298 J. Chem. Eng.

Data 1991, 36, 298-303

Liquid-Liquid Equilibria for Ternary Systems Formed by Acetonitrile,

Water, and Aromatic Hydrocarbons

Serglo DI Cave and Barbara Mazzarotta'

Department of Chemical Engineering, University of Rome "La Sapienza ", Via Eudossiana 18, I-00 184 Rome, Italy

Table I. Refractive Indexes n of the Chemicals

Llquld-llquld equlllbrla were measured at 25 and 45 OC for
the ternary systems formed by acetonltrlle water+ an + n(D,20 "c)
aromatlc hydrocarbon (benzene, toluene, m-xylene, chemical this work lit."
p-xylene, ethylbenzene). Type I Isotherms were acetonitrile 1.3430 1.34411
observed for the flve systems both at 25 and at 45 O C . benzene 1.5011 1.501 12
The blnodal curves and tie lines may be predlcted by toluene 1.4961 1.496 93
using the NRTL model on the basls of parameters found n-xylene 1.4973 1.497 22
p-xylene 1.4958 1.495 82
from blnary data. ethylbenzene 1.4959 1.495 88
"Riddick, J. A.; Bunger, W. B.; Sakano, T. K. In Organic Sol-
uents, 4th ed.; Weissberger, A., Ed.; Techniques of Chemistry, Vol.
I ntroductlon 11; John Wiley & Sons: New York, 1986.

The presence of an azeotrope renders difficult the recovery Table 11. Ternary Liquid-Liquid Equilibrium Data (xUe
of acetonitrile from aqueous solutions by distillation. Liquid Mass Fraction) and Percentage Standard Deviations
extraction with an aromatic hydrocarbon as solvent has been (lOOa(xU),Equation 6) for the Acetonitrile (1)-Water
suggested as an alternative way to perform this separation ( 7). (2)-Benzene (3) System
The extraction of acetonitrile from aqueous mixtures has tie lines
been studied for a number of organic solvents (2-5). binodal curve aqueous phase organic phase
I n the present work, ternary liquid-liquid equilibria (LLE)were x,2 xw3 XW2 x w3 XW2 xw3
measured at 25 and 45 OC for the five ternary systems con- At 25 "C
+ + +
taining acetonitrile (1) water (2) benzene (3), toluene (3), 0.006 0.773 0.875 0.002 0.006 0.728
+ + +
m-xylene (3), p-xylene (3), or ethylbenzene (3). The 0.006 0.719 0.855 0.003 0.013 0.656
information reported in the literature concerning these system 0.012 0.657 0.827 0.003 0.019 0.578
appears to be limited. I n particular, Hartwig et al. (2)reported 0.022 0.529 0.809 0.004 0.029 0.492
three tie lines for the system acetontrile-water-benzene at 25 0.038 0.445 0.800 0.004 0.034 0.465
0.048 0.376 0.783 0.005 0.041 0.400
OC. Francis (3) reported in small triangular graphs the binodal 0.063 0.305 0.754 0.005 0.059 0.316
curves for the systems acetonitrile-water-benzene and ace- 0.070 0.281 0.746 0.005 0.069 0.282
tonitrile-water-toluene at 25 OC. Rao et al. ( 4 ) reported com- 0.088 0.230 0.735 0.006 0.105 0.195
plete LLE data for the system acetonitrile-water-toluene at 30 0.117 0.177 0.720 0.006 0.137 0.151
OC, and Rao et al. (5)reported similar data for the system 0.139 0.143
0.167 0.117
acetonitrile-water-mixture of xylenes at 31 O C .
0.211 0.086
0.251 0.068
0.339 0.045
Experimental Sectlon 0.451 0.027
0.544 0.015
Materiels. All chemicals used in this work had a minimum 0.641 0.009
stated purity from the manufacturer (Farmitalia Carlo Erba) as
A t 45 "C
follows: bidistilled water; acetonitrile, 99.8% ; benzene, toluene, 0.001 0.950 0.960 0.0005 0.001 0.946
and ethylbenzene, 99.5%; m-xylene and p-xylene, 99%. No 0.002 0.862 0.910 0.002 0.003 0.847
further purification was done. The actual refractive indexes of 0.010 0.750 0.850 0.004 0.009 0.752
the chemicals are given in Table I. 0.019 0.637 0.833 0.005 0.018 0.650
-ratus and procekrre. The measurements were carried 0.030 0.536 0.809 0.006 0.027 0.555
out in a glass equllibrium cell, 250 mL in capacity, where the 0.048 0.444 0.794 0.006 0.041 0.466
0.080 0.328 0.748 0.009 0.065 0.369
temperature was maintained constant to within f0.1 OC. First 0.110 0.258 0.735 0.010 0.084 0.306
of all, the limits of the Immiscibility region were determined. To 0.139 0.207 0.690 0.015 0.113 0.247
this end, binary mixtures were prepared from known amounts 0.168 0.175
of two miscible components and fixed volumes (0.05 mL each 0.205 0.143
time) of the third one were added until a second liquid phase 0.248 0.114
appeared. After each addition the solution was vigorously 0.268 0.104
0.377 0.066
stirred and then allowed to settle for about 10 min. The ac- 0.470 0.042
curacy of the method was fO.OO1 mass fraction. The mea- 0.578 0.026
surements proved to be reproducible to within fO.OO1 mass 0.681 0.016
fraction. 100cr(x,) = 2.5
The tie lines were then determined by preparing a ternary
mixture from weighed amounts of the three components, basis, by using the graduation on the wall of the cell: samples
equilibrating the system by agltatlon with a magnetic stlrrer for of the light and of the heavy liquids were withdrawn, and thelr
at least 2 h, and allowing it to settle for 20 h. The relative densities were measured by means of a pycnometer to allow
amounts of the two phases were measured, on a volumetric the conversion of the results from a volume to a mass basis.

