Introducing COSMO-based models, including F-SAC
Prof. Rafael de Pelegrini Soares, D.Sc.
FED. UNIV. OF RIO GRANDE DO SUL - BRAZILCHEMICAL ENGINEERING DEPARTMENT
LAB. VIRTUAL DE PREDICAO DE PROPRIEDADEShttp://www.enq.ufrgs.br/labs/lvpp
January 16, 2017
IntroductionSome Challenges
Conclusions
COSMOCOSMO-RS or COSMO-SACSigma profile
Map
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
COSMOCOSMO-RS or COSMO-SACSigma profile
Our Group: Virtual Laboratory for Property Predictions (LVPP)
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
COSMOCOSMO-RS or COSMO-SACSigma profile
COSMO method (solutes alone)
The COSMOa method was originally developed for thecomputation of solvation effects
The method belongs to the class of dielectric continuum models
,the cavities are discretized into segments or patches
aA Klamt and G Schuurmann. In: J. Chem. Soc., Perkin Trans. 2 (1993), pp. 799–805 Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
COSMOCOSMO-RS or COSMO-SACSigma profile
COSMO method (solutes alone)
The COSMOa method was originally developed for thecomputation of solvation effects
The method belongs to the class of dielectric continuum models,the cavities are discretized into segments or patches
aA Klamt and G Schuurmann. In: J. Chem. Soc., Perkin Trans. 2 (1993), pp. 799–805 Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
COSMO-RS – Surface contacting theory (mixtures)
In the COSMO-RSa methods we rely onCOSMO computations
Based on these pure substancecomputations, the mixture behavior ispredicted (γi )
The COSMO-SACb formulation follows thesame idea
aAndreas Klamt. In: The J. of Phys. Chem. 99.7 (1995),pp. 2224–2235
bShiang-Tai Lin and Stanley I. Sandler. In: Ind. Eng.Chem. Res. 41.5 (2002), pp. 899–913
COSMO-RS – Surface contacting theory (mixtures)
In the COSMO-RSa methods we rely onCOSMO computations
Based on these pure substancecomputations, the mixture behavior ispredicted (γi )
The COSMO-SACb formulation follows thesame idea
aAndreas Klamt. In: The J. of Phys. Chem. 99.7 (1995),pp. 2224–2235
bShiang-Tai Lin and Stanley I. Sandler. In: Ind. Eng.Chem. Res. 41.5 (2002), pp. 899–913
IntroductionSome Challenges
Conclusions
COSMOCOSMO-RS or COSMO-SACSigma profile
COSMO-RS – Surface contacting theory
For each surface pair contact, there is an energy change
This results in different behavior for different substancesin solution
Clearly, there are many different possible contactingarranges and a statistical thermodynamics treatment isnecessary
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
COSMO-RS – Surface contacting theory (mixtures)
Based on these pure substance computations,the mixture behavior is predicted (γi )
“It is always desirable to express the properties of a solutionin terms that can be calculated completely from theproperties of the pure components.” – J. M. Prausnitz.Molecular thermodynamics of fluid-phase equilibria. Third.Prentice-Hall, 1999.
Chloroform/tetrahydrofuran at 30◦C
IntroductionSome Challenges
Conclusions
COSMOCOSMO-RS or COSMO-SACSigma profile
Sigma profile – p(σ)
For a statistical thermodynamicstreatment (without using MD),the 3D apparent surface chargesare projected into a simplehistogram
These pure compounddistributions, known as sigmaprofiles – p(σ), are the basis forcomputing the activity coefficientsin mixture
“It is always desirable to express the properties of a solution in terms that can be calculated completelyfrom the properties of the pure components.” – J. M. Prausnitz. Molecular thermodynamics offluid-phase equilibria. Third. Prentice-Hall, 1999.
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
COSMOCOSMO-RS or COSMO-SACSigma profile
Sigma profile – p(σ)
For a statistical thermodynamicstreatment (without using MD),the 3D apparent surface chargesare projected into a simplehistogram
These pure compounddistributions, known as sigmaprofiles – p(σ), are the basis forcomputing the activity coefficientsin mixture
“It is always desirable to express the properties of a solution in terms that can be calculated completelyfrom the properties of the pure components.” – J. M. Prausnitz. Molecular thermodynamics offluid-phase equilibria. Third. Prentice-Hall, 1999.
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Sigma profile database
The quantum chemistry calculations for generating the sigmaprofiles represent the most time-consuming aspect ofCOSMO-based methods
There are several different quantum chemistry packagesimplementing the COSMO method, e.g.: Gaussian, Turbomole,MOPAC, DMol3 and GAMESS.
