1176 J. Chem. Eng.

Data 1996, 41, 1176-1180

Isobaric Vapor-Liquid Equilibria of the Water + 1-Propanol System

at 30, 60, and 100 kPa

Carmen Gabaldón, Paula Marzal, Juan B. Montón,* and Manuel A. Rodrigo

Departamento de Ingenierı́a Quı́mica, Facultad de Quı́mica, Universitat de València, 46100 Burjassot, Valencia,
Spain

Isobaric vapor-liquid equilibria for the water + 1-propanol system are reported at 30, 60, and 100 kPa.
The results were found to be thermodynamically consistent according to Van Ness-Byer-Gibbs, Kojima,
and Wisniak methods. The system shows a minimum boiling azeotrope, and the azeotropic composition
is scarcely shifted with pressure. Results were compared with literature values. The data were correlated
with Margules, Van Laar, Wilson, NRTL, and UNIQUAC liquid-phase activity coefficient models.

Introduction
Distillation is the most common operation for the sepa-
ration of liquid binary and multicomponent mixtures. The
correct design of distillation columns requires the avail-
ability of accurate and, if possible, thermodynamically
consistent vapor-liquid equilibria (VLE) data and the use
of generalized methods to predict the properties of the
mixtures.
The present work is part of a project studying the effect
of pressure on the behavior of the azeotropic point in
mixtures in which at least one component is an alcohol.
For this purpose, the azeotropic system water + 1-propanol
was selected. For this system, isobaric and isothermal VLE
data sets have been found in the literature, the majority
of these included in the DECHEMA Chemistry Data Series
(Gmehling and Onken, 1977, 1981) VLE compilation, and
other more recently presented sources (Munday et al., 1980;
Morrison et al., 1990; Zielkiewicz and Konitz, 1991).
Among these, seven isobaric and six isothermal sets of data
have been reported in the DECHEMA as consistent ac-
cording to the point-to-point thermodynamic consistency Figure 1. Variation of the azeotropic composition for the system
test of Van Ness-Byer-Gibbs (1973), modified by Freden- water (1) + 1-propanol (2): (O) this work; (b) literature data
slund (1977). However, a detailed study on these data (points numbered as indicated in Table 6).
reveals differences with respect to the azeotropic values. Table 1. Physical Properties of Chemicals: Densities d,
The azeotropic points are plotted in Figure 1, showing how Refractive Indexes n, and Boiling Points Tb
azeotropic compositions tend to be scattered. The object
d(293.15 K)/g cm-3 n(293.15 K) Tb (100.00 kPa)/K
of our study is the accurate measurement of VLE and the
determination of azeotropic data for the isobaric system compd exptl lit. exptl lit. exptl lit.
at three pressures (30, 60, and 100 kPa) not previously water 0.998 06 0.998 20a 1.3334 1.3330a 372.78 372.77a
reported. The contribution presented in this paper includes 1-propanol 0.803 66 0.803 75a 1.3853 1.3855a 369.75 369.95a
VLE data that fulfill the thermodynamic consistency tests a TRC Thermodynamic Tables Hydrocarbons, 1994.
(Van Ness-Byer-Gibbs, 1973; Kojima et al. 1990; Wisniak,
1993).
using the apparatus described below. The experimental
Experimental Section values of these properties are listed in Table 1 together
with the literature values (TRC, 1994).
Chemicals. Milli-Q water and 1-propanol high-purity Apparatus and Procedure. The equilibrium vessel
grade purchased from Aldrich Chemical were used without used in this work is an all-glass, dynamic recirculating still
further purification. The purities of all chemicals, checked described by Walas (1985), equipped with a Cottrell pump.
by gas chromatography (GC), were as follows: water, The still (Labodest model), manufactured by Fischer Labor
100.00 mass %; 1-propanol, 99.80 mass %. Experimental und Verfahrenstechnik (Bonn, Germany) is capable of
densities of the pure liquids at 293.15 K were measured handling pressures from 0.25 to 400 kPa and temperatures
in a digital precision densimeter, Anton Paar DMA55, with up to 523.15 K. In the boiler, vapor is generated by
a thermostated bath controlled to 0.01 K. The refractive external heating. The Cottrell pump ensures the intimate
indexes at 293.15 K were measured using an Abbe refrac- contact between the liquid and vapor phases and also with
tometer (ATAGO 3T). The accuracies in density and the temperature sensing element. The equilibrium tem-
refractive index measurements were 0.000 01 g cm-3 and perature is measured with a digital Ditel thermometer with
(0.0002, respectively. The boiling points were determined an accuracy of (0.01 K. For the pressure measurement,

