SPE-187163-MS

A Statistical Mechanics Model for PVT Behavior in Nanopores

Y. B. Coskuner, X. Yin, and E. Ozkan, Colorado School of Mines

Copyright 2017, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Annual Technical Conference and Exhibition held in San Antonio, Texas, USA, 9-11 October 2017.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
Recent studies on phase behavior in nanopore confinement reveal inconsistent and contradicting results
about the shift of the phase diagram and critical point. This paper focuses on statistical mechanics and
molecular simulation to model the PVT behavior in confinement. Grand Canonical Monte-Carlo simulation
is used to observe the effect of confinement on phase behavior of pure methane in 2 nm. The model takes
into consideration the effects of the intermolecular forces between fluid particles and between fluid particles
and solid surface. Density of methane is obtained from simulations at different pressures under isothermal
conditions to determine the phase transition point. Results are compared with the published studies and the
differences are discussed. It is shown that the size of the simulation box significantly affects the results
of molecular simulation. As a result, some of the conclusions drawn in the literature about the shift of the
critical point are questioned.

Introduction
Unconventional phase behavior in nanoporous media have been recognized as an important factor in
modeling and predicting the behavior of unconventional reservoirs. In this study, we examine the phase
behavior under severe confinement (less than 10nm). Recently, surface forces have been noted to be more
dominant than the capillary forces in pores under 60-nm radius (Meyer et al. 2009). Similarly, intermolecular
forces become comparable to capillary forces under 10-nm pore sizes. Unfortunately, conventional PVT
cell measurements cannot take into account the capillaryand surface-force effects and the technology for
the direct measurement of PVT properties in nanopores is not yet available. Therefore, theoretical models
based on equilibrium thermodynamics (e.g., Travalloni et al. 2010, Sapmanee 2011, Firincioglu et al. 2012,
Honarpour et al. 2012, Teklu et al. 2014), experimental models on nanofluidics chips (e.g., Wang et al 2014
and Parsa et al. 2015), and simulation studies based on molecular dynamics (e.g., Makimura et al. 2011 and
Pitakbunkate et al. 2016) have been mostly used in the recent studies of phase behavior in unconventional
reservoirs.
In addition to the differences caused by the inherent assumptions and limitations of the different
approaches used to study phase behavior in confinement, even the results of the studies using the same
approach do not seem to agree. In principle, there is an agreement on the shift of the bubble point and dew
point curves; however, the discrepancies in the magnitude of the shift are surmounting. Another area of
2 SPE-187163-MS

disagreement is the shift of the critical point. Some studies start from a shifted critical point to predict the
shift of the rest of the phase envelope (Sapmanee 2011, Teklu et al. 2014) while some others use a fixed
critical point and shift the rest of the phase envelope (Firincioglu et al. 2013). Even for the cases a shift of
the critical point is presumed or allowed, there are large differences in the expected magnitude of the shift.
In this paper, we use the Grand Canonical Monte-Carlo (GCMC) simulation based on statistical
mechanics to calculate the impact of intermolecular forces on PVT behavior. The tool used in the study
is RASPA, which is a molecular simulation software for adsorption and diffusion in flexible nanoporous
materials. A modified equation of state presented by Teklu et al. (2014) is also used to account for the effect
of capillary forces on PVT behavior and to estimate the bubble point suppression. While equation of state
takes macroscopic thermodynamic properties into account, molecular simulation considers microscopic
properties such as inter-particle interactions.
Below, we first introduce the background of our discussions, which intends to highlight the differences
of the results obtained by different approaches and shed some light or grow the skepticism on the causes of
the differences. Then we present the methodology used in the GCMC simulation. Finally, we present our
results to demonstrate the sensitivity of molecular simulation to the size of the simulation box and conclude
with comments on the shift of the critical point observed or used in the previous studies.

