VMG Property Package Thermodynamic Models 3.7
VMG Property Package Thermodynamic Models 3.7
May 2002
By
Thermodynamic Models 8
Introduction ...............................................................................................................................................8
Basics of Fluid Phase Thermodynamics ...............................................................................................8
Fugacities and Standard States ...............................................................................................................9
Thermal Calculations.............................................................................................................................11
Transport Properties ...............................................................................................................................12
Choosing Thermodynamic Models ......................................................................................................12
Reference Definitions ............................................................................................................................14
Enthalpy....................................................................................................................................14
Entropy .....................................................................................................................................14
Helmholtz Energy Departure ................................................................................................15
Suggested Aggregations ........................................................................................................................15
Complete Packages.................................................................................................................15
Partial Packages ......................................................................................................................16
Creating A Property Package Selection ..............................................................................16
Activity Coefficient Models .................................................................................................................18
Introduction .............................................................................................................................18
Liquid Phase Modeling Using Activity Coefficient Models ...........................................18
Gas Phase Modeling Using Activity Coefficient Models ................................................18
Liquid Phase Standard State Fugacity.................................................................................19
Understanding Literature Interaction Parameters..............................................................22
Understanding Virtual Materials Matrix Notation............................................................24
Understanding Interaction Parameter Access ....................................................................25
Non Condensable Components ............................................................................................25
Comments on activity coefficient models ..........................................................................26
Understanding Virtual Materials UNIFAC.........................................................................................36
Handling of interaction parameters for activity coefficient models ...............................36
The UNIFAC Method ............................................................................................................36
UNIFAC Application .............................................................................................................37
VMG Implementation of UNIFAC......................................................................................37
When it is used ?.....................................................................................................................37
What UNIFAC version is used ? .........................................................................................37
How is it used ? .......................................................................................................................38
How can I find from where my interaction parameters are coming from ? ..................38
Understanding the error information ...................................................................................39
Warnings and hints.................................................................................................................40
Equations Of State..................................................................................................................................41
Virial Equation Of State ........................................................................................................42
Ideal gas properties.................................................................................................................58
General description of a cubic equation of state ...............................................................60
Redlich-Kwong .......................................................................................................................62
Index 169
Copyright Notice
The copyright in this manual and its associated computer program are the property of
Virtual Materials Group, Inc. All rights reserved. Both this manual and the computer
program have been provided pursuant to a License Agreement containing restrictions
on use.
Virtual Materials Group, Inc. reserves the right to make changes to this manual or its
associated computer program without obligation to notify any person or
organization. Companies, names and data used in examples herein are fictitious
unless otherwise stated.
No part of this manual may be reproduced, transmitted, transcribed, stored in a
retrieval system, or translated into any other language, in any form or by any means,
electronic, mechanical, magnetic, optical, chemical, manual or otherwise, or
disclosed to third parties without the prior written consent of Virtual Materials
Group, Inc., 657 Hawkside Mews NW, Calgary, AB, Canada, T3G 3S1
1999, 2000, 2001, 2002 Virtual Materials Group, Inc., All rights reserved
Introduction
At the heart of Virtual Materials property package system are mathematical models
designed to represent the actual behavior of many different fluid mixtures. This
section provides brief overview of fluid phase thermodynamics, details on how new
property packages can be created by aggregation and finally details on the specifics
of each property package and calculation methods for all property packages.
T I = T II = ... = T np 1
P I = P II = ... = P np 2
i
I
= i
II
= ... = np
i 3
f i np ( )
r
fi I ( I
) = fi II ( rII ) = ... = r
np 4
fi
i = 5
Pxi
And the activity coefficient:
i = i
pure
6
i
Using equations 5 and 6 we can write the equilibrium relationships in the way
normally found in textbooks:
I I
i i x P= i x P = ...
II II
i i
np np
x P
i 7
I I
i i x f i I , pure = i
II
xiII f i II , pure = ... = i
np np np , pure
x f
i i 8
Also note that activity coefficients can also be used to describe any phase, including
the vapor phase. This has not found any significant practical application.
Fugacity coefficients in turn are related to pressure, volume and temperature and
composition by the following equation:
∞
∂P RT
RT ln = ∫ − dV − RT ln Z 9
∂n i
i
V T , n j ≠i V
With:
PV
Z= 10
RT
Equations of state such as Soave-Redlich-Kwong and Peng-Robinson can be used for
the calculation of fugacity coefficients for the liquid and vapor phase and therefore
they provide a compact way for the calculation of thermodynamic equilibrium.
Before the development of advanced mixing rules for equations of state, simple
equations of state frequently proved inadequate for modeling solutions with strong
interactions in the liquid phase, such as aqueous mixtures of polar compounds. In
those cases, Equation 7 is used for modeling the gas phase while Equation 8 is used
for modeling the liquid phase.
It is common practice to assume that the partial molar volume of component i and
the molar volume of component i are identical, and independent of composition and
pressure.
Note that the assumption of composition independence for the partial molar volume
is clearly incorrect, albeit normally used. This assumption has serious consequences
at conditions close to the critical point of the mixture. Fortunately this inaccuracy is
usually immaterial for low-pressure vapor-liquid calculations.
Equation 13 is usually applied to the liquid phase, and diverse fugacity coefficient
models give rise to diverse expressions for the saturation fugacity coefficient. The
term:
P
vi
∫ RT dP
POYi = e Psat,i
14
f i = Himix xi 15
Aisolvent
ln H isolvent = + Bsolvent
i lnT + C solvent
i T + D solvent
i 16
T
Note that the Henry coefficients are dependent on the solvent, and equation 16 holds
for a binary mixture.
nc
ln H imix = ∑x j
j =1, i ≠ inert
ln H ij 17
Note that contrary to what one finds for condensable components, the standard state
fugacity is a function of the solvents present in the mixture, and is not a pure
component property.
Thermal Calculations
Thermodynamic equilibrium defines the state of a fluid. Nevertheless, knowledge of
enthalpies, entropies and heat capacities is fundamental for the modeling of
equipment and energy balances. These quantities are defined using rigorous
relationships reproduced below
∂P V
∂P
H = H + ∫ T
0
+V dV 18
∞ ∂T V , ni ∂V T , ni
∂P
V
R
S = S + ∫ − dV + R ln Z
0
19
∞ ∂T V , ni V
∂P
2
T
V
∂ P
2
∂T V
C p = C p + T ∫ 2 dV −
0
−R 20
∞
∂T V ∂P
∂V T
Therefore, with model and ideal gas heat capacities all thermal properties of interest
can be calculated.
Related physical properties are calculated according to the following definitions:
G = H − TS
U = H − pV
A = U − TS
Note that isomer separations are usually problematic, and it is recommended that a check on
Virtual Materials vapor pressure prediction be performed on each isomer (Agarwal et al,
2000).
Refinery
Refinery processes are conveniently modeled using APR or REFINERY-SRK. Both
models include special correlations for modeling hydrogen. APR will provide more
accurate density predictions than REFINERY-SRK.
Chemicals
The basic question on chemical processing is the pressure you are interested at, as it
will determine if an equation of state or activity coefficient model will be necessary.
High Pressure
If vapor-liquid equilibrium is important, GE-PR is usually the best choice.
Association Alert: If your system contain carboxylic acids at high pressure then GE-
PR is not adequate and you should use an activity coefficient model with Hayden-
If your model has low pressure distillation and high pressure intermediate equipment
(for example compressors and reactors), a good model can be created by using SRK,
REFINERY-SRK, PR, APR for the gas phase and Wilson, NRTL or UNIQUAC for
the liquid phase. The RK model is provided for compatibility with older calculations,
and should not be used for new calculations as it is significantly less accurate than
the more modern SRK and PR equations. See the methods AddAggregatePkgFromId
and AddAggregatePkgFromName for more details.
If you are performing initial design work, and you have little or no information about
the mixture of interest, a useful method is PSRK (predictive SRK). It combines a
modification of the SRK equation of state designed to accurately model the vapor
pressure of polar and non-polar substances and the UNIFAC method, with specially
determined atomic groups for common industrial gases. It can be used for low and
high-pressure conditions.
Low Pressure
Low-pressure equilibrium calculations are usually complex due to large polar
interactions between chemicals in the liquid phase and sometimes we will find truly
anomalous behavior in the gas phase when carboxylic acids are present
Carboxylic Acids Are Present
In this case the use of the Hayden-O'Connell model for the gas phase is essential.
Failure of doing so will result in wrong phase equilibrium and physical properties
such as gas density. Usually activity coefficient models such as Wilson, NRTL or
UNIQUAC will work well.
Systems That may phase split
If your mixture can phase split (for example, it contains water and hydrocarbons or
heavy alcohols) and no carboxylic acids usually NRTL and UNIQUAC will be
adequate with ideal gas law for the gas phase. If you have carboxylic acids make
sure you use Hayden-O'Connell for the gas phase
Systems that do not phase split
Usually the best model for the liquid phase for systems which do not phase split is
Wilson, with ideal gas for the gas phase (or Hayden-O'Connell if the system has
carboxylic acids).
