Properties and Methods in PREDICT Properties and Methods in PREDICT
Inputs Required: Tb, Structure
Expected Error: This is an improvement of Lydersen's method. Joback tested this method on 409 compounds. The average error was 4.8 K or 0.8% with a standard deviation of 7 K. This is only slightly worse than Ambrose but this method is much easier to use. This error will be higher with unreliable Tb.
Reference: Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/1984
Inputs Required: Tb, Structure
Expected Error: RPS states errors are "Usually < 2%; up to 5% for higher molecular weight (>100) nonpolar materials, errors uncertain for molecules with multifunctional polar groups."
Reference: Lyderson, A. L.: Estimation of Critical Properties of Critical Properties of Organic Compounds, Univ. Wisconsin Coll. Eng., Eng. Exp. Stn. Rep. 3, Madison, Wis., April 1955. RPS pp. 12-15. Si Groups: Myers, K. H.,"Thermodynamic and Transport Property Prediction Methods for Organimetallic Compounds", MS Thesis, Penn State, 1990
Inputs Required: Tb, Structure
Expected Error: DIPPR found errors with method to be generally < Lydersen. Especially for alcohols, anhydrides, fluorides and ketone. On the average errors were approximately 1 %
Reference: Ambrose, D., "Correlation and Estimation of Vapour-Liquid Critical Properties: I. Critical Temperatures of Organic Compounds", NPL Report Chem 92, September 1978. Si Groups: Myers, K. H.,"Thermodynamic and Transport Property Prediction Methods for Organimetallic Compounds", MS Thesis, Penn State, 1990
Inputs Required: Structure
Expected Error: Both Klincewicz-Reid and the original author found the error to be at least 5%. This is higher than the others but notice that only the structure and not Tb is required.
Reference: Fedors, R. F., Chem. Eng. Comm., 16, 149 (1982).
Inputs Required: Tb, MW, Structure
Expected Error: The original investigator compares these two methods to Lydersen, Ambrose and Fedors. The result is an easier method than Ambrose with comparable errors.
Reference: Klincewicz, K. M., R. C. Reid, AIChE J., 30 (1): 137 1984).
Inputs Required: Tb, MW, # of atoms
Expected Error: The original investigator compares these two methods to Lydersen, Ambrose and Fedors. The simple method does not require structure and gives errors in the range of Lydersen, and higher than Ambrose.
Reference: Klincewicz, K. M., R. C. Reid, AIChE J., 30 (1): 137 1984).
Inputs Required: Structure, # of Atoms
Expected Error: This is an improvement of Lydersen's method. Joback tested this method on 392 compounds. The average error was 2.1 bar or 5.2% with a standard deviation of 3.2 bar. This is only slightly worse than Ambrose but this method is much easier to use.
Reference: Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/1984
Inputs Required: MW, Structure
Expected Error: RPS states errors are "Usually < 4% up to 10% for higher molecular weight (>100) nonpolar materials, errors uncertain for molecules with multifunctional polar groups."
Reference: Lyderson, A. L.: Estimation of Critical Properties of Critical Properties of Organic Compounds, Univ. Wisconsin Coll. Eng., Eng. Exp. Stn. Rep. 3, Madison, Wis., April 1955. RPS pp. 12-15 Si Groups Myers, K. H.,"Thermodynamic and Transport Property Prediction Methods for Organimetallic Compounds", MS Thesis, Penn State, 1990
Inputs Required: MW, Structure
Expected Error: DIPPR found this method to be on the average better than Lydersen, but several families, such as Aldehydes and Anhydrides were were much worse. Halides, acids and alcohols were better, expect errors approximately 4%.
Reference: Ambrose, D., "Correlation and Estimation of Vapour-Liquid Critical Properties: I. Critical Pressures and Critical Volumes of Organic Compounds", NPL Report Chem 98, May 1979. Si Groups Myers, K. H.,"Thermodynamic and Transport Property Prediction Methods for Organimetallic Compounds", MS Thesis, Penn State, 1990.