Background
Recently, Teklu et al. (2014) studied the effect of confinement on phase behavior of Bakken oil using Peng-
Robinson Equation of State (EOS), which was modified for capillary pressure. They also implemented the
critical temperature and pressure shifts in nanopores using the method presented by Zarragoicoechea &
Kuz (2004). In the work of Zarragoicoechea & Kuz (2004) it was discussed that the critical points were
depended on the ratio between Lennard-Jones size parameter and the pore-throat radius. The Lennard Jones
size parameter is given by

(1)

and the shifts in the critical temperature and pressure are calculated from the following equations,
respectively:

(2)

and

(3)

Pitakbunkate et al. (2016) studied intermolecular interactions and their effect on phase behavior of
confined methane-ethane mixture using GCMC simulation. They observed the shift in the critical points
and compared their simulation results to those calculated by the method of Zarragoicoechea & Kuz (2004).
Pitakbunkate et al. (2016) used MUSIC, which is a molecular simulation software developed by Chempath et
al. (2013). [In our study, we use RASPA developed by the same research group. RASPA is a more advanced
tool using the most recent methods available in literature (Dubbeldam et al. 2016).]
Figure 1 presents the phase envelopes of a binary mixture (30.02% methane and 69.98% ethane) in 5
nanometer pore confinement computed by Pitakbunkate et al. (2016) from molecular simulation and by the
equilibrium thermodynamics algorithm provided by Firincioglu et al. (2013). Also, shown in Figure 1 is the
phase envelope computed from the Peng-Robinson EOS under bulk conditions. In Figure 1, we observe that
the GCMC simulations of Pitakbunkate et al. (2016) show a large shift in critical point but a low bubble-
SPE-187163-MS 3

point suppression; on the other hand, equilibrium thermodynamics calculations used by Firincioglu et al.
(2013) do not consider a shift in the critical point but yield a much larger shift of the bubble-point curve
than the molecular simulations.

Figure 1—Comparison of phase behavior of the binary mixture (30.02% methane and 69.98%
ethane) in bulk and confined environment. EOS with capillary force does not give roots for volume
below 15 Fahrenheit, that’s why we observe that yellow line starts from 15 F. Moreover, Equilibrium
thermodynamics does not give a proper dew point line; hence, only the bubble point line is put on the graph.

To explain the large discrepancies observed in Figure 1, we studied the phase envelope of pure methane in
2-nm confinement using RASPA with the μVT (constant chemical potential, constant volume and constant
temperature) ensembles. The μVT ensembles are very useful to observe adsorption and to determine the
phase transition point of confined fluids. In μVT simulations, it is assumed that the ensemble is connected
to a particle bath with infinite number of particles, which can move around until the system reaches an
equilibrium. In other words, at fixed chemical potential, the density of the system fluctuates by insertion or
deletion of particles until equilibrium condition is satisfied.
The GCMC simulation is a well-established technique to model the molecular movements in confined
environments. Ismail et al. (2014) studied methane and n-butane adsorption in confined environment and
showed that the GCMC technique yields consistent results with experiments below 1,000 psi. Although the
GCMC simulation is considered as an accurate and powerful method, it has some deficiencies. Having a
high-density fluid such as liquid phase of a confined hydrocarbon, insertion of a particle into the ensemble
is rejected with high probability because it is hard to find a large enough cavity for particle insertion (Yau
et al. 1994). To alleviate this problem, a cavity biased grand canonical Monte Carlo method is developed
by Mezei et al. (1980). RASPA uses this method for μVT ensembles to increase the accuracy.

Methodology
To imitate the confinement in a nanopore, we used two graphite sheets which are separated by a distance
equal to the size of the pore. Surfaces of shale pores are likely to have a complex mineralogical composition;
therefore, to simplify the calculations, we assumed to consider hydrocarbon fluids in kerogen pores, which
could be better represented by the graphite sheets.
Recently, Pitakbunkate et al. (2016) showed that, as expected, when the pore size increases, the interaction
between the wall surface and fluid particles decreases. Therefore, we observe less change in phase envelope
due to molecular interactions above 10 nanometers separation and the shift of the bubble-point and dew-
point lines is dominated by the capillary pressure (Firincioglu et al. 2012, Teklu, et al. 2014, Jin &
Firoozabadi 2016, Sandoval et al. 2015, Didar & Akkutlu 2013).
In a recent study, Takemura & Kitao (2007) noted that differences in the size of the simulation box
significantly affect the translational diffusion of the protein molecules in molecular simulation. In our study,
4 SPE-187163-MS

therefore; to increase the resolution of our simulations, we used large graphite slabs (100.54 Å by 174.141
Å) with 2 nanometer separation Figure 2. Although using large graphite slabs increases the resolution of
molecular simulation, having a larger volume requires larger number of molecules to reach the equilibrium
and causes a dramatic increase in the computational time.

