J. of Supercritical Fluids 31 (2004) 27–32
Styrene–carbon dioxide phase equilibria at high pressures Mesut Akgün∗ , Didem Emel, Nil Baran, Nalan A. Akgün, Sennur Deniz, Salih Dinçer∗ Yildiz Technical University, Chemical Engineering Department, Davutpasa Campus, No:127, 34210 Esenler, Istanbul, Turkey Received 12 March 2003; received in revised form 13 August 2003; accepted 30 September 2003
Abstract High pressure vapor–liquid equilibrium data for the binary system styrene and carbon dioxide were measured using the static method. Experiments were carried out at temperatures and pressures between 333.15 and 348.15 K, and 6 and 13 MPa, respectively. Experimental results were correlated with the Peng–Robinson equation of state using conventional mixing rules with one interaction parameter. © 2003 Elsevier B.V. All rights reserved. Keywords: Vapor–liquid equilibria; Equation of state; High pressure; Styrene; Carbon dioxide
1. Introduction Styrene is an important raw material, which is usually produced by dehydrogenation of ethylbenzene. As the reaction is not completed, the product contains about 65% styrene and 35% ethylbenzene, which are usually separated by vacuum distillation. Separation of these components by supercritical fluids, having liquid-like densities and good solvent properties, can be preferred to vacuum distillation as an alternative process because of low operating temperatures and leaving no solvent residue in the product [1]. Not only separation processes but also chemical reactions are accomplished in supercritical CO2 . For example, Lin and Akgerman [2] carried out catalytic hydroformu∗ Corresponding author. Tel.: +90-212-449-1925; fax: +90-212-449-18-95. E-mail address:
[email protected] (M. Akgün).
lation of styrene in supercritical CO2 . They indicated that the catalyst activity is higher in the supercritical region [2]. Polymerization of styrene in supercritical CO2 has been extensively applied recently, since CO2 is separated from the product easily and being nontoxic environmentally [3–7]. Design and development of the processes mentioned above depend on the knowledge of the phase behavior of the system styrene and CO2 . Jiang et al. [8] have measured styrene solubility in supercritical CO2 at pressures of 8–24 MPa and temperatures of 323, 328 and 333 K using a dynamic method. Suppes and Mc Hugh [9] have measured the phase behavior of styrene–CO2 in the pressure range of 2.69–16.24 MPa for the temperatures of 308, 328, 353 and 373 K using a synthetic method. Tan et al. [1] also have measured the phase behavior of styrene–CO2 system in the pressure range of 1.4–8.5 MPa for the temperatures of 308, 318 and 328 K.
0896-8446/$ – see front matter © 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.supflu.2003.09.019
28
M. Akgün et al. / J. of Supercritical Fluids 31 (2004) 27–32
Nomenclature a, b aij AAD EoS fˆ i kij N P PR R T V x, y
EoS parameters mixing rule parameter average absolute deviation equation of state fugacity of a component i adjustable binary interaction parameter number of data points pressure Peng–Robinson universal gas constant temperature molar volume liquid, vapor mole fractions
Greek letters α temperature-dependent equation of state parameter κ Peng–Robinson equation parameter ˆφ fugacity coefficient ω acentric factor Superscripts cal calculated exp experimental liq liquid phase vap vapor phase Subscripts b boiling point c critical property i, j components i and j
In this work, vapor–liquid equilibria (VLE) of the styrene–CO2 binary were measured at 333.15– 348.15 K and 6–13 MPa. The results were correlated with the Peng–Robinson equation of state (PR-EOS) using the van der Waals mixing rules with one interaction parameter.
2. Theoretical study The general phase equilibrium relation for a liquid– gas mixture [10] is:
liq vap fˆ i (T, P, xi ) = fˆ i (T, P, yi )
(1)
or liq vap φˆ i xi = φˆ i yi
(2)
liq vap where φˆ i and φˆ i are the fugacity coefficients for liquid and vapor phases, respectively, and are calculated by the Peng–Robinson-EoS described below [11]:
P=
RT a − V − b V(V + b) + b(V − b)
(3)
The parameters a and b are obtained from the following expressions using critical properties [11,12]: a = 0.457235
R2 Tc2 α Pc
(4)
b = 0.077796
RTc Pc
(5)
α= 1+κ 1−
T Tc
2
κ = 0.37464 + 1.54226ω − 0.2692ω2
(6) (7)
For the mixture calculations, the Van der Waals mixing rules are used: a= (8) xi xj aij i
j
xi bi
(9)
√ aij = (1 − kij ) ai aj
(10)
b=
i
where
Here kij denotes the adjustable binary parameter accounting for the interactions between i and j components. Its use improves the agreement between experimental and predicted results. kij was determined by minimizing the average absolute deviation (AAD) [13] defined as: Ny Nx exp cal 1 1 yexp − ycal x − x AAD = + Ny xexp yexp Nx (11) where N is the number of data points.