Reference Definitions
Enthalpy
Ideal gas enthalpy equal to zero at 101.325 kPa and 0 K
Entropy
Ideal gas entropy equal to zero at 101.325 kPa and 1 K
Suggested Aggregations
Virtual Materials property packages can be combined in a very large number of
ways. Nevertheless, for practical applications only a relatively small subset of
aggregations is necessary to emulate the functionality normally found in process
calculation software.
Complete Packages
The following packages contain information to perform all phase, thermal and
transport calculations:
Activity Coefficient Models with Ideal Gas Phase
IdealLiquid/Ideal/HC
Margules/Ideal/HC
Margules4/Ideal/HC
MargulesSour
vanLaar/Ideal/HC
Wilson/Ideal/HC
TKWilson/Ideal/HC
NRTL/Ideal/HC
UNIQUAC/Ideal/HC
Chao-Seader
Grayson-Streed
Equations of State
RK
SRK
Refinery-SRK
PSRK
ZSRK
Peng-Robinson
Advanced Peng-Robinson
Ge-Peng-Robinson
NGL Peng-Robinson
Steam95
Partial Packages
The following packages contain information for specific phase calculations or for
calculations of a few specialized physical properties.
Equations of State for Gas Phase Only
HaydenOConnell
Using this suggested division of model types a character or graphical user interface
can be easily created as shown below:
In this simple interface created using Visual Basic, the two lists on the left are made
visible if the Activity Coefficient button is clicked on, and the gas and liquid phase
lists are created according to the lists described above. Then, with the user selection
for gas and liquid phase the appropriate aggregation is done. For example, for the
Introduction
Virtual Materials Property Package supports all commonly used activity coefficient
methods. It also properly handles non-condensable gases as well as estimation of
missing interaction parameters.
i
l
xi = i
v
yi 21
The fugacity coefficients for the gas phase are calculated based on the
thermodynamic model chosen for the gas phase. The fugacity coefficients for the
liquid phase are calculated based on the definition of fugacity:
i
l
xi P = i xi f i ol 22
Therefore the fugacity coefficient for the liquid phase is given by:
f i ol
i
l
= i
23
P
f i ol = i
sat
Poyi Pvi 24
The Poynting correction is approximated by using the pure component molar volume
at 25 C instead of using the partial molar volume of component i in the mixture:
v l ( P − Pvi )
Poyi = exp i 25
RT
For non-condensable components the standard state fugacity is calculated using
Henry constants:
f i ol = H imix 26
ln H imix = ∑x ln H ij
j
j =1 , j ≠ NonCond
27
Margules
bij
Aij = aij + + c ij ln T 28
T
nc
Ai = ∑ Aij x j 29
j =1
nc
Ri = ∑ A ji x j 30
j =1
ln i = (1 − x i )[ Ai + 2(Ri − Ai )x i ] 31
1 nc
(1 − xi ) ∑
Bi = x j A ji 31c
j =1
nc
1
(1 − x i ) ∑
Di = x j Dij 31d
j =1
Dij = D ji 31e
ln i
2
[
= (1 − xi ) Ai + 2 xi ( Bi − Ai − Di ) + 3Di x i2 ] 31f
van Laar
The van Laar equation used by Virtual Materials is identical to the one used by
Monsanto with the N coefficient set to 1 (Null, 1980)
bij
Aij = aij + + c ij ln T 32
T
Dij = D ji = d ij 33
nc
∑A x
i =1, i ≠ j
ij j
Ai =
(1 − xi )
34
nc
∑A
i =1, i ≠ j
ji xj
Bi =
(1 − xi )
35
nc
∑D
i =1,i ≠ j
ij xj
Di =
(1 − xi )
36
Ai x i
Zi =
Ai x i + Bi (1 − xi )
37
Ai Bi 2
= Ai (1 − Z i2 )1 + 2Z i
ln − 1 + D i Z i2 − Z i 38
3
i
Ai Bi
Wilson
bij
Aij = aij + + c ij ln T 39
T
nc nc
x Λ
ln i = − ln ∑ x j Λ ij + 1 − ∑ nc k ki 41
j=1 k =1
∑ x j Λ kj j =1
NRTL
bij
Aij = aij + + c ij ln T 42
T
ij = ji = d ij = d ji 43
ij = Aij 44
ln Gij = − ij ij 45
nc nc
∑ ji G ji x j nc x j G ij
∑x k G kj
kj
ln =−
j =1
+ ∑ nc − k =1nc 46
i nc ij
∑G
j =1
ji xj j =1
∑
k =1
G kj x k
∑k =1
xk G kj
UNIQUAC
bij
Aij = aij + + c ij ln T 47
T
ln ij = Aij 48
ri xi
i = nc
49
∑r x
j =1
j j
qi x i
i = nc
50
∑q
j =1
j xj
li =
Z
(ri − qi ) − (ri − 1) 51
2
Z = 10 52
nc
Z
ln c
i = ln i
xi
+ q ln
2 i
i
+ li −
xi
i
∑x l
j =1
j j 53
i
k =1
∑ k kj
ln i = ln c
i + ln i
r
55
nc nc x Λ nc v lj nc vil / v lj
ln = − ln ∑ x j Λ ij − ∑ nc k ki + ln ∑ x j l + ∑ x k
i
j =1 v i k =1 nc v l x
j=1 k =1
∑ x j Λ kj
∑
j j
l
j =1 j =1 v k
58
The molar volumes for components i and j are calculated based on the component
molecular weight and liquid density at 25 C. No temperature dependency is assumed.
Margules or Margules4
a ij = Aij
60
a ji = A ji
bij = b ji = c ij = c ji = 0 61
van Laar
a ij = Aij
62
a ji = A ji
bij = b ji = c ij = c ji = d ij = d ji = 0 63
Wilson
Bij
a ij = −
R
B ji
a ji = −
R
64
Aij
bij = −
R
A ji
b ji = −
R
c ij = c ji = 0 65
d ij = d ji = ij 68
UNIQUAC
Bij
a ij = −
R
B ji
a ji = −
R
69
Aij
bij = −
R
A ji
b ji = −
R
c ij = c ji = 0 70
Sanity Check: Interaction parameters are a function of vapor pressures. If you are
using literature interaction parameters make sure the vapor pressures are
compatible. If you fail in checking this detail your results may be incorrect.
Margules
The oldest activity coefficient model (1895) is a very simple yet flexible model for
the representation of activity coefficients.
Advantages
• Can represent maximum and minimum in activity coefficient curves
versus composition
• Binary interaction parameters correspond to infinite dilution activity
coefficients
• Can be used to model VLE and LLE
Disadvantages
• Empirical extension to multicomponent mixtures
Margules4
Extension of Margules equation by Wohl (1946) with more flexible composition
dependency on activity coefficient.
Disadvantages
• Empirical extension to multicomponent mixtures
MargulesSOUR
The MargulesSOUR property package combines the flexibility of the Margules
property package with the chemical reactions happening in liquid phase when CO2,
H2S and NH3 are present in the system. MargulesSOUR also contains tuned
interaction parameters for Urea and Water and can be used for modeling the
behavior of low pressure Urea production processes.
Chemical Basis
MargulesSOUR is an extension of the API's original sour method (API, 1978)
allowing for multicomponent calculations outside the component set originally
defined by API. MargulesSOUR allows the user to include light gases such as
hydrogen and methane into the calculations as well as heavier hydrocarbons, and
consequently the model can predict multiphase equilibrium, important for synthetic
fuels and refinery applications.
The following liquid phase chemical reactions are taken into account when using
MargulesSOUR:
H 2 S ⇔ HS − + H +
HCO3− + NH 3 ⇔ H 2 NCOO − + H 2 O
CO2 + H 2O ⇔ HCO3− + H +
HCO3− ⇔ CO3− − + H +
NH 3 + H + ⇔ NH 4+
H 2 S ⇔ HS − + H +
HS − ⇔ S −− + H +
H 2 O ⇔ OH − + H +
Advantages
• Same as original Margules plus.
• Accurate modeling of low pressure sour systems with Urea present
• Can be used to model VLE and LLE found in synthetic fuels
processing
Boiler Plate
MargulesSOUR was extensively validated against experimental data as summarized
in the table below. The system is always assumed aqueous.
A complete validation for this property package can be obtained from Virtual
Materials Group by requesting Technical Bulletin 057 (VMG, 2001)
This model is not recommended for ammonia concentrations in the liquid phase
above approximately 30 mole %.
Reactions are NOT taken into account if the water concentration in the liquid phase
is less than 50 mole %.
Reactions are NOT taken into account if the component list does not include mole
fractions of NH3, or CO2 or H2S equal or larger than 1.0e-12.
Urea is modeled as a very low volatility compound, and the user should avoid bubble
pressure or bubble temperature calculations using pure Urea.
van Laar
Originally created by van Laar in 1910 based on the van der Waals equation. The
original van der Waals equation does not provide a reliable model for liquids, and
therefore the model generated using parameters based on van der Waals theory are
unreliable. Therefore the mathematical structure of van Laar's equation is preserved
but the parameters are determined from experimental data.
