ∆AB ) x2voo [exp(AB/kT) - 1]κAB
the ideal gas at the same T and F. If the Helmholtz free (1 - η)3
energy of the fluid is known, all other thermodynamic (Huang and Radosz) (10)
properties, such as pressure, chemical potential, etc.,
can be calculated using standard thermodynamic equa- In eqs 9 and 10, the two pure-component parameters
tions.1 for association are introduced: the energy of association
For the hard-sphere term of eq 1, the Carnahan- AB/k and the volume of association κAB. Although in
Starling expression is used, so that principle these parameters can be determined from
spectroscopic data for the pure liquid, in practice they
are fitted to experimental liquid density and vapor
ahs 4η - 3η2
)m (2) pressure data, as the other three parameters.
RT (1 - η)2 For the dispersion term in eq 1, a number of different
expressions were proposed. Chapman et al.5 used the
where m is the number of spherical segments per expression originally proposed by Cotterman et al.:13
molecule and η is the reduced density evaluated from
the expression

o
adisp
RT
)m a
kT 1 (
uo disp uo disp
+ a
kT 2 ) (11)
η ) 0.74048Fmv (3)
where
where vo is the close-packed hard-core volume of the
fluid which is calculated from the temperature-inde- adisp
1 ) -11.604η - 6.132η2 - 2.871η3 + 13.885η4
pendent soft-core volume of the fluid, according to one (12)
of the following two expressions:
adisp
2 ) -2.575η + 13.463η2 - 29.992η3 + 21.470η4
o oo o o
v ) v [1 + 0.2977/(u /kT)]/[1 + 0.33163/(u /kT) + (13)
(0.0010477 + 0.025337(m - 1)/m)/(uo/kT)2] whereas Huang and Radosz used the following polyno-
mial expression originally proposed based on molecular
(Chapman et al.) (4) dynamics simulation data for the square-well fluid:14
 Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002 955

( )( )
adisp 4 9
u i η j used. The model was applied to a limited number of pure

RT
)m ∑ ∑Dij kT
i)1 j)1 0.74048
(14) compounds and mixtures successfully. However, more
than one binary parameters were adjusted for some
mixtures that weaken predictive capabilities.
with u/k ) (uo/k)(1 + e/kT) and e/k ) 10 K for all Jackson et al.4,17 proposed the so-called hard-sphere
molecules except a few small molecules (for details, see SAFT (SAFT-HS), which is a simplification compared
ref 10). to the original SAFT. The simplification consists of a
SAFT extension to mixtures is straightforward: The simple van der Waals dispersion term used so that
equation of Mansoori et al.15 is used for the hard-sphere
mixtures, whereas the chain and the association terms adisp uo
in TPT are extended to mixtures rigorously. Finally, the )m η (21)
RT kT
dispersion term is applied to mixtures assuming the van
der Waals one-fluid theory approximation. In this For mixtures, one-fluid van der Waals mixing rules are
approach, mixing rules are only needed for the disper- used. SAFT-HS was used to predict generalized phase
sion term and are as follows: diagrams for model water-n-alkane mixtures17 and also
to correlate the liquid-liquid equilibria (LLE) of water-
m) ∑i ximi (15) alkylpolyoxyethylene surfactant mixtures exhibiting
closed-loop behavior.18
Fu and Sandler19 proposed a different simplification
∑i ∑j xixjmimjvoij for the dispersion term of SAFT using an expression
proposed by Lee et al.:20
vo ) (Chapman et al.) (16)
( ∑i ximi)2 adisp
RT {
) -mZM ln 1 +
η
0.74048
exp [ ( ) ]}
u
2kT
-1 (22)

uijo where ZM ) 36. This simplified SAFT was applied to a

uo
∑i ∑j xixjmimj voij
k
number of pure components (hydrocarbons, polar or-
o ganic compounds, carboxylic acids, and water) and used
v ) to correlate experimental vapor-liquid equilibrium
∑i ximi)2
k (VLE) data of binary mixtures. Simplification of the
(
dispersion term in SAFT provides a minor improvement
because the equation is of higher order anyway, and for
uij associating fluids (either pure component or mixtures),

u
∑i ∑j xixjmimj k voij the most computationally intensive term is, by far, the
association term. Furthermore, the simplified SAFT did
) (Huang and Radosz) (17) not provide considerable improvement in accuracy
∑i ∑j
k
xixjmimjvoij compared to the original SAFT.
An empirical modification of SAFT consists of replac-
ing the hard-sphere, chain, and dispersion terms in eq
1 with a cubic EoS and retaining only the association
[21((v )
3
voij ) o 1/3
ii + (vojj)1/3) ] (18) term from TPT-1. The resulting EoS is, of course,
noncubic. Despite its simplicity, such an approach

( )
provides accurate correlation of VLE and LLE of highly
uoij uoii uojj 1/2
) (1 - kij) (Chapman et al.) (19) nonideal fluid mixtures.21-23
k k k Müller and Gubbins24 developed a modified SAFT for

( )
pure water that accounts explicitly for repulsive and
uij uii ujj 1/2
dispersion interactions using an accurate Lennard-
) (1 - kij) (Huang and Radosz) (20) Jones equation,25 for dipole-dipole interactions using
k k k
a Padé approximant for the perturbation expansion, and
where xi is the mole fraction of component i and kij is a for association using the TPT expression. This model
binary adjustable parameter. Equations 15 and 18 are was further applied to pure alkanes and alcohols and
used both by Chapman et al.5 and Huang and Radosz.11 to binary VLE mixtures of water, alkanes, and alcohols
The elegance and the strong statistical mechanics and will be referred to as SAFT-LJ.26,27 To improve
basis of TPT and the accuracy of SAFT for a number of accuracy, two binary parameters were used for the
different systems made SAFT a very popular model in dispersion energy and hard-core volume of unlike spe-
the academic and industrial communities. Although cies. The introduction of the second parameter does not
very similar, the SAFT version of Huang and Radosz10 seem to provide considerable improvement over the
has gained considerably more popularity than that of original SAFT with a single binary parameter.
Chapman et al.5 In the rest of the paper, when we refer Banaszak et al.28 proposed a SAFT model that ac-
to the original SAFT, we imply the Huang and Radosz counts explicitly for heteronuclear chains and is referred
model, although this is not fully correct. to as copolymer SAFT. Chain heterogeneity is accounted
Numerous modifications and improvements of the explicitly by the chain term of the equation, whereas
theory were proposed and are presented here briefly. the hard-core and dispersion terms are the same as
Walsh et al.16 modified the dispersion term of SAFT in those in the original SAFT.
order to account explicitly for multipolar interactions, Blas and Vega29 developed a SAFT model (referred
due to permanent dipole and quadrupole moments, to as soft-SAFT) with a Lennard-Jones reference fluid30
respectively. A Padé approximant using the second- and instead of the hard-sphere fluid in the original SAFT
third-order term in the perturbation expansion was and chain and association terms similar to the original
956 Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002