M. Akgün et al. / J. of Supercritical Fluids 31 (2004) 27–32
29
Table 1 Thermophysical properties of styrene and CO2 [8] Components
MW (g/mol)
Tb (K)
Tc (K)
Pc (MPa)
ω
Styrene Carbon dioxide
104.1 44.0
418.3 194.7
647.0 304.2
3.99 7.38
0.257 0.225
3. Experimental 3.1. Materials Styrene (ACROS) was used without further purification, and its purity was more than 99% as determined by gas chromatographic (GC) analysis. The main impurity is hydroquinone used as inhibitor. CO2 (purity 99.9%) was obtained from HABAS, Istanbul. 3.2. Experimental procedure The vapor–liquid equilibrium data of the system styrene–CO2 were taken as a function of pressure and temperature using the static method. The physical properties of styrene and CO2 studied in this work are listed in Table 1. The experiments were carried out at 333.15, 338.15, 343.15 and 348.15 K and pressures from 6 to 13 MPa. The experimental apparatus used in this study was described in detail elsewhere [14], and is briefly described here. Fig. 1 shows a schematic diagram of the experimental apparatus used to take phase equi-
librium data using the static method. The equilibrium cell has an internal volume of 115 ml, and is heated by a heating tape using a PID controller (±1 K). The temperature measurements were performed on the surface of the well insulated equilibrium cell. CO2 was compressed to the system pressure by a syringe pump (ISCO-model 260D) which could read pressures within ±0.01 MPa. The equilibrium cell was loaded with styrene (40 ml) in each run, and mixed by a tiny magnetic stirring bar. Our preliminary trials showed that the system reached equilibrium around 30 min. Thus stirring the system for 30 min to reach equilibrium at the desired operating conditions, samples of the vapor phase and the liquid phase were collected in sample tubes. The sample size is 2 ml in the vapor phase, and 0.5 ml in the liquid phase. The temperature was maintained constant all along the sample line by insulation of the line. While the styrene in the vapor and liquid samples were separately trapped into the collection bottle containing methanol, the amount of CO2 in the samples was measured by a wet-test meter. After the experiment was terminated, the expansion valves
Fig. 1. Experimental setup for VLE measurements (1) CO2 cylinder, (2) syringe pump, (3) equilibrium cell, (4) digital pressure display, (5) magnetic stirrer, (6) sample tube, (7) cold trap, (8) wet-test meter (9) temperature controller with PID, (10) CO2 vent, (11) heating tape.
30
M. Akgün et al. / J. of Supercritical Fluids 31 (2004) 27–32 Table 2 Vapor–liquid equilibrium data for styrene–CO2 system
were washed twice with methanol, and the concentration of styrene in each phase was determined by gas chromatographic (GC) analysis. 3.3. Analytical procedure For the analysis a gas chromatograph was used Unicam Model 610 GC with FID detector. The separation was obtained by using a capillary column (EC-WAX Carbowax, 30 m × 0.32 mm i.d., film thickness 0.25 m). Argon was used as the carrier gas. The GC was temperature-programmed between 60 and 150 ◦ C at 5 ◦ C/min. The concentration of styrene was computed from GC peak areas using the calibration curve for styrene–methanol mixture. Using the concentration of styrene and the amount of styrene–methanol mixture in the collection bottle, the amount of styrene in CO2 present in the sampling tubes was determined for each run.
4. Results and discussion The vapor–liquid equilibrium data were measured for the styrene–CO2 system at 333.15, 338.15, 343.15 and 348.15 K, respectively. The experimental data are given in Table 2, and the resulting plots as a function of pressure at each temperature are illustrated in Fig. 2. The experiments were repeated several times with different loadings to get both vapor and liquid phase mole fractions. The reproducibility of phase compositions was within ±0.12%. Styrene is a liquid at experimental conditions, and vapor–liquid equilibrium data can be taken and correlated at the pressures under the critical point of the mixture. Many equations of state give satisfactory results under the critical point of the mixture. Various modeling procedures using an EoS have been proposed in literature to predict the phase behavior of monomer–CO2 systems. Computationally and
P (MPa)
xexp.
xcal.
yexp.
ycal.