0021-9568/91/1736-0298%02.50/0 0 1991 American Chemical Society

 Journal of Chemical and Engit?8erirlngData, Vd. 36,No. 3, 1991 299

Table 111. Ternary Liquid-Liquid Equilibrium Data (xwj, Table IV. Ternary Liquid-Liquid Equilibrium Data ( x ~ ,
Mass Fraction) and Percentage Standard Deviations Mass Fraction) and Percentage Standard Deviations
( 100u(xw),Equation 6) for the Acetonitrile (1)-Water (lOOu(xw), Equation 6) for the Acetonitrile (1)-Water
(2)-Toluene (3) System (2)-m-X~lene(3) System
tie lines tie lines
binodal curve aqueous phase organic phase binodal curve aqueous phase organic phase
XW2 xw3 XWZ xw3 XW2 W9 XW2 xw3 XW2 xw3 XW2 xu3
At 25 O C At 25 O C
0.006 0.797 0.923 0.001 0.005 0.872 0.005 0.716 0.880 0.0005 0.0005 0.971
0.015 0.609 0.882 0.002 0.007 0.812 0.015 0.613 0.848 0.001 0.002 0.894
0.023 0.489 0.847 0.003 0.010 0.698 0.026 0.510 0.800 0.002 0.006 0.750
0.041 0.400 0.836 0.003 0.012 0.660 0.033 0.456 0.757 0.003 0.017 0.598
0.062 0.293 0.827 0.003 0.013 0.645 0.041 0.405 0.726 0.004 0.030 0.485
0.085 0.240 0.820 0.004 0.015 0.611 0.051 0.354 0.677 0.005 0.056 0.330
0.114 0.178 0.796 0.004 0.023 0.527 0.055 0.351 0.622 0.008 0.104 0.188
0.137 0.153 0.790 0.004 0.028 0.500 0.076 0.262
0.139 0.144 0.771 0.005 0.040 0.400 0.113 0.196
0.161 0.121 0.754 0.005 0.049 0.336 0.151 0.130
0.192 0.107 0.744 0.006 0.054 0.311 0.206 0.088
0.222 0.072 0.742 0.006 0.059 0.293 0.275 0.059
0.249 0.069 0.738 0.006 0.075 0.246 0.292 0.050
0.280 0.065 0.340 0.042
0.295 0.049 0.341 0.039
0.378 0.037 0.358 0.039
0.400 0.030 0.374 0.036
0.474 0.021 0.450 0.021
0.581 0.013 0.555 0.010
0.691 0.007
At 45 "C
At 45 "C 0.004 0.889 0.891 0.003 0.005 0.902
0.001 0.904 0.917 0.001 0.001 0.910 0.007 0.828 0.844 0.003 0.008 0.817
0.007 0.795 0.890 0.002 0.006 0.833 0.011 0.745 0.802 0.005 0.009 0.749
0.013 0.685 0.835 0.003 0.009 0.721 0.021 0.623 0.748 0.007 0.019 0.619
0.026 0.588 0.803 0.004 0.018 0.642 0.038 0.500 0.713 0.009 0.033 0.510
0.039 0.479 0.782 0.005 0.027 0.563 0.066 0.376 0.688 0.010 0.050 0.440
0.050 0.406 0.758 0.006 0.036 0.485 0.074 0.343 0.658 0.010 0.078 0.330
0.079 0.331 0.733 0.007 0.053 0.398 0.086 0.311 0.629 0.011 0.110 0.252
0.111 0.252 0.712 0.008 0.075 0.331 0.114 0.247 0.596 0.012 0.140 0.202
0.141 0.202 0.662 0.010 0.113 0.239 0.148 0.191
0.168 0.167 0.640 0.010 0.140 0.200 0.179 0.159
0.207 0.136 0.221 0.126
0.241 0.113 0.248 0.109
0.310 0.080 0.309 0.080
0.374 0.060 0.443 0.037
0.477 0.034 0.489 0.026
0.582 0.018 0.599 0.014
0.691 0.010 0.676 0.012
100u(xw) = 2.4 100u(rw) = 2.5