Advantages
• Success reported by Monsanto (Null, 1980) in using the equation for
process simulation.
• Binary interaction parameters correspond to infinite dilution activity
coefficients.
• Can be used to model VLE and LLE
Disadvantages
• Maximum or minimum in activity coefficients versus composition can
not be predicted for binary systems.
• Walas (1985) reports poor fits for the water / 3-hydroxy-2-butanone
system when infinite dilution activity coefficients are used directly in
the van Laar equation using Null's modification, which corresponds to
the model used by Virtual Materials.
• Usually the poorest liquid model for alcohol / hydrocarbon systems
Wilson
The Wilson equation provides a simple, semi-theoretical framework for the
correlation of VLE data. The basis of the equation is the idea that liquids exist with
microstructures inside, where relative concentrations of molecules are determined by
their respective energies of interaction. This idea is usually known as the "local
composition" model and is the basis for the more modern activity coefficient models.
Advantages
• Good representation of activity coefficient behavior for polar and non-
polar mixtures, usually superior to other models.
• Simple extension for multicomponent mixtures
Disadvantages
• Can not represent liquid-liquid behavior
TK-Wilson
Same advantages as Wilson, with the additional advantage of being able to represent
liquid-liquid behavior.
Disadvantages
• Interaction parameters determined for the Wilson equation can not be
directly used in the TK-Wilson equation.
Tip
The interaction parameters determined for the Wilson equation can not be
directly used in the TK-Wilson equation. Nevertheless, a simple procedure
for converting Wilson interaction parameters into TK-Wilson interaction
parameters can be constructed using infinite dilution activity coefficients
(Walas, 1985):
NRTL
The NRTL model was basically developed to offer a solution to Wilson's original
equation inability to represent liquid-liquid equilibrium.
Advantages
• Capable of representing VLE and LLE
• Usually of comparable accuracy to Wilson and superior to other models
for aqueous / organic systems
• Simple extension to multicomponent systems
Disadvantages
• Need 3 parameters per binary pair
• Has a tendency to predict liquid-liquid phase split when none occurs in
reality.
Tip
Many times the non-random term for NRTL ( ij ) can be set based on the
chemical characteristics of the binary. Renon and Prausnitz (1968)
recommend the following values:
Recommended ij = 0.3
1
Based on small number of binary systems
2
Based on small number of binary systems
3
Based on small number of binary systems
Advantages
• Takes into account differences in size and shape and in principle can
be applied to mixtures with components of widely different molecular
sizes.
• Applicable to VLE and LLE
• Binary parameters are usually well behaved from a multicomponent,
multiphase flash calculation point of view.
• Basis for UNIFAC method
Disadvantages
• Need two pure component parameters, van der Waals area and volume.
• Not as good as Wilson or NRTL in fitting experimental data.
Aqueous Organics
Aqueous Organics
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
Margules van Laar Wilson NRTL UNIQUAC
Model
0.45
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
Margules van Laar Wilson NRTL UNIQUAC
Model
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
Margules van Laar Wilson NRTL UNIQUAC
Model
0.3
0.25
0.2
0.15
0.1
0.05
0
Margules van Laar Wilson NRTL UNIQUAC
Model
C4-C6 hydrocarbons
C4-C6 hydrocarbons
0.4
0.35
0.3
0.25
0.2
0.15
0.1
0.05
0
Margules van Laar Wilson NRTL UNIQUAC
Model
0.3
0.25
0.2
0.15
0.1
0.05
0
Margules van Laar Wilson NRTL UNIQUAC
Model
Aromatics
Aromatics
0.3
0.25
0.2
0.15
0.1
0.05
0
Margules van Laar Wilson NRTL UNIQUAC
Model
Summary
System Type Best Model
Aqueous Organics NRTL
Alcohols Wilson
Alcohols and Phenols Wilson
Alcohols, ketones, ethers Margules
C4-C6 Hydrocarbons Wilson
C7-C18 Hydrocarbons Wilson
Aromatics Margules
Assumption 1
The logarithm of the activity coefficient is assumed to be the sum of two
contributions:
Combinatorial given by the differences in size and shape between the molecules
in the mixture
Residual given by the energy interactions between the molecules in the
mixture
This is essentially the UNIQUAC model, and the logarithm of the activity coefficient
can then be written as Equation 72:
ln i = ln i
c
+ ln r
i 72
Assumption 2
The residual part of the activity coefficient is modeled by group contribution, which
is assumed to be the sum of individual contributions of each solute group in the
solution less the sum of individual contributions in the pure component
(normalization necessary to guarantee unit activity coefficients for pure
components). This is represented by equation 2:
where the summation is extended over all groups in the mixture. Γk is the residual
(i)
activity coefficient of group k in the mixture, and Γk is the activity coefficient of
group k in a mixture of groups corresponding to pure component i.
Assumption 3
The individual group contributions in any mixture containing atomic groups are
assumed to be a function of group concentration and temperature only.
UNIFAC Application
The UNIFAC method is a useful tool to get VLE calculations off the ground quickly,
and it provides accurate estimations for many industrially important applications.
Nevertheless, it is not a replacement for experimental data, and should never be
treated as such. For example, it is not possible to test for a priori accuracy when
using UNIFAC. Usually, if you are exploring a new process or idea it is useful to
use UNIFAC to perform preliminary screenings, but if you are serious about design
and your thermodynamic model had UNIFAC estimated interaction parameters a
critical review of the results is fundamental, and likely you will need to find the
necessary data in the literature or have it measured. Agarwal (Agarwal et al., 2000)
provide detailed guidelines on how to access VLE models and estimations.
When it is used ?
UNIFAC is called when interaction parameters between two molecules are missing.
If UNIFAC groups are available interaction parameters are estimated and a record of
the estimation is created.
2.50E+00
2.00E+00
1.50E+00
lnG1
lnG2
1.00E+00
5.00E-01
0.00E+00
0.00 0.20 0.40 0.60 0.80 1.00
Ethanol Mole Fraction
With the activity coefficient curves the interaction parameters for the actual
thermodynamic model of interest can be generated. In the example below the
interaction parameters correspond to Virtual Materials Margules:
2.50E+00
2.00E+00
1.50E+00 lnG1
lnG2
Marg-1
1.00E+00
Marg-2
5.00E-01
0.00E+00
0.00 0.20 0.40 0.60 0.80 1.00
Ethanol Mole Fraction
iIdx = 0;
jIdx = 1;
SeaGetBinaryPairInformation(&pkgNumber, "SeaStdNRTL", &iIdx, &jIdx,
&stringsize, refinfo, &retCode);
printf("\n Binary %d %d",iIdx, jIdx);
printf("\n %s", refinfo);
iIdx = 0;
jIdx = 2;
SeaGetBinaryPairInformation(&pkgNumber, "SeaStdNRTL", &iIdx, &jIdx,
&stringsize, refinfo, &retCode);
printf("\n Binary %d %d",iIdx, jIdx);
printf("\n %s", refinfo);
iIdx = 1;
jIdx = 2;
SeaGetBinaryPairInformation(&pkgNumber, "SeaStdNRTL", &iIdx, &jIdx,
&stringsize, refinfo, &retCode);
printf("\n Binary %d %d",iIdx, jIdx);
printf("\n %s", refinfo);
100 np
c
− UNIFAC
avg j = ∑
np i =1
ij ij
UNIFAC
74
ij
ssq = ∑∑ (ln )
2 np
c
ij − ln UNIFAC
ij 76
j =1 i =1
3.00E+00
2.50E+00
2.00E+00
lnG1
lnG2
1.50E+00
Wilson-1
Wilson-2
1.00E+00
5.00E-01
0.00E+00
0.00 0.20 0.40 0.60 0.80 1.00
Ethanol Mole Fraction
Keep in mind that mixtures which present phase splitting (for example benzene and
water) will have very large infinite dilution activity coefficients, and the chosen
Virtual Materials liquid phase model may not be able to fit the activity coefficients
for the entire concentration range, specially at the infinite dilution regions.
The assessment of the quality of UNIFAC generated activity coefficients is rather
difficult as ideally one would have to go back and compare the UNIFAC predicted
behavior against experimental data. Since UNIFAC's greatest usefulness is related to
estimating the behavior of mixtures when no data is available, there is not way of
guaranteeing a priori the accuracy of UNIFAC generated parameters. Nevertheless,
the items below offer a few pointers related to UNIFAC and its applicability. It
should be stressed that it is always recommended to replace interaction
parameters estimated using UNIFAC by interaction parameters determined
from experimental data.
The recommended temperature range for UNIFAC is between 250 to 425 K (Hansen
et al., 1991)
Conditions removed from the critical point.
Should not be applied for mixtures containing more than 10 functional groups
Detailed performance data for UNIFAC at conditions close to infinite dilution can be
found in Fredenslund et al. (1977) table 6.1
Usually UNIFAC performs best at intermediary mole fractions, removed from the
infinite dilution region.
Equations Of State
Virtual Materials has a comprehensive set of equations of state designed for process
calculations as the main property package (e.g. refinery applications, natural gas
applications) as well as to support activity coefficient models.