Figure 2—Graphite Sheet Structure

The chemical potential of the particle bath is calculated by an EOS and used as input in the GCMC
simulation (Frenkel & Smit 2002). For a given temperature and pressure, EOS gives the fugacity coefficient
and the chemical potential is computed from (Koretsky 2013)
(4)

Lennard Jones Potential

To calculate the interaction between two particles, 12-6 Lennard-Jones model (Lennard-Jones 1931) is used
in the simulations. The potential energy is calculated by

(5)

which is a summation of the repulsive and attractive forces shown in Figure 3. To reduce the computational
time, the potential is calculated for particles having a separation larger than a specific cut off distance (Allen
& Tildesley 1989). Because this truncation may be significant, the following expressions must be used to
add a tail correction to the potential:

Figure 3—Change in Lennard-Jones Potential with changing distance between two particles (retrieved from Jiang, 2014)

(6)

and
SPE-187163-MS 5

(7)

where ε = depth of potential well, r = distance between particles, ρ = density of particles and rc= cut off
distance. As a force field, we used transferable potentials for phase equilibria (TraPPE) which gives the
Lennard Jones Potential parameters of pseudo atoms presented in Table 1 (Martin & Siepmann 1998). To
calculate the interaction between two different particles, we apply the following Lorentz-Berthelot mixing
rules (Schnabel, T. 2008)

(8)

and
(9)

Table 1—TraPPE Force Field Parameters (Martin & Siepmann 1998)

Pseudo atom ϵ/kB[K] σ[Å]

CH4 148 3.73

CH3 (ethane) 98 3.75

CH3 (n-alkane) 98 3.75

CH2 46 3.95

C 28 3.40

Particle Movements in GCMC Simulation

In the GCMC molecular simulation, four types of molecular movements are considered: namely, rotation,
translation, deletion, and insertion. Orientation of non-spherical molecules, configuration of molecules
in the simulation box, and number of particles have an effect on the potential calculation. Figure 4
is an illustration of these movements. For each movement, the probability of the molecules’ changing
configuration in the simulation box is determined by the Metropolis algorithm (Chib & Greenberg 1995).
For the accuracy of the Monte Carlo simulation, the ratio of the accepted and rejected moves must be around
1, which requires the appropriate selection of the rotation angle and displacement rate. The workflow for
each motion is well documented by Ismail et al. (2014) and Frenkel & Smit (2002). Below, we provide a
summary.

Figure 4—Summary of Molecular Movements in Grand Canonical Monte Carlo Simulation.

Deletion and insertion continues until "μparticle bath = μconfinement" condition is satisfied
6 SPE-187163-MS

To start the simulations, first, a random particle is selected and displaced by Δx. A random number is
generated and checked if the displacement is larger than the acceptable probability of the movement. The
same algorithm is used for rotational motion. After giving rotation to a particle, its acceptable probability
is calculated by
(10)
Particle insertion and deletion probabilities are also calculated, respectively, by

(11)

and

(12)

where, Λ is De Broglie Wave-length, U is Lennard-Jones Potential, N is number of particles, μ is chemical

potential, r represents position of a particle and β is reciprocal temperature. If the insertion condition (Eq.
11) is accepted, we insert a particle into a random location. Similarly, if the deletion condition (Eq. 12) is
accepted, we remove the random particle from the confined environment.
As illustrated in Figure 4, the particle movement and the deletion and insertion of the particles in the
GCMC simulation continues until the chemical potential of the simulation box and the chemical potential of
the particle bath reach an equilibrium. The graphite sheet which is represented as simulation box in Figure
4 is assumed to be infinitely long. To represent this situation, periodic boundary conditions are applied in
directions at which we do not have a pore wall.

Analyzing Results
The GCMC simulation gives the density of the phases simulated at a given temperature and pressure.
The density data is important to determine the phase transition points. For a gas or liquid phase, as the
pressure increases at constant temperature, the density increases smoothly with pressure until a jump in the
density is observed as an indication of phase transition. This phase transition point is clearly observed at
lower temperatures; however, as the temperature increases and approaches to the critical point, molecular
simulations fail to determine the phase transition precisely and the density output of the simulations starts
showing fluctuations. As noted earlier, using a larger simulation box, the precision may be increased at the
cost of drastically increased computational time.

