Fluid Phase Equilibria 296 (2010) 46–52

Contents lists available at ScienceDirect

Fluid Phase Equilibria

journal homepage: [Link]/locate/fluid

New analytic apparatus for experimental determination of vapor–liquid

equilibria and saturation densities
Luis A. Galicia-Luna a,∗ , Octavio Elizalde-Solis b
a
Laboratorio de Termodinámica, SEPI-ESIQIE, Instituto Politécnico Nacional, UPALM, Edif. Z, Secc. 6, 1ER piso, Lindavista, C.P. 07738 México D. F., Mexico
b
Departamento de Ingeniería Química Petrolera, SEPI-ESIQIE, Instituto Politécnico Nacional, UPALM, Edif. 8, 2◦ piso, Lindavista, C.P. 07738 México D. F., Mexico

a r t i c l e i n f o a b s t r a c t

Article history: A new experimental apparatus for performing simultaneous determination of high-pressure vapor–liquid
Received 19 December 2009 equilibria (VLE) and saturated densities was developed in this work. The experimental methodology
Received in revised form 27 February 2010 was verified by measuring these properties for the carbon dioxide + 1-propanol and carbon dioxide + 2-
Accepted 2 March 2010
propanol systems from 313 to 363 K. The apparatus is based on the static-analytic method for VLE
Available online 17 March 2010
determinations and was slightly modified by coupling a vibrating U-tube densitometer to obtain satu-
rated densities for both vapor and liquid phases. VLE measurements agreed with previous literature data
Keywords:
and were correlated with the Peng–Robinson equation of state coupled to the Wong–Sandler mixing
Saturated density
Densitometer
rules. Saturation densities at temperatures above 313 K have not been published up to now.
Vapor–liquid equilibria © 2010 Elsevier B.V. All rights reserved.

1. Introduction ide + 2-propanol systems in the temperature range of 313–363 K.

These binary mixtures are of great interest to enhance (solvent + co-
High-pressure phase equilibria and density data for systems solvent) supercritical extractions or solubility studies of specific
containing carbon dioxide are important for developing or testing compounds. Experimental results were in agreement with the data
semi-empirical, theoretical equations of state and in the process published in literature [21–28]. The Peng–Robinson equation of
design using supercritical extraction technology [1–6]. Experimen- state [29] using the classical and Wong–Sandler mixing rules [30]
tal phase equilibria and saturated density data at high pressures was used to correlate the VLE data.
are reported in literature. However, papers on the simultaneous
determination of high-pressure vapor–liquid or vapor liquid–liquid
2. Experimental
equilibria and saturated densities are scarce [7–15].
Our laboratory has different apparatus to obtain accurate
2.1. Materials
and reliable experimental data [16–19]. Experimental volumetric
properties [17] based on the synthetic method [20] and phase equi-
1-Propanol (CAS number [71-23-8]) with a certified purity of
librium measurements [19] are obtained in different apparatuses.
99.7% and 2-propanol (CAS number [67-63-0]) with a certified
Therefore, the aim of this work was to develop a new experimental
purity of 99.5% were supplied by Aldrich Chemical Co. The purity
apparatus suitable to perform simultaneous phase equilibria and
of carbon dioxide (CAS number [124-38-9]) is 99.995% purchased
saturated density determinations in order to avoid several exper-
from Infra and was used as received. Alkanols were used after
iments. The new apparatus is based on the static-analytic method
degassing. An 831 KF coulometer (831, Metrohm) was used to
and consists basically of an equilibrium cell coupled with vibrat-
determine the initial water content for alkanols. It was 1.65 × 10−4
ing tube densitometer. A circulating gear pump is used to send
for 1-propanol and 5.51 × 10−5 for 2-propanol in mass fraction.
vapor or liquid samples from the equilibrium cell to the densito-
Physical properties for each compound are presented in Table 1
meter; in consequence the apparatus only needs one vibrating tube
[31].
densitometer.
The reliability of the apparatus for VLE and saturated densities
was verified with carbon dioxide + 1-propanol and carbon diox- 2.2. Experimental apparatus

Experiments were performed in a new apparatus based on the

static-analytic method coupled to a densitometer to determine
∗ Corresponding author. Tel.: +52 55 5729 6000x55133; fax: +52 55 5586 2728. VLE and saturated densities. The main parts of the apparatus are
E-mail address: lgalicial@[Link] (L.A. Galicia-Luna). constituted by the equilibrium cell, circulating gear pump, vibrat-

