Article

Cite This: ACS Earth Space Chem. 2019, 3, 2569−2581 http://pubs.acs.org/journal/aesccq

Modeling Binary Mixtures of Water + Light Hydrocarbon Using the

Perturbed-Chain Statistical Associating Fluid Theory with Induced
Association: Improvement in Describing All Equilibrium Phases
Sugata P. Tan,*,† Jeffrey S. Kargel,† Steven D. Vance,‡ and Rosaly M. Lopes‡
†
Planetary Science Institute, Tucson, Arizona 85719, United States
‡
Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California 91109, United States
*
S Supporting Information
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

ABSTRACT: Hydrophobic effects, including low mutual solubil-

ities, are well-documented between water and hydrocarbons.
However, as deduced from quantum chemical calculations and
Downloaded via UPPSALA UNIV on July 18, 2020 at 02:10:43 (UTC).

molecular simulations, an attractive interaction as a result of

electrostatic interactions and induction by water on hydrocarbon
molecules exists and must be included to adequately model water−
hydrocarbon systems. As shown in this work, the inclusion of these
interactions allows for the perturbed-chain statistical associating
fluid theory equation of state (EOS) to perform well in all
equilibrium phases. Without these corrections, the EOS usually
performs well only for the phase used for parametrization. Because
the EOS will also be used in near-freezing water temperatures, the
parameters of water are carefully derived to be able to reproduce
the density anomaly of water and behavior when extrapolated to the supercooled region. The improvements and calculations
will be useful for the assessment of the occurrences of potentially habitable water-poor and water-rich environments within the
ice crusts and rocky crusts of extraterrestrial bodies.
KEYWORDS: water−hydrocarbon mixtures, PC-SAFT, induced association, aqueous solubility, mutual solubility, supercooled water,
Titan

■ INTRODUCTION
Water and hydrocarbons are ubiquitous in the universe. While
they are known to be almost immiscible with each other, their
mixtures nevertheless extensively exist in nature. For example,
natural gas on Earth contains water (moisture), which is
undesired because it can cause corrosion of pipelines. For
engineering purposes, the saturation water content in gaseous
hydrocarbons is of great interest. In a very different context, in
cold extraterrestrial worlds where liquid water only exists under
the surface, hydrocarbon condensates on the surface result
from photolytic reactions in the atmosphere or potentially
from cryovolcanic eruptions. On Titan, the largest of Saturn’s
moons, hydrocarbons exist in a liquid phase in lakes and seas,
which are fed by hydrocarbon rain. These surface reservoirs
may even connect to subsurface alkanifers, constituting an
alkanology analogous to hydrology on Earth.1 It is likely that
liquid hydrocarbons meet water at some depth, either forming
nearly immiscible mixtures or solid clathrate hydrates. It also
seems likely that most of the surface hydrocarbons were
originally released as outgassed ethane and methane from
Titan’s interior,2 with further photolytic alterations in the
atmosphere, consistent with predictions of ethane and methane
clathrates suffusing the icy lithosphere.3 Many of the physical
water−hydrocarbon phenomena pertinent to industry and
biology must have analogues in geochemistry, geophysics, and
potential biology on Titan and other hydrocarbon icy worlds,
such as clathrate hydrate formation, oil−water phase
segregation in the subsurface ocean, toxicity, and metabolism.
As a result of the very low miscibility occurring in water−
hydrocarbon mixtures, experimental measurements are always
challenging. Experimental data are often mutually inconsis-
tently found in the literature. Even for simple systems, such as
water−methane and water−ethane mixtures, measurements
are subject to substantial errors.4,5 Evaluations of the data are
typically made available to check their consistency [e.g., by
International Union of Pure and Applied Chemistry (IUPAC)
in their Solubility Data Series6]. Consistent and accurate
thermodynamic models are therefore always desired to
describe these important systems. Modeling that applies
equations of state (EOS) is preferred, because it is easy to
implement and computationally fast and applies the additional
requirement of thermodynamic consistency.
Received: August 20, 2019
Revised: October 10, 2019
Accepted: October 14, 2019
Published: October 14, 2019

