Editorial

pubs.acs.org/jced

Guidelines for the Analysis of Vapor−Liquid Equilibrium Data

ABSTRACT: It is current practice to represent and analyze vapor−liquid equilibrium data through the use of state conditions
(temperature and pressure) and phase compositions. However, these representations do not reveal and identify the accuracy of
the data for important purposes, such as separation by distillationand this is particularly a problem for the important low-
concentration regions. The intent of this editorial is to present the guidelines of the Journal of Chemical & Engineering Data that
vapor−liquid equilibrium data should be evaluated through the use of charts and related analyses that bring to light the low-
concentration regions, for example, K-values and relative volatilities.

Figures are essential to communicating the content and

See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

accuracy of measured and estimated data in the Journal of

Chemical & Engineering Data (JCED). Ambrose1 captured the
crux of the idea when he wrote: “The eye can see at a glance
what may not be easy to deduce from columns of numbers and
Downloaded via 179.33.111.91 on December 1, 2020 at 01:17:24 (UTC).

a few statistical parameters.”

Diky2 recently demonstrated that modifying the mole
fraction x1 to log10(x1/x2) = log10(x1/(1 − x1)) for liquid−
liquid and solid−liquid equilibria of binary mixtures enables
“stretching” of both ends of the mole fraction, and hence “is
probably the most efficient way to communicate such
information to a human.” It is sufficient to visualize and
analyze just the low mole fraction in liquid−liquid and solid−
liquid equilibria because the other phase usually has the
analogous concentration at a relatively high value, and, in fact,
the value of Diky’s proposed figure is that it highlights both low
mole fractions in a single chart. Vapor−liquid equilibrium
(VLE) data are inherently different because both the vapor and
liquid concentrations are simultaneously at low values, and it is
the ratio of the concentrations that is relevant to industrial
separation processes such as distillation, and especially
distillations that produce ultrapure products.
Mathias3 investigated the benzene + chloroform + acetone
system and showed the improved effectiveness of K-value and
relative-volatility charts over the customary Txy diagram to
quantitatively estimate the data accuracy, and to establish
uncertainty estimates for engineering design and analysis.
Representative ideas and results from this paper are reproduced
here to highlight the key views. Figure 1 compares calculations
from the NRTL-RK property option (nonrandom two liquid
(NRTL) activity-coefficient model and Redlich−Kwong
equation of state for the vapor phase) with the data of Kojima
et al.4 for the acetone + chloroform binary mixture at 101.3 kPa
using a Txy diagram. The figure indicates that the agreement
between model and data is good, and that both the model and
data clearly depict the maximum-boiling azeotrope, but charts
such as Figure 1 are unable to provide quantitative uncertainty
estimates, and are particularly weak in this respect at the two
pure ends. Figure 2 compares calculated relative volatilities for
the same acetone + chloroform binary at 101.3 kPa with data
from Kudryavtseva and Kusarev,5 Kogan and Deizenrot,6
Kojima et al.,4 and Segura et al.7 The data from all four
sources agree well (except for two outliers from Kogan and
Diezenrot, which are readily identified), and the chart clearly
demonstrates that the relative standard uncertainty of the
model for the relative volatility is better than 0.05; note that 51
out of 62 points or 82% are within the 5% corridor, and the
Figure 1. Txy diagram of the acetone−chloroform binary mixture at
101.3 kPa. Comparison of model calculations to the data of Kojima et
al.4 Figure reprinted with permission from ref 3. Copyright 2016
Elsevier.
standard uncertainty corresponds to 68%. Figure 1 is effective
in showing the existence of the azeotrope, but Figure 2 is far
superior since it also identifies the azeotrope (relative volatility
equal to unity), and, in addition, enables quantification of
model and data uncertainties “at a glance,” and is an effective
way to “communicate such information to a human.” Figure 3
shows a comparable way to visualize information analogous to
Figure 2 for ternary and higher mixtures; it presents percentage
differences between acetone K-values from model and data
plotted against the mole fraction of acetone; note that Mathias3
also presented analogous charts to Figure 3 for benzene and
chloroform as the chosen component. We expect the most
interesting and relevant comparisons at low concentrations.
Indeed, review of Figure 3 shows that the model K-values agree
with the data of Kojima et al.4 to within ±5%, but the
deviations from the data of Reinders and de Minjer8 are much
larger, suggesting greater uncertainty of the latter data set. The
relatively high uncertainty of the Reinders and de Minjer8 data
is likely a consequence of their indirect analytical approach in
which the compositions were deduced through measurements
of the mixture density and refractive index rather than by direct
composition measurements. Kojima et al.4 did not seem to
recognize the large uncertainty of the Reinders and de Minjer8
K-values since their paper simply stated: “The predicted vapor
compositions of our parameters agree with Reinders’ data in the
Published: August 10, 2017