0378-3812/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.fluid.2010.03.004
 L.A. Galicia-Luna, O. Elizalde-Solis / Fluid Phase Equilibria 296 (2010) 46–52 47

Fig. 1. Schematic diagram of the experimental apparatus for simultaneous determination of vapor–liquid equilibria and saturated density data. AB: air bath; CGP: circulating
gear pump; VTD: vibrating tube densitometer; EC: equilibrium cell; ELD: electric lift device; GC: gas chromatograph; LB: liquid bath; MR: magnetic rod; MSD: magnetic
stirring device; PI: pressure indicator; SI: sampler-injector; SP: syringe pump; TI: temperature indicator; UC: universal counter.

ing U-tube densitometer, ROLSITM [32] sampler-injector, magnetic 2.3.1. Calibration section
stirring device and the solvent feeding system. The schematic dia- Platinum temperature probes (100-, Specitec) are connected
gram of this apparatus is presented in Fig. 1. to a digital indicator, TI (1529 Chub E4, Fluke) for temperature read-
The stainless steel cell of 40 cm3 has two sapphire windows in ings. Calibration of the temperature probes was performed in a
the front and back side that allow observing the coexisting phases calibration system (F300S, Automatic Systems Laboratories) fitted
in the studied mixture and to follow the position of the ROLSITM with a 25- reference probe (162CE, ±0.005 K certified accuracy,
(SI). A pressure transducer placed outside the oven is connected Rosemount). The temperature uncertainty for the platinum probes
to the cap of the equilibrium cell. Two platinum probes are placed was estimated to be ±0.05 K.
on two wells at the top and bottom right side. The ROLSI inserted Pressure of the mixture was monitored with a pressure trans-
through the cap of the cell takes out mixture samples from the ducer (PDCR 4010-A093, Druck) connected to a digital indicator,
cell and which are sent to the gas chromatograph (GC) through a PI (DPI 145, Druck). It was calibrated against a dead weight bal-
thermo-regulated tubing circuit. ance (Model 5304 class S2, Desgranges & Huot). The estimated
Two holes were made on the left side of the cell to connect the uncertainty for pressure measurements were found to be within
vibrating tube densitometer and the circulating gear pump. A plat- 0.008 MPa.
inum probe is placed in the middle part of the densitometer. After Samples from the equilibrium cell are analyzed using a gas chro-
several test aiming to establish the correct position of the densit- matograph (GC 6890, Agilent). The GC is equipped with a stainless
ometer manually, it was decided to be placed below the lower or steel packed column (PORAPAK Q 80/100, 4 ft × 1/8 in., Alltech) and
above the upper level of the cell by means of an electric lift device. a thermal conductivity detector. Helium was used as carrier gas at
The feeding system consists of the solvent supply tank, the syringe a flow rate of 30 ml/min at 433 K. The temperature of the injec-
pump (100DM, Isco) and the vacuum pump. tor port and the thermal conductivity detector (TCD) was 463 and
473 K, respectively. Known amounts of pure standards of carbon
dioxide, 1-propanol and 2-propanol were injected using calibrated
2.3. Experimental procedure
syringes for the TCD calibration within accuracy of 1%.
The vibrating U-tube densitometer (hastelloy C-276, DMA
The experimental procedure for the vapor–liquid equilibrium
60/512P, Anton Paar) has an internal volume of about 1 cm3 and
measurements was previously published by Galicia-Luna et al. [19].
the oscillation period of the sample is displayed in a universal
This procedure was modified to obtain saturated density measure-
counter (53131A, Agilent). The densitometer was calibrated using
ments simultaneously. It consists of the following parts:
the method described by Galicia-Luna et al. [20]. Water and nitro-
gen are used as reference fluids. Densities for water were calculated
1. Calibration of the platinum temperature probes, pressure trans- with the equation of state developed by Wagner and Pruß [33]. In
ducer, vibrating U-tube densitometer and gas chromatograph the case of the nitrogen densities, the equation proposed by Span et
detector. al. [34], was used for the calculations. Experimental densities were
2. Cell loading. obtained according with the following equation:
3. Setting up the experimental conditions.
4. Vapor–liquid equilibrium and saturation density measurements. F (p, T ) = H2 O (p, T )
  
F2 (p, T ) − H
2 (p, T ) H2 O (p, T ) − N2 (p, T )
2O
Table 1 + 2 2 (p, T )
(1)
H (p, T ) − N
Critical properties and acentric factor of pure compounds from Ref. [31]. 2O 2