© 2019 American Chemical Society 2569 DOI: 10.1021/acsearthspacechem.9b00229

ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry
Unless the EOS can account for molecular association, such
as hydrogen bonding of water molecules, modeling of water−
hydrocarbon mixtures fails to describe the aqueous solubility of
hydrocarbons.7 Perturbed-chain statistical associating fluid
theory (PC-SAFT) is an EOS approach that includes
association in its building blocks, which makes it suitable for
modeling aqueous systems. It is also known to be accurate in
representing both vapor and liquid phases and has good
predictability as a result of its self-consistent underlying
molecular thermodynamics.8,9 As a result of its well-established
accuracy in modeling hydrocarbons, PC-SAFT is being applied
in planetary science. It has been coupled with a solid-phase
EOS, known as CRYOCHEM, which has been applied to
describe Titan’s atmospheric and surface condensates10 and
Pluto’s solid solutions.11 PC-SAFT was also applied to
investigate the phase behavior of fluids on Titan12−15 and
the solvation of nitrogen compounds in Titan’s seas.16 This
paper marks an extension of PC-SAFT to enable modeling of
planetary water−hydrocarbon systems.
Modeling water−hydrocarbon mixtures remains an active
area of research. Various EOS have been used; readers are
referred to recent papers and the references therein for a
deeper discussion.7,17,18 The difficulties in these modeling
efforts comprise two main issues: (1) few EOS can
simultaneously describe all equilibrium phases comparably
well, and (2) many EOS fail to reproduce the minimum
aqueous solubility of hydrocarbons with varying temperature.
There is not a single SAFT-based EOS that can simultaneously
resolve both issues. The first EOS that can capture the
solubility minimum was soft-SAFT,19 which loses accuracy in
the aqueous phase. Haarmann et al.7 applied PC-SAFT to
present the mutual solubility of water−n-alkane mixtures
starting from n-pentane to n-pentadecane and briefly discussed
the three-phase equilibrium of water−propane and water−n-
butane mixtures. The minimum aqueous solubility is well-
reproduced. However, for all mixtures discussed in the paper,
the description of aqueous phases is still substantially worse
than that of the hydrocarbon-rich phases.
It is common to fit the EOS binary parameters, which
correct the molecular dispersive interaction energy, to
experimental data for binary mixtures. For water−hydrocarbon
mixtures, the phase chosen for the fitting determines where the
model can be properly applied. In general, previous work used
the composition data for hydrocarbon phases to fit the model.
Consequently, they describe the saturated water content in
hydrocarbon phases quite well, but their prediction of the
aqueous solubility of hydrocarbons is less accurate.7 On the
other hand, if one is interested only in modeling the aqueous
solubility, then the composition data for the aqueous phase are
better used for parametrization, which results in a different
binary parameter. The latter parametrization is adopted in this
work because application for planetary bodies, such as Titan,
finds aqueous solubility to be of most interest, for example, in
studies of subsurface ice-crusted oceans. However, a few
kilometers below Titan’s surface, depending upon the thermal
characteristics of its icy lithosphere, hydrocarbons may coexist
with liquid water, creating water-poor liquids analogous to oil
slicks/plumes or water polluted with hydrocarbons on Earth.20
Hence, there is interest in consistent modeling for both the
aqueous and hydrocarbon-rich phases.
The failure in describing equilibrium phases equally well
raises a question of whether the interactions between water
and hydrocarbon molecules have been adequately represented
2570
Article
in the EOS or if there is still some missing physics to be taken
into account. Because hydrocarbon molecules are non-
associating, it is common to assume that they interact
attractively with water molecules only through a van-der-
Waals-type dispersive interaction.
However, from a PC-SAFT model 21 and molecular
dynamics simulations,22 the effective water Lennard−Jones
energy was found much larger than that commonly used in
simulation water models for describing the solubility of water
in alkanes. This prompted the same research group to perform
ab initio calculations,23 which revealed that nonpolar
molecules, such as n-alkanes can be electrostatically induced
by the presence of water molecules, consistent with other
independent works.24,25 They even conclude that the electro-
static and induction effects must be included to predict
correctly the water content in alkanes.26
The cartoon in Figure 1, which is adapted with permission
from ref 23, shows methane and water molecules at their
Figure 1. Methane and water molecules at their minimum energy
configuration. The broken line represents the electrostatic cross
interaction. This figure was adapted with permission from ref 23.
Copyright 2016 AIP Publishing.
minimum energy configuration obtained by a quantum
chemical calculation. There is an electrostatic type of cross
interaction between the molecules, where water donates a
proton to methane.
This weak interaction as a result of induction was never
accounted for in previous modeling works. Therefore, it is the
purpose of this work to improve the performance of the EOS
in describing water−hydrocarbon mixtures in all equilibrium
phases by including the induced interaction in PC-SAFT. With
application of the same binary parameters, the EOS may be
used to calculate the saturated water content useful for
engineering as well as the aqueous solubility of hydrocarbons
and water solubility in hydrocarbon-rich liquids that are useful
in planetary studies, such as that of Titan, or for environmental
work, e.g., in evaluating water polluted by hydrocarbons.
As a result of the same physics that governs the induction for
polar and nonpolar molecules, to a first approximation, the
approach chosen in this work is the simplest method
introduced by Kleiner and Sadowski,27 which is readily
incorporated into the original PC-SAFT framework. Their
method, which incorporates an interaction that they referred to
as induced association, was originally applied for polar
molecules induced by water and recently used for various
groups of polar compounds.28 The application of this approach
to water−hydrocarbon systems is introduced in this work by
assigning induced associating sites to the nonpolar molecules
that cross-associate with water molecules, as illustrated in
Figure 1.
Because PC-SAFT will also be used at near-freezing water
temperatures, the EOS parameters of water are carefully
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry
derived to be able to reproduce the density anomaly of water at
277 K. The new set of parameters can describe this anomaly
and can be well-extrapolated down to the supercooled region.
In this paper, we discuss mainly the first three light n-alkanes
and ethylene as a result of their abundance on Titan,29,30,12
which is relevant to the National Aeronautics and Space
Administration (NASA) project that supported the work.
Acetylene, which is also an abundant light hydrocarbon on
Titan, is not included because there are no experimental data
for saturated water content in the vapor phase, which are
needed to derive the induced association. Nevertheless, a brief
discussion of the aqueous solubility of acetylene is given in the
Supporting Information that accompanies this paper.
■
MODELING METHOD
EOS. In PC-SAFT, a molecule is modeled to consist of
spherical segments that make a chain via covalent bonds. The
chain molecules interact with each other via a short-range
dispersive interaction and longer range associating interactions,
such as that between water molecules via hydrogen bonds.
The building block of this EOS is the residual Helmholtz
energy that sums the above existing molecular interactions
a ̃R = a ̃HC + a ̃D + a ̃ A (1)
where the superscripts of the energy terms HC, D, and A
denote the hard-chain reference fluid, dispersion between
chains, and association between associating sites on the
interacting molecules. Other properties, including pressure
and densities, can be derived from eq 1.
PC-SAFT EOS has five major parameters for each chemical
species, namely, molecular segment number (m), segment
diameter (σ), and segment energy (ε) for all molecules and
two other parameters for associating molecules, i.e., the
association energy (εAB) and association volume (κAB) between
a proton-acceptor association site A and a proton-donor site B.
To describe mixtures, as a result of the imperfection of the
model potential energy used for the dispersive interaction
between segments, the cross segment energy parameter
between segments of molecules i and j needs a correction
using a binary interaction parameter kij.
εij = εiεj (1 − kij) (2)
All parameters are derived from experimental phase equili-
brium data. This makes the EOS both theoretically consistent
and, at the same time, accurate to describe real chemical
systems. However, as pointed out in the Introduction, the EOS
must apply two different kij values to represent the equilibrium
phases of hydrocarbon−water binary mixtures equally well.
One is fitted to the aqueous solubility of the hydrocarbon, and
the other is fitted to the water content in the hydrocarbon-rich
phase (liquid or vapor). As shown later, we will use the same kij
for all equilibrium phases, which is impossible to do without
correcting the missing interaction, i.e., by accounting for the
induction by water molecules on the hydrocarbon.
For phase equilibria calculations, the fugacity of each
component must be equal across the equilibrium phases at
temperature T and pressure P. For a binary aqueous mixture in
three-phase equilibrium (superscripts V = vapor, W = aqueous
phase, and H = hydrocarbon-rich liquid)
V W H
fi ̂ (y, T , P) = fi ̂ (x , T , P) = fi ̂ (q, T , P)
i = w, h (3a)
2571
Article
yi φî V (y, T , P) = xiφî W (x , T , P) = qiφî H(q, T , P)
i = w, h (3b)
where y = {yw, yh}, x = {xw, xh}, and q = {qw, qh} are the
compositions in vapor, aqueous phase, and hydrocarbon-rich
liquid, respectively. Subscripts w and h refer to water and the
hydrocarbon component, respectively. The fugacity coefficient
φ is derived from eq 1 using the EOS. The formulation of each
energy term in eq 1 and the fugacity coefficient can be found
elsewhere.8,31 For two-phase equilibria, the fugacity of the non-
existent phase must be removed from eq 3.
In studies of solubility, such as for water−hydrocarbon
binary mixtures in this work, xh and qw are the mutual
solubility, where xh is the solubility of the hydrocarbon in the
aqueous phase and qw is the solubility of water in the
hydrocarbon-rich liquid. This mutual solubility can be in two-
phase liquid−liquid equilibrium or three-phase vapor−liquid−
liquid equilibrium. For two-phase vapor−liquid equilibrium,
there are the aqueous solubility xh and the content of water in
vapor yw. As a result of the simple relationship between
solubility and heat of solution,32 accuracy in describing the
solubility xh will result in good estimates for the heat of
solution and the heat capacity of solution.
It is also worth noting that eq 3 can be extended to
multicomponent mixtures, where all pure-component and
binary parameters are carried over to the mixtures to make
good predictions. This extension is a subject of our current
research, of which the results will be presented in a subsequent
publication. However, to illustrate the EOS capability, an
example of calculations for ternary mixtures of water−
methane−ethane is provided in the Supporting Information.
Molecular Association and Polarity. Associating mole-
cules, such as water and alcohols, are modeled to have two
types of associating sites, i.e., proton-acceptor and proton-
donor sites, that can associate with each other between the
same chemical species (self-association) or different chemical
species (cross-association). Sites of the same type do not
associate. Therefore, water molecules self-associate with each
other and cross-associate with alcohol molecules. The cross
parameters between two associating molecules i and j are
calculated using the Wolbach−Sandler mixing rules.33
1 AB
εijAB = (εii + εjjAB)
2 (4a)
i jj y
zz
κiiABκjjAB jjj 1 zz
z
3
jj (σi + σj) zz
σσ
i j
k2 {
κijAB =
(4b)
The interaction between polar molecules, on the other hand,
may or may not be modeled using association. Polar molecules
having dipoles can be modeled to have dipolar interactions
among themselves and, thus, the need for an additional
interaction energy term in eq 1.34 Alternatively, a simplified
approach using association can be equally accurate, e.g., for
ketones and esters, by assuming two associating sites in each
molecule, one site for each type.35 The advantage of this
association model is that eq 1 does not need an extra energy
term for dealing with polar molecules.
In a different setting, the presence of associating molecules
may introduce induced association to other molecules in their
surroundings. Such an induction has been successfully
modeled by assigning one or more induced sites on the
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry Article