The tie lines were then determined by applying the lever rule, To provide a fit for the ternary LLE, the thermodynamic re-
the Hmits of the ImmiscIbHity region being known. The accuracy lationship
of the method was f0.002 mass fraction. The observed re-
producibility of the reported L E data has been estimated to be (T/X/))=l = (7, X,)j=* i = 1, 2, 3 (1)
within f0.002 mass fraction. was applied. The activity coefficients y, in the liquid aqueous
0' = 1) and organic 0' = 2)phases were calculated by means
Results and Discussion of the NRTL equation (6)

The experimentaldata are reported In Tables 11-VI. Type

I isotherms were observed in all cases, acetonitrile being
completely miscible with water and with all the aromatic hy-
drocarbons. The immiscibility region decreases by increasing
the temperature from 25 to 45 OC. 11
The comparison with llterature data Is restricted to the sys- (2)
tems containing benzene and toluene and, only qualitatively, to
those containing m-xylene and p-xylene. Our data compare where 1) = (g, - g,)/RT, GI = exp(-a)r,), gq = g], and aq=
reasonably with the data measured by Hartwlg et al. (2) at 25 a,. The parameters gu- gu and gq - g, were assumed to be
O C for the system acetonitrile-water-benzene and by Rao et linearly dependent on the temperature T , according to the
al. (4) at 30 OC for the system acetontblle-water-toluene and following relationships (7):
with the graphs reported by Francis (3)for both of these sys- gl - gy = al 4- bl ( ( T / K ) - 273.15) (3)
tems at 25 OC. With regard to the systems Containing xylenes,
our binodal curves are fairly close to that reported by Rao et
al. (5) for the system acetonitrile-water-mixed xylenes at 31
g4 - g, = a, + b, ((f/K) - 273.15) (4)
OC,while some discrepancies appear in the slopes of the tie The NRTL model presents the advantage that predlctlons of
Ilnes. ternary phase equilibria can be carried out on the basis of
300 Journal of Chemlcal and Engineering De&, Vol. 36,No. 3, 199 1