S0021-9568(96)00146-X CCC: $12.00 © 1996 American Chemical Society

 Journal of Chemical and Engineering Data, Vol. 41, No. 5, 1996 1177
Table 2. Antoine Coefficients A, B, and C Table 4. Vapor-Liquid Equilibrium Data, Liquid-Phase
Mole Fraction x1, Vapor-Phase Mole Fraction y1,
Antoine coefficients
Temperature T, and Activity Coefficients γi for the Water
component temp range/K A B C (1) + 1-Propanol (2) System at 60 kPa
water 274-373 16.5700 3984.92 -39.724 x1 y1 T/K γ1 γ2
1-propanol 303-370 16.0353 3415.56
0.000 0.000 356.78
-70.733

Table 3. Vapor-Liquid Equilibrium Data, Liquid-Phase 0.009 0.031 356.35 3.845 0.995
Mole Fraction x1, Vapor-Phase Mole Fraction y1, 0.045 0.123 354.53 3.318 1.009
Temperature T, and Activity Coefficients γi for the Water 0.103 0.245 352.44 3.107 1.011
(1) + 1-Propanol (2) System at 30 kPa 0.168 0.340 350.70 2.839 1.028
0.231 0.408 349.51 2.603 1.051
x1 y1 T/K γ1 γ2 0.294 0.461 348.63 2.397 1.083
0.361 0.501 347.63 2.212 1.158
0.000 0.000 341.08
0.417 0.527 347.74 2.005 1.197
0.008 0.027 340.49 3.658 1.008
0.487 0.551 347.50 1.814 1.306
0.066 0.184 337.88 3.393 1.016
0.557 0.571 347.52 1.642 1.443
0.137 0.315 335.92 3.057 1.015
0.621 0.585 347.53 1.508 1.631
0.210 0.405 334.33 2.757 1.041
0.686 0.594 347.55 1.385 1.925
0.304 0.481 332.82 2.425 1.110
0.747 0.602 347.73 1.279 2.323
0.377 0.514 332.26 2.145 1.194
0.796 0.605 347.80 1.203 2.850
0.434 0.545 332.06 1.994 1.243
0.833 0.609 347.91 1.152 3.430
0.510 0.568 331.93 1.780 1.371
0.869 0.612 348.14 1.099 4.294
0.578 0.583 331.93 1.613 1.537
0.892 0.620 348.28 1.078 5.070
0.659 0.597 331.84 1.453 1.848
0.925 0.630 348.51 1.046 7.036
0.725 0.606 331.94 1.335 2.229
0.946 0.646 348.99 1.028 9.154
0.779 0.612 331.99 1.252 2.725
0.960 0.672 349.92 1.014 10.992
0.823 0.615 332.08 1.186 3.360
0.970 0.704 350.93 1.009 12.655
0.857 0.619 332.17 1.141 4.098
0.981 0.760 352.78 0.998 14.956
0.889 0.622 332.27 1.100 5.211
0.988 0.806 354.43 0.983 17.840
0.914 0.629 332.51 1.070 6.523
0.993 0.872 356.06 0.992 18.838
0.936 0.638 332.83 1.044 8.418
0.996 0.923 357.30 0.996 18.830
0.954 0.661 333.43 1.033 10.646
0.999 0.974 358.02 1.019 24.684
0.964 0.691 334.27 1.028 11.896
1.000 1.000 359.14
0.977 0.739 335.90 1.007 14.523
0.983 0.796 337.06 1.023 14.519
0.989 0.851 338.46 1.020 15.326
(1995) were used. Antoine constants for both compounds
0.994 0.891 339.50 1.015 19.565 are summarized in Table 2.
0.997 0.935 340.52 1.015 22.240 The experimental VLE data for the binary system at (30,
1.000 1.000 342.33 60, and 100) kPa are given in Tables 3-5. The T-x-y
diagrams at the three pressures are shown in Figure 2. It
a digital manometer with an accuracy of (0.01 kPa is used. can be observed that the system presents a minimum
The temperature probe was calibrated against the ice and boiling azeotrope and the azeotropic composition is scarcely
steam points of distilled water. High-purity (>99.9 mass shifted with pressure.
%) hexane vapor pressures were used for the manometer The azeotropic composition, pressure, and temperature
calibration. are listed in Table 6 along with thermodynamically con-
In each VLE experiment, the pressure was fixed and sistent literature data from the DECHEMA Chemistry
remained constant using a vacuum pump, and the heating Data Series (Gmehling and Onken, 1977, 1981). In all
and shaking systems of the liquid mixture were turned on. cases, the azeotropic point has been calculated from the
The system was kept at the boiling point at least for 30 experimental data by fitting the function z ) x1/y1 to a
min to ensure that the steady state was reached. At this polynomial function of x1 and solving for x1 at z ) 1.
moment, 0.2 cm3 samples of liquid and condensed vapor of Variation of azeotropic pressure with composition is shown
the Cottrell pump were taken for analysis. in Figure 1. Azeotropic pressure is shifted to lower values
Analysis. All the samples were analyzed by using a as the water mole fraction increases. The values obtained
Varian Star 3400 CX gas chromatograph with a thermal in this work are in agreement with the tendency shown
conductivity detector. The GC response was treated with by the majority of the consistent data previously reported,
a Star Chromatography Station. The chromatographic contributing to reduce the moderate divergence among the
column (2 m × 1/8 in.) was packed with Porapak P. The different data.
carrier gas was helium flowing at 50 cm3 min-1, and the Comparison of VLE data can be also performed by
column temperature was 383.15 K. The GC was calibrated representing the value of GE/RT at constant composition
with gravimetrically prepared standard solutions. The vs 1/T. Such a plot for 0.5 mole fraction is presented in
uncertainty of composition measurements was estimated Figure 3. The temperature dependence of GE may be
to be (0.001. At least two analyses were made for each established by means of HE data according to the Gibss-
liquid and each condensed vapor sample. Helmholtz equation
Results and Discussion
Vapor pressures P0i were calculated with the Antoine
equation
( ∂(GE/T)
∂(1/T)) P,n
) HE (2)