Figure 2. Performance of PC-SAFT for water with the current parameters. The vapor pressure is well-described by most models, as shown in panel
a, but the density anomaly is not, as shown in panel b. Below 273.15 K, the extrapolation enters the supercooled region. Symbols are data from the
National Institute of Standards and Technology (NIST).41 Calculations using parameters from Haarmann et al.7 are added for comparison.

Table 1. PC-SAFT Parametersa

kwh (eq 6)
species m σ (Å) ε/k (K) NSb (nA, nB) εAB/k (K) κAB c0 c1 c2
water 1.46223 eq 5a 142.3355 (2, 2) 2375.217 eq 5b
CH4 1.0000 3.7039 150.03 (1, 0) 0 eq 5b 0.27908 4.19823 87.39782
C2H6 1.6114 3.5245 190.9926 (0.5, 0) 0 eq 5b c
C2H4 1.6018 3.4029 176.5698 (3, 0) 0 eq 5b 0.20144 8.04737 62.88397
C3H8 2.0022 3.6209 207.7721 (1, 0) 0 eq 5b 0.20185 8.57 59.37154
a
In columns 4 and 6, k is the Boltzmann constant. bThe number of associating sites NS in terms of (nA, nB) to refer to proton-acceptor (A) and
proton-donor (B) types, respectively. ckwh(T) = −0.042063 + 5.68741 × 10−4T.

polar molecules that do not self-associate but can cross-
associate with sites of the opposite type on the associating
molecules.27 PC-SAFT with induced association was used to
model this kind of system, including mixtures of polar
molecules with alcohols or water,27 in which the induced
associating sites are assumed to consist of one proton-acceptor
site and one proton-donor site. To exclude self-association of
these polar molecules, εAB is set to zero but with non-zero κAB
equal to that of the associating component.27 Therefore, the
EOS does not need extra pure-component parameters.
However, the number of induced sites (NS) needs to be
estimated on a case-by-case basis, as described later.
The polarity of non-associating molecules may also arise
from the presence of either proton-acceptor (e.g., in acetone)
or proton-donor (e.g., in chloroform) sites but not both.27
Acetone and chloroform, which are polar but do not self-
associate, cross-associate with water molecules. Other examples
are aqueous solutions of CO2 and SO2, which have been
successfully described using this induced association ap-
proach.36 Both molecules are modeled to have four and five
proton-acceptor sites, respectively, corresponding to their
electron lone pairs at oxygen and sulfur atoms. Similarly,
SAFT-γ-Mie EOS assigns association sites, some of which are
inactive in pure fluid but active in mixtures.37 These sites are
referred to as “unlike induced” sites, which are effectively the
2572
same as the induced association sites discussed above. This
approach has been applied to describe acetone−water and
carboxylic acid−water binary mixtures.
Interestingly, as described in the Introduction (see Figure
1), nonpolar molecules, such as hydrocarbons, can also be
induced electrostatically by the presence of associating
molecules, such as water. Nevertheless, no attempts have
been made to model this kind of system using the induced
association as that applied to polar molecules, even though the
physics of induction is the same regardless of the polarity of
the induced molecules. This work is the first such attempt, of
which the results will be discussed in the next section.
To improve the EOS performance for water−hydrocarbon
mixtures, components of the mixtures must be accurately
described by the EOS. Otherwise, effects of the missing weak
interaction in mixtures could be masked by the inaccuracies in
describing the pure components. The conventional approach
of adjusting the binary parameter between water and
hydrocarbon (kwh) may compensate for these inaccuracies
but, consequently, loses the chance to elucidate the missing
interaction. While PC-SAFT is known to model hydrocarbons
well, it struggles in describing water, as all other EOS do.
Therefore, water deserves our attention, as discussed next.
Water. The density anomaly of liquid water at 277 K is the
most challenging feature for any EOS to reproduce. PC-SAFT
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry
Figure 3. PC-SAFT description of water in this work, with extrapolation down to 239.15 K in the supercooled region: (a) saturated vapor pressure
and (b) specific volume. Symbols are experimental data.42,43 Calculations using parameters from Haarmann et al.7 are added for comparison.
with its original water parameters9 cannot capture this
anomaly. There have been many studies using various
parameters, but none satisfactorily describes liquid water at
low temperatures. SAFT138 and soft-SAFT19 can reproduce
the anomaly using constant parameters but not at the correct
temperature. A recent review of modeling water molecules
using SAFT-based EOS is available.17 If the segment diameter
parameter (σ) is made temperature-dependent, the anomaly
can be well-represented.39,36
In SAFT-based EOS, each water molecule has been modeled
to have two to four associating sites.17 However, it is best to
model a water molecule with four associating sites, i.e., two
proton-donor sites at the hydrogen atoms and two proton-
acceptor sites corresponding to the electron lone pairs at the
oxygen atom.40 This four-site model is used in the present
work.
Because we will also deal with near-freezing temperatures in
this work, we make the association volume parameter (κAB)
temperature-dependent as well. This is necessary to improve
the EOS performance in the low-temperature region, as shown
in Figure 2. The parameters that were fitted to water-saturated
pressure and liquid density are given in Table 1 and eqs 5a and
5b. The average relative deviation (ARD) of the fitting is
0.04% for pressure and 0.02% for the density in the range of
273−393 K. With this high accuracy, the EOS is ready for
fixing the missing interaction. The density anomaly is shown in
Figure 2b as the minimum molar volume calculated to be at
277 K with the maximum density underestimated by 0.05%.
Another benefit of using these parameters is that the EOS
performance can be extrapolated to fit the experimental data in
the supercooled region without any additional treatment, as
shown in Figure 3. This enables the EOS to deal with
metastable supercooled water as well as aqueous liquid phases
near the eutectic that may exist below the freezing point of
pure water, such as that of some aqueous solutions of
hydrocarbons in this work or in clathrate hydrate systems.
As an additional note, for readers who model water
molecules using an association scheme, it may be of interest
2573
Article
to examine the EOS performance using the new parameter set
for matching the non-bonded site fraction with the
spectroscopic data available in the literature. A brief discussion
on this issue is given in the Supporting Information.
Parameters. The EOS parameters for pure compounds
used in this work are tabulated in Table 1, while the
temperature-dependent parameters of water, which are needed
for representing accurate phase equilibrium, including the
density anomaly, are as follows:
2735.913 1.277 × 106
σw(T ) = 4.815729 − +
T T2
2.66991 × 108 2.116541 × 1010
− + (Å)
T 3
T4 (5a)
AB
κ ww (T ) = ( −6.683097 + 0.8480913T ) × 10−4 (5b)
Note that the association parameters for hydrocarbon
molecules in Table 1 are εAB = 0 for the association energy
and equal to that of water for the association volume κAB to
accommodate induced association.
The binary interaction parameters between water and
hydrocarbon, except for ethane, are functions of the temper-
ature as follows:
k wh(T ) = c0 − c1 exp( −T /c 2) (6)
where coefficients c0, c1, and c2 are listed in the last columns of
Table 1. This exponential function is consistent with the work
by NguyenHuynh,18 which pointed out that the binary
parameters of aqueous hydrocarbons are constant at high
temperatures.
■
RESULTS AND DISCUSSION
While the main results were obtained by applying the
association induced by water on the hydrocarbon molecules,
those obtained without the induced association are also
presented here for comparison. The latter interaction accounts
for only self-association between water molecules, so that the
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry Article