Basic Expressions
BP
z = 1+ 77
RT
B = ∑∑ yi y j Bij (T ) 78
i j
P
ln i = 2∑ y j Bij − B 79
j RT
note that unless expressly noted in the summation symbol, all summations are done
over all apparent species.
The combined metastable and bound contributions are available for Lennard-Jones
and Stockmayer potentials as a function of temperature, dipole moment and
structural parameters:
∆H ij
Bijmetastable + Bijbound = boij Aij exp( ) 82
Tij*
boij = 1.26184 3
ij 83
7243.8 i j
*
=
ij
( ij / k ) ij3 86
T
Tij* =
( ij / k )
87
The contribution coming from polar associating substances is added separately using
a similar form as the bound and metastable contribution:
1500
Bijchemical = boij Eij 1 − exp ij
88
T
The value of Eij is determined based on the value of the association parameter ij :
650
Eij = exp − 4.27 for < 4.5 89
ij ij / k + 300
ij
42800
Eij = exp
ij
− 4.27 for ≥ 4.5 90
ij / k + 22400
ij
A non-polar and a polar part give the free part of the second virial coefficient:
3 2.1 2.1
B free, polar, ij = − boij *'
ij 0.75 − *' + *'2 + *'3 5 93
T T T
4
Equation reported in Prausnitz et al (1980) is incorrect (+ sign in 1.015 coefficient)
5
Equation reported in Prausnitz et al (1980) is incorrect (0.75 instead of 0.74)
1 1
*'
= * − 1.6 ij 97
Tij Tij
( '
/ k ) = ( ii / k ) 1 − c1 1 −
(1 + c1 )
ii
99
2
ii = '
ii (1 + c2 )
1/ 3
100
( / k ) = Tci 0.748 + 0.91
'
−
0.4 ii
101
2 + 20
ii ii
ii
1
Tci 3
ii = (2.44 − ii )1.0133
'
Pci
102
1.7941E + 07 4
= ii
for ≥ 1.45 104
1.882 ii
ii
0.03 + ii
ii ii
16 + 400
c1 = ii
105
10 + 400 ii
3
c2 = 106
10 + 400 ii
ij =
1
2
( ii + jj ) 109
( / k )3
2
2 4
'
= i jj ii 7
≥ 2 and =0
( / k )' '6
for i j 110
ij ij
'
=0 for all other values 111
16 + 400
ci' = ij
112
10 + 400 ij
3
c2' = 113
10 + 400 ij
( ij / k ) = ( ij / k ) (1 + 'c '2 )
'
114
ij = '
ij (1 − ' '
2 c ) 115
Databases
The implementation of Hayden-O'Connell model required the creation of a new
interaction parameter database, SeaHaydenOConnell.mdb, to hold solvation
parameters and the Yaws.mdb database was modified to include an association
parameter as defined by Hayden and O'Connell. Association parameters for 92 pure
components are included as well as 4186 solvation parameters for binaries.
6
note that formula A-27 in Prausnitz et al (1980) is incorrect
7
note that formula A-29a in Prausnitz et al (1980) is incorrect
RT
P= 117
V −B
Using the ideal gas state at system P and T as the reference for the Helmholtz energy
we will have:
V
∆ A = RT ln − RT ln Z 118
V −B
RT dB V
∆S = − − R ln + R ln Z 119
V − B dT V −B
∆H = ∆A + T∆S + RT ( Z − 1) 120
and therefore
+ RT (Z − 1)
RT dB
∆H = − 121
V − B dT
where delta represents the difference between the actual and ideal gas property.
The following auxiliary functions are useful for the calculation of heat capacities:
∂P R RTB '
= + 122
∂T v V − B (V − B )
2
∂2 P 2 RT (B' )
2
2 RB ' RTB ''
2 = + + 123
∂T v (V − B ) (V − B)3 (V − B )2
2
2
∂P
∂T V
∆C p = ∆Cv − T 126
∂P
∂V T
dB
B' = 127
dT
d 2B
B '' = 128
dT 2
Assuming that we have the following chemical reaction in the vapor phase:
i + j ⇔ ij 129
fij
Kij = 130
fi f j
it is assumed that the gas mixture obeys the Lewis-Randall rule (i.e. it behaves like
an ideal mixture of real gases) and therefore one can write, for a mixture of real
gases obeying the second virial equation applied to physical interactions only:
where the superscript T means total species in the gas phase (i.e., all monomers,
dimers and solvates are present).
The fugacity coefficient should not take into account any "chemical" contribution
because we are now looking at an actual mixture of substances with their mole
fractions determined by chemical equilibria. Therefore the fugacity coefficients are
given simply by:
BiF P
ln T
i = 132
RT
The apparent fugacity coefficients to be used in VLE calculations are given by:
zi iT
i = 133
yi
In this report y always makes reference to apparent mole fractions (i.e. specified
ignoring dimerization) and z always makes reference to actual mole fractions (i.e.
calculated taking into account all chemical reactions).
− BijD (2 − )= 1 zij T
Kij =
ij ij
T T
134
RT P zi z j i j
− PBijD (2 − ) T T
Cij =
ij i j
T
135
RT ij
z ij = Cij zi z j 136
∑y i =1 137
i
∑ zi + ∑∑ zij = 1
i i j
138
and the following relationship between apparent and actual species holds:
i
z z
zi 1 + ∑ ij + ∑ ji
yi = j zi j = i zi
139
1 k
1 + ∑ ∑ zkj + ∑ z jk
2 k j j =k
where the summations are done over the number of apparent species excepted when
noted by the upper index.
i
i
i = ∑ zij + ∑ z ji = zi ∑ Cij z j + ∑ C ji z j 140
j i=j j j =i
with equations 64 and 65 one can create an iteration scheme for the calculation of
actual mole fractions:
1
yi 1 + ∑ n
k
zin+1 = 2 k 141
i
1 + ∑ Cij z nj + ∑ C ji z nj
j j =i
Assuming that the apparent number of moles is always 1, the actual number of moles
is given by:
BF P
z = nT + nT 143
RT
where the second virial coefficient is calculated according to Lewis and Randall's
assumption:
Since now we are assuming ideal mixture, the calculations for enthalpy, entropy and
heat capacity departure have to be modified accordingly:
RT dB F
∆H = −nT
P
+ RT ( Z − 1) 145
V − B dT
F
Note that the major contribution to the fluid enthalpy comes from the formation of
stable hydrogen bonds in the gas phase:
i
∆H D = −nT RT ∑∑ zij∆hijD 146
i j
∆ H = ∆H P + ∆H D 147
RT dBF V
∆S = nT − − R ln + R ln Z
P
148
V − B dT V −B
F F
∆H D
∆S =D
149
T
'
∂P R RTB F
= +
(V − B F )2
151
∂T v V − B
F
∂2P
2 =
2 RB F
'
+
2 RT B F( )+
' 2
RTB F
''
( ) (V − B ) (V − B )
152
∂T v V − BF
2 F 3 F 2
∂P RT
=−
(V − B F )2
153
∂V T
∂2P
V
∆ C = T ∫ 2 dV − R = −
F RT
( )
2B F ' + TB F '' + T B
F'
2
− R 154
v
∂T V − BF V − BF
∞
The isobaric heat capacity departure is then calculated using equation 51.
The additional heat capacity coming from the chemical reaction is calculated by
numerical differentiation of the enthalpy of dimerization:
∂∆H D ∆∆H D
∆C pD = ≅ 155
∂T ∆T
I
RD = 156
m
2
I= mR 2 157
5
2
RD = R 158
5
which corresponds to having the radius of gyration proportional to the cubic root of
the critical volume. This assumption holds well for normal alkanes as shown in the
figure below:
Radius of Gyration
y = 1.06656E-09x - 3.84667E-10
6.00E-10
R2 = 9.98728E-01
5.00E-10
4.00E-10
RD
3.00E-10
2.00E-10
1.00E-10
0.00E+00
0.4 0.5 0.6 0.7 0.8 0.9
vc^1/3
1
RD = 1.06656 E − 09V − 3.84667 E − 10
c
3
159
where the critical volume is expressed in m3/kmol and the radius of gyration is
expressed in m.
For example, for benzene the critical volume is 0.258 m3/kmol and the estimated RD
is 2.94e-10 m, while the accepted value is 3.00e-10 m.