SAFT. Soft-SAFT is similar to the SAFT-LJ without the MacDowell and Vega41 developed an EoS for pure
multipolar term. Soft-SAFT was applied successfully to fluids and mixtures of alkanes from TPT-1 that accounts
pure n-alkanes, 1-alkenes, and 1-alcohols and to binary explicitly for the intramolecular interactions in chain
and ternary n-alkane mixtures, including the critical fluids. In this way, the model accounts explicitly for the
region.29 In the case of mixtures, two binary parameters different conformations assumed by the chain molecules
were used, even for n-alkane mixtures. Here again, the and calculations become considerably more computa-
introduction of the second parameter is not really tionally demanding. Qualitative agreement was ob-
justified (and, in fact, is very close to unity). Even with tained with experimental data for the pure alkane
no binary parameters, accurate predictions are obtained critical constants without any macroscopic parameter
for the mixtures examined. fitting. To obtain quantitative agreement with experi-
Chen et al.31 proposed a SAFT-LJ similar to soft- mental data for mixture excess properties, macroscopic
SAFT with the exception that the radial distribution parameter adjustment was needed.42 In its present
function for the reference spherical fluid was reparam- form, the model cannot be used for engineering calcula-
etrized. Unlike the original SAFT and other SAFT tions, because of its relatively high computing require-
versions where it has a constant value, the spherical ments.
segment parameter for pure n-alkanes increases as a Recently, Gross and Sadowski43 presented a perturbed-
function of the molecular weight, and this is a limitation chain SAFT (PC-SAFT) with a hard-chain reference
for the extension of the model to polymers. fluid by applying the Barker-Henderson perturbation
theory truncated at the second-order term. The integral
Jackson and co-workers32 improved SAFT by replac- expressions for the first- and second-order terms were
ing the hard-sphere reference fluid with a more realistic refitted to appropriate Taylor series expansions in
intermolecular potential model such as the square-well, density using pure n-alkane data. In its current formu-
Sutherland, or Yukawa potentials. In all cases the lation, PC-SAFT is applicable to nonassociating com-
attractive part of the potential is of variable range (an ponents only. Pure-component parameters were re-
extra pure component parameter λ is introduced), and gressed for a large number of components. PC-SAFT is
so this model is referred to as SAFT-VR. The chain and more accurate than the original SAFT for pure compo-
association terms of the equation are modified to ac- nents and binary mixtures. Preliminary calculations
count for the structure of the new reference fluid. SAFT- presented for polyethylene mixtures are also encourag-
VR with the variable square-well fluid correlated accu- ing.
rately the pure n-alkane and n-perfluoroalkane phase
equilibrium over a wide temperature range. Recently, TPT-Based EoS for Model Systems
Galindo et al.33 extended SAFT-VR to aqueous electro-
lyte systems. Long-range Coulombic ion-ion interac- TPT refers to the original thermodynamic model of
tions are calculated with the restricted primitive model Wertheim that was used as the basis for the develop-
using the mean-spherical approximation. With one ment of numerous theoretical EoS for model fluids. In
transferable fitted parameter per ion, the vapor pres- other words, the distinction between TPT and SAFT
sure and liquid density for a number of single-salt here is that the former refers to an equation for model
aqueous solutions and one mixed-salt system of strong fluids whereas the latter refers to an equation for real
electrolytes were correlated. SAFT-VR was extended fluids. Despite this distinction, most of the TPT-based
recently to heteronuclear chain molecules but has not models can be easily extended to real fluids, as will be
been applied to real fluids yet.34 made clear in the following paragraphs.
Adidharma and Radosz35 proposed a modified SAFT Walsh and Gubbins44 modified TPT-1 to chains made
(referred to as SAFT1) that resembles SAFT-VR with of fused hard spheres by incorporating information from
the square-well reference fluid. In SAFT1, unlike SAFT- the scaled particle theory. Different molecular geom-
VR, a truncation term is added to the second-order etries were considered, and theoretical predictions were
perturbation expansion for the square-well fluid. Fur- in good agreement with molecular simulation results.
thermore, SAFT1 accounts explicitly for heteronuclear Similarly, Jackson and co-workers45-47 proposed a TPT-
chain molecules. SAFT1 is more accurate than the 1-based theory for heteronuclear hard chains made of
original SAFT for n-alkanes and n-alkane-polyolefin tangent or fused spheres. The model was also general-
mixtures. A number of different analytical approxima- ized to mixtures. In this spirit, Shukla and Chapman48
tions to the square-well fluid implemented in SAFT1 proposed a theory for fluid mixtures of heteronuclear
hard chains.
were tested.36
Banaszak et al.49 extended TPT-1 to chains of sticky
Pfohl and Brunner37 modified SAFT in order to spheres and to chains of square-well spheres. Further-
account for the convexity of nonspherical molecules more, Banaszak et al.50 and Johnson et al.51 used TPT-1
based on the BACK EoS.38 An extra pure component for the development of an EoS for chains of tangent
parameter was introduced to describe the convex body. freely jointed Lennard-Jones spheres and mixtures of
The SAFT-BACK EoS improves prediction in the vicin- Lennard-Jones spheres.52 Good agreement was obtained
ity of the pure-component critical point, and so it can with simulation results for the compressibility factor,
be used for supercritical fluid mixture calculations. second virial coefficient, and internal energy of Lennard-
An improved SAFT for the critical region of fluids was Jones chains of variable length. The Johnson et al.
proposed by Kiselev and Ely39,40 by incorporating cross- model was extended to associating Lennard-Jones
over functions into the EoS so that critical exponents chains.53
are correctly reproduced. The simplified crossover SAFT Ghonasgi and Chapman54-56 developed a TPT-1 theory
provides an accurate description of the pure fluid and for hard spheres, hard chains, Lennard-Jones spheres,
binary mixture phase equilibrium including the critical and dispheres with a variable number of associating
point vicinity without any additional pure-component sites ranging from one to four. Furthermore, Ghonasgi
parameters compared to the original SAFT. et al.57 extended this model to hard chains that associate
 Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002 957

intramolecularly. In this way, hard rings can be formed.