MW (kg/kmol) Tc (K) pc (MPa) ω where F (p, T ) is the studied fluid density, H2 O (p, T ) and N2 (p, T )
Carbon dioxide 44.010 304.12 7.374 0.225 are the reference fluid densities, in kg m−3 , i (p, T ) are the vibrating
1-Propanol 60.096 536.78 5.175 0.629 periods in s−1 for each fluid at the same conditions of temperature
2-Propanol 60.096 508.30 4.762 0.665
and pressure.
48 L.A. Galicia-Luna, O. Elizalde-Solis / Fluid Phase Equilibria 296 (2010) 46–52

Table 2
Associated uncertainties in the experimental densities for the two binary mixtures using a vibrating tube densitometer.
2 2
Associated uncertainty Source Estimate Divisor/distribution Influence (∂F /∂z) u(z) (kg2 m−6 )
√
−3
u(H2 O ) (kg m ) Reference equation 0.060 2√3 rectangular 0.017 3.88E−04
u(N2 ) (kg m−3 ) Reference equation 0.034 2 3 rectangular 0.010 7.29E−05
√
u(H2 O ) (ms) Resolution 1.00E−06 2 3 rectangular 2.89E−07 3.68E−03
Repeatability 1.05E−05 1 normal 1.05E−05
√
u(N2 ) (ms) Resolution 1.00E−06 2 3 rectangular 2.89E−07 3.02E−02
Repeatability 3.12E−05 1 normal 3.12E−05
√
u(F ) (ms) Resolution 1.00E−06 2 3 rectangular 2.89E−07 5.59E−03
Repeatability 2.50E−06 1 normal 2.50E−06

u(F ) (kg m−3 ) Combination 0.04

U(F ) (kg m−3 ) Expansion 0.202 k = 2 normal 0.20

According to the law of propagation of errors described in of liquid in the densitometer. The magnetic rod is also activated
Ref. [35] from N.I.S.T., uncertainties for the experimental densities for stirring the mixture to achieve the equilibrium conditions in a
based on Eq. (1) is given by: minor time.
 2  2  2  2  2
∂F ∂F ∂F ∂F ∂F
u(F ) = u2 (H2 O ) + u2 (N2 O ) + u2 (F ) + u2 (H2 O ) + u2 (N2 ) (2)
∂H2 O ∂N2 ∂F ∂H2 O ∂N2