Figure 4. Typical issues in modeling water−hydrocarbon mixtures, illustrated using the water−methane mixture. (a) Only if the induced
association is included can the equilibrium phases be described equally well. (b) Minimum solubility can be captured when using EOS that
accounts for association, such as PC-SAFT. Symbols are experimental data.44,45

Figure 5. Performance of PC-SAFT with and without the induced association at high (T/P): (a) aqueous solubility of methane used to fit kwh and
(b) predicted saturated water content in vapor. If kwh is fitted to yw, the predicted aqueous solubility overestimates the data by an order of
magnitude and, thus, cannot be shown in panel a. Symbols are experimental data.44,45

binary interaction parameters (kwh) are different from those in
Table 1 and given in the Supporting Information. Both
approaches derive kwh from the aqueous solubility data for
hydrocarbons. In fact, only this property of the mixtures is
sensitive to kwh. The saturated water content of the
hydrocarbon phase is predicted from the phase equilibrium
calculations. For the current approach with induced associa-
tion, the parametrization also needs to evaluate the number of
induced sites assigned to hydrocarbon molecules, NS, such that
the EOS can reproduce the water content in hydrocarbon-rich
phases at acceptable accuracies.
2574
To illustrate the main issues mentioned in the Introduction,
Figure 4 shows a typical situation for an EOS in modeling
water−hydrocarbon mixtures. In Figure 4a, without induced
association, the EOS performance in the equilibrium phases
depends upon which data are fitted for the parametrization. If
the water content in vapor is used (yw), then the EOS
overestimates the aqueous solubility data (dash-dotted curves).
Conversely, if the aqueous solubility data are used (xh), then
the EOS underestimates the water content in vapor (dashed
curves). There is no way to reconcile these phase
compositions, e.g., by adjusting kwh. However, the issue is
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry Article

Figure 6. Prediction of PC-SAFT with induced association at low (T/P) using kwh as in Figure 5: (a) aqueous solubility of methane and (b)
saturated water content in vapor. Symbols are experimental data.46−50

Figure 7. Performance of PC-SAFT with the induced association at low (T/P): (a) aqueous solubility of ethane used to fit kwh and (b) predicted
saturated water content in vapor. The calculation without induced association (dashed curves) is hardly different from that with it (solid curves).
Symbols are experimental data.53−56

resolved if the induced association is included in the EOS
(solid curves). While the binary parameter kwh is used to fit xh,
NS is estimated for the EOS to also match yw. There is only a
value of NS that enables consistent fitting of both xh and yw.
The second issue is the minimum aqueous solubility of
hydrocarbons at moderate temperature, as shown in Figure 4b.
PC-SAFT, which includes the molecular association, can
reproduce the minimum, with or without accounting for the
induced association. Therefore, it is evident here that the effect
of the induced association is negligible in the aqueous phase,
where the self-association among water molecules is dominant.
2575
Water−Methane Mixture. The performance of PC-SAFT
in describing water−methane mixtures in a wide range of
temperature/pressure (T/P) is shown in Figure 5 (high T/P)
and Figure 6 (low T/P). As shown in Figure 5b, the inclusion
of induced association substantially improves the prediction on
the water content in vapor. Of course, without the induced
association as previously mentioned, the binary parameter kwh
may be fitted to the vapor composition (yw) and found
constant (kwh = −0.12), of which the behavior could be
mistakenly thought to be a good model. In fact, in this case, the
calculated aqueous solubility of methane is an order of
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry Article

Figure 8. Performance of PC-SAFT with and without the induced association at three-phase equilibrium of the water−ethane mixture: (a) phase
compositions and (b) three-phase pressure. Symbols are experimental data.56,57

Figure 9. Performance of PC-SAFT with and without the induced association at three-phase equilibrium: (a) water content in propane-rich phases
and (b) aqueous solubility of propane. Symbols are experimental data.56−59 Calculations using parameters from Haarmann et al.7 are added for
comparison.