Table V. Ternary Liquid-Liquid Equilibrium Data (xwi, Table VI. Ternary Liquid-Liquid Equilibrium Data (xwi,
Mass Fraction) and Percentage Standard Deviations Mass Fraction) and Percentage Standard Deviations (100
(lOOu(xw),Equation 6) for the Acetonitrile (1)-Water a(x,), Equation 6) for the Acetonitrile (1)-Water
( 2 h - X y l e n e (3) System (2)-Ethylbenzene (3) System
tie lines tie lines
binodal curve aqueous phase organic phase binodal curve aqueous phase organic phase
xu2 xu3 xu2 xw3 XW2 xw3
At 25 O C At 25 O C
0.002 0.906 0.852 0.002 0.004 0.862 0.003 0.714 0.923 0.001 0.0005 0.920
0.006 0.778 0.819 0.003 0.006 0.772 0.006 0.629 0.897 0.002 0.001 0.906
0.011 0.687 0.796 0.004 0.008 0.704 0.012 0.572 0.877 0.002 0.001 0.868
0.018 0.587 0.751 0.005 0.023 0.553 0.023 0.474 0.833 0.003 0.001 0.802
0.031 0.472 0.735 0.005 0.031 0.465 0.027 0.461 0.798 0.003 0.002 0.743
0.045 0.382 0.725 0.005 0.040 0.410 0.041 0.377 0.771 0.004 0.004 0.667
0.071 0.281 0.710 0.005 0.058 0.319 0.055 0.319 0.742 0.004 0.011 0.559
0.091 0.227 0.703 0.006 0.066 0.289 0.063 0.292 0.723 0.005 0.029 0.433
0.114 0.180 0.681 0.007 0.090 0.222 0.107 0.203 0.708 0.006 0.040 0.376
0.139 0.144 0.664 0.008 0.100 0.210 0.119 0.168 0.693 0.006 0.069 0.278
0.165 0.117 0.642 0.009 0.110 0.190 0.141 0.139 0.643 0.006 0.136 0.154
0.206 0.088 0.147 0.138
0.233 0.074 0.170 0.122
0.261 0.062 0.196 0.095
0.316 0.046 0.227 0.072
0.399 0.031 0.253 0.068
0.483 0.017 0.271 0.057
0.586 0.009 0.305 0.048
0.696 0.005
At 45 "C
At 45 O C 0.004 0.906 0.918 0.002 0.006 0.855
0.002 0.903 0.948 0.002 0.003 0.965 0.008 0.762 0.830 0.003 0.008 0.751
0.009 0.793 0.909 0.003 0.003 0.927 0.018 0.612 0.776 0.004 0.015 0.652
0.013 0.683 0.850 0.004 0.006 0.843 0.054 0.419 0.714 0.006 0.053 0.409
0.020 0.595 0.806 0.005 0.011 0.741 0.087 0.304 0.683 0.009 0.068 0.353
0.035 0.503 0.753 0.008 0.019 0.625 0.089 0.300 0.639 0.010 0.093 0.282
0.051 0.421 0.713 0.008 0.034 0.508 0.115 0.244 0.595 0.013 0.156 0.184
0.078 0.330 0.688 0.008 0.047 0.433 0.122 0.228 0.520 0.022 0.217 0.125
0.109 0.254 0.645 0.008 0.076 0.335 0.142 0.202
0.134 0.212 0.610 0.010 0.106 0.262 0.160 0.179
0.163 0.174 0.600 0.010 0.134 0.213 0.170 0.167
0.206 0.136 0.171 0.166
0.258 0.105 0.203 0.133
0.332 0.070 0.249 0.107
0.499 0.023 0.266 0.095
0.599 0.011 0.290 0.089
0.695 0.009 0.319 0.078
100u(xw)= 1.6 0.334 0.072
0.400 0.051
0.471 0.032
parameters determined from the constituting binary systems. 0.489 0.028
For each binary system a proper value of aywas assumed (6) 0.505 0.026
(Table VIII). The values of the parameters a, a,, b,, and b] 0.563 0.017
were determined as follows. 0.613 0.011
For the systems water-aromatic hydrocarbons, they were 100o(xu) = 3.2
directly obtained by applying eqs 1 and 2 to recommended LLE
data (8) at two different temperatures; for the system water- Table VII. Weighting Factors, FR FT,F ,,Used in Deriving
p-xylene, the data were found only at one, temperature and, the Binary NRTL Parameters from VLd Data. Eouation 5
consequently, the parameters b, and b, were assumed equal component i-j
to zero.
acetonitrile (1)-benzene (3) 1 1 15
For the systems acetonitrile-aromatic hydrocarbons and acetonitrile (1)-toluene (3) 1 1 30
water-acetonitrlle, the values of the parameters were obtained acetonitrile (1)-m-xylene (3) 1 1 60
by submitting to regression vapor-liquid equilibrium (VLE) data acetonitrile (1)-p-xylene (3) 1 1 60
of each examined binary system: in particular, two sets of acetonitrile (1)-ethylbenzene (3) 1 1 60
isobaric VLE data at 28.0 and 101.3 kPa ( 9 ) ,for the system acetonitrile (1)-water (2) 0.75 1 30
acetonitrile-aromatic hydrocarbons, and two sets of VLE data,
one isobaric at 101.3 kPa ( 7 0 )and the other isothermal at 60 The values of the NRTL parameters for each binary system
OC ( 7 I), for the system acetronttrile-water. The calculation are summarized in Table V I I I . The ternary LLE were then
p r o d w e , described in detail elsewhere (S),was based on the predicted, and the results were converted into mass fractions
minimization of the objective function, S: in order to allow a comparison with the experimental data.
N
Tables 11-VI also report the standard devations a(x,) of the
s= fitting of the ternary equilibria, calculated as
m=l
II~y,CY"xp - Y1m,calc)12 +
M 3 2
4xw) = [ cc -
n x b v ~ , e w Xw*,calc)2/6M10.5
k = l / = 7 /=1
(6)
The weighting factors Fp, FT, and FYI used for each binary
system are given in Table V I I . where M is the total number of tie lines of both isotherms. The
 Journal of Chemical and Engineerlng Data, Vol. 36,No. 3, 199 1 301

