Ind. Eng. Chem. Res. 2002, 41, 1051-1056
1051
Experimental Determination of Vapor-Liquid Equilibrium Data for Asymmetric Systems Hergen Gardeler,† Kai Fischer,† and Ju 1 rgen Gmehling*,‡ Laboratory for Thermophysical Properties (LTP GmbH), Institute at the University of Oldenburg, and Department of Industrial Chemistry, University of Oldenburg, P.O. Box 2503, D-26111 Oldenburg, Germany
Isothermal vapor-liquid equilibrium data for the six following asymmetric binary systems were measured: ethane + decane at 411 and 444 K, propane + dodecane at 419 and 458 K, propane + 1-decanol at 408 and 448 K, ethane + 1-decanol at 448 K, ethylene + 1-decanol at 308 and 318 K, and carbon dioxide + dodecane at 318 K. A static apparatus equipped with rapid on-line sampler-injectors (ROLSIs) was used. With this device, very small samples can be withdrawn from the vapor and liquid phases in the equilibrium cell and transferred directly into the carrier gas stream for gas chromatographic analysis. The experimental data are compared to literature data and to the results predicted with the PSRK group contribution equation of state. Introduction
Table 1. Suppliers and Purities of the Chemicals
For the synthesis and design of extraction processes with supercritical fluids, an exact knowledge of the content of the liquid or solid high-boiling substance in the supercritical phase, as well as its variation with temperature and pressure, is essential. Apart from direct experimental data, only reliable predictive equations of state can provide this information. Unfortunately, up to now, the calculated phase equilibrium behavior for asymmetric systems has not been reliable enough for the design of extraction processes with supercritical fluids, so that one has to rely on correlated experimental data. The data measured in this work belong to the experimental part of a systematic study. The objective of this study is to define rules, recommendations, and limitations for thermodynamic approaches for calculating phase equilibrium behavior in supercritical fluid extraction processes. Simple cubic equations of state are suitable models for correlating the phase equilibria over a large temperature and pressure range.1-3 However, their applicability for a reliable estimate of the small solubilities of high-boiling substances in supercritical fluids, and of the influence of co- or antisolvents on the solubility, has not been analyzed and reviewed until now. The experimental data were obtained using a thermostated pressure cell equipped with two rapid on-line sampler-injectors (ROLSIs).4 With this device, very small samples can be withdrawn from the vapor and liquid phases in the equilibrium cell and transferred directly into the carrier gas stream for gas chromatographic analysis. Although a large amount of VLE (vapor-liquid equilibrium) and gas solubility data are available for the selected systems ethane + decane,5-10 propane + dodecane,11-13 carbon dioxide + dodecane,11,14-19 and ethylene + 1-decanol,20 only a limited number of them * Corresponding author. E-mail: gmehling@ tech.chem.uni-oldenburg.de. Tel.: ++49 (0)441 798 3831. Fax: ++49 (0)441 798 3330. † Laboratory for Thermophysical Properties (LTP GmbH). ‡ Department of Industrial Chemistry.