333.15 (K) 6.03 7.58 9.02 9.84 10.32 10.65 11.00 11.10 11.20
0.508 0.632 0.714 0.783 0.818 0.859 0.906 0.918 0.930
0.506 0.619 0.723 0.786 0.825 0.854 0.889 0.900 0.912
0.998 0.996 0.993 0.991 0.989 0.987 0.983 0.977 0.900
0.996 0.995 0.993 0.990 0.988 0.984 0.979 0.976 0.971
338.15 (K) 6.24 7.42 7.75 9.42 10.28 11.23 11.57 11.84 11.90
0.470 0.560 0.579 0.717 0.778 0.812 0.874 0.895 0.932
0.491 0.572 0.593 0.703 0.761 0.830 0.857 0.882 0.888
0.994 0.995 0.995 0.993 0.989 0.984 0.981 0.967 0.895
0.996 0.995 0.994 0.992 0.989 0.982 0.977 0.971 0.969
343.15 (K) 6.00 7.82 8.72 9.56 10.54 11.46 11.80 12.70 12.93 13.10
0.435 0.562 0.627 0.660 0.720 0.770 0.830 0.890 0.924 0.930
0.449 0.564 0.619 0.670 0.729 0.787 0.810 0.878
0.996 0.995 0.995 0.994 0.992 0.984 0.979 0.967 0.960 0.912
0.995 0.994 0.993 0.991 0.988 0.983 0.980 0.962
348.15 (K) 6.29 6.96 7.98 9.05 10.11 11.20 12.70 13.42
0.428 0.484 0.553 0.612 0.667 0.754 0.827 0.864
0.444 0.484 0.544 0.605 0.665 0.726 0.815 0.865
0.996 0.995 0.994 0.993 0.991 0.986 0.977 0.954
0.994 0.994 0.993 0.991 0.989 0.985 0.973 0.958
Table 3 The values of AAD at different temperatures for kij = 0.045 and kij = 0.0 333.15 (K) kij AAD
0.045 0.023
338.15 (K) 0.0 0.066
0.045 0.035
343.15 (K) 0.0 0.092
0.045 0.019
348.15 (K) 0.0 0.110
0.045 0.017
0.0 0.109
M. Akgün et al. / J. of Supercritical Fluids 31 (2004) 27–32
31
Fig. 2. Vapor–liquid equilibrium data for styrene–CO2 system: kij = 0.045 for all temperatures using PR-EoS.
thermodynamically the most consistent method is fitting to experimental equilibrium data with the PR-EoS [11] using two adjustable parameters. Usually the second adjustable parameter, accounting for the size effect, is taken equal to zero, because predictions using only one adjustable parameter, kij , yielded, in general, satisfactory results. Since temperature dependence of kij is weak, all the data obtained in this work were well correlated by PR-EoS using one parameter, kij , at a constant average value of 0.045, and the calculated results are shown in Fig. 2 together with the experimental data. Table 3 compares AAD values at different temperatures for kij = 0.045, and when kij = 0 so that geometric mean mixing rule is assumed for aij calculation in Eq. (10). The results show that deviations are much larger when kij = 0. However, the calculated results for kij = 0 could not be plotted in Fig. 2 because of the overlapping of curves. It can be concluded that the correlation of the experimental data of this work by the PR-EoS with one adjustable parameter is quite acceptable. Jiang et al. [8] gave only vapor phase solubility data for the styrene–CO2 system measured in the dynamic method at 323–333 K. However, these solubility data cannot be compared exactly with our VLE data, because they are given in g/l units. Other data for vapor–liquid equilibria of the styrene–CO2 system were reported at different conditions [1,9]. Table 4
Table 4 Comparison of vapor–liquid equilibrium data for styrene–CO2 system of this work with literature at different conditionsa Xcal.
yexp.
ycal.