Inputs Required: MW, Structure
Expected Error: The original authors compare this method with Lydersens and Ambrose, they find that this method is easier to use and nearly as accurate as the Ambrose method (expect errors approximately 5-7 %).
Reference: Klincewicz, K. M., R. C. Reid, AIChE J., 30 (1): 137 1984).
Inputs Required: MW, # of atoms
Expected Error: The original authors compare this method with Lydersens and Ambrose. This simple method doesn't usestructure and has errors in the 10+ % range.
Reference: Klincewicz, K. M., R. C. Reid, AIChE J., 30 (1): 137 1984).
Inputs Required: Structure
Expected Error: This is an improvement of Lydersen's method and is better than Ambrose's. Joback tested this method on 310 compounds. The average error was 7.5 cc/mol or 2.3% with a standard deviation of 13.2 cc/mol. This is better than Ambrose method and easier to use.
Reference: Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/1984
Inputs Required: Structure
Expected Error: RPS states errors are "Usually < 4% up to 10% for higher molecular weight (> 100) nonpolar materials, errors uncertain for molecules with multifunctional polar groups." Error for Si, 3.8% for 19 silanes, 4.8% for 12 siloxanes, better than Fedors.
Reference: Lyderson, A. L.: Estimation of Critical Properties of Critical Properties of Organic Compounds, Univ. Wisconsin Coll. Eng., Eng. Exp. Stn. Rep. 3, Madison, Wis., April 1955. RPS pp. 12-15 Si Groups Fit by Dragon Technology, using data from the following ref: Myers, K. H.,"Thermodynamic and Transport Property Prediction Methods for Organimetallic Compounds", MS Thesis, Penn State, 1990.
Inputs Required: Structure
Expected Error: Klincewicz and Reid list the error of this method to be approximately 3% for nearly 400 compounds tested. Error for Si, 3.3% for 19 silanes, 3.5% for 12 siloxanes, < Fedors or Lydersen.
Reference: Ambrose, D., "Correlation and Estimation of Vapour-Liquid Critical Properties: I. Critical Pressures and Critical Volumes of Organic Compounds", NPL Report Chem 98, May 1979. Si Groups Fit by Dragon Technology, using data from the following ref: Myers, K. H.,"Thermodynamic and Transport Property Prediction Methods for Organimetallic Compounds", MS Thesis, Penn State, 1990.
Inputs Required: MW, Structure
Expected Error: The original authors tested their method on nearly 400 compounds and found the average error to be approximately 3 %.
Reference: Klincewicz, K. M., R. C. Reid, AIChe J., 30 (1): 137 1984).
Inputs Required: MW, # of atoms
Expected Error: Error will be > the more complex method by the same author, the average error for the complex method is approximately 3 %.
Reference: Klincewicz, K. M., R. C. Reid, AIChe J., 30 (1): 137 1984).
Inputs Required: Structure
Expected Error: The original author reports errors to be slightly < the Lydersen method, on 160 compounds tested the average error was 3 %. DIPPR found errors in the range of 1-5 % on 80 compounds.
Reference: Fedors, R. F., AIChE J., 25(1): 202 (1979). Si Groups Myers, K. H.,"Thermodynamic and Transport Property Prediction Methods for Organimetallic Compounds", MS Thesis, Penn State, 1990.
Inputs Required: Tc, Pc, w
Expected Error: The compressibility factor is evaluated at the critical point and then the Vc is calculated. Pitzer claims error are < 2% in the Z calc. The error in Vc will depend greatly on the error in Tc and Pc.
Reference: Pitzer, K. S., et al., J.A.C.S., 77: 3433 (1955).
Inputs Required: Tc, Pc, w, x
Expected Error: The compressibility factor is evaluated at the critical point and then the Vc is calculated. The authors tested 19 POLAR compounds and found an error of < 2%. The error in Vc will depend greatly on the error in Tc and Pc and on the complexity of the compound's polarity.
Reference: Halm, R. L., L. I. Stiel, AIChE J., 16(1): 3 (1970).