magnitude higher than the scale in Figure 5a and, thus, not
shown in the figure.
With the inclusion of induced association, kwh is exponential
with the temperature and approaches a constant at high
temperatures (see Table 1). The constant kwh at high
temperatures is consistent with a previous study.18 For
aqueous methane, the number of induced sites is to be set
to unity, NS = 1. Other numbers do not work to reconcile the
EOS performance in both phases. As shown in Figure 1, the
induced site on methane is of proton-acceptor type.
The performance at lower T/P is consistent with that at high
T/P (using the same kwh), as shown in Figure 6. The hydrate
2576
phase replaces the aqueous phase at a pressure of about 60 bar4
at 283 K, so that the vapor−hydrate boundary is even at lower
pressures at temperatures lower than 283 K. The data by Folas
et al.46 in Figure 6a at these low temperatures, therefore, may
be all or partly in equilibrium with the hydrate phase, instead
of the liquid phase. Consequently, the observed inaccuracy of
PC-SAFT at this region is due to the inability of the EOS to
account for the hydrate phase. Extension of the EOS with
hydrate phases is beyond the scope of this work and is the
subject of subsequent future work.
Water−Ethane Mixture. For this binary mixture, it is
known that the experimental aqueous solubility data at high T/
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry Article

Figure 10. Performance of PC-SAFT for the isothermal phase diagram of the water−propane mixture at 360.93 K: (a) aqueous phase and (b)
propane-rich phases. The legend is the same as in panel a. Lw, aqueous phase; LC3, propane-rich liquid; and V, vapor. Mole fractions: x (in Lw), q
(LC3), and y (in V). Symbols are experimental data.58 Calculations using parameters from Haarmann et al.7 are added for comparison.

Figure 11. Performance of PC-SAFT for the aqueous solubility of propane in the aqueous phase. The complete series of experimental data are
taken from Chapoy et al.60 In panel a, a set of data (filled squares)61 is shown inconsistent with the others, while in panel b, two sets of data (filled
diamonds and triangles)58 are consistent with those of Chapoy et al. Calculations using parameters from Haarmann et al.7 are added for
comparison in panel a.