However, different packages lead to different sigma profiles a,requiring specific model parametrizations b
aEric Mullins et al. In: Ind. Eng. Chem. Res. 45.12 (June 2006), pp. 4389–4415.doi: 10.1021/ie060370h
bR.P. Gerber and R. de P. Soares. In: Braz. J. of Chem. Eng. 30 (1 2013),pp. 1–11
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Sigma profile database
The quantum chemistry calculations for generating the sigmaprofiles represent the most time-consuming aspect ofCOSMO-based methods
There are several different quantum chemistry packagesimplementing the COSMO method, e.g.: Gaussian, Turbomole,MOPAC, DMol3 and GAMESS.
However, different packages lead to different sigma profiles a,requiring specific model parametrizations b
aEric Mullins et al. In: Ind. Eng. Chem. Res. 45.12 (June 2006), pp. 4389–4415.doi: 10.1021/ie060370h
bR.P. Gerber and R. de P. Soares. In: Braz. J. of Chem. Eng. 30 (1 2013),pp. 1–11
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
VT-2005 database
The freely available database known as VT-2005 contains 1432compoundsa, mostly solvents and small molecules
The VT-2006 contains 206 pharmaceutical-related molecules
VT-databases were constructed using DMol3 (Materials Studio)which is very expensive. . .
aEric Mullins et al. In: Ind. Eng. Chem. Res. 45.12 (June 2006), pp. 4389–4415.doi: 10.1021/ie060370h
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
VT-2005 database
The freely available database known as VT-2005 contains 1432compoundsa, mostly solvents and small molecules
The VT-2006 contains 206 pharmaceutical-related molecules
VT-databases were constructed using DMol3 (Materials Studio)which is very expensive. . .
aEric Mullins et al. In: Ind. Eng. Chem. Res. 45.12 (June 2006), pp. 4389–4415.doi: 10.1021/ie060370h
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Alternatives available
Alternative database with nearly 1000 compoundsa usingMOPAC (semi-empirical – could lead to poor results)
GAMESS (freely available, including source code) can be usedb
aRenan P. Gerber and Rafael de P. Soares. In: Ind. Eng. Chem. Res. 49.16 (Aug.2010), pp. 7488–7496
bShu Wang et al. In: Fluid Phase Equilib. 276.1 (2009), pp. 37–45
Suggestion
Assemble an extensive and freely available database using the rigorous methodsavailable in GAMESS
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Alternatives available
Alternative database with nearly 1000 compoundsa usingMOPAC (semi-empirical – could lead to poor results)
GAMESS (freely available, including source code) can be usedb
aRenan P. Gerber and Rafael de P. Soares. In: Ind. Eng. Chem. Res. 49.16 (Aug.2010), pp. 7488–7496
bShu Wang et al. In: Fluid Phase Equilib. 276.1 (2009), pp. 37–45
Suggestion
Assemble an extensive and freely available database using the rigorous methodsavailable in GAMESS
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Alternatives available
Alternative database with nearly 1000 compoundsa usingMOPAC (semi-empirical – could lead to poor results)
GAMESS (freely available, including source code) can be usedb
aRenan P. Gerber and Rafael de P. Soares. In: Ind. Eng. Chem. Res. 49.16 (Aug.2010), pp. 7488–7496
bShu Wang et al. In: Fluid Phase Equilib. 276.1 (2009), pp. 37–45
Suggestion
Assemble an extensive and freely available database using the rigorous methodsavailable in GAMESS
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Reliability: The good. . .
There are several very good results in the literature, for instance Eugene Paulechkaet al. In: J. Chem. Eng. Data (2015)
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
The good, the bad and the ugly. . .
There are also bad and ugly results around, e.g. Hans Grensemann andJurgen Gmehling. In: Ind. Eng. Chem. Res. 44.5 (2005), pp. 1610–1624, foracetone(1)/diisopro-pylether and perfluorohexane(1)/n-hexane, respectivelly:
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Precision of COSMO-based models
COSMO-based models have exceptional theoretical features
However, very often empirical modifications are needed for improving theagreement with experimental data, for instance:
Corrections in the water apparent surface charge1
Empirical scaling factors2 or empirically scaled surface areas3
Special parametrizations for different families, e.g. alcohols4
. . .