Inputs Required: Tc, Pc, Vapor Pressure or Heat of Vaporization Data Point
Types of Compounds: Non-Polar, "Normal Fluids"
Expected Error: For "Normal Fluids" the resulting Pv's and Hv's calculated from this w can be accurate to < 2% error, if the Tc, Pc and original data used to calculate w were sufficiently accurate.
Reference: Pitzer, K. S., et al., J.A.C.S., 77: 3433 (1955).
Inputs Required: w, Tc, Pc, Two Vapor Pressure &/or Heat of Vaporiz. Data Point
Expected Error: For simple polar fluids the Pv's and Hv's calculated from this will can be accurate to < 5% error, if the Tc, Pc and original data used to calculate w were sufficiently accurate. For Complex Polar Cmpds the error will be greater.
Reference: Halm, R. L., L. I. Stiel, AIChE J., 13(2): 351 (1967).
Inputs Required: Tc, Pc, Vapor Pressure Data Point, or w
Expected Error: DIPPR found that at pressures > 100 mmHg the expected errors in vapor pressure should be in the range of 2-3 %. At lower pressures the error can be 20-30%. In general the errors will depend on the accuracy of the Tc, Pc and accuracy of vapor pressure data point used. Reference: Riedel, L., Chem. Ing. Tech., 26: 83 (1954).
Inputs Required: Structure
Types of Compounds: Polar, Non-Polar, Plus Si,B,Sn,Pb,Cr,Zn,Ti,V,Cd,Al,Ge,Hg,P
Expected Error: Joback and RPP report errors on 28 organic compounds, Benson's method shows an average error of 0.9 kcal/mol (approximately 2%) with a std dev of 1.0 and a max error of 3.3 kcal/mol, compared with Joback average error at 2.2 kcal/mol (approximately 4%) & max error of 8 kcal/mol. DIPPR reports the error for Si, using new groups at about 1.9% , B compounds at 4.8% and Al compounds at 2.1%.
Reference: Benson,S.W.:"Thermochemical Kinetics",Wiley,New York, 1968,Chap. 2. Benson,S.W.,F.R.Cruickshank,D.M.Golden,G.R.Haugen,H.E.O'Neal, A.S.Rodgers,R.Shaw and R. Walsh: Chem Rev., 69:279 (1969). Si, B & Al Myers, K. H.,R. P. Danner, "Documentation of the Basis for Selection of Prediction Methods for Organometallic Compounds" Doc. Report No. 802-90, DIPPR, AIChE, New York, 1990.
Inputs Required: Structure
Expected Error: Joback tested this method on 379 compounds. The average error was 2.2 kcal/mol or 15.2% with a standard deviation of 4.3 kcal/mol. If 6 compounds are removed, the error is only 9.2%. This is only slightly worse than Benson but is much easier to use. More complicated compounds will show large errors, as high as 10 to 20 kJ/mol.
Reference: Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/1984
Inputs Required: Structure
Expected Error: Comparisons with experimental data by RPS, indicate errors are generally < 5%.
A complete group contribution method with a considerably better accuracy than the simple bond contribution method. This method compares favorably with other group contribtuion methods for heat of formation estimation. The only method better is that of Benson, which is also much more complex to use.
Reference: Franklin, J.L., IEC, 41:1070(1949) ,J. Chem.Phys.,21:2029 (1953)
Inputs Required: Structure
Reference: Reid, R., J. Praustniz, T. Sherwood, "The Properties of Gases and Liquids", 3rd Ed., McGraw-Hill, New York, 1977, pp.222-258.
Inputs Required: Structure
Expected Error: Joback tested this method on 328 compounds. The average error was 2.0 kcal/mol or 15.7% with a standard deviation of 4.3 kcal/mol, when one compound was removed (2,2-dimethylpentane). Joback claims to not need symmentry terms that further complicate other methods, because DelGf is modeled as a linear function of the sum of group contributions. Others, such as Benson, model the entropy of formation, requireing symmentry effects. Errors are too high for reliable equilibrium (Keq) calculations.