Results and Discussion

To check the validity of the molecular simulation used in this study, first we run a simulation for pure
methane under bulk conditions and compare the results to those from Peng-Robinson EOS. Initially, we
used a simulation box of 30x30x30 Å with periodic boundary condition. As shown in Figure 5, although we
obtained a continuous phase envelope trend below 173° K, when the temperature increased above 173° K,
molecular simulation did not converge very well. To overcome this problem, the simulation box size was
increased to 70x70x70 Å and the results shown in Figure 6 were obtained. Having a bigger simulation box
enabled accurate calculations until 182° K but did not completely fix the problem when approaching the
critical point. Due to the limitations of calculation time and the sufficient evidence provided by Figs. 5 and
6, we did not further increase the simulation box size.
SPE-187163-MS 7

Figure 5—(30×30×30 Å) Density vs. pressure graph of methane in bulk condition. We observe
a good match for initial temperature values; however, as the temperature approaches
critical point, due to the simulation box size, the accuracy of the simulation decreases

Figure 6—(70×70×70 Å) After Increasing the simulation box size, we get better data until 182 Kelvin but above this
temperature, because simulations does not work properly near critical point, we did not observe phase transition

Figure 7 shows the results in Figure 6 (molecular simulation for pure methane in bulk conditions with
increased simulation box size) in the form of a P-T diagram. For comparison purposes, we also show the bulk
P-T diagram computed from the Peng-Robinson EOS and the bulk critical point. As discussed earlier, we
obtain a well-defined phase-separation line until 182° K. Above 182° K, phase transition cannot be predicted
accurately because of the size of the simulation box and the proximity to the critical point. However,
our molecular simulations follow the EOS results and extrapolation connects the simulation results to the
critical point. This discussion proves the point that increasing the size of the simulation box, the accuracy
of molecular simulation can be increased. We now proceed to evaluate the consequences of the size of the
simulation box on the estimation of critical point from molecular simulations.

Figure 7—Comparison of Molecular Simulation and Equation of State for Pure Methane. "molecular simulation
1" series used 30×30×30 Å simulation box and "molecular simulation 2" series used 70×70×70 Å simulation box.
One can see that above 173 K 30×30×30 Å box size does not give a phase transition point. Because near critical
point the simulation does not converge, we extrapolated the results to the bulk critical point by a dotted line
8 SPE-187163-MS

Pitakbunkate et al. (2016) analyzed the density data from their GCMC simulations to predict the phase
behavior of methane in 5-nm confinement using a simulation box with 42.53x41.90x50 Å. They used the
fact that a jump would be observed in the fluid density when the phase change occurred as a result of the
pressure change at constant temperature. When they could no longer see a jump in the density, they took this
pressure as the critical point for the given temperature. Based on this approach, in Figure 8, they predicted a
significant shift of the critical temperature of methane in 5-nm confinement to 175° K. They also presented
their results in the form of critical temperature and pressure shift as a function of the graphite slab separation
Figure 9. As we demonstrated in Figure 5 through Figure 7, however, such a critical point determination
could be erroneously made due to the inappropriate size of the simulation box (it is reasonable to expect
that the limit of the simulation box size; that is, the limit of the simulation time, would be reached before
the critical point is accurately observed).

Figure 8—Molecular simulation of phase behavior of Methane in 5-nm confinement (retrieved from Pitakbunkate et al. 2016)

Figure 9—The relationship between the graphite sheet separation and the critical
pressure & temperature of methane (retrieved from Pitakbunkate et al. 2016)

In Figure 10, we present the density of methane in 2-nm confinement as a function of pressure at fixed
temperature computed from our GCMC simulations. We have selected the 2-nm separation to increase the
speed of computations while using a larger simulation box size (100.54x174.141x20 Å) for more accurate
SPE-187163-MS 9

results. We recall that, in Figure 7, Pitakbunkate et al. (2016) predicted the critical temperature around 175°
K for methane in 5-nm confinement. Based on their results in Figure 8, the critical temperature should have
been at 130° K for 2-nm confinement, which is even lower than 175° K. The difference between our results
and those of Pitakbunkate et al. (2016) should be attributed to the different simulation-box sizes used in the
two studies (theoretically, our results should be more accurate as we use a larger simulation box to improve
the accuracy of molecular simulations).

Figure 10—Density vs. pressure graph of pure methane in 2-nm confinement. Until 182° K, we observe phase transition points.
However, above 182° K, the simulation does not give any phase transition data due to the proximity to the critical point.