component
supplier
final purity (%)
decane dodecane 1-decanol carbon dioxide ethane ethylene propane
Sigma Aldrich Sigma Aldrich Sigma Aldrich A. Lu¨bke KG Messer Griesheim Messer Griesheim Messer Griesheim
99.9 99.9 99.9 99.995 99.95 99.95 99.5
provide the important information about the vapor phase composition.5,20 The system ethane + decane was selected as a test system to verify the results obtained with our equipment by comparing them with those of Reamer and Sage.5 Only heat of mixing data21 and LLE (liquid-liquid equilibrium) data22-24 are available for the system ethane + 1-decanol. No data were found in the literature for the system propane + 1-decanol. SLE (solid-liquid equilibrium) data are available for the systems ethane + decane25 and propane + dodecane.26 They are important because the liquid-gas saturated phase can be stable at temperatures much lower than the melting temperature of high-boiling substances at high pressure. This effect can also be used in a separation process competing with supercritical extraction. For the generation of very fine particles, both the liquid (PGSS, particles from gas saturated solutions) and the supercritical (RESS, rapid expansion of supercritical solutions) phases can be rapidly expanded. Experimental Section Materials. All liquids used for the measurements were dried over molecular sieve 4A. Furthermore, the chemicals were distilled and degassed as described by Fischer and Gmehling.27 To remove isomeric impurities, a distillation column equipped with 2 m of laboratory packing (Sulzer DX) was used. The water content was checked by Karl Fischer titration and found to be lower than 35 ppm. Final purities were checked by gas chromatography. The gases were used as supplied without any further purification. Table 1 lists the suppliers and final purities of the chemicals. Apparatus and Procedure. For the measurements, a static apparatus as shown in Figure 1 was used. It can be applied in a pressure range up to 25 MPa and at
10.1021/ie0103456 CCC: $22.00 © 2002 American Chemical Society Published on Web 08/24/2001
1052
Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002
Figure 1. Static apparatus with sampling devices for phase analysis. Table 2. Experimental P-x-y Data for the System Ethane (1) + Decane (2) at 410.95 K x1
y1
P/MPa
x1
y1
P/MPa
0.1024 0.1368 0.1547 0.1832 0.2400
0.95079 0.96476 0.96772 0.97605 0.97799
0.947 1.295 1.497 1.876 2.552
0.3166 0.4490 0.5232 0.6008 0.6795
0.98071 0.98240 0.97898 0.97628 0.96964
3.543 5.418 6.702 7.960 9.175
Table 3. Experimental P-x-y Data for the System Ethane (1) + Decane (2) at 444.25 K
Figure 2. Pneumatic controlled sampling device for microsamples (ROLSI).
temperatures from 20 to 200 °C. The apparatus consists of an equilibrium cell with a magnetically driven stirrer. The cell is mounted in a thermostated air bath and equipped with two windows for visual observation. Visual observation is particularly important at higher pressures, where the system should be checked to determine whether or not the supercritical state of the mixture has been reached and to observe the formation of additional phases, which might occur under certain conditions. The pressure inside the cell is measured with a pressure transmitter (model 891.20.501, WIKA), which is regularly calibrated with a dead weight balance pressure gauge (model 21000, D&H). The temperature is monitored using a Pt 100 thermometer (model 1560, Hart Scientific) with a probe mounted inside the stainless steel body of the equilibrium cell. The compounds are charged into the equilibrium cell using piston injectors (model 2200-801, Ruska) or a thermostated pressure vessel. The equilibrium cell is equipped with two rapid online sampler-injectors (ROLSIs), one at the bottom and one at the top of the cell, which allow very small samples to be withdrawn (e.g., 0.1 σ(xi) ) 0.002 + 0.01xi in the range 0.1 > xi > 0.01 σ(xi) ) 0.05xi in the range xi < 0.01 Results The experimental data are listed in Tables 2-11 and shown graphically in Figures 3-9. The PSRK model28 was used to predict the VLE behavior and to compare it with the experimental values. The lines in the figures give the predicted results, and the symbols represent the experimental data. Because of the low concentration of the high-boiling component in the gas phase, the data for the systems propane + 1-decanol (Figure 6), ethane + 1-decanol (Figure 7), ethylene + 1-decanol (Figure 8), and carbon dioxide + dodecane (Figure 9) are shown on a logarithmic concentration scale. The discontinuities of the lines reflect the resolution of the calculated values in the program output. In the case of measurements at temperatures much higher (more than 100 K) than the critical temperature of the supercritical solvent, the solubility change of the high-boiling compounds with pressure is not very large, as the data for the systems ethane + decane, propane + dedecane, propane + 1-decanol, and ethane + 1-decanol show. However, at temperatures not far (up to ca. 30 K above) from the
1054
Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002
Figure 5. Experimental [this work at 419.15 K (O,)) and at 457.65 K (b,()] and calculated [PSRK (s)] P-x-y data for the system propane (1) + dodecane (2).