1.000
0.950
Compressibility Factor
0.900
0.850
0.800
Z-exp
0.750
Z-VMG
0.700
0.650
0.600
0.550
0.500
250.00 300.00 350.00 400.00 450.00 500.00 550.00
Temperature (K)
1.000
0.950
Compressibility Factor
0.900
0.850
0.800
Z-exp
0.750
Z-VMG
0.700
0.650
0.600
0.550
0.500
0.10 1.00 10.00 100.00 1000.00 10000.00
Pressure (kPa)
400.00
Heat Capacity (kJ/kmol.K)
350.00
300.00
Cp-exp
250.00
Cp-VMG
200.00
150.00
100.00
350.00 400.00 450.00 500.00 550.00
Temperature (K)
400.00
300.00
Cp-exp
250.00
Cp-VMG
200.00
150.00
100.00
20.00 40.00 60.00 80.00 100.00 120.00
Pressure (kPa)
1.04
1.02
Fugacity Coefficient
1.00
0.98 phiEthylene
phiEthanol
0.96
0.94
0.92
0.90
0 0.2 0.4 0.6 0.8 1
Ethylene mole fraction
1
0.99
0.98
Fugacity Coefficient
0.97
0.96
phiAcetone
0.95
phiCC6one
0.94
0.93
0.92
0.91
0.9
0 0.2 0.4 0.6 0.8 1
Acetone mole fraction
2.00
1.80
Fugacity Coefficient
1.60
1.40
1.20
phiAcetal
1.00
0.80 phiAcetic
0.60
0.40
0.20
0.00
0 0.2 0.4 0.6 0.8 1
Acetaldehyde mole fraction
Polar / Polar with weak association (acetaldehyde) and dimerization (acetic acid)
0.45
0.43
0.41
Fugacity Coefficient
0.39
0.37
phiFormic
0.35
phiAcetic
0.33
0.31
0.29
0.27
0.25
0 0.2 0.4 0.6 0.8 1
Formic Acid Mole Fraction
Zfactor
Z =1 161
Volume
RT
V = 162
P
Mass Density
PMwavg
= 163
RT
Fugacity Coefficient
i =1 164
Activity Coefficient
i =1 165
Enthalpy
nc T
H = H 0 = ∑ yi ∫ C opi dT 167
i =1 0
Entropy
nc T
C opi nc
H = S 0 = ∑ yi ∫ dT − R ∑ yi ln yi 168
i =1 1
T i =1
Ideal Gas Cp
Same as Cp
Residual Enthalpy
∆H = 0 171
Residual Entropy
∆S = 0 172
Residual Cv
∆ Cv = 0 173
Residual Cp
∆C p = 0 174
Isothermal Compressibility
1
Iso = 176
P
(dPdV)T
∂P RT
=− 2 177
∂V T V
Mechanical Zfactor
Undefined, set to EMPTY code (-12321)
(dVdT)P
∂V 1
= 180
∂T P T
Poynting Factor
POYi = 1 181
Saturated Fugacity
f i sat = Pi sat 182
Vapor Pressure
Defined by property package empirical vapor pressure equation (MAS)
aP
A= 186
(RT )2
bP
B= 187
RT
Basic Model u w ac b
RK 1 0 Ω A R 2Tci2.5 ΩB RTci
Pci Pci
SRK 1 0 Ω A R 2Tci2 ΩB RTci
Pci Pci
PR 2 -1 Ω A R 2Tci2 ΩB RTci
Pci Pci
a (
2 Z + B u − u2 − 4w ) − RT ln (Z − B )
2 Z + B (u + u − 4w )
∆A = ln 188
b u − 4w2 2
da 2Z + B(u − u − 4w )2
∆S = R ln (Z − B ) −
1
u − 4 w dT 2Z + B(u + u − 4w )
ln 189
2 2
b
∆H = ∆A + T∆S + RT ( Z − 1) 190
∂2P
V
∆Cv = T ∫ 2 dV − R 191
∞
∂T
2
∂P
∂T V
∆C p = ∆Cv − T 192
∂P
∂V T
∂P RT da 1
= − 193
∂T V V − b dT V + ubV + wb
2 2
∂P RT 2a(V + b)
=− 2 +
(V − b) (V + ubV + wb 2 )2
194
∂V T 2
1 1 ∂Z
Iso = − (Isothermal Compressibility) 196
P Z ∂P T
1 ∂V
= (coefficient of thermal expansion) 197
V ∂ T P
Redlich-Kwong
The Redlich-Kwong (Redlich and Kwong, 1949) equation of state was an important
technical advancement over the van der Waals equation of state and was extensively
used for process calculations, specially in combination with the Chao-Seader and
Grayson-Streed models for distillation of heavy hydrocarbons.
Equation Settings
Omega A Omega B Alpha (T) Mechanical Zc
0.42748 0.08664 1 1
T 3
Mixing Rules
a = ∑∑ (1 − kij ) aia j xi x j
b = ∑ xibi 198
kij1
k ij = k ijo + + k ij2 ln T
T
Soave-Redlich-Kwong (SRK)
The Soave-Redlich-Kwong (Soave, 1972) equation of state accurately models the
vapor pressures of light hydrocarbons and light gases at pressures at or above
atmospheric.
Equation Settings
Omega A Omega B Alpha (T) Mechanical Zc
0.42748 0.08664
(1 + f (1 −
wi Tri ))
2 1
3
Mixing Rules
a = ∑∑ (1 − k ij ) ai a j xi x j
b = ∑ xi bi 200
k ij1
k ij = k ij + + kij ln T
o 2
Refinery-SRK
The SRK implementation suggested by the API (API, 1992) uses the basic structure
defined by the original SRK equation of state (Soave, 1972) with the modifications
described below.
Equation Settings
Omega A Omega B Alpha (T) Mechanical Zc
0.42747 0.08664
(1 + f (1 −
wi Tri ))
2
1
3
Mixing Rules
a = ∑∑ (1 − kij ) aia j xi x j
b = ∑ xibi 202
kij1
k ij = k ijo + + k ij2 ln T
T
Equation Settings
Omega A Omega B Alpha (T) Mechanical Zc
0.42747 0.08664 (T ) 2 1
3
T
i =1− 205
Tc, i
The constants c1, c2 and c3 are determined using the Mathias-Copeman procedure
(Mathias and Copeman, 1983) in such a way as to minimize the error between the
experimental and calculated vapor pressures of pure components.
The attractive term for the PSRK equation of state is defined as:
g oE nc ai RT nc
b
a = b + ∑ xi + ∑ x b
i 206
A1 i =1 bi A1 i =1 i
The repulsive term is calculated using the standard formulation used by SRK:
RTc , i
bi = 0.08664 208
Pc ,i
nc
b = ∑ xi bi 209
i =1
8
Main Group Name
9
Sub Group Name
10
Main Group Code
11
Sub Group Code
The PSRK method is quite useful for preliminary design of chemical systems with
inerts and high and low pressures, such as the synthesis and separation sections of a
chemical facility. Please note that the SRK predicted volume is translated using
Peneloux method, and the Peneloux volume translation constants are calculated in
such a way as to match the liquid molar volume as predicted by VMG's Rackett
equation at 60 F for components with a critical temperature above 60 F or to match
the normal boiling point for components with critical temperatures below 60 F. The
only exception is water, since it uses the liquid density of water at 60 F from Perry's
Chemical Engineer's Handbook 6th Edition, pag. 3-75.
Association Alert: PSRK does not handle systems with strong association in the gas
phase, such as systems containing carboxylic acids.
Polymer Alert: PSRK does not handle well mixtures of highly asymmetric
mixtures, such as solvents and polymers.
UNIFAC Alert: PSRK has the same limitations as the UNIFAC method. As a
guideline, it should be used for preliminary design only.
Equation Settings
Omega A Omega B Alpha (T) Mechanical Zc
Mixing Rules
a = ∑∑ aij x i x j
211
b = ∑ xi bi
Component i Component j
Nitrogen Methane
Carbon Dioxide
Ethane
Hydrogen Sulfide
Propane
Methane Carbon Dioxide
Ethane
Hydrogen Sulfide
Propane
Technical Bulletin Available. If you need more details on this property package
please contact Virtual Materials Group and request Technical Bulletin 037.
Phase Equilibria Alert: The ZSRK model should be used for natural gas
calculations only. For detailed phase equilibria of complex mixtures such as the ones
found in downstream processing or refinery you should consider using REFINERY-
SRK or APR.
Custody Alert: The ZSRK model was designed using data from 0 to 1500 psia,
temperatures from 70 to 250 F with an average error in compressibility factor of
0.3% and a maximum error of 1.1% where the proportions of nitrogen, carbon
dioxide, hydrogen sulfide and hydrocarbons are in the usual range of natural gas
mixtures. Considerably higher errors can be expected for highly sour (i.e. high
content of carbon dioxide and/or hydrogen sulfide).
Peng-Robinson (PR)
The Peng-Robinson (Peng and Robinson, 1976) equation of state was designed to
provide better liquid densities than the SRK equation of state as well as better vapor
pressures for hydrocarbons in the gasoline range.
Equation Settings
Omega A Omega B Alpha (T) Mechanical Zc
0.457235 0.0777969
(1 + f (1 −
wi Tri ))
2 0..307
Mixing Rules
a = ∑∑ (1 − kij ) aia j xi x j
b = ∑ xibi 214
kij1
k ij = k ijo + + k ij2 ln T
T
Advanced-Peng-Robinson (APR)
The Advanced Peng-Robinson model has all the characteristics of the Peng-
Robinson model plus volume translation for accurate liquid phase density estimation
using the Peneloux (1980) idea as modified by Mathias et al. (1991).