Reference: Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/1984
Inputs Required: Structure
Expected Error: RPS, indicate errors are should be < 5 kcal/mole
Reference: van Krevelen, D., Chermin, H., Chem Eng Sci, 1:66(1951),1:238(1952)
Inputs Required: Structure
Reference: Reid, R., J. Praustniz, T. Sherwood, "The Properties of Gases and Liquids", 3rd Ed., McGraw-Hill, New York, 1977, pp.222-258.
Inputs Required: Structure
Expected Error: This is as good as prediction of this basic property can get. Joback tested this method on 438 compounds. The average error was 12.9 K or 3.6% with standard a deviation of 17.9 K. Tb is used in the prediction of many other properties. An error in Tb will be propagated to other properties calculated with it. It is recommended to get an experimental value for this very important and fundamental property. If prediction is absolutely necessary, this is the best method.
Reference: Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/1984
Inputs Required: MW, Structure
Expected Error: This Method should only be used as a last resort and should be considered as only a rough APPROXIMATION. This Method Combines the Rackett Liquid Density, the Tyn & Calus estimation of the liquid density at the normal boiling point and the Lydersen Tc, Pc, Vc.
Reference: Perry's Chemical Engineering Handbook, 6th Edition, McGraw-Hill, 1984, p. 3-267, Tyn, M.T, W. Calus, Processing 21(4): 16 (1975). Rackett, H. G., J. Chem. Eng. Data, 15: 514 (1970)
Types of Compounds: Non-Polar "Normal Fluids" (Group Cont. Must be Available)
Inputs Required: Structure, MW, Vapor Pressure/Temperature Data Point
Temperature Range: 0.25 ó Tr ó 1.0 (Data cannot be outside of this range
Reference: Lyderson, A. L.: Estimation of Critical Properties of Critical Properties of Organic Compounds, Univ. Wisconsin Coll. Eng., Eng. Exp. Stn. Rep. 3, Madison, Wis., April 1955. RPS pp. 12-15. Ambrose, D., "Correlation and Estimation of Vapour-Liquid Critical Properties: I. Critical Temperatures of Organic Compounds", NPL Report Chem 92, September 1978. Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/84. Pitzer, K. S., et al., J.A.C.S, 77: 3433 (1955). Carruth, G., R. Kobayashi, IEC, Fund., 11(4): 509 (1972).
Inputs Required: Tc, Pc, w
Temperature Range: 0.25 < = Tr < = 1.0
Expected Error: Errors are usually < 2% between the normal boiling point and the critical, expect errors approximately 5% at lower temperatures. This error is depend- ent on the quality of Tc, Pc and w. The method is only good for "normal fluid" A "normal fluid" is non-polar, i.e.no dipole, quadrapole or hydrogen bondings.
Reference: Pitzer, K. S., et al., J.A.C.S, 77: 3433 (1955). Carruth, G., R. Kobayashi, IEC, Fund., 11(4): 509 (1972).
Inputs Required: Tc, Pc, w, x
Temperature Range: 0.44 < = Tr < = 1.0
Expected Error: A fourth parameter (x) extends the eariler Pitzer correlation to polar compounds. The ERROR using this equation for simply polar compounds should be < 5%. The error depends on the quality of Tc, Pc, x & w. The error will be higher for compounds with complex forces,like hydrogen bonds
Reference: Halm, R. L., L. I. Stiel, AIChE J., 13(2): 351 (1967).
Inputs Required: Tc, Pc, riedel factor
Temperature Range: Tr < = 1.0, and it should not be used for Pressures < 10 mmHg
Expected Error: DIPPR found that at pressures > 100 mmHg the expected errors should be in the range of 2-3 %. At lower pressures the error can be 20-30%. In general the errors will depend on the accuracy of the Tc, Pc and riedel factor used.
Reference: Riedel, L., Chem Ing. Tech., 26: 83 (1954)
Inputs Required: Tc, Pc, w, for polar MW
Temperature Range: Tr < = 1.0
Expected Error: The original authors tested 113 non-polar compounds and found the average error with this method to be approximately 1 %. For 25 polar compounds the error was approximately 1.5%. This accuracy depends greatly on the goodness of Tc & Pc.