Table VIII. NRTL Parameters uij, aij,aji, bij,and b,i, Equations 2-4, f o r the Binary Systems
component i-j aij aij/(cal mol-') aji/(cal mol-') bij/(cal mol-l K-l) bji/(cal mol-' K-l)
water (2)-benzene (3) 0.2 2692.74 3307.26 -11.501 10.607
water (2)-toluene (3) 0.2 2631.38 3783.84 -0.874 22.374
water (2)-m-xylene (3) 0.2 2895.39 4856.16 -9.432 9.604
water (2)-p-xylene (3) 0.2 2556.90 5149.00 0.0 0.0
water (2)-ethylbenzene (3) 0.2 3025.01 4667.66 -14.187 18.360
acetonitrile (1)-benzene (3) 0.3 143.62 548.74 3.290 -3.305
acetonitrile (1)-toluene (3) 0.3 378.41 36.83 9.491 -2.421
acetonitrile (1)-m-xylene (3) 0.3 -566.06 1140.48 16.615 -1 1.468
acetonitrile (1)-p-xylene (3) 0.3 -208.13 798.70 14.872 -8.981
acetonitrile (1)-ethylbenzene (3) 0.3 363.69 632.87 3.408 -4.182
acetonitrile (1)-water (2) 0.3 604.06 797.80 -2.421 6.920

a b

+ +
Flguro 1. Ternary liquid-liquid equilibria for the acetonitrile (1) water (2) benzene (3) system at 25 and 45 OC, x,, mass fractions: (a) A,
experimental blnodal curve, and A, experimental tie Ilne; (b) calculated binodal curve and tie lines.

prediction of binodal curves and of tie lines appears to be The distribution coefficients of acetonitrile between solvent
satisfactory, as shown in Figure 1 for the system acetonitrile- and aqueous layers are shown in Figure 2. The order of
water-benzene. Similar diagrams are obtained for the other solvents is benzene > toluene > m-xylene = p-xylene =
four systems. ethylbenzene.
The binodal curve data and the tie-line data indicate that ail The selectivities are reported in Figure 3, as mass fractions
the aromatic hydrocarbons investigated can be used in the of acetonitrile on a solvent-free basis. Ethylbenzene and the
liquid-liquid extraction of acetonitrile. The immiscibility region xylenes appear to be more selective than toluene and benzene.
slightly increases for the solvents studied in the order benzene By increasing the temperature from 25 to 45 O C , both se-
> toluene > m-xylene = p-xylene = ethylbenzene. lectivity and distribution coefficient values slightly decrease, and
302 Jownel of Chemical and EnglneerJng Data, Vot. 36,No. 3, 1991