Taking into account both binary systems, associated uncertain-

ties attributed to each variable from Eq. (2) are summarized as
the maximum value in Table 2. In consequence the estimated 2.3.4. Vapor–liquid equilibrium and saturated density
uncertainty of experimental densities was found to be within measurements
0.20 kg m−3 . Isothermal VLE measurements were performed when equilib-
rium was reached (after 5 h) by considering when pressure and
2.3.2. Cell loading temperatures in the cell and densitometer became stable within
Either 1-propanol or 2-propanol was loaded into the equilib- uncertainty of the instrument. At each temperature and pressure,
rium cell and was placed inside the air bath which contains the at least five consecutive samples of vapor or liquid are taken out
ROLSI. The inlet of the circulating gear pump (GAH, Micropump) with the SI and quantified in the gas chromatograph. The compo-
was connected to the lower hole of the cell and the outlet was con- sition of each phase is estimated with the averages of the samples
nected to the lower hole of the densitometer. The upper hole of with a repeatability of ±1%.
densitometer was connected to the upper hole of the equilibrium The vapor phase is initially analyzed based on the above proce-
cell. The pressure transducer was inserted through the cell cap, and dure. After the last five samples, it was observed that the oscillating
three thermometers were inserted in the two wells of the cell and period of the vapor phase in the densitometer became stable. There-
the other in the densitometer. Finally, the feeding system was con- fore the circulating pump was turned off and four measurements of
nected to the equilibrium cell through a three-way valve. Alkanol vibrating periods (at saturated conditions) were registered as well
was degassed for 30 min under stirring by the magnetic rod (MR) as the pressure and temperature. Afterwards, the capillary and the
placed at the bottom of the cell and driven by the MSD. The air densitometer were taken down to the lower level of the cell and
was also evacuated from the rest of the apparatus. Both steps were the flow direction was inverted by the frequency controller. Conse-
done by using a vacuum pump which is connected on the three-way quently, the liquid phase goes from the cell, circulates through the
valve. gear pump and densitometer and returns in the top of the cell. This
allows the complete filling with liquid in the densitometer. Mole
2.3.3. Setting up the experimental conditions fraction for the liquid phase is also estimated by GC analysis and
Carbon dioxide is introduced to the main parts of the apparatus the liquid density by turning off the gear pump when equilibrium
(equilibrium cell, circulating pump and densitometer) to check for was reached.
leaks. It flows from the supply tank and through the syringe pump. After determination of mole fraction and saturated densities
An opening valve located inside the oven is closed to isolate the of vapor and liquid phases, carbon dioxide is again added to the
main parts of the apparatus. system by increasing pressure using the syringe pump. Isothermal
An air bath (Spame) controls the temperature of the cell and the measurements were performed at pressure changes up to near the
head of the pump, while the densitometer is thermo-regulated by critical point of the mixture. Mole fraction uncertainties for the
a liquid bath (9512, PolyScience) and must be initially positioned vapor phase were calculated to be within ±1%. Liquid mole fraction
above the upper level of the cell by means of the electric lift device. uncertainties were found to be less than 2%.
As first test, several temperature settings were carried out to have
the same temperature in the equilibrium cell and vibrating tube 3. Results and modeling
densitometer. Then, the cell, pump and densitometer were set to
the first temperature, 313 K. 3.1. Vapor–liquid equilibria
Finally the pump controlled by a frequency inverter (CFW-08,
Weg) is turned on for circulating the mixture in the following way: Isothermal vapor–liquid equilibria for the carbon dioxide + 1-
vapor from the upper side of the cell flows to the densitometer; then propanol were measured at temperatures of 313.24 and 363.17 K
it enters through the pump and returns at the bottom of the cell. and are presented in Table 3. The VLE are compared with the data
The position of the densitometer and the vapor flow direction was published by Yaginuma et al. [21], Secuianu et al. [22], and Suzuki
found to be the appropriate after several tests to avoid the presence et al. [23] in Fig. 2 at 313 K. Agreement is found when comparing
 L.A. Galicia-Luna, O. Elizalde-Solis / Fluid Phase Equilibria 296 (2010) 46–52 49

Table 3 Table 4
Vapor–liquid equilibria and saturated densities for the carbon dioxide (1) + 1- Vapor–liquid equilibria and saturated densities for the carbon dioxide (1) + 2-
propanol (2) system. propanol (2) system.

p (MPa) x1 y1 m (kg/m3 ) p (MPa) x1 y1 m (kg/m3 )

Liquid Vapor Liquid Vapor

T = 313.24 K T = 313.22 K
3.313 0.2020 0.9963 807.0 70.7 2.402 0.1267 0.9904 783.1 48.7
4.363 0.2572 0.9961 812.6 95.7 3.600 0.2066 0.9923 789.6 74.6
5.284 0.3141 0.9961 817.6 124.4 4.662 0.3029 0.9924 799.3 107.8
6.169 0.4054 0.9957 823.3 161.5 5.533 0.3845 0.9917 803.6 138.3
6.501 0.4405 0.9946 824.4 178.3 5.929 0.4501 0.9910 806.7 155.2
6.999 0.5256 0.9933 826.3 207.3 6.483 0.5540 0.9901 805.4 185.5
7.255 0.5773 0.9923 825.7 227.7 6.950 0.6472 0.9885 794.9 216.2
7.567 0.6774 0.9913 822.0 255.5 7.361 0.7539 0.9870 783.1 242.6
7.883 0.7843 0.9893 786.8 298.9 7.788 0.9020 0.9824 703.7 303.2

T = 363.17 K T = 333.03 K
3.924 0.1483 0.9668 752.0 68.1 2.577 0.1370 0.9767 761.0 46.7
5.932 0.2294 0.9701 754.9 113.3 3.783 0.2077 0.9793 762.7 78.4
8.016 0.3137 0.9688 754.8 165.5 5.384 0.3001 0.9821 765.0 117.2
9.868 0.4007 0.9615 753.6 232.6 6.823 0.4023 0.9790 767.5 166.7
11.728 0.4863 0.9494 749.7 304.6 7.512 0.4645 0.9765 768.5 202.2
12.564 733.0 355.9 8.974 0.6377 0.9687 770.7 284.8
13.575 0.6808 0.9096 680.6 435.6 9.639 0.7582 0.9523 771.4 355.2
14.004 0.7308 0.8836 638.5 484.3 10.145 0.8297 0.9265 771.4 483.0