Similarly, Sear and Jackson58 and Ghonasgi and Chap-
man59 extended the theory to systems where both inter-
and intramolecular associations occur.
In TPT-1 the hard sphere is used as the reference
fluid and chain molecules are formed through bonding
of hard spheres. A more accurate approach is to incor-
porate information for the hard disphere into the
reference fluid. In this way, a hard monomer-dimer
reference fluid is used. Chang and Sandler,60 Ghonasgi
and Chapman,61 and Sadus62 independently developed
the TPT-D EoS that results in more accurate predictions
for the compressibility factor of hard-sphere chains
compared to TPT-1. An improved TPT-D was proposed
recently for tangent and fused hard-sphere chain flu-
ids.63 Sadus64 extended TPT-D to mixtures of hard-
sphere chains. Tavares et al.65 extended the TPT-D EoS Figure 2. Pure water coexisting densities (b), superheated
densities at 10 MPa (×), and supercritical densities at 35 (O), 50
to square-well chain fluids by incorporating structural (0), 75 (4), and 100 MPa (]). Experimental data (points)114 and
information for the square-well dimer into the reference SAFT predictions (lines).
fluid. Furthermore, Shukla and Chapman66 extended
TPT-D EoS to heteronuclear hard-sphere chain fluids.
They showed that TPT-D is considerably more accurate
than TPT-1 and slightly more accurate than second-
order TPT (TPT-2) for pure heteronuclear hard-chain
fluids.
Jog and Chapman67 extended TPT-D to dipolar hard-
sphere chain fluids. The model fluid consists of hard-
sphere chains with a dipole on every other spherical
segment. Polar interactions are calculated through a
third-order Padé approximant to the perturbation ex-
pansion. The agreement of the polar TPT-D with
simulation results is good, and so this model can form
the basis for a SAFT EoS for real polar pure fluids and
mixtures.
Escobedo and de Pablo68 and Zhou et al.69 realized
that the compressibility factor of tangent hard chains
varies linearly with the chain length at constant reduced
density above a minimum chain length. As a result, a Figure 3. Water-n-heptane density at 673 K. Experimental data
(points)115 and SAFT predictions (lines).
two-reference fluid approach can be used from a linear
combination of the compressibility factors of two refer-
ence chains. Zhou et al.69 showed that a 4-mer-8-mer oxygen. Huang and Radosz10 assigned three associating
reference fluid results in an accurate equation for hard sites per molecule, whereas in subsequent work, Gub-
chains, while Escobedo and de Pablo68 further extended bins and co-workers16,24 used a four-site model for water.
this approach by developing a 2-mer-∞-mer reference Economou and Tsonopoulos72 performed an extensive
fluid EoS, applicable also for mixtures. Escobedo and comparison of three-and four-site models for water in
de Pablo70 generalized their EoS for square-well chains order to calculate water-hydrocarbon LLE and con-
and Lennard-Jones chains. Recently, Shukla and Chap- cluded that the four-site model is relatively more
man71 applied the two-reference fluid TPT-1 to hard accurate.
heteronuclear chain molecules. The advantage of using a model with a strong
statistical mechanics basis is that it can be extrapolated
Applications for Complex Fluids accurately to conditions away from the region where the
model parameters are regressed. In Figure 2, experi-
Despite the modifications and improvements of SAFT mental data and SAFT predictions are shown for the
over the past several years, the original SAFT of Huang coexisting density, superheated density at 10 MPa, and
and Radosz10,11 remains the most popular version for supercritical density at 35, 50, 75, and 100 MPa of
pure and mixture fluid thermodynamic and phase water. The pure water parameters were fitted to the
equilibrium calculations in academia and in industry. saturated liquid density and vapor pressure data in the
Representative results are presented here for two temperature range 278-641 K.23 Although it overpre-
classes of fluid mixtures where SAFT is expected to have dicts the pure water critical point because of mean-field
a competitive advantage over traditional EoS: aqueous limitations of the model, SAFT predicts accurately the
systems and polymers. In all cases, the original SAFT superheated and supercritical density.
of Huang and Radosz10,11 is used. As a result, SAFT is a suitable model for supercritical
Aqueous Systems. SAFT accounts explicitly for fluid applications. In such applications, a model is
hydrogen-bonding interactions (eq 7) that affect ther- expected to provide accurate estimates of the thermo-
modynamic properties and phase equilibrium consider- dynamic properties as a function of temperature, pres-
ably. In SAFT, water is modeled as a sphere with a sure, and composition. In Figure 3, experimental data
number of associating sites that correspond to the and SAFT predictions are shown for the water-n-
hydrogen atoms and the lone pairs of electrons in the heptane mixture density at 673 K at different pressure
958 Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002

Table 1. Comparison of Polymer PVT Data Correlation

with Various EoSa
EoS % rms deviation in volume average ∆v × 104 (mL/g)
SAFT 0.13 10.8
FOV 19.4
LFT 28.8
MCM 5.5
PHSC 0.14
a Calculations for FOV, LFT, and MCM are from Rodgers,80 and
those for PHSC are from Song et al.81

and composition values. With the exception of the

density at 25 MPa and 0.9 water mole fraction which is
very close to the pure water critical point, model
predictions are very accurate.
Accurate calculation of water-hydrocarbon phase
equilibrium is of extreme importance for many processes Figure 4. Propylene-PEP phase equilibria for different PEP
in the refining and petrochemical industry and for molecular weights at constant composition (15 wt %). Experimen-
environmental control. At the same time, it is a very tal data (points)83 and SAFT correlation (lines). SAFT predicts
challenging scientific problem because of the difference hourglass behavior for molecular weight values above approxi-
mately 12 000. For lower molecular weights, SAFT predicts both
in the nature of intermolecular forces exhibited between LCST (in excellent agreement with experimental data) and UCST
water molecules and hydrocarbon molecules. SAFT and (shown here for a molecular weight of 5900).
SAFT-LJ were shown to be successful in the correlation
of high-pressure water-methane and water-ethane conceptually. It consists of a hard-sphere chain term and
phase equilibria,11,23,73 and SAFT-HS predicted ac- a perturbation expansion for the dispersion interactions.
curately the water-n-alkane critical lines.74 On the Although the actual mathematical expressions are
other hand, the model is less accurate for the calculation different in the two models, they provide similar results
of the low hydrocarbon solubility in water at low when applied to real chain fluids.
pressure.72 This inaccuracy may be attributed to the Pure polymer parameters for LFT, FOV, and other
failure of the model to account correctly for the change polymer EoS fitted to melt PVT data are used for
in the water structure in the vicinity of hydrocarbon polymer mixture calculations with these models. In the
molecules (which affects hydrocarbon solubility consid- case of SAFT, polymer parameters regressed from PVT
erably) and for the intermolecular interactions between data do not provide a good description of mixture phase
water molecules and hydrocarbon molecules. In other equilibria, even with the use of large binary interaction
words, better mixing rules are needed for these systems. parameters. This shortcoming is attributed to the voo
Using an asymmetric mixing rule for the dispersion and uo/k values obtained from the PVT data. An attempt
term, limited improvement was obtained.72 Interest- to include isothermal compressibility data in the fit of
ingly, if one replaces the hard-sphere, chain, and disper- polymer parameters did not result in a significant
sion terms in SAFT with a much simpler cubic EoS, improvement.82
then an improved correlation is obtained for the solubil- On the other hand, SAFT pure-component parameters
ity of several hydrocarbons in water.23 for small molecules in a homologous series vary smoothly
Polymer Systems. SAFT accounts for chain con- with the molecular weight.10 As a result, they can be
nectivity rigorously, and so it is accurate for long-chain extrapolated to high molecular weight values and used
molecules such as polymers. Most of the SAFT applica- for mixtures containing heavy oil fractions, polymers,
tions to real fluid systems appearing in the literature etc., and so the model becomes significantly more
are related to high-pressure polymer phase equilibria general. This behavior was first realized by Economou
where conventional thermodynamic models fail, al- and co-workers83 for high-pressure binary and ternary
though significant work has been done also for low- mixtures of alternating poly(ethylene-propylene) (PEP)
pressure VLE. Most of the high-pressure work can be in different R-olefin solvents. A typical example is shown
found in the review papers of Folie and Radosz75 and of in Figure 4 where the pressure-temperature phase
Kirby and McHugh.76 diagram is presented for a 15 wt % monodisperse PEP
Pure polymer EoS parameters are usually calculated dissolved in propylene. Experimental data were mea-
by fitting melt PVT data over a wide temperature and sured for four different PEP molecular weight values.
pressure range (typically 0-200 MPa). SAFT param- Using a single PEP molecular weight dependent but
eters were calculated for 25 different polymers including temperature-independent binary interaction parameter,
several linear and branched polyolefins, polyacrylates, SAFT correlates the lower critical solution temperature
polystyrenes, polysiloxanes, etc., by fitting literature (LCST) phase behavior for the low molecular weight
PVT data.77 SAFT correlation of these data was com- mixtures and the hourglass behavior for the high
pared with correlations obtained from other polymer molecular weight mixtures. Furthermore, SAFT is able
EoS such as the Flory-Orwoll-Vrij (FOV) EoS,78 the to predict the phase boundaries for the intermediate
LFT,2 the modified cell model (MCM),79 and the per- molecular weights and also a theoretical upper critical
turbed hard-sphere chain EoS (PHSC).81 This compari- solution temperature (UCST) curve for the low molec-
son is summarized in Table 1. The relatively simple ular weight mixtures, exhibited at low temperatures.
MCM is the most accurate model in this case. Interest- SAFT correlation for this system is by far more accurate
ingly, LFT, used widely for polymer mixtures phase than a correlation with the FOV EoS.84 In addition,
equilibria, is the least accurate model. SAFT and PHSC SAFT was shown to be more accurate than LFT for
correlations are very similar overall and on a polymer polyethylene-n-alkane high-pressure phase equilibria.85
by polymer basis. PHSC is very similar to SAFT On the basis of the work of Economou and co-workers,83
 Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002 959