0.7 - 0.7

0.6 - A
0.6 LA
a
: ?
0.5 - * A ; 0.5
a
r'
fl ' 8
L
m 8
c 8 c M
a t a
U
0.4 - e
L 0
0.4
*d 8 ad
c 8 c L
a ea A ld
M 0.3 - m A
M 3.3
k
b
0 a 0
rt
c A c
*d 0.2 - A
,+ 0.2
L
d r. m
3 a 3 A*

x 0.1 - A
A A
b
a
x 0.1
8 .
A
b
A
A
e

0 1 I
L

I I 1
L. 0
A

0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4

xw l in aqueous phase xwl 1n aqueous phase

+ +
Flgura 2. Experlmenteldlstrlbutbn coeffldents of acetontMie (1) water (2) solvent (3) system at (a) 25 O C and (b) 45 O C (xwl, mass fractkns).
Solvent: 0 , benzene; W, toluene; L , m-xylene; L, p-xylene; A, ethylbenzene.

a
0
*d

a
M
I.
0
c
0.9-

0.8-
t
A A A A A ,

0 A
a

I
0.9

0.01
0
b

ad

X
ri
3
0.7
0 0.1 0.2 0.3 0.4
0.7 I
0 0.1 0.2 0.3 0.4
xw 1 in aqueous phase Xwl in aqueous phase

Flgura 3. Experimental selectivity data for acetonltrlle (1) + water (2) + solvent (3)system at (a) 25 O C and (b) 45 O C (X,,, mass fractions on
a solvent-free basis). Solvent: 0,benzene; 0 , toluene; h , m-xylene; 4 , p-xylene; A, ethylbenzene.

the differences between the solvents become less evident. 4 parameter of the NRTL equation, eq 2
b, parameter, eq 3, cai mol-' K-'
Oloreary
parameter, eq 3, cal moi-'
4 parameter, eq 4, cal mol-' K-'
a1 M total number of tie lines referred to both isotherms
4 parameter, eq 4, cal mol-' of each ternary system LLE data
FP weighting factor for pressure deviations, eq 5 N total number of vapor-liquid equllibrium data of each
FT weighting factor for temperature deviations, eq 5 binary system VLE data
FYI
weighting factor for vapor-phase composition devi- P pressure, kPa
ations of component i,eq 5 R gas constant, cai mol-' K-'
@ energy of interactlon between an I-j pair of mole- S objective function, eq 5
tules, cal mol-' T temperature, K
 J. Chem. Eng. Data 1991, 36,303-307 303

liquid-phase mole fraction of component i Literature Cited

liquid-phase mass fraction of component i
Toiiefson. E. L.; Decker, R. M.; Johnson, C. B. Can. J. Chem. €ng.
liquid-phase mass fraction of component i on a 1970, 48, 219.
soivent-free basis Hartwig, G. M.; Hood, G. C.; Maycock, R. L. J. Fhys. Chem. 1955,
59, 52.
vapor-phase mole fraction of component i Francis, A. W. J . Chem. Eng. Lkta 1965, 70, 145.
parameter of the NRTL equatlon, eq 2 Rao, D. S.;Rao, K. V.; Raviprasad, A,; Chiranjivi, C. J. Chem. Eng.
actlvity coefficient of component i Data 1979. 2 4 , 241.
Rao, C. V. S.R.; Rao, K. V.; Raviprasad, A.; Chiranjivi, C. J. Chem.
parameter of the NRTL equation, eq 2 f n g . Data 1978, 23, 23.
) standard deviation, mass fraction Renon, H.; Prausnitz, J. M. AIChf J . 1968, 14, 135.
Renon, H.; Asseiineau, L.; Cohen, G.; Raimbault, C. Calculsur ordfnet-
Subscripts eur des equilibres liquide-vapeur et liquide-liqukle; Technip: Paris,
1971: D 16.
calc calculated (8) Sorensen. J. M.; Arti, W. Liquid-Liquid Equilibrium Data Collection:
Binary Systems; DECHEMA: Frankfurt Main, FRG, 1979; Vol. 5. Part
eXP experimental I , p 341.
iJ component indexes (1, acetonitrile; 2, water; 3, (9) Di Cave, S.;Mazzarotta, 6. J. Chem. f n g . Data, preceding paper in
aromatic hydrocarbon) this issue.
(10) Volpicelii, G. Chim. Ind. (Milan) 1967, 49, 720.
1 phase index (1, aqueous: 2, organic) (11) Sugi, H.: Katayama, T. J. Chem. Eng. Jpn. 1978, 7 7 , 167.
k tie-line index
m running variable, eq 5 Received for review May 21, 1990. Revised December 5, 1990. Accepted
PJ running variables, eq 2 March 11, 1991.