P51,52 are considered tentative, as stated in a review by IUPAC,
which states that their data underestimate the recommended
data when extrapolated to low pressures.5 Therefore, only low
T/P data are presented here. A brief discussion on the
inconsistencies of the experimental data of this mixture is given
in the Supporting Information.
The EOS performance at low T/P is shown in Figure 7. As a
result of the short range of the temperature in this region, kwh
is approximated to be linear with the temperature (see Table
1). If reliable data are available in the whole range of T/P, the
2577
linearity may be a lower part of an exponential function, thus
the same behavior as the other hydrocarbons in Table 1. In this
low-pressure range, the discrepancy between including and
omitting the induced association is small in both vapor and
aqueous phases.
However, the improvement is apparent in the three-phase
equilibrium, as shown in Figure 8, when using NS = 0.5, which
indicates the very weak induced association in the water−
ethane mixture. Following water−methane, the induced site is
also assigned to be of proton-acceptor type. The description of
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry
Figure 12. Performance of PC-SAFT with and without the induced association: (a) aqueous solubility of ethylene and (b) predicted saturated
water content in vapor. The aqueous solubility data52 were used to fit kwh (the curves with and without induced association almost coincide with
each other; not shown for clarity). Other experimental data in panel a are from Bradbury et al.63
ethane-rich liquid and vapor is much better with the induced
association than without it.
Water−Propane Mixture. This aqueous binary mixture
has been used to evaluate previous models in the literature,
particularly regarding the three-phase equilibrium involving
vapor, aqueous phase, and propane-rich liquid. As shown in
Figure 9, the inclusion of induced association undoubtedly
improves the EOS performance in all phases. The presence of a
solubility minimum in the aqueous phase is well-captured by
our model. A recent model by Haarmann et al.7 can also do the
same but with less accuracy. The apparent large deviation at
the lowest temperature in Figure 9b is likely because the datum
was an extrapolation to the hydrate-phase boundary.58
The EOS performance in describing the two-phase regions is
illustrated by the isothermal diagram shown in Figure 10. After
the inclusion of induced association, the prediction for vapor
and propane-rich liquid is improved in bordering the two-
phase vapor−liquid and liquid−liquid regions, respectively. For
this binary mixture, NS = 1 with the site type of proton
acceptor, as it is for the water−methane mixture. The EOS
performance for the aqueous solubility of propane is shown in
Figure 11, which is based on the experimental data series by
Chapoy et al.60
Water−Ethylene Mixture. Unlike the n-alkanes described
above, the saturated water content in the vapor phase exhibits a
minimum with pressure. This phenomenon can be modeled
using PC-SAFT only if the binary mixture has stronger induced
association, perhaps as a result of π electrons in ethylene
molecules. The π electrons, which are weak proton acceptors,
can have a weak hydrogen bond with a strong proton donor.62
The stronger association for ethylene is reflected by NS = 3 as
listed in Table 1, with the site type of proton acceptor. The
other two proton acceptors may be assigned to the two carbon
atoms of ethylene in a fashion similar to that of methane in
Figure 1.
As shown in Figure 12b by the water vapor content
predictions, the improvement using induced association is
2578
Article
dramatic. Meanwhile, the minimum aqueous solubility with
varying temperature can be estimated from Figure 12a, in
which the curve for 344 K is the lowest, as also the case for the
water−methane mixture in Figure 5a.
In summary, for all cases of water−light hydrocarbon
mixtures above, the binary parameter kwh was fitted to the
aqueous solubility of hydrocarbon, while the water content is
not sensitive to it. However, the latter was much more affected
by the inclusion of induced association. This means that the
induction effect is observable only if the number of water
molecules is small, such as that in hydrocarbon-rich vapor or
liquid. In the aqueous phase, where water is dominant, the
induced association is too weak to show observable effects.
Emborsky et al.,64 who also used PC-SAFT, removed the
association of water when water is surrounded by numerous
nonpolar molecules in the hydrocarbon-rich phases, at the
expense of inaccuracies in representing pure water.
Finally, it is also worth noting that the number of induced-
association sites, NS, cannot be arbitrarily chosen; numbers
other than those listed in Table 1 just do not work. To
illustrate the situation, the crosses in Figure 8 represent the
calculations using NS = 1 (instead of 0.5) for the water−ethane
mixture. While the location of xh is easily adjusted by changing
kwh (the connected three crosses with the smallest kwh for the
top one), that of yw and qw virtually remain the same upon
changes of kwh. The latter two mole fractions substantially
change only if a different NS is used. Nevertheless, there is no
obvious trend of NS from one hydrocarbon to the other.
Lacking the physical basis, as is now the case, NS is treated
rather as a free parameter that is a constant depending upon
the hydrocarbon species induced by water. Therefore, instead
of making baseless speculations, this situation calls for further
investigation that is beyond the scope of this work. The
present work suggests that, in addition to the well-known weak
hydrogen bond62 O−H···π in the case of ethylene−water,
there is also a bond as a result of the induced association O−
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry
H···C−H in Figure 1 for methane−water and perhaps other
new types of bonds in the other mixtures.
■
CONCLUSIONS AND REMARKS
The association induced by the presence of water on light
hydrocarbon molecules is required to reconcile the modeling
of all equilibrium phases in the water−hydrocarbon mixtures.
Without it, equilibrium phase compositions cannot be modeled
consistently using the same binary parameter. To show a
prospect that this improvement may also be valid for heavier
hydrocarbons, we provide the EOS performance for the three-
phase equilibrium of the water−n-butane mixture in the
Supporting Information. Future work may proceed with even
heavier molecules to investigate whether the induced
association is valid for any water−hydrocarbon system.
The effect of induced association on the modeling is only
apparent in hydrocarbon-rich phases where the number of
water molecules is small. In the aqueous phase, it is
overwhelmed by the hydrogen-bond self-association between
water molecules, so that the aqueous solubility of hydro-
carbons may be modeled without it but at the expense of
inaccurate representation of hydrocarbon-rich phases. The
same situation occurs when modeling the saturated water
content in the vapor or hydrocarbon-rich liquid, where the
EOS still works without induced association but loses the
ability to predict the aqueous phase. Therefore, as also justified
by theory and simulation, the inclusion of cross association
between water and hydrocarbons, such as induced association,
is crucial to obtain consistencies in all equilibrium phases. This
way, the EOS with a set of parameters can be used for
multiphase applications.
However, despite the substantial qualitative improvements
in the hydrocarbon-rich phases, as evident in the results
presented in this paper, the quantitative description of the EOS
in these phases still has room for further refinement. With the
promising results in this paper, a further rigorous development
of this induced association approach for water−hydrocarbon
systems is strongly encouraged for future work. For this
purpose, in addition to the improvements in the EOS
development, better experimental data are absolutely needed,
e.g., ethane as discussed above and in the Supporting
Information, for which the data are not consistent across the
phases at low versus high pressures.
The model for water−hydrocarbon mixtures in this work
enables the assessment of the occurrences of potentially
habitable hydrocarbon-bearing water-poor and water-rich
environments within the ice crusts, sub-ice oceans, seafloors,
and sub-oceanic rocky crusts of icy ocean worlds. Hydro-
carbons in Titan’s icy lithosphere, for example, may undergo
liquid−gas and solid−liquid phase transitions as a result of the
temperature increases by solid convection,1 while liquid water
is also present below the surface, perhaps because of enhanced
tidal heating.65 When these compounds meet, the composition
and densities of the equilibrium phases provide indispensable
information for the larger applications of this work on the
habitability of hydrocarbon worlds, such as Titan and beyond.
Applications may also be developed for hydrocarbon−water
systems in the martian crust, in chondritic asteroids, and in the
cold-region crust of Earth, such as Greenland and Antarctica
and in the Arctic Ocean. Extension to include clathrate hydrate
phases is a necessary step for some anticipated applications,
whereas other applications may require consideration of
equilibria with saline solutions.
2579
Article
■
*
ASSOCIATED CONTENT
S Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acsearthspace-
chem.9b00229.
(1) Parameters for PC-SAFT without induced associa-
tion, (2) performance of PC-SAFT in describing
aqueous solubility of acetylene, (3) brief discussion on
the data inconsistency of water−ethane mixtures, (4)
example of ternary mixture calculations, (5) perform-
ance of PC-SAFT in describing the three-phase
equilibria of the water−n-butane mixture, and (6) brief
discussion on the fraction of non-bonded sites in the
PC-SAFT model for water (PDF)
■ AUTHOR INFORMATION
Corresponding Author
*E-mail: stan@psi.edu.
ORCID
Sugata P. Tan: 0000-0003-2049-3706