For clarification, we also comment on the accuracy of the approach used by Pitakbunkate et al. (2016) to
select the critical point from density vs. pressure plots. In Figure 8, Pitakbunkate et al. (2016) select 175° K
as the critical temperature based on the assumption that at this temperature, the jump in the density turns into
a continuous change as a function of pressure. Based on their figure, it can even be argued that the continuous
change in the density may start at as low as 172° K. If we combine these results with the conclusions from
Figure 9, we should expect a lower critical temperature for 2-nm confinement (Figure 9 shows critical
temperature as 130° K at 2 nm). Returning back to our results in Figure 10 for 2-nm confinement, if we
applied the approach used by Pitakbunkate et al. (2016), we would select a critical temperature higher than
174° K. This is in contradiction with the predictions of Pitakbunkate et al. (2016).
As a final remark, we also comment on the differences between the results of the critical-point-shift
method of Zarragoicoechea & Kuz (2004) and our GCMC simulations shown in Figure 11. The results in
Figure 11 indicate that the critical-point-shift method creates a significant deviation from the bulk phase
behavior and the deviation increases as the confinement decreases. Our GCMC simulations for 2-nm
confinement, on the other hand, yield a much smaller deviation from the bulk phase behavior and appear to
be extrapolating to the bulk critical point. Zarragoicoechea & Kuz (2004) developed their model using a fluid
that does not interact with the walls; on the other hand, in our study, RASPA takes the fluid wall interaction
into account, which may be the reason for such a difference. It is known that fluid-wall interactions, attractive
or repulsive, change the critical properties of confined fluids (Votyakov et al., 1999).
10 SPE-187163-MS

Figure 11—Comparison of P-T diagrams captured by our molecular simulation and critical point shift algorithm of
Zarragoicoechea & Kuz (2004). As indicated by the dashed line, the extrapolation of the 2 nm-results connects with the
bulk critical point. The purple line is the results of Pitakbunkate et al. (2016) for 5-nm and is indented for comparison.

For comparison purposes, the results of Pitakbunkate et al. (2016) for 5-nm confinement are also shown
in Figure 11. As expected from the discussions of this paper, due to the smaller size of their simulation box,
their results do not show a consistent trend.

Conclusions
In this study, we focused on the molecular simulation of phase behavior in confinement and addressed the
differences of the results reported in the literature. Using bulk conditions, we showed that the accuracy of
molecular simulations was a strong function of the size of the simulation box and larger box sizes improved
the accuracy but also prohibitively increased the computational times. The results of this paper also indicated
that the large critical-point shifts alluded in various studies may be a result of the accuracy of the method
used and the approach used to determine the critical point. Considering the discussions given here, it is
suggested that the results of the existing phase-behavior studies should be used with caution and more
qualitatively than quantitatively.

Acknowledgements
The authors would like to thank to Dr. Snurr and Dr. Dubbledam for granting access to RASPA, Dr. Sum for
his help on molecular simulations, Dr. Kazemi and Dr. Kaiser for their help to use the computers in Golden
High Performance Computer Laboratory. This research has been conducted under UREP (Unconventional
Reservoir Engineering Project) at the Colorado School of Mines. Financial support of the UREP members
is acknowledged. Parts of this work will be used for the partial fulfillment of Y. B. Coskuner’s MS Degree
requirements at the Colorado School of Mines.

Nomenclature
CBGCMC = Cavity Biased Grand Canonical Monte Carlo
EOS = Equation of State
f = Fugacity Coefficient
GCMC = Grand Canonical Monte Carlo
kB = Boltzmann Constant, J·K−1
SPE-187163-MS 11

Nm = Nanometer
N = Number of Particles
P = Pressure, Pascal
Pcb = Critical Pressure at Bulk Condition
r = Distance Between Two Particles, Angstrom
rp = Pore Throat Radius, Å
T = Temperature, Kelvin
Tcb = Critical Temperature at Bulk Condition
U = Lennard Jones Potential
U tail = Lennard Jones Potential Tail Correction
V = Volume
μ = Chemical Potential
μparticle bath = Chemical Potential of Fluid in a Particle Bath
μconfinement = Chemical Potential of Confined Fluid
Å = Angstrom
ϵ = Depth of the Potential Well
σ = Finite Distance Between Particles at Which Inter Particle Potential is Zero
σLJ = Lennard Jones Size Parameter, Å
Λ = Thermal de Broglie Wave-Length
= Relative Critical Temperature Shift, Dimensionless
= Relative Critical Pressure Shift, Dimensionless