HE being the molar excess enthalpy. The HE variation with

B temperature at equimolar composition has been adjusted
ln(P0i /kPa) ) A - (1)
(C + T/K) from literature HE data (Christensen et al., 1984). In
Figure 3, two solid lines showing this tendency have been
The Antoine constants for water were those given in plotted, delimiting the experimental data.
DECHEMA Chemistry Data Series (Gmehling and Onken, To calculate the liquid-phase activity coefficients, γi, the
1977). For 1-propanol, the values given in Aucejo et al. Poynting factor was considered as unity at the experimen-
1178 Journal of Chemical and Engineering Data, Vol. 41, No. 5, 1996

Table 6. Variation of the Azeotropic Composition and

Temperature with Pressure for the System Water (1) +
1-Propanol (2)
P/kPa x1 T/K ref
1 6.4 0.604 303.15 Udovenko and Mazanko, 1972a
2 18.7 0.591 323.07 Vrevskii, 1910a
3 26.7 0.583 329.83 Smirnova, 1959a
4 30.0 0.585 331.90 this work
5 31.2 0.574 333.15 Schreiber et al., 1971a
6 40.1 0.583 339.09 Vrevskii, 1910a
7 53.3 0.576 345.07 Smirnova, 1959a
8 60.0 0.575 347.52 this work
9 66.7 0.550 350.66 Goelles and Still, 1979b
10 73.3 0.571 352.95 Vrevskii, 1910a
11 80.0 0.567 354.83 Smirnova, 1959a
12 100.0 0.569 360.29 this work
13 101.1 0.579 360.57 Dawe et al., 1973a
14 101.3 0.567 360.81 Smirnova, 1959a
15 101.3 0.565 360.77 Kojima et al., 1968b
16 110.1 0.567 363.15 Ratcliff and Chao, 1969a
a Data from the DECHEMA Chemistry Data Series (Gmehling

and Onken, 1977). b Data from the DECHEMA Chemistry Data

Series (Gmehling and Onken, 1981).