1H. Grensemann and J. Gmehling. In: Ind. Eng. Chem. Res. 44.5 (2005), pp. 1610–1624.2Renan P. Gerber and Rafael de P. Soares. In: Ind. Eng. Chem. Res. 49.16 (Aug. 2010),
pp. 7488–7496.3Jens Reinisch et al. In: Fluid Phase Equilib. 310.1-2 (2011), pp. 7–10.4Robert Franke, Bernd Hannebauer, and Sebastian Jung. In: Fluid Phase Equilib. 340 (2013),
pp. 11–14.Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
F-SAC: Functional-Segment Activity Model
In the F-SAC model, the quantumchemistry computation is replaced byan artificial sigma-profilea
Reduced predictive power
But hopefully, with increasedresolution and with less parametersthan UNIFAC-type models
aR. de P. Soares and R. P. Gerber. In: Ind. Eng.Chem. Res. 52 (2013), pp. 11159–11171
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
F-SAC: Functional-Segment Activity Model
In the F-SAC model, the quantumchemistry computation is replaced byan artificial sigma-profilea
Reduced predictive power
But hopefully, with increasedresolution and with less parametersthan UNIFAC-type models
aR. de P. Soares and R. P. Gerber. In: Ind. Eng.Chem. Res. 52 (2013), pp. 11159–11171
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
F-SAC: Functional-Segment Activity Model
F-SAC and COSMO-SAC model equations are identical, the difference is in thesigma profileIn the F-SAC model there is a neutral and two charged peaks per functional groupSimply add the sigma-profiles of functional groups to get the sigma-profile ofmolecules
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
Why don’t just use UNIFAC instead?
According to UNIFAC (Do) revision 5 – Antje Jakob et al. In: Ind. Eng. Chem. Res.45.23 (2006), pp. 7924–7933.
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
F-SAC parameters
For the 24 grupos considered, only 152 parameterswere calibrated.
Only the hydrogen-bonding? energies are pairwise
All other parameters are for pure groups alone,reducing the total number of parameters
In order to represent the same molecules in mixtures,778 parameters are used in UNIFAC (Do).
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
VLE predictions for non-associating mixtures
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x1, y1
0.075
0.100
0.125
0.150
0.175
0.200
0.225
0.250
Pre
ssu
re[b
ar]
Exp. 298.2 K
F-SAC
UNIFAC(Do)
COSMO-SAC
(a) Chloroform/n-heptane
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x1, y1
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
Pre
ssu
re[b
ar]
Exp. 323.15 K
F-SAC
UNIFAC(Do)
COSMO-SAC
(b) Acetone/cyclohexane
R. de P. Soares and R. P. Gerber. In: Ind. Eng. Chem. Res. 52 (2013), pp. 11159–11171
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
VLE predictions for non-associating mixtures
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x1, y1
2.00
2.25
2.50
2.75
3.00
3.25
3.50
3.75
Pre
ssu
re[b
ar]
Exp. 283.6 K
F-SAC
UNIFAC(Do)
COSMO-SAC
(c) Dimethyl ether/1-butene
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x1, y1
0.65
0.70
0.75
0.80
0.85
0.90
0.95
1.00
1.05
1.10
Pre
ssu
re[b
ar]
Exp. 323.15 K
F-SAC
UNIFAC(Do)
COSMO-SAC
(d) Methyl acetate/1-hexene
R. de P. Soares and R. P. Gerber. In: Ind. Eng. Chem. Res. 52 (2013), pp. 11159–11171
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
F-SAC parameters for Hydrogen-Bonding mixtures (association)
HB-acceptor is red (positive), HB-donor is blue (negative)5
5R. de P. Soares, R.P. Gerber, et al. In: Ind. Eng. Chem. Res. 52 (2013), pp. 11172–11181.Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
VLE predictions for associating mixtures
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x1, y1
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70P
ress
ure
[ba
r]Exp. 298.15 K
F-SAC
UNIFAC(Do)
(e) diethyl-ether/chloroform
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x1, y1
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
Pre
ssu
re[b
ar]
Exp. 380.15 K
Exp. 368.15 K
F-SAC
UNIFAC(Do)
(f) toluene/3-methyl,1-butanol
R. de P. Soares, R.P. Gerber, et al. In: Ind. Eng. Chem. Res. 52 (2013), pp. 11172–11181
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
IDAC comparison for associating mixtures (water excluded)
-2 -1 0 1 2 3 4 5 6Logarithm of Experimental IDAC
-2
-1
0
1
2
3
4
5
6
Log
arith
m o
f M
odel
ID
AC
(g) F-SAC
-2 -1 0 1 2 3 4 5 6Logarithm of Experimental IDAC
-2
-1
0
1
2
3
4
5
6
Log
arith
m o
f M
odel
ID
AC
(h) UNIFAC (Do)
R. de P. Soares, R.P. Gerber, et al. In: Ind. Eng. Chem. Res. 52 (2013), pp. 11172–11181
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
IDAC comparison for mixtures with water
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Logarithm of Experimental IDAC