U(p) = 0.008 MPa, U(T) = 0.05 K, U() = 0.20 kg m−3 , U(x1 ) = 1%, U(y1 ) = 2%. T = 348.25 K
1.801 0.0736 0.9418 741.2 38.9
3.182 0.1432 0.9593 743.9 62.6
with Ref. [21,22], but some deviations can be found with the data 5.073 0.2469 0.9635 746.5 100.5
6.522 0.3234 0.9658 748.6 139.3
presented by Suzuki et al. [23] at pressures lower than about 6 MPa.
8.011 0.4042 0.9613 750.9 184.1
For the carbon dioxide + 2-propanol system, the VLE data are 9.362 0.5132 0.9519 752.9 242.1
listed in Table 4. These measurements were performed at 313.22, 9.833 0.5658 0.9470 753.6 277.7
333.03, 348.25 and 362.92 K. The VLE data for this system was in 11.153 0.7177 0.9165 745.3 388.2
agreement with the literature data at 313 K [24–27] and 333 K 11.300 0.7550 0.8944 646.5 401.5
11.570 0.8241 0.8858 568.0 445.2
[16,24,27,28]. However some deviations can be observed when
comparing against the published data by Yaginuma et al. [25] at T = 362.92 K
2.580 0.1016 0.9255 641.7 47.2
313 K and Khalil et al. [24] and Radosz [28] at 333 K (Fig. 3).
4.014 0.1668 0.9393 690.6 73.3
The Peng–Robinson equation of state (PR EoS) [29]: 5.255 0.2312 0.9431 726.4 101.4
6.835 0.3079 0.9467 728.3 136.6
RT a(T )
p= − (3) 8.248 0.3719 0.9435 728.7 182.4
v−b v(v + b) + b(v − b) 9.567 0.4681 0.9341 726.5 223.3
10.921 0.5653 0.9213 704.5 284.7
is expressed in terms of pressure p, temperature T and v the molar 11.678 0.6285 0.9032 671.6 338.3
volume. The parameters a(T) and b are obtained from the critical 12.423 0.7165 0.8545 590.4 399.1
temperature Tc and critical pressure pc and the acentric factor ω of U(p) = 0.008 MPa, U(T) = 0.05 K, U() = 0.20 kg m−3 , U(x1 ) = 1%, U(y1 ) = 2%.
each pure component.
Eq. (3) was coupled with classical and Wong–Sandler [30] mix-
ing rules to correlate the VLE of the binary systems.

Fig. 3. Vapor–liquid equilibria for the carbon dioxide + 2-propanol system at 313 K:
Fig. 2. Vapor–liquid equilibria for the carbon dioxide + 1-propanol system at 313 K: () Khalil et al. [24]; () Suzuki et al. [27]; () Yaginuma et al. [25]; () Lazzaroni et
() Yaginuma et al. [21]; (♦) Secuianu et al. [22]; () Suzuki et al. [23]; (䊉) this work; al. [26]; (䊉) this work. 333 K: () Elizalde-Solis et al. [16]; () Khalil et al. [24]; ()
() this work at 363 K. Suzuki et al. [27]; (♦) Radosz [28]; () this work.
50 L.A. Galicia-Luna, O. Elizalde-Solis / Fluid Phase Equilibria 296 (2010) 46–52

Table 5
Deviations in pressure and vapor mole fraction from correlations using the Peng–Robinson equation of state.

System T (K) Classical mixing rule Wong–Sandler mixing rule—NRTL parameters

k12 p (%) y (mol/mol) k12  12 (kJ mol−1 )  21 (kJ mol−1 ) p (%) y

CO2 + 1- 313.24 0.108 5.16 0.001 0.457 4.8896 −0.0903 2.10 0.001
propanol 363.17 0.093 9.45 0.004 0.425 5.4102 −0.5357 0.98 0.004

313.22 0.128 7.79 0.002 0.369 2.7759 1.4438 0.53 0.001

CO2 + 2- 333.03 0.111 5.54 0.004 0.471 4.4388 −0.2863 1.89 0.014
propanol 348.25 0.111 8.65 0.009 0.455 4.0120 0.1628 1.61 0.007
362.92 0.113 6.59 0.008 0.434 4.8788 −0.3149 1.13 0.005

