ArcPlot representation of vapor pressures

Traditionally, experimental vapor pressures p as a function of temperature T are represented such that ln p on the ordinate axis is plotted against 1/T on the abscissa axis. This is, however, rather an insensitive representation that is often not capable to reveal systematic errors in the data. As shown by Thodos [1], vapor pressure temperature dependence ln p0 vs. 1/T is in fact concave over temperature range from the triple point temperature to temperatures well above the normal boiling temperature (Fig. 2 in ref [1]),  A simple transformation of vapor pressure data devised in 1981 by Karel Aim (Institute of Chemical Process Fundamentals, Academy of Sciences of the Czech Republic, https://www.icpf.cas.cz) [2,3] stresses out the differences among different vapor pressure data sets and helps to exclude the  data that systematically deviate from others. The scale of the vapor curve is set so that the terminal point [Tmax, pmax] is assigned the coordinate [1,1] and the beginning point [Tmin, pmin] the coordinate [0,0]. This is done by the following definition of the x and y coordinates

It should be stressed that utilization of the above coordinates enables to assess the quality of vapor pressure  datasets (and even individual data points) without the necessity to select any particular model, thus avoiding least-squares analysis.

Comparison of the sensitivity of ln p = f(1/T) and proposed representation is shown in a figure bellow. It is obvious that obvious outliers can be easily idetified and rejected.  On the other hand, arc plot merely highlights differences among various vapor pressure data sets, thus other means (e.g. SimCor) are necessary to decide which data should be retained or rejected in the final correlation.

The representation presented here is similar to that proposed in 1998 in Utrecht [4,5] but its application is simpler. 

Similar approach can be used also for sublimation pressures, however when the temperature/pressure range is narrow, the arc is rather flat and sometimes not easy to interpret (see Figure1 and Figure S4 in our recent paper [6] and related discussion).

Recently, "arc plot representation" was applied in our laboratory for evaluation of quality of chromatographic retention factors k [7] in an attempt to determine vapor pressures of supercooled liquids by indirect chromatographic method.  While X-coordinate remains the same, the Y-coordinate is defined as

List of selected vapor pressure data sources is available here

REFERENCES:

[1] Thodos, G., Vapor Pressures of Normal Saturated Hydrocarbons. Ind. Eng. Chem. 1950, 42, 1514-1526.

[2] K. Aim, Program for the Correlation of Saturated Vapor Pressure Data, Rep. No. 1/81 ed., Inst. Chem. Process Fundam., Prague, 1981.

[3] M. Čenský, V. Roháč, K. Růžička, M. Fulem, K. Aim. Fluid Phase Equilib. 298 (2010) 192-198.

[4] P.R. van der Linde, J.G. Blok, H.A.J. Oonk, J. Chem. Thermodyn. 30 (1998) 909-917.

[5] Oonk, H. A. J.; van der Linde, P. R.; Huinink, J.; Blok, J. G., Representation and assessment of vapour pressure data; a novel approach applied to crystalline 1-bromo-4-chlorobenzene, 1-chloro-4-iodobenzene, and 1-bromo-4-iodobenzene. J. Chem. Thermodyn. 1998, 30, 897-907.

[6] Mahnel, T.; Štejfa, V.; Maryška, M.; Fulem, M.; Růžička, K., Reconciled thermophysical data for anthracene. The Journal of Chemical Thermodynamics 2019, 129, 61-72.

[7] Bohumír Koutek, Tomáš Mahnel, Karel Řehák, Václav Pokorný, Michal Fulem, Květoslav Růžička, Regression against Temperature of Gas-Liquid Chromatography Retention Factors. Van't Hoff Analysis.
Journal of Chemical & Engineering Data, 65, 2020,  https://doi.org/0.1021/acs.jced.0c00119