Equation Settings
Omega A Omega B Alpha (T) Mechanical Zc
0.457235 0.0777969
(1 + f (1 −
wi Tri ))
2 0..307
Mixing Rules
a = ∑∑ (1 − kij ) aia j xi x j
b = ∑ xibi 217
kij1
k ij = k ijo + + k ij2 ln T
T
Volume translation
The Peng-Robinson model does not provide an accurate enough model for liquid
densities, especially for polar compounds. Therefore, a volume translation as
suggested by Peneloux (1980) and further modifications suggested by Mathias et al.
(1991) is used in APR In addition, as special mixing rule for the calculation of
excess volume translations is used, thus allowing the calculation of accurate mixture
densities. The Mathias modification for the calculation of liquid densities is given by
the following equations:
0.41
V corr = V + s + f c 218
0.41 +
V2 ∂P
=− 219
RT ∂V T
f c = Vc − (3.946b + s ) 220
r
smix = ∑ si xi + s Excess (T , x ) 221
i
NGL Peng-Robinson
The Peng-Robinson (Peng and Robinson, 1976) equation of state was designed to
provide better liquid densities than the SRK equation of state as well as better vapor
pressures for hydrocarbons in the gasoline range. The GPA-PR model implements
the exact definition of the Peng-Robinson equation of state as specified by the GPA
RR-28 (Robinson and Peng, 1978) , including the GPA recommended methods for
the estimation of pure component physical properties for oil fractions.
Equation Settings
Omega A Omega B Alpha (T) Mechanical Zc
0.457235 0.0777969
(1 + f (1 −
wi Tri ))
2 0..307
Mixing Rules
a = ∑∑ (1 − kij ) aia j xi x j
b = ∑ xibi 224
kij1
k ij = k ijo + + k ij2 ln T
T
Huron-Vidal-Peng-Robinson (GEPR)
The Huron-Vidal Peng-Robinson model has a different mixing rule designed to
incorporate a Gibbs Excess model into the equation of state as described by Huron
and Vidal (1979) and in principle model mixtures with strong non-idealities in the
liquid phase. It also incorporates the volume translation from Advanced Peng-
Robinson as discussed in the Advanced Peng-Robinson section.
Equation Settings
Omega A Omega B Alpha (T) Mechanical Zc
0.457235 0.0777969 (1 + f (T ))2 0..307
( ) (
f (T ) = 1 + Ai 1 − Tri + Bi 1 − Tri ) + C (1 −
2
i Tri )
3
225
The constants A, B and C are determined for each individual compound by fitting the
vapor pressure over the range defined by the vapor pressure correlation.
Extrapolation beyond the critical point is done using the method suggested by
Mathias (1980).
Mixing Rules
nc a
a = b ∑ i xi − G∞E
i=1 bi 226
b = ∑ xi bi
Where GE is given by the NRTL model and is a numerical constant for Peng-
Robinson.
Volume translation
Same as in Advanced Peng-Robinson
Environmental Alert: GEPR can provide accurate LLE predictions for water and
hydrocarbon rich phase as long as interaction parameters are provided. If you are
making environmental impact studies make sure the appropriate interaction
parameters are used
Steam95
The Steam95 property package provides convenient access to rigorous physical
properties of water as defined by NIST/ASME (Harvey et al., 1997). In addition to
providing accurate physical properties for water, Virtual Materials Group extended
the original NIST/ASME formulation to include other components, including
hydrocarbons and water decantation.
Fugacity Calculations
It is assumed that the components in a mixture when using the Steam95 property
package form an ideal solution with an ideal gas phase:
ln i
v
=0
227
ln i
l
= ln Pvi − ln P
If the water partial pressure in the system exceeds the water vapor pressure, a pure
water phase is formed, assuming that water will have a solubility equal to the
solubility of water in kerosene as defined in the API handbook. Fugacity coefficients
of components with critical temperatures below 25 C in the hydrocarbon phase are
calculated using the Grayson-Streed (1963) correlation. Accurate solubilities of
industrially important light gases in water are computed using the Harvey and
Levelt-Sengers (1990) correlation. The gases are Hydrogen, Helium, Neon, Argon,
Krypton, Xenon, Nitrogen, Oxygen, Methane and Ethane.
Thermal Calculations
Calculations for enthalpies, entropies and heat capacities are conducted assuming
that the solution is ideal. If M is any of the physical properties above, then M for the
mixture is defined as:
nc
M = ∑ M i xi 228
i =1
where M for water is calculated using the NIST/ASME formulation and Advanced
Peng-Robinson for all other components by computing M for pure component i at
system P, T and desired phase.
Volumetric Calculations
Calculations for compressibility factors, volumes and densities are defined using
equation 216:
where Zi for water is calculated using the NIST/ASME formulation and Advanced
Peng-Robinson for all other components by computing V for pure component i at
system P, T and desired phase.
This property package should not be used for environmental studies since it always
predicts a virtually pure water phase for subcooled water/hydrocarbon systems,
except for systems containing Methane or Ethane.
Chao-Seader
The Chao-Seader property package is Virtual Materials extension of the classic
Chao-Seader (Chao and Seader, 1961) model used for the simulation of hydrocarbon
mixtures. The Chao-Seader property package is included in the Virtual Materials
Property Package system for completeness and compatibility with older simulations.
It is recommended that the Grayson-Streed, Advanced Peng-Robinson or
RefinerySRK models be used for new calculations instead of Chao-Seader.
Fugacity Calculations
The gas phase fugacities are calculated using the Redlich-Kwong equation of state
and the liquid phase fugacities are calculated using the Chao-Seader correlation,
If the water partial pressure in the system exceeds the water vapor pressure, a pure
water phase is formed, assuming that water will have a solubility equal to the
solubility of water in kerosene as defined in the API handbook. Accurate solubilities
of industrially important light gases in water are computed using the Harvey and
Levelt-Sengers (1990) correlation. The gases are Hydrogen, Helium, Neon, Argon,
Krypton, Xenon, Nitrogen, Oxygen, Methane and Ethane. The solubility of H2S in
water is approximated using Virtual Materials H2S Henry law.
Thermal Calculations
Identical to Steam95
Volumetric Calculations
Identical to Steam95.
This property package should not be used for environmental studies since it always
predicts a virtually pure water phase for subcooled water/hydrocarbon systems,
except for systems containing Methane or Ethane.
Grayson-Streed
The Grayson-Streed property package is Virtual Materials extension of the classic
Grayson-Streed (Grayson and Streed, 1963) model used for the simulation of
hydrocarbon mixtures.
Thermal Calculations
Identical to Steam95.
Volumetric Calculations
Identical to Steam95.
This property package should not be used for environmental studies since it always
predicts a virtually pure water phase for subcooled water/hydrocarbon systems,
except for systems containing Methane or Ethane.
T < 260 C
For hydrocarbons expect methane 0.5 < Tri < 1.3 and mixture critical pressure < 0.8
For systems containing methane and/or hydrogen, molal average Tr < 0.93, and
methane mole fraction < 0.3. Mole fraction of other dissolved gases < 0.2
Viscosity
The mixture viscosity is calculated using an empirical correlation for the liquid and
gas phase.
Special Components
The viscosity of water is calculated using the recommended NIST formulation for
the steam tables (1997). If at the pressure, temperature and phase specified the NIST
formulation can not calculate the desired value of the property values from DIPPR
(1999) are used.
i
o
= Yaws
i (T ) 231
i
p
= i
p
(T , P ) 232
The Yaws expression for viscosity is defined in the Yaws Database documentation.
The pressure correction term is calculated using the Reichenberg method (Reid et al.,
1987)
i = Yaws
i (T ) 234
Thermal conductivity
The mixture thermal conductivity is calculated using an empirical correlation for the
liquid and gas phase.
Special Components
The thermal conductivity of water is calculated using the recommended NIST
formulation for the steam tables (1997). If at the pressure, temperature and phase
specified the NIST formulation can not calculate the desired value of the property
values from DIPPR (1999) are used.
k i = k iYaws (T ) 236
The pressure correction term is calculated using the Stiel and Thodos method (Reid
et al., 1987) while the A ij term is calculates using the Mason and Saxena method
(Reid et al., 1987)
−1
1 1
= 2 + 238
ij
k k
i j
And the factors are calculated according to Li's method (Reid et al, 1987)
Surface Tension
The mixture surface tension is calculated using an empirical correlation for the liquid
phase.
Special Components
The surface tension of water is calculated using the recommended NIST formulation
for the steam tables (1997). If at the pressure, temperature and phase specified the
NIST formulation can not calculate the desired value of the property values from
DIPPR (1999) are used.
i = Yaws
i (T ) 240
YawsThermal
The YawsThermal abstract property package provides enhanced enthalpies and heat
capacities for the liquid phase by using empirical expressions for liquid heat
capacity. The following definitions are important to understand how thermal
properties are calculated.
Liquid Enthalpy
The liquid enthalpy is calculated as follows:
H L = H 0 + ∆H 241
nc
∆ H = − ∑ xi ∆H i
L L
242
i =1
∆H iL = ci + ∫ C piL dT 243
0
with the constant ci calculated to match exactly the enthalpy of the gas phase less
than the enthalpy of vaporization of component i at the reference temperature.