Figure 2. Temperature T vs composition (x1, y1) for the system

water (1) + 1-propanol (2) at constant pressure, P ) 30, 60, and Figure 3. Comparison of GE(x ) 0.5)/RT at various tempera-
100 kPa: (O) experimental; (s) NRTL (R12 adjustable) model. tures: (O) this work; (b) literature data (points numbered as
indicated in Table 6).
Table 5. Vapor-Liquid Equilibrium Data, Liquid Phase
Mole Fraction x1, Vapor Phase Mole Fraction y1, and Curl equations (1957) with the correction proposed by
Temperature T, and Activity Coefficients γi for the Water Tsonopoulos (1974). For every experimental pressure-
(1) + 1-Propanol (2) System at 100 kPa
temperature-composition condition, the calculated fugacity
x1 y1 T/K γ1 γ2 coefficients were close to unity (0.9998 maximum, 0.9666
0.000 0.000 369.75 minimum) and were not considered in the analysis of the
0.013 0.025 369.40 2.174 0.994 VLE data. So, the experimental liquid-phase activity
0.038 0.097 367.84 3.056 1.003 coefficients γi were calculated from
0.083 0.194 366.07 2.989 1.006
0.160 0.317 363.80 2.760 1.018 Pyi
0.226 0.390 362.46 2.530 1.041 γi ) (3)
0.283 0.437 361.63 2.337 1.072 xiP0i
0.356 0.487 360.82 2.136 1.124
0.426 0.522 360.47 1.939 1.192
0.500 0.550 360.33 1.750 1.295 The thermodynamic consistency of the VLE experimental
0.583 0.572 360.28 1.564 1.480 data was checked by the following methods: the point-to-
0.648 0.585 360.27 1.440 1.701 point test of Van Ness-Byer-Gibbs (1973), the infinite
0.708 0.595 360.32 1.338 1.997 dilution test proposed by Kojima et al. (1990), modified by
0.763 0.603 360.44 1.252 2.400
Jackson and Wilsak (1995), and the L-W method of
0.817 0.608 360.62 1.171 3.047
0.853 0.613 360.78 1.124 3.720 Wisniak (1993).
0.883 0.619 360.93 1.090 4.574 For the point-to-point test of Van Ness-Byer-Gibbs a
0.911 0.627 361.12 1.062 5.841 four-parameter Legendre polynomial was used for the
0.926 0.632 361.43 1.041 6.844 excess Gibbs free energy. The temperature dependence of
0.945 0.643 361.77 1.024 8.812 the excess Gibbs free energy was allowed for by the term
0.957 0.663 362.41 1.018 10.369
∆H/RT2 dT from the Gibbs-Duhem equation. Heat of
0.967 0.683 363.30 1.003 12.265
0.978 0.739 365.06 1.004 14.126 mixing data were taken from the literature (Christensen
0.985 0.794 366.87 1.000 15.232 et al., 1984). The selected objective function to minimize
0.991 0.854 368.71 0.999 16.756 was the sum of the squared relative deviations in the total
0.996 0.905 370.23 0.996 23.145 pressure. The consistency criteria in this test are that the
0.999 0.984 371.95 1.015 24.348 mean absolute deviation between calculated and measured
1.000 1.000 372.78
mole fractions of component 1 in the vapor phase, MAD-
tal conditions. The fugacity coefficients were calculated on (y), is less than 0.01, and that deviations scatter randomly
the basis of the virial equation of state, with the second about zero. To check the absence of bias in the data, a
virial coefficient being estimated by means of the Pitzer graphical inspection of the error in y1 is performed, y1
 Journal of Chemical and Engineering Data, Vol. 41, No. 5, 1996 1179
Table 7. Thermodynamic Consistency Tests
Kojima
Van Ness-Byer-Gibbs error for dilute error for dilute Wisniak
P/kPa MAD(y) BIAS component 1 (%) component 2 (%) L W D
30 0.0098 -0.0054 20.4 21.4 8.23 8.66 2.58
60 0.0071 -0.0007 6.1 1.8 8.79 9.27 2.65
100 0.0070 0.0003 5.4 0.2 9.16 9.67 2.74