333.15 K [this work] 6.03 0.508 7.58 0.632 9.02 0.714 9.84 0.783 10.32 0.818
0.506 0.619 0.723 0.786 0.825
0.998 0.996 0.993 0.991 0.989
0.996 0.995 0.993 0.990 0.988
328 K [1] 5.58 6.97 8.31
0.354 0.477 0.614
0.505 0.615 0.723
0.997 0.996 0.994
0.997 0.996 0.995
328.15 K [9] 7.19 8.95 9.84
0.512 0.735 0.883
0.632 0.775 0.859
348.15 K [this work] 6.96 0.484 9.05 0.612 12.70 0.827
0.484 0.605 0.815
353.15 K [9] 6.87 9.41 12.37
0.456 0.595 0.750
P (MPa)
a
xexp.
0.3662 0.5121 0.735
CO2 phase compositions are given in the table. Calculated results are obtained by using Peng–Robinson EoS with a constant kij = 0.045.
32
M. Akgün et al. / J. of Supercritical Fluids 31 (2004) 27–32
shows some of the comparable data given in literature together with the data of this work and model calculation results using Peng–Robinson EoS with a constant kij = 0.045 for all data points. Although experimental and model calculated vapor compositions for all data agree quite well, corresponding experimental and model calculated liquid compositions of other researchers deviate somewhat. However, model calculations for our data do not yield such deviations. Apparently a constant value of kij for a wide range of experimental conditions is not suitable for model calculations. In conclusion, Table 4 reflects the relationship between different sets of data satisfactorily, and considering the different experimental conditions used for the different sets of data reported, the agreement between different sets of data are acceptable.
Acknowledgements This work was supported by DPT (Turkish State Planning Organization, project no 22-DPT-07-01-01) and YTU-BAPK (Yildiz Technical University, Organization of Scientific Research Project, project no 21-07-01-02).
References [1] C.S. Tan, S.J. Yarn, J.H. Hsu, Vapor-liquid equilibria for the systems carbon dioxide-etilbenzene and carbon dioxidestyrene, J. Chem. Eng. Data 36 (1991) 23.
[2] B. Lin, A. Akgerman, Styrene hydroformylation in supercritical carbon dioxide: rate and selectivity control, Ind. Eng. Chem. Res. 40 (2001) 1113. [3] J.R. Combes, Z. Guan, J.M. DeSimone, Homogeneous free-radical polymerizations in carbon dioxide. 3. Telomerization of 1,1-difluoroethylene in supercritical carbon dioxide, Macromolecules 27 (1994) 865. [4] V.P. Saraf, E. Kiran, Free radical polymerization of styrene in supercritical fluids, Polymer 31 (1) (1990) 687. [5] H. Shiho, J.M. DeSimone, Preparation of micron-size polystyrene particles in supercritical carbon dioxide, J. Polym. Sci. 37 (1999) 2429. [6] H. Shiho, J.M. DeSimone, Dispersion polymerizations of styrene in supercritical carbon dioxide utilizing random copolymers containing a fluorinated acrylate for preparing micron-size polystyrene particles, J. Polym. Sci. 38 (2000) 1146. [7] D.A. Canelas, J.M. DeSimone, Dispersion polymerizations of styrene in carbon dioxide stabilized with poly(styrene-bdimethylsiloxane), Macromolecules 30 (19) (1997) 5673. [8] C. Jiang, Q. Pan, Z. Pan, Solubility behavior of solids and liquids in compressed gases, J. Supercrit. Fluids 12 (1998) 1. [9] J.G. Suppes, A.M. Mc Hugh, Phase behavior of the carbon dioxide-styrene system, J. Chem. Eng. Data 34 (1989) 310. [10] S.I. Sandler, Chemical and Engineering Thermodynamics, second ed., Wiley, New York, 1989. [11] D.Y. Peng, D.B. Robinson, A new two constant equation of state, Ind. Eng. Chem. Fundam. 15 (1976) 59. [12] S. Raeissi, C.J. Peters, Bubble-point pressures of the binary system carbon dioxide + linalool, J. Supercrit. Fluids 20 (2001) 221. [13] Y.J. Sheng, P.C. Chen, Y.P. Chen, D.S.H. Wong, Calculations of solubilities of aromatic compounds in supercritical carbon dioxide, Ind. Eng. Chem. Res. 31 (1992) 967. [14] M. Akgün, N. Akgün, S. Dinçer, Phase behaviour of essential oil components in supercritical carbon dioxide, J. Supercrit. Fluids 15 (1999) 117.