References
Al Ismail, M., & Horne, R. (2014). Modeling Adsorption of Gases in Nanoscale Pores Using Grand Canonical Monte
Carlo Simulation Techniques. Paper SPE-170948, presented at the SPE Annual Technical Conference & Exhibition,
27-29 October, Amsterdam, The Netherlands. http://dx.doi.org/10.2118/170948-ms
Allen, M., & Tildesley, D. (1989). Computer simulation of liquids (2nd ed.). Oxford [England]: Clarendon Press.
Carl Y. H. Jiang,(2014) A New Approach to Model Adsorption in Heterogeneous Phase System with Monte Carlo Method,
American Journal of Materials Science, Vol. 4 No. 1, 25–38. 10.5923/j.materials.20140401.05.
Chempath, S., Düren, T., Sarkisov, L., & Snurr, R. (2013). Experiences with the publicly available multipurpose simulation
code, Music. Molecular Simulation, 39 (14–15), 1223–1232. http://dx.doi.org/10.1080/08927022.2013.819103
Chib, S., & Greenberg, E. (1995). Understanding the Metropolis-Hastings Algorithm. The American Statistician, 49 (4),
327. http://dx.doi.org/10.2307/2684568
Dubbeldam, D., Calero, S., Ellis, D., & Snurr, R. (2015). RASPA: molecular simulation software for
adsorption and diffusion in flexible nanoporous materials. Molecular Simulation, 42 (2), 81–101. http://
dx.doi.org/10.1080/08927022.2015.1010082
Firincioglu, T., Ozkan, E., & Ozgen, C. (2012). Thermodynamics of Multiphase Flow in Unconventional Liquids-Rich
Reservoirs. Paper SPE-159869, presented at the SPE Annual Technical Conference and Exhibition, 8-10 October, San
Antonio, Texas, USA http://dx.doi.org/10.2118/159869-ms
Firincioglu, T., Ozgen, C., & Ozkan, E. (2013). An Excess-Bubble-Point-Suppression Correlation for Black Oil
Simulation of Nano-Porous Unconventional Oil Reservoirs. Paper SPE-166459, presented at the SPE Annual Technical
Conference & Exhibition, 30 September-2 October, New Orleans, Louisiana, USA http://dx.doi.org/10.2118/166459-
MS
Frenkel, D., & Smit, B. (2012). Understanding molecular simulation (1st ed.). San Diego, CA, USA: Academic Press.
Honarpour, M. M., Nagarajan, N. R., Orangi, A., Arasteh, F., and Yao, Z. (2012). Characterization of Critical Fluid, Rock,
and Rock-Fluid Properties-Impact on Reservoir Performance of Liquid-rich Shales, paper SPE 158042 presented at
the SPE Annual Technical Conference and Exhibition held in San Antonio, Texas, USA, 8-10 October 2012
Lennard-Jones, J. (1931). Cohesion. Proceedings of The Physical Society, 43 (5), 461–482. http://
dx.doi.org/10.1088/0959-5309/43/5/301
Jin, Z., & Firoozabadi, A. (2016). Thermodynamic Modeling of Phase Behavior in Shale Media. SPE Journal, 21 (01),
190–207. http://dx.doi.org/10.2118/176015-PA
Koretsky, M. (2013). Engineering and chemical thermodynamics. Hoboken, NJ: Wiley.
12 SPE-187163-MS