Molecular Weight
Arbitrarily set to 30,000.0 kg/kmol
Critical Temperature
Set equal to 1800.0 K (Orbey et al., 1998)
Critical Volume
Calculated using the definition of critical volume of ideal fluid with Zc of 0.27:
RZ cTc
Vc = Equal to 4.04031240 m3/kmol
Pc
Acentric Factor
Orbey et al. (1998) set all Mathias-Copeman constants to zero. In order to have
Peng-Robinson behaving in the same way the function:
Antoine Constants
The vapor pressure for solids is set arbitrarily to 1e-20 kPa when calculated using the
Antoine expression. The temperature limits for the Antoine expression are set
between 10 to 5000 K, and all the constants are set to zero, except the A constants
which is set to –4.605170e+01
Mathias-Copeman Constants
A, B and C set equal to 0.0 (Orbey et al., 1998)
45
40
35
30
25
20
15
10
0
250 300 350 400 450 500 550 600 650
-CH2- Mer ideal gas molar heat capacity (kJ/kmol.K) as a function of temperature
Solubility Parameter
The solubility parameter is set to the value reported to n-C12, equal to 1.60000E+04
in SI units
Rackett Constants
Estimated using the standard liquid density values and estimated critical properties
SeaLiquidVolume
Estimated using the Rackett Constants
Dipole Moment
Set to zero
Surface Tension
The surface tension is set to a constant value equal to 1 N/m
Gas Viscosity
The gas viscosity is set to a constant equal to 1e-4 Pa.s
Solid Density
The solid density is estimated by using the density of low-density polyethylene at
298.15 K, equal to 880 kg/m3.
Density
The solid phase density is computed by averaging the molar volumes using the
following formula:
∑ mw x i i
= i
244
mwi
∑
i
xi
i
Volume
The solid phase molar volume is computed by averaging the molar volumes using
the following formula:
mwi
V =∑ xi 245
i i
Compressibility
The solid compressibility is calculated using the definition of compressibility factor:
PV
Z= 246
RT
Solid Enthalpy
The solid enthalpy is calculated using the solid heat capacity without any pressure
correction using the formulae below:
H = ∑ H i xi 248
i
( Ai + BiT + CiT 2 )
( )
T
dT = Ai ln T + Bi (T − 1) + Ci T 2 − 1
1
Si = ∫ 249
1
T 2
S = ∑ Si xi − R ∑ xi ln xi 250
i i
Cp = ∑ Cpi xi 252
i
Fugacity Coefficient
Fugacity coefficients for solid components in the solid phase are set to 1.0e-38.
Solid Fugacity
Solid fugacities are calculated using the definition of fugacity below.
f i = i xi P . The fugacity coefficients are computed using the values defined in the
Fugacity Coefficient item.
Activity Coefficients
Since the fugacity coefficient of a component is independent of composition as
defined in the Fugacity Coefficient item, it follows that the activity coefficients are
equal to 1.0
f i = i P . The fugacity coefficients are computed using the values defined in the
Fugacity Coefficient item.
Residual Enthalpy
Residual enthalpies are calculated using the difference between the solid enthalpy
and the ideal gas enthalpy. Note that no equation of state is used in the calculations.
Residual Cp
Residual enthalpies are calculated using the difference between the solid Cp and the
ideal gas Cp. Note that no equation of state is used in the calculations.
Residual Cv
Residual enthalpies are calculated using the difference between the solid Cv and the
ideal gas Cv. Note that no equation of state is used in the calculations.
Solid Viscosity
Solid viscosity is set to 10,000 poise, and is independent of composition.
Surface Tension
Solid Surface tension is set to 1 N/m and is independent of composition
Thermal Conductivity
The solid thermal conductivity is calculated using the formulae below:
= ∑ i xi 254
i
Speed of sound
The speed of sound is set to 1950 m/s, corresponding to the value in polyethylene
referent to plane longitudinal wave in bulk material (Lide, 1990)
Isothermal Compressibility
1
The isothermal compressibility is set to since the compressibility factor is
P
assumed to be pressure independent.
∂P
Calculation
∂V T
The derivative is set to 1e38 to reflect the incompressible nature of the solid phase
Virtual Materials physical property system can defined pure component properties
that do not have correspondent in the pure component database. From the pure
component physical property estimator these pure component properties are as valid
as pure component properties coming from the pure component database, and some
keywords you find in the items below may not be in the pure component database
keyword definition.
StandardLiquidDensity
If LiquidDensity@298 is available then StandardLiquidDensity is set to
LiquidDensity@298
Else if RackettLiquidDensity parameters are available then calculate the liquid
density using the Rackett parameters and the following criteria:
If CriticalTemperature < 298.15 then use T = NormalBoilingPoint for calculations
Else use T = 298.15 K
Else
Use the Katz-Firoozabadi correlation to estimate the liquid density at 60 F.
LiquidDensityRefT
If StandardLiquidDensity was determined using LiquidDensity@298 or component
is sub-critical at 298.15 then LiquidDensityRefT = 298.15
Else if component is supercritical at 298.15 then LiquidDensityRefT =
NormalBoilingPoint
Else LiquidDensityRefT = 288.705556 K
MolecularWeight
Calculate using Twu's method (Twu, 1985)
12
Value correspond to all ideal gas enthalpy of formation constants unknown
13
Value correspond to all ideal gas Gibbs free energy of formation constants unknown
14
Default value for activity coefficient models
CriticalTemperature
Calculate using the Lee-Kesler correlation (Lee and Kesler, 1975)
CriticalPressure
Calculate using the Lee-Kesler correlation (Lee and Kesler, 1975)
CriticalVolume
Calculate using the API method 14A3.1 (API, 1992)
CriticalCompressibility
PcVc
Zc = 255
RT c
AcentricFactor
Calculate using the Lee-Kesler correlation (Lee and Kesler, 1975)
SRKAcentricFactor
Set to be identical to AcentricFactor
Poly5IdealGasCpA
Note that all necessary Calculated using the Lee-Kesler correlation (Lee and Kesler, 1975)
polynomial coefficients will
also be calculated
Antoine5LogA or DIPPRAntoine0
Note that all necessary Calculated using the Lee-Kesler correlation (Lee and Kesler, 1975)
Antoine coefficients will also
be calculated
PengRobinsonMathiasCopemanA
Calculated using Antoine5Log coefficients and VIRTUAL MATERIALS GROUP
proprietary regression method.
COSTALDCharacteristicVolume
Calculated to match liquid density at reference temperature
vanderWaalsArea
If chemical structure is available calculate vanderWaalsArea using UNIFAC
contributions.
Else estimate vanderWaalsArea using proprietary molecular weight correlation
UNIQUACR
Calculate based on vanderWaalsVolume and polyethylene mer value
UNIQUACQ
Calculate based on vanderWaalsArea and polyethylene mer value
UNIQUACL
l = 5.0 (R - Q) - (R - 1.0) 256
SolubilityParameter
Estimate using enthalpy of vaporization from Pitzer correlation (Reid et al, 1987)
∆U vapourization
= 257
V
∆ U vapourization = ∆ H vapourization − RT 258
RackettLiquidDensityA
Note that the remaining Calculated using the LiquidDensityRefT, StandardLiquidDensity and
Rackett parameters will also CriticalVolume
be calculated
SeaLiquidVolumeA
Note that the remaining Fast saturated liquid volume correlation for dynamic calculations. Use Rackett
SeaLiquidVolume parameters correlation to generate table of density versus temperature points.
will also be calculated
2
V = A + B1 − + C 1 −
T T
259
Tc Tc
DipoleMoment
Set to zero by default
SurfaceTensionA
Note that the remaining Estimated using the Brock and Bird correlation (Reid et al, 1987)
SurfaceTension parameters
will also be calculated
GasViscosityA
Note that the remaining Estimated using Chung's low pressure viscosity estimation method (Reid et al, 1987)
GasViscosity parameters will
also be calculated
LiquidViscosityA
Note that the remaining Estimated using the Letsou-Stiel correlation (Reid et al, 1987)
LiquidViscosity parameters
will also be calculated
GasThermalConductivityA
Note that the remaining Estimated using Chung's low pressure thermal conductivity method (Reid et al,
GasThermalConductivityPara 1987)
meters will also be calculated
IdentificationNumber
Arbitrary number identifies a component. The user is encouraged to use the CASN
number as a more appropriate identifier. Appendix A has a complete list of names,
CASN and identification numbers used by Virtual Materials
Formula
Chemical Formula.