The classical mixing rules are expressed as follows: for pressure. In the case of vapor phase mole fraction

am = xi xj (ai aj )1/2 (1 − kij ) (4) 1 
NP
 ex cal 
y = y − y  (14)
NP i i
i j
i=1

bm = xi bi (5)
3.2. Saturated density
i

where kij is the interaction parameter for the binary mixture. Experimental densities in the saturated region for liquid and
The Wong–Sandler mixing rules are represented as: vapor phase were measured simultaneously by coupling a vibrating
 tube densitometer to the equilibrium cell. Saturated densities for
xi xj (b − (a/RT ))ij the carbon dioxide + 1-propanol system are presented in Table 3
bm = i j exc
(6)
1− x (a /bi RT ) − (G /CRT ) at 313.24 and 363.17 K and in Table 4 for the carbon dioxide + 2-
i i i
propanol mixture at 313.22, 333.03, 348.25 and 362.92 K. Although
with different temperature control equipment was used, the differences
(bi − (ai /RT )) + (bj − (aj /RT )) between the temperature of the equilibrium cell and that of the
a
b− = (1 − kij ) (7) densitometer were found to be within the estimated temperature
RT ij 2
uncertainty.
and Experimental densities for the carbon dioxide + 1-propanol sys-
tem were compared and agree satisfactory with the published data
 a Gexc reported by Yaginuma et al. [21] at 313 K and are shown in Fig. 4.
i
am = bm xi + (8)
bi C In the case of the saturated densities for the carbon dioxide + 2-
i propanol system, the data were compared with the data reported
√ √ by Yaginuma et al. [25] and Lazzaroni et al. [26] at 313 K in Fig. 5. Sig-
For the PR EoS C = ln( 2 − 1)/ 2. The excess Gibbs energy Gexc
nificant deviations from the liquid density trend were found when
is calculated using the NRTL model [36]:
comparing our results and those published in Ref. [25] against the

Gexc   g x
j ji j i j
data presented in Ref. [26]. From Fig. 6 it can be seen that liq-
= xi  (9) uid saturated densities for carbon dioxide + 2-propanol are lower
RT g x
i k ki k than densities obtained for the carbon dioxide + 1-propanol sys-
tem at about the same pressure. For the vapor saturated densities,
ıij
ij = (10) these are higher than the measured saturated densities for CO2 + 1-
RT propanol system at constant pressure.
gij = exp(−˛ij ij ) (11) Saturation densities for the two binary systems at 363 K were
computed with Eq. (3) using Eqs. (6) and (8). These results are
The nonrandomness parameter ˛ji for these systems was fixed
to 0.3. The interaction parameters  12 ,  21 and k12 were obtained by
VLE calculation algorithm using the Levenberg–Marquardt method
[37] by minimizing the following objective function:
⎡ 2 2 ⎤

NP
Nc
yijcal − yijex pcal − pex
OF = ⎣ ex +
j j
⎦ (12)
yij pex
j
j=1 i=1

NP is the number of data points, Nc is the number of components, y

is the vapor mole fraction, and the superscripts cal and ex denotes
the calculated and experimental values, respectively.
The PR EoS coupled to the Wong–Sandler mixing rule correlates
with agreement the bubble point data for the VLE of carbon diox-
ide + 1-, and 2-propanol systems. Deviations for bubble pressure
and vapor phase mole fraction are shown in Table 5.
These were calculated by

NP  ex

100  p − pcal 
i i
p = (13)
NP pex
i
Fig. 4. Saturated densities for the carbon dioxide + 1-propanol system at 313 K: (䊉)
i=1 this work; () Yaginuma et al. [21]. 363 K: () this work.
 L.A. Galicia-Luna, O. Elizalde-Solis / Fluid Phase Equilibria 296 (2010) 46–52 51

Table 6
Deviations in pressure and saturation density from the correlations using the PR EoS.

System T (K) Classical mixing rule Wong–Sandler mixing rule—NRTL parameters

k12 p (%) L (%) V (%) k12  12 (kJ mol−1 )  21 (kJ mol−1 ) p (%) L (%) V (%)

CO2 + 1- 313.24 0.098 6.3 0.8 14.5 0.498 3.1895 0.7589 3.5 2.2 4.9
propanol 363.17 0.121 8.9 6.2 16.7 0.506 1.4108 1.4330 5.1 3.8 10.2