Notes
The authors declare no competing financial interest.
■ ACKNOWLEDGMENTS
This work was supported by NASA Astrobiology Institute
Grant 17-NAI8_2-0017: “Habitability of Hydrocarbon
Worlds: Titan and Beyond”. J.K. and S.V. also acknowledge
support from NASA Grant 13-13NAI7_2-0024. Part of this
research was carried out at the Jet Propulsion Laboratory,
California Institute of Technology, under a contract with
NASA.
■
REFERENCES
(1) Vance, S.; Sotin, C.; Choukroun, M.; Mitchell, K. Titan’s
subsurface alkanology. Proceedings of the 43rd Lunar and Planetary
Science Conference; The Woodlands, TX, March 19−23, 2012;
Abstract 2939.
(2) Miller, K. E.; Glein, C. R.; Waite, J. H. Contributions from
accreted organics to Titan’s atmosphere: New insights from cometary
and chondritic data. Astrophys. J. 2019, 871, 59.
(3) Choukroun, M.; Sotin, C. Is Titan’s shape caused by its
meteorology and carbon cycle? Geophys. Res. Lett. 2012, 39, L04201.
(4) Chapoy, A.; Coquelet, D.; Richon, D. Corrigendum to “Revised
solubility data and modeling of water in the gas phase of the
methane/water binary system at temperatures from 283.08 to
318.12K and pressures up to 34.5MPa” [Fluid Phase Equilibria 214
(2003) 101−117]. Fluid Phase Equilib. 2005, 230, 210−214.
(5) Hayduk, W. IUPAC Solubility Data Series; Pergamon Press:
Oxford, U.K., 1982; Vol. 8: Ethane, p 16.
(6) National Institute of Standards and Technology (NIST). IUPAC
Solubility Data Series; NIST: Gaithersburg, MD, 2015; https://srdata.
nist.gov/solubility/IUPAC/iupac.aspx (accessed June 17, 2019).
(7) Haarmann, N.; Enders, S.; Sadowski, G. Modeling binary
mixtures of n-alkanes and water using PC-SAFT. Fluid Phase Equilib.
2018, 470, 203−211.
(8) Gross, J.; Sadowski, G. Perturbed-Chain SAFT: An equation of
state based on a perturbation theory for chain molecules. Ind. Eng.
Chem. Res. 2001, 40, 1244−1260.
(9) Gross, J.; Sadowski, G. Application of the perturbed-chain SAFT
equation of state to associating systems. Ind. Eng. Chem. Res. 2002, 41,
5510−5515.
(10) Tan, S. P.; Kargel, J. S. Multiphase-equilibria analysis:
Application in modeling the atmospheric and lacustrine chemical
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry
systems of Saturn’s moon Titan. Fluid Phase Equilib. 2018, 458, 153−
169.
(11) Tan, S. P.; Kargel, J. S. Solid-phase equilibria on Pluto’s surface.
Mon. Not. R. Astron. Soc. 2018, 474, 4254−4263.
(12) Tan, S. P.; Kargel, J. S.; Marion, G. M. Titan’s atmosphere and
surface liquid: New calculation using Statistical Associating Fluid
Theory. Icarus 2013, 222, 53−72.
(13) Tan, S. P.; Kargel, J. S.; Jennings, D. E.; Mastrogiuseppe, M.;
Adidharma, H.; Marion, G. M. Titan’s liquids: Exotic behavior and its
implications on global fluid circulation. Icarus 2015, 250, 64−75.
(14) Cordier, D.; García-Sánchez, F.; Justo-García, D. N.; Liger-
Belair, G. Bubble streams in Titan’s seas as a product of liquid N2+
CH4+ C2H6 cryogenic mixture. Nat. Astron. 2017, 1, 0102.
(15) Cordier, D.; Liger-Belair, G. Bubbles in Titan’s Seas:
Nucleation, Growth, and RADAR Signature. Astrophys. J. 2018, 859,
26.
(16) Stevenson, J. M.; Fouad, W. A.; Shalloway, D.; Usher, D.;
Lunine, J.; Chapman, W. G.; Clancy, P. Solvation of nitrogen
compounds in Titan’s seas, precipitates, and atmosphere. Icarus 2015,
256, 1−12.
(17) Vega, L. F.; Llovell, F. Review and new insights into the
application of molecular-based equations of state to water and
aqueous solutions. Fluid Phase Equilib. 2016, 416, 150−173.
(18) NguyenHuynh, D. Accurate modeling of multiphase behavior
of aqueous systems. I. Alkanes, alkenes, cycloalkanes, alcohols,
aromatics. Fluid Phase Equilib. 2018, 473, 201−219.
(19) Vega, L. F.; Llovell, F.; Blas, F. J. Capturing the Solubility
Minima of n-Alkanes in Water by Soft-SAFT. J. Phys. Chem. B 2009,
113, 7621−7630.
(20) King, G.; Kostka, J.; Hazen, T.; Sobecky, P. Microbial responses
to the deepwater horizon oil spill: From coastal wetlands to the deep
sea. Annu. Rev. Mar. Sci. 2015, 7, 377−401.
(21) Fouad, W. A.; Ballal, D.; Cox, K. R.; Chapman, W. G.
Examining the consistency of water content data in alkanes using the
perturbed-chain form of the statistical associating fluid theory
equation of state. J. Chem. Eng. Data 2014, 59, 1016−1023.
(22) Ballal, D.; Venkataraman, P.; Fouad, W. A.; Cox, K. R.;
Chapman, W. G. Isolating the non-polar contributions to the
intermolecular potential for water-alkane interactions. J. Chem. Phys.
2014, 141, 064905.
(23) Asthagiri, D.; Ballal, D.; Venkataraman, P.; Fouad, W. A.; Cox,
K. R.; Chapman, W. G. Response to “Comment on ‘Isolating the non-
polar contributions to the intermolecular potential for water-alkane
interactions”’. J. Chem. Phys. 2016, 144, 137102.
(24) Jiang, H.; Economou, I. G.; Panagiotopoulos, A. Z. Phase
equilibria of water/CO2 and water/n-alkane mixtures from polar-
izable models. J. Phys. Chem. B 2017, 121, 1386−1395.
(25) McDaniel, J. G.; Yethiraj, A. Comment on “Isolating the non-
polar contributions to the intermolecular potential for water-alkane
interactions. J. Chem. Phys. 2016, 144, 137101.
(26) Asthagiri, D.; Valiya Parambathu, A.; Ballal, D.; Chapman, W.
G. Electrostatic and induction effects in the solubility of water in
alkanes. J. Chem. Phys. 2017, 147, 074506.
(27) Kleiner, M.; Sadowski, G. Modeling of Polar Systems Using
PCP-SAFT: An Approach to Account for Induced-Association
Interactions. J. Phys. Chem. C 2007, 111, 15544−15553.
(28) Haarmann, N.; Enders, S.; Sadowski, G. Heterosegmental
Modeling of Long-Chain Molecules and Related Mixtures using PC-
SAFT: 1. Polar Compounds. Ind. Eng. Chem. Res. 2019, 58, 2551−
2574.
(29) Strobel, D. F.; Atreya, S. K.; Bézard, B.; Ferri, F.; Flasar, F. M.;
Fulchignoni, M.; Lellouch, E.; Mü ller-Wodarg, I. Atmospheric
structure and composition. In Titan from Cassini-Huygens; Brown,
R. H., Lebreton, J.-P., Waite, J. H., Eds.; Springer: Dordrecht,
Netherlands, 2009; Chapter 10, pp 235−257 and the references
therein, DOI: 10.1007/978-1-4020-9215-2_10.
(30) Niemann, H. B.; Atreya, S. K.; Demick, J. E.; Gautier, D.;
Haberman, J. A.; Harpold, D. N.; Kasprzak, W. T.; Lunine, J. I.;
Owen, T. C.; Raulin, F. Composition of Titan’s lower atmosphere and
2580
Article
simple surface volatiles as measured by the Cassini−Huygens probe
gas chromatograph mass spectrometer experiment. J. Geophys. Res.
2010, 115, E12006.
(31) Tan, S. P.; Adidharma, H.; Radosz, M. Generalized procedure
for estimating the fractions of nonbonded associating molecules and
their derivatives in thermodynamic perturbation theory. Ind. Eng.
Chem. Res. 2004, 43, 203−208.
(32) Tsonopoulos, C. Thermodynamic analysis of the mutual
solubilities of hydrocarbons and water. Fluid Phase Equilib. 2001, 186,
185−206.
(33) Wolbach, J. P.; Sandler, S. I. Using molecular orbital
calculations to describe the phase behavior of cross-associating
mixtures. Ind. Eng. Chem. Res. 1998, 37, 2917−2928.
(34) Tumakaka, F.; Sadowski, G. Application of the Perturbed-
Chain SAFT equation of state to polar systems. Fluid Phase Equilib.
2004, 217, 233−239.
(35) von Solms, N.; Michelsen, M. L.; Kontogeorgis, G. M. Applying
association theories to polar fluids. Ind. Eng. Chem. Res. 2004, 43,
1803−1806.
(36) Tan, S. P.; Yao, Y.; Piri, M. Modeling the Solubility of SO2 +
CO2 Mixtures in Brine at Elevated Pressures and Temperatures. Ind.
Eng. Chem. Res. 2013, 52, 10864−10872.
(37) Sadeqzadeh, M.; Papaioannou, V.; Dufal, S.; Adjiman, C. S.;
Jackson, G.; Galindo, A. The development of unlike induced
association-site models to study the phase behaviour of aqueous
mixtures comprising acetone, alkanes and alkyl carboxylic acids with
the SAFT-γ Mie group contribution methodology. Fluid Phase Equilib.
2016, 407, 39−57.
(38) Tan, S. P.; Adidharma, H.; Radosz, M. Statistical Associating
Fluid Theory Coupled with Restricted Primitive Model to Represent
Aqueous Strong Electrolytes. Ind. Eng. Chem. Res. 2005, 44, 4442−
4452.
(39) Held, C.; Cameretti, L. F.; Sadowski, G. Modeling aqueous
electrolyte solutions. Part 1. Fully dissociated electrolytes. Fluid Phase
Equilib. 2008, 270, 87−96.
(40) Kontogeorgis, G. M.; Tsivintzelis, I.; von Solms, N.; Grenner,
A.; Bøgh, D.; Frost, M.; Knage-Rasmussen, A.; Economou, I. G. Use
of monomer fraction data in the parametrization of association
theories. Fluid Phase Equilib. 2010, 296, 219−229.
(41) Lemmon, E. W.; McLinden, M. O.; Friend, D. G.
Thermophysical Properties of Fluid Systems. In NIST Chemistry
WebBook, NIST Standard Reference Database Number 69; Linstrom, P.
J., Mallard, W. G., Eds.; National Institute of Standards and
Technology (NIST): Gaithersburg, MD, 2017; DOI: 10.18434/
T4D303 (accessed Oct 2018).
(42) Murphy, D. M.; Koop, T. Review of the vapour pressures of ice
and supercooled water for atmospheric applications. Q. J. R. Meteorol.
Soc. 2005, 131, 1539−1565.
(43) Zheleznyi, B. V. Density of supercooled water. Russ. J. Phys.
Chem. 1969, 43, 1311.
(44) Culberson, O. L.; McKetta, J. J. Phase equilibria in
hydrocarbon-water systems IIIThe solubility of methane in water
at pressures to 10,000 psia. JPT, J. Pet. Technol. 1951, 3, 223−226.
(45) Olds, R. H.; Sage, B. H.; Lacey, W. N. Composition of the dew-
point gas of the methane−water system. Ind. Eng. Chem. 1942, 34,
1223−1227.
(46) Folas, G. K.; Froyna, E. W.; Lovland, J.; Kontogeorgis, G. M.;
Solbraa, E. Data and prediction of water content of high-pressure
nitrogen, methane and natural gas. Fluid Phase Equilib. 2007, 252,
162−174.
(47) Chapoy, A.; Mohammadi, A. H.; Richon, D.; Tohidi, B. Gas
solubility measurement and modeling for methane−water and
methane−ethane−n-butane−water systems at low temperature
conditions. Fluid Phase Equilib. 2004, 220, 111−119.
(48) Chapoy, A.; Coquelet, C.; Richon, D. Corrigendum to “Revised
solubility data and modeling of water in the gas phase of the
methane/water binary system at temperatures from 283.08 to 318.12
K and pressures up to 34.5 MPa. Fluid Phase Equilib. 2005, 230, 210−
214.
DOI: 10.1021/acsearthspacechem.9b00229
ACS Earth Space Chem. 2019, 3, 2569−2581
ACS Earth and Space Chemistry Article