Name
Common name of chemical compound
MolecularWeight
Compound molecular weight in kg/kmol
FreezingPoint
Compound Freezing Point in K
NormalBoilingPoint
Compound Normal Boiling Point at 1 atm in K
CriticalTemperature
Compound critical temperature in K
CriticalPressure
Compound critical pressure in kPa
CriticalDensity
Compound critical density in kg/m3
CriticalCompressibility
Compound critical compressibility
AcentricFactor
Compound Acentric Factor as defined by Pitzer
SolidDensity
Compound solid density at specific temperature given by SolidDensityTemperature
LiquidDensity@298
Compound liquid density at 25 C
LowerExplosionLimitVol
Lower explosive limit in air expressed in volume %
UpperExplosionLimitVol
Upper explosive limit in air expressed in volume %
ExplosionLimitStatus
Upper and lower explosive limit in air status, experimental or estimated
FlashPointTemperature
Flash point temperature in air, K
FlashPointStatus
Flash point status, experimental or estimated
AutoIgnitionTemperature
Auto ignition temperature, K
AutoIgnitionTemperatureStatus
Auto ignition temperature status, experimental or estimated
PermissibleExposureLimitOSHAppm
Permissible exposure limit in air in workplace of the Occupational Safety and Health
Administration, ppm (volume)
PermissibleExposureLimitOSHAmgm3
Permissible exposure limit in air in workplace of the Occupational Safety and Health
Administration, mg/m3
RecommendedExposureLimitNIOSHmgm3
Permissible exposure limit in air in workplace of the National Institute for
Occupational Safety and Health, mg/m3
ImmediatelyDangerousToLifeHealthppm
Threshold limit value in air in workplace of the American Conference of
Governmental Industrial Hygienists, ppm (volume)
ImmediatelyDangerousToLifeHealthmgm3
Threshold limit value in air in workplace of the American Conference of
Governmental Industrial Hygienists, mg/m3
EnthalpyOfCombustionAt298
Enthalpy of combustion at 25 C. It is defined as the net increase in heat content when
a substance in its standard state at ambient conditions (77 F and 1 atm) undergoes
complete oxidation. The products of combustion are CO2(gas), H2O(gas), F2(gas),
Cl2(gas), Br2(gas), I2(gas), N2(gas), SO2(gas), H3PO4(solid) and
SiO2(crystobalite)
EnthalpyOfCombustionState
State of compound going under combustion (gas, liquid, solid)
CASN
The Chemical Abstracts Chemical Abstracts Service Number
Service Number is an
universal identifier for
chemical species. It is used by
Virtual Materials to toggle
special handling for
components such as water.
RadiusofGyration
Radius of gyration in m
SolubilityParameter
Hildebrand's solubility parameter in (J/m3)^0.5
vanderWaalsArea
van der Waals area in m2/kmol
vanderWaalsVolume
van der Waals volume in m3/kmol
RefractiveIndex
Refractive index at 25 C
ChemicalFamily
Chemical family definition using the American Institute of Chemical Engineers
codes defined by the DIPPR project 801.
The family codes are defined below.
Family Chemical Type Family Chemical Type
Code Code
AA n-Alkanes US Unsaturated
Aliphatic Esters
MA Methyl Alkanes RS Aromatic Esters
DA Dimethyl AT Aliphatic Ethers
Alkanes
OA Other Alkanes OT Other
Ethers/Diethers
CA Cyclo Alkanes CT Epoxides
GA Alkyl Cyclo TT Peroxides
Pentanes
HA Alkyl Cyclo AH C1/C2 Aliphatic
Hexanes Chlorides
RA Multiring Cyclo HH C3 & Higher
Alkanes Aliphatic
Chlorides
AE 1-Alkenes RH Aromatic
Chlorides
BE 2,3,4-Alkenes VH C,H,Br
Compounds
ME Methyl Alkenes WH C, H, I
Compounds
EE Ethyl and FH C, H, F
Higher Alkenes Compounds
CE Cyclo Alkenes PH C, H
Multihalogen
C 0p = A + BT + CT 2 + DT 3 + ET 4 260
B
ln P = A + + C ln T + DT + ET 2 262
T
Surface Tension
The surface tension is calculated using a simple 3 constant equation. The necessary
coefficients are accessed as follows:
Virtual Materials Keyword Coefficient
SurfaceTensionA A
SurfaceTensionB B
SurfaceTensionN N
C p = A + BT + CT 2 + DT 3 265
∆H of = A + BT + CT 2 266
∆G of = A + BT + CT 2 267
B C D
C op = A + + 2 + 3 268
T T T
C p = A + BT + CT
s 2
B
ln = A+ + CT + DT 2 269
T
Gas Viscosity
Gas viscosities are calculated using a simple polynomial. The necessary coefficients
are accessed as follows:
Virtual Materials keyword Coefficients
GasViscosityA A
GasViscosityB B
GasViscosityC C
GasViscosityTmin Minimum Temperature
GasViscosityTmax Maximum Temperature
= A + BT + CT 2 270
= A + BT + CT 2 272
= A + BT + CT 2 273
= A + BT + CT 2 274
B C
ln S = A + + 275
T T2
ln S = A + BX + CX 2 276
T in K and D in m2/s
D = A + BT + CT 2 277
T in K and D in m2/s
B
log D = A + 278
T
T in K, B in 1/K
M
B = A1 −
T
279
Tc
UNIFAC structure
The UNIFAC structure is defined in the UNIFACStruct field as a string using the
following shorthand:
(Number Of Groups)i (Group Type)i .
Example: Ethanol
1 1 1 2 1 15
The definition of UNIFAC groups currently supported by Virtual Materials can be
found below.
Electric Charge
The electric charge property holds the electric charge of an ion and can be accessed
using the keyword ElectricCharge. This property is usually defined only for
electrolyte based property packages such as UREA++.
Nomenclature
A, B, C, D Henry coefficient constants
A Modified reduced dipole moment
A Helmholtz energy
B Second Virial Coefficient, m3/kmol
bo Characteristic molecular volume
c1, c2 Functions of non-polar acentric factor
C Auxiliary quantity defined by equation 59
Superscripts
bound Contribution to the second virial coefficient fro m bonding
potential energy
chemical Contribution to the second virial coefficient from
chemical reactions (dimerization)
D Contribution to the second virial coefficient from
chemical and reaction interactions
D Dimerization interactions
F, free Contribution to the second virial coefficient from physical
interactions
I,II,… Phase
Subscripts
c Critical
free Contributions from non chemical factors for second virial
coefficient
i,j,k Component index
ij Associated substance ij (ij i is the same as j we have a
dimer)
non-polar Non-polar contribution to the second virial coeffic ient
np Number of Phases
polar Polar contribution to the second virial coefficient
ref Reference value
* Reduced by ( ij / k ) or ( ij /k) 3
ij
Greek Symbols
ij Kroenecker delta
Characteristic energy
Fugacity coefficient
Activity coefficient
Bibliography
Abrams, D.S. and Prausnitz, J.M.; AIChE Journal, 21, 116, 1975
Agreda, V.H. and Zoeller, J.R.; "Acetic acid and its derivatives"; Marcel Dekker,
Inc. (1993)
Agarwal, R.K.; Vieler, A.; Li, Y-K.; Santollani, O.J. and Satyro, M.A.; Accepted for
publication on CEP (2000)
American Petroleum Institute; "Technical Data Book – Petroleum Refining"; Fifth
Edition, May 1992
American Petroleum Institute; "A New Correlation of NH3, CO2 and H2S Volatility
Data from Aqueous Sour Water Systems", API Publication 955, March 1978
Bokis, C.P.; Orbey, H. and Chen, C.C.; "Properly Model Polymer Processes"; CEP
April, 1999
Carlson, E.C.; "Don't Gamble with Physical Properties for Simulations"; CEP,
October (1996)
Carroll, J.J; "What's Henry's Law"; CEP, September (1991)
Carroll, J.J.; "Use Henry's Law for Multicomponent Mixtures"; CEP, August (1992)
Carroll, J.J.; "Henry's Law – A Historical View"; Journal of Chemical Education; 70,
n. 2; February (1993)
Carroll, J.J.; "Henry's Law Revisited"; CEP; January (1999)
Chao, K.C. and Seader, J.D.; “A General Correlation of Vapor-Liquid Equilibria in
Hydrocarbon Mixtures”; AIChE Journal December 1961
http://www.edl-inc.com/Plastic%20expansion%20rates.htm
Fisher, K. and Gmehling, J.; “Further development, status and results of the PSRK
method for the prediction of vapor-liquid equilibria and gas solubilities”; Fluid Phase
Equilibria, 121(1996) 185-206
Fredenslund, Aa., Gmehling, J. and Rasmussen, P.; "Vapor-Liquid Equilibria Using
UNIFAC – A Group-Contribution Method"; Elsevier Scientific Publishing
Company; Amsterdam, 1977
Gmehling, J. and Onken, U.; "Vapor-Liquid Equilibrium Data Collection",
DECHEMA, 1977 onwards.
Gmehling, J.; Li, J. and Fisher, K.; “Further developments of the PSRK model for
the prediction of gas solubilities and vapor-liquid equilibria at low and high
pressures II”; Fluid Phase Equilibria 141(1997) 113-127
Grayson, H.G. and Streed, C.W.; "Vapor-Liquid equilibrium for high temperature,
high pressure hydrogen-hydrocarbon mixtures"; Sixth World Petroleum Congress
Proceedings, Sect III, 233-245, (1963)