Table 8. Correlation Parameters for Activity Coefficients, Activity Coefficients at Infinite Dilution
γ∞i , and Mean Absolute Deviations MAD(y) and MAD(T)
P/kPa model A12 A21 R12 γ∞1 γ∞2 MAD(y) MAD(T)
30.00 Margules 0.9388a 2.6214a 2.56 13.75 0.0291 0.64
Van Laar 1.2133a 2.9060a 3.36 18.28 0.0109 0.33
Wilson 5451.42b 4019.48b 4.19 25.24 0.0107 0.37
NRTLc 8103.12b -5.61b 0.30a 3.35 17.19 0.0115 0.35
NRTLd 1048.05b 2190.39b -1.56a 4.16 18.91 0.0062 0.29
UNIQUAC 1867.76b 194.98b 3.12 19.00 0.0117 0.39
60.00 Margules 0.9000a 2.6011a 2.46 13.48 0.0297 0.74
Van Laar 1.1578a 2.9373a 3.18 18.86 0.0101 0.31
Wilson 5442.05b 4521.85b 3.85 26.47 0.0088 0.31
NRTLc 8738.57b -198.20b 0.30a 3.16 17.42 0.0120 0.35
NRTLd 977.32b 2313.63b -1.56a 3.76 19.29 0.0064 0.21
UNIQUAC 2127.32b 37.79b 3.00 19.11 0.0113 0.37
100.00 Margules 0.8164a 2.5933a 2.26 13.37 0.0299 0.86
Van Laar 1.1010a 2.9219a 3.01 18.58 0.0099 0.31
Wilson 5457.66b 4505.32b 3.60 23.93 0.0109 0.34
NRTLc 9218.80b -414.77b 0.30a 2.96 17.00 0.0117 0.34
NRTLd 874.58b 2399.76b -1.57a 3.39 18.78 0.0062 0.21
UNIQUAC 2132.80b 37.71b 2.95 17.98 0.0109 0.33
a Dimensionless. b J mol-1. c With recommended R12 value. d With adjustable R12.

Figure 4. Deviation between calculated and measured vapor-

phase mole fraction (y1) vs liquid-phase mole fraction (x1) for the
water (1) + 1-propanol (2) system at constant pressure, P ) 60
kPa. Figure 5. Infinite dilution test for the water (1) + 1-propanol (2)
binary system at constant pressure, P ) 60 kPa: (O) ln γ1; (]) ln
residuals being plotted vs x1. This plot for the 60 kPa data γ2; (0) GE/RTx1x2; (s) NRTL (R12 adjustable) model; (- - -) polyno-
mial regression; (+) excluded points.
is shown, as an example, in Figure 4. Similar plots were
obtained at 30 kPa and at 100 kPa. Experimental VLE The L-W method of Wisniak requires the evaluation of
data were found thermodynamically consistent according the integrals L and W as described in the paper by Wisniak
to this test, with the values of MAD(y) and sum of y1 (1993), and values of the deviation D defined as
residuals (BIAS) as listed in Table 7.
The application of the infinite dilution test of Kojima |L - W|
(1990) includes the calculation of the excess Gibbs free D ) 100 (4)
L+W
energy from the experimental data, and the extrapolation
to infinite dilution using smoothing functions. The ex- less than 3-5 indicate thermodynamic consistency. As can
trapolated values are then compared with those extrapo- be observed in Table 7, experimental VLE data were also
lated of ln γ1 and ln γ2 at infinite dilution, and thermody- found consistent according to the Wisniak test.
namic consistency is achieved if the values agree to within The activity coefficients were correlated with the Mar-
30%. The results at 60 kPa are shown in Figure 5 where gules, Van Laar, Wilson, NRTL, and UNIQUAC equations
GE/RTx1x2, ln γ1, and ln γ2 have been plotted vs x1. (Gmehling and Onken, 1977). To fit the binary param-
Deviations of the extrapolated values at the three pressures eters, a nonlinear optimization method was used to mini-
(Table 7) were within the method tolerance. mize the function
1180 Journal of Chemical and Engineering Data, Vol. 41, No. 5, 1996

