Molecular dynamics simulations of volumetric thermophysical properties of natural gases 417

18
X

Molecular dynamics simulations of volumetric

thermophysical properties of natural gases
Santiago Aparicio1 and Mert Atilhan2
1Department
of Chemistry. University of Burgos
Spain
2Department of Chemical Engineering. Qatar University

Qatar

1. Introduction
The accurate knowledge of thermophysical properties of natural gas mixtures is of great
importance for practical purposes for the gas industry from exploration stages to final
customer use (Jaescke et al., 2002; Wagner & Kleinrahm, 2004; Gallagher, 2006). Two main
properties are required by the oil and natural gas industry: i) phase equilibria and ii)
pressure–density–temperature (PρT) data. The large impact of PρT, volumetric, data on
production, processing and transportation of natural gas is well–known (Hall & Holste,
1990; Husain, 1993; Wagner & Kleinrahm, 2004; Bluvshtein, 2007). Although the required
accuracy for the considered properties varies depending on the purpose for which they are
used (performance analysis or gas sales, Mokhatab et al. 2006), a high degree of accuracy is
frequently required for most of the applications. Thermophysical properties of natural gas
systems must be accurately known for national and international custody transfer
considering that flowmeters measurements applied to custody transfer are used to buy and
sell natural gas between pipeline companies. It should be remarked that custody transfer
operations relies on accurate density data, and thus, the economic impact of accurate density
measurements is very large both for the calculation of energy content of natural gas and for
the flow rate obtained usually from orifice meters (which are used in about 90 % of the
metering stations). Inaccuracies in density measurements may lead to very large economical
losses or profits for gas producers or buyers, and thus, to conflicts between companies or
even countries. Therefore, density values with accuracies below the 0.1 % level are
commonly required.
Experimental measurements using highly accurate densimeters, mainly magnetic
suspension type, may lead to measurements with a level of uncertainty below the required
0.1 % limit (Wagner & Kleinrahm, 2004; Patil et al., 2007), and thus, this would be the best
and most reliable option to obtain density data. Nevertheless, natural gas composition
changes remarkably from one to other reservoir (because of age and deep), moreover, the
advances in exploration and drilling technologies allow to explore and produce from non-
conventional reservoirs (such as ultra-deep ones) which characteristics are very different to
the traditional ones. Therefore, considering that to carry out experimental density

www.intechopen.com
418 Natural Gas

measurements is very costly, both in time and economical resources, and having in mind
that it is not possible to measure all possible compositions of natural gas in the wide
pressure–temperature ranges required for production, transportation and processing
purposes, thus, accurate predictive models are required by the gas industry. The current
industrial standard model for custody transfer purposes is AGA8-DC92 (Starling & Savidge,
1992); this is a complex multiparametric equation of state which has been used for years as
an international standard. Nevertheless, in spite of the common use of AGA8-DC92 in the
gas industry it presents several problems. Any predictive model must be validated with
reliable experimental data obtained on a limited number of samples that have well defined
compositions. None of the samples used for the AGA8-DC92 validation are in the extended
region, and thus, the application of this model for mixtures with large concentrations of
CO2, N2, or long alkanes (as the natural gases from non-conventional reservoirs; Babusiaux,
2004) may be problematic. Moreover, the accuracy of AGA8-DC92 model decreases
remarkably on going from region 1 (265 to 335 K, 0 to 12 MPa, deviations < 0.1 %) to regions
2 ( 12 to 17 MPa, 211 to 394 K, deviation < 0.3 %) and 3 (17 to 70 MPa, 144 to 477 K, deviation
< 0.5 %) of model validity, and thus the model performance decreases remarkably when
pressure and temperature increases. Recent studies have reported larger deviations than the
claimed ones for AGA8-DC92 (Patil et al., 2007). Therefore, two main conclusions may be
extracted: i) performance of AGA8-DC92 has to be systematically analyzed using a
collection of carefully selected mixtures and ii) probably a new model has to be proposed as
international standard for custody transfer and transmission purposes.
Properties of natural gases, including PρT behavior, are a reflection of intermolecular forces
rising by the simultaneous presence of very different types of molecules in these complex
multicomponent fluids, and their evolution with pressure and temperature. Therefore
macroscopic properties of these fluids, such as density, are a consequence of the microscopic
structure and behavior of the involved molecules in the considered phases. These
relationships between microscopic and macroscopic behavior may be developed through
the principles of statistical mechanics if an accurate knowledge of the forces acting between
the involved molecules were available. Hence, if we want to develop reliable and accurate
models to predict natural gas density, we should get a deeper insight into the microscopic
structure for complex gas mixtures and their relationships with PρT behavior. For this
purpose, computational chemistry methods, mainly classical molecular dynamics and
Monte Carlo approaches, are very useful tools allowing i) to infer microscopic structural and
energetic features and ii) to predict macroscopic relevant properties, such as density, as a
function of pressure and temperature.
We report in this work a computational study in which the ability of classical molecular
dynamics simulation methods to predict PρT behavior of complex natural gas mixtures is
analyzed. Computational predictions are compared with available highly accurate
experimental density data. The possibility of using this approach to predict density values
with an acceptable degree of accuracy at moderate computational costs is analyzed, the
weaknesses and strengths of the method together with possible future directions are
considered.

www.intechopen.com
Molecular dynamics simulations of volumetric thermophysical properties of natural gases 419