Figure 7. Experimental [this work at 448.15 K (O,))] and calculated [PSRK (s)] P-x-y data for the system ethane (1) + 1-decanol (2).
Figure 6. Experimental [this work at 408.15 K (O,)) and at 448.15 K (b,()] and calculated [PSRK (s)] P-x-y data for the system propane (1) + 1-decanol (2).
critical temperature of the solvent, the solubility change with pressure exceeds a factor of 100, as the data for the systems ethylene + 1-decanol and carbon dioxide + dodecane demonstrate. A quantitative prediction of the VLE is not possible for the selected systems with the PSRK model, but the calculated results can be used as rough estimates, and they describe the correct phenomenology.
System Ethane + Decane. This system was chosen as a test system, because Reamer and Sage5 published reliable VLE data including the experimental vaporand liquid-phase compositions. Our results agree well with the literature data at both 411 (Figure 3) and 444 K (Figure 4). The system exhibits type I behavior according to the classification of van Konynenburg and Scott.29 The PSRK model predicts a higher pressure of the boiling point curve and the critical point and a slightly greater solubility of decane in ethane. In Figure 10, the calculated Henry coefficients are compared with literature data. It can be seen that the calculated Henry coefficients are slightly larger than the experimental values. System Propane + Dodecane. This system is also of type I. The results of the PSRK model perfectly agree with the boiling point curve, as it can be seen in Figure 5. However, the solubility of dodecane in propane is larger than calculated at both temperatures, 419 and 458 K. In Figure 10, it can be seen that the calculated Henry coefficients are slightly larger than the experimental values. System Propane + 1-Decanol. Because the solubility of 1-decanol in propane is quite low, the phase diagram is also plotted on a logarithmic scale in addition to the normal scale of the concentration axis in Figure 6. At 408 and 448 K, the system seems to be similar to the systems ethane + decane and propane + decane,
Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002 1055
Figure 8. Experimental [this work at 308.15 K (O,)) and at 318.15 K (b,()] and calculated [PSRK (s)] P-x-y data for the system ethylene (1) + 1-decanol (2).
but in fact, it is known30 that the system is of type III, IV, or V according to the classification of van Konynenburg and Scott. The PSRK model predicts a slightly higher pressure of the boiling point curve and a slightly greater solubility of 1-decanol in propane. System Ethane + 1-Decanol. This system also belongs to type III, IV, or V. The corresponding VLLE behavior has been observed experimentally22-24 between 298 and 303 K. This phenomenon is at least qualitatively predicted with PSRK. In Figure 7, it can be seen that the boiling point curve calculated with PSRK gives a lower pressure than was measured and a greater solubility of 1-decanol in ethane than observed. System Ethylene + 1-Decanol. The behavior and class of this system similar to that of the ethane + 1-decanol system. The prediction of the VLE data at 308 and 318 K shown in Figure 8 differ from the experimental data especially at high pressure. The pressure of the critical point is probably very much underestimated. System Carbon Dioxide + Dodecane. For this system, it has not been clear until now whether it belongs to type II, III, or IV. Schneider discussed the critical behavior for different CO2 + alkane systems.30 The predicted results agree well with the experimental data, as shown in Figure 9. In Figure 10, the calculated Henry coefficients and their temperature dependence
Figure 9. Experimental [this work at 318.15 K (O,))] and calculated [PSRK (s)] P-x-y data for the system carbon dioxide (1) + dodecane (2).
Figure 10. Experimental [literature data for ethane in decane (b),8,9 propane in dodecane ((),11-13 and carbon dioxide in dodecane (2)11,14-16] and calculated [PSRK (s)] Henry coefficients.