calculations of polymer mixtures and, to a lesser extent,

of aqueous systems. For supercritical fluid applications,
the simplified crossover SAFT seems quite promis-
ing.39,40 Some of the theoretical improvements proposed
for TPT-1 and reviewed here also deserve further
exploitation for real fluids. The two-reference fluid
approach68-70 as well as the models that account
explicitly for polar interactions are quite promising.67
Despite the current popularity of SAFT models,
significantly more work is needed in order that these
models become standard tools for process simulation.
The limitation of the model for prediction of hydrocarbon
solubility in water should be primarily attributed to the
mixing rules. In their current formulation, mixing rules
do not account for the change in water microscopic
structure initiated by the hydrocarbon molecule inser-
tion. If a model accounts properly for these important
Figure 5. Weight fraction Henry’s constant of ethylene ((), molecular processes, then model predictions are in good
n-butane (9), n-hexane (2), n-octane (4), benzene (×), and toluene agreement with experimental data, as was shown for
(O) in low-density polyethylene at 1 atm. Experimental data
(points)116 and SAFT correlation (lines) using a single temperature-
Monte Carlo simulation using appropriate molecular
independent binary interaction parameter. models.73,112
In terms of polymer applications, a severe limitation,
SAFT was further applied to a number of other high- especially for the case of polar polymers, is the method
pressure polymer mixtures.85-105 currently used to estimate polymer model parameters.
At the same time, SAFT was applied successfully to A possible explanation for the failure of polymer pa-
low-pressure polymer phase equilibria. In Figure 5, the rameters obtained from PVT data regression when
weight fraction Henry’s law constant for ethylene, applied to polymer mixtures is that SAFT does not
n-butane, n-hexane, n-octane, benzene, and toluene in account fully for the intramolecular structure which
low-density polyethylene at 1 atm and over a wide becomes important as the molecular size increases.
temperature range is shown. SAFT with a temperature- Although the extrapolation of n-alkane parameters to
independent binary interaction parameter provides high molecular weight values works well for polyolefins,
accurate correlation of the experimental data. it is certain that it is impractical for other types of
In all such applications, polymer parameters were polymers. Simultaneous regression of other pure poly-
estimated using generalized correlations based on the mer thermodynamic data in addition to PVT data is an
polymer molecular weight only. As a result, the effect option that has not been fully tested yet. Alternatively,
of chemical structure, chain architecture, branching, polymer-solvent mixture phase equilibrium data can
and morphology of the polymer on the phase behavior be incorporated in the regression so that the pure
of a mixture is calculated implicitly through the binary polymer and the binary interaction parameters are
interaction parameter(s). This is certainly a limitation calculated simultaneously.113 Sadowski and co-work-
of the model that has not been resolved yet. To improve ers113 claim that the so-determined polymer parameters
SAFT agreement with experimental data, especially in for a polymer-solvent mixture work well for other
the case of polar polymer or copolymer mixtures, a solvent mixtures and can be regarded as characteristic
second binary parameter is introduced in the mixing for the given polymer.
rule for m (eq 15) or voij (eq 18).76,90 However, this
empirical approach has limited applicability, and the Literature Cited
resulting binary parameters cannot be generalized.
(1) Prausnitz, J. M.; Lichtenthaler, R. N.; Gomes de Azevedo,
A more elegant and accurate approach is to use the E. Molecular Thermodynamics of Fluid Phase Equilibria, 3rd ed.;
copolymer SAFT,28 especially for modeling copolymer Prentice-Hall: Englewood Cliffs, NJ, 1999.
and/or branched polymer solutions. This approach was (2) Sanchez, I. C.; Lacombe, R. H. An Elementary Molecular
used successfully for polyethylene and other polyolefin Theory of Classical Fluids. Pure Fluids. J. Phys. Chem. 1976, 80,
copolymers in different n-alkane and other organic 2352.
solvents including supercritical solvents by Radosz and (3) Donohue, M. D.; Prausnitz, J. M. Perturbed Hard Chain
co-workers.106-110 Alternatively, SAFT1 with an ad- Theory for Fluid Mixtures: Thermodynamic Properties for Mix-
ditional pure-component parameter for the range of tures in Natural Gas and Petroleum Technology. AIChE J. 1978,
square-well potential was used to correlate polyolefin 24, 849.
copolymer-light solvent phase equilibria.102-104,111 (4) Jackson, G.; Chapman, W. G.; Gubbins, K. E. Phase
Equilibria of Associating Fluids. Spherical Molecules with Multiple
Bonding Sites. Mol. Phys. 1988, 65, 1.
Conclusions (5) Chapman, W. G.; Gubbins, K. E.; Jackson, G.; Radosz, M.
New Reference Equation of State for Associating Liquids. Ind. Eng.
The pioneering work of Wertheim in the mid-1980s Chem. Res. 1990, 29, 1709.
resulted in a family of engineering EoS developed in the (6) Wertheim, M. S. Fluids with Highly Directional Attractive
1990s that includes SAFT,4,5,10 polar-SAFT,16 SAFT- Forces. I. Statistical Thermodynamics. J. Stat. Phys. 1984, 35, 19.
HS,17 simplified SAFT,19 SAFT-LJ,24,26 copolymer SAFT,28 (7) Wertheim, M. S. Fluids with Highly Directional Attractive
soft-SAFT,29 SAFT-VR,32 SAFT1,35 SAFT-BACK,37 cross- Forces. II. Thermodynamic Perturbation Theory and Integral
over SAFT,39,40 and PC-SAFT.43 SAFT-based models Equations. J. Stat. Phys. 1984, 35, 35.
have received an impressive acceptance in academia and (8) Wertheim, M. S. Fluids with Highly Directional Attractive
in industry as the leading models for phase equilibrium Forces. III. Multiple Attraction Sites. J. Stat. Phys. 1986, 42, 459.
960 Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002