Reference: Gomez-Nieto, M., G. Thodos, IEC, Fund.,17: 45(1978), 16: 254(1977) Can. J. Chem. Eng., 55: 445 (1977).
Inputs Required: Tc, Pc, w
Temperature Range: 0.30 < = Tr < = 1.0
Expected Error: Errors are usually < 2% between the normal boiling point and the critical, expect errors approximately 5% at lower temperatures. This error is depend- ent on the quality of Tc, Pc and w. The method is only good for "normal fluid" A "normal fluid" is non-polar, i.e.no dipole, quadrapole or hydrogen bondings.
Reference: Pitzer, K. S., et al., J.A.C.S, 77: 3433 (1955). Carruth, G., R. Kobayashi, IEC, Fund., 11(4): 509 (1972).
Inputs Required: Tc, Pc, w, x
Temperature Range: 0.56 < = Tr < = 0.72
Expected Error: A fourth parameter (x) extends the eariler Pitzer correlation to polar compounds. The ERROR using this equation for simply polar compounds should be < 5%. The error depends on the quality of Tc, Pc, x & w. The error will be higher for compounds with complex forces,like hydrogen bonds.
Reference: Halm, R. L., L. I. Stiel, AIChE J., 13(2): 351 (1967).
Inputs Required: Tc, Hv(ref) @ T(ref) or Tb and Pc
Temperature Range: T < Tc, not recommended for T < Tb
Expected Error: The constant valuse in the equation has been choosen as a constant value of 0.38. This is generally satisfactory for engineering calculations in the region Tb < T < Tc (Expect errors in the 10% range).
Reference: Watson, K. M., Ind. Eng. Chem., 35: 398 (1943).
Inputs Required: Tc, Tb and Pc
Temperature Range: T = Tb, Expanded to T < Tc with the Watson eq.
Expected Error: Generally approximately 2%
Reference: Riedel, L., Chem. Ing. Tech., 26:679 (1954).
Inputs Required: Tc, Tb and Pc
Temperature Range: T = Tb, Expanded to T < Tc with the Watson eq.
Expected Error: Generally approximately 2%
Reference: Chen, N. H., J. Chem. Eng. Data., 15: 592 (1923).
Inputs Required: Structure
Temperature Range: T = Tb, Expanded to T < Tc with the Watson eq.
Expected Error: Joback reported the average absolute error for 328 organic compounds to be 303 cal/gmole with a standard deviation of 429 cal/gmole. The average percent error was 3.88%.
Reference: Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/1984.
Inputs Required: Tc, w, & either Pc or a reference density data point
Temperature Range: 0.2 < = Tr < 1.0
Expected Error: Very Accurate!, the original authors found only .22% average error when testing 26, inorganics, organics; halogens,alcohols, nitriles, ethers, aromatics, alkanes, alkenes, etc.
Reference: Gunn, R., T. Yamada, AIChE J., 17: 1341 (1971).
Inputs Required: Tc, Pc, w, x
Temperature Range: 0.56 < = Tr < = 1.0
Expected Error: Joffe-Zudkevitch found errors of approximately 2 % with Tr < 0.85 and approximately 4% with Tr > 0.85 when testing this method. This depends on the quality of Tc & Pc.
Reference: Halm, R. L., L. I. Stiel, AIChE J., 16(1): 3 (1970).
Inputs Required: Tc, Pc and a reference density data point
Temperature Range: Tr < 1.0
Expected Error: Errors reported by Spencer are < 1 % for non-polar compounds and approximately 2 % for alcohols and acids. Rackett does not use the reference density and therfore has higher error, again it all depends on accuracy of Tc an Pc
Reference: Rackett, H. G., J. Chem. Eng. Data 15: 514 (1970). Spencer, C., R. Danner, J. Chem Eng. Data, 17 236 (1972).
Inputs Required: Tc, Pc, w, reference density data point for polar only
Temperature Range: 0.3 < = Tr < = 1.0 (Non-Polar); 0.42 < = Tr < = 1.0 (Polar)
Expected Error: The original authors report errors of < 1% for tests with Polar compounds and approximately 0.5 % error with non-polar compounds. This again will depend to a great extent on the accuracy of Tc and Pc.