are compared with literature data. They also agree well, as already shown for the carbon dioxide + hexadecane system.31 Conclusion The static, analytic apparatus for VLE determination presented in this work was shown to be applicable for
1056
Ind. Eng. Chem. Res., Vol. 41, No. 5, 2002
asymmetric systems and under conditions of elevated pressure. Ten data sets were measured for six binary systems, and the results for the test system ethane + decane agree well with literature data. The capillary microsamplers and the analytic method allow for a precise determination of composition over 4-5 logarithmic decades. All of the data were compared to the results predicted with the PSRK group contribution equation of state. The degree of agreement changes for the different systems. The gas solubility behavior is usually described quite well, but in the vicinity of the critical point, the reliability is not as good. This means that the solubility of high-boiling compounds in supercritical fluids cannot be predicted with the desired precision up to now. Acknowledgment The authors thank the Deutsche Forschungsgemeinschaft DFG for financial support and Dominique Richon, Ecole des Mines de Paris, Fontainebleau, France, for supplying the ROLSI microsamplers. Literature Cited (1) Orbey, H.; Sandler, S. I. A Comparison of Various Cubic Equation of State Mixing Rules for the Simultaneous Description of Excess Enthalpies and Vapor-Liquid Equilibria. Fluid Phase Equilib. 1996, 121, 67-83. (2) Orbey, H.; Sandler, S. I. On the combination of equation of state and excess free energy models. Fluid Phase Equilib. 1995, 11, 53-70. (3) Orbey, H.; Sandler, S. I. Analysis of Excess Free Energy Based Equations of State Models. AIChE J. 1996, 42, 2327-2334. (4) Guilbot, P.; Valtz, A.; Legendre, H.; Richon, D. Rapid online sampler-injector: A reliable tool for HAT-HP sampling and on-line GC analysis. Analysis 2000, 28, 426-431. (5) Reamer, H. H.; Sage, B. H. Phase Equilibria in Hydrocarbon Systems. Volumetric and phase behavior of the ethane-n-decane system. J. Chem. Eng. Data 1962, 7, 161-168. (6) Reamer, H. H.; Lower, J. H.; Sage, B. H. Diffusion Coefficients in Hydrocarbon Systems. The Ethane-n-Decane System in the Liquid Phase. J. Chem. Eng. Data 1964, 9, 54-59. (7) Bufkin, B. A.; Robinson, R. L.; Estrera, S. S.; Luks, K. D. Solubility of Ethane in n-Decane at Pressures to 8.2 MPa and Temperatures from 278 to 411 K. J. Chem. Eng. Data 1986, 31, 421-423. (8) Monfort, J. P.; Arriaga, J. L. Chromatographic Determination with an Exponential Dilutor of Henry Constants of Hydrocarbons Mixtures. Chem. Eng. Commun. 1980, 7, 17-25. (9) Jadot, R. Determination de constantes de Henry par chromatographie. J. Chim. Phys. Phys.-Chim. Biol. 1972, 6, 10361040. (10) Singh, H.; Lucien, F. P.; Foster, N. R. Critical Properties for Binary Mixtures of Ethane Containing Low Concentrations of n-Alkane. J. Chem. Eng. Data 2000, 45, 131-135. (11) Hayduk, W.; Walter, E. B.; Simpson P. Solubility of Propane and Carbon Dioxide in Heptane, Dodecane, and Hexadecane. J. Chem. Eng. Data 1972, 17, 59-61. (12) Gonzales, R.; Murrieta-Guevera, F.; Parra, O.; Trejo, A. Solubility of propane and butane in mixtures of n-alkanes. Fluid Phase Equilib. 1987, 34, 69-81. (13) Sciamanna, S. F.; Lynn, S. Solubility of Hydrogen Sulfide, Sulfur Dioxide, Carbon Dioxide, Propane, and n-Butane in Poly(glycol ethers). Ind. Eng. Chem. Res. 1988, 27, 492-499.