(49) Rigby, M.; Prausnitz, J. M. Solubility of water in compressed

nitrogen, argon, and methane. J. Phys. Chem. 1968, 72, 330−334.
(50) Chapoy, A.; Mohammadi, A. H.; Tohidi, B.; Richon, D.
Estimation of water content for methane+ water and methane+
ethane+ n-butane+ water systems using a new sampling device. J.
Chem. Eng. Data 2005, 50, 1157−1161.
(51) Culberson, O. L.; McKetta, J. J. Phase Equilibria in
Hydrocarbon-Water Systems. IIThe Solubility of Ethane in Water
at Pressures to 10,000 psi. JPT, J. Pet. Technol. 1950, 2, 319−322.
(52) Anthony, R. G.; McKetta, J. J. Phase Equilibrium in the
Ethylene-Water System. J. Chem. Eng. Data 1967, 12, 17−20.
(53) Mohammadi, A. H.; Chapoy, A.; Tohidi, B.; Richon, D.
Measurements and thermodynamic modeling of vapor-liquid
equilibria in ethane-water systems from 274.26 to 343.08 K. Ind.
Eng. Chem. Res. 2004, 43, 5418−5424.
(54) Althaus, K. Fortschritt-Berichte; VDI: Düsseldorf, Germany,
1999; Reihe 3, 350 (in German). Oellrich, L. R.; Althaus, K. GERG-
Water Correlation (GERG Technical Monograph TM14) Relation-
ship between Water Content and Water Dew Point Keeping in
Consideration the Gas Composition in the Field of Natural Gas;
Fortschritt-Berichte; VDI: Düsseldorf, Germany, 2000; Reihe 3, 679 (in
English).
(55) Chapoy, A.; Coquelet, C.; Richon, D. Measurement of the
water solubility in the gas phase of the ethane+ water binary system
near hydrate forming conditions. J. Chem. Eng. Data 2003, 48, 957−
966.
(56) Mokraoui, S.; Coquelet, C.; Valtz, A.; Hegel, P. E.; Richon, D.
New solubility data of hydrocarbons in water and modeling
concerning vapor-liquid-liquid binary systems. Ind. Eng. Chem. Res.
2007, 46, 9257−9262.
(57) Song, K. Y.; Kobayashi, R. The water content of ethane,
propane and their mixtures in equilibrium with liquid water or
hydrates. Fluid Phase Equilib. 1994, 95, 281−298.
(58) Kobayashi, R.; Katz, D. Vapor-liquid equilibria for binary
hydrocarbon-water systems. Ind. Eng. Chem. 1953, 45, 440−446.
(59) Poettmann, F. H.; Dean, M. R. Water content of propane. Pet.
Refin. 1946, 25, 125−128.
(60) Chapoy, A.; Mokraoui, S.; Valtz, A.; Richon, D.; Mohammadi,
A. H.; Tohidi, B. Solubility measurement and modeling for the system
propane−water from 277.62 to 368.16 K. Fluid Phase Equilib. 2004,
226, 213−220.
(61) Azarnoosh, A.; McKetta, J. J. The solubility of propane in water.
Pet. Refin. 1958, 37, 275−278.
(62) Desiraju, G. R. Hydrogen bridges in crystal engineering:
Interactions without borders. Acc. Chem. Res. 2002, 35, 565−573.
(63) Bradbury, E. J.; McNulty, D.; Savage, R. I.; McSweeney, E. E.
Solubility of ethylene in water. Effect of temperature and pressure.
Ind. Eng. Chem. 1952, 44, 211−212.
(64) Emborsky, C. P.; Cox, K. R.; Chapman, W. G. Correlation and
Prediction of Water Content in Alkanes Using a Molecular Theory.
Ind. Eng. Chem. Res. 2011, 50, 7791−7799.
(65) Kalousova, K.; Soucek, O.; Tobie, G.; Choblet, G.; Cadek, O.
Ice melting and downward transport of meltwater in two-phase flow
in Europa’s ice shell. J. Geophys. Res. Planets 2014, 119, 532−549.

2581 DOI: 10.1021/acsearthspacechem.9b00229

ACS Earth Space Chem. 2019, 3, 2569−2581