Reference: Bhirud, V. L., AIChE J., 24: 1127 (1978), AIChE J., 24: 880 (1978).
Inputs Required: Tc, Pc, and riedel factor or a reference surface tension data point
Temperature Range: Tr < 1.0
Expected Error: Errors are expected to be < 5% for non-polar compounds This will depend on the quality of Tc, Pc and riedel factor.
Reference: Brock, J. R., R. B. Bird,AIChE J., 1: 174 (1955).
Inputs Required: Tc, Pc, w, x
Temperature Range: Tr < 1.0
Expected Error: Errors for polar compounds should be in the range of 5%. As with a corresponding states method this error depends greatly on the accuracy of Tc, Pc, w and x.
Reference: Hakim,D.I., D.Stienberg, and L.I.Stiel: IEC Fund., 10: 174 (1974).
Inputs Required: Tc, Structure
Temperature Range: Tr < = 0.8
Expected Error: Morris found an average error of 12% when testing data on 70 organic compounds. RPS found that the errors can very widely and are generally lower at lower temperatures and for lower members of a homologous family.
Reference: Morris, P., MS Thesis, Poly. Tech. Brooklyn, Brooklyn, N.Y., 1964.
Inputs Required: A Single Liquid Viscosity reference data point.
Temperature Range: Tr < 0.8
Expected Error: This equation represents an analytical solution to a graphic by Gambill. This method is based on the empirical fact that the temperature variation of the liquid viscosity is as given in the equation below. The error of this method is about average for liquid viscosity methods, about 20%. It should not be used with highly polar fluids,emulsions or suspensions.
Reference: Gambill, W. R., Chem. Eng., 66 (3):123 (1959).
Inputs Required: Tc, Pc, MW, w
Temperature Range: 0.76 < = Tr < = 0.98
Expected Error: This is the most accurate method to estimate liquid viscosity for non-polar compounds in this temperature range. The original investigators report only errors of approximately 3% on 12 compounds tested. For lower temperatures, and with considerable accuracy loss, one point from this correlation could be used with the Gambill correlation. The errors depend on the quality of Tc & Pc.
Reference: Letsou, A., L. I. Stiel, AIChE J., 19: 409 (1973).
Inputs Required: Structure, MW
Temperature Range: T < = Tc
Expected Error: Joback calculated the errors on 36 compounds, the average error was 18% as compared to 15% for the Morris method. All liquid viscosity methods report high errors.
Reference: Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/1984.
Inputs Required: Tc, Pc, MW, Polar - Vc
Temperature Range: Polar Tr < 2.5, H-Bonded Tr < 2.0
Expected Error: RPS Tested many compounds with an average error of approximately 2% for non-polar and approximately2.5% for polar compounds. Error depends on the quality of Tc,Pc.
Reference: Yoon,P.,G. Thodos, AIChE J.,16: 300 (1970); Stiel,L.E.,G. Thodos, AIChE J.,8: 229 (1962)
Inputs Required: Tc, MW, Structure
Temperature Range: Unlimited
Expected Error: RPS Tested many compounds with an average error of approximately 2%. This error will depend on the quality of Tc.
Reference: Reichenberg, D., AIChE J., 19: 854 (1973); AIChE J., 21: 181 (1975).
Inputs Required: Tc, w, Cpv
Temperature Range: 273 K < T < = Tc
Expected Error: Errors are generally < 10% but the errors will depend on the accuracy of Cpv and Tc.
Reference: Bondi, A., IEC, Fund., 5: 443 (1966).
Inputs Required: Tc, w, x, Cpv
Temperature Range: T > 273 K, and 0.40 < = Tr < = 0.96(Non-Polar), 0.44 < = Tr < = 0.94 (Polar)
Expected Error: The orig. authors and RPS report errors of about 5% but this will depend on the accuracy of Cpv and Tc.
Reference: Yuan, T. F., L. I. Stiel, IEC, Fund., 9: 393 (1970).
Inputs Required: Structure
Temperature Range: T < Tb
Expected Error: RPS Expect the error for this method will be approximately 5%.