(9) Wertheim, M. S. Fluids with Highly Directional Attractive for Chain Molecules with Attractive Potentials of Variable Range.
Forces. IV. Equilibrium Polymerization. J. Stat. Phys. 1986, 42, J. Chem. Phys. 1997, 106, 4168.
477. (33) Galindo, A.; Gil-Villegas, A.; Jackson, G.; Burgess, A. N.
(10) Huang, S. H.; Radosz, M. Equation of State for Small, SAFT-VRE: Phase Behavior of Electrolyte Solutions with the
Large, Polydisperse, and Associating Molecules. Ind. Eng. Chem. Statistical Associating Fluid Theory for Potentials of Variable
Res. 1990, 29, 2284. Range. J. Phys. Chem. B 1999, 103, 10272.
(11) Huang, S. H.; Radosz, M. Equation of State for Small, (34) McCabe, C.; Gil-Villegas, A.; Jackson, G.; Del Rı́o, F. The
Large, Polydisperse, and Associating Molecules: Extension to Thermodynamics of Heteronuclear Molecules Formed from Bonded
Fluid Mixtures. Ind. Eng. Chem. Res. 1991, 30, 1994. Square-Well (BSW) Segments Using the SAFT-VR Approach. Mol.
(12) Chapman, W. G.; Jackson, G.; Gubbins, K. E. Phase Phys. 1999, 97, 551.
Equilibria of Associating Fluids. Chain Molecules with Multiple (35) Adidharma, H.; Radosz, M. Prototype of an Engineering
Bonding Sites. Mol. Phys. 1988, 65, 1057. Equation of State for Heterosegmented Polymers. Ind. Eng. Chem.
(13) Cotterman, R. L.; Schwarz, B. J.; Prausnitz, J. M. Molec- Res. 1998, 37, 4453.
ular Thermodynamics for Fluids at Low and High Densities. Part (36) Adidharma, H.; Radosz, M. A Study of Square-Well
I: Pure Fluids Containing Small or Large Molecules. AIChE J. Statistical Associating Fluid Theory Approximations. Fluid Phase
1986, 32, 1787. Equilib. 1999, 161, 1.
(14) Alder, B. J.; Young, D. A.; Mark, M. A. Studies in Molecular (37) Pfohl, O.; Brunner, G. 2. Use of BACK to Modify SAFT in
Dynamics. X. Corrections to the Augmented van der Waals Theory Order to Enable Density and Phase Equilibrium Calculations
for the Square Well Fluid. J. Chem. Phys. 1972, 56, 3013. Connected to Gas-Extraction Processes. Ind. Eng. Chem. Res.
(15) Mansoori, G. A.; Carnahan, N. F.; Starling, K. E.; Leland, 1998, 37, 2966.
T. W. Equilibrium Thermodynamic Properties of the Mixture of (38) Boublik, T. Hard Convex Body Equation of State. J. Chem.
Hard Spheres. J. Chem. Phys. 1971, 54, 1523. Phys. 1975, 63, 4084.
(16) Walsh, J. M.; Guedes, H. J. R.; Gubbins, K. E. Physical (39) Kiselev, S. B.; Ely, J. F. Crossover SAFT Equation of
Theory for Fluids of Small Associating Molecules. J. Phys. Chem. State: Application to Normal Alkanes. Ind. Eng. Chem. Res. 1999,
1992, 96, 10995. 38, 4993.
(17) Green, D. G.; Jackson, G. Theory of Phase Equilibria for (40) Kiselev, S. B.; Ely, J. F. Simplified Crossover SAFT
Model Aqueous Solutions of Chain Molecules: Water + Alkane Equation of State for Pure Fluids and Fluid Mixtures. Fluid Phase
Mixtures. J. Chem. Soc., Faraday Trans. 1992, 88, 1395. Equilib. 2000, 174, 93.
(18) Garcia-Lisbona, M. N.; Galindo, A.; Jackson, G.; Burgess, (41) MacDowell, L. G.; Vega, C. Vapor-Liquid Equilibria of
A. N. An Examination of the Cloud Curves of Liquid-Liquid Linear and Branched Alkanes from Perturbation Theory. J. Chem.
Immiscibility in Aqueous Solutions of Alkyl Polyoxyethylene Phys. 1998, 109, 5681.
Surfactants Using the SAFT-HS Approach with Transferable (42) Vega, C.; MacDowell, L. G.; López-Rodrı́guez, A. Excess
Parameters. J. Am. Chem. Soc. 1998, 120, 4191. Properties of Mixtures of n-Alkanes from Perturbation Theory. J.
(19) Fu, Y.-H.; Sandler, S. I. A Simplified SAFT Equation of Chem. Phys. 1999, 111, 3192.
State for Associating Compounds and Mixtures. Ind. Eng. Chem. (43) Gross, J.; Sadowski, G. Perturbed-Chain SAFT: An Equa-
Res. 1995, 34, 1897. tion of State Based on a Perturbation Theory for Chain Molecules.
(20) Lee, K. H.; Lombardo, M.; Sandler, S. I. The Generalized Ind. Eng. Chem. Res. 2001, 40, 1244.
van der Waals Partition Function. II. Application to the Square- (44) Walsh, J. M.; Gubbins, K. E. A Modified Thermodynamic
Well Fluid. Fluid Phase Equilib. 1985, 21, 177. Perturbation Theory Equation for Molecules with Fused Hard
(21) Suresh, S. J.; Elliott, J. R., Jr. Multiphase Equilibrium Sphere Cores. J. Phys. Chem. 1990, 94, 5115.
Analysis via a Generalized Equation of State for Associating (45) Archer, A. L.; Jackson, G. Theory and Computer Simula-
Mixtures. Ind. Eng. Chem. Res. 1992, 31, 2783. tions of Heteronuclear Diatomic Hard-Sphere Molecules (Hard
(22) Kontogeorgis, G. M.; Voutsas, E. C.; Yakoumis, I. V.; Dumbbells). Mol. Phys. 1991, 73, 881.
Tassios, D. P. An Equation of State for Associating Fluids. Ind. (46) Amos, M. D.; Jackson, G. BHS Theory and Computer
Eng. Chem. Res. 1996, 35, 4310. Simulations of Linear Heteronuclear Triatomic Hard-Sphere
(23) Voutsas, E. C.; Boulougouris, G. C.; Economou, I. G.; Molecules. Mol. Phys. 1991, 74, 191.
Tassios, D. P. Water/Hydrocarbon Phase Equilibria Using the (47) Amos, M. D.; Jackson, G. Bonded Hard-Sphere (BHS)
Thermodynamic Perturbation Theory. Ind. Eng. Chem. Res. 2000, Theory for the Equation of State of Fused Hard-Sphere Polyatomic
39, 797. Molecules and Their Mixtures. J. Chem. Phys. 1992, 96, 4604.
(24) Müller, E. A.; Gubbins, K. E. An Equation of State for (48) Shukla, K. P.; Chapman, W. G. SAFT Equation of State
Water from a Simplified Intermolecular Potential. Ind. Eng. Chem. for Fluid Mixtures of Hard Chain Copolymers. Mol. Phys. 1997,
Res. 1995, 34, 3662. 91, 1075.
(25) Kolafa, J.; Nezbeda, I. The Lennard-Jones FluidsAn (49) Banaszak, M.; Chiew, Y. C.; Radosz, M. Thermodynamic
Accurate Analytic and Theoretically-Based Equation of State. Perturbation Theory: Sticky Chains and Square-Well Chains.
Fluid Phase Equilib. 1994, 100, 1. Phys. Rev. E 1993, 48, 3760.
(26) Kraska, T.; Gubbins, K. E. Phase Equilibria Calculations (50) Banaszak, M.; Chiew, Y. C.; O’Lenick, R.; Radosz, M.
with a Modified SAFT Equation of State. 1. Pure Alkanes, Thermodynamic Perturbation Theory: Lennard-Jones Chains. J.
Alkanols, and Water. Ind. Eng. Chem. Res. 1996, 35, 4727. Chem. Phys. 1994, 100, 3803.
(27) Kraska, T.; Gubbins, K. E. Phase Equilibria Calculations (51) Johnson, J. K.; Müller, E. A.; Gubbins, K. E. Equation of
with a Modified SAFT Equation of State. 2. Binary Mixtures of State for Lennard-Jones Chains. J. Phys. Chem. 1994, 98, 6413.
n-Alkanes, 1-Alkanols, and Water. Ind. Eng. Chem. Res. 1996, 35, (52) Banaszak, M.; Chiew, Y. C.; Radosz, M. Mixing Rules for
4738. Binary Lennard-Jones Fluid Structures. Fluid Phase Equilib.
(28) Banaszak, M.; Chen, C. K.; Radosz, M. Copolymer SAFT 1995, 111, 161.
Equation of State. Thermodynamic Perturbation Theory Extended (53) Müller, E. A.; Vega, L. F.; Gubbins, K. E. Theory and
to Heterobonded Chains. Macromolecules 1996, 29, 6481. Simulation of Associating Fluids: Lennard-Jones Chains with
(29) Blas, F. J.; Vega, L. F. Prediction of Binary and Ternary Association Sites. Mol. Phys. 1994, 83, 1209.
Diagrams Using the Statistical Associating Fluid Theory (SAFT) (54) Ghonasgi, D.; Chapman, W. G. Theory and Simulation of
Equation of State. Ind. Eng. Chem. Res. 1998, 37, 660. Pàmies, J. Associating Fluids with Four Bonding Sites. Mol. Phys. 1993, 79,
C.; Vega, L. F. Vapor-Liquid Equilibria and Critical Behavior of 291.
Heavy n-Alkanes Using Transferable Parameters from the Soft- (55) Ghonasgi, D.; Chapman, W. G. Theory and Simulation of
SAFT Equation of State. Ind. Eng. Chem. Res. 2001, 40, 2532. Associating Chain Fluids. Mol. Phys. 1993, 80, 161.
(30) Johnson, J. K.; Zollweg, J. A.; Gubbins, K. E. The Lennard- (56) Ghonasgi, D.; Chapman, W. G. Theory and Simulation for
Jones Equation of State Revisited. Mol. Phys. 1993, 78, 591. Associating Hard Chain Fluids. Mol. Phys. 1994, 83, 145.
(31) Chen, C.-k.; Banaszak, M.; Radosz, M. Statistical Associat- (57) Ghonasgi, D.; Perez, V.; Chapman, W. G. Intramolecular
ing Fluid Theory Equation of State with Lennard-Jones Reference Association in Flexible Hard Chain Molecules. J. Chem. Phys.
Applied to Pure and Binary n-Alkane Systems. J. Phys. Chem. B 1994, 101, 6880.
1998, 102, 2427. (58) Sear, R. P.; Jackson, G. Thermodynamic Perturbation
(32) Gil-Villegas, A.; Galindo, A.; Whitehead, P. J.; Mills, S. J.; Theory for Association into Chains and Rings. Phys. Rev. E 1994,
Jackson, G.; Burgess, A. N. Statistical Associating Fluid Theory 50, 386.
 Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002 961