313.22 0.128 7.8 3.1 14.4 0.512 3.1070 1.3285 5.3 2.2 3.6
CO2 + 2- 333.03 0.109 5.6 3.2 14.0 0.483 4.7698 −0.2117 6.7 3.0 6.1
propanol 348.25 0.108 8.7 4.5 8.7 0.445 5.5489 −0.1212 5.8 7.1 6.0
362.92 0.122 7.9 4.0 13.8 0.434 5.3886 −0.2011 4.6 6.6 4.6

Fig. 5. Saturated densities for the carbon dioxide + 2-propanol system at 313 K: ()
Fig. 7. Saturated densities at 363 K: () carbon dioxide + 1-propanol; () carbon
this work; (䊉) Yaginuma et al. [25]; () Lazzaroni et al. [26].
dioxide + 2-propanol. Solid lines represents the correlation with PR EoS coupled to
the Wong–Sandler mixing rules.
shown in Fig. 7. The behavior at lower temperatures is the same.
Deviations for densities in both phases listed in Table 6 were cal- late the experimental VLE data; however deviations in pressure and
culated using Eq. (15). density could not be lowered and are about of the same magnitude

NP  ex
 as the results shown in Table 6.
100   − cal 
i i
 = (15)
NP iex 4. Conclusions
i=1

As can be observed these representations were not satisfactory Simultaneous determination of the vapor–liquid equilibria and
using either the classical or Wong–Sandler mixing rules. Param- saturated densities were performed in a new static-analytic appa-
eters were optimized by adding the terms of the calculated and ratus coupled to a vibrating tube densitometer. The experimental
experimental vapor and liquid densities in Eq. (12). The Patel-Teja methodology was verified by measuring the VLE and saturated
EoS [38] and the Wong–Sandler mixing rule was also used to corre- densities for the carbon dioxide + 1-propanol and carbon diox-
ide + 2-propanol systems from 313 to 363 K. The literature data
agree satisfactorily with the obtained results for both properties.
In the case of the saturated properties at temperatures higher than
313 K, these contribute with new data. Correlation of the VLE data
using the PR EoS and the WS model were found to be satisfactory.
However for calculated saturation densities, deviations are not in
agreement with the experimental data.

List of symbols

a,b parameters given by Eq. (3)

Gexc excess Gibbs energy
k interaction parameter
NP number of data points from Eq. (12)
Nc number of components from Eq. (12)
p pressure
R universal gas constant
T temperature
Fig. 6. Saturated densities for the carbon dioxide + 1-propanol: () Yaginuma et al.
[21] and (䊉) this work at 313 K; () this work at 363 K. Carbon dioxide + 2-propanol: v molar volume
() Yaginuma et al. [25] and () this work at 313 K; () this work at 363 K. ω acentric factor
52 L.A. Galicia-Luna, O. Elizalde-Solis / Fluid Phase Equilibria 296 (2010) 46–52