2. Literature Review
A detailed analysis of the literature shows that the available studies on the use of molecular
modelling with predictive purposes for thermophysical properties of natural gas like
mixtures are scarce. We will not report here the available literature on the use of Gibbss
ensemble Monte Carlo methods for the prediction of phase equilibria, both for pure
compounds or for the involved binary mixtures, because this is not the object of this work.
Moreover, most of the studies use Monte Carlo approach instead of the classical molecular
dynamics methods proposed in this work. In this section, we will analyze the relevant
studies available in the open literature in which computational methods, both classical
molecular dynamics and Monte Carlo approaches, are used to predict thermophysical
properties of natural gas mixtures. Results of literature analysis are reported in a
chronological ordering.
Saager and Fischer, 1989, reported a study on NVT molecular dynamics simulations of PVT
and thermal properties of pure methane up to 1000 MPa.
Duan et al., 1992, reported a wide study in which PVT properties of pure methane are
predicted in the 273 – 2000 K and 100 – 20000 bar ranges using NVE molecular dynamics
simulation methods together with a united atom approach. Results show deviations within
the 1.5 % range.
Yoshida & Uematsu, 1996, published a study reporting the results for the prediction of PVT
properties of natural gases by Monte Carlo molecular simulation. They studied light
mixtures composed by methane, ethane, propane, CO2 and N2 in the conditions of natural
gas transportation in pipelines. They reported deviations lower than 1.5 % in pressure,
although density prediction analysis is not carried out in a straightforward manner.
Duan et al., 1996, used molecular dynamics to simulate the PVT properties of the ternary
mixture methane + CO2 + N2, leading to results with low deviations compared with the
general equation of state proposed by the authors.
Neubauer et al., 1999, reported a study in which NPT Monte Carlo method was applied for
the simulation of volumetric properties of natural gas mixtures, both in the single phase and
two – phase conditions. A united atom approach was used leading to density deviations up
to 5 % in the high pressure region, decreasing with decreasing pressure and increasing
temperature. These too large deviations, that obviously hinder the application of the
proposed method with purely predictive purposes, are a consequence of the poor
representation of the composition of the gas during the simulation.
Errington et al., 1999, reported a NPT Monte Carlo study on the properties of the n-alkanes
homologous series, developing a united atom intermolecular potential and showing its
validity for alkanes up to C78. Saturated densities are predicted within a 2 % limit and
liquid densities for the longer n-alkanes to 1 %.
Dysthe et al., 1999, used equilibrium molecular dynamics together with the Green-Kubo
formalism to predict transport coefficients of multicomponent natural gas like mixtures
including alkanes up to C4 , N2 and He, both in the gas and liquid phases. Simulations were
performed in the NVT ensemble with a united atom approach for alkanes leading to
viscosity deviations of 7 % and 11 % for the gaseous and liquid states, respectively.
Escobedo and Chen, 2001, developed a NPT Monte Carlo study for the prediction of Joule –
Thomson inversion curves for several fluids, including pure methane and a gas condensate
mixture (with alkanes up to C7), leading to reliable predictions.

www.intechopen.com
420 Natural Gas

Lagache et al., 2001, reported a study in which NPT Monte Carlo simulations were
performed to compute second order derivatives of the Gibbs energy for simple alkanes up
to butane and for the methane – ethane binary mixture. The authors used a united atom
potential. Results show that predicted data are in fair agreement with experimental values,
even for complex properties such as Joule – Thomson coefficient for which deviations below
the 10 % are obtained.
Ungerer, 2003, reported a wide review in which the use of Monte Carlo and molecular
dynamics methods is analyzed for several relevant fields in the petroleum and gas industry.
Lagache et al., 2004, reported a study in which NPT Monte Carlo was used for the prediction
of density and other relevant properties (thermal expansivity, isothermal compressibility,
isobaric heat capacity and Joule – Thomson coefficient) of methane, ethane and two
mixtures including heavy components up to 35 carbon atoms. The authors use a united
atom approach to model the involved molecules. Density, and compressibility factor, show
deviations up to 3 %, whereas for the remaining studied properties deviations are lower
than 10 %. The authors show the importance of the mixtures characterization and
representation to obtain accurate results.
Ungerer et al., 2004, reported a NPT Monte Carlo study on the prediction of relevant
properties, including density, for H2S – rich gases, showing that although reported results
provide valuable information on the understanding of the complex mixed fluids, this
computational approach does not lead to the high accuracy required for process design
purposes for the studied acid gases.
Ungerer et al. (Ungerer et al., 2006; Ungerer et al., 2006) reported a wide and useful study on
the application of Monte Carlo methods for oil and gas production and processing purposes.
They showed some of the results previously reported by Lagache et al., 2004, and claimed
again to the remarkable importance of an adequate characterization of studied mixtures to
obtain accurate results for single phase and equilibria properties.
Bessieres et al., 2006, reported a study in which NPT Monte Carlo simulations were used to
predict the Joule–Thomson inversion curve of pure methane using and united atom
approach. Results show very accurate predictions with deviations below the 1 % limit.
Vrabec et al., 2007, carried out a study on the performance of molecular simulation methods
for the prediction of Joule–Thomson inversion curves for light natural gas mixtures.
Reported results show deviations usually within the 5 % range, larger for high
temperatures, but being competitive with most state-of-the-art EOS in predicting Joule -
Thomson inversion curves.
In a review work, Ungerer et al., 2007, analyzed the weaknesses and strengths of using
molecular simulation for the predication of thermophysical properties of complex fluids,
including natural gas like mixtures. The authors claim that one of the main limitations of the
computational approach is the availability of potentials and force field parameters tested in
wide pressure–temperature ranges.
The main conclusions obtained from the analysis of the available open literature are:
i) Monte Carlo approach is used in an exclusive basis when thermophysical properties of
natural gas mixtures are under study.
ii) United atom potentials are the most common option.
iii) Studies in wide pressure–temperature ranges and for multicomponent mixtures are very
scarce, and thus the performance of the proposed approaches is not clear.

www.intechopen.com
Molecular dynamics simulations of volumetric thermophysical properties of natural gases 421

Thus, in this work we report results using the molecular dynamics approach, all-atoms
potential, and analysis in wide P-T ranges for selected pure and mixed fluids. This
methodology was considered because of the absence of similar studies in the open literature
to analyze its validity for natural gas industry production, transportation and processing
purposes.