Loucks, R., Reed, R., Ruppel, S., & Jarvie, D. (2009). Morphology, Genesis, and Distribution of Nanometer-Scale Pores
in Siliceous Mudstones of the Mississippian Barnett Shale. Journal Of Sedimentary Research, 79 (12), 848–861. http://
dx.doi.org/10.2110/jsr.2009.092
Makimura, D., Kunieda, M., Liang, Y., Matsuoka, T., Takahashi, S., & Okabe, H. (2011). Application of Molecular
Simulations to CO2-EOR: Phase-Equilibria and Interfacial Phenomena. Paper SPE-14774, presented at International
Petroleum Technology Conference, 15-17 November, Bangkok, Thailand http://dx.doi.org/10.2523/iptc-14774-MS
Martin, M. G. & Siepmann, J. I. (1998). Transferable Potentials for Phase Equilibria. 1. United-Atom Description of n-
Alkanes, The Journal of Physical Chemistry B, 102, 2569–2577. 10.1021/jp972543
Meyer, V., Creux, P., Graciaa, A., Franco, F., & Luck, F. (2009). Non-classical Nucleation Model for Cold Production of
Heavy Oil. Journal Of Canadian Petroleum Technology, 48 (04), 49–56. http://dx.doi.org/10.2118/09-04-49
Mezei, M. (1980). A cavity-biased (T, V, μ) Monte Carlo method for the computer simulation of fluids. Molecular Physics,
40 (4), 901–906. http://dx.doi.org/10.1080/00268978000101971
Pitakbunkate, T., Balbuena, P., Moridis, G., & Blasingame, T. (2016). Effect of Confinement on Pressure/
Volume/Temperature Properties of Hydrocarbons in Shale Reservoirs. SPE Journal, 21 (02), 621–634. http://
dx.doi.org/10.2118/170685-PA
Parsa, E., Yin, X, Ozkan, E. 2015. Direct Observation of the Impact of Nanopore Confinement on Petroleum Gas
Condensation. Paper SPE 175118, presented at the at the SPE Annual Technical Conference and Exhibition held in
Houston, Texas, USA, 28-30 September 2015.
Rahmani Didar, B., & Akkutlu, I. (2013). Pore-Size Dependence of Fluid Phase Behavior and the Impact on Shale Gas
Reserves. Paper URTeC-1624453 presented at Unconventional Resources Technology Conference, Denver, Colorado,
12-14 August 2013. http://dx.doi.org/10.1190/urtec2013-183
Sandoval, D., Yan, W., Michelsen, M., & Stenby, E. (2015). Phase Envelope Calculations for Reservoir Fluids in the
Presence of Capillary Pressure. Paper SPE-175110 presented at SPE Annual Technical Conference And Exhibition,
Houston, Texas, USA, 28-30 September, http://dx.doi.org/10.2118/175110-MS
Sapmanee, K. (2011). Effects of Pore proximity on Behavior and Production Prediction of Gas/Condensate. Master Thesis,
Mewbourne School of Petroleum and Geological Engineering, University of Oklahoma, 2011.
Schnabel, T. (2008). Molecular modeling and simulation of hydrogen bonding pure fluids and mixtures (1st ed.). Berlin:
Logos-Verl.
Takemura, K., & Kitao, A. (2007). Effects of Water Model and Simulation Box Size on Protein Diffusional Motions. The
Journal Of Physical Chemistry B, 111 (41), 11870–11872. http://dx.doi.org/10.1021/jp0756247
Teklu, T., Alharthy, N., Kazemi, H., Yin, X., Graves, R., & AlSumaiti, A. (2014). Phase Behavior and Minimum
Miscibility Pressure in Nanopores. Paper 168865-PA, SPE Reservoir Evaluation & Engineering, 17 (03), 396–403.
http://dx.doi.org/10.2118/168865-pa
Travalloni, L., Castier, M., Tavares, F. W., Sandler, S. I. (2010). Critical behavior of pure confined fluids from an extension
of the van der Waals equation of state. J. of Supercritical Fluids 55 (2010) 455–461.
Wang, L., Parsa, E., Gao, Y., Ok, J., Neeves, K., Yin, X., & Ozkan, E. (2014). Experimental Study and Modeling of
the Effect of Nanoconfinement on Hydrocarbon Phase Behavior in Unconventional Reservoirs. Paper SPE-169581
presented at SPE Western North American And Rocky Mountain Joint Meeting. http://dx.doi.org/10.2118/169581-ms
Votyakov, E., Tovbin, Y., MacElroy, J., & Roche, A. (1999). A Theoretical Study of the Phase Diagrams of Simple Fluids
Confined within Narrow Pores†. Langmuir, 15 (18), 5713–5721. http://dx.doi.org/10.1021/la9813634
Yau, D., Liem, S., & Chan, K. (1994). A contact cavity‐biased method for grand canonical Monte Carlo simulations. The
Journal of Chemical Physics, 101 (9), 7918–7924. http://dx.doi.org/10.1063/1.468218
Zarragoicoechea, G., & Kuz, V. (2004). Critical shift of a confined fluid in a nanopore. Fluid Phase Equilibria, 220 (1),
7–9. http://dx.doi.org/10.1016/j.fluid.2004.02.014