( )
γexptl - γcalcd 2 Liquid Equilibrium Data: Data Evaluation and Model Testing.
F) ∑∑
n i γexptl
(5) Fluid Phase Equilib. 1995, 103, 155-197.
Kojima, K.; Moon, H. M.; Ochi, K. Thermodynamic Consistency Test
i,n of Vapor-Liquid Equilibrium data. Fluid Phase Equilib. 1990, 56,
269-284.
where n is the number of data points. For the NRTL Morrison, J. F.; Baker, J. C.; Meredith, H. C., III; Newman, K. E.;
Walter, T. D.; Massie, J. D.; Perry, R. L.; Cummings, P. T.
model, two alternatives were selected: using the recom- Experimental Measurements of Vapor-Liquid Equilibrium in Alcohol/
mended value for R12 of 0.30 (Renon and Prausnitz, 1968) Water/Salt Systems. J. Chem. Eng. Data 1990, 35, 395-404.
or adjusting R12 in the fitting process. The parameters A12, Munday, E. B.; Mullins, J. C.; Edie, D. D. Vapor Pressure Data for
A21, and R12 for the correlation equations, mean absolute Toluene, 1-Pentanol, 1-Butanol, Water, and 1-Propanol and for the
Water and 1-Propanol System from 273.15 to 323.15 K. J. Chem.
deviations, and activity coefficients at infinite dilution Eng. Data 1980, 25, 191-194.
γ∞i are given in Table 8. The systems show positive Pitzer, K. S.; Curl, R. F. The Volumetric and Thermodynamic Proper-
deviations from ideality, as can be observed from the ties of Fluids III. Empirical Equation for the Second Virial Coef-
ficient. J. Am. Chem. Soc. 1957, 79, 2369-2370.
activity coefficients at infinite dilution. At the three Renon, H.; Prausnitz, J. M. Local Compositions in Thermodynamic
pressures, Van Laar, Wilson, NRTL with R12 ) 0.30, and Excess Functions for Liquid Mixtures. AIChE J. 1968, 14, 134-
UNIQUAC models yield similar deviations between ex- 144.
perimental and calculated vapor compositions and tem- TRC Thermodynamic Tables Hydrocarbons; Thermodynamic Research
Center, The Texas A&M University System: College Station, TX,
peratures, while Margules shows greater deviations in both 1994.
variables. Better results were obtained with the NRTL Tsonopoulos, C. An Empirical Correlation of the Second Virial Coef-
equation using adjustable R12. ficients. AIChE J. 1974, 20, 263-272.
Van Ness, H. C.; Byer, S. M.; Gibbs, R. E. Vapor-Liquid Equilibrium:
Literature Cited Part I. An Appraisal of Data Reduction Methods. AIChE J. 1973,
19, 238-244.
Aucejo, A.; González-Alfaro, V.; Montón J. B.; Vázquez, M. I. Isobaric Walas, S. M. Phase Equilibria in Chemical Engineering; Butter-
Vapor-Liquid Equilibria of Trichloroethylene with 1-Propanol and worth: London, 1985.
2-Propanol at 20 and 100 kPa. J. Chem. Eng. Data 1995, 40, 332- Wisniak, J. A New Test for the Thermodynamic Consistency of Vapor-
335. Liquid Equilibrium. Ind. Eng. Chem. Res. 1993, 32, 1531-1533.
Christensen, C.; Gmehling, J.; Rasmussen, P.; Weidlich, U. Heats of Zielkiewicz, J.; Konitz, A. (Vapor + liquid) equilibria of (N,N-dimeth-
Mixing Data Collection; Chemistry Data Series; DECHEMA: Frank- ylformamide + water + propan-1-ol) at the temperature 313.15 K.
furt/Main, 1984; Vol. III, Part 1. J. Chem. Thermodyn. 1991, 23, 59-65.
Fredenslund, Aa.; Gmehling, J.; Rasmussen, P. Vapor-Liquid Equi-
libria Using UNIFAC; Elsevier: Amsterdam, 1977.
Gmehling, J., Onken, U., Eds. Vapor-Liquid Equilibrium Data Col-
lection; Chemistry Data Series; DECHEMA: Frankfurt/Main, 1977; Received for review April 24, 1996. Accepted July 1, 1996.X
Vol. 1, Part 1. Financial support by Generalitat Valenciana (Grant GV-1006/93) is
Gmehling, J., Onken, U., Eds. Vapor-Liquid Equilibrium Data Col- gratefully acknowledged.
lection; Chemistry Data Series; DECHEMA: Frankfurt/Main, 1981; JE9601467
Vol. 1, Part 1a.
Jackson, P. L.; Wilsak, R. A. Thermodynamic Consistency Tests Based
on the Gibbs-Duhem Equation Applied to Isothermal, Binary Vapor- X Abstract published in Advance ACS Abstracts, August 1, 1996.