3. Computational Methods
Classical molecular dynamics simulations were carried out using the TINKER molecular
modeling package (Ponder, 2004). All simulations were performed in the NPT ensemble; the
Nosé–Hoover method (Hoover, 1985) was used to control the temperature and pressure of
the simulation system. The motion equations were solved using the Verlet Leapfrog
integration algorithm (Allen & Tildesley, 1989). Long-range electrostatic interactions were
treated with the smooth particle mesh Ewald method (Essmann, 1995).
The simulated systems consist of cubic boxes, with the number of molecules and
compositions, for mixed fluids, reported in Table 1, to which periodic boundary conditions
were applied in the three directions to simulate an infinite system. The composition of the
mixed fluid selected to test the performance of the computational approach for
multicomponent natural gas–like mixtures resembles the one reported by Patil et al., 2007,
for which very accurate and reliable density data obtained through magnetic suspension
densitometers are reported. The number of molecules used for the simulation of pure
compounds was selected to obtain systems with 4000 – 5000 total atoms leading to
reasonable computing times. For the mixture we have selected a total number of molecules
(1000) that allow the representation of all the involved species, even those that appear at
very low mole fraction but which effect on the mixed fluid behavior is important.
The simulations were performed using a cutoff radius of L/2 Å for the non bonded
interactions, L being the initial box side. Initial boxes generated using the PACKMOL
program (Martínez & Martínez, 2005) were minimized according to the MINIMIZE program
in TINKER package to a 0.01 kcal mol-1 Å-1 rms gradient. Long simulation times are needed
for computing the properties of these fluids and procedures have to be designed carefully to
avoid the presence of local minima. Therefore several heating and quenching steps in the
NVT ensemble up to 600 K were performed after which a 100 ps NVT equilibration
molecular dynamics simulation was run at the studied temperature; finally, from the output
NVT simulation configuration, a run of 1 ns (time step 1 fs) in the NPT ensemble at the
studied pressure and temperature was run, from which the first 0.5 ns were used to ensure
equilibration (checked through constant energy) and the remaining 0.5 ns for data collection.
n-Alkanes were described according to the so called Optimized Potential for Liquid
Simulations (all atom version) OPLS–AA (Jorgensen et al., 1996), eqs. 1-4. Parameters for CO2
and N2 were obtained from the literature (Shi & Maginn, 2008; Lagache et al., 2005), with
bonds, angles and non-bonded interactions treated using eqs. 1,2 and 4. A Lennard- Jones 6-
12 potential, eq. 4, was used to describe the interactions between sites which are separated
more than three bonds, if they are in the same molecule (intramolecular interactions), and
for interactions between sites belonging to different molecules (intermolecular interactions).
Non-bonded interactions between 1-4 sites are scaled with a 0.5 factor. Lorentz-Berthelot
mixing rules are applied for Lennard–Jones terms between different sites, eqs. 5-6. The used
forcefield parameters are reported in Table 2.

www.intechopen.com
422 Natural Gas

Ebond   k r  r 
2 (1)

 k θ  θ 
r eq
bonds

E angle 
2 (2)
θ eq
angles

Ebond  1  cos    2 1  cos 2   3 1  cos 3 

V1 V V (3)

2 2 2
 qiq j e   ij  ij6 
Enonbonded     4 ij  12  6 
2 12 (4)

j 
 rij rij 
 

rij  
i  j
i

 ij 
(5)

2
 ij   i  j (6)

Thermophysical properties were obtained from the statistical analysis of the results in the
last 0.5 ns together with fluctuation formulae for the NPT ensemble, eqs. 7-9. We have
selected five relevant properties to analyze the performance of the proposed computational
approach (Lagache et al., 2004): density (ρ, because of its importance for natural gas
engineering purposes), thermal expansion coefficient (αP), isothermal compressibility (βT),
isobaric heat capacity (CP) and Joule – Thomson coefficient (μ, calculated according to eq.
10).

P  V   K    PV 
1 (7)

k B T 2V

T  V 2
1 (8)

k BTV

CP   K    PV 2
1 (9)
2
k BT


v
T P  1
(10)

CP

In eqs. 7-9, <> stands for ensemble averages, K for kinetic energy, P for potential energy, P
for pressure, T for temperature, V for volume and kB for the Boltzmann constant. The
notation δX (where X stands for any of the quantities reported in eqs. 7-9) denotes X-<X>. In
eq.10, v stands for the molar volume.

www.intechopen.com
Molecular dynamics simulations of volumetric thermophysical properties of natural gases 423

Pure fluids simulations Mixed fluid simulations

Experimental mole fraction Mole
Component N (Patil et al., 2007) N fraction
Methane 1000 0.90991 910 0.91000
Ethane 500 0.02949 29 0.02900
Propane 500 0.01513 15 0.01500
i-Butane 300 0.00755 8 0.00800
n-Butane 300 0.00755 8 0.00800
i-Pentane 300 0.00299 3 0.00300
n-Pentane 300 0.00304 3 0.00300
CO2 1000 0.02031 20 0.02000
N2 1000 0.00403 4 0.00400
Table 1. Molecular representation of studied systems used for molecular dynamic
simulations. N stands for the number of molecules used in the simulation. For mixed fluid
simulations the experimental compositions reported by Patil et al., 2007, are reported
together with the composition used in our simulations for comparative purposes

alkanes
req / Å kr / kcal mol-1
C-C 1.529 268.0
H-C 1.090 340.0
θeq / deg kθ / kcal mol-1
H-C-H 107.8 33.00
H-C-C 110.7 37.50
C-C-C 112.7 58.35
V1 / kcal mol-1 V2 / kcal mol-1 V3 / kcal mol-1
H-C-C-H 0.000 0.000 0.318
H-C-C-C 0.000 0.000 0.366
C-C-C-C 1.740 -0.157 0.279
q / e- /Å ε / kcal mol-1
C (in CH4) -0.240 3.500 0.066
C (in RCH3) -0.180 3.500 0.066
C (in R2CH2) -0.120 3.500 0.066
C (in R3CH3) -0.060 3.500 0.066
H 0.060 2.500 0.030
CO2
req / Å kr / kcal mol-1
C-O 1.160 1030.0
θeq / deg kθ / kcal mol-1
O-C-O 180.0 56.00
q / e- /Å ε / kcal mol-1
C 0.700 3.143 0.053
O -0.350 3.424 0.157
N2
req / Å
N-N 1.098
q / e- /Å ε / kcal mol-1
N 0.000 3.798 0.142
Table 2. Forcefield parameters used along this work obtained from the literature (Jorgensen
et al., 1996; Shi & Maginn, 2008; Lagache et al., 2004)

www.intechopen.com
424 Natural Gas

4. Molecular Dynamics Simulations Results

4.1 Methane
As a first proof of the ability of the computational approach proposed in this work to
reproduce accurately the properties of pure methane (the main component of natural gas
mixtures) we have calculated the phase equilibria diagram of this fluid. Although the
objective of this work is not devoted to the analysis of phase equilibria, these results may be
considered as a previous test of the model. Phase equilibria was calculated using Monte
Carlo method from the coupled–decoupled (Martin & Siepmann, 1999), dual cut-off (Vlugt
et al., 1998), configurational bias (Siepmann & Frenkel, 1992) simulations, using the MF-CPN
strategy (Martin & Frischknecht, 2006), which were performed using MCCCS Towhee code
in the NVT–Gibbs ensemble (Errington & Panagiotopoulos, 1998). Simulated phase
equilibria in comparison with literature experimental results are reported in Fig. 1. Critical
properties were calculated using the Ising scaling law (Smit and Williams, 1990) with a
scaling exponent of 0.32.