(59) Ghonasgi, D.; Chapman, W. G. Competition between (86) Gregg, C. J.; Stein, F. P.; Radosz, M. Phase Behavior of
Intermolecular and Intramolecular Association in Flexible Hard Telechelic Polyisobutylene (PIB) in Subcritical and Supercritical
Chain Molecules. J. Chem. Phys. 1995, 102, 2585. Fluids. 1. Inter- and Intra- Association Effects for Blank, Mono-
(60) Chang, J.; Sandler, S. I. An Equation of State for the Hard- hydroxy, and Dihydroxy PIB (1K) in Ethane, Propane, Dimethyl
Sphere Chain Fluid: Theory and Monte Carlo Simulation. Chem. Ether, Carbon Dioxide, and Chlorodifluoromethane. Macromol-
Eng. Sci. 1994, 49, 2777. ecules 1994, 27, 4972.
(61) Ghonasgi, D.; Chapman, W. G. A New Equation of State (87) Gregg, C. J.; Stein, F. P.; Radosz, M. Phase Behavior of
for Hard Chain Molecules. J. Chem. Phys. 1994, 100, 6633. Telechelic Polyisobutylene (PIB) in Subcritical and Supercritical
(62) Sadus, R. J. Equations of State for Hard-Sphere Chains. Fluids. 2. PIB Size, Solvent Polarity, and Inter- and Intra-
J. Phys. Chem. 1995, 99, 12363. Association Effects for Blank, Monohydroxy, and Dihydroxy PIB
(63) Kushwaha, K. B.; Khanna, K. N. SAFT-D Theory for Hard (11K) in Ethane, Propane, Carbon Dioxide, and Dimethyl Ether.
Sphere and Hard Disc Chain Fluids. Mol. Phys. 1999, 97, 907. Macromolecules 1994, 27, 4981.
(64) Sadus, R. J. A Simplified Thermodynamic Perturbation (88) Gregg, C. J.; Stein, F. P.; Radosz, M. Phase Behavior of
TheorysDimer Equation of State for Mixtures of Hard-Sphere Telechelic Polyisobutylene (PIB) in Subcritical and Supercritical
Chains. Macromolecules 1996, 29, 7212. Fluids. 3. Three-Arm-Star PIB (4K) as a Model Trimer for
(65) Tavares, F. W.; Chang, J.; Sandler, S. I. Equation of State Monohydroxy and Dihydroxy PIB (1K) in Ethane, Propane,
for Square-Well Chain Fluid Based on the Dimer Version of Dimethyl Ether, Carbon Dioxide, and Chlorodifluoromethane. J.
Wertheim’s Perturbation Theory. Mol. Phys. 1995, 86, 1451. Phys. Chem. 1994, 98, 10634.
(66) Shukla, K. P.; Chapman, W. G. TPT2 and SAFTD Equa- (89) Gregg, C. J.; Stein, F. P.; Radosz, M. Phase Behavior of
tions of State for Mixtures of Hard Chain Copolymers. Mol. Phys. Telechelic Polyisobutylene in Subcritical and Supercritical Fluids.
2000, 98, 2045. 4. SAFT Association Parameters from FTIR for Blank, Monohy-
(67) Jog, P. K.; Chapman, W. G. Application of Wertheim’s droxy, and Dihydroxy PIB 200 in Ethane, Carbon Dioxide, and
Thermodynamic Perturbation Theory to Dipolar Hard Sphere Chlorodifluoromethane. J. Phys. Chem. B 1999, 103, 1167.
Chains. Mol. Phys. 1999, 97, 307. (90) Lee, S.-H.; LoStracco, M. A.; McHugh, M. A. High-Pressure
(68) Escobedo, F. A.; de Pablo, J. J. Chemical Potential and Molecular Weight-Dependent Behavior of Co(polymer)-Solvent
Equations of State of Hard Core Chain Molecules. J. Chem. Phys. Mixtures: Experiments and Modeling. Macromolecules 1994, 27,
1995, 103, 1946. 4652.
(69) Zhou, Y.; Smith, S. W.; Hall, C. K. Linear Dependence on (91) Hasch, B. M.; McHugh, M. A. Calculating Poly(ethylene-
Chain Length for the Thermodynamic Properties of Tangent Hard- co-acrylic acid)-Solvent Phase Behavior with the SAFT Equation
Sphere Chains. Mol. Phys. 1995, 86, 1157. of State. J. Polym. Sci., Polym. Phys. 1995, 33, 715.
(70) Escobedo, F. A.; de Pablo, J. J. Simulation and Prediction (92) Chen, S.-j.; Banaszak, M.; Radosz, M. Phase Behavior of
of Vapour-Liquid Equilibria for Chain Molecules. Mol. Phys. 1996, Poly(ethylene-1-butene) in Subcritical and Supercritical Pro-
87, 347. pane: Ethyl Branches Reduce Segment Energy and Enhance
(71) Shukla, K. P.; Chapman, W. G. A Two-Fluid Theory for Miscibility. Macromolecules 1995, 28, 1812.
Chain Fluid Mixtures from Thermodynamic Perturbation Theory. (93) Hasch, B. M.; Lee, S.-H.; McHugh, M. A. Strengths and
Mol. Phys. 1998, 93, 287. Limitations of SAFT for Calculating Polar Copolymer-Solvent
(72) Economou, I. G.; Tsonopoulos, C. Associating Models and Phase Behavior. J. Appl. Polym. Sci. 1996, 59, 1107.
Mixing Rules in Equations of State for Water/Hydrocarbon (94) Lee, S.-H.; Hasch, B. M.; McHugh, M. A. Calculating
Mixtures. Chem. Eng. Sci. 1997, 52, 511. Copolymer Solution Behavior with Statistical Associating Fluid
(73) Errington, J. R.; Boulougouris, G. C.; Economou, I. G.; Theory. Fluid Phase Equilib. 1996, 117, 61.