(14) King, M. B.; Al-Najjar, H. The solubilities of carbon dioxide, hydrogen sulphide, and propane in some normal alkane solvents. Chem. Eng. Sci. 1977, 32, 1241-1246. (15) Makranczy, J.; Megyery-Balog, K.; Rusz, L.; Patyi, L. Solubility of gases in normal alkanes. Hung. J. Ind. Chem. 1976, 4, 269-280. (16) Henni, A.; Jaffer, S.; Mather, A. E. Solubility of N2O and CO2 in n-Dodecane. Can. J. Chem. Eng. 1996, 74, 554-557. (17) Stewart, W. C.; Nielsen, R. F. Phase equilibria for mixtures of carbon dioxide and several normal saturated hydrocarbons. Bull. Miner. Ind. Exp. Stn., Pa. State Univ.) 1953, 62, 19-29. (18) Schneider, G. Druckeinfluss auf die Entmischung flu¨ssiger Systeme. IV. Entmischung flu¨ssiger n-Alkan-CO2 Systeme bis -60 °C und 1500 bar. Messungen zum Problem der sog. “Entmischung in der Gasphase”. Ber. Bunsen-Ges. Phys. Chem. 1966, 70, 10-16. (19) Hottovy, J. D.; Luks, K. D.; Kohn, J. P. Three-Phase Liquid-Liquid-Vapor Equilibria Behavior of Certain Binary CO2-n-Paraffin Systems. J. Chem. Eng. Data 1981, 26, 256-258. (20) Todd, D. B.; Elgin, J. C. Phase Equilibria in Systems with Ethylene above Its Critical Temperature. AIChE J. 1955, 1, 2027. (21) McFall, T. A.; Post, M. E.; Christensen, J. J.; Izatt, R. M. The excess enthalpies of eight (ethane + alcohol) mixtures at 298.15 K. J. Chem. Thermodyn. 1981, 13, 441-446. (22) Lam, D. H.; Jangkamolkulchai, A.; Luks, K. D. LiquidLiquid-Vapor Phase Equilibrium Behavior of Certain Binary Ethane + n-Alkanol Mixtures. Fluid Phase Equilib. 1990, 59, 263277. (23) Lam, D. H.; Luks, K. D. Multiphase Equilibrium Behavior of the Mixture Ethane + Methanol + 1-Decanol. J. Chem. Eng. Data 1991, 36, 307-311. (24) Patton, C. L.; Luks, K. D. Multiphase Equilibrium Behavior of the Mixture Ethane + 1-Decanol + n-Docosane. Fluid Phase Equilib. 1993, 85, 181-190. (25) Kohn, J. P.; Luks, K. D.; Liu, P. H. Three-Phase SolidLiquid-Vapor Equilibria of Binary-n-Alkane Systems (Ethanen-Octane, Ethane-n-Decane, Ethane-n-Dodecane). J. Chem. Eng. Data 1976, 21, 360-362. (26) Tiffin, D. L.; Kohn, J. P.; Luks, K. D. Three-Phase SolidLiquid-Vapor Equilibria of the Binary Hydrocarbon Systems Ethane-2-Methylnaphthalene, Ethane-Naphthalene, Propanen-Decane, and Propane-n-Dodecane. J. Chem. Eng. Data 1979, 24, 98-100. (27) Fischer, K.; Gmehling J. P-x and γ∞ Data for the Different Binary Butanol-Water Systems at 50 °C. J. Chem. Eng. Data 1994, 39, 309-315. (28) Horstmann, S.; Fischer, K.; Gmehling, J. PSRK group contribution equation of state: Revision and extension III. Fluid Phase Equilib. 2000, 167, 173-186. (29) van Konynenburg, P. H.; Scott, R. L. Critical Lines and Phase Equilibria in Binary van der Waals Mixtures. Phil. Trans. R. Soc. London, Ser. A 1980, 298, 495-540. (30) Schneider, G. M. High-Pressure Phase Diagrams and Critical Properties of Fluid Mixtures Chemical Thermodynamics; McGlashan, M. L., Ed.; Specialist Periodical Reports; Chemical Society: London, 1978; Vol. 2, pp 105-146. (31) Horstmann, S.; Fischer, K.; Gmehling, J.; Kola´r, P. Experimental determination of the critical line for (carbon dioxide + ethane) and calculation of various thermodynamic properties for (carbon dioxide + n-alkane) using the PSRK model. J. Chem. Thermodyn. 2000, 32, 451-464.
Received for review April 18, 2001 Revised manuscript received June 18, 2001 Accepted June 22, 2001 IE0103456