200
Tc = 190.17 K
ρc = 0.1722 g cm-3
190

180
T/K

170

160

150
0 0.1 0.2 0.3 0.4
/ g cm-3
Fig. 1. Phase equilibria for pure methane. Black line shows values obtained from the NIST
webbook reference data (Linstrom & Mallard, 2009), red points show values calculated in
this work using Monte Carlo simulations (together with error bars), black point shows
reference critical point from NIST webbook (Linstrom & Mallard, 2009), the calculated point
at the highest temperature shows the simulated critical point obtained from Monte Carlo
results and Ising scaling law (with numerical values reported in red within the Figure)

Results reported in Fig. 1 show a fair agreement between experimental reference data (NIST
webbook, Linstrom & Mallard, 2009) and simulated values. Critical temperature is just a
0.23 % lower and critical density a 5.53 % larger than the experimental values. Therefore the
proposed forcefield parameterization reproduces accurately the phase equilibria of pure
methane.

www.intechopen.com
Molecular dynamics simulations of volumetric thermophysical properties of natural gases 425

To obtain reliable density values from molecular dynamics simulations, the evolution of
calculated densities with simulation time was analyzed (together with energy values) to
assure that stable configurations are obtained and to check if longer simulation times are
required. Therefore in Fig. 2 we report some representative results of these analyses for pure
methane as a function of pressure and temperature, we report selected results for the lowest,
medium and larger densities studied.

385

380

375

250 K / 100 MPa

370

265

260
 / kg m-3

255

450 K / 100 MPa

250

95

90

85

450 K / 20 MPa
80
0 100 200 300 400 500
t / ps
Fig. 2. Representation of calculated densities for pure methane, ρ, obtained from NPT
molecular dynamics simulations as a function of simulation time. The results are obtained
for the last 500 ps of the simulations. Symbols: continuous lines show density values, dashed
lines average values obtained from the simulation and dotted line experimental values
obtained from NIST webbook reference data (Linstrom & Mallard, 2009)

Results reported in Fig. 2 show that density values converge rapidly, and thus no longer
simulation times are required (total simulation time is 1000 ps). In some cases, as it is
reported in Fig. 2 for 450 K / 20 MPa simulations (and usually appears for the lower
temperature simulations), density convergence is slower, nevertheless density also converge

www.intechopen.com
426 Natural Gas

for this temperature, and thus statistical analysis was performed with the data once
convergence is reached. Once density convergence is reached, fluctuations are in the ± 1.5 %
range of the average density value.
Calculated densities of pure methane are reported in Table 3 in comparison with available
reference data. Simulations are carried out in wide pressure–temperature ranges (250 to 450
K and 20 to 100 MPa), relevant for natural gas engineering purposes, to analyze the global
performance of the approach. These pressure–temperature ranges were used for all the
simulations along this work. Deviations between experimental and simulated densities are
in the 1 – 2 % range for the studied pressure – temperature range, and thus, slightly larger
than the accuracy required for many natural gas engineering purposes.

T/K P / MPa ρref / kg m-3 ρsim / kg m-3 standard deviation    sim 

 ref
ref
100
of ρsim / kg m-3
250 20 223.23 219.30 0.84 1.76
250 60 332.38 336.69 1.98 -1.30
250 100 373.51 379.78 2.01 -1.68
350 20 119.41 117.94 0.64 1.23
350 60 255.39 251.25 1.93 1.62
350 100 312.26 308.73 1.19 1.13
450 20 84.72 85.36 0.30 -0.76
450 60 202.23 198.83 1.19 1.68
450 100 264.73 257.96 1.10 2.56
Table 3. Comparison between reference, ρref (Linstrom & Mallard, 2009), and NPT molecular
dynamics simulated, ρsim, density values, for methane as a function of pressure and
temperature

T / K P / MPa 103 αP / K-1 103 βT / MPa-1 CP / J mol-1 K-1 μ / K MPa-1

ref sim ref sim ref sim ref sim
250 20 7.8345 7.05
29.943 27.72 66.246 64.42 1.0400 0.87
250 60 2.6297 2.934.1285 4.66 50.537 48.85 -0.32718 -0.26
250 100 1.8203 1.992.1330 2.40 48.036 47.05 -0.48724 -0.45
350 20 4.4143 4.85
46.945 46.09 50.698 49.73 1.4443 1.91
350 60 2.5420 2.557.6472 8.49 50.924 49.92 -0.13605 -0.14
350 100 1.7380 1.883.4090 3.58 49.162 48.84 -0.40934 -0.36
450 20 2.7308 2.52
46.794 40.89 49.556 48.21 0.87455 0.52
450 60 2.1109 1.91
10.145 10.96 52.622 50.95 -0.075505 -0.22
450 100 1.5561 1.514.5467 4.03 52.218 50.02 -0.34786 -0.40
Table 4. Comparison between reference, ref (Linstrom & Mallard, 2009, and Lemmon et al.,
2002), and NPT molecular dynamics simulated properties, sim, for methane as a function of
pressure and temperature. Properties: thermal expansion coefficient (αP), isothermal
compressibility (βT), isobaric heat capacity (CP) and Joule–Thomson coefficient (μ)

The remaining properties obtained from molecular dynamics simulations are reported in
Table 4. Average absolute deviation, between reference and simulated data, are 7.70 %, 9.19
%, 2.53 % and 37.62 % for thermal expansion coefficient, isothermal compressibility, isobaric
heat capacity and Joule – Thomson coefficient, respectively. Therefore, deviations obtained
for these properties obtained from fluctuation analysis of simulations are too large,

www.intechopen.com
Molecular dynamics simulations of volumetric thermophysical properties of natural gases 427

especially for Joule – Thomson coefficient. We report in Fig. 3 a comparison between

experimental and simulated values of isobaric heat capacity and Joule – Thomson
coefficient, from the reported results it may be inferred that trends of the properties with
pressure and temperature are properly reproduced by simulations, with the larger
deviations obtained for Joule – Thomson coefficient at the lower pressures. Deviations
obtained in this work are in agreement with those reported by Lagache et al. (2004) using
Monte Carlo approach with a different force field parameterization. Therefore, the main
problems rising from molecular dynamics simulations are for derived properties, especially
for Joule – Thomson coefficient, as we may expect.