© 2017 American Chemical Society 2231 DOI: 10.1021/acs.jced.7b00582

J. Chem. Eng. Data 2017, 62, 2231−2233
Journal of Chemical & Engineering Data
Figure 2. Relative volatilities of the acetone + chloroform system at
101.3 kPa. Comparison of model results to the data of Kudryavtseva
and Susarev,5 Kogan and Diezenrot,6 Kojima et al.,4 and Segura et al.7
The chart also shows (dashed lines) ± 5% deviations from the
calculated relative volatility. Figure reprinted with permission from ref
3. Copyright 2016 Elsevier.
Figure 3. Percentage errors of model K-values of acetone in
comparison to the ternary VLE data of Reinders and de Minjer8 and
Kojima et al.4 for the acetone + chloroform + benzene ternary mixture
at 101.3 kPa. Figure reprinted with permission from ref 3. Copyright
2016 Elsevier.
accuracy of 0.003 mole fraction on the average;” however, a
chart such as Figure 3 renders the conclusion obvious “at a
glance.”
The use of Txy and analogous charts is, unfortunately, the
norm in VLE data evaluation. Figure 4 presents another
example where the calculated ethanol K-values for the water +
methanol + ethanol ternary mixture at 101.3 kPa are compared
to two sets of experimental data, from Kurihara, Nakamichi,
and Kojima,9 and Resa et al.10 The model used is the NRTL-
RK property option in Aspen Plus V8.811 with out-of-the-box
parameters. Figure 4 is similar to Figure 3 in that the data of
Kurihara, Nakamichi, and Kojima9 obviously have relatively low
scatter and agree with the model to about ±5%, while the data
of Resa et al.10 have larger uncertainty, particularly at low
ethanol mole fractions. The experimental procedure used by
Resa et al. is similar to that of Reinders and de Minjer,8 and this
2232
Editorial
Figure 4. Percentage errors of model K-values of ethanol in
comparison to the ternary VLE data of Kurihara, Nakamichi, and
Kojima,9 and Resa et al.10 for the water + methanol + ethanol ternary
mixture at 101.3 kPa. It is noted that the y-axis range has been
restricted to ±40% in order to clearly reveal the data accuracy. There
are six points from Resa et al. for which the percentage error exceeds
40%.
is further evidence that indirect concentration measurements
have poorer accuracy at low concentrations. It is noted here
that Resa et al. did report activity coefficients, and comparison
of calculated to experimental activity coefficients will give a
chart equivalent to Figure 4 since vapor-phase nonideality is not
important at the low pressure of 101.3 kPa.
The editors of JCED do not stipulate that any particular
model should be used, but the analysis in this editorial clearly
demonstrates the value of using a phase-equilibrium model as
an aid to data analysis. The editors of JCED also do not wish to
prejudge particular experimental and estimation procedures as
these constantly evolve and improve over time due to the skill
and creativity of researchers. However, the editors of JCED
emphasize the following guidelines for the analysis of measured
and estimated VLE data:
1. Data should be analyzed through the use of an
appropriate model.
2. Data analysis should be performed such that the
sensitivity remains valid even when the measured
concentrations are low.
3. Data analysis should include distribution coefficients
such as K-values and relative volatilities, similar to those
in Figure 2, Figure 3, and Figure 4 of this Editorial.
4. Charts should be designed such that the key results can
be “seen at a glance by a human.”
Paul M. Mathias*
Fluor Corporation, 3 Polaris Way, Aliso Viejo, California
92698, United States
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: Paul.Mathias@Fluor.com.
ORCID
Paul M. Mathias: 0000-0001-5781-9525
Notes
Views expressed in this editorial are those of the author and not
necessarily the views of the ACS.
DOI: 10.1021/acs.jced.7b00582
J. Chem. Eng. Data 2017, 62, 2231−2233
Journal of Chemical & Engineering Data

■
Editorial

REFERENCES
(1) Ambrose, D. The Evaluation of Vapor Pressure Data, unpublished
manuscript, 1985.
(2) Diky, V. An Efficient Way of Visualization of Mutual Solubility
Data in the Whole Range of Compositions. J. Chem. Eng. Data 2017,
DOI: 10.1021/acs.jced.7b00174.
(3) Mathias, P. M. Effect of VLE Uncertainties on the Design of
Separation Sequences by Distillation - Study of the Benzene-
Chloroform-Acetone System. Fluid Phase Equilib. 2016, 408, 265−272.
(4) Kojima, K.; Tochigi, K.; Kurihara, K.; Nakamichi, M. Isobaric
Vapor- Liquid Equilibria for Acetone + Chloroform + Benzene and the
Three Constituent Binary Systems. J. Chem. Eng. Data 1991, 36, 343−
345.
(5) Kudryavtseva, L. S.; Susarev, M. P. Liquid-Vapor Equilibrium in
Chloroform-Hexane and Acetone-Chloroform Systems. Zh. Prikl.
Khim. (Leningrad) 1963, 36, 1231−1237.
(6) Kogan, L. V.; Deizenrot, I. V. Apparatus for Determination of
Liquid-Vapor Equilibrium in Systems with Nonideal Vapor Phases. Zh.
Prikl. Khim. (Leningrad) 1975, 48, 2757−2759.
(7) Segura, H.; Mejia, A.; Reich, R.; Wisniak, J.; Loras, S. Isobaric
Vapor-Liquid Equilibria and Densities for the Binary Systems Oxolane
+ Ethyl 1,1-Dimethylethyl Ether, Oxolane + 2-Propanol and Propan-2-
One + Trichloromethane. Phys. Chem. Liq. 2003, 41, 283−301.
(8) Reinders, W.; de Minjer, C. H. Vapour-Liquid Equilibria in
Ternary Systems. II. The System Acetone-Chloroform-Benzene. Rec.
Trav. Chim. 1940, 49, 369−391.
(9) Kurihara, K.; Nakamichi, M.; Kojima, K. Isobaric Vapor-Liquid
Equilibria for Methanol + Ethanol + Water and the Three Constituent
Binary Systems. J. Chem. Eng. Data 1993, 38, 446−449.
(10) Resa, J. M.; Goenaga, J. M.; Gonzalez-Olmos, R.; Iglesias, M.
Measurement and Modeling of Phase Equilibria for Ethanol + Water +
Methanol at Isobaric Condition. J. Chem. Eng. Data 2006, 51, 2114−
2120.
(11) For further details on Aspen Plus property options, refer to
documentation available from Aspen Technology, Inc.

2233 DOI: 10.1021/acs.jced.7b00582

J. Chem. Eng. Data 2017, 62, 2231−2233