Reference: Missenard, F. A., C. R., 260) 5521 (1965).
Inputs Required: Structure
Temperature Range: T=300K, T=400K, T=500K, T=600K, T=800K, T=1000K Data at these descrete points are regressed to a cubic in temperature.
Expected Error: RPP tested a total of 27 compounds at 298 K and 800 K and found an average error of 1.1%.
Reference: Benson,S.W.:"Thermochemical Kinetics",Wiley,New York, 1968,Chap. 2. Benson,S.W.,F.R.Cruickshank,D.M.Golden,G.R.Haugen,H.E.O'Neal, A.S.Rodgers,R.Shaw and R. Walsh: Chem Rev., 69:279 (1969).
Inputs Required: Structure
Temperature Range: 298 K to 1000 K
Expected Error: Joback tested this method on 28 compounds, 8 of which were not included in the development. The average error was 1.4% with a standard deviation of 2.0%. This compares with 1.1% mean error and 1.6% sts dev for those same compounds using Benson's method. The maximum error of the 8 cmpds not used in the correlation was 3.5% with a mean error of 1.3%.
Reference: Joback, K. G.:SM Thesis in Chem. Engin., MIT, Cambridge, MA, 6/1984.
Inputs Required: Empirical Formula
Temperature Range: 300 K to 1500 K
Expected Error: The authors tested this method on 2500 data points. The error for all compounds ranged from 6% at low temperatures to 2.3% at 1500 K. For only compounds not specifically included, the errors were similar at low temp but only went down to about 3.5% at high temperatures. For a diverse list of 28 compounds used by RPS, this method had 3.2% error, Benson 1.1%, Joback 1.4%.
Reference: B. K. Harrison, W. H. Seaton, IEC Res., 27,1536-1540, (1988).
Inputs Required: Structure
Temperature Range: T > 273 Kelvin
Expected Error: The original authors report an average error at 300 K of approximately 3% for many organics tested of 3%. While over the range of 400 to 1500 K the error was found to be < 2%.
Reference: Rihani, D. N., L. K. Doraiswamy, Ind. Eng. Chem. Fund., 4: 17(1965).
Inputs Required: Tc, either (kl Ref) or (MW, Tb, ld@0øC, Cpl@0øC, # atoms)
Temperature Range: Tr < 1.0
Expected Error: Errors vary widely, RPS found approximately 13 % for 32 polar & non polar
Reference: Missenard, A., Conductivite thermique des solids, liquides, gaz et de leurs melanges, Editions Eyrolles, Paris, 1965; Missenard, A., C. R., 260(5): 5521 (1965). Riedel, L., Chem.Ing.Tech., 21:349 (1949); 23: 59, 321, 465 (1951).
Inputs Required: Tc, MW, Tb
Temperature Range: Tr < 1.0
Expected Error: Errors vary widely, this method should be used only as an approximate estimation. It's usefulness is it's lack of inputs.
Reference: Sato, K., Reid, R., Sherwood, T., Prausnitz, J., "Properties of Gases and Liquids", 3rd Ed., McGraw-Hill, 1977, p 519. Riedel, L., Chem.Ing.Tech., 21:349 (1949); 23: 59, 321, 465 (1951).
Inputs Required: Tc, Pc, MW, Cp
Temperature Range: T > 273 K or Tr > 0.355
Expected Error: RPS found the errors to be in the 5-10% range for tested cmpds The errors will depend on the accuracy of the Tc and Pc used.
Reference: Misic, D., G. Thodos, AIChE J., 7: 264 (1961). Misic, D., G. Thodos, J. Chem. Eng. Data., 9: 540 (1963).
Inputs Required: MW, gas viscosity, Cpø
Temperature Range: T > 273 Kelvins
Expected Error: RPS found the errors to be approximately 14% for 19 polar and non-polar cmpds. The error was somewhat lower for non-polar compounds only.
Reference: Sato, K., Reid, R., Sherwood, T., Prausnitz, J., "Properties of Gases and Liquids", 3rd Ed., McGraw-Hill, 1977, p 474.