68 2
(a) (b)
64
1.5

60
CP / J mol-1 K-1

 / K MPa-1
1
56
0.5
52

0
48

44 -0.5
20 40 60 80 100 20 40 60 80 100
P / MPa P / MPa

Fig. 3. Isobaric heat capacity (CP) and Joule–Thomson coefficient (μ) of pure methane as a
function of pressure and temperature. Symbols, NPT molecular dynamics simulated
properties, and lines, reference data (Linstrom & Mallard, 2009). Color code: black (250 K),
blue (350 K) and red (450 K)

4.2 Ethane
Density predictions for ethane are reported in Table 5, deviations between experimental and
simulated data are larger for ethane than for methane (Tables 3 and 5), and for methane
these deviations are larger than the ones required for natural gas engineering purposes.
Thus inaccuracies of simulations increase with increasing chain length of n-alkane, in spite
of using an all-atoms force field approach, this could lead to increasing inaccuracies for
multicomponent mixtures rich in large n-alkanes. Derived properties are reported in Table
6, large deviations are obtained for all the studied properties, especially for Joule–Thomson
coefficient, as we obtained in the previous section for methane. Average absolute deviation,
between reference and simulated data, are 8.86 %, 10.62 %, 3.06 % and 17.798 % for thermal
expansion coefficient, isothermal compressibility, isobaric heat capacity and Joule–Thomson
coefficient, respectively.

www.intechopen.com
428 Natural Gas

T/K P / MPa ρref / kg m-3 ρsim / kg m-3 standard deviation    sim 

 ref
ref
of ρsim / kg m-3 100

250 20 486.74 471.42 1.22 3.15

250 60 529.91 510.23 1.81 3.71
250 100 557.16 537.81 1.95 3.47
350 20 343.55 329.85 0.85 3.99
350 60 445.96 436.74 1.74 2.07
350 100 489.08 478.33 1.39 2.20
450 20 202.36 195.12 0.94 3.58
450 60 370.02 360.81 1.22 2.49
450 100 429.48 420.75 1.40 2.03
Table 5. Comparison between reference, ρref (Linstrom & Mallard, 2009), and NPT molecular
dynamics simulated, ρsim, density values, for ethane as a function of pressure and
temperature

T / K P / MPa 103 αP / K-1 103 βT / MPa-1 CP / J mol-1 K-1 μ / K MPa-1

ref sim ref sim ref sim ref sim
250 20 2.5063 2.84
3.1016 2.66 77.586 75.04 -0.29735 -0.25
250 60 1.6200 1.39
1.5345 1.63 71.927 70.09 -0.46941 -0.55
250 100 1.2873 1.32
1.0391 1.02 70.138 68.85 -0.52185 -0.54
350 20 4.8378 4.35
16.779 14.61 98.734 94.76 0.61455 0.50
350 60 1.8187 1.67
3.1585 3.56 81.503 78.87 -0.30069 -0.36
350 100 1.3102 1.35
1.7358 1.47 78.897 76.32 -0.42192 -0.43
450 20 4.6465 5.24
44.489 38.01 99.313 96.23 1.6323 2.17
450 60 1.8816 2.06
5.5582 6.23 91.100 88.78 -0.13673 -0.07
450 100 1.2809 1.36
2.5866 2.44 88.511 85.24 -0.33508 -0.33
Table 6. Comparison between reference, ref (Linstrom & Mallard, 2009, and Lemmon et al.,
2002), and NPT molecular dynamics simulated properties, sim, for ethane as a function of
pressure and temperature. Properties: thermal expansion coefficient (αP), isothermal
compressibility (βT), isobaric heat capacity (CP) and Joule–Thomson coefficient (μ)

4.3 CO2 and N2

Results for carbon dioxide and nitrogen are reported in Tables 7 – 10. Although these gases
are minor component of common natural gas mixtures, their effect on the thermodynamic
properties of natural gas mixtures is relevant, and thus, the accurate reproduction of their
properties through simulations is required to obtain reliable results. Density is simulated for
both gases with low deviations, average absolute deviation, between reference and
simulated data, are 1.70 % and 1.71 % for carbon dioxide and nitrogen respectively.
Nevertheless, these deviations are larger than 1 %, and thus, as for the results reported for n-
alkanes in previous sections, simulations lead to deviations too large for engineering
purposes. Derived properties reported in Tables 8 and 10, show too large deviations, leading
to average absolute deviations, between reference and simulated data, of 9.52 and 9.11 %
(for thermal expansion coefficient, for carbon dioxide and nitrogen, respectively), 13.09 and
13.17 % (for isothermal compressibility), 3.54 and 3.32 % (for isobaric heat capacity) and
70.57 and 40.3 % (for Joule–Thomson coefficient).

www.intechopen.com
Molecular dynamics simulations of volumetric thermophysical properties of natural gases 429

T/K P / MPa ρref / kg m-3 ρsim / kg m-3 standard deviation    sim 

 ref
ref
of ρsim / kg m-3 100

250 20 1105.5 1082.45 1.40 2.09

250 60 1183.6 1161.16 1.10 1.90
250 100 1235.6 1215.04 1.55 1.66
350 20 614.18 603.76 1.04 1.70
350 60 923.92 910.12 1.47 1.49
350 100 1027.7 1013.22 1.33 1.41
450 20 285.14 280.48 1.25 1.63
450 60 689.59 677.97 1.01 1.69
450 100 847.00 832.01 1.38 1.77
Table 7. Comparison between reference, ρref (Linstrom & Mallard, 2009), and NPT molecular
dynamics simulated, ρsim, density values, for carbon dioxide as a function of pressure and
temperature

T / K P / MPa 103 αP / K-1103 βT / MPa-1 CP / J mol-1 K-1 μ / K MPa-1

ref simref sim ref sim ref sim
250 20 3.1348 2.83
2.3550 2.04 83.318 80.11 -0.10335 -0.15
250 60 2.1531 1.96
1.2889 1.13 76.168 73.35 -0.22540 -0.26
250 100 1.7459 1.60
0.9062 0.79 73.521 70.81 -0.27302 -0.31
350 20 10.574 9.49
37.489 32.62 115.34 111.89 1.6780 1.51
350 60 2.7954 2.54
3.7943 3.31 74.349 72.10 -0.013846 -0.07
350 100 1.9221 1.75
1.9368 1.70 69.370 67.18 -0.20202 -0.25
450 20 4.5519 4.04
54.620 47.14 64.820 62.38 2.4962 2.06
450 60 2.8988 2.60
8.1235 7.02 69.816 67.12 0.27833 0.16
450 100 1.9076 1.75
3.3676 2.92 66.107 63.57 -0.11128 -0.18
Table 8. Comparison between reference, ref (Linstrom & Mallard, 2009, and Lemmon et al.,
2002), and NPT molecular dynamics simulated properties, sim, for carbon dioxide as a function
of pressure and temperature. Properties: thermal expansion coefficient (αP), isothermal
compressibility (βT), isobaric heat capacity (CP) and Joule–Thomson coefficient (μ)