Vapor-Liquid Equilibria in the Carbon Dioxide + Ethanol and

Carbon Dioxide l-Butanol Systems+
Davld W. Jennlngs, RongJwyn Lee, and Amyn S. Teja’
School of Chemical Engineering, Georgia Institute of Technology, Atlanta, Georgia 30332-0 700

Vapor-liquid equilibrium data for CO, +

ethanol mlxtures
+
systems. The COP ethanol system was chosen as a test
at 314.5, 325.2, and 337.2 K and for CO, +
1-butanol
system to check the validity of the measurements from the new
apparatus, since a number of Investigators have reported data
mlxtures at 314.8, 325.3, and 337.2 K have been
for this system at the condfins of interest. The onty previously
measured by using a hlgh pressure flow apparatus. The
+
reported data for the CO, I-butanol system are those of King
pressure In the experiments ranged from 4.633 to 11.776
MPa. Our results for the CO, +
ethanol system are in
et ai. (73)at 313.15 and 383.15 K. The C02 +
1-butanol
system was studied in the present work because of a need for
good agreement with the recent results of Surukl el ai.
data at temperatures other than those previously reported.
and wlth the results of Panaglotopoulos. However, the
results for the CO, +
l-butanol system are slgnlflcantiy
Experlmental Sectlon
dlfferent from those reported by King et ai.
Experlmental Apparatus. I n the flow apparatus, two or
I ntroductlon more components are mixed thoroughly and continuously at
constant temperature until vapor-liquid equilibrium is achieved.
+
C02 alcohol systems are of interest because of their im- The vapor and liquid phases are then separated and analyzed
portance as supercriticalfluid/cosolvent pairs in the separation to obtain the concentrations of the phases in equilibrium. A
of biomaterlals ( 7 - 4 ) . A knowledge of vapor-liquid equilibrium schematlc diagram of the apparatus used in this work Is shown
in these systems is therefore necessary for evaluating models in Figure 1, and a more detailed description is given below. The
for the extraction of biomaterials with supercritical fluid/co- apparatus is similar to those described by Simnick et ai. (74),
solvent pairs. In spite of these interests, however, only the CO, Thies and Paulaitis (75), Inomata et ai. (76), and Radosz et ai.
+ +
methanol (5-8) and COP ethanol (8-7 7 ) systems have ( 77).
+
been studied extensively. Surprisingly, the C02 I-propanol Liquid carbon dioxide was obtained from a cylinder equipped
+ +
(8, 7 7), COP 2-propanol ( 7 7- 72), and COP 1-butanol ( 73) with an eductor tube and was compressed to the desired
systems have received very little attention, and COP higher+ pressure by using a Milton Roy metering pump (Model 396-89).
alcohol systems have not been studied at all. The inlet tube and the pump head were covered with an ice
We have constructed a flow apparatus to measure high- bath to ensure that the carbon dioxide remained liquid during
pressure vapor-liquid equilibria in systems containing COP + the compression process. A regulator was used instead of the
alcohol and biomaterials in liquid solutions. A flow apparatus pump when the desired pressure was lower than the carbon
was chosen because large sample sizes could be generated dioxide cylinder pressure. Liquid ethanol and I-butanol were
for analysis, and this can be particularly valuable when dealing compressed to the desired pressure by wlng another MlAon Roy
with components present in dilute concentrations. We are metering pump (Model 396-57).
+ +
currently evaluating the CO, ethanol, COP propanol, and The C02 and alcohol were pumped through a section of
COP + l-butanol systems for the supercritical extraction of 0.125-in.-o.d. X 0.069-in.-l.d. stainless steel tubing, tublng colls,
pyrrolizidine alkaloids. We have therefore measured vapor- and two 0.1875-in.-o.d. X 0.132-In.4.d. X 7-In. static mixers
liquid equilibria in the C02 + ethanol and COP + I-butanol placed inside a constant-temperature air bath. Vapor-liquid

002 1-956019 1I1 736-0303$02.50/0 0 1991 American Chemical Society