Panagiotopoulos, A. Z.; Theodorou, D. N. Molecular Simulation of (95) Lee, S.-H.; LoStracco, M. A.; McHugh, M. A. Cosolvent
Phase Equilibria for Water-Methane and Water-Ethane Mix- Effect on the Phase Behavior of Poly(ethylene-co-acrylic acid)-
tures. J. Phys. Chem. B 1998, 102, 8865. Butane Mixtures. Macromolecules 1996, 29, 1349.
(74) Galindo, A.; Whitehead, P. J.; Jackson, G.; Burgess, A. N. (96) Albrecht, K. L.; Stein, F. P.; Han, S. J.; Gregg, C. J.;
Predicting the High-Pressure Phase Equilibria of Water + n- Radosz, M. Phase Equilibria of Saturated and Unsaturated
Alkanes Using a Simplified SAFT Theory with Transferable Polyisoprene in Sub- and Supercritical Ethane, Ethylene, Propane,
Intermolecular Interaction Parameters. J. Phys. Chem. 1996, 100, Propylene, and Dimethyl Ether. Fluid Phase Equilib. 1996, 117,
6781. 84.
(75) Folie, B.; Radosz, M. Phase Equilibria in High-Pressure (97) Folie, B.; Gregg, C.; Luft, G.; Radosz, M. Phase Equilibria
Polyethylene Technology. Ind. Eng. Chem. Res. 1995, 34, 1501. of poly(ethylene-co-vinyl acetate) Copolymers in Subcritical and
(76) Kirby, C. F.; McHugh, M. A. Phase Behavior of Polymers Supercritical Ethylene and Ethylene-Vinyl Acetate Mixtures. Fluid
in Supercritical Fluid Solvents. Chem. Rev. 1999, 99, 565. Phase Equilib. 1996, 120, 11.
(77) Economou, I. G. Unpublished results, 1994. (98) Byun, H.-S.; Hasch, B. M.; McHugh, M. A.; Mähling, F.-
(78) Flory, P. J.; Orwoll, R. A.; Vrij, A. Statistical Thermody- O.; Busch, M.; Buback, M. Poly(ethylene-co-butyl acrylate). Phase
namics of Chain Molecule Liquids. I. An Equation of State for Behavior in Ethylene Compared to the Poly(ethylene-co-methyl
Normal Paraffin Hydrocarbons. J. Am. Chem. Soc. 1964, 86, 3507. acrylate)-Ethylene System and Aspects of Copolymerization
(79) Dee, G. T.; Walsh, D. J. Equations of State for Polymer Kinetics at High Pressures. Macromolecules 1996, 29, 1625.
Liquids. Macromolecules 1988, 21, 811. (99) Han, S. J.; Gregg, C. J.; Radosz, M. How the Solute
(80) Rodgers, P. A. Pressure-Volume-Temperature Relation- Polydispersity Affects the Cloud-Point and Coexistence Pressures
ships for Polymeric LiquidssA Review of Equations of State and in Propylene and Ethylene Solutions of Alternating Poly(ethylene-
Their Characteristic Parameters for 56 Polymers. J. Appl. Polym. co-propylene). Ind. Eng. Chem. Res. 1997, 36, 5520.
Sci. 1993, 48, 1061. (100) Lora, M.; Rindfleisch, F.; McHugh, M. A. Influence of the
(81) Song, Y.; Lambert, S. M.; Prausnitz, J. M. A Perturbed Alkyl Tail on the Solubility of Poly(alkyl acrylates) in Ethylene
Hard-Sphere-Chain Equation of State for Normal Fluids and and CO2 at High Pressures: Experiments and Modeling. J. Appl.
Polymers. Ind. Eng. Chem. Res. 1994, 33, 1047. Polym. Sci. 1999, 73, 1979.
(82) Economou, I. G.; Heidman, J. L. Unpublished results, 1997. (101) Lora, M.; McHugh, M. A. Phase Behavior and Modeling
(83) Chen, S.-J.; Economou, I. G.; Radosz, M. Density-Tuned of the Poly(methyl methacrylate)-CO2-Methyl methacrylate
Polyolefin Phase Equilibria. 2. Multicomponent Solutions of System. Fluid Phase Equilib. 1999, 157, 285.
Alternating Poly(ethylene-propylene) in Subcritical and Super- (102) Chan, A. K. C.; Radosz, M. Fluid-Liquid and Fluid-
critical Olefins. Experiment and SAFT Model. Macromolecules Solid-Phase Behavior of Poly(ethylene-co-hexene-1) Solutions in
1992, 25, 4987. Sub- and Supercritical Propane, Ethylene, and Ethylene + Hex-
(84) Chen, S.-J.; Radosz, M. Density-Tuned Polyolefin Phase ene-1. Macromolecules 2000, 33, 6800.
Equilibria. 1. Binary Solutions of Alternating Poly(ethylene- (103) Chan, A. K. C.; Adidharma, H.; Radosz, M. Fluid-Liquid
propylene) in Subcritical and Supercritical Propylene, 1-Butene and Fluid-Solid Transitions of Poly(ethylene-co-octene-1) in Sub-
and 1-Hexene. Experiment and Flory-Patterson Model. Macro- and Supercritical Propane Solutions. Ind. Eng. Chem. Res. 2000,
molecules 1992, 25, 3089. 39, 3069.
(85) Xiong, Y.; Kiran, E. Comparison of Sanchez-Lacombe and (104) Chan, A. K. C.; Adidharma, H.; Radosz, M. Fluid-Liquid
SAFT Model in Predicting Solubility of Polyethylene in High- Transitions of Poly(ethylene-co-octene-1) in Supercritical Ethylene
Pressure Fluids. J. Appl. Polym. Sci. 1995, 55, 1805. Solutions. Ind. Eng. Chem. Res. 2000, 39, 4370.
962 Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002