T/K P / MPa ρref / kg m-3 ρsim / kg m-3 standard deviation    sim 

 ref
ref
of ρsim / kg m-3 100

250 20 268.10 262.99 1.62 1.91

250 60 523.04 513.14 1.51 1.89
250 100 631.37 621.81 1.44 1.51
350 20 178.03 174.46 1.00 2.01
350 60 400.90 394.65 1.74 1.56
350 100 520.38 513.04 1.25 1.41
450 20 136.50 134.03 1.04 1.81
450 60 325.58 320.51 1.91 1.56
450 100 442.22 434.70 1.58 1.70
Table 9. Comparison between reference, ρref (Linstrom & Mallard, 2009), and NPT molecular
dynamics simulated, ρsim, density values, for nitrogen as a function of pressure and
temperature

www.intechopen.com
430 Natural Gas

T/K P / MPa 103 αP / K-1103 βT / MPa-1 CP / J mol-1 K-1 μ / K MPa-1

ref simref sim ref sim ref sim
250 20 5.4355 5.04
41.524 36.13 40.966 39.42 0.91540 0.70
250 60 2.9791 2.73
7.0050 6.11 40.680 39.20 -0.33604 -0.44
250 100 2.0939 1.93
3.2581 2.81 39.685 38.29 -0.53276 -0.61
350 20 3.1609 2.86
43.877 38.05 34.185 33.13 0.48940 0.00
350 60 2.3437 2.13
9.6497 8.39 36.591 35.54 -0.34318 -0.51
350 100 1.7733 1.60
4.4900 3.92 36.671 35.73 -0.55688 -0.67
450 20 2.2621 2.02
44.586 38.50 32.184 31.14 0.11451 -0.61
450 60 1.8502 1.66
11.083 9.61 34.323 33.16 -0.41970 -0.67
450 100 1.4925 1.35
5.3610 4.66 34.852 33.53 -0.59686 -0.75
Table 10. Comparison between reference, ref (Linstrom & Mallard, 2009, and Lemmon et al.,
2002), and NPT molecular dynamics simulated properties, sim, for nitrogen as a function of
pressure and temperature. Properties: thermal expansion coefficient (αP), isothermal
compressibility (βT), isobaric heat capacity (CP) and Joule–Thomson coefficient (μ)

4.4 Lean Natural Gas Mixtures

Results for the multicomponent natural gas mixture studied in this work are reported in
Table 11. Deviations are in the 2–4 % range with an average absolute deviation from
experimental results of 2.55 %, which is too large in comparison with reference methods
used by the natural gas industry for density predictions. There are not available
experimental data for derived properties for the studied mixture. Therefore, although
thermal expansion coefficient, isothermal compressibility, isobaric heat capacity and Joule–
Thomson coefficient were calculated using molecular dynamics simulations for the studied
mixture, these results are not reported here because it is not possible to infer their accuracy.
Nevertheless, we may expect that considering that density deviations reported in Table 11
for the studied mixture are larger than those reported in previous sections for methane,
ethane, carbon dioxide and nitrogen, deviations of derived properties for the mixture should
be also larger than for the studied pure fluids.

T/K P / MPa ρref / kg m-3 ρsim / kg m-3 standard deviation of    sim 

 ref
ref
ρsim / kg m-3 100

270 3.483 31.673 32.39 5.87 -2.26

270 17.261 206.727 201.32 1.22 2.62
270 34.543 298.051 288.81 0.89 3.10
290 3.453 28.415 29.05 6.02 -2.23
290 17.271 175.201 171.27 1.07 2.24
290 34.506 274.016 265.64 0.70 3.06
305 3.484 26.894 27.38 6.24 -1.81
305 20.707 185.026 181.45 0.38 1.93
305 34.472 257.481 251.40 0.97 2.36
340 3.450 23.264 24.27 5.98 -4.32
340 20.699 151.983 147.85 0.94 2.72
340 34.486 224.033 219.71 0.96 1.93
Table 11. Comparison between reference, ρref (Patil et al., 2007), and NPT molecular dynamics
simulated, ρsim, density values, for the natural gas like mixture (composition reported in
Table 1) as a function of pressure and temperature

www.intechopen.com
Molecular dynamics simulations of volumetric thermophysical properties of natural gases 431

5. Conclusions
Results reported in this work show preliminary conclusions obtained in the first stages of a
wide computational study that we are carrying out on the performance of molecular
simulation approaches for the prediction of thermophysical properties of multicomponent
natural gas like mixtures. The main conclusions that may be inferred from these initial
results may be resumed in:
i) Density for the main components of natural gases is predicted with deviations
in the 1 – 2 % range for the 250 – 450 K / 20 – 100 MPa studied ranges. These
deviations are low considering the purely predictive character of the
considered approach but are too large for natural gas engineering purposes.
ii) Derived properties (thermal expansion coefficient, isothermal compressibility,
isobaric heat capacity and Joule–Thomson coefficient), for the main
components of natural gases are predicted with deviations up to 10 % (even
larger for Joule–Thomson coefficient for some fluids). Therefore, although
these are the deviations commonly obtained using molecular dynamics
approach for many complex fluids, they are too large to use this approach for
industrial purposes. Fortunately, deviations for isobaric heat capacity, which is
a property with remarkable importance for natural gas engineering purposes,
is predicted with deviations usually lower than 5 %.
iii) Predictions for the studied multicomponent natural gas like mixture lead to
analogous results than those mentioned in conclusions i and ii: low deviations
for density which lead to probably larger deviations for the remaining
properties, but for all of them too large for industrial purposes. Moreover, it
should be remarked that the studied mixture is composed of n-alkanes only up
to C5, and thus, heavier mixtures containing larger alkanes should lead to
even larger deviations.
Therefore, considering the results reported in this work, molecular dynamics approach is
not able to lead to predictions that may be used for natural gas production, processing or
transportation purposes in the present situation. Nevertheless, we think that results
reported in this work, and by other authors in the literature (Lagache et al., 2001; Lagache et
al. 2004), are very promising, and the use of molecular simulation approach to predict
complex natural gas like mixtures properties should not be discarded. Therefore studies are
being carried out by the authors with the next main objectives:
i) Improvement of forcefield parameterizations.
ii) Comparison of the performance of several force fields.
iii) Improvements of all atoms force fields approach in comparison with united
atoms approach, does the increase of computational time lead to a remarkable
improvement in the accuracy of predictions?
iv) Decrease the errors in density predictions below the 1 % limit, in wide
pressure – temperature ranges, and thus, leading to results competitive with
the methods commonly used by the gas industry nowadays.
v) Analyze the predictions for heavy mixtures containing long alkanes.
vi) Comparison of the performance of Monte Carlo and molecular dynamics
approaches.
vii) Analyze from a molecular viewpoint the structural factors that govern these
complex mixtures behavior using these computational tools.