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Log
arith
m o
f M
odel
ID
AC
(i) F-SAC
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20Logarithm of Experimental IDAC
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Log
arith
m o
f M
odel
ID
AC
(j) UNIFAC (Do)
R. de P. Soares, R.P. Gerber, et al. In: Ind. Eng. Chem. Res. 52 (2013), pp. 11172–11181
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
FSAC: Hydrocarbon-water mutual solubilities
275 300 325 350 375 400 425 450 475
Temperature [K]
0.00001
0.0001
0.001
0.01
0.1M
ola
r S
olu
bili
tyExp. S1
Exp. S2
F-SAC*
UNIFAC(PSRK)
UNIFAC(Do)
UNIFAC-LLE
(k) n-Hexane-Water
275 300 325 350 375 400 425 450 475 500 525 550 575
Temperature [K]
0.0001
0.001
0.01
0.1
Mo
lar
So
lub
ility
Exp. S1
Exp. S2
F-SAC*
UNIFAC(PSRK)
UNIFAC-LLE
(l) Toluene-Water
L.F.K. Possani et al. In: Fluid Phase Equilib. 384 (2014), pp. 122–133
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
Predicting high-pressure VLE with SCMR
The F-SAC can be combined with SRK with Mathias-Copeman α function andthe Self-Consistent Mixing Rule6 (SCMR) for propane-benzene, no parameter isadjusted for the mixture effects:
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0x1, y1
0
5
1 0
1 5
2 0
2 5
3 0
3 5
4 0
4 5
5 0
5 5
6 0
6 5
Pre
ssu
re[b
ar]
Exp. 310.93 K
Exp. 344.26 K
Exp. 377.59 K
Exp. 477.59 K
SRK(MC)+SCMR(FSAC)
6Paula B. Staudt and Rafael de P. Soares. In: Fluid Phase Equilibria 334 (2012), pp. 76–88.
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Extension for electrolytes: eF-SAC
Ions considered spherical with a given charge (+1,-1, +2, -2, etc.)
Then the only parameter is the surface area (waterparameters from previous works)
To avoid direct contact between ions, theirinteraction energy is assumed high
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
eF-SAC σ-profiles
Group Q+k (A
2) Q−
k (A2) σ+
k (eA−2
)
Li+ 0 36,96 -0,027Na+ 0 29,05 -0,034K+ 0 47,13 -0,021Rb+ 0 40,17 -0,025Cs+ 0 41,64 -0,024F− 27,12 0 0,037Cl− 39,19 0 0,026Br− 30,11 0 0,033I− 38,69 0 0,026
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Results: σ-Profile for eF-SAC
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Results: γ±
NaFaq
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Results: γ±
LiIaq
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Results: γ±
NaBraq
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Results: γ±
LiBraq
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
Sigma profile databaseReliability
Results: γ±
CsIaq
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
ConclusionsLinks and more info
Conclusions
COSMO-based models have exceptional theoretical features
An extensible, open, and rigorous sigma-profile database needs to be developed
Currently, results can be considered semi-quantitative
Agreement with experimental data is usually obtained by means of empiricalcorrections
The variation known as F-SAC usually produces better results but depends onfitted parameters
There are certanly many other opportunities for improvement and new fields forapplication
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
ConclusionsLinks and more info
Conclusions
COSMO-based models have exceptional theoretical features
An extensible, open, and rigorous sigma-profile database needs to be developed
Currently, results can be considered semi-quantitative
Agreement with experimental data is usually obtained by means of empiricalcorrections
The variation known as F-SAC usually produces better results but depends onfitted parameters
There are certanly many other opportunities for improvement and new fields forapplication
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
ConclusionsLinks and more info
Conclusions
COSMO-based models have exceptional theoretical features
An extensible, open, and rigorous sigma-profile database needs to be developed
Currently, results can be considered semi-quantitative
Agreement with experimental data is usually obtained by means of empiricalcorrections
The variation known as F-SAC usually produces better results but depends onfitted parameters
There are certanly many other opportunities for improvement and new fields forapplication
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models
IntroductionSome Challenges
Conclusions
ConclusionsLinks and more info
Thank you!
Molecules database at http://code.google.com/p/jcosmo/
Download the F-SAC demonstration code athttp://www.enq.ufrgs.br/labs/lvpp
F-SAC is already available in the iiSE process simulator and being integrated intoDWSim (open source) process simulator
Contact: [email protected]
Prof. Rafael de Pelegrini Soares, D.Sc. Some challenges for COSMO-based models