(105) Kinzl, M.; Luft, G.; Adidharma, H.; Radosz, M. SAFT Scattering Probe of Cloud-Point Pressure and Critical Polymer
Modeling of Inert-Gas Effects on the Cloud-Point Pressures in Concentration. Fluid Phase Equilib. 2000, 173, 149.
Ethylene Copolymerization Systems: Poly(ethylene-co-vinyl ac- (112) Boulougouris, G. C.; Voutsas, E. C.; Economou, I. G.;
etate) + Vinyl Acetate + Ethylene and Poly(ethylene-co-hexene- Theodorou, D. N.; Tassios, D. P. Henry’s Constant Analysis for
1) + Hexene-1 + Ethylene with Carbon Dioxide, Nitrogen, or Water and Nonpolar Solvents from Experimental data, Macro-
n-Butane. Ind. Eng. Chem. Res. 2000, 39, 541. scopic Models and Molecular Simulation. J. Phys. Chem. B 2001,
(106) Pan, C.; Radosz, M. Copolymer SAFT Modeling of Phase 105, 7792.
Behavior in Hydrocarbon-Chain Solutions: Alkane Oligomers, (113) Tumakaka, F.; Gross, J.; Sadowski, G. Modeling of
Polyethylene, Poly(ethylene-co-olefin-1), Polystyrene, and Poly- Polymer Phase Equilibria Using Perturbed-Chain SAFT. 9th
(ethylene-co-styrene). Ind. Eng. Chem. Res. 1998, 37, 3169. International Conference on Properties and Phase Equilibria for
(107) Pan, C.; Radosz, M. Phase Behavior of Poly(ethylene-co- Product and Process Design, Japan, 2001.
hexene-1) Solutions in Isobutane and Propane. Ind. Eng. Chem. (114) NIST chemistry webbook, NIST Standard Reference
Res. 1999, 38, 2842. Database, Number 69, Feb 2000 Release.
(108) Pan, C.; Radosz, M. Modeling of Solid-Liquid Equilibria (115) Abdulagatov, I. M.; Bazaev, A. R.; Gasanov, R. K.; Bazaev,
in Naphthalene, Normal-akane and Polyethylene Solutions. Fluid E. A.; Ramazanova, A. E. Measurement of PVTx Properties of
Phase Equilib. 1999, 155, 57. n-heptane in Supercritical Water. J. Supercrit. Fluids 1997, 10,
(109) Han, S. J.; Lohse, D. J.; Radosz, M.; Sperling, L. H. Short 149.
Chain Branching Effect on the Cloud-Point Pressures of Ethylene (116) Maloney, D. P.; Prausnitz, J. M. Solubilities of ethylene
Copolymers in Subcritical and Supercritical Propane. Macromol- and other organic solutes in liquid, low-density polyethylene in
ecules 1998, 31, 2533. the region 124-300 °C. AIChE J. 1976, 22, 74.
(110) Chen, A.-Q.; Radosz, M. Phase Equilibria of Dilute Poly- Received for review March 5, 2001
(ethylene-co-1-butene) Solutions in Ethylene, 1-Butene, and Revised manuscript received June 25, 2001
1-Butene + Ethylene. J. Chem. Eng. Data 1999, 44, 854. Accepted June 27, 2001
(111) Chan, A. K. C.; Russo, P. S.; Radosz, M. Fluid-Liquid
Equilibria in Poly(ethylene-co-hexene-1) + Propane: A Light IE0102201