www.intechopen.com
432 Natural Gas

6. References
Allen, M. P. & Tildesley, D. J. (1989). Computer Simulation of Liquids,Clarendon Press, ISBN
0198556454, Oxford.
Babusiaux, D. (2004). Oil and Gas Exploration and Production: Reserves, Costs, Contracts,
Editions Technip, ISBN 2710808404, Paris.
Bessieres, D.; Randzio, S. L.; Piñeiro, M. M.; Lafitte, Th. & Daridon, J. L. (2006). A Combined
Pressure-controlled Scanning Calorimetry and Monte Carlo Determination of the
Joule−Thomson Inversion Curve. Application to Methane. J. Phys. Chem. B, 110, 11,
February 2006, 5659-5664, ISSN 1089-5647.
Bluvshtein, I. (2007). Uncertainties of gas measurement. Pipeline & Gas Journal, 234, 5, May
2007, 28-33, ISSN 0032-0188.
Bluvshtein, I. (2007). Uncertainties of measuring systems. Pipeline & Gas Journal, 234, 7, July
2007, 16-21, ISSN 0032-0188.
Duan, Z.; Moller, N. & Weare, J. H. (1992). Molecular dynamics simulation of PVT
properties of geological fluids and a general equation of state of nonpolar and
weaklyu polar gases up to 2000 K and 20000 bar. Geochim. Cosmochim. Acta, 56, 10,
October 1992, 3839-3845, ISSN 0016- 7037.
Duan, Z.; Moller, N. & Weare, J. H. (1996). A general equation of state for supercritical fluid
mixtures and molecular dynamics simulation of mixture PVTx properties. Geochim.
Cosmochim. Acta, 60, 7, April 1996, 1209-1216, ISSN 0016- 7037.
Dysthe, D. K., Fuch, A. H.; Rousseau, B. & Durandeau, M. (1999). Fluid transport properties
by equilibrium molecular dynamics. II. Multicomponent systems. J. Chem.
Phys., 110, 8, February 1999, 4060-4067, ISSN 0021-9606.
Errington, J.R. & Panagiotopoulos, A. Z. (1998). A Fixed Point Charge Model for Water
Optimized to the Vapor−Liquid Coexistence Properties. J. Phys. Chem. B, 102, 38,
September 1998, 7470-7475, ISSN 1089-5647.
Errington, J. & Panagiotopoulos, A. Z. (1999). A New Intermolecular Potential Model for the
n-Alkane Homologous Series. J. Phys. Chem. B, 103, 30, July 1999, 6314-6322, ISSN
1089-5647.
Escobedo, F. A. & Chen, Z. (2001). Simulation of isoenthalps curves and Joule – Thomson
inversion of pure fluids and mixtures. Mol. Sim., 26, 6, June 2001, 395-416, ISSN
0892-7022.
Essmann, U. L.; Perera, M. L.; Berkowitz, T.; Darden, H.; Lee,H. & Pedersen, L. G. (1995) J.
Chem. Phys., 103, 19, November 2005, 8577-8593, ISSN 0021-9606.
Gallagher, J. E. (2006). Natural Gas Measurement Handbook, Gulf Publishing Company, ISBN
1933762005, Houston.
Hall, K. R. & Holste, J. C. (1990). Determination of natural gas custody transfer properties.
Flow. Meas. Instrum., 1, 3, April 1990, 127-132, ISSN 0955-5986.
Hoover, W. G. (1985). Canonical dynamics: Equilibrium phase-space distributions. Phys.
Rev. A, 31, 3, March 1985, 1695-1697, ISSN 1050-2947.
Husain, Z. D. (1993). Theoretical uncertainty of orifice flow measurement, Proceedings of 68th
International School of Hydrocarbon Measurement, pp. 70-75, May 1993, publ,
Oklahoma City.
Jaescke, M.; Schley, P. & Janssen-van Rosmalen, R. (2002). Thermodynamic research
improves energy measurement in natural gas. Int. J. Thermophys., 23, 4, July 2002,
1013-1031, ISSN 1572-9567.

www.intechopen.com
Molecular dynamics simulations of volumetric thermophysical properties of natural gases 433

Jorgensen, W. L.; Maxwell, D. S. & Tirado–Rives, J. (1996). Development and testing of the
OPLS All-Atom force field on conformational energetics and properties of
organic liquids. J. Am. Chem. Soc., 118, 45, November 1996, 11225-11236, ISSN
0002-7863.
Lagache, M.; Ungerer, P., Boutin, A. & Fuchs, A. H. (2001). Prediction of thermodynamic
derivative properties of fuids by Monte Carlo simulation. Phys. Chem. Chem.
Phys., 3, 8, February 2001, 4333-4339, ISSN 1463-9076.
Lagache, M. H.; Ungerer, P.; Boutin, A. (2004). Prediction of thermodynamic derivative
properties of natural condensate gases at high pressure by Monte Carlo simulation.
Fluid Phase Equilibr., 220, 2, June 2004, 211-223, ISSN 0378-3812.
Lemmon, E. W.; McLinden, M. O.; Huber, M. L. NIST Standard Reference Database 23,
Version 7.0, National Institute of Standards and Techcnology, Physical and
Chamical Properties Division, Gaithersburg, MD, 2002.
Linstrom, P. J. & Mallard, W.G. Eds. (2009). NIST Chemistry WebBook, NIST Standard
Reference Database Number 69, National Institute of Standards and Technology,
Gaithersburg. Available at http://webbook.nist.gov
Martínez, J. M. & Martínez, L. (2003). Packing optimization for automated generation of
complex system's initial configurations for molecular dynamics and docking. J.
Comput. Chem., 24, 7, May 2003, 819-825, ISSN 0192-8651.
Martin, M.G. & Frischknecht, A. L. (2006). Using arbitrary trial distributions to improve
intramolecular sampling in configurational-bias Monte Carlo. Mol. Phys., 104, 15,
July 2006, 2439-2456, ISSN 0026-8976.
Martin, M.G. & Siepmann, J.I. (1999). Novel Configurational-Bias Monte Carlo Method for
Branched Molecules. Transferable Potentials for Phase Equilibria. 2. United-
Atom Description of Branched Alkanes. J. Phys. Chem. B, 103, 21, May 1999,
4508-4517, ISSN 1089-5647.
Mokhatab, S.; Poe, W. A. & Speight, J. G. Handbook of Natural Gas Transmission and
Processing, Gulf Professional Publishing, ISBN 0750677767, Burlington.
Neubauer, B.; Tavitian, B.; Boutin, A.; Ungerer, P. (1999). Molecular simulations on
volumetric properties of natural gas. Fluid Phase Equilibr., 161, 1, July 1999, 45-62,
ISSN 0378-3812.
Patil, P.; Ejaz, S.; Atilhan, M.; Cristancho, D.; Holste, J. C. & Hall, K. R. (2007). Accurate
density measurements for a 91 % methane natural gas-like mixture. J. Chem.
Thermodyn., 39, 8, August 2007, 1157-1163, ISSN 0021-9614.
Ponder, J. W. (2004). TINKER: Software tool for molecular design. 4.2 ed, Washington
University School of Medicine.
Saager, B. & Fischer, J. (1990). Predictive power of effective intermolecular pair potentials:
MD simulation results for methane up to 1000 MPa. Fluid Phase Equilibr., 57, 1-2,
July 1990, 35-46, ISSN 0378-3812.
Shi, W. & Maginn, E. (2008). Atomistic Simulation of the Absorption of Carbon Dioxide and
Water in the Ionic Liquid 1-n-Hexyl-3-methylimidazolium
Bis(trifluoromethylsulfonyl)imide ([hmim][Tf2N]. J. Phys. Chem. B, 112, 7,
January 2008, 2045-2055, ISSN ISSN 1089-5647.
Siepmann, J.I. & Frenkel, D. (1992). Configurational bias Monte Carlo: a new sampling
scheme for flexible chains. Mol. Phys., 75, 1, January 1992, 59-70, ISSN 0026-8976.