y vapor fraction [6] F.F. Betancourt-Cárdenas, L.A. Galicia-Luna, A.L. Benavides, E. Scholl-
x mole fraction Paschinger, Mol. Phys. 106 (2008) 113–126.
[7] D. Kodama, J. Miyazaki, M. Kato, T. Sako, Fluid Phase Equilib. 219 (2004) 19–23.
z variable from Eq. (2) [8] D. Kodama, R. Sato, A. Haneda, M. Kato, J. Chem. Eng. Data 50 (2005) 122–124.
[9] D. Kodama, T. Yagihashi, T. Hosoya, M. Kato, J. Chem. Eng. Data 54 (2009)
1046–1048.
Greek letters
[10] M. Stievano, N. Elvassore, J. Supercrit. Fluids 33 (2005) 7–14.
 density [11] I. Tsivintzelis, D. Missopolinou, K. Kalogiannis, C. Panayiotou, Fluid Phase Equi-
 vibrating period from Eq. (1) lib. 224 (2004) 89–96.
 ij interaction NRTL parameter [12] D. Missopolinou, I. Tsivintzelis, C. Panayiotou, Fluid Phase Equilib. 238 (2005)
204–209.
˛ji nonrandomness parameter [13] M. Kato, H. Tanaka, H. Yoshikawa, J. Chem. Eng. Data 44 (1999) 116–117.
[14] Y. Higashi, J. Chem. Eng. Data 44 (1999) 328–332.
[15] M.A. Barrufet, S. Rahman, J. Chem. Eng. Data 42 (1997) 120–123.
Subscripts [16] O. Elizalde-Solis, L.A. Galicia-Luna, L.E. Camacho-Camacho, Fluid Phase Equilib.
F fluid phase given by Eq. (1) 259 (2007) 23–32.
H2 O water given by Eq. (1) [17] L.A. Zúniga-Moreno, F.F. Galicia-Luna, Betancourt-Cárdenas, Fluid Phase Equi-
lib. 236 (2005) 193–204.
N2 nitrogen given by Eq. (1)
[18] A. Zúniga-Moreno, L.A. Galicia-Luna, L.E. Camacho-Camacho, J. Chem. Eng. Data
c critical 50 (2005) 1030–1037.
m mixture [19] L.A. Galicia-Luna, A. Ortega-Rodriguez, D. Richon, J. Chem. Eng. Data 45 (2000)
i,j components 265–271.
[20] L.A. Galicia-Luna, D. Richon, H. Renon, J. Chem. Eng. Data 39 (1994) 424–431.
[21] R. Yaginuma, T. Nakajima, H. Tanaka, M. Kato, Fluid Phase Equilib. 144 (1998)
Superscripts 203–210.
[22] C. Secuianu, V. Feroiu, D. Geana, J. Chem. Eng. Data 53 (2008) 2444–2448.
cal calculated values [23] K. Suzuki, H. Sue, M. Itou, R.L. Smith, H. Inomata, K. Arai, S. Saito, J. Chem. Eng.
ex experimental values Data 35 (1990) 63–66.
[24] W. Khalil, C. Coquelet, D. Richon, J. Chem. Eng. Data 52 (2007) 2032–2040.
[25] R. Yaginuma, T. Nakajima, H. Tanaka, M. Kato, J. Chem. Eng. Data 42 (1997)
Acknowledgment 814–816.
[26] M.J. Lazzaroni, D. Bush, J.S. Brown, C.A. Eckert, J. Chem. Eng. Data 50 (2005)
Authors wish to express their gratitude to Consejo Nacional de 60–65.
[27] T. Suzuki, N. Tsuge, K. Nagahama, Fluid Phase Equilib. 67 (1991) 213–226.
Ciencia y Tecnologia and Instituto Politécnico Nacional (MEXICO) [28] M. Radosz, J. Chem. Eng. Data 31 (1986) 43–45.
for the financial support of this research. [29] D.Y. Peng, D.B. Robinson, Ind. Eng. Chem. Fundam. 15 (1976) 59–64.
[30] D.S.H. Wong, S.I. Sandler, AIChE J. 38 (1992) 671–680.
[31] B.E. Poling, J.M. Prausnitz, J.P. O’Connell, The Properties of Gases and Liquids,
References 5th ed., McGraw-Hill, New York, 2001, pp. A6, A8.
[32] P. Guilbot, A. Valtz, H. Legendre, D. Richon, Analusis 28 (2000) 426–431.
[1] G. Brunner, Gas Extraction. An Introduction to Fundamentals of Supercritical [33] W. Wagner, A. Pruß, J. Phys. Chem. Ref. Data 31 (2002) 387–535.
Fluids and the Application to Separation Processes, Springer, New York, 1994. [34] R. Span, E.W. Lemmon, R.T. Jacobsen, W. Wagner, Int. J. Thermophys. 19 (1998)
[2] M.A. McHugh, V.J. Krukonis, Supercritical Fluid Extraction Principles and Prac- 1121–1132.
tice, 2nd ed., Butterworth-Heinemann, USA, 1994. [35] B.N. Taylor, C.E. Kuyatt, Guidelines for evaluating and expressing the uncer-
[3] Z.Q. Hu, [Link]. Yang, Y.-G. Li, Fluid Phase Equilib. 205 (2003) 1–15. tainty of NIST measurement results. NIST Technical Note 1297, 1994 edition.
[4] F.F. Betancourt-Cárdenas, L.A. Galicia-Luna, S.I. Sandler, Mol. Phys. 105 (2007) [36] H. Renon, J.M. Prausnitz, AIChE J. 14 (1968) 135–144.
2987–2998. [37] W.H. Press, S.A. Teukolsky, W.T. Vetterling, B.P. Flannery, Numerical Recipes in
[5] F.F. Betancourt-Cárdenas, L.A. Galicia-Luna, S.I. Sandler, Fluid Phase Equilib. 264 Fortran 77, 2nd ed., Cambridge University Press, USA, 1999, pp. 678–683.
(2008) 174–183. [38] N.C. Patel, A.S. Teja, Chem. Eng. Sci. 37 (1982) 463–473.