www.intechopen.com
434 Natural Gas

Smit, B. & Williams, C. P. (1990). Vapour-liquid equilibria for quadrupolar Lennard-Jones

fluids. J. Phys. Condens. Matter, 2, 18, May 1990, 4281-4288, 0953-8984.
Starling, K.E. & Savidge, J.L. (1992) Compressibility Factors of Natural Gas and Other Related
Hydrocarbon Gases, AGA transmission Measurement Committee Report 8,
American Gas Association, 1992.
Ungerer, P. (2003). From Organic geochemistry to statistical thermodynamics: the
development of simulation methods for the petroleum industry. Oil & Gas Science
and Technology – Rev. IFP, 58, 2, May 2003, 271-297, ISSN 1294-4475.
Ungerer, P.; Wender, A.; Demoulin, G.; Bourasseau, E. & Mougin, P. (2004). Application of
Gibbs Ensemble and NPT Monte Carlo Simulation to the Development of
Improved Processes for H2S-rich Gases. Mol. Sim., 30, 10, August 2004, 631-648,
ISSN 0892-7022.
Ungerer, P.; Lachet, V. & Tavitian, B. (2006). Properties of natural gases at high pressure. In:
Applications of molecular simulation in the oil and gas industry. Monte Carlo methods.,
162-175, Editions Technip, ISBN 2710808587, Paris.
Ungerer, P.; Lachet, V. & Tavitian, B. (2006). Applications of molecular simulation in oil and
gas production and processing. Oil & Gas Science and Technology – Rev. IFP., 61, 3,
May 2006, 387-403, ISSN 1294-4475.
Ungerer, P.; Nieto-Draghi, C.; Rousseau, B.; Ahunbay, G. & Lachet, V. (2007). Molecular
simulation of the thermophysical properties of fluids: From understanding
toward quantitative predictions. J. Mol. Liq., 134, 1-3, May 2007, 71-89, ISSN
0167-7322.
Vlugt, T. J. H.; Martin, M.G.; Smit, B.; Siepmann, J.I. & Krishna, R. (1998). Improving the
efficiency of the configurational-bias Monte Carlo algorithm. Mol. Phys., 94, 4,
July 1998, 727-733, ISSN 0026-8976.
Vrabec, J.; Kumar, A. & Hasse, H. (2007). Joule–Thomson inversion curves of mixtures by
molecular simulation in comparison to advanced equations of state: Natural gas as
an example. Fluid Phase Equilibr., 258, 1, September 2007, 34-40, ISSN 0378-3812.
Wagner, W. & Kleinrahm, R. (2004). Densimeters for very accurate density measurements of
fluids over large ranges of temperature, pressure, and density. Metrologia, 41, 2,
March 2004, S24-S29, ISSN 0026-1394.
Yoshida, T.; Uematsu, M. (1996). Prediction of PVT properties of natural gases by molecular
simulation. Transactions of the Japan Society of Mechanical Engineers, Series B, 62, 593, ,
278-283, ISSN 03875016.

www.intechopen.com
 Natural Gas
Edited by Primož PotoÄÂ​nik

ISBN 978-953-307-112-1
Hard cover, 606 pages
Publisher Sciyo
Published online 18, August, 2010
Published in print edition August, 2010

The contributions in this book present an overview of cutting edge research on natural gas which is a vital
component of world's supply of energy. Natural gas is a combustible mixture of hydrocarbon gases, primarily
methane but also heavier gaseous hydrocarbons such as ethane, propane and butane. Unlike other fossil
fuels, natural gas is clean burning and emits lower levels of potentially harmful by-products into the air.
Therefore, it is considered as one of the cleanest, safest, and most useful of all energy sources applied in
variety of residential, commercial and industrial fields. The book is organized in 25 chapters that cover various
aspects of natural gas research: technology, applications, forecasting, numerical simulations, transport and
risk assessment.

How to reference
In order to correctly reference this scholarly work, feel free to copy and paste the following:

Santiago Aparicio and Mert Atilhan (2010). Molecular Dynamics Simulations of Volumetric Thermophysical
Properties of Natural Gases, Natural Gas, Primož PotoÄÂ​nik (Ed.), ISBN: 978-953-307-112-1, InTech,
Available from: http://www.intechopen.com/books/natural-gas/molecular-dynamics-simulations-of-volumetric-
thermophysical-properties-of-natural-gases

InTech Europe InTech China

University Campus STeP Ri Unit 405, Office Block, Hotel Equatorial Shanghai
Slavka Krautzeka 83/A No.65, Yan An Road (West), Shanghai, 200040, China
51000 Rijeka, Croatia
Phone: +385 (51) 770 447 Phone: +86-21-62489820
Fax: +385 (51) 686 166 Fax: +86-21-